Ultrasonography in the diagnosis of appendicitis: evaluation by meta-analysis

We wanted to review the usefulness of ultrasonography (US) for the diagnosis of appendicitis and to evaluate the diagnostic accuracy of US according to patients' and researchers' characteristics.

The relevant Korean articles published between 1985 and 2003 were included in this study if the patients had clinical symptoms of acute appendicitis. The histopathologic findings were the reference standard and the data were presented for 2 x 2 tables. Articles were excluded if patients had no sonographic signs of appendicitis according to graded-compression US. Two reviewers independently extracted the data on study characteristics. The Hasselblad method was used to obtain the combined estimates of sensitivity and specificity for the performance of US.

Twenty-two articles (2,643 patients) fulfilled all inclusion criteria. The estimate of d calculated by combining the sensitivity and specificity was 2.0054 (95% confidence interval [CI]: 1.8553, 2.1554) by a random effects model. The overall sensitivity and specificity (95% CI) were 86.7% (85.4 to 88.0), and 90.0% (88.9 to 91.2), respectively. According to the subgroup meta-analysis by patients' characteristics, the d estimate (95% CI) of dominantly younger age, male, and highly clinical suggestive group for US was 2.2388 (1.8758 to 2.6019), 2.7131 (2.2493 to 3.1770), and 2.4582 (1.7387 to 3.1777), respectively. Also, according to subgroup meta-analysis by researchers' characteristics, the d value (95% CI) for US done by diagnostic radiologists and gray-scale was 2.0195 (1.7942 to 2.2447) and 2.2630 (1.8444 to 2.6815).

This evidence suggests that US may be useful for the diagnosis of acute appendicitis, especially when patients are younger age, male, and highly clinical suggestive.