SARS-CoV-2 in First Trimester Pregnancy

A Cohort Study

N. la Cour Freiesleben; P. Egerup; K.V.R. Hviid; E.R. Severinsen; A.M. Kolte; D. Westergaard; L. Fich Olsen; L. Prætorius; A. Zedeler; A.-M.H. Christiansen; J.R. Nielsen; D. Bang; S. Berntsen; J. Ollé-López; A. Ingham; J. Bello-Rodríguez; D.M. Storm; J. Ethelberg-Findsen; E.R. Hoffmann; C. Wilken-Jensen; F.S. Jørgensen; H. Westh; H.L. Jørgensen; H.S. Nielsen


Hum Reprod. 2021;36(1):40-47. 

In This Article


We found that pregnant women with SARS-CoV-2 infection in the first trimester did not have a significantly different nuchal translucency thickness measured at their first trimester scan. Furthermore, there was no significantly increased risk of pregnancy loss in women with SARS-CoV-2 infection in the first trimester before the time of the double test.

Of the 36 women with early pregnancy loss, before the double test was taken (Cohort 2), none had SARS-CoV-2 antibodies. It is well known that early pregnancy loss is predominantly related to intrinsic embryonic inborn errors (Ouyang et al., 2016; Pylyp et al., 2018). However, we included Cohort 2 to minimize a risk of bias that could potentially exist if there were a high number of women with SARS-CoV-2 antibodies in this cohort. Significantly more women with antibodies reported symptoms of COVID-19 compared to women without antibodies. None of the women had been hospitalized for COVID-19.

The first case of COVID-19 in Denmark was confirmed 27 February 2020. At the beginning of the epidemic in Denmark, it was only individuals requiring hospitalization who were tested for SARS-CoV-2 with a respiratory specimen. Citizens suspected of COVID-19 but not requiring admission were asked to remain at home and self-quarantine and were not tested. Only 53% of the women with SARS-CoV-2 antibodies in our study reported symptoms of COVID-19 in pregnancy. This corresponds well to a comparable study from Italy where 6 out of 14 (42.8%) first trimester pregnant women with SARS-CoV-2 antibodies referred previous symptoms of COVID-19 (Cosma et al., 2020). Symptoms of COVID-19 are very similar to symptoms of other viral infections and 26% of the women without SARS-CoV-2 antibodies in our study reported similar symptoms. Therefore, serological testing in appointed risk groups, such as pregnant women, is a valuable tool to identify previous infections and to evaluate whether infection in pregnant women requires additional vigilance during the pregnancy. Our findings suggest that pregnant women in their first trimester are not at increased risk of severe COVID-19. This is similar to what has been reported for pregnant women in the third trimester in Wuhan, China (Chen et al., 2020) and in the first trimester in Turin, Italy (Cosma et al., 2020).

In general, the study population had normal BMI and the vast majority were non-smokers. Our results and conclusion may therefore not apply directly to populations with higher BMI, higher frequency of smoking and associated higher frequency of lifestyle diseases. People with lifestyle diseases, diabetes, obesity and individuals who smoke are at higher risk of developing more severe COVID-19 if infected (Guo et al., 2020; Petrakis et al., 2020). Our study does not rule out the possibility that more severe COVID-19 might lead to a higher risk of adverse outcomes for the developing fetus.

The frequency of participants with previous SARS-CoV-2 infection in pregnancy was relatively low and steady over the study period (Supplementary Figure S1). This is most likely a result of the extended measures implemented by the Danish government at an early stage of the epidemic to limit the transmission of the virus. Measures included closing the national borders, banning of group gatherings of more than 10 people, closure of all educational facilities, implementing work-from-home measures for all non-critical government and state employees, and recommending that private employees also work from home where feasible. However, in addition to the general societal changes, the relatively low occurrence of SARS-CoV-2 antibodies among participants could also be due to pregnant women taking additional precautionary measures such as self-quarantine and limiting social contacts even before the implementation of official governmental restrictions. As per 23 May 2020, it was estimated that the seroprevalence of people with SARS-CoV-2 antibodies in the Danish population was 1.1% (95% CI 0.5–1.8) (SSI, 2020a), but the admission area covered by Hvidovre Hospital had a higher than average incidence of SARS-CoV-2 infected individuals at the time (SSI, 2020b). It is therefore not surprising that the prevalence of SARS-CoV-2 antibodies in the study population is higher than 1.1%.

There is still uncertainty concerning the accuracy of the various tests for SARS-CoV-2 antibodies (Infantino et al., 2020a). One study from China used a sandwich enzyme-linked immunosorbent assay (Xiang et al., 2020) and found the sensitivity and specificity to be 77.3/100% for IgM and 83.3/95% for IgG, respectively. We used iFlash 1800 with its IgM/IgG kit, which has previously shown highly accurate results (Infantino et al., 2020b). We used the reference values for positive and negative test results suggested by Harritshøj et al. (2020) who tested the diagnostic accuracy of the SARS-CoV-2 antibody assays. According to Harritshøj et al. YHLO iFlash assays showed acceptable IgG sensitivity and acceptable and high IgM and IgG specificity. It is very likely that more assays will be developed in the future and as reviewed by Infantino et al. (2020a), more studies are needed to validate the serological assays, especially for use as screening tools for asymptomatic individuals. Additionally, our study includes a risk of bias due to the low prevalence of positive samples in the study population (Christopher Sempos, 2020).

Despite a high participant rate, it is a potential limitation of the study that not all invited women participated in the study. By the end of 28 May 2020, a total of 337 women had not responded to our study invitation. It could potentially introduce selection bias if the none-respondents were different in terms of rates of SARS-CoV-2 antibodies and pregnancy loss. The blood sample for antibody analyses from women in Cohort 2 were drawn at the day of pregnancy loss, and we found none with SARS-CoV-2 antibodies. This provides some certainty that we did not overlook a potential effect of SARS-CoV-2 at the earlier stage of pregnancy before the double test is taken and if any relationship exists it is possibly very low.