We assessed the risk of bias (high, moderate, or low) in each included study following EPOC criteria (for observational studies) and the Cochrane Handbook for Systematic Reviews of Interventions (24). For each RCT, we examined sequence generation, allocation concealment, blinding, incomplete outcome data, selective outcome reporting, and other potential biases. For quasi-randomized trials, we used the GRADE (Grades of Recommendation, Assessment, Development, and Evaluation) “risk of bias” framework, which is reported in the eligibility criteria, exposure measurement, outcome, confounding, and attrition rates (25).

For observational studies, we used the Dawn and Black quality assessment tool. This tool contains items such as clarity and external and internal validity (bias, exposure, and confounding) and culminates in a numerical score (26, 27).

The unit of analysis was the individual participants in each study. The analysis of this review was limited to the analytical method used in the trial report (e.g., intent to treat, per protocol, or a modification of either type). Studies were combined for meta-analysis only when identical modern contraceptive devices, tools, or drugs; dosages; and regimens could be compared. We calculated the rate ratios (RR) and hazard ratios (HR) with a 95% confidence interval (95% CI) for each dichotomous outcome. We created forest plots for three-time points when possible: less than 5 years, more than 5 years, more than 10 years. If outcomes were reported for multiple reference periods, we reported the outcomes for the first and last reference periods only.

Subgroup analysis was conducted by comparing different types of contraception, dose, and route of administration when possible. We also conducted a sensitivity analysis to test the robustness of any results that appeared to be based on heterogeneous combinations by examining the effect of deleting each study. Finally, sensitivity analyses were conducted based on rates of loss to follow-up and any study with a rate of loss to follow-up over 20% was excluded.

We conducted a meta-analysis if two data points or more presented themselves for each comparison and outcome. We visually examined heterogeneity by comparing study designs, target populations, and primary outcome measures across included studies. We assessed the homogeneity of trials combined in a meta-analysis using both fixed-effects and random-effects models. The classical measure of heterogeneity is Cochran's Q, which was calculated as the weighted sum of squared differences between individual study effects and the pooled effect across studies. Q is distributed as a chi-square statistic, and the alpha level is set at 0.10. We then used the I² score to identify the magnitude of heterogeneity. Any score of I² above 50% was investigated for the clinical and methodological diversity of the studies.

GRADE profiler (GRADEpro 2020) was used to import data from Review Manager (Revman) 5.3 to create a “Summary of findings” table. A summary of the intervention effect and a measure of quality for each of the above outcomes was produced using the GRADE approach, which involves five considerations (study limitations, consistency of effect, imprecision, indirectness, and publication bias) to assess the quality of the evidence for each outcome.

Figure 2 reflects the quality assessment of included studies and is presented below.

Supplementary Appendix 2 -Table S3 shows the quality assessment of observational studies using Dawn and Black scoring system. We considered the overall quality of evidence to be moderate for our review (mean: 13.24 ± 2.25, median = 14 min = 5, max = 16).

Contraceptive users had a significantly lower risk of endometrial cancer incidence when compared to non-users (0.56, 95% CI 0.50, 0.63). We identified six studies on oral contraceptives (33, 35, 37, 40, 41, 57), two studies on hormonal IUS (36), one study on progestin-only products (37), and one study for any contraceptives (oral and non-oral) (37). The rate ratio for oral contraceptive users was 0.53 (95% CI 0.45, 0.63), for hormonal IUS 0.58 (95% CI 0.42, 0.79), for progestin-only products 0.73 (95% CI 0.23, 2.32) and for any contraception 0.61 (95% CI 0.49, 0.76). All hormonal contraception methods had a significant downward trend for endometrial incidence, except progestin-only products had no significant results (Figure 5).

Three studies investigated the associated risk incidence of endometrial cancer by comparing varying dosing regimens to non-users (37, 43, 44).

A significant protective effect against endometrial cancer incidence was observed among the following dosing regimens: Gestodene 20–35 mg ethinylestradiol (RR 0.27, 95% CI 0.12–0.61), LNG-IUS (Supplementary Appendix 3 - Figure S2).

Progestogen Only Non-oral (RR 0.22, 95% CI 0.13–0.37), Alpha-hydroxyprogesterone Caproate and 5 mg Estradiol Valerate (RR 0.27, 95% CI 0.12–0.61), and any combined oral pills (RR 0.66, 95% CI 0.48–0.91). A different cohort with 116, 429 females observed protective effects of estrogen or progestin formulations against endometrial cancer (HR 0.77; 95% CI 0.69–0.85) (60) (Figure 8).

Cervical cancer was studied by six studies on oral contraceptives (33, 35, 38, 40, 41, 57), one study on hormonal IUS (38), and one study for any contraceptives (oral and non-oral) (38) (Figure 7). The rate ratio of cervical cancer was higher among non- users compared to hormonal contraceptive users when we pooled the results (1.28, 95% CI 1.21, 1.35). When stratifying the results, a significant increased risk was observed for OCP users (RR = 1.43, 95% CI 1.31, 1.56) and any users of oral and non-oral (RR = 1.19, 95% CI 1.10–1.29), while women using IUS showed no significant effect on cervical cancer risk incidence (RR = 0.76, 95% CI 0.52–1.11). In a comparison between non-users and hormonal IUD users, the risk ratio was 0.58(0.42, 0.79). Any contraceptive users reported one study only (37). These results had high heterogeneity (I² = 80%) (Figure 6).

When comparing dosing regimens, a nonsignificant downward trend in cervical cancer incidence was observed by Rosenblatt et al., 2007 among Hydroxyprogesterone Caproate and 5 mg Estradiol Valerate users (RR 0.54, 95% CI 0.07–4.17), while an upward trend was observed by Iversen et al., 2017 for any combined oral contraceptive users (RR 1.31, 95% CI 0.82–1.94).

The association between gynecological cancer and the protective effect of tubal ligation was reported by four studies (47–50). Tubal ligation reduces the hazards of gynecological cancer risk by 20% (HR = 0.80, 95% CI 0.66–0.87) (Supplementary Appendix 3 - Figure S3).

Three studies assess the hazards of gynecological cancer among OCP users and non-users (22, 47, 60) (Figure 7). Compared to non-users, OCP use reduces the hazard of gynecological cancers of the ovaries and endometrium by 17% (HR 0.83; 95% CI 0.74–0.93) (Figure 8).

Four studies assessed the association between hormonal contraception usage and cancer mortality (42, 51–53). Compared to non-users, a reduced incidence of gynecological cancer mortality was observed among HC users (RR 0.60, 95% CI 0.51–0.71). When stratifying the results by gynecological cancer type, a 63% increase in cervical cancer mortality was observed among HC users (RR 1.63, 95% CI 1.07–2.49). Similarly, a cohort study containing 121,577 participants found a slight protective effect when comparing ever users of OCs to never users (HR 0.87; 95% CI 0.77–0.98) (59) (Supplementary Appendix 3 - Figure S4).

Breast cancer was studied by four studies on oral contraceptives (35, 41, 57), and one study on hormonal IUS (43) (Figure 8). A pooled non-significant rate ratio was estimated (1.00, 95% CI 0.94, 1.06) for OCP and IUS users. Dosing regimens comparing HC users to non-users found no significant difference in breast cancer incidence among Hydroxyprogesterone Caproate and 5 mg Estradiol Valerate, LNG-IUS Progestogen-Only Non-Oral, and any Combined Oral Pill users with a pooled effect of 1.02 (95% CI 0.93–1.11) (35, 43, 44) (Figure 8).

Only one study observed the association between tubal ligation and breast cancer risk (47). No significant effect was observed among women who underwent tubal ligation and those who have not (RR 1.15, 95% CI 0.86–1.54). Three studies investigated regimen dosing to understand breast cancer risk among HC users (35, 43, 44). Compared to non-users, no significant difference in breast cancer incidence was observed among women using alpha- hydroxyprogesterone caproate and 5 mg estradiol valerate, LNG-IUS Progestogen-Only Non- Oral, and any combined oral pills (RR 1.02, 95% CI 0.93–1.11).

Additionally, breast cancer mortality was not significantly different between OCP users and non-users (1.00, 95% CI 0.89, 1.12) (Supplementary Appendix 3 - Figure S5).

Ovarian cancer risk among BRCA 1 and 2 mutation carriers using OCs was investigated within two cohort studies including 9,753 women (54, 58) (Figure 9). Among BRCA 1 and 2 mutation carriers, ever users of oral contraceptives are associated with a decreased risk of ovarian cancer (HR = 0.55, 95% CI 0.44, 0.69) (Figure 9).

Among BRCA 1 and 2 carriers using OCs for less than 5 years, no significant effect was observed (HR = 0.76; 95% CI 0.57, 1.02) Supplementary Appendix 3 - Figure S6.

A significant protective effect against ovarian cancer is observed among BRCA 1 and 2 Carriers who use oral contraceptives between 4 and 9 years (HR = 0.45; 95% CI 0.34, 0.61) (Supplementary Appendix 3 - Figure S7).

For BRCA 1 carriers, OCPs have a protective effect, while no difference in effect is detected among BRCA 2 carriers. BRCA 1 and 2 mutation carriers using oral contraceptives for greater than 10 years have a significant protective effect against the risk of ovarian cancer (HR = 0.37; 95% CI 0.26, 0.54). (Supplementary Appendix 3 - Figure S8).

The risk of breast cancer association from OC use was observed among 11,432 BRCA 1 or 2 gene mutation carriers with two cohort studies conducted in the United Kingdom and Netherlands (55, 56) (Figure 10). Among BRCA 1 and 2 mutation carriers, ever users of oral contraceptives are associated with an increased risk of breast cancer compared to never users (HR = 1.39, 95% CI 1.15, 1.67).

Among BRCA 1 and 2 carriers using OCs for less than 5 years, an increased risk in breast cancer was observed compared to never users (HR = 1.33; 95% CI 1.07, 1.66) (Supplementary Appendix 3 - Figure S9).

A significant increase in risk of breast cancer is observed among BRCA 1 and 2 Carriers who use oral contraceptives between 4 and 9 years compared to never users (HR = 1.29; 95% CI 1.03, 1.61), however, the risk between carriers using OCs for less than 5 years and between 4 and 9 years are very similar as you compare them (Supplementary Appendix 3 - Figure S10).

BRCA 1 and 2 mutation carriers using oral contraceptives for greater than 10 years have an increased risk of breast cancer (HR = 1.47; 95% CI 1.18, 1.85). Long term OC use of greater than 10 years has a higher risk for mutation carriers as compared to those using them for less than 10 years (Supplementary Appendix 3 - Figure S11).

Duration of hormonal contraception use among a sample of over 21 million women were accounted for in 6 cohort studies measuring the rate ratio of gynecological cancers among users and non-users to examine a dose-response relationship (10, 37, 39–41, 57) (Supplementary Appendix 3 - Figures S12, S13).

Among short-term hormonal contraception users who used HCs for less than 5 years, a significant protective effect was observed for ovarian (RR = 0.63; 95% CI 0.52, 0.75) and endometrial cancer (RR = 0.75; 95% CI 0.59, 0.94) outcomes, however, a harmful effect was observed for cervical cancer (RR = 2.55; 95% CI 1.17, 5.59) (Supplementary Appendix 3 - Figure S12).

Among long-term HC users who used HCs for at least 8 years or more, a similar pattern to short-term users was observed. A significant reduction in ovarian (RR = 0.36; 95% CI 0.27, 0.49) and endometrial cancers (RR = 0.26; 95% CI 0.19, 0.36) was observed. A harmful effect was observed for cervical cancer, and this risk increased from the initial risk that short-term users have (RR = 2.55 vs. 5.40; 95% CI 2.89, 10.89) (Supplementary Appendix 3 - Figure S13).

There were three cohort studies with 275,988 participants examined for the association between duration of HC use and breast cancer incidence (Supplementary Appendix 3 - Figures S14, S15) (40, 41, 57). No significant relationship among short-term and long-term HC users was observed for breast cancer incidence.

There are significant reductions in the rate of ovarian and endometrial cancer incidence and mortality among contraceptive users. Short- and long-term methods showed a protective effect against ovarian and endometrial cancers. When comparing dosing regimens, HCs containing both progestins and estrogens have protective effects against ovarian and endometrial cancer. Conversely, HC users are at an increased risk of cervical cancer incidence and mortality, and this risk persists regardless of duration. Among healthy women with no genetic mutations, no significant difference in breast cancer incidence and mortality was reported among users and non-users of HC. Among mutation carriers (BRCA1/2) using OCs, an increased risk in breast cancer incidence was observed and this risk persists among short- and long-term users. For ovarian and cervical cancer, our case-control findings matched our cohort studies, however, for breast cancer, a significant risk was observed among OCP users.

Contraceptive use as a primary exposure for our review was self-reported in a majority of our studies. Recall bias and information bias are a major concern for studies that relied on self- reports of contraceptive use as this can potentially underestimate the true effect of contraceptive use on our outcomes of interest. Future studies should focus on techniques such as recording the on-time injection checklist or electronic pill count.

Highlighting the mechanism of action in each scenario outlined above is crucial. For example, the inhibition of ovulation is a key factor in providing protection against ovarian cancer, shedding light on why specific methods like DMPA may not confer the same level of protection as COCs. Nevertheless, it's important to note that this manuscript, while acknowledging the importance of theoretical mechanisms, refrains from delving extensively into the intricacies beyond the causality of the effects on different types of cancer.

Our results on contraceptives as a protective effect against ovarian cancer are confirmatory to the initial findings of Havrilesky 2013 (61). A previous study (Pragout et al.) reported an insignificant increased risk of cervical cancer, while our study found a significantly (28%) higher risk among users than non-users (62). In the systematic review conducted by Asthana, the overall odds ratio of invasive cancer on OC use was estimated as 1.51 (95% CI 1.35, 1.68) with 19 included studies, which is not very different from our study (63). In Pragout et al., a slight but significant increase in breast cancer risk among users was observed; however, in our study, no significant differences were observed between users and non-users, except for women who are BRCA1/2 carriers. A study (Conz et al.) systematically reviewed the relationship between use of the Levonorgestrel-releasing intrauterine system (LNG-IUS) and breast cancer risk with 26 studies. They found an increased risk in LNG-IUS users (pooled OR = 1.16, 95% CI 1.06–1.28) (64). However, we did not detect a significant risk in LNG-IUS users. Our study supports the previous findings from Pragout 2018 on the significant reduction of endometrial cancer. However, there is conflicting information about the roles IUD and tubal ligation play in the risk factor. For BRCA1/2 mutation carriers, Huber examined OC users’ risk of ovarian cancer and breast cancer compared to non-users in a systematic review (65). It reported a risk reduction in ovarian cancer, and a significant risk elevation in breast cancer, which is aligned with our findings.

Our study contributes nuanced insights into the complex relationship between contraceptive use and various types of cancer, offering both confirmatory evidence and novel findings that warrant further exploration and validation in future research endeavors.

Our disagreement with other studies can also be explained by the adopted methodology that we employed a comprehensive approach by searching across 19 databases to ensure inclusivity. Papers were included without geographical or language restrictions. The population of interest encompassed women of reproductive age (14–49 years). Interventional studies were included if they investigated modern contraceptive methods endorsed by the WHO, such as short-acting hormonal contraception, long-term contraception, one-time barrier contraception, permanent contraception, and emergency contraception. Observational studies considered contraceptive use of all types as the main exposure, excluding those investigating contraception alongside other medications or modalities. Comparison groups consisted solely of non-users of contraception. The primary outcome of interest was mortality and morbidity related to reproductive tract cancers. Eligible study designs included parallel, or cluster randomized controlled trials, controlled clinical trials, controlled before and after studies, interrupted time series studies, cohort or longitudinal analyses, regression discontinuity designs, and case-control studies.

First, we searched the databases for all studies from their inception. Thus, we included studies with both outdated and up-to-date contraceptive methods, with broad coverage. Second, for each outcome category, our studies covered diseases and conditions commonly seen in women of reproductive age. For some of these outcomes, we have studies that yielded large pooled sample sizes and low loss to follow-up rates, allowing for higher statistical power, narrower confidence intervals, and more credible results. The prospective study design eliminated the possible influence of differential recall of contraception use.

Cancer is a difficult outcome to measure as there is a long latency period between the time of exposure and the detection of an outcome. During these long latency periods, the methods and dosing regimens a woman uses can change over time. Our study did not account for the age at first use of contraceptives, time since last use, environmental exposures, or demographic characteristics. Included studies lacked information on specific HC regimens and dosing and HC methods (patch/ring, implants, and injectables) were often left out of analysis due to a lack of available data. There were limited studies examining the association between varying HC methods and dosing regimens for cervical and breast cancer outcomes. Most studies were on OC users, with limited studies looking at the association of long-term contraceptive types. As long-term contraceptive methods become more popular, it is essential to have more information on these methods. Heterogeneity was a significant issue in this study.

This review provides new insights into the relationship between modern contraceptive use and cancer risk, focusing on the effects of birth control pills, and long-acting reversible contraceptives. We found a significant reduction in the risk of developing ovarian and endometrial cancers among contraceptive users, highlighting the potential benefits of these contraceptive methods. However, the review also reveals an increased risk of cervical cancer among contraceptive users, which persists regardless of the duration of use. These findings suggest the need for additional research into the link between hormonal contraceptive use and cancer incidence. The authors recommend that healthcare providers consider various factors such as age, physical and mental status, family history, and individual preferences when selecting contraceptive methods to maximize benefits and minimize risks.

The findings of this research offer a comprehensive body of evidence on the utilization of contraception and its impact on various stakeholders, such as healthcare professionals, pharmaceutical companies, pharmacists, and others. The results indicate that aside from its primary role as a contraceptive, contraception may also have preventive or therapeutic properties for certain diseases or conditions. Nevertheless, the study highlights potential adverse effects that must be considered. Healthcare providers should evaluate several situational factors, such as age, physical and mental health status, and family history, to choose the most suitable contraceptive method for their female patients, ensuring maximum benefits and minimal risks. Recommendations include discussions about the woman's overall health and changes to medical history, continued monitoring and a complete physical examination at each follow-up visit checking for symptoms of blood pressure changes, weight changes, headaches, or abnormal vaginal bleeding (66).