mRNA Covid-19 vaccines in pregnancy: A systematic review

Objective Pregnancy is a known risk factor for severe Coronavirus disease 2019. It is important to develop safe vaccines that elicit strong maternal and fetal antibody responses. Methods Registries (ClinicalTrials.gov, the WHO Clinical Trial Registry, and the European Union Clinical Trial Registry) and databases (MEDLINE, ScienceDirect, Cochrane Library, Proquest, Springer, medRxiv, and bioRxiv) were systematically searched in June 20–22, 2021, for research articles pertaining to Covid-19 and pregnancy. Manual searches of bioRxiv and medRxiv were also conducted. Inclusion criteria were studies that focused on Covid-19 vaccination among pregnant women, while review articles and non-human studies were excluded. Infection rate, maternal antibody response, transplacental antibody transfer, and adverse events were described. Results There were 13 observational studies with a total of 48,039 pregnant women who received mRNA vaccines. Of those, three studies investigated infection rate, six studies investigated maternal antibody response, seven studies investigated antibody transfer, three studies reported local adverse events, and five studies reported systemic adverse events. The available data suggested that the mRNA-based vaccines (Pfizer–BioNTech and Moderna) can prevent future SARS-CoV-2 infection. These vaccines did not show clear harm in pregnancy. The most commonly encountered adverse reactions were pain at the injection site, fatigue, and headache, but these were transient. Antibody responses were rapid after the first vaccine dose. After the booster, antibody responses were stronger and associated with better transplacental antibody transfer. Longer intervals between first vaccination dose and delivery were also associated with higher antibody fetal IgG and a better antibody transfer ratio. Conclusions The SARS-CoV-2 mRNA vaccines are encouraged for pregnancy. These vaccines can be a safe option for pregnant women and their fetuses. Two vaccine doses are recommended for more robust maternal and fetal antibody responses. Longer latency is associated with higher fetal antibody responses. Further research about its long-term effect on pregnancy is needed. Systematic review registration PROSPERO (CRD42021261684).


Introduction
Since the beginning of the Covid-19 pandemic, extensive efforts have been made to end this global disaster. One of the most effective approaches is vaccination against severe acute respiratory syndrome Coronavirus 2 (SARS-CoV-2). The efficacy and safety of Coronavirus disease 2019 (Covid- 19) vaccines have been demonstrated in adults across a range of demographics [1], but their impact on pregnant women remain unclear due to insufficient information being available. In fact, pregnant women are still at a higher risk of acquiring viral respiratory infections and severe pneumonia due to the unique physiological changes in their immune and cardiopulmonary systems [2,3]. Although most pregnant women suffer only mild to moderate symptoms, SARS-CoV-2 infection is more severe in pregnant women than in others, with increased risks of hospital admission, intensive care unit stay, and death [4].
Despite their higher risk of SARS-CoV-2 infection, pregnant and lactating women were not included in any initial Covid-19 vaccine trials, resulting in a lack of data to guide vaccine decision-making in these populations [5]. A previous study revealed that most pregnant women with Covid-19 admitted to hospital were asymptomatic, which allows these undetected patients to transmit the virus to others [6][7][8]. This shows that efforts to prevent SARS-CoV-2 infection, such as by vaccination, are critical for investigations on this population. For this reason, we systematically review the latest evidence on Covid-19 vaccination to summarize its efficacy, immunogenicity, and safety profile in pregnant women.

Methods
This systematic review adhered to PRISMA (Preferred Reporting Items for Systematic Review and Meta-Analysis) 2020 guidelines [9], and has been registered in the PROSPERO database (CRD42021261684). the paper was written in English. Our primary outcomes included infection rate, maternal antibody titer, and local and systemic adverse events. Our secondary outcomes included neonatal outcome, cord blood antibody titer, and placental transfer ratio. We excluded review articles, irrelevant studies, non-human studies, and duplicates.

Search strategy and selection of studies
We conducted comprehensive keyword searches, on June 20-22, 2021, to find articles published in trial registries (ClinicalTrials.gov, the WHO Clinical Trial Registry, and the EU Clinical Trial Registry) and databases (MEDLINE, ScienceDirect, Cochrane Library, Proquest, and Springer). We added one eligible study on July 13, 2021. Our research is limited to studies involving humans reported in english. Manual searches, including in bioRxiv and medRxiv, and a bibliographical search were also conducted to obtain additional evidence. The following keywords were used: "[(SARS-CoV-2) OR (Covid- 19)] AND [(pregnancy) OR (pregnant)] AND [(vaccine) OR (vaccination)]." Details of the search strategies are available in (S1 File). We exported all studies retrieved from the electronic searches into Mendeley reference manager to remove duplicates and perform screening. The two review authors (NRP and IAW) independently screened the titles and abstracts of the articles to identify potentially eligible studies and subsequently screened the full texts independently. Any disagreements between the two review authors were resolved by discussion until a consensus was reached. Excluded studies are described in the PRISMA flow diagram alongside the reasons for their exclusion (Fig 1).

Data extraction
Three authors (NRP, IAW, and DSB) independently extracted relevant data using a structured and standardized form from each selected study. The following information was extracted: first authors' names and publication year, study design, country of origin, sample size, gestational age at first vaccination, sample size, age, sample collection, vaccine type, and outcomes (infection rate, maternal titer antibody, cord blood titer antibody, placental transfer ratio, and local and systemic adverse events). Any disagreements between the review authors were resolved by discussion until a consensus was reached.

Study selection
The search strategy yielded 4,218 records. After screening the titles and abstracts, 48 potentially eligible articles were reviewed. After full-text assessment, 13 studies were included in this systematic review. The process of study selection in this review is described in the PRISMA flow diagram (Fig 1), along with the reasons for exclusion.

Quality assessment
Nine cohort studies were assessed using the NOS assessment tool and considered to be of high quality (S1 Table), while the assessment of the quality of case reports, case series, and crosssectional studies using the JBI critical appraisal checklist is summarized in S2-S4 Tables.

Study characteristics
A total of 13 observational studies with 48,039 pregnant women who received Covid-19 vaccines were included in this systematic review. Among these studies, 10 (6 cohort, 1 cross-sectional, 1 case series, and 2 case reports) were conducted in the USA, while the other three (all cohort studies) were carried out in Israel in 2021. The detail of study characteristics was presented in Table 1.

Patient characteristics
The median [interquartile range (IQR)] of the mean or median gestational age across the studies was 32 (31.5-33.2) weeks. All pregnant women reported receiving an mRNA vaccine, either Pfizer-BioNTech or Moderna vaccine, except for four pregnant women who received a vaccine of an unknown type. The median or mean gestational age of the participants at the first vaccination were mostly 30 weeks or more, except in two studies, which reported values of 23.2 weeks and �28 weeks. Some studies compared vaccinated pregnant women with unvaccinated pregnant women, either naturally infected or not infected, or vaccinated non-pregnant women.

Outcomes
Vaccine efficacy was described in terms of the infection rate, which was measured as the proportion of individuals infected with Covid-19. It was reported at different timeframes across the studies. Goldshtein et al. [24] grouped it into �10 days, 11-27 days, and � 28 days. Meanwhile, it was grouped into �14 days and >14 days by Shimabukuro et al. [12], and Trimesters I, II, and III by Theiler et al. [21]. Immunogenicity was measured in terms of maternal antibody response, fetal antibody response, and transplacental antibody transfer. Safety outcomes were measured as the adverse events, maternal outcomes, and neonatal outcomes. Adverse events were divided into local and systemic, local adverse events included injection-site pain and soreness, while systemic adverse events included fatigue, headache, myalgia, chills, fever, and nausea. Maternal outcomes were divided into pregnancy outcomes and delivery outcomes. The detail of outcomes of the individual studies was presented in Tables 2 and 3.

Maternal antibody response
Maternal antibody responses were investigated by six observational studies [13-16, 19, 22]. Vaccination induced immunoglobulin (Ig)G and IgM production in 71% (87/122) of pregnant women, 16% (19/122) of pregnant women produced only IgG, while in 13% (16/122) neither IgG nor IgM was detectable [16]. Vaccination provided a rapid immunological response after the first dose, while infection provided a gradual immunological response. Moreover, the administration of a second dose further increased the IgG level among vaccinated women [13,22]. Spike and receptor-binding domain (RBD) IgG titers rose rapidly after the first dose (p < 0.0001 and p < 0.01, respectively); in addition, after receiving the booster, they became higher than after the first dose (p < 0.05 and p < 0.001, respectively) [13]. Vaccination elicited IgG responses against spike (S)1, S2, and RBD, but not nucleocapsid (N) protein. Meanwhile, infection elicited all IgG responses against S1, S2, RBD, and N protein. Although S1 IgG, S2 IgG and RBD IgG responses were observed in both vaccinated and infected pregnant women, the S1 IgG and RBD IgG levels in vaccinated pregnant women were higher (p = 0.0009 and p = 0.0045, respectively), while S2 IgG and N IgG were higher in infected pregnant women (p = 0.0016 and p < 0.0001, respectively) [22]. Meanwhile,  reported that spike IgG titer was higher upon vaccination than upon natural infection in pregnant women [13].
Antibody responses in pregnant and non-pregnant women were evaluated by Collier et al. (2021). They reported the median IgG levels in vaccinated and infected pregnant women. The RBD IgG titers were 27,601 and 1,321, while the neutralizing antibody titers were 910 and 148, respectively. The median RBD IgG titers in vaccinated and infected non-pregnant women were 37,839 and 771, while for the neutralizing antibody they were 901 and 193, respectively [14].

Transplacental antibody transfer
Seven observational studies investigated transplacental antibody transfer [14,16,17,19,20,22,23]. A prospective case series reported that IgG was detected in 89% (25/28) of cord blood, but no cases had detectable IgM [20]. Moreover, antibody against SARS-CoV-2 RBD and neutralizing antibody were observed in cord blood. In vaccinated women, the maternal and cord blood RBD IgG levels were 14,953 AU and 19,873 AU, while for the neutralizing antibody they were 1,016 AU and 324 AU, respectively [14]. IgG against S protein was also detected in cord blood, with a concentration of 193 (111-260) AU/mL, and its transfer ratio was 0.44 (0.25-0.61). Furthermore, the concentration of IgG against RBD was 3,494 (1,817-6,163) AU/mL and its transfer ratio was 0.34 (0.27-0.56) [19]. Two different case reports described a mother who received two doses of BNT162b2 vaccine and a mother who received one dose of Moderna vaccine; they reported that SARS-CoV-2-specific IgG was detected in maternal blood and cord blood at a titer 1:25,600, while the cord blood IgG concentration was determined to be 1.31 U/mL [17,23].

Local adverse events
Three observational studies reported local adverse events [12,13,18]. Among pregnant women, injection-site pain was the most common adverse event for both the Pfizer-BioNTech and Moderna vaccines. Following the Pfizer-BioNTech vaccination, as many as 84% (7,602/ 9,052) in the first dose and 89% (5,886/6,638) in the second dose experienced injection-site pain. Meanwhile, for the Moderna vaccine, 93% (7,360/7,930) and 96% (5,388/5,635) experienced injection-site pain following the first and second doses, respectively [12]. It was also reported that 88% (73/84) of pregnant women experienced injection-site soreness following the first vaccination and 57% (44/84) did so following the second dose. Additionally, 75% (12/ 16) of non-pregnant women experienced injection-site soreness after the first and second doses of the vaccine [13].
In pregnant and non-pregnant women, sore arms or pain were observed in 97% (37/38) of pregnant women and 90% (894/991) of non-pregnant women following the Pfizer-BioNTech and Moderna vaccinations [18].

Systemic adverse events
Systemic adverse events were reported in five observational studies [12-14, 18, 24]. Following first and second vaccinations with the Pfizer-BioNTech vaccine, the six most common systemic adverse events were fatigue [27% (2, 909/5,635)]. Numerically, the incidence of each event was higher for the second dose. Moreover, the Moderna vaccine had more systemic adverse events than the Pfizer-BioNTech vaccine [12]. Seizure was reported in a woman who received the mRNA vaccine (p = 0.0369), but this patient had a history of seizure disorder and the anticonvulsant level in the blood was borderline low [18].

Main findings
Two studies demonstrated the efficacy of the Pfizer-BioNTech and Moderna vaccines for preventing future SARS-CoV-2 infection in pregnant women [21,24]. Following vaccinations with the Pfizer-BioNTech and Moderna vaccines, the vast majority of pregnant women had injection-site pain or soreness [12,13,18]. The most common systemic adverse events were fatigue, headache, chills, myalgia, fever, and nausea [12,18]. There were stark differences in the adverse events reported between the study by Goldshtein et al. [24] and other studies [12,13,18]. This was likely due to the different methodologies used. That study from Israel used International Classification of Diseases codes, followed by manual chart review for codes for suspected adverse events. That could have resulted in smaller numbers of complaints than in self-reported methods such as the use of v-safe software. The incidence of these systemic adverse events was higher after the second dose than after the first dose [12][13][14]. Numerically, more individuals experienced systemic adverse events in the Moderna vaccine group than in the Pfizer-BioNTech group [12]. The rates of adverse events did not differ significantly between pregnant and non-pregnant women, with the exception of seizures. However, the woman affected by seizures was known to have a history of seizure disorder and the anticonvulsant level was measured as borderline low [21]. Interestingly, the maternal and neonatal outcomes did not differ between vaccinated and unvaccinated pregnant women [21,24].
Maternal antibody responses have been reported to develop following vaccination. Upon vaccination, antibody responses were rapidly established, while through infection they formed gradually [22]. The IgG and IgM titers against SARS-CoV-2 were significantly increased after vaccination. The response was also increased after a booster vaccination was given [13,15]. Although the vast majority of pregnant women exhibited IgG seroconversion, IgM seroconversion was observed in a minority of pregnant women [20]. After vaccination, IgG against S1, S2, and RBD formed, while IgG against S1, S2, RBD, and N protein formed following natural infections. Moreover, S1 and RBD IgG levels were found to be higher in vaccinated pregnant women. Meanwhile, S2 and N IgG levels were found to be higher in naturally infected pregnant women [22]. Furthermore, RBD IgG and neutralizing antibody levels were higher in vaccinated individuals than in naturally infected ones [14].
Transplacental antibody transfer was also reported. Cord blood antibody and maternal antibody levels were reported to be approximately equal [20]. Additionally, latency and number of doses were correlated with the intensity of transplacental antibody transfer [16,20]. The longer the latency, the better the transplacental antibody transfer and the higher the IgG. The offspring of mothers receiving two doses of vaccine were also shown to have higher IgG levels [20]. Finally, there was no difference in maternal-fetal IgG response between infected and vaccinated cases [22].

Findings in other studies
Randomized controlled trials showed that two-dose regimens of both Moderna and Pfizer-BioNTech vaccines provided excellent efficacy at preventing Covid-19 illness, being 94.1% and 95%, respectively, with no associated safety concerns [25,26]. As reported in a pregnancy cohort [12], more frequent systemic adverse reactions after the second dose of vaccine were also found in non-pregnant individuals [25,26]. However, these adverse reactions were only transient, being resolved within a few days [25,26].
The reported spontaneous abortion rate of 12.6% of completed pregnancies within the vaccinated group was still within an acceptable range, since their underlying medical conditions were unknown [27]. For example, Center for Disease Control and Prevention surveillance in 2015 reported an abortion rate of 188 in 1,000 live births [28]. Some vaccinations, including the tetanus toxoid, reduced diphtheria toxoid, and acellular pertussis (Tdap) and killed influenza vaccines, administered to pregnant women showed no increased risk of adverse outcomes [29,30]. The Tdap vaccine was most often given in the first trimester of pregnancy [29]. Meanwhile, the majority of participants in Covid-19 vaccination studies were in the third trimester of pregnancy [15,17,[19][20][21][22][23]. The optimal timing of administering Covid-19 vaccine during pregnancy has remained unclear. Although longer latency was associated with better transplacental antibody transfer [20,22], data from studies on influenza vaccine showed better immunogenicity and protection upon administration in the third trimester rather than in earlier trimesters [31].
Passive immunity in neonates can potentially protect against SARS-CoV-2 infection. However, this passive immunity may change due to placental sieving [32], depending on the gestational age at first vaccination or infection [20,22]. In Covid-19 infection, poor transplacental antibody transfer was exclusively observed only in the third trimester of pregnancy, even though the maternal antibody response was significantly higher [33]. Interestingly, lower transplacental antibody transfer upon natural viral infection, such as in Zika and Dengue Viruses [34,35], compared to vaccinations, such as in pertussis and influenza vaccines [36][37][38], was also observed in SARS-CoV-2 [20,39].
High vaccination coverage is important to achieve a sufficient threshold for herd immunity in a population, which provides indirect protection for susceptible individuals from infected hosts [40]. This threshold varies across different infections and populations [40][41][42]. Additionally, a lower threshold would require higher vaccine efficacy [43,44]. This issue is important since antibody resistance among SARS-CoV-2 variants of concern was reported [45]. Moreover, mutations in some variants of concern changed the transmissibility, severity, and treatment efficacy of Covid-19, especially for neutralizing monoclonal antibody treatments due to immune escape [46][47][48][49][50][51].

Strengths and limitations
This systematic review used the recent available evidence to describe the efficacy, safety, and immunogenicity of Covid-19 mRNA vaccine in pregnancy. All studies included in this review were assessed as being of high quality. However, they were all observational studies due to the lack of reports about RCT on Covid-19 vaccination for pregnant women. These studies reported only on mRNA-type vaccines. Additionally, some major design differences included the number of patients, sample collection methods, and outcome definitions. Moreover, all available studies that were included were only from the United States and Israel.

Conclusion
This study suggests that mRNA vaccines, especially Pfizer-BioNTech and Moderna vaccines, can reduce the risk of future SARS-CoV-2 infections. These vaccines can induce antibody responses for pregnant women and their fetuses. Pregnant women should be given two doses of vaccine for more robust maternal and fetal antibody responses. Longer latency was associated with a more robust fetal antibody response. The majority of pregnant women who received the vaccination, as either the first or the second dose, would experience injection-site pain. Furthermore, the second dose of the vaccine would produce more systemic adverse events than the first one, and the administration of the Moderna vaccine is more often associated with systemic adverse events. Biologically speaking, we may conclude that vaccination does not affect pregnancy, delivery, or neonatal outcomes in the short term.