A late maternal death is defined as the death of a woman from direct or indirect obstetric causes, more than 42 days but less than one year after the end of pregnancy. Late maternal deaths are often classified into four main categories: cardiovascular causes, thromboembolism, cancer, and suicide. Pregnancy can trigger or aggravate disease, or cause specific diseases, such as peripartum cardiomyopathy, which typically presents 1-3 months postpartum. Pre-existing mental disorders, the need to cope with the demands of caring for a newborn baby, pregnancy loss, and intimate partner violence can also increase a woman’s vulnerability to suicide risk.
To examine reported late maternal deaths we obtained data on deaths by ICD-10 code from the WHO Mortality Database, a compilation of mortality data by age, sex, and cause of death, as reported annually by countries from their civil registration systems. We identified late maternal deaths using the ICD-10 codes O96 (Death from any obstetric cause occurring more than 42 days but less than one year after delivery) and O97 (Death from sequelae of obstetric causes). We restricted the dataset to country/year estimates between 1995-2017 with non-zero maternal deaths, totalling 1484 observations from 114 countries.
According to these data, late maternal deaths accounted for an increasing proportion of reported global maternal deaths, rising to over 15% by 2017. This increase may reflect improved identification of late maternal deaths over time. We also calculated the proportion of late maternal deaths (out of all maternal deaths) by income group (not enough data were available for low income countries). We see a gradient by income group, with a higher proportion of late maternal deaths in high income countries, likely because the risk of death from other major obstetric complications is lower.
A study of seven countries in the Americas using the regional mortality database of PAHO between 1999 and 2013 found that the proportion of late maternal deaths and deaths from sequelae of obstetric causes ranged between 2.40% (CI 0.85% – 5.48%) and 18.68% (CI 17.06% – 20.47%).
Other reported estimates vary by setting:
Argentina: A study found higher rates of underreporting for late maternal deaths (15.4% under-registration, n=26) than for early maternal deaths (9.5% under-registration, n=95).
Australia: A study identified 173 maternal deaths in 1994-2001, of which 97 were classified as occurring within 42 days of pregnancy ending, and 76 were classified as occurring between 43 and 365 days after pregnancy. Most (73/76) of these deaths were classified as indirect maternal deaths with the most common causes of deaths reported as suicide (n=23), cardiac disorders (n=16) or accident/violence (n=16). Another study found that of 129 reported maternal deaths, 37 were early deaths and 92 occurred late, with 67% of deceased women having a mental health diagnosis or issue related to substance abuse, and trauma (including suicide, accidental injury, vehicle accidents, and homicides) accounting for 73% of all maternal deaths (early and late).
Brazil: A population-based study including all the late maternal deaths from the case series of the Maternal Mortality Committees in the city of São Paulo and the state of Paraná found that late maternal deaths accounted for 13.4% of all maternal deaths in São Paulo and 12.1% in Paraná, with direct obstetric causes accounting for 32.1% of late maternal deaths in São Paulo and 42.1% in Paraná. Death occurred between 43 and 69 days postpartum in 44% of the cases in São Paulo and 39.5% in Paraná.
Canada: A study of late maternal deaths in Canada (excluding Quebec and Manitoba) estimated that the mortality rate was 9.3 per 100,00 deliveries (95% CI 8.5-10.3), and reported the following late maternal mortality rates by cause:
|Diagnosis||Rate per 100,000 deliveries (95% CI)|
|Diseases of circulatory system||2.08 (1.63-2.67)|
|Diseases of nervous system||1.62 (1.22-2.14)|
|Diseases of respiratory system||1.59 (1.20-2.11)|
|Injury, poisoning, external causes of death||1.55 (1.17-2.07)|
|Diseases of blood and blood-forming organs||1.49 (1.11-1.99)|
|Endocrine, nutritional and metabolic diseases, immunity disorders||0.99 (0.69-1.42)|
|Infectious and parasitic diseases||0.86 (0.59-1.26)|
Another study in Canada found that of 120 late maternal, 16 had direct obstetric causes (including 9 deaths by suicide), while 100 were due to indirect obstetric causes (including 46 related to malignancy), and 4 were due to undetermined causes.
Georgia: A study found 23 maternal deaths, of which 15 (65.2%) were early and 8 (34.8%) were late maternal deaths.
Italy: A study identified 277 early maternal deaths and 543 late maternal deaths, the majority of which (38.8%) were caused by malignant neoplasms, followed by violent deaths (suicide: 10.0%, homicide: 3.7%, other: 2.0%), road accidents (9.6%), cardiac deaths (8.9%), and neurological deaths (3.9%).
Jamaica: Of 903 pregnancy-related deaths identified in 1998-2015, 81% were early maternal deaths (n=735) and 19% were late maternal deaths (n=168), with the most common causes of deaths being malignancy, cardiovascular disease, HIV/AIDS, and cardiomyopathy.
Kenya: A study in two Nairobi slums estimated a maternal mortality rate of 706 per 100,000 live births based on 29 maternal deaths, and also identified 22 late maternal deaths, implying a late maternal mortality rate of around 530 per 100,000 live births.
Mali: A cohort study with complete follow-up data for 4,717 women recorded 15 maternal deaths, yielding an overall MMR of 327, and identified 5 additional late maternal deaths.
Mexico: A study in Mexico found that late maternal deaths are more likely to be underreported. Of 9,043 maternal deaths identified by a review process there was a 13% increase in overall identified deaths and a threefold increase in the proportion of maternal deaths classified as late maternal deaths (from 2.1% to 6.9%).
Morocco: A study found that of 69 maternal deaths, 9 were late maternal deaths, of which 5 occurred in a health facility while 4 occurred at home.
South Africa: A study found that 84 women died from ‘early maternal death’, while 100 died from ‘late maternal death’.
Sweden: A study of maternal deaths in 1980-1988 identified 58 early maternal deaths and 76 late maternal deaths, predominantly caused by malignancy, stroke, heart disease, and suicide.
Tanzania: A study found that 111 women died within 42 days of childbirth, while 229 maternal deaths occurred within 365 days, suggesting that half of maternal deaths were late maternal deaths - however, these estimates exclude deaths during pregnancy.
USA: A study in Colorado found that of maternal deaths in 2011, 30% resulted from self-harm, with overdose and suicide leading causes. Deaths were equally distributed throughout the first postpartum year.
Although most of the estimates we identified reported the proportion of maternal deaths due to late maternal deaths, we did obtain some estimates of the incidence of late maternal deaths, which varied by setting. Here we summarize the reported (or inferred) estimates.
|Estimate (per 100,000 births)||Country||Source|
Estimates of case fatality rates are scarce. However, a study in South Africa of 269 peripartum women presenting with CVD in pregnancy or within 6-months postpartum estimated a CFR of 5.9% in Group 1 (treated by a dedicated cardiac-obstetric) team, and 0.9% in Group 2 (treated with additional interventions, including early postpartum follow, referral to dedicated specialist clinics, and increased beta-blocker therapy). Mortality risks for women not enrolled in a trial are likely higher, and may be higher for other causes of late maternal death as well.
We model the probability of late maternal death for all causes combined (both direct and indirect), which likely vary by country. Estimated mortality rates from late maternal deaths vary from 5.4 to 530 per 100,000 births, and estimates of CFR from CVD with specialist care ranged between 1-5%. Assuming a higher CFR of around 10% for non-specialist care, we set wide priors for incidence, centered around 1% incidence (1 per 100 births), which we model over 10 months. We assume that late maternal deaths are uniformly distributed during the year following pregnancy, consistent with data from Metz 2016. We assume that the incidence of late maternal events decreases by income group. Given an incident late maternal complication, we simulate mortality risks conditional on the site of treatment (accounting for referral probabilities) and quality of care. While there is likely substantial heterogeneity in CFR by cause (e.g. suicide attempt versus neoplasm), for simplicity we model one aggregate CFR, and set priors centered around 10% for women at Home, decreasing to around 1% for women in CEmOC facilities. We assume that case fatality rates decrease with higher level delivery sites and enforce these constraints when sampling parameters (Home > Home-SBA > non-EmOC > BEmOC > CEmOC).
We use a monthly cycle in the model, so we model ‘early maternal deaths’ as occurring within 2 months of delivery. For the next 10 months we simulate the risk of late maternal death, with a constant risk per month. If a late complication/event occurs, the risk of mortality is simulated based on a woman’s treatment site, accounting for referral probabilities and quality of care.
GMatH (Global Maternal Health) Model - Last updated: 28 November 2022
© Copyright 2020-2022 Zachary J. Ward