Ultrasonography in diagnosing clinically occult groin hernia: systematic review and meta-analysis

To provide an updated systematic review on the performance of ultrasonography (US) in diagnosing clinically occult groin hernia.

A systematic search was performed in MEDLINE and Embase. Methodological quality of included studies was assessed. Accuracy data of US in detecting clinically occult groin hernia were extracted. Positive predictive value (PPV) was pooled with a random effects model. For studies investigating the performance of US in hernia type classification (inguinal vs femoral), correctly classified proportion was assessed.

Sixteen studies were included. In the two studies without verification bias, sensitivities were 29.4% [95% confidence interval (CI), 15.1-47.5%] and 90.9% (95% CI, 70.8-98.9%); specificities were 90.0% (95% CI, 80.5-95.9%) and 90.6% (95% CI, 83.0-95.6%). Verification bias or a variation of it (i.e. study limited to only subjects with definitive proof of disease status) was present in all other studies. Sensitivity, specificity, and negative predictive value (NPV) were not pooled. PPV ranged from 58.8 to 100%. Pooled PPV, based on data from ten studies with low risk of bias and no applicability concerns with respect to patient selection, was 85.6% (95% CI, 76.5-92.7%). Proportion of correctly classified hernias, based on data from four studies, ranged between 94.4% and 99.1%.

Sensitivity, specificity and NPV of US in detecting clinically occult groin hernia cannot reliably be determined based on current evidence. Further studies are necessary. Accuracy may strongly depend on the examiner's skills. PPV is high. Inguinal and femoral hernias can reliably be differentiated by US.

• Sensitivity, specificity and NPV of ultrasound in detecting clinically occult groin hernia cannot reliably be determined based on current evidence. • Accuracy may strongly depend on the examiner's skills. • PPV of US in detection of clinically occult groin hernia is high [pooled PPV of 85.6% (95% confidence interval, 76.5-92.7%)]. • US has very high performance in correctly differentiating between clinically occult inguinal and femoral hernia (correctness of 94.4- 99.1%).