Can middle cerebral artery Doppler predict adverse outcomes in growth-restricted fetuses? A systematic review and meta-analysis.

BACKGROUND

Intrauterine growth restriction (IUGR) is a major contributor to perinatal morbidity and mortality. While umbilical artery (UA) and cerebroplacental ratio are established in surveillance, the isolated prognostic value of the middle cerebral artery pulsatility index (MCA PI) remains uncertain.

OBJECTIVES

To evaluate the diagnostic accuracy of MCA PI for predicting adverse perinatal outcomes in IUGR pregnancies and to compare its performance with UA Doppler.

METHODS

We performed a PRISMA-based systematic review and meta-analysis. PubMed/MEDLINE, Embase, and Web of Science were searched from inception to May 2025; only English-language studies were included. We reconstructed 2 × 2 tables and pooled sensitivity/specificity using a bivariate random-effects HSROC (Reitsma) model (AUC derived from the fitted curve). We also derived LR+/LR- and post-test probability curves across plausible prevalence values, assessed influence with leave-one-out ΔAUC, and repeated analyses excluding the most influential studies. In paired cohorts, MCA vs UA was compared using the relative diagnostic odds ratio (rDOR; MCA/UA) with the same exclusions applied to both tests.

RESULTS

Thirteen studies (n = 2983) were included. For MCA PI predicting the composite adverse outcome (11 studies), pooled sensitivity was 0.553 and specificity 0.664, with AUC 0.642 (95% CI 0.517-0.735). After excluding two influential studies, pooled sensitivity and specificity were 0.654 and 0.590, with AUC 0.643 (95% CI 0.535-0.741) (9 studies). Pooled likelihood ratios suggested modest shifts in post-test probability (primary: LR+ 1.53 [95% CI 1.08-2.16], LR- 0.70 [0.51-0.97]; sensitivity: LR+ 1.77 [1.29-2.44], LR- 0.61 [0.43-0.87]). In paired cohorts (8 studies), overall discrimination was similar (HSROC AUC 0.657 for UA vs 0.658 for MCA), with rDOR 0.995 (95% CI 0.172-5.773); excluding the same two influential cohorts from both tests yielded rDOR 0.570 (95% CI 0.261-1.243).

CONCLUSION

In IUGR pregnancies, isolated MCA PI shows moderate accuracy for predicting a heterogeneous composite adverse outcome. Likelihood-ratio-based translation indicates modest post-test probability changes across plausible prevalence values. Evidence for individual endpoints (e.g., 5-minute Apgar <7) was limited and estimates were imprecise due to few contributing studies. Further prospective studies with standardized outcome definitions, prespecified MCA PI thresholds/timing, and consistent 2 × 2 reporting are needed.