Risk of major gynecologic surgery before age 40 among daughters of young mothers

STUDY QUESTION Is being born of a young mother associated with worse gynecologic health, as indicated by a bilateral oophorectomy or hysterectomy before age 40? SUMMARY ANSWER Daughters of mothers younger than 25 did not have reduced parity but did have a higher risk of having bilateral oophorectomy or hysterectomy before age 40, particularly if their mother was younger than 20 years at their birth. WHAT IS KNOWN ALREADY Three recent studies have reported lower fecundability among daughters of mothers younger than 20 years; adverse socioeconomic conditions may explain part of that association. STUDY DESIGN, SIZE, DURATION This study reports cumulative, primarily retrospective, accrual of outcomes up to age 40 among 41 450 women recruited into the US-based Sister Study between 2003 and 2009. PARTICIPANTS/MATERIALS, SETTING, METHODS The analysis sample included women ≥41 years at the time of the latest follow-up and <66 years at recruitment. Using log-binomial regression, we estimated adjusted relative risks (RRs) of having major gynecologic surgery (bilateral oophorectomy or hysterectomy) before age 40 by age of the participant’s mother (G1) when she gave birth to the participant (G2). All models were adjusted for father’s age at G2’s birth, daughter’s self-identified race/ethnicity, and year of birth. We assessed possible effect modification by stratifying the analyses by self-reported G2’s family income level during childhood (poor-low, medium-high) and G2’s educational level (categorized as below bachelor’s degree and bachelor’s degree or higher) and, in the following step, by G2’s age at first birth. MAIN RESULTS AND THE ROLE OF CHANCE Compared with daughters born to mothers aged 30–34, daughters of mothers <20 and 20–24 years had an RR of 1.74 (95% CI 1.51, 2.00) and 1.35 (1.22, 1.50), respectively, of major gynecologic surgery before age 40. Although lower childhood income, G2 education, and giving birth before age 25 were strongly associated with outcome risk, the RRs changed little after accounting for those factors. LIMITATIONS, REASONS FOR CAUTION This is a descriptive study of a proxy indicator of poor gynecologic health. Furthermore, all information was self-reported and, for nearly all women, recalled after the event. The measures used for socioeconomic status may have been insufficient. WIDER IMPLICATIONS OF THE FINDINGS Daughters of younger mothers did not have reduced parity but appeared to have a higher risk of major gynecologic surgery before age 40. This study adds to prior evidence that daughters of young mothers have worse gynecologic health. STUDY FUNDING/COMPETING INTEREST(S) This research was supported in part by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences (Z01-ES044005, Z01-ES102245, and Z01-ES103086). The authors report no conflict of interest. TRIAL REGISTRATION NUMBER N/A.

young mother, early hysterectomy, gynecologic health, fibroids, intergenerational factors

Three recent studies have reported that daughters of mothers younger than 20 years have lower fecundability, compared with daughters of mothers ≥25 (Reynolds et al., 2020; Basso et al., 2021, 2022). Being small at birth, which, like preterm birth, is more common among children of adolescent mothers (Jeha et al., 2015; Amjad et al., 2019; DeMarco et al., 2021), has been linked to conditions that affect fertility, such as endometriosis (Upson et al., 2015; Gao et al., 2019), lower AMH levels (Dior et al., 2021), PCOS and hyperandrogenism (Sadrzadeh et al., 2017; Petraitiene et al., 2020), and early menopause (Bjelland et al., 2020). Birth before term has also been associated with a higher risk of uterine fibroids (D’Aloisio et al., 2010, 2012). While it is possible that these associations stem from biological mechanisms, children of adolescent mothers have a higher probability of experiencing social deprivation (Kahn and Anderson, 1992; Meade et al., 2008), an established risk factor for adverse health outcomes, as well as early age at natural menopause (Wise et al., 2002; Hardy and Kuh, 2005; Mishra et al., 2009; Schoenaker et al., 2014) and other conditions that may affect fertility, such as hypertension (Leng et al., 2015), type 2 diabetes (Agardh et al., 2011), and obesity (Newton et al., 2017). In this study, we first examined whether mother’s age at birth was associated with a daughter’s risk of undergoing hysterectomy or bilateral oophorectomy before age 40 (as a marker of poor gynecologic health) and then explored how childhood and adult indicators of socioeconomic status (SES) affected the association.

Study population This analysis was carried out in a subset of women taking part in the Sister Study, a cohort study of 50 884 women enrolled from the USA, including Puerto Rico, between 2003 and 2009. Women were eligible to enroll if they had not had a diagnosis of breast cancer, were between 35 and 74 years old, and had at least one half- or full-sister diagnosed with breast cancer. Participants answered different questionnaires (both computer-assisted interviews and self-administered) including a family history questionnaire with questions about the mother’s pregnancy. Every year, participants provide updates about changes in health and, approximately every 3 years, they are invited to complete comprehensive follow-up questionnaires. More details on the Sister Study are provided elsewhere (Sandler et al., 2017). Questionnaires are available on the Sister Study Website: https://sisterstudy.niehs.nih.gov/English/researchers.htm. This analysis, based on Sister Study data release 10.1, is restricted to participants who were ≥41 years at the latest comprehensive follow-up and younger than 66 at baseline, had not been adopted, and had filled out the family history questionnaire, yielding 42 137 eligible participants. We further excluded 1.7% of women with missing values in key variables, leaving 41 450 for analysis (Fig. 1). The Sister Study is overseen by the Institutional Review Board (IRB) of the National Institutes of Health. The IRB of the McGill University Health Centre approved the current project. Exposure, outcome, covariates Age of the mother (G1) when the participant (G2) was born was asked in the baseline family history questionnaire. (To avoid confusion between generations, G2 will refer to study participants and G1 to their mothers). In the analysis sample, G1’s age was reported in years by 92.1% and in 5-age categories by 5.5%. For 2.4%, it was imputed based on other information in the questionnaire (e.g. reported current age or age at death when the questionnaire was filled out or from information provided by a sister who also took part in the study). We examined G1’s age in categories, as shown in Table 1. Table 1. | G1 age at G2’s birth | ||||| |---|---|---|---|---|---| | <20 | 20–24 | 25–29 | 30–34 | ≥35 | | | n: | 1872 | 9619 | 12 498 | 9787 | 7674 | | G2 characteristics | % | % | % | % | % | | Race/ethnicity | ||||| | Non-Hispanic white | 69.4 | 82.6 | 86.3 | 85.6 | 83.0 | | Non-Hispanic Black | 19.5 | 9.4 | 6.9 | 7.4 | 9.2 | | Hispanic | 6.8 | 5.2 | 4.3 | 4.5 | 5.6 | | Other | 4.3 | 2.8 | 2.6 | 2.5 | 2.3 | | Family income growing upa | ||||| | Well off | 1.9 | 5.0 | 7.5 | 7.2 | 6.9 | | Medium | 52.6 | 62.5 | 63.6 | 62.7 | 57.6 | | Low | 33.6 | 25.8 | 22.8 | 23.5 | 26.6 | | Poor | 12.0 | 6.7 | 6.1 | 6.7 | 8.9 | | Participant’s educationb | ||||| | ≤High school | 21.4 | 15.5 | 13.1 | 13.3 | 13.3 | | Some college | 42.9 | 36.8 | 31.9 | 30.7 | 32.5 | | Bachelor degree | 19.5 | 25.7 | 28.7 | 29.7 | 29.1 | | MSc or PhD | 16.2 | 22.0 | 26.3 | 26.3 | 25.1 | | Age at menarchec | ||||| | <11 years | 9.8 | 7.0 | 6.8 | 6.7 | 6.6 | | 11 to <14 years | 69.7 | 69.1 | 69.7 | 70.0 | 71.4 | | ≥14 years | 20.5 | 23.9 | 23.6 | 23.2 | 22.0 | | Parityc | ||||| | 0 | 13.4 | 16.8 | 19.7 | 20.8 | 21.7 | | 1 | 16.7 | 16.2 | 15.1 | 15.6 | 14.9 | | ≥2 | 69.9 | 67.0 | 65.2 | 63.6 | 63.5 | | Age at first birthc | ||||| | ≤19 years | 32.6 | 17.2 | 13.9 | 13.0 | 13.6 | | 20–24 years | 36.9 | 39.0 | 34.3 | 33.0 | 32.9 | | 25–29 years | 7.4 | 11.8 | 14.6 | 16.0 | 16.4 | | ≥30 years | 10.1 | 16.5 | 20.6 | 23.2 | 24.0 | | Oophorectomy before age 40d | 6.5 | 4.8 | 3.9 | 3.6 | 3.9 | | Hysterectomy before age 40e | 13.3 | 9.2 | 7.5 | 6.1 | 6.8 | | Gynecologic disorders, any age | ||||| | Endometriosis | 20.5 | 18.9 | 18.2 | 17.5 | 17.2 | | Fibroids | 46.9 | 40.7 | 39.1 | 38.3 | 38.2 | | Cancer before age 40 | ||||| | Cervix, uterus, or ovary | 1.6 | 1.3 | 1.0 | 1.0 | 1.1 | | Other | 0.9 | 1.0 | 0.9 | 0.9 | 0.9 | | Sister diagnosed with ovarian or breast cancer before age 40f | 31.3 | 25.5 | 19.7 | 15.8 | 12.9 | | Prophylactic mastectomy or positive or indeterminate for BRCA1 or BRCA2 | 1.7 | 2.2 | 2.3 | 2.0 | 1.7 | Family income level during childhood was dichotomized as ‘poor/low income’ and ‘medium/high income’. Education was dichotomized as ‘less than college degree’ (≤High school or some college) and ‘college degree or higher’ (bachelor’s degree, MSc, or PhD). Variables with missing values: age at menarche (n: 30), parity (n: 38), age at first birth (n: 8, among participants with non-missing parity). With or without hysterectomy (only 37 out of 1716 had bilateral oophorectomy without hysterectomy). Information on oophorectomy was missing for 5 (who reported a hysterectomy before age 40). Hysterectomy leaving at least one ovary. To be eligible for participation in the Sister Study women had to have a sister who had been diagnosed with breast cancer. The women (G2) born to a mother (G1) under 20 were typically the older of the two sisters, while those born to a mother over 35 were typically the younger of the two, hence the linear decline across the G1 age categories in the proportion with a sister who had been diagnosed with ovarian or breast cancer under age 40. The outcome we studied in the G2 daughters was having had major gynecologic surgery (hysterectomy or bilateral oophorectomy) before the age 40. We chose this outcome because, regardless of the indication, such drastic surgery at a relatively young age would typically have been undertaken because of a serious reproductive tract medical problem (see Table 2). Such surgeries are also memorable events for which information was asked and probably reliably provided at baseline and at subsequent follow-ups. As secondary outcomes, we examined endometriosis and fibroids, which were reported at baseline and in follow-up interviews. Table 2. | Age at hysterectomy | |||||| |---|---|---|---|---|---|---| | <40 years | ≥40 years | ||||| | All | Parous | No births | All | Parous | No births | | | n | 4869 | 4110 | 756 | 7776 | 6510 | 1260 | | Abnormal bleeding | 76.2 | 76.6 | 74.0 | 78.0 | 78.9 | 72.7 | | Pelvic pain | 75.0 | 73.9 | 81.5 | 60.3 | 59.3 | 65.5 | | Both of the above | 58.4 | 58.0 | 60.6 | 48.4 | 48.7 | 46.5 | | Urinary incontinence | 13.3 | 14.4 | 7.2 | 18.4 | 19.8 | 10.9 | | Uterine prolapse | 16.7 | 18.8 | 5.0 | 15.7 | 17.9 | 3.3 | | Other | 21.6 | 20.8 | 26.0 | 21.7 | 20.3 | 29.2 | | No symptom reportedb | 8.2 | 7.8 | 9.9 | 17.8 | 16.6 | 23.7 | Percentages are calculated excluding from the denominator participants who reported no symptoms. 100% of participants with hysterectomy before age 40 (all were reported at enrollment) and 79.8% of those with hysterectomy at a later age. Most of these did not report that they had no symptom, but skipped the question. We approached this as a descriptive analysis, recognizing the potential for reverse causation between the variables examined (e.g. low parents’ SES during G2’s childhood may be both a cause and a consequence of G1’s giving birth at an early age; G2’s education is both a cause and a consequence of the age at which she first gave birth). Thus, we used modeling (and stratification) to describe how the association between G1’s age at birth and G2’s gynecological surgery before age 40 changed when accounting for potential outcome predictors associated with being born to a young mother. All models were adjusted for the following ‘base covariates’: daughter’s self-identified race/ethnicity, father’s age when the participant was born, and G2’s year of birth. The former two were categorized as in Table 1; the latter was modeled as a cubic spline with 3 knots. Except for father’s age, these factors are also potential risk factors for gynecologic surgery before age 40. Father’s age was included because it tends to be correlated with mother’s age and the interpretation of an association with the outcome as maternal age would not be clear without such adjustment. As indicators of childhood and adult SES, we used self-reported level of family income during G2’s childhood (poor/low and medium/high) and G2’s level of education (less than college degree and college degree or higher). In sensitivity analyses, we evaluated whether including the following prenatal and childhood/adolescence factors changed the relative risk (RR) estimates between young G1 age and the primary outcome: prenatal exposure to maternal smoking or diethylstilbestrol (DES), birthweight <5 lbs (∼2.5 kg), having been breastfed or having received soy formula as an infant, having had times without enough to eat during childhood, whether the father was listed as a legally responsible adult in the household when the G2 participant was 13 (categorized as in Supplementary Table S1), G2 age at menarche (categorized as in Table 1), and whether the G2 participant had herself given birth before age 20. Statistical analysis We used log-binomial regression models to estimate adjusted RRs and 95% confidence intervals for each category of maternal age. We also estimated the marginal predicted percents for each category of maternal age, obtained by summing across the entire sample the fitted probability for the outcome with that maternal age category imposed, retaining for each participant their covariates and the corresponding model-based estimated coefficients. (Subtracting each marginal predicted probability from that of the reference category yields the estimated risk difference and the ratio of the two corresponds to the RR). As the reference, we used G1’s age of 30–34 because it was the category with the lowest crude risk of gynecologic surgery before age 40 in G2. We first examined the association between mother’s age at G2’s birth and daughter’s gynecologic surgery before age 40, after checking that the estimates for young G1 age at birth were similar for pre-40 bilateral oophorectomy with or without hysterectomy (only 37 reported the latter) and pre-40 hysterectomy with retention of at least one ovary. To evaluate the contribution of SES, we estimated RRs and marginal probabilities in stratified analyses, first by family income level during G2’s childhood and then further divided by G2 education, yielding four subsets of childhood and adult SES. (In further analyses, instead of stratifying, we adjusted for childhood SES and adult education.) For each of the four subsets, we checked whether the estimated RRs changed when the early-life factors listed in the previous section were added to the models (singly or in groups of two). For the latter analysis, we used multiple imputation by chained equations (MICE) to generate 20 imputed datasets, as there was a high proportion of ‘don’t know’ and missing answers to the questions about whether the participant’s mother had smoked or taken DES during pregnancy. Daughters born to teenage mothers tended to themselves experience earlier menarche (Table 1); thus, based on the results of the above analysis, we re-estimated the association between G1 mother’s age and G2 gynecologic surgery before age 40 after stratifying by G2 age at first birth to allow for the possibility that early birth is to some extent a heritable phenotype. We restricted these analyses to participants with a first birth before age 30 (80.7% of the analytic sample gave birth at least once, of whom 79.7% had a first birth before age 30). We chose this threshold to avoid reverse causality, which would be an issue if the problems leading to early hysterectomy also interfered with the participants’ ability to have children. We then compared RRs in three groups: all women, those with two live births (the most common), and those with no births. All models were adjusted for base covariates and for proxies of SES, plus parity (in the analysis including all women) or age at first birth (in the analysis of women with two live births). We then excluded participants who may have been more likely to be monitored or to have a prophylactic oophorectomy (8033 with a sister diagnosed with breast or ovarian cancer before she (the sister) was age 40, plus 1125 who were either themselves diagnosed with cancer before age 40 or who, at any age, had a prophylactic mastectomy or tested positive or undetermined for a BRCA1 or BRCA2 mutation). In this subset, we ran the base model among all women and after restricting to participants with medium-high family income, a college degree or higher, and no pregnancy before the age 20. In imposing those exclusions for possible prophylaxis, we cast a wide net based on our retrospective data. Note that the G2 participants who had been born to young mothers were typically the oldest of the sisters, and the majority of the family events that might have led them to undergo prophylactic surgery occurred after they had passed 40, including a positive test for BRCA1 or BRCA2 (the testing for which first became available in 1996), and a sister’s or the mother’s diagnosis with breast or ovarian cancer. Thus, because of the relative timing of the events, any tendency in this cohort toward undergoing more prophylactic surgeries should have had little impact on the outcome we are analyzing, which was almost always pre-enrollment. Lastly, we examined the cumulative incidence of endometriosis and fibroids (regardless of age at diagnosis) as a function of G1’s age at G2’s birth. We show the results from (i) a model adjusted for base covariates, (ii) model (i) plus adjustment for family income in childhood and education level and, (iii), model (ii) plus additional early-life factors. For fibroids, we used Poisson regression with robust standard errors (Cummings, 2009) because that is a common outcome and log-binomial regression failed to converge.

Compared with participants born when the mother was 30–34 years old, daughters of mother <20 years were more likely to be non-white, to report low family income growing up, and to have a lower level of education (Table 1). They were also more likely to have given birth at least once, and more often before age 20. Supplementary Table S1 shows that daughters of mothers <20 were more likely to have been breastfed and to report that their family experienced times without enough to eat. They were less likely to have been prenatally exposed to smoking and DES and to report the father as a legally responsible adult in the household when they were 13. Overall, 4906 (11.8%) of the 41 450 women reported bilateral oophorectomy or hysterectomy before age 40; 1679 reported that they had had both (5 reported hysterectomy but no information on oophorectomy), 3185 had hysterectomy with at least one ovary retained, and 37 had only oophorectomy. Daughters born to mothers younger than 25 years had a substantially higher risk of each endpoint; daughters of mothers 25–29 and ≥35 also had a slightly higher risk of hysterectomy before age 40. In further analyses, we used ‘either surgery by age 40’ as the outcome and did not show estimates for father’s age (although it was retained in the model). Table 2 shows the distribution of pre-hysterectomy symptoms among women who reported a hysterectomy at baseline (i.e. all those who had surgery before age 40 and 79.8% of those with surgery at a later age). Of those who reported at least one symptom, abnormal bleeding and pelvic pain were the most common. A higher proportion of women with hysterectomy before age 40, compared to those with later surgery, reported pelvic pain, particularly if they had not given birth. Overall, endometriosis (at any age) was reported by 18.2% of participants (39.9% of those with surgery before age 40, 29.5 of those with later surgery, and 10.3% of those with no surgery). Uterine fibroids were reported by 39.5% (48.7% of those with surgery before age 40, 63.8% of those with later surgery, and 29.2% of those with no surgery). Among women with either surgery before age 40, 67.4% reported one or the other condition, versus 72.1% of the 9530 with surgery at a later age, and 33.8% of the 26 262 with no surgery. (Percentages include reports both at baseline (the majority) and during follow-up.) Compared with daughters born to mothers aged 30–34, daughters of mothers <20 and 20–24 years had a RR of 1.74 (95% CI 1.51, 2.00) and 1.35 (1.22, 1.50), respectively, of major gynecologic surgery before age 40 (Table 3). Although lower childhood income, G2 education, and giving birth before age 25 were strongly associated with outcome risk, the RRs changed little after accounting for those factors (see below). Table 3. | Bilateral oophorectomya | Hysterectomyb | Either surgery | ||||||| |---|---|---|---|---|---|---|---|---|---| | n: 41 445 | n: 41 450 | n: 41 450 | ||||||| | G1 age at G2’s birth (years) | RR | 95% CI | Margin (%)a | RR | 95% CI | Margin (%)a | RR | 95% CI | Margin (%)a | | Mother’s | ||||||||| | <20 | 1.71 | 1.32, 2.22 | 6.2 | 1.77 | 1.48, 2.11 | 11.2 | 1.74 | 1.51, 2.00 | 17.4 | | 20–24 | 1.33 | 1.11, 1.60 | 4.9 | 1.37 | 1.20, 1.56 | 8.7 | 1.35 | 1.22, 1.50 | 13.5 | | 25–29 | 1.08 | 0.93, 1.27 | 4.0 | 1.19 | 1.07, 1.33 | 7.5 | 1.15 | 1.05, 1.26 | 11.5 | | 30–34 | 1.00 | Reference | 3.6 | 1.00 | Reference | 6.3 | 1.00 | Reference | 10.0 | | ≥35 | 1.01 | 0.84, 1.20 | 3.7 | 1.14 | 1.00, 1.30 | 7.2 | 1.09 | 0.98, 1.20 | 10.9 | | Father’s | ||||||||| | <25 | 0.99 | 0.82, 1.19 | 4.0 | 1.11 | 0.98, 1.27 | 8.4 | 1.07 | 0.97, 1.19 | 12.3 | | 25–29 | 1.02 | 0.89, 1.18 | 4.1 | 1.01 | 0.94, 1.12 | 7.6 | 1.02 | 0.94, 1.10 | 11.7 | | 30–34 | 1.00 | Reference | 4.0 | 1.00 | Reference | 7.5 | 1.00 | Reference | 11.5 | | 35–39 | 0.99 | 0.84, 1.17 | 4.0 | 0.99 | 0.88, 1.12 | 7.5 | 0.99 | 0.90, 1.09 | 11.5 | | ≥40 | 1.19 | 0.98, 1.45 | 4.8 | 1.00 | 0.87, 1.15 | 7.5 | 1.07 | 0.96, 1.20 | 12.4 | Model additionally included race/ethnicity and year of birth (modeled as a cubic spline) of G2 participants. With or without hysterectomy (only 37 out of 1716 had bilateral oophorectomy without hysterectomy). Information on oophorectomy was missing for 5 (who reported a hysterectomy before age 40). Hysterectomy before age 40 leaving at least one ovary (reported by 3185). In analyses stratified by family income in childhood (Fig. 2, first row of models, Supplementary Table S2), the RRs are similar for medium/high (1.58 for <20 and 1.27 for 20–25) and low income (1.64 for <20 and 1.39 for 20–25) for G1 age 0.20), although the absolute risk (marginal predicted percent) is higher in the latter. The RRs changed little (but the confidence intervals grew wider, reflecting the smaller numbers) when the analysis was further stratified by education (Fig. 2, second and third rows of models), but the marginal predicted percents differed substantially across levels of education. When we added G2 early-life factors to the models (Supplementary Table S3), again the RRs stayed largely unchanged, with considerable overlap among the confidence intervals. When we stratified the analysis by G2’s age at first birth (among participants who gave birth before age 30), the RRs were again very similar across strata, but the absolute risk was substantially higher in participants who had themselves given birth before age 25 (Fig. 3, left panel). We excluded women who had only given birth after the age 30 to avoid reverse causality due to the development of the problems that could lead them to elect pre-40 hysterectomy. Adjusting for family income in childhood and education had minimal impact on the estimates. As shown in Table 4, after accounting for proxies of SES and parity (for all) or age at first birth (for women with two live births), RRs were similar among all participants, those with two live births (the most common number of children among participants), and nulliparas. Table 4. | Alla | Two live births | Nulliparas | ||||||| |---|---|---|---|---|---|---|---|---|---| | n: 41 412 | n: 15 950 | n: 8011 | ||||||| | G1 age at G2’s birth | RR | 95% CI | Margin (%)a | RR | 95% CI | Margin (%)a | RR | 95% CI | Margin (%)a | | Base model | ||||||||| | <20 | 1.75 | 1.52, 2.02 | 17.5 | 2.08 | 1.68, 2.58 | 20.4 | 1.63 | 1.10, 2.41 | 12.9 | | 20–24 | 1.36 | 1.22, 1.50 | 13.6 | 1.49 | 1.27, 1.74 | 14.6 | 1.38 | 1.07, 1.77 | 11.0 | | 25–29 | 1.15 | 1.05, 1.25 | 11.5 | 1.28 | 1.11, 1.46 | 12.5 | 1.13 | 0.91, 1.40 | 9.0 | | 30–34 | 1.00 | Reference | 10.0 | 1.00 | Reference | 9.8 | 1.00 | Reference | 8.0 | | ≥35 | 1.09 | 0.98, 1.20 | 10.9 | 1.15 | 0.97, 1.36 | 11.3 | 1.26 | 0.99, 1.60 | 10.0 | | +Childhood SES, education | +parityb | + age at 1st birthc | ||||||| | <25 | 1.51 | 1.31, 1.73 | 15.5 | 1.59 | 1.28, 1.96 | 16.6 | 1.47 | 1.00, 2.16 | 11.9 | | 25–29 | 1.27 | 1.15, 1.40 | 13.1 | 1.31 | 1.12, 1.53 | 13.7 | 1.31 | 1.02, 1.68 | 10.6 | | 30–34 | 1.13 | 1.03, 1.23 | 11.6 | 1.21 | 1.06, 1.38 | 12.6 | 1.12 | 0.90, 1.38 | 9.1 | | 35–39 | 1.00 | Reference | 10.3 | 1.00 | Reference | 10.4 | 1.00 | Reference | 8.1 | | ≥40 | 1.08 | 0.98, 1.20 | 11.2 | 1.15 | 0.98, 1.35 | 12.0 | 1.26 | 1.00, 1.59 | 10.2 | For all participants, models were Poisson, as binomial regression did not converge for the model adjusted for childhood SES, education, and parity. Parity was missing for 38 women (not included in any of the models) and was categorized as 0, 1, 2, and ≥3. Age at first birth was categorized as <20, 20–24, 25–29, and ≥30 years. When excluding participants who were more likely to have undergone a prophylactic oophorectomy or to have been closely monitored (because of personal or family cancer history, or because they had had a prophylactic mastectomy or a positive or inconclusive BRCA1 or BRCA2 genetic test), the association with young maternal age was still present, but as expected because of the reduction in sample size, the estimates were more imprecise (Supplementary Table S4). As shown in Fig. 4, the cumulative risk of endometriosis decreased with increasing mother’s age, and also after adjusting for SES and other early-life risk factors, although the association was very modest. For uterine fibroids, only the daughters of mothers <20 years had a slightly elevated risk, and adjustment for various risk factors barely changed the estimates.

In this study, daughters of mothers younger than 25 years had a higher risk of having had hysterectomy or bilateral oophorectomy before age 40, particularly if born when the mother was younger than 20 years. Low SES and, especially, having themselves given birth before age 25 predicted a higher absolute risk, although the estimated RRs were remarkably consistent across sub-analyses, including among nulliparas. Despite their clinical importance, bilateral oophorectomy and hysterectomy before the age 40 can occur for several reasons, making it difficult to hypothesize a mechanism for the association observed in this study. We did not ask about the specific indication for surgery, but the two most common reasons, especially for surgery before age 40, are likely to have been endometriosis and fibroids, often characterized by pelvic pain and abnormal bleeding: the most frequently reported pre-surgery symptoms in this study, especially among women with hysterectomy before age 40. The association of endometriosis with reduced fecundability (Filip et al., 2020) would provide a possible explanation for the previous findings among daughters of young mothers (Reynolds et al., 2020; Basso et al., 2021, 2022). Although endometriosis at any age was only weakly associated with mother’s age at birth, there was a clear trend that persisted after adjusting for SES and several early-life factors previously reported as being associated with endometriosis (Upson et al., 2015; Gao et al., 2019). Mother’s age <20 years was also weakly associated with fibroids. Among Sister Study participants, a modest association between mother’s age <20 and uterine fibroids before age 35 has been reported for white non-Hispanic women (D’Aloisio et al., 2010) and a stronger one has been reported among Black women (D’Aloisio et al., 2012); in both instances, the association was only seen among firstborn daughters. (In this study, the proportion of daughters of mothers <20 who had surgery before age 40 was 19.7% among firstborn children and 19.8% among non-firstborn children). One puzzling observation was that, among those who had either hysterectomy or bilateral oophorectomy before age 40, about a third allegedly had both. We suspect that many of those surgeries might in fact have removed the uterus and spared one of the ovaries. These data are based on self-report with no medical confirmation, with some participants answering questions many years after the surgery, and some may have forgotten that an ovary was retained. The loss of a uterus before the age 40 may simply be much more memorable than would be the simultaneous retention of an ovary. A study that queried women 2 weeks prior to their hysterectomy confirmed that many have an inaccurate understanding about what will be removed (Kassem et al., 2019). Strengths of this study include the detailed information on family factors, pregnancy history and health-related conditions, which was updated at several follow-ups. The sample permitted several subgroup analyses, allowing us to check the consistency of our findings across several potential effect modifiers. This study also has several limitations. All information, including that on age at surgery, was self-reported, for the most part at the baseline questionnaire. To reduce the potential for inaccurate recall, we restricted the analysis to women younger than 66 at baseline. Regarding mother’s age, imputed based on other information in the questionnaire for <3%, an add-on study of 1788 mothers of participants aged 35–59 years at cohort entry, suggested that the daughter’s report was reasonably good, within a year in 92.6% of instances, and within 2 years in 96.7% (unpublished results). In addition, errors in mother’s age are unlikely to correlate with the outcome. We defined SES based on two levels of self-reported family income during childhood and on two levels of education in adulthood. While these were admittedly crude markers of SES, they predicted absolute risk of gynecologic surgery before age 40, but their association with the outcome was weaker when the participant’s age at her own first birth was taken into account. We did not consider G1 education as an indicator of SES, because not pursuing a higher education may be a consequence of early parenthood. Indeed, the highest education in the household when the participant was 13 was ≤high school for 74.7% of those born to a mother younger than 20 years, and only 8.9% had completed college. Women taking part in the Sister Study are more educated, wealthier, and more were white than women in the underlying US population, as is the case in many volunteer cohorts (Sandler et al., 2017), but it is unclear whether this type of selection would distort an association between mother’s age at delivery and gynecologic surgery. The prevalence of hysterectomy (with or without bilateral oophorectomy) before age 40 in this study population was 11.7%, higher than the 7.5% and 7.1% reported in the 2006 and 2016 Behavioral Risk Factor Surveillance System, respectively (Harvey et al., 2022). However, women in the current analysis were older, and the prevalence of hysterectomy before age 40 was higher among older birth cohorts. For example, in the Sister Study cohort, it was 14.1% in the 50.2% born before 1953, compared with 9.4% among those born later. All Sister Study participants had a sister diagnosed with breast cancer, which could raise questions about generalizability if, for example, that history made them more likely to have had a hysterectomy or bilateral oophorectomy prior to age 40 prophylactically. However, the majority of the diagnoses in their sisters occurred after the participant had turned 40, so could only have influenced later surgical decisions and not the outcome we have analyzed. Moreover, the association with young maternal age was still present after excluding women with evident reasons for concern, plus those who themselves had a cancer diagnosis before age 40, a prophylactic mastectomy, or a positive or indeterminate BRCA1 or BRCA2 test result. Furthermore, the estimates were virtually the same when the outcome was hysterectomy leaving at least one ovary (Table 3).

In this descriptive analysis, daughters of mothers <25 years had a higher risk of gynecological surgeries before age 40, even after accounting for indicators of adverse socioeconomic conditions that were strongly associated with risk of the outcome. This study adds to previous evidence that daughters of young mothers have worse gynecologic health (Reynolds et al., 2020; Basso et al., 2021, 2022). Supplementary Material Contributor Information Clarice R Weinberg, Biostatistics and Computational Biology Branch, National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, NC, USA. Olga Basso, Department of Obstetrics and Gynecology, McGill University and Research Institute of McGill University Health Centre, Montreal, Canada; Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Canada. Aimee A D’Aloisio, Social & Scientific Systems, DLH Holdings Corporation, Durham, NC, USA. Dale P Sandler, Epidemiology Branch, National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, NC, USA. Data availability Data (and computing code) used in this analysis may be requested following posted procedures on the Sister Study website (https://sisterstudy.niehs.nih.gov/English/coll-data.htm). Authors’ roles O.B. had the idea for this analysis; she carried out the literature search, designed the study, analyzed the data, interpreted the results, and wrote the first draft of the paper. C.R.W. is the Sister Study co-principal investigator (PI); she contributed to data acquisition, study design, data analysis and interpretation, critically revised the article, and drafted the current revision. A.A.D. is a data manager for the Sister Study; she provided analytic datasets and contributed to the quality control and interpretation of the data, and critically revised the article. D.P.S. is the Sister Study PI; she contributed to the data acquisition, study design and interpretation of the data, and critically revised the article. Funding Intramural Research Program of the NIH (Z01-ES044005, Z01-ES102245, and Z01-ES103086). Conflict of interest The authors declare that they have no conflicts of interest.

- Agardh E, Allebeck P, Hallqvist J, Moradi T, Sidorchuk A. Type 2 diabetes incidence and socio-economic position: a systematic review and meta-analysis. Int J Epidemiol 2011;40:804–818. [DOI] [PubMed] [Google Scholar] - Amjad S, MacDonald I, Chambers T, Osornio‐Vargas A, Chandra S, Voaklander D, Ospina MB. Social determinants of health and adverse maternal and birth outcomes in adolescent pregnancies: a systematic review and meta‐analysis. Paediatr Perinatal Epidemiol 2019;33:88–99. [DOI] [PubMed] [Google Scholar] - Basso O, Magnus MC, Arge LA, Haberg SE. Parents’ age at birth and daughters’ time to pregnancy: a study within the Norwegian mother, father and child cohort. Hum Reprod 2022;37:1896–1906. [DOI] [PubMed] [Google Scholar] - Basso O, Willis SK, Hatch EE, Mikkelsen EM, Rothman KJ, Wise LA. Maternal age at birth and daughter’s fecundability. Hum Reprod 2021;36:1970–1980. [DOI] [PMC free article] [PubMed] [Google Scholar] - Bjelland EK, Gran JM, Hofvind S, Eskild A. The association of birthweight with age at natural menopause: a population study of women in Norway. Int J Epidemiol 2020;49:528–536. [DOI] [PMC free article] [PubMed] [Google Scholar] - Cummings P. Methods for estimating adjusted risk ratios. Stata J 2009;9:175–196. [Google Scholar] - D’Aloisio AA, Baird DD, DeRoo LA, Sandler DP. Association of intrauterine and early-life exposures with diagnosis of uterine leiomyomata by 35 years of age in the Sister Study. Environ Health Perspect 2010;118:375–381. [DOI] [PMC free article] [PubMed] [Google Scholar] - D’Aloisio AA, Baird DD, DeRoo LA, Sandler DP. Early-life exposures and early-onset uterine leiomyomata in black women in the Sister Study. Environ Health Perspect 2012;120:406–412. [DOI] [PMC free article] [PubMed] [Google Scholar] - DeMarco N, Twynstra J, Ospina MB, Darrington M, Whippey C, Seabrook JA. Prevalence of low birth weight, premature birth, and stillbirth among pregnant adolescents in Canada: a systematic review and meta-analysis. J Pediatr Adolesc Gynecol 2021;34:530–537. [DOI] [PubMed] [Google Scholar] - Dior UP, Karavani G, Soloveichick V, Friedlander Y, Hochner H. Early-life factors and adult anti-Müllerian hormone levels. J Assist Reprod Genet 2021;38:3019–3025. [DOI] [PMC free article] [PubMed] [Google Scholar] - Filip L, Duică F, Prădatu A, Crețoiu D, Suciu N, Crețoiu SM, Predescu D-V, Varlas VN, Voinea S-C. Endometriosis associated infertility: a critical review and analysis on etiopathogenesis and therapeutic approaches. Medicina 2020;56:460. [DOI] [PMC free article] [PubMed] [Google Scholar] - Gao M, Allebeck P, Mishra GD, Koupil I. Developmental origins of endometriosis: a Swedish cohort study. J Epidemiol Community Health 2019;73:353–359. [DOI] [PMC free article] [PubMed] [Google Scholar] - Hardy R, Kuh D. Social and environmental conditions across the life course and age at menopause in a British birth cohort study. BJOG 2005;112:346–354. [DOI] [PubMed] [Google Scholar] - Harvey SV, Pfeiffer RM, Landy R, Wentzensen N, Clarke MA. Trends and predictors of hysterectomy prevalence among women in the United States. Am J Obstet Gynecol 2022;227:e611.e2. [DOI] [PMC free article] [PubMed] [Google Scholar] - Jeha D, Usta I, Ghulmiyyah L, Nassar A. A review of the risks and consequences of adolescent pregnancy. J Neonatal-Perinat Med 2015;8:1–8. [DOI] [PubMed] [Google Scholar] - Kahn JR, Anderson KE. Intergenerational patterns of teenage fertility. Demography 1992;29:39–57. [PubMed] [Google Scholar] - Kassem Z, Coleman CM, Bossick AS, Su W-T, Sangha R, Wegienka G. Patient perceptions of planned organ removal during hyterectomy. J Patient Cent Res Rev 2019;6:28–35. [DOI] [PMC free article] [PubMed] [Google Scholar] - Leng B, Jin Y, Li G, Chen L, Jin N. Socioeconomic status and hypertension: a meta-analysis. J Hypertens 2015;33:221–229. [DOI] [PubMed] [Google Scholar] - Meade CS, Kershaw TS, Ickovics JR. The intergenerational cycle of teenage motherhood: an ecological approach. Health Psychol 2008;27:419–429. [DOI] [PubMed] [Google Scholar] - Mishra GD, Cooper R, Tom SE, Kuh D. Early life circumstances and their impact on menarche and menopause. Womens Health (Lond) 2009;5:175–190. [DOI] [PMC free article] [PubMed] [Google Scholar] - Newton S, Braithwaite D, Akinyemiju TF. Socio-economic status over the life course and obesity: systematic review and meta-analysis. PLoS One 2017;12:e0177151. [DOI] [PMC free article] [PubMed] [Google Scholar] - Petraitiene I, Valuniene M, Jariene K, Seibokaite A, Albertsson-Wikland K, Verkauskiene R. Sex hormones, gonad size, and metabolic profile in adolescent girls born small for gestational age with catch-up growth. J Pediatr Adolesc Gynecol 2020;33:125–132. [DOI] [PubMed] [Google Scholar] - Reynolds TS, Lynch CD, Hade EM, Allain DC, Westman JA, Toland AE. Maternal age at delivery and fertility of the next generation. Paediatr Perinat Epidemiol 2020;34:629–636. [DOI] [PubMed] [Google Scholar] - Sadrzadeh S, Hui E, Schoonmade L, Painter R, Lambalk C. Birthweight and PCOS: systematic review and meta-analysis. Hum Reprod Open 2017;2017:hox010. [DOI] [PMC free article] [PubMed] [Google Scholar] - Sandler DP, Hodgson ME, Deming-Halverson SL, Juras PS, D’Aloisio AA, Suarez LM, Kleeberger CA, Shore DL, DeRoo LA, Taylor JA et al. ; Sister Study Research Team. The Sister Study cohort: baseline methods and participant characteristics. Environ Health Perspect 2017;125:127003. [DOI] [PMC free article] [PubMed] [Google Scholar] - Schoenaker DA, Jackson CA, Rowlands JV, Mishra GD. Socioeconomic position, lifestyle factors and age at natural menopause: a systematic review and meta-analyses of studies across six continents. Int J Epidemiol 2014;43:1542–1562. [DOI] [PMC free article] [PubMed] [Google Scholar] - Upson K, Sathyanarayana S, Scholes D, Holt VL. Early-life factors and endometriosis risk. Fertil Steril 2015;104:964–971.e5. [DOI] [PMC free article] [PubMed] [Google Scholar] - Wise L, Krieger N, Zierler S, Harlow B. Lifetime socioeconomic position in relation to onset of perimenopause. J Epidemiol Commun Health 2002;56:851–860. [DOI] [PMC free article] [PubMed] [Google Scholar] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Data Availability Statement Data (and computing code) used in this analysis may be requested following posted procedures on the Sister Study website (https://sisterstudy.niehs.nih.gov/English/coll-data.htm).