Endoscopy versus radiography for the diagnosis of acute bacterial rhinosinusitis.

The aim of the study was to estimate the sensitivity and specificity of endoscopy and the diagnostic value of clinical criteria for acute bacterial rhinosinusitis (ABRS). The hospital records of 117 consecutive patients who had symptoms compatible with ABRS and who underwent sinus radiography and flexible nasendoscopy were retrieved. A positive diagnosis was entertained when radiography demonstrated air fluid level, complete opacification or at least 6 mm mucosal thickening and/or endoscopy revealed purulent material within the drainage area of the sinuses. Using a modified version of the Hui and Walter procedure, the sensitivity and specificity of nasendoscopy was calculated against sinus radiography, a standard reference test, with known estimated sensitivity (76%) and specificity (79%). The findings show that 40 patients (34%) had positive endoscopy and radiography. Twenty patients (17%) had positive endoscopy and negative radiography, and vice versa in 15 (13%). In 42 patients (36%) both modalities were negative. Endoscopy yielded a sensitivity of 80% (95% CI, 0.55-1.00) and a specificity of 94% (95% CI, 0.33-1.00). Although estimates seem better than those of radiography, the confidence intervals are quite wide and no firm conclusion is drawn. Neither single nor combination of symptoms, concurrent with two major and one minor factor advocated by the AAO-HNS Task Force on Rhinosinusitis, were associated with a positive diagnosis of ABRS. The data suggest that endoscopy is no less effective than radiography in predicting ABRS, also that clinical criteria are of rather limited diagnostic value, and thus should be supported by other diagnostic tests. Endoscopy is recommended as a first line diagnostic tool for the diagnosis of ABRS in routine ENT practice. If the outcome is negative, radiography may follow.