The C-VIPER is an international, non-interventional, post-marketing, cohort study designed to collect prospective safety and birth outcome data among people vaccinated with a COVID-19 vaccine during pregnancy or within 30 days prior to the first day of their last menstrual period (LMP). As of November 2022, ethics committee approval to conduct the C-VIPER has been obtained in Australia, Austria, Colombia, Cyprus, Germany, Liberia, Malta, Mexico, Philippines, Portugal, USA, and Zambia. In addition, confirmation that approval is not required has been obtained in Bulgaria, Canada, Croatia, Denmark, Greece, Hong Kong, Italy, Malaysia, New Zealand, Nigeria, Poland, Romania, Singapore, Slovakia, Slovenia, South Africa, South Korea, and the UK. As vaccines become available to pregnant people in additional countries, the study may expand. Countries contribute information when specific vaccines are authorized in-country and have been recommended for use in this population.

Recruitment into the study takes place via the dedicated website (https://c-viper.pregistry.com), links to which are included on other relevant websites worldwide. Awareness of the study is increased by advertising on social media channels frequented by pregnant people and by disseminating C-VIPER data to a wider scientific audience, including attendees at relevant meetings and the readership of peer-reviewed medical journals. Particular efforts are made to enroll participants from minority communities from social media and parenting platforms with high minority engagement, as well as through targeting geographical areas with high minority population densities. The study is sponsored by Pregistry and administered online via a web-based data collection tool. Participants can withdraw from the C-VIPER at any time. Although participants receive automatic reminders from the web-based system and from C-VIPER staff to complete modules, losses to follow-up are anticipated to occur. These losses will be censored from the cohort at the last time of contact. An attempt will be made to collect a minimum set of de-identified demographic information from all participants who are lost to follow-up in order to assess data representativeness in final analyses relative to the initial population. It is also anticipated that some participants in the unexposed group will subsequently receive a vaccination dose, at which point they will be censored from the study.

Participant retention is facilitated with short easy-to-complete questionnaire modules. Participants can also sign up for optional e-mails, texts, and phone calls during the study. All study materials are available in the languages spoken in the countries covered by the C-VIPER.

The target study population consists of pregnant people who are at least 18 years of age. The exposed group includes pregnant people who have received at least one dose of a COVID-19 vaccine during the time period from 30 days prior to the first day of the LMP until the end of the pregnancy. The study is intended to include both individuals receiving a primary vaccination course as well as those receiving booster vaccinations. Participants are asked to upload documentary proof of vaccination, and reminders are sent to those who fail to provide it.

The exposed group will be compared to a second cohort of pregnant people who are participants in the Pregistry International Pregnancy Exposure Registry (PIPER, ClinicalTrials.gov NCT05352256, EUPAS46841). These subjects have not received a COVID-19 vaccine during pregnancy (although they may have received it more than 30 days prior to the first day of the LMP). The PIPER, which is also international in scope, follows the same methods of recruitment and data collection as the C-VIPER.

Study subjects self-enroll using the C-VIPER website and can enroll at any time during their pregnancy. Participants who enroll before an outcome of interest is known are considered prospective enrollees for that outcome. Conversely, participants who enroll after an outcome is recognized or suspected are considered retrospective enrollees for that outcome, with the exception of pregnancy loss before 20 weeks gestation.

Enrollment occurs during the first 2 years of the study, with follow-up of pregnancies and infants continuing in years 3 and 4. Data analysis and preparation of publications will take place during the final year. The C-VIPER may be discontinued if other methods of gathering appropriate information become achievable or are deemed preferable, or if the feasibility of collecting sufficient information diminishes to unacceptable levels because of poor enrollment or a lack of funding.

The C-VIPER aims to recruit a minimum of 500 pregnancies for each COVID-19 vaccine brand. As vaccine exposure in early pregnancy is of particular interest, the aim is to enroll at least 150 subjects vaccinated during the first trimester of gestation. The exposed to unexposed ratio is 1:2. Therefore, for each vaccine brand, the unexposed group will consist of approximately 1000 pregnancies. These sample sizes give sufficient power to detect relative risks of 3.0 or higher for outcomes originating in early pregnancy with frequencies of around 3%. Including a larger proportion of participants exposed later in pregnancy allows smaller effects to be detectable for outcomes originating in late pregnancy, assuming a loss to follow-up of around 50%. For a common outcome such as spontaneous abortion (i.e., the spontaneous loss of a pregnancy prior to 20 weeks of gestation), which occurs in around 15–20% of pregnancies [19], a two-fold increase in risk could be detected, based on a minimum of 150 exposed pregnancies.

Details of the timing of data collection are presented in Fig. 1. Data are collected mainly through a series of structured web-based questionnaires (available from the corresponding author upon request). Every effort is made to ensure participant confidentiality, and personal identifying data are available to few C-VIPER personnel on a need basis. These data are kept securely in a separate file from the exposure and outcome data and are linked to each participant by a six-digit alphanumeric code. At enrollment, all participants must complete an initial COVID-19 vaccination module. Updates are then collected monthly until pregnancy loss or delivery, including exposure to additional vaccine doses. At this point, the delivery module records the pregnancy outcome. In the case of a live birth, data are collected at 3-monthly intervals during the first 12 months of the infant’s life, or until the infant’s death, should this occur during the first 12 months after delivery.

Timing of modules for collection of key variables. B baseline module, C-VIPER COVID-19 Vaccines International Pregnancy Registry, COVID-19 coronavirus disease 2019, D delivery module, F follow-up modules during pregnancy, LMP last menstrual period, PIPER Pregistry International Pregnancy Exposure Registry, V Vaccination module; 3, 6, 9, 12: maternal and infant outcomes modules at 3, 6, 9, and 12 months of age

In order to verify the self-reported data collected by the study and reduce potential exposure misclassification, participants are encouraged to upload copies of relevant supporting documents. A photograph of the vaccination certificate is requested, as well as medical records for the mother and the infant, from which all identifying information has been redacted. These records can include the results of SARS-CoV-2 tests, vaccination certificates, reports from healthcare professionals, photographic evidence of medication use, the delivery hospital discharge report, pediatric reports if applicable, and any other healthcare records the participant considers relevant. In case of a report of a serious AE (i.e., death, a life-threatening AE, inpatient hospitalization or prolongation of existing hospitalization, a persistent or significant incapacity or substantial disruption of the ability to conduct normal life functions, or a congenital anomaly/birth defect), the study participant may receive a Medical Records Release Authorization Form and, if signed, the relevant healthcare provider is contacted to request medical records. These records will be used to validate maternally reported diagnoses. They also allow for adjudication of outcomes by the study dysmorphologist. In the event of discordance between the medical and maternal reports of COVID-19 vaccine administration, the subject will be classified as exposed to vaccination.

The semi-structured questionnaires have been thoroughly tested in Pregistry’s previous International Registry of Coronavirus Exposure in Pregnancy (IRCEP) study (ClinicalTrials.gov NCT04366986, EUPAS37360), which used a similar data collection system [20]. Following a pilot study with non-study volunteers, and comments from IRCEP participants, improvements were made, particularly to the wording of questions or to clarify instructions before the questionnaires were used in the C-VIPER and PIPER studies. The answers at baseline and the questions in the subsequent modules are interconnected to provide a personalized experience.

Details of the exposures measured by the C-VIPER and PIPER studies are provided in Table 1. The main exposure of interest in the C-VIPER is COVID-19 vaccination. Participants are asked to provide the number of vaccinations received, name of the vaccine or manufacturer, and date of vaccinations(s). Co-administration of another vaccine and AEs experienced during the 48 h following vaccination are also recorded. The specific gestational timing of vaccination is also of interest, as the etiologically relevant period for each outcome measure will vary.

The C-VIPER and the PIPER collect data on a wide range of outcomes, full details of which are given in Table 2. The selection of obstetric and neonatal outcomes of interest and their definitions are based on guidelines published by the Global Alignment of Immunization safety Assessment in pregnancy (GAIA) network [21, 22]. Information on maternal and infant medical conditions, new COVID-19 vaccination doses, use of medications, environmental exposures, and results of SARS-CoV-2 tests are collected during pregnancy and until 12 months after delivery for all live births. In addition, infant motor, cognitive, language, social-emotional, and mental health milestones are assessed at 6, 9, and 12 months of age using the Caregiver-Reported Early Development Instrument [23]. This instrument has been adapted to allow parents to report their child’s developmental milestones using a brief set of yes/no questions. It was designed to be used within large-sample studies and contains items that are not affected by culturally specific contexts. It has been validated in more than 25 countries [24, 25] and is available in the languages used by the C-VIPER study. The Caregiver-Reported Early Development Instrument Short Form, which takes approximately 5 min to complete, is used in the 6- and 9-month modules, while the Long Form, which takes 15–20 min to complete, is used in the 12-month module.

The baseline modules collect maternal demographic information, data on reproductive history, health-related behaviors, and pre-pregnancy health. Information collected by the pregnancy and birth outcome modules includes, among other variables, the number of fetuses present, health status throughout pregnancy, concomitant medications, pregnancy outcome (livebirth, miscarriage, stillbirth, elective termination), gestational length, and mode of delivery. Variables included in the delivery module concern infant demographics and health complications, feeding (breastmilk only; mostly breastmilk, some formula; mostly formula, some breastmilk; bank/donated breastmilk only; some bank/donated breastmilk, some formula; formula only; neither breastmilk nor formula), and vaccination history. Infants’ weight, length, and medical history are recorded at 3, 6, 9, and 12 months of age, and mothers are requested to provide redacted medical records for their infant, if available (medical records will also be solicited from healthcare professionals after maternal consent is obtained). This length of follow-up will allow the capture of structural and functional malformations that do not become clinically apparent until several months after birth as well as the ascertainment of developmental milestones.

Unexposed controls will be matched 2:1 to the vaccinated on country and gestational age at enrollment (time zero of follow-up). As the C-VIPER and PIPER studies may not cover the exact same time period, matching by calendar time of enrollment is not always possible. The frequency of outcomes of interest will be estimated for exposed and unexposed participants, using absolute risks, risk differences and unadjusted relative risks with 95% confidence intervals. Kaplan–Meier survival curves will present cumulative risk for outcomes that are dependent on gestational age. Adjusted analyses will be conducted using propensity score stratification, and relative risks will be calculated from generalized linear models when appropriate.

For each outcome, a directed acyclic graph, informed by expert knowledge and available literature, will be used to identify relevant confounders, competing events, and censoring. The baseline characteristics of the two study groups will be compared using the standardized mean difference, with an absolute difference >0.2 being considered an indicator of substantial imbalance. Unbalanced risk factors will be controlled through the use of propensity score; only characteristics collected at baseline will be considered as potential confounders. The baseline module collects information on characteristics that existed before vaccination, thus avoiding adjustment for factors in the causal pathway [26].

The optimal group for analysis for each specific outcome is participants enrolled before the outcome is known. Early enrollment reduces potential selection bias and allows the collection of reliable prospective information on vaccine exposure and other characteristics unaffected by potential outcomes. Analyses of malformations will be restricted to participants who enroll before the results from informative tests, such as ultrasound screening after 12 weeks of gestation, are known. Analysis of early pregnancy losses will be restricted to participants who enrolled before 20 weeks of gestation, while those for preterm delivery will be restricted to participants enrolled before 37 weeks of gestation.

In most cases, enrollment in the study and information on covariables will be collected before the outcome occurs, to avoid selection and recording bias [27, 28]. However, as retrospective enrollment is permitted, enrollment and recording of information on exposure and baseline covariates may occur after some outcomes have been identified. To assess the impact of retrospective enrollment, the primary analysis for each outcome will be restricted to prospectively enrolled participants.

Some participants will receive more than one dose of vaccine during the pregnancy being studied. Secondary analyses will explore the effect of two or more doses. The effects of exposure to the first dose, compared to subsequent doses, will also be examined. The effects of homologous or heterologous vaccination will be examined in analyses on an intention-to-treat, per protocol, or combined approach. In the intention-to-treat approach, exposure is defined by the first vaccine brand used, regardless of the brand for any subsequent doses. In the per-protocol approach, participants will be censored from the study when they receive a second vaccination dose with a brand that is different from the first dose received. In the combined approach, a participant is considered to be exposed if at least one dose of any approved or authorized vaccine is received during pregnancy. Therefore, participants receiving both homologous and heterologous vaccine regimens will be included in analyses.

Participants who enroll into the unexposed group and subsequently receive a COVID-19 vaccination during pregnancy will be censored from the study on the vaccination date. An intention-to-treat approach that would keep such participants in the unexposed group is inappropriate for this study, which focusses on vaccine safety, as it may bias results towards the null. Weights will be used to adjust for potential informative censoring.

The C-VIPER is overseen by a scientific advisory committee, made up of experts in maternal-fetal medicine, infectious diseases, epidemiology, and biostatistics. Members of the scientific advisory committee receive interim reports from the C-VIPER coordinator. The responsibilities of the scientific advisory committee include decisions regarding emerging safety issues, resolution of operational issues, and development of strategies to increase awareness of the study.

Several marketing authorization holder companies have contracted with Pregistry to fund analyses of the C-VIPER in order to fulfill their post-authorization safety study requirements. These companies have access to a dedicated online portal that allows their staff to see all available de-identified case data for pregnancies exposed to their COVID-19 vaccine in near real time. Serious AEs occurring during the study period in exposed participants or their offspring are reported expeditely to the relevant marketing authorization holder company (see “Funding”) within agreed timelines in accordance to country-specific regulations via an electronic gateway. Non-serious AEs are reported in a non-expedited way. Safety analyses of data occurring in participants vaccinated with products of marketing authorization holder companies not financially contributing to the C-VIPER will be published in the medical literature.