NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Lin JS, Webber EM, Beil TL, et al. Fecal DNA Testing in Screening for Colorectal Cancer in Average-Risk Adults [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2012 Feb. (Comparative Effectiveness Reviews, No. 52.)

Cover of Fecal DNA Testing in Screening for Colorectal Cancer in Average-Risk Adults

Fecal DNA Testing in Screening for Colorectal Cancer in Average-Risk Adults [Internet].

Show details


Literature Yield

The literature search yielded a total of 336 citations from electronic database searches and outside sources (Figure 2). Based on the review of title and abstracts, we subsequently reviewed 34 full-text articles for their eligibility. We included 12 articles in total, three diagnostic accuracy studies that met inclusion criteria for Key Question 2, three analytic validity studies for Key Question 4, and six studies of acceptability or preference of testing for Key Question 5. Two studies for Key Question 2 also reported adherence to testing and therefore are discussed with Key Question 5 results. We found no studies that addressed clinical utility (Key Question 1), intervals of screening (Key Question 3), or specific harms of screening (Key Question 6).

This flow chart summarizes the literature yield and selection of articles. There were 279 citations identified through published literature searches. An additional 57 citations were identified from outside sources such as reference lists. Of these 336, 302 citations were excluded at the title/abstract level. The full-text of the 34 remaining citations were examined for inclusion. Of these, 22 studies were excluded for the following reasons: study design, population, test not currently available, lack of appropriate comparator, analytic validity not vimentin methylation. One article was excluded for different reasons for different key questions. In total, there were no studies included for key questions 1, 3, or 6. There were 3 studies included for key question 2, 3 studies included for key question 4, and 6 studies included for key question 5.

Figure 2

Literature flow diagram. * 1 article was excluded for different reasons for different key questions † 2 articles from KQ2 reported adherence to testing (and therefore are also discussed with KQ5)

Results of Included Studies

Key questions 1 to 3. Benefits of fecal DNA testing

We found no studies that evaluated the effectiveness of fecal DNA screening on CRC incidence or mortality, alone or in combination with other screening tests (Key question 1). We found three fair- to poor-quality studies in screening populations (n=5662) that evaluated the absolute test performance of fecal DNA testing for CRC screening, compared to colonoscopy (reference standard).33,34,47 Two of these studies also concomitantly evaluated guaiac-based fecal occult blood testing (FOBT) (Key Question 2).33,34 We found no studies that evaluated the test performance of fecal DNA testing across different screening intervals (Key Question 3). We excluded 65 initial test-validation studies that were not conducted in screening populations, the majority of which were case-control studies in patients with CRC and healthy controls. Only three of these 65 studies were prospective cohort studies conducted in high-risk patients (Table 4a and b). Six of the excluded case-control studies evaluated vimentin methylation independently, which is the basis for the only commercial fecal DNA test currently available (Table 1a and b).

Table 4a. DNA markers evaluated in excluded studies focusing on test development and/or initial validation of fecal DNA testing for colorectal cancer: Case control studies.

Table 4a

DNA markers evaluated in excluded studies focusing on test development and/or initial validation of fecal DNA testing for colorectal cancer: Case control studies.

Table 4b. DNA markers evaluated in excluded studies focusing on test development and/or initial validation of fecal DNA testing for colorectal cancer: Cohort studies in high risk patients.

Table 4b

DNA markers evaluated in excluded studies focusing on test development and/or initial validation of fecal DNA testing for colorectal cancer: Cohort studies in high risk patients.

Despite the availability of numerous initial validation studies of fecal DNA testing, we found only three studies that examined the test accuracy of fecal DNA testing in screening populations.33,34,47 In summary, two fair- to poor-quality diagnostic accuracy studies in screening cohorts of average-risk patients undergoing colonoscopy have evaluated multi-marker fecal DNA testing (two different versions) by Exact Sciences.33,34 The third study was rated as poor-quality and was a smaller cohort study evaluating the test accuracy of KRAS mutations.47 Due to the evolution of fecal DNA tests, none of these studies evaluated tests currently on the market. Two studies (n analyzed=5004) evaluated a multi-marker fecal DNA test that was a prototype to a later version that was clinically available as PreGen Plus™.33,34 These two studies found somewhat different estimates of sensitivities to detect CRC (25 percent [95% CI, 5 to 57 percent] versus 51.6 percent [95% CI, 34.8 to 68.0]). There were some differences in study populations, but it is unclear if these differences should affect test performance. Sensitivities for advanced adenomas were very low in both studies. A smaller (n=217) subset analysis evaluating a different multi-marker fecal DNA test and a smaller (n=441) study evaluating a single marker (KRAS) were both poor quality.

The most recent study by Ahlquist and colleagues published in 2008 was an NCI and industry-funded diagnostic accuracy study conducted in a large cohort (n enrolled=4482) of 50 to 80 year olds at average-risk for CRC (Table 5).34 Due to cancellations, protocol violations, incomplete colonoscopy or incomplete stool samples, 718 persons (16 percent) were excluded, thus 3764 participants were included in the analysis. Based on limited characteristics (i.e., age, sex, and race/ethnicity), included participants did not appear to be different from the overall enrolled study population. The study evaluated a pre-commercial stool DNA test (SDT-1, pre-commercial version of PreGen Plus), and a second, different multi-marker panel (SDT-2) assay, that evolved after the study had started. The SDT-1 evaluation ended after reviewing interim results on the first 2497 participants, which, per the author, was precipitated by a decision to add Hemoccult SENSA™ as a comparator. The evaluation of SDT-2 was a smaller nested case-control study in which the SDT-2 test was run on a subset of patients (n analyzed=217). This subset included all cancers (n=19), high-grade dysplasia (n=20), adenomas ≥ 2 cm (n=53), a random subset of 1-2 cm adenomas (n=50), and normal controls (n=75). All participants received a colonoscopy for validation of CRC-related findings. SDT-1 was a multi-target fecal DNA test that included 21 mutations in the KRAS, APC, and TP53 genes, along with markers for microsatellite-instability (MSI) and long DNA. SDT-2 was a different multi-target fecal DNA test that included mutations in the KRAS, APC genes as well as vimentin gene methylation. Guaiac-based FOBTs, both Hemoccult II™ and Hemoccult SENSA, were concomitantly evaluated. Only a subset of patients, however, were advised about dietary and medication restrictions. Therefore, results reported for FOBT (n=3764) are problematic and not discussed further in the results. Test performance outcomes for SDT-1 were rated fair quality despite reporting of an unplanned interim analysis as the final results. However, test performance outcomes for SDT-2 were rated as poor quality (see below).

Table 5. Patient characteristics for studies of diagnostic accuracy of fecal DNA testing (KQ2).

Table 5

Patient characteristics for studies of diagnostic accuracy of fecal DNA testing (KQ2).

The cohort for SDT-1 evaluation (n analyzed=2497) had a mean age of 60.4 years, were 46.0 percent male, and 92.7 percent White (Table 5). Overall test positivity was 5.2 percent. The sensitivity for one-time fecal DNA testing with SDT-1 was 25 percent (95% CI, 5 to 57) for CRC, 19 percent (95% CI, 5 to 42) for advanced adenomas and 20 percent (95% CI, 14 to 26) for CRC and advanced adenomas. Specificity for any CRC or advanced adenomas was 96 percent (95% CI, 95 to 97) (Table 6). The subset who were included in the SDT-2 evaluation (n analyzed=217) were generally older than those evaluated in the overall cohort (mean age 66.4), 50.2 percent male and 92.6 percent White (Table 5). Because only a selected subset of persons had SDT-2 testing, overall test positivity was artificially elevated at 35 percent. Weighted sensitivity for one-time fecal DNA testing with SDT-2 was 40 percent (95% CI, 32 to 49) for detection of CRC and advanced adenomas, which was twice the sensitivity of SDT-1 (Table 6). Sensitivity for detection of CRC was 58 percent (95% CI, 36 to 80). Although specificity was not reported, 16 percent of patients with normal colonoscopy had an abnormal SDT-2 result, and 26 percent of patients 65 years or older with a normal colonoscopy had an abnormal SDT-2 result. Thus, enhanced sensitivity with SDT-2 likely had a cost in terms of decreased specificity.

Table 6. Diagnostic accuracy of fecal DNA testing in screening populations (KQ2).

Table 6

Diagnostic accuracy of fecal DNA testing in screening populations (KQ2).

In general, the strength of this study was that it was a large cohort in an average-risk screening population, which recruited from 22 academic and regional health care systems. However, important study limitations impacted our ability to interpret the results for SDT-2 test performance. These include the small sample size, exclusion of all patients with protocol violations, inadequate samples, or colonoscopy, limited sampling of controls despite weighting sensitivity for proportion of screen relevant neoplasia in the entire population, and inability to accurately estimate test specificity (Table 7). Although the lack of adherence to dietary and medication restrictions should, in theory, not decrease sensitivity of FOBT, the sensitivities for Hemoccult II and Hemoccult SENSA reported in this study were much lower than more generally accepted estimates for sensitivities.103 In addition to concerns about the generalizability of the cohort studied (i.e., inclusion of a mostly White population, n eligible not reported), the single most important limitation is that neither SDT-1 or SDT-2 were ever available for clinical use and are both different from the currently available test ColoSure™.

Table 7. Limitations and quality concerns for diagnostic accuracy studies of fecal DNA testing (KQ2).

Table 7

Limitations and quality concerns for diagnostic accuracy studies of fecal DNA testing (KQ2).

The study by Imperiale and colleagues was another industry-funded fair-quality large cohort study (n included=4404) evaluating the same pre-commercial stool DNA test (SDT-1, pre-commercial version of PreGen Plus) as was evaluated in the study by Ahlquist and colleagues (Table 5).33 In this study, researchers compared SDT-1 with 3-card non-rehydrated Hemoccult II average-risk, asymptomatic patients who all underwent colonoscopy. Of the 5486 enrolled participants, 1082 (20 percent) were excluded due to incomplete testing. Baseline characteristics for enrolled participants were not reported in comparison with included participants. About 50 percent more patients (641 vs. 426) did not provide an adequate sample for fecal DNA testing as compared to Hemoccult II, which may signal differences in feasibility or acceptability to patients. Although all included patients had colonoscopy and Hemoccult II testing, only a subset received SDT-1 testing. A subset (n=2507) of the 4404 that completed testing were analyzed. This subset included all subjects with an invasive cancer (n=31) or advanced adenoma (n=403), along with a random subgroup with minor polyps (n=648) and normal findings (n=1423). The analyzed subgroup was similar with respect to age, sex, race/ethnicity, and family history of CRC, as compared to the overall cohort (Table 5). As a group, this cohort included slightly older participants (69.5 yrs), slightly more men (44.5% male), and slightly more non-White individuals (13%) than the cohort studied by Ahlquist and colleagues. Two patients (one rectal carcinoid tumor, one cloacogenic tumor) were excluded from analyses and reported results. As compared to the study by Ahlquist and colleagues, the comparative results for Hemoccult II performance are of fair (as opposed to poor) quality. In this study, investigators reported that subjects were given proper dietary and medication instructions, cards were returned for non-rehydrated analysis consistent with manufacturer's instructions, and results for Hemoccult II are given for the same subgroup of persons who received fecal DNA testing.

Of those tested with SDT-1, 8.2 percent were test-positive on the fecal DNA panel and 5.8 percent had a positive Hemoccult II. One-time fecal DNA testing was more sensitive for CRC than Hemoccult II (51.6 percent, [95% CI, 34.8, 68.0] and 12.9 percent [95% CI, 5.1, 28.9], respectively). This estimate of sensitivity is much higher than the sensitivity for CRC reported for SDT-1 in the study by Ahlquist and colleagues (Table 6). Sensitivity for advanced adenomas was similarly poor for fecal DNA testing (15.1 percent, [95% CI, 12.0 to 19.0] and for Hemoccult II (10.7 percent, [95% CI, 8.0 to 14.1]). While specificity for CRC or CRC and advanced adenomas did not differ significantly between fecal DNA and Hemoccult II, power to detect a difference was limited since the full sample was not tested (Table 6).

In general, this was a fair-quality study conducted in a large, average-risk screening population. This population was drawn from 81 private-practice and university-based settings. Study investigators were blinded, except for the gastroenterologists who had access to FOBT results. This study had several limitations, however, that impacted both its internal and external validity (Table 7). These limitations include: poor precision in the estimates of test performance characteristics due to sample size issues, exclusion of 20 percent of the study population for incomplete testing data (and unknown comparability of patient characteristics for excluded participants), inclusion of a mostly older population (in which three-quarters of the study population was over 65 years of age), inclusion of a mostly White population, number of eligible participants not reported, conduct of the fecal DNA testing centrally at a single lab, and an unusually low estimate for Hemoccult II sensitivity (compared to conventional understanding of Hemoccult II performance). Most importantly, the version of the test evaluated in this study was never commercially available, and the included markers for evaluation are different (no overlap) from the currently available test ColoSure.

The third included diagnostic accuracy study was rated as poor quality and was an analysis from a population-based cohort study that examined baseline stool samples for a single mutation of the KRAS gene in 441 older adults (aged 50 to 75 years) within a larger cohort study (n=9953) (Table 5).47 This subgroup represented those who opportunistically received their reference colonoscopy within two years of the DNA testing. The included subgroup was similar to the overall study population, except that more participants in this group reported a first-degree relative with CRC. The fecal test had zero percent sensitivity, testing positive in none of the 31 participants with advanced colorectal neoplasia (seven patients with invasive CRC) (Table 6). The highest rate of mutant KRAS was reported in participants with a negative colonoscopy (7.5 percent). Important study limitations included bias in the spectrum of patients self-selecting for colonoscopy, and the lag-time between stool collection and clinical diagnosis that could have affected test performance (Table 7).

Key Question 6. Harms of fecal DNA testing

We found no studies that specifically evaluated the harms of fecal DNA testing. Hypothesized harms other than harms from diagnostic inaccuracy (false positives and false negatives) include psychological harms (anxiety or worry) around testing. It is unclear if the psychological harms around genetic based testing are qualitatively different than psychological harms of other stool based testing like guaiac-based or immunochemical fecal occult blood testing.

The downstream effects of false positive results primarily include the harms of unnecessary diagnostic interventions (i.e., colonoscopies). Therefore, the harms of fecal DNA testing inaccuracy would be relative to comparative specificity of other stool based testing, but less than colonoscopies itself as a screening option. Based on the included studies, the specificity of SDT-1 and FOBT were not statistically significantly different, although the studies had limited power to detect a difference.33 Although the specificity of SDT-2 was not reported, SDT-2 had a positivity rate of 16 percent (95% CI, 8 to 24) in persons with normal colonoscopies, and the positivity rate increase with age.34,104 This positivity rate is much higher than that reported for FOBT test positivity, but FOBT test performance data from this study is of poor quality (for the quality reasons discussed above). As stated above, these data have poor applicability given that SDT-1 and SDT-2 are not clinically available and are very different from currently available testing.

Key Question 4. Analytic validity of fecal DNA testing

We found three poor-quality studies that evaluated the analytic validity of currently available fecal DNA assays, specifically a single-marker test for methylated vimentin.29,31,80 We did not identify any additional information by searching grey literature or from requesting non-published literature from LabCorp via the Scientific Information Packet. For analytic validity, we specifically looked for analytic accuracy of the test, repeatability or reproducibility, analytic sensitivity or lower limit of detection of the test, or analytic specificity (Table 8).

Table 8. Definitions for analytic validity.

Table 8

Definitions for analytic validity.

In summary, these three studies show that technological advances in analytic factors (i.e., test methods and performance of procedures) can improve the analytic sensitivity of assays to detect methylated vimentin in stool samples (Table 9). None of the studies evaluated the repeatability, reproducibility, or analytic specificity of testing. These three studies were generally of poor quality, and it is unclear if the exact technological advances evaluated in these studies are applicable to the currently available test for methylated vimentin (ColoSure).

Table 9. Analytic validity of fecal DNA testing (KQ4).

Table 9

Analytic validity of fecal DNA testing (KQ4).

Two studies evaluated technological advances aimed at improving the analytic sensitivity of testing. The most recent study by Li and colleagues was aimed at testing methyl-BEAMing to methylated vimentin in plasma and stool from colon cancer patients.31 Methyl-BEAMing (Beads, Emulsion, Amplification, and Magnetics) is a method for performing methylation specific PCR in compartments created by the emulsion that contain individual strands of DNA to allow digital enumeration of the PCR products. This technique's accuracy was compared with next generation sequencing (sequencing by synthesis) in stool samples from patients with CRC (n=5) and the analytic sensitivity was compared to methylation specific PCR without methyl-BEAMing. The study found that enumeration of methylation by either next generation sequencing or methyl-BEAMing produced essentially the same result in both samples with low and high fractions of methylated vimentin (Table 9). Another experiment within this study showed that methyl-BEAMing enhanced overall technical sensitivity for detecting methylated vimentin DNA by at least 62-fold in artificial samples created by mixing DNA from peripheral blood lymphocytes (unmethylated) with DNA from CRC cell lines (methylated). The overall quality of both experiments was poor given the small sample sizes and incomplete reporting of results. Accuracy was only assessed in five samples, and only reported for four of the five samples. It is unclear if the experiment for the analytic sensitivity was replicated. The applicability of this experiment is also poor given that the accuracy study was conducted in plasma samples, rather than stool samples, and methyl-BEAMing does not appear to be used in the assay evaluated by Ahlquist and colleagues (included in Key Question 2), or in the currently available methylated vimentin test.

The second study by Zou and colleagues tested a method known as methyl-binding domain (MBD) enrichment to see if it could increase the analytic sensitivity for detecting tumor-specific methylated markers in patient stools.80 Two sets of experiments showed that the assay with MBD enrichment had a lower limit of detection of methylated vimentin in normal stool aliquots with added DNA from CRC cell lines, and in stool samples from patients (n=8) with known CRC with tissue positive for methylated vimentin (Table 9). The lower limit of detection of methylated vimentin with MBD enrichment was 10 ng from the normal stool with added DNA from cell lines, and 4 ng from stool samples from patients with CRC. This study, however, was rated as poor quality because it had very small sample sizes. We cannot assess the applicability of this study because it is unclear if MBD enrichment was used in the assay evaluated by Ahlquist and colleagues (included in Key Question 2), or is used in the currently available methylated vimentin test.

The earliest study by Chen and colleagues is essentially a proof-of-concept study aimed at testing the technical limits of the sensitivity of detecting DNA methylation.29 It showed that methylation-specific PCR could detect as little as 25-50 pg of input methylated DNA, which corresponds to a detection limit for the assay of approximately 15 methylated cells (Table 9). This study, however, was rated as poor quality mainly because the experimental results were only reported for one sample. More importantly, this study has poor applicability to current tests as it was conducted in tumor tissue and not conducted in stool samples. This lower limit of detection is an order of magnitude lower than the Zou study conducted in stool samples.

Key Question 5. Acceptability and adherence of testing

We found six fair- to poor-quality studies that evaluated the acceptability,27,106-110 and two diagnostic accuracy studies (from Key Question 2) that reported the adherence to fecal DNA testing.33,34 From very limited evidence, it appears that fecal DNA testing, in the form of a single whole-stool sample, is generally acceptable, although an important test attribute for acceptability appears to be the test's accuracy. In one fair-quality diagnostic accuracy study fecal DNA adherence was lower than adherence to Hemoccult II. No studies have evaluated the relative acceptability or adherence of fecal DNA tests to FIT tests. This is an unfortunate omission, as FIT is the most similar to fecal DNA testing in that it is a non-invasive stool based test that does not require any dietary or medication restrictions. Unlike other stool-based testing, however, fecal DNA testing is currently a single whole-stool sample that theoretically may be preferable to serial card based testing.

It is likely that future fecal DNA testing will be sufficiently different both in the format/collection (no longer a single whole stool sample) and in test accuracy, such that this currently available evidence on acceptability and adherence to fecal DNA testing will no longer be relevant.

Acceptability of testing. We found five fair- to poor-quality studies that evaluated the acceptability of fecal DNA testing (Table 10). Two fair- to poor-quality studies were conducted by Marshall and colleagues.109,110 In these two studies, authors used a cross-sectional survey and modeling to measure Canadian preferences for CRC screening tests109 or to measure patient and physician preferences for CRC screening tests in both Canada and the United States.110 These two studies were rated as fair quality in assessing patient preferences on attributes of testing, but poor quality in assessing patient or physician preference (or willingness to pay) for type of testing. In both studies, adults without a history of CRC (n=2135) were surveyed to elicit preferences on key attributes of available screening tests (e.g., process, preparation, pain, sensitivity, specificity, frequency, followup, complication risk and cost). The studies found that patient's preferred tests that were non-invasive, required no pain or preparation, were highly accurate, did not require repeated measurements over time, and caused no complications. One study used modeling to rank the importance of attributes and found that the relative importance was (in order of most to least importance): sensitivity, specificity, preparation, process, pain.109 The second study also surveyed 100 United States and 100 Canadian practicing primary care physicians.110 The most preferred attribute for primary care physicians was high sensitivity, other important attributes included the pain, specificity, complication risk, preparation, and testing frequency. The major limitations of the survey portion of the study include was a 52 percent response rate in one study 109 and no reported response rate in the second study.110 Both studies used modeling to rank the types of tests (in order of most preferred to least preferred), however, the results from these models are not reliable due to substantial modeling limitations (Table 10).

Table 10. Patient preferences and acceptability of fecal DNA testing (KQ5).

Table 10

Patient preferences and acceptability of fecal DNA testing (KQ5).

One fair-quality test manufacturer-funded study by Schroy and colleagues was designed to compare the perceptions of fecal DNA, Hemoccult II, and colonoscopy in a screening population (Table 10).106 In this study, participants (n=4840) in the diagnostic accuracy study by Imperiale and colleagues received a 25-item questionnaire on the three different screening tests that they completed as part of the study. Eighty-four percent (4042/4840) of all participants who completed all three screening tests returned the mailed questionnaire. This study evaluated the prototype to PreGen Plus a multi-marker panel requiring one single whole stool sample (mailed into laboratory by patient). Overall, 45 percent of respondents preferred fecal DNA testing, 32 percent preferred Hemoccult II, 15 percent preferred colonoscopy, and eight percent had no preference (p<0.001). However, on the individual measures (ease of instruction, simplicity of collection, comfort, invasiveness, embarrassment, anxiety of prep, anxiety of test, accuracy, and likeliness to repeat test) there was no meaningful difference between fecal DNA testing and Hemoccult II. The authors concluded that fecal DNA testing was rated more favorably (for simplicity of collection, comfort, anxiety of prep, accuracy, and likeliness to repeat) based on statistical significance alone. However, the difference was a fraction of a point on a five-point scale. Although colonoscopy was perceived to be more accurate than fecal DNA testing, respondents rated colonoscopy as less favorable in terms of invasiveness, anxiety (around prep and test), and less likely to repeat the test. Although this study was rated as fair quality, a few limitations of this study make it difficult to generalize study findings—no mention of missing data, participants received financial compensation for the study, and participants had to adequately complete all three screening tests to be included in the study.

The other three poor-quality studies provide little additional information given their methodological weaknesses (Table 10). One case-control diagnostic accuracy study (n=162) by Itzkowitz also evaluated patient preferences in a survey, however, no response rate for this survey was reported and patients likely knew if they had CRC or a normal colonoscopy prior to doing the fecal DNA test. Another study by Schroy and colleagues was a cross-sectional survey (n=263) of participants at average-risk for CRC designed to assess patient preferences for CRC screening methods.108 The survey was conducted as a structured interview after receiving patient education about the different screening modalities, including fecal DNA testing, FOBT, colonoscopy, flexible sigmoidoscopy with or without FOBT, and double contrast barium enema (DCBE). Unfortunately, as stated by the authors, during the education component the accuracy of fecal DNA testing was incorrect (it stated that the ability to predict precancerous polyps was medium to high). While the study also reported on participants' willingness to pay out of pocket for testing, these patients were not presented with the costs of each test to inform their decision. The last study was a convenience sampling survey designed to examine patients' screening experiences with fecal DNA testing. Berger and colleagues analyzed returned questionnaires that were included with the PreGen Plus collection kit for the first 2 years it was commercially available. However, only 18 percent of persons who received the PreGen Plus test kit returned the survey.

Adherence to testing. Two studies included from Key Question 2 examining the diagnostic accuracy of fecal DNA testing reported completion of fecal DNA testing in comparison to FOBT and colonoscopy.33,34 We found no additional studies that specifically examined adherence to fecal DNA testing. Both of the studies addressing adherence evaluated fecal DNA testing requiring a single whole stool sample mailed to the laboratory. In the fair-quality study by Imperiale and colleagues (see Key Question 2 results), 11.7 percent (641/5486) did not complete the fecal DNA test versus 7.8 percent (426/5486) did not complete Hemoccult II, and 14 percent (770/5486) did not complete colonoscopy.33 In the other fair-to poor-quality study by Ahlquist and colleagues (see Key Question 2 results), the adherence to testing was much higher, only 1.8 percent (68/3834) did not complete the stool testing (assumed both fecal DNA and FOBT) within the allotted time (120 days).34 Authors report that 4.3 percent (171/4005) did not have an adequate colonoscopy (did not reach cecum or view over 90 percent of the colorectum). It is unlikely that these completion rates can be generalized to practice given that participants were given financial compensation for participating in both studies. Additionally, completion of the fecal DNA and FOBT testing were done simultaneously in one study (patients collected three whole-stool specimens in a plastic bucket and promptly smeared stool onto both window of the Hemoccult II and Hemoccult SENSA cards and mailed all specimens into the lab).34

PubReader format: click here to try


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...