The diagnostic performance of musculoskeletal ultrasound in gout: A systematic review and meta-analysis

Abstract

Background: Musculoskeletal ultrasound is widely used in diagnosing gout, but its accuracy is debatable. We conducted a systematic review and meta-analysis to quantitatively evaluate the value of ultrasound in the diagnosis of gout.

Methods: We systematically searched for publications using Cochrane Library, PubMed/Medline and Embase and manually screened the references of eligible articles for additional relevant publications. Studies were included in this systematic review if they assessed the diagnostic accuracy of ultrasound in gout compared to that of the gold standard, demonstration of monosodium urate crystals in joint fluid or tophi. We then conducted quantitative analyses by extracting data from each study and calculating the pooled sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR) and diagnostic odds ratio (DOR). The summary receiver operating characteristic curves (sROCs) were constructed to obtain the Q*-index and the area under the curve (AUC).

Results: Thirteen studies were included in this meta-analysis. The diagnostic performances of three distinctive ultrasonographic features of gout, double contour sign (DCS), the presence of tophi and the snowstorm sign, were evaluated. For person-based evaluations, the pooled sensitivity, specificity, DOR, AUC and Q* were as follows: for the DCS, 66% (95% confidence interval (CI) 62%-69%), 92% (95% CI 90%-94%), 25.91 (95% CI 11.80-56.89), 0.8163 and 0.7503, respectively; for the presence of tophi, 56% (95% CI 52%-60%), 94% (95% CI 92%-96%), 21.11 (95% CI 7.84-56.89), 0.8928 and 0.8236, respectively; for the snowstorm sign, 31% (95% CI 27%-36%), 91% (95% CI 88%-93%), 4.54(95% CI 3.13-6.58), 0.5946 and 0.5712, respectively; and for simultaneous consideration of these ultrasonographic features, 80% (95% CI 76%-83%), 83% (95% CI 79%-86%), 19.03 (95% CI 13.97-25.93), 0.889 and 0.8197, respectively. For the joint-/location-based evaluations, the pooled sensitivity, specificity, DOR, AUC and Q* were as follows: for the DCS, 75% (95% CI 68%-80%), 65% (95% CI 59%-70%), 16.90 (95% CI 5.10-56.03), 0.871 and 0.8014, respectively; and for the presence of tophi, 48% (95% CI 40%-57%), 96% (95% CI 91%-99%), 30.20 (95% CI 9.23-98.87), 0.8776 and 0.8081, respectively.

Conclusions: In this meta-analysis, relatively high specificity but modest or low sensitivity were demonstrated in the diagnosis of gout using each of the three ultrasonographic features for person-based evaluations. Simultaneous consideration of these ultrasound findings may improve the diagnostic sensitivity. However, the double contour sign alone is weak in the differentiation of gout and non-gout for joint-/location-based evaluations. Further well-designed studies are still needed to support the current findings.

Fig 1. A flow chart summarizing the study selection process for this meta-analysis.

Fig 2. Results of the methodological quality evaluations using the QUADAS-1 tool.

Green indicates that the criterion is satisfied. Yellow means that it is unclear whether the criterion is satisfied or not. Red indicates that the study did not meet the criterion.