Does Adjusting for Causal Intermediate Confounders Resolve the Perinatal Crossover Paradox?

Wen Wei Loh, Cande V. Ananth

Adjusting for causal intermediates such as abruption, SGA, and chorioamnionitis is inadequate to eliminate unmeasured PTB-perinatal mortality confounding. The paradox of preeclampsia's protective direct effect on mortality remains unresolved. Sensitivity analyses to unmeasured confounding are effective in bolstering conclusions from causal mediation analyses and should be more widely applied.