Sir, In their very interesting review, Moretti and Pizzi examined the anatomical and physiological features, safety and accuracy of ultrasonography of optic nerve sheath diameter (ONSD). They pooled studies reporting ONSD and invasive intracranial pressure (ICP) considering invasive measurements as the reference standard and using two different models: they calculated performance estimates (sensitivity, specificity, etc.) using a bivariate model and determined the area under the summary receiver-operating characteristic (SROC) curve for raised ICP detection using a fixed-effect model. The most appropriate method for pooling data from the evaluation of diagnostic tests is the bivariate model. It accounts for study size and incorporates the negative correlation between sensitivity and specificity as a result of differences in test positive or negative threshold between studies and heterogeneity. Articles on ONSD do not report the 2 2 table to be used in the bivariate model but it could be calculated from results section and figures presenting the correlation between ONSD and ICP. We recently performed a meta-analysis using a bivariate model with the same six studies. Our results slightly differ from Moretti and Pizzi’s results, showing a pooled sensitivity of 0.90 (0.80–0.95) and a pooled specificity of 0.85 (0.73–0.93) for raised ICP detection. The area under the SROC curve was, in our meta-analysis, 0.94 (0.91–0.96). Moreover, in Moretti and Pizzi’s article, it is not clear how many studies were finally included for estimating the pooled sensitivity or specificity as they quoted in the legend of Table 2 ‘Pooled performance estimates from the studies comparing ONSD with ICP where the true/ false-positive/-negative results could be calculated’. Furthermore, the authors should also precisely state as to how many patients from Soldatos et al. study were included for the calculation of sensitivity and specificity. Table 1 from Moretti and Pizzi’s article indicates that 50 patients were included in this study, but only 32 patients actually had an invasive measure of ICP in Soldatos’s study. We agree that ultrasonography of ONSD is a good add-on test for the detection of raised ICP. However, physicians should keep in mind that 10–14% of patients with raised ICP would not be detected. From the available evidence, ultrasonography of ONSD is not intended to replace ICP monitoring using invasive devices, but may be very useful, when combined with a set of clinical and radiological signs, in cases where no specific recommendations on ICP monitoring exist or when ICP monitoring is not feasible. It may also help physicians to decide to transfer patients or to start a specific treatment for raised ICP before placing invasive ICP monitoring devices. An individual data meta-analysis could be very useful to confirm these results and to estimate the best cut-off value.