Study population and setting
This prospective cohort included 3374 pregnant women who were tested for SARS-CoV-2 by PCR (outpatient, emergency department, and inpatient testing) and subsequently delivered in the Parkland Health and Hospital System in Dallas Country, Texas, between March 18 and August 22, 2020 (universal inpatient testing started May 18, 2020). All hospitalizations among included women were dichotomized as obstetric (non-COVID related) or COVID-related. The authors used demographic and clinical information from medical records and the primary outcome was a composite of preterm birth (medically indicated or spontaneous), preeclampsia with severe features (organ failure), or cesarean section due to fetal indication after 20 weeks. Neonatal SARS-CoV-2 infection, tested at 24 and 48 hours by PCR, was a secondary outcome. They compared women who delivered during the study period with and without a positive SARS-CoV-2 test with t-tests or Wilcoxon tests for continuous covariates and Chi-squared tests for categorical covariates. They presented an unadjusted relative risk after determining that none of the demographic or clinical factors should confound the relationship between SARS-CoV-2 diagnosis and the primary outcome. In a secondary analysis, outcomes were compared by COVID-19 severity (asymptomatic, mild, moderate, severe, or critical) among participants who tested positive for SARS-CoV-2.
Summary of Main Findings
Of the 3374 pregnant women (mean age 27.6 years) who met their inclusion criteria, 252 (7.5%) tested positive for SARS-CoV-2; the majority of whom (172, 68%) tested positive on the labor and delivery ward. Spontaneous abortion occurred in seven (3%) women who tested positive and 87 (3%) women who tested negative for SARS-CoV-2; these participants were excluded. They assessed neonatal outcomes in 248 neonates with mothers with and 3015 neonates with mothers without SARS-CoV-2. Of the 245 women who tested positive for SARS-CoV-2 included in the analysis, 16 (7%) tested positive in their second trimester and 227 (93%) tested positive in their third trimester. Although there were no differences in age, parity, body mass index, hypertension, or diabetes in women who delivered with or without a history of SARS-CoV-2, Hispanic women were overrepresented among women who tested positive for SARS-CoV-2 (91% among those who tested positive versus 73% in those who tested negative). There was no substantive difference in the primary outcome between women with (21%) and without (23%) SARS-CoV-2 (RR 0.94; 95% CI: 0.73,1.21) and the other neonatal outcomes were similar regardless of maternal SARS-CoV-2 history. Preterm birth was 11% in each group, and there were no stillbirths in women with SARS-CoV-2, however, among patients with SARS-CoV-2, the likelihood of preterm birth increased with COVID-19 severity. There were six neonates (3% of the 188 tested) who tested positive for SARS-CoV-2.
This study is one of the largest published longitudinal cohorts of SARS-CoV-2 in pregnancy, and includes a comparison group of women who delivered during the same time period in the same facility, but tested negative for SARS-CoV-2. Though testing was not universal among pregnant women during the first several months of the study, the authors reported that there were no demographic differences between pregnant women who did versus did not get tested for SARS-CoV-2 during the study period.
It is possible that there was misclassification of SARS-CoV-2 exposure at the time of delivery, as the majority of the study period preceded universal testing of pregnant women; the authors do not present data on how many women were tested more than once, how they tested on repeated testing, or how they handled women with more than one test in their analysis. If women who tested negative at the time of delivery but were previously infected had worse outcomes, they would have biased these results towards no effect. Though the authors showed similarity on many of the demographic and clinical factors between those with and without SARS-CoV-2, an adjusted analysis would have further strengthened the inferences. In particular, adjusting for race/ethnicity, while an insufficient proxy for likely exposure to systemic racism, would nonetheless provide some adjustment for the underlying factors that increase risk of SARS-CoV-2 infection (such as through housing conditions, occupation, etc.) and birth outcomes (such as differential access to care and physician treatment). Finally, this study was underpowered to assess for differences depending on when during pregnancy a woman became infected with SARS-CoV-2, with the sample predominantly infected in the third trimester.
This is one of the largest longitudinal studies of SARS-CoV-2 in pregnancy to include a comparison group of pregnant women drawn from the same source population who tested negative.
This review was posted on: 12 February 2021