In our population-based case-control study, we found the majority of cases had an SMM intrapartum, and the most common SMM were transfusion, hysterectomy, and respiratory failure. Maternal age of 40 years and older, the presence of a preexisting medical condition, a multiple birth, and a prior cesarean delivery were the strongest risk factors for SMM. In addition, nulliparity and being a minority, particularly Black, were also strong risk factors for SMM.
Our results are consistent with a large US population-based study in which transfusion was also the most common SMM (48.4%), followed by eclampsia (14.0%) and hysterectomy (11.9%)3
; notably eclampsia was observed in only 7.8% of our cases. Another population-based study from the Netherlands combined antepartum, intrapartum, and postpartum SMM and found the largest SMM rates for eclampsia, major obstetric hemorrhage, and uterine rupture, respectively;7
a population-based Canadian study reported rates of intrapartum SMM were largest for transfusion, puerperal sepsis, hysterectomy, and cardiac events.5
Our study did not assess severe hemorrhage and uterine rupture as unique SMM categories, although transfusion is a surrogate for severe hemorrhage and uterine rupture commonly is associated with hysterectomy.
Women with a preexisting condition, as compared to those without, were at two times the risk of SMM. This observation is similar to two case-control studies in which women with a history of a chronic medical condition12
were at two times the risk of SMM, and women with a previous or preexisting illness13
were 2.5 times more likely to have an ICU admission during pregnancy. Although the case definitions and ascertainment methods differed in these studies, preexisting conditions were consistently identified as strong risk factors for SMM. It has been well established that conditions such as those listed in lead to a number of adverse health and pregnancy outcomes.14,15
We also observed that women with a multiple birth were at 2.5 times the risk of SMM as compared to women with a singleton birth. Prior studies note that the risk of SMM in women with a multiple birth is 2–5 times greater than in women with singleton births,5,7,13,16
most likely related to increased risks of hypertensive disease, diabetes, hemorrhage and operative vaginal and cesarean deliveries.17–21
Additionally, we found that prior cesarean delivery was strongly associated with SMM. Prior cesarean delivery is known to be associated with preeclampsia, placenta previa, placenta accrete, placental abruption,22
uterine rupture, postpartum infection, transfusion, and admission to the intensive care unit.23
Black race was identified as another strong risk factor for SMM, which is consistent with previous studies noting 2 to 7.5 times the risk of SMM among Black women as compared to white women.3,24,25
This increased risk may result from a greater prevalence of disease due to genetic factors or underlying poor health, more severe disease, less access to prenatal care, or a combination of these factors.26
Hypertension and gestational diabetes (GDM), differ by race and ethnicity and the severity of these morbidities may be increased in minority women.27
Maternal obesity in pregnancy also differs by race/ethnicity, with minorities at increased risk of overweight and obesity, which is associated with a number of pregnancy complications.27
However, we simultaneously controlled for factors that could confound the association between race and SMM [preexisting conditions, BMI, other measures of SES (education and payer source)] and the increased risk associated with maternal race/ethnicity persisted, indicating that this finding is likely not an artifact of improper adjustment. These results may reflect genetic differences or disparities in access to and utilization of medical care.
Lastly, we observed an increased risk of SMM at parity extremes. Nulliparous women had nearly two times the risk of SMM as women with one prior delivery. This finding has been observed in other studies12,28
but not consistently.29
Nulliparous women have been observed to be at increased risk of GDM, pregnancy-induced hypertension, premature rupture of membranes (>24 hours), postpartum hemorrhage, and third-degree tears.30
We also observed an increased risk for multiparous women with three or more prior births; pregnancy complications and adverse pregnancy outcomes are associated with increasing parity.31,32
Some of the important demographic risk factors for SMM (also classified as “near miss”) identified by our study are also risk factors for maternal mortality. In a study of maternal deaths in New York City, women ≥35 years were at increased risk of maternal mortality as compared to women 15–19.33
Another US study noted increasing parity was associated with risk of maternal mortality.34
The risk for pregnancy-related death was observed to increase with increasing live birth order.
Our study had several limitations. As discussed by Callaghan et al, conventional obstetric ICD-9 codes often are not specific and do not provide information about the severity of the condition, both of which are limitations to identifying “near-miss” morbidities.3
However, our case definition focused on codes that could identify severe complications by using codes for conditions/procedures occurring as a result of only severe
complications (e.g. cerebrovascular accident, blood transfusion, or artificial ventilation in the case of severe preeclampsia).3
Although the selected SMM codes could still result in misclassification of cases as controls, the rarity of these morbidities and the very large number of controls in this study would likely result in this misclassification having little or no impact on our results. We further strengthened our definition by restricting cases to women who had a hospital stay of at least three days or who were transferred from another facility, as a shorter stay without transfer would be inconsistent with the severity of a true SMM. In addition, although our definition of preexisting conditions using hospitalizations within the five years prior to the index delivery may have not captured all women with these conditions, using this method ensured the condition was present before the pregnancy and was significant enough to result in hospitalization. Compared to using birth certificate data, hospital discharge data are more accurate for obtaining preexisting medical conditions.35
An additional limitation of our preexisting condition definition is that women who delivered before 1992 had less than five years of data because CHARS became available in 1987. Nevertheless, because this information would be missing equally for cases and controls, misclassification is likely non-differential and would attenuate observed associations. We performed a sensitivity analysis with an alternate definition of preexisting conditions derived from birth certificates and results were almost identical.
Although this study investigated risk factors at the level of the patient, both provider and health care system factors may also play a critical role in the development of SMM.26
In a study investigating the preventability of maternal mortality and severe morbidity, 45% of near-miss morbidities were deemed preventable; provider factors were identified as the source of preventability in approximately 93% of these instances, including failure to identify high-risk status, lack of referral to a tertiary care center, and in the greatest proportion incomplete or inappropriate management.24
Patient factors only accounted for 13% and system factors 47% of the cases of preventable near-miss morbidities. The majority of risk factors we identified cannot be modified at the level of the individual, including age, race, parity, multiple birth, and prior cesarean delivery, which suggests that improvements at the provider or system-level may be the key to reducing SMM. For example, if preeclampsia is identified early and proper treatment is initiated, progression to eclampsia is rare; however 738 women in our population were diagnosed with eclampsia, likely indicating some error in identification or management. Further studies investigating the specific provider and system factors that contribute to preventable SMM are necessary develop interventions that reduce the risk of SMM.