Study design and setting

CRONOS is a multicentric, prospective, observational study established by the German Society of Perinatal Medicine (DGPM) in April 2020 to rapidly provide data to counsel women with SARS-CoV-2 infection during their pregnancy. Information on the study is available at www.dgpm-online.org and from the German Clinical Trials Register (DRKS00021208); part of the study results have been published recently. Ethical approval was obtained for the study (University Hospital Schleswig-Holstein in Kiel, file number D 451/20, and separate for each study side respectively). Women with clinically confirmed SARS-CoV-2 infection were eligible for inclusion.

All German maternity hospitals were invited to participate in the CRONOS registry. By August 24, 2021, obstetricians and neonatologists from 157 German hospitals and from the Kepler University Hospital in Linz, Austria, confirmed to participate. Of these, 115 hospitals actively provided data to CRONOS. These maternity units attended 224,647 deliveries in 2020, accounting for 29.1% of the births in Germany. Participating hospitals were asked to register all women with SARS-CoV-2 infection independent of the time point of infection during pregnancy.

Data capture and study variables

For collecting data, a reporting form was developed using the cloud-based electronic data capture platform of the service provider castoredc.com (Amsterdam, Netherlands). After the patients had given informed consent, information on the demographic characteristics, comorbidities, previous and current pregnancy characteristics, SARS-CoV-2-specific symptoms and treatments, pregnancy- and birth-specific events, and neonatal outcomes were entered by each treating hospital in the data capture platform.

Cohort

The data presented here were collected between April 3, 2020 and August 24, 2021 by 115 hospitals comprising a total cohort of 2819 cases. During the review of the registry and the plausibility check, duplicate entries were suspected in 10 cases; 8 of the cases were confirmed and excluded after contact with the entering hospital. In 98 cases, the week of gestation at the time of infection remained unknown, and they were excluded, as this was considered mandatory information. In 63 other cases, it remained unclear whether women had been infected during or before their pregnancy; these were also excluded. From the remaining 2,650 women with confirmed SARS-CoV-2 infection during their pregnancy, 1812 (68.3%) were symptomatic, 709 (26.8%) were asymptomatic, and data on symptoms were not provided in 129 (4.9%) cases ( Figure ). The final study cohort for the analyses consisted of patients with COVID-19 (symptomatic SARS-CoV-2 infection) with valid information on whether or not medically assisted reproduction (MAR) had been conducted before the present pregnancy (n=1485). Of these, 1331 (89.6%) had a confirmed SARS-CoV-2 infection via a viral RNA detected by polymerase chain reaction testing, 30 (2.0%) via a detection of maternal SARS-CoV-2 antibodies, 45 (3.0%) via an antigen testing, and for 75 cases, (5.1%) no information about the exact diagnostic test of SARS-CoV-2 infection was available.

Flowchart of the study cohort

Ziert et al. Adverse pregnancy outcomes by mode of conception in women with COVID-19. Am J Obstet Gynecol 2022.

Statistical analyses

For evaluating whether pregnancies after fertility treatment involve a greater risk of adverse maternal and neonatal outcomes than spontaneous pregnancies, a stepwise statistical analysis strategy was performed. Firstly, the baseline data were analyzed to identify statistically significant differences in the baseline risks between pregnancy with MAR and spontaneous conceptions ( Table 1 ). Secondly, it was evaluated whether COVID-19 associated clinical outcomes ( Table 2 ) and maternal and neonatal complications differed in a statistically significant way between MAR pregnancies and spontaneous conceptions ( Supplemental Table 1 ). Thirdly, for those maternal and neonatal outcomes with statistically significant differences between the 2 study groups, separate multivariate models were calculated ( Supplemental Tables 2 and 3 ).

For comparing the categorical variables between pregnancies after MAR and spontaneous conceptions, the absolute and relative frequencies are presented for each group separately. The statistical significance was tested for the categorical baseline variables by using the chi-square tests or Fisher exact tests, and univariate (multinominal) logistic regression models were applied for the categorical maternal and neonatal outcomes. The continuous variables are shown as means and standard deviations for each group. The statistical significance was tested for the continuous baseline variables by using independent t tests and univariate analysis of covariance for continuous maternal and neonatal outcomes.

In addition to P values, risk estimators (odds ratios [OR] or mean difference [MD]) and corresponding 95% confidence intervals (95% Cl) were calculated for a comparison of maternal and neonatal complications. For the calculation of adjusted estimators in multivariate (multinominal) logistic regression models, the baseline variables that significantly differed between both the groups (confounders) or those that are known risk factors for adverse outcomes were included in the models.

Statistical analyses were performed using Statistical Package for Social Sciences (SPSS; version 28; IBM Corp, Armonk, NY) for Windows (Microsoft, Redmond, WA). SPSS, by default, conducts analyses by dropping cases for which there are missing values, so the sample sizes may differ in the statistical analyses. Inferential statistics were used in a descriptive manner. Thus, neither global nor local significance levels were determined, and no adjustment for multiplicity was applied. However, P values <.05 were considered to be statistically significant.

Principal findings

We report that in women with MAR pregnancies, the risk of COVID-19-associated adverse outcomes, for example, the need for inpatient treatment, pneumonia, oxygen ventilation, maternal death, and delivery and also stillbirth and neonatal death following birth within 4 weeks of the onset of COVID-19 was comparable with pregnancies after spontaneous conceptions. MAR was also not the primary risk factor of adverse maternal or neonatal outcomes in pregnancies affected by COVID-19. Instead, other factors, such as maternal age, nulliparity, BMI >30, or multiple gestation were the key drivers. However, MAR conceptions were associated with descriptively higher risks of gestational diabetes mellitus, peripartum hemorrhage, cervical insufficiency, cesarean delivery, preterm birth, and admission to NICU.

Results in the context of what is known

Several cohort studies report an association between COVID-19 in pregnancy and substantially increased maternal and neonatal morbidity and mortality than pregnant women without a diagnosis of COVID-19. This involves preeclampsia, ^{, ,} gestational hypertension, maternal death, stillbirth, preterm delivery, ^, and poor fetal growth, among others. ^, The risks are significantly higher in women with comorbidities, eg, diabetes mellitus, hypertension, and obesity; the risks are also higher in older age.

Insight on the outcomes of pregnancies achieved through MAR and COVID-19 is sparse. The first data were provided by the ESHRE COVID-19 Working Group. It collected 80 cases from 32 countries, including 67 live births, 10 miscarriages, 2 stillbirths, and 1 maternal death. One-third of the reported cases had an asymptomatic infection, whereas 31.4% were treated in the hospital. The authors concluded that infections in pregnancies after MAR do not lead to a higher risk of adverse outcomes than those after spontaneous conceptions. This is somehow in line with our results, which demonstrate that the conception mode is not the primary risk factor for adverse obstetrical and neonatal complications in COVID-19-affected women. Particularly encouraging is also the fact that comparable incidences of complications were reported for those directly associated with COVID-19 (eg, pneumonia, ICU admission, and death).

Engels Calvo et al reported a higher incidence of preeclampsia and cesarean delivery in a cohort of symptomatic and asymptomatic SARS-CoV-2-infected women after IVF than with spontaneous pregnancies. This is in contrast to our observations for both the outcomes. One main difference in the analysis is the adjustment for important confounders, which included maternal age and clinical presentation in the Spanish study. In the cohort, 36 out of 74 women conceived with donor oocytes, which was not a covariate in the multinomial logistic regression model but is one of several well-known risk factors for preeclampsia. In addition, the higher rate of cesarean delivery in IVF patients than in spontaneous conceptions can particularly be explained by the increased rate of multiple births in the Spanish study. Our multinominal logistic regression model included the clinical characteristics (eg, age, nulliparity, and multiple gestation), which significantly differed between MAR pregnancies and spontaneous conceptions and also known potential confounders (eg, BMI >30) for the respective outcome. The models developed suggest that the (still) elevated though not statistically significant risk of MAR pregnancies would be further reduced if additional risk factors were considered. In our cohort of spontaneous conceptions, a significant number of stillbirths (2.5%) occurred within 4 weeks of COVID-19 onset, which is higher than the expected stillbirth rate of approximately 0.4% in Germany and requires further attention. None of these cases were present in the MAR group, which may be because of the low number of registered pregnancies on the one hand and the possibly better monitoring and earlier delivery of these high-risk pregnancies per se on the other.

Compared with spontaneous conceptions, adverse maternal and neonatal outcomes are more common among conceptions after MAR independent of the onset of COVID-19 in pregnancy. ^, The main driver for the adverse outcomes is the higher risk of multiple gestations in IVF. As the field moves toward single embryo transfer, and as the rate of multiple gestations decreases, it also becomes clear that other factors, eg, the choice of a programmed protocol in frozen-thawed embryo transfer cycles or subfertility itself make a significant contribution to a higher incidence of adverse outcomes. Subfertile women carry risk factors for pregnancy complications more often, eg, higher age, obesity, or metabolic alterations. ^{, ,} Beside an expected higher rate of multiple gestations, we confirmed these observations in our infertile cohort, which was older and more often obese. The rate of nulliparity, which is a risk factor for hypertensive disorders of pregnancy, was also higher. The results from our multivariate models demonstrate that MAR itself is not the primary risk factor for adverse outcomes of pregnancies affected by COVID-19, for example, gestational diabetes mellitus, peripartum hemorrhage, cervical insufficiency, cesarean delivery, preterm birth, and admission to NICU, but that these women enter pregnancy with a higher baseline risk. In this context, maternal age, multiple gestation, and BMI >30 were the main predictors of obstetrical and neonatal complications.

Clinical implications

Our findings are of clinical importance during an ongoing pandemic with a so far unknown end. They will help advise couples seeking fertility treatment and will provide reassurance that the fertility treatment itself will not add to the potential risk of adverse maternal and neonatal outcomes when a patient is affected by COVID-19. Nevertheless, our data clearly show again that other, potentially avoidable, risk factors (eg, multiple gestation) lead to a poorer outcome.

Strengths and limitations

In this study, we took advantage of a well-supervised prospective registry study using a standardized electronic clinical report form. Items specifically targeting MAR were incorporated and linked to COVID-19 and obstetrical and neonatal outcomes. A particular strength of our study is the differentiation between the methods of conception such as IVF or ICSI and OI with or without IUI, which ensures the quality of the data entries. Because asymptomatic cases are most likely considered as “incidental SARS-CoV-2 infection” cases requiring hospitalization for other reasons, we focused our analysis on symptomatic women. In our experience this approach avoids including women who present to the hospital primarily because of obstetrical complications, eg, spontaneous preterm delivery, fetal growth restriction, or preeclampsia.

Our study has a variety of limitations, including a relatively small sample size of MAR pregnancies, and therefore, differing cohort sizes. These factors make it difficult to reliably show statistical correlations. Nevertheless, our cohort reflects the reality in Germany, with approximately 4% of births occurring through MAR. Because of the cohort design as a registry study, no conclusion can be drawn about the actual incidences of complications in COVID-19-affected women in both the conception groups. Similarly, owing to the currently low case numbers, no final statement can be made regarding the risk factors for severe maternal courses and neonatal infections. Beside, a comparison to women without COVID-19 is not possible, so the direct impact on outcomes (COVID-19 vs no SARS-CoV-2 infection) cannot be derived.

Conclusion and future directions

COVID-19 during pregnancy after MAR was not associated with a substantially higher risk of COVID-19-associated adverse outcomes. However, maternal and neonatal morbidity was increased compared with outcomes after spontaneous conceptions. The data provide important information for counseling couples who seek fertility treatment and to provide reassurance about the risks that are not primarily driven by MAR but other individual risk factors compared with spontaneous conceptions with COVID-19 during pregnancy. With the growing number of SARS-CoV-2 infections worldwide, further research is needed to confirm our observations and to elucidate the impact of different SARS-CoV-2 variants and of infections after previous immunization on obstetrical and neonatal outcomes.