key: cord-0733293-gydzq4n3
authors: Wesselink, Amelia K; Hatch, Elizabeth E; Rothman, Kenneth J; Wang, Tanran R; Willis, Mary D; Yland, Jennifer; Crowe, Holly M; Geller, Ruth J; Willis, Sydney K; Perkins, Rebecca B; Regan, Annette K; Levinson, Jessica; Mikkelsen, Ellen M; Wise, Lauren A
title: A prospective cohort study of COVID-19 vaccination, SARS-CoV-2 infection, and fertility
date: 2022-01-20
journal: Am J Epidemiol
DOI: 10.1093/aje/kwac011
sha: 4c3a4ab67133245b85ee492bf03cb00a67e3d743
doc_id: 733293
cord_uid: gydzq4n3

Some reproductive-aged individuals remain unvaccinated against COVID-19 due to concerns about potential adverse effects on fertility. We examined the associations of COVID-19 vaccination and SARS-CoV-2 infection with fertility among couples trying to conceive spontaneously using data from an internet-based preconception cohort study. We enrolled 2,126 self-identified females residing in the U.S. or Canada during December 2020-September 2021 and followed them through November 2021. Participants completed questionnaires every 8 weeks on sociodemographics, lifestyle, medical factors, and partner information. We fit proportional probabilities regression models to estimate associations between self-reported COVID-19 vaccination and SARS-CoV-2 infection in both partners with fecundability, the per-cycle probability of conception, adjusting for potential confounders. COVID-19 vaccination was not appreciably associated with fecundability in either partner (female FR=1.08, 95% CI: 0.95, 1.23; male FR=0.95, 95% CI: 0.83, 1.10). Female SARS-CoV-2 infection was not strongly associated with fecundability (FR=1.07, 95% CI: 0.87, 1.31). Male infection was associated with a transient reduction in fecundability (FR=0.82, 95% CI: 0.47, 1.45 for infection within 60 days; FR=1.16, 95% CI: 0.92, 1.47 for infection >60 days). These findings indicate that male SARS-CoV-2 infection may be associated with a short-term decline in fertility and that COVID-19 vaccination does not impair fertility in either partner.

with 82% having received at least one dose of any vaccine. 4 Vaccination rates were lower among reproductive-aged adults, with approximately 60% of adults aged 18-39 years fully vaccinated. 4 Safety is an important factor in individual decision-making. Concern about possible side effects is a top reported reason for remaining unvaccinated 5 and, among reproductive-aged adults, there is particular concern about the potential effects of vaccination on fertility. [6] [7] [8] The hypothesis that COVID-19 vaccination may impair female fertility originated with a blog post that claimed the similarity between a SARS-CoV-2 surface glycoprotein and syncytin-1 (an envelope protein essential for formation of the placenta 9 vaccination and fertility are still limited, but do not indicate a harmful association. Although pregnant individuals were ineligible for the initial COVID-19 vaccine trials, the rate of unintended pregnancies occurring during the trials did not differ substantially between vaccinated and control groups. [14] [15] [16] In clinical trials for the AstraZeneca vaccine (ChAdOx1 nCoV- 19) , fertility rates were similar in participants who received the vaccine (n=50 pregnancies) vs. the placebo (n=43 pregnancies). 17 In three separate studies of female patients undergoing in vitro fertilization, no meaningful association was found between COVID-19

vaccination status and implantation rates, 18 stimulation characteristics, 19 embryological outcomes, 19 or ovarian follicular function. 20 Likewise, a limited number of studies have evaluated the association of COVID-19 vaccination on male fertility. Two studies in couples undergoing fertility treatments 19, 21 and one in the general population 22 found no appreciable difference in semen volume, sperm concentration, or motility measures before and after COVID-19 vaccination.

In contrast to data on COVID-19 vaccination, which do not indicate adverse associations with fertility, infection with SARS-CoV-2 has been associated with reproductive dysfunction. 23 Recent SARS-CoV-2 infection has been associated with poor sperm quality, including abnormal morphology, decreased concentration, lower motility, and increased DNA fragmentation; [24] [25] [26] [27] [28] [29] [30] [31] these findings may result from COVID-19 disease-associated fever and inflammation. 32, 33 SARS-CoV-2 infection has also been associated with impaired Leydig cell function 34 and dysregulation of the hypothalamic-pituitary-gonadal axis. 35 

cycles, decreased menstrual volume, and prolonged menstrual cycles, 36, 37 although these studies lacked an uninfected comparison group. Studies of patients undergoing fertility treatment report that SARS-CoV-2 infection is largely unrelated to treatment outcomes. 38, 39 However, in an observational study among reproductive-aged females, recent SARS-CoV-2 infection was associated with lower concentrations of anti-Müllerian hormone and higher concentrations of testosterone and prolactin. 40 Here, we examine the associations of female and male COVID-19 vaccination and SARS-CoV-2 infection with fecundability, the per-cycle probability of conception, in a North American prospective cohort study of couples trying to conceive.

Pregnancy Study Online (PRESTO) is an internet-based prospective preconception cohort study of couples residing in the U.S. and Canada. 41 

On female and male baseline questionnaires and female follow-up and early pregnancy questionnaires, we asked, "Have you ever received a COVID-19 vaccination?" and, if "yes,"

participants reported the vaccine brand ("Moderna," "Pfizer," "Johnson & Johnson," or "Other,"

with a text box to enter the brand) and dates of first and second doses. Beginning in June 2021,

we also asked female participants on all questionnaires if their partners had received a COVID-19 vaccination, as well as the dates of vaccination and vaccine brand.

On female and male baseline questionnaires and female follow-up and early pregnancy questionnaires, we asked participants if they had ever tested positive for SARS-CoV-2, and if so, the date they tested positive. On female questionnaires, we asked if their partners had ever tested positive for SARS-CoV-2, and if so, the date they tested positive. For both vaccination and infection, we prioritized male partner data from the male baseline questionnaire (available for 25% of couples); otherwise, we relied on female report of male exposures.

We collected menstrual cycle information on the baseline and follow-up questionnaires. At baseline, participants reported how long they had been trying to conceive (in menstrual cycles), their last menstrual period (LMP) date, typical menstrual cycle length, and whether their cycles were regular (i.e., can usually predict date of next period within a few days). On follow-up questionnaires, we asked for number of cycles since previous questionnaire, LMP dates for each

cycle, and length of the most recent cycle. On follow-up questionnaires, participants also reported whether they were currently pregnant, had initiated fertility treatment, or had experienced any pregnancy losses since their previous questionnaire. Those who conceived reported how the pregnancy was confirmed (e.g., urine test, blood test, ultrasound). We asked non-pregnant participants if they were still trying to conceive.

For each menstrual cycle during follow-up, we identified the first day of menses. If participants did not provide information on number and dates of cycles since the previous questionnaire, we estimated LMP date(s) that occurred between questionnaires using information on time between

reported LMP dates, length of the most recent menstrual cycle, and typical cycle length. 42 

In this analysis, we included PRESTO participants who enrolled between December 14, 2020

(when COVID-19 vaccines first became available in U.S.) and September 22, 2021 (Web Figure   1 ; n=2,679). We followed participants through November 11, 2021. We excluded 91 individuals with implausible baseline dates for LMP. We restricted to those who had been trying to conceive for ≤6 cycles at enrollment to reduce the potential for reverse causation, which could occur if fertility concerns influence decisions about vaccination. The final analytic sample included 2,126 couples. Analyses of male partner vaccination and fecundability were restricted to the 1,369 couples for whom these data were available from either partner. compared participants who had ever tested positive for SARS-CoV-2 by the first day of the menstrual cycle with those who had never tested positive. We fit proportional probabilities regression models (i.e., log-binomial models adjusting for cycle number at risk) to estimate fecundability ratios (FRs) and 95% confidence intervals (CIs). The FR represents the per-cycle probability of conception comparing exposed and unexposed individuals. We followed couples until pregnancy (regardless of outcome) or the occurrence of a censoring event (i.e., initiation of fertility treatment, cessation of pregnancy attempt, loss to follow-up, or 12 cycles of pregnancy attempt), whichever came first. To examine the association between time since vaccination or infection with fecundability, we fit restricted cubic splines.

In multivariable-adjusted models, we adjusted for the following female baseline variables: age We also fit models adjusting for confounding using fine stratification by propensity score. 45, 46 Use of propensity scores to control confounding is as effective as stratification or regression modeling, and offers the ability to improve validity by excluding individuals who are outside the mutual range of propensity scores for exposed and unexposed. 47 We fit a logistic regression model of cycle-specific vaccination status (or infection status) regressed on covariates to calculate propensity scores (i.e., predicted probabilities of exposure). The propensity score models included the following variables that are either associated with both exposure and excluded individuals who were outside the overlapping range of propensity scores for exposed and unexposed. We then divided the data set into 50 strata of propensity scores based on the distribution of propensity scores in exposed individuals and developed weighted regression models to derive an adjusted exposure association. Exposed individuals were assigned weights of 1; unexposed individuals were assigned weights as follows:

This weighting scheme generates a pseudo-population in which confounder balance is achieved within each stratum, and thus, in the population overall. We then calculated the marginal measures of association in the weighted population to estimate the average treatment effect among the treated.

In sensitivity analyses, we defined vaccination date as dose date plus 14 days to assess the association with a full immune response to the dose. We also stratified by vaccination brand, 

reverse causation, we stratified by attempt time at study entry (<3 vs. 3-6 cycles) and restricted to participants without a history of infertility. Finally, for vaccination analyses, we restricted to participants who never tested positive for SARS-CoV-2 to control for potential confounding by infection.

We used multiple imputation with fully conditional specification to impute missing data. We generated 20 imputed datasets and combined estimates across analytic datasets. Missingness was generally low: no participants were missing vaccination status or brand, and covariate missingness ranged from 0% (age) to 2% (household income).

Most female participants in our analysis had high educational attainment (83% with ≥16 years), high household income (57% with income ≥$100,000 USD/year) and private health insurance (employment-based or purchased privately; 86%). Most participants self-identified as non-Hispanic white (85%). A large proportion worked in the health care industry (25%). Around 37% had a previous live birth, and 9% reported a history of infertility.

Vaccination prevalence was similar among female and male participants. Seventy-three and 74%

had received at least one dose of COVID-19 vaccine by the LMP date of the final observed cycle, respectively. Vaccinated individuals were more likely to have higher education and income, reside in the U.S., work in the health care industry, and perform night or rotating shift work, and were less likely to be parous, report history of infertility, and have irregular menstrual

cycles than unvaccinated individuals (Table 1) . We observed few differences in participant characteristics by vaccine brand (Web Table 1 ).

COVID-19 vaccination was not appreciably associated with fecundability in either partner ( Findings were similar after adjustment for potential confounders using fine stratification on propensity scores (Table 2) . After trimming non-overlapping propensity scores and re-weighting across 50 propensity score strata, the distribution of propensity scores was similar across exposure groups (Web Figure 2) , and we achieved reasonable balance of covariates by exposure status (Web Figure 3) . the association between COVID-19 vaccination and menstruation. Two retrospective reports 62, 63 show that high proportions of menstruating adults report irregular cycles and heavy bleeds postvaccination, and that breakthrough bleeding was common among individuals taking genderaffirming hormones or long-acting reversible contraception and among post-menopausal individuals. However, these studies were likely enriched with individuals who noticed a change in their cycles and so cannot be used to estimate associations with menstruation. Results from our study indicate that even if vaccines do have short-term effects on menstruation, there is likely little or no subsequent effect on fertility.

In our study, vaccinated participants were trying to conceive between 0 and 11 months after vaccination (mean=3.5 months). Therefore, at this time, we cannot draw conclusions about longterm effects of vaccination on fertility. There are two possible sources of long-term effects of vaccination: the components of the vaccine and the immune response to vaccination.

Components of the vaccine have documented safety profiles, 1-3 and any potential allergic reactions attributable to vaccine ingredients would be observed within approximately 15-30 minutes of vaccination. 64 The innate (rapid, non-specific) phase of the immune response takes place over several days and triggers the adaptive phase (slower, highly specific), which occurs over several weeks. 65 Beyond this point, antibody concentrations plateau or slowly decline, and the risk of severe immunization-related complications drops dramatically. Enrollment in PRESTO is ongoing, and we will continue to monitor long-term associations of COVID-19 vaccination and fecundability; however, it is unlikely that adverse effects on fertility could arise many months after vaccination.

with several studies indicating short-term declines in sperm quality after SARS-CoV-2 infection. [24] [25] [26] [27] [28] [29] [30] Fever is a known determinant of impaired spermatogenesis, and effects on sperm concentration, motility, and morphology can persist for 3-4 months (i.e., the duration of spermatogenesis). 33 Fever is one of the most common symptoms of SARS-CoV-2 infection; 32 therefore, fever could explain our finding of an acute decline in fertility among men with recent SARS-CoV-2 infection. Although fever is also a side effect of vaccination, it is much less common than for infection. 14-16 The fertility decline could also be related to immune response and inflammation in the testes and epididymis, which have been observed in hospitalized COVID-19 patients. 27 Erectile dysfunction is also more common among males following SARS-CoV-2 infection. 66 Due to a lack of data on COVID-19 symptoms or disease severity, we could not assess this hypothesis. Regardless, we did not observe any association between SARS-CoV-2 infection and fecundability that persisted beyond 60 days.

We adjusted for a broad range of sociodemographic, lifestyle, medical, occupational, and reproductive factors that could confound the association between COVID-19 vaccination or SARS-CoV-2 infection and fecundability. We adjusted for confounding using traditional regression modeling as well as propensity score stratification. As in any non-experimental study, uncontrolled confounding is possible.

Loss to follow-up was low in our cohort (82% completed at least one questionnaire, and of those, only 3% were subsequently lost to follow-up) and was similar by vaccination status. Therefore, it is unlikely that differential loss to follow-up was an important source of bias.

We relied on self-report to assess COVID-19 vaccination status, which may have resulted in some misclassification. In addition, for couples in which the male partner did not complete his questionnaire, we relied on female report of male vaccination status. We expect that any misclassification was infrequent and non-differential with respect to fecundability. Validation studies of influenza vaccination in the past year found 97% agreement between vaccination status based on self-report and medical records. 67 Because length of the recall interval was shorter for COVID-19 vaccination in this study and recipients received vaccination cards, we anticipate little exposure misclassification.

We assessed history of SARS-CoV-2 infection by asking participants if they had ever tested positive for SARS-CoV-2. We also relied on female report of male infection for nearly 75% of couples. Underestimation of the true incidence of SARS-CoV-2 infection is probable because most participants were likely not testing regularly throughout the follow-up period. Given the high specificity of antigen and PCR tests for SARS-CoV-2, 68 we anticipate that our exposure definition had very high specificity but potentially low sensitivity. If misclassification of SARS-CoV-2 infection was unrelated to fecundability, there should be minimal to no bias in relative measures of association. 69 We calculated fecundability using self-reported information on LMP dates, typical menstrual cycle length, and pregnancy status. We also estimated LMP dates that occurred between followup questionnaires. To the extent that any of these variables were ascertained with error, outcome misclassification may have occurred. In previous work from this cohort, LMP dates

prospectively-reported on a menstrual charting app and retrospectively-reported on follow-up questionnaires were within 1 day for 93% of participants. 41 Because we did not have daily urinary measures of human chorionic gonadotropin, we likely missed some conceptions ending in early loss. However, 96% of the cohort used home pregnancy tests, and the median weeks' gestation at pregnancy detection was 4.0 (interquartile range: 3.7-4.4), indicating that participants are testing early for pregnancy.

Several features of PRESTO make it an ideal setting in which to assess the relation of COVID-

conceive without use of fertility treatment is challenging, given that individuals often do not publicize their intentions or interact with health care providers. Our study has successfully recruited couples during preconception using advertising on social media, with internet-based data collection and follow-up. 41 Our internet-based methods allowed us to continue enrolling couples throughout the COVID-19 pandemic, as participation required no face-to-face interaction with study staff. We prospectively followed couples every two months and collected time-varying data on COVID-19 vaccination and SARS-CoV-2 infection. Finally, our cohort is more geographically and socioeconomically diverse than most other preconception cohorts 70 and represents the largest study on these associations to date.

Our study was limited to pregnancy planners enrolled through the internet. Although both pregnancy planning status and internet access are related to sociodemographic characteristics such as income and education, we do not expect our associations to vary by these characteristics. a Fecundability is the per-cycle probability of conception. FRs >1 indicate an exposure associated with improved fecundability (or shorter time-to-pregnancy), whereas FRs <1 indicate an exposure associated with reduced fecundability (or longer time-to-pregnancy). b Adjusted for female age, educational attainment, household income, current smoker, private health insurance, hours/week of work, rotating shift work, night shift work, body mass index, intercourse frequency, doing something to improve chances of conception, sleep duration, Perceived Stress Scale score, Major Depression Inventory score, Pap smear in past three years, history of infertility, parity, irregular menstrual cycles, menstrual cycle length, geographic region of residence, last method of contraception, occupation in health care industry, race/ethnicity, and days since 12/14/2020. Analysis of COVID-19 vaccination status was adjusted for ever having tested positive for SARS-CoV-2 and analysis of SARS-CoV-2 infection was adjusted for COVID-19 vaccination. c Propensity scores were developed to predict the odds of vaccination (see Appendix). We adjusted for propensity score using fine stratification weighting and calculated the Mantel-Haenszel summary FR d Individuals included in "second dose" are also included in "first dose." Those who received the Johnson & Johnson vaccine are included in the sample of "first dose" and "second dose". 

Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine

Efficacy and Safety of the mRNA-1273 SARS-CoV-2 Vaccine

Safety and Efficacy of Single-Dose Ad26.COV2.S Vaccine against Covid-19

Demographic Characteristics of People Receiving COVID-19 Vaccinations in the United States

An Uncertain Public -Encouraging Acceptance of Covid-19 Vaccines

COVID-19 vaccine hesitancy linked to increased internet search queries for side effects on fertility potential in the initial rollout phase following Emergency Use Authorization

United States internet searches for "infertility" following COVID-19 vaccine misinformation

American Society for Reproductive Medicine. A survey of fertility patients' attitudes towards the COVID-19 vaccine

Functional characterization of the human placental fusogenic membrane protein syncytin 2

Addressing anti-syncytin antibody levels , and fertility and breastfeeding concerns, following BNT162B2 mRNA vaccination

No evidence of fetal defects or antisyncytin-1 antibody induction following COVID-19 mRNA vaccination

Food & Drug Administration; 2020. (Vaccines and Related Biological Products Advisory Committee Briefing Document

Drug Administration; 2020. (Vaccines and Related Biological Products Advisory Committee Briefing Document)

Fertility rates and birth outcomes after ChAdOx1 nCoV-19 (AZD1222) vaccination

SARS-CoV-2 spike protein seropositivity from vaccination or infection does not cause sterility

Ovarian follicular function is not altered by SARS-CoV-2 infection or BNT162b2 mRNA COVID-19 vaccination

BNT162b2 mRNA Covid-19 vaccine does not impair sperm parameters

Sperm Parameters Before and After COVID-19 mRNA Vaccination

Effects of COVID-19 and mRNA vaccines on human fertility

Semen impairment and occurrence of SARS-CoV-2 virus in semen after recovery from COVID-19

Impaired semen parameters in patients with confirmed SARS-CoV-2 infection: A prospective cohort study

Impaired spermatogenesis in COVID-19 patients

Evaluation of sex-related hormones and semen characteristics in reproductive-aged male COVID-19 patients

Investigation of SARS-CoV-2 in semen samples and the effects of COVID-19 on male sexual health by using semen analysis and serum male hormone profile: A cross-sectional, pilot study

Sperm quality and absence of SARS-CoV-2 RNA in semen after COVID-19 infection: a prospective, observational study and validation of the SpermCOVID test

Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China

History of febrile illness and variation in semen quality

Effect of SARS-CoV-2 infection upon male gonadal function: A single center-based study

Atrophy and Hypothalamic Pathology in COVID-19: Possibility of the Incidence of Male Infertility and HPG Axis Abnormalities

SARS-CoV-2 infection and subsequent changes in the menstrual cycle among participants in the Arizona CoVHORT study

Analysis of sex hormones and menstruation in COVID-19 women of child-bearing age

Investigating the impact of asymptomatic or mild SARS-CoV-2 infection on female fertility and in vitro fertilization outcomes: A retrospective cohort study

Disease in Reproductive-Aged Women in Wuhan, China: An Observational Study

Design and Conduct of an Internet-Based Preconception Cohort Study in North America: Pregnancy Study Online

Seasonal patterns in fecundability in North America and Denmark: a preconception cohort study

A global measure of perceived stress

Quality of life instruments in depression

A Propensityscore-based Fine Stratification Approach for Confounding Adjustment When Exposure Is Infrequent

Alternative approaches for confounding adjustment in observational studies using weighting based on the propensity score: a primer for practitioners

A review of the application of propensity score methods yielded increasing use, advantages in specific settings, but not substantially different estimates compared with conventional multivariable methods

Lack of effects on female fertility and prenatal and postnatal offspring development in rats with BNT162b2, a mRNA-based COVID-19 vaccine

Vaccination during Pregnancy and First-Trimester Miscarriage

Receipt of mRNA Covid-19 Vaccines and Risk of Spontaneous Abortion

Preliminary Findings of mRNA Covid-19

Vaccine Safety in Pregnant Persons

Maternal and Neonatal Morbidity and Mortality Among Pregnant Women With and Without COVID-19 Infection: The INTERCOVID Multinational Cohort Study

Pregnancy as a risk factor for severe coronavirus disease 2019 using standardized clinical criteria

Disease severity, pregnancy outcomes, and maternal deaths among pregnant patients with severe acute respiratory syndrome coronavirus 2 infection in Washington State

Clinical manifestations, risk factors, and maternal and perinatal outcomes of coronavirus disease 2019 in pregnancy: living systematic review and meta-analysis

Gonadal dysfunction in systemic diseases

59. Lamb AR. Experiences with Prophylactic Typhoid Vaccination: Its Effect on Menstruation

Menstrual abnormalities after hepatitis B vaccine

No association between HPV vaccine and reported post-vaccination symptoms in Japanese young women: Results of the Nagoya study

Effect of COVID-19 vaccination on menstrual periods in a retrospectively recruited cohort

Characterizing menstrual bleeding changes occurring after SARS-CoV-2 vaccination

Janeway's Immunobiology, Ninth edition

Validation of self-reported influenza vaccination in the current and prior season

Analytical Sensitivity and Specificity of Four Point of Care Rapid Antigen Diagnostic Tests for SARS-CoV-2 Using Real-Time Quantitative PCR

The Preconception Period analysis of Risks and Exposures Influencing health and Development (PrePARED) consortium