Skip to main content
Download PDF

Risk factors for human papillomavirus infection, cervical intraepithelial neoplasia and cervical cancer: an umbrella review and follow-up Mendelian randomisation studies

BMC Medicine volume 21, Article number: 274 (2023) Cite this article

Abstract

Background

Persistent infection by oncogenic human papillomavirus (HPV) is necessary although not sufficient for development of cervical cancer. Behavioural, environmental, or comorbid exposures may promote or protect against malignant transformation. Randomised evidence is limited and the validity of observational studies describing these associations remains unclear.

Methods

In this umbrella review, we searched electronic databases to identify meta-analyses of observational studies that evaluated risk or protective factors and the incidence of HPV infection, cervical intra-epithelial neoplasia (CIN), cervical cancer incidence and mortality. Following re-analysis, evidence was classified and graded based on a pre-defined set of statistical criteria. Quality was assessed with AMSTAR-2. For all associations graded as weak evidence or above, with available genetic instruments, we also performed Mendelian randomisation to examine the potential causal effect of modifiable exposures with risk of cervical cancer. The protocol for this study was registered on PROSPERO (CRD42020189995).

Results

We included 171 meta-analyses of different exposure contrasts from 50 studies. Systemic immunosuppression including HIV infection (RR = 2.20 (95% CI = 1.89–2.54)) and immunosuppressive medications for inflammatory bowel disease (RR = 1.33 (95% CI = 1.27–1.39)), as well as an altered vaginal microbiome (RR = 1.59 (95% CI = 1.40–1.81)), were supported by strong and highly suggestive evidence for an association with HPV persistence, CIN or cervical cancer. Smoking, number of sexual partners and young age at first pregnancy were supported by highly suggestive evidence and confirmed by Mendelian randomisation.

Conclusions

Our main analysis supported the association of systemic (HIV infection, immunosuppressive medications) and local immunosuppression (altered vaginal microbiota) with increased risk for worse HPV and cervical disease outcomes. Mendelian randomisation confirmed the link for genetically predicted lifetime smoking index, and young age at first pregnancy with cervical cancer, highlighting also that observational evidence can hide different inherent biases. This evidence strengthens the need for more frequent HPV screening in people with immunosuppression, further investigation of the vaginal microbiome and access to sexual health services.

Peer Review reports

Background

Although persistent infection with high-risk human papillomavirus (hrHPV) is causally associated with cervical cancer, only a fraction of women that get infected with HPV develop persistence, high-grade CIN and if not detected and treated cervical cancer. Cancer promotion in some individuals is likely to be explained by a complex interplay between the host system, the HPV virus and behavioural, environmental, or comorbid factors. Genetic variation and predisposition to cervical cancer only explain a small amount of the difference in underlying risk between individuals [1]. HrHPV is the most common sexually transmitted infection with over 70% of women being infected during their lifetime [2]. Most hrHPV infections are cleared by incompletely understood immune response, the epidemiological and lifestyle factors leading to hrHPV persistence, and especially neoplastic progression, are not fully understood.

Over the last three decades, many epidemiological studies have investigated the risk factors associated with hrHPV persistence and development of cervical cancer including immunosuppression [3,4,5,6], concomitant sexual infections [7,8,9], risky sexual behaviour [10, 11] and tobacco smoking [12]. However, most behavioural, environmental, or comorbid exposures are not suitable for investigation by randomised design trials and the evidence base is therefore subject to inherent biases. For some reported associations, a wide range in the magnitude of the effect size has been observed and studies have reported opposing directions of effect for the same exposure, such as for early age of first pregnancy [13, 14] or tobacco smoking [4, 15]. Determining the true strength of an association from p-values alone can be misleading, and selective reporting of positive results can also lead to an overestimation of the strength of an association. Previous umbrella reviews have demonstrated that despite numerous reported significant associations across differing scientific specialties, very few survive more rigorous analyses [16,17,18,19,20,21]. Although umbrella reviews offer a further appraisal of the evidence, they cannot infer causality if the underlying studies are observational in design. To explore the potential causal relationships of identified exposures, Mendelian randomisation (MR) can be useful and complementary to traditional observational studies, as genetic instruments, where available, can control for a degree of unknown confounding.

We performed an umbrella review of systematic reviews and meta-analyses exploring the association between modifiable risk factors and environmental exposures and hrHPV infection, cervical intraepithelial neoplasia (CIN) and cervical cancer incidence and mortality. We further conducted a MR analysis, to assess the strength and validity and potential causal effect of previously reported estimates.

Methods

Search strategy and selection criteria

We searched PubMed, Ovid MEDLINE and Embase Classic, and the Cochrane database for systematic reviews to investigate the association between non-genetic behavioural, environmental, or comorbid risk factors and incidence or prevalence of hrHPV, CIN or cervical cancer, cervical cancer mortality, and regression or progression of disease. All articles were screened at least in duplicate (SB, TD and AA) using pre-defined search terms (Additional file 1 - Supplementary Methods). We further hand-searched the reference lists of included papers for and the proceedings of relevant conferences for unpublished data (SB, TD and AA) (Fig. 1). The protocol for this study is available on PROSPERO (CRD42020189995). We included all systematic reviews and meta-analyses of observational or interventional studies on behavioural, environmental or co-morbid risk factors affecting hrHPV, CIN or cervical cancer incidence or mortality - including outcomes concerning disease regression and progression. We excluded studies investigating genetic risk factors. We also excluded meta-analyses that did not report the necessary study-specific data including the relative risk (RR), 95% confidence intervals (CI), the number of cases/controls or total population (where we could not retrieve data from original studies). Where more than one meta-analysis examined the same exposure-outcome pair, we chose the meta-analysis containing the largest number of cohort studies. If the same number of cohort studies were included, we included the more recently published meta-analysis.

Fig. 1
figure 1

PRISMA flow chart of study-selection

Full size image

Data extraction

We extracted the individual study level data for each study within each meta-analysis including the number of cases and controls or total population and the maximally adjusted relative risk (further details provided in Additional file 1 - Supplementary Methods). All the data extraction was performed at least in duplicate (TD, AA, SB and HGS), with any discrepancies resolved by discussion with a third investigator (IK).

Data analysis and evaluating the strength of evidence by grading criteria

For each exposure and outcome pair, we calculated the summary effect and the 95% CI using fixed and random effects methods [22]. The heterogeneity between studies was assessed with Cochran’s Q test [23] and the I2 statistic [24] with 95% CI [25]. To further account for heterogeneity between studies we calculated 95% prediction intervals [26, 27] for the summary random effect estimates. We assessed whether smaller studies gave higher risk estimates than larger studies - an indication of publication bias, true heterogeneity, or chance. To assess small study effects, we used Egger’s regression asymmetry test (P ≤ 0.10) [28] and whether random effects summary estimates are larger than the point estimate of the largest study in the meta-analysis. We assessed excess significance bias by evaluating whether the observed number of studies with nominally statistically significant results (p < 0.05) in the published literature is different from the expected number of studies [25]. Finally, we used a credibility ceilings threshold to account that a single observational study cannot give more than a maximum certainty [29]. We graded the strength of evidence into strong, highly suggestive, suggestive and weak using criteria as previously described [16, 17] and are outlined in the supplementary material (Additional file 2 - Supplementary Table 1).

Evaluation of the quality of included meta-analyses

We used the AMSTAR-2 criteria [30] to assess the quality of the evidence. Evidence was graded based on the strength and validity as according to previous published umbrella reviews [16, 17]. The four grades range from ‘high’ to ‘critically low’, with ‘high’ evidence having zero or one non-critical weakness in the study, and ‘critically low’ having more than one critical flaw, with or without a non-critical weakness (Additional file 1 - Supplementary Methods).

Sensitivity analyses

The main analysis was restricted to cohort studies only. We conducted a sensitivity analysis which included all eligible observational studies, i.e. meta-analyses including both cohort and case–control studies. These were then graded using the same criteria as for the main analyses.

In the event of multiple meta-analyses reporting on the same exposure-outcome associations, we selected the one with the largest number of studies to prevent duplication of the original studies. The concordance between the included and duplicate meta-analyses was explored in a sensitivity analysis (Additional file 1 - Supplementary Methods).

Mendelian randomisation

To investigate the potential causal effects of proposed environmental risk factors on cervical cancer, we conducted a two-sample MR analysis, which uses genetic variants with known effects on the risk factor, as a proxy for the exposure [31, 32].

For all exposures graded as weak evidence or above in the umbrella review, we searched the GWAS Catalog [33] to identify relevant GWAS studies providing female-specific summary-level genetic data. We used data from a previously published genome-wide association study to obtain associations of SNPs with risks of cervical cancer [1].

We used inverse variance weighted (IVW) MR as the main analysis to estimate the causal effect on risk for cervical cancer [34]. For exposures with significant effects in the main IVW MR analysis (multiple testing adjusted using Benjamini and Hochberg’s false discovery rate [35]), we performed sensitivity analyses using a range of robust MR approaches to circumvent possible violations of the instrumental variable assumptions (weighted median [36], MR-Egger [37], and Mendelian Randomisation Pleiotropy RESidual Sum and Outlier (MR-PRESSO) [38] (Additional file 1 - Supplementary Methods)).

Role of the funding source

The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

Results

The literature search was performed in November 2020 and identified 4722 systematic reviews or meta-analytical papers. Following sequential title, abstract and full-text screening, 56 meta-analysis papers met the inclusion criteria. After exclusion of duplicate meta-analyses, 50 papers remained [4, 7, 10, 12, 14, 39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83], which included 171 separate meta-analyses of 1513 individual primary studies, (447 were cohort studies and 1056 case–control) (Fig. 1); no randomised controlled trials were identified.

Characteristics of the meta-analyses

A total of 11 outcomes were identified, which related to either HPV infection, CIN and/or cervical cancer: cervical disease regression (hrHPV clearance; CIN regression), disease incidence and prevalence (hrHPV incidence; hrHPV prevalence; CIN incidence; CIN prevalence; cancer incidence), disease persistence (hrHPV persistence; CIN persistence), disease progression (CIN progression) and cancer mortality.

Fifty exposures were identified, which belonged to eight broad categories: immunocompromise (HIV infection, inflammatory bowel disease (IBD) on medication, rheumatoid arthritis (RA), systemic lupus erythematous (SLE), transplant recipients); co-infection and vaginal microbiome (VMB), (Epstein-Barr virus (EBV) infection, Chlamydia trachomatis, Trichomonas, U. urealyticum, Mycoplasmas, Candida albicans, vaginal Lactobacillus spp., herpes simplex virus (HSV) infection); anthropometric measures (body mass index (BMI), height); lifestyle and/or behavioural factors (smoking, alcohol intake, number of sexual partners); medical co-morbidities and/or medication use (retinoid use); gynaecological and obstetric factors (parity, age of first pregnancy, pregnancy, in vitro fertilisation (IVF), gestational diabetes mellitus (GDM), vaginal douching); hormonal medication use (combined oral contraceptive pill (COCP), intrauterine devices (IUD), injectable contraception); dietary, vitamin or antioxidant intake (vitamins A, E, C, lycopene, folate and carotenoids, vegetable/fruit, zinc, copper and selenium).

Quality assessment

Twelve per cent of all the included papers were graded as high quality (1/50, 2%) [54] or moderate quality (5/50, 10%) [7, 41, 42, 60, 63], with Helm et al. [54] being the only research study that adequately fulfilled all the major components of the AMSTAR2 questionnaire. On the contrary, 14% (7/50) and 74% (37/50) of meta-analytical papers were graded as low or critically low quality, respectively (Additional file 3 - Supplementary Table 2). Papers characterised as ‘low’ or ‘critically low’ quality failed to meet criteria related to protocol, literature search, description of excluded studies and risk of bias assessment. Specifically, only 18% (9/50) listed the excluded studies and reason for exclusion. Additionally, 38% (19/50) did not use a pre-registered protocol, while 20% (10/50) did not report screening the literature adequately (either because they searched only one electronic database and/or did not provide a search strategy). Study selection and data extraction were performed in duplicate in 40% (20/50) and 42% (21/50), respectively. Most provided satisfactory methods for the risk of bias assessment (80%, 40/50), statistical analysis (76%, 38/50), interpretation of the main results (70%, 35/50) and investigations of small study effects (66%, 33/50).

Main analysis

Of the 171 meta-analyses, 87 included two or more cohort studies and were included in the main analysis and assessed a total of 39 exposures across the eight categories (Table 1, Fig. 2).

Table 1 Summary of evidence grading for meta-analysis of risk factors associated with HPV infection and pre-invasive and invasive cervical cancer outcomes — cohort studies onlya
Full size table
Fig. 2
figure 2

Distribution of studies across evidence grade for all exposures of either increased or decreased risk (y-axis) by exposure category (X-axis) from the main analysis (summary random effects for cohort studies only) and outcome a HPV infection; b cervical intraepithelial neoplasia; and c cervical cancer

Full size image

Summary effect size

Applying p < 0.05 as a threshold for the level of significance, the summary fixed effects estimates were significant in 62% of cohort studies (54/87) and the summary random effects in 60% (52/87) of the meta-analyses (Additional file 4 - Supplementary Table 3). When using p < 0.001 as a cut-off, 53% (45/87) and 39% (34/87) of the studies presented significant summary fixed and random effects estimates, respectively. Where a cut-off of p < 10−6 was applied, 32% (28/87) and 19% (16/87) of the meta-analyses produced significant summary results in the fixed and random effects model, respectively. Out of the 16 meta-analyses with a random p < 10−6, 14 exposures were associated with an increased risk of either hrHPV incidence or persistence, or the increased risk of progression to LSIL or HSIL, CIN or cervical cancer (vaginal dysbiosis, HIV + , IBD on immunosuppression, Chlamydia trachomatis infection and co-infection with hrHPV, smoking and rheumatoid arthritis), while two exposures were associated with a decreased risk for hrHPV clearance (HIV + ; HIV + with low CD4 + cell count).

Heterogeneity between studies

The Cochrane’s Q test for heterogeneity was significant at p ≤ 0.10 in 27 of 87 meta-analyses (31%). Twenty-five studies (29%) presented a high heterogeneity (I2 = 50–75%) and nine (10%) very high (I2 > 75%), which included 5 different exposures (HIV + ; HIV + on treatment; bacterial vaginosis; Chlamydia trachomatis; pregnancy). When calculating the 95% prediction intervals, the null hypothesis was excluded for nine associations (IBD on immunosuppression – cervical cancer incidence; vaginal dysbiosis – hrHPV incidence; vaginal dysbiosis – progression from normal to LSIL or HSIL; HIV + – hrHPV incidence; HIV + – CIN2 + treatment failure; HIV + – CIN1 + treatment failure; smoking – hrHPV incidence; smoking – cervical cancer incidence; COCP – cervical cancer incidence) (Additional file 5 - Supplementary Table 4).

Small study effects

In three meta-analyses, there was evidence of small study effects (Egger’s test of P < 0.10 and detection of more conservative effects in the largest study of a meta-analysis compared with the summary random effects estimate): Chlamydia trachomatis infection – hrHPV incidence, HIV + – CIN1 + treatment failure, and gestational diabetes mellitus – cervical cancer incidence (Additional file 5 - Supplementary Table 4).

Excess significance

When using the largest study estimate as the plausible effect size, evidence of excess significance was observed in eight meta-analyses (9%) of varying exposures including bacterial vaginosis, Chlamydia trachomatis infection, HIV positivity, RA, BMI, and current pregnancy. Using the fixed or random effect estimate as the plausible effect sizes, five and four meta-analyses presented excess significance bias respectively (Additional file 5 - Supplementary Table 4).

Grading the evidence

Of the 87 meta-analyses included in the main analysis, three met criteria to be graded as strong evidence, a further three as highly suggestive, nine as suggestive, while thirty-seven meta-analyses were of weak evidence; the remaining meta-analyses showed null associations (Additional file 6 - Supplementary Table 5). Out of the meta-analyses with strong and highly suggestive evidence, an increased risk of hrHPV and HPV incidence was associated with HIV positivity (strong evidence, N = 2323, RR 2.20, 95% CI 1.89–2.54, p = 3.01 × 10−26, I2 = 22 and highly suggestive evidence, N = 1151, RR 3.1, 95% CI = 2.17–4.4, random p = 3.75 × 10−10, I2 = 82, respectively), while a decreased risk of hrHPV clearance was also associated with HIV positivity (highly suggestive evidence, N = 2977, RR 0.53, 95% CI = 0.43–0.65, p = 4.32 × 10−10, I2 = 73). An increased risk for progression from normal to LSIL or HSIL was related to vaginal dysbiosis (strong evidence, N = 27,405, RR = 1.6, 95% CI = 1.4–1.81, p = 5.34 × 10−12, I2 = 26) and an increased risk for cervical cancer incidence was associated with people with IBD using immunosuppressive medications (strong evidence, N = 10,829, RR = 1.33, 95% CI = 1.89–2.54), p = 7.78 × 10−37, I2 = 0) and with HIV positivity (highly suggestive evidence, N = 1160, RR = 5.82, 95% CI = 2.98–11.34, p = 2.34 × 10−07, I2 = 86) (Fig. 3).

Fig. 3
figure 3

Forest plot of effect estimates and 95% confidence intervals for all exposure-outcome pairs for all outcomes (HPV, CIN (cervical intraepithelial neoplasia) and cervical cancer) in the main analysis (summary random effects for cohort studies only), which graded as strong or highly suggestive evidence (n = 6)

Full size image

Two of the meta-analyses [41, 42] that met criteria for strong evidence were of moderate quality on AMSTAR assessment while the third [64] was evaluated as critically low quality. The most common reason for lower quality on AMSTAR for those studies was no report of the funding source, no description of the excluded studies, partial description of the included studies and for the search strategy. Similarly, most of the meta-analyses that were graded as highly suggestive evidence were graded as critically-low quality on AMSTAR assessment for the above reasons as well as for poor study design.

Sensitivity analyses

Of the 87 meta-analyses in the main analysis, 37 (43%) retained nominal statistical significance (P < 0.05) with a credibility ceiling of 5%. With ceilings of 10%, 15% and 20%, twenty-two (25%), fourteen (16%) and four (5%) studies remained significant respectively (Additional file 7 - Supplementary Table 6).

When also including case–control studies, 11/171 meta-analyses met the criteria for strong evidence, including eight additional exposure-outcome pairs (Additional file 8 - Supplementary Table 7). Increased risk of hrHPV incidence was associated with Chlamydia trachomatis infection (N = 5049, RR = 2.32, 95% CI = 2.02–2.65, p = 3.45 × 10−34, I2 = 0%), while increased risk of CIN was related to multiple sexual partners CIN (N = 5638, RR = 1.97, 95% CI = 1.8–2.15, p =  < 1 × 10−100, I2 = 0%). An increased risk of cervical cancer incidence was associated with not only Chlamydia trachomatis infection (N = 3392, RR = 2.19, 95% CI = 1.74–2.74, p = 1.03 × 10−11, I2 = 47%), but also with Trichomonas Vaginalis infection (N = 7715, RR = 2.09, 95% CI = 1.69–2.6, p = 2.33 × 10−11, I2 = 34%) and oral contraception use (N = 5839, RR = 2.13, 95% CI = 1.87–2.42, p = 4.37 × 10−31, I2 = 1.4%). Meanwhile, co-infection of hrHPV with Chlamydia Trachomatis seems to increase the risk of cervical cancer by more than four times (N = 1086, RR = 4.37, 95% CI = 2.75–6.96, p = 4.59 × 10−10, I2 = 44%), while in a subgroup analysis Chlamydia trachomatis infection appears to increase the risk of squamous cell cervical cancer incidence (N = 3198, RR = 2.09, 95% CI = 1.79–2.44, p = 6.87 × 10−21, I2 = 32%) (evidence for adenocarcinoma was weak only).

Of the eight additional associations that met strong criteria in the sensitivity analysis, five were graded as weak or did not present nominally statistically significant association at p < 0.05 when only cohort studies were included. The remaining three included only case–control studies and were not assessed in the main analysis. All three strong associations from the main analysis remained strong when case–control studies were included. 14 studies met criteria for highly suggestive evidence, 27 met criteria for suggestive and 69 studies were classified as weak evidence only; the remaining 51 meta-analyses did not meet criteria for weak evidence (Additional file 8 - Supplementary Table 7).

We identified duplicate meta-analyses meeting inclusion criteria for six exposure-outcome pairs. For all duplicates the same direction of effect was observed in both the magnitude and significance of the summary associations, between included and excluded studies (Additional file 9 - Supplementary Table 8).

Mendelian randomisation

In the MR analysis, 11 exposures had genetic instruments available to perform MR (Additional file 10 - Supplementary Table 9), namely smoking using a lifetime smoking index (which is a composite score that captures the lifetime smoking exposure by taking into account smoking status, as well as smoking duration, heaviness and cessation) [84], reproductive behaviour in women as measured by age at first pregnancy [85], number of sexual partners [86], lupus [87], rheumatoid arthritis, IBD, alcohol consumption [88], BMI, GDM, parity (number of living births) and height [89]. No hrHPV infection phenotypes were identified; however, genetic instruments for the aggregate CIN3 and cervical cancer phenotypes were available. The characteristics of the GWAS studies from which we selected the genetic instruments can be found in the supplement (Additional file 11 - Supplementary Table 10) and in the originally published studies.

The strongest associations with risk of cervical cancer were observed for genetically predicted lifetime smoking index (OR = 2.46, 95% CI = 1.64–3.69) and number of sexual partners (OR = 1.95, 95% CI = 1.44–2.63) (Fig. 4 and Additional file 12 - Supplementary Table 11). We additionally identified a protective effect for older age at first pregnancy (OR = 0.80, 95% CI = 0.68–0.95), while genetically predicted liability to rheumatoid arthritis increased the risk of cervical cancer (OR = 1.10 95% CI = 1.05–1.15) (Fig. 5). The associations of genetically predicted lifetime smoking index, age at first pregnancy, liability to rheumatoid arthritis, and number of sexual partners, with cervical cancer risk were supported in sensitivity analysis (Additional file 13 - Supplementary Tables 12–13). Genetically predicted age of first pregnancy, and lifetime smoking index, were independently associated with cervical cancer when controlling for genetically predicted number of sexual partners. Genetically predicted liability to SLE, IBD, GDM, and genetically predicted alcohol consumption, BMI, nulliparity and height all showed no significant association with CIN3 or cervical cancer (Additional file 13 - Supplementary Tables 12–13).

Fig. 4
figure 4

Forest plot demonstrating inverse variance weighted Mendelian randomisation results for all identified known environmental risk or protective factors for cervical cancer with available GWAS, to determine effect sizes by OR and 95% CI (x-axis; n = 11)

Full size image
Fig. 5
figure 5

Summary of results from umbrella review and Mendelian randomisation: associations with strong evidence in the main analysis (left): HIV for human papillomavirus incidence, vaginal dysbiosis for development of cervical intraepithelial neoplasia, and inflammatory bowel disease on immunosuppressive therapy for cervical cancer and Mendelian randomisation results supporting an association with cervical cancer incidence (right: rheumatoid arthritis, smoking, number of sexual partners and older age at first pregnancy (protective effect))

Full size image

Discussion

We present an umbrella review of 87 meta-analyses of observational studies on life course exposures including co-morbidities, environmental and behavioural exposures and risk of HPV infection, CIN and invasive cervical cancer (ICC). HIV, vaginal dysbiosis (Lactobacillus spp. depletion) and immunosuppressive medications in women with IBD were all supported by strong evidence, while smoking, Chlamydia infection and bacterial vaginosis were supported by highly suggestive or suggestive evidence. Sensitivity analyses also identified strong or highly suggestive evidence for an association between Trichomonas infection, increased number of sexual partners, medium- to long-term COCP usage, high parity, earlier age at first pregnancy, low vegetable intake, increased vitamin C or selenium intake, and HPV, CIN, or ICC.

We identified meta-analyses with strong, highly suggestive, and suggestive evidence for an increased risk of hrHPV and cervical cancer incidence in HIV-positive women, when compared to the general population. There was also highly suggestive evidence that HIV positivity reduces the risk of hrHPV clearance, whereas suggestive evidence exists that antiretroviral treatment (ART) increases CIN regression rates. The included studies were graded as low and critically low in the AMSTAR assessment while there were no genetic instruments to perform an MR analysis. Our findings support that of a recent study published jointly by the WHO and the International Agency for Research on Cancer (IARC), which reported that women living with HIV have a six-fold higher risk of developing cervical cancer [6]. Immunosuppression is considered the primary mechanism of many HIV-related diseases, and of HPV persistence [90], while genetic studies of women with CIN3 and cervical cancer have identified more frequent mutations in the HLA region coding for MHC class 2 cell production and T-cell activation [1]. Early ART initiation is thought to improve HPV clearance and CIN regression through maintenance of high levels of CD4 + T-cells [58].

The evidence for women with other forms of chronic immunosuppression and cervical cancer was more limited. We found strong evidence in the main analysis that women with IBD on immunosuppression are at increased risk for high-grade CIN and cervical cancer compared to the general population. The quality of the study scored moderate on AMSTAR while MR analysis didn’t show any significant association. It is unclear, though, whether IBD alone is associated with cervical cancer regardless of medication.

We found weak evidence for an increased risk of cervical cancer for other autoimmune diseases, including a meta-analysis on SLE and CIN. Multiple studies have suggested that SLE alone is associated with hrHPV acquisition [91], CIN [91,92,93,94,95,96] and that the use of immunosuppressants in SLE patients is associated with an increased incidence of cervical dysplasia [96, 97]. The risk associated with RA was also weak. People living with rheumatoid arthritis have been found to be at significantly higher risk of other cancers including lymphoma and lung cancer, compared to the general population [70, 98]. As well as the immune dysfunction induced by RA itself, it is likely that immunosuppressive medications like steroids and disease-modifying antirheumatic drugs (DMARDS) which can decrease the number and function of T-cells [99] play a role in the risk of cancer. Although MR did not support a causal association for genetically predicted liability to IBD and SLE to CIN3 and cervical cancer, there was nominal evidence of causality for genetically predicted liability to RA (OR = 1.1, 95% CI = 1.05–1.15), it was not plausible however to assess the role of the disease alone in separation to immunosuppressive medications however. The studies for both SLE and RA were graded as critically low in the AMSTAR assessment mainly due to the quality of literature search, description of the included studies, funding, and statistical methods for the MA.

We found strong and suggestive evidence that a Lactobacillus spp. deplete VMB, as well as cervicovaginal infection (including Chlamydia trachomatis and Trichomonas spp.) increase the risk of hrHPV acquisition, CIN incidence and cancer incidence, including progression from normal cytology to HSIL. Evidence was particularly strong for progression to LSIL or HSIL [42] due to the large number of individuals from cohort studies (N = 460,746). The evidence of the association between chlamydia and hrHPV incidence [67] was downgraded to suggestive in the main analysis. The highest score in the AMSTAR assessment was moderate and that was for the study with strong evidence for altered VMB and CIN. The rest of the studies were graded as low or critically low. Lactobacillus spp. maintain a low pH and the acidic environment is essential for the function of the cervical epithelial barrier [100,101,102] and Chlamydia infection can disrupt the cervical epithelium, allowing increased HPV entry to basal epithelium [103,104,105]. Studies have demonstrated that VMB diversity increases with advancing CIN disease severity [106], while lactobacillus-depleted VMB is associated with a significantly lower chance of regression of untreated CIN1 when compared to Lactobacillus-dominant VMB [107]. There were insufficient genetic instruments to assess the VMB or vaginal infection with Chlamydia or Trichomonas with MR.

Although smoking is widely considered to be a strong risk factor for cervical cancer, the evidence was graded as suggestive between smoking and increased HPV incidence and as only weak between active smoking and ICC. The meta-analyses of ICC were considered to present only weak evidence of association due to the small number of cases, relatively large p-values and the presence of small study effects. All the studies investigating smoking scored critically low in AMSTAR, mainly due to study design, risk of bias assessment, heterogeneity and small study bias. Cigarette smoking is widely accepted as a strong carcinogen which hampers cellular immunity at the cervix [108] and this analysis highlights the inherent issues that can arise from the quality of observational research. Particularly, that the study of a rare outcome such as cervical cancer can be difficult in a cohort design and that larger numbers of cases can only be achieved in a case–control study. MR suggests a true causal effect that may have been obscured in meta-analyses based on cohort studies. For cigarette smoking, our MR suggested a strong causal effect of lifetime smoking index on an increased risk of cervical cancer (OR = 2.46, 95% CI = 1.64–3.69), this was still apparent when controlling for other risky behaviours.

There were no meta-analyses based on cohort studies for the evaluation of the sexual and reproductive history on cervical cancer. The association of multiple sexual partners [10, 70] with CIN and cervical cancer was supported with strong and highly suggestive evidence in a sensitivity analysis of 6 meta-analyses, and confirmed in the MR analysis, although it is challenging to instrument such a variable, which might largely reflect a general propensity for risky behaviours. The quality of the studies was low in AMSTAR. While an increased number of sexual partners is thought to increase hrHPV exposure, particularly at an early age, the mechanism regarding parity and early age of pregnancy is less well understood. It is possible that an early age of first pregnancy affects the transformation zone, increasing its vulnerability to infection [109, 110] or the immunosuppressed pregnant state increases vulnerability. However, high parity and early age of first pregnancy may be surrogate markers for increased HPV exposure at an earlier age. We explored the independence of age of first pregnancy and sexual behaviour via a multivariate MR. We observed that a genetically predicted young age of first pregnancy was independently associated with cervical cancer when controlling for risky behaviours such as genetically predicted higher number of sexual partners.

While BMI was linked to increased cervical cancer mortality possibly due to other obesity-related complications, there was no link to any other HPV or cervical cancer outcome, which was confirmed on MR. Our study provided only weak evidence that the use of COCP increases the risk of invasive cervical cancer, although the evidence became strong for medium to long-term COCP use when case–control studies were included as well [109].

We used a well-established methodology for this analysis [16, 17, 21, 111]. A lack of evidence does not infer the absence of an association; however, where a weak evidence grade was assigned, this may suggest the need for further good-quality studies. This is particularly true in the context of associations that are widely thought to be causal, as downgrading of evidence results from presence of biases in the evaluated literature, not from suspected absence of association. As causality cannot be inferred from observational research, and a lack of randomised research was observed from this systematic review, we performed an MR wherever genetic instruments were available. This is the first MR study in the field of cervical cancer risk and brings new insights to the possible causality from examined exposures.

Possible limitations should be considered in the interpretation of our findings. This review relies on the previously published meta-analyses and literatures searches performed by the authors of those studies. Some literature may have been missed; however, the assessment of duplicate analyses did not highlight any discordant results which minimises this risk. Additionally, we studied exposures against several outcomes linked to the development of cervical cancer, these results were also consistent across meta-analyses. Although the overall number of studies included was large, for some associations, the number of studies and participants was small, limiting our ability to assess for the presence of small study effects and excess significance due to low power. It is likely that this would result in a more conservative estimate and the true association may be more severe. Additionally, most studies in the umbrella were graded as low or critically low according to the AMSTAR2 criteria, suggesting a high risk of bias within this evidence. Although we assessed for risk of bias, the statistical tests are unable to explain the definitive presence or the likely source of bias. Furthermore, MR analyses were underpowered for some exposures (such as for IBD), and sample overlap in some of the associations was relatively large (such as the number of sexual partners) [112].

Conclusions

There is consistently strong and highly suggestive evidence that HIV positivity reduces HPV clearance rates and increases the risk of HPV infection and cervical cancer development. Vaginal Lactobacillus spp. depletion and immunosuppressive medications for women with IBD are also strongly associated, with a suggestion that other forms of immunosuppression increase risk of cervical cancer. Our results suggest that the presence of Chlamydia trachomatis and Trichomonas infections influence the development of cervical cancer in the presence of hrHPV and prompt treatment should be prioritised. In conservatively managed HPV infections and CIN, additional screening and treatment for concomitant bacterial infection and bacterial vaginosis should be considered [113, 114]. This strengthens the call for more evidence on the role of probiotics in preventing HPV persistence and cervical cancer. While for cigarette smoking, we found highly suggestive evidence in case–control studies only, this is likely secondary to small numbers of cervical cancers and smoking cessation should still be recommended to women with cervical abnormalities. Young age of first pregnancy was independently associated with cervical cancer when controlling for other risky behaviours including a higher number of sexual partners and smoking.

This evidence highlights the importance of preventative strategies including the provision of sexual health and family planning services, with early initiation of ART in HIV-positive women, alongside cervical screening. The strong interaction between HIV and cervical cancer necessitates the prioritisation of HPV vaccination in populations where HIV prevalence is high in initiatives to increase global access.

Availability of data and materials

All data generated or analysed during this study are included in this published article [and its supplementary files].

Abbreviations

AMSTAR:

A MeaSurement Tool to Assess systematic Reviews

ART:

Antiretroviral treatment

BMI:

Body mass index

CD4:

Cluster of differentiation 4

CI:

Confidence interval

CIN:

Cervical intraepithelial neoplasia

CIN1:

Cervical intraepithelial neoplasia grade 1

CIN1 + :

Cervical intraepithelial neoplasia grade 1 or higher

CIN2 + :

Cervical intraepithelial neoplasia grade 2 or higher

CIN3:

Cervical intraepithelial neoplasia grade 3

COCP:

Combined oral contraceptive pill

DMARDS:

Disease-modifying antirheumatic drugs

EBV:

Epstein-Barr virus

GDM:

Gestational diabetes mellitus

GWAS:

Genome-wide association study

HIV:

Human immunodeficiency virus

HLA:

Human leukocyte antigens

HPV:

Human papillomavirus

HrHPV:

High-risk human papillomavirus

HSIL:

High-grade squamous intraepithelial lesion

HSV:

Herpes simplex virus

IARC:

International agency for research on cancer

IBD:

Inflammatory bowel disease

ICC:

Invasive cervical cancer

IUD:

Intrauterine devices

IVF:

In vitro fertilisation

IVW:

Inverse variance weighted

LSIL:

Low-grade squamous intraepithelial lesion

MHC:

Major histocompatibility complex

MR:

Mendelian randomization

MR-PRESSO:

Mendelian randomization - pleiotropy residual sum and outlier

OR:

Odds ratio

PROSPERO:

International prospective register of systematic reviews

RA:

Rheumatoid arthritis

RR:

Risk ratio

SLE:

Systemic lupus erythematosus

SNPs:

Single nucleotide polymorphisms

VMB:

Vaginal microbiome

WHO:

World Health Organisation

References

  1. Bowden SJ, Bodinier B, Kalliala I, Zuber V, Vuckovic D, Doulgeraki T, et al. Genetic Variation in cervical preinvasive and invasive disease: a genome-wide association study. Lancet Oncol. 2021;In Press.

  2. Bowden SJ, Kyrgiou M. Human papillomavirus. Obstet Gynaecol Reprod Med. 2020;30(4):109–18.

    Article  Google Scholar 

  3. Mitra A, MacIntyre DA, Marchesi JR, Lee YS, Bennett PR, Kyrgiou M. The vaginal microbiota, human papillomavirus infection and cervical intraepithelial neoplasia: what do we know and where are we going next? Microbiome. 2016;4(1):58.

  4. Appleby P, Beral V, Berrington de González A, Colin D, Franceschi S, Goodill A, et al. Carcinoma of the cervix and tobacco smoking: collaborative reanalysis of individual data on 13,541 women with carcinoma of the cervix and 23,017 women without carcinoma of the cervix from 23 epidemiological studies. Int J Cancer. 2006;118(6):1481–95.

    Article  CAS  PubMed  Google Scholar 

  5. Chaturvedi AK, Madeleine MM, Biggar RJ, Engels EA. Risk of human papillomavirus-associated cancers among persons with AIDS. J Natl Cancer Inst. 2009;101(16):1120–30.

    Article  PubMed  PubMed Central  Google Scholar 

  6. Stelzle D, Tanaka LF, Lee KK, Ibrahim Khalil A, Baussano I, Shah ASV, et al. Estimates of the global burden of cervical cancer associated with HIV. Lancet Glob Health. 2021;9(2):e161–9.

    Article  CAS  PubMed  Google Scholar 

  7. Zhu H, Shen Z, Luo H, Zhang W, Zhu X. Chlamydia trachomatis infection-associated risk of cervical cancer: a meta-analysis. Medicine (Baltimore). 2016;95(13):e3077.

    Article  PubMed  Google Scholar 

  8. Ji Y, Ma XX, Li Z, Peppelenbosch MP, Ma Z, Pan Q. The burden of human papillomavirus and chlamydia trachomatis coinfection in women: a large cohort study in inner Mongolia, China. J Infect Dis. 2019;219(2):206–14.

    Article  CAS  PubMed  Google Scholar 

  9. Yang M, Li L, Jiang C, Qin X, Zhou M, Mao X, et al. Co-infection with trichomonas vaginalis increases the risk of cervical intraepithelial neoplasia grade 2–3 among HPV16 positive female: a large population-based study. BMC Infect Dis. 2020;20(1):642.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Liu ZC, Liu WD, Liu YH, Ye XH, Chen SD. Multiple sexual partners as a potential independent risk factor for cervical cancer: a meta-analysis of epidemiological studies. Asian Pac J Cancer Prev. 2015;16(9):3893–900.

    Article  PubMed  Google Scholar 

  11. International Collaboration of Epidemiological Studies of Cervical Cancer. Cervical carcinoma and sexual behavior: collaborative reanalysis of individual data on 15,461 women with cervical carcinoma and 29,164 women without cervical carcinoma from 21 epidemiological studies. Cancer Epidemiol Biomarkers Prev. 2009;18(4):1060–9.

  12. Haverkos HW, Soon G, Steckley SL, Pickworth W. Cigarette smoking and cervical cancer: part I: a meta-analysis. Biomed Pharmacother. 2003;57(2):67–77.

    Article  PubMed  Google Scholar 

  13. Hinkula M, Pukkala E, Kyyrönen P, Laukkanen P, Koskela P, Paavonen J, et al. A population-based study on the risk of cervical cancer and cervical intraepithelial neoplasia among grand multiparous women in Finland. Br J Cancer. 2004;90(5):1025–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. International Collaboration of Epidemiological Studies of Cervical Carcinoma. Cervical carcinoma and reproductive factors: collaborative reanalysis of individual data on 16,563 women with cervical carcinoma and 33,542 women without cervical carcinoma from 25 epidemiological studies. Int J Cancer. 2006;119(5):1108–24.

    Article  Google Scholar 

  15. Berrington De González A, Sweetland S, Green J. Comparison of risk factors for squamous cell and adenocarcinomas of the cervix: a meta-analysis. Br J Cancer. 2004;90(9):1787–91.

    Article  PubMed  PubMed Central  Google Scholar 

  16. Kalliala I, Markozannes G, Gunter MJ, Paraskevaidis E, Gabra H, Mitra A, et al. Obesity and gynaecological and obstetric conditions: umbrella review of the literature. BMJ. 2017;359:j4511.

    Article  PubMed  PubMed Central  Google Scholar 

  17. Kyrgiou M, Kalliala I, Markozannes G, Gunter MJ, Paraskevaidis E, Gabra H, et al. Adiposity and cancer at major anatomical sites: umbrella review of the literature. BMJ. 2017;356:j477.

    Article  PubMed  PubMed Central  Google Scholar 

  18. Papadimitriou N, Markozannes G, Kanellopoulou A, Critselis E, Alhardan S, Karafousia V, et al. An umbrella review of the evidence associating diet and cancer risk at 11 anatomical sites. Nat Commun. 2021;12(1):4579.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Markozannes G, Aretouli E, Rintou E, Dragioti E, Damigos D, Ntzani E, et al. An umbrella review of the literature on the effectiveness of psychological interventions for pain reduction. BMC Psychol. 2017;5(1):31.

    Article  PubMed  PubMed Central  Google Scholar 

  20. Li X, Meng X, Timofeeva M, Tzoulaki I, Tsilidis KK, Ioannidis JPA, et al. Serum uric acid levels and multiple health outcomes: umbrella review of evidence from observational studies, randomised controlled trials, and Mendelian randomisation studies. BMJ (Clinical research ed). 2017;357:j2376.

  21. Raglan O, Kalliala I, Markozannes G, Cividini S, Gunter MJ, Nautiyal J, et al. Risk factors for endometrial cancer: an umbrella review of the literature. Int J Cancer. 2019;145(7):1719–30.

    Article  CAS  PubMed  Google Scholar 

  22. DerSimonian RLN. Meta-analysis in clinical trials. Control Clin Trials. 1986;7(3):177–88.

    Article  CAS  PubMed  Google Scholar 

  23. Wg C. The combination of estimates from different experiments. Biometrics. 1954;10:101–29.

    Article  Google Scholar 

  24. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327(7414):557–60.

    Article  PubMed  PubMed Central  Google Scholar 

  25. Ioannidis JP, Trikalinos TA. An exploratory test for an excess of significant findings. Clin Trials. 2007;4(3):245–53.

    Article  PubMed  Google Scholar 

  26. IntHout J, Ioannidis JP, Rovers MM, Goeman JJ. Plea for routinely presenting prediction intervals in meta-analysis. BMJ Open. 2016;6(7):e010247.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Riley RD, Higgins JP, Deeks JJ. Interpretation of random effects meta-analyses. BMJ. 2011;342:d549.

    Article  PubMed  Google Scholar 

  28. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315(7109):629–34.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Salanti G, Ioannidis JP. Synthesis of observational studies should consider credibility ceilings. J Clin Epidemiol. 2009;62(2):115–22.

    Article  PubMed  Google Scholar 

  30. Shea BJ, Reeves BC, Wells G, Thuku M, Hamel C, Moran J, et al. AMSTAR 2: a critical appraisal tool for systematic reviews that include randomised or non-randomised studies of healthcare interventions, or both. BMJ. 2017;358:j4008.

    Article  PubMed  PubMed Central  Google Scholar 

  31. Burgess S, Foley CN, Zuber V. Inferring causal relationships between risk factors and outcomes from genome-wide association study data. Annu Rev Genomics Hum Genet. 2018;19:303–27.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Davies NM, Holmes MV, Davey SG. Reading Mendelian randomisation studies: a guide, glossary, and checklist for clinicians. BMJ. 2018;362:k601.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Buniello A, MacArthur JAL, Cerezo M, Harris LW, Hayhurst J, Malangone C, et al. The NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted arrays and summary statistics 2019. Nucleic Acids Res. 2019;47(D1):D1005–12.

    Article  CAS  PubMed  Google Scholar 

  34. Burgess S, Butterworth A, Thompson SG. Mendelian randomization analysis with multiple genetic variants using summarized data. Genet Epidemiol. 2013;37(7):658–65.

    Article  PubMed  PubMed Central  Google Scholar 

  35. Benjamini Y, Hochberg Y. Controlling the false discovery rate - a practical and powerful approach to multiple testing. J Royal Statist Soc Series B. 1995;57:289–300.

    Google Scholar 

  36. Bowden J, Davey Smith G, Haycock PC, Burgess S. Consistent estimation in Mendelian randomization with some invalid instruments using a weighted median estimator. Genet Epidemiol. 2016;40(4):304–14.

    Article  PubMed  PubMed Central  Google Scholar 

  37. Bowden J, Davey Smith G, Burgess S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol. 2015;44(2):512–25.

    Article  PubMed  PubMed Central  Google Scholar 

  38. Verbanck M, Chen CY, Neale B, Do R. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet. 2018;50(5):693–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. WCRF. World Cancer Research Fund International Systematic Literature Review. The Associations between Food, Nutrition and Physical Activity and the Risk of Cervical Cancer. 2018.

  40. IARC. Cervical cancer and hormonal contraceptives: collaborative reanalysis of individual data for 16 573 women with cervical cancer and 35 509 women without cervical cancer from 24 epidemiological studies. Lancet. 2007;370(9599):1609–21.

    Article  Google Scholar 

  41. Allegretti JR, Barnes EL, Cameron A. Are patients with inflammatory bowel disease on chronic immunosuppressive therapy at increased risk of cervical high-grade dysplasia/cancer? A meta-analysis. Bowel Dis. 2015;21(5):1089–97.

    Article  Google Scholar 

  42. Brusselaers N, Shrestha S, van de Wijgert J, Verstraelen H. Vaginal dysbiosis and the risk of human papillomavirus and cervical cancer: systematic review and meta-analysis. Am J Obstet Gynecol. 2019;221(1):9–18 e8.

    Article  PubMed  Google Scholar 

  43. Cao D, Shen K, Li Z, Xu Y, Wu D. Association between vitamin C Intake and the risk of cervical neoplasia: a meta-analysis. Nutr Cancer. 2016;68(1):48–57.

    Article  CAS  PubMed  Google Scholar 

  44. Cao S, Gan Y, Dong X, Lu Z. Herpes simplex virus type 2 and the risk of cervical cancer: a meta-analysis of observational studies. Arch Gynecol Obstet. 2014;290(6):1059–66.

    Article  CAS  PubMed  Google Scholar 

  45. Castellsagué X, Díaz M, Vaccarella S, de Sanjosé S, Muñoz N, Herrero R, et al. Intrauterine device use, cervical infection with human papillomavirus, and risk of cervical cancer: a pooled analysis of 26 epidemiological studies. Lancet Oncol. 2011;12(11):1023–31.

    Article  PubMed  Google Scholar 

  46. Chen S, Shen L, Luo S, Lan X, Wang L. Association between serum iron levels and the risk of cervical cancer in Chinese: a meta-analysis. J Int Med Res. 2020;48(3):300060519882804.

    Article  CAS  PubMed  Google Scholar 

  47. Cortessis VK, Barrett M, Brown Wade N, Enebish T, Perrigo JL, Tobin J, et al. Intrauterine device use and cervical cancer risk: a systematic review and meta-analysis. Obstet Gynecol. 2017;130(6):1226–36.

    Article  PubMed  Google Scholar 

  48. de Lima MAP, Neto PJN, Lima LPM, Goncalves Junior J, Teixeira Junior AG, Teodoro IPP, et al. Association between Epstein-Barr virus (EBV) and cervical carcinoma: a meta-analysis. Gynecol Oncol. 2018;148(2):317–28.

    Article  PubMed  Google Scholar 

  49. Debeaudrap P, Sobngwi J, Tebeu PM, Clifford GM. Residual or recurrent precancerous lesions after treatment of cervical lesions in human immunodeficiency virus-infected women: a systematic review and meta-analysis of treatment failure. Clin Infect Dis. 2019;69(9):1555–65.

    Article  PubMed  PubMed Central  Google Scholar 

  50. Gillet E, Meys JF, Verstraelen H, Bosire C, De Sutter P, Temmerman M, et al. Bacterial vaginosis is associated with uterine cervical human papillomavirus infection: a meta-analysis. BMC Infect Dis. 2011;11:10.

    Article  PubMed  PubMed Central  Google Scholar 

  51. Gillet E, Meys JF, Verstraelen H, Verhelst R, De Sutter P, Temmerman M, et al. Association between bacterial vaginosis and cervical intraepithelial neoplasia: systematic review and meta-analysis. PLoS ONE. 2012;7(10):e45201.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  52. Grulich AE, van Leeuwen MT, Falster MO, Vajdic CM. Incidence of cancers in people with HIV/AIDS compared with immunosuppressed transplant recipients: a meta-analysis. Lancet. 2007;370(9581):59–67.

    Article  PubMed  Google Scholar 

  53. He D, Wang Z, Huang C, Fang X, Chen D. Serum selenium levels and cervical cancer: systematic review and meta-analysis. Biol Trace Elem Res. 2017;179(2):195–202.

    Article  CAS  PubMed  Google Scholar 

  54. Helm CW, Lorenz DJ, Meyer NJ, Rising WW, Wulff JL. Retinoids for preventing the progression of cervical intra-epithelial neoplasia. Cochrane Database Syst Rev. 2013;6(6):CD003296.

  55. Hu X, Li S, Zhou L, Zhao M, Zhu X. Effect of vitamin E supplementation on uterine cervical neoplasm: a meta-analysis of case-control studies. PLoS One. 2017;12(8):e0183395.

  56. Josyula S, Lin J, Xue X, Rothman N, Lan Q, Rohan TE, et al. Household air pollution and cancers other than lung: a meta-analysis. Environ Health. 2015;14:24.

    Article  PubMed  PubMed Central  Google Scholar 

  57. Kaderli R, Schnuriger B, Brugger LE. The impact of smoking on HPV infection and the development of anogenital warts. Int J Colorectal Dis. 2014;29(8):899–908.

    Article  PubMed  Google Scholar 

  58. Kelly H, Weiss HA, Benavente Y, de Sanjose S, Mayaud P, Qiao YL, et al. Association of antiretroviral therapy with high-risk human papillomavirus, cervical intraepithelial neoplasia, and invasive cervical cancer in women living with HIV: a systematic review and meta-analysis. Lancet HIV. 2018;5(1):e45–58.

    Article  PubMed  Google Scholar 

  59. Lee PN, Thornton AJ, Hamling JS. Epidemiological evidence on environmental tobacco smoke and cancers other than lung or breast. Regul Toxicol Pharmacol. 2016;80:134–63.

    Article  CAS  PubMed  Google Scholar 

  60. Li LL, Zhou J, Qian XJ, Chen YD. Meta-analysis on the possible association between in vitro fertilization and cancer risk. Int J Gynecol Cancer. 2013;23(1):16–24.

    Article  PubMed  Google Scholar 

  61. Liang Y, Chen M, Qin L, Wan B, Wang H. A meta-analysis of the relationship between vaginal microecology, human papillomavirus infection and cervical intraepithelial neoplasia. Infect Agent Cancer. 2019;14:29.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  62. Liu G, Sharma M, Tan N, Barnabas R. HIV-positive women have higher risk of HPV infection, precancerous lesions, and cervical cancer: a systematic review and meta-analysis. Aids. 2018.

  63. Liu P, Xu L, Sun Y, Wang Z. The prevalence and risk of human papillomavirus infection in pregnant women. Epidemiol Infect. 2014;142(8):1567–78.

    Article  CAS  PubMed  Google Scholar 

  64. Looker KJ, Ronn MM, Brock PM, Brisson M, Drolet M, Mayaud P, et al. Evidence of synergistic relationships between HIV and Human Papillomavirus (HPV): systematic reviews and meta-analyses of longitudinal studies of HPV acquisition and clearance by HIV status, and of HIV acquisition by HPV status. J Int AIDS Soc. 2018;21(6):e25110.

    Article  PubMed  PubMed Central  Google Scholar 

  65. Manhart LE, Koutsky LA. Do condoms prevent genital HPV infection, external genital warts, or cervical neoplasia? A meta-analysis. Sex Transm Dis. 2002;29(11):725–35.

    Article  PubMed  Google Scholar 

  66. Myung SK, Ju W, Kim SC, Kim H, Korean Meta-analysis Study G. Vitamin or antioxidant intake (or serum level) and risk of cervical neoplasm: a meta-analysis. BJOG. 2011;118(11):1285–91.

    Article  PubMed  Google Scholar 

  67. Naldini G, Grisci C, Chiavarini M, Fabiani R. Association between human papillomavirus and chlamydia trachomatis infection risk in women: a systematic review and meta-analysis. Int J Public Health. 2019;64(6):943–55.

    Article  PubMed  Google Scholar 

  68. Peng Y, Wang X, Feng H, Yan G. Is oral contraceptive use associated with an increased risk of cervical cancer? An evidence-based meta-analysis. J Obstet Gynaecol Res. 2017;43(5):913–22.

    Article  PubMed  Google Scholar 

  69. Plummer M, Herrero R, Franceschi S, Meijer CJ, Snijders P, Bosch FX, et al. Smoking and cervical cancer: pooled analysis of the IARC multi-centric case–control study. Cancer Causes Control. 2003;14(9):805–14.

    Article  PubMed  Google Scholar 

  70. Simon TA, Thompson A, Gandhi KK, Hochberg MC, Suissa S. Incidence of malignancy in adult patients with rheumatoid arthritis: a meta-analysis. Arthritis Res Ther. 2015;17:212.

    Article  PubMed  PubMed Central  Google Scholar 

  71. Smith JS, Green J, de Gonzalez AB, Appleby P, Peto J, Plummer M, et al. Cervical cancer and use of hormonal contraceptives: a systematic review. Lancet. 2003;361(9364):1159–67.

    Article  PubMed  Google Scholar 

  72. Tamarelle J, Thiebaut ACM, de Barbeyrac B, Bebear C, Ravel J, Delarocque-Astagneau E. The vaginal microbiota and its association with human papillomavirus, Chlamydia trachomatis, Neisseria gonorrhoeae and Mycoplasma genitalium infections: a systematic review and meta-analysis. Clin Microbiol Infect. 2019;25(1):35–47.

    Article  CAS  PubMed  Google Scholar 

  73. Tomita LY, Horta BL, da Silva LLS, Malta MB, Franco EL, Cardoso MA. Fruits and vegetables and cervical cancer: a systematic review and meta-analysis. Nutr Cancer. 2021;73(1):62–74.

    Article  PubMed  Google Scholar 

  74. Wang H, Ma Y, Li R, Chen X, Wan L, Zhao W. Associations of cervicovaginal lactobacilli with high-risk human papillomavirus infection, cervical intraepithelial neoplasia, and cancer: a systematic review and meta-analysis. J Infect Dis. 2019;220(8):1243–54.

    Article  PubMed  Google Scholar 

  75. Wang Y, Yan P, Fu T, Yuan J, Yang G, Liu Y, et al. The association between gestational diabetes mellitus and cancer in women: a systematic review and meta-analysis of observational studies. Diabetes Metab. 2020;46(6):461–71.

    Article  CAS  PubMed  Google Scholar 

  76. Xie Y, Wang J, Zhao X, Zhou X, Nie X, Li C, et al. Higher serum zinc levels may reduce the risk of cervical cancer in Asian women: a meta-analysis. J Int Med Res. 2018;46(12):4898–906.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  77. Yang S, Zhao W, Wang H, Wang Y, Li J, Wu X. Trichomonas vaginalis infection-associated risk of cervical cancer: a meta-analysis. Eur J Obstet Gynecol Reprod Biol. 2018;228:166–73.

    Article  PubMed  Google Scholar 

  78. Ye H, Song T, Zeng X, Li L, Hou M, Xi M. Association between genital mycoplasmas infection and human papillomavirus infection, abnormal cervical cytopathology, and cervical cancer: a systematic review and meta-analysis. Arch Gynecol Obstet. 2018;297(6):1377–87.

    Article  PubMed  Google Scholar 

  79. Zard E, Arnaud L, Mathian A, Chakhtoura Z, Hie M, Touraine P, et al. Increased risk of high grade cervical squamous intraepithelial lesions in systemic lupus erythematosus: a meta-analysis of the literature. Autoimmun Rev. 2014;13(7):730–5.

    Article  PubMed  Google Scholar 

  80. Zeng XT, Xiong PA, Wang F, Li CY, Yao J, Guo Y. Passive smoking and cervical cancer risk: a meta-analysis based on 3,230 cases and 2,982 controls. Asian Pac J Cancer Prev. 2012;13(6):2687–93.

    Article  PubMed  Google Scholar 

  81. Zhang J, Thomas AG, Leybovich E. Vaginal douching and adverse health effects: a meta-analysis. Am J Public Health. 1997;87(7):1207–11.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  82. Zhang M, Shi M, Zhao Y. Association between serum copper levels and cervical cancer risk: a meta-analysis. Biosci Rep. 2018;38(4):161.

    Article  Google Scholar 

  83. Zhang X, Dai B, Zhang B, Wang Z. Vitamin A and risk of cervical cancer: a meta-analysis. Gynecol Oncol. 2012;124(2):366–73.

    Article  CAS  PubMed  Google Scholar 

  84. Wootton RE, Richmond RC, Stuijfzand BG, Lawn RB, Sallis HM, Taylor GMJ, et al. Evidence for causal effects of lifetime smoking on risk for depression and schizophrenia: a Mendelian randomisation study. Psychol Med. 2019;50(14):2435–43.

  85. Barban N, Jansen R, de Vlaming R, Vaez A, Mandemakers JJ, Tropf FC, et al. Genome-wide analysis identifies 12 loci influencing human reproductive behavior. Nat Genet. 2016;48(12):1462–72.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  86. Karlsson Linnér R, Biroli P, Kong E, Meddens SFW, Wedow R, Fontana MA, et al. Genome-wide association analyses of risk tolerance and risky behaviors in over 1 million individuals identify hundreds of loci and shared genetic influences. Nat Genet. 2019;51(2):245–57.

    Article  PubMed  Google Scholar 

  87. Bentham J, Morris DL, Graham DSC, Pinder CL, Tombleson P, Behrens TW, et al. Genetic association analyses implicate aberrant regulation of innate and adaptive immunity genes in the pathogenesis of systemic lupus erythematosus. Nat Genet. 2015;47(12):1457–64.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  88. Clarke TK, Adams MJ, Davies G, Howard DM, Hall LS, Padmanabhan S, et al. Genome-wide association study of alcohol consumption and genetic overlap with other health-related traits in UK Biobank (N=112 117). Mol Psychiatry. 2017;22(10):1376–84.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  89. NealeLab. UK Biobank GWAS Results. 2019. Available from: http://www.nealelab.is/uk-biobank/.

  90. Stanley M. Immune responses to human papillomavirus. Vaccine. 2006;24(Suppl 1):S16–22.

    Article  PubMed  Google Scholar 

  91. Tam LS, Chan PK, Ho SC, Yu MM, Yim SF, Cheung TH, et al. Natural history of cervical papilloma virus infection in systemic lupus erythematosus - a prospective cohort study. J Rheumatol. 2010;37(2):330–40.

    Article  PubMed  Google Scholar 

  92. Dhar JP, Kmak D, Bhan R, Pishorodi L, Ager J, Sokol RJ. Abnormal cervicovaginal cytology in women with lupus: a retrospective cohort study. Gynecol Oncol. 2001;82(1):4–6.

    Article  CAS  PubMed  Google Scholar 

  93. Bateman H, Yazici Y, Leff L, Peterson M, Paget SA. Increased cervical dysplasia in intravenous cyclophosphamide-treated patients with SLE: a preliminary study. Lupus. 2000;9(7):542–4.

    Article  CAS  PubMed  Google Scholar 

  94. Blumenfeld Z, Lorber M, Yoffe N, Scharf Y. Systemic lupus erythematosus: predisposition for uterine cervical dysplasia. Lupus. 1994;3(1):59–61.

    Article  CAS  PubMed  Google Scholar 

  95. Nath R, Mant C, Luxton J, Hughes G, Raju KS, Shepherd P, et al. High risk of human papillomavirus type 16 infections and of development of cervical squamous intraepithelial lesions in systemic lupus erythematosus patients. Arthritis Rheum. 2007;57(4):619–25.

    Article  CAS  PubMed  Google Scholar 

  96. Ognenovski VM, Marder W, Somers EC, Johnston CM, Farrehi JG, Selvaggi SM, et al. Increased incidence of cervical intraepithelial neoplasia in women with systemic lupus erythematosus treated with intravenous cyclophosphamide. J Rheumatol. 2004;31(9):1763–7.

    CAS  PubMed  Google Scholar 

  97. Klumb EM, Araújo ML Jr, Jesus GR, Santos DB, Oliveira AV, Albuquerque EMN, et al. Is higher prevalence of cervical intraepithelial neoplasia in women with lupus due to immunosuppression? J Clin Rheumatol. 2010;16(4):153–7.

  98. Smitten AL, Simon TA, Hochberg MC, Suissa S. A meta-analysis of the incidence of malignancy in adult patients with rheumatoid arthritis. Arthritis Res Ther. 2008;10(2):R45.

    Article  PubMed  PubMed Central  Google Scholar 

  99. Love T, Solomon DH. The relationship between cancer and rheumatoid arthritis: still a large research agenda. Arthritis Res Ther. 2008;10(3):109.

    Article  PubMed  PubMed Central  Google Scholar 

  100. Breshears LM, Edwards VL, Ravel J, Peterson ML. Lactobacillus crispatus inhibits growth of Gardnerella vaginalis and Neisseria gonorrhoeae on a porcine vaginal mucosa model. BMC Microbiol. 2015;15(1):276.

    Article  PubMed  PubMed Central  Google Scholar 

  101. Gong Z, Luna Y, Yu P, Fan H. Lactobacilli inactivate Chlamydia trachomatis through lactic acid but not H2O2. PLoS ONE. 2014;9(9):e107758.

    Article  PubMed  PubMed Central  Google Scholar 

  102. Graver MA, Wade JJ. The role of acidification in the inhibition of Neisseria gonorrhoeae by vaginal lactobacilli during anaerobic growth. Ann Clin Microbiol Antimicrob. 2011;10(1):8.

    Article  PubMed  PubMed Central  Google Scholar 

  103. Silins I, Ryd W, Strand A, Wadell G, Tornberg S, Hansson BG, et al. Chlamydia trachomatis infection and persistence of human papillomavirus. Int J Cancer. 2005;116(1):110–5.

    Article  CAS  PubMed  Google Scholar 

  104. Grieshaber SS, Grieshaber NA, Miller N, Hackstadt T. Chlamydia trachomatis causes centrosomal defects resulting in chromosomal segregation abnormalities. Traffic. 2006;7(8):940–9.

    Article  CAS  PubMed  Google Scholar 

  105. Borgdorff H, Gautam R, Armstrong SD, Xia D, Ndayisaba GF, van Teijlingen NH, et al. Cervicovaginal microbiome dysbiosis is associated with proteome changes related to alterations of the cervicovaginal mucosal barrier. Mucosal Immunol. 2016;9(3):621–33.

    Article  CAS  PubMed  Google Scholar 

  106. Mitra A, MacIntyre DA, Paraskevaidi M, Moscicki AB, Mahajan V, Smith A, et al. The vaginal microbiota and innate immunity after local excisional treatment for cervical intraepithelial neoplasia. Genome Med. 2021;13(1):176.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  107. Mitra A, MacIntyre DA, Ntritsos G, Smith A, Tsilidis KK, Marchesi JR, et al. The vaginal microbiota associates with the regression of untreated cervical intraepithelial neoplasia 2 lesions. Nat Commun. 2020;11(1):1999.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  108. Muñoz N, Franceschi S, Bosetti C, Moreno V, Herrero R, Smith JS, et al. Role of parity and human papillomavirus in cervical cancer: the IARC multicentric case-control study. Lancet. 2002;359(9312):1093–101.

    Article  PubMed  Google Scholar 

  109. Castellsague X, Munoz N. Chapter 3: Cofactors in human papillomavirus carcinogenesis--role of parity, oral contraceptives, and tobacco smoking. J Natl Cancer Inst Monogr. 2003;(31):20–8.

  110. Smith JS, Green J, Berrington de Gonzalez A, Appleby P, Peto J, Plummer M, et al. Cervical cancer and use of hormonal contraceptives: a systematic review. Lancet. 2003;361(9364):1159–67.

    Article  PubMed  Google Scholar 

  111. Whelan E, Kalliala I, Semertzidou A, Raglan O, Bowden S, Kechagias K, et al. Risk factors for ovarian cancer: an umbrella review of the literature. Cancers (Basel). 2022;14(11):2708.

    Article  CAS  PubMed  Google Scholar 

  112. Burgess S, Davies NM, Thompson SG. Bias due to participant overlap in two-sample Mendelian randomization. Genet Epidemiol. 2016;40(7):597–608.

    Article  PubMed  PubMed Central  Google Scholar 

  113. Tainio K, Athanasiou A, Tikkinen KAO, Aaltonen R, Cardenas J, Hernandes, et al. Clinical course of untreated cervical intraepithelial neoplasia grade 2 under active surveillance: systematic review and meta-analysis. BMJ. 2018;360:k499.

  114. Skorstengaard M, Lynge E, Suhr J, Napolitano G. Conservative management of women with cervical intraepithelial neoplasia grade 2 in Denmark: a cohort study. BJOG. 2020;127(6):729–36.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

This work was supported by Wellcome Trust Imperial 4i/NIHR BRC clinical PhD fellowship (P77712 to S.B.) and the Academy of Finland (to IK).

Funding

This work was supported by Wellcome Trust Imperial 4i/NIHR BRC Clinical PhD Fellowship (P77712 to S.B.) and the Academy of Finland (to I.K.)

Author information

Author notes

  1. Sarah J. Bowden and Triada Doulgeraki are joint first authors.

  2. Ilkka Kalliala and Maria Kyrgiou are joint senior authors.

Authors and Affiliations

  1. Department of Metabolism, Digestion and Reproduction and Department of Surgery and Cancer, Institute of Reproductive and Developmental Biology, Faculty of Medicine, Imperial College London, Hammersmith Hospital campus, London, W12 0HS, UK

    Sarah J. Bowden, Antonios Athanasiou, Harriet Grout-Smith, Konstantinos S. Kechagias, Laura Burney Ellis, Ilkka Kalliala & Maria Kyrgiou

  2. Queen Charlotte’s and Chelsea – Hammersmith Hospital, Imperial College Healthcare NHS Trust, London, UK

    Sarah J. Bowden, Triada Doulgeraki, Laura Burney Ellis & Maria Kyrgiou

  3. Department of Hygiene and Epidemiology, University of Ioannina School of Medicine, Ioannina, Greece

    Emmanouil Bouras, Georgios Markozannes & Konstantinos K. Tsilidis

  4. Department of Epidemiology and Biostatistics, School of Public Health, Imperial College London, London, UK

    Georgios Markozannes, Verena Zuber, Marc Chadeau-Hyam & Konstantinos K. Tsilidis

  5. Department of Surgery and Cancer, Institute of Reproductive and Developmental Biology, Faculty of Medicine, Imperial College London, London, UK

    James M. Flanagan

  6. Department of Obstetrics and Gynaecology, University of Helsinki and Helsinki University Hospital, Helsinki, Finland

    Ilkka Kalliala

Authors
  1. Sarah J. Bowden
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Triada Doulgeraki
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Emmanouil Bouras
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Georgios Markozannes
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Antonios Athanasiou
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Harriet Grout-Smith
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Konstantinos S. Kechagias
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Laura Burney Ellis
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Verena Zuber
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Marc Chadeau-Hyam
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. James M. Flanagan
    View author publications

    You can also search for this author in PubMed Google Scholar

  12. Konstantinos K. Tsilidis
    View author publications

    You can also search for this author in PubMed Google Scholar

  13. Ilkka Kalliala
    View author publications

    You can also search for this author in PubMed Google Scholar

  14. Maria Kyrgiou
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

SB, IK and MK conceived of the study. SB, IK and MK wrote the protocol. SB and IK developed the search strategy and ran the search. TD, AA, HGS and SB performed the searches and data extraction. KSK and LE performed the quality assessment. IK, GM and EB performed the statistical analyses. All authors (SB, TD, GM, EB, KSK, LE, VZ, KT, JF, MCH, IK, MK) contributed to the data analysis and drafting of the manuscript. All authors reviewed the final manuscript.

Corresponding author

Correspondence to Sarah J. Bowden.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1: Supplementary Methods.

Additional details including electronic search terms and data analysis pipeline of Umbrella review.

Additional file 2: Supplementary Table 1.

Statistical criteria for grading of the evidence.

Additional file 3: Supplementary Table 2.

Summary quality assessment for all included systematic reviews using the ASMTAR 2 tool (A Measurement Tool to Assess Systematic Reviews 2).

Additional file 4: Supplementary Table 3.

Description of 87 meta-analyses investigating risk factors associated with HPV, cervical pre cancer and cancer outcomes - only cohort studies included.

Additional file 5: Supplementary Table 4.

Evaluation of heterogeneity, small study effects and excess significance bias in the 87 meta-analyses investigating the risk factors associated with HPV, cervical pre cancer and cancer outcomes - only cohort studies included.

Additional file 6: Supplementary Table 5.

Details of evidence grading for meta-analysis of risk factors for HPV, cervical precancer and cancer outcomes – _only cohort studies included*.

Additional file 7: Supplementary Table 6.

Sensitivity analysis using credibility ceilings when the association is non-significant of the 87 studies investigating the risk factors associated with HPV, cervical pre cancer and cancer outcomes - only cohort studies included.

Additional file 8: Supplementary Table 7.

Details of evidence grading for meta-analyses of risk factors for HPV, pre-invasive and invasive cervical cancer incidence, progression, regression, or mortality - all study types included (sensitivity analysis).

Additional file 9: Supplementary Table 8.

Excluded duplicate studies and studies selected in their place to be included in the analysis.

Additional file 10: Supplementary Table 9.

Mendelian randomisation (MR) analysis; exposure-outcome pairs included in the main analysis, cohorts only.

Additional file 11: Supplementary Table 10.

Main GWAS study characteristics.

Additional file 12: Supplementary Table 11.

Two-sample inverse variance weighted mendelian randomization full results of the analyses of risk factors on cervical cancer.

Additional file 13: Supplementary Table 12.

Results from Mendelian randomisation sensitivity analyses of risk factors on cervical cancer. Supplementary Table 13. Results from multivariable (MV) Mendelian randomisation (MR) sensitivity analyses of risk factors on cervical cancer.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Bowden, S.J., Doulgeraki, T., Bouras, E. et al. Risk factors for human papillomavirus infection, cervical intraepithelial neoplasia and cervical cancer: an umbrella review and follow-up Mendelian randomisation studies. BMC Med 21, 274 (2023). https://doi.org/10.1186/s12916-023-02965-w

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12916-023-02965-w

Keywords

BMC Medicine

ISSN: 1741-7015

Contact us