Reproductive Health Outcomes among Adolescent and Young Adult Cancer Patients: A Systematic Review and Meta-Analysis

Simple Summary Reproductive health is an important consideration for adolescents and young adults (AYAs, aged 15–39 years) with cancer. Our systematic review and meta-analysis synthesized the current literature on the impacts of AYA cancer on reproductive health outcomes. We searched EMBASE and MEDLINE from 1 January 2000 to 26 January 2022 to capture observational studies exploring impacts of AYA cancer on reproductive health outcomes compared to controls. A total of 21 studies were included, and 62 outcomes were explored across all studies. We classified these outcomes in a sex-based framework as fetal/neonatal (n = 26), maternal (n = 11), fetal/neonatal-maternal (n = 23), and maternal-paternal (n = 2). Our analysis showed significantly higher preterm birth, gestational diabetes, and use of fertility treatment in AYA cancer patients compared to controls. Additionally, there is a higher risk of preterm birth and low APGAR score at birth for AYA cancer patients who receive radiation compared to controls. This review provides evidence of impacts of AYA cancer on reproductive health outcomes. Abstract Background: This systematic review and meta-analysis (SRMA) aimed to synthesize the current literature on the impacts of adolescent and young adult (AYA, ages 15–39 years) cancer on reproductive health outcomes. Methods: EMBASE and Medline were searched from 1 January 2000 to 26 January 2022 for observational studies that included individuals with AYA cancer and controls which evaluated reproductive health outcomes. We used random effects models and 95% confidence intervals to obtain pooled measures of associations between AYA cancer, cancer treatment, and reproductive health outcomes. Results: The search identified 8625 articles; 21 were included. 62 reproductive outcomes were assessed and classified according to a sex-based framework as fetal/neonatal (n = 26), maternal (n = 11), fetal/neonatal-maternal (n = 23), and maternal-paternal (n = 2). Meta-analyses of crude estimates showed significant associations between AYA cancer and outcomes including preterm birth (pooled odds ratio [pOR] 1.31; 95% CI: 1.22, 1.42), gestational diabetes (pOR 1.43; 95% CI: 1.03, 1.99), and fertility treatment (pOR 2.66; 95% CI 1.71, 4.11). We also found higher odds of preterm birth (pOR 1.65; 95% CI: 1.21, 2.26) and low APGAR score at birth (pOR 2.03; 95% CI: 1.32, 3.13) among AYA cancer patients who received radiation compared to controls. Conclusions: Our SRMA quantified impacts of AYA cancers and treatments on several reproductive health outcomes.


Introduction
The incidence of cancer diagnosed among adolescents and young adults (AYAs), that is individuals between 15 and 39 years, is increasing at an alarming rate, with studies reporting a nearly 30% increase from 1973 to 2015 [1]. The impacts of cancer and its treatment on endocrine and reproductive organ function often persist long after diagnosis and treatment [2,3]. As such, these impacts can lead to long-term psycho-oncologic challenges across multiple domains including mental [4][5][6][7], psychosocial [4,5], and reproductive and sexual health [8,9]. Due to improvements in treatments translating to higher remission rates [10], reproductive health has become an important consideration among individuals diagnosed with cancer during adolescence/young adulthood as they consider family planning.
Despite the growing literature on reproductive health outcomes in AYA cancer patients-who we define as individuals across the care continuum from diagnosis to treatment to survivorship [11]-limited synthesis precludes understanding of current evidence and knowledge gaps. In 2018, Gerstl et al. conducted a systematic review and meta-analysis of 17 studies evaluating reproductive health outcomes in females diagnosed with cancer between 0 and 25 years of age [12]. Pooled analyses showed 79% of female cancer patients experienced a live birth, of which 22% were preterm births. Moreover, females who received chemotherapy alone had a pooled estimated rate of 18% of experiencing a live birth compared with 10% of females who received radiation alone. Overall, stillbirth rates were similar for cancer patients aged 0-25 (0.01%; 95% CI: 0.00, 0.002) and controls (0.01%; 95% CI: 0.006, 0.01). Low birthweight (<2500 g) was slightly higher in cancer patients aged 0-25 (10%; 95% CI: 0.09, 0.11) than controls (6%; 95% CI: 0.05, 0.07). Despite these findings, limitations related to sex (including only females) and age (individuals 0 to 25 years) of this prior systematic review limit the ability to extrapolate findings to all AYA cancer patients [13]. To expand on, comprehensively assess, and quantify the impacts of AYA cancer on reproductive health outcomes, we conducted a systematic review and meta-analysis to synthesize reproductive health outcomes evaluated in both male and females across the entire age range of AYA cancer patients (i.e., 15 to 39 years).

Search Methods
We conducted a systematic review and meta-analysis that adhered to the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols 2020 guidelines (PROS-PERO registration number: CRD42022313343) [14]. We incorporated principles of sex-and gender-based analyses (SGBA) [15,16] throughout conduct and reporting, which accounts for the influence of sex and gender on differences in health. In collaboration with a research librarian, we developed a literature search strategy to identify peer-reviewed, published manuscripts relating to the impact of AYA cancers on reproductive health outcomes (Tables 1 and 2). Searches were conducted in the following databases: (1) EMBASE Ovid and (2) Ovid MEDLINE(R) and Epub Ahead of Print, In-Process, In-Data-Review & Other Non-Indexed Citations, Daily and Versions. We used database-relevant terms and keywords mapping to the following concepts: (1) AYA age range (e.g., "young adult" OR "teen" OR "youth", etc.); (2) cancer and cancer treatment (e.g., "chemotherapy" OR "radiation" OR "cancer treatment", etc.); and (3) reproductive health outcomes (e.g., "reproductive health" OR "stillbirth" OR "preeclampsia", etc.). Limits were added to the search to restrict results to human studies published from 1 January 2000 to 26 January 2022. Bibliographies of included studies were hand searched for additional studies that met the criteria for this review. (Chemotherap* or "cancer treatment*" or radiation or brachytherap* or "antineoplastic agent*" or "antitumor* drug*" or "antitumor* agent*" or antineoplastics* or "anticancer* agent**" or "anticancer* drug*" or "early detection of cancer" or "oncolog* surger*").ti,ab,kw.

Study Screening and Inclusion
Search results were uploaded onto Covidence [17], where duplicates were automatically removed. The screening was completed by two reviewers (NO and MDV). In order to be eligible for inclusion, studies had to fulfill all of the following criteria: (1) used an observational study design; (2) primarily included individuals diagnosed with cancer from 15 to 39 years and a comparator (control) group of individuals without cancer (e.g., matching AYA cancer patients with controls in databases or national surveys based on sociodemographic and/or clinical factors); and (3) evaluated reproductive health outcomes (e.g., stillbirth, gestational diabetes, preterm birth, etc.). Studies that focused on pre-cancerous lesions or pregnancy-associated cancers (those diagnosed and/or treated during pregnancy) were excluded. To ensure a comprehensive capture of studies, we did not place limits on lower or upper end of age ranges but did require that the majority of study participants were between 15 to 39 years of age, which for our purposes, we defined a priori as ≥80%. No restrictions were placed on geography, language, or availability of full text.

Data Extraction and Quality Assessment
We extracted information on study characteristics (publication year, country, study design, data source, sample size, and follow-up timeline) and AYA cancer exposure (definition of exposure, type of cancer, type of treatment, inclusion/exclusion criteria, age at diagnosis, subgroup analyses, age at diagnosis, and age at study). Where feasible, we pooled reported age across studies using StataSE 17 [18] (e.g., for studies that reported mean and standard deviation). Of particular relevance to our SGBA-informed approach [15] is reported information on sex (i.e., a set of biological attributes traditionally associated with sex chromosome status) and/or gender (i.e., socially and culturally constructed roles) in included studies. Specifically, we extracted information on reported sex/gender variable(s) (e.g., sex, gender, both, and neither), corresponding groups (e.g., male/female and men/women), and definitions (where relevant/provided). Key to our systematic review and meta-analysis is reproductive health outcomes, which we define as outcomes relating to conditions of male and female reproductive systems during all life stages [19]. Aside from extracting information on reproductive health outcomes assessed, we further characterized these according to who is impacted by the outcome (mother, fetus/newborn, or father) and when the outcome was assessed (before pregnancy, during pregnancy, intrauterine, delivery, and after delivery). Finally, we extracted available measures such as counts, proportions, and rates of reproductive health outcomes and measures of associations (e.g., crude and/or adjusted odds ratio and relative risk).

Analysis
For our meta-analyses, we computed random effects models for reproductive health outcomes that were reported by at least two studies. This was accomplished by pooling, where reported, proportions of crude events reported in each study and obtaining crude odds ratios (OR) and respective 95% confidence intervals (CI). Where possible, we conducted stratified analyses to evaluate impacts of cancer treatments. As crude events were rarely reported according to type of cancer treatment, a generic inverse-variance approach was used to obtain pooled estimates. Heterogeneity was assessed using the chi-squared test, with p < 0.10 indicating significant heterogeneity as opposed to p < 0.05, as the test is low in power when studies have small sample sizes or are few in number [23]. As an added measure, we also used the I 2 test for inconsistency and interpreted it according to Cochrane's recommendations with (1) 0-40% indicating little to no heterogeneity; (2) 30-60% indicating moderate heterogeneity; (3) 50-90% indicating substantial heterogeneity; and (4) ≥75% indicating considerable heterogeneity [23]. Forest plots and funnel plots were constructed for all pooled analyses. All analyses were conducted using RevMan5 [24].

Search Results
Our search strategy resulted in 8625 original citations from 1 January 2000 to 26 January 2022 ( Figure 1). The main reasons for excluding 120 citations in full-text screening were: incorrect study design (n = 19); lack of a comparator group (n = 29); and participants not representative of AYA age range (n = 34). We also excluded studies that focused on pre-cancerous lesions or pregnancy-associated cancers (n = 22). Screening resulted in a total of 20 studies eligible for inclusion, and handsearching yielded one study, resulting in a total of 21 included studies.  [23]. Forest plots and funnel plots were constructed for all pooled analyses. All analyses were conducted using RevMan5 [24].

Study Characteristics
Altogether, studies included a total of 102,041 AYA cancer patients. Age is an important consideration; the majority of studies (n = 18) reported age at cancer diagnosis, and all reported age at the time of the study. However, we observed variation in the reporting of age, including mean and standard deviation or proportion according to varying age categories. The pooled AYA cancer age at diagnosis was 31.42 (95% CI: 29.49, 33.36), and pooled AYA cancer age at the time of study was 32.59 (95% CI: 31.09, 34.10). The majority of studies (n = 15) studied only females. On inspection, six of these studies conflated sex and gender terminology (i.e., authors would refer to sex but use female and woman interchangeably). Six studies included both females and males, and of these, three conflated sex and gender terminology (i.e., authors would refer to sex but use male/female and men/women interchangeably). Finally, quality assessment of included studies resulted in a "Good" ranking on all cohort studies, with scores ranging from 6 to 8, and a "Poor" ranking (score = 3) on the one cross-sectional study.

Reproductive Health Outcomes
Impacts of AYA cancer were reported on a total of 62 reproductive health outcomes across the 21 included studies. As all studies reported sex as male and female, we categorized outcomes according to who is impacted and when the outcome is assessed. This led to the development of a sex-based framework for conceptualizing reproductive health outcomes as: (1) fetal/neonatal outcomes affecting the fetus or baby and assessed intrauterine, at delivery, and after delivery (n = 26); (2) maternal outcomes affecting the birth mother (with cancer) and assessed before pregnancy, during pregnancy, and after delivery (n = 11); (3) fetal/neonatal-maternal outcomes that affect both fetus/baby and birth mother and assessed during pregnancy, delivery, and after delivery (n = 23); and (4) maternal-paternal outcomes that may affect either birth mother (with cancer) or birth father (with cancer) and assessed before pregnancy and after delivery (n = 2). Figure 2 illustrates this framework, and Table 4 lists all 62 extracted outcomes, corresponding studies, and reported crude and adjusted measures of association where available.   Cancer registry often linked to other databases for sociodemographic and/or perinatal information. b "Any" cancer includes but is not limited to: thyroid, breast, blood and leukemia, lymphoma, gynecologic (cervix, uterus, and ovary), intestines, gall bladder, pancreas, bone, soft tissue tumor of bone/fat, and/or skin. c The Newcastle-Ottawa Scale for assessing the quality of non-randomized studies. d The Surveillance, Epidemiology, and End Results (SEER) Program provides information on cancer statistics in the United States.
mother and assessed during pregnancy, delivery, and after delivery (n = 23); and (4) maternal-paternal outcomes that may affect either birth mother (with cancer) or birth father (with cancer) and assessed before pregnancy and after delivery (n = 2). Figure 2 illustrates this framework, and Table 4 lists all 62 extracted outcomes, corresponding studies, and reported crude and adjusted measures of association where available.

Meta-Analysis
There were 17 reproductive health outcomes that were evaluated by at least two or more included studies, thereby enabling a meta-analysis to obtain pooled measures of associations.

Maternal Outcomes (n = 3)
Meta-analysis was feasible for three maternal outcomes, with pooled results showing that preeclampsia and gestational diabetes were significantly associated with AYA cancer diagnosis, and gestational hypertension was not. The impact of AYA cancer (n = 9967) on preeclampsia compared to those without AYA cancer (n = 926,338) was explored by six studies, with meta-analysis yielding a pOR of 1.29 (95% CI: 1.01, 1.64) ( Figure 6A). There was evidence for substantial to considerable heterogeneity among studies (Chi-squared statistic: 15.66, p = 0.008; I 2 = 68%). The odds of gestational diabetes in AYA cancer patients (n = 9879) compared to controls (n = 787,751) was evaluated in six studies in this review; pooling showed that there are higher odds of gestational diabetes in AYA cancer patients (pOR 1.43; 95% CI: 1.03, 1.99). ( Figure 6B). Additionally, heterogeneity was substantial to considerable in this outcome (Chi-squared statistic: 28.88, p = <0.0001; I 2 = 83%). Lastly, meta-analysis did not indicate higher likelihood of gestational hypertension (pOR 1.51; 95% CI 0.46, 4.91) ( Figure 6C) in AYA cancer patients (n = 290) compared to controls (n = 78,338). Heterogeneity was substantial in this outcome (Chi-squared statistic: 2.60, p = 0.11; I 2 = 62%). Funnel plots for all maternal health outcomes were also generated and indicate various levels of publication bias ( Figure 5K-M).

Maternal-Paternal Outcomes (n = 1)
The likelihood of requiring fertility treatments in female and male AYA cancer patients (n = 50,358) compared to controls (n = 2,599,602) was explored by six studies (Figure 8), and there was a significantly higher odds of requiring fertility treatments in male and female AYA cancer patients compared to male and female controls (pOR 2.66; 95% CI 1.71, 4.11). Heterogeneity was considerable in this outcome as well (Chi-squared statistic: 318.71, p < 0.00001; I 2 = 98%). The funnel plot for this outcome indicates publication bias ( Figure 5Q).

Maternal-Paternal Outcomes (n = 1)
The likelihood of requiring fertility treatments in female and male AYA cancer pa tients (n = 50,358) compared to controls (n = 2,599,602) was explored by six studies (Figur 8), and there was a significantly higher odds of requiring fertility treatments in male an female AYA cancer patients compared to male and female controls (pOR 2.66; 95% CI 1.71 4.11). Heterogeneity was considerable in this outcome as well (Chi-squared statistic 318.71, p < 0.00001; I 2 = 98%). The funnel plot for this outcome indicates publication bia ( Figure 5Q).

Impact of Treatment
Where feasible, we also assessed the impacts of AYA cancer treatments on the fo lowing reproductive health outcomes: low birthweight, caesarian delivery, preterm birth low APGAR score at birth, and small for gestational age. AYA cancer patients had signi icantly higher odds of having a newborn with low birthweight across both chemotherap (pOR 1.75; 95% CI: 1.15, 2.67) ( Figure 9B) and radiation (pOR 1.67; 95% CI: 1.28, 2.18) (Fig  ure 10B) compared to AYA controls. Caesarean delivery followed a similar trend, wit significantly higher odds in AYA cancer patients compared to controls across both chem otherapy (pOR 1.28; 95% CI: 1.06, 1.54) ( Figure 9C) and radiation therapy (pOR 1.35; 95% CI 1.02, 1.79) ( Figure 10C). When considering radiation, preterm birth (pOR 1.65; 95% C 1.21, 2.26) (Figures 9A and 10A) and low APGAR score at birth (pOR 2.03; 95% CI: 1.32 3.13) (Figures 9E and 10E) were significantly higher in AYA cancer patients compared t controls. Small for gestational age remained unchanged across treatments, with neithe chemotherapy nor radiation indicating higher likelihood in AYA cancer patients com pared to controls ( Figures 9D and 10D). Details regarding heterogeneity and publicatio bias can be found in Figures 9-11.

Impact of Treatment
Where feasible, we also assessed the impacts of AYA cancer treatments on the following reproductive health outcomes: low birthweight, caesarian delivery, preterm birth, low APGAR score at birth, and small for gestational age. AYA cancer patients had significantly higher odds of having a newborn with low birthweight across both chemotherapy (pOR 1.75; 95% CI: 1.15, 2.67) ( Figure 9B) and radiation (pOR 1.67; 95% CI: 1.28, 2.18) ( Figure 10B) compared to AYA controls. Caesarean delivery followed a similar trend, with significantly higher odds in AYA cancer patients compared to controls across both chemotherapy (pOR 1.28; 95% CI: 1.06, 1.54) ( Figure 9C) and radiation therapy (pOR 1.35; 95% CI 1.02, 1.79) ( Figure 10C). When considering radiation, preterm birth (pOR 1.65; 95% CI: 1.21, 2.26) (Figures 9A and 10A) and low APGAR score at birth (pOR 2.03; 95% CI: 1.32, 3.13) (Figures 9E and 10E) were significantly higher in AYA cancer patients compared to controls. Small for gestational age remained unchanged across treatments, with neither chemotherapy nor radiation indicating higher likelihood in AYA cancer patients compared to controls (Figures 9D and 10D). Details regarding heterogeneity and publication bias can be found in Figures 9-11.  Cancers 2023, 15, x FOR PEER REVIEW 20 of 28 Figure 9. Forest plots of meta-analyses of chemotherapy and reproductive health outcomes (n = 5).

Discussion
This systematic review and meta-analysis aimed to synthesize current evidence on the impact of AYA cancer on reproductive health outcomes. Altogether, we included 21 studies that reported on 62 reproductive health outcomes across 102,041 AYA cancer patients. A key contribution is the development of a sex-based framework for organizing and conceptualizing reproductive health outcomes, categorizing them into fetal/neonatal (n = 26), maternal (n = 11), fetal/neonatal-maternal (n = 23), and maternal-paternal (n = 2) outcomes. Meta-analyses that were feasible showed associations between AYA cancer and eight reproductive health outcomes: fetal/neonatal outcomes of preterm birth ( . These findings align with the current literature on the implications of cancer treatment and diagnosis on reproductive health outcomes, while providing quantitative evidence regarding the size and direction of the impact. Given the number of reproductive health outcomes extracted (n = 62) and the variability in reporting, our developed sex-based framework for organizing and conceptualizing outcomes indicates areas research has covered, how reproductive health outcomes interrelate with each other, and gaps in the current literature. Across the entire framework, a large number of outcomes (n = 45) were reported in single studies. To investigate the impact of AYA cancer on outcomes that have been reported by single studies, we require more literature exploring these outcomes. This would facilitate pooling across studies in order to estimate the true impact of AYA cancer on the outcome. Furthermore, in our meta-analysis, we found unconfirmed associations between AYA cancer and nine outcomes, including six fetal/neonatal outcomes (small for gestational age, neonatal mortality, perinatal death, sex ratio, low APGAR score at birth, and spontaneous abortions), one maternal outcome (gestational hypertension), and two fetal/neonatal-maternal outcomes (premature ruptured membranes and antepartum hemorrhage). However, this lack of statistical significance may be driven by small sample sizes for rare outcomes across studies as well as a small number of studies in the meta-analyses, which is reflected by wide confidence intervals [46]. Therefore, there is a need for more studies that explore these outcomes. Categorization of who is impacted in this framework also identifies gaps. Particularly, although our review identified two maternal-paternal outcomes (need for fertility treatment and birth rate), we did not identify any studies that specifically evaluated the impact of male AYA cancer on paternal reproductive health outcomes. A 2021 systematic review and meta-analysis by Pizzol et al. summarized the impact of cancer treatment on ejaculatory dysfunction across all males with cancer in cross-sectional and case-control studies [47]. This review established that cancer treatment involving the lower spinal cord can impact ejaculatory function (prevalence of 6.8 to 68.7%). However, AYA cancer patients were underrepresented in the review. Considering the large psychosocial and economic impact of these outcomes for male AYA cancer patients [48], there is a need for more cohort studies to adequately assess these impacts compared to control populations.
Key findings of our meta-analysis are pooled estimates that quantify the association between AYA cancer and four fetal/neonatal outcomes (preterm birth, very preterm birth, low birthweight, and congenital anomalies), two maternal outcomes (preeclampsia and gestational diabetes), one fetal/neonatal-maternal outcome (caesarean delivery), and one maternal-paternal outcome (fertility treatment). Furthermore, when we further evaluated cancer treatment, we found associations between radiation exposure and preterm birth, low birthweight, caesarian delivery, and low APGAR score at birth. Chemotherapy was associated with low birthweight and caesarian delivery. Previous research has shown that radiation has a significant impact on fertility, especially when damage occurs to the pelvic or cranial regions [49,50]. Our findings align with previous research, as our data extraction found a higher risk of premature ovarian failure (OR 3.12; 95% CI: 1.70, 5.72) [9], and pooling use of fertility treatment indicated a higher risk (pOR 2.66; 95% CI 1.71, 4.11) in AYA cancer patients. Treatment to the pelvic region can result in damage to germ cells, while treatment to the cranial regions can alter the production of sex hormones from the hypothalamic-pituitary axis. These can result in issues with fertility, such as premature ovarian failure and increased usage of fertility treatment in AYA cancer patients. However, the majority of included studies did not report reproductive health outcomes according to treatment type, and as such, we were limited in our meta-analysis of the impact of cancer treatment. Additionally, further research is needed in comparing different types of radiation and/or chemotherapy, types of cancer, the impact of dosage, as well as location of treatment on reproductive health outcomes. Nonetheless, in quantifying associations between AYA cancer and, where feasible, cancer treatment and reproductive health outcomes according to our framework, our review provides empirical evidence to guide reproductive health care and decision making for both providers and patients. Indeed, our review supports the need for oncofertility counselling both prior to and after receiving treatment for AYA cancer patients. As treatment types can impact reproductive health outcomes, this should be taken into consideration during treatment plan development. This is reflected in an included study in our review by Medica et al. [38], who found a significantly higher use of emergency contraception in AYA cancer patients (OR 2.09; 95% CI: 1.82, 1.39). A higher frequency of emergency contraception usage suggests a need for more contraceptive counselling in AYA cancer patients. Therefore, discussions regarding family planning and the potential risks of adverse reproductive health outcomes across the developed framework is warranted. This will allow AYA cancer patients to make informed decisions regarding their reproductive health and family planning, which may reduce the anxiety associated with this process after cancer. In addition to oncofertility counselling, there is a need for closer obstetrical follow-up during pregnancy and delivery. Our review provides evidence of higher risk of gestational diabetes and preeclampsia, which negatively impact pregnancy. There is also a higher risk of fetal reproductive health outcomes such as preterm birth, very preterm birth, and low birthweight, which require management both before and after delivery. By providing closer follow-up, both the pregnant person and the fetus can receive care efficiently.
Given the inquiry into reproductive health, an important consideration in our systematic review is sex (a set of biological attributes traditionally associated to sex chromosome status) and gender (socially and culturally constructed roles). All of the included studies in our review reported sex in a binary fashion (i.e., male and female) and did not report gender or sexual orientation as a sociodemographic factor. This may be a limitation of data sources used, as the majority of the included studies relied on administrative health data, where gender-diverse data are not collected. As development of our framework for organizing and conceptualizing reproductive health outcomes was informed by studies included in the systematic review, it is important to note that it is sex-based. Indeed, it is also important that future research on AYA cancer and reproductive health is guided on principles of SGBA [15]. Among studies included in our systematic review, we noted instances of conflation of sex and gender when referring to sex. For example, the term "woman" was used when referring to the sex variable collected in a database, hence conflating sex and gender. Representation and inclusion of gender is integral to the external validity of research as well as the safety and care of the target population. Research has shown that trans and non-binary folks experience significant health disparities due to lack of access to appropriate care, financial barriers, and minimal cultural competency from healthcare providers [51]. Similar trends are seen for those of non-heterosexual status [52], who are at a higher risk for certain cancers in adolescence and young adulthood [53]. Therefore, data regarding the impact of AYA cancer on reproductive outcomes stratified according to sex, gender, and sexual orientation is imperative to measure the unique impact of these factors. Although data limitations may preclude the ability to incorporate SGBA [54] when evaluating the impact of AYA cancer and treatments on reproductive health, it is important to be aware of the intersectionality between sex, gender, and sexual identities and the potential impacts on outcomes. Future studies should intentionally include valid measures of sex, gender, and sexual orientation given that each of these have known, and potentially distinct, impacts on sexual and reproductive health measures.
Strengths and limitations of our work warrant discussion. Our search strategy was developed in collaboration with a research librarian. We applied a systematic approach to categorizing reported reproductive health outcomes, resulting in the aforementioned conceptual framework that guided our meta-analysis. It is important to comment on the heterogeneity we observed across meta-analyses that we were able to conduct. The heterogeneity in our meta-analyses ranged from 0% to 98%. Given our focus on observational epidemiologic studies, this was anticipated [23]. Heterogeneity is largely explored as methodological and clinical differences in how studies were executed. In terms of methodological heterogeneity, 20 of the included studies were cohort studies, and one was a cross-sectional study. All studies were conducted in high-resource countries and utilized a form of administrative health data. However, there was large diversity in sample sizes, which can be a driver of heterogeneity. Additionally, we expect a large degree of clinical heterogeneity, as participants of observational studies are not randomized to reduce confounding and selection bias, nor is there consistency across the included studies in terms of diversity of participants, interventions, or outcomes measured [23]. In order to further explore the source of heterogeneity, it is recommended to conduct subgroup analyses and stratify by the study feature in question [23]. This was not feasible in our review, as the included studies did not report their results according to characteristics of potential heterogeneity (e.g., intervention, type of cancer, clinical setting, sociodemographic factors, etc.). Furthermore, sensitivity analyses (that is, excluding individual studies and observing the impact on heterogeneity) were not feasible, as many of our outcomes were reported by a small number of studies. Removal of studies would have resulted in even wider confidence intervals and lower accuracy of our results [23].

Conclusions
Altogether, our systematic review and meta-analysis provide a comprehensive synthesis of reproductive health outcomes among AYA cancer patients. Findings have implications for supporting the need for oncofertility counseling before and after treatment so patients and their families can make informed decisions. Guidelines for specific obstetrical followup during and after pregnancy is also warranted to address the entire continuum before, during, and after pregnancy, which are impacted by AYA cancer status. This review also informs future research to address reproductive health for AYA cancer patients.