The effect of human papillomavirus vaccination during adolescence or early adulthood on incidence of spontaneous abortion in adulthood

Abstract

While human papillomavirus (HPV) infection has been associated with spontaneous abortion (SAB), there are limited data on HPV vaccination and SAB risk. We analyzed data from Pregnancy Study Online (PRESTO), a North American preconception cohort study (2013–2025). We followed 11,770 females aged 21–45 years for 12 months or until pregnancy, whichever came first. On baseline questionnaires, participants reported their HPV vaccination status and age at HPV vaccination. Participants reported pregnancy outcomes on follow-up questionnaires. We used Cox regression to estimate hazard ratios (HRs) and 95 % confidence intervals (CIs) adjusted for potential confounders. The prevalence of HPV vaccination was 52.9 %, and 20.4 % of participants experienced SAB. There was no appreciable association between HPV vaccination and SAB overall (HR = 0.95, 95 % CI: 0.87–1.04) or by age at vaccination (<18 years: HR = 1.03, 95 % CI: 0.92–1.16; ≥18 years: HR = 0.90, 95 % CI: 0.81–1.00). Results did not vary appreciably by timing of SAB (<6 vs. ≥6 weeks).

1. Background

Infection with human papillomavirus (HPV) is an established cause of cervical cancer [1]. Though the prevalence of HPV infection in the United States (U.S.) has declined rapidly in recent years [1], HPV is still the most common sexually transmitted infection (STI), with the prevalence of disease-associated HPV types (6, 11, 16, and 18) being 19.9 % among females aged 15–59 years in 2018 [2].

HPV vaccines (Gardasil, Cervarix) became available in the U.S. and Canada in 2006 [4], and have shown high efficacy in immunizing against HPV strains linked to genital warts, cervical intraepithelial neoplasia, and selected gynecologic cancers (6, 11, 16, and 18). The Centers for Disease Control and Prevention (CDC) recommends two doses of the HPV vaccine for those aged 11–12 years [3] and, in 2019, the Advisory Committee on Immunization Practices (ACIP) recommended “catch-up” HPV vaccination for all unvaccinated individuals through age 26 years [4]. The most recent World Health Organization (WHO) guidelines now recommend a one- or two-dose schedule for females up to age 20 years, and two doses six months apart for females ≥21 years [5]. As of 2023, 57.3 % of U.S. adolescents aged 13–15 years had received at least 2 doses of the HPV vaccine [6]. In the ten years following the roll-out of the quadrivalent HPV vaccine, the prevalence of HPV infection types 6, 11, 16, and 18 declined 86 % among females age 14–19 years and 71 % among females age 20–24 years in the U.S [1]. Moreover, there is accumulating evidence that HPV vaccination is effective in reducing rates of cervical precancer [7] and cancer [8] among reproductive-aged U.S. females.

Spontaneous abortion (SAB), defined as the loss of a fetus before 20 weeks of gestation, occurs in ~20 % of recognized U.S. pregnancies [9,10]. Female HPV infection has been associated with SAB in some [11–13] but not all [14] studies. Compelling evidence supporting HPV as a cause of SAB comes from a 1997 case-control study showing a higher prevalence of HPV-16 comparing first-trimester spontaneous and electively-aborted products of conception [13]. If active HPV infection is a putative cause of SAB, HPV vaccination would be expected to protect against SAB by reducing the likelihood of active HPV infection during early pregnancy. To our knowledge, however, there have been no studies of HPV vaccination at the recommended ages (i.e., 11–12 years) and subsequent risk of SAB in adult females. However, there have been several studies of periconceptional or gestational HPV vaccination and SAB risk [15,16], including a 2024 systematic review and meta-analysis of 11 studies [16]. While quadrivalent-HPV vaccination during the periconceptional or gestational period showed little association with SAB, bivalent-HPV vaccination during the period from 45 days before last menstrual period (LMP) to LMP was associated with higher SAB risk (RR = 1.59, 95 % CI: 1.04–2.45) as was nonavalent-HPV vaccination during the period from 90 days before LMP to 45 days after LMP (RR = 2.04, 95 % CI: 1.28–3.24) [16]. The vast majority of these studies included females vaccinated at ages greater than 21 years.

We hypothesize that HPV vaccination earlier in life (<18 years) would be associated with reduced SAB risk via a reduction in the likelihood of active HPV infection during early pregnancy. Studies that enroll participants during preconception and follow them prospectively are better able to capture early pregnancy losses and evaluate differences in association by timing of SAB, which are important if etiologic factors differ by gestational week at risk [17,18]. In a North American preconception cohort study, we examined the association between HPV vaccination during adolescence or early adulthood and SAB incidence in adulthood among non-contracepting couples. We stratified the data the calendar year at birth, history of STIs, and history of abnormal Pap test. We also evaluated differences in effect according to timing of SAB (<6 vs. ≥6 weeks of gestation).

2. Methods

2.1. Study population

Pregnancy Study Online (PRESTO) is an ongoing preconception cohort study of pregnancy planners in the U.S. and Canada [16]. Participants were enrolled between June 2013 and May 2025 and were recruited primarily through advertisements on social media (e.g., Facebook and Instagram) and health-related websites. Eligible primary participants (hereafter referred to as “female”) were aged 21–45 years, assigned female at birth, attempting to conceive without the use of fertility treatments, and in a relationship with one partner. Female participants completed a baseline questionnaire and follow-up questionnaires every 2 months for 12 months or until pregnancy, whichever came first. Participants who conceived completed questionnaires during early (~8 weeks) and late (~32 weeks) pregnancy and at 6 months postpartum. Boston University Medical Campus institutional review board approved the study protocol, and all participants provided informed consent.

2.2. Assessment of HPV vaccination and covariates

At baseline, female participants provided comprehensive data on socio-demographic factors (age, education, income, marital status, occupation); geographic region of residence; contraceptive, reproductive, and medical history, medication use; anthropometrics (height and weight); behavioral factors (cigarette smoking, alcohol consumption, physical activity); lifetime number of sexual partners; and male partner characteristics (age, education, height, weight, smoking). Participants were asked “has a doctor ever told that you had an abnormal Pap smear?” and, if yes, how many abnormal Pap tests and their age at first abnormal Pap test. Participants also reported whether they had been vaccinated against HPV, how many shots they received, and at what age. Among the vaccinated, 83 % reported three doses, 11.9 % two doses, and 4.9 % one dose. Given recent data demonstrating efficacy with a single dose [5], we defined participants as “vaccinated” if they reported receiving at least one HPV vaccine prior to enrollment.

2.3. Assessment of spontaneous abortion (SAB)

On follow-up questionnaires, female participants reported the date of their last menstrual period (LMP), whether they were currently pregnant, and whether they had experienced a pregnancy loss (miscarriage, blighted ovum, induced abortion, or ectopic pregnancy) since their last questionnaire. They also reported when and how often they used home pregnancy tests and the results of each test. Those with an index pregnancy reported their LMP date, due date, date of their first positive pregnancy test, and how many gestational weeks they were at the time of questionnaire completion. SABs occurring after the early pregnancy questionnaire were identified on the late pregnancy or postpartum questionnaires. Among participants lost to follow-up, we collected additional pregnancy outcome data by contacting participants via email or phone, and/or by linking with birth registries in selected states (<2 %) [19]. Among participants with SAB, we ascertained how many weeks their pregnancy lasted and on what date their pregnancy ended. We prioritized participants’ self-reported gestational weeks at loss. Among those with missing gestational week at loss, we estimated gestational age as: (pregnancy end date – [pregnancy due date – 280 days])/7. If data on gestational week at loss and due date were unavailable, we estimated gestational age as: (pregnancy end date – LMP date)/7.

2.4. Exclusions

Of the 21,804 eligible female participants with the opportunity to complete follow-up (June 2013–March 2025), we excluded 51 participants who started their baseline questionnaire >60 days after their eligibility screener or finished their baseline questionnaire >60 days after starting it; 4214 participants who did not conceive during the study period; 5762 participants with no follow-up data; and 7 participants with implausible data, leaving 11,770 participants followed for SAB in the analytic sample.

2.5. Data analysis

We evaluated the association of HPV vaccination (ever vs. never) and age at vaccination (<18 and ≥ 18 years) with incidence rates of SAB. We used Cox proportional regression models to estimate hazard ratios (HRs) and 95 % confidence intervals (CIs). The HR represents the instantaneous rate of SAB among the vaccinated divided by that among the unvaccinated participants. We selected confounders based on the literature and the drawing of a causal diagram. We controlled for calendar year at enrollment (continuous), age at conception (<25, 25–29, 30–34, ≥35 years), race and ethnicity (Hispanic, non-Hispanic White, Non-Hispanic Black, Non-Hispanic Asian, mixed race or some other race), smoking (current, past, never), education (≤12, 13–15, 16, ≥17 years), parental education (≤12, 13–15, ≥16 years), annual household income (<100,000, 100,000–149,000, ≥150,000 USD), region of residence (U.S. Midwest, Northeast, South, West; Canada), and history of STIs. We avoided controlling for history of SAB or infertility, or lifetime number of sexual partners, because these are unlikely causes of HPV vaccination and their adjustment might represent a form of overcontrol.

We stratified analyses by calendar year at enrollment, history of STIs or pelvic inflammatory disease (PID), and history of abnormal Pap smear before HPV vaccination. We also stratified on timing of SAB (<6 vs. ≥6 weeks), consistent with earlier studies [15]. Finally, we repeated our analyses after restricting the unvaccinated group to those with ≤5 years since their first abnormal Pap test, a subgroup with a higher likelihood of active HPV infection. Data on HPV vaccination (ever vs. never) and age at vaccination were missing for 0.2 % and 0.7 % of participants, respectively. Data on gestational week at loss were missing for 102 SABs (4.2 %). Missing covariate data ranged from 0.0 % (age, race, and region) to 6.3 % (age at abnormal Pap test). We used fully conditional specification methods to multiply impute missing exposure, covariate, and outcome data [20], combining beta coefficients and standard errors across 20 imputed datasets [18].

2.6. Quantitative Bias analysis (QBA)

In the absence of a national immunization registry that collects HPV vaccine information, there is no single reliable source of HPV vaccine data [21]. Studies have evaluated the accuracy of self-reported HPV vaccination among U.S. adolescent girls [22] and reproductive-aged females from other countries [23,24]. A study most comparable to ours examined U.S. females aged 18–26 and found 91 % sensitivity and 76 % specificity for self-reported HPV vaccination versus electronic medical records [25]. We conducted a QBA using inputs from the latter study [25], assuming non-differential misclassification of HPV vaccination [26].

3. Results

A total of 2407 (20.4 %) participants experienced SAB, with 1032 (42.9 %) of SABs occurring before 6 weeks’ gestation. Table 1 shows baseline characteristics of the analytic sample. The prevalence of HPV vaccination was 52.9 % (25.5 % at ages <18 years, 27.4 % at ages ≥18 years). Median age at vaccination was 18 years (interquartile range [IQR]: 15–22 years). Median time since vaccination was 11 years (IQR: 8–14 years). HPV vaccination was positively associated with calendar year at enrollment, income, education (personal and parental), and residence in the Northeastern U.S., and inversely associated with age, smoking, parity, irregular menses, and history of SAB or infertility. Participants vaccinated before age 18 were appreciably less likely to have STIs (especially genital warts) or an abnormal Pap test. A history of abnormal Pap test among unvaccinated participants was 28.8 %. Participants who were and were not lost-to-follow-up differed with respect to current cigarette smoking (8 % vs. 5 %), educational attainment (≥16 years: 76 % vs. 83 %), and race/ethnicity (non-Hispanic White: 80.8 % vs. 85.1 %), but were similar in terms of HPV vaccination status (50.3 % vs. 53.1 %), age (30.0 vs. 30.5 years), gestational weeks at first positive pregnancy test (4.1 vs. 4.1), parity status (31.4 % vs. 32.9 %), BMI (27.5 vs. 26.7 kg/m²), and history of STIs (11.8 % vs. 12.9 %).

Table 1.

Baseline characteristics of 11,770 female participants by HPV vaccination and age at first vaccination, PRESTO, 2013–2025.

Characteristica	Unvaccinated	Vaccinated age < 18 years	Vaccinated age ≥ 18 years
Number of participants (%)	5548 (41.1)	2997 (25.4)	3225 (27.4)
Female age (years, mean)	31.2	28.8	30.9
Male partner age (years, mean)	32.3	32.0	32.4
Calendar year at enrollment
2013–2019	67.4	25.5	52.0
2020–2025	32.6	74.5	48.0
Non-Hispanic White (%)	84.9	85.0	84.0
Hispanic (%)	6.5	7.4	6.1
Non-Hispanic Black (%)	2.5	1.4	2.0
Non-Hispanic Asian (%)	2.5	1.6	3.3
Non-Hispanic multiracial or other race (%)	3.7	4.6	4.6
Household income <$50,000 (%)	15.2	9.8	8.6
Parental education: highest level < 16 years (%)	41.6	29.6	29.5
Participant education: <16 years (%)	22.5	13.1	11.5
Geographic region of residence (%)
Midwest	21.1	23.6	20.7
Northeast	19.4	27.7	27.1
South	24.0	23.2	22.3
West	15.2	17.6	20.2
Canada	20.3	7.9	9.7
Perceived Stress Scale-10 (score)	16.0	15.6	15.5
Major Depression Inventory (score)	10.7	10.0	9.7
Body mass index (kg/m2, mean)	27.1	26.6	26.3
Multivitamin or prenatal vitamin use (%)	84.1	89.3	89.1
Former cigarette smoker (%)	13.6	7.4	9.6
Current cigarette smoker (%)	6.6	4.2	4.4
Sleep duration <6 h (%)	5.2	3.4	2.4
Parous (%)	37.5	30.5	26.6
Irregular menses (%)	14.7	11.9	10.8
History of spontaneous abortion (%)	28.4	23.4	22.5
History of subfertility or infertility (%)	15.9	12.4	10.8
Number of sex partners (mean)	7.1	7.9	8.0
History of STIs (including PID) (%)	12.6	11.4	13.6
Genital warts	2.9	0.6	2.6
Genital herpes	3.4	3.6	4.8
Chlamydia	6.9	7.6	7.2
History of abnormal Pap test before HPV vaccination (%)b	28.8	0.6	11.4
Gestational week at first positive pregnancy test (mean)	4.1	4.1	4.1
Conceived index pregnancy using fertility treatment (%)	7.3	8.4	7.5

PID = pelvic inflammatory disease, STI = sexually transmitted infection.

^aAll characteristics, except for female age, are standardized to the female age distribution of cohort at baseline.

^bRepresents any history of abnormal Pap test among those not vaccinated.

Overall, there was little association between HPV vaccination and SAB incidence (HR = 0.95, 95 % CI: 0.87–1.04) (Table 2). Results did not differ markedly by age at vaccination (age < 18 years: HR = 1.03, 95 % CI: 0.92–1.16; age ≥ 18 years: HR = 0.90, 95 % CI: 0.81–1.00). Associations were relatively uniform across strata of calendar year at enrollment and history of STI/PID (Table 3). HRs were 0.92 (95 % CI: 0.80–1.06) for the 1032 SABs occurring before 6 weeks’ gestation and 0.98 (95 % CI: 0.87–1.10) for the 1375 SABs occurring ≥6 weeks’ gestation. When we excluded the 388 participants (6.17 %) with an abnormal Pap test before their HPV vaccination, results were similar to the overall results (HR = 0.96, 95 % CI: 0.88–1.06). The overall bias-adjusted HR accounting for systematic and random error was 0.88 (95 % CI 0.82–0.96) (Supplemental Fig. 1) [25].

Table 2.

HPV vaccination and spontaneous abortion among pregnancy planners in PRESTO (2013–2025).

Exposure	Number of participants	Gestational weeks	SABs	Incidence rate	Unadjusted HR (95 % CI)	Adjusted HR (95 % CI)a
Unvaccinated	8604	71,312	1191	0.0167	1.00 (Reference)	1.00 (Reference)
Vaccinated	10,858	78,767	1216	0.0154	0.93 (0.85–1.01)	0.95 (0.87–1.04)
Vaccinated < age 18	5435	36,907	595	0.0161	0.96 (0.86–1.06)	1.03 (0.92–1.16)
Vaccinated ≥ age 18	5423	41,860	621	0.0148	0.90 (0.82–0.99)	0.90 (0.81–1.00)

SAB = spontaneous abortion.

^aAdjusted for female age, calendar year, education, income, geographic region of residence, race, ethnicity, history of cigarette smoking, history of sexually transmitted infection or pelvic inflammatory disease, and parents’ highest level of education.

Table 3.

HPV vaccination and spontaneous abortion stratified by calendar year at enrollment and history of sexually transmitted infections or pelvic inflammatory disease.

Exposure	Gestational weeks	SABs	Incidence rate	Adjusted HR (95 % CI)a	Gestational weeks	SABs	Incidence rate	Adjusted HR (95 % CI)a
	Calendar year <2020				Calendar year ≥2020
Unvaccinated	49,430	705	0.0143	1.00 (Reference)	21,882	449	0.0205	1.00 (Reference)
Vaccinated	36,238	427	0.0118	0.92 (0.81–1.05)	42,530	768	0.0181	0.98 (0.86–1.11)
	History of STI or PID				No history of STI or PID
Unvaccinated	9266	156	0.0168	1.00 (Reference)	62,046	998	0.0161	1.00 (Reference)
Vaccinated	9.871	162	0.0164	1.00 (0.77–1.29)	68,896	1033	0.0150	0.95 (0.86–1.04)

PID = pelvic inflammatory disease, SAB = spontaneous abortion, STI = sexually transmitted infection.

^aAdjusted for calendar year, age, education, income, geographic region of residence, race, ethnicity, history of cigarette smoking, history of sexually transmitted infection or pelvic inflammatory disease, and parents’ highest level of education.

4. Discussion

In the present North American preconception cohort study, HPV vaccination was not appreciably associated with SAB incidence overall or by age at vaccination. Findings were generally similar across strata of calendar year and history of STI/PID, and there was little difference by timing of SAB. The association was similar after excluding vaccinated participants with an abnormal Pap test before vaccination; we hypothesized that the vaccine could not have prevented HPV in these participants. To our knowledge, there are no studies of HPV vaccination during adolescence or early adulthood and SAB incidence in adulthood. Our findings agree with most epidemiologic studies demonstrating little effect of periconceptional or early pregnancy HPV vaccination on SAB [16]. However, the latter studies differ from ours because they examined vaccine safety during pregnancy (i.e., when HPV vaccination is contraindicated).

Study limitations include our reliance on self-reported data for HPV vaccination and SAB (i.e., misclassification), lack of data on type of HPV vaccine, potential unmeasured confounding by factors such as childhood socioeconomic position, lack of information about active HPV infection, and variability in vaccine prevalence over calendar time. We accounted for time effects by controlling for and stratifying by calendar year at enrollment. QBA results indicated that exposure misclassification likely resulted in bias toward the null; therefore, the true association between HPV vaccination and SAB is likely more inverse than that observed in our study. Although cohort retention did not differ markedly by vaccination status, we cannot rule out selection bias due to differential loss-to-follow-up. SAB incidence was consistent with nationally representative data (~20 %) [9]. Given the study population comprised pregnancy planners, results may not generalize to all reproductive-aged U.S. females. Future studies with data on active HPV infection could evaluate the extent to which HPV vaccination reduces SAB incidence by protecting against HPV-associated SAB in early pregnancy.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.vaccine.2025.128152.

Data availability

Given the very sensitive nature of the outcome (miscarriage), we do not have permission from our participants to share these data.