Abdullah et al., 1 and felt obliged to make the following comments. First, it was not clear whether the article was a prospective or a retrospective study. Although the authors mention that there were 3000 consecutive patients examined in a period of five years, giving the impression that this was a prospective study, the protocol of the study is similar to that of Goletti et al., 2 including the algorithm. Actually they had similar results by performing ultrasound within 30 minutes of patients' arrival, with an average of four minutes to perform it. Was this a retrospective or prospective analysis?Second, when evaluating any test, a gold standard is needed to compare the results. Although laparotomy was used as the gold standard for the operated patients, the authors did not define the false-positive findings for the nonoperated patients, who also need a gold standard (CT scan, for example). This is important when studying the specificity. Specificity is defined as the true-negative results divided by the summation of true-negative and falsepositive. The authors mentioned that ultrasound has a specificity of 97% in detecting splenic injury. Since there were 2927 true-negative results, a simple calculation shows that there were 91 false-positive splenic injuries. How were these defined, and what was their clinical course?Third, the authors cited only four references relating to the use of ultrasound in blunt abdominal trauma, which is a controversial topic. One of these was wrongly cited, 2 as the authors mentioned that sensitivity, specificity and accuracy in ultrasound in detecting hemoperitoneum approach 86.6%, 100% and 97% respectively. The original reference shows that "the overall sensitivity of US in detecting free fluid collection in that study was 98%, with a specificity of 99% and a positive predictive value of 100%." Extensive literature has been published over the last four years regarding the role of ultrasound in blunt abdominal trauma which could have been evaluated. I notice that the time between finishing the study (April 1995), and its submission (May 1995), was less than two months. This is a very short period compared to this extensive experience (3000 patients). It is actually the largest series in the literature reported to date. The second largest series was reported by Glaser et al., 3 and was a retrospective analysis of 10 years' duration, involving 1151 patients. Writing a clinical study should take a lot of consideration, including critical evaluation of the information available.The fourth point is that the inclusion criteria of the study were not mentioned. This is essential in defining the study population. The reported injury percentage of this study (true-positive and false-negative findings) is the lowest in the literature at 2% (73/3000), compared to other studies which varied between 6% to 54%.3-7 Was ultrasound actually indicated in all these patients?My last point is that it seemed that some of the statistical principles are not clear to the authors. The mean (average) is a s...