Study population and setting
The study included 8,207 pregnant women and 83,205 non-pregnant women aged 15-44 years with laboratory-confirmed SARS-CoV-2 infection reported to the United States national COVID-19 surveillance system from January 22 to June 7, 2020.
Summary of Main Findings
Hispanic and Black women were disproportionately represented, accounting for 46% and 22% of COVID-19 cases among the 8,207 pregnant women respectively. Data on underlying comorbidities were only available for 23% of pregnant and 35% of non-pregnant women. Among women with available data, chronic lung disease, cardiovascular disease, and diabetes were more prevalent among pregnant than non-pregnant women. Independent of age, underlying comorbidities, and race/ethnicity, pregnant women had a higher likelihood than non-pregnant women of hospitalization (adjusted risk ratio (aRR)=5.4, 95% confidence interval (CI): 5.1–5.6), intensive care unit admission (aRR=1.5, 95% CI: 1.2–1.8) and mechanical ventilation (aRR=1.7, 95% CI: 1.2–2.4). However, mortality rates were similar between pregnant and non-pregnant women infected with SARS-CoV-2 (aRR = 0.9, 95% CI: 0.5–1.5).
The main strength of this study was in its sample size and geographic reach. It included all reported laboratory-confirmed cases with data on pregnancy status during the study period across the United States — from all 50 states, the District of Columbia, and New York City.
The study findings are severely limited because the possibility that cases may have been underreported for certain groups of women (e.g., by race, presence of symptoms, and/or healthcare access, etc.). Criteria for SARS-CoV-2 testing varied significantly by state and over the time period. Furthermore, testing was more likely to capture symptomatic cases, and pregnant women were more likely to be screened/tested than non-pregnant women. The study was also limited because the surveillance system was missing an extensive amount of data on outcomes and key covariates. For example, the majority of the confirmed cases among women aged 15-44 lacked data on pregnancy status and were hence excluded; counting missing outcome data as non-outcomes may differentially undercount events among non-pregnant women, and many of the analyses were conducted on significantly smaller samples after excluding participants with missing covariate data, all of which could introduce significant bias. The associations seen between pregnancy and outcomes could be artifacts of these biases. The definition of hospitalization did not distinguish between admission due to COVID-19 or obstetric care, and the threshold for admission was likely to be lower for pregnant women, making this a poor indicator of disease severity.
This study drew on a national surveillance system to compare COVID-19 severity outcomes between pregnant women and women of reproductive age who were not pregnant in the US. The study and findings were limited by considerable missing data.
This review was posted on: 4 July 2020