The use of fertility treatments among reproductive-aged women after cancer.

In this study, 1,036 cancer survivors and 1,026 comparison women were included prior to matching. The average age at participant interview was 37.5 years among cancer survivors and 37.8 years among comparison women. The median time from cancer diagnosis to the interview amongst the cancer survivors was 7 years (interquartile range 5–11 years). Demographic characteristics of cancer survivors and comparison women are presented in Table 1 , stratified by whether individuals visited a provider for help becoming pregnant. Comparison women were less likely to be white and more likely to report being in a married or committed relationship and earning an annual income greater than $50,000. In contrast to their cancer-free counterparts, cancer survivors were more likely to report a history of no children (32.6% vs. 19.5%) and less likely to report a history of at least two children (44.1% vs. 62.4%). Among cancer survivors, breast cancer was the most common type of cancer, and the average age at cancer diagnosis was 29.3 years old. Cancer treatments of chemotherapy and radiation were commonly reported among survivors, with 58.0% of survivors undergoing chemotherapy and 4.5% receiving total body or pelvic radiation. Only 25.5% of cancer survivors reported meeting their desired family size prior to cancer diagnosis. Cancer survivors who visited a doctor for help becoming pregnant were less likely to have met their desired family size at diagnosis (17.8%) compared to cancer survivors who did not visit a doctor for help becoming pregnant (27.1%). Cancer survivors who visited a provider for help becoming pregnant were more likely to be white, have completed college, be married or in a committed relationship at the time of study interview, and make more than $50,000 per year. No difference in type of cancer or receipt of chemotherapy or radiation was seen between cancer survivors who visited a doctor for help with pregnancy compared to survivors who did not. Similar demographic patterns were seen when comparison women who visited a doctor for help getting pregnant were compared with those who did not, except comparison women were more likely to have met their desired family size at the time of the interview regardless of whether they visited a doctor for help getting pregnant or not. Overall, cancer survivors did not have an increased likelihood of ever visiting a doctor for help becoming pregnant compared to cancer-free comparison women ( Table 2 ). Almost 17% of cancer survivors visited a doctor for help becoming pregnant compared to 15.8% of the cancer-free comparison women. Approximately 15–16% of comparison women and cancer survivors who did not receive gonadotoxic treatment visited a doctor for help becoming pregnant compared to 19.0% of survivors who received gonadotoxic treatment. Approximately 10% of both cancer survivors and comparison women in our population chose to pursue treatment. Use of each fertility treatment (surgery, hormones or medications, IUI, or IVF) was similar among cancer survivors and comparison women. Cancer survivors who received gonadotoxic therapy were as likely to pursue IUI (2.5%) or IVF (0.6%) as comparison women (3.7% and 1.9%, respectively), although the sample size was small among these groups. There were 73 cancer survivors and 56 comparison women who reported visiting a doctor for help becoming pregnant but did not ultimately pursue fertility treatment. The most common reasons for not pursuing fertility treatment among cancer survivors were wanting to attempt (38.4%) or successfully conceiving (43.8%) naturally, feeling that treatment was too expensive (24.8%), and feeling like it was the wrong time in life to get pregnant or have children (23.3%); these reasons were similar for comparison women ( Supplementary Table S1 ) After matching on birth year, adjusted estimates of the hazard of visiting a doctor for help becoming pregnant were similar among cancer survivors compared to cancer-free women (hazard ratio [HR] 1.16, 95% SI 0.78–1.74) ( Table 3 ). The adjusted HR [aHR] for pursuing any treatment to help become pregnant was close to the null when cancer survivors were compared with cancer-free women (aHR 0.88, 95% SI 0.46–1.56). The hazard for pursuing surgery to help become pregnant was higher among cancer survivors compared to cancer-free women (HR 1.55, 95% SI 0.67–3.71), though the estimate was imprecise, and the sample size was too small to calculate an adjusted estimate. Reported reasons for surgical interventions among cancer survivors and cancer-free women are presented in the Supplementary Table S2 . Endometriosis was the most commonly reported cause of surgical intervention in our population. No meaningful difference was seen in the aHR for using hormonal fertility treatment or medications comparing cancer survivors with cancer-free women (aHR 0.86, 95% SI 0.46–1.63). The aHR for history of cancer and IUI was 1.26 (95% SI 0.40–5.88). IVF usage was rare among both the cancer survivors and comparison women in our population. Only 13 (1.3%) cancer survivors and 19 (1.9%) comparison women reported pursuing IVF. Compared with cancer-free women, utilization of hormones or medications to help get pregnant was similar among women who received gonadotoxic therapy (aHR 0.81, 95% SI 0.14–3.01). The relation between receipt of gonadotoxic therapy and pursuing any fertility treatment was null (aHR 0.88, 95% SI 0.23–2.96). Similar results were obtained for all outcomes of interest in analyses where women were censored when they met their desired family size ( Supplementary Table S3 ). Among the cancer survivors, 39.9% reported that they would likely raise fewer children than they desired compared to 27.8% of comparison women. Cancer survivors were more likely than comparison women to report feeling they would raise fewer children than desired because of health reasons (49.0% vs. 22.3%), concern for a baby’s health (25.4% vs. 16.7%), and concern about living long enough to raise their children (35.2% vs. 13.1%) ( Supplementary Table S4 ). Cancer survivors were also more likely to report feeling that they were unable to get pregnant (43.4%) or were concerned about their ability to get pregnant (39.9%) than comparison women (24.1% and 26.6%, respectively).

The Furthering the Understanding of Cancer, Health, and Survivorship in Adult (FUCHSIA) Women’s Study is a population-based cohort study exploring the impact of cancer treatment on fertility among adult female cancer survivors residing in the state of Georgia. Eligible cancer survivors were identified from the state-wide Georgia Cancer Registry. Comparison women were identified from a purchased marketing list from InfoUSA and were frequency-matched to cancer survivors on age and region of residence within Georgia. Cancer survivors were eligible to participate if they were 20–35 years old at first diagnosis of a reportable invasive cancer or ductal carcinoma in situ occurring between the years of 1990–2009 and were more than two years from cancer diagnosis ( 27 , 28 ). Recruitment for the FUCHSIA Women’s Study occurred from 2012–2013. At recruitment, all women had to be between 22–45 years of age, have a working telephone, and speak English. For this analysis, participants were restricted to white and Black/African American women because of the small sample size for other races (n=108). Cancer survivors who underwent a hysterectomy, bilateral oophorectomy, or tubal ligation before or at the time of their cancer diagnosis were excluded. All recruited women provided verbal consent to participate in this study and then completed a structured computer assisted telephone interview. The institutional review boards of Emory University and the Georgia Department of Public Health approved this study. During the interview, women were asked whether they had ever visited a doctor or medical professional for help becoming pregnant. Additionally, cancer survivors were asked if they had been counseled about fertility preservation around the time of their cancer diagnosis. Those who responded “Yes” were then asked whether this occurred before, during, or after treatment. If a woman reported visiting a doctor for help becoming pregnant, she was then asked the type of doctor visited, age at the visit, and the reason for her visit. This information was collected for each reported visit to a doctor for help becoming pregnant. Women were also asked whether they had ever pursued treatment for help becoming pregnant and if they underwent specific types of fertility treatments, including surgery, medication or hormone use (including shots), artificial or intrauterine insemination (IUI), or in vitro fertilization (IVF). Cancer survivors who previously utilized fertility preservation with oocyte or embryo cryopreservation prior to cancer treatment were asked if they ever tried to fertilize their frozen oocytes or use their embryos for an embryo transfer. If a woman reported yes to ever pursuing a specific type of fertility treatment, she was then asked her age at each reported treatment, whether she became pregnant, and if that pregnancy lasted at least 20 weeks. Participants were also asked their expectations on the number of children they would raise and whether this would be fewer or more children than they want. Those who felt they would raise fewer children were presented with a list of possible reasons why they thought they would raise fewer children than they want and asked to select all that applied to them. Additional information was collected on covariates of interest including income, education, insurance status, pregnancy history, cancer treatment, and cancer type. Age at cancer diagnosis and type of cancer information were obtained from the cancer registry and verified with medical records. Use of chemotherapy and radiation was abstracted from participant medical records. Exposure to gonadotoxic cancer treatment was defined as receipt of alkylating chemotherapy, total body irradiation, or radiation to the abdomen or pelvis. Participant demographic characteristics were examined among cancer survivors and comparison women. We assessed the lifetime use of fertility care by cancer survivors and comparison women, describing whether women visited a doctor for help becoming pregnant, pursued any fertility treatment, or pursued specific fertility treatments. We fit unadjusted and adjusted Cox proportional hazards models to compare time to visiting a doctor for fertility treatment after cancer diagnosis for cancer survivors versus age-matched comparison women. To ensure comparable follow up, each cancer survivor was randomly matched with replacement on year of birth to a cancer-free woman in our study population. Then the cancer-free women were then assigned a proxy-diagnosis age equal to the actual age at cancer diagnosis of their matched cancer survivor. The follow-up period began at age of cancer diagnosis for survivors or assigned proxy-age of diagnosis for comparison women and lasted until censored at the date of their study interview or an earlier age if they underwent a hysterectomy, bilateral oophorectomy, or tubal ligation prior to study interview. We did not censor for cancer recurrence because this was on the causal pathway between cancer diagnoses and our outcome of pursuit of fertility treatment. To ensure stable estimates, we repeated the analytic matching procedure and refit each model 1,000 times. Thus, instead of being matched to only one comparison woman, each cancer survivor had the opportunity to be matched to other eligible comparison women in the study population. To summarize the results across these replications, we report the median hazard ratio and the 95% simulation interval (SI) based on the 2.5 and 97.5 percentiles of the replication results. The 95% SI differs from the 95% confidence interval because the SI reflects the observed distribution of the hazard ratios across replications rather than being calculated based on the parameter estimate and the standard error. Thus, the SI represents the variability in the HR as a result of sampling from the comparison group. Potential confounders were identified from the existing literature and based on the construction of a causal diagram for this study. Age was accounted for through matching and additional covariates in the adjusted models included maternal race, income, residence, education, and parity. Parity was calculated based on the number of live births a woman had at the time of her cancer diagnosis or proxy-diagnosis age. We fit additional Cox regression models for other outcomes of interest, including the pursuit of any fertility treatment and the pursuit of specific individual fertility treatments (surgery, hormones or medications, or IUI). Analyses were performed restricting cancer survivors to those who received gonadotoxic treatment. In supplementary analyses repeated for all outcomes, women were also censored at the time they reported meeting their desired family size. Women who met their desired family size before cancer diagnosis were excluded from these analyses, even if they became pregnant later. We only fit models for outcomes where there was sufficient sample size. Thus, we did not fit adjusted or unadjusted models for use of IVF in our population. Additionally, we did not fit adjusted models for having had surgery. When the analysis was restricted to cancer survivors receiving gonadotoxic therapy, models were suppressed for those who underwent surgery or IUI due to sample size.

This study is a large population-based study of the relation between cancer survivorship and fertility care. Overall, use of fertility treatments among our population was low. Visiting a doctor for help getting pregnant and choosing to undergo surgery, use hormones or medications, or receive IUI did not differ between cancer survivors and women without a history of cancer. Our results in an adult population support prior findings of low utilization of medications for the treatment of infertility and low overall use of fertility treatments among a study of childhood cancer survivors ( 10 ). Similarly, the low use of fertility treatments in our population, even among those who underwent known gonadotoxic therapy, supports prior conclusions that most female cancer survivors do not pursue fertility care despite reporting wanting biologic children ( 22 ). Economic, racial, and geographical disparities in access to treatments for infertility are well documented ( 27 , 29 ). Unlike some states ( 30 ), Georgia does not have a fertility treatment mandate requiring insurers to cover fertility preservation or fertility treatments in infertile women, which may have further limited access to treatments in our study population ( 31 ). Nearly one-quarter of cancer survivors reported not pursuing any fertility treatment due to cost. The proportion of survivors who would be concerned about cost of treatment may be even higher in the subset of women who would require more expensive treatments, such as IVF, to conceive. While there may be a modest update to referral patterns for fertility preservation or employer-based insurance coverage in Georgia over the past 10 years, access to services through state-mandated insurance has not changed. Other common reasons for not pursuing fertility treatment among cancer survivors included feeling like it was not the right time in life to get pregnant or have children (23.3%) and citing another (unspecified) reason (24.7%) ( Supplementary Table 1 ). Of survivors who reported feeling that they would raise fewer children than they wanted, a significant proportion reported having concerns about their health or concerns about living long enough to raise children. Thus, it is possible that cancer survivors have unique reasons for not pursuing treatment, such as fear of cancer recurrence with hormones used to conceive or fear of mortality. In addition to cost and access barriers to fertility treatments, female cancer survivors report not receiving enough information about fertility treatments from providers ( 21 , 22 , 32 – 34 ). Adequate discussions regarding fertility preservation and treatment options in cancer survivors is also linked to decreased regret and decisional conflict in this population, regardless of whether women choose to pursue fertility care ( 35 – 38 ). Further counseling and education surrounding fertility options is needed among young adult female cancer survivors, both before and after cancer treatment ( 39 , 40 ). Unlike most previous studies of fertility care among cancer survivors, our study recruited both cancer survivors and cancer-free women outside of a fertility clinic setting. This recruitment method allowed the inclusion of both cancer survivors and comparison women from the general population who might not visit a fertility clinic. Further, while many prior studies focused solely on the use of ART, we were able to investigate the use of additional fertility treatments, including hormone or medication usage and surgery. Due to the extensiveness of the interview, we were able to capture a woman’s entire reproductive history as opposed to focusing on the success of individual visits to a fertility clinic. As both age and cancer treatment may delay a woman’s ability to meet her reproductive goals and increase her risk of infertility, it was important to incorporate both in our analysis. A women’s parity and whether she has already met her reproductive goals prior to cancer diagnosis, may influence her pursuit of fertility care after treatment. Unlike previous analyses on this topic, we utilized a Cox model which accounted for time to fertility care outcomes of interest, taking both a woman’s age at diagnosis and her parity at that point in her reproductive history into account when estimating the relationship between cancer history and use of fertility care. For this analysis, we excluded cancer survivors who had had a hysterectomy, bilateral oophorectomy, or tubal ligation prior to or at the time of cancer diagnosis. We acknowledge that there are family-building options for those who have undergone these surgeries, and that these options are often even more involved and expensive than some traditional fertility treatments, as they may require the use of donor oocytes and/or a gestational carrier. When we evaluated the cancer survivors who were excluded based on prior hysterectomy, bilateral oophorectomy, or tubal ligation, we found that only 2 women reported seeing a doctor for help becoming pregnant after cancer diagnosis, and only 1 of these pursued treatment (reversal of a tubal ligation). Despite being excluded from this analysis, all cancer survivors who reported a hysterectomy, bilateral oophorectomy, or tubal ligation prior to or at the time of cancer treatment were asked about whether they attempted to use alternative family building methods, such as donor oocytes or a gestational carrier; however, none reported pursuing these options. Despite the large overall sample size of this study, several types of fertility treatments were rare among this population leading to a small sample size for multiple outcomes of interest. This reduced sample size may have led to imprecision in our estimates, which should be interpreted with caution. However, these questions have not been adequately addressed in the literature and very rarely investigated among the general population. Additionally, as Georgia does not require health insurers to cover infertility treatment, our results may not be applicable to women living in states where coverage is mandated ( 41 ). A further limitation of our study is the applicability of these results to current fertility care practices. Our study population underwent initial cancer treatment before 2009. Fertility care may have been less common at this time, and women may not have been as informed about fertility treatment options compared to women diagnosed today. In particular, oocyte cryopreservation was not widely utilized until after the American Society for Reproductive Medicine Practice Committee issued a statement removing the experimental label in 2012 ( 42 ). In a contemporary population, we might expect increased utilization of different types of fertility treatment among both cancer survivors and women in the general population. In the overall population where we frequency-matched on age, cancer survivors were less likely to have met their desired family size. When we focused on the post-cancer timeframe, our study suggests that cancer survivors diagnosed in their reproductive years were not more likely to pursue fertility treatments after cancer despite adjusting for age and parity at diagnosis. Additional study of fertility practices among cancer survivors is needed to more effectively counsel women diagnosed with cancer regarding their fertility care options and identify a plan to help them meet their reproductive goals after cancer treatment. In addition to patient-centered fertility care counseling being offered widely, reducing barriers to fertility preservation and post-cancer fertility treatment may be needed in this population in order for these women to meet their reproductive goals and achieve their desired family size.

Over the past several decades, improvements in cancer detection and treatment have led to increased survival among young adult cancer patients ( 1 ). Standard cancer treatments, such as some types of chemotherapy and radiation, have been shown to adversely affect female fertility ( 2 – 9 ) and are associated with increased risks of infertility, amenorrhea, diminished ovarian reserve ( 5 , 10 – 13 ). Further, a cancer diagnosis during the reproductive years may delay or impair their ability to have biological children or meet their desired family size ( 14 , 15 ). For those survivors who do pursue having biologic children after cancer treatment, their reproductive window may be shortened. Thus, cancer treatments could lead to an increased need for fertility treatment among survivors. However, the utilization of these treatments among reproductive aged cancer survivors is not well known. The American Society for Reproductive Medicine and the American Society of Clinical Oncology recommend providers discuss the risk of infertility in all patients treated for cancer during their reproductive years prior to cancer treatment ( 16 , 17 ). Although recent efforts in cancer centers have led to greater counseling and uptake of fertility preservation prior to treatment, historically, counseling and referral for oocyte cryopreservation has been low in this population ( 18 , 19 ). Nonetheless, gonadotoxic treatment often results in reduced fertility and a potentially greater need for fertility care services after cancer treatment ( 20 , 21 ). Further, despite recommendations for providers to refer young adults with a cancer diagnosis to fertility specialists, the access to fertility care among survivors after treatment remains well below recommended guidelines, even among those who complete preservation prior to treatment ( 22 ). Few studies have examined adult female cancer survivors’ use of fertility care following cancer recovery. A cross-sectional study of women with childhood, adolescent, and young adult cancers found 15% of women pursued fertility care after cancer treatment ( 22 ). Another study of female childhood cancer survivors reported that survivors sought fertility treatment as often as cancer-free women but experienced increased time to pregnancy ( 10 ). Other studies have investigated the success of specific types of fertility treatments, such as assisted reproductive technology (ART), among female cancer patients at ART clinics ( 23 – 26 ). However, very few studies have examined how often adult female cancer survivors in the general population, beyond the ART clinic population, seek care for infertility compared to cancer-free women. Additionally, a paucity of data is available on the types of fertility treatments they choose to pursue beyond ART, such as surgery or the use of hormones and medications. A better understanding of the fertility care options used by cancer survivors in the general population is crucial to identify areas where additional counseling or education may be needed before and after cancer treatment. The purpose of our study was to evaluate the pursuit of fertility treatments among adult female cancer survivors of reproductive age in a population-based setting.