Prediction of Maternal Acute End Organ Injury or Death Using Pre-Pregnancy and Early Pregnancy Factors: Population-Based Study in Ontario

Introduction: Severe maternal morbidity or mortality affects 1-2 in 100 births in North America. We attempted to improve the early prediction of maternal end-organ injury or death using commonly available data.

Methods: This population-based cohort study used linked administrative health data in Ontario, Canada, April 1, 2006 to March 31, 2014. We included women aged 18-60 years with a livebirth or stillbirth. The primary outcome was a composite of maternal end-organ injury or death arising between 20 weeks’ gestation and 42 days after birth. A clinical prediction model (CPM) was derived for the primary outcome, based on early- and pre-pregnancy variables. A second CPM, limited to multiparous women, also further used variables from a previous pregnancy. CPM discrimination was expressed as C-statistic and 95% CI.

Results: Of 634,290 births included in the study, 1969 women experienced the primary outcome (3.1 per 1000 births), including 62 deaths (0.10 per 1000). Variables significantly predictive of end-organ injury or death included maternal world region of origin, chronic medical conditions (e.g., renal disease), nulliparity, and obstetrical/perinatal issues (e.g., abnormal placentation) – with an overall model C-statistic of 0.67 (95% CI 0.66-0.68). The C-statistic predicting maternal death was 0.70 (95% CI 0.66-0.74). Among multiparas, C-statistic was 0.61 (95% CI 0.59-0.62) when limited variables in the index pregnancy alone, rising to 0.69 (95% CI 0.67-0.70) adding pre-pregnancy predictors, and 0.71 (95% CI 0.69-0.73) also including variables from a previous pregnancy.

Conclusion: A CPM using information commonly available by early pregnancy may aid in identifying women appropriate for targeted preventive, monitoring and specialty referral strategies. Enhancement of this CPM using direct clinical measures, and by external validation, is recommended.