Impact of cancer treatment on risk of infertility and diminished ovarian reserve in women with polycystic ovary syndrome

Abstract

Objective

To compare markers of fertility and ovarian reserve between cancer survivors and cancer-free women with and without PCOS.

Design

FUCHSIA Women’s Study–a population-based cohort study.

Setting

Not applicable.

Patients

Female cancer survivors (n=1,090) aged 22–45 years, diagnosed between ages 20–35 years, and at least 2 years post-diagnosis; 369 participated in a clinic visit. Three hundred seventy-four reproductive-aged women without cancer also completed a clinic visit.

Intervention(s)

None.

Main Outcome Measure(s)

Infertility, time to first pregnancy after cancer diagnosis, and measures of ovarian reserve (anti-Müllerian hormone [AMH] and antral follicle count [AFC]).

Results

Seventy-eight (7.2%) cancer survivors reported a PCOS diagnosis, with 41 receiving gonadotoxic treatment. Survivors with PCOS exposed to gonadotoxic treatment (odds ratio [OR] 2.3, 95% confidence interval [CI] 1.2, 4.5) and unexposed (OR 3.4, 95% CI 1.7, 6.9) were more likely to report infertility than unexposed survivors without PCOS, and were more likely to have fewer children than desired (exposed: OR 2.1, 95% CI 1.0, 4.2; unexposed: OR 3.0, 95% CI 1.4, 6.8). After adjusting for age, comparison women with PCOS had the highest markers of ovarian reserve (AMH: 2.43 ng/mL, 95% CI: 1.22, 4.82; AFC: 20.7, 95% CI: 15.3, 27.8) and cancer survivors without PCOS treated with gonadotoxic agents had the lowest levels (AMH: 0.19 ng/mL, 95% CI: 0.14, 0.26; AFC: 7.4, 95% CI: 6.4, 8.5).

Conclusion

Despite having higher AMH and AFC on average after cancer treatment, cancer survivors with PCOS were less likely to meet their reproductive goals compared to survivors without PCOS.

INTRODUCTION

Polycystic ovary syndrome (PCOS) affects approximately 5–10% of women, making it the most common endocrine abnormality in women of reproductive age (1, 2). Characterized by oligoovulation or anovulation, clinical or biological evidence of hyperandrogenism, and polycystic ovaries on ultrasound (3), PCOS is one of the most common causes of female infertility (4). Recent work in the field of oncofertility (5) has shown that some cancer treatments can diminish or deplete ovarian reserve in multiple different types of cancers (6–9). However, there is little information about fertility and ovarian reserve in women with PCOS who had cancer during their reproductive years.

Women diagnosed with PCOS who undergo treatment for cancer may have different reproductive outcomes than women without PCOS. PCOS is associated with higher anti-Müllerian hormone (AMH) levels and antral follicle count (AFC) as assessed by transvaginal ultrasound (10–13). Since AMH and AFC are correlated with the primordial follicle pool, these markers provide an estimate for ovarian reserve (14). Cancer treatment, especially gonadotoxic chemotherapy and radiation, is known to decrease a woman’s ovarian reserve. However, it is unclear whether women with PCOS may be somewhat protected from gonadotoxic therapy because they have a higher ovarian reserve before treatment, and if so, how this impacts their risk for infertility. Given that women with PCOS are already at a higher risk of infertility compared with those without PCOS, it is important to determine how a coexisting cancer diagnosis may affect this high-risk group. The primary objective of this study was to compare markers of fertility and ovarian reserve between cancer survivors with and without PCOS. Secondarily, cancer survivors were compared to women with and without PCOS who have no history of cancer.

METHODS

Study Population

We used data from the Furthering Understanding of Cancer, Health, and Survivorship in Adult (FUCHSIA) Women’s Study. The FUCHSIA Women’s Study is a population-based study examining the effect of cancer treatment during the reproductive years on future fertility. Eligibility criteria included: female; diagnosed with a reportable malignant cancer (15) or ductal carcinoma in situ (DCIS) between the ages of 20–35; diagnosed between 1990–2009; age 22–45 at the time of enrollment in the study between 2012–2013; and at least 2 years since cancer diagnosis at enrollment. Eligible cancer survivors were identified in collaboration with the Georgia Cancer Registry (GCR). A total of 1,282 cancer survivors completed the telephone interview. Women were excluded from this analysis if they had a hysterectomy or bilateral oophorectomy before their cancer diagnosis or as part of their initial cancer treatment because they would not be able to become pregnant after cancer treatment. Women with a uterus and at least one ovary were invited to participate in a sub-study to assess clinical markers of fertility; 369 cancer survivors completed a clinic visit.

Comparison women with no history of cancer were invited to participate in the telephone interview to represent the general population. Comparison women were told that the study would compare the health of women who survived cancer to the health of women who did not have cancer. Women were recruited using a commercial list that was frequency-matched to the cancer survivors based on age and location of residence in the state of Georgia. Comparison women were eligible to participate if they were 22–45 years at the time of recruitment. Eligible comparison women were also invited to participate in the sub-study to measure markers of ovarian reserve; 376 comparison women came to clinic.

The Institutional Review Boards of Emory University and the Georgia Department of Public Health approved this study.

Procedures

All study participants completed a detailed telephone interview about their reproductive histories, including questions regarding demographic characteristics, pregnancy history, medical conditions, lifestyle factors, periods of infertility, and reproductive goals. Cancer survivors were asked about cancer diagnosis and treatments received. During the interview, women were asked whether they had ever been diagnosed with polycystic ovary syndrome (PCOS); if they responded yes, they were then asked whether a doctor diagnosed their PCOS, the age at which it was diagnosed, and what signs or symptoms they experienced. Women who did not answer the question about a history of PCOS (survivor: n=13; comparison: n=2) were excluded. We asked women whether they had a period of time during which they had regular (at least three times per month), unprotected sex with a male partner for 6 months or longer but did not get pregnant. Those who responded in the affirmative were asked for the age at which this subfertile period occurred, the length of time for which it continued, whether they were actively trying to get pregnant during this time, and whether they got pregnant at the end of this time period. Further, for each pregnancy that a woman reported, she was asked how long she was having regular unprotected sex before the pregnancy. Whether or not a woman had met her reproductive goals was calculated by subtracting the number of children a woman had given birth to from the reported number of children the woman desired.

Information regarding cancer diagnosis, cancer type, and treatment for cancer survivors, including exposure to gonadotoxic treatment, was abstracted from medical records and the GCR, and compared with self-reported exposures. Gonadotoxic treatment was defined as receipt of systemic chemotherapy, total body irradiation, or radiation to the abdomen or pelvis. Participants who had evidence of administered chemotherapy in their medical records or who did not have records available but reported chemotherapy exposure that was confirmed by the GCR, were considered exposed to gonadotoxic chemotherapy treatment. Those who had total body irradiation or radiation to the abdomen or pelvis documented in their medical record or who did not have records available but self-reported radiation and the GCR indicated radiation for cervical, ovarian, colon, endometrial, kidney, placental, or vaginal cancer were considered exposed to gonadotoxic radiation treatment.

Clinic visits took place at participating reproductive clinics across the state of Georgia. Clinic visits included a blood draw and a transvaginal ultrasound. A trained sonographer performed transvaginal ultrasounds and measured ovarian volume for each ovary and antral follicle count (AFC, follicle sizes 2–10 mm). Sonographers were also asked to note whether polycystic ovaries were seen on ultrasound, defined by the Rotterdam criteria as 12 or more follicles measuring 2–9 mm in diameter in at least one ovary (3). For women who had two ovaries for whom only one ovary was visualized, AFC in the single, visualized ovary was doubled. Blood was drawn to measure serum anti-Müllerian hormone (AMH). Serum AMH levels were measured in duplicate by an enzyme-linked immunosorbent assay (ELISA) (UltraSensitive AMH/MIS ELISA, Ansh Labs, Webster, TX). For participants whose AMH was undetectable by the UltraSensitive assay, AMH was measured in duplicate using the Ansh Labs picoAMH ELISA (Ansh Labs, Webster, TX) with an assay sensitivity of 0.006 ng/mL. Low AMH (16) and AFC (17) were defined as being below the age-specific 25^th percentile for women in the general population.

Statistical Analysis

Descriptive statistics were used to examine the study population of cancer patients, stratified by history of PCOS and gonadotoxic exposure. For the main analysis, the proportions reporting infertility were compared among cancer survivors with and without PCOS who had been exposed or unexposed to gonadotoxins. Infertility was defined as having 12 months of regular, unprotected intercourse with a male partner after the age of 20 without a resultant pregnancy. Infertility both before and after cancer diagnosis was examined. Covariates that were considered to potentially confound the relationship between PCOS and infertility were age at the telephone interview, cancer type, age at cancer diagnosis, time since diagnosis, cancer treatment, whether or not the woman had met her reproductive goals by cancer diagnosis, being childless at diagnosis, and whether or not she had procedures or treatment to preserve fertility prior to or during cancer treatment. Separate logistic regression models were fit to estimate the odds ratio (OR) of having a period of infertility comparing those with and without PCOS for both before and after cancer diagnosis. The ORs for additional fertility and reproductive outcomes (visiting a doctor for help becoming pregnant, being childless at the time of the interview, and having fewer children than desired by the time of the interview) were also calculated comparing cancer survivors with and without PCOS stratified by gonadotoxic treatment.

For each self-reported pregnancy, women were asked the length of time they were having regular, unprotected sex prior to becoming pregnant. A Cox regression model was used to estimate the hazard ratio (HR) for time to pregnancy among cancer survivors who reported a positive pregnancy test after cancer diagnosis. For survivors reporting multiple pregnancies after cancer diagnosis, the time to first post-cancer pregnancy was used. This model was adjusted for age at cancer diagnosis, race, and BMI.

AMH values that were below the limit of detection were assigned a value of (limit of detection)/√2. Models were fit to examine the relationship between log-transformed AMH, history of cancer, and PCOS. Women were stratified based on cancer survivorship and PCOS status. The cancer survivors were further stratified based on whether or not they received gonadotoxic agents during cancer treatment. The model was adjusted for age at the time of the clinic visit. A negative log binomial model was used to evaluate whether the mean total AFC values were different for cancer survivors and comparison women with and without PCOS. The negative log binomial model was also adjusted for age at clinic visit and stratified on receipt of gonadotoxic treatment.

SAS 9.4 was used for all statistical analyses (SAS Institute, Cary, N.C.).

RESULTS

Characteristics of Study Population

A total of 1,282 cancer survivors participated in the study. One-hundred ninety-two survivors were excluded because they did not answer the question about a history of PCOS during the interview (n=13) or had a hysterectomy or bilateral oophorectomy before or in the year following cancer diagnosis (n=179). Of the 1,090 cancer survivors included in the analysis, 78 (7.2%) reported being diagnosed with PCOS by a medical professional. Among the survivors diagnosed with PCOS, 52 (66.7%) reported being diagnosed before or during the year of their cancer diagnosis and 26 (33.3%) reported being diagnosed after their cancer diagnosis. The median age of PCOS diagnosis was 26. The most common symptoms reported during the interview by cancer survivors with PCOS were irregular or no menstrual cycles (n=65, 83.3%); increased hair growth on face, back, or stomach (n=40, 51.3%); infertility (n=34, 43.6%); insulin resistance, diabetes, or pre-diabetes (n=30, 38.5%); and high testosterone levels (n=21, 26.9%). Sixty-two women with PCOS (79.5%) reported experiencing two or more of the above symptoms.

The characteristics of the cancer survivors included in the analysis stratified by PCOS status and exposure to gonadotoxic treatment are presented in Table 1. Cancer survivors with and without PCOS were similar in regards to age at cancer diagnosis, race, and receipt of radiation and chemotherapy. Of survivors exposed to gonadotoxic treatment, 91.2% of those with PCOS and 92.8% of those without PCOS were exposed to alkylating agents. The proportion of cancer survivors with reproductive cancer was higher in those with versus without PCOS, while the proportion of survivors with breast cancer was lower. Absence of menses was more common among women with gonadotoxic treatment for women both with and without PCOS. Women with PCOS were slightly younger at the time of the interview compared to women without PCOS. Survivors with PCOS had higher body mass indices (BMI) compared to survivors without PCOS (mean BMI 31.5 vs. 26.6, respectively). Use of fertility preservation was low in all groups but was highest in those exposed to gonadotoxic treatment, regardless of PCOS status. Both groups desired a median of two children.

Table 1.

Characteristics of a Cohort of Young Adult Female Cancer Survivors by PCOS Status and Receipt of Gonadotoxic treatment^a,b

	PCOS with Gonadotoxic Treatment (n=41)		PCOS without Gonadotoxic Treatment (n=37)		No PCOS with Gonadotoxic Treatment (n=572)		No PCOS without Gonadotoxic Treatment (n=440)
	No.	%	No.	%	No.	%	No.	%
Age at cancer diagnosis
20–24	8	19.5	4	10.8	86	15.0	69	15.7
25–29	9	22.0	16	43.2	161	28.2	167	37.9
30–35	24	58.5	17	46.0	325	56.8	204	46.4
Age at interview
22–29	3	7.3	2	5.4	43	7.5	27	6.1
30–39	28	68.3	26	70.3	299	52.3	247	56.1
40–45	10	24.4	9	24.3	230	40.2	166	37.7
Race/Ethnicity
White	32	78.1	23	62.2	361	63.6	332	76.2
Black	8	19.5	9	24.3	181	31.9	85	19.5
Other race	1	2.4	5	13.5	26	4.6	19	4.4
Missing					4		4
Body Mass Index
Underweight	0	0.0	0	0.0	4	0.70	11	2.5
Normal	16	40.0	7	19.4	275	48.3	210	47.8
Overweight	8	20.0	5	13.9	146	25.6	122	27.8
Obese	16	40.0	24	66.7	145	25.4	96	21.9
Cancer Type
Breast	16	39.0	6	16.2	286	50.0	62	14.1
Lymphomas	13	31.7	0	0.0	146	25.5	15	3.4
Reproductivec	1	2.4	7	18.9	19	3.3	33	7.5
Other	11	26.8	24	64.9	121	21.2	330	75.0
Menses after treatmentd
Had menses	37	90.2	35	94.6	496	86.7	434	98.6
Menses absent	4	9.8	2	5.4	76	13.3	6	1.4
Had fertility preserving treatmente
Yes	7	17.1	0	0.0	70	12.2	4	0.9
No	34	82.9	37	100.0	502	87.8	436	99.1

^aCancer survivors who had a hysterectomy and bilateral oophorectomy prior to cancer diagnosis, and those who did not answer the interview question regarding a history of PCOS were excluded.

^bCancer survivors were considered exposed to gonadotoxic treatment if they had received systemic chemotherapy, total body irradiation, or radiation to the abdomen or pelvis.

^cReproductive cancers: uterine, ovarian, and cervical

^dMenstrual status assessed by participant’s response to the questions, “Did your menstrual periods stop during your cancer treatment?” and “For how long did your period stop?” Women who reported their period stopping and never returning are classified as having absent menses.

^eFertility preserving treatment includes receiving drugs to suppress ovarian function (such as Lupron, Synarel, or Zoladex), undergoing egg retrieval for freezing eggs or embryos, surgery to remove and freeze the ovaries or ovarian tissue, or surgery, such as moving the ovaries out of the field of cancer treatment radiation.

Infertility and Time to Pregnancy Among Cancer Survivors

A higher proportion of cancer survivors with PCOS reported ever having a period of infertility (55.1%) compared to survivors without PCOS (33.2%). After separating infertile periods before and after cancer diagnosis, survivors with PCOS compared to those with no PCOS were still more likely to experience infertility both before (43.6 vs. 21.7%, respectively) and after (25.7% vs. 15.3%, respectively) cancer diagnosis. After adjusting for age at diagnosis, BMI, and race, cancer survivors with PCOS were more likely to report infertility prior to diagnosis (adjusted odds ratio [aOR] 3.0, 95% confidence interval [CI]: 1.8, 5.1), as well as after diagnosis (aOR 2.2, 95% CI: 1.2, 4.1) compared to survivors without PCOS. When the groups were stratified on exposure to gonadotoxic treatment (Table 2), the odds ratios (OR) for infertility before diagnosis among survivors with PCOS with and without gonadotoxic exposure were similar, using survivors without PCOS who were unexposed to gonadotoxic treatments as the reference group. However, the OR for infertility after cancer diagnosis was larger for women with PCOS who had never been exposed to gonadotoxic treatment than for women with PCOS who had been exposed to gonadotoxins as part of their cancer treatment. Women with PCOS both with (OR 2.3, 95% CI 1.1, 4.8) and without (OR 3.9, 95% CI 1.9, 7.8) gonadotoxic exposure were more likely to have ever visited a doctor or health professional for help becoming pregnant than women without PCOS; this relation remained true for visiting a doctor after cancer in those with PCOS who were not exposed to gonadotoxins (OR 2.2, 95% CI 0.9, 5.0), but those with PCOS and gonadotoxic exposure were not more likely to have visited a doctor after cancer treatment than those without PCOS who were unexposed to gonadotoxic treatment (OR 0.8, 95% CI 0.3, 2.5). The majority of survivors with PCOS who reported infertility after cancer (10/18, 55.6%) reported having regular menstrual cycles at the interview. The vast majority of survivors with PCOS both with (87.1%) and without (93.7%) gonadotoxic exposure who still had a uterus and at least one ovary reported still menstruating at the time of the interview, with over half in both groups reporting regular cycles (Supplemental Table 1). Survivors with PCOS were more likely to be childless (gonadotoxic exposure: OR 3.0, 95% CI: 1.6, 5.8; no gonadotoxic exposure: 1.6, 95% CI 0.8, 3.2) and more likely to have fewer kids than they desired (gonadotoxic exposure: OR 2.1, 95% CI: 1.0, 4.2; no gonadotoxic exposure: 3.0, 95% CI 1.4, 6.8) by the time of the interview compared with women without PCOS unexposed to gonadotoxins. Fifty percent of cancer survivors with PCOS agreed with the statement, “I will probably raise fewer children than I want” compared with 36.5% of survivors without PCOS. Of those who agreed with the statement, 66.7% of survivors with PCOS stated that they were unable to get pregnant or were concerned about their ability to get pregnant compared with 58% of survivors without PCOS.

Table 2.

Odds ratios for analysis of the association between PCOS, receipt of gonadotoxic treatment, and fertility outcomes in female survivors of young adult cancers

			Unadjusted		Adjustedb
	N	%a	OR	95% CI	OR	95% CI
Ever Reported Infertility
PCOS with Gonadotoxic Treatment	20	48.8	2.3	(1.2, 4.5)	2.4	(1.2, 4.8)
PCOS without Gonadotoxic Treatment	23	62.2	3.4	(1.7, 6.9)	3.5	(1.7, 7.3)
No PCOS with Gonadotoxic Treatment	199	34.8	1.2	(0.9, 1.6)	1.1	(0.9, 1.5)
No PCOS without Gonadotoxic Treatment	137	31.1	1.0	Referent	1.0	Referent
Reported infertility prior to cancer diagnosis
PCOS with Gonadotoxic Treatment	16	39.0	2.9	(1.5, 5.7)	3.2	(1.6, 6.7)
PCOS without Gonadotoxic Treatment	18	48.6	3.8	(1.9, 7.5)	3.4	(1.6, 7.1)
No PCOS with Gonadotoxic Treatment	135	23.6	1.3	(1.0, 1.8)	1.2	(0.8, 1.6)
No PCOS without Gonadotoxic Treatment	85	19.3	1.0	Referent	1.0	Referent
Reported infertility after cancer diagnosisc
PCOS with Gonadotoxic Treatment	7	20.6	1.5	(0.6, 3.7)	–	–
PCOS without Gonadotoxic Treatment	11	30.6	2.6	(1.2, 5.5)	–	–
No PCOS with Gonadotoxic Treatment	80	15.8	1.1	(0.8, 1.6)	–	–
No PCOS without Gonadotoxic Treatment	58	14.6	1.0	Referent	–	–
Visited a doctor for help becoming pregnant
PCOS with Gonadotoxic Treatment	12	29.3	2.3	(1.1, 4.8)	2.5	(1.2, 5.3)
PCOS without Gonadotoxic Treatment	15	40.5	3.9	(1.9, 7.8)	4.0	(1.9, 8.5)
No PCOS with Gonadotoxic Treatment	88	15.4	1.0	(0.7, 1.5)	1.1	(0.8, 1.6)
No PCOS without Gonadotoxic Treatment	66	15.0	1.0	Referent	1.0	Referent
Visited a doctor for help becoming pregnant after cancer treatmentd
PCOS with Gonadotoxic Treatment	4	10.8	0.8	(0.3, 2.5)	–	–
PCOS without Gonadotoxic Treatment	8	22.2	2.2	(0.9, 5.0)	–	–
No PCOS with Gonadotoxic Treatment	50	9.4	0.7	(0.5, 1.1)	–	–
No PCOS without Gonadotoxic Treatment	48	11.6	1.0	Referent	–	–
Childless at interviewe
PCOS with Gonadotoxic Treatment	24	58.5	3.0	(1.6, 5.8)	2. 3	(0.9, 6.0)
PCOS without Gonadotoxic Treatment	16	43.2	1.6	(0.8, 3.2)	0.8	(0.2, 3.4)
No PCOS with Gonadotoxic Treatment	157	27.4	0.8	(0.6, 1.1)	0.9	(0.6, 1.4)
No PCOS without Gonadotoxic Treatment	141	32.0	1.0	Referent	1.0	Referent
Had fewer children than desired by interviewf
PCOS with Gonadotoxic Treatment	19	46.3	2.1	(1.0, 4.2)	2.0	(0.7, 5.7)
PCOS without Gonadotoxic Treatment	28	75.7	3.0	(1.4, 6.8)	1.7	(0.4, 7.2)
No PCOS with Gonadotoxic Treatment	310	54.2	1.0	(0.8, 1.3)	1.0	(0.6, 1.5)
No PCOS without Gonadotoxic Treatment	231	52.5	1.0	Referent	1.0	Referent

OR = odds ratio; 95% CI = 95% confidence interval

^aPercentages are out of total number of cancer survivors with PCOS who received gonadotoxic treatment (n=41), survivors with PCOS who did not receive gonadotoxic treatment (n=37), survivors without PCOS who received gonadotoxic treatment (n=572), and survivors without PCOS who did not receive gonadotoxic treatment (n=440) unless otherwise stated

^bAdjusted for age at diagnosis, race, and BMI. Adjustment suppressed when cell size was below 5.

^cPercentages are out of cancer survivors with cancer survivors with PCOS who received gonadotoxic treatment (n=34), survivors with PCOS who did not receive gonadotoxic treatment (n=36), survivors without PCOS who received gonadotoxic treatment (n=506), and survivors without PCOS who did not receive gonadotoxic treatment (n=398) who were capable of getting pregnant (still had a uterus and at least one ovary) after cancer and who answered questions regarding infertility after cancer. Women who reported a hysterectomy or bilateral oophorectomy prior to or during the year of cancer diagnosis were excluded.

^dPercentages are out of cancer survivors with PCOS who received gonadotoxic treatment (n=37), survivors with PCOS who did not receive gonadotoxic treatment (n=36), survivors without PCOS who received gonadotoxic treatment (n=535), and survivors without PCOS who did not receive gonadotoxic treatment (n=415) who were capable of getting pregnant (still had a uterus and at least one ovary) after cancer. Women who reported a hysterectomy or bilateral oophorectomy prior to or during the year of cancer diagnosis were excluded.

^eChildless: not having given birth to a child by the time of the interview

^fFewer children than desired: calculated by subtracting the number of children women gave birth to from the total number they reported they desired

A similar proportion of cancer survivors with (n=26, 33.3%) and without (n=368, 36.4%) PCOS reported a pregnancy after cancer diagnosis. Among survivors who reported having a pregnancy after cancer diagnosis, those with PCOS took longer to get pregnant compared to survivors without PCOS (HR 0.6, 95% CI: 0.4, 1.0) (Supplemental Figure 1). In a model adjusted for age at diagnosis, race, and BMI, the results were similar (adjusted HR 0.7; 95% CI: 0.4, 1.1). When the unadjusted model was restricted to survivors who received gonadotoxic treatment, the relationship between PCOS and time to pregnancy was null (HR 1.1, 95% CI: 0.4, 2.6), but the number of women with PCOS who became pregnant after gonadotoxic treatment was small.

Clinical Markers of Ovarian Reserve Among Cancer Survivors and Comparison Women with and without PCOS

Three-hundred and sixty-nine cancer survivors participated in a clinic visit; however, AFC and AMH on one participant could not be assessed. Consequently, 368 survivors were included in the analysis; 26 of them (7.1%) had a history of PCOS. There were 376 women with no history of cancer who participated in a clinic visit as a comparison group; however, two comparison women did not answer questions regarding PCOS and were thus excluded. Among the 374 comparison women in the analysis, 27 (7.2%) had a history of PCOS. The survivors with PCOS who visited clinic were slightly younger than the other women, but all groups were similar with respect to race (Table 3). Survivors and comparison women with PCOS had a higher BMI than women without PCOS. One survivor with PCOS and one survivor without PCOS did not have AMH levels drawn due to difficulty obtaining intravenous access. Without taking into account exposure to gonadotoxic treatment, the geometric mean (95% CI) AMH levels were 1.64 (0.91, 2.98) ng/mL for survivors with PCOS and 0.31 (0.24, 0.40) ng/mL for survivors with no history of PCOS. All cancer survivors with PCOS had a detectable AMH level; 16.4% of the survivors with no history of PCOS had AMH levels below the limit of detection. Half of the cancer survivors with PCOS had low age-specific AMH levels compared with 67.0% of the cancer survivors without PCOS. Of the 26 survivors with PCOS, 38.5% had polycystic ovarian morphology on ultrasound and 23.1% had low AFC. Of the 342 survivors without PCOS, 13.2% had polycystic ovarian morphology on ultrasound and 44.4% had low AFC.

Table 3.

Characteristics of women who participated in a clinic visit by cancer survivorship and PCOS status^a

	Cancer Survivors				Comparison Women
	With PCOS (n=26)		Without PCOS (n=342)		With PCOS (n=27)		Without PCOS (n=347)
	No.	%	No.	%	No.	%	No.	%
Age at clinic visit
23–36	13	50.0	121	35.4	7	25.9	88	25.4
37–40	7	26.9	92	26.9	9	33.3	114	32.9
41–47	6	23.1	129	37.7	11	40.7	145	41.8
Race/Ethnicityb
White	15	57.7	216	63.5	17	63.0	217	62.9
Black	10	38.5	104	30.6	9	33.3	112	32.5
Other race	1	3.9	20	5.9	1	3.7	16	4.6
Body Mass Index
Underweight	0	0.0	5	1.5	0	0.0	6	1.7
Normal	3	11.5	130	38.0	6	22.2	108	31.1
Overweight	3	11.5	98	28.7	1	3.7	100	28.8
Obese	20	76.9	109	31.9	20	74.1	133	38.3
Cancer Type
Breast	5	19.2	99	28.9	—	—	—	—
Lymphomas	4	15.4	67	19.6	—	—	—	—
Reproductivec	3	11.5	16	4.7	—	—	—	—
Other	14	53.8	160	46.8	—	—	—	—
Received chemotherapy
Yes	10	38.5	196	57.3	—	—	—	—
No	16	61.5	146	42.7	—	—	—	—
Received radiation
Yes	14	53.8	180	52.6	—	—	—	—
No	12	46.2	162	47.4	—	—	—	—
Exposed to gonadotoxic treatmentd
Yes	10	38.5	196	57.3	—	—	—	—
No	16	61.5	146	42.7	—	—	—	—
Menses after treatmente
Had menses	24	92.3	312	91.2	—	—	—	—
Menses absent	2	7.7	30	8.8	—	—	—	—
Ever Reported Infertilityf
Yes	16	64.0	110	33.5	18	72.0	102	30.8
No	9	36.0	218	66.5	7	28.0	231	69.2
Low anti-Mullerian hormone (AMH) levelg
Yes	13	50.0	229	67.0	9	33.3	179	51.6
No	13	50.0	113	33.0	18	66.7	168	48.8
Low antral follicle count (AFC)h
Yes	6	23.1	152	44.4	5	18.5	82	23.6
No	20	76.9	190	55.6	22	81.5	265	76.4

^aWomen who did not answer the interview question regarding a history of PCOS were excluded. Percentages are out of 26 cancer survivors with PCOS, 342 cancer survivors without PCOS, 27 comparison women with PCOS, and 347 comparison women without PCOS, unless otherwise specified.

^bPercentages are out of cancer survivors with PCOS (n=26), cancer survivors without PCOS (n=340), comparison women with PCOS (n=27), and comparison women without PCOS (n=345) who answered questions regarding race and ethnicity.

^cReproductive cancers: uterine, ovarian, and cervical

^dCancer survivors were considered exposed to gonadotoxic treatment if they had received systemic chemotherapy, total body irradiation, or radiation to the abdomen or pelvis.

^eMenstrual status assessed by participant’s response to the questions, “Did your menstrual periods stop during your cancer treatment?” and “For how long did your period stop?” Women who reported their period stopping and never returning are classified as having absent menses.

^fPercentages are out of cancer survivors with PCOS (n=25), cancer survivors without PCOS (n=328), comparison women with PCOS (n=25), and comparison women without PCOS (n=333) who answered questions regarding periods of infertility.

^gLow AMH defined as being below the age-specific 25^th percentile for the general population (16)

^hLow AFC defined as being below the age-specific 25^th percentile for the general population (17)

The geometric mean (95% CI) AMH levels were 1.68 (0.79, 3.58) ng/mL for comparison women with PCOS and 0.92 (0.76, 1.10) ng/mL for comparison women without PCOS. Two comparison women with PCOS (7.4%) and 11 comparison women (3.2%) without PCOS had undetectable AMH levels. A lower proportion of comparison women with and without PCOS had low age-specific AMH levels and AFC compared with cancer survivors with and without PCOS (Table 3).

Table 4 depicts the predicted geometric mean and 95% CI for AMH levels using the stratified model adjusted for age at the time of clinic visit. The comparison women with PCOS had the highest predicted AMH levels (2.43 ng/mL, 95% CI: 1.22, 4.82) followed by the cancer survivors with PCOS who did not receive gonadotoxic treatment (1.82 ng/mL, 95% CI: 0.75, 4.42); however, due to small sample size of women with PCOS who came to clinic, the estimates are not precise. The lowest predicted geometric mean levels were in the cancer survivors with (0.92 ng/mL, 95% CI: 0.29, 2.92) and without (0.19 ng/mL, 95% CI: 0.14, 0.26) PCOS who had been exposed to gonadotoxic treatment.

Table 4.

Adjusted estimates and 95% confidence interval for geometric mean values of anti-Müllerian hormone (AMH) and antral follicle count (AFC) among cancer survivors and comparison women with and without polycystic ovary syndrome (PCOS) stratified on exposure to gonadotoxic treatment^a and adjusted for age at clinic visit.

^aCancer survivors were considered exposed to gonadotoxic treatment if they had received systemic chemotherapy, total body irradiation, or radiation to the abdomen or pelvis.

^bAdjusted for age at clinic visit. Adjusted estimate for a woman who was 39 years old at the time of the clinic visit.

The AFC data showed a similar pattern when the examining cancer survivorship status, PCOS, and exposure to gonadotoxic agents (Table 4). The comparison women with PCOS had the highest predicted AFC (20.7, 95% CI: 15.3, 27.8) followed by the cancer survivors with PCOS who were not exposed to gonadotoxins (19.4, 95% CI: 13.1, 28.8). The lowest predicted AFC values belonged to the cancer survivors with (11.2, 95% CI: 6.8, 18.4) and without (7.4, 95% CI: 6.4, 8.5) PCOS who had been exposed to gonadotoxic agents during cancer treatment.

DISCUSSION

Cancer survivors with PCOS were much less likely to meet their reproductive goals by the time of the interview compared to survivors without PCOS despite having higher AMH and AFC on average. Survivors with PCOS were more likely to be childless at the time of the interview even though they had similar reproductive goals as survivors without PCOS. Yet, cancer survivors with PCOS were less likely to have age-specific low AFC or AMH, despite similar cancer diagnoses and treatments. This suggests that having PCOS may be, to some extent, protective against developing diminished ovarian reserve after cancer treatment. However, this preserved ovarian reserve does not appear to translate into improved fertility in women with PCOS, as cancer survivors with PCOS were more likely to report a period of infertility both before and after cancer diagnosis. Other lifestyle and biological factors, particularly the increased likelihood of anovulation in women with PCOS (3), may be at play for cancer survivors with PCOS and should be taken into consideration for those who desire children after cancer.

Data from the clinic portion of the study suggest that cancer survivors with PCOS may not be at risk of reduced ovarian reserve so long as they are not exposed to gonadotoxic treatment. The comparison women and survivors with PCOS had the highest AMH and AFC, although small sample size for these groups led to wide confidence intervals around the AMH and AFC estimates. Furthermore, cancer survivors with PCOS who were not exposed to gonadotoxins had higher AMH and AFC than comparison women without PCOS, suggesting that having cancer alone does not drastically diminish ovarian reserve in women with PCOS. However, cancer survivors who were exposed to gonadotoxic treatment fared worse with regards to ovarian reserve regardless of whether or not they had PCOS. Even in cancer survivors with PCOS, those who were exposed to gonadotoxic agents had lower AMH and AFC compared to both comparison women and cancer survivors without PCOS who had not been exposed to gonadotoxins.

Some treatment strategies for PCOS-related infertility aim to reduce the number of developing follicles in the ovaries. Laparoscopic ovarian drilling or diathermy (LOD) has been shown to decrease both AMH and AFC levels (13, 18–20) while making women with PCOS more likely to ovulate and therefore conceive in the post-LOD period. The underlying physiological mechanism of action of LOD involves a decrease in ovarian and systemic androgen concentrations via focal destruction of the ovarian stroma, likely contributing to an increase in follicle-stimulating hormone (FSH) levels and subsequent ovulation (21). While LOD is no longer a first-line treatment to assist women with PCOS in conceiving, it remains a viable option for women with PCOS who fail medical management alone (22). One study found that out of 74 women with PCOS who underwent LOD, 47 achieved pregnancy (63.5%), 20 of which were spontaneous (23). A 2010 review of 6 randomized controlled trials including a total of 439 patients with PCOS found that women who received LOD had similar pregnancy rates compared to those who received gonadotropin treatment (24). Additionally, despite oligoovulation being a hallmark symptom of PCOS, follicle loss in women with polycystic ovaries leads to more regular menstrual cycles as women with PCOS age (25). Follicle loss, through destruction by LOD or through the process of natural aging, restores ovulatory function and may improve fertility in women with PCOS.

Similarly, some cancer treatments, particularly some chemotherapeutic agents, have been found to decrease follicles and decrease AMH and AFC in cancer survivors (6–9), which might be hypothesized to be beneficial to women with PCOS. In our study, a smaller proportion of cancer survivors with PCOS had low age-specific AMH and AFC compared to survivors without PCOS, but cancer survivors with PCOS were more likely to report infertility than those without PCOS. Among those with PCOS, the odds ratios for reporting infertility after cancer and for visiting a doctor for help getting pregnant were smaller among those who received gonadotoxic treatment compared to those who did not. However, women with PCOS both exposed and unexposed to gonadotoxic treatment were still more likely to have fewer children than desired compared to those without PCOS. While different chemotherapy regimens may have differing degrees of impact on ovarian reserve, the overwhelming majority of survivors in our study who were exposed to systemic chemotherapy received an alkylating agent, which has been shown to be especially gonadotoxic due to its nonspecific mechanism of action and affinity for rapidly dividing cells (26). Thus, it does not appear that gonadotoxic treatment substantially improves fertility in cancer survivors with PCOS given that survivors with PCOS were less likely to have met their reproductive goals by the interview.

Survivors with PCOS were more likely to experience infertility both before and after cancer compared to survivors without PCOS. Since survivors with PCOS were less likely to have met their reproductive goals at diagnosis, and women of reproductive age with PCOS are less likely to use contraception and more likely to be trying to conceive (27), they may be more likely to report infertility after cancer than survivors without PCOS who met their reproductive goals before cancer diagnosis or who were contracepting after cancer. However, the odds ratio for experiencing infertility comparing women with PCOS to those without decreased after cancer compared to before cancer. While cancer survivors with and without PCOS become more similar after cancer, it does not necessarily indicate that survivors with PCOS have improved fertility after cancer. A more likely explanation is that the chance of experiencing infertility increases in cancer survivors without PCOS following cancer treatment (28) and with increasing age (29). Among cancer survivors who had a pregnancy after cancer, it took those with PCOS longer to get pregnant. These results may underestimate the difference in time trying to achieve pregnancy because the survival analysis only included women who were ultimately able to become pregnant. If cancer survivors with PCOS are disproportionately affected by infertility that never results in a pregnancy compared to cancer survivors without PCOS (as is seen in the literature regarding women with and without PCOS with no history of cancer), then the HR among women achieving pregnancy may underestimate the actual difference.

Our study is not without limitations. First, PCOS status was assessed via self-report and may therefore be susceptible to recall error. We attempted to verify PCOS status by explicitly asking whether this diagnosis was made by a doctor, the type of doctor who made the diagnosis, and the PCOS symptoms that each woman with PCOS experienced. However, questions regarding PCOS symptoms were not asked to women who did not report being diagnosed by a physician with PCOS. Therefore, we cannot evaluate the prevalence of potentially undiagnosed or misdiagnosed PCOS in our cohort. For women who participated in the clinic visit, we were able to evaluate the proportion of women who had polycystic ovaries on ultrasound (12 or more follicles measuring 2–9 mm in at least one ovary). However, this has limitations as well because there may be women with undiagnosed PCOS whose cancer treatment affected their ovaries to a degree to which they no longer had polycystic ovarian morphology despite potentially having other PCOS sequelae. Despite this limitation, possibly having some women with undiagnosed PCOS misclassified into the group without PCOS would likely cause the results to trend towards the null. Therefore, any differences found in the analysis would likely be an underestimation of the true difference.

This study has many strengths. One strength is the large number of cancer survivors who we were able to include in our analysis, among whom a similar proportion had been diagnosed with PCOS as is seen in the general population (1, 2). We were able to include survivors with a history of many different types of cancer, allowing us to evaluate whether exposure to different treatments had differential influences on fertility in women with and without PCOS. We were able to include medical record data for most women to confirm receipt of different types of treatments, including gonadotoxic exposures. Another strength is our ability to incorporate data from the clinic visits to supplement data from the interview, allowing us to draw conclusions not only in regards to reproductive outcomes, but also in regards to ovarian reserve and reproductive potential. We were also able to compare the cancer survivors to a large group of comparison women without a history of cancer, allowing us to further evaluate the extent to which ovarian reserve in cancer survivors with PCOS may be affected by cancer diagnosis and exposure to gonadotoxic treatments.

To the best of our knowledge, this is the first study examining infertility in women with PCOS after cancer treatment. Despite being less likely to have diminished ovarian reserve after cancer treatment, women with a history of PCOS were still at risk for not meeting their reproductive goals and for taking longer to become pregnant. Women with PCOS may be a unique population when it comes to fertility counseling after cancer treatment, who need counseling that addresses both PCOS-related infertility as well as infertility related to cancer treatment. Additional research is needed regarding whether fertility treatment success rates differ for cancer survivors with and without PCOS.

Capsule.

Cancer survivors with PCOS had higher ovarian reserve but were less likely to meet their reproductive goals compared to survivors without PCOS.