Primary HPV Infection in Postmenopausal Women: A Cohort Study to Enable Risk-Stratified Management

Abstract

Purpose

This study aims both to identify the independent risk factors for high-grade cervical intraepithelial neoplasia (HSIL) and to develop and validate a clinical risk stratification model.

Patients and Methods

This study retrospectively enrolled 400 postmenopausal women with HR-HPV high-risk human papillomavirus (HR-HPV) positivity and biopsy-confirmed cervical intraepithelial neoplasia. Univariate and multivariate logistic regression analyses were performed to identify independent predictors of HSIL, and a predictive model was constructed. This model was subsequently validated using an independent cohort of 172 patients.

Results

Multivariate analysis identified older age, higher parity, HPV 16/18, and hypertension as independent risk factors for HSIL. The regression model based on these factors demonstrated good discrimination in the validation set, with an area under the receiver operating characteristic curve (AUC) of 0.805 in the training set and 0.778 in the validation set. Risk stratification according to the model revealed a significant ascending trend in HSIL prevalence across the defined risk groups. The diagnostic performance for HSIL was evaluated for key predictors, showing an AUC of 0.769 for HPV 16/18 genotype alone and 0.792 for its combination with age.

Conclusion

A predictive model was constructed to effectively stratify the risk of HSIL in postmenopausal women with an initial HR-HPV infection. This tool facilitates the identification of high-risk individuals and provides a basis for individualized clinical management.

Introduction

Cervical cancer, with a substantial burden of close to 660,000 incident cases and up to 350,000 fatalities in 2022, remains a major global public health challenge.¹ It was primarily caused by persistent high-risk human papillomavirus (HR-HPV) infection.^1,2 Notably, epidemiological data reveal a second peak of HR-HPV incidence among postmenopausal women, likely driven by both systemic immunosenescence and localized genital tract changes associated with estrogen deficiency, such as mucosal atrophy and impaired cellular immunity.^3–5

With expanding HR-HPV-based screening, more postmenopausal women are being identified with their first HR-HPV infection.⁶ However, clinical management strategies for this population remain poorly defined.⁷ Current screening guidelines, largely based on data from younger cohorts, lead to high rates of colposcopy referral in postmenopausal women.⁸ This approach is problematic due to anatomical changes-cervical atrophy, the inward migration of the squamocolumnar junction, and often unsatisfactory visualization of transformation zone, which reduce the accuracy and tolerability of colposcopy while increasing patient discomfort and anxiety.^9,10

Furthermore, evidence regarding the natural history and optimal management of HR-HPV infections in women over age 65 is particularly limited. Most guidelines recommend discontinuing routine screening in this age group after adequate prior negative screens, resulting in a lack of prospective data and inconsistent clinical follow-up practices.^11,12 Given increasing life expectancy, there is a pressing need for individualized, evidence-based management strategies.

To address these gaps, the present study aimed to determine the prevalence of high-grade squamous intraepithelial lesions or worse (HSIL) and to identify key risk factors associated with prevalent disease. By analyzing existing clinical data, this study aims to develop a foundational risk assessment tool that could support more individualized management strategies, help avoid unnecessary procedures, and ultimately contribute to improving cervical cancer prevention efforts in this underserved population.

Materials and Methods

Study Design and Population

This was a retrospective, observational cohort study conducted at Women’s Hospital School of Medicine Zhejiang University. A total of 572 patients were included in this study, with 400 patients as training set, 172 patients as validation set. The study protocol was approved by the Institutional Review Board of Women’s Hospital School of Medicine Zhejiang University (Approval No: IRB-20230137-R). We included postmenopausal women aged ≥50 years with natural menopause ≥1 year, who were diagnosed with HR-HPV (HPV 16/18/31/33/35/39/45/51/52/56/58/59/66/68 infection in Jun.2023–Jun.2025, and who underwent colposcopy-directed biopsy confirming cervical intraepithelial neoplasia (CIN). All patients provided informed consent for the use of their clinical data.

Inclusion and Exclusion Criteria

Inclusion criteria were as follows: (1) age ≥50 years; (2) natural menopause ≥1 year; (3) initial HR-HPV positive test during the study period; (4) baseline colposcopy and histopathological diagnosis of CIN; and (5) complete medical records without missing key variables.¹³

Exclusion criteria included: (1) previous history of cervical treatments such as loop electrosurgical excision procedure (LEEP), conization, or hysterectomy; (2) diagnosis of other malignancies; (3) use of hormone replacement therapy after menopause; and (4) co-infection with syphilis or hepatitis B virus; (5) patients with any documented prior HPV test (positive or negative) in the hospital electronic records.

Sample Size and Power Considerations

As this was a retrospective observational study, the sample size was determined by the number of eligible patients presenting during the study period (Jun.2023–Jun.2025). A total of 572 patients met the inclusion criteria and were included in the analysis. The cohort was randomly divided into a training set (n = 400) and a validation set (n = 172) to support model development and internal validation.

Although no prospective power calculation was performed, post-hoc analysis indicated that with a sample size of 400 in the training set and an assumed odds ratio of 1.8 for key predictors, the study had over 85% power to detect significant associations at α = 0.05, supporting adequate statistical power for the analyses performed.

Data Collection

Demographic and clinical variables collected included age, parity, number of abortions, duration of menopause, HR-HPV genotype results, ThinPrep cytologic test (TCT) results, and colposcopic biopsy pathology reports. Comorbid conditions such as diabetes, autoimmune diseases, hypertension which is defined as ≥130/80 mmHg, and condyloma acuminatum were also documented.

Group Stratification

According to the 2014 World Health Organization (WHO) classification of tumors of the female reproductive organs, CINs were histologically categorized into two groups: low-grade squamous intraepithelial lesions (LSIL, classified as CIN I and CIN II); high-grade squamous intraepithelial lesions (HSIL, classified as CIN III and above).¹⁴ The distribution of TCT categories and the statistical comparisons between the LSIL and HSIL groups in training set and in the validation set were shown in Supplemental Table 1 and Supplemental Table 2, respectively.

Statistical Analysis

In the data statistics, the infection of HPV 16 and 18 was coded as 1, and others were coded as 2. Continuous variables were presented as medians with interquartile ranges and compared using the Mann–Whitney U-test. Categorical variables were expressed as frequencies and percentages and analyzed using the chi-square test or Fisher’s exact test, as appropriate.

Univariate logistic regression analyses were performed to identify factors associated with HSIL. TCT was excluded from the final model due to its substantial overlap with the histopathological endpoint, to avoid collinearity. Variables with p < 0.05 in univariate analysis were entered into a multivariate logistic regression model to identify independent risk factors. Results were reported as odds ratios (OR) with 95% confidence intervals (CI).

A risk stratification model was developed and verified based on the final multivariate model. The diagnostic performance of significant predictors (eg, HR-HPV genotype and age) for discriminating HSIL was evaluated using receiver operating characteristic (ROC) curve analysis. Calibration of the model was assessed by plotting a calibration curve in the validation set.

All statistical analyses were performed using SPSS version 27 (IBM Corp., Armonk, NY, USA), and a two-sided p-value < 0.05 was considered statistically significant.

Results

Study Population and Baseline Characteristics

A total of 400 postmenopausal women with HR-HPV infection and histologically confirmed CIN were included in training set. The cohort was stratified into two groups: 200 (50.0%) women with low-grade lesions (CIN1-2) and 200 (50.0%) with HSIL (CIN3 and above).

The baseline characteristics of the entire cohort are summarized in Table 1. The median age was 59 years (IQR: 55–64). The median parity and number of abortions were 2 (IQR: 1–2) and 1 (IQR: 0–2), respectively. The majority of patients had no history of diabetes (94.5%), autoimmune diseases (99.8%), or condyloma acuminatum (99.5%). Hypertension was present in 91 patients (22.8%).

Table 1.

Baseline Data

Baseline Data	Patients (N=400)
Age	59 (55, 59, 64)
Parity	2 (1, 2, 2)
Abortion	1 (0, 1, 2)
Duration	1 (0, 1, 2)
HPV
non-HPV 16/18	237 (59.3%)
HPV 16/18	163 (40.8%)
Diabetes
Yes	378 (94.5%)
No	22 (5.5%)
Autoimmune
Yes	399 (99.8%)
No	1 (0.3%)
Hypertension
Yes	309 (77.3%)
No	91 (22.8%)
Condyloma acuminatum
Yes	398 (99.5%)
No	2 (0.5%)
Pathology
LSIL	200 (50%)
HSIL	200 (50%)

Association Between HR-HPV Genotype and Clinical Parameters

We further analyzed the differences between patients infected with different HR-HPV genotypes (Table 2). Patients were categorized based on the presence or absence of specific high-risk genotypes (HPV 16/18+; HPV 16 or18). The two groups were comparable in terms of age, parity, duration of menopause, and the prevalence of diabetes, autoimmune diseases, hypertension, and condyloma acuminatum (all p > 0.05). However, a significant difference was observed in the distribution of pathological grades. The proportion of HSIL was significantly higher in the HPV 16/18+ group (110/163, 67.5%) compared to the HPV 16/18- group (90/237, 38.0%), and this difference was statistically significant (p < 0.001).

Table 2.

Data Analysis of Different HPV Types

Variable	HPV 16/18-	HPV 16/18+	U/χ2/F	P
Age	59 (50, 59, 87)	60 (50, 60, 81)	1.483	0.138
Parity	1 (0, 1, 6)	2 (0, 2, 7)	1.911	0.056
Abortion	1 (0, 1, 8)	1 (0, 1, 6)	−2.214	0.027
Duration	1 (0, 1, 3)	1 (0, 1, 3)	1.030	0.303
Diabetes
Yes	226	152	0.825	0.364
No	11	11
Autoimmune
Yes	236	163	0.689	1.000
No	1	0
Hypertension
Yes	188	121	1.425	0.233
No	49	42
Condyloma acuminatum
Yes	236	162	0.071	0.790
No	1	1
Pathology
LSIL	147	53	33.641	< 0.001
HSIL	90	110

Univariate Analysis of Factors Associated with HSIL

Univariate analysis was performed to identify factors associated with the development of HSIL (Table 3). Several factors were significantly associated with a higher risk of HSIL. These included older age (median 60 vs 58 years, p < 0.001), higher parity (median 2 vs 1, p < 0.001), higher number of abortions (median 1 vs 1, p < 0.001), longer duration of menopause (median 1 vs 1, p < 0.001), and the presence of hypertension (32.0% vs 13.5%, p < 0.001). Crucially, infection with HPV 16/18 was strongly associated with HSIL, with 110 (68.8%) of the HSIL+ group being HPV 16/18+ compared to 53 (22.2%) in the HSIL- group (p < 0.001). A history of diabetes was also more common in the HSIL+ group (8.0% vs 3.0%, p = 0.028).

Table 3.

Univariate Analysis of Factors Associated with HSIL

Variable	HSIL-	HSIL+	χ2/F	P
Age	58 (50, 58, 77)	60 (50, 60, 87)	4.741	< 0.001
Parity	1 (0, 1, 7)	2 (0, 2, 6)	4.511	< 0.001
Abortion	1 (0, 1, 8)	1 (0, 1, 5)	−3.627	< 0.001
Duration	1 (0, 1, 3)	1 (0, 1, 3)	4.184	< 0.001
HPV 16/18
Yes	53	110	33.641	< 0.001
No	147	90
Diabetes
Yes	194	184	4.810	0.028
No	6	16
Autoimmune
Yes	200	199	1.003	0.317
No	0	1
Hypertension
Yes	173	136	19.474	< 0.001
No	27	64
Condyloma acuminatum
Yes	198	200	2.010	0.156
No	2	0

Multivariate Logistic Regression Analysis

Variables with p < 0.05 from the univariate analysis were incorporated into a multivariate logistic regression model to identify independent risk factors for HSIL (Table 4). The analysis confirmed that age (OR = 1.061, 95% CI: 1.020–1.103, p = 0.003), parity (OR = 1.409, 95% CI: 1.097–1.811, p = 0.007), infection with HPV 16/18 (OR = 3.246, 95% CI: 2.079–5.068, p < 0.001), and hypertension (OR = 2.464, 95% CI: 1.426–4.256, p = 0.001) were independent predictors of HSIL.

Table 4.

Multivariate Analysis of Factors Associated with HSIL

Variable	B	SE	Wald	P	OR (95% CI)
Age	0.059	0.020	32.490	0.003	1.061 (1.020~1.103)
Parity	0.343	0.128	7.179	0.007	1.409 (1.097~1.811)
HPV	1.178	0.227	26.847	< 0.001	3.246 (2.079~5.068)
Hypertension	0.092	0.279	10.453	0.001	2.464 (1.426~4.256)

A risk stratification model was constructed based on this multivariate model (Figure 1A and Table 5), with the area under the curve (AUC) of 0.805. The regression equation was as follows: Logit(P) = −4.794 + (0.059 × Age) + (0.343 × Parity) + (1.178 × HPV status) + (0.902 × Hypertension status), where HPV status and Hypertension were coded as 1 for positive and 0 for negative.

Figure 1.

Development and validation of the predictive model for HSIL. (A) Receiver Operating Characteristic (ROC) curve of the multivariate logistic regression model in the training set. (B) ROC curve of the same model applied to the independent validation set.

Table 5.

Performance of the Prediction Model in the Training and Validation Sets

Sets	AUC (95% CI)	P	Sensitivity	Specificity	Youden Index
Training set	0.805 (0.762~0.848)	<0.001	56.41	83.46	0.40
Validation set	0.778 (0.681~0.876)	<0.001	61.54	93.98	0.56

Model Validation

The multivariate model was subsequently validated in an independent, validation set of 172 postmenopausal patients who met the same inclusion and exclusion criteria (Figure 1B and Table 5). The model demonstrated good discriminatory power in the validation set, with an area under the ROC curve of 0.778 (95% CI: 0.681–0.876). The calibration curve demonstrated a high agreement between the predicted probabilities and the actual risks (Supplementary Figure 1).

Risk Stratification

For clinical application, a risk stratification system was established based on the predicted probabilities from the logistic regression model. Patients were categorized into three distinct risk groups: low-risk, intermediate-risk, and high-risk. The prevalence of HSIL in these groups was 18.5%, 33.5%, and 48%, respectively, showing a significant ascending trend (Figure 2A and Table 6, p < 0.001). This trend was consistently validated in the independent validation set, where the HSIL prevalence across the low-, medium-, and high-risk groups was 0%, 15.9%, and 66.7%, respectively (Figure 2B and Table 6, p < 0.001). This stratification effectively identified a subgroup of patients with a very high likelihood of possessing HSIL.

Figure 2.

Risk stratification and prevalence of HSIL. (A and B) Distribution of patients across the three risk categories (low, intermediate, and high) derived from the logistic regression model in training set (A) and validation set (B).

Table 6.

Distribution of the Binary Variable Across Different Risk Groups

	Variable	Low Risk	Medium Risk	High Risk	P
Training set	LSIL	107	56	37	<0.001
	HSIL	37	67	96
Validation set	LSIL	3	122	8	<0.001
	HSIL	0	23	16

Diagnostic Performance of the Risk Model

The diagnostic value of the significant predictors from the multivariate model, specifically HR-HPV genotype and age, for discriminating HSIL was evaluated using ROC curve analysis (Figure 3A and Table 7). The AUC for the alone or combination of HR-HPV and age was 0.769 (95% CI: 0.722–0.815), 0.773 (95% CI: 0.726–0.818), and 0.792 (95% CI: 0.748–0.835) demonstrating good diagnostic performance. The model exhibited robust and reproducible performance in the validation set. Notably, the combined model achieved an even higher AUC of 0.820 (95% CI: 0.747–0.893), with a sensitivity of 79.5% and a specificity of 75.9% (Figure 3B and Table 7).

Figure 3.

Diagnostic performance of predictors for discriminating HSIL. (A and B) ROC curves comparing the diagnostic value of individual predictors and their combination in training set (A) and validation set (B).

Table 7.

Diagnostic Performance of Individual Predictors and Their Combination

	Groups	AUC (95% CI)	P	Sensitivity	Specificity	Youden Index
Training set	HPV	0.769 (0.722~0.815)	<0.001	70.00	78.00	0.48
	Age	0.773 (0.726~0.818)	<0.001	63.00	78.50	0.42
	Combine	0.792 (0.748~0.835)	<0.001	81.50	66.00	0.48
Validation set	HPV	0.755 (0.666~0.845)	<0.001	74.36	76.69	0.62
	Age	0.724 (0.623~0.825)	<0.001	61.50	79.70	0.41
	Combine	0.820 (0.747~0.893)	<0.001	79.50	75.90	0.55

Discussion

This retrospective cohort study provides a critical investigation into the risk factors for high-grade cervical lesions among postmenopausal women experiencing their first detected HR-HPV infection. Our findings illuminate that this population is not a monolith of low risk but comprises distinct subgroups with varying disease potentials. The core discovery of our research is that a combination of easily obtainable clinical variables-specific HR-HPV genotype, older age, higher parity, and the presence of hypertension-can effectively stratify the risk of HSIL. We successfully developed and validated a multivariate model based on these factors, which demonstrated good discriminatory performance. This risk stratification tool holds significant promise for refining clinical management strategies, potentially guiding more personalized and effective follow-up for postmenopausal women who present with an initial HR-HPV positive test.

The strong independent association between HPV 16/18 and HSIL in our cohort aligns with the well-established oncogenic potential of certain viral types, particularly HPV 16 or18, in the general population.^15,16 However, its pronounced significance in postmenopausal women underscores a critical point: the fundamental biology of HPV-driven carcinogenesis remains potent even in later life. The decline in immune surveillance associated with aging, known as immunosenescence, may create a permissive environment for viral persistence and progression.^4,5 Our findings suggest that genotyping is not merely an academic exercise but a clinical necessity in postmenopausal women. A positive HR-HPV test in this demographic should prompt immediate genotyping, as the presence of a high-risk genotype, per our model, confers a more than three-fold increase in the odds of having an underlying HSIL.

While prior literature indicates links between hypertension and poorer cervical cancer outcomes, our analysis identified older age and hypertension as factors associated with the endpoint in this cohort.^17,18 This unconventional finding highlights their potential predictive utility in our context, though the nature of this association requires further clarification. A particularly intriguing and less conventional finding from our study is the identification of older age and hypertension as independent risk factors. While the correlation between advancing age and a second peak of HPV incidence has been noted epidemiologically, its link to disease severity in a cohort already confined to HPV-positive women is highly suggestive.^3,19 Beyond immune-senescence, the cumulative lifetime exposure to carcinogenic insults and the progressive accumulation of somatic genetic alterations likely contribute to this age-related risk increase.^20,21 The association with hypertension, however, is more novel in the context of cervical carcinogenesis. We hypothesize that hypertension may not be a direct cause but rather a surrogate marker for broader systemic aging and diminished vascular and tissue health. Chronic hypertension is linked to microvascular dysfunction and a state of chronic, low-grade inflammation, which could theoretically compromise local tissue repair and immune responses in the cervical epithelium, facilitating the progression of HPV-related lesions.^22–26 Alternatively, hypertension may be a component of a metabolic syndrome phenotype, which has been previously suggested to influence cancer risk through hormonal and inflammatory pathways.^27,28 This finding warrants further investigation into the potential role of cardiovascular health in gynecological neoplasia.

Furthermore, our results confirm that higher parity is a significant risk factor, which is consistent with the extensive literature in younger women linking multiparity to an increased risk of cervical cancer.^29–31 The proposed mechanisms include hormonal changes promoting carcinogenesis and trauma to the cervix during childbirth, which might facilitate HPV entry and integration. Our study extends this established association to the postmenopausal population, indicating that reproductive history continues to exert a long-term influence on cervical cancer risk.

The clinical implications of our study are substantial. Current management guidelines for HR-HPV positive women, largely derived from younger populations, often lead to a one-size-fits-all approach of immediate colposcopy for postmenopausal women.⁸ This is problematic given the anatomical challenges in this group, such as cervical atrophy and a non-visible transformation zone, which can make colposcopy less accurate, more uncomfortable, and sometimes unsatisfactory.^9,10 Our risk stratification model offers a pathway towards a more nuanced and individualized algorithm. For instance, a postmenopausal woman with a first-time HR-HPV infection involving a high-risk genotype, advanced age, and comorbidities like hypertension could be justifiably prioritized for expedited colposcopic assessment. Conversely, a woman of a similar age with a lower-risk HPV type, lower parity, and no hypertension might be managed more conservatively with a period of observation and repeat testing, thereby avoiding an immediate invasive procedure and its associated anxiety and potential morbidity. This approach aligns with the broader goals of value-based healthcare-maximizing benefit while minimizing harm and resource utilization.

Several limitations of our study must be acknowledged. First, its retrospective and single-center design inherently introduces the potential for selection bias. Of note, the multivariable model we constructed has the risk of overfitting, that is, the model that performs well on the current dataset may not be well generalized to other independent populations. Second, we only relied on histologically but not p16 staining confirmed CIN as our endpoint, and we cannot rule out the possibility that some women in the low-grade group might have had occult higher-grade disease, although this is a universal challenge in cervical pathology. Third, our analysis was limited to variables available in the electronic medical records. Data on other potential confounders, such as detailed sexual history, smoking status, body mass index, or specific hormonal levels, were not available and their influence remains unexplored. Future studies incorporating a wider array of molecular and lifestyle factors could further refine the predictive model.

Conclusion

In conclusion, this study identifies a distinct risk profile for HSIL among postmenopausal women with a first-time HR-HPV infection, characterized by HPV 16/18, advanced age, higher parity, and hypertension. The developed and validated risk-stratification model demonstrates good discriminatory ability and can assist clinicians in personalizing management. By distinguishing women at higher risk of underlying significant disease, the model may support more efficient allocation of diagnostic resources and reduce unnecessary procedures for lower-risk individuals. These findings contribute to improving cervical cancer prevention in this growing population segment. Further multicenter, prospective studies are recommended to confirm the model’s generalizability and to evaluate its impact on clinical decision-making and patient outcomes.

Funding Statement

This work was supported by Zhejiang Provincial Natural Science Foundation of China under Grant No. LTGY23H160023.

Data Sharing Statement

All the results are presented in the article. Further inquiries can be directed to the corresponding authors.

Ethics Statement

The research protocol was approved by the Ethics Committee of Women’s Hospital, Zhejiang University School of Medicine (Approval Number: IRB-20230137-R). All experiments and procedures were performed according to the Declaration of Helsinki (as revised in 2013). All patients provided informed consent for the use of their clinical data.

Disclosure

The authors report no conflicts of interest in this work.