Severe acute respiratory syndrome coronavirus 2 and preterm birth, clinical implications—A register‐based cohort study from Sweden

Abstract

Introduction

Severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) infection during pregnancy is associated with an increased risk of preterm birth (PTB). The aim of the study was to investigate the effect of timing of SARS‐CoV‐2 infection during pregnancy on PTB risk and whether the risk is related to the interval from infection onset. We also studied whether the association is driven primarily by spontaneous or iatrogenic PTB, as well as to what extent disease severity, viral variant, background characteristics, and vaccination status affect the risk for PTB.

Material and methods

This is a Swedish population–based register study including all singleton pregnancies registered in the Swedish Pregnancy Register with estimated date of birth between March 1, 2020 and May 31, 2022 (n = 233,335). Data on SARS‐CoV‐2 test positivity, vaccination status, and disease severity were retrieved from mandatory Swedish health registers. Cox regression analyses with time‐varying covariates were performed to investigate the association of testing positive for SARS‐CoV‐2 during pregnancy and risk of PTB.

Results

There was a significant (p = 0.047) interaction between gestational age at the debut of SARS‐CoV‐2 infection and hazard ratio (HR) for PTB within 2 weeks. The adjusted HRs (aHRs and 95% confidence interval [CI]) for PTB within 2 weeks of a positive SARS‐CoV‐2 test were 6.0 (3.9–9.2), 6.1 (4.4–8.4), and 3.2 (2.7–3.7) for women with a positive test at 22–27, 28–31, or 32–36 gestational weeks, respectively. Two weeks or more after the start of a SARS‐CoV‐2 infection, only a slightly increased hazard for PTB could be observed (aHR, 1.2; 95% CI, 1.0–1.4, p = 0.016). Excluding women admitted for SARS‐CoV‐2 infection, the aHR for PTB within 2 weeks was aHR = 2.8 (95% CI, 2.4–3.3). The aHR for preterm birth among vaccinated infected women compared to non‐infected was 1.6 (95% CI, 1.2–2.2). The corresponding aHR for women who had a SARS‐CoV‐2 infection without a previous vaccination was 3.2 (95% CI, 2.3–4.3). Thus, vaccinated infected women had a significantly lower risk for PTB than had women without vaccination before infection (p for homogeneity = 0.007). The viral variants Alpha and Delta were associated with a greater risk of PTB compared with Omicron, but the risk increase was found during the entire study period.

Conclusions

Pregnant women are at particularly high risk of PTB within 14 days after a SARS‐CoV‐2 infection, and should be counselled on PTB symptoms to ensure that they seek medical care without delay. The increased PTB risk is partly mitigated by vaccination against SARS‐CoV‐2.

Abbreviations

aHR

adjusted hazard ratio

CI

confidence interval

GW

gestational week

HR

hazard ratio

PTB

preterm birth

SARS‐CoV‐2

severe acute respiratory syndrome corona virus 2

1. INTRODUCTION

Severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) infection during pregnancy is associated with an increased risk of preterm birth (PTB) [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]. There are several knowledge gaps regarding different aspects of this association, and available data are contradictory. It is unclear how gestational age at infection affects the risk of PTB [2, 5], whether the risk is only increased during the period of infection or whether it remains increased throughout pregnancy [11, 12]. It is also unclear whether SARS‐CoV‐2‐associated PTB onset is primarily spontaneous or iatrogenic [13]. Neonatal morbidity and mortality are foremost associated with birth at earlier preterm gestational ages and these PTBs are more often caused by infection, compared with PTB during the later preterm period [14]. Identifying a time span associated with increased risk of PTB, and determining whether there are certain gestational weeks (GWs) with especially critical consequences of infection, would be highly relevant for pregnant women. Furthermore, this would enable the provision of appropriate medical surveillance and information concerning the infection‐associated risks.

Current knowledge of the association between SARS‐CoV‐2 infection during pregnancy and PTB is based on highly heterogeneous studies in terms of study design (single‐center vs. population‐based), testing strategies (universal vs. nonuniversal), data collection (self‐reported outcome, patient charts, registers), study population (background characteristics, PTB risk, vaccination status, severity of infection), infection time point, and definition of outcome. This heterogeneity hinders comparability and interpretation of the scientific evidence. Sweden has one of the lowest PTB rates in the world, with a pre‐pandemic prevalence of 5%–6% [15]. The population‐based Swedish healthcare registers, with mandatory registration of all positive SARS‐CoV‐2 tests, vaccination status, and pregnancy outcome recorded by healthcare professionals, provide a unique opportunity to study in detail how SARS‐CoV‐2 infection is associated with PTB risk. This facilitates guidance to pregnant women, as well as to healthcare professionals.

In this nationwide, population‐based cohort study, we aimed to address clinically relevant knowledge gaps regarding the association between SARS‐CoV‐2 infection during pregnancy and PTB. We investigated the association between timing of infection and PTB, whether there is an increased PTB risk after infection during certain GWs, and whether the risk is related to the interval from infection onset. We also studied whether the association is driven primarily by spontaneous or iatrogenic PTB, as well as to what extent disease severity, viral variant, background characteristics, and vaccination status affect the risk.

2. MATERIAL AND METHODS

For this register‐based cohort study, data on pregnancy outcome were obtained from the Swedish Pregnancy Register (SPR), a certified National Quality Register with automatic data transfer from standardized electronic prenatal, birth, and neonatal records. The SPR contains data from more than 95% of all deliveries taking place in Sweden and covers the entire period from the first antenatal visit until the follow‐up visit at 8–12 weeks postpartum [16]. Data on infection status, including SARS‐CoV‐2 positivity and time point for testing, were retrieved from the mandatory Swedish Register for Surveillance of Communicable Diseases (SmiNet) [17]. In Sweden, public self‐testing was available from June 2020. Earlier tests were preferably used for people with severe COVID‐19 and healthcare personnel due to a lack of testing and analysis materials. Pregnant women were identified as a risk group for severe disease in April 2020 [18]. The pregnant women were advised to test for SARS‐CoV‐2 if they experienced symptoms suggestive of COVID‐19 and report a positive test immediately to their midwife to determine whether thromboprophylaxis was warranted [19]. Most women were tested for SARS‐CoV‐2 upon admission to the hospital for pregnancy complications or birth [20]. Viral variant was defined as the dominant variant at the time point of infection [21]. Vaccination is based on data from the mandatory National Vaccination Register [22]. A woman was characterized as vaccinated after receiving at least one dose. The severity of COVID‐19 infection was categorized as (I) positive test only, (II) hospitalization due to COVID‐19, (III) intubation and/or intensive care due to COVID‐19, and (IV) ECMO treatment or death due to COVID‐19. This definition is based on the WHO classification [23] of the severity of COVID‐19 and has adopted a Swedish concept. Categories III and IV were analyzed as one category.

Disease severity was defined based on data from the National Patient Register (NPR) [24] and the Swedish Intensive Care Register (SIR) [25]. The NPR is a mandatory health register containing data related to inpatient admissions and outpatient specialist visits. NPR was the source of hospitalization data concerning COVID‐19. Linkage between the registers was based on the unique national personal identification numbers accorded to all residents of Sweden.

2.1. Study population

All women registered in the SPR with singleton pregnancies with estimated date of birth between March 1, 2020 and May 31, 2022 were included in the study. Date of birth was estimated based on first‐ or second‐trimester ultrasound in more than 97% of all pregnancies. Inclusion was based on estimated date of birth to prevent overinclusion of post‐term birth in the beginning of the study period and overinclusion of PTB in the end of the study period.

2.2. Exposure

SARS‐CoV‐2 positivity during any week of pregnancy before 37 full GWs, confirmed via polymerase chain reaction (PCR) (99.1%) and/or antigen testing (0.9%) as registered in the SmiNet, was the main studied exposure. Further exploratory analyses were performed, with GWs at infection, severity of SARS‐CoV‐2 infection, interval from infection to PTB, viral variant, and vaccination status as additional exposure variables.

2.3. Outcomes

The primary outcome of this study was PTB, defined as birth between 22 and 37 full GWs. PTB was further subdivided into extremely preterm (22 to less than 28 weeks), very preterm (28 to less than 32 weeks), and moderate‐to‐late preterm (32 to less than 37 weeks), according to the WHO classification [23]. Iatrogenic PTB were excluded in a separate analysis.

2.4. Confounders

Possible confounder variables were defined a priori [26, 27, 28, 29] and based on data from the SPR: maternal age, country of birth, educational level, body mass index (BMI), and parity.

2.5. Statistics

Chi‐square tests were used to evaluate heterogeneity between groups displayed in descriptive tables. Moving averages (over 3 GW classes) were used to smooth graphs in descriptive figures. For descriptive purposes, the risk for birth less than 14 and 14–20 days after SARS‐CoV‐2 infection at a certain GW was compared with the risk among SARS‐CoV‐2‐free ongoing pregnancies at a certain GW (mid‐week). Based on those data, the breaking point for the interval between SARS‐CoV‐2 infection and PTB was chosen to 14 days. For the main results, adjusted hazard ratios (aHRs) for preterm birth <14 days and ≥14 days after SARS‐CoV‐2 infection, respectively, were computed using time‐varying covariates considering the gestational age at the first day of the SARS‐CoV‐2 infection. Interactions were checked between SARS‐CoV‐2 and gestational age at infection and BMI, respectively. Adjustments were made for maternal age (continuous), parity (primiparity, multiparity), BMI (continuous), educational level (ordinal, treated as continuous), year of delivery (classes), and maternal country of birth (within or outside Europe). Multiple imputation was used to replace missing information on educational level, smoking, or BMI—the rate of missingness for these variables is presented in Table 1. Tests of homogeneity of HRs over k strata (e.g., SARS‐CoV‐2 without previous vaccination vs. previous vaccination, or over BMI strata) were performed where the HRs were weighted by precision (1/Standard Error of HR _k ) and compared to the Chi‐squared (k‐1) distribution. Statistical analyses were performed using IBM SPSS Statistics, version 27.0 (IBM Corp) and Gauss (Aptech Systems Inc.).

TABLE 1.

Maternal and pregnancy characteristics by the presence of a positive SARS‐CoV‐2 test before gestational week 37. Singleton pregnancies in Sweden with estimated date of birth between March 1, 2020 and May 31, 2022.

	SARS‐CoV‐2 infection during pregnancy
	Yes, n = 20,222		No, n = 213,113
	n	(%)	n	(%)	p value a
Year of conception					<0.001
2019	835	(4.1)	64,296	(30.2)
2020	9870	(48.8)	95,661	(44.9)
2021	9517	(47.1)	53,156	(24.9)
Dominant virus variant period at positive test
Original	6755	(33.4)
Alpha	5239	(25.9)
Delta	2008	(9.9)
Omicron BA 1 and 2	6220	(30.8)
Age					<0.001
<20 years	90	(0.4)	1433	(0.7)
20–24 years	1015	(5.0)	11,688	(5.5)
25–29 years	4411	(21.8)	46,842	(22.0)
30–34 years	6703	(33.1)	70,066	(32.9)
35–39 years	3382	(16.7)	35,578	(16.7)
40+ years	957	(4.7)	9835	(4.6)
Parity					<0.001
Primipara	7858	(38.9)	91,776	(43.1)
Multipara	12,361	(61.1)	121,262	(56.9)
Maternal country of birth					<0.001
Sweden	13,380	(66.2)	140,698	(66.0)
Other Europe	1339	(6.6)	14,731	(6.9)
Outside Europe	4000	(19.8)	41,010	(19.2)
BMI (kg/m2)					<0.001
<18.5	342	(1.7)	4614	(2.2)
18.5–25	9888	(48.9)	108,863	(51.1)
25–29	5686	(28.1)	57,007	(26.7)
≥30	3591	(17.8)	33,941	(15.9)
Not known	715	(3.5)	8689	(4.1)
Smoking					<0.001
No	18,648	(92.2)	194,978	(91.5)
Yes	549	(2.7)	7098	(3.3)
Not known	1025	(5.1)	11,037	(5.2)
Region of delivery					<0.001
Region East	5484	(27.1)	53,591	(25.1)
Region Mid	3271	(16.2)	32,778	(15.4)
Region SouthEast	1873	(9.3)	21,680	(10.2)
Region South	3061	(15.1)	35,467	(16.6)
Region West	3617	(17.9)	36,586	(17.2)
Region North	1413	(7.0)	16,337	(7.7)
Educational level					<0.001
<Elementary school	291	(1.4)	3383	(1.6)
Elementary school	851	(4.2)	9968	(4.7)
High school	6167	(30.5)	63,135	(29.6)
College/University	9775	(48.3)	102,218	(48.0)
Not known	3138	(15.5)	34,409	(16.1)

Abbreviations: BMI, body mass index; SARS‐CoV‐2, severe acute respiratory syndrome coronavirus 2.

^a Chi‐square analyses testing overall heterogeneity.

3. RESULTS

Of the study population of 233,335 women, 20,222 (8.7%) had a positive SARS‐CoV‐2 test during pregnancy. The prevalence of SARS‐CoV‐2 infection among pregnant women increased during the study period. In the total study population, 10,712 (4.6%) gave birth preterm. Of the 914 women who gave birth preterm after SARS‐CoV‐2 infection, 684 (74.8%) had a spontaneous and 230 (25.2%) had an iatrogenic PTB (see Figure 1 and Table S2). The occurrence of stillbirth was equivalent in the infected and noninfected groups (n = 52, 0.3% vs n = 610, 0.3%). Table 1 shows maternal characteristics by SARS‐CoV‐2 infection status during pregnancy. Statistically significant differences in age, parity, country of birth, smoking, BMI, maternal country of birth, and Swedish birth region were found between women who had a positive SARS‐CoV‐2 test during pregnancy and those who had not, but the differences were minor and of hardly any clinical relevance.

FIGURE 1.

Flowchart of the study population and the incidence of preterm birth in the severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) exposure groups. GW, gestational week; PCR, polymerase chain reaction; PTB, preterm birth.

The respective rates of vaccination before onset of a positive SARS‐CoV‐2 test during pregnancy and of admission to hospital due to COVID‐19 by year of conception are presented in Table S1. Vaccination against SARS‐CoV‐2 became available in the spring of 2021. Among women with a positive SARS‐CoV‐2 test during pregnancy, the proportion requiring inpatient care or admission to the intensive care unit (ICU) decreased over time (Table S1).

Table S2 shows gestational duration by timing of positive SARS‐CoV‐2 test during pregnancy. Among 4222 women who had a positive SARS‐CoV‐2 test at 22–27 GWs, 26 (0.6%) delivered extremely preterm (22–27 GWs). The corresponding rate in pregnancies without a positive SARS‐CoV‐2 test was 553/203,335 (0.3%). Among 2869 women with a positive SARS‐CoV‐2 test at 28–31 GWs, 38 delivered very preterm (1.3%). The corresponding rate in pregnancies without a positive SARS‐CoV‐2 test was 0.4%. Finally, among 3509 women who had a positive SARS‐CoV‐2 test at 32–36 weeks, 6.8% delivered at 32–36 weeks, compared with 3.9% among women with no positive SARS‐CoV‐2 test during pregnancy. The majority of PTBs were spontaneous. Table S1 shows the number of women with a positive SARS‐CoV‐2 test during pregnancy by timing of the positive test, gestational duration, and admission to hospital or ICU.

Figure 2 shows the risk of birth at 0–13 days or at 14–20 days, respectively, after a positive SARS‐CoV‐2 test, compared with the corresponding risk among all pregnancies without a positive SARS‐CoV‐2 test that were ongoing at mid‐week of the respective GW. As the figure shows, the PTB risk was markedly increased within 2 weeks after a positive SARS‐CoV‐2 test, compared with pregnancies without a positive SARS‐CoV‐2 test. However, more than 2 weeks or more after a positive SARS‐CoV‐2 test, the PTB rate was similar to that among pregnancies without a positive SARS‐CoV‐2 test.

FIGURE 2.

Risk for birth within 14 days or 14–20 days, respectively, by timing of severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) infection, compared to ongoing pregnancies (counted from mid‐week). Moving averages.

Cox regression analyses with time‐varying covariates were performed to investigate the overall risk of PTB in pregnancies with a positive SARS‐CoV‐2 test, compared with ongoing pregnancies without a positive SARS‐CoV‐2 test. Within 14 days after a positive SARS‐CoV‐2 test, a significantly increased PTB risk was found (p < 0.001), and there was a significant interaction between a positive test and GW at positive SARS‐CoV‐2 test (p = 0.047). Two or more weeks after a positive SARS‐CoV‐2 test, only a modest increase in aHR for PTB was observed, albeit statistically significant (aHR, 1.2; 95% CI, 1.0–1.4), and no interaction with GW at infection was detected. The resulting HRs and aHRs, respectively, for PTB by gestational duration at positive test are displayed in Figure 3. This figure shows that the PTB risk was higher after infection earlier in pregnancy, compared with closer to term. Figure 3 also shows the actual HRs for PTB, obtained by dividing the observed rates of PTB within 2 weeks after a positive SARS‐CoV‐2 test by the rate of PTB within 2 weeks among ongoing pregnancies without a positive SARS‐CoV‐2 test (at each GW). Women who had a positive SARS‐CoV‐2 test but did not require hospital admission were at lower risk of giving birth preterm within 2 weeks than were women who did require hospitalization. However, the former were still at increased risk of delivering preterm, compared to women without a positive SARS‐CoV‐2 test (aHR, 2.8; 95% CI, 2.4–3.3). The HR did not vary by timing of positive SARS‐CoV‐2 test (see Figure 3).

FIGURE 3.

Hazard ratio (HR) for preterm birth within 2 weeks after severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) infection. Adjusted for maternal age, parity, educational level, year of birth, and maternal country of birth. (A) Observed number: risk for birth within 2 weeks after SARS‐CoV‐2 infection versus risk for birth within 2 weeks among all ongoing SARS‐CoV‐2‐free pregnancies at mid‐week.

The aHR for PTB among pregnancies conceived during 2021 was computed, stratified for vaccine status. An increased risk of PTB was detected among women who had a positive SARS‐CoV‐2 test after vaccination, compared to women without a positive SARS‐CoV‐2 test during pregnancy (aHR, 1.6; 95% CI, 1.1–2.1). The corresponding aHR among women who had a positive SARS‐CoV‐2 test without previous vaccination was 3.2 (95% CI, 2.3–4.3). The difference between the respective HR estimates (a positive SARS‐CoV‐2 test after vaccination, compared with a positive SARS‐CoV‐2 test with no previous vaccination) was statistically significant (p for homogeneity = 0.007).

Cox regression analyses were performed to detect a possible interaction between maternal BMI and risk of SARS‐CoV‐2‐related PTB. No significant heterogeneity over BMI strata was detected (p for homogeneity = 0.86).

Table 2 shows the HRs and aHRs, for PTB at 22–27, 28–31, and 32–36 GWs, respectively. The HR estimates correspond quite well with the estimates shown in Figure 2. A clear interaction between the GW of a positive SARS‐CoV‐2 test and the risk of PTB within 2 weeks, compared with pregnancies without a positive SARS‐CoV‐2 test, is demonstrated. The aHR varied from 6.0 for PTB at 22–27 GW to 3.2 for PTB at 32–36 GWs. The estimates changed marginally when iatrogenic PTB were excluded. Thus, the association was mainly driven by spontaneous PTB. The risk of PTB (all types) after a positive SARS‐CoV‐2 test decreased during the pandemic from 2020 to 2022. Table 2 also shows that the PTB risk was higher when the Alpha and Delta variants were dominant (aHR, 10.2 and 7.8 at 22–27 GWs to 4.0 and 2.4 at 32–36 GWs) compared with during the Omicron period (aHR, 3.4 to 1.8). Women who were admitted to hospital due to COVID‐19 were at markedly higher risk of delivering preterm (aHR, 20.8 at 22–27 GWs to 7.7 at 32–36 GWs) than were women without a positive SARS‐CoV‐2 test during pregnancy. For women with a positive SARS‐CoV‐2 test but without the need of hospitalization, the corresponding aHRs were significantly lower (aHR, 4.4 to 2.6), but still statistically significant.

TABLE 2.

Hazard ratio (HR) for preterm birth within 2 weeks of verified SARS‐CoV‐2 infection by timing of preterm birth. Adjustments were made for maternal age, parity, educational level, and year of birth, if not otherwise specified.

	Preterm birth class
	22–27 weeks		28–31 weeks		32–36 weeks
	HR	95% CI	HR	95% CI	HR	95% CI
Crude	5.3	3.5–8.1	5.9	4.3–8.1	3.1	2.6–3.6
Adjusted	6.0	3.9–9.2	6.1	4.4–8.4	3.2	2.7–3.7
Spontaneous preterm births (adjusted)	9.9	4.7–21.0	7.0	3.8–12.6	3.5	2.6–4.8
Dominant virus variant period at positive test (adjusted a )
Original	4.8	2.0–11.7	8.0	4.6–13.8	5.7	4.5–7.2
Alpha	10.2	5.2–19.8	7.9	4.2–14.8	4.0	3.0–5.4
Delta	7.8	2.5–24.5	14.0	7.4–26.2	2.4	1.4–4.0
Omicron BA 1 and 2	3.4	1.4–8.2	2.0	0.9–4.5	1.8	1.4–2.4
No hospital admission for SARS‐CoV‐2 during pregnancy (adjusted)	4.4	2.7–7.4	3.6	2.3–5.6	2.6	2.2–3.2
Cases admitted for COVID‐19 (adjusted)	20.8	9.9–43.9	27.4	17.2–43.6	7.7	5.7–10.3

Abbreviations: CI, confidence interval; HR, hazard ratio; SARS‐CoV‐2, severe acute respiratory syndrome coronavirus 2.

^a Not adjusted for year of birth.

4. DISCUSSION

This nationwide, population‐based cohort study found a statistically significant and clinically relevant association of SARS‐CoV2 infection and PTB with aHR between 3.2 and 6.1, depending on gestational age at PTB. The increased PTB risk after SARS‐CoV‐2 infection during pregnancy was limited to 2 weeks after infection onset and the association was driven by spontaneous PTB. Furthermore, there was a significant interaction between a positive SARS‐CoV‐2 test and GW at a positive test. The aHR was doubled at earlier (22–27 w), compared with later (32–36 w), gestational ages. Women with a positive SARS‐CoV‐2 test after SARS‐CoV‐2 vaccination were not at an increased risk of PTB, compared to those without a positive test during pregnancy, but were at significantly lower risk than women who had a SARS‐CoV‐2 infection without vaccination.

Previous studies partly support our finding of increased PTB risk within 2 weeks after infection onset, especially at earlier preterm gestational ages. The CRONOS study and a study by Piekos et al. reported increased PTB risk among women with COVID‐19 during early pregnancy, compared with later gestational ages [2, 7]. However, Neelam et al. reported increased PTB risk mainly after third‐trimester infection [30]. Few previous studies have specifically investigated the infection‐to‐PTB interval; two found increased PTB risk within the first 4 weeks after infection. Darling et al. found an increased risk of both iatrogenic and spontaneous PTB during the first 3 days and during 4–30 days after infection onset [11]. CRONOS found increased PTB risk during the first 4 weeks after infection onset, compared with thereafter [7]. These studies were mostly based on patients testing positive due to universal testing at birth. Some asymptomatic women might have been diagnosed because of admission for preterm labor, that is, reverse causality. Smith et al. found that women delivering preterm had higher odds of having had a symptomatic SARS‐CoV‐2 infection within 30 days before birth, compared to earlier during pregnancy. This study was based on digital self‐reporting with the risk of self‐selection bias [11, 12].

The increased PTB risk in our study was clearly driven by spontaneous PTB. Data from prior studies are conflicting regarding whether the SARS‐CoV‐2‐associated PTB risk is due to spontaneous or iatrogenic PTB. One Internet‐based cohort study found a predominantly iatrogenic PTB risk increase following severe COVID‐19 at later preterm gestational ages, although the risk of spontaneous PTB was also elevated [12]. A Swedish study considering testing strategies found that a positive SARS‐CoV‐2 test within 10 days before birth was associated with increased risk of iatrogenic, but not of spontaneous, PTB. This was related to universal testing; however, nonuniversal testing was associated with an increased risk of both PTB types [31].

Routine management of SARS‐CoV‐2‐infected pregnant women differed among countries and during different periods of the pandemic. In Sweden, SARS‐CoV‐2 infection was never an indication for inducing labor, unless the woman was critically ill. In China and other countries, induction of labor and cesarean section in infected women was more common [32, 33]. It is also plausible that iatrogenic PTB was more frequent in the early pandemic stages due to a lack of knowledge about the virus and protective measures such as thromboprophylaxis and vaccination.

The pathogenesis behind spontaneous PTB is not fully understood, but it is well known that infection or infection‐associated inflammation is associated with PTB, especially early PTB [14]. Intrauterine transmission of SARS‐CoV‐2 is rare [34], and the effect on PTB rates is likely related to the systemic inflammation caused by the primary infection. The infection usually starts in the host's airways, triggering a delayed inflammatory reaction which, through the spread of IFN‐I and IFN‐III, induces inflammatory responses in peripheral organs that can persist for 7–10 days [35]. This inflammatory response may also affect the placenta, cervix, and uterus, which might explain the findings of increased PTB risk within 14 days after infection onset, and the particularly increased risk of early PTB [36].

Admission to hospital for SARS‐CoV‐2 infection markedly increased the risk of PTB within 2 weeks after infection onset. Several studies report an increased risk after symptomatic SARS‐CoV‐2 infection [12, 37], although other studies did not confirm this [38]. A population‐based study with data from the five Nordic countries during March 1 to June 30, 2020 found that the overall PTB rate was higher in women admitted due to COVID‐19 than in women giving birth during 2018. These data were based on different testing strategies, likely affecting results [6].

In our study, the increased PTB risk after SARS‐CoV‐2 infection persisted after excluding women admitted to the hospital, suggesting the existence of risk‐influencing factors related to the infection itself, regardless of severity. These findings correlate with findings for other viral diseases with general infectious symptoms, such as influenza, and support infection and inflammation as one of the main causes of PTB, especially during early gestation [39, 40].

Our results show an increased risk of PTB after SARS‐CoV‐2 infection during the Alpha and Delta periods compared with during the Omicron period. This corresponds well to previous studies, where specifically the Delta variant is identified as being associated with worse pregnancy outcomes compared with both pre‐Delta and Omicron variants [10, 41, 42].

Previous studies, including one in the Swedish setting, have established that vaccination before or during pregnancy does not increase PTB risk [43, 44]. A registry study from Scotland, including 18,399 women vaccinated against SARS‐CoV‐2 during pregnancy, found no increased PTB rate compared with the general pregnant population [45]. Several cohort studies confirm these results [46, 47]. A few studies report that the PTB risk decreases after vaccination [2, 48]. An Australian cohort study found a significant reduction in PTB (overall, spontaneous, and iatrogenic) rates among vaccinated, compared with unvaccinated, pregnant women [48]. Our data suggest that vaccination before SARS‐CoV‐2 infection protects against SARS‐CoV‐2‐associated PTB, further supporting the recommendation that pregnant women be vaccinated.

The major strength of this study is the comprehensive population‐based dataset based on high‐quality Swedish healthcare and quality registers, with close to 100% of the Swedish pregnant population included in the dataset, all exposure and outcome measures recorded by healthcare professionals, and low risk for loss to follow‐up. Despite the comprehensive dataset, there are only few observations in some of the subgroups and meta‐analysis with comparable datasets is needed to study associations for these groups.

As pregnant women were recommended to test even in case of mild symptoms and report positive test results to their midwife for assessment of thrombosis risk, pregnant women in Sweden can be assumed to have comparably few unregistered symptomatic infection episodes. Our dataset therefore captures not only positive tests of women admitted for labor or pregnancy complications, but also women with mild infection. Thus, the risk of reverse causality is attenuated compared to studies based on infection status at admission to the hospital only. Further, classifying infected women as not infected may attenuate any true association between SARS‐CoV‐2 infection and the risk of PTB.

Our results indicate an increased risk of PTB during the Delta variant period. The risk was still elevated during the Omicron period, which is also confirmed by other studies [49]. These findings suggest that viral virulence does not necessarily correlate with the risk of PTB. Therefore, PTB should be studied as an independent outcome across different SARS‐CoV‐2 variants.

As in all observational studies, the results might have been affected by residual confounding. However, we adjusted for relevant confounders and their temporal relationship to exposure.

The results reflect a Swedish context and might not be fully generalizable to other countries. There are differences that make the Swedish example extra interesting to analyze: the maternity care program in Sweden is free of charge, with approximately 99% participation. Sweden has among the lowest worldwide levels of maternal and perinatal mortality. During the pandemic, Sweden adopted an internationally controversial strategy; lockdowns were avoided, and the authorities mostly relied on the population voluntarily complying with recommendations and restrictions [50].

5. CONCLUSIONS

In this nationwide, population‐based study from Sweden, we found an increased risk of spontaneous PTB after SARS‐CoV‐2 infection during pregnancy within the first 2 weeks after infection onset. The PTB risk was significantly higher after infection earlier in pregnancy, compared with closer to term. Especially, viral variants Alpha and Delta were associated with PTB risk. Vaccination counteracted the increase in PTB risk.

Pregnant women are at particularly high risk of PTB within 14 days after a SARS‐CoV‐2 infection, and should be counselled on PTB symptoms to ensure they seek medical care without delay, especially during the earlier preterm gestational period. Further, these results support the recommendation that pregnant women be vaccinated against SARS‐CoV‐2.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

ETHICS STATEMENT

This study was approved by the Swedish Ethical Review Authority (May 21, 2021, number 2021‐00877).

Supporting information

Supporting Information

Supporting Information

DATA AVAILABILITY STATEMENT

Research data are not shared.