Key Findings and Strength of Evidence

The key findings and strength of evidence are summarized below and in Table 6.

Table 6

Summary of findings and strength of evidence.

Findings in Relation to What Is Already Known

Over the past 25 to 30 years, gout diagnosis has been an area of some controversy. Efforts have been aimed at determining whether the assessment of MSU crystals in synovial fluid aspirated from joints is really the gold standard, validating algorithms comprising various combinations of clinical and laboratory criteria, and validating the use of ultrasound and DECT imaging.

The focus of this report was on evaluating the validity and safety of existing diagnostic methods for use in a primary, urgent, and emergent care setting, where the majority of gout patients are first seen and diagnosed. Patients who present in these settings with an inflamed joint and who have not had a prior diagnosis of gout (or another rheumatic condition) are almost certainly having an acute attack, which may be the first or the latest of a number of attacks. Thus they may be in an early stage of the disease or at least will be less advanced in the disease process than patients seen in the rheumatology setting. Important considerations in diagnosing gout in these patients include ensuring criteria are sensitive enough to diagnose less advanced disease and specific enough to rule out other conditions, including septic arthritis and calcium pyrophosphate deposition disease.

Applicability

Two factors may reduce the applicability of this review.

First, of the studies we identified that assessed the validity of clinical diagnostic algorithms and imaging for the diagnosis of gout, most included at least some participants who had already had a definitive diagnosis. Relatively few studies enrolled only participants with an inflamed joint or even suspected gout but no definitive diagnosis. The present review excluded studies of individuals with a prior gout diagnosis; however, we identified no studies that limited inclusion only to patients presenting with a first attack, and almost no studies considered the duration of the disease or the number of prior attacks in their assessments.

Second, all imaging studies were conducted in a rheumatology setting, usually an academic rheumatology department. Patients seen in this setting may have more advanced disease than those seen in a primary care setting, or may have comorbidities that add complexity to their treatment.

Implications for Clinical and Policy Decisionmaking

The findings of this review provide some evidence to support the further development and validation of diagnostic algorithms based on a combination of clinical signs and symptoms for the diagnosis of gout in the primary care setting. The review further supports the use of imaging modalities (ultrasound and DECT) in cases where a definitive diagnosis cannot be made from signs and symptoms alone.

Limitations of the Comparative Effectiveness Review Process

Assessing the comparative validity of diagnostic tests in systematic reviews presents a number of challenges that are not faced with comparative effectiveness reviews of treatment strategies. These limitations are magnified by several issues surrounding tests for gout and the natural history of the disease itself. To increase applicability to the specific patient population and health care settings of interest, we limited included studies to those that enrolled previously undiagnosed patients; in doing so, we excluded a number of studies on the use of ultrasound and DECT for monitoring patients with chronic gout or hyperuricemia. Previous systematic reviews on the use of ultrasound and DECT included studies that enrolled patients with asymptomatic hyperuricemia and studies of patients with definitive gout diagnoses in various stages of the disease, as well as studies of patients with suspected gout but without a definitive diagnosis (their findings were similar to ours).

Our searches were aimed at identifying studies on gout diagnosis. Searches that identified studies on gout would be expected to identify studies on the differential diagnosis of gout, septic arthritis, calcium pyrophosphate deposition disease, and other such conditions. If a study was aimed at diagnosing patients with a mono- or oligo-arthritis, the chance that the word “gout” would appear is nearly 100%, as that would be one possible diagnosis. However, we might have overlooked an occasional study on differential diagnosis of inflammatory joint conditions that was applicable to gout.

In addition, our consideration of unpublished literature was limited. We were unable to obtain information from manufacturers of microscopes and imaging equipment used to diagnose gout. In addition, we did not include conference proceedings as sources of data but cited several in discussing our findings in the context of what is known about gout diagnosis.

Limitations of the Evidence Base

The literature that addresses the diagnosis of gout has numerous limitations that make it difficult to draw firm conclusions. These limitations can be divided into three categories: study volume, design, and reporting quality. We have already addressed some of the issues in the discussion above. Few studies have attempted to address the diagnosis of gout. Almost no studies have examined the impact of diagnostic test accuracy on decision-making (decisions to order further testing or to initiate particular treatments) or any clinical or patient centered outcomes, and almost no studies addressed adverse events potentially associated with diagnostic testing. Most studies of gout address management issues or monitoring of patients with chronic gout.

Of the diagnostic studies we did identify, few studies limited enrollment to gout suspects or patients with a monoarthritis or some other clinical signs or symptoms that might suggest gout. Many studies enrolled only patients with known gout and included no control group.

Even studies that enrolled patients who were gout suspects or included a control group and employed blinded assessment systematically failed to limit enrollment to patients in their first attack or with recent onset or did not stratify findings by duration of the condition (as would be ascertained by asking, “How long have you been having these attacks?”). The lack of stratification by duration of condition affects the sensitivity and specificity of both clinical diagnostic criteria and imaging techniques.

Most studies also fail to stratify by other relevant factors, such as time since the onset of the current or most recent flare, sex, and comorbidities. The time since onset of the current flare definitely affects the presence of crystals as well as clinical signs and symptoms.

No studies tested the validity of combining a diagnostic algorithm comprising clinical signs and symptoms with an imaging test, compared with clinical signs and symptoms or imaging alone.

As described above, issues concerning the use of synovial fluid MSU crystal identification as the reference standard abound. Taking the validity of the reference standard at face value, some studies assessed MSU in a fraction of participants only (e.g., those for whom synovial fluid could be aspirated, those most suspected of having gout, or those willing to undergo the test), using the ACR criteria or individual clinical judgment as the reference standard for the remaining participants. The technical problems with aspiration and analysis have been assessed and described extensively and include inconsistencies introduced by patient factors (e.g., the time lapse from the start of the flare to aspiration), sample handling factors (storage duration and temperature), and practitioner skills in aspiration and analysis. Finally, failure to report important study design details in publications is a further limitation. Studies tended to be vague regarding blinding of assessors and the time lapse between implementation of the index test and reference standard (and the sequence of tests), a critical detail considering the short duration of gout attacks.

Research Gaps

In a 2013 commentary, Dalbeth,¹⁷ noted that thus far, none of the current diagnostic (classification) criteria have been adequately validated. Efforts to validate the existing classification criteria have either failed to enroll patients prospectively (i.e., before a definitive diagnosis has been made) or have been limited to very small numbers of patients. The ongoing Study for Updated Gout classification cRiteria (SUGAR) project is validating gout classification criteria to improve case ascertainment for recruitment into research studies and for epidemiological purposes. As we suggested above in describing limitations of the research base, promising algorithms for diagnosis in the primary care setting, such as the Diagnostic Rule and the CGD have been validated, but need additional validation in larger, broader populations.

In addition, specific elements of the criteria, such as hyperuricemia, require additional testing. Most clinical diagnostic and classification criteria for gout include hyperuricemia as a criterion.^{16, 22, 31, 39, 41, 42} However, a 1994 study concluded that serum urate was not a valid criterion for diagnosing gout, as there was no lower level below which gout was not a possibility (and no upper limit beyond which it is a certainty).⁵² The 2011 Postgraduate Medicine criteria also excluded hyperuricemia as an element of its clinical diagnostic criteria for that reason,⁴⁹ and the new 2014 ACR/EULAR criteria include hyperuricemia but state that it should not be the sole criterion on which a diagnosis of gout is made.⁵⁰ Thus, further assessment of the effect of hyperuricemia on the sensitivity and specificity of the clinical diagnostic algorithms may be needed.

Patient-level factors that influence test behavior have also been understudied: These include the influence of duration of a flare; number and identity of joints involved; and patient age, sex, and comorbidities. A 2010 systematic review on the diagnosis of gout in women noted that clinical features and risk factors of gout in women differ from those in men.⁴⁴ Women have later onset, are more likely to be taking diuretics, have more CVD and renal comorbidity, are less likely to drink alcohol, are less likely to have podagra (more involvement of other joints), are more likely to have polyarticular gout, and have less frequent recurrent attacks. These findings suggest the need for different clinical diagnostic criteria for women. Likewise, a number of the clinical diagnostic criteria, including the Diagnostic Rule and the new 2014 ACR/EULAR criteria, include cardiovascular comorbidities as a criterion. The sensitivity and specificity of this criterion may need to be established across a broad group of populations.

The findings of Park and colleagues on the effects of gout misdiagnosis³⁷ suggest that studies are needed on differential diagnosis of gout and other inflammatory joint conditions, particularly septic arthritis and calcium pyrophosphate deposition disease. We identified two recent studies that assessed the validity of a simple laboratory test for the differential diagnosis of gout from septic arthritis. Neither study met our inclusion criteria because the gout diagnosis was made prior to the studies. A 2014 study conducted in Germany analyzed multiple inflammatory markers in serum and synovial fluid drawn from patients seen in a hospital emergency room; gout and septic arthritis were ascertained by synovial fluid aspiration with MSU crystal identification and culture, respectively. Among the markers assayed (e.g., serum UA, synovial fluid white blood cells, synovial fluid total protein), synovial fluid lactate had the greatest diagnostic potential to differentiate septic arthritis from gout, followed by glucose and serum uric acid concentrations.⁵³ A 2014 study conducted in an academic orthopedics department in China found that serum and synovial fluid procalcitonin can discriminate between septic arthritis and the non-infectious forms of arthritis (gout, rheumatoid arthritis, and osteoarthritis) in the knee, but that synovial fluid procalcitonin is much more sensitive;⁵⁴ unfortunately, this assessment would still require joint aspiration. Response to colchicine, which has been suggested as a diagnostic criterion for gout, also does not distinguish gout from other crystal arthopathies. Ultrasound and DECT show some evidence of distinguishing gout from calcium pyrophosphate deposition disease, but further work is needed.

Finally, studies are needed that assess the incremental value of ultrasound and DECT imaging over the use of a clinical diagnostic algorithm or even MSU analysis alone. One study assessed the potential additive value of DECT in patients with uncertain diagnosis: the findings suggested DECT may be a useful adjunct to clinical algorithms among patients with disease of longer duration but not those with new onset gout (first attack and symptom duration less than 6 months).²⁸ Another study purported to assess the added value of ultrasound in a clinical diagnostic algorithm, but this study fell short of actually achieving that outcome.³⁶ This information will be necessary in determining the importance and the practicality of setting a guideline for referring patients for imaging in making a diagnosis of gout. Of potential utility would be an appropriateness assessment study that creates a panel of possible clinical scenarios of inflammatory joint presentation with the goal of eliciting the most appropriate diagnostic workup for the primary/urgent/emergency care setting.

Conclusions

This review highlights the need for further, broader validation of promising diagnostic algorithms in primary care settings, where the majority of patients with signs and symptoms suggestive of gout, but no definitive gout diagnosis, are likely to be seen. A clinical algorithm with high diagnostic accuracy can ideally form part of a diagnostic decision tree, with referral of more clinically challenging cases to rheumatologists for more invasive tests or imaging. Research is needed to assess the incremental value of synovial fluid MSU crystal analysis and imaging over that of a diagnostic clinical algorithm.