Introduction

Cervical cancer, primarily caused by infection with oncogenic human papillomavirus (HPV) on the cervix [1], is preventable through HPV vaccination, which prevents HPV infection, and cervical screening, which enables early detection and treatment of cervical precancerous lesions. In 2020, the World Health Organization (WHO) launched efforts to accelerate the elimination of cervical cancer as a public health issue [2]. However, women with mental illness present a unique challenge to this agenda, as they not only face an increased risk of cervical cancer but also tend to have reduced participation in cervical screening [3–6].

This heightened risk among women with mental illness cannot be solely attributed to lower screening rates. Even among those who participate in screening, women with specialist diagnoses of mental disorders still carry a 2-fold risk of developing cervical precancerous lesions [3]. This suggests that mental illness likely influences the risk of cervical cancer also through other mechanisms apart from differential screening participation, including modulating the risk of infection with HPV, the primary cause of cervical cancer.

Several factors may contribute to the elevated risk of HPV infection among women with mental illness, encompassing abnormal levels of immune biomarkers [7], engagement in more risky sexual behaviors [8], higher smoking rate [9], limited knowledge about HPV infection [10], and a higher chance of having experienced sexual abuse [11], compared to their counterparts. Despite these concerns, population-based studies examining HPV infection disparities between women with and without mental illness remain limited.

To address this gap, we conducted a study using data from a substantial prospective cohort of women participating in HPV-based cervical screening in Stockholm, Sweden. Our hypothesis was that women with mental illness would exhibit an increased risk of HPV infection.

Methods

Results

The study cohort included 337,116 women with a median follow-up of 2.21 years, corresponding to 779,142 person-years (Fig 1). The proportion of person-time exposed to a specialist diagnosis of mental disorder was slightly lower in the study cohort compared with the alternative general female population in Stockholm (14.20% and 16.40%). A similar pattern was found for use of psychotropic medications (42.37% and 44.41%) (Table 1). Depression, anxiety, and stress-related disorders were the most common diagnoses among all mental disorders (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Flowchart of the study design.

https://doi.org/10.1371/journal.pmed.1004372.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Prevalence of mental illness in the study cohort and the alternative female population of the greater Stockholm region.

https://doi.org/10.1371/journal.pmed.1004372.t001

Within the study cohort, women with mental illness were younger and had lower educational level, compared to women without mental illness. Women with filled prescriptions of psychotropic medications were older, but also had lower educational level, compared to women without such use. Country of birth, HPV vaccination, and maternal history of CIN3+ were generally similar between women with or without a diagnosis of mental disorder or filled prescriptions of psychotropic medication (S3 Table).

Within the study cohort, 3,263 women were tested positive for high-risk HPV during follow-up. A total of 333,853 women were censored, among whom 57,109 had at least 1 negative test results during follow-up and 276,744 did not have a test (Fig 1). Among women with at least 2 HPV tests during the study period, the incidence of high-risk HPV infection was higher among women with a diagnosis of mental disorder (HR = 1.45; 95% CI [1.34, 1.57]; p < 0.001) or filled prescription of psychotropic medication (HR = 1.67; 95% CI [1.55, 1.79]; p < 0.001), compared to women without such (Table 2). The association was similar for psychiatric disorders (HR = 1.44; 95% CI [1.32, 1.57]; p < 0.001) and neurodevelopmental disorders (HR = 1.47; 95% CI [1.19, 1.83]; p < 0.001). Analyses on individual diagnoses showed an increased infection rate of high-risk HPV among women with a diagnosis of depression, anxiety, stress-related disorder, alcohol-related disorder, substance abuse (excluding tobacco or alcohol use), ADHD, or intellectual disability (Table 2). An increase in infection rate was suggested but not statistically significant for tobacco-related disorders, whereas no risk alteration was noted for psychotic disorder and autism spectrum disorder. Analyses of individual psychotropic medications showed similarly increased infection rate of high-risk HPV for sedatives and hypnotics, anxiolytics, antidepressants, and antipsychotics (Table 2). Results from unadjusted models were similar (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. HRs with 95% CIs of any high-risk HPV infection in relation to diagnosis of mental disorder or use of psychotropic medication, using interval censored data.

https://doi.org/10.1371/journal.pmed.1004372.t002

Among all women within the study cohort, results from the right censored data analysis were largely similar to the ones obtained in the interval censored analysis, although the estimated HRs were slightly decreased (Fig 2 and Table 3). We also found similar results between HPV-16/18 and other high-risk HPV types and between adjusted and unadjusted models (Fig 2 and Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. HRs with 95% CIs of any high-risk HPV infection in relation to any diagnosis of mental disorder or any use of psychotropic medication.

*Tobacco- and alcohol-related disorders were excluded. ADHD, attention-deficit hyperactivity disorder; CI, confidence interval; CIN3+, cervical intraepithelial neoplasia grade 3 or worse; HPV, human papillomavirus; HR, hazard ratio. HR-HPV, high risk-HPV, including 14 types: 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, and 68. HRs were adjusted for age, country of birth, educational level, HPV vaccination status, and maternal history of CIN3+.

https://doi.org/10.1371/journal.pmed.1004372.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Adjusted HRs with 95% CIs for infection with HPV16/18 or other high risk-HPV in relation to diagnosis of mental disorder or use of psychotropic medication.

https://doi.org/10.1371/journal.pmed.1004372.t003

Women with mental illness were more likely to miss a subsequent screening (i.e., loss to follow-up) compared with women without mental illness (Table 4). The differential loss to follow-up was noted across all diagnoses of mental disorders and all kinds of psychotropic medications, although women with substance abuse (excluding alcohol and tobacco use) had the highest rate of loss to follow-up, whereas the result was not statistically significant for tobacco-related disorder and intellectual disability.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Differential loss to follow-up between women with and without mental illness.

https://doi.org/10.1371/journal.pmed.1004372.t004

HRs were adjusted for age, country of birth, educational level, HPV vaccination status, and maternal history of CIN3+.

Stratification analysis by age did not suggest any difference in the association of mental illness with infection rate of high-risk HPV between age groups (S4 Table). Finally, compared to women with neither a diagnosis of mental disorder nor use of psychotropic mediation, use of psychotropic medication without a specialist diagnosis was associated with a 1.43 times increased infection rate of any high-risk HPV (95% CI [1.33 to 1.55]; p < 0.001), whereas women with a specialist diagnosis of mental disorder had a 1.54 times increased infection rate of high-risk HPV (95% CI [1.40, 1.69]; p < 0.001) (S5 Table). Sensitivity analysis using 12 months as the grace period showed similar results as the main analysis (S6 Table).

Discussion

In this population-based cohort study, we found an increased rate of incident infection with high-risk HPV, either HPV-16/18 or other high-risk HPV types, among women with a specialist diagnosis of mental disorder or use of psychotropic medications without a specialist diagnosis. A statistically significant association was noted for the majority of psychiatric disorders and psychotropic medications studied, except tobacco-related disorder, psychotic disorder (specialist diagnosis or use of antipsychotics), and neurodevelopmental disorders. Women with mental illness were also more likely to miss the following screening.

The increased infection rate of high-risk HPV was found among women with a specialist diagnosis of mental disorder (i.e., relatively severe mental illness) as well as among women that used psychotropic medications without a specialist diagnosis (i.e., likely relatively milder mental illness). This is in line with our previous study [3] including a population of 4,112,598 women born during 1940 to 1995, who ever resided in Sweden during 1968 to 2018, showing an increased risk of precancerous cervical lesions among both groups of women. The 337,116 women we included in this study were at age 30 to 64, ever lived in Stockholm, and had a negative HPV test, during August 2014 to December 2019. Therefore, the population in this study is a subset of the previous study. The increased infection rate was similar between HPV-16/18 and other high-risk HPV types and was consistent across age groups, indicating a potentially common mechanism linking together mental illness and different high-risk HPV types at different ages. Moreover, the increased infection rate of high-risk HPV was independent of HPV vaccination and maternal history of CIN3+, suggesting that mechanisms other than differential vaccination uptake and genetic susceptibility to cervical cancer might contribute. For example, it has been suggested that women with mental illness are more likely to contract sexually transmitted diseases, have sexual debut at an early age [8], and smoke [9]. Several illegal drugs, including cannabis, may exhibit immunosuppressive effects, potentially increasing the susceptibility to viral infections [24]. Notably, a higher proportion of individuals with mental illness use these substances compared to the general population [25]. In addition, women with mental illnesses are more likely to have experienced adverse life events such as childhood sexual abuse [26], a traumatic life event previously shown to be associated with both the development of mental ill health and an elevated risk of HPV infection [11,27]. Although the precise mechanisms remain unknown, immune alterations might have contributed, in acquiring, clearing, and reactivating of HPV infection [26,27]. Although not all factors are modifiable, targeted interventions are possible and feasible in order to minimize such disparities.

The present study showed that, even in a cohort of women who participated in cervical screening and had a negative test of high-risk HPV to start with, women with mental illness were still more likely to miss the following screening and thus less likely to have their HPV infection detected. Therefore, HRs observed in this study might be an underestimate of the real association between mental illness and HPV infection, partly explaining the lack of statistically significant findings in the right-censored analysis for several mental illnesses, including psychotic disorder, substance abuse, and neurodevelopmental disorders. This is further consistent with our previous finding that women with psychotic disorder, substance abuse, and neurodevelopmental disorders have the lowest degree of screening participation [3].

A few studies have examined the risk of HPV infection in relation to mental illness, primarily substance abuse, and reported conflicting results [28–30]. Two studies [28,30] found no association between drug abuse and the risk of HPV infection, whereas 1 study reported an increased risk of infection with high-risk HPV in relation to cocaine use [29]. These studies all used questionnaire data and had limited sample size. For example, although 1 study [30] did suggest a higher prevalence of HPV infection among women with drug abuse (37.7%) than women without drug abuse (21.9%), no statistically significant association was found after adjusting for lifestyle factors, including age at onset of sexual activity, number of sexual partners, uptaking of cervical screening via Pap smear, and current smoking. It is unclear, however, whether these lifestyle factors should be adjusted for as they are likely mediating the causal pathways between mental illness and HPV infection. In contrast to these studies, we focused on specialist diagnosis of mental disorder, as a proxy for severe mental illness, as well as use of psychotropic medication without specialist diagnosis, as a proxy for less severe mental illness, and extended the existing evidence base from substance abuse to all major psychiatric disorders as well as neurodevelopmental disorders.

Strengths of our study include the large sample size and comprehensive data on clinical diagnoses of mental disorders, filled prescription of psychotropic medications, HPV screening results, and mother–daughter linkage from the whole female population of the greater Stockholm area. Acquired from national registers, these data are therefore less affected by selection and information bias. Importantly, the 337,116 women included in the present study are a subset of the 4 million women included in our previous study [3] and the findings corroborated with each other. However, there are a few limitations. First, as we could only detect HPV infection through screening, we might have missed infections acquired and cleared between 2 screening rounds, or infections undetected due to unattended screening. Additionally, as the eligible participants of the study must have had a negative test result for HPV infection, the study participants were likely at a lower risk of HPV infection compared to other women. As a result of both, the overall infection rate of high-risk HPV infection appeared lower in the present study, compared to previous studies [31,32]. Given that we have found women with mental illness are less likely to attend for the next screening, they tend to have longer follow-up after their last negative test, during which they are assumed to be infection-free; thus, the infection rate among women with mental illness and the overall association might be underestimated. To address this, we conducted an interval-censored analysis on women with at least 2 HPV tests, restricting the follow-up as between the first and last tests, which showed similar but slightly stronger results. However, the right-censored analysis provided the opportunity to study the entire sample of the present study, instead of only women with at least 2 HPV tests during the study, and the differential adherence to HPV screening in relation to mental illness. Second, women included in the study cohort were only a subset of the female population in Stockholm, as they must have at least 1 negative test result of HPV during the study period. However, we found no substantial difference in the proportion of mental illness between the study cohort and the entire female population in Stockholm. This argues against substantial selection bias and suggests that our results are generalizable. Lastly, we had a relatively short follow-up as HPV-based screening was only introduced in 2014 in Stockholm. As a result, the majority (82%) of women in our study did not have had the time to receive a new screening. However, Stockholm was among the earliest regions to initiate HPV-based screening in Europe [33], and our study represents the largest study to date with the longest follow-up concerning women regularly screened for HPV infection. Regardless, follow-up studies with longer follow-up or larger sample size (e.g., national data) are needed to validate these findings with better statistical power, especially for subgroup analysis.

Based on our previous study of the entire Swedish female population [3], women with mental illness participate less in cervical screening and experience a higher risk of cervical cancer. The fact that there is an increased risk of infection with high-risk HPV in relation to mental illness among women who participate in screening indicates that both factors—lower screening attendance and higher risk of acquiring high-risk HPV infection—contribute to the increased risk of invasive cervical cancer in relation to mental illness. We ended the present study in 2019 as the COVID-19 pandemic impacted greatly the HPV screening program, due to reduced testing capacity and screening participation. Fortunately, self-collected HPV tests have been introduced to enhance cervical screening since July 2020 in Sweden [34].

In conclusion, we observed an increased infection rate of high-risk HPV among women with a specialist diagnosis of psychiatric disorder or neurodevelopmental disorder or with filled prescriptions of psychotropic medications, despite that they were less likely to attend screening compared with other women. This timely information may guide interventions to achieve the WHO cervical cancer elimination agenda.

Supporting information