To determine whether predefined maternal early warning triggers (MEWTs) can predict pregnancy morbidity.In a retrospective case–control study, obstetric patients admitted to the intensive care unit (ICU) between 2012 and 2013 at seven pilot US hospitals were compared with control patients who had a normal delivery outcome. Six MEWTs were assessed.The case and control groups each contained 50 patients. Hemorrhage (15/50, 30%), sepsis (12/50, 24%), cardiac dysfunction (8/50, 16%), and pre-eclampsia (6/50, 12%) were the most common reasons for ICU admission. Significant associations were recorded between ICU admission and tachycardia (OR 5.0, 95% CI 2.1–11.7), mean arterial pressure less than 65 mm Hg (OR 4.5, 95% CI 1.9–10.8), temperature of at least 38 °C (OR 44.1, 95% CI 13.0–839.1), and altered mental state (OR 44.1, 95% CI 13.1–839.0). Two or more triggers were persistent for 30 minutes or more in 36 (72%) ICU patients versus 2 (4%) controls (OR 61.7, 95% CI 13.2–288.0). Earlier medical intervention might have led to a lesser degree of maternal morbidity for 31 (62%) ICU patients with at least one MEWT.Persistent MEWTs were present in most obstetric ICU cases. Retrospectively, MEWTs in this cohort seemed to separate normal obstetric patients from those for whom ICU admission was indicated; their use might reduce maternal morbidity.