Cesarean and maternal mortality: Risk factor and indication bias

To evaluate the influence of cesarean section on maternal death and discern the effect of indication bias on this association.

A retrospective analysis of case series. All deaths confirmed by the official Epidemiological Surveillance Committee as maternal deaths and involving termination of pregnancy by cesarean section in Antioquia, Colombia, during the period 2011-2020 were included. Two researchers independently analyzed all the records, and, by consensus, classified the relationship as directly related, associated, and no association.

The maternal mortality rate for Antioquia was 40.7 deaths per 100 000 live births over the period studied, 48.6 per 100 000 cesarean sections, and 17.9 per 100 000 vaginal deliveries, a crude relative risk of cesarean of 2.7 (95% CI 2.1-3.7); 58.9% of deaths had no association, 13.5% were directly related, and 21.5% were associated with cesarean section. In 6.5% of cases, the association could not be assessed. There is an overlap between the maternal conditions and indications for cesarean section and the assessment of the association between death and cesarean section. Reasons for considering no association were that the underlying cause of death occurred after cesarean section (2), occurred in perimortem women (6), or occurred in women who were critically ill due to obstetric complications (2), underlying medical conditions (26), or pregnancy-related conditions (27).

To control for indication bias, it is necessary to identify the influence of cesarean section on the chain of events leading to death in each case because there is no predictable and consistent pattern of indications for cesarean section that might also cause death.