Immunogenicity and reactogenicity of heterologous COVID-19 vaccination in pregnant women

ABSTRACT

This open-labeled non-inferiority trial evaluated immunogenicity and reactogenicity of heterologous and homologous COVID-19 vaccination schedules in pregnant Thai women. 18–45-year-old pregnant women with no history of COVID-19 infection or vaccination and a gestational age of ≥12 weeks were randomized 1:1:1 into three two-dose primary series scheduled 4 weeks apart: BNT162b2-BNT162b2 (Group 1), ChAdOx1-BNT162b2 (Group 2), and CoronaVac-BNT162b2 (Group 3). Serum antibody responses, maternal and cord blood antibody levels at delivery, and adverse events (AEs) following vaccination until delivery were assessed. The 124 enrolled participants had a median age of 31 (interquartile range [IQR] 26.0–35.5) years and gestational age of 23.5 (IQR 18.0–30.0) weeks. No significant difference in anti-receptor binding domain (RBD) IgG were observed across arms at 2 weeks after the second dose. Neutralizing antibody geometric mean titers against the ancestral Wuhan strain were highest in Group 3 (258.22, 95% CI [187.53, 355.56]), followed by Groups 1 (187.47, 95% CI [135.15, 260.03]) and 2 (166.63, 95% CI [124.60, 222.84]). Cord blood anti-RBD IgG was correlated with, and equal to or higher than, maternal levels at delivery (r = 0.719, P < .001) and inversely correlated with elapsed time after the second vaccination (r = −0.366, P < .001). No significant difference in cord blood antibody levels between groups were observed. Local and systemic AEs were mild-to-moderate and more frequent in Group 2. Heterologous schedules of CoronaVac-BNT162b2 or ChAdOx1-BNT162b2 induced immunogenicity on-par with BNT162b2-BNT162b2 and may be considered as alternative schedules for primary series in pregnant women in mRNA-limited vaccine settings.

KEYWORDS:

• Heterologous schedules
• Homologous schedules
• BNT162b2
• ChAdOx1
• CoronaVac
• Thailand

View responses to this article:

Heterologous COVID-19 vaccination in pregnant women: Comment

Introduction

COVID-19 has infected more than 685 million individuals and caused more than 6.84 million deaths globally as of April 2023.¹ Pregnant women are one of the high-risk groups for severe COVID-19.^2–4 Increased rates of intensive care unit (ICU) admission, maternal death, stillbirths, and preterm births were recorded in unvaccinated pregnant women compared to non-pregnant women or non-infected pregnant women.^3–8 This susceptibility is further exacerbated by comorbidities (i.e., hypertension, diabetes, obesity) that increase the risk of complications during pregnancy (i.e., preeclampsia, preterm births, cesarean delivery).³ Despite this evident risk, this sub-population is typically excluded from COVID-19 vaccination studies.

With a growing body of evidence supporting vaccination in pregnant women,^4,8–10 the World Health Organisation (WHO), American College of Obstetricians and Gynaecologists (ACOG), Society for Maternal-Fetal Medicine (SMFM), and Royal College of Obstetricians and Gynaecologists (RCOG) now recommend pregnant women get vaccinated – regardless of gestational age – to protect themselves and their infants against COVID-19.^11–13 Out of the nine vaccines approved by the WHO, three are widely used and have demonstrated efficacy and safety in healthy individuals,^14–18 including: chimpanzee, adenovirus-vectored ChAdOx1-nCoV-19 (Oxford/AstraZeneca^™); BNT162b2 (Pfizer/BioNTech^™), an mRNA vaccine with full-length SARS-CoV-2 spike protein; and inactivated, whole-virus CoronaVac (Sinovac/Biotech^™). A recent systematic review and meta-analysis indicates that this efficacy and safety extends to pregnant women, with no reported adverse pregnancy outcomes.^4,9,19

Current COVID-19 vaccine studies in pregnant women emphasize the effects of mRNA vaccines (i.e., BNT162b2 and mRNA-1273 [Moderna^™]).^4,8–10,20 There is limited data regarding the use of other vaccine types, despite the wider use of viral vector and inactivated vaccines in developing countries.¹⁰ Heterologous schedules of ChAdOx1-BNT162b2 or CoronaVac-BNT162b2 generate similar or higher immunogenicity as homologous mRNA vaccination in healthy adults,^21,22 and are less reactogenic than homologous BNT162b2 or ChAdOx1.^21–25 The sequence of heterologous regimens is crucial. We and others have previously shown that BNT162b2, when administered as a second dose following CoronaVac- or ChAdOx1-primed heterologous schedules, maximized immune response in heterologous primary series.^21,22 Heterologous regimens also help address mRNA vaccine shortages in resource limited settings and provide flexible vaccine schedules.²² WHO recommends heterologous scheduling of emergency use listed vaccines in such settings with those available vaccine platforms.²² The implementation of such strategies, and the influence these regimens have on pregnant women’s altered immune systems, requires further evaluation in pregnant cohorts.²⁶

This study sought to evaluate the immunogenicity and reactogenicity of heterologous and homologous COVID-19 vaccination regimens of CoronaVac, ChAdOx1, and BNT162b2 in pregnant Thai women. We hypothesized that heterologous vaccination schedules would induce immunogenicity on-par with standard homologous schedules.

Materials and methods

Study design and participants

This prospective, randomized, open-labeled non-inferiority trial was conducted between August 2021 and 2022 at the Faculty of Medicine, Siriraj Hospital – a tertiary care and referral center in Bangkok, Thailand. 18–45-year-old women with singleton pregnancies at a gestational age ≥12 weeks were included in the study. Participants with a history of: smoking, drinking, drug-abuse; preeclampsia, placental abnormalities, premature labor, or fetal growth restriction; abnormal vaginal bleeding more than twice during their current pregnancy; prior SARS-CoV-2 infection, COVID-19 vaccination before screening; prior blood, blood component, plasma, immunoglobulin, antiviral, or antibody transfusion ≤90 days of the study; immunocompromised status, immunosuppressant use; underlying diseases or concomitant disorders (i.e., diabetes mellitus type 1 or 2, hypertension); or with severe vaccine or drug allergies were excluded from the study. Post-enrollment, participants with high-risk pregnancies that were deemed inappropriate to continue the study were discontinued by principal investigators. This study was approved by its institutional ethics committee (COA no. Si 481/2021) and conducted according to the Belmont Report, Declaration of Helsinki, and International Council on Harmonization’s Good Clinical Practice.

It was registered under Thai Clinical Trials on the 17th of July, 2021 (TCTR20210717002, https://clinicaltrials.gov/).

Study procedure

Following written informed consent, participants were randomized 1:1:1 using Sealed Envelope™ and allocated into three vaccine arms, each designated to receive a two-dose (first-second dose) primary vaccination series four weeks apart: BNT162b2-BNT162b2 (Group 1), ChAdOx1-BNT162b2 (Group 2), and CoronaVac-BNT162b2 (Group 3). Vital signs were measured, and physical examinations performed, for participants prior to each dose. All were monitored for 30 minutes after each vaccination for immediate adverse reactions (AEs). Immunogenicity was determined through blood samples drawn from the participants at baseline (before vaccination), four weeks after the first vaccination, and two weeks after the second vaccination. Additional blood samples from mothers and their umbilical cords were collected at the time of delivery. Participants were instructed to self-monitor and digitally record local and systemic AEs for seven days after each vaccination. Solicited local (injection-site specific) AEs included pain or tenderness, swelling or induration, and erythema. Solicited systemic AEs included fever (oral temperature > 38°C), headache, fatigue or malaise, myalgia, arthralgia, and nausea or vomiting. The severity of solicited AEs were graded 1–4 according to the United States Food and Drug Administration’s (USFDA) Toxicity Grading Scale for Healthy Adult and Adolescent Volunteers Enrolled in Preventive Vaccine Clinical Trials.²⁷ Any AEs that exacerbated preexisting conditions and resulted in a significant clinical deterioration in patients’ conditions were monitored and reported.

Chemiluminescent microparticle immunoassays

Chemiluminescent microparticle immunoassays (CMIA) were used to assess blood samples for SARS-CoV-2 spike protein anti-receptor binding domain (anti-RBD) IgG. This was performed using SARS-CoV-2 IgG II Quant (Abbott, Illinois, USA) detection assays that linearly measured antibody levels between 21.0 and 40,000.0 arbitrary units per mL (AU/mL). These units are subsequently converted to binding antibody units per mL (BAU/mL) per the WHO’s international standards through the equation $\mathrm{B}\mathrm{A}\mathrm{U}/\mathrm{m}\mathrm{L}=0.142\times \mathrm{A}\mathrm{U}/\mathrm{m}\mathrm{L}$. Antibody levels ≥ 50 AU/mL, or 7.1 BAU/mL, were defined as seropositive.

Microneutralization assays

Neutralizing antibodies (NAb) titers against SARS-CoV-2 were measured at baseline and two weeks after the second dose for all participants. Microneutralization assays were performed using ancestral Wuhan, Omicron BA.1, BA.2, and BA.5 strains. SARS-CoV-2 Wuhan/WA1/2020, Omicron BA.1, BA.2, and Vero-E6 cells were provided by Professor Florian Krammer and Dr. Juan Manuel Carreno Quiroz from the Icahn School of Medicine at Mount Sinai, NY, USA. The Omicron BA.5 variant was isolated from clinical samples collected at Siriraj Hospital, Mahidol University using Vero-E6 cell lines. Omicron spike genes were confirmed through sequencing. Viruses were passaged in Vero-E6 cells and stored at −80°C. Serum was heat inactivated at 56°C for 30 minutes. Heat inactivated sera were serially diluted 1:10–10,240 before adding 100 TCID₅₀ of SARS-CoV-2 in MEM media with 0.5% BSA and incubation at room temperature for 1 hour. Residual viral infectivity in the serum-virus mixture was assessed in quadruplicate wells of Vero-E6 cells incubated in serum-free media at 37°C, 5% CO₂. Viral cytopathic effect was read on the fifth day. NAb titers were calculated using the Reed-Muench method.^28–30 Non-linear regression curve fitting was performed to calculate 50% inhibitory dilution (ID₅₀). All samples were analyzed in a blinded manner. The limit of detection of ID₅₀ was at 10.

Statistical analyses

Non-inferiority of heterologous schedules were demonstrated through lower bounds of 95% confidence intervals (CIs) of the ratio of anti-RBD SARS-CoV-2 IgG antibody geometric mean concentrations (GMC) two weeks after the second vaccination being at least 0.67, as recommended by the WHO.³¹ Based on the parameters of the same immunogenicity study performed in healthy adults,²² 50 participants per group would provide 90% statistical power to determine non-inferiority of anti-RBD IgG GMCs between groups (assuming an equal ratio between two groups and considering a 30% dropout rate). AEs were recorded as frequencies (%) with two-sided 95% CIs. Anti-RBD IgG levels were reported as GMCs and geometric mean ratios (GMRs), both with 95% CIs, and compared with GMCs observed in healthy adults under the similar regimen from our previous study in the same setting (defined as reference cohort).²² NAb levels were illustrated as geometric mean titers (GMT) with 95% CIs. Participants with anti-RBD IgG seropositive at baseline were excluded from the analysis. Statistical differences in immunological endpoints between and across study arms were analyzed using t-tests and one-way ANOVA, respectively. Statistical differences in AE endpoints were assessed using a Chi-square test or Fisher’s exact test. The Clopper-Pearson correlation coefficient was used to evaluate correlation between maternal and fetal anti-RBD IgG levels. All statistical analyses were performed using GraphPad^™ Prism 9 (v9.2.0, 283; GraphPad^™ Software, CA, USA) and STATA (v17; StataCorp^™, LP, College Station, TX, USA) with a statistical significance of P ≤. 05. Anti-RBD IgG values of 0 were replaced with values of 0.01. ID₅₀ values <10 were replaced by 10.

Results

One hundred and fifty participants were enrolled 1:1:1 into Groups 1, 2, and 3, respectively (Figure 1). Twenty-six participants were excluded due to baseline anti-RBD IgG seropositivity, and 124 subjects included in subsequent analyses. The median (interquartile range [IQR]) age was 31 (26.0–35.5) years, gestational age was 23.5 (18.0–30.0) weeks, and BMI was 23.90 (21.2–27.6) kg/m². No significant differences in baseline characteristics were observed between groups (Table 1). Antenatal screening revealed the following underlying conditions: thalassemia trait (n = 25), non-insulin-dependent gestational diabetes (n = 5), and hepatitis B inactive carrier (n = 2). Eighty-nine participants had deliveries at the study center, 10 were delivered before 37 weeks of gestation (all were ≥35 weeks), and six weighed <2,500 grams. None experienced miscarriages.

Figure 1. Consort flow diagram. One hundred and sixty-five participants were screened and 150 determined eligible to participate in the study. Participants were randomized 1:1:1 into three study arms (50 each) and assessed pre-vaccination (baseline), four weeks after the first vaccination, two weeks after the second vaccination, and upon delivery (after excluding baseline seropositive patients).

Table 1. Participant characteristics by COVID-19 vaccine regimen.

		Vaccination type
First vaccination — Second vaccination	Total	BNT162b2 — BNT162b2	ChAdOx1 — BNT162b2	CoronaVac — BNT162b2	P-value^a
Numbers of included participants	n = 124	n = 44	n = 37	n = 43
Age [median (IQR)], years	31.00 (26.00, 35.50)	31.00 (28.00, 36.00)	33.00 (26.00, 35.00)	29.00 (25.00, 34.00)	.2431
Gestational age [median (IQR)], weeks	23.50 (18.00, 30.00)	25.00 (20.50, 31.50)	24.00 (18.00, 30.00)	20.00 (16.00, 29.00)	.7315
BMI [median (IQR)], kg/m^Citation2	23.90 (21.15, 27.55)	25.95 (22.15, 28.10)	23.60 (22.00, 27.80)	23.60 (20.90, 26.70)	.2893
Underlying disease, n (%)
Hepatitis B inactive carriers	2 (1.61)	0	2 (5.41)	0	0
Thalassemia trait	25 (20.16)	11 (25.00)	7 (18.92)	7 (16.28)	.5154
Gestational diabetes	5 (4.03)	2 (4.55)	1 (2.70)	3 (6.98)	.3012
Number of participants at delivery	n = 89	n = 36	n = 26	n = 27	…
Gestational age at delivery [median (IQR)], week	38.00 (37.00, 39.00)	38.00 (37.00, 39.00)	38.00 (37.00, 39.00)	38.00 (38.00, 39.00)	.8890
Preterm delivery (<37^b weeks of gestation), n (%)	10 (11.24)	4 (11.11)	5 (19.23)	1 (3.70)	.2016
Abortion or miscarriage, n (%)	0	0	0	0	0
Low birth weight (<2,500 g), n (%)	6 (6.74)	3 (8.33)	2 (7.69)	1 (3.70)	.7487

^aChi-squared test and Kruskal-Wallis tests were used to determine p-values. ^bAll had a gestational age between 35–36 weeks. Abbreviations: IQR, Interquartile Range; BMI, Body Mass Index.

Maternal anti-SARS-CoV-2 RBD IgG and NAb responses

Four weeks after the first vaccination, anti-RBD IgG GMC was highest among Group 1 recipients (107.14 BAU/mL, 95% CI [79.41, 144.55]), followed by Group 2 (84.85 BAU/mL, 95% CI [60.13, 119.74]), and Group 3 (11.83 BAU/mL, 95% CI [9.05, 15.45]) recipients (P < .0001, Figure 2a, Supplementary Table S1). A similar trend was observed two weeks following the second vaccination, with IgG GMC recorded as 2072.84 BAU/mL (95% CI [1691.21, 2540.58]) for Group 1, 1754.61 BAU/mL for Group 2 (95% CI [1334.10, 2307.66]), and 1609.68 BAU/mL for Group 3 (95% CI [1334.83, 1941.13]) (P = .2379, Figure 2a, Supplementary Table S1). No significant difference in anti-RBD IgG levels was observed compared to healthy non-pregnant adults in the reference study.²²

Figure 2. Maternal humoral immune responses four weeks after the first dose and two weeks after the second dose. (a) Illustrates anti-SARS-CoV-2 receptor binding domain (RBD) IgG geometric mean concentrations (GMCs), with references of the same vaccine regimens in healthy, non-pregnant adults for comparison.²² (b) Illustrates geometric mean titers (GMTs) of neutralizing antibodies (NAbs) from 50% inhibitory dilution (ID₅₀) by microneutralization assays against the ancestral Wuhan strain and Omicron BA.1, BA.2, and BA.5 subvariants. Titers below the lower limit of detection (LLOD of 1:10) were replaced with a value of 10. Error bars represent geometric means with 95% confidence intervals (CIs). Abbreviations: BAU, Binding Antibody Units; ID₅₀, 50% Inhibitory Dilution.

Two weeks after completing the second vaccination, all vaccination regimens generated high NAb GMTs against the ancestral Wuhan strain, with no statistical difference in titers observed across cohorts (P = .1284). Titers were highest in Group 3 (258.22, 95% CI [187.53, 355.56]), followed by Group 1 (187.47, 95% CI [135.15, 260.03]) and Group 2 (166.63, 95% CI [124.60, 222.84]) (Figure 2b, Supplementary Table S1). As expected, low NAb GMTs and seropositivity were observed against Omicron BA.1 (12.07, 95% CI [11.24, 12.97]; 22.13% seropositivity]), BA.2 (16.14, 95% CI [14.41, 18.07]; 48.36%), and BA.5 (10.43, 95% CI [10.09, 10.77]; 5.74%) across all three study arms (Figure 2b, Supplementary Table S1).

Maternal serum and cord blood anti-SARS-CoV-2 RBD IgG levels at delivery

The median duration between the second vaccination and date of delivery across the three groups was 9 (3–15) weeks (n = 89). There were no significant differences in anti-RBD IgG GMCs in maternal serum at delivery (P = .9356) and cord blood (P = .7987) across groups (Figure 3a, Supplementary Table S1). However, anti-RBD IgG GMC of cord blood was significantly higher than maternal serum at delivery for Group 1 (563.11 vs 362.73 BAU/mL, P < .001) and Group 3 (494.89 vs 360.73 BAU/mL, P = .03), respectively. This was not seen for Group 2 (488.33 vs 401.15 BAU/mL, P = .252). Anti-RBD IgG levels measured in maternal serum and cord blood were strongly correlated (r = 0.719, P < .001, Figure 3b). Anti-RBD IgG levels were inversely correlated with the time since the last vaccination for both maternal serum (r = −0.562, P < .001, Figure 3c) and cord blood (r = −0.366, P < .001, Figure 3d).

Figure 3. (a) Comparative geometric mean concentrations (GMCs) of maternal and cord blood anti-SARS-CoV-2 RBD IgG at delivery by vaccine regimen. (b) Pearson’s correlation coefficient (PCC) of umbilical cord anti-RBD IgG level against maternal anti-RBD IgG at delivery. (c) PCC of maternal anti-RBD IgG at time of delivery against time since the second vaccination. (d) PCC of umbilical cord blood anti-RBD IgG at the time of delivery against time since the second vaccination. (r) represents PCC.

Local and systemic AEs

All AEs were reportedly mild or moderate (Figure 4, Supplementary Figure S1) and self-resolved within 2–3 days. After the first vaccination, a significant difference in frequency and severity was observed for both local (P = .0013) and systemic reactions (P = .0036), particularly injection site reactions (P = .0013) and headaches (P < .0001) (Figure 4, Supplementary Figure S1, Supplementary Table S2) between arms. This was particularly observed after ChAdOx1 administration for both moderate local (35.14%, 22.73%, and 13.95% for ChAdOx1, BNT162b2, and CoronaVac, respectively) and moderate systemic reactions (54.05%, 11.36%, and 9.30% for ChAdOx1, BNT162b2, and CoronaVac, respectively) (Supplementary Figure S1, Supplementary Table S2). After the second dose of BNT162b2 across groups, a significant difference in frequency and severity was observed only for specific systemic reactions in the heterologous ChAdOx1-BNT162b2 arm including: fatigue (P = .0123), headaches (P = .0189), and rashes (P = .0039) (Supplementary Figure S1, Supplementary Table S3). We found no association between AEs and gestational age.

Figure 4. Self-reported local (a) and systemic (b) adverse events (AEs) within the first seven days after each vaccination by vaccine regimen.

Discussion

To our knowledge, this is the first study to assess immunogenicity and reactogenicity of various heterologous primary COVID-19 vaccination series in pregnant women. It is also one of few studies that explored ChAdOx1 administration in pregnant women. We found heterologous vaccination regimens of CoronaVac-BNT162b2 or ChAdOx1-BNT162b2 generated similar immunogenicity to homologous BNT162b2-BNT162b2, as previously noted in non-pregnant healthy adult cohorts.²² Both heterologous and homologous vaccine arms induced robust NAbs against the ancestral Wuhan strain, but limited response against Omicron subvariants. While ChAdOx1-BNT162b2 regimens induced higher reactogenicity, heterologous regimens were safe and induced no concerned pregnancy outcomes. Cord blood and maternal serum antibody levels at delivery were not different across arms.

Homologous mRNA vaccine regimens were previously reported as safe and immunogenic in pregnant women.^32,33 It is recommended by both the US and EU for pregnant women due to the high vaccine efficiency it confers in non-pregnant adult populations.²⁶ Previous studies reported that, for this same population, heterologous ChAdOx1-BNT162b2 vaccination induced similar or higher humoral and cellular immune responses than homologous ChAdOx1-ChAdOx1 or BNT162b2-BNT162b2 regimens.³⁴ Our findings extend this humoral response to pregnant women administered ChAdOx1-BNT162b2 or CoronaVac-BNT162b2 vaccine regimens. Some studies report lower anti-RBD IgG levels in pregnant populations.^9,20,22,32 This could be from the relatively immunosuppressed state of pregnant women themselves.⁹ However, we noted no difference in anti-RBD antibody responses between pregnant and non-pregnant healthy adult cohorts. Similar conclusions were drawn by Gray et al. (as cited in²⁰) and Bookstein et al.⁹

Limited NAbs generated against Omicron subvariants support the need for boosters in pregnant women. A previous report indicated higher neonatal and maternal NAb concentrations post-booster dose compared to primary series administered within the first trimester.²⁰ Healthy Thai adults who received heterologous COVID-19 vaccination also demonstrated anamnestic antibody responses subsequent to BNT162b2 boosting.²² Real-world studies confirmed the effectiveness of heterologous regimens with boosters in Thailand.^35,36 Additionally, previous studies also suggest that administering CoronaVac after ChAdOx1 or BNT162b2 priming resulted in poorer immune responses.^22,37 This highlights the importance of vaccine administration sequencing toward heterologous scheduling efficacy and need for stringent clinical trial evaluation.

Previously reported positive correlations between maternal serum, cord blood, breast milk anti-RBD IgG levels are consistent with the strong correlation reported in this study.^4,20,26,38 We found higher cord blood anti-RBD IgG levels than maternal levels at delivery, with statistical significance in Groups 1 and 3. This indicates that there is vertical transplacental transfer of SARS-CoV-2 antibodies after COVID-19 vaccination, conferring primary protection to new-born babies against infection.^39,40 The degree of infant protection depends on the placental transfer ratio, which increases with gestational age.^20,26 Cord blood and maternal antibody concentrations are influenced by the interval between delivery and vaccination.²⁰ The inverse correlation between cord blood and elapsed time from vaccination in our study further supports booster administration during the third trimester of pregnancy to optimize fetal antibody transfer and ensure sufficient neonatal protection.²⁰ The duration of this protection requires further evaluation.²⁶

COVID-19 vaccinations are safe for healthy pregnant women and are not associated with adverse pregnancy outcomes.^41–43 Similar local and systemic AEs were observed in pregnant sub-populations compared to non-pregnant populations.^3,44 Fetal complications were consistent with expected ranges described in previous literature,^{3,20,41,44,45} and no additional AEs were observed specifically for pregnant women.⁹ Badell et al.²⁰ noted AEs in pregnant cohorts to be typically mild-to-moderate, occurring on the day of vaccination and resolving a couple days thereafter. While no significant differences in AE frequency were reported across first, second, and third trimesters,⁹ a greater frequency and severity has been recorded after the second dose (i.e., uterine contractions, injection-site pain, headaches, myalgia, and fatigue).^9,20 Consistent with previous studies in non-pregnant adults in Thailand and other countries,^22,46–48 higher systemic AEs such as fever and fatigue were more commonly reported following ChAdOx1 vaccination compared with BNT162b2. Nevertheless, no serious AEs (SAEs) or concerning adverse pregnancy outcomes were observed in our study. Our findings support flexible vaccination regimens and contribute to the growing body of evidence that heterologous vaccination schedules during pregnancy are safe and immunogenic. However, there is some evidence that suggests a potential risk between ChAdOx1 administration and vaccine-induced immune thrombotic thrombocytopenia (VITT) at a very low frequency.^49,50 ChAdOx1 has since been discontinued for use in routine vaccination practices in Thailand. With this, we discourage ChAdOx1 administration in pregnant women.

This study has some limitations. First, the unblinded nature of our study risks observer bias, particulary for reactogenicity. However, immunogenicity would not be affected by an open-labeled design. Second, we excluded women with high-risk pregnancies and restricted our study to two heterologous regimens which had been found to be highly immunogenic in non-pregnant adults. These two heterologous regimens have also been used routinely in Thailand. Our results may not be generalizable to women with high-risk pregnancies or other heterologous COVID-19 regimens. Third, we did not evaluate cellular immune responses, which may be altered due to their pregnant status. Fourth, we may not have been able to study rare AEs due to the limited number of enrolled subjects. Fifth, several participants were excluded from analysis due to prior SARS-CoV-2 infection or lost to follow up during delivery. This reduced the statistical power from 90% to 85%. Our data must be interpreted with caution due to the smaller sample size. However, the number of evaluable subjects by the time of delivery is still sufficient to demonstrate differences in immunogenicity. Lastly, we did not include non-pregnant, healthy adults in this study as a control. However, our previous study in non-pregnant, healthy adults vaccinated under similar regimens and settings served as a reference for anti-RBD IgG response.²²

To conclude, our findings that heterologous primary vaccine series of CoronaVac-BNT162b2 and ChAdOx1-BNT162b2 offers immunogenicity on-par with homologous BNT162b2-BNT162b2 regimens grants flexibility to both vaccine management and utilization in countries with limited mRNA vaccine supplies. This strongly supports homologous or heterologous COVID-19 vaccination recommendations during pregnancy to confer protection to both mother and child. With the exception of ChAdOx1 vaccination, the lack of SAEs and adverse pregnancy outcomes also further alleviates concerns of heterologous regimens in pregnant women. Further multicenter trials with large cohorts are required to verify the results of this study as well as determine long-term infant immunity and safety.

Author contributions

Conceptualization, Methodology, and Supervision, Kulkanya Chokephaibulkit; Data Curation and Investigation, Chenchit Chayachinda, Kanokwaroon Watananirun, Chayawat Phatihattakorn, Sanitra Anuwutnavin, Suvimol Niyomnaitham, Wanatpreeya Phongsamart, Keswadee Lapphra, Orasri Wittawatmongkol, Supattra Rungmaitree, Kobporn Boonnak, Patimaporn Wongprompitak, Sansnee Senawong; Formal Analysis, Kulkanya Chokephaibulkit, Laddawan Jansarikit; Software and Visualization, Laddawan Jansarikit; Writing – Original Draft, Avishek Upadhya, Kulkanya Chokephaibulkit; Writing – Review & Editing, all authors.

Informed consent statement

All participants signed paper informed consent forms prior to participanting in the clinical study.

Institutional review board ststament

This study was approved by its institutional ethics committee (COA no. Si 481/2021) and conducted according to the Belmont Report, Declaration of Helsinki, and International Council on Harmonization’s Good Clinical Practice.

Acknowledgments

We would like to thank Professor Florian Krammer and Dr. Juan Manuel Carreno, from the Icahn School of Medicine, Mount Sinai, New York, USA for their contributions to the microneutralization assays.

Disclosure statement

No potential conflicts of interest were reported by the authors.

Data availability statement

Deidentified individual patient information and other datasets, study protocols, and documents generated and/or analyzed during this study will be made available after publication and shared to researchers with a methodologically sound proposal to achieve their aims upon request, by e-mail, to the corresponding author.

Supplementary data

Supplemental data for this article can be accessed on the publisher’s website at https://doi.org/10.1080/21645515.2023.2228670.

Additional information

Funding

This work was funded by the National Research Council of Thailand [grant number: N35A640318]. The funder had no role in the study’s design, collection, analysis and interpretation of data, write up of the report, or decision to submit the article for publication.