Effect of HPV Adult Vaccination on Serum Anti-Müllerian Hormone Levels: Paired Measurements in a Retrospective Cohort

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear authors,

In the current manuscript, you have evaluated whether HPV vaccination was associated with short-term changes in anti-Müllerian hormone (AMH). You retrospectively compared 158 vaccinated and 106 unvaccinated control females. Results showed that the baseline AMH was similar between both groups, and the unadjusted AMH decline was smaller in vaccinated women (median ΔAMH −0.13 vs. −0.27 ng/mL; p = 0.015; median percent change −10.9% vs. −20.6%; p = 0.006). In the adjusted mixed-effects model, controls showed an estimated AMH decline of −27.6% per year (95% CI −35.5% to −18.7%; 33 p < 0.001). Thus, HPV vaccination was not associated with evidence of clinically meaningful short-term impairment in ovarian reserve as assessed by AMH. Observed differences in AMH alterations were modest and should be interpreted cautiously, given residual confounding, measurement variability, and reduced precision in restricted-cohort analyses. This is a very good and interesting point. However, I have a few comments:

• Why are the numbers of cases and controls not equal?
• Why was the AMH measured for the controls and the vaccinated females, despite being a retrospective study (Indication for AMH testing)?
• Why the median follow-up differs substantially between the 2 groups (6.7 vs 8.9 months)
• Investigations related to the ovarian reserve (such as hormonal contraception use, marital status, presence of PCOS, and cycle phase at testing) are not clarified.
• Sensitivity analyses of the group difference in annual AMH change need to be revised
• References 1, 2, 19, 20, 23 are of a different format.
• Other comments in the attached manuscript

Best regards,

Comments for author File: Comments.pdf

Author Response

1. Why are the numbers of cases and controls not equal?

Answer: Thank you for this comment. We did not perform strict 1:1 matching because eligibility required paired AMH measurements and application of prespecified exclusion criteria (e.g., documented PCOS/endometriosis and recorded hormonal contraceptive use), which reduced the pool of eligible controls identified from routine general gynecology clinics. Instead, we addressed comparability using a covariate-adjusted linear mixed-effects model on log(AMH) with explicit handling of follow-up time (time×group interaction), and we supported the main analysis with prespecified sensitivity analyses (Table 4).

 

2. Why was the AMH measured for the controls and the vaccinated females, despite being a retrospective study (Indication for AMH testing)?

Answer: AMH testing was not protocolized for research purposes; it was performed as part of routine clinical care. In the vaccinated cohort, AMH was obtained in real-world practice before dose 1 and again after completing the 3-dose series; in controls, AMH was measured during routine general gynecology evaluation, reflecting common practice in our setting (e.g., ovarian reserve assessment/counseling or routine gynecologic evaluation). We clarified this in the Methods section and discussed the implications for selection/healthcare utilization patterns in the Limitations. 

 

3. Why the median follow-up differs substantially between the 2 groups (6.7 vs 8.9 months)

Answer: Follow-up timing differed because repeat AMH testing occurred in routine care rather than at standardized time points. In vaccinated women, the post-vaccination AMH was often obtained in closer temporal proximity to completion of the 3-dose series, whereas in controls repeat testing was driven by individualized clinical follow-up in general gynecology clinics. Importantly, our primary analysis explicitly accounted for heterogeneous follow-up by modeling time (years) and using a time×group interaction in the mixed-effects framework, and we performed prespecified follow-up–restricted sensitivity analyses (≥3 months; 3–18 months) to assess robustness.

 

4. Investigations related to the ovarian reserve (such as hormonal contraception use, marital status, presence of PCOS, and cycle phase at testing) are not clarified.

Answer: We agree that these factors are important. During data extraction, we excluded women with documented PCOS/endometriosis and those with recorded hormonal contraceptive use. In the previous version, these exclusions were not explicitly stated, as we had listed “uncontrolled endocrine disorders potentially affecting ovarian function (e.g., uncontrolled thyroid dysfunction, hyperprolactinemia)”; we have now clarified the exclusion criteria accordingly.
Because cycle timing was not standardized in this retrospective, real-world dataset, we did not analyze cycle-dependent hormones (FSH/estradiol). We also did not analyze AFC because it is operator-dependent and ultrasound examinations were performed by multiple clinicians. We therefore focused on AMH as the primary longitudinal marker and acknowledge potential residual confounding due to incomplete documentation inherent to record-based analyses. Marital status was not uniformly recorded in the electronic records and therefore could not be included reliably.

 

5. Sensitivity analyses of the group difference in annual AMH change need to be revised

Answer: We revised the sensitivity analysis section to present the prespecified robustness checks in a clearer and more consistent manner (Table 4; Figure 2), reporting effect estimates with 95% confidence intervals and p-values and minimizing interpretation in the Results.

 

6. References 1, 2, 19, 20, 23 are of a different format.

Answer: The relevant references have been revised and formatted according to the journal’s required style.

 

7. Other comments in the attached manuscript

Answer: We carefully reviewed and corrected the additional annotated comments in the attached manuscript, including minor wording/formatting issues, to improve clarity and consistency throughout the text, tables, and figure legends.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

To correctly assess the impact of HPV vaccination on AM H, it is also necessary to consider:

1) whether the women enrolled in the study used contraceptives  (hormonal ,steroids and progestins) during the observation period and

2) whether there were  any spontaneous abortions.

3) Comments on these issues should be added in Discussion section

 

Author Response

To correctly assess the impact of HPV vaccination on AM H, it is also necessary to consider:

 

1. whether the women enrolled in the study used contraceptives (hormonal, steroids and progestins) during the observation period and
2. whether there were any spontaneous abortions.
3. Comments on these issues should be added in Discussion section.

 

Answer: During data extraction, we excluded women with documented PCOS/endometriosis and those with recorded hormonal contraceptive use. In the previous version, these exclusions were not explicitly stated, as we had listed “uncontrolled endocrine disorders potentially affecting ovarian function (e.g., uncontrolled thyroid dysfunction, hyperprolactinemia)”; we have now clarified the exclusion criteria accordingly. We addressed these issues by applying prespecified exclusions and clarifying them in the Methods: women with recorded hormonal contraceptive use were excluded, and women who were pregnant at testing or had a pregnancy occurring between AMH measurements (including spontaneous abortion) were stated. We also added/expanded discussion of residual confounding due to incomplete documentation in routine records.

Reviewer 3 Report

Comments and Suggestions for Authors

I was invited to revise the paper entitled "Effect of HPV Adult Vaccination on Serum AMH Levels: Paired Measurements in a Retrospective Cohort". it was a retrospective cohort study aimed to evaluate the impact of hpv vaccination on antimullerian hormone.

Observations:

• Title: AMH should be reported as extended text;
• Introduction section lacks in several points: HPV infection is linked also to anal cancer, oropharyngeal cancer, vulvar cancer and penis cancer. Authors should report data on HPV epidemiology;
• Authors did not reported in introduction section the type of vaccine available in Turkey with the relative proposed schedule;
• Authors did not performed any matching procedure between study groups. It could led to biased results;
• Authors should add information on possible comorbidities that could impact AMH levels.

Author Response

1. Title: AMH should be reported as extended text;

Answer: We revised the title and ensured that the abbreviation “AMH” is expanded to “anti-Müllerian hormone” at first mention, with consistent terminology throughout the manuscript.

 

2. Introduction section lacks in several points: HPV infection is linked also to anal cancer, oropharyngeal cancer, vulvar cancer and penis cancer. Authors should report data on HPV epidemiology;

Answer: We expanded the Introduction to note that persistent high-risk HPV infection is causally linked not only to cervical cancer but also to other anogenital and oropharyngeal malignancies (anal, vulvar, vaginal, penile, and oropharyngeal cancers). We also added epidemiologic context regarding HPV prevalence in reproductive-age populations and the global burden of HPV-attributable disease with geographic variation.

 

3. Authors did not reported in introduction section the type of vaccine available in Turkey with the relative proposed schedule;

Answer: We added a concise description of the vaccine type and schedule relevant to our setting and cohort. Specifically, exposure corresponded to the 9-valent HPV vaccine (Gardasil 9®) administered as a 3-dose series in adult women, reflecting real-world practice during the study period.

 

4. Authors did not performed any matching procedure between study groups. It could led to biased results;

Answer: We agree and clarified this point in the Methods. Because strict 1:1 matching was not feasible after requiring paired AMH measurements and applying exclusions, we addressed between-group differences using covariate-adjusted mixed-effects modeling with explicit handling of follow-up time (time×group interaction) and performed prespecified sensitivity analyses to assess robustness.

 

5. Authors should add information on possible comorbidities that could impact AMH levels.

Answer: We clarified comorbidity-related exclusions and confounding control. Women with documented PCOS/endometriosis and those with recorded hormonal contraceptive use were excluded; BMI was not routinely recorded and could not be adjusted for without substantial missingness and recording bias. We added these details to the Methods and acknowledged residual confounding in the Limitations. 

Reviewer 4 Report

Comments and Suggestions for Authors

Review of the article "Effect of HPV Adult Vaccination on Serum AMF Levels: Paired Measurements in a Retrospective Cohort"

The HPV vaccine was developed to prevent infection with highly oncogenic HPV strains that cause cervical cancer. Although some have claimed that both HPV infection and vaccination are associated with lower pregnancy rates, the global literature contains very few studies examining the relationship between HPV vaccination and fertility.

This makes this study relevant, and the results obtained are significant from both theoretical and practical perspectives.

 

There are several comments regarding the presented work:

There are several comments regarding the presented work:

1. It is recommended to indicate that AMH levels are a key marker of ovarian reserve in women of reproductive age, with an explanation as to why FSH and estradiol levels were not included in the analysis to fully assess hormonal status.
2. Include in the inclusion and exclusion criteria: endometriosis, polycystic ovary syndrome, and use of combined oral contraceptives.
3. It is also recommended to analyze AMH levels in groups based on the reproductive age period (early, optimal, late), given that AMH tends to decrease with increasing age.
4. There are a few studies in the international literature examining the effect of HPV vaccination on fertility, the results of which are consistent with those obtained in the presented work. However, have the authors encountered documented opposite results (where vaccination had an effect on fertility)?
5. Replace links 1, 3, 10, 11, 16 and 22 with more relevant ones.

Author Response

There are several comments regarding the presented work:

1. It is recommended to indicate that AMH levels are a key marker of ovarian reserve in women of reproductive age, with an explanation as to why FSH and estradiol levels were not included in the analysis to fully assess hormonal status.

Answer: We added an explicit rationale: because cycle timing was not standardized in this retrospective cohort, cycle-dependent hormones (FSH/estradiol) were not suitable for analysis. Additionally, AFC is operator-dependent and was performed by multiple clinicians; therefore, we focused on AMH as the primary longitudinal marker for comparison. 

 

2. Include in the inclusion and exclusion criteria: endometriosis, polycystic ovary syndrome, and use of combined oral contraceptives.

Answer: During data extraction, we excluded women with documented PCOS/endometriosis and those with recorded hormonal contraceptive use. In the previous version, these exclusions were not explicitly stated, as we had listed “uncontrolled endocrine disorders potentially affecting ovarian function (e.g., uncontrolled thyroid dysfunction, hyperprolactinemia)”. We revised the Methods to explicitly classify endometriosis/endometrioma, PCOS, and hormonal contraceptive use among the exclusion criteria applied to both cohorts. 

 

3. It is also recommended to analyze AMH levels in groups based on the reproductive age period (early, optimal, late), given that AMH tends to decrease with increasing age.

Answer: We agree that AMH declines with age and that effect modification by reproductive age period is an important consideration. We examined the feasibility of age-period stratified analyses (early/optimal/late); however, stratification would substantially reduce the effective sample size within each subgroup—particularly in the early subgroup (control n = 29)—leading to limited statistical power and imprecise estimates (wide confidence intervals), especially for formal interaction testing (time×group×age period). Therefore, we retained age as an adjustment variable in the primary mixed-effects model and emphasized this as a limitation, noting that age-related effect modification cannot be excluded.

 

4. There are a few studies in the international literature examining the effect of HPV vaccination on fertility, the results of which are consistent with those obtained in the presented work. However, have the authors encountered documented opposite results (where vaccination had an effect on fertility)?

Answer: We addressed this in the Discussion by noting that early concerns were driven mainly by case reports/small series and passive surveillance signals, which cannot establish causality, whereas the most informative evidence comes from large population-based studies that have not found an increased risk of primary ovarian insufficiency after HPV vaccination. We emphasize that the overall body of evidence does not support a clinically meaningful adverse effect on ovarian function or fertility, while acknowledging the rarity of outcomes and the importance of ongoing surveillance. 

 

5. Replace links 1, 3, 10, 11, 16 and 22 with more relevant ones.

Answer: We updated the requested citations and replaced the specified links with more relevant, peer-reviewed sources and/or authoritative guidance documents to better support the corresponding statements in the Introduction/Discussion. The reference list has been updated accordingly and reformatted consistently.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The paper should be accepted for publication

Author Response

Thank you very much for your positive evaluation of our revised manuscript. We appreciate your time and consideration. No additional changes were requested in this round; we have carefully proofread the manuscript and ensured consistency of wording, formatting, and cross-references throughout.

Reviewer 3 Report

Comments and Suggestions for Authors

Authors responded to all my previous comments. However I’m skeptikal about the lack of a matching procedure. The methodology proposed by Authors however could led to biased results. In addition, the lack of several other covariates heavily impacts results.

Author Response

Comment 1: Authors responded to all my previous comments. However I’m skeptikal about the lack of a matching procedure. The methodology proposed by Authors however could led to biased results. In addition, the lack of several other covariates heavily impacts results.

Response 1: Thank you for your thoughtful comment. We agree that lack of matching and incomplete covariate information can introduce bias in observational, record-based studies. In our dataset, strict 1:1 matching was not feasible because eligibility required paired AMH measurements and prespecified exclusions, which reduced the pool of eligible controls. We therefore addressed comparability using covariate-adjusted mixed-effects modeling with explicit handling of follow-up time and prespecified sensitivity analyses. In addition, we reported standardized mean differences (SMDs) to transparently summarize between-group baseline differences and added an SMD column in Table 1. Changes/additions were highlighted in green.