Abstract
Objective:
To estimate the association between pregnancy-associated death or severe maternal morbidity and infant outcomes.
Methods:
We conducted a retrospective cohort study using Massachusetts statewide data from 1999-2021. The dataset included hospital records longitudinally linked to births and maternal and infant death records. The primary exposures were pregnancy-associated death (deaths during pregnancy or in the year postpartum), severe maternal morbidity and pregnancy-associated death following severe maternal morbidity. The main outcomes were infant death in the first year and, for those infants who survived, hospitalization in their first year of life. Bivariate and robust Poisson regression analysis via generalized estimating equations regression were used to estimate the association between the exposures and outcomes.
Results:
Of 1,617,054 live births in Massachusetts between 1999 and 2020, there were 474 pregnancy-associated deaths. Pregnancy associated death ratios were highest among those 40 years or older (49.3 per 100,000), who were non-Hispanic Black (43.0 per 100,000), who used public insurance (51.1 per 100,000) or who had a parity greater than or equal to 4 (80.6 per 100,000). Among those experiencing severe maternal morbidity (745.3 per 100,000), those who had a hospital encounter associated with opioid use (721.2 per 100,000) or a documented prepregnancy comorbidity (200.7 per 100,000) had the highest pregnancy-associated death ratios. In cases of pregnancy-associated death, the infant mortality rate per 1,000 live births was 55.0 (95% CI 34.9-75.2) compared with 4.0 (95% CI 3.9-4.1) when the mother survived. When the pregnancy-associated death followed severe maternal morbidity, the infant mortality rate was 87.9 per 1,000 live births (95% CI 29.7-146.1). Following a pregnancy-associated death, when a full-term infant survived to age 1, there was a 35% greater likelihood of rehospitalization in the first year of life (aRR 1.35; 95% CI 1.01-1.82) than when the mother did not die.
Conclusion:
Pregnancy-associated death was associated with infant death and worse health of surviving children in the first year of life, further demonstrating the far-reaching consequences of maternal deaths, and the clear link between maternal and infant health.
Precis
Pregnancy-associated death and severe maternal morbidity were associated with infant death and, when infants survived, hospitalization in the first year of life.
Introduction
Aside from a surge during the COVID pandemic, pregnancy-related death ratios in the U.S. have remained relatively steady in the past decade,1 albeit at ratios far above those of other wealthy countries.2 Severe maternal morbidity (SMM), a measure used to identify serious but non-fatal complications encompassing 20 different critical conditions that can be documented in hospital records at birth, has been increasing steadily in the U.S. since 2010.3 While these disturbing women’s health outcomes have drawn considerable attention from the clinical community4, policymakers5 and the public,6 far less attention has been paid to the relationship between pregnancy- associated deaths and the outcomes of the infants born from those pregnancies.
The U.S. has for decades also fared poorly in international comparisons of infant health.7 While there have been a number of studies examining the influence of a maternal death on infant and family health in the global context,8–10 where maternal death ratios are generally much higher, studies in the U.S. have typically focused on maternal and infant mortality as joint outcomes of inadequate care.11,12 To better understand infant outcomes in the U.S. following a pregnancy-associated death or pregnancy complicated by SMM, we used a longitudinally linked population database to examine infant deaths and hospitalizations following these pregnancy outcomes compared with births to a living mother.
Methods
This was a retrospective cohort study using data from the Massachusetts Pregnancy to Early Life Longitudinal (PELL) Data System . Details of this dataset are described elsewhere.13,14 The Massachusetts Department of Public Health (DPH) Institutional Review Board deemed this cohort study exempt from the informed consent requirement because all data were deidentified, and it was not considered human participant research. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
PELL longitudinally links birth and fetal death records to corresponding maternal hospital encounters (admissions, outpatient stays, and emergency department (ED) visits) in Massachusetts. PELL data are housed at the DPH and include over 98% of birth certificates and fetal death certificates (when fetus weighed at least 350 gms or death occurred at 20+ weeks’ gestation) linked to hospital discharge records from January 1, 1998, through December 31, 2021. PELL records are linked using a deterministic and probabilistic matching program (LinkPro version 3.0; InfoSoft Inc)15 and maternal and infant hospital records are linked to birth certificate data on index deliveries according to facility code, medical record number, date of delivery, sex, zip code, and birthweight. To create the longitudinal linkage (before or after delivery), a unique non-missing encrypted Social Security number and a unique non-missing combination (concatenated) of medical record numbers and hospital numbers of mothers and infants are used. This longitudinal link allows the examination of hospital encounters among birthing individuals at any time during the study period, including hospital visits during pregnancy. Maternal records are also linked to state death records as well as birth, death and hospitalization records for their infants. To enhance the likelihood of tracking both prenatal and postpartum hospitalizations, we included only live births to Massachusetts residents in Massachusetts hospital births in our analytic sample.
The primary exposures in this study were SMM and pregnancy-associated death. Pregnancy-associated death was defined as a death that occurred during pregnancy or up to 1 year postpartum regardless of cause.16 The SMM measure was based on an algorithm developed as part of an interagency collaboration between the Health Resources and Services Administration, Centers for Disease Control and Prevention, Agency for Healthcare Research and Quality, and Alliance for Innovation on Maternal Health (version 07-01-2021).4 Since transfusion has been excluded from national SMM measures due to poor specificity in the absence of other SMM indicators,17,18 we excluded transfusion and focused on 20 SMM conditions or procedures identified through International Classification of Diseases, Ninth Revision and International Statistical Classification of Diseases and Related Health Problems, Tenth Revision codes across the study period (Appendix 1, available online at http://links.lww.com/xxx).19 We examined instances of SMM during pregnancy and the delivery hospitalization.13
We applied the standard definition of pregnancy-associated death ratio, which includes deaths during pregnancy and up to one year postpartum over total births in the year of death.20 Since we were examining the relationship between a mother’s death and infant survival and health, we opted to use the broader category of pregnancy-associated deaths rather than limiting cases to pregnancy-related deaths.
As noted above, the majority of the cohort consisted of live births from 1999-2020. However, we included 8 cases in which a birth occurred in 1998, but the mother died in 1999 for the calculation of the pregnancy-associated death ratio over the time period of interest from 1999-2020. Our measure of SMM included conditions that appeared during pregnancy,13 and therefore we began our study period for all other cases in 1999 to allow us to review hospitalizations during pregnancy and likewise, since we were interested in infant hospitalizations in the year following the birth, our study period ends in 2020. Finally, two key variables (comorbidity and opioid use) involved hospital encounters in the year prior to pregnancy and in those cases, the study period was limited to 2000-2020. Data related to ED visits became available in 2002.
Our outcomes were infant mortality rates, defined as the number of deaths to live born infants before their first birthday per 1,000 live births,21 and, for those infants who did not die, re-hospitalization in the first year of life. Infant death records included age at death and cause of death while hospitalization records included age at rehospitalization and diagnosis or treatment associated with hospitalization. We utilized the hospital discharge data in PELL to document the timing of and reasons for infant hospitalizations after a pregnancy-associated death. The pregnancy period was defined by subtracting the obstetrical estimate of gestational age from the date of birth. We ascertained demographic and perinatal health characteristics of pregnant women available on birth certificates including prepregnancy or gestational diabetes. Race and ethnicity (Hispanic, non-Hispanic Black, non-Hispanic White) were self-identified by birthing individuals as part of the parent’s worksheet for the birth certificate. A category of non-Hispanic other (non-Hispanic Asian or Pacific Islander, - non-Hispanic American Indian and Alaskan Natives, and unknown) included groups with numbers of infant deaths too small to analyze separately. Race and ethnicity data were collected because of their demonstrated relationship with maternal outcomes. Our measure of comorbidities in the year prior to pregnancy was based on the algorithm developed by Bateman and colleagues.22 Given Massachusetts’ experience with opioid misuse as a major contributor to pregnancy associated deaths,23 we excluded the specific opioid use codes from the comorbidity measure and a separate variable was created based on any hospital encounter (admission, observational stay or emergency room visit) with opioid use,24 in the year prior to a pregnancy.
The analysis was conducted using SAS, version 9.4 (SAS Institute Inc). Given that outcomes were relatively rare, we used robust Poisson regression analysis via generalized estimating equations with exchangeable working correlation to take into account multiple deliveries by the same women, and as the primary means to assess the association of pregnancy-associated deaths, SMM and infant hospitalizations in models that were unadjusted and adjusted for relevant covariates chosen a priori based on the literature. The covariates included maternal age, race and ethnicity25, education, health insurance at delivery26, parity, chronic or gestational diabetes20 and whether there was a hospital encounter for comorbidity22 or opioid use in the year prior to pregnancy. Given the central role of gestational age at birth in infant health, we stratified the final model by prematurity and included gestational age in the stratified models. We generated unadjusted and adjusted risk ratios (aRRs) with 95% confidence intervals (CI) from these models.
Results
Overall, 1,617,054 Massachusetts live births were eligible for analysis between 1999 and 2020 (Appendix 2, available online at http://links.lww.com/xxx). These births involved 491 live births to 474 women who died during pregnancy or up to 1 year after birth from 1999-2020, including 8 who died in 1999 but had given birth in 1998.
The characteristics of the women and infants involved in these deliveries are presented in Table 1 along with the pregnancy-associated mortality ratio and infant mortality rate by subgroup. The overall pregnancy-associated death ratio (per 100,000 live births) was 29.3 while the infant mortality rate (per 1,000 live births) was 4.06. When compared with individuals who delivered and did not have a pregnancy-associated death, pregnancy-associated death was more common following a diagnosis of SMM (745.3 per 100,000 live births), a history of opioid use (721.2 per 100,000) or other comorbidity (200.7 per 100,000) (all p < .01). In terms of demographic characteristics, the pregnancy-associated mortality ratio was highest among those with a parity of 4 or greater (80.6 per 100,000 live births), who used public insurance (51.1 per 100,000), who were over 40 years of age (49.3 per 100,000), who were non-Hispanic Black race ethnicity (43.0 per 100,000), who had chronic or gestational diabetes (41.9 per 100,000) and who had less than a college degree (41.7 per 100,000). Infant mortality followed a similar pattern regarding maternal characteristics with the highest rates in births to a mother with SMM (15.69 per 1,000 live births), a history of opioid use (10.98 per 1,000), who was non-Hispanic Black (8.22 per 1,000), experienced a pregnancy hospitalization for a comorbidity (7.54 per 1,000) or was less than 25 years old (5.82 per 1,000).
Table 1.
Pregnancy Associated Death Ratios and Infant Mortality Rates by Subgroups, Massachusetts, 1999-2020.
| Total Live Births 1999-2020 | Pregnancy Associated Deaths (PAD) | Pregnancy Associated Mortality Ratios (Per 100,000 live births) | Infant Deaths | Infant Mortality Rates (Per 1,000 live births) | |
|---|---|---|---|---|---|
| Overall Cohort | 1,617,054 | 474 | 29.31 | 6,566 | 4.06 |
| MATERNAL | |||||
| Race and Ethnicity | |||||
| Hispanic | 251,683 | 75 | 29.80 | 1,306 | 5.19† |
| Non-Hispanic Black | 144,188 | 62 | 43.00† | 1,185 | 8.22† |
| Non-Hispanic White* | 1,058,329 | 295 | 27.87 | 3,479 | 3.29 |
| Other NH / Unknown | 162,854 | 42 | 25.79 | 596 | 3.66† |
| Maternal Age | |||||
| <25 | 308,958 | 99 | 32.04 | 1,798 | 5.82† |
| 25-29* | 388,262 | 113 | 29.10 | 1,573 | 4.05 |
| 30-34 | 541,192 | 134 | 24.76 | 1,723 | 3.18† |
| 35-39 | 307,614 | 93 | 30.23 | 1,086 | 3.53† |
| 40+ | 71,028 | 35 | 49.28† | 386 | 5.43† |
| Maternal Education | |||||
| < College (≤ HS, GED, some college) | 891,446 | 372 | 41.73† | 4,516 | 5.07† |
| College and Above* | 710,469 | 86 | 12.10 | 1,864 | 2.62 |
| Unknown | 15,139 | 16 | 105.69† | 186 | 12.29† |
| Health Insurance at Delivery | |||||
| Private* | 932,074 | 134 | 14.38 | 2,922 | 3.13 |
| Public | 653,064 | 334 | 51.14† | 3,486 | 5.34† |
| Self pay or unknown | 31,916 | 6 | 18.80 | 158 | 4.95† |
| Parity | |||||
| 1* | 718,359 | 154 | 21.44 | 2,999 | 4.17 |
| 2 | 555,014 | 139 | 25.04 | 1,907 | 3.44† |
| 3 | 225,648 | 89 | 39.44† | 918 | 4.07† |
| 4+ | 112,943 | 91 | 80.57† | 714 | 6.32† |
| Unknown | 5,090 | 1 | 19.65 | 28 | 5.50 |
| Comorbidity in year prior to pregnancy (exclude opioid use)‡ | |||||
| No* | 1,484,400 | 344 | 23.17 | 5,808 | 3.91 |
| Yes | 53,304 | 107 | 200.74† | 402 | 7.54† |
| Opioid use 1 year prior to pregnancy‡ | |||||
| No* | 1,531,326 | 405 | 26.45 | 6,140 | 4.01 |
| Yes | 6,378 | 46 | 721.23† | 70 | 10.98† |
| Chronic or gestational diabetes | |||||
| No* | 1,500,107 | 425 | 28.33 | 6,151 | 4.10 |
| Yes | 116,947 | 49 | 41.9† | 415 | 3.55† |
| SMM (during pregnancy orbirth) | |||||
| No* | 1,605,649 | 389 | 24.23 | 6,387 | 3.98 |
| Yes | 11,405 | 85 | 745.29† | 179 | 15.69† |
| INFANT | 491 | ||||
| Plurality | |||||
| Singleton* | 1,548,757 | 456 | 5,152 | 3.33 | |
| Multiple | 68,297 | 35 | 1,414 | 20.70† | |
| Gestational Age | |||||
| <= 32 weeks | 34,531 | 55 | 4,101 | 118.76† | |
| 33-34 weeks | 29,440 | 26 | 248 | 8.42† | |
| 35-36 weeks | 79,839 | 50 | 389 | 4.87† | |
| 37-38 weeks | 344,796 | 130 | 663 | 1.92† | |
| 39 weeks* | 471,363 | 102 | 499 | 1.06 | |
| 40+ weeks | 657,085 | 128 | 666 | 1.01† | |
| 5-minute Apgar Score | |||||
| 0-3 | 5,982 | 18 | 2,961 | 494.98† | |
| 4-6 | 16,306 | 25 | 888 | 54.46† | |
| 7-10* | 1,588,465 | 441 | 2,571 | 1.62 | |
| Unknown | 6,301 | 7 | 146 | 23.17† | |
| Birthweight | |||||
| <1500 gms | 19,577 | 28 | 3,721 | 190.07† | |
| 1500-2499 gms | 100,298 | 95 | 848 | 8.45† | |
| 2500-3999 gms* | 1,332,906 | 327 | 1,747 | 1.31 | |
| 4000+ | 159,739 | 41 | 133 | 0.83† | |
| Unknown | 4,534 | 0 | 117 | 25.81† |
Reference group. Bivariate Generalized Estimating Equations (GEEs) were used to assess the associations between characteristics and pregnancy associated woman deaths, as well as between characteristics and infant deaths.
p< 0.01
2000-2020 to allow examination of pre-pregnancy hospital contacts
In terms of infant variables, more than one-fourth (26.7%) of infants born in the setting of pregnancy-associated deaths were premature compared with 8.8% when the mother survived and one in nine (11.2%) of the pregnancy associated deaths involved a birth at less than 33 weeks compared with 2.1% among all births. Infant mortality was associated with birth at 32 weeks or less (118.76 deaths per 1,000 live births) and multiple gestation (20.70 per 1,000 live births) as well as low birth weight and lower Apgar scores. Seven of the 27 infant deaths occurred prior to the mother’s death but following a live birth, while an additional ten occurred on the same day. Only 3 infant deaths occurred more than a week after the mother’s death (Appendix 3, available online at http://links.lww.com/xxx).
We explored the relationship between SMM, pregnancy-associated death and infant death in Table 2. In cases where the mother survived and did not experience severe morbidity, the infant mortality rate was low (4.0 per 1,000 births; 95% C.I. 3.9-4.1). If SMM followed by a pregnancy-associated death occurred, the infant mortality rate was almost 22 times higher at 87.9 (95% CI 29.7-146.1). If only SMM were present, the infant mortality rate was 15.1 (95% C.I. 12.9-17.4), while if there were a pregnancy-associated death without SMM, the rate was 47.5 (95% C.I. 26.7-68.3). The small number of infant deaths precluded a multivariable analysis of these outcomes.
Table 2.
Infant Deaths (per 1,000 live births) Related to Severe Maternal Morbidity and Pregnancy-Associated Deaths, Massachusetts, 1999-2020
| Exposure | Infant Deaths (per 1,000 livebirths with 95% CI) | Number of infant deaths |
|---|---|---|
| Unexposed (n=1,605,257) | 4.0 (3.9-4.1) | 6,369 |
| SMM only (n=11,314) | 15.1 (12.9-17.4) | 171 |
| Pregnancy-associated death only (n=400) | 47.5 (26.7-68.3) | 19 |
| SMM followed by pregnancy-associated death (n=91) | 87.9 (29.7-146.1) | 8 |
| Pregnancy associated death overall (n=491) | 55.0 (34.8-75.2) | 27 |
| SMM Overall (n=11,405) | 15.7 (13.4-18.0) | 179 |
We also summarized the causes of infant death (day of life 0-365) when there was a pregnancy-associated death (Appendix 4, available online at http://links.lww.com/xxx). Among the 15 neonatal deaths (day of life 0-28), 11 occurred following delivery at less than 34 weeks’ gestation. Hypoxic ischemic encephalopathy and congenital malformation of the heart were the only specific causes of death listed more than once among premature infants. Among the 12 post-neonatal deaths (day of life 29-365), 4 infants had an “ill defined and unknown cause,” one was associated with birth asphyxia, one involved a congenital malformation, while the remainder of deaths were injury-related, either by fire (n=3), assault (n=2) or a motor vehicle accident (n=1). (Appendix 4, available online at http://links.lww.com/xxx)
Overall, 8.3% of infants whose mother did not experience a pregnancy-associated death had a hospital admission in the first year of life, while 15.1% of the infants whose mother experienced a pregnancy-associated death were hospitalized (Table 3). After controlling for demographic and medical risk factors selected a priori excluding gestational age, infants of women who died experienced a 44% greater risk of hospitalization in their first year of life (aRR 1.44 95% CI 1.15-1.81) and those born to mothers with severe morbidity a 57% greater risk (aRR 1.57 95% CI 1.49-1.65). Other factors strongly associated with infant hospitalization included maternal prepregnancy comorbidities (aRR 1.24 95% CI 1.21-1.27), giving birth at less than 25 years of age (aRR 1.23; 95% C.I. 1.21-1.25) and parity 4 or higher (aRR 1.21; 95% C.I. 1.18-1.24). When we stratified by prematurity, the relationship between pregnancy-associated death and hospitalization was no longer observed among the preterm births but was still significant (aRR 1.35 95% C.I. 1.01-1.82) among term births. Infants born in the setting of maternal SMM experienced a higher risk of rehospitalization whether born prematurely (aRR 1.12 95% C.I. 1.05-1.19) or at term (aRR 1.10 95% C.I. 1.02-1.18).
Table 3.
Relative risk ratiosa (RR) of hospitalizations in the first year among infants born after a pregnancy-associated death, Massachusetts, 2000-2022
| Total | Hospitalized in the first year of life | Unadjusted RR (95% CI) | Adjusted RR (95% CI)* (overall cohort) | Stratified adjusted RR (95% CI)* (GA < 37 weeks) | Stratified adjusted RR (95% CI)* (GA>=37 weeks) | ||
|---|---|---|---|---|---|---|---|
| N | N | % | |||||
| Total number of infants who survived until age 1 | 1,610,495 | 133,494 | 8.3 | ||||
| Pregnancy-associated Death | |||||||
| No | 1,610,031 | 133,424 | 8.3 | ref | ref | ref | ref |
| Yes | 464 | 70 | 15.1 | 1.82 (1.46-2.27) | 1.44 (1.15-1.81) | 1.01 (0.73-1.39) | 1.35 (1.01-1.82) |
| SMM (pregnancy or birth) | |||||||
| No | 1,599,269 | 131,982 | 8.3 | ref | ref | ref | ref |
| Yes | 11,226 | 1,512 | 13.5 | 1.62 (1.54-1.70) | 1.57 (1.49-1.65) | 1.12 (1.05-1.19) | 1.10 (1.02-1.18) |
| Race and Ethnicity | |||||||
| Hispanic | 250,378 | 26,673 | 10.7 | 1.37 (1.35-1.39) | 1.13 (1.11-1.15) | 0.94 (0.92-0.97) | 1.15 (1.13-1.17) |
| Non-Hispanic Black | 143,005 | 12,064 | 8.4 | 1.09 (1.07-1.11) | 0.94 (0.92-0.96) | 0.76 (0.73-0.79) | 0.93 (0.91-0.95) |
| Non-Hispanic White | 1,054,853 | 81,418 | 7.7 | ref | ref | ref | ref |
| Other NH / Unknown | 162,259 | 13,339 | 8.2 | 1.06 (1.05-1.08) | 1.06 (1.04-1.08) | 0.90 (0.86-0.93) | 1.09 (1.07-1.11) |
| Maternal Age | |||||||
| <25 | 307,162 | 32,477 | 10.6 | 1.39 (1.37-1.42) | 1.23 (1.21-1.25) | 1.16 (1.12-1.20) | 1.28 (1.25-1.30) |
| 25-29 | 386,689 | 32,135 | 8.3 | 1.10 (1.09-1.12) | 1.04 (1.03-1.06) | 1.06 (1.03-1.09) | 1.06 (1.04-1.07) |
| 30-34 | 539,471 | 40,377 | 7.5 | ref | ref | ref | ref |
| 35-39 | 306,530 | 22,970 | 7.5 | 1.00 (0.98-1.02) | 0.99 (0.97-1.01) | 0.96 (0.93-1.00) | 0.96 (0.95-0.98) |
| 40+ | 70,643 | 5,535 | 7.8 | 1.04 (1.02-1.07) | 1.01 (0.99-1.04) | 0.96 (0.91-1.01) | 0.94 (0.91-0.97) |
| Maternal Education | |||||||
| < College (≤ HS, GED, some college) | 886,935 | 83,601 | 9.4 | 1.36 (1.35-1.38) | 1.13 (1.12-1.15) | 1.09 (1.06-1.12) | 1.11 (1.09-1.13) |
| College and Above | 708,607 | 48,670 | 6.9 | ref | ref | ref | ref |
| Unknown | 14,953 | 1,223 | 8.2 | 1.18 (1.12-1.25) | 1.02 (0.96-1.08) | 1.07 (0.96-1.18) | 0.94 (0.88-1.01) |
| Health insurance at delivery | |||||||
| Private | 929,154 | 66,703 | 7.2 | ref | ref | ref | ref |
| Public | 649,583 | 64,694 | 10.0 | 1.37 (1.36-1.39) | 1.14 (1.12-1.15) | 1.11 (1.08-1.15) | 1.14 (1.12-1.15) |
| Self pay/Unknown | 31,758 | 2,097 | 6.6 | 0.92 (0.88-0.96) | 0.92 (0.88-0.97) | 0.96 (0.89-1.04) | 0.85 (0.81-0.90) |
| Parity | |||||||
| 1 | 715,363 | 57,142 | 8.0 | ref | ref | ref | ref |
| 2 | 553,108 | 44,585 | 8.1 | 1.01 (1.00-1.02) | 1.04 (1.03-1.05) | 1.06 (1.03-1.08) | 1.05 (1.04-1.07) |
| 3 | 224,733 | 19,995 | 8.9 | 1.10 (1.09-1.12) | 1.12 (1.10-1.14) | 1.10 (1.06-1.14) | 1.14 (1.12-1.16) |
| 4+ | 112,229 | 11,297 | 10.1 | 1.23 (1.20-1.25) | 1.21 (1.18-1.24) | 1.16 (1.11-1.21) | 1.20 (1.17-1.23) |
| Unknown | 5,062 | 475 | 9.4 | 1.16 (1.07-1.27) | 1.20 (1.10-1.32) | 0.67 (0.57-0.79) | 1.27 (1.15-1.42) |
| Comorbidity in year prior to pregnancy (exclude opioid use) † | |||||||
| No | 1,478,592 | 120,474 | 8.2 | ref | ref | ref | ref |
| Yes | 52,902 | 6,414 | 12.1 | 1.45 (1.41-1.48) | 1.24 (1.21-1.27) | 1.06 (1.01-1.11) | 1.19 (1.16-1.23) |
| Opioid use 1 year prior to pregnancy † | |||||||
| No | 1,525,186 | 126,037 | 8.3 | ref | ref | ref | ref |
| Yes | 6,308 | 851 | 13.5 | 1.57 (1.47-1.68) | 1.13 (1.05-1.21) | 1.01 (0.91-1.13) | 1.05 (0.97-1.15) |
| Chronic or gestational diabetes | |||||||
| No | 1,493,963 | 122,413 | 8.2 | ref | ref | ref | ref |
| Yes | 116,532 | 11081 | 9.5 | 1.15 (1.13-1.17) | 1.15 (1.13-1.17) | 1.01 (0.97-1.05) | 1.07 (1.04-1.09) |
| Gestational Age | |||||||
| <= 32 weeks | 30,430 | 10,195 | 33.5 | 5.30 (5.19-5.40) | 2.32 (2.26-2.37) | ||
| 33-34 weeks | 29,194 | 6,723 | 23.0 | 3.63 (3.54-3.72) | 1.59 (1.55-1.64) | ||
| 35-36 weeks | 79,451 | 11,353 | 14.3 | 2.26 (2.21-2.30) | ref | ||
| 37-38 weeks | 344,133 | 31,514 | 9.2 | 1.44 (1.42-1.47) | 1.42 (1.40-1.44) | ||
| 39 weeks | 470,865 | 32,192 | 6.8 | 1.08 (1.06-1.09) | 1.07 (1.05-1.08) | ||
| 40+ weeks | 656,422 | 41,517 | 6.32 | ref | ref | ||
Crude and adjusted relative risk approximated using general estimating equations with a log link and Poisson distribution to take into account multiple deliveries by the same women; multivariate models adjusted for the variables in the table. Model adjusted for pregnancy-associated death, SMM, maternal age, race ethnicity, education, insurance at delivery, parity, prenatal comorbidity, prenatal opioid hospital utilization, pre-existing or gestational diabetes and gestational age.
Limited to 2000-2020 livebirths
SMM – severe maternal morbidity
GA – gestational age
We described the timing of infant hospitalizations after a pregnancy-associated death with 67.1% occurring in the first month after the birth hospitalization, compared with 49.9% in cases in which the mother survived (data not shown). We compared the reasons for infant hospitalization in the first year among surviving infants of those whose mother did and did not experience a pregnancy-associated death (Appendix 5, available online at http://links.lww.com/xxx) and there was substantial overlap in ICD codes across the two groups with 20 of the top 30 codes found in both groups.
Discussion
This study of infant outcomes associated with pregnancy-associated death and SMM reaffirms the intrinsic relationship between maternal and infant health. Pregnancy associated death is a tragedy in and of itself, but its consequences can extend to a family and community. In this longitudinal, linked statewide study, when a pregnancy- associated death occurred, the likelihood of an infant death was more than 13 times greater than when the mother survived the first year after birth. In deliveries in which there was both a pregnancy-associated death and SMM, the results were even more dire, with an infant death rate almost 22 times greater than in births without either outcome. For those infants who survived the first year of life after their mother’s death, there was a 35% greater risk of hospitalization among term births and a 10% greater risk if their mother experienced one of the conditions or procedures associated with SMM.
Prior studies of infant death have focused on a variety of potential influences, from race and ethnicity,27 to maternal age,28 genetic factors,29 maternal health conditions30 maternal behaviors such as smoking31 or substance use,32 and state policy decisions.33 As noted above, there have also been an array of studies in global health examining the influence of a maternal death on family structure and infant survival. What has not been as widely studied are the implications of a maternal death for infant survival and health in the U.S. context. Our findings are generally consistent with prior research in finding that higher infant mortality rates were associated with older maternal age, being on public insurance, prior health status and opioid use, but both severe maternal morbidity and pregnancy-associated death were far more significant.
This research suggests the need to broaden the scope of the U.S. maternal mortality crisis to emphasize the importance of protecting maternal health to preserve infant health. This is hardly a new conceptualization – in an exhaustive 1925 study of infant mortality in eight cities, researchers found an infant mortality rate four times higher (450 per 1,000) when the mother died than when she survived the year after birth (109 per 1,000) with 72% of the infant deaths in the first month. 34 In the century since that study, access to and quality of prenatal care has increased, providing the opportunity for identification of risk and enhanced care for pregnant women. However, such care often does not carry through into the postpartum period with dire consequences.35 While the potential to prevent both maternal and infant deaths far surpasses what clinicians could have imagined in 1925, the challenge is to manifest that potential in a system that recognizes the most effective way to improve infant health is a determined focus on improving women’s health.
This study is subject to several limitations. It is based on data from a single state which is not necessarily representative of the U.S. Massachusetts annually has the lowest infant mortality rate of any U.S. state for the period 2007-202236 as well as the lowest rate of uninsured in the U.S.37 and hence the experience there may not be characteristic of the nation as a whole. The PELL database relies on the linkage of vital statistics to administrative data and while the linkage rate is high (98%), cases that were not linked may not be randomly distributed. We excluded transfusions and examined SMM during pregnancy and at birth and hence our results may not be comparable to other national studies of SMM that have used the CDC algorithm that includes transfusions, and limit their analysis to SMM at delivery.13 Our data do not include non-birth hospitalizations outside of Massachusetts.
In addition, based on the data available, we were not able establish a direct causal pathway between pregnancy-associated death and infant death, since seven of the infant deaths (out of 27) occurred before the death of the mother and ten were on the same day. In these cases, we were not able to ascertain whether the death of the mother led to the death of the infant or if these infants might have died regardless of the maternal outcome. In reviewing the causes of infant death, it was not possible to disentangle the chain of events to determine which event came first. For example, in three cases, encephalopathy and ischemic encephalopathy38 were listed as the cause of death and in two cases birth asphyxia was the listed as the cause. These infant deaths may have been related to antepartum or intrapartum events unrelated to the maternal death, or may have been downstream consequences of significant maternal illness which also compromised the pregnancy and ultimately resulted in maternal death.
The causes of infant death and hospitalization in these cases reflect the overlapping and shared health outcomes affecting the maternal-infant dyad. When we consider opportunities for intervention, it is critical to underscore the dyad as at risk when advocating for prevention of maternal morbidity and mortality, as well as, adverse infant health outcomes.
Supplementary Material
Funding Source:
NIH/National Inst on Minority Health and Health Disparities R01: MD016026
Footnotes
Financial Disclosure
The authors did not report any potential conflicts of interest.
Each author has confirmed compliance with the journal’s requirements for authorship.
Contributor Information
Eugene Declercq, Boston University School of Public Health.
Chia-Ling Liu, Evalogic Services, Inc.
Howard J. Cabral, Boston University School of Public Health.
Ndidiamaka Amutah-Onukagha, Tufts University School of Medicine.
Sunah Hwang, University of Colorado Anschutz Medical Campus.
Hafsatou Diop, Massachusetts Department of Public Health.
References
- 1.Pregnancy Mortality Surveillance System. CDC, 2025. (Accessed Jume 2, 2025, at https://www.cdc.gov/maternal-mortality/php/pregnancy-mortality-surveillance-data/.) [Google Scholar]
- 2.Insights into the U.S. Maternal Mortality Crisis: An International Comparison. 2024. (Accessed Accessed June 2, 2025,, at https://www.commonwealthfund.org/publications/issue-briefs/2024/jun/insights-us-maternal-mortality-crisis-international-comparison.)
- 3.Reid L Trends in Severe Maternal Morbidity Complications by Patient Characteristics, 2016–2021. Rockville, MD: Agency for Healthcare Research and Quality; 2024. [Google Scholar]
- 4.Mahoney JR. The Alliance for Innovation in Maternal Health Care: A Way Forward. Clinical Obstetrics & Gynecology 2018;61:400–10. doi: 10.1097/GRF.0000000000000363. [DOI] [PubMed] [Google Scholar]
- 5.White House. White House Blueprint for Addressing the Maternal Health Crisis. Washington, D.C. 2022. [Google Scholar]
- 6.Lost Mothers. ProPublica, 2024. (Accessed January 21, 2025, 2025, at https://www.propublica.org/series/lost-mothers.)
- 7.Singh GK, Yu SM. Infant Mortality in the United States, 1915-2017: Large Social Inequalities have Persisted for Over a Century. Int J MCH AIDS 2019;8:19–31. doi: 10.21106/ijma.271. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Naha SK, Arpon MEI, Siddique RT, Ripa FR, Hasan MN, Uddin MJ. A study of association between maternal tetanus toxoid immunization and neonatal mortality in the context of Bangladesh. PLoS One 2025;20:e0316939. doi: 10.1371/journal.pone.0316939. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Zhang J, Li H, Wei B, et al. Association between maternal health service utilization and under-five mortality rate in China and its provinces, 1990-2017. BMC Pregnancy Childbirth 2024;24:326. doi: 10.1186/s12884-024-06437-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Tlou B, Sartorius B, Tanser F. Effect of timing of mother’s death on child survival in a rural HIV hyper-endemic South African population. BMC Public Health 2018;18:1237. doi: 10.1186/s12889-018-6152-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Park RY, Bilinski A, Parks RM, Flaxman S. Trends in Maternal, Fetal, and Infant Mortality in the US, 2000-2023. JAMA Pediatr 2025;179:765–72. doi: 10.1001/jamapediatrics.2025.0440 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Dagher RK, Linares DE. A Critical Review on the Complex Interplay between Social Determinants of Health and Maternal and Infant Mortality. Children (Basel) 2022;9. doi: 10.3390/children9030394. [DOI] [Google Scholar]
- 13.Declercq ER, Cabral HJ, Cui X, et al. Using Longitudinally Linked Data to Measure Severe Maternal Morbidity. Obstet Gynecol 2022;139:165–71. doi: 10.1097/AOG.0000000000004641. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Diop H, Declercq ER, Liu CL, et al. Leveraging a Longitudinally Linked Dataset to Assess Recurrence of Severe Maternal Morbidity. Womens Health Issues 2024. doi: 10.1016/j.whi.2024.06.002. [DOI] [Google Scholar]
- 15.LinkPro v3.0. InfoSoft, Inc;, 2020. [Google Scholar]
- 16.Reference Guide for Pregnancy-Associated Death Identification. CDC, 2025. (Accessed July 10, 2025,, at https://www.cdc.gov/maternal-mortality/php/mmrc/reference-guide.html.) [Google Scholar]
- 17.Main EK, Abreo A, McNulty J, et al. Measuring severe maternal morbidity: validation of potential measures. Am J Obstet Gynecol 2016;214:643.e1–.e10. doi: 10.1016/j.ajog.2015.11.004. [DOI] [Google Scholar]
- 18.Kuklina EV, Ewing AC, Satten GA, et al. Ranked severe maternal morbidity index for population-level surveillance at delivery hospitalization based on hospital discharge data. PLoS One 2023;18:e0294140. doi: 10.1371/journal.pone.0294140. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Hirai AH, Owens PL, Reid LD, Vladutiu CJ, Main EK. Trends in Severe Maternal Morbidity in the US Across the Transition to ICD-10-CM/PCS From 2012-2019. JAMA Netw Open 2022;5:e2222966. doi: 10.1001/jamanetworkopen.2022.22966. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Declercq ER, Cabral HJ, Liu CL, et al. Prior Hospitalization, Severe Maternal Morbidity, and Pregnancy-Associated Deaths in Massachusetts From 2002 to 2019. Obstet Gynecol 2023;142:1423–30. doi: 10.1097/AOG.0000000000005398. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Ely DM, Driscoll AK. Infant Mortality in the United States, 2022: Data From the Period Linked Birth/Infant Death File2024 Jul 25. Report No.: 1551–8922 (Print) doi: 10.15620/cdc/157006 [DOI] [Google Scholar]
- 22.Bateman BT, Mhyre J, Hernandez-Diaz S, et al. Development of a Comorbidity Index for Use in Obstetric Patients. Obstet Gynecol 2013;122:957–65. doi: 10.1097/AOG.0b013e3182a603bb. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Maternal Mortality and Morbidity Review in Massachusetts. Substance Use among Pregnancy-Associated Deaths — Massachusetts, 2005–2014. Boston: 2018. [Google Scholar]
- 24.Somerville NJ, Nielsen TC, Harvey E, et al. Obstetric Comorbidity and Severe Maternal Morbidity Among Massachusetts Delivery Hospitalizations, 1998-2013. Matern Child Health J 2019;23:1152–8. doi: 10.1007/s10995-019-02796-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Hoyert DL. Maternal Mortality Rates in the United States, 2023. NCHS Health E Stats. Hyattsville (MD): National Center for Health Statistics (US) A; 2024. [Google Scholar]
- 26.Mohamed MA, Venkatesh K, Abdelatif D, Nandakumar V, Hoffman H. Healthcare system variability and inpatient maternal mortality in the United States. J Neonatal Perinatal Med 2024;17:543–53. doi: 10.3233/NPM-240025. [DOI] [PubMed] [Google Scholar]
- 27.Powers DA. Black-white differences in maternal age, maternal birth cohort, and period effects on infant mortality in the US (1983-2002). Soc Sci Res 2013;42:1033–45. doi: 10.1016/j.ssresearch.2013.03.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Gill SK, Broussard C, Devine O, Green RF, Rasmussen SA, Reefhuis J. Association between maternal age and birth defects of unknown etiology: United States, 1997-2007. Birth Defects Res A Clin Mol Teratol 2012;94:1010–8. doi: 10.1002/bdra.23049 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Kingsmore SF, Henderson A, Owen MJ, et al. Measurement of genetic diseases as a cause of mortality in infants receiving whole genome sequencing. NPJ Genom Med 2020;5:49. doi: 10.1038/s41525-020-00155-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Declercq EP, MacDorman MP, Cabral HP, Stotland NMD. Prepregnancy Body Mass Index and Infant Mortality in 38 U.S. States, 2012-2013. Obstet Gynecol 2016;127:279–87. doi: 10.1097/AOG.0000000000001241. [DOI] [PubMed] [Google Scholar]
- 31.Ratnasiri AWG, Lakshminrusimha S, Dieckmann RA, et al. Maternal and infant predictors of infant mortality in California, 2007-2015. PLoS One 2020;15:e0236877. doi: 10.1371/journal.pone.0236877. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Hauck FR, Blackstone SR. Maternal Smoking, Alcohol and Recreational Drug Use and the Risk of SIDS Among a US Urban Black Population. Front Pediatr 2022;10:809966. doi: 10.3389/fped.2022.809966. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Evans MI, Ryan GF, Britt DW, Macedonia CR. The Mortality of Politics: An American Paradox. Fetal Diagn Ther 2025;52:65–80. doi: 10.1159/000541912. [DOI] [PubMed] [Google Scholar]
- 34.Woodbury R Causal factors in infant mortality. A statistical study based on investigations in eight cities. Washington, D.C.: US Dept. of Labor. Children’s Bureau; 1925. [Google Scholar]
- 35.Vogel JP, Jung J, Lavin T, et al. Neglected medium-term and long-term consequences of labour and childbirth: a systematic analysis of the burden, recommended practices, and a way forward. Lancet Glob Health 2024;12:e317–e30. doi: 10.1016/S2214-109X(23)00454-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Centers for Disease Control and Prevention. CDC WONDER Online Database. In: USDHHS C, NCHS, Div. of Vital Statistics, public-use data. , ed. Hyattsville, MD: 2024. [Google Scholar]
- 37.Terlizzi E, Cohen R. Geographic Variation in Health Insurance Coverage: United States, 2020. Hyattsville, MD: 2022 February 11, 2022. [Google Scholar]
- 38.Roto S, Nupponen I, Kalliala I, Kaijomaa M. Risk factors for neonatal hypoxic ischemic encephalopathy and therapeutic hypothermia: a matched case-control study. BMC Pregnancy Childbirth 2024;24:421. doi: 10.1186/s12884-024-06596-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.