"Gestational weight gain and modifiable risk factors of severe maternal morbidity"

Epidemiology/Public Health

Committee:
Lisa M. Bodnar, PhD, MPH, RD (advisor)

Katherine P. HImes, MD, MS
Division of Maternal-Fetal Medicine, Department of Obstetrics, Gynecology, and Reproductive Sciences

Maria M. Brooks, PhD
Departments of Epidemiology and Biostatistics

Kathleen M. McTigue, MD, MPH, MS
Departments of Medicine and Epidemiology

Abstract:

Pregnancy-related maternal mortality is a growing public health concern as its incidence has doubled over the past two decades, yet it remains a rare outcome (<1 per 10,000 pregnancies) and is therefore difficult to study, particularly in individual healthcare institutions. However, severe maternal morbidity is 70 times more common than maternal death and shares the same etiologies and risk factors. Our goal was to quantify the burden that modifiable risk factors place on severe maternal morbidity, with a focus on gestational weight gain because of its amenability to intervention during pregnancy.

            We first studied the association between total gestational weight gain and risk of severe maternal morbidity using a cohort of 84,241 singleton delivery hospitalizations. The cumulative incidence of severe maternal morbidity was 1.9/100 delivery hospitalizations. We used multivariable logistic regression using generalized estimating equations and found an J-shaped distribution in risk of severe maternal morbidity as gestational weight gain increased. Overall, we found that the highest risk was among those who experienced very high weight gain. A z-score of +2 (31kg at 40 weeks gestation among normal weight women) was associated with 1.0 (0.46, 1.5)) excess cases of severe maternal morbidity per 100 delivery hospitalizations compared with a z-score of 0 (16kg at 40 weeks among normal weight). Very low weight gain was associated with an increased risk, as well, though the magnitude of association was smaller (a z-score of -2 (7kg at 40 weeks among normal weight) was associated with 0.14 (-0.07, 0.35) excess cases).

Next, we examined the association between weight gain in the first half of pregnancy and risk of severe maternal morbidity at delivery hospitalization. Among a retrospective cohort of 4,774 singleton delivery hospitalizations for which serial weight gain measurements were abstracted, the cumulative risk of severe maternal morbidity was 2.1/100 delivery hospitalizations, with the lowest risk among those who gained more or less than 1 standard deviation outside the mean. After adjusting for known confounders, and found a similar pattern. In our curvilinear models where we modeled weight gain z-score as a restricted cubic spline, we found that the risk of severe maternal morbidity was lowest among those with gestational weight gain z-scores outside 1 standard deviation of the mean. A z-score of +2 (13kg at 19 weeks gestation among normal weight women) was associated with 1.1 (-0.18, 2.4) fewer cases per 100 delivery hospitalizations compared with a z-score of 0 (5.0kg at 19 weeks). A z-score of -2 (0.1kg at 19 weeks) was associated with 0.80 (-0.74, 2.3) fewer cases compared with the same referent.

            Finally, in a retrospective cohort of 86,260 delivery hospitalizations, we calculated the population attributable fraction of eight, known, modifiable risk factors of severe maternal morbidity (prepregnancy body mass index, gestational weight gain, smoking during pregnancy, marital status, maternal education, preexisting hypertension, preexisting diabetes, and maternal age). After adjusting for known confounders, we found that optimizing all these risk factors concurrently could prevent 36% (626 cases) of the severe maternal morbidity in this sample. Though optimizing individual risk factors had nominal impacts, gestational weight gain >1 SD above the mean, overweight and obesity, maternal age ≥35 years, preexisting hypertension, and lack of a college degree had population attributable fractions ranging from 4.5% to 13%. Preexisting diabetes has a smaller association (1.6%) and smoking and marital status were not associated with any notable reduction in cases.

These findings together support that in the overall population, optimizing individual-level risk factors would likely have modest impacts on reducing risk of severe maternal morbidity, but improving overall health before and during pregnancy should still be encouraged to help reduce the risk of severe events. While suboptimal or excessive, total gestational weight gain is associated with increased risk, the most meaningful results are among those with extremely high or low weight gain, of which there are few women. Furthermore, we found a much different pattern of risk between early weight gain and severe maternal morbidity, where the lowest risk was among those who gained outside 1 standard deviation of the mean. The diverging results between the two cohorts are likely due to differences in the distribution of the phenotype of severe maternal morbidity as well as differences in maternal characteristics of the women who experienced severe maternal morbidity in the two samples. Women who had serial weight gain measurements and faced severe maternal morbidity tended to be older, more educated, married, have a lower incidence of preterm birth, less preexisting hypertension or diabetes, less likely to smoke during pregnancy, and utilize private insurance as their primary method of payment compared with those who faced severe maternal morbidity and had total weight gain measurements.  Because the relationship is not consistent between samples or causes of severe maternal morbidity, together with the relatively low population attributable fraction we observed, gestational weight gain does not appear to be a high-leverage risk factor in of itself, but it is likely part of a constellation of risk factors that contribute to increasing risk of severe maternal morbidity. Overall, this topic is of great public health importance because the incidence of severe maternal morbidity is higher in the United States than other, comparable countries and more women are entering pregnancy with comorbidities, increasing the risk for adverse pregnancy outcomes. Additional research should confirm and extend our findings, but in the long term, there must be structural changes at the policy and population levels to provide affordable access to quality healthcare if any meaningful reductions in these outcomes are going to be achieved.