Accuracy of liver function tests for predicting adverse maternal and fetal outcomes in women with preeclampsia : a systematic review

Background. Liver function tests are routinely performed in women as part of a battery of investigations to assess severity at admission and later to guide appropriate management. Objective. To determine the accuracy with which liver function tests predict complications in women with preeclampsia by a systematic review. Data. We conducted electronic searches without language restrictions in (1951-2010), (1980-2010) and the Cochrane Library (2009). Methods of Study Selection. Primary articles that evaluated the accuracy of liver function tests in predicting complications in women with preeclampsia were chosen. Data was extracted by two reviewers independently. A bivariate model estimated area under the curve, sensitivity and specificity. Results. There were 13 primary articles including a total of 3 497 women assessing maternal (30 2x2 tables) and fetal (19 2x2 tables) outcomes. For predicting adverse maternal outcome, the point estimates of specificity were >70% in 18 tables with 0.79 (95% CI 0.51, 0.93). For predicting adverse fetal outcomes the specificity of the test was >70% in 2x2 tables. Sensitivity of the test was poor for both maternal and fetal outcomes. Conclusion. In women with preeclampsia, function tests performed better in predicting adverse maternal than fetal outcomes. The presence of increased liver enzymes was associated with an increased probability of maternal and fetal complications, but normal liver enzyme levels did not rule out disease, as specificity was often higher than sensitivity.