Objective: To assess the validity of the clinical detection of knee hydrarthrosis compared to ultrasonography (US). To assess the differences of results for both clinical and US tests when they are done by independent investigators. To assess the effect of a previous clinical evaluation on the US diagnosis.
Method: The clinical and US examinations were scored using a 3 point scale. In 50 symptomatic patients, 82 knees underwent clinical and US evaluations by the same physician. Two independent observers clinically and echographically tested 22 other knees (11 patients). Finally, in another series of 20 knees (10 patients), the first investigator performed clinical testing before the US evaluation, while the other did the US with no previous clinical evaluation.
Results: Between the clinical and US evaluations, agreement was moderate (kappa = 0.508), but statistically significant (p<0.001). The relationship between the 2 investigators for clinical evaluation was slightly weaker (kappa = 0.446, p = 0.032). Possible (score 1) or definite (score 2) clinical knee hydrarthrosis sensitivity was 79/100% and specificity 25/78%, respectively. In contrast, US studies from 2 independent observers were strongly related (kappa = 0.902) and were not modified by a previous clinical evaluation.
Conclusion: Clinical knee hydrarthrosis detection is less accurate and reproducible than US detection.