Forty-four studies on pre-eclampsia (n=169,637 participants) including 35 cohort studies and nine case-control studies. Patient selection criteria, description of index and reference tests and blinding of the reference standard were poorly reported. The case-control studies were excluded from the meta-analysis.
Eighty-six studies assessed small for gestational age (n=382,005) including 61 cohort studies and 25 case-control studies. Only 40 studies included an adequate description of the index test. Performance of the reference standard, blinding of the reference standard and use of treatment between the index test and reference standard were poorly reported. The case-control studies and eight studies in which thresholds for small for gestational age were not defined were excluded from the meta-analysis.
Prediction of pre-eclampsia: All but one of the tests evaluated showed poor diagnostic performance for the prediction of pre-eclampsia, especially for ruling out pre-eclampsia. Results were reported for the most accurate threshold:
Maternal serum alpha fetoprotein greater than 2.0 multiples of median (MoM): positive likelihood ratio 2.36 (95% confidence interval (CI): 1.46 to 3.83); negative likelihood ratio 0.96 (95% CI: 0.95 to 0.98); 16 studies, 10 contributed to pooled estimate.
Maternal serum human chorionic gonadotrophin greater than 2.0MoM; positive likelihood ratio 2.45 (95% CI: 1.57 to 3.84); negative likelihood ratio 0.89 (95% CI: 0.83 to 0.96); 21 studies, 11 contributed to pooled estimate.
Maternal serum unconjugated estriol greater 0.5 MoM: positive likelihood ratio 1.50 (95% CI: 1.02 to 2.19); negative likelihood ratio 0.99 (95% CI: 0.97 to 1.00); four studies, two contributed to pooled estimate.
Maternal serum pregnancy associated plasma protein A less than 5th centile: positive likelihood ratio 2.10 (95% CI: 1.57 to 2.81); negative likelihood ratio 0.95 (95% CI: 0.93 to 0.98); (16 studies, five contributed to pooled estimate).
Maternal serum inhibin A greater than 2.79 MoM: positive likelihood ratio 19.52 (95% CI: 8.33 to 45.79); negative likelihood ratio 0.30 (95% CI: 0.13 to 0.68); (six studies, one study contributed to summary estimate).
Performance was considerably worse for all other thresholds evaluated with summary positive likelihood ratio <8.
Prediction of small for gestational age: Tests also showed poor performance for the prediction of small for gestational age. Results were reported for the most accurate threshold for predicting birth weight <10th centile:
Maternal serum alpha fetoprotein less than 10th centile: positive likelihood ratio 8.80 (95% CI: 5.57 to 13.91); negative likelihood ratio 0.02 (95% CI: 0.00 to 0.34); 30 studies, one study contributed to summary estimate.
Maternal serum human chorionic gonadotrophin greater than 2.0MoM: positive likelihood ratio 1.74 (95% CI: 1.48 to 2.04); negative likelihood ratio 0.95 (95% CI: 0.93 to 0.96); 22 studies, seven studies contributed to pooled estimate.
Maternal serum unconjugated estriol less than 0.75 MoM: positive likelihood ratio 2.54 (95% CI: 1.54 to 4.19); negative likelihood ratio 0.75 (95% CI: 0.63 to 0.89); seven studies, two studies contributed to pooled estimate.
Maternal serum pregnancy associated plasma protein A less than 1st centile: positive likelihood ratio 4.36 (95% CI: 3.27 to 5.80); negative likelihood ratio 0.97 (95% CI: 0.96 to 0.98); 10 studies, one study contributed to pooled estimate.
Maternal serum inhibin A greater than 2.0 MoM: positive likelihood ratio 4.45 (95% CI: 3.92 to 5.06); negative likelihood ratio 0.92 (95% CI: 0.91 to 0.93); one study.
Triple test (serum alpha fetoprotein, human chorionic gonadotrophin and unconjugated estriol: positive likelihood ratio 1.07 and 2.71; negative likelihood ratio 0.98 and 1.19; two studies.
Subgroup analysis was only possible for incidence of disease and no significant differences were found between any of the subgroups investigated.