The anatomic snuff box (ASB) tenderness and the longitudinal thumb compression (LTC) tests were both evaluated in 8 studies ensuring power to perform statistical analyses. The summary receiver operating characteristic curves showed that sensitivity was higher for ASB tenderness. For ASB tenderness, 8 articles studying 1164 patients were included. The sensitivity was relatively homogeneous, with a range between 0.87 and 1.00. The specificity ranged between 0.03 and 0.98 indicating high heterogeneity. For LTC, 8 reports studying 961 patients were included. The test had heterogeneous results for sensitivity (range between 0.48 and 1.00) and specificity (range between 0.22 and 0.97) (table 2). Because of the high heterogeneity we refrained …show more content…
Power to perform a stratified analysis for the relevant analyses was too low. The other 23 tests were evaluated in too few studies (1 to 4) to estimate all 5 parameters needed to fit the analysis model. Scaphoid tubercle tenderness (STT) showed reasonably consistent and high sensitivities in 4 papers studying 879 patients, with a range of 0.82 to 1.00. This was also noticed for painful ulnar deviation (PUD) in 4 papers studying 394 patients where sensitivity ranged from 0.67 to 1.00. Rhemrev et al studied grip strength and range of motion (supination/pronation and flexion/extension). Both were significantly decreased in patients with a fractured scaphoid compared to those without fractured scaphoids.25 Three studies combined multiple tests to improve diagnostic accuracy. One study combined ASB tenderness, STT, and LTC to reach a sensitivity of 1.00 and a specificity of 0.74. (23) This would result in a post-test fracture probability of 64%. Two studies described diagnostic accuracy for combining tests but also used predicted probabilities. Multivariate logistic regression analysis was performed to calculate the optimal combination of predictors. Predicted probabilities of a scaphoid fracture derived from regression