Worse outcomes of pregnancy in COVID-19 infection during parturition may be due to referral bias: analysis in a prospective cohort of 963 pregnancies


The initial studies of COVID-19 suggested that pregnant women have more severe infection with an increased risk of preterm birth, preterm rupture of membranes, and even maternal deaths. Later studies and systematic reviews showed different results. When pregnant women were universally screened, severe disease rates mirrored those of the normal population. Most studies on outcomes have not controlled for either preexisting maternal risk factors or those acquired during pregnancy. In addition, there is still a gray area in understanding how COVID-19 infection around the time of delivery affects pregnant women. Thus, we analyzed if the apparent high risk of severe COVID-19 in referral centers was confounded owing to other concomitant risk factors.

Study Design

In our cohort from a single tertiary referral hospital in India, all pregnant women coming for delivery or with labor pain were universally screened for SARS-CoV-2 infection using reverse transcriptase polymerase chain reaction performed on oronasopharyngeal samples. The patients who left the hospital before delivery were excluded. Of 963 pregnant women, 127 were COVID-19 positive. They were compared using tests for proportion in terms of maternal complications (cesarean deliveries, antepartum and postpartum hemorrhage, preterm and prelabor rupture of membrane, puerperal sepsis, and mortality) and neonatal outcomes (appearance, pulse, grimace, activity, and respiration scores; low birthweight; intensive care requirement; neonatal COVID-19 infection; neonatal sepsis; and death). The generalized linear models (GLMs) were then built to assess the contribution of various maternal risk factors and COVID-19 positivity on these outcomes.


The age, gravida, parity, gestational diabetes, and pregnancy-induced hypertension (PIH) rates were similar between the COVID-19 positive and negative cohorts ( Supplemental Table 1 ). The COVID-19 cohort had an overrepresentation of various other pregnancy risk factors ( Table ). Furthermore, the COVID-19 cohort had higher cesarean deliveries (87 [68%] vs 445 [53.3%] in the negative cohort; P =.02), higher postpartum hemorrhage (6 [4.7%] vs 1 [0.1%]; P <.001), and higher maternal mortality (2 [1.6%] vs 1 [0.1%]; P =.048) ( Supplemental Table 2 ). Among neonatal outcomes, Apgar score was lower at 1 minute (mean [standard deviation], 7.20 (1.63) in COVID-19 vs 7.54 (1.69) in the controls; P =.035) and at 5 minutes (mean [standard deviation], 8.27 (1.72) in COVID-19 vs 9.14 (1.74) in controls; P <.001) ( Supplemental Table 3 ). In the first GLM model on the mode of delivery, the significant predictors were previous cesarean deliveries, COVID-19 positivity, presence of PIH, and gestational diabetes ( Supplemental Table 4 ). In the second GLM model, bad maternal outcomes were only associated with the presence of PIH ( Supplemental Table 5 ). In the third GLM model, bad neonatal outcomes were associated with the presence of PIH or 1 of the 7 other factors for high-risk pregnancy ( Supplemental Table 6 ). Thus, the associations found on univariate analysis reflect a possible referral bias where the high-risk patients were being referred if they were COVID-19 positive than if they were negative.


Comorbidities and pregnancy risk factors

Condition COVID-19 (n=127) Controls (n=836) P value
Prepregnancy comorbidity 26 (20.3) 159 (19.0) .72
Hypertensive disease of pregnancy 9 (7.0) 64 (7.7) >.99
Gestational diabetes 6 (4.7) 32 (3.8) .62
Other pregnancy-related risk
None 799 (95.70) 117 (91.40) NA
Twin pregnancy 7 (0.80) 1 (0.80) <.01 a
Breech presentation 10 (1.20) 8 (6.30) <.01 a
Intrauterine growth restriction 3 (0.40) 1 (0.80) <.01 a
In-vitro fertilization 2 (0.20) 0 .017 a
Rhesus-negative pregnancy 8 (1.00) 1 (0.80) <.001 b
Thalassemia 2 (0.20) 0 .017 a
Obstetrical cholestasis 4 (0.5) 0 <.001 b

Data are presented as number (percentage).

Mohini. Outcomes of COVID-19 positive deliveries. Am J Obstet Gynecol 2022 .

a Significant

b Highly significant.


This study reiterates that COVID-19 infection does not pose additional risk to pregnancy outcomes by itself. Earlier systematic reviews were hampered by the high heterogeneity of the reported cohorts. This was compounded by duplicate reporting of the same patients in different cohorts, variable inclusion criteria of systematic reviews, and scarce and missing data. More recent systematic reviews have shown that maternal deaths and neonatal outcomes were similar in deliveries conducted in COVID-19 mothers compared with non–COVID-19 mothers. The limitations of our study include the fact that we do not have the indications for cesarean deliveries in the cohort and that it was carried out in a tertiary center that would receive more complicated cases. It brings to light that COVID-19–positive mothers being treated at tertiary care centers have higher rates of cesarean delivery and higher morbidity and mortality, possibly owing to the extra underlying risk factors arising from a referral bias.

Supplementary Materials

Supplemental Table 1

Demographics of the cohort

Characteristic COVID-19 positive (n=127) COVID-19 negative (n=836) P value
Mean Standard deviation Mean Standard deviation
Age 28.18 4.64 27.71 4.23 .29
Gravida 1.69 0.98 1.72 0.87 .74
Parity 0.42 0.58 0.51 0.59 .11
Period of gestation at delivery 36.94 3.14 37.42 2.90 .10
Body mass index 28.00 1.33 28.12 1.35 .34

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Feb 23, 2022 | Posted by in OBSTETRICS | Comments Off on Worse outcomes of pregnancy in COVID-19 infection during parturition may be due to referral bias: analysis in a prospective cohort of 963 pregnancies

Full access? Get Clinical Tree

Get Clinical Tree app for offline access