 |
|
|
The Agency for Healthcare Research and Quality (AHRQ), through its Evidence-based Practice Centers (EPCs), sponsors the development of evidence reports and technology assessments to assist public- and private-sector organizations in their efforts to improve the quality of health care in the United States. The reports and assessments provide organizations with comprehensive, science-based information on common, costly medical conditions and new health care technologies. The EPCs systematically review the relevant scientific literature on topics assigned to them by AHRQ and conduct additional analyses when appropriate prior to developing their reports and assessments.
To bring the broadest range of experts into the development of evidence reports and health technology assessments, AHRQ encourages the EPCs to form partnerships and enter into collaborations with other medical and research organizations. The EPCs work with these partner organizations to ensure that the evidence reports and technology assessments they produce will become building blocks for health care quality improvement projects throughout the Nation. The reports undergo peer review prior to their release.
AHRQ expects that the EPC evidence reports and technology assessments will inform individual health plans, providers, and purchasers as well as the health care system as a whole by providing important information to help improve health care quality.
We welcome written comments on this evidence report. They may be sent to: Director, Center for Practice and Technology Assessment, Agency for Healthcare Research and Quality, 6010 Executive Blvd., Suite 300, Rockville, MD 20852.
Carolyn Clancy, M.D.
Acting Director
Agency for Healthcare Research and Quality
Robert Graham, M.D.
Director, Center for Practice and Technology Assessment
Agency for Healthcare Research and Quality
The authors of this report are responsible for its content. Statements in the report should not be construed as endorsement by the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services of a particular drug, device, test, treatment, or other clinical service.
Objectives. We conducted a systematic review of published evidence on four common musculoskeletal disorders affecting workers; carpal tunnel syndrome (CTS), cubital tunnel syndrome, epicondylitis, and de Quervain's disease. This report is a “Best Evidence” synthesis in which we address the best available evidence, not the best possible evidence. We addressed 13 key questions regarding their diagnosis, treatment, and costs.
Search Strategy. To identify information for this report, we searched 31 databases, relevant web sites, four U.S. government datasets, hand-searched the reference lists of all studies retrieved for this evidence report, searched Current Contents-Clinical Medicine weekly, and reviewed over 1,600 documents maintained in ECRI's collections.
Selection Criteria. To be selected for evaluation, a published study had to enroll patients diagnosed with one of the four relevant disorders. All controlled trials were retrieved, regardless of year of publication or whether they were described as randomized or prospective. Other studies were evaluated only if they were published in 1980, or later, and included 10 or more patients. Only English-language articles were retrieved. After retrieval, documents were examined to ensure that they did not contain flaws (e.g. confounding, incomparable study groups) precluding interpretation of results.
Data Collection and Analysis. Data about trial design, patient signs, symptoms, comorbidities, characteristics, and treatments, treatment outcomes and diagnostic measurements were abstracted from articles meeting inclusion criteria using electronic forms. Data were meta-analyzed when possible. Other analyses included corrections for patient attrition, statistical power analyses, multiple regression analyses, effect size computation, determinations of statistically significant differences between patient characteristics and verification of diagnostic test characteristics.
Main Results
The literature describing these disorders is often of poor quality, with few studies addressing any given issue. The evidence currently available suggests the following tendencies:
Two diagnostic tests for CTS, distal motor latency and palmar sensory latency, appear to have high specificity and low-to-moderate sensitivity.
Patients who have undergone surgery for CTS are predominantly middle aged and female. It is not possible to determine the characteristics of those undergoing surgery for the other three conditions.
Studies comparing open and endoscopic carpal tunnel release show a small but statistically significant advantage for endoscopic release, despite a higher rate of complications and reoperation compared to open release.
CTS patients benefit, but may not recover fully or permanently after steroid injection into the carpal tunnel.
Published data do not support the use of neurolysis, ligament reconstruction, or ultrasound for most CTS patients.
Laser therapy does not appear to be an effective treatment for epicondylitis.
Patients with epicondylitis who were treated with acupuncture had better global outcomes and greater pain relief than patients given sham acupuncture.
Conclusions. Published literature describing the diagnosis, treatment and impact of worker-related upper-extremity disorders is diffuse and generally of low quality, making it difficult to come to firm evidence-based conclusions. There are trends in available data, but it is often difficult to quantify them.
This report is a systematic evaluation of the evidence pertaining to a broad range of issues related to the diagnosis and treatment of worker-related upper extremity disorders (WRUEDs). For the purposes of this report, “worker-related” is defined as a disorder that affects workers, not as a disorder necessarily caused by work. Four disorders are the focus of this report; carpal tunnel syndrome, cubital tunnel syndrome, epicondylitis, and de Quervain's disease.
The first two disorders are the result of nerve entrapment. Carpal tunnel syndrome is the result of increased pressure on the median nerve in the carpal tunnel of the wrist, resulting in sensory and motor disturbances in the parts of the hand innervated by this nerve. Cubital tunnel syndrome results from increased pressure on the ulnar nerve in the cubital tunnel of the elbow, resulting in sensory and motor disturbances in the parts of the forearm and hand innervated by this nerve. The second two disorders are the result of stress to the tendons of the elbow and wrist, respectively. All four disorders can lead to pain, loss of function, and long-term disability.
The overall prevalence of carpal tunnel syndrome in the United States may be as high as 1.9 million people, and each year there are 300,000–500,000 operations for the condition. Epicondylitis has been reported to affect 4.23 individuals per 1,000 adults per year in the U.S. The prevalence of cubital tunnel syndrome and de Quervain's disease has not been established.
In this evidence report, the Evidence-based Practice Center (EPC) assessed the published literature describing the effects of these disorders, before and after treatment, on patients, particularly workers. They did this by examining the literature pertaining to 13 key questions.
This report addresses 13 questions regarding worker-related disorders of the upper extremity. Eleven of these are condition-specific. Therefore, the EPC individually address them for each of the four above-mentioned disorders. Two questions are not condition-specific. Therefore, the EPC addressed them only once. The 11 condition-specific Key Questions addressed in this evidence report are:
Question 1: What are the most effective methods and approaches for the early identification and diagnosis of worker-related musculoskeletal disorders of the upper extremity?
Question 2: What are the specific indications for surgery for worker-related musculoskeletal disorders of the upper extremity?
Question 3: What are the relative benefits and harms of various surgical and nonsurgical interventions for persons with worker-related musculoskeletal disorders of the upper extremity?
Question 4: Is there a relationship between specific clinical findings and specific treatment outcomes among patients with worker-related musculoskeletal disorders of the upper extremity?
Question 5: Is there a relationship between duration of symptoms and specific treatment outcomes among patients with worker-related musculoskeletal disorders of the upper extremity?
Question 6: Is there a relationship between factors such as patients' age, gender, socioeconomic status and/or racial or ethnic grouping and specific treatment outcomes among patients with worker-related musculoskeletal disorders of the upper extremity?7
Question 7: What are the surgical and nonsurgical costs or charges for treatment of worker-related musculoskeletal disorders of the upper extremity?
Question 8: For persons who have had surgery for worker-related musculoskeletal disorders of the upper extremity, what are the most effective methods for preventing the recurrence of symptoms, and how does this vary depending on subject characteristics or other underlying health problems?
Question 9: What instruments, if any, can accurately assess functional limitations in an individual with a worker-related disorder of the upper extremity?
Question 10: What are the functional limitations for an individual with a worker-related musculoskeletal disorder of the upper extremity before treatment?
Question 11: What are the functional limitations of an individual with a worker-related musculoskeletal disorder of the upper extremity after treatment?
The two Key Questions that are not condition-specific are:
Question 12: What are the cumulative effects on functional abilities among individuals with more than one worker-related musculoskeletal disorder of the upper extremity in the same limb?
Question 13: What level of function can patients achieve in what period of time when they are required to change hand dominance as a result of injury to their dominant hand?
A panel of nine Technical Experts was employed to assist in defining the scope of this evidence report, developing its questions, and developing the criteria for retrieving and including articles.
To identify information for this evidence report, the EPC searched 31 electronic databases, the World Wide Web, and four U.S. government databases. In addition to these searches, researchers also reviewed the bibliographies and reference list of all studies included in this evidence report, searched Current Contents®/Clinical Medicine on a weekly basis, and routinely reviewed over 1,600 journals and supplements maintained in ECRI's collections.
To be included in this evidence report, an article had to meet a set of a priori retrieval criteria and a set of a priori question-specific inclusion criteria. EPC designed broad retrieval criteria to ensure comprehensive retrieval. They retrieved an article whenever there was uncertainty about whether it met the retrieval criteria. They also retrieved articles when an abstract was not present in the search results, but when the title of the article suggested that it was relevant. The criteria for article retrieval are briefly summarized below:
The patients had to have been diagnosed with a worker-related disorder of the upper extremity.
All controlled trials, regardless of whether they were described as randomized or prospective, were retrieved, regardless of year of publication.
Case series and other reports were evaluated only if published in 1980 or later and included 10 or more patients.
Only English-language articles were retrieved.
Once an article was retrieved, it was examined to determine whether it met the question-specific criteria. The major criteria are briefly summarized below; additional question-specific inclusion criteria, which are not listed here, were also applied:
The study could not have a serious design flaw that precluded interpretation of the results.
The study must have addressed one of the key questions and have included patients with one of the WRUEDs of interest.
For studies addressing Key Question 3, the study must have been a controlled trial.
The study must have reported on at least one of the seven key outcomes addressed in this assessment. The outcomes are: pain, function, quality of life, ability to return to work, ability to return to activities of daily living, harms, and global outcome.
A global outcome is any score that attempts to encompass the overall success or failure of the treatment. It may be a numerical rating of overall symptom relief or patient satisfaction, a categorical rating such as excellent, good, fair or poor, or a dichotomous rating such as the answer to the question “Would you undergo this procedure again?”
Data from all articles that met our inclusion criteria were abstracted using electronic data abstraction forms. Separate data abstraction forms were designed for entering data about basic trial design information; patient signs, symptoms, comorbidities, characteristics, and treatments; reporting of treatment outcomes; surgical complications; and nerve conduction measurements.
The EPC employed a variety of statistical methods in this evidence report. Meta-analyses of studies of treatments were conducted using Hedges' d as a measure of each study's effect size, and then computing the precision-weighted summary d from the combined results of all studies. Hedges' d is the difference between the means of any study's two groups expressed in standard deviation units. Researchers employed two tests for heterogeneity, the Q statistic and each study's standardized residual. The EPC researchers regarded the data as heterogeneous if the results of either test were statistically significant.
Diagnostic test meta-analyses were performed according to the method of Littenberg and Moses. The researchers took the the mean threshold as the best estimate of a single threshold, and the values of sensitivity and specificity at the mean threshold as the single best global estimate of test effectiveness. Before using the results of a meta-analysis of diagnostics, they verified that there was no statistically significant heterogeneity among the results of the included articles using the Q statistic. If heterogeneity was detected, they removed any subgroups that caused the heterogeneity from the analysis. If there were no subgroups in the analysis, or those subgroups did not cause the heterogeneity, They looked for data points that were outliers, and reported the meta-analytic results with and without exclusion of these outliers.
The EPC performed numerous other statistical computations in addition to those involved in performing meta-analyses. Briefly, these were:
Corrections for patient attrition.
Statistical power analyses.
Multiple regression for certain questions when such results were of interest.
Computations of effect sizes for all studies, when possible, even when no meta-analysis was performed.
Determinations of whether there were statistically significant differences between the characteristics of patients in any given study.
Computation of pretreatment effect sizes.
Verification of diagnostic test characteristics.
Question 1: What are the most effective methods and approaches for the early identification and diagnosis of carpal tunnel syndrome?
The evidence base on most individual diagnostic tests for carpal tunnel syndrome is small, even though the total number of articles on CTS diagnosis is large. This is because many different tests have been described. Nerve conduction tests are most frequently reported in the literature, but there is great diversity in their methods.
The results of our analyses may overestimate the specificity of nerve conduction measurements in typical practice. This is because the trials we examined used healthy, asymptomatic persons as controls. In clinical practice, the test would be used on workers believed to be at risk for CTS or persons suspected of having CTS. Under these conditions, the false positive rate would be higher, and the specificity correspondingly lower.
The most frequently reported nerve conduction tests were distal motor latency and palmar sensory latency. For both tests, clinicians chose thresholds that yielded high specificity (a low incidence of false-positive results). The EPC's meta-analyses of distal motor latency studies found the sensitivity of the test to be 57% to 66% and the specificity to be 98%. Meta-analysis of palmar sensory latency studies found a sensitivity of 76% and a specificity of 98%.
Clinical signs and symptoms are also used in the diagnosis of CTS. They attempted to use our meta-analysis techniques to obtain summary values for the sensitivity and specificity of two such signs: Tinel's sign and Phalen's maneuver. In both cases, there was heterogeneity in the published results that could not be explained by differences in patient selection or by single outlier studies. Therefore, they did not calculate summary measurements for sensitivity or specificity. The sensitivity of Phalen's maneuver was lower than its specificity, and two trials reported sensitivity of 80% to 90%. All of the studies of Tinel's sign found that its sensitivity was lower than its specificity, and none found a sensitivity of 75 percent or greater. There was too much heterogeneity in the results for us to conclude that one test was superior to the other, or to compare these tests to nerve conduction testing.
Regarding sensory tests, composite nerve conduction tests, and imaging tests, there was insufficient evidence for the EPC to perform meta-analyses of clinical trial results.
Their well-designed study suggests that nerve conduction measurement may be able to identify some workers at risk of developing CTS in the future. By itself, this evidence is not sufficient for the EPC to conclude that nerve conduction screening for CTS is effective.
Question 2: What are the specific indications for surgery for carpal tunnel syndrome?
Patients who have undergone surgery for carpal tunnel syndrome are predominantly middle aged and female.
Because of underreporting, no firm evidence-based conclusions can be drawn regarding the signs, symptoms, neuroelectrical characteristics, and comorbidities of these patients.
Question 3: What are the relative benefits and harms of various surgical and nonsurgical interventions for persons with carpal tunnel syndrome?
Meta-analysis of studies comparing open and endoscopic carpal tunnel release show a small but statistically significant advantage to endoscopic release in global treatment outcome. In addition, the data show a trend toward faster return to work and to activities of daily living among patients receiving endoscopic release. However, these findings must be viewed only as trends in currently available data. This is because they are based on a meta-analysis that contained a number of non-randomized, non-blinded studies. Data from these studies also suggest that endoscopic release has a higher complication rate and a higher rate of reoperation compared to open release. The higher reoperation rates likely arise because of incomplete transection of the transverse carpal ligament. Exact complication rates cannot be determined from presently available data. Presently available data also do not allow one to reach firm evidence-based conclusions about the relative effects of open and endoscopic surgery on the ability of patients to perform daily functions.
Meta-analysis of global outcomes demonstrates a potential benefit from not performing neurolysis. Available return to work data also shows a trend toward an advantage to not performing neurolysis. There is insufficient data to determine the effect of neurolysis on pain and function. The available evidence suggests there is little or no benefit from performing neurolysis along with surgical release of the carpal tunnel. The possibility remains that neurolysis may be helpful in special cases, such as in the presence of marked scarring or neural adhesion, but no available evidence specifically documents the benefits and harms of neurolysis among such patients.
Results of four studies suggest that injection of steroid into the carpal tunnel yields superior global outcomes compared to no treatment, placebo or oral steroids. However, relief from steroid treatments is not complete. Carpal tunnel injection was significantly better than intramuscular injection at a 1 month followup time. Because no further time points were reported, researchers are unable to determine whether this difference persists beyond this time. There are no data available that indicate whether any type of steroid may be superior to any other, or whether any particular dose is optimum. Although the effects of steroid injection may wear off over time, there is no information indicating the expected duration of relief for the average patient, or whether any patients can expect to experience permanent relief.
Two double-blinded randomized controlled trials suggest that oral steroids may lead to a reduction in symptoms of CTS. However, the effects of oral steroids are short-lived and may not be sufficient for patient satisfaction. The effects of higher steroid doses or longer treatment regimens have not been examined in published controlled trials.
A single published randomized controlled trial indicates that oral tenoxicam (a NSAID) and trichlormethiazide (a diuretic) do not reduce the symptoms of CTS under the dosing regimens described. Further trials are needed to confirm this observation, and to test the effects of additional drugs and dosing regimens.
Results of a single study suggest that manual therapy may have some use in the treatment of carpal tunnel syndrome. This study suggests that carpal bone mobilization provides pain relief, improves function, and delays or eliminates the need for surgery among patients with carpal tunnel syndrome. However, this small study was unblinded. Results from neurodynamic mobilization show a similar trend, but because of a lack of statistical power one cannot conclude that this trend is real. For the same reason, differences in effectiveness between these two treatment groups cannot be determined. A large, blinded, randomized controlled trial is necessary to confirm these results.
A larger, more statistically powerful study found no difference between the effects of a physical therapy program and home exercise instructions on pain or function. However, patients receiving physical therapy returned to work faster than those instructed to exercise at home.
Although these studies indicate a trend toward some forms of physical therapy having an effect on carpal tunnel syndrome, their small size and design difficulties make it difficult to arrive at a firm evidence-based conclusion.
Only one study meeting inclusion criteria addresses the use of ultrasound for carpal tunnel syndrome. Because of this, and because of its associated design and analysis difficulties, one cannot reach a firm evidence-based conclusion.
Splint use was addressed only by a single trial that had design difficulties. Because of this, one cannot reach a firm evidence-based conclusion about splint use. There may be conditions under which splints offer an advantage and conditions under which they do not, but this is not addressed by available evidence.
The results of one study suggest that suboptimal outcomes are obtained when patients receive ligament reconstruction. However, this trial was neither randomized nor blinded, so one cannot draw firm evidence-based conclusions from it.
Although the low statistical power of the one relevant study prevents any solid conclusion from being drawn, this study does not support the therapeutic effectiveness of Vitamin B6. This is because it showed a trend toward a greater percentage of improved patients in the placebo group.
Question 4: Is there a relationship between specific clinical findings and specific treatment outcomes among patients with carpal tunnel syndrome?
The only clinical finding variable shown by more than one study to significantly predict treatment outcomes was electrodiagnostic testing. Patients with mildly impaired or normal results of electrodiagnostic tests had longer sick leaves and were less likely to be satisfied with the results of treatment. This finding was statistically significant in three of the four studies examining this relationship.
This apparent lack of consistency of results could indicate that, although the relationship between electrodiagnostic tests and treatment outcomes is statistically significant, it may not be substantial. The possibility that this relationship is small is supported by the results of stratified studies that examined the relationship between electrodiagnostic test results and global outcomes. Six of seven studies did not find a statistically significant relationship.
Question 5: Is there a relationship between duration of symptoms and specific treatment outcomes among patients with carpal tunnel syndrome?
The majority of available evidence is less than optimal because it consists primarily of retrospective studies. The highest quality study (prospective with multiple regression analysis) suggested that there was no statistically significant correlation between duration of symptoms and global outcome after surgery. One prospective and two retrospective stratified studies found similar results. Two retrospective studies (one performing multiple regressions, one stratified) found a statistically significant relationship between shorter duration of symptoms and symptom resolution or patient satisfaction after surgery. The retrospective nature of these trials could have created bias that influenced these findings. An additional high quality prospective study is needed before firm conclusions can be reached.
Question 6: Is there a relationship between factors such as patients' age, gender, socioeconomic status and/or racial or ethnic grouping and specific treatment outcomes among patients with carpal tunnel syndrome?
The available evidence suggests that patients who are not receiving workers' compensation tend to return to work faster than those receiving such compensation. This is suggested by one of two “multiple regression” studies of this relationship and by a combination of 10 prospective and retrospective stratified studies. Evidence of a relationship does not constitute evidence of causality.
Some evidence also suggests that patients who are not receiving workers' compensation have better global outcomes, but this evidence is derived exclusively from retrospective studies. Therefore, these latter findings require confirmation.
Available evidence suggests that there is no strong relationship between gender, employment status, or hand dominance and return to work or global outcomes.
There is insufficient evidence to arrive at a firm evidence-based conclusion on the relationship between type of work, presence of diabetes, or age and patient outcomes.
Question 7: What are the surgical and nonsurgical costs or charges for treatment of carpal tunnel syndrome?
According to the Medicare Provider Analysis and Review (MEDPAR) database, which covers hospital inpatient services, average total charges per patient for the DRG (diagnosis-related group) of carpal tunnel release are $8,185.24 (calculated by dividing total charges by number of discharges). This DRG includes open and endoscopic release.
The Median Costs for Hospital Outpatient Services Dataset contains median costs for services that are reimbursed under Medicare for the hospital outpatient prospective payment system. The reported median cost for endoscopic release of the transverse carpal ligament is $849.84 (cost of open release was not reported by this database). The reported median cost for application of a short arm static splint is $72.69.
Question 8: For persons who have had surgery for carpal tunnel syndrome, what are the most effective methods for preventing the recurrence of symptoms, and how does this vary depending on subject characteristics or other underlying health problems?
No controlled trials have been published that report on the efficacy or effectiveness of any technique for the prevention of recurrence of carpal tunnel syndrome. In the absence of controlled trials, no analysis may be performed and no evidence-based conclusions may be drawn.
Question 9: What instruments, if any, can accurately assess functional limitations in an individual with carpal tunnel syndrome?
Three prospective cohort trials have indicated that the SF-36 is not a useful instrument for assessing functional limitations in individuals with carpal tunnel syndrome. The SF-36 was reported to be unresponsive to treatment and unable to predict ability to work.
Four prospective cohort trials have indicated that the Levine CTS-I may be a useful instrument for assessing functional limitations in individuals with carpal tunnel syndrome. This instrument was reported to be responsive to treatment, and to have concurrent validity as measured by grip and pinch strength. However, the studies that addressed the Levine CTS-I did not examine its internal reliability, content validity, or its ability to predict how well patients could perform activities of daily living. In addition, the Levine CTS-I has been reported by one study to be unable to predict ability to work.
No other instrument has been evaluated by more than one study. It is difficult to reach an evidence-based conclusion as to the usefulness of the other instruments evaluated in this report due to the limited evidence base.
Question 10: What are the functional limitations for an individual with carpal tunnel syndrome before treatment?
There is some evidence to suggest that most untreated patients with carpal tunnel syndrome have mild to moderate functional difficulties before treatment. However, this evidence is derived from only two studies comprised of a total of 51 patients. This is too few patients and too few studies to allow one to reach a firm evidence-based conclusion.
Question 11: What are the functional limitations of an individual with carpal tunnel syndrome after treatment?
Although studies of non-surgical therapies suggested that most patients experience only mild difficulty with functional activities after treatment, it is unclear whether the results of these two studies are generalizable to the larger patient population.
Studies with surgical outcomes suggested that most patients report no-to-moderate difficulty with functional activities (mean 1.4–2.6 on the Levine CTS-I) after surgery.
Although there were no statistically significant differences between specific patient groups, there was a trend toward more difficulty with functional activities among workers' compensation patients in surgical studies. This trend was based on the results of two studies.
The available data are insufficient to determine a cutoff point on measuring scales above which patients are unable to work.
Question 1: What are the most effective methods and approaches for the early identification and diagnosis of cubital tunnel syndrome?
One test for cubital tunnel syndrome, ulnar motor nerve conduction velocity at the elbow, was commonly mentioned by reviewers. Three studies reported high specificity and low sensitivity for this test. Due to the small number of studies, however, one cannot draw quantitative conclusions about the effectiveness of the test. There are insufficient data to permit firm evidence-based conclusions about the effectiveness of this or any other tests for cubital tunnel syndrome.
Question 2: What are the specific indications for surgery for cubital tunnel syndrome?
Thirty-two studies of patients who received surgery for cubital tunnel syndrome were identified. The mean age of patients who received surgery for cubital tunnel syndrome was 46 years.
The patients were slightly more likely to be male (62% male).
On average, patients had symptoms 10 to 24 months before receiving surgical treatment.
Question 3: What are the relative benefits and harms of various surgical and nonsurgical interventions for persons with cubital tunnel syndrome?
One randomized controlled trial of 52 patients found that medial epicondylectomy was superior to anterior transposition in relieving pain and in improving global outcome scores. The results of this study are suggestive, but one cannot arrive at a strong conclusion from the results of only one trial. There is insufficient evidence to determine the relative effectiveness of other surgical treatments.
There are insufficient data available to determine the rates of surgical complications for any of the described surgical procedures.
Question 4: Is there a relationship between specific clinical findings and specific treatment outcomes among patients with cubital tunnel syndrome?
The only clinical finding variable shown by more than one study to significantly predict treatment outcomes was severity of symptoms. This correlation was statistically significant in four out of seven studies that examined it. The studies that did not find a statistically significant correlation may have been underpowered. Therefore, currently available evidence tentatively suggests that there is a correlation between having less severe symptoms and having a higher global outcome score after surgical treatment for cubital tunnel syndrome.
There are insufficient data to reach evidence-based conclusions about the relationships between other clinical findings and treatment outcomes.
Question 5: Is there a relationship between duration of symptoms and specific treatment outcomes among patients with cubital tunnel syndrome?
Currently available evidence does not suggest a clear-cut relationship between the duration of symptoms before treatment and the success of surgery.
There are insufficient data available to reach evidence-based conclusions about the relationship between symptom duration and other treatment outcomes.
Question 6: Is there a relationship between factors such as patients' age, gender, socioeconomic status and/or racial or ethnic grouping and specific treatment outcomes among patients with cubital tunnel syndrome?
The available data do not suggest a substantial correlation between the age, sex, or workers' compensation status of the patient and the success of surgery.
Two studies that used multiple regression to examine relationships between patient characteristics and treatment outcomes found that patients whose cubital tunnel syndrome is caused by an acute trauma have better outcomes after surgical treatment than patients with cubital tunnel syndrome from other causes. However, three studies that stratified by etiology found no statistically significant relationship between cause and patient outcomes. The studies that used multiple regression techniques are of better quality than the stratified studies. Thus, current data suggest that there may be a correlation between etiology and patient outcomes, but this cannot be regarded as definitive.
Question 7: What are the surgical and nonsurgical costs or charges for treatment of cubital tunnel syndrome?
According to Medicare Provider Analysis and Review (MEDPAR), average total charges per patient for the DRG (diagnosis-related group) of major shoulder/elbow procedures with comorbidities or complications are $9,008.94 (calculated by dividing total charges by number of discharges).
For the DRG shoulder, elbow or forearm procedures, except major joint procedures, without comorbidities or complications, average total charges per patient are $7729.16.
For the DRG peripheral and cranial nerve and other nerve procedures without complications or comorbidities, the average total per patient charges are $14,357.65 (with complications or comorbidities the charges are $24,288).
The Median Costs for Hospital Outpatient Services Dataset contains median costs for services that are reimbursed under Medicare for the hospital outpatient prospective payment system. The reported median cost for a decompression fasciotomy of the forearm and/or wrist is $603.85. The reported median cost for application of a long-arm splint is $80.48.
Question 8: For persons who have had surgery for cubital tunnel syndrome, what are the most effective methods for preventing the recurrence of symptoms, and how does this vary depending on subject characteristics or other underlying health problems?
None of the included studies addressed this question.
Question 9: What instruments, if any, can accurately assess functional limitations in an individual with cubital tunnel syndrome?
None of the included studies addressed this question.
Question 10: What are the functional limitations for an individual with cubital tunnel syndrome before treatment?
None of the included studies addressed this question.
Question 11: What are the functional limitations of an individual with cubital tunnel syndrome after treatment?
None of the included studies addressed this question.
Question 1: What are the most effective methods and approaches for the early identification and diagnosis of epicondylitis?
There are insufficient data to permit evidence-based conclusions about the effectiveness of any tests for epicondylitis. This is because the evidence base is small and heterogeneous.
Question 2: What are the specific indications for surgery for epicondylitis?
Nineteen studies of patients who received surgery for epicondylitis were identified. Due to a lack of reported data, few trends or characteristics of patients who received surgery could be identified. A typical patient who received surgery for epicondylitis was middle-aged and equally likely to be male or female.
Question 3: What are the relative benefits and harms of various surgical and nonsurgical interventions for persons with epicondylitis?
Seven double-blinded randomized controlled trials compared laser therapy to sham laser therapy as treatment for epicondylitis. A meta-analysis of the results of the four studies that reported “success of treatment” did not reveal a statistically significant difference in outcome between laser and sham-treated patients.
The four studies that reported the effect of laser treatment on pain also did not find a statistically significant difference in outcome between laser and sham treated patients. However, EPC researchers were unable to perform a meta-analysis of the outcome pain and, because all of these studies were small, their individual results cannot be taken as definitive proof that laser therapy has no effect on the pain of epicondylitis.
Only one study examined work status of patients after laser treatment. This study was also small, and it failed to find a statistically significant effect of laser treatment on work status. The results of all seven small randomized double-blinded controlled trials are consistent with the results of our meta-analysis, and suggest that if there is an effect of laser therapy on epicondylitis, it is not large.
Two randomized controlled trials of a total of 62 patients compared oral naproxen to oral diflunisal. One study reported no statistically significant difference in outcomes when comparing patients treated with the two different drugs, and did not find a consistent trend in favor of one drug. The other study reported that diflunisal treatment consistently resulted in better outcomes. For two outcomes, pain and function, the difference reached statistical significance. Further studies are necessary to resolve discrepancies between these studies.
Two randomized controlled trials of 82 patients in total compared ultrasound treatment to phonophoresis of hydrocortisone as a therapy for epicondylitis. Neither study found a statistically significant difference between treatment groups for any of the outcomes. When interpreting these results, it is important to keep in mind that both studies may have been too small to be able to detect clinically relevant differences between treatment groups.
Three randomized controlled trials of 220 patients in total compared ultrasound treatment to sham ultrasound treatment or no treatment as a therapy for epicondylitis. All three of the studies reported a trend towards better outcomes in the groups treated with ultrasound. However, this difference reached statistical significance in only one of the studies. Although low statistical power may explain the negative results of the two “nonsignificant” studies, further research is required to demonstrate this.
Simply wearing an elbow brace is reported by two crossover studies to have no effect on pain. Because these two studies were of less than optimal design, further studies are necessary before a conclusion may be reached.
Two randomized controlled trials of a total of 134 patients evaluated the effect of acupuncture on epicondylitis. Both studies reported patients treated with acupuncture had better global outcomes and greater pain relief than patients treated with sham acupuncture at relatively short (2 weeks) followup times. Although only two studies evaluated this treatment, both were well-designed. It is possible to tentatively conclude that acupuncture is an effective palliative treatment for epicondylitis.
Two randomized controlled trials of a total of 203 patients compared oral NSAIDs to injections of corticosteroids. One study did not find a statistically significant difference between the groups. The other study reported that patients treated with injections of corticosteroids had better outcomes than the patients treated with oral NSAIDs. Design differences may explain the discrepancy between these studies' results, and further study is required to resolve this issue.
One double-blinded randomized controlled trial reported that patients treated with placebo had a trend towards better outcomes than patients treated with topical DMSO; however, this trend did not reach statistical significance. This study also reported that topical DMSO application caused clinically significant skin irritation. However, this trial was based on only 51 patients, so further studies are necessary before a definitive evidence-based conclusion can be reached.
One randomized controlled trial of 128 patients compared oral diclofenac to placebo. The group treated with diclofenac had statistically significantly less pain than the placebo group, but the NSAID treatment had no statistically significant effect on hand/arm function, number of days of missed work, or global outcome. Oral NSAIDs were reported to occasionally cause gastrointestinal side effects. In the absence of a very large effect, it is difficult to reach a firm evidence-based conclusion from the results of a single trial of moderate size.
One double-blinded randomized controlled trial and one double blinded randomized crossover trial, of a total of 47 patients, compared topical diclofenac to placebo. One of the studies reported no statistically significant differences between the two groups for any of the outcomes. The other study reported that the group treated with the NSAID may have had some statistically significant benefit from the treatment. Researchers were unable to determine whether the differences in results between studies were due to differences in statistical power. Further studies are necessary to resolve discrepancies between these studies.
One randomized controlled trial of 40 patients compared topical diclofenac to topical salicylate, and reported that diclofenac was more effective for treating epicondylitis. Topical NSAIDs were reported to occasionally cause mild skin rashes. Further studies are necessary before a definitive evidence-based conclusion can be reached.
One randomized double-blinded study reported that injections of glucosamines are effective in treating the symptoms of epicondylitis in the short term (less than 6 months) as measured by global outcome and patient-reported pain. However, injections of glucosamines were found to have a high rate of side effects—40% of patients experienced pain at the site of injection, and 6% developed hematomas at the site of injection. Further studies are necessary before a definitive evidence-based conclusion about the clinical utility of this treatment can be reached.
One randomized double-blinded study reported that injections of methylprednisolone plus lidocaine were statistically significantly more effective at treating pain than injections of lidocaine. Further studies are necessary before a definitive evidence-based conclusion can be reached.
One randomized double-blinded study reported that injections of lignocaine plus triamcinolone were statistically significantly more effective at treating pain than injections of lignocaine or injections of lignocaine plus hydrocortisone. Further studies are necessary before a definitive evidence-based conclusion can be reached.
One randomized double-blinded study reported that injections of triamcinolone plus bupivacaine were more successful at treating epicondylitis than injections of triamcinolone plus lidocaine. Further studies are necessary before a definitive evidence-based conclusion can be reached.
One study reported a trend towards more successful treatment of epicondylitis after injections of methylprednisolone than after injections of hydrocortisone. However, this study was of less than optimal design, which makes it problematic to come to a definitive evidence-based conclusion on the basis of its results.
One study reported no difference in rates of successful treatment or number of work-days missed after treatment with injections of methylprednisolone as compared to injections of betamethasone plus lidocaine. This study had sufficient statistical power to have detected relatively small differences between treatment groups. However, design flaws in this study make it problematic to come to a definitive evidence-based conclusion on the basis of its results.
One study reported that wearing a brace regularly over the course of several months is not as effective in treating epicondylitis as is physiotherapy, but a different study reported that wearing a brace regularly in addition to physiotherapy may be more effective than physiotherapy alone. Further studies of these therapies are necessary before one can reach definitive evidence-based conclusions.
One retrospective case-controlled study compared fasciectomy, wide fasciectomy plus anconeus transfer, and re-operation of failed fasciectomy to include an anconeus transfer. However, because this was a single study of suboptimal design, one cannot reach a firm evidence-based conclusion about the relative efficacy of these procedures.
One non-parallel historically controlled trial reported that simple denervation led to statistically significantly better global outcome and greater pain relief than denervation plus decompression. However, because this was a single study of suboptimal design, one cannot reach a firm evidence-based conclusion about the relative efficacy of these procedures.
A single double-blinded randomized controlled trial of 30 patients reported that there were no statistically significant differences in the signs and symptoms of epicondylitis between patients treated with pulsed electromagnetic field therapy and patients receiving sham treatment. When interpreting the results of this trial, it must be kept in mind that the small size of the trial may have prevented the results from reaching statistical significance.
One randomized controlled trial reported that patients treated with extracorporeal shock wave therapy had statistically significantly greater improvements in pain and arm function than patients given sham treatment. However, it is difficult to reach firm evidence-based conclusions from the results of this trial because the lack of blinding and lack of intent-to-treat analysis of this trial may have affected its results.
One randomized controlled trial reported that patients treated with injections of corticosteroids had better outcomes than patients treated with manipulations and deep friction massage. Incomplete data and methods reporting from this trial make it problematic to reach any definitive evidence-based conclusions from its results.
One randomized controlled trial of 76 patients reported that patients treated with injections of corticosteroids had better outcomes than patients treated with braces or immobilization. Partly because of the small size of this trial, further studies are necessary before a definitive evidence-based conclusion can be reached.
One randomized controlled trial of 63 patients reported that patients treated with acupuncture had better outcomes than patients treated with corticosteroid injections. However, the results of this study may have been affected by patient selection bias because it enrolled only patients previously found to be unresponsive to injections of corticosteroids.
Two randomized controlled trials, one comparing transcutaneous electrical nerve stimulation, ultrasound, phonophoresis, and injections of steroids, the other comparing physical therapy to ultrasound, reported no statistically significant differences between treatment groups. However, both trials may have been too small to be able to have detected clinically meaningful differences between treatment groups.
Five randomized controlled trials evaluated various combinations of therapies for the treatment of epicondylitis. One trial of 18 patients found that patients treated with manipulation plus a home exercise program had fewer difficulties in performing activities of daily living than patients treated with a combination of ultrasound, physiotherapy, and home exercise. The other four trials did not find statistically significant differences between treatment groups. However, these studies were small, which may have prevented them from detecting clinically important differences between the treatment groups.
Question 4: Is there a relationship between specific clinical findings and specific treatment outcomes among patients with epicondylitis?
One study reported that the site of pain could be used to predict response to treatment, one reported that the severity of pain could be used to predict response to treatment, and one reported that the timing of onset of symptoms (acute vs. gradual) did not correlate with the response to treatment. Because only one study addressed each outcome, it is difficult to reach firm evidence-based conclusions from the available data.
Question 5: Is there a relationship between duration of symptoms and specific treatment outcomes among patients with epicondylitis?
Seven studies examined whether duration of symptoms correlated with treatment outcomes. Only one of the four studies that employed multiple regression found a statistically significant relationship between symptom duration and outcomes, and this study was retrospective. One of three studies that stratified patients according to their duration of symptoms found a statistically significant correlation with treatment outcomes. As this study was also retrospective, evidence suggesting a relationship is contradictory and weak. Two prospective studies that employed multiple regression did not find such a relationship. Both were of patients who had received ultrasound. However, currently available evidence about use of ultrasound in patients with epicondylitis or de Quervain's disease does not allow firm evidence-based conclusions. A lack of treatment effectiveness could obscure potential relationships between symptom duration and treatment-related outcomes. Therefore, one cannot draw firm evidence-based conclusions from currently available data.
Question 6: Is there a relationship between factors such as patients' age, gender, socioeconomic status and/or racial or ethnic grouping and specific treatment outcomes among patients with epicondylitis?
Three studies that used multiple regression found no statistically significant correlation between gender or age and response to treatment, suggesting that there is no strong relationship between these variables and patient outcomes.
One study found no statistically significant correlation between certain hobbies and response to treatment. However, it is difficult to reach evidence-based conclusions from the results of a single study.
The only study that examined co-morbidities reported that patients with co-existent ulnar neuropathy had significantly poorer outcomes than patients without ulnar neuropathy. However, it is difficult to reach evidence-based conclusions from the results of a single study.
Question 7: What are the surgical and nonsurgical costs or charges for treatment of epicondylitis?
According to Medicare Provider Analysis and Review (MEDPAR), average total charges per patient for the DRG (diagnosis-related group) of major shoulder/elbow procedures with comorbidities or complications are $9,008.94 (calculated by dividing total charges by number of discharges).
For the DRG shoulder, elbow or forearm procedures, excepting major joint procedures, without comorbidities or complications, average total charges per patient are $7729.16.
The Median Costs for Hospital Outpatient Services Dataset contains median costs for services that are reimbursed under Medicare for the hospital outpatient prospective payment system. The reported median cost for strapping of the elbow or wrist is $62.61. (cost of open release was not reported by this database).
Question 8: For persons who have had surgery for epicondylitis, what are the most effective methods for preventing the recurrence of symptoms, and how does this vary depending on subject characteristics or other underlying health problems?
No controlled trials addressed this question. Therefore, it was not possible to perform a reliable analysis, and one cannot draw firm evidence-based conclusions from the available data.
Question 9: What instruments, if any, can accurately assess functional limitations in an individual with epicondylitis?
Three studies evaluated two different instruments (PRFEQ and F-VAS) as ways to measure functional limitations of patients with epicondylitis. Neither assessment instrument has been shown to be a useful instrument for evaluating functional limitations in persons with epicondylitis. However, it is difficult to reach firm evidence-based conclusions about the instruments evaluated in this report due to the limited evidence base.
Question 10: What are the functional limitations for an individual with epicondylitis before treatment?
This question is addressed by only two studies comprised of a total of 82 patients. Although these studies suggest that epicondylitis patients have an average level of functional difficulty between 30% – 40% (mild to moderate) on functional status scales, the low number of studies and patients makes it difficult to arrive at an evidence-based answer to this question.
Question 11: What are the functional limitations of an individual with epicondylitis after treatment?
There were no studies that met the inclusion criteria for this question. Therefore, it cannot be answered in an evidence-based fashion.
Question 1: What are the most effective methods and approaches for the early identification and diagnosis of de Quervain's disease?
None of the included studies addressed this question.
Question 2: What are the specific indications for surgery for de Quervain's disease?
Two of the three studies that addressed this question reported that surgery was performed only on patients who did not benefit from conservative (non-operative) treatment. However, with so few studies and so many unreported patient characteristics, one cannot assume that the present data are representative of the larger patient population with de Quervain's disease.
Question 3: What are the relative benefits and harms of various surgical and nonsurgical interventions for persons with de Quervain's disease?
Although one study found that corticosteroid plus lidocaine injection produced more treatment success than immobilization splints among de Quervain's patients, there were design problems with this study. Because of these problems, and because only one study addressed this question, it is difficult to reach firm evidence-based conclusions concerning the effectiveness of any treatment for de Quervain's disease.
Question 4: Is there a relationship between specific clinical findings and specific treatment outcomes among patients with de Quervain's disease?
This question was addressed by only one relatively small retrospective study. This study found no relation between presence of a septated first dorsal compartment and treatment outcome. However, it is difficult to reach evidence-based conclusions from the results of a single study of suboptimal design.
Question 5: Is there a relationship between duration of symptoms and specific treatment outcomes among patients with de Quervain's disease?
This question was addressed by only one relatively small retrospective study. This study found no relation between duration of symptoms and treatment outcome. However, it is difficult to reach evidence-based conclusions from the results of a single study of suboptimal design.
Question 6: Is there a relationship between factors such as patients' age, gender, socioeconomic status and/or racial or ethnic grouping and specific treatment outcomes among patients with de Quervain's disease?
This question was addressed by only one relatively small retrospective study. This study found no relation between age, gender or occupational status and treatment outcome. However, it is difficult to reach evidence-based conclusions from the results of a single study of suboptimal design.
Question 7: What are the surgical and nonsurgical costs or charges for treatment of de Quervain's disease?
According to the Medicare Provider Analysis and Review (MEDPAR) database, which covers hospital inpatient services, average total charges per patient for the DRG (diagnosis-related group) of hand or wrist procedures (excepting major joint procedures) without complications or comorbidities are $7,408.14 (calculated by dividing total charges by number of discharges).
The Median Costs for Hospital Outpatient Services Dataset contains median costs for services that are reimbursed under Medicare for the hospital outpatient prospective payment system. The reported median cost for application of a short arm static splint is $72.69.
Question 8: For persons who have had surgery for de Quervain's disease, what are the most effective methods for preventing the recurrence of symptoms, and how does this vary depending on subject characteristics or other underlying health problems?
None of the included studies addressed this question.
Question 9: What instruments, if any, can accurately assess functional limitations in an individual with de Quervain's disease?
None of the included studies addressed this question.
Question 10: What are the functional limitations for an individual with de Quervain's disease before treatment?
None of the included studies addressed this question.
Question 11: What are the functional limitations of an individual with de Quervain's disease after treatment?
None of the included studies addressed this question.
Question 12: What are the cumulative effects on functional abilities among individuals with more than one worker-related musculoskeletal disorder of the upper extremity in the same limb?
There were no studies that met the inclusion criteria for this question. Therefore, it cannot be answered in an evidence-based fashion.
Question 13: What level of function can patients achieve in what period of time when they are required to change hand dominance as a result of injury to their dominant hand?
The studies of the ability of training to improve use of the non-dominant hand do not allow one to determine the degree to which this training provides the patient with employment opportunities or allows resumption of normal activities. These studies also lack long-term followup data. Evidence from two studies suggests that some learning and training in the use of the non-dominant hand is possible, and statistically significant improvement can be accomplished in 2 to 6 months of training. For some activities, statistically significant improvement can be accomplished within 1 week.
In general, the literature addressing WRUEDs is of uneven quality. Well-designed studies on many aspects of WRUEDs are needed. Prospective, randomized double-blinded controlled trials are widely considered to provide the highest quality of evidence for treatment effectiveness. Results of non-randomized trials can be affected by differences in the characteristics of the patient groups, rather than the treatment applied. Uncontrolled trials do not allow one to ascertain whether patients improve in the absence of treatment, and they do not allow one to accurately gauge the magnitude of any change that occurs after treatment. Blinding of patients and evaluators to treatments avoids the potential for placebo effects and previously held beliefs about the effectiveness of treatments to impact on the results of trials.
Studies of diagnostic tests do not necessarily need not be randomized or contain control groups. In the absence of a “gold standard” test, longitudinal studies are the most desirable for assessing diagnostic tests for WRUEDs. In these studies, patients are first given the diagnostic test, and then they are followed for a period of time to determine whether they develop symptoms of a WRUED. Repeating the tests at regular intervals during the trial could yield insights into the etiology of the conditions as well as measure test-retest variability. If a “gold standard” test were developed, then single-arm cross-sectional studies that compared the results of the “gold standard” test to the results of the test under investigation would be appropriate. In such studies, in order to obtain the most useful information, it is important to select a patient population that closely resembles the general population on whom the diagnostic test would ultimately be used.
Worker-related upper-extremity disorders (WRUEDs) result in pain, disability, and loss of productivity. This report is a systematic analysis of the evidence pertaining to thirteen key questions and four specific disorders. These disorders are considered worker-related not because they are necessarily caused by working, but because they effect workers.
Although a wide variety of WRUEDs have been described in the medical literature, this report is limited to four. They are:
Carpal tunnel syndrome
Cubital tunnel syndrome
Epicondylitis
De Quervain's disease
This report addresses 13 questions regarding worker-related disorders of the upper extremity. Eleven of these are condition specific. Therefore, we individually address them for each of the disorders we consider. Questions 12 and 13 are not condition-specific. Therefore, they are answered only once. The questions we address are:
Condition-Specific Questions:
Question #1: What are the most effective methods and approaches for the early identification and diagnosis of worker-related musculoskeletal disorders of the upper extremity?
Question #2: What are the specific indications for surgery for worker-related musculoskeletal disorders of the upper extremity?
Question #3: What are the relative benefits and harms of various surgical and nonsurgical interventions for persons with worker-related musculoskeletal disorders of the upper extremity?
Question #4: Is there a relationship between specific clinical findings and specific treatment outcomes among patients with worker-related musculoskeletal disorders of the upper extremity?
Question #5: Is there a relationship between duration of symptoms and specific treatment outcomes among patients with worker-related musculoskeletal disorders of the upper extremity?
Question #6: Is there a relationship between factors such as patients' age, gender, socioeconomic status and/or racial or ethnic grouping and specific treatment outcomes among patients with worker-related musculoskeletal disorders of the upper extremity?
Question #7: What are the surgical and nonsurgical costs or charges for treatment of worker-related musculoskeletal disorders of the upper extremity?
Question #8: For persons who have had surgery for worker-related musculoskeletal disorders of the upper extremity, what are the most effective methods for preventing the recurrence of symptoms, and how does this vary depending on subject characteristics or other underlying health problems?
Question #9: What instruments, if any, can accurately assess functional limitations in an individual with a worker-related disorder of the upper extremity?
Question #10: What are the functional limitations for an individual with a worker-related musculoskeletal disorder of the upper extremity before treatment?
Question #11: What are the functional limitations of an individual with a worker-related musculoskeletal disorder of the upper extremity after treatment?
Non-Condition-Specific Questions:
Question #12: What are the cumulative effects on functional abilities among individuals with more than one worker-related musculoskeletal disorder of the upper extremity in the same limb?
Question #13: What level of function can patients achieved in what period of time when they are required to change hand dominance as a result of injury to their dominant hand?
Carpal tunnel syndrome (CTS) results from compression of the median nerve as it passes through the carpal tunnel from the wrist to the hand. This leads to progressive sensory and motor disturbances.
Symptoms of CTS include paresthesia (tingling), anesthesia (numbness), diminished or altered sensation (hypoesthesia or dysesthesia) in the affected area of the hand; pain in the hand and arm, and/or the impairment of motor function, particularly of the abilities to grip and grasp.2 Usually the symptoms appear first (and worst) at nighttime.3 In about 1% of cases, permanent nerve damage results, resulting in impaired use of the hands.4 Continued denervation can lead to atrophy of the innervated muscle.5
) that is bordered on
three sides by the carpal bones and on the other by the flexor
retinaculum (or transverse carpal ligament). The median nerve shares the
carpal tunnel with nine flexor tendons that displace the nerve to the
superficial (palm-most) side of the tunnel, directly against the
transverse carpal ligament (See figure
2). The nerve is the softest and most sensitive element in
the tunnel. Anything that decreases the size of the tunnel or increases
the size of its contents can cause CTS. This may include space-occupying
lesions, arthritis, trauma, edema, and dislocation of the lunate bone.Carpal tunnel syndrome is often idiopathic. The most common attributed cause of CTS is tenosynovitis or hypertrophy of the tendon sheaths of the finger flexor tendons due to overuse, often from the repetitive hand motions associated with certain occupations.6, 7 Assemblers, cashiers, and secretaries are among those most prone to the disease, with data-entry keyers, typists, and office clerks also at high risk.4 It is not clear, however, whether occupational activities cause or merely contribute to development of CTS.8 Female sex, middle age, diabetes, alcoholism, hypothyroidism, obesity, pregnancy, menopause, and the use of birth control pills are all associated with CTS.9
CTS is associated with several conditions. Rheumatoid involvement in the wrist joint may lead to carpal tunnel compression.3 Bone growth due to acromegaly may lead to shrinking of the carpal tunnel and median nerve compression.10 Patients receiving hemodialysis may develop CTS because of edema or amyloid deposits in the carpal tunnel.7, 11 Tissue deposits due to gout may also cause or exacerbate CTS.12
Carpal tunnel syndrome may be exacerbated by other nerve injuries, such as at the neck, shoulder, elbow, or by generalized peripheral neuropathies. This phenomenon, known as double-crush syndrome,13 has not been definitively established to exist, and remains controversial.14 Comorbidities causing peripheral neuropathy such as diabetes or thyroid disturbances may both exacerbate CTS and interfere with its diagnosis.15–18 CTS associated with pregnancy, childbirth and lactation may resolve spontaneously.19
The overall prevalence of CTS in the United States may be as high as 1.9 million people, and each year there are 300,000–500,000 operations for the condition, at a total cost of more than $2 billion.20 There are no widely accepted figures for the fraction of cases requiring surgery. Estimates range from nearly half of all CTS patients with occupational disease to a “small percentage” of all patients.20
The incidence of CTS is higher in women than in men, and differences in carpal tunnel volume between men and women may contribute to these differences.21 Idiopathic CTS occurs in women three to five times more frequently than in men.22 Many of the occupations associated with CTS are held disproportionately by women, and several of the causal medical conditions are found more often in women than in men.20 In addition, the prevalence for men generally increases steadily with increasing age while, for women, the prevalence peaks dramatically during middle age (45–55 years of age) and then levels off.23, 24
About 60% of cases are seen in patients between 40 and 60 years of age.25 Whites have been reported to have a 1.8 times higher prevalence of carpal tunnel syndrome than do non-whites.26
The U.S. Bureau of Labor Statistics reported 29,937 cases of CTS that resulted in work days lost in 1996, and the National Institute of Occupational Safety and Health (NIOSH) reported that, in 1993, CTS occurred at a rate of 5.2 per 10,000 full-time workers. This syndrome required the longest recuperation period of all conditions that result in lost work days, with a median of 30 work days lost.4 A study of all surgeries performed to treat carpal tunnel syndrome in Wisconsin from July 1990 to March 1993 found that 75% of the individuals had only one surgery, 24.7% had two surgeries, and 0.3% had three or more surgeries. Workers' Compensation paid for 26.1% of these surgeries.23
Diagnosis of carpal tunnel syndrome is complicated by the fact that there is no “gold standard” method for verifying its presence or absence.27 A variety of diagnostic instruments have been used by investigators including clinical signs, sensory tests, nerve conduction studies, and imaging tests. It is not known which modality or combination of modalities are optimal for the diagnosis of carpal tunnel syndrome.
Most clinical tests to diagnose carpal tunnel syndrome involve specific maneuvers that elicit pain, numbness, or tingling in the median-nerve portion of the wrist. For example, in Phalen's test, the patient places both elbows on a horizontal surface with the forearms vertical, and allows the wrists to flex by gravity. If the patient feels numbness or tingling within one minute, the test is positive.28 In Tinel's test, the examiner taps lightly on the palmar aspect of the wrist, over the carpal tunnel. If the patient feels tingling, the test is positive.29
Sensory tests for carpal tunnel syndrome typically involve measurement of a patient's threshold for detection of a sensory stimulus. For example, in the Semmes-Weinstein test, the examiner touches the patient with monofilaments, and the test is positive if the patient's sensitivity to the monofilaments falls outside normal limits.30 Another example is the two-point discrimination test in which the examiner touches two closely-spaced prongs to the patient's fingers. The test is positive if the patient cannot discriminate the prongs when they are 5 millimeters apart.31
Nerve conduction tests are also used to diagnose CTS. In such tests, electrodes are placed in two locations along a nerve; the nerve is stimulated from one electrode, and the impulse is recorded from the other electrode. Tests can be performed on either the median nerve, ulnar nerve, or radial nerve, and can assess either motor or sensory function. The placement of electrodes in sensory nerve conduction tests can be either orthodromic (in which stimulating electrodes are placed distal to recording electrodes) or antidromic (in which stimulating electrodes are placed proximal to recording electrodes). Other aspects of the nerve impulse can also be measured such as latency, amplitude, and velocity. Some investigators compare two or more nerve conduction tests in an attempt to assist the diagnosis of carpal tunnel syndrome (e.g., compute a difference between two latencies). We refer to these comparisons as composite nerve conduction tests.
Imaging tests for carpal tunnel syndrome include magnetic resonance imaging (MRI), computed tomography (CT), scan x-ray film, and ultrasound. Using these methods, investigators attempt to measure the size of anatomical areas within the carpal tunnel or other areas that may be affected by carpal tunnel syndrome.
Nonsurgical interventions that have been used to treat CTS include wrist splints, avoidance of precipitating activities, anti-inflammatory drugs, vitamin B6, diuretics, ultrasound, injection of anti-inflammatory steroids and physical therapy.17, 32–36 Treatment of comorbid conditions contributing to CTS may also be effective.37, 38
The standard surgery for CTS is the transection of the transverse carpal ligament.39 This transection may be accomplished by endoscopic or open surgery. For virtually all patients it is an outpatient procedure performed in an ambulatory surgical center under regional anesthesia, but a few patients request general anesthesia. A variety of endoscopic techniques have been reported.40–46 Variations in technique include the specific types of equipment used and whether the technique requires one or two incisions. No published evidence is available quantifying the relative advantages and disadvantages of the various methods.
Additional procedures, such as ligament repair or neural surgery may also be used. Ligament reconstruction involves the reattachment of the transected ends of the transverse carpal ligament in such a way that the overall ligament is lengthened. This results in an enlargement of the carpal tunnel and relief of the pressure on the median nerve.47–49
Neural surgery for CTS (external or internal neurolysis or epineurotomy) is generally performed immediately following the division of the transverse carpal ligament. The term “neurolysis” is used to encompass several different procedures.50 These include removal of adhesions from the connective tissue surrounding the nerve (the epineurium), relieving pressure within the epineurium by means of a longditudinal incision, or removal of a segment of epineurium. There is confusion due to the nonstandard usage of terms, compounded by the different subspecialties and nationalities of surgeons. The common goal in all techniques is to remove adhesions and scar tissue to decompress the nerve and allow it to glide freely.
Patients with cubital tunnel syndrome are affected by a weak grip, lack of hand coordination, hand clumsiness, and numbness, paresthesia, and pain in the hand, particularly in the fourth and fifth digits. These symptoms are thought to be caused by compression of the ulnar nerve at multiple sites in the area of the elbow, where the nerve passes through an anatomically restricted area called the cubital tunnel.
Patients presenting with cubital tunnel syndrome usually complain of a weak grip, hand clumsiness and lack of coordination, and dropping of objects. Numbness and paresthesia in the fourth and fifth digits may also be present, in particular after prolonged flexion of the elbow.51 Pain in the hand may be present, but is neither as severe or as common as in carpal tunnel syndrome.52 The medial aspect of the elbow may be painful.53 Severe cases may present with atrophy of the intrinsic muscles and clawing of the fourth and fifth fingers.51
Upon examination, patients with cubital tunnel syndrome are positive for Tinel's sign (tingling in the fingers after tapping over the ulnar nerve at the elbow), and the ulnar nerve may feel swollen and hard upon palpation.52 In addition, patients have diminished sensation in the fourth and fifth digits (pin-prick or Semmes-Weinstein monofilament testing), weak intrinsic hand muscles, a progressive inability to separate the fingers, and a loss of power grip and dexterity.53 Patients with more advanced cases may exhibit a positive Wartenberg's sign (upon extension of the fingers abduction of the fifth digit occurs) and/or a positive Froment's sign (patient cannot pinch between the index finger and thumb without flexion of the distal phalanx of the thumb).53
Electrodiagnostic tests can be used to confirm a lesion of the ulnar nerve, and to help locate the exact site of compression. Two examples of such tests are motor and sensory conduction velocities across the elbow.54, 55 For motor conduction velocity, stimulating electrodes are placed above and below the elbow, and a recording electrode is placed on the abductor digit minimi (a muscle in the hand that is innervated by the ulnar nerve).54 The measured latencies, along with the measured distances between stimulating and recording electrodes, are used to compute the motor conduction velocity in the across-elbow portion.54 For sensory conduction velocity, the ulnar nerve can be stimulated below the elbow and recorded above the elbow (this placement of electrodes is termed orthodromic because the stimulating electrode is distal to the recording electrode).54 Alternatively, the electrodes can be reversed to yield an antidromic sensory measurement.55 Regardless of whether orthodromic or antidromic placement is employed, the latencies and distances are used to calculate the sensory conduction velocity across the elbow.54, 55
Cubital tunnel syndrome can be confused with compression of nerves at other points. Cervical root lesions, such as compression of the eighth cervical root by a bulging disc, may produce symptoms similar to that of cubital tunnel syndrome.56 Other nerve compression disorders that may produce symptoms similar to that of cubital tunnel syndrome included compression of the medial components of the brachial plexus (thoracic outlet syndrome), compression of the ulnar nerve at the wrist in Guyon's canal (ulnar tunnel syndrome), and compression of the ulnar nerve at more than one point.56
The ulnar nerve carries nerve fibers from the eighth cervical and first thoracic nerves. It passes down the upper arm medial to the brachial artery, then passes through the intermuscular septum and travels towards the elbow near the medial head of the triceps. At the elbow, the ulnar nerve passes behind the medial epicondyle of the humerus in a groove between it and the heads of the flexor carpi ulnaris, the cubital tunnel. The ulnar nerve then enters the forearm between the two heads of the flexor carpi ulnaris muscle and enters the hand.57–59 It is not until the ulnar nerve passes between the two heads of the flexor carpi ulnaris muscle that it begins supplying motor and sensory innervation. It supplies motor innervation to the muscles of the forearm and hand, and sensory innervation to the medial half of the hand, the palm, and the fourth and fifth digits.57
The groove that the ulnar nerve passes through at the elbow is referred to as the cubital tunnel. This tunnel is bounded by the medial epicondyle of the humerus anteriorly (See Figure 3
), the ulnohumeral ligament laterally,
and posteromedially, a fibrous arcade of fascial strands that extends
from the olecranon to the medial epicondyle, bridging the two heads of
the flexor carpi ulnaris muscle.57,
58 Under normal conditions, the capacity of the ulnar tunnel is
greatest during elbow extension. Flexion of the elbow decreases the
volume of the cubital tunnel by tightening the arcuate ligament, bulging
of the medial elbow ligament, and contraction of the flexor carpi
ulnaris muscle.58
Inside the cubital tunnel, the motor fibers to the flexor carpi ulnaris and flexor digitorum profundus are located deep inside the ulnar nerve, while the motor fibers to the hand muscles and sensory fibers to the fingers are located more superficially. This peripheral location places these fibers to the hand at increased risk of damage from compression, and accounts for their early involvement in the development of cubital tunnel syndrome.56
Cubital tunnel syndrome is caused by compression of the ulnar nerve within or near the cubital tunnel. The site of entrapment of the ulnar nerve in the region of the elbow can occasionally occur in locations other than the cubital tunnel, including proximal to the elbow by the medial head of the triceps (the arcade of Struthers), at the elbow by the arcuate ligament, or in the mid-forearm by the flexor carpi ulnaris muscle.53 Chronic reduction in volume of the cubital tunnel results in compression damage and focal ischemia of the nerve. Compression of the ulnar nerve within the cubital tunnel is most often due to constriction of the nerve by the overlying fibrous arcade. Compression can be caused by repetitive trauma, inflammation, idiopathic thickening of Osborne's band, arthritis, hematomas, tumors, bone fragments, and idiopathic persistent epitrochleoanconeus muscle.57, 59 Fractures, dislocations, and direct blunt trauma near the elbow can cause acute compression of the ulnar nerve.59 Cubital tunnel syndrome can be precipitated by general anesthesia, and is thought to be related to compression of the ulnar nerve caused by poor limb positioning, tourniquets, and/or blood pressure cuffs.58, 59 Systemic diseases such as diabetes, kidney disease, amyloidosis, acromegaly, alcoholism, hemophilia, and leprosy can contribute to the development of cubital tunnel syndrome.58
In many patients, no precipitating event can be identified. Compression of the ulnar nerve can be the end result of a pathological cycle of chronic irritation of the nerve. Mild irritation of the nerve can causeinflammation and swelling. These processes restrict movement of the nerve through the cubital tunnel. Failure of the ulnar nerve to slide smoothly during elbow flexion and extension causes the nerve to be stretched, and to rub against surrounding surfaces, damaging the nerve and surrounding tissues, leading to more inflammation, swelling, and the formation of adhesions between the nerve and surrounding tissues, which further restricts nerve movement. Eventually this process leads to chronic compression of the nerve.59 Activities thought to result in repetitive trauma to the ulnar nerve include habitual leaning on the elbow, sleeping with the arms flexed, or performing repetitive elbow flexion-extension motions.
The incidence and prevalence of this disorder has not been established. In Connecticut, 3% of claims for Workers' Compensation for occupational disorders of the upper extremity were reported to be for cubital tunnel syndrome.60 Cubital tunnel syndrome affects men 1.3 to 3 times more often than women.61, 62 Thin women (BMI<22) are reported to have a greater prevalence of cubital tunnel syndrome than heavier women. No association between BMI and cubital tunnel syndrome has been reported for men.61
The choice of how to treat cubital tunnel syndrome is based upon the severity of symptoms upon presentation. Mild cases are usually treated by minimizing elbow flexion through behavioral changes and splinting, minimizing direct pressure on the elbow using pads and pillows, and reducing inflammation with non-steroidal anti-inflammatory drugs (NSAIDs). If symptoms are severe, or do not respond to conservative treatment, then surgery may be performed.63
Surgical techniques used to relieve the compression of the ulnar nerve can be divided into three categories: decompression, epicondylectomy, and transposition of the ulnar nerve.
Decompression is the simplest of the procedures and usually involves cutting the tissues that form the roof of the cubital tunnel.64 The tissues commonly cut during decompression are the medial intermuscular septum, the arcade of Struthers, the superficial fascia, and the deep flexor pronator aponeurosis. Decompression can be performed through an open incision or by endoscopic techniques.65 Cutting the tissues in this fashion is thought to relieve the compression on the nerve that is causing the problem.
Medial epicondylectomy consists of removal of the medial epicondyle, and reattachment of the flexor-pronator muscle groups to the site of removal.66 Decompression is usually performed at the same time. Removal of the epicondyle is thought to allow greater anterior migration of the ulnar nerve upon elbow flexion.63
Transposition of the ulnar nerve describes several different procedures, all of which reposition the ulnar nerve outside of the cubital tunnel, anterior to the medial epicondyle.67 Moving the nerve in this fashion is thought to decrease or eliminate nerve tension and avoid further irritation and compression of the nerve.67 Subcutaneous transposition refers to shifting the ulnar nerve and forming a sling of fascia to hold it in place.68 The nerve can also be placed in a trough inside the flexor-pronator muscle mass (intramuscular transposition). Submuscular transposition (the Learmonth procedure) involves detaching the flexor-pronator muscle mass from the medial epicondyle, moving the ulnar nerve anteriorly and underneath the flexor-pronator muscle to lie on the brachialis fascia near the median nerve, and then re-attaching the flexor-pronator muscles to the epicondyle. Sometimes when using this technique the flexor-pronator muscle is elongated to prevent tension from being placed on the underlying ulnar nerve.69
Patients with epicondylitis experience pain at the elbow. The pain is localized over the affected epicondyle, and becomes severe upon use of the affected muscles when grasping objects.
The chief complaint of patients affected by epicondylitis is an insidious onset of elbow pain. The pain is described as dull and aching when at rest, but becomes sharp and severe upon use of the affected muscles when grasping objects.70 There is tenderness upon palpation over the affected epicondyle. In severe cases, the afflicted person may complain of grip weakness. Upon resisting wrist extension (flexion, for medial epicondylitis), severe pain occurs at the affected epicondyle.53
Diagnosis of epicondylitis is reached by clinical exam and history. In addition to pain upon resisted wrist extension, other clinical signs of epicondylitis include pain upon resisted supination of the forearm, reduced grip strength, and pain upon resisted extension of the middle finger.71–73 In clinically diagnosed cases that do not improve with conservative management, MRI of the elbow has been used to clarify the diagnosis and assess the degree of tendon disease.74
Epicondylitis refers to pain in the area where the muscles of the forearm attach to the epicondyle of the elbow, pain that is worsened by use of these muscles. Epicondylitis is divided into two distinct syndromes: lateral and medial epicondylitis. Lateral epicondylitis, also referred to as tennis elbow, refers to pain in the attachment of the extensor muscles, most commonly the insertion of the extensor carpi radialis brevis tendon, into the lateral epicondyle. Medial epicondylitis, also referred to as golfer's elbow, refers to pain in the attachment of the flexor muscles of the forearm to the medial epicondyle. Lateral epicondylitis is more common than medial epicondylitis.75
A tendon attaches muscle to bone or fascia. The power of the muscle contraction is transmitted down the tendon and causes the attached bone to move. The site of attachment of the tendon to the bone is thus subject to considerable force with each contraction of the muscle.76 Tendonitis and tenosynovitis refer to disorders of the tendon and the synovial membrane of the tendon sheath, respectively. Although historically inflammation was thought to be the pathology underlying tendonitis, chronic degenerative changes in the tendon and synovial tissue appear to be the predominant pathological processes.53, 77
The exact pathology that underlies epicondylitis is not known.70 The problem appears to be confined to the tendinous and fascial attachments to the bone (See Figure 4
). The tendons become dull, gray, friable, and edematous. The
normal tendon fibers become disrupted by invading fibroblasts and
granulation tissue.78 Adhesions may form between the tendon and surrounding tissues.
The extensor carpi radialis brevis tendon appears to be most often
affected because it is intimately attached to the joint capsule, and
because of this proximity adhesions readily form between it and the
joint.Lateral epicondylitis is thought to be a degenerative process caused by overuse of the wrist extensors. Repetitive strong synergic and fixator action of the wrist extensors during gripping are believed to result in minor trauma to the muscle attachment to the epicondyle.75 Continued muscle use prevents healing. Medial epicondylitis is thought to be a similar process affecting the flexor, rather than the extensor, muscles. Forceful, repetitive motions of the forearm are thought to be the initial precipitating factor.79
Epicondylitis has been reported to affect 4.23 individuals per 1000 adults per year in the U.S.80 The mean age of diagnosis is 45 years, and men and women appear to be equally affected.80 Lateral epicondylitis is six times more common than medial epicondylitis.80 Individuals who have been diagnosed with carpal tunnel syndrome have a greater prevalence of lateral epicondylitis than do those without carpal tunnel syndrome.81 Persons who engage in forceful, repetitive forearm work such as mechanics, butchers, and construction workers have a higher prevalence of the condition than the general population.82
Initial treatment of epicondylitis usually involves rest and massage. In addition, a number of conservative therapies are used to treat epicondylitis. These are briefly described below.
Pharmacologic treatments for epicondylitis include NSAIDs, either taken orally or applied topically, topical dimethyl sulfoxide (DMSO), injections of glucocorticoid steroids, injections of anesthetics, and oral glucosamines.
Rest, ice, massage, physiotherapy, manipulations, splints, braces, and exercise programs are commonly used when treating epicondylitis.
Other treatments for epicondylitis include acupuncture, low level red or infrared lasers, ultrasound, phonophoresis, transcutaneous electrical nerve stimulation (TENS), extracorporal shock-wave therapy (ESWT), and pulsed electromagnetic fields (PEMF).
| Category | Type of surgery |
|---|---|
| Denervation | Complete denervation |
| Partial lateral denervation | |
| Partial ventral denervation | |
| Nerve decompression | Decompression of thePIN |
| Decompression of the radial nerve | |
| Combination of denervation and decompression of the PIN | |
| Lengthening of the ERCB | Distal lengthening of the ECRB |
| Proximal lengthening of the ERCB | |
| Removal of tissues | Incision of the ERCB |
| Partial resection of the annular ligament (Bosworth technique) | |
| Epicondylar osteotomy | |
| Epicondylectomy and excision of the distal portion of the annular ligament | |
| Excision of subtendinious pathological tissue | |
| Excision of the subcutaneous tissue | |
| Excision of the radiohumoral bursa | |
| Fasciectomy of the common extensor origin | |
| Fasciectomy plus anconeous transfer | |
| Debriding of the elbow join |
Adapted from Wilhem et al.84
PIN = posterior interosseus nerve
ERCB = extensor carpi radialis brevis tendon
De Quervain's disease is characterized by pain localized on the radial border of the wrist that may also radiate into the thumb and forearm.85 The pain is usually worsened by abduction and/or extension of the thumb.53 Other symptoms may include weakness of the thumb and loss of grip. Range of motion of the wrist and thumb is usually unaffected or only slightly limited.85
De Quervain's disease is a stenosis (thickening) of the fibrous sheath of the first extensor compartment of the extensor retinaculum.86 This compartment surrounds two tendons, the extensor pollicis brevis and the abductor pollicis longus (See Figure 5
). In the past, de Quervain's disease
has been described as a type of stenosing tenosynovitis of the hand and
wrist. Because recent studies have shown that there is no inflammatory
process associated with de Quervain's disease, some experts believe that
the term tenosynovitis is not accurate for describing this
condition.53,
86
Possible etiologic factors include acute trauma, recurrent trauma, or an underlying collagen disease.87
De Quervain's disease appears most frequently in the 30 to 50 year age group and has been reported to be 10 times more common among women than men.85 Work occupations commonly associated with this condition include musicians, weavers, typists, nurses, knitters, golfers, switchboard operators, and manual workers.53, 85 However, there is disagreement among experts as to whether these types of work cause de Quervain's disease or merely exacerbate the symptoms.53, 86 Anatomic variations of the first extensor compartment have also been reported to be associated with de Quervain's disease.86
Diagnosis of de Quervain's disease is usually accomplished by the Finkelstein test. While the patient flexes the thumb within the palm while holding it tightly with the other fingers, the examiner performs an ulnar deviation of the patient's wrist. Intense pain on the styloid process of the radius indicates a positive test. The pain disappears after the thumb is released and extended.85 Additional diagnostic criteria include patient-reported pain at the radial wrist and tenderness to palpation at the radial wrist.53
A number of conservative therapies have been used to treat de Quervain's disease. These include workplace modification, hand rest, neutral wrist splinting with a thumb spica, anti-inflammatory medication, and iontophoresis.53 If these therapies fail, injection of cortisone may be used to supplement splinting and anti-inflammatory medication.
Persistent pain after four to six weeks of conservative therapy is usually considered an indication for surgery.85, 87 This procedure consists of unroofing the retinaculum to release the abductor pollicis longus and extensor pollicis brevis tendon sheaths.87 As noted earlier, anatomic variation exists in that these tendon sheaths may be contained in one or two compartments. Reported complications of surgery include radial sensory nerve injury and painful surgical scarring.88
This evidence report is concerned with worker-related upper extremity disorders. The term “worker-related” implies a disorder that affects workers, not a disorder caused by work. In this report, we address four specific disorders: (1) carpal tunnel syndrome, (2) cubital tunnel syndrome, (3) epicondylitis, and (4) de Quervain's disease. This list of disorders was determined during discussions among ECRI, the Agency for Healthcare Research and Quality (AHRQ), the organizations that nominated this topic to AHRQ, and a panel of technical experts. Below, we provide further details about the nominating organizations and technical experts.
Technical Experts were employed to assist in defining the scope of this evidence report, developing its questions, and developing the criteria for retrieving and including articles. Seven organizations were solicited to nominate individuals who could serve as Technical Experts. All solicitations were pre-approved by AHRQ. All seven organizations nominated an individual. Thus, the Expert Panel was comprised of individuals from the American Association of Electrodiagnostic Medicine, the American Academy of Neurology, the American Academy of Physical Medicine and Rehabilitation, the American Physical Therapy Association, the Association for Repetitive Motion Syndromes, the American Association of Neurological Surgeons, and the American Academy of Orthopedic Surgeons. The participation of these individuals and organizations does not imply their endorsement of the findings of this evidence report.
To determine the specific questions that this evidence report would address, a multidisciplinary team was assembled. This team included ECRI research staff, AHRQ project staff, representatives from the organizations that nominated this topic to AHRQ (the Social Security Administration and the American College of Occupational and Environmental Medicine), and the Technical Experts. The key questions for this report were decided during three conference telephone calls between ECRI, AHRQ, the experts, and the nominating organizations, as well as subsequent discussions between ECRI, AHRQ, and the nomination organizations.
The final set of key questions is comprised of 13 questions, 11 of which are separately addressed for the four above-mentioned disorders. The remaining two questions are not disorder specific. This evidence report is correspondingly organized. Thus, we first address each of the 11 questions for each disorder, beginning with carpal tunnel syndrome, and conclude by addressing the two questions that are not disorder-specific.
The 11 condition specific questions that we address in this report are:
Question #1: What are the most effective methods and approaches for the early identification and diagnosis of worker-related musculoskeletal disorders of the upper extremity?
Question #2: What are the specific indications for surgery for worker-related musculoskeletal disorders of the upper extremity?
Question #3: What are the relative benefits and harms of various surgical and nonsurgical interventions for persons with worker-related musculoskeletal disorders of the upper extremity?
Question #4: Is there a relationship between specific clinical findings and specific treatment outcomes among patients with worker-related musculoskeletal disorders of the upper extremity?
Question #5: Is there a relationship between duration of symptoms and specific treatment outcomes among patients with worker-related musculoskeletal disorders of the upper extremity?
Question #6: Is there a relationship between factors such as patients' age, gender, socioeconomic status and/or racial or ethnic grouping and specific treatment outcomes among patients with worker-related musculoskeletal disorders of the upper extremity?
Question #7: What are the surgical and nonsurgical costs or charges for treatment of worker-related musculoskeletal disorders of the upper extremity?
Question #8: For persons who have had surgery for worker-related musculoskeletal disorders of the upper extremity, what are the most effective methods for preventing the recurrence of symptoms, and how does this vary depending on subject characteristics or other underlying health problems?
Question #9: What instruments, if any, can accurately assess functional limitations in an individual with a worker-related disorder of the upper extremity?
Question #10: What are the functional limitations for an individual with a worker-related musculoskeletal disorder of the upper extremity before treatment?
Question #11: What are the functional limitations of an individual with a worker-related musculoskeletal disorder of the upper extremity after treatment?
The two questions that are not condition specific are:
Question #12: What are the cumulative effects on functional abilities among individuals with more than one worker-related musculoskeletal disorder of the upper extremity in the same limb?
Question #13: What level of function can one achieve in what period of time when one is required to change hand dominance as a result of injury to his or her dominant hand?
The scope of this report can be illustrated by a causal pathway. More specifically, this pathway illustrates the key questions and the relationships among them. It also illustrates items that are beyond the scope of this evidence report. This pathway is shown in Figure 6
. The rectangles in this figure depict the primary clinical “events”,
from presentation of a patient (who has certain characteristics that may be at
least partly diagnostic and/or prognostic) to the outcomes that the patient
experiences (e.g., improves/does not improve). That this, in fact, is a pathway
that proceeds in a certain chronological order is depicted by solid arrows that
connect the rectangles in Figure 6.
Because these arrows connect two rectangles, they are termed “links.” The
numbers next to each link represent the numbers of the Key Questions that
address that link. Key Question 7 is not shown in the pathway because it is
concerned with costs and, therefore, is not part of the clinical pathway.
The dashed lines in the figure “overarch” several rectangles. We have drawn these lines as dashed because they do not depict the sequence of events in the clinical pathway. In general, these lines portray Key Questions about how patient characteristics (including clinical findings) may influence a patient's movement through the clinical pathway or whether these characteristics influence outcomes.
. Therefore,
rectangles not connected by solid or dashed lines are beyond the scope of this
evidence report.Our searches for information were designed to produce a comprehensive dataset. Therefore, we searched a number of electronic databases and other sources. These are described below.
We searched 31 electronic databases. These databases were:
CISILO Database (International Occupational Safety and Health Information Centre) (through November 2000)
The Cochrane Database of Systematic Reviews (through 2000, Issue 4)
The Cochrane Registry of Clinical Trials (through 2000, Issue 4)
The Cochrane Review Methodology Database (through 2000, Issue 4)
CRISP (Computer Retrieval of Information on Scientific Projects) (through November 16, 2000)
Cumulative Index to Nursing and Allied Health (CINAHL) (1988 through September 29, 2000)
Current Contents (through December 2000)
The Database of Reviews of Effectiveness (Cochrane Library) (through 2000, Issue 4)
DIRLINE (through September 27, 2000)
ECRI Health Devices Alerts (1977 through January 2001)
ECRI Health Devices Sourcebase (through January 2001)
ECRI Healthcare Standards (1975 through January 2001)
ECRI International Health Technology Assessment (IHTA) (1990 through January 2001)
ECRI Library Catalog (through January 2001)
ECRI TARGET (ECRI's database of emerging technologies; through January 2001)
Embase (Excerpta Medica) (1974 through December 12, 2000)
ERIC (Educational Resources Information Center) (searched June 28, 2000)
Health and Psychosocial Instruments (HAPI) (through January 30, 2001)
Health Services Research Projects (HSRPROJ) (through September 27, 2000)
HealthSTAR (Health Services, Technology, Administration, and Research) (1990 through September 26, 2000)
LocatorPlus (through January 2001)
NIOSHTIC (through November 3, 2000)
Old Medline (1957 -1965) (searched September 27, 2000)
PsycINFO (1967 through January 22,2001)
PubMed (1966 through January 22, 2001)
Rehabdata (through November 2000)
SciSearch (through November 13, 2000)
U.K. National Health Service (NHS) Economic Evaluation Database (NHS EED) (through January 2001)
U.S. Health Care Financing Administration (HCFA) (through January 2001)
U.S. National Guidelines Clearinghouse (NGC) (through January 2001)
U.S. National Institutes of Health Web site (NIH) (through January 2001)
To further ensure that this evidence report was comprehensive, we also searched the World Wide Web using various resources and search engines including AltaVista, NorthernLight, and Google. These resources included:
American Academy of Orthopedic Surgeons http://www3.aaos.org
American College of Occupational and Environmental Medicine (ACOEM) http://www.acoemwebapps.org/gov/welcomeNS.asp
Association for Repetitive Motion Syndromes (ARMS) http://www.certifiedpst.com/arms/
Canadian Centre for Occupational Health and Safety (CCOHS) http://www.ccohs.ca/
Centre for Clinical Effectiveness http://www.med.monash.edu.au/publichealth/cce/
Development Evaluation Committee http://www.hta.nhsweb.nhs.uk/rapidhta/main.htm
ErgoWeb http://www.ergoweb.com/
HCUPnet http://www.ahcpr.gov/data/hcup/hcupnet.htm
Medscape http://www.Medscape.com
National Institute for Occupational Safety and Health (NIOSH) http://www.cdc.gov/niosh/homepage.html
NHS Centre for Reviews and Dissemination http://www.york.ac.uk/inst/crd/welcome.htm
Safety and Health Statistics, Bureau of Labor Statistics http://stats.bls.gov/oshhome.htm
SUM Search http://sumsearch.uthscsa.edu/searchform4.htm
TRIP Database http://www.tripdatabase.com/
In addition to the above searches, we also reviewed the bibliographies and reference lists of all studies included in this evidence report, searched Current Contents—Clinical Medicine on a weekly basis, and routinely reviewed over 1,600 journals and supplements maintained in ECRI's collections.
We searched four additional U.S. government datasets solely to obtain information about costs. These were:
2001 Physician Fee Schedule. This Health Care Financing Administration (HCFA) dataset contains fees and limiting charges for physician services under Medicare in 2001.
Median Costs for Hospital Outpatient Services Dataset. This HCFA dataset contains median costs, by HCPCS codes, for services reimbursed under the hospital outpatient prospective payment system. The data are calculated based on 1996 hospital outpatient claims.
Medicare Provider Analysis and Review (MEDPAR). This HCFA dataset contains information for 100% of Medicare beneficiaries using hospital inpatient services. The data are provided by state and then by diagnostic related group (DRG) for all short stay and inpatient hospitals for fiscal years 1990-1996. Data include total charges, covered charges, Medicare reimbursement, total days, number of discharges, and average total days.
Hospital Outpatient Prospective Payment System. This HCFA dataset contains rules for payment of outpatient services provided by hospitals or affiliated organizations under hospital control. The system is based on ambulatory payment classifications (APCs). This classification system groups services both clinically and by resource utilization.
The systematic nature of the searches for information for an evidence report is a means of diminishing reviewer bias during the preparation of a report. This systematic nature is reflected in our strategies for searching PubMed/Medline and HCUPnet for ICD-9 procedure codes and CPT codes, diagnostic related groupings (DRGs), ambulatory related groupings (ARGs), and HCPS codes. These strategies are detailed in Appendix A.
To be included in this evidence report, an article had to meet two sequentially applied sets of a priori criteria. The first set determined whether a full article would be retrieved. The second set, which was based on major study design flaws and certain elements specific to each question, determined whether a retrieved article would be included in the report. To facilitate comprehensive article retrieval, the retrieval criteria were designed to be broad.
The abstracts of articles identified by our searches were reviewed against the retrieval criteria to determine whether we would retrieve an article identified by our searches. This task was independently performed by six research analysts, each of whom individually worked on different questions. We retrieved an article whenever there was uncertainty about whether it met the retrieval criteria. We also retrieved articles when an abstract was not present in the search results, but when the title of the article suggested that it was relevant.
The criteria for article retrieval were:
All patients, or a separately reported subset of patients in any given article, had to be diagnosed with a worker-related disorder of the upper extremity. No restrictions were placed on the patient populations in clinical trials of conservative or surgical treatments that were retrieved for this analysis. For studies addressing condition-specific key questions, patients had to be diagnosed with the specific disorder of interest.
All controlled trials were retrieved, regardless of whether they were described as randomized or prospective. There was no cutoff date for year of publication. Included in the retrieved articles were those that compared a treatment to a placebo, sham, or untreated group and those that compared two or more treatments.
Case series and other reports were evaluated only if published in 1980 or later. This was an arbitrary cut-off date set to exclude case series using obsolete techniques and outdated patient selection criteria.
Case series had to enroll 10 or more patients. Studies with less than 10 patients are unlikely to be representative of the range of patients with the disorder being evaluated.
Only English-language articles were retrieved.
Once an article was retrieved, it was examined to determine whether it suffered from a major design flaw and whether it met certain question-specific criteria. When an article was excluded, the research analysts entered a unique article identifier and the reason(s) for exclusion into an electronic data abstraction form (DAF).
When an article was included, the unique identifier and details about the studies results, design, and enrolled patient population were entered in these forms. Additional details about the DAFs are provided below.
Many of our exclusions were made because an article contained a significant design flaw. To avoid redundancy, we do not list these flaws here. Rather, we provide a listing of the major design flaws used to exclude articles in the sections of this report in which we evaluate the quality of the literature. Below, we provide the inclusion criteria that are unique to each question:
Question 1. What are the most effective methods and approaches for the early identification and diagnosis of worker-related musculoskeletal disorders of the upper extremity?
Studies meeting the retrieval criteria were included:
Only if they reported sensitivity and specificity or provided sufficient data to allow us to compute these measures of test performance.
If they did not use obsolete tests (e.g., first- and second-generation CT scanners).
Regardless of whether they were prospective or retrospective.
Regardless of whether they contained a concurrent control group. Use of controlled and particularly randomized controlled studies is exceedingly rare in the evaluation of any diagnostic test. Often, such controls are not needed because the patients can validly serve as their own controls.
Question 2. What are the specific indications for surgery for worker-related musculoskeletal disorders of the upper extremity?
To address this question, we tabulated the characteristics of patients enrolled in clinical studies. Doing so does not require any particular study design, and this is reflected in our inclusion criteria. Thus, among the studies that met the retrieval criteria, we included:
Controlled trials and case series of surgical patients
Studies in which not all patients received surgery were included, but only if characteristics of patients receiving surgery were reported on separately.
Studies that did not exclusively enroll patients with co-morbidities not routinely encountered during routine clinical practice (e.g., patients with amyloidosis).
Question 3. What are the relative benefits and harms of various surgical and nonsurgical interventions for persons with worker-related musculoskeletal disorders of the upper extremity?
Among studies meeting the retrieval criteria we included:
Controlled studies, regardless of whether they were randomized or blinded.
Studies that were not exclusively dedicated to comparing highly similar treatment variations (such as incision shape).
Studies that reported on at least one of the seven key outcomes addressed in this assessment. The outcomes are: pain, function, quality of life, ability to return to work, ability to return to activities of daily living, harms, and global outcome.
Question 4. Is there a relationship between specific clinical findings and specific treatment outcomes among patients with worker-related musculoskeletal disorders of the upper extremity?
We evaluated controlled trials and case series that attempted to correlate patient-oriented outcomes with specific clinical findings, patient characteristics or duration of symptoms. It is not feasible to conduct randomized controlled trials that address this question because, by definition, one cannot fully randomize patients with different pretreatment clinical findings into different groups. Therefore, the inclusion criteria adopted for this question were:
Studies that evaluated the relationship of pretreatment clinical findings and outcomes using multiple linear or logistic regression.
Studies that statistically compared the outcomes of patients stratified across some pretreatment clinical finding.
Studies reporting patient-level data were included when the data were presented in enough detail to allow us to perform independent multiple regression analyses.
Studies that reported on at least one of the seven key outcomes addressed in this assessment. The outcomes are: pain, function, quality of life, ability to return to work, ability to return to activities of daily living, harms, and global outcome.
Studies that examined a simple correlation between a given pretreatment variable and outcomes were included, even if they did not attempt to control for the effects of other predictor variables. However, we only included such studies if there were at least three studies that attempted to correlate the same outcome with the same predictor variable. We adopted the arbitrary criterion of requiring three correlational studies because, when taken individually, interpretation of such studies is difficult. This is because they do not contain information about potential inter-variable multicolinearity.
Question 5. Is there a relationship between duration of symptoms and specific treatment outcomes among patients with worker-related musculoskeletal disorders of the upper extremity?
The criteria used for this question were identical to those used for Question 4.
Question 6. Is there a relationship between factors such as patients' age, gender, socioeconomic status and/or racial or ethnic grouping and specific treatment outcomes among patients with worker-related musculoskeletal disorders of the upper extremity?
The criteria used for this question were identical to those used for Question 4.
Question 7. What are the surgical and nonsurgical costs or charges for treatment of worker-related musculoskeletal disorders of the upper extremity?
Cost and charge information from large national databases was included.
Question 8. For persons who have had surgery for worker-related musculoskeletal disorders of the upper extremity, what are the most effective methods for preventing the recurrence of symptoms, and how does this vary depending on subject characteristics or other underlying health problems?
Controlled trials of any design (RCTs, prospective non-randomized, and retrospective) were included.
Question 9. What instruments, if any, can accurately assess functional limitations in an individual with a worker-related disorder of the upper extremity?
For inclusion in this question, a study meeting the retrieval criteria had to be:
A case series or controlled study that measured the validity, response to treatment, or test-test reliability of the assessment instrument.
A study not exclusively devoted to measuring the internal consistency of an instrument. Although internal consistency is important in instrument development, it does not directly address the ability of an instrument to predict functional limitations.89
A study of an instrument designed to evaluate patient function. Instruments that only evaluated symptoms or that were primarily designed to aid in diagnosis were not included.
A study of an instrument that enrolled patients with one of the four specific disorders of interest.
Question 10. What are the functional limitations for an individual with a worker-related musculoskeletal disorder of the upper extremity before treatment?
In addressing this question, we tabulate functional limitations. Answering this question does not require randomized controlled trials. Therefore, our inclusion criteria for studies meeting the retrieval criteria were:
All studies, regardless of design
Studies that measured functional disability using one of the instruments identified in Question 9
Studies that exclusively enrolled patients with one of four conditions of interest.
Studies reporting on functional ability using portions of these instruments or minor variations of these instruments were included as well.
Study must not have enrolled patients who received prior treatment.
Question 11. What are the functional limitations of an individual with a worker-related musculoskeletal disorder of the upper extremity after treatment?
This question is similar to Question 10 and, therefore, identical inclusion criteria were employed except for the one requiring that patients must not have had prior treatment. To be included for Question 11, the study must have been of patients who received prior treatment.
Question 12. What are the cumulative effects on functional abilities among individuals with more than one worker-related musculoskeletal disorder of the upper extremity in the same limb?
The criteria for this question were identical to those used for Question 11, except that the study must have reported data on the patient population relevant to Question 12.
Question 13. What level of function can one achieve in what period of time when one is required to change hand dominance as a result of injury to his or her dominant hand?
This question also does not depend on randomized controlled trials. Therefore, we included any retrieved study, regardless of design, that employed any test of functional ability in patients required to change hand dominance as a result of injury to the dominant hand.
Data from all articles that met our inclusion criteria were abstracted using electronic data abstraction forms. These forms were created using Microsoft Access. Using this software, separate data abstraction forms were designed for entering data about basic trial design information; patient signs, symptoms, comorbidities, characteristics, and treatments; reporting of treatment outcomes; surgical complications; and nerve conduction measurements. The data abstraction forms are presented in the appendix B.
The abstraction form for trial information contained information on trial design, purpose, author, year of publication, general diagnosis of patient condition, a specific description of the treatment outcomes examined, inclusion and exclusion criteria, and other important information with which to judge the quality of the trial. One record containing a unique trial identification number appears for each trial entered in the database.
The abstraction form for patient characteristics and treatments was designed to contain information on each patient group within a trial. A separate record containing a unique patient group identification number appears for each patient group within a trial. This form contained entries for treatment given to the patient group, stratification of patient groups based on pretreatment characteristics, number of patients in the group, specific descriptions of patient treatment, and patient characteristics such as age, dropouts, signs, symptoms, disease severity and duration of symptoms prior to treatment.
Abstraction forms with similar design were created to contain information on treatment outcomes. Separate abstraction forms were needed for dichotomous, categorical and continuous outcome data. These forms contained entries for the patient group identification number, number of patients reporting the outcomes, and time the outcome was measured. A separate record was entered for each patient group and each follow up time for which an outcome was reported.
Special forms were designed for symptoms, comorbidities, complications, and results of diagnostic tests.
Because diagnostic trials differ from treatment trials in many important ways, several special forms were used in the abstraction of diagnostic data, and irrelevant sections of the other data abstraction forms were not completed.
One clinical trial information form and one diagnostic clinical trial information form were completed for each study; not all of the fields in the clinical trial information form were relevant to the diagnostic studies. One patient groups—diagnostics and characteristics form was completed for each patient group or subgroup in each study. Most articles from which we abstracted data reported on two groups; some reported more. One diagnostic test information form was completed for each diagnostic test result reported in each study. Because separate forms were completed for each test parameter reported (e.g. distal motor latency v. distal sensory latency), most studies required more than one form and several required 30 or more forms. One study reported 57 different tests.90
| Question # | Carpal Tunnel | Cubital Tunnel | Epicondylitis | De Quervain's |
|---|---|---|---|---|
| 1 | 189 | 20 | 10 | 0 |
| 2 | 145 | 32 | 19 | 3 |
| 3 | 44 | 3 | 50 | 1 |
| 4 | 12 | 11 | 3 | 1 |
| 5 | 5 | 14 | 7 | 1 |
| 6 | 21 | 15 | 6 | 1 |
| 8 | 0 | 0 | 0 | 0 |
| 9 | 8 | 0 | 3 | 0 |
| 10 | 2 | 0 | 2 | 0 |
| 11 | 12 | 0 | 0 | 0 |
| Code | Definition |
|---|---|
| WRUED groups | |
| Symptoms/presented | Patients had unspecified symptoms of the disorder being studied, or were referred for diagnosis of suspected WRUED |
| Simple signs/symptoms | Patients included if they had specified symptoms of the disorder, but other tests such as nerve conduction tests were not used for patient selection |
| Simple NCS | Patients included if they had abnormal results in a specific nerve conduction test or tests (no more than three tests in selection algorithm) |
| Complex objective standard | A specified algorithm with more than three nerve conduction studies or combining specific NCS tests with specific symptoms |
| Unspecified (diagnosed) | Authors reported that all patients had been diagnosed with the disorder in question, but did not detail how the diagnosis was defined |
| Other | Details reported in separate database field |
| Control groups | |
| Healthy volunteers | Subjects drawn from hospital or community populations, and not being evaluated for other upper extremity disorders |
| Workers at risk | Asymptomatic individuals considered to be at risk for WRUED |
| Unrelated disease | Subjects were being evaluated or treated for known abnormalities of the hand or wrist unrelated to WRUEDs |
| Contralateral arm | Unaffected contralateral extremity of persons with diagnosed WRUED |
| Other | Details reported in separate database field |
| Test group | Included tests |
|---|---|
| Imaging tests | Radiography (film x-ray), computed tomography, MRI, ultrasound |
| Nerve conduction | Amplitude, latency, and velocity of signal conduction in median and ulnar nerves |
| Composite nerve conduction | Differences and ratios of nerve conduction test results |
| Signs and symptoms | Phalen's maneuver, reverse Phalen maneuver, Tinel's sign, Durkin (carpal compression) test, sensory diagrams |
| Sensory tests | Semmes-Weinstein monofilament test, vibrometry, current perception threshold |
| Reporting level | Definition |
|---|---|
| Patient-level | Results for each patient reported individually. This includes studies where patient-level results were reported in a graph rather than a table. Where possible, ECRI research analysts |
| Counts | Sufficient data to yield a two-by-two truth table relating test results to another condition (usually patient's assignment to disease or control group) |
| Summary statistics | Mean and standard deviation of results for all patients in the group |
| Agreement or difference | Statistics reporting agreement or difference between results of one test and another, but not the results themselves |
| Technical criteria | Accuracy, precision, and reproducibility of the test results, but not the results themselves. |
| Characteristic | Definition |
|---|---|
| Longitudinal data | Study reported repeated measurements on the same subjects, from which information on the progression of the condition can possibly be derived |
| Early diagnosis | Study reported that it was intended to identify early-stage disease. For purposes of this assessment, we relied on the authors' own definitions of “early diagnosis” and did not try to validate that validate that description. |
| Screening study | Study included at least one group of subjects that can be considered a screening population (e.g. asymptomatic individuals whose work entails repetitive movements). |
For the two questions that were not condition specific, Questions 12 and 13, we included 0 and 2 articles, respectively. Question 7 is not depicted in the above table because we addressed it using information from a national database, not published articles.
Because this is a “best evidence” synthesis, we incorporated studies that represented the best available evidence, not the best possible evidence. Therefore, not all evidence that we included is of equal quality.
The quality of studies of treatments that we evaluated can be ranked according to the following hierarchy:
Randomized controlled trials
Other prospective controlled trials
Retrospective controlled trials, including those with historical control groups
Prospective case series
Retrospective case series
The hierarchy, like any evidence hierarchy, is only a rough guide. As noted above, randomized controlled trials are not necessary for some of the questions (among which are questions about diagnostics) that we addressed. In such cases, this hierarchy is not applicable. Therefore, for these questions, we discuss the dimensions along which we evaluated the quality of the literature when we address that question. These discussions appear in the appropriate Internal Validity sections under each of these questions.
Meta-analyses of studies of treatments were conducted using Hedges' d as a measure of each study's effect size, and then computing the precision-weighted summary d from the combined results of all studies.91 Hedges' d is the difference between the means of any study's two groups expressed in standard deviation units. We performed meta-analyses on data from studies of treatments only when four or more controlled studies of a given treatment reported the same outcome. We did not perform meta-analyses of smaller data sets because of the high potential for publication bias to affect their results.
For computation of effect sizes derived from dichotomous outcomes, we converted the odds ratio to Hedges' d as described by Hasselblad and Hedges.92 For computation of effect sizes derived from rating scale data, we calculated a mean for each group as described by Torgenson (his equations 71–78).93 An advantage of this method is that it does not assume that all patients employ exactly the same boundaries for each category in a rating scale.
We employed two tests for heterogeneity, the Q statistic and each study's
standardized residual. We regarded the data as heterogeneous if the results
of either test was statistically significant. When we detected
heterogeneity, we analyzed the data for sources of heterogeneity. It was not
always possible to find a source, particularly when there were only a small
number of studies in the meta-analyses. These models were computed using a
modified method of moments.94 To further assist in interpreting the results of our meta-analyses,
we present the results of our fixed effects models in terms of Forrest plots
and as a pair of normal curves. Each curves represents the distribution of
results in a study's two groups. The difference between the means of these
two normal curves represents d, the effect size. We quantified the degree of
the non-overlap of these two curves using the ![[union or logical sum]](corehtml/pmc/pmcents/x222A.gif)
statistics described by
Cohen95, and have expressed these results in terms of the overlap between
these curves.
Diagnostic test meta-analyses were done according to the method of Littenberg and Moses.1 Meta-analyses of diagnostic studies were performed only when there were 10 or more retrieved trials of a given test. We adopted this criterion to ensure that this evidence report would focus on the tests for which there is the greatest research interest. We have taken the mean threshold as the best estimate of a single threshold, and the values of sensitivity and specificity at the mean threshold as the single best global estimate of test effectiveness.
Before using the results of a meta-analysis, we verified that there was no statistically significant heterogeneity among the results of the included articles. This was accomplished, using the Q statistic, as described by Hasselblad and Hedges.92 The presence of heterogeneity indicates that something other than threshold is affecting sensitivity and specificity, and that the points on an ROC curve are not derived from the same population of sensitivity/specificity pairs. If heterogeneity was detected, we removed any subgroups that caused the heterogeneity from the analysis. If there were no subgroups in the analysis, or those subgroups did not cause the heterogeneity, we looked for data points that were outliers, and reported the meta-analytic results with and without exclusion of these outliers.
Meta-analysis results of diagnostic tests are reported both in table and graphical form. Tables list each study in the meta-analysis, its 2 × 2 data, and any special steps ECRI had to take in abstracting that data. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) are also reported in those tables, along with confidence intervals on each of these ratios, calculated according to Wilson's method.96 Finally, the prevalence of WRUED cases in each study is reported. The last row of the results table provides the sensitivity/specificity at mean threshold results of the meta-analysis, along with the sensitivity and specificity of the points representing the 95% confidence interval on the mean threshold point. Summary values for PPV and NPV are not calculated in the meta-analysis because they are dependent on disease prevalence. Meta-analysis results graphs include the summary ROC itself, the confidence interval and the sensitivity/specificity data points for each included article. The diagonal line in each graph represents the performance of a test that worked no better than chance.
Some investigators based their diagnostic thresholds on results obtained in a control population of individuals without the condition, typically setting a threshold at 2.0 or 2.5 standard deviations from the mean test score of the controls. When the actual number of positive and negative results in the control subjects was reported in the article, we used that data in the meta-analysis. In cases where these numbers were not reported, we assumed a normal distribution of test results in the control subjects, and calculated the theoretical number of false positives and true negatives based on the one-tailed normal distribution. If the threshold was two standard deviations from the mean, one expects false positive results in 2.275% of controls; if the threshold was 2.5 standard deviations, then false positives should make up 0.621% of the control group. The appropriate percentage was multiplied by the number of control subjects and rounded to the closest whole number of patients to get counts for the 2 × 2 table. If the number of controls given the study test was not reported, the article was excluded from analysis even though we knew test specificity from the reported threshold. This is because actual counts of false positives and true negatives are needed to obtain confidence intervals on specificity and the predictive values.
We performed numerous other statistical computations in addition to those involved in performing meta-analyses. We describe these calculations and the logic behind them in our considerations of the appropriate questions. Briefly, these calculations included:
Corrections for attrition; Following all patients for the duration of a study is difficult, particularly when the study is relatively long term. It is possible that in some studies, poor outcomes among patients lost to followup could overturn the results of a study, including those of a well-designed randomized controlled trial. Therefore, wherever possible, we made conservative assumptions about outcomes in patients who were not accounted for in an effort to determine how robust reported results were. This approach is preferable to one that ignores attrition and to one that discards such studies that exceed an arbitrary attrition level. The former approach could lead to incorrect conclusions and the latter can lead to information loss.
Statistical power analyses; Studies that do not contain a sufficient number of patients cannot detect statistically significant differences between groups, even when these differences are clinically meaningful. Therefore, whenever possible, we computed the minimum between-group difference that any given controlled study had the power to detect.
Multiple regression; For certain questions, the results of multiple regressions were of interest, but such analyses were not conducted by the authors. We therefore conducted these analyses when t-patient-level data were available.
Computations of effect sizes for all studies, when possible, even when no meta-analysis was performed. Results of statistical tests (p-values) do not convey information about the magnitude of an effect. To provide an idea about this magnitude, we computed effect sizes for all controlled studies, wherever such computations were possible.
Determinations of whether there were statistically significant differences between the characteristics of patients in any given study. Although studies may report that they were randomized, it is sometimes the case that the randomization protocol was not adequately followed or the study was not truly randomized. These departures from randomization can manifest themselves in pretreatment between-group differences in patient characteristics.
Computation of pretreatment effect sizes. Departures from randomization can also manifest themselves as a statistically significant difference in the outcome between groups prior to the administration of treatment. For example, if the pain levels experienced by patients were significantly different before treatment, one might suspect that the study was not truly randomized.
Verification of 2 × 2 tables reported in studies of diagnostic tests. Because peer-reviewed published articles often contain errors in reported results, we attempted to verify the calculations in each article. If an error was found, we corrected the data and included it in the analysis. If we could not verify the 2 × 2 table, the article was excluded. These exclusions are documented in the text of this report.
To select peer-reviewers for the draft evidence report, ECRI prepared a list of 30 potential reviewers. This list was submitted to AHRQ, which approved all reviewers. Letters inviting these individuals to review were then mailed. Fifteen individuals responded to these letters, 12 individuals agreed to review the draft evidence report, and 9 individuals returned reviews.
Upon receipt of reviews, ECRI revised the draft report accordingly. ECRI also prepared a document describing the disposition of all substantive reviewer comments and supplied this document to AHRQ for review and approval.
Our response to this question is comprised of a subsection on early diagnosis and a subsection on studies of diagnosis of carpal tunnel syndrome, in general. These two subsections follow our evaluation of the internal validity and generalizability of the available relevant literature. Following these two subsections, is a subsection on screening.
The subsection on early diagnosis is the most direct answer to this question, and in it we examine all articles described by their authors as pertaining to early diagnosis of these conditions. However, there are only a few such articles, and we therefore expand our response to diagnosis in general on the grounds that a “good” diagnostic method may also be a “good” method for making an early diagnosis. Ultimately, though, this reasoning is inferential, and conclusive evidence about whether a “good” diagnostic method is also useful for making an early diagnosis can only be derived by studies that directly address this issue.
The evaluation the diagnostic tests we consider is, as with any such test, greatly complicated by the absence of an independent “gold standard” test for any of the upper extremity disorders we address .27 With no independent reference standard whose results are definitive, clinical trials of diagnostic tests for these disorders generally report differences in test results between a group of patients believed to have the condition and a group believed not to have it. Because determinations of who has and does not have the disorder are imperfect (for example, persons who do not have CTS may have symptoms of another condition that mimics CTS), it is impossible for such studies to draw accurate conclusions on how well any test performs.
The definitions of the groups being compared in these studies can also affect results by introducing spectrum effects to the study population. Criteria for selecting patients withWRUEDs may result in inclusion of only clear-cut cases of the condition, thus excluding mild cases that would be harder to diagnose. Selection criteria for patients without WRUEDs may result in inclusion of only those in ideal health, excluding those with early-stage cases of an upper extremity disorder. Together, these spectrum effects amplify the differences that are found in these studies. Thus, their results may not be applicable to the population most likely to get a test in routine practice: persons in high risk groups or with questionable symptoms.
| Test | Definition |
|---|---|
| Closed fist test101 | The patient makes a fist. If the patient feels tingling within one minute, the test is positive. |
| Combined Phalen's and Durkan's test102 | With the patient's elbow extended, the forearm in supination, and the wrist flexed to 60 degrees, the examiner uses one thumb to apply pressure over the carpal tunnel. If the patient feels tingling or numbness within 30 seconds, the test is positive. |
| Decreased muscle strength103 | Maximum force exerted by the patient on a measurement device. |
| Durkan compression test104 | This test is also called the carpal compression test. With the patient's wrist in a neutral position and the forearm supinated, the examiner uses his/her thumbs to compress the wrist at the median nerve. If the patient feels numbness or tingling within 30 seconds, the test is positive. |
| Flick test105 | The patient is asked: “What do you do with your hands when your symptoms are at their worst?” If the patient shakes or flicks the hands, the test is positive. |
| Gilliat tourniquet test106 | The examiner inflates a blood pressure monitor on the patient's arm proximal to the elbow. If the patient feels numbness or tingling within one minute, the test is positive. |
| Grip strength107 | Force measured when patient squeezes a measurement device using the whole hand. |
| Hypesthesia103 | Also called hypoesthesia. It refers to decreased sensitivity to touch. |
| Pain on VAS108 | Pain as measured by a visual analog scale in which the patient rates the subjective degree of pain by placing a mark on a graphical bar. |
| Paresthesia in APB109 | Tingling in the abductor pollicus brevis muscle of the hand. |
| Phalen's test28 | This test is also called the wrist flexion test. The patient places both elbows on a horizontal surface with the forearms vertical, and allows the wrists to flex by gravity. If the patient feels numbness or tingling within one minute, the test is positive. |
| Pinch strength107 | Force measured when patient squeezes a measurement device using the thumb and a finger |
| Symptoms measured systematically29 | Any symptoms of carpal tunnel such as pain, tingling, or numbness, as measured by a questionnaire or a hand diagram. |
| Symptoms during ultrasound110 | Whether the patient experiences carpal tunnel symptoms when the wrist is stimulated with an ultrasound transducer. |
| Reverse Phalen's test111 | This test is also called the wrist extension test. The patient extends both wrists and fingers. If the patient feels numbness or tingling within two minutes, the test is positive. |
| Thenar atrophy103 | The degree of wasting in the thenar muscle of the hand. |
| Thenar weakness31 | The degree of weakness in the thenar muscle of the hand. |
| Tinel's test29 | This test is also called Hoffman-Tinel's test. The examiner taps lightly on the medial aspect of the wrist. If the patient feels tingling, the test is positive. |
Sources: Massy-Westrop112 and ECRI review of clinical trial articles
| Test | Definition |
|---|---|
| Current perception113 | Whether the patient's threshold for perception of electrical current is within normal limits. |
| Moving two-point discrimination107 | The examiner touches two closely-spaced prongs to patient's fingers and moves them distally. The test is positive if the patient cannot discriminate the prongs when they are 4–6 millimeters apart. |
| Object identification114 | The patient blindly feels wooden shapes and is asked to identify them. |
| Pinprick sensation109 | Whether the patient has normal pinprick-induced sensation. |
| Pressure measurement115 | Whether the patient's threshold for perception of pressure is within normal limits. |
| Ridge threshold116 | The patient places an index finger on a circular disc that has a small ridge. If the patient's threshold for detection of the ridge is abnormal, the test is positive. |
| Semmes-Weinstein monofilament30 | This test is also called the von Frey hairs test. The examiner touches the patient with a series of standardized nylon monofilaments, and records the smallest monofilament the patient can detect the presence of. |
| Static two-point discrimination31 | The examiner touches two closely-spaced prongs to patient's fingers and holds them still. The test is positive if the patient cannot discriminate the prongs when they are 5 millimeters apart. |
| Temperature measurement117 | Whether the patient's threshold for perception of temperature, heat pain or cold pain is within normal limits. |
| Tuning fork30 | The examiner hits a metal tuning fork which vibrates, and the patient's threshold for detection of vibration is determined. If the threshold falls outside of normal limits, the test is positive. |
| Vibrometer118 | An instrument vibrates at varying frequencies, and the patient's threshold for detection of vibration is determined. If the threshold falls outside of normal limits, the test is positive |
Sources: Massy-Westrop112 and ECRI review of clinical trial articles
| Test | Definition |
|---|---|
| Nerves tested | |
| Median nerve | The central nerve that is believed to be impaired in carpal tunnel syndrome. It innervates the thumb, index, middle, and ring fingers. |
| Ulnar nerve | The nerve on the medial side of the arm that innervates the ring and little fingers. Some researchers compare median and ulnar nerve conduction tests to diagnose carpal tunnel syndrome. |
| Radial nerve | The nerve on the lateral side of the arm that innervates the thumb. Some researchers compare median and radial nerve conduction tests to diagnose CTS. |
| Motor or sensory | Whether the test assesses motor or sensory nerve function. |
| Orthodromic or antidromic | The relative placement of the stimulating and recording electrodes. If the stimulating electrode is distal to the recording electrode (i.e., the stimulator is further from the torso), the test is orthodromic. Conversely, if the stimulating electrode is proximal to the recording electrode, (i.e., the stimulator is closer to the torso), the test is antidromic. These terms apply to sensory tests but not to motor tests. |
| Electrode placement sites | |
| Abductor pollicus brevis muscle (APB) | A muscle in the hand that is used to record median motor parameters. |
| Abductor digiti minimi (ADM) | A muscle in the hand that is used to record ulnar motor parameters. |
| Parameters Measured | |
| Latency | The time in milliseconds (ms) between stimulation and recording of an electrical impulse. |
| Onset latency | The time in milliseconds (ms) between stimulation and recording of an electrical impulse when measured to the beginning of the action potential. |
| Peak latency | The time in milliseconds (ms) between stimulation and recording of an electrical impulse when measured to the largest amplitude of the action potential. |
| Velocity | Speed of nerve conduction in meters per second (m/s) |
| Amplitude | Size of the action potential in microvolts (uV) |
| Presence/absence | Whether the nerve action potential was recordable. In severe cases, some action potentials may not be recordable. |
| Inching test | A series of nerve conduction tests designed to locate specific areas of nerve slowing. It can be performed orthodromically or antidromically. Electrodes are placed in 9–12 locations which are each a small distance (e.g., 1 cm) apart. By stimulating a fixed site (e.g., the middle finger) and recording at several locations (e.g., 9 evenly-spaced locations along the wrist), researchers can measure the nerve latencies and velocities for each segment along the nerve. |
| Test | Definition |
|---|---|
| Film | Plain film radiograph (x-ray). |
| CT | Computed tomography scan. No articles reported use of obsolete (first- or second-generation CT scanners). |
| MRI | Magnetic resonance imaging scan. No articles reported use of obsolete or prototype MR scanners |
| Ultrasound | Ultrasonic imaging |
Some investigators compare two or more nerve conduction tests in an attempt to assist the diagnosis of CTS (e.g., compute a difference between two latencies). We refer to these comparisons as composite nerve conduction tests. One potential advantage of composite nerve conduction tests is that they can compare two measurements in the same individual, thereby controlling for the effect of age on single nerve conduction tests.97
Imaging tests for carpal tunnel syndrome include radiography (conventional film x-ray), computed tomography (CT) scan, magnetic resonance imaging (MRI), and ultrasound. Using these methods, investigators attempt to measure the size of anatomical features such as the carpal tunnel or the median nerve. Radiologists may also look for qualitative signs of CTS, such as bowing of the flexor retinaculum or a flattened shape of the carpal tunnel.98 CTS may also manifest itself through changes in the appearance of the image, such as changes MR signal intensity of the median nerve. One cannot generalize that CTS will always be represented by an increase in signal intensity, because the relative contrast of different tissues is a function of the specific MR pulse sequence used.99 Within a given study, if the same pulse sequence is used, the effect on appearance of normal and abnormal tissue is expected to be consistent.
Many different measurements are possible from a single image. Some of them may be useful in diagnosis of CTS while others are of no use at all. Furthermore, radiologists may take several of these measurements into account when judging an image as positive or negative for CTS. When assessing imaging tests for CTS, one must be specific as to the particular image parameter or combination of parameters being used, and avoid generalization from effectiveness of one imaging measurement to effectiveness of another. Because they were so numerous, we did not tabulate all imaging measurements reported in clinical trial articles, but instead we tabulated the use of each imaging modality (x-ray, CT, MRI, or ultrasound).
Imaging tests, particularly film radiography, may be used to rule out other causes of hand and wrist symptoms, such as fractures or osteoarthritis 100 and thus may have a role in differential diagnosis of CTS, even if they are not themselves tests for CTS.
As noted above, the vast majority of CTS diagnostic trials compared groups of patients with known or suspected disorders and groups of healthy normal controls. Therefore it is worth summarizing the difficulties with such studies:
Potential spectrum bias because the controls are required to be asymptomatic, and subjects with unrelated upper extremity disorders are excluded. In routine practice, the spectrum of negative cases is likely to include patients with abnormalities that might mimic the condition being tested for, thereby reducing test specificity and positive predictive value.
Potential spectrum bias when severe or obvious cases are selected for in patient inclusion criteria, and patients with mild disorders are excluded. In routine practice, the spectrum of patients with CTS is likely to include mild cases that may not be detected by the diagnostic test, thereby reducing sensitivity and negative predictive value.
The converse of the above spectrum bias, where inclusion criteria are designed to study patients with mild disorders. Studies of patients with only mild disease will underestimate test performance.
Potential age bias arising from selection of young hospital or laboratory workers as controls rather than persons of the same ages as CTS sufferers. Where possible, we recorded mean ages of CTS and control groups in each study, and identified studies in which the mean ages of the groups differed by 5.0 years or more.
Potential sex bias arising from different sex distributions in the patient group and the control group. Where possible, we recorded the sex distributions of CTS and control groups in each study, and identified studies in which the percentage of females differed by 20 percentage points or more.
Articles were included in this analysis if they reported counts of positive and negative test results for at least one test, and they included ten or more patients. Having sufficient data from each included study to complete the 2 × 2 diagnostic truth table is important, because sensitivity and specificity must be measured simultaneously, using the same diagnostic threshold. Otherwise, the threshold could be shifted to favor the reported statistic at the expense of the unreported one.
Not all of the articles we examined are addressed in this evidence report. However, data from the articles we did not address are provided in the evidence tables in the appendix. We included articles in these evidence tables, regardless of their level of reporting, if their authors described them as screening studies or studies on “early diagnosis” of CTS.
| Author | Reason for Exclusion |
|---|---|
| Ikegaya119 | Special patient population (dialysis) |
| Tackmann120 | No diagnostic data |
| Jordan121 | Reported only statistical significance of results |
| Sivri122 | Special patient population (arthritis), only 2 cases of CTS |
| Stolp-Smith123 | Special patient population (pregnant women), only 5 cases of CTS |
| Dlabalová124 | All patients post-surgery for CTS |
| Lazaro125 | All patients post-surgery for CTS |
| Nakamichi126 | All patients post-surgery for CTS |
| Williams127 | Discrepancies in reported results; 2 × 2 table could not be accurately reproduced by ECRI. |
| Mossman128 | Published as letter rather than full paper; 2 × 2 table could not be accurately reproduced by ECRI. |
| Westerman129 | Discrepancies in reported number of patients, unexplained exclusions of patients. |
| Herrick130 | Combined results from CTS patients and patients with other conditions. |
| MacDermid131 | Combined results from CTS patients and patients with other conditions. |
| Gerrning132 | Combined results from CTS patients and patients with other conditions. |
| Byl133 | Combined results from CTS patients and patients with other conditions. |
| Palmer134 | Combined results from CTS patients and patients with other conditions. |
After these exclusions, 189 articles remained for analysis, with a total of 38,087 participants in these studies. The majority of studies (110 or 58%) were conducted outside the United States, and almost all of the studies (184 or 97%) were done at a single center.
In order to be included in meta-analyses of diagnostic trial results, articles had to report sufficient data to permit calculation of sensitivity and specificity for the test in question. In other words, counts of positive and negative test results had to be reported, percentages had to be reported with sufficient data on numbers of patients and controls for us to recalculate the 2 × 2 table, or results for each individual patient had to be reported. Patient-level data were reported in 19 of the 189 articles, and counts for at least some patient groups were reported in 131. Only summary statistics (typically group means) were reported in 39 articles. Even though sensitivity and specificity were not reported in these articles, they were included in the analysis because they met other criteria, such as reporting “early diagnosis” of CTS or an intent to evaluate diagnostic tests in a screening population. In 129 of the articles (68%), it was possible to determine sensitivity and specificity for at least one test from the reported data; in 79 of the articles, the authors themselves reported sensitivity and specificity.
| Study characteristic | Number of studies reporting (percentage) | Specifics (percentage) |
|---|---|---|
| Whether trial was funded by a for-profit institution | 24 (13%) | For-profit funding: 3 (2%) |
| No for-profit funding: 21 (11%) | ||
| Was selection of patients prospective or retrospective? | 75 (40%) | Prospective: 58 (28%) |
| Retrospective: 17 (9%) | ||
| Patient inclusion criteria | 185 (98%) | See Table 46 |
| Patient exclusion criteria | 87 (46%) | See Table 46 |
| Was sex distribution of patients reported? | 131 (69%) | aPercentage female: 61.5% |
| Was the percentage of females in the patient group within 20 percentage points of the control group? | 89 (47%) | Yes: 65 (34%) |
| No, patients were = 20% more female: 21 (11%) | ||
| No, controls were =20% more female: 3 (2%) | ||
| Were patient ages reported? | 123 (65%) | aMean age 48.1 years |
| Was the mean patient age within 5 years of the mean control age? | 89 (47%) | Yes: 52 (28%) |
| No, patients were = 5 years older: 36 (19%) | ||
| No, controls were =5 years older: 1 (1%) | ||
| Was duration of patients' condition reported? | 18 (10%) | a, bMean duration 28.1 months |
| Were patient comorbidities reported? | 46 (24%) | NA |
| Was the test operator blinded? | 13 (7%) | Yes: 13 (7%) |
| Was the test reader blinded? | 23 (12%) | Yes: 23 (12%) |
| Were there multiple test readers? | 7 (4%) | 2 readers: 4 (2%) |
| 3 readers: 2 (1%) | ||
| 4 readers: 1 (1%) | ||
| What was the method for multiple test readers? | 4 (57% of studies reporting multiple readers) | Independent: 2 (1%) |
| Mean: 1 (1%) | ||
| Consensus: 1 (1%) | ||
| Was the test compared to an independent reference standard? | 38 (20%) | Yes: 38 (20%) |
| Were all patients given the test and the reference standard? | 28 (15%) | Yes: 28 (15%) |
Key:
NA—not applicable
Calculated on a per-patient basis (i.e., weighted by number of patients in each study reporting this characteristic)
| Article | Funded by for-profit institution? | Inclusion criteria reported? | Exclusion criteria reported | Method of diagnosis reported | Patient selection | Comorbidity reported | aPercent female | Possible sex bias | aMean age | Possible age bias | aMean duration of condition | Test operator blinded | Test reader blinded | Multiple readers | Method for multiple readers | Independent reference standard | Were patients given both test and reference |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Distal Motor Latency: Unspecified Diagnosis Patient Group | |||||||||||||||||
| Rosén, 1993 138 | NR | Yes | Yes | Yes | NR | NR | 75% | P | 41 | No | NR | NR | NR | NR | NR | No | No |
| Marin, 1983 139 | NR | Yes | NR | NR | NR | NR | 86% | P | 49 | P | 13 | NR | NR | NR | NR | No | No |
| Kimura, 1979 140 | NR | Yes | Yes | Yes | NR | NR | 75% | No | 48 | No | NR | NR | NR | NR | NR | No | No |
| Loong, 1972 141 | NR | Yes | NR | NR | NR | NR | 100% | No | 43.7 | MNR | 12.7 | NR | NR | NR | NR | No | No |
| Plaja, 1971 142 | NR | NR | Yes | NR | Retrospective | NR | NR | GNR | NR | MNR | NR | NR | NR | NR | NR | No | No |
| Distal Motor Latency: Symptoms/Presented Patient Groups | |||||||||||||||||
| Murthy, 1999 143 | NR | Yes | NR | Yes | NR | NR | NR | GNR | NR | ANR | NR | NR | NR | NR | NR | No | No |
| Atroshi, 1996 136 | No | Yes | NR | NR | Prospective | Yes | 69% | No | 52 | P | 24 | NR | NR | NR | NR | No | No |
| Kuntzer, 1994 144 | NR | Yes | Yes | NR | Prospective | NR | 80% | P | 51 | P | NR | NR | NR | NR | NR | No | No |
| Chang, 1991 145 | NR | Yes | Yes | NR | NR | Yes | 79% | GNR | 42.3 | No | NR | NR | NR | NR | NR | No | No |
| Cioni, 1989 146 | NR | Yes | Yes | NR | NR | NR | 16% | C | 46.4 | P | NR | NR | NR | NR | NR | No | No |
| Messina, 1980 120 | NR | Yes | NR | NR | NR | NR | NR | GNR | 45.1 | No | NR | NR | NR | NR | NR | No | No |
| Melvin, 1972 147 | NR | Yes | NR | NR | NR | NR | NR | GNR | NR | ANR | NR | NR | NR | NR | NR | No | No |
| Loong, 1971 148 | NR | Yes | Yes | NR | NR | Yes | 100% | No | NR | ANR | 7.6 | NR | NR | NR | NR | No | No |
| Palmar Sensory Latency: Symptoms/Presented Patient Groups | |||||||||||||||||
| Murthy, 1999 143 | NR | Yes | NR | Yes | NR | NR | NR | GNR | NR | ANR | NR | NR | NR | NR | NR | No | No |
| Girlanda, 1998 149 | NR | Yes | Yes | NR | NR | Yes | 93% | GNR | 39 | ANR | 48 | NR | NR | NR | NR | No | No |
| Chang, 1991 145 | NR | Yes | Yes | NR | NR | Yes | 79% | GNR | 42.3 | No | NR | NR | NR | NR | NR | No | No |
| Jackson, 1989 150 | No | Yes | Yes | NR | NR | Yes | 82% | No | 52.6 | P | NR | NR | NR | NR | NR | No | No |
| Escobar, 1985 151 | NR | Yes | Yes | NR | NR | Yes | 70% | No | NR | ANR | NR | NR | NR | NR | NR | No | No |
| Phalen's Maneuver: All Patient Groups | |||||||||||||||||
| Szabo, 1999 152 | No | Yes | NR | Yes | Prospective | NR | 76% | No | NR | ANR | NR | NR | Yes | NR | NR | No | No |
| Fertl, 1998 153 | NR | Yes | Yes | Yes | Prospective | NR | 83% | P | 55.5 | P | NR | Yes | Yes | NR | NR | No | No |
| Gerr, 1998 31 | NR | Yes | Yes | Yes | NR | NR | 72% | No | 46.6 | P | NR | NR | NR | NR | NR | Yes | No |
| Ghavanini, 1998 154 | NR | Yes | Yes | Yes | Prospective | NR | 81% | No | 40 | No | 15 | NR | NR | NR | NR | No | No |
| Tetro, 1998 102 | No | Yes | Yes | Yes | Prospective | NR | 64% | No | 49.3 | No | NR | NR | NR | NR | NR | No | No |
| González del Pino, 1997 104 | NR | Yes | NR | Yes | Prospective | NR | 81% | No | 50 | No | 37.9 | NR | NR | 3 | NR | Yes | Yes |
| De Smet, 1995 101 | NR | Yes | NR | Yes | NR | NR | 88% | C | 49.2 | C | NR | NR | NR | NR | NR | No | No |
| Werner, 1994 111 | NR | Yes | NR | Yes | NR | NR | NR | GNR | NR | ANR | NR | NR | NR | NR | NR | No | No |
| Durkan, 1991 155 | No | Yes | NR | Yes | NR | NR | NR | GNR | 45 | ANR | NR | NR | NR | NR | NR | Yes | Yes |
| Gellman, 1986 106 | No | Yes | NR | Yes | NR | Yes | 74% | GNR | NR | ANR | NR | NR | NR | NR | NR | Yes | Yes |
| Tinel's Sign: All Patient Groups | |||||||||||||||||
| Szabo, 1999 152 | No | Yes | NR | Yes | Prospective | NR | 76% | No | NR | ANR | NR | NR | Yes | NR | NR | No | No |
| Gerr, 1998 31 | NR | Yes | Yes | Yes | NR | NR | 72% | No | 46.6 | P | NR | NR | NR | NR | NR | Yes | No |
| Ghavanini, 1998 154 | NR | Yes | Yes | Yes | Prospective | NR | 81% | No | 40 | No | 15 | NR | NR | NR | NR | No | No |
| Tetro, 1998 102 | No | Yes | Yes | Yes | Prospective | NR | 64% | No | 49.3 | No | NR | NR | NR | NR | NR | No | No |
| González del Pino, 1997 104 | NR | Yes | NR | Yes | Prospective | NR | 81% | No | 50 | No | 37.9 | NR | NR | 3 | NR | Yes | Yes |
| De Smet, 1995 101 | NR | Yes | NR | Yes | NR | NR | 88% | C | 49.2 | C | NR | NR | NR | NR | NR | No | No |
| Durkan, 1991 155 | No | Yes | NR | Yes | NR | NR | 74% | GNR | 45 | ANR | NR | NR | NR | NR | NR | Yes | Yes |
| Seror, 1987 156 | NR | Yes | Yes | Yes | NR | NR | 79% | No | 56.8 | No | NR | NR | NR | NR | NR | No | No |
| Gellman, 1986 106 | No | Yes | NR | Yes | NR | Yes | NR | GNR | NR | ANR | NR | NR | NR | NR | NR | Yes | Yes |
| Gelmers, 1979 29 | NR | Yes | Yes | Yes | NR | NR | 81% | No | 57 | No | NR | NR | NR | NR | NR | Yes | No |
| Stewart, 1978 157 | NR | Yes | Yes | Yes | NR | Yes | 81% | No | 55 | No | NR | NR | NR | NR | NR | Yes | No |
Key:
Percent female, mean age, and mean duration of condition for CTS patients
Possible sex bias: No—proportion women in epicondylitis group within 20% of proportion of women in control group; P—Patients were more likely to be female; C—Controls were more likely to be female; GNR—Genders not reported for both groups; NC—Study did not contain a separate control group
Possible age bias: No—mean age of epicondylitis group within 5 years of mean age of control group; P—Patients were older than controls; C—Controls were older than patients; ANR—Ages not reported for both groups; NC—Study did not contain a separate control group
Method for multiple test readers: Indep—Independent
The design of most studies raised the possibility of age bias in which patients were markedly older than controls. Some nerve conduction measurements become slower as people age,97 thus if patients are older than controls, the study will overestimate the effectiveness of some nerve conduction tests. For this analysis, we defined age bias as a difference of five years or more between the mean age of patients and the mean age of controls. If a study reported ages of more than one group of carpal tunnel patients or more than one group of controls, we used the ages that implied the least amount of age bias in the study. This conservative approach tends to underestimate the amount of age bias in the studies.
Of 189 carpal tunnel studies we examined, 35 did not include a separate control group and 65 failed to report mean or median ages for one or both groups. That left 89 studies for which we could determine whether there was an age bias. Of these 89 studies, 52 had no age bias according to our definition. In 36 studies, patients were five years or more older than controls, while in one study135, controls were five years or more older than patients. In only 12 articles were all patient groups within one year of the controls in mean age. This suggests that there is little use of age-matching to ensure that age bias does not affect results, even though it is known that results of some diagnostic tests are affected by age.
Figure 7
plots each study using
the mean age of controls on the horizontal axis and the mean age of
patients on the vertical axis. The solid diagonal line represents the
points at which patients and controls had the same age. The dashed
diagonal lines represent the points at which patients and controls were
five years apart. The plot shows that patients tended to be older than
controls. Whereas patients were older than controls in 76 studies, the
reverse was true in only 11 studies (in the remaining two studies, the
group means were the same).
A similar analysis was done for possible sex bias. We arbitrarily defined potential sex bias as a difference of 20 or more percentage points in the proportions of females in the patient group and in the control group. As with the age bias analysis, when a study had more than one carpal tunnel group or more than one control group, we used a conservative approach by selecting groups that minimized potential sex bias. This approach will underestimate the amount of potential sex bias.
Of 189 carpal tunnel diagnostic studies recorded in the database, 35 did not contain a separate control group, and 65 did not report the sex distribution for one or both of the CTS and control groups. There were 89 studies for which we could determine whether there was a sex bias. Note that these were not the same 89 studies for which we could determine age bias; 21 studies reported age but not sex, and 21 studies reported sex but not age.
Of these 89 studies, 65 did not meet our definition of possible sex bias. In 21 studies, the percentage of females in the CTS group was 20 or more percentage points higher than the control group. In 3 studies, the percentage of females in the CTS group was 20 or more percentage points lower than in the control group.
Figure 8
plots the sex
distribution of each study, using the percentage of females in the
control group on the horizontal axis and the percentage of females in
patient group on the vertical axis. The plot shows that the percentage
of females tended to be higher in patient groups than in control groups.
The percentage of females in the patient group was greater than the
percentage of females in the control group in 63 of the 89 studies. The
reverse was true in only 13 studies. There were 13 studies in which the
percentages were equal.
We defined studies as sex-matched if the proportion of women in each patient groups differed two percentage points or less from the proportion of women in the control group. Using this definition, 20 of the 89 studies (22%) could be called sex-matched. To the extent that sex affects the diagnostic tests for CTS, there is a potential for sex bias in the results. Despite this possible bias, few studies controlled for differing proportions of men and women in their CTS and control groups. These differences, and age differences in patient and control group, are components of the evaluation of diagnostic clinical trial results.
Use of multiple readers was not widely reported, and where there were multiple readers reported, only 4 of 7 articles reported how they arrived at conclusions. This could affect the internal validity of the conclusions in studies where multiple readers interpreted each test and then met with each other to resolve their differences in interpretation. This practice can reduce interobserver variability and thus may overestimate the true performance of tests which normally are interpreted by just one person.
| Study characteristic | Number of studies reporting (percentage) | Specifics (percentage) |
|---|---|---|
| Years in which study was conducted | 39 (21%) | NA |
| Number of centers | 189 (100%) | Single: 184 (97%) |
| Multiple (<5): 4 (2%) | ||
| Multiple (>5): 1 (1%) | ||
| Country in which study was conducted | 189 (100%) | USA: 79 (42%) |
| Other: 110 (58%) | ||
| Patient inclusion criteria | 185 (98%) | See Table 46 |
| Patient exclusion criteria | 87 (46%) | See Table 46 |
| Were patient comorbidities reported? | 46 (24%) | NA |
| Was sex distribution of patients reported? | 131 (69%) | aPercentage female: 61.5% |
| Were patient ages reported? | 123 (65%) | aMean age 48.1 years |
| Was duration of patients' condition reported? | 18 (10%) | a, bMean duration 28.1 months |
| Did all patients have previous conservative treatment? | 1 (1%) | Yes: 1 (1%) |
| Did any patients have previous surgical treatment? | 6 (3%) | Yes: 6 (3%) |
| Adequate reporting of study's source of patients | 29 (15%) | NA |
| Was there a potential selection bias for easy cases? | 58 (31%) | Yes: 58 (31%) |
| Was there a potential selection bias for hard cases? | 40 (21%) | Yes: 40 (21%) |
Key:
NA—not applicable
Calculated on a per-patient basis (i.e., weighted by number of patients in each study reporting this characteristic)
| Article | Years in which trial was conducted | Number of centers | Country where trial was conducted | Are patient comorbidity reported? | Percent female | Mean age | Mean duration of condition | Did all patients have previous conservative treatment? | Did any patients have previous surgical treatment? | Source of patients adequately described and generalizable to broader clinical practice? | Potential selection bias for easy cases? | Potential selection bias for difficult cases? |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Distal Motor Latency: Unspecified Diagnosis Patient Group | ||||||||||||
| Rosén, 1993 138 | 1986-1987 | Single | Sweden | No | 75% | 41 | NR | No | No | Yes | No | No |
| Marin, 1983 139 | NR | Single | USA | No | 86% | 49 | 13 | No | No | No | Yes | No |
| Kimura, 1979 140 | 1978 | Single | USA | No | 75% | 48 | NR | No | No | No | No | Yes |
| Loong, 1972 141 | NR | Single | Singapore | No | 100% | 43.7 | 12.7 | No | No | No | No | No |
| Plaja, 1971 142 | NR | Single | Spain | No | NR | NR | NR | No | No | No | Yes | No |
| Distal Motor Latency: Symptoms/Presented Patient Groups | ||||||||||||
| Murthy, 1999 143 | NR | Single | India | No | NR | NR | NR | No | No | No | Yes | No |
| Atroshi, 1996 136 | NR | Single | Sweden | Yes | 69% | 52 | 24 | Yes | No | No | Yes | No |
| Kuntzer, 1994 144 | NR | Single | Switzerland | No | 80% | 51 | NR | No | No | Yes | No | No |
| Chang, 1991 145 | NR | Single | Taiwan | Yes | 79% | 42.3 | NR | No | No | No | No | No |
| Cioni, 1989 146 | NR | Single | Italy | No | 16% | 46.4 | NR | No | No | No | No | No |
| Messina, 1980 120 | NR | Single | Italy | No | NR | 45.1 | NR | No | No | No | No | No |
| Melvin, 1972 147 | NR | Single | USA | No | NR | NR | NR | No | No | No | No | No |
| Loong, 1971 148 | NR | Single | Singapore | Yes | 100% | NR | 7.6 | No | No | No | No | No |
| Palmar Sensory Latency: Symptoms/Presented Patient Groups | ||||||||||||
| Murthy, 1999 143 | NR | Single | India | No | NR | NR | NR | No | No | No | Yes | No |
| Girlanda, 1998 149 | NR | Single | Italy | Yes | 93% | 39 | 48 | No | No | No | No | Yes |
| Chang, 1991 145 | NR | Single | Taiwan | Yes | 79% | 42.3 | NR | No | No | No | No | No |
| Jackson, 1989 150 | NR | Single | Canada | Yes | 82% | 52.6 | NR | No | No | No | No | No |
| Escobar, 1985 151 | NR | Single | USA | Yes | 70% | NR | NR | No | No | No | No | No |
| Phalen's Maneuver: All Patient Groups | ||||||||||||
| Szabo, 1999 152 | 1993-1996 | Single | USA | No | 76% | NR | NR | No | No | No | No | No |
| Fertl, 1998 153 | 1997 | Single | Austria | No | 83% | 55.5 | NR | No | No | Yes | No | No |
| Gerr, 1998 31 | NR | Single | USA | No | 72% | 46.6 | NR | No | No | No | No | No |
| Ghavanini, 1998 154 | NR | Single | Iran | No | 81% | 40 | 15 | No | No | No | No | No |
| Tetro, 1998 102 | 1995-1997 | Single | USA | No | 64% | 49.3 | NR | No | No | Yes | No | No |
| González del Pino, 1997 104 | 1992-1995 | Single | Spain | No | 81% | 50 | 37.9 | No | No | No | Yes | No |
| De Smet, 1995 101 | NR | Single | Belgium | No | 88% | 49.2 | NR | No | No | No | No | No |
| Werner, 1994 111 | NR | Single | USA | No | NR | NR | NR | No | No | No | No | No |
| Durkan, 1991 155 | 1987-1990 | Single | USA | No | NR | 45 | NR | No | No | No | No | No |
| Gellman, 1986 106 | 1982-1984 | Single | USA | Yes | 74% | NR | NR | No | No | No | Yes | No |
| Tinel's Sign: All Patient Groups | ||||||||||||
| Szabo, 1999 152 | 1993-1996 | Single | USA | No | 76% | NR | NR | No | No | No | No | No |
| Gerr, 1998 31 | NR | Single | USA | No | 72% | 46.6 | NR | No | No | No | No | No |
| Ghavanini, 1998 154 | NR | Single | Iran | No | 81% | 40 | 15 | No | No | No | No | No |
| Tetro, 1998 102 | 1995-1997 | Single | USA | No | 64% | 49.3 | NR | No | No | Yes | No | No |
| González del Pino, 1997 104 | 1992-1995 | Single | Spain | No | 81% | 50 | 37.9 | No | No | No | Yes | No |
| De Smet, 1995 101 | NR | Single | Belgium | No | 88% | 49.2 | NR | No | No | No | No | No |
| Durkan, 1991 155 | 1987-1990 | Single | USA | No | 74% | 45 | NR | No | No | No | No | No |
| Seror, 1987 156 | NR | Single | France | No | 79% | 56.8 | NR | No | No | No | No | No |
| Gellman, 1986 106 | 1982-1984 | Single | USA | Yes | NR | NR | NR | No | No | No | Yes | No |
| Gelmers, 1979 29 | NR | Single | Netherlands | No | 81% | 57 | NR | No | No | No | Yes | No |
| Stewart, 1978 157 | NR | Single | Canada | Yes | 81% | 55 | NR | No | No | No | Yes | No |
Key:
NR—not reported
In this literature, reporting of patient comorbidities was particularly bad. Only 46 of the articles (24%) reported any comorbidities at all. Duration of patients' conditions was reported in only 18 studies (10%) even though this variable is an indicator of condition severity.
Ninety-eight CTS diagnostic articles (52%) reported patient selection criteria that had the potential to bias studies towards including more easy cases (e.g. including only cases of severe CTS) or more difficult cases to diagnose (e.g. including only cases where other diagnostic tests were equivocal). These criteria represent potential for bias but not conclusive proof of bias, thus we did not exclude such studies. Instead, we used potential selection bias in our analyses of homogeneity, by separately analyzing the homogeneity of studies with and without these potential biases. Generalizability of study results is also affected by the possible spectrum bias arising from study designs where patients with known CTS are compared to healthy volunteers, and the absence of a “gold standard” test for diagnosis of CTS.
Incomplete reporting of important study design and patient characteristics prevents one from ruling out selection biases and other confounding factors as the cause of clinical trial results. The quality of this evidence base is not sufficient to permit us to draw reliable conclusions from a single study. Meta-analysis and heterogeneity analysis can be used to try and identify the effects of these study variables on study results.
Because there is no broad agreement among clinicians of what constitutes and “early” diagnosis of CTS, we accepted any studies so described by their authors as studies of early identification of the condition.
| Article | Patient selection criteria relevant to early detection | Symptoms and normal NCS? | Authors's proposed method for early detection | Sensory NCS? |
|---|---|---|---|---|
| Seror, 2000 158 | Symptoms, but normal needle examination, normal DML (<4 ms) and normal palm-to-wrist orthodromic SCV (>45 m/s). |
 | Orthodromic sensory inching test from the middle finger. |
|
| Girlanda, 1998 149 | Symptoms, but no weakness, no muscle atrophy, and normal DML (<4 ms). |
| Combination of nerve conduction tests:a) Difference between median and ulnar orthodromic SCV from ring finger to wrist, and b) Ratio of orthodromic SCV from middle finger to palm and orthodromic SCV from palm to wrist |
|
| Seror, 1998 159 | Symptoms, but normal DML (<4 ms) and normal palm-to-wrist orthodromic SCV (>45 m/s). |
| Orthodromic sensory inching test from the middle finger. |
|
| Terzis, 1998 162 | Symptoms, but normal DML (<4.2 ms) |
| Combination of orthodromic sensory nerve conduction tests from the ring finger. |
|
| Bronson, 1997 163 | Symptoms, but normal DML (<4 ms) and normal needle examination. |
| Comparison of DMLs using five different wrist positions. | □ |
| Murata, 1996 164 | Workers at risk | □ | Ratio of:a) Antidromic SCV from wrist to index finger, and b) Antidromic SCV from palm to index finger |
|
| Padua, 1996 165 | Symptoms, but no signs of severe CTS (e.g., absent SNAP at the wrist). |
| Ratio of:a) Orthodromic SCV from middle finger to palm, and b) Orthodromic SCV from palm to wrist |
|
| Young, 1995 166 | Workers at risk | □ | Total score on a grading scale that included seven clinical signs, four symptoms, and DML ≥4.45 ms. | □ |
| Johnson, 1993 167 | Workers at risk | □ | Track changes in DML over time | □ |
| Uncini, 1993 160 | Symptoms, but normal DML (<4.2 ms) and normal SCV from index finger to wrist (>45 m/s) |
| Difference between: a)Median orthodromic latency between ring finger and wrist, and b)Ulnar orthodromic latency between ring finger and wrist |
|
| Jetzer, 1991 168 | Workers at risk | □ | Vibrometry | □ |
| Luchetti, 1991 169 | Symptoms, but normal motor function, sensory function, quantitative sensory examination, cutaneous trophism, DSL (NR), and DML (NR). |
| Antidromic inching test to the middle finger |
|
| Charles, 1990 170 | Clinical diagnosis of CTS by referring physician, and at least one of the following: a)DML ≥4.5 ms; b) Orthodromic SCV from index finger <45 m/s; c) Difference ≥0.5 ms between median and ulnar sensory antidromic latencies to the ring finger | □ | Difference between: a)Median antidromic latency between ring finger and wrist, and b) Ulnar antidromic latency between ring finger and wrist |
|
| Palliyath, 1990 171 | Symptoms, but “very little electrophysiological changes on routine tests for CTS” (p 307). |
| Duration of relative refractory period and absolute refractory period. | □ |
| Cioni, 1989 146 | Symptoms | □ | Orthodromic SCV from ring finger to wrist |
|
| Jackson, 1989 150 | Symptoms. Patients were stratified into three groups, and one group represented mild CTS as defined by normal NCS (based on four tests) and normal needle examination. |
| Combination of two nerve conduction tests: a)Difference between median and radial antidromic sensory latencies from wrist to thumb, and b) Difference between median and ulnar antidromic sensory latencies from wrist to ring finger |
|
| Uncini, 1989 161 | Symptoms, but normal DML (≤4.2 ms) and SNAPs were present with normal amplitude. |
| Difference between:a) Median orthodromic latency between ring finger and wrist, and b) Ulnar orthodromic latency between ring finger and wrist |
|
| Wongsam, 1983 172 | Symptoms suggesting early CTS. | □ | Ratio of:a) Antidromic latency from wrist to middle fingerb) Antidromic latency from palm to middle finger |
|
Key:
DML—Distal motor latency
DSL—Distal sensory latency
ms—Milliseconds
m/s—Meters per second
SCV—Sensory conduction velocity
SNAP—Sensory nerve action potential
NR—Not reported
Thirteen of the 18 studies (72%) proposed sensory nerve conduction test(s) for the early diagnosis of carpal tunnel syndrome. As with the selection criteria, however, there was little agreement regarding test specifics. Two studies by Seror158, 159 each proposed the orthodromic sensory inching test for the early detection of CTS. Two studies by Uncini160, 161 each proposed the difference between median and ulnar orthodromic sensory latencies from the ring finger for the early detection of CTS. None of the other nine studies of sensory nerve conduction proposed the same specific tests or combination of tests. Therefore, studies of the early detection of CTS utilize the same general categories of nerve conduction tests, but there is wide variability in the specific tests employed. Furthermore, there are insufficient studies of any specific test to permit meta-analysis for drawing conclusions on whether it is effective for early identification of CTS. For this reason, we proceed to examine diagnostic tests for carpal tunnel syndrome, in general.
Our evaluation of methods for diagnosing CTS is primarily meta-analytic. To identify diagnostic tests of CTS for which meta-analyses were appropriate, we performed several tabulations. These tabulations were restricted to studies that met each of the following three criteria: 1) Study included a carpal tunnel syndrome group; 2) Study included a normal group; 3) Study was not a screening study. There were 138 studies that met all of these criteria.
| Legend: | ||||
|---|---|---|---|---|
| First entry in cell—Total number of articles | ||||
| Second entry in cell—Number of articles with derivable sensitivity and specificity | ||||
| Sign/symptom | Complex objective standard | Simple nerve conduction | Symptoms/ presented | Unspecified diagnosis |
| Closed fist test | 0, 0 | 1, 1 | 1, 1 | 0, 0 |
| Combined Phalen's/Durkan test | 1, 1 | 0, 0 | 0, 0 | 0, 0 |
| Decreased muscle strength | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| Durkan compression | 5, 5 | 1, 1 | 3, 3 | 1, 1 |
| Flick sign | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| Gilliat tourniquet | 1, 1 | 1, 1 | 1, 1 | 1, 1 |
| Grip strength | 0, 0 | 0, 0 | 0, 0 | 1, 0 |
| Hypesthesia | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| Pain on VAS | 0, 0 | 0, 0 | 1, 1 | 1, 1 |
| Paresthesia in APB | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| Phalen's/reverse Phalen's | 7, 7 | 2, 1 | 6, 6 | 3, 3 |
| Pinch strength | 0, 0 | 0, 0 | 0, 0 | 1, 0 |
| Symptoms measured systematically | 3, 3 | 0, 0 | 2, 2 | 1, 0 |
| Symptoms during ultrasound | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| Thenar atrophy | 0, 0 | 0, 0 | 2, 2 | 0, 0 |
| Thenar weakness | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| Tinel's | 9, 9 | 2, 1 | 3, 3 | 2, 2 |
| Legend: | ||||
|---|---|---|---|---|
| First entry in cell—Total number of articles | ||||
| Second entry in cell— Number of articles with derivable sensitivity and specificity | ||||
| Sensory test | Complex objective standard | Simple nerve conduction | Symptoms/ presented | Unspecified diagnosis |
| Object identification | 0, 0 | 0, 0 | 0, 0 | 1, 0 |
| Pinprick sensation | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| Pressure measurement | 0, 0 | 0, 0 | 1, 1 | 1, 0 |
| Ridge threshold | 0, 0 | 0, 0 | 0, 0 | 1, 0 |
| Semmes-Weinstein filament | 1, 1 | 0, 0 | 0, 0 | 4, 1 |
| Temperature measurement | 0, 0 | 0, 0 | 1, 1 | 2, 1 |
| Texture discrimination | 0, 0 | 0, 0 | 0, 0 | 1, 0 |
| Tuning fork | 1, 1 | 0, 0 | 1, 1 | 0, 0 |
| Two-point discrimination (moving or static) | 2, 2 | 0, 0 | 2, 2 | 1, 0 |
| Vibrometer | 2, 2 | 0, 0 | 5, 5 | 1, 0 |
| Legend: Nerve tested: MED-median, RAD-radial, ULN-ulnar, MOT-motor, SEN-sensory | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Configuration (not applicable to motor nerve tests): OR-orthodromic, AN-antidromic | |||||||||
| Stimulation electrode placement: ELB-elbow, FOR-forearm, WR-wrist, PAL-palm, TH-thumb, IN-index finger, MI-middle finger, RI-ring finger, LI-little finger, APB-abductor policis brevis, ADM-abductor digiti minimi, OTH-other | |||||||||
| Recording electrode placement (see D for abbreviations) | |||||||||
| Measured parameter: LAT-latency, PRE-presence/absence of signal, AMP-amplitude, VEL-velocity, INCH-inching, OTH-other | |||||||||
| Blank cells—Not reported or not applicable | |||||||||
| First entry in cell—Total number of articles | |||||||||
| Second entry in cell—Number of articles with derivable sensitivity and specificity | |||||||||
| Shaded cells—Ten or more articles reporting sensitivity and specificity. | |||||||||
| Nerve Conduction Test | Patient selection type | ||||||||
| Nerve Tested | Nerve Tested | Configuration | Stimulation | Recording | Parameter | Complex objective standard | Simple nerve conduction | Symptoms/ presented | Unspecified diagnosis |
| MOT | LAT | 1, 1 | 0, 0 | 0, 0 | 1, 1 | ||||
| MOT | WR | OTH | LAT | 0, 0 | 0, 0 | 0, 0 | 1, 1 | ||
| SEN | LAT | 1, 1 | 0, 0 | 0, 0 | 1, 1 | ||||
| SEN | OR | TH | WR | PRE | 0, 0 | 0, 0 | 1, 1 | 0, 0 | |
| MED | OTH | 0, 0 | 0, 0 | 2, 2 | 0, 0 | ||||
| MED | MOT | 0, 0 | 0, 0 | 0, 0 | 1, 0 | ||||
| MED | MOT | AMP | 0, 0 | 1, 0 | 0, 0 | 0, 0 | |||
| MED | MOT | LAT | 2, 1 | 1, 0 | 2, 2 | 2, 1 | |||
| MED | MOT | OTH | 1, 1 | 1, 0 | 2, 1 | 0, 0 | |||
| MED | MOT | VEL | 0, 0 | 1, 0 | 1, 1 | 1, 0 | |||
| MED | MOT | APB | AMP | 1, 1 | 0, 0 | 0, 0 | 0, 0 | ||
| MED | MOT | APB | LAT | 1, 1 | 0, 0 | 0, 0 | 0, 0 | ||
| MED | MOT | ELB | APB | AMP | 1, 0 | 0, 0 | 1, 1 | 1, 1 | |
| MED | MOT | ELB | APB | LAT | 1, 1 | 0, 0 | 0, 0 | 1, 1 | |
| MED | MOT | ELB | APB | OTH | 1, 1 | 0, 0 | 1, 1 | 0, 0 | |
| MED | MOT | ELB | APB | VEL | 1, 0 | 0, 0 | 1, 1 | 2, 2 | |
| MED | MOT | ELB | IN | AMP | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | MOT | ELB | IN | LAT | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | MOT | ELB | IN | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | MOT | ELB | WR | AMP | 1, 0 | 0, 0 | 0, 0 | 0, 0 | |
| MED | MOT | ELB | WR | LAT | 0, 0 | 0, 0 | 0, 0 | 0, 0 | |
| MED | MOT | ELB | WR | VEL | 2, 1 | 0, 0 | 3, 3 | 1, 1 | |
| MED | MOT | FOR | VEL | 1, 1 | 0, 0 | 0, 0 | 1, 1 | ||
| MED | MOT | FOR | APB | AMP | 1, 1 | 0, 0 | 0, 0 | 0, 0 | |
| MED | MOT | FOR | APB | LAT | 1, 1 | 0, 0 | 0, 0 | 0, 0 | |
| MED | MOT | FOR | APB | VEL | 0, 0 | 0, 0 | 1, 1 | 0, 0 | |
| MED | MOT | FOR | PAL | AMP | 1, 1 | 0, 0 | 0, 0 | 0, 0 | |
| MED | MOT | FOR | PAL | LAT | 1, 1 | 0, 0 | 0, 0 | 0, 0 | |
| MED | MOT | FOR | WR | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | MOT | PAL | APB | AMP | 1, 1 | 0, 0 | 1, 0 | 2, 1 | |
| MED | MOT | PAL | APB | LAT | 0, 0 | 0, 0 | 0, 0 | 2, 1 | |
| MED | MOT | PAL | IN | AMP | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | MOT | PAL | IN | LAT | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | MOT | PAL | IN | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | MOT | WR | AMP | 0, 0 | 0, 0 | 1, 1 | 0, 0 | ||
| MED | MOT | WR | LAT | 2, 2 | 1, 0 | 1, 1 | 0, 0 | ||
| MED | MOT | WR | PRE | 1, 1 | 0, 0 | 0, 0 | 0, 0 | ||
| MED | MOT | WR | VEL | 1, 1 | 0, 0 | 0, 0 | 0, 0 | ||
| MED | MOT | WR | APB | AMP | 2, 1 | 0, 0 | 9, 7 | 9, 6 | |
| MED | MOT | WR | APB | LAT | 4, 4 | 3, 2 | 21, 17 | 24, 21 | |
| MED | MOT | WR | APB | OTH | 2, 1 | 1, 0 | 1, 1 | 2, 2 | |
| MED | MOT | WR | APB | PRE | 0, 0 | 0, 0 | 3, 3 | 1, 1 | |
| MED | MOT | WR | APB | VEL | 0, 0 | 0, 0 | 2, 1 | 5, 5 | |
| MED | MOT | WR | IN | AMP | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | MOT | WR | IN | LAT | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | MOT | WR | IN | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | MOT | WR | OTH | AMP | 1, 0 | 0, 0 | 1, 1 | 1, 1 | |
| MED | MOT | WR | OTH | LAT | 1, 1 | 1, 1 | 8, 8 | 3, 3 | |
| MED | MOT | WR | OTH | OTH | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | MOT | WR | OTH | VEL | 1, 0 | 0, 0 | 1, 1 | 0, 0 | |
| MED | MOT | WR | PAL | AMP | 0, 0 | 0, 0 | 1, 1 | 0, 0 | |
| MED | MOT | WR | PAL | LAT | 0, 0 | 0, 0 | 1, 1 | 0, 0 | |
| MED | MOT | WR | PAL | OTH | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | MOT | WR | PAL | VEL | 0, 0 | 0, 0 | 1, 0 | 0, 0 | |
| MED | MOT | WR | TH | LAT | 0, 0 | 0, 0 | 2, 0 | 0, 0 | |
| MED | MOT | WR | TH | VEL | 0, 0 | 0, 0 | 1, 1 | 0, 0 | |
| MED | SEN | 0, 0 | 0, 0 | 0, 0 | 1, 0 | ||||
| MED | SEN | LAT | 3, 2 | 0, 0 | 0, 0 | 1, 0 | |||
| MED | SEN | OTH | 0, 0 | 0, 0 | 1, 0 | 1, 0 | |||
| MED | SEN | VEL | 0, 0 | 1, 0 | 0, 0 | 0, 0 | |||
| MED | SEN | WR | AMP | 1, 1 | 0, 0 | 0, 0 | 0, 0 | ||
| MED | SEN | WR | LAT | 1, 1 | 0, 0 | 0, 0 | 0, 0 | ||
| MED | SEN | AN | AMP | 0, 0 | 0, 0 | 0, 0 | 1, 1 | ||
| MED | SEN | AN | LAT | 1, 1 | 0, 0 | 1, 1 | 1, 1 | ||
| MED | SEN | AN | VEL | 1, 1 | 0, 0 | 1, 1 | 0, 0 | ||
| MED | SEN | AN | ELB | IN | AMP | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| MED | SEN | AN | ELB | IN | OTH | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| MED | SEN | AN | ELB | MI | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| MED | SEN | AN | ELB | PAL | INCH | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| MED | SEN | AN | ELB | WR | VEL | 0, 0 | 0, 0 | 2, 2 | 1, 1 |
| MED | SEN | AN | FOR | IN | LAT | 0, 0 | 0, 0 | 1, 0 | 0, 0 |
| MED | SEN | AN | FOR | RI | LAT | 0, 0 | 0, 0 | 1, 0 | 0, 0 |
| MED | SEN | AN | FOR | TH | LAT | 0, 0 | 0, 0 | 1, 0 | 0, 0 |
| MED | SEN | AN | PAL | IN | AMP | 1, 1 | 0, 0 | 2, 1 | 0, 0 |
| MED | SEN | AN | PAL | IN | LAT | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| MED | SEN | AN | PAL | IN | PRE | 1, 1 | 0, 0 | 0, 0 | 0, 0 |
| MED | SEN | AN | PAL | IN | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| MED | SEN | AN | PAL | MI | 0, 0 | 0, 0 | 1, 0 | 0, 0 | |
| MED | SEN | AN | PAL | MI | AMP | 0, 0 | 0, 0 | 2, 1 | 0, 0 |
| MED | SEN | AN | PAL | MI | LAT | 0, 0 | 0, 0 | 2, 1 | 0, 0 |
| MED | SEN | AN | PAL | MI | OTH | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| MED | SEN | AN | PAL | MI | VEL | 0, 0 | 0, 0 | 1, 1 | 2, 2 |
| MED | SEN | AN | WR | LAT | 0, 0 | 0, 0 | 1, 1 | 0, 0 | |
| MED | SEN | AN | WR | IN | AMP | 3, 2 | 0, 0 | 6, 5 | 5, 4 |
| MED | SEN | AN | WR | IN | LAT | 1, 1 | 0, 0 | 11, 9 | 5, 3 |
| MED | SEN | AN | WR | IN | OTH | 2, 1 | 0, 0 | 2, 2 | 0, 0 |
| MED | SEN | AN | WR | IN | PRE | 1, 1 | 0, 0 | 2, 2 | 2, 2 |
| MED | SEN | AN | WR | IN | VEL | 0, 0 | 0, 0 | 3, 2 | 0, 0 |
| MED | SEN | AN | WR | MI | AMP | 0, 0 | 0, 0 | 4, 3 | 0, 0 |
| MED | SEN | AN | WR | MI | INCH | 1, 1 | 0, 0 | 1, 1 | 0, 0 |
| MED | SEN | AN | WR | MI | LAT | 0, 0 | 0, 0 | 2, 1 | 0, 0 |
| MED | SEN | AN | WR | MI | PRE | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| MED | SEN | AN | WR | MI | VEL | 0, 0 | 0, 0 | 3, 3 | 1, 1 |
| MED | SEN | AN | WR | OTH | VEL | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| MED | SEN | AN | WR | PAL | AMP | 0, 0 | 0, 0 | 1, 0 | 0, 0 |
| MED | SEN | AN | WR | PAL | LAT | 0, 0 | 1, 0 | 1, 1 | 0, 0 |
| MED | SEN | AN | WR | PAL | VEL | 0, 0 | 0, 0 | 3, 2 | 2, 2 |
| MED | SEN | AN | WR | RI | AMP | 0, 0 | 0, 0 | 1, 0 | 0, 0 |
| MED | SEN | AN | WR | RI | LAT | 0, 0 | 0, 0 | 3, 2 | 3, 2 |
| MED | SEN | AN | WR | RI | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| MED | SEN | AN | WR | TH | AMP | 1, 1 | 0, 0 | 2, 1 | 0, 0 |
| MED | SEN | AN | WR | TH | LAT | 1, 1 | 0, 0 | 3, 2 | 0, 0 |
| MED | SEN | AN | WR | TH | VEL | 0, 0 | 0, 0 | 1, 1 | 1, 1 |
| MED | SEN | OR | AMP | 0, 0 | 1, 0 | 0, 0 | 0, 0 | ||
| MED | SEN | OR | LAT | 0, 0 | 1, 0 | 0, 0 | 0, 0 | ||
| MED | SEN | OR | WR | AMP | 0, 0 | 0, 0 | 2, 2 | 0, 0 | |
| MED | SEN | OR | WR | LAT | 0, 0 | 0, 0 | 1, 1 | 2, 2 | |
| MED | SEN | OR | WR | VEL | 0, 0 | 0, 0 | 2, 2 | 0, 0 | |
| MED | SEN | OR | IN | AMP | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | IN | LAT | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | IN | OTH | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | IN | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | IN | PAL | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| MED | SEN | OR | IN | WR | AMP | 4, 3 | 0, 0 | 7, 5 | 2, 2 |
| MED | SEN | OR | IN | WR | LAT | 1, 1 | 0, 0 | 8, 7 | 3, 3 |
| MED | SEN | OR | IN | WR | OTH | 2, 2 | 0, 0 | 2, 1 | 1, 1 |
| MED | SEN | OR | IN | WR | PRE | 1, 1 | 0, 0 | 4, 4 | 0, 0 |
| MED | SEN | OR | IN | WR | VEL | 4, 3 | 1, 1 | 8, 7 | 3, 3 |
| MED | SEN | OR | MI | AMP | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | MI | LAT | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | MI | OTH | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | MI | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | MI | MI | AMP | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| MED | SEN | OR | MI | MI | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| MED | SEN | OR | MI | PAL | AMP | 1, 0 | 0, 0 | 0, 0 | 0, 0 |
| MED | SEN | OR | MI | PAL | VEL | 1, 0 | 0, 0 | 2, 2 | 0, 0 |
| MED | SEN | OR | MI | WR | AMP | 2, 1 | 0, 0 | 3, 3 | 4, 4 |
| MED | SEN | OR | MI | WR | INCH | 1, 1 | 0, 0 | 0, 0 | 2, 2 |
| MED | SEN | OR | MI | WR | LAT | 0, 0 | 0, 0 | 4, 3 | 0, 0 |
| MED | SEN | OR | MI | WR | OTH | 1, 1 | 0, 0 | 1, 1 | 0, 0 |
| MED | SEN | OR | MI | WR | PRE | 1, 1 | 0, 0 | 2, 2 | 1, 1 |
| MED | SEN | OR | MI | WR | VEL | 3, 2 | 0, 0 | 5, 5 | 5, 5 |
| MED | SEN | OR | OTH | VEL | 1, 1 | 0, 0 | 0, 0 | 0, 0 | |
| MED | SEN | OR | OTH | WR | AMP | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| MED | SEN | OR | OTH | WR | LAT | 0, 0 | 0, 0 | 2, 2 | 0, 0 |
| MED | SEN | OR | OTH | WR | VEL | 0, 0 | 0, 0 | 2, 2 | 1, 1 |
| MED | SEN | OR | PAL | WR | AMP | 0, 0 | 0, 0 | 2, 2 | 1, 1 |
| MED | SEN | OR | PAL | WR | LAT | 1, 1 | 1, 1 | 11, 11 | 1, 1 |
| MED | SEN | OR | PAL | WR | OTH | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| MED | SEN | OR | PAL | WR | PRE | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| MED | SEN | OR | PAL | WR | VEL | 0, 0 | 0, 0 | 7, 7 | 7, 6 |
| MED | SEN | OR | RI | AMP | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | RI | LAT | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | RI | OTH | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | RI | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | RI | WR | AMP | 3, 2 | 0, 0 | 3, 2 | 1, 1 |
| MED | SEN | OR | RI | WR | LAT | 1, 1 | 1, 1 | 4, 3 | 1, 1 |
| MED | SEN | OR | RI | WR | OTH | 1, 1 | 0, 0 | 1, 1 | 0, 0 |
| MED | SEN | OR | RI | WR | PRE | 1, 1 | 0, 0 | 1, 1 | 2, 2 |
| MED | SEN | OR | RI | WR | VEL | 2, 1 | 0, 0 | 3, 3 | 2, 2 |
| MED | SEN | OR | TH | AMP | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | TH | LAT | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | TH | OTH | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | TH | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| MED | SEN | OR | TH | ELB | PRE | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| MED | SEN | OR | TH | MI | VEL | 0, 0 | 0, 0 | 0, 0 | 0, 0 |
| MED | SEN | OR | TH | PAL | VEL | 0, 0 | 0, 0 | 0, 0 | 0, 0 |
| MED | SEN | OR | TH | WR | AMP | 1, 1 | 0, 0 | 3, 3 | 2, 2 |
| MED | SEN | OR | TH | WR | LAT | 0, 0 | 0, 0 | 3, 3 | 0, 0 |
| MED | SEN | OR | TH | WR | OTH | 1, 1 | 0, 0 | 1, 1 | 0, 0 |
| MED | SEN | OR | TH | WR | PRE | 1, 1 | 0, 0 | 1, 1 | 1, 1 |
| MED | SEN | OR | TH | WR | VEL | 1, 1 | 0, 0 | 5, 5 | 2, 2 |
| MED | SEN | OR | WR | ELB | AMP | 2, 1 | 0, 0 | 0, 0 | 1, 1 |
| MED | SEN | OR | WR | ELB | OTH | 1, 1 | 0, 0 | 0, 0 | 0, 0 |
| MED | SEN | OR | WR | ELB | PRE | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| MED | SEN | OR | WR | ELB | VEL | 2, 1 | 0, 0 | 0, 0 | 1, 1 |
| MED | Transcarpal | AMP | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |||
| MED | Transcarpal | LAT | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |||
| RAD | SEN | AN | FOR | TH | LAT | 0, 0 | 0, 0 | 1, 0 | 0, 0 |
| RAD | SEN | AN | WR | TH | AMP | 1, 1 | 0, 0 | 0, 0 | 0, 0 |
| RAD | SEN | AN | WR | TH | LAT | 1, 1 | 0, 0 | 2, 2 | 2, 0 |
| RAD | SEN | AN | WR | TH | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| RAD | SEN | OR | TH | WR | AMP | 1, 0 | 0, 0 | 1, 1 | 0, 0 |
| RAD | SEN | OR | TH | WR | LAT | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| RAD | SEN | OR | TH | WR | PRE | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| RAD | SEN | OR | TH | WR | VEL | 1, 0 | 0, 0 | 2, 2 | 1, 1 |
| ULN | MOT | LAT | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |||
| ULN | MOT | OTH | 1, 1 | 0, 0 | 1, 0 | 0, 0 | |||
| ULN | MOT | ELB | ADM | LAT | 1, 1 | 0, 0 | 0, 0 | 0, 0 | |
| ULN | MOT | ELB | ADM | OTH | 1, 1 | 0, 0 | 0, 0 | 0, 0 | |
| ULN | MOT | ELB | OTH | AMP | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| ULN | MOT | ELB | OTH | PRE | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| ULN | MOT | ELB | OTH | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 | |
| ULN | MOT | ELB | WR | VEL | 1, 1 | 0, 0 | 1, 1 | 1, 1 | |
| ULN | MOT | WR | LAT | 1, 1 | 0, 0 | 1, 1 | 0, 0 | ||
| ULN | MOT | WR | VEL | 1, 1 | 0, 0 | 0, 0 | 0, 0 | ||
| ULN | MOT | WR | ADM | AMP | 0, 0 | 0, 0 | 1, 0 | 2, 1 | |
| ULN | MOT | WR | ADM | LAT | 2, 2 | 1, 1 | 4, 2 | 5, 4 | |
| ULN | MOT | WR | ADM | OTH | 1, 1 | 0, 0 | 0, 0 | 0, 0 | |
| ULN | MOT | WR | ADM | VEL | 0, 0 | 0, 0 | 1, 0 | 0, 0 | |
| ULN | MOT | WR | APB | LAT | 1, 1 | 0, 0 | 0, 0 | 0, 0 | |
| ULN | MOT | WR | OTH | AMP | 0, 0 | 0, 0 | 1, 1 | 0, 0 | |
| ULN | MOT | WR | OTH | LAT | 0, 0 | 1, 1 | 3, 3 | 4, 3 | |
| ULN | MOT | WR | OTH | PRE | 0, 0 | 0, 0 | 1, 1 | 0, 0 | |
| ULN | MOT | WR | PAL | AMP | 0, 0 | 0, 0 | 1, 1 | 0, 0 | |
| ULN | MOT | WR | PAL | LAT | 0, 0 | 0, 0 | 1, 1 | 0, 0 | |
| ULN | SEN | OTH | 0, 0 | 0, 0 | 1, 0 | 0, 0 | |||
| ULN | SEN | WR | AMP | 1, 1 | 0, 0 | 0, 0 | 0, 0 | ||
| ULN | SEN | WR | LAT | 1, 1 | 0, 0 | 0, 0 | 0, 0 | ||
| ULN | SEN | AN | FOR | LI | LAT | 0, 0 | 0, 0 | 1, 0 | 0, 0 |
| ULN | SEN | AN | FOR | RI | LAT | 0, 0 | 0, 0 | 1, 0 | 0, 0 |
| ULN | SEN | AN | PAL | LI | LAT | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| ULN | SEN | AN | WR | LI | AMP | 0, 0 | 0, 0 | 2, 2 | 1, 1 |
| ULN | SEN | AN | WR | LI | LAT | 0, 0 | 0, 0 | 2, 2 | 1, 1 |
| ULN | SEN | AN | WR | LI | VEL | 0, 0 | 0, 0 | 3, 3 | 0, 0 |
| ULN | SEN | AN | WR | PAL | LAT | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| ULN | SEN | AN | WR | RI | LAT | 0, 0 | 0, 0 | 2, 2 | 4, 2 |
| ULN | SEN | AN | WR | RI | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| ULN | SEN | OR | LI | WR | AMP | 2, 1 | 0, 0 | 4, 3 | 3, 3 |
| ULN | SEN | OR | LI | WR | LAT | 1, 1 | 0, 0 | 3, 2 | 1, 1 |
| ULN | SEN | OR | LI | WR | OTH | 1, 1 | 0, 0 | 1, 0 | 0, 0 |
| ULN | SEN | OR | LI | WR | PRE | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| ULN | SEN | OR | LI | WR | VEL | 2, 1 | 0, 0 | 3, 2 | 3, 3 |
| ULN | SEN | OR | OTH | VEL | 1, 1 | 0, 0 | 0, 0 | 0, 0 | |
| ULN | SEN | OR | OTH | WR | VEL | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| ULN | SEN | OR | PAL | WR | AMP | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| ULN | SEN | OR | PAL | WR | LAT | 0, 0 | 1, 1 | 6, 6 | 0, 0 |
| ULN | SEN | OR | PAL | WR | VEL | 0, 0 | 0, 0 | 2, 2 | 1, 1 |
| ULN | SEN | OR | RI | WR | AMP | 2, 1 | 0, 0 | 2, 1 | 2, 2 |
| ULN | SEN | OR | RI | WR | LAT | 1, 1 | 1, 1 | 3, 2 | 1, 1 |
| ULN | SEN | OR | RI | WR | PRE | 0, 0 | 0, 0 | 1, 1 | 1, 1 |
| ULN | SEN | OR | RI | WR | VEL | 2, 1 | 0, 0 | 2, 2 | 3, 3 |
| ULN | SEN | OR | WR | ELB | AMP | 1, 1 | 0, 0 | 0, 0 | 0, 0 |
| ULN | SEN | OR | WR | ELB | OTH | 1, 1 | 0, 0 | 0, 0 | 0, 0 |
| ULN | SEN | OR | WR | ELB | VEL | 1, 1 | 0, 0 | 0, 0 | 0, 0 |
| Legend: | ||||||||
|---|---|---|---|---|---|---|---|---|
| Blank cells—Not reported or not applicable | ||||||||
| First entry in cell—Total number of articles | ||||||||
| Second entry in cell—Number of articles with derivable sensitivity and specificity | ||||||||
| Composite test type | Patient selection group | |||||||
| Nerve for test 1 | Nerve for test 2 | Motor or sensory | Unit of nerve test | Type composite | Complex objective standard | Simple nerve conduction | Symptoms/ presented | Unspecified diagnosis |
| Median | Median | Motor | Amplitude | Difference | 0, 0 | 0, 0 | 0, 0 | 1, 0 |
| Median | Median | Motor | Amplitude | Ratio | 1, 1 | 0, 0 | 1, 0 | 0, 0 |
| Median | Median | Motor | Latency | Difference | 0, 0 | 0, 0 | 2, 2 | 2, 2 |
| Median | Median | Motor | Latency | Ratio | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| Median | Median | Motor | Velocity | Difference | 1, 0 | 0, 0 | 0, 0 | 0, 0 |
| Median | Median | Sensory | Amplitude | Difference | 1, 1 | 0, 0 | 2, 2 | 0, 0 |
| Median | Median | Sensory | Amplitude | Ratio | 1, 1 | 0, 0 | 1, 0 | 1, 1 |
| Median | Median | Sensory | Latency | Difference | 1, 1 | 0, 0 | 6, 5 | 1, 1 |
| Median | Median | Sensory | Latency | Ratio | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| Median | Median | Sensory | Velocity | Difference | 0, 0 | 0, 0 | 2, 2 | 1, 1 |
| Median | Median | Sensory | Velocity | Ratio | 0, 0 | 0, 0 | 4, 4 | 2, 2 |
| Median | Radial | Sensory | Latency | Difference | 1, 1 | 0, 0 | 3, 3 | 2, 0 |
| Median | Radial | Sensory | Velocity | Difference | 0, 0 | 0, 0 | 0, 0 | 1, 1 |
| Median | Radial | Sensory | Velocity | Ratio | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| Median | Ulnar | Motor | Latency | Difference | 1, 1 | 2, 2 | 3, 3 | 5, 4 |
| Median | Ulnar | Motor | Other | Difference | 1, 1 | 0, 0 | 0, 0 | 0, 0 |
| Median | Ulnar | Sensory | Amplitude | Ratio | 0, 0 | 0, 0 | 2, 2 | 1, 1 |
| Median | Ulnar | Sensory | Latency | Difference | 1, 1 | 1, 1 | 10, 9 | 5, 3 |
| Median | Ulnar | Sensory | Velocity | Difference | 0, 0 | 0, 0 | 1, 1 | 1, 1 |
| Median | Ulnar | Sensory | Velocity | Ratio | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| Radial | Median | Sensory | Velocity | Ratio | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| Radial | Radial | Sensory | Latency | Difference | 0, 0 | 0, 0 | 1, 0 | 0, 0 |
| Ulnar | Median | Sensory | Velocity | Difference | 1, 0 | 0, 0 | 0, 0 | 0, 0 |
| Ulnar | Median | Sensory | Velocity | Ratio | 0, 0 | 0, 0 | 1, 1 | 0, 0 |
| Other Difference | 3, 1 | 0, 0 | 3, 3 | 1, 1 | ||||
| Other Ratio | 0, 0 | 0, 0 | 3, 2 | 1, 1 | ||||
| Other Composite | 5, 4 | 0, 0 | 9, 8 | 4, 2 | ||||
| Legend: | ||||
|---|---|---|---|---|
| First entry in cell—Total number of articles | ||||
| Second entry in cell—Number of articles with derivable sensitivity and specificity | ||||
| Imaging test | Complex objective standard | Simple nerve conduction | Symptoms/presented | Unspecified diagnosis |
| CT | 0, 0 | 0, 0 | 0, 0 | 2, 0 |
| MRI | 2, 0 | 2, 0 | 1, 1 | 5, 2 |
| Ultrasound | 1, 0 | 1, 0 | 1, 0 | 3, 3 |
As an initial criterion for conducting meta-analyses, we required that a minimum of 10 studies that reported a specific test in a specific population had derivable sensitivity and specificity. In other words, the second number in the table cell was required to be 10 or more. We adopted this criterion to ensure that our analysis would focus on the diagnostic tests that are the subject of greatest research interest. When there was a minimum of 10 articles, we proceeded with a meta-analysis even if one or more articles were subsequently excluded because it did not report sensitivity and specificity for the particular test being analyzed (or for other reasons discussed below).
Ideally, a meta-analysis of a test includes only studies that use the same definition of what is to be diagnosed. However, the absence of a gold standard for defining carpal tunnel syndrome resulted in there being as many different definitions of the condition (and therefore of positive cases) as there were studies. Therefore, we could only combine study results by permitting different authors to use different definitions of CTS. Testing for heterogeneity of results helps reduce, but does not eliminate the possibility that different definitions affected study results.
Distal Motor Latency: Patients with Unspecified Diagnosis of CTS v. Normal Controls
| Study | Reason for Exclusion |
|---|---|
| Pease, 1990177 | Did not report sensitivity and specificity for distal motor latency test |
| Seror, 1998159 | Did not report sensitivity and specificity for distal motor latency test |
| Rossi, 1994178 | Did not report sensitivity and specificity for distal motor latency test |
| Seror, 1995179 | Did not report sensitivity and specificity for distal motor latency test |
| Lang, 1995109 | Did not report sensitivity and specificity for distal motor latency test |
| Tzeng, 1990180 | Did not report sensitivity and specificity for distal motor latency test |
| Mondelli, 2001181 | Did not report sensitivity and specificity for distal motor latency test |
| Simovic, 1997182 | Did not report distal motor latency results for control subjects |
| Simovic, 1999183 | Did not report distal motor latency results for control subjects |
| Resende, 2000184 | Did not report distal motor latency results for control subjects |
| Lauritzen, 1991185 | Did not report distal motor latency results for control subjects |
| Loscher, 2000175 | Did not report distal motor latency results for CTS patients |
| Bronson, 1997163 | Selective reporting of distal motor latency results |
| So, 1989173 | Reported combination test of distal motor latency and other nerve conduction measurements |
| Charles, 1990170 | Discrepancy in reported threshold |
| Resende, 2000174 | No diagnostic threshold reported |
Two of the five studies included in the meta-analysis140, 175 did not report counts of normal and abnormal results in the control subjects, but because their thresholds were based on two standard deviations from the mean, we estimated the number of false-positive results by multiplying the number of patients in the control group by the probability that a result would be two or more standard deviations above the mean (0.02275 based on the normal distribution). We also recalculated the results from the study by Rosén176, which reported a histogram of latency results and did not report a 2 × 2 table for their specified threshold. In the other included articles, there were no discrepancies between the sensitivity and specificity figures reported by the authors and the figures calculated by ECRI and used in the meta-analysis.
| Study | TP | FN | FP | TN | Sen. 95% CI | Spec. 95% CI | PPV 95% CI | NPV 95% CI | Prev. |
|---|---|---|---|---|---|---|---|---|---|
| aKimura140 | 105 | 67 | 3 | 119 | 61.0% | 97.5% | 97.2% | 64.0% | 58.5% |
| 53.4% 68.2% | 92.9% 99.2% | 92.0% 99.1% | 56.7% 70.7% | ||||||
| Marin139 | 9 | 5 | 0 | 12 | 64.3% | 100% | 100% | 70.6% | 53.8% |
| 38.3% 83.9% | 75.0% 100% | 69.2% 100% | 46.4% 86.9% | ||||||
| Loong141 | 17 | 10 | 0 | 30 | 63.0% | 100% | 100% | 75.0% | 47.4% |
| 43.9% 78.7% | 88.2% 100% | 81.0% 100% | 59.5% 86.0% | ||||||
| Plaja142 | 16 | 7 | 0 | 20 | 69.6% | 100% | 100% | 74.1% | 53.5% |
| 48.7% 84.6% | 83.3% 100% | 80.0% 100% | 54.9% 87.0% | ||||||
| bRosén138 | 12 | 29 | 0 | 50 | 29.3% | 100% | 100% | 63.3% | 45.1% |
| 17.4% 44.8% | 92.6% 100% | 75.0% 100% | 52.0% 73.3% | ||||||
| Meta-analysis results (mean threshold) | 57.1% | 97.9% | |||||||
| 49.1% 64.8% | 97.1% 98.5% | ||||||||
Key:
TP-true positive, FN-false negative, FP-false positive, TN-true negative
Sen.-sensitivity, Spec-specificity, PPV-positive predictive value, NPV-negative predictive value, Prev.-prevalence of CTS
Confidence intervals on sensitivity, specificity, PPV, NPV calculated by Wilson method96
Counts for control group (false positive, true negative) estimated by ECRI from threshold reported by authors (mean + 2 SD)
Results calculated by ECRI from published histogram
. No statistically
significant heterogeneity was found in the results (Q = 0.33, p = 0.99).
The results clustered in a small portion of the graph, suggesting there
was good agreement among clinicians in how this test is used and how
effective it is. The sensitivity and specificity at mean threshold, our
best estimate of the effectiveness of the test, was 57.1% sensitivity,
97.9% specificity.
The section of the summary ROC curve above sensitivity = 70% is an extrapolation from the actual data. It represents thresholds that are much lower than the thresholds used in the published trials and as such, may not represent an accurate description of clinical events.
Distal Motor Latency: Patients with Symptoms of CTS v. Normal Controls
| Study | Reason for Exclusion |
|---|---|
| Jackson, 1989150 | Did not report sensitivity and specificity for distal motor latency test |
| Sener, 2000186 | Did not report sensitivity and specificity for distal motor latency test |
| Schwartz, 1979187 | Did not report sensitivity and specificity for distal motor latency test |
| Escobar, 1985151 | Did not report sensitivity and specificity for distal motor latency test |
| Preston, 1992188 | Did not report distal motor latency results for control subjects |
| Kimura, 1985189 | Did not report distal motor latency results for control subjects |
| Cherniak, 1996190 | Used distal motor latency for patient selection |
| Sheean, 1995191 | Used distal motor latency for patient selection |
| Foresti, 1996192 | Discrepancies in reported results |
| Study | TP | FN | FP | TN | Sen. | Spec. | PPV | NPV | Prev. |
|---|---|---|---|---|---|---|---|---|---|
| 95% CI | 95% CI | 95% CI | 95% CI | ||||||
| a, bChang145 | 17 | 26 | 0 | 40 | 39.5% | 100% | 100% | 60.6% | 51.8% |
| 26.1% 54.7% | 90.9% 100% | 81.0% 100% | 48.3% 71.7% | ||||||
| Kuntzer144 | 47 | 53 | 1 | 69 | 47.0% | 98.6% | 97.9% | 56.6% | 58.8% |
| 37.3% 56.9% | 92.1% 99.8% | 88.8% 99.6% | 47.5% 65.2% | ||||||
| aMurthy143143 | 38 | 19 | 2 | 72 | 66.7% | 97.3% | 95.0% | 79.1% | 43.5% |
| 53.5% 77.7% | 90.5% 99.3% | 83.2% 98.6% | 69.5% 86.3% | ||||||
| Cioni146 | 300 | 75 | 0 | 56 | 80.0% | 100% | 100% | 42.7% | 87.0% |
| 75.6% 83.8% | 93.3% 100% | 98.7% 100% | 34.4% 51.5% | ||||||
| bMessina120 | 34 | 6 | 1 | 39 | 85.0% | 97.5% | 97.1% | 86.7% | 50.0% |
| 70.6% 93.0% | 86.8% 99.6% | 85.1% 99.5% | 73.5% 93.8% | ||||||
| Melvin147 | 13 | 4 | 0 | 24 | 76.5% | 100% | 100% | 85.7% | 41.5% |
| 52.2% 90.6% | 85.7% 100% | 76.5% 100% | 68.1% 94.4% | ||||||
| Loong148 | 13 | 9 | 0 | 60 | 59.1% | 100% | 100% | 87.0% | 26.8% |
| 38.4% 77.0% | 93.8% 100% | 76.5% 100% | 76.8% 93.1% | ||||||
| cAtroshi136 | 25 | 18 | 8 | 52 | 58.1% | 86.7% | 75.8% | 74.3% | 41.7% |
| 43.0% 71.9% | 75.6% 93.2% | 58.6% 87.3% | 62.7% 83.2% | ||||||
| Meta-analysis results (mean threshold) | 66.0% | 98.3% | |||||||
| 55.7% 75.0% | 97.4% 98.9% | ||||||||
Key:
TP-true positive, FN-false negative, FP-false positive, TN-true negative
Sen.-sensitivity, Spec-specificity, PPV-positive predictive value, NPV-negative predictive value, Prev.-prevalence of CTS
Confidence intervals on sensitivity, specificity, PPV, NPV calculated by Wilson method
Counts for control group (false positive, true negative) estimated by ECRI from threshold reported by authors (mean + 2 or 2.5 SD)
Results calculated by ECRI from published graph
Outlier (excluded from meta-analysis results): see text
Note: One outlier136 was excluded (see text).
), the
sensitivity/specificity at mean threshold was 66.0%/98.3%. Including the
outlier changed the results by less than a percentage point: the
sensitivity/specificity at mean threshold was 65.0%/97.7%.
The results of this meta-analysis are very similar to the results for the meta-analysis of DML with patient groups with unspecified diagnosis of CTS. The results of both meta-analyses suggest that this test has very high specificity, but only moderate sensitivity.
Palmar Sensory Latency: Patients with Symptoms of CTS v. Normal Controls
| Study | Reason for Exclusion |
|---|---|
| Gerr, 1998 31 | Did not report sensitivity and specificity for palmar sensory latency test |
| Foresti, 1996 192 | Did not report sensitivity and specificity for palmar sensory latency test |
| Eisen, 1993 193 | Did not report sensitivity and specificity for palmar sensory latency test |
| Mills, 1985 194 | Did not report sensitivity and specificity for palmar sensory latency test |
| Kim, 1983 195 | Did not report sensitivity and specificity for palmar sensory latency test |
| Andary, 1996 196 | Palmar sensory latency results used as patient selection criterion |
| Study | TP | FN | FP | TN | Sen. 95% CI | Spec. 95% CI | PPV 95% CI | NPV 95% CI | Prev. |
|---|---|---|---|---|---|---|---|---|---|
| a, bChang145 | 26 | 17 | 0 | 40 | 60.5% 45.3% 73.9% | 100% 90.9% 100% | 100% 86.7% 100% | 70.2% 57.1% 80.6% | 51.8% |
| cJackson150 | 91 | 40 | 1 | 37 | 69.5% 60.9% 76.8% | 97.4% 86.2% 99.5% | 98.9% 93.9% 99.8% | 48.1% 37.0% 59.3% | 77.5% |
| aMurthy143 | 55 | 2 | 2 | 72 | 96.5% 87.8% 99.1% | 97.3% 90.5% 99.3% | 96.5% 87.8% 99.1% | 97.3% 90.5% 99.3% | 43.5% |
| aEscobar151 | 32 | 8 | 2 | 102 | 80.0% 64.9% 89.6% | 98.1% 93.1% 99.5% | 94.1% 80.5% 98.4% | 92.7% 86.1% 96.3% | 27.8% |
| cGirlanda149 | 38 | 37 | 1 | 89 | 50.7% 39.4% 61.9% | 98.9% 93.8% 99.8% | 97.4% 86.5% 99.6% | 70.6% 62.0% 78.0% | 45.5% |
| Meta-analysis results (mean threshold) | 75.8% 68.8% 81.6% | 97.7% 96.8% 98.4% | |||||||
Key:
TP-true positive, FN-false negative, FP-false positive, TN-true negative
Sen.-sensitivity, Spec-specificity, PPV-positive predictive value, NPV-negative predictive value, Prev.-prevalence of CTS Confidence intervals on sensitivity, specificity, PPV, NPV calculated by Wilson method
Counts for control group (false positive, true negative) estimated by ECRI from threshold reported by authors (mean + 2 or 2.5 SD)
Results calculated by ECRI from published graph
Results calculated by ECRI from published percentages
.
Like DML, palmar sensory latency has very high specificity. The normal volunteers studied in these trials rarely had abnormal results. This finding, however, does not reveal the test performance on persons with suspected CTS. To address that issue, a computation of sensitivity is required. The sensitivity/specificity at mean threshold was 75.8%/97.7%, and it is clear that the test has some ability to identify persons with symptoms of CTS. Although the summary ROC can be extrapolated to a point where sensitivity and specificity are both quite high (i.e., 96%, 96% respectively), in actual practice it is likely that only specificity is so high. Sensitivity was lower than specificity in all five studies.
Phalen's Maneuver: Combined CTS Groups v. Normal Controls
There were no clinical signs or symptoms for which at least 10 articles reported sensitivity and specificity in a specific patient population. Therefore, we loosened the inclusion criteria by first combining the four patient selection categories, and then requiring a total of 20 or more sensitivity/specificity articles. Because none of the signs and symptoms data met that loosened criterion, we again lowered the threshold to a total of 15 studies or more. Two tests met that criterion: Phalen's maneuver and Tinel's sign. We proceeded to attempt meta-analysis of these data, recognizing that combining patient selection groups could cause heterogeneity of study results that could prevent meta-analysis.
| Study | Reason for Exclusion |
|---|---|
| Koris, 1988 198 | Did not report specificity of Phalen's maneuver |
| Brahme, 1997 199 | Did not report specificity of Phalen's maneuver |
| Lang, 1995 109 | Did not report specificity of Phalen's maneuver |
| Glass, 1995 28 | Reported results for only 22 of 159 affected hands |
| Gerr, 1994 197 | Duplicate publication |
| Study | TP | FN | FP | TN | Sen. 95% CI | Spec. 95% CI | PPV 95% CI | NPV 95% CI | Prev. |
|---|---|---|---|---|---|---|---|---|---|
| De Smet101 | 57 | 9 | 4 | 77 | 86.4% 75.8% 92.7% | 95.1% 87.8% 98.1% | 93.4% 84.1% 97.5% | 89.5% 81.1% 94.5% | 44.9% |
| Durkan155 | 32 | 14 | 8 | 42 | 69.6% 54.9% 81.1% | 84.0% 71.2% 91.8% | 80.0% 64.9% 89.6% | 75.0% 62.0% 84.6% | 47.9% |
| Gellman106 | 45 | 18 | 10 | 40 | 71.4% 59.0% 81.3% | 80.0% 66.7% 88.9% | 81.8% 69.4% 89.9% | 69.0% 55.9% 79.6% | 55.8% |
| a, bGerr31 | 48 | 67 | 4 | 119 | 41.7% 33.0% 51.1% | 96.7% 91.8% 98.8% | 92.3% 81.5% 97.0% | 64.0% 56.7% 70.7% | 48.3% |
| bGhavanini154 | 34 | 40 | 17 | 41 | 45.9% 34.9% 57.4% | 70.7% 57.7% 81.0% | 66.7% 52.7% 78.2% | 50.6% 39.7% 61.4% | 56.1% |
| González del Pino 104 | 174 | 26 | 20 | 180 | 87.0% 81.5% 91.0% | 90.0% 84.9% 93.5% | 89.7% 84.5% 93.3% | 87.4% 82.0% 91.3% | 50.0% |
| aSzabo152 | 65 | 22 | 5 | 95 | 74.7% 64.4% 82.8% | 95.0% 88.7% 97.9% | 92.9% 84.1% 97.0% | 81.2% 73.0% 87.3% | 46.5% |
| Tetro1021 | 58 | 37 | 16 | 80 | 61.1% 50.8% 70.4% | 83.3% 74.4% 89.6% | 78.4% 67.5% 86.4% | 68.4% 59.3% 76.2% | 49.7% |
| Fertl153 | 50 | 23 | 3 | 36 | 68.5% 56.9% 78.2% | 92.3% 79.3% 97.4% | 94.3% 84.4% 98.1% | 61.0% 48.0% 72.6% | 65.2% |
| cWerner111 | 17 | 14 | 0 | 20 | 54.8% 37.5% 71.1% | 100% 83.3% 100% | 100% 81.0% 100% | 58.8% 41.9% 73.9% | 60.8% |
| Meta-analysis results (mean threshold) | NA | NA | |||||||
Key:
TP-true positive, FN-false negative, FP-false positive, TN-true negative
Sen.-sensitivity, Spec-specificity, PPV-positive predictive value, NPV-negative predictive value, Prev.-prevalence of CTS Confidence intervals on sensitivity, specificity, PPV, NPV calculated by Wilson method
NA—Results not valid because of excessive heterogenity in study results
Results calculated by ECRI from published percentages
Errors in published results corrected by ECRI
Tested reverse Phalen's maneuver
| Group | Q (p-value) for larger group |
|---|---|
| All articles (N = 10) | 71.4 (p <0.000001) |
| Patients selected with complex objective standard (N = 6) v. other selection | 59.4 (p <0.000001) |
| Reverse Phalen's maneuver (N = 1) v. conventional | 70.8 (p <0.000001) |
| Not funded by for-profit device or drug manufacturer (N = 4) v. not reported | 58.5 (p <0.000001) |
| Reported both inclusion and exclusion criteria (N = 4) v. reported only inclusion criteria | 20.5 (p = 0.001) |
| Prospective patient selection (N = 5) v. not reported | 58.7 (p <0.000001) |
| Comorbidity reported (N = 1) v. not reported | 69.9 (p <0.000001) |
| Sex ratios of patients, controls within 20% of each other (N = 5) v. possible sex bias | 58.5 (p <0.000001) |
| Mean ages of patients, controls within 5 years (N = 3) v. possible age bias | 15.4 (p = 0.017) |
| Duration of condition reported (N = 2) v. not reported | 48.4 (p <0.000001) |
| Independent reference standard (N = 4) v. no independent reference standard reported | 48.2 (p <0.000001) |
| Patients given both study test and reference test (N = 3) v. did not do so | 49.3 (p <0.000001) |
| Studies done in USA (N = 6) v. other countries | 58.1 (p <0.000001) |
| Potential selection bias for easy cases (N = 4) v. no bias or not reported | 49.3 (p <0.000001) |
Q—Q-statistic, with probability that variability in study results [D, logit (sensitivity) + logit (specificity)] is the result of random variability within a homogeneous sample of studies.
The variability of results is shown in Figure 12
; sensitivity/specificity covered a large range. We
can only conclude that Phalen's maneuver has some ability to distinguish
CTS patients from normal controls; the data are too heterogeneous to
estimate sensitivity or specificity.
Tinel's Sign: Combined CTS Groups v. Normal Controls
| Study | TP | FN | FP | TN | Sen. 95% CI | Spec. 95% CI | PPV 95% CI | NPV 95% CI | Prev. |
|---|---|---|---|---|---|---|---|---|---|
| De Smet101 | 14 | 17 | 0 | 81 | 45.2% 28.9% 62.5% | 100% 95.3% 100% | 100% 77.8% 100% | 82.7% 73.8% 89.0% | 27.7% |
| Durkan155 | 26 | 20 | 10 | 40 | 56.5% 42.0% 70.0% | 80.0% 66.7% 88.9% | 72.2% 55.7% 84.3% | 66.7% 53.8% 77.5% | 47.9% |
| Gellman106 | 29 | 37 | 3 | 47 | 43.9% 32.4% 56.2% | 94.0% 83.5% 98.0% | 90.6% 75.4% 96.8% | 56.0% 45.1% 66.3% | 56.9% |
| Gelmers29 | 20 | 27 | 11 | 32 | 42.6% 29.3% 57.0% | 74.4% 59.4% 85.2% | 64.5% 46.6% 79.1% | 54.2% 41.4% 66.5% | 52.2% |
| a, bGerr31 | 8 | 50 | 2 | 121 | 13.8% 7.1% 25.2% | 98.4% 94.1% 99.6% | 80.0% 48.4% 94.5% | 70.8% 63.4% 77.2% | 32.0% |
| Ghavanini154 | 24 | 52 | 9 | 49 | 31.6% 22.1% 42.9% | 84.5% 72.8% 91.7% | 72.7% 55.4% 85.1% | 48.5% 38.8% 58.3% | 56.7% |
| González del Pino104 | 42 | 87 | 6 | 194 | 32.6% 24.9% 41.2% | 97.0% 93.5% 98.6% | 87.5% 75.0% 94.2% | 69.0% 63.3% 74.3% | 39.2% |
| aSeror156 | 63 | 37 | 18 | 22 | 63.0% 53.0% 72.0% | 55.0% 39.5% 69.6% | 77.8% 67.4% 85.6% | 37.3% 25.9% 50.3% | 71.4% |
| Stewart157 | 23 | 28 | 15 | 37 | 45.1% 32.0% 58.9% | 71.2% 57.4% 81.8% | 60.5% 44.4% 74.6% | 56.9% 44.6% 68.5% | 49.5% |
| aSzabo152 | 56 | 31 | 1 | 99 | 64.4% 53.7% 73.8% | 99.0% 94.4% 99.8% | 98.2% 90.5% 99.7% | 76.2% 68.0% 82.8% | 46.5% |
| aTetro102 | 70 | 25 | 9 | 87 | 73.7% 63.8% 81.6% | 90.6% 82.9% 95.1% | 88.6% 79.5% 94.0% | 77.7% 68.9% 84.5% | 49.7% |
| Meta-analysis results (mean threshold) | NA | NA | |||||||
TP-true positive, FN-false negative, FP-false positive, TN-true negative
Sen.-sensitivity, Spec-specificity, PPV-positive predictive value, NPV-negative predictive value, Prev.-prevalence of CTS Confidence intervals on sensitivity, specificity, PPV, NPV calculated by Wilson method
NA—Results not valid because of excessive heterogenity in study results
Results calculated by ECRI from published percentages
Errors in published results corrected by ECRI
The heterogeneity is evident in Figure 13
. Sensitivity/specificity results are widely dispersed in
the graph, and there is no pattern of results that is obvious on
inspection. The data suggest that Tinel's sign has some ability to
diagnose CTS, but the sensitivity and specificity of the test are
uncertain. However, the sensitivity of the test appears to be low.
| Group | Q (p-value) for larger group |
|---|---|
| All articles (N = 11) | 59.1 (p <0.000001) |
| Patients selected with complex objective standard (N = 9) v. other selection | 46.1 (p <0.000001) |
| Not funded by for-profit device or drug manufacturer (N = 5) v. not reported | 10.7 (p = 0.057) |
| Reported both inclusion and exclusion criteria (N = 6) v. reported only inclusion criteria | 30.2 (p = 0.000013) |
| Prospective patient selection (N = 4) v. not reported | 16.6 (p = 0.011) |
| Comorbidity reported (N = 2) v. not reported | 51.4 (p <0.000001) |
| Mean ages of patients, controls within 5 years (N = 6) v. possible age bias | 37.8 (p <0.000001) |
| Possible sex bias (N = 3) vs. sex ratios of patients, controls within 20% of each other (N = 8) | 52.8 (p <0.000001) |
| Duration of condition reported (N = 2) v. not reported | 50.6 (p <0.000001) |
| Independent reference standard (N = 6) v. no independent reference standard reported | 16.5 (p = 0.005545) |
| Patients given both study test and reference test (N = 3) v. did not do so | 51.6 (p <0.000001) |
| Studies done in USA (N = 5) v. other countries | 22.3 (p = 0.000454) |
| Potential selection bias for easy cases (N = 4) v. no bias or not reported | 41.9 (p <0.000001) |
Q—Q-statistic, with probability that variability in study results [D, logit (sensitivity) + logit (specificity)] is the result of random variability within a homogeneous sample of studies.
| Article | N | Population | Symptoms | Positive NCS | Symptoms & Positive NCS |
|---|---|---|---|---|---|
| Workers-at-risk screening studies for carpal tunnel syndrome | |||||
| Kearns, 2000 204 | 45 | Pork processors | NR | NR | NR |
| Missere, 1999 205 | 45 | Meat manufacturers | NR | a 28.9% | NR |
| Nathan, 1998 202 | 283 | Steel mill workers, food processors, electronics workers, and plastics workers | 12.9% | 43.0% | 8.2% |
| Tan, 1998 206 | 64 | Carpet weavers | NR | NR | NR |
| Werner, 1998 207 | 119 | Automobile parts manufacturers | NR | 27% | b 20.2% |
| 98 | Furniture manufacturers | NR | 26% | b 10.2% | |
| 77 | Paper containers manufacturers | NR | 34% | b 14.3% | |
| 64 | Automobile parts manufacturers | NR | 30% | b 17.2% | |
| 164 | Clerical insurance workers | NR | 15% | b 11.0% | |
| 202 | Spark plugs manufacturers | NR | 28% | b 9.4% | |
| Franzblau, 1997 208 | 148 | Automobile parts manufacturers | 41% | NR | NR |
| Jeng, 1997 209 | 27 | Food processors | 48.8% | 34.1% | 22.0% |
| Werner, 1997 210 | 59 | Manufacturing workers and clerical workers | 11.1% | 45.4% | 5.6% |
| Bingham, 1996 211 | 1021 | Applicants for jobs in meat packers, plastics assemblers, food processors, furniture manufacturers, or grocery warehousing workers | c 6.0% | a 17.4% | c 1.8% |
| Murata, 1996 164 | 27 | Data entry operators | NR | 37% | NR |
| Pierre-Jerome, 1996 212 | 24 | Floor cleaners | NR | NR | NR |
| Werner, 1995 213 | 167 | Automobile parts manufacturers | 19.8% | 24.6% | 9.0% |
| Young, 1995 166 | 157 | Poultry processors | 70%b | 31% | NR |
| Franzblau, 1994 113 | 84 | Automobile parts manufacturers | 21.4% | 19.3% | 8.40% |
| Kirschberg, 1994 214 | 112 | Poultry processors | 22.3% | 29.5% | 17.0% |
| Nathan, 1994 215 | 101 | Japanese furniture factory workers | a, b4.5% | b17.8% | b2.0% |
| 316 | Steel mill workers, food processors, electronics workers, and plastics workers | a, b23.4% | b22.0% | b8.3% | |
| Nilsson, 1994 216 | 61 | Office workers | NR | 33% | NR |
| 58 | Truck assemblers | NR | 40% | NR | |
| 56 | Platers | NR | 55% | NR | |
| Werner, 1994 217 | 130 | Automobile parts manufacturers | 27.7% | d 20.2% | NR |
| Johnson, 1993 167 | 184 | Poultry processors | a, b 37.3% | a, b 19.2% | a, b 6.0% |
| Nathan, 1993 218 | 737 | Steel mill workers, meat/food processors, electronics workers, plastics workers, aluminum reduction workers, and cable plant workers. | a, b51.0% | a, b 33.6% | a, b19.8% |
| Grant, 1992 219 | 63 | Manufacturing plant workers | a 25.4% | NR | NR |
| Jetzer, 1991 168 | 39 | Computer assemblers | NR | NR | NR |
| 100 | Meat processors | NR | NR | NR | |
| 284 | Keyboard operators | NR | NR | NR | |
| General population screening studies for carpal tunnel syndrome | |||||
| Atroshi, 1999 220 | 2466 | General population | 14.4% | c 22.3% | c 6.6% |
| Ferry, 1998 221 | 648 | General population | 18.5% | 17.4% | 7.7% |
| DeKrom, 1990 222 | 500 | General population | 13.8% | NR | c 7.8% |
Key
NR-Not reported
NCS-Nerve conduction studies
Based on hands instead of participants
Calculated by ECRI based on information reported in the article
Estimated by ECRI based on information reported in the article
Prevalence of positive NCS in the study by Werner217 was based on 129 participants .
| Author, Year | Clinical findings | Nerve conduction studies | Comments | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| SYM | CLN | OTH CLN | DML | DSL | PAL | SEN DIF | MOT DIF | OTH NCS | ||
| Bland, 2000 200 | □ | □ | □ |
| □ | □ | □ | □ |
| If tests equivocal, authors measured sensory potential or inching test |
| Kearns, 2000 204 | □ | □ | □ | □ | □ | □ | □ | □ | □ | NR |
| Atroshi, 1999 220 |
|
| □ | □ | □ | □ |
| □ | □ | |
| Missere, 1999 205 | □ | □ | □ | □ | □ | □ | □ | □ |
| |
| Ferry, 1998 221 | □ | □ | □ | □ | □ | □ | □ | □ | □ | NR |
| Nathan, 1998 202 |
| □ | □ | □ |
|
| □ | □ |
| |
| Rosen, 1998 201 | □ | □ | □ | □ | □ | □ | □ | □ | □ | NR |
| Tan, 1998 206 | □ | □ | □ | □ | □ | □ | □ | □ | □ | NR |
| Werner, 1998 207 |
| □ | □ | □ | □ | □ |
| □ | □ | |
| Franzblau, 1997 208 | □ | □ | □ | □ | □ | □ | □ | □ | □ | NR |
| Jeng, 1997 209 |
| □ | □ |
|
| □ |
| □ | □ | |
| Werner, 1997 210 |
| □ | □ | □ | □ | □ |
| □ | □ | |
| Bingham, 1996 211 | □ | □ | □ | □ | □ | □ | □ | □ | □ | NR |
| Murata, 1996 164 | □ | □ | □ | □ | □ | □ | □ | □ | □ | NR |
| Pierre-Jerome, 1996 212 | □ | □ | □ | □ | □ | □ | □ | □ | □ | NR |
| Werner, 1995 213 |
| □ | □ | □ | □ | □ |
| □ | □ | |
| Young, 1995 166 | □ | □ | □ | □ | □ | □ | □ | □ | □ | NR |
| Franzblau, 1994 113 |
| □ | □ | □ | □ | □ |
| □ | □ | |
| Kirschberg, 1994 214 |
|
|
|
| □ |
|
| □ |
| |
| Nathan, 1994 215 |
| □ | □ | □ |
|
| □ | □ |
| |
| Nilsson, 1994 216 | □ | □ | □ | □ | □ | □ | □ | □ | □ | NR |
| Werner, 1994 217 | □ | □ | □ | □ | □ | □ | □ | □ | □ | NR |
| Johnson, 1993 167 | □ | □ | □ | □ | □ | □ | □ | □ | □ | NR |
| Nathan, 1993 218 |
| □ | □ | □ |
|
| □ | □ |
| |
| Grant, 1992219 | □ | □ | □ |
|
| □ |
|
| ? | |
| Jetzer, 1991 168 |
| □ | □ | □ | □ | □ | □ | □ | □ | Or positive NCS (tests not reported) |
| DeKrom, 1990 222 |
| □ | □ |
| □ | □ |
| □ | □ | |
| Welch, 1973 223 | □ | □ | □ | □ | □ | □ | □ | □ | □ | NR |
| Totals | 12 | 2 | 1 | 5 | 5 | 4 | 9 | 1 | 6 | |
Key
SYM—Were positive symptoms included in the author's method of diagnosis?
CLN—Was a positive clinical exam included in the author's method of diagnosis?
OTH CLN—Were other clinical findings included in the author's method of diagnosis?
DML—Was distal motor latency included in the author's method of diagnosis?
DSL—Was distal sensory latency included in the author's method of diagnosis?
PAL—Was palmar sensory latency included in the author's method of diagnosis?
SEN DIF—Was the difference between median and ulnar sensory studies included in the author's method of diagnosis?
MOT DIF—Was the difference between median and ulnar motor studies included in the author's method of diagnosis?
OTH NCS—Were other nerve conduction studies included in the author's method of diagnosis?
NR—Method of diagnosis was not reported
| Legend: | |
|---|---|
| First entry in cell—Total number of articles | |
| Second entry in cell—Number of articles with derivable sensitivity and specificity | |
| Sign/symptom | Number of articles reporting |
| Clinical exam and history | 1, 0 |
| Durkan compression | 1, 1 |
| Flick sign | 1, 1 |
| Flick: Does shaking alleviate night symptoms? | 1, 1 |
| Gilliat tourniquet | 1, 1 |
| Grip strength | 2, 0 |
| Hypalgesia | 1, 0 |
| Hyperpathia | 1, 0 |
| Lateral pinch strength | 1, 0 |
| Luthy's test | 1, 1 |
| Night symptoms | 1, 1 |
| Opponens pollicus weakness | 1, 1 |
| Phalen's/reverse Phalen's | 3, 2 |
| Right or left hand worse? Or bilateral? | 1, 1 |
| Signs | 1, 0 |
| Symptoms measured systematically | 15, 7 |
| Symptoms | 2, 0 |
| Symptoms and signs | 1, 0 |
| Thenar atrophy | 1, 1 |
| Thenar weakness | 1, 1 |
| Three-point pinch strength | 1, 0 |
| Tinel's | 3, 2 |
| When are symptoms worse? | 1, 1 |
| Which fingers are worst affected? | 1, 1 |
| Legend: | |
|---|---|
| First entry in cell—Total number of articles | |
| Second entry in cell—Number of articles with derivable sensitivity and specificity | |
| Sensory test | Number of articles reporting |
| Current perception | 1, 1 |
| Gap detection test | 1, 1 |
| Semmes-Weinstein monofilament | 1, 0 |
| Tactile discrimination | 1, 1 |
| Vibrometer | 6, 3 |
| Legend: | ||||||
|---|---|---|---|---|---|---|
| Nerve tested: MED-median, RAD-radial, ULN-ulnar | ||||||
| Nerve tested: MOT-motor, SEN-Sensory | ||||||
| Configuration (not applicable to motor nerve tests: OR-orthodromic, AN-antidromic | ||||||
| Stimulation/measurement sites: ELB-elbow, FOR-forearm, WR-wrist, PAL-palm, IN-index finger, MI-middle finger, RI-ring finger, LI-little finger, APB-abductor policis brevis, ADM -abductor digiti minimi, OTH-other | ||||||
| Measured parameter: LAT-latency, AMP-amplitude, VEL-velocity, INCH-inching, OTH-other | ||||||
| Blank cells—characteristic not reported | ||||||
| First entry in cell—Total number of articles | ||||||
| Second entry in cell—Number of articles with derivable sensitivity and specificity | ||||||
| Numeric entries—Total number of articles, articles from which sensitivity and specificity can be calculated | ||||||
| Nerve tested | Configuration | Stimulation site | Measurement site | Parameter measured | Number of articles reporting | |
| MED | MOT | LAT | 2, 0 | |||
| MED | MOT | FOR | APB | LAT | 1, 1 | |
| MED | MOT | WR | APB | LAT | 4, 2 | |
| MED | MOT | WR | APB | VEL | 1, 1 | |
| MED | MOT | WR | OTH | AMP | 1, 0 | |
| MED | MOT | WR | OTH | LAT | 3, 2 | |
| MED | MOT | WR | OTH | VEL | 1, 0 | |
| MED | SEN | AMP | 1, 0 | |||
| MED | SEN | LAT | 4, 0 | |||
| MED | SEN | OTH | 1, 1 | |||
| MED | SEN | AN | LAT | 1, 1 | ||
| MED | SEN | AN | PAL | IN | VEL | 1, 1 |
| MED | SEN | AN | PAL | MI | AMP | 1, 1 |
| MED | SEN | AN | PAL | MI | VEL | 1, 1 |
| MED | SEN | AN | WR | IN | AMP | 2, 2 |
| MED | SEN | AN | WR | IN | LAT | 5, 3 |
| MED | SEN | AN | WR | IN | VEL | 1, 1 |
| MED | SEN | AN | WR | MI | AMP | 1, 1 |
| MED | SEN | AN | WR | MI | INCH | 3, 1 |
| MED | SEN | AN | WR | MI | VEL | 1, 1 |
| MED | SEN | AN | WR | OTH | LAT | 3, 1 |
| MED | SEN | AN | WR | PAL | VEL | 2, 2 |
| MED | SEN | AN | WR | RI | LAT | 1, 1 |
| MED | SEN | OR | IN | WR | LAT | 1, 1 |
| MED | SEN | OR | IN | WR | VEL | 1, 0 |
| MED | SEN | OR | PAL | WR | LAT | 5, 2 |
| MED | SEN | OR | WR | ELB | VEL | 1, 1 |
| ULN | MOT | LAT | 1, 0 | |||
| ULN | MOT | WR | ADM | LAT | 1, 0 | |
| ULN | SEN | LAT | 2, 0 | |||
| ULN | SEN | AN | LAT | 1, 1 | ||
| ULN | SEN | AN | WR | LI | AMP | 2, 2 |
| ULN | SEN | AN | WR | LI | LAT | 4, 2 |
| ULN | SEN | AN | WR | RI | LAT | 1, 1 |
| ULN | SEN | OR | LI | WR | LAT | 1, 0 |
| ULN | SEN | OR | LI | WR | VEL | 1, 0 |
| ULN | SEN | OR | PAL | WR | LAT | 3, 2 |
| Legend: | |||||
|---|---|---|---|---|---|
| Nerves: MED—median, ULN—Ulnar | |||||
| Measured parameter: LAT-latency, VEL-velocity | |||||
| First entry in cell—Total number of articles | |||||
| Second entry in cell—Number of articles with derivable sensitivity and specificity | |||||
| First nerve | Second nerve | Motor or Sensory | Parameter Measured | Combination | Number of articles reporting |
| MED | MED | SEN | VEL | Ratio | 1, 1 |
| MED | ULN | MOT | LAT | Difference | 2, 0 |
| MED | ULN | SEN | LAT | Difference | 11, 6 |
| ULN | MED | SEN | LAT | Difference | 1, 0 |
| Other composite | 7, 3 | ||||
| Legend: | |
|---|---|
| First entry in cell—Total number of articles | |
| Second entry in cell—Number of articles with derivable sensitivity and specificity | |
| Imaging modality | Number of articles reporting |
| CT | 1, 0 |
| MRI | 1, 0 |
| Ultrasound | 1, 1 |
The presence of symptoms and the presence of a positive nerve conduction test appeared to be independent of each other in the screening studies. Figure 14
plots the
prevalence of symptoms on the horizontal axis and the prevalence of
positive nerve conduction tests on the vertical axis. We could only plot
the 15 studies that reported both variables. The correlation between
symptoms and nerve conduction was 0.21 (r2 = 0.04) and was
not statistically different from zero. Because two of the 15 studies
screened a general population, we recomputed the correlation after
removing these two studies. The correlation was 0.16 (r2 =
0.02) and was not statistically different from zero. The weak
association between symptoms and abnormal nerve conduction suggests that
a high incidence of CTS symptoms in workers at risk does not necessarily
imply that those same workers will have a high incidence of abnormal
nerve conduction.
| Article | Method of diagnosis used to determine patient condition |
|---|---|
| Bland, 2000200 | Median and ulnar sensory conduction (velocity?), DML to APB. Sensory potential or segmental study of conduction used if previous tests equivocal. Threshold 2.5 SD from the mean. |
| Kearns, 2000204 | Not reported |
| Atroshi, 1999220 | Two definitions: 1) Symptoms and positive clinical exam. Symptoms were pain, numbness and/or tingling in 2 or more of the first 4 fingers at least twice weekly during the preceding 4 weeks, as stated on a questionnaire. Clinical exam required the presence of nocturnal and/or activity-related numbness and/or tingling involving the palmar aspects of at least 2 of the first 4 fingers. The presence of median nerve sensory and/or motor deficit was supportive of the diagnosis but not necessary. 2) Symptoms and positive clinical exam and positive nerve conduction. Included the same definitions as above, and in addition required a difference of 0.8 ms or more between the median sensory latency (middle finger to wrist) and the ulnar sensory latency (little finger to wrist). |
| Missere, 1999205 | SCV <42.5 m/s as measured by the nerve conduction inching test. |
| Ferry, 1998221 | Not reported |
| Nathan, 1998202 | Symptoms and abnormal nerve conduction. Symptoms defined as positive when the patient has either one of two sets of symptoms: 1) Two or more specific CTS symptoms such as numbness, tingling, nocturnal awakening occurring at least twice per month in the median nerve distribution. 2) One specific CTS symptoms and two or more nonspecific symptoms such as pain, tightness, clumsiness occurring at least twice per month in the median nerve distribution. NCS was defined as abnormal when a patient had any of the following three abnormalities: 1) Maximum latency difference = 0.4 ms in the orthodromic inching test. 2) Antidromic wrist-to-digit sensory latency >3.6 ms. 3) Orthodromic palm to wrist sensory latency >2.2 ms |
| Rosen, 1998201 | Not reported |
| Tan, 1998206 | Not reported |
| Werner, 1998207 | Nerve conduction abnormality defined as a difference >0.5 ms between median and ulnar antidromic sensory latencies to index and little fingers, respectively. Symptom abnormality defined as numbness, tingling, burning, or pain in the wrist, fingers, or hand. |
| Franzblau, 1997208 | Not reported |
| Jeng, 1997209 | Two definitions: One required both symptoms and abnormal conduction, and the other required either symptoms or abnormal nerve conduction :Symptoms: tingling, numbness, pain, perceived weakness, and clumsiness.Nerve conduction was abnormal on any of the following three tests: 1) DML >4.5 ms. 2) Antidromic sensory latency from index finger >3.7 ms. 3) Difference between median palm-to-wrist latency and ulnar palm-to-wrist latency >0.5 ms. |
| Werner, 1997210 | Difference between median and ulnar sensory latency >0.5 ms, and symptoms. |
| Bingham, 1996211 | Not reported |
| Murata, 1996164 | Not reported |
| Pierre-Jerome, 1996212 | Not reported |
| Werner, 1995213 | Symptoms and abnormal NCS. Positive symptoms were defined as any of the following: numbness, tingling, buning, pain, or nocturnal paresthesia in the hand. Abnormal CTS was defined as a difference greater than 0.5 ms between the median and ulnar sensory antidromic latencies. |
| Young, 1995166 | Not reported |
| Franzblau, 1994113 | Symptoms and abnormal nerve conduction. Positive symptoms was defined as having both 1) numbness, tingling, burning, or pain in the fingers, hand, wrist, or forearm and 2) nocturnal occurrence of above symptoms. Abnormal nerve conduction was defined as a difference >0.5 between median sensory antidromic wrist-to-index latency and ipsilateral ulnar sensory antidromic wrist-to-little-finger latency. |
| Kirschberg, 1994214 | Clinical CTS: One or more of the following 7 findings: 1) nocturnal paresthesia of the hand, relieved by shaking; 2) sensory symptoms in the specific distribution of the median nerve; 3) specific median nerve sensory loss; 4) positive Phalen's sign; 5) Positive Tinel's sign; 6) Thenar atrophy; 7) Thenar weakness. Electrodiagnostic CTS (using Mayo Clinic criteria) involved any of the following 4 findings: 1) Median DML >4.6 ms; 2) Median palmar sensory latency >2.2 ms; 3) Difference >0.2 ms between median and ulnar palmar latencies; 4) Difference >1.8 ms between median and ulnar latencies. |
| Nathan, 1994215 | Symptoms and abnormal nerve conduction. Symptoms defined as positive when the patient has either one of two sets of symptoms: 1) Two or more specific CTS symptoms such as numbness, tingling, nocturnal awakening occurring at least twice per month in the median nerve distribution 2) One specific CTS symptom and two or more nonspecific symptoms such as pain, tightness, clumsiness occurring at least twice per month in the median nerve distribution. NCS was defined as abnormal when a patient had any of the following three abnormalities: 1) Maximum latency difference = 0.4 ms in the orthodromic inching test. 2) Antidromic wrist-to-digit sensory latency >3.6 ms 3) Orthodromic palm to wrist sensory latency >2.2 ms |
| Nilsson, 1994216 | Not reported |
| Werner, 1994217 | Not reported |
| Johnson, 1993167 | Not reported |
| Nathan, 1993218 | Symptoms and abnormal nerve conduction. Symptoms defined as positive when the patient has either one of two sets of symptoms: 1) Two or more specific CTS symptoms such as numbness, tingling, nocturnal awakening occurring at least twice per month in the median nerve distribution 2) One specific CTS symptoms and two or more nonspecific symptoms such as pain, tightness, clumsiness occurring at least twice per month in the median nerve distributionNCS was defined as abnormal when a patient had any of the following three abnormalities: 1) Maximum latency difference = 0.4 ms in the orthodromic inching test. 2) Antidromic wrist-to-digit sensory latency >3.6 ms 3) Orthodromic palm to wrist sensory latency >2.2 ms |
| Grant, 1992219 | Median DML >4.5 ms or median DSL >3.5 ms or median-ulnar DML difference >1.2 ms or median-ulnar DSL difference >0.5 ms |
| Jetzer, 1991168 | Symptoms and either positive EMG or recent prior carpal tunnel surgery. |
| DeKrom, 1990222 | Nocturnal paresthesia at least twice a week and either DML >4.5 ms or a difference >0.4 ms between median and ulnar antidromic latencies to the ring finger. |
| Welch, 1973223 | Not reported |
| Article | N | Population | Selection | Followup |
|---|---|---|---|---|
| Kearns, 2000204 | 45 | Porkprocessors | Starting employment | 42–83 days, mean 64 |
| Nathan, 1998202203218 | 283 | Various manufacturing and clerical | Randomly-selected workers | 11 years |
| Werner, 1997210 | NR, though over 700 | Various manufacturing and clerical | NCS positive workers and matched controls | 10 to 24 months |
| Johnson, 1993167 | 184 | Meat processors | Mostly new employees | Not reported, but few followed more than 3 months |
Kearns204 measured nerve conduction in new workers at a pork processing plant. Tests were done before the workers started employment and after two months' employment, though the actual time of the followup test ranged from 42 days to 83 days. Only the nerve conduction tests were done; no symptoms were reported and the authors cautioned that the study was not intended to identify workers who developed CTS. Therefore, this study cannot be used to base conclusions of nerve conduction measurement as a screening test for CTS.
Nathan et al. performed the longest longitudinal study on nerve conduction measurements: 11 years. Two articles202, 203 reported on the same group of subjects: 471 workers from a variety of manufacturing and clerical jobs. Their initial testing was in 1984, with subsequent testing in 1989 (316 subjects followed)203 and 1994-95 (283 subjects)202. Both inching tests and sensory latency measurements were reported in the latest article, though several other nerve conduction tests were also done.
| MLD result | Future CTS | No CTS | Threshold | Sensitivity | Specificity | PPV | NPV |
|---|---|---|---|---|---|---|---|
| <0.28 ms | 3 | 129 | 0.28 ms | 90.9% 76.1% 96.9% | 29.9% 25.7% 34.5% | 9.0% 6.4% 12.7% | 97.7% 93.4% 99.2% |
| 0.28–0.35 ms | 11 | 211 | 0.36 ms | 57.6% 40.5% 73.0% | 78.9% 74.7% 82.5% | 17.3% 11.2% 25.6% | 96.0% 93.4% 97.7% |
| 0.36–0.43 ms | 7 | 56 | 0.44ms | 36.4% 22.0% 53.7% | 91.9% 88.8% 94.1% | 25.5% 15.1% 39.8% | 95.0% 92.4% 96.7% |
| 0.44–0.51 ms | 5 | 20 | 0.52 ms | 21.2% 10.5% 38.1% | 96.5% 94.3% 97.9% | 31.8% 16.1% 53.1% | 94.1% 91.5% 96.0% |
| >0.51 ms | 7 | 15 | |||||
Data from Nathan et al., 1998202
Future CTS—Patients developed CTS during the 11-year followup periof
No CTS—Patients did not develop CTS during followup period.
| Article | SGN | SEN | NCS | CMP | IMG | OTH | Centers | CTS groups | CTS pts. | Neg. groups | Neg. subjects | Prospective or retrospective | Level of reporting | Could sensitivity & specificity be determined? |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Finsen, 2001224 |
| □ | □ |
| □ | □ | Single | 1 | 68 | 0 | 0 | Prospective | Counts | No: thresholds not reported |
| Mondelli, 2001181 | □ | □ |
| □ | □ |
| Single | 1 | 20 | 1 | 19 | Not reported | Counts | Calculated by ECRI |
| Atroshi, 2000225 | □ | □ |
| □ | □ | □ | Single | 1 | 262 | 1 | 125 | Prospective | Summary | No: only summary statistics reported |
| Bland, 2000200 |
| □ | □ | □ | □ |
| Single | 1 | 8223 | 1 | 3533 | Retrospective | Counts | Reported by authors |
| Cuturic, 2000226 | □ | □ |
| □ | □ |
| Single | 1 | 19 | 1 | 16 | Prospective | Patient level | Calculated by ECRI |
| Kearns, 2000204 | □ | □ |
|
| □ | □ | Single | 1 | 45 | 0 | 0 | Prospective | Summary | No: only summary statistics reported |
| Loscher, 2000175 | □ | □ |
|
| □ | □ | Single | 2 | NR | 1 | 87 | Prospective | Counts | Reported by authors |
| Montagna, 2000227 | □ | □ | □ | □ | □ |
| Single | 1 | 30 | 1 | 15 | Not reported | Counts | Reported by authors |
| Nakamichi, 2000228 | □ | □ |
| □ |
| □ | Single | 1 | 125 | 1 | 200 | Not reported | Summary | No: only summary statistics reported |
| Raudino, 2000229 |
| □ | □ | □ | □ | □ | Single | 1 | 83 | 0 | 0 | Not reported | Counts | Reported by authors |
| Resende, 2000184 | □ | □ |
|
| □ | □ | Single | 1 | 32 | 1 | 20 | Not reported | Patient level | Calculated by ECRI |
| Resende, 2000174 | □ | □ |
| □ | □ | □ | Single | 1 | 20 | 1 | 20 | Not reported | Patient level | Calculated by ECRI |
| Sener, 2000186 |
| □ |
| □ | □ |
| Single | 1 | 31 | 1 | 21 | Not reported | Counts | Calculated by ECRI |
| Seror, 2000158 | □ | □ |
|
| □ | □ | Single | 1 | 20 | 1 | 20 | Not reported | Counts | Reported by authors |
| Stalberg, 2000230 | □ | □ | □ |
| □ | □ | Single | 1 | 136 | 1 | 32 | Not reported | Counts | Reported by authors |
| Weber, 2000108 |
|
| □ |
| □ | □ | Single | 1 | 53 | 1 | 26 | Not reported | Counts | Reported by authors |
| Atroshi, 1999220 |
| □ |
|
| □ | □ | Single | 1 | 2466 | 0 | 0 | Prospective | Counts | No: only one patient group |
| Burke, 1999231 |
| □ |
|
| □ | □ | Multiple (<5) | 1 | 186 | 0 | 0 | Prospective | Counts | Calculated by ECRI |
| Duncan, 1999232 | □ | □ | □ | □ |
| □ | Single | 1 | 68 | 1 | 36 | Prospective | Counts | Reported by authors |
| Kabiraj, 1999233 | □ | □ |
|
| □ | □ | Single | 1 | 31 | 1 | 38 | Not reported | Counts | Calculated by ECRI |
| Lee, 1999234 | □ | □ | □ | □ |
| □ | Single | 1 | 50 | 1 | 28 | Prospective | Counts | Reported by authors |
| Missere, 1999205 | □ | □ |
| □ |
| □ | Single | 1 | 45 | 0 | 0 | Not reported | Counts | Reported by authors |
| Mongale, 1999235 | □ | □ | □ | □ |
| □ | Single | 1 | 8 | 2 | 16 | Not reported | Summary | No: only summary statistics reported |
| Murthy, 1999143 | □ | □ |
|
| □ |
| Single | 1 | 84 | 1 | 37 | Not reported | Counts | Reported by authors |
| Rudolfer, 1999236 | □ | □ | □ |
| □ | □ | Single | 1 | 937 | 0 | 0 | Retrospective | Counts | Calculated by ECRI |
| Sander, 1999237 | □ | □ |
|
| □ |
| Single | 1 | 59 | 1 | 34 | Prospective | Counts | Reported by authors |
| Simovic, 1999183 | □ | □ |
|
| □ | □ | Single | 2 | 66 | 1 | 19 | Prospective | Counts | Reported by authors |
| Szabo, 1999152 |
|
|
| □ | □ |
| Single | 1 | 50 | 2 | 100 | Prospective | Counts | Reported by authors |
| Thonnard, 1999117 | □ |
|
|
| □ |
| Single | 1 | 11 | 1 | 10 | Prospective | Summary | No: only summary statistics reported |
| Wang, 1999238 | □ | □ |
|
| □ | □ | Single | 1 | 12 | 1 | 12 | Prospective | Summary | No: only summary statistics reported |
| Aurora, 1998239 | □ | □ |
| □ | □ | □ | Single | 1 | 19 | 1 | 20 | Not reported | Summary | No: only summary statistics reported |
| Ferry, 1998221 |
| □ |
|
| □ | □ | Single | 1 | 648 | 0 | 0 | Prospective | Counts | Reported by authors |
| Fertl, 1998153 |
| □ |
| □ | □ |
| Single | 1 | 47 | 1 | 20 | Prospective | Counts | Reported by authors |
| Gerr, 199831 |
|
|
|
| □ |
| Single | 1 | 60 | 1 | 59 | Not reported | Counts | Reported by authors |
| Ghavanini, 1998154 |
|
|
| □ | □ | □ | Single | 1 | 74 | 1 | 58 | Prospective | Counts | Reported by authors |
| Girlanda, 1998149 |
| □ |
|
| □ | □ | Single | 1 | 41 | 1 | 45 | Not reported | Counts | Reported by authors |
| Kabiraj, 1998240 | □ | □ |
|
| □ | □ | Single | 1 | 72 | 1 | 65 | Retrospective | Summary | No: only summary statistics reported |
| Kleindienst, 1998241 | □ | □ | □ | □ |
| □ | Single | 1 | 77 | 1 | 18 | Prospective | Summary | No: only summary statistics reported |
| Luchetti, 1998242 | □ | □ | □ | □ | □ |
| Single | 1 | 39 | 1 | 12 | Not reported | Summary | No: only summary statistics reported |
| Nathan, 1998202 |
| □ |
|
| □ | □ | Single | 1 | 283 | 0 | 0 | Prospective | Counts | No: only one patient group |
| Rosen, 1998201 | □ |
| □ | □ | □ | □ | Single | 2 | 34 | 1 | 60 | Prospective | Counts | Reported by authors |
| Scelsa, 1998243 | □ | □ |
| □ | □ | □ | Single | 2 | 63 | 1 | 25 | Prospective | Counts | Reported by authors |
| Seror, 1998159 | □ | □ |
| □ | □ | □ | Single | 1 | 85 | 1 | 80 | Not reported | Counts | Reported by authors |
| Smith, 1998244 | □ | □ |
| □ | □ | □ | Single | 1 | 82 | 0 | 0 | Prospective | Counts | Calculated by ECRI |
| Tan, 1998206 | □ | □ |
| □ |
| □ | Single | 1 | 64 | 1 | 56 | Not reported | Summary | No: only summary statistics reported |
| Terzis, 1998162 | □ | □ |
| □ | □ | □ | Single | 1 | 72 | 1 | 43 | Not reported | Counts | Reported by authors |
| Tetro, 1998102 |
|
|
| □ | □ |
| Single | 1 | 64 | 1 | 50 | Prospective | Counts | Reported by authors |
| Werner, 1998207 |
| □ |
|
| □ | □ | Multiple (>5) | 1 | 727 | 0 | 0 | Prospective | Counts | No: only one patient group |
| Wilson, 1998245 | □ | □ |
|
| □ | □ | Single | 1 | 23 | 1 | 14 | Not reported | Summary | No: only summary statistics reported |
| Bak, 1997246 | □ | □ | □ |
|
| □ | Single | 1 | 20 | 0 | 0 | Prospective | Counts | No: no control group |
| Brahme, 1997199 |
| □ | □ | □ |
| □ | Single | 1 | 20 | 1 | 15 | Prospective | Counts | Reported by authors |
| Bronson, 1997163 | □ | □ |
|
| □ | □ | Single | 1 | 22 | 1 | 16 | Prospective | Patient level | Calculated by ECRI |
| Del Pino, 1997104 |
| □ | □ | □ | □ | □ | Single | 1 | 180 | 1 | 100 | Prospective | Counts | Reported by authors |
| Dellon, 1997107 |
|
| □ | □ | □ | □ | Single | 1 | 72 | 2 | 94 | Not reported | Counts | No: inconsistent thresholds |
| Franzblau, 1997208 |
| □ | □ | □ | □ | □ | Single | 1 | 148 | 0 | 0 | Prospective | Summary | No: only summary statistics reported |
| Guglielmo, 1997247 | □ | □ |
|
| □ | □ | Single | 1 | 198 | 1 | 69 | Prospective | Summary | No: only summary statistics reported |
| Gunnarsson, 1997248 |
| □ |
|
| □ | □ | Single | 1 | 100 | 0 | 0 | Prospective | Counts | Reported by authors |
| Horch, 1997249 | □ | □ | □ | □ |
| □ | Single | 1 | 19 | 1 | 17 | Not reported | Summary | No: only summary statistics reported |
| Jeng, 1997209 |
|
|
|
| □ |
| Single | 1 | 27 | 0 | 0 | Prospective | Counts | Reported by authors |
| Kaneko, 1997250 | □ | □ |
|
| □ | □ | Single | 1 | 15 | 3 | 66 | Not reported | Summary | No: only summary statistics reported |
| King, 1997114 | □ |
| □ | □ | □ | □ | Single | 1 | 29 | 1 | 100 | Not reported | Summary | No: only summary statistics reported |
| Pierre-Jerome, 1997251 | □ | □ |
| □ |
| □ | Single | 1 | 27 | 1 | 28 | Prospective | Summary | No: only summary statistics reported |
| Radack, 1997252 | □ | □ | □ | □ |
| □ | Single | 1 | 161 | 1 | NR | Retrospective | Counts | Reported by authors |
| Rosecrance, 1997253 | □ | □ | □ |
| □ |
| Single | 1 | 28 | 1 | 25 | Not reported | Summary | No: only summary statistics reported |
| Simovic, 1997182 | □ | □ |
|
| □ | □ | Single | 1 | 107 | 1 | 15 | Retrospective | Counts | Reported by authors |
| Werner, 1997210 | □ | □ |
|
| □ | □ | Single | 2 | 108 | 0 | 0 | Retrospective | Counts | No: incomplete reporting |
| Andary, 1996196 | □ | □ |
|
| □ | □ | Single | 1 | 81 | 1 | 17 | Prospective | Counts | Reported by authors |
| Atroshi, 1996136 | □ | □ |
| □ | □ | □ | Single | 1 | 36 | 2 | 60 | Prospective | Counts | Reported by authors |
| Bingham, 1996211 |
| □ |
|
| □ | □ | Single | 1 | 1021 | 0 | 0 | Prospective | Counts | No: only one patient group |
| Checkosky, 1996254 | □ |
| □ | □ | □ | □ | Single | 1 | 24 | 1 | 20 | Not reported | Patient level | Reported by authors |
| Cherniak, 1996190 | □ |
|
|
| □ |
| Single | 1 | 49 | 1 | 10 | Not reported | Counts | Reported by authors |
| Foresti, 1996192 | □ | □ |
|
| □ | □ | Single | 1 | 100 | 1 | 25 | Prospective | Counts | Reported by authors |
| Ghavanini, 1996255 | □ | □ |
|
| □ | □ | Single | 1 | 50 | 1 | 50 | Not reported | Summary | No: only summary statistics reported |
| Kleindienst, 1996256 | □ | □ | □ | □ |
| □ | Single | 1 | 55 | 1 | 18 | Not reported | Counts | Reported by authors |
| Murata, 1996164 |
| □ |
|
| □ | □ | Single | 1 | 27 | 1 | 19 | Not reported | Counts | Calculated by ECRI |
| Padua, 1996165 | □ | □ |
|
| □ |
| Single | 1 | 43 | 1 | 36 | Not reported | Counts | Reported by authors |
| Pierre-Jerome, 1996212 |
| □ |
| □ |
| □ | Single | 1 | 24 | 1 | 19 | Prospective | Summary | No: only summary statistics reported |
| Britz, 1995257 |
| □ |
|
|
|
| Single | 1 | 32 | 1 | 5 | Prospective | Patient level | No: results not reported for controls |
| De Smet, 1995101 |
| □ | □ | □ | □ | □ | Single | 2 | 50 | 2 | 55 | Not reported | Counts | Reported by authors |
| Gerr, 1995118 | □ |
| □ | □ | □ | □ | Single | 2 | 60 | 1 | 59 | Not reported | Counts | Reported by authors |
| Glass, 199528 |
| □ |
| □ | □ | □ | Single | 1 | 82 | 1 | 24 | Not reported | Counts | Calculated by ECRI |
| Golovchinsky, 1995258 | □ | □ |
|
| □ |
| Single | 1 | 571 | 0 | 0 | Retrospective | Counts | Reported by authors |
| Hamanaka, 1995259 | □ | □ |
|
| □ |
| Single | 2 | 647 | 1 | 31 | Retrospective | Counts | Calculated by ECRI |
| Hansson, 1995137 | □ | □ |
|
| □ | □ | Single | 2 | 30 | 1 | 10 | Not reported | Counts | Reported by authors |
| Kothari, 1995260 | □ | □ |
| □ | □ | □ | Single | 1 | 59 | 1 | 30 | Not reported | Summary | No: only summary statistics reported |
| Lang, 1995109 |
|
|
| □ | □ |
| Single | 1 | 23 | 1 | 16 | Prospective | Counts | Reported by authors |
| Lesser, 1995261 | □ | □ |
|
| □ |
| Single | 1 | 45 | 1 | 20 | Not reported | Counts | Reported by authors |
| Nakamichi, 1995262 | □ | □ | □ | □ |
| □ | Single | 1 | 15 | 1 | 15 | Not reported | Patient level | Calculated by ECRI |
| Seradge, 1995263 | □ | □ | □ | □ | □ |
| Single | 1 | 72 | 1 | 21 | Not reported | Summary | No: only summary statistics reported |
| Seror, 1995179 | □ | □ |
|
| □ | □ | Single | 3 | 75 | 1 | 40 | Not reported | Counts | Reported by authors |
| Shafshak, 1995264 | □ | □ |
| □ | □ |
| Single | 2 | 36 | 2 | 36 | Not reported | Counts | No: no diagnostic results reported |
| Sheean, 1995191 | □ | □ |
|
| □ | □ | Single | 1 | 49 | 1 | NR | Not reported | Counts | Calculated by ECRI |
| Tassler, 1995115 | □ |
| □ |
| □ | □ | Single | 1 | 14 | 1 | 13 | Retrospective | Counts | Reported by authors |
| Valls-Sole, 1995265 | □ | □ |
|
| □ | □ | Single | 1 | 18 | 1 | 15 | Prospective | Summary | No: only summary statistics reported |
| Werner, 1995213 |
|
|
|
| □ | □ | Single | 1 | 167 | 0 | 0 | Not reported | Counts | Reported by authors |
| Young, 1995166 |
|
|
| □ | □ | □ | Single | 1 | 157 | 0 | 0 | Prospective | Counts | No: only one patient group |
| Clifford, 1994266 | □ | □ |
|
| □ | □ | Single | 1 | 20 | 1 | 10 | Not reported | Summary | No: only summary statistics reported |
| Durkan, 1994267 |
| □ | □ | □ | □ | □ | Single | 1 | 30 | 1 | 25 | Not reported | Counts | Calculated by ECRI |
| Franzblau, 1994113 |
|
|
|
| □ | □ | Single | 1 | 83 | 0 | 0 | Prospective | Counts | Reported by authors |
| Gerr, 1994197 |
|
| □ | □ | □ | □ | Single | 2 | NR | 1 | NR | Not reported | Counts | Reported by authors |
| Kirschberg, 1994214 |
| □ |
|
| □ | □ | Single | 1 | 112 | 0 | 0 | Retrospective | Counts | No: only one patient group |
| Kuntzer, 1994144 | □ | □ |
|
| □ |
| Single | 1 | 100 | 1 | 70 | Prospective | Counts | Reported by authors |
| Nathan, 1994215 |
| □ |
|
| □ | □ | Multiple (<5) | 2 | 417 | 0 | 0 | Retrospective | Counts | No: no control subjects |
| Nilsson, 1994216 | □ | □ |
| □ | □ | □ | Single | 3 | 175 | 0 | 0 | Prospective | Counts | Reported by authors |
| Para, 1994103 |
| □ |
|
| □ |
| Single | 2 | 51 | 1 | 12 | Not reported | Counts | Reported by authors |
| Rossi, 1994178 | □ | □ |
|
| □ | □ | Single | 1 | 62 | 1 | 27 | Not reported | Counts | Reported by authors |
| Werner, 1994217 |
|
|
|
| □ | □ | Single | 1 | 130 | 0 | 0 | Prospective | Counts | Calculated by ECRI |
| Werner, 1994111 |
| □ |
|
| □ | □ | Single | 1 | 31 | 1 | 20 | Not reported | Counts | Calculated by ECRI |
| Eisen, 1993193 | □ | □ |
|
| □ | □ | Single | 1 | NR | 1 | NR | Not reported | Counts | Reported by authors |
| Johnson, 1993167 |
| □ |
| □ | □ | □ | Single | 1 | 184 | 0 | 0 | Prospective | Summary | No: only summary statistics reported |
| Nakamichi, 1993268 | □ | □ | □ | □ |
| □ | Single | 1 | 128 | 0 | 0 | Not reported | Counts | No: only one patient group |
| Nathan, 1993218 |
| □ |
|
| □ | □ | Single | 2 | 1125 | 1 | 45 | Prospective | Counts | Reported by authors |
| Rodriquez, 1993269 | □ | □ |
| □ | □ |
| Single | 1 | 10 | 1 | 8 | Prospective | Patient level | Calculated by ECRI |
| Rosen, 1993270 | □ |
|
| □ | □ | □ | Single | 2 | 62 | 2 | 71 | Not reported | Counts | Calculated by ECRI |
| Rosén, 1993138 | □ | □ |
|
| □ | □ | Single | 1 | 28 | 3 | 86 | Not reported | Counts | Calculated by ECRI |
| Uncini, 1993160 | □ | □ |
|
| □ | □ | Single | 1 | 70 | 1 | 47 | Not reported | Counts | Reported by authors |
| Buchberger, 1992271 | □ | □ | □ | □ |
| □ | Multiple (<5) | 1 | 18 | 1 | NR | Not reported | Counts | Reported by authors |
| Grant, 1992219 | □ |
|
| □ | □ | □ | Single | 1 | 22 | 1 | 47 | Not reported | Counts | Calculated by ECRI |
| Imaoka, 1992272 | □ | □ |
| □ | □ | □ | Single | 1 | 42 | 1 | 32 | Not reported | Counts | Calculated by ECRI |
| Kindstrand, 1992273 | □ | □ | □ | □ | □ |
| Single | 1 | 94 | 1 | 127 | Prospective | Patient level | Calculated by ECRI |
| Preston, 1992188 | □ | □ |
|
| □ | □ | Single | 1 | 8 | 1 | NR | Not reported | Counts | Calculated by ECRI |
| Tchou, 1992274 | □ | □ | □ | □ | □ |
| Single | 1 | 61 | 1 | 40 | Not reported | Patient level | Reported by authors |
| Buchberger, 1991275 | □ | □ | □ | □ |
| □ | Single | 1 | 25 | 1 | 14 | Not reported | Summary | No: only summary statistics reported |
| Chang, 1991145 | □ | □ |
|
| □ | □ | Single | 1 | 43 | 1 | 40 | Not reported | Counts | Calculated by ECRI |
| Durkan, 1991155 |
| □ | □ | □ | □ | □ | Single | 1 | 31 | 1 | 50 | Not reported | Counts | Reported by authors |
| Jetzer, 1991168 |
|
| □ | □ | □ |
| Single | 3 | 323 | 1 | 284 | Prospective | Counts | No: no control subjects |
| Katz, 1991276 |
|
|
|
| □ |
| Single | 1 | 78 | 0 | 0 | Not reported | Counts | Reported by authors |
| Lauritzen, 1991185 | □ | □ |
|
| □ |
| Single | 1 | 38 | 1 | 23 | Not reported | Counts | Calculated by ECRI |
| Luchetti, 1991169 |
| □ |
|
| □ | □ | Single | 1 | 14 | 0 | 0 | Retrospective | Patient level | No: only one patient group |
| Radwin, 1991116 | □ |
| □ | □ | □ | □ | Single | 1 | 12 | 1 | 15 | Not reported | Patient level | No: no diagnostic threshols used |
| Charles, 1990170 | □ | □ |
|
| □ | □ | Single | 1 | 158 | 2 | 90 | Not reported | Counts | Reported by authors |
| De Krom, 1990222 |
| □ |
|
| □ | □ | Single | 1 | 50 | 0 | 0 | Prospective | Counts | Calculated by ECRI |
| Fitz, 1990277 | □ | □ |
|
| □ | □ | Single | 1 | 36 | 1 | 44 | Not reported | Counts | Calculated by ECRI |
| Gilliatt, 1990278 | □ | □ |
|
| □ | □ | Single | 1 | 10 | 1 | 15 | Not reported | Counts | Calculated by ECRI |
| MacDonell, 199090 | □ | □ |
| □ | □ | □ | Single | 1 | 34 | 1 | 12 | Not reported | Counts | Reported by authors |
| Merchut, 1990279 | □ |
|
| □ | □ | □ | Single | 1 | 23 | 1 | 54 | Not reported | Counts | Reported by authors |
| Palliyath, 1990171 | □ | □ |
| □ | □ | □ | Single | 1 | 10 | 1 | 11 | Not reported | Summary | No: only summary statistics reported |
| Pease, 1990177 | □ | □ |
| □ | □ | □ | Single | 1 | 21 | 1 | 16 | Not reported | Counts | Calculated by ECRI |
| Rojviroj, 1990280 | □ |
| □ | □ | □ | □ | Single | 1 | 33 | 1 | 16 | Prospective | Counts | Reported by authors |
| Tzeng, 1990180 | □ | □ |
| □ | □ | □ | Single | 1 | 84 | 1 | 50 | Not reported | Counts | Calculated by ECRI |
| Uncini, 1990135 | □ | □ |
|
| □ | □ | Single | 1 | 35 | 1 | 39 | Not reported | Summary | No: only summary statistics reported |
| Winn, 1990281 | □ |
| □ | □ | □ | □ | Single | 2 | 61 | 0 | 0 | Prospective | Summary | No: only summary statistics reported |
| Braun, 1989282 |
|
| □ | □ | □ | □ | Single | 1 | 40 | 0 | 0 | Not reported | Counts | No: no diagnostic thresholds reported |
| Cioni, 1989146 | □ | □ |
|
| □ | □ | Single | 1 | 307 | 1 | 54 | Not reported | Counts | Reported by authors |
| Jackson, 1989150 | □ | □ |
|
| □ |
| Single | 1 | 123 | 1 | 38 | Not reported | Counts | Reported by authors |
| Meyers, 1989283 | □ | □ | □ | □ | □ |
| Single | 1 | 14 | 1 | 19 | Not reported | Counts | Calculated by ECRI |
| So, 1989173 | □ | □ |
| □ | □ |
| Single | 1 | 22 | 2 | 35 | Not reported | Counts | Reported by authors |
| Szabo, 1989284 | □ |
| □ | □ | □ | □ | Single | 1 | 22 | 0 | 0 | Not reported | Summary | No: only summary statistics reported |
| Uncini, 1989161 | □ | □ |
|
| □ | □ | Single | 1 | 32 | 1 | 33 | Not reported | Summary | No: only summary statistics reported |
| De Léan, 1988285 | □ | □ |
| □ | □ | □ | Single | 1 | 150 | 0 | 0 | Not reported | Counts | Calculated by ECRI |
| Koris, 1988198 |
|
| □ | □ | □ | □ | Single | 1 | 21 | 1 | 15 | Prospective | Counts | Reported by authors |
| Molitor, 1988110 |
| □ | □ | □ | □ |
| Single | 1 | 19 | 1 | NR | Not reported | Counts | Calculated by ECRI |
| Mortier, 1988286 | □ | □ |
|
| □ | □ | Single | 1 | 116 | 1 | 102 | Retrospective | Counts | Reported by authors |
| Pease, 1988287 | □ | □ |
|
| □ | □ | Single | 1 | 25 | 1 | 23 | Not reported | Summary | No: only summary statistics reported |
| Carroll, 1987288 | □ | □ |
|
| □ | □ | Single | 1 | 101 | 1 | 50 | Not reported | Counts | Reported by authors |
| Jessurun, 1987289 | □ | □ | □ | □ |
| □ | Multiple (<5) | 1 | 24 | 1 | 10 | Not reported | Summary | No: only summary statistics reported |
| Johnson, 1987290 | □ | □ |
|
| □ | □ | Single | 1 | 20 | 1 | 78 | Not reported | Counts | Calculated by ECRI |
| Liang, 1987291 | □ | □ | □ | □ |
| □ | Single | 1 | 68 | 2 | 139 | Not reported | Summary | No: only summary statistics reported |
| Macleod, 1987292 |
| □ |
| □ | □ | □ | Single | 1 | 111 | 1 | 125 | Not reported | Summary | No: only summary statistics reported |
| Seror, 1987156 |
| □ | □ | □ | □ | □ | Single | 1 | 62 | 1 | 20 | Not reported | Counts | Reported by authors |
| Borg, 1986293 |
|
|
| □ | □ | □ | Single | 1 | 22 | 0 | 0 | Not reported | Counts | Calculated by ECRI |
| Gellman, 1986106 |
| □ |
|
| □ | □ | Single | 1 | NR | 2 | NR | Not reported | Counts | Reported by authors |
| Escobar, 1985151 | □ | □ |
| □ | □ | □ | Single | 1 | 23 | 1 | 55 | Not reported | Counts | Calculated by ECRI |
| Kimura, 1985189 | □ | □ |
|
| □ |
| Single | 1 | 438 | 1 | 148 | Not reported | Counts | Reported by authors |
| Mills, 1985194 | □ | □ |
|
| □ | □ | Single | 1 | 47 | 2 | 49 | Not reported | Counts | Calculated by ECRI |
| Borg, 1984294 | □ |
| □ | □ | □ | □ | Single | 3 | 45 | 0 | 0 | Prospective | Patient level | Calculated by ECRI |
| Pryse-Phillips, 1984105 |
| □ | □ | □ | □ | □ | Single | 1 | 212 | 4 | 184 | Retrospective | Counts | Reported by authors |
| Satoh, 1984295 |
| □ |
| □ | □ | □ | Single | 1 | 14 | 0 | 0 | Retrospective | Patient level | No: only one patient group |
| Szabo, 198430 |
|
| □ | □ | □ | □ | Single | 1 | 20 | 0 | 0 | Prospective | Counts | No: only one patient group |
| Goddard, 1983296 | □ | □ |
| □ | □ | □ | Single | 1 | 24 | 1 | 49 | Not reported | Counts | Calculated by ECRI |
| Kim, 1983195 | □ | □ |
|
| □ | □ | Single | 1 | 39 | 1 | 33 | Not reported | Counts | Reported by authors |
| Marin, 1983139 | □ | □ |
|
| □ | □ | Single | 1 | 14 | 1 | 12 | Not reported | Counts | Calculated by ECRI |
| Wongsam, 1983172 | □ | □ |
| □ | □ | □ | Single | 1 | 15 | 2 | 56 | Not reported | Summary | No: only summary statistics reported |
| Johnson, 1981297 | □ | □ |
|
| □ | □ | Single | 1 | 18 | 1 | 37 | Not reported | Summary | No: only summary statistics reported |
| Dekel, 198021 | □ | □ | □ | □ |
| □ | Single | 1 | 26 | 1 | 33 | Prospective | Patient level | No: could not extract 2 × 2 counts from graph |
| Messina, 1980120 | □ | □ |
|
| □ | □ | Single | 1 | 40 | 1 | 40 | Not reported | Counts | Reported by authors |
| Gelmers, 197929 |
| □ |
|
| □ | □ | Single | 1 | 47 | 1 | 43 | Not reported | Counts | Reported by authors |
| Kimura, 1979140 | □ | □ |
|
| □ | □ | Single | 1 | 105 | 1 | 61 | Not reported | Counts | Calculated by ECRI |
| Schwartz, 1979187 | □ | □ |
|
| □ | □ | Single | 1 | 20 | 1 | 10 | Not reported | Counts | Calculated by ECRI |
| Stewart, 1978157 |
| □ | □ | □ | □ | □ | Single | 1 | 37 | 1 | 38 | Not reported | Counts | Reported by authors |
| Eisen, 1977298 | □ | □ |
|
| □ | □ | Single | 1 | 30 | 3 | 101 | Not reported | Patient level | Calculated by ECRI |
| Sedal, 1973299 | □ | □ |
| □ | □ | □ | Single | 1 | 214 | 1 | 34 | Retrospective | Counts | Reported by authors |
| Welch, 1973223 | □ |
| □ | □ | □ | □ | Single | 1 | 428 | 1 | 111 | Not reported | Summary | No: only summary statistics reported |
| Casey, 1972300 | □ | □ |
| □ | □ | □ | Single | 1 | 16 | 2 | 112 | Not reported | Patient level | Calculated by ECRI |
| Loong, 1972141 | □ | □ |
|
| □ | □ | Single | 1 | 18 | 1 | 30 | Not reported | Patient level | Calculated by ECRI |
| Melvin, 1972147 | □ | □ |
| □ | □ | □ | Single | 1 | 17 | 1 | 24 | Not reported | Counts | Calculated by ECRI |
| Buchthal, 1971301 | □ | □ |
| □ | □ |
| Single | 1 | 22 | 1 | 10 | Not reported | Counts | Calculated by ECRI |
| Loong, 1971148 | □ | □ |
|
| □ | □ | Single | 1 | 15 | 1 | 30 | Not reported | Patient level | Calculated by ECRI |
| Plaja, 1971142 | □ | □ |
| □ | □ |
| Single | 1 | 56 | 1 | 20 | Retrospective | Counts | Reported by authors |
| Article | Disorder type | Patient selection | N patients | % female | Mean age | Age of youngest | Age of oldest | Duration of condition before treatment (months) | Shortest duration (months) | Longest duration (months) | Are patient comorbidities reported? |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Finsen, 2001224 | CTS | Unspecified diagnosis | 68 | 74 | 48 | 21 | 86 | Yes | |||
| Mondelli, 2001181 | Normal | Healthy volunteers | 19 | NR | 51.9 | 31 | 72 | No | |||
| Mondelli, 2001181 | CTS | Unspecified diagnosis | 20 | 80 | 52.8 | 35 | 75 | No | |||
| Atroshi, 2000225 | CTS | Symptoms/ presented | 262 | 57 | 52 | No | |||||
| Atroshi, 2000225 | Normal | Healthy volunteers | 125 | 55 | 51 | No | |||||
| Bland, 2000200 | CTS | Complex objective standard | 4690 | 65 | 57 | No | |||||
| Bland, 2000200 | CTS | Symptoms/ presented | 8223 | 66 | 53 | 10 | 98 | No | |||
| Bland, 2000200 | Normal | Other | 3533 | 67 | 49 | No | |||||
| Cuturic, 2000226 | CTS | Unspecified diagnosis | 19 | 0 | 43 | 29 | 62 | No | |||
| Cuturic, 2000226 | Normal | Healthy volunteers | 16 | 0 | 41 | 26 | 58 | No | |||
| Kearns, 2000204 | CTS | Workers at risk | 45 | 4 | Yes | ||||||
| Loscher, 2000175 | Normal | Healthy volunteers | 87 | NR | 47 | 15 | 86 | No | |||
| Loscher, 2000175 | CTS | Unspecified diagnosis | NR | No | |||||||
| Loscher, 2000175 | CTS | Other | NR | No | |||||||
| Montagna, 2000227 | Cubital tunnel syndrome | Unspecified diagnosis | 10 | NR | No | ||||||
| Montagna, 2000227 | Normal | Healthy volunteers | 15 | NR | No | ||||||
| Montagna, 2000227 | CTS | Unspecified diagnosis | 30 | NR | No | ||||||
| Nakamichi, 2000228 | CTS | Simple nerve conduction | 125 | 100 | 56 | 40 | 70 | No | |||
| Nakamichi, 2000228 | Normal | Healthy volunteers | 200 | NR | 57 | 40 | 70 | No | |||
| Raudino, 2000229 | CTS | Complex objective standard | 83 | 82 | 48.9 | 19 | 82 | 26.9 | 1 | 180 | Yes |
| Resende, 2000174 | CTS | Unspecified diagnosis | 20 | NR | No | ||||||
| Resende, 2000174 | Normal | Healthy volunteers | 20 | NR | 21 | 55 | No | ||||
| Resende, 2000184 | Normal | Healthy volunteers | 20 | 100 | 36 | 20 | 54 | No | |||
| Resende, 2000184 | CTS | Unspecified diagnosis | 32 | 100 | 44 | 25 | 59 | No | |||
| Sener, 2000186 | CTS | Symptoms/ presented | 31 | NR | 46 | 26 | 70 | Yes | |||
| Sener, 2000186 | Normal | Healthy volunteers | 21 | NR | 38 | 18 | 60 | Yes | |||
| Seror, 2000158 | Normal | Healthy volunteers | 20 | 75 | 43 | 20 | 67 | No | |||
| Seror, 2000158 | CTS | Complex objective standard | 20 | 75 | 47 | 32 | 76 | No | |||
| Stalberg, 2000230 | CTS | Symptoms/ presented | 136 | NR | No | ||||||
| Stalberg, 2000230 | Normal | Healthy volunteers | 32 | NR | 21 | 62 | No | ||||
| Weber, 2000108 | CTS | Symptoms/ presented | 53 | 79 | 45 | No | |||||
| Weber, 2000108 | Normal | Healthy volunteers | 26 | 85 | 37 | No | |||||
| Burke, 1999231 | CTS | Symptoms/ presented | 186 | NR | No | ||||||
| Atroshi, 1999220 | Normal | Other | 2466 | NR | No | ||||||
| Duncan, 1999232 | CTS | Complex objective standard | 68 | 74 | 54 | Yes | |||||
| Duncan, 1999232 | CTS | Complex objective standard | NR | Yes | |||||||
| Duncan, 1999232 | Normal | Healthy volunteers | 36 | 64 | 44 | Yes | |||||
| Kabiraj, 1999233 | Normal | Healthy volunteers | 38 | 50 | 20 | 79 | No | ||||
| Kabiraj, 1999233 | CTS | Complex objective standard | 31 | 68 | 28 | 85 | No | ||||
| Lee, 1999234 | Normal | Healthy volunteers | 28 | 54 | 22 | 47 | No | ||||
| Lee, 1999234 | CTS | Unspecified diagnosis | 50 | 74 | 32 | 81 | No | ||||
| Missere, 1999205 | CTS | Workers at risk | 45 | 0 | 37.7 | No | |||||
| Mongale, 1999235 | Normal | Healthy volunteers | 9 | 100 | 39 | 26 | 50 | No | |||
| Mongale, 1999235 | Normal | Healthy volunteers | 7 | 0 | 39 | 27 | 58 | No | |||
| Mongale, 1999235 | CTS | Unspecified diagnosis | 8 | 100 | 43 | 24 | 54 | No | |||
| Murthy, 1999143 | CTS | Symptoms/ presented | 84 | NR | No | ||||||
| Murthy, 1999143 | Normal | Healthy volunteers | 37 | NR | No | ||||||
| Rudolfer, 1999236 | CTS | Symptoms/ presented | 937 | NR | No | ||||||
| Sander, 1999237 | Normal | Healthy volunteers | 34 | NR | 41 | 26 | 71 | No | |||
| Sander, 1999237 | CTS | Complex objective standard | 59 | NR | 49 | 29 | 73 | No | |||
| Simovic, 1999183 | CTS | Other | 12 | NR | Yes | ||||||
| Simovic, 1999183 | Normal | Healthy volunteers | 19 | 63 | 40 | 25 | 68 | Yes | |||
| Simovic, 1999183 | CTS | Unspecified diagnosis | 54 | NR | Yes | ||||||
| Szabo, 1999152 | Normal | Healthy volunteers | 50 | 66 | 18 | 59 | No | ||||
| Szabo, 1999152 | CTS | Complex objective standard | 50 | 76 | 20 | 73 | 2 | 240 | No | ||
| Szabo, 1999152 | Unrelated disease | Other | 50 | 80 | 28 | 72 | 0 | 180 | No | ||
| Thonnard, 1999117 | CTS | Unspecified diagnosis | 11 | 73 | 52 | No | |||||
| Thonnard, 1999117 | Normal | Healthy volunteers | 11 | 73 | 53 | No | |||||
| Wang, 1999238 | CTS | Complex objective standard | 12 | 92 | 46 | 30 | 65 | No | |||
| Wang, 1999238 | Normal | Healthy volunteers | 12 | 42 | 37 | 28 | 59 | No | |||
| Aurora, 1998239 | CTS | Symptoms/ presented | 19 | NR | 52.8 | No | |||||
| Aurora, 1998239 | Normal | Healthy volunteers | 20 | NR | 32.9 | No | |||||
| Ferry, 1998221 | Normal | Other | 648 | 56 | 46.9 | No | |||||
| Fertl, 1998153 | Normal | Healthy volunteers | 20 | 60 | 42 | 25 | 77 | No | |||
| Fertl, 1998153 | CTS | Symptoms/ presented | 47 | 83 | 55.5 | 21 | 78 | No | |||
| Gerr, 199831 | Normal | Healthy volunteers | 59 | 69 | 38.2 | No | |||||
| Gerr, 199831 | CTS | Symptoms/ presented | 60 | 72 | 46.6 | No | |||||
| Ghavanini, 1998154 | CTS | Complex objective standard | 26 | 100 | 37 | 20 | 50 | 9 | 1 | 36 | No |
| Ghavanini, 1998154 | CTS | Symptoms/ presented | 74 | 81 | 40 | 20 | 50 | 15 | 1 | 60 | No |
| Ghavanini, 1998154 | Normal | Healthy volunteers | 58 | 76 | 36.7 | 20 | 50 | No | |||
| Ghavanini, 1998154 | CTS | Complex objective standard | 26 | 69 | 41 | 20 | 50 | 19.4 | 1 | 48 | No |
| Ghavanini, 1998154 | CTS | Complex objective standard | 22 | 73 | 42 | 30 | 50 | 19 | 4 | 60 | No |
| Girlanda, 1998149 | CTS | Symptoms/ presented | 41 | 93 | 39 | 24 | 65 | 48 | 1 | 180 | Yes |
| Girlanda, 1998149 | Normal | Healthy volunteers | 45 | NR | Yes | ||||||
| Kabiraj, 1998240 | CTS | Symptoms/ presented | 72 | NR | No | ||||||
| Kabiraj, 1998240 | Normal | Healthy volunteers | 65 | 45 | 39.8 | 20 | 75 | No | |||
| Kleindienst, 1998241 | CTS | Complex objective standard | NR | No | |||||||
| Kleindienst, 1998241 | CTS | Other | NR | No | |||||||
| Kleindienst, 1998241 | CTS | Complex objective standard | NR | No | |||||||
| Kleindienst, 1998241 | CTS | Other | NR | No | |||||||
| Kleindienst, 1998241 | Normal | Healthy volunteers | 18 | 83 | 51 | 43 | 59 | No | |||
| Kleindienst, 1998241 | CTS | Complex objective standard | NR | No | |||||||
| Kleindienst, 1998241 | CTS | Unspecified diagnosis | 77 | 82 | 54 | 22 | 79 | No | |||
| Luchetti, 1998242 | CTS | Unspecified diagnosis | 39 | 79 | 31 | 26 | 45 | No | |||
| Luchetti, 1998242 | Normal | Healthy volunteers | 12 | 83 | 27 | 24 | 36 | No | |||
| Nathan, 1998202 | CTS | Workers at risk | 283 | 45 | 35.2 | No | |||||
| Rosen, 1998201 | Normal | Healthy volunteers | 60 | NR | No | ||||||
| Rosen, 1998201 | CTS | Workers at risk | 20 | 5 | 46 | 26 | 65 | No | |||
| Rosen, 1998201 | CTS | Unspecified diagnosis | 14 | 100 | 53 | 33 | 78 | No | |||
| Scelsa, 1998243 | CTS | Other | 21 | 48 | 46 | 10 | 69 | No | |||
| Scelsa, 1998243 | CTS | Unspecified diagnosis | 42 | 76 | 50 | 25 | 85 | No | |||
| Scelsa, 1998243 | Normal | Healthy volunteers | 25 | 44 | 42 | 23 | 63 | No | |||
| Seror, 1998159 | CTS | Unspecified diagnosis | 85 | 74 | 46 | 25 | 83 | No | |||
| Seror, 1998159 | Normal | Healthy volunteers | 80 | 64 | 42 | 22 | 68 | No | |||
| Smith, 1998244 | CTS | Symptoms/ presented | 82 | 61 | 44 | 17 | 88 | 14 | 1 | 120 | No |
| Tan, 1998206 | CTS | Workers at risk | 64 | 63 | 22 | 28 | No | ||||
| Tan, 1998206 | Normal | Healthy volunteers | 56 | 57 | 21 | 29 | No | ||||
| Terzis, 1998162 | CTS | Unspecified diagnosis | 72 | 92 | 49.6 | No | |||||
| Terzis, 1998162 | Normal | Healthy volunteers | 43 | 84 | 48.3 | No | |||||
| Tetro, 1998102 | Normal | Healthy volunteers | 50 | 74 | 46.9 | 22 | 79 | No | |||
| Tetro, 1998102 | CTS | Complex objective standard | 64 | 64 | 49.3 | 21 | 83 | No | |||
| Werner, 1998207 | CTS | Workers at risk | 727 | 54 | 42 | 25 | 69 | Yes | |||
| Wilson, 1998245 | Normal | Healthy volunteers | 14 | NR | 52 | 33 | 76 | No | |||
| Wilson, 1998245 | CTS | Complex objective standard | 23 | NR | 59 | 24 | 76 | No | |||
| Bak, 1997246 | CTS | Symptoms/ presented | 20 | 55 | Yes | ||||||
| Brahme, 1997199 | CTS | Unspecified diagnosis | 20 | 90 | 37 | 21 | 61 | No | |||
| Brahme, 1997199 | Normal | Healthy volunteers | 15 | 47 | 35 | 22 | 60 | No | |||
| Bronson, 1997163 | Normal | Other | 16 | 56 | 29.5 | 21 | 44 | Yes | |||
| Bronson, 1997163 | CTS | Unspecified diagnosis | 22 | 73 | 34.4 | 21 | 59 | Yes | |||
| Del Pino, 1997104 | Normal | Healthy volunteers | 100 | 78 | 49 | 37 | 67 | No | |||
| Del Pino, 1997104 | CTS | Complex objective standard | 180 | 81 | 50 | 16 | 84 | 37.9 | 1 | 216 | No |
| Dellon, 1997107 | CTS | Unspecified diagnosis | 72 | NR | Yes | ||||||
| Dellon, 1997107 | Cubital tunnel syndrome | Unspecified diagnosis | 42 | NR | Yes | ||||||
| Dellon, 1997107 | Normal | Other | 52 | 62 | Yes | ||||||
| Franzblau, 1997208 | CTS | Workers at risk | 148 | 57 | 44.2 | Yes | |||||
| Guglielmo, 1997247 | CTS | Symptoms/ presented | 198 | 60 | 46 | 13 | 84 | No | |||
| Guglielmo, 1997247 | Normal | Healthy volunteers | 69 | 57 | 40.3 | 20 | 86 | No | |||
| Gunnarsson, 1997248 | CTS | Symptoms/ presented | 100 | NR | No | ||||||
| Horch, 1997249 | Normal | Healthy volunteers | 17 | 71 | 43.4 | 24 | 58 | No | |||
| Horch, 1997249 | CTS | Simple nerve conduction | 19 | 63 | 49.7 | 25 | 67 | No | |||
| Jeng, 1997209 | CTS | Workers at risk | 27 | 52 | 40.2 | 23 | 57 | No | |||
| Kaneko, 1997250 | CTS | Unspecified diagnosis | 15 | 87 | 40 | 54 | Yes | ||||
| Kaneko, 1997250 | Normal | Healthy volunteers | 46 | 22 | 25 | 45 | Yes | ||||
| Kaneko, 1997250 | Cubital tunnel syndrome | Unspecified diagnosis | 10 | 20 | 45 | 56 | Yes | ||||
| Kaneko, 1997250 | Combined WRUEDs | Unspecified diagnosis | 10 | 50 | 40 | 62 | Yes | ||||
| King, 1997114 | CTS | Unspecified diagnosis | 29 | 62 | No | ||||||
| King, 1997114 | Normal | Healthy volunteers | 100 | 50 | No | ||||||
| Pierre-Jerome, 1997251 | Normal | Healthy volunteers | 28 | 100 | 45.1 | 26 | 67 | No | |||
| Pierre-Jerome, 1997251 | CTS | Simple nerve conduction | 27 | 100 | 51.9 | 16 | 78 | 36 | 12 | 72 | No |
| Radack, 1997252 | CTS | Complex objective standard | NR | No | |||||||
| Radack, 1997252 | Normal | Unrelated disease | NR | No | |||||||
| Radack, 1997252 | CTS | Symptoms/ presented | 161 | 53 | 37.4 | 13 | 86 | ||||
