Parental infertility and sexual maturation in children

Abstract

BACKGROUND

The reproductive health of children born of infertile couples may be affected by infertility treatment or factors associated with infertility. We examined sexual maturation in children of parents with infertility.

METHODS

We used data from a follow-up of 3382 girls and 2810 boys born between 1984 and 1987 in the Aalborg–Odense Birth Cohort. We had mothers’ report of time to pregnancy (TTP) and infertility treatment (at the time, mostly hormonal) from the pregnancy questionnaire administered in 1984–1987, and the children’s report of their own sexual maturation from the follow-up questionnaire administered in 2005, when they were between 18 and 21 years old. Many reported age only in year when they had the events related to sexual maturation, and for each event, we imputed the month based on the median month at each year of age among those reporting both years and months.

RESULTS

In girls, the mean age at menarche was 13.3 years and, in boys, the mean age at appearance of acne, voice break, regular shaving and first nocturnal emission were 14.5, 14.5, 17.2 and 14.7 years, respectively. We saw no significant differences in age at these events among children born of either fertile (with TTP of 0–12 months and no treatment), untreated infertile (with TTP of more than 12 months and no treatment) or treated infertile couples (with a history of examination or treatment for infertility).

CONCLUSIONS

Our data suggest no significant association between parental infertility or hormonal treatment and timing of sexual maturation in the offspring.

Introduction

Development of the endocrine system and maturation of the endocrine-control systems are influenced by intrauterine hormone concentrations (Rojas-Marcos et al., 2005), and it is possible that hormonal profiles in pregnant women depend on their ability to conceive. Besides age-related ovarian failure, infertility is most likely the result of genetic predisposition in combination with environmental factors (e.g. hormones or endocrine disruptors, other reproductive toxicants, infections). Although infertility treatment has been associated with increased risk of hypospadias in the offspring (Silver et al., 1999; Zhu et al., 2006; Elizur and Tulandi, 2008), this association may reflect the underlying fertility problems rather than the treatment, as fathers of boys with hypospadias showed an increased frequency of reproductive health problems, including low semen quality (Asklund et al., 2007).

Altered timing of puberty is included in the US Environmental Protection Agency’s programs for children’s health-risk assessment, and children with early puberty are at increased risk of accelerated skeletal maturation and short adult height, early sexual debut, psychosocial problems and, possibly, even development of reproductive tract cancers later in life (Golub et al., 2008). Early development of breasts or pubic hair has been suspected as a side effect of assisted reproductive technology (ART) (Rojas-Marcos et al., 2005), but one study reported no association between ICSI and Tanner stages (physical measurements of development of external primary and secondary sex characteristics, including breasts, genitalia and pubic hair) by the age of 8 years (Belva et al., 2007). The early age at assessment may, however, explain the negative findings. In this study, we examined whether parental infertility was associated with age at appearance of puberty indicators, using a cohort design.

Materials and Methods

Study population

We used data from the Aalborg–Odense Birth Cohort (Olsen et al., 1989; Magnusson et al., 2005), which started as a health campaign (‘Healthy Habits for Two’) aimed at all pregnant women giving birth at the hospitals in Aalborg and Odense, Denmark, between April 1984 and April 1987 (Olsen et al., 1989). Midwives handed out a self-administered questionnaire at the routine antenatal visit in the 36th gestational week, and a total of 11 980 pregnant women (87% of the eligible) in the two regions took part in the study. We have data on parents’ socio-demographic, environmental and lifestyle factors from the questionnaire.

In 2005 (when the children were between 18 and 21 years old), a follow-up questionnaire on health and lifestyle, including questions on sexual maturation, was administered through the internet to 9811 adolescents who were still alive and living in the country. They were introduced to the online questionnaire by a letter mailed to their address, and two reminders were sequentially sent if they had not responded to the questionnaire after 4 weeks. The second reminder also included a paper version of the questionnaire. A total of 6195 adolescents (including 913 who used the paper version) responded (63%). We excluded three adolescents who, probably by mistake, answered that they were females whereas their civil registration number indicated that they were males.

Time to pregnancy and infertility treatment

In the pregnancy questionnaire, mothers were asked if their pregnancy was planned (yes, no or do not know), if they had ever been examined or treated for infertility (yes or no) and for how long they had tried to become pregnant before succeeding (three categories: 0–6, 7–12 and >12 months). We defined as fertile couples those who planned their pregnancy and conceived within 12 months without having been examined or treated for infertility. Couples who planned their pregnancy but took longer than 12 months to conceive without having been examined or treated for infertility were labeled as untreated infertile, whereas couples who reported having been examined or treated for infertility were considered as treated infertile. Because the question on infertility examination or treatment was not restricted to the pregnancy under study, this group included all women with a history of infertility examination or treatment. Couples who did not plan their pregnancy and who had not been examined or treated for infertility were classified as a group called ‘unplanned pregnancy’. The latter group also included 36 couples with planned pregnancies but without information on time to pregnancy (TTP) or who reported that they conceived while using contraception, 26 couples who did not provide information on treatment and 232 couples who did not provide information on pregnancy planning.

Sexual maturation

In the follow-up questionnaire, girls were asked the following question about timing of menarche ‘Have your periods started?’ with three answer categories: ‘have not started yet’, ‘yes’ and ‘do not know’. If they answered ‘yes’, they were asked to provide their age (in years and months) at menarche. Boys were asked about five events related to puberty: ‘Do you have pubic hair?’, ‘Have you had acne?’, ‘Has your voice broken?’, ‘Have you started to shave regularly?’ and ‘Have you had your first nocturnal emission (wet dream)?’. There were three or four answer categories: ‘no’, ‘yes’, ‘prefer not to answer’ and ‘do not remember’. Those who answered ‘yes’ were also asked to provide age (in years and months) at which the event first occurred (except for the appearance of pubic hair, which we did not include in our analysis).

A number of boys answered that they had experienced the events in question but did not provide the age at which the events started (Table I). The proportions of boys who provided an age for 1, 2, 3 or 4 events were 16.0%, 22.6%, 32.5% and 18.6%, respectively, and 10.3% of boys did not provide an age for any of these events. Furthermore, about three-quarter of the boys, as well as about half of the girls, provided age only in years. There appeared to be little difference in the distribution of age in years for these events between those who reported age only in years and those who reported age in years and months. For each event, we, therefore, imputed the month based on the median month at each year of age among those who reported age in years and months, as the reports in years only would affect the mean age (and SD) for each event. We used these data in the analysis.

Table I.

Questionnaire responses according to pregnancy planning status and TTP

	TTP 0–12 months		TTP >12 months		Examined or treated for infertility		Unplanned
	n	%	n	%	n	%	n	%
Total	3896	100.0	497	100.0	496	100.0	1303	100.0
Girls	2068	53.1	296	59.6	286	57.7	732	56.2
Boys	1828	46.9	201	40.4	210	42.3	571	43.8
Girls
Menarche
Yes, age provided	1940	93.8	273	92.2	268	93.7	688	94.0
Yes, no age provided	111	5.4	23	7.8	15	5.2	41	5.6
No	8	0.4	0	0.0	2	0.7	1	0.1
Do not know or no answer	9	0.4	0	0.0	1	0.3	2	0.3
Boys
Acne
Yes, age provided	1153	63.1	128	63.7	147	70.0	376	65.8
Yes, no age provided	291	15.9	25	12.4	30	14.3	97	17.0
No	284	15.5	34	16.9	28	13.3	75	13.1
Prefer not to answer or no answer	100	5.5	14	7.0	5	2.4	23	4.0
Voice break
Yes, age provided	1095	59.9	130	64.7	132	62.9	339	59.4
Yes, no age provided	613	33.5	49	24.4	63	30.0	190	33.3
No	93	5.1	18	9.0	12	5.7	25	4.4
Prefer not to answer or no answer	27	1.5	4	2.0	3	1.4	17	3.0
Regular shaving
Yes, age provided	1380	75.5	159	79.1	163	77.6	426	74.6
Yes, no age provided	254	13.9	19	9.5	26	12.4	74	13.0
No	187	10.2	22	10.9	21	10.0	70	12.3
Prefer not to answer or no answer	7	0.4	1	0.5	0	0.0	1	0.2
First nocturnal emission
Yes, age provideda	606	33.2	56	27.9	72	34.3	190	33.3
Yes, no age provided	794	43.4	94	46.8	87	41.4	252	44.1
No	367	20.1	43	21.4	41	19.5	103	18.0
Prefer not to answer or no answer	61	3.3	8	4.0	10	4.8	26	4.6

^aIncluding 851 (92.1%) boys who answered yes to the question ‘Have you had your first nocturnal emission?’ and provided the age at first nocturnal emission and 73 (7.9%) boys who answered no to the question but provided the age at first nocturnal emission. There were no differences in the distribution of age at first nocturnal emission between these two groups of boys.

Statistical analysis

Age at each event was approximately normally distributed, and we calculated mean age (SD) for each event. We used Cox regression to compare the age at the occurrence of the event for children born of untreated or treated infertile couples with that of children born of fertile couples (the reference category) in STATA 9.1 (StataCorp, College Station, TX, USA). Follow-up started at birth and ended at the age at which the event happened or at the end of follow-up, whichever came first. Participants who had not experienced the event at the time when they answered the questionnaire were also included in the Cox regression models, and follow-up was censored at the age at which they answered the questionnaire in 2005. We used the ‘efron’ option, which provides approximation of the exact marginal log-likelihood, to treat tied events in the calculation of the log partial likelihood (STATA, 2003). Since 2.1% of the women contributed two children to the cohort, we used a robust variance estimator, similar to the so-called ‘sandwich’ methods, to calculate confidence intervals (Li and Wei, 1989; STATA, 2003). We further checked if age at the event was associated with TTP by restricting our analysis to children born of untreated couples with planned pregnancy (fertile couples and untreated infertile couples), and treating categories of TTP as ordinal numbers in the Cox regression model. In addition, we used as a cut-off point the integer year (avoiding no reporting in month) nearest in terms of distribution to the 10% cut-off at the low or high end of the age distribution for each event to define early or late event, respectively, and these dichotomized outcomes were analyzed with logistic regression.

Potential confounders included maternal age at delivery (<25, 25–29, 30+ years), smoking during pregnancy (yes, no), pre-pregnancy BMI (<18.5, 18.5–24, 25+ kg/m²), municipality of residence at the time of delivery (Odense–Aalborg, other places), and family’s socioeconomic status (skilled workers, unskilled or semi-skilled workers, unemployed, students) based on the highest ranking between the mother and the father at the time of delivery (Olsen and Frische, 1993). Parity is a potential confounder for infertility treatment but not for infertility per se, and we thus calculated estimates with and without adjustment for parity. Since adjustment for parity did not change the estimates; however, we present results without adjustment for parity. We also checked if further adjustment for the children’s BMI changed the estimates among those who provided the information on weight and height in the follow-up questionnaire. Children’s age at the time of questionnaire administration was similar across the exposure groups and was thus not included as a covariate.

Results

The study included 3382 girls and 2810 boys. The proportion of girls who provided age at menarche was 93.7%, and the proportions of boys who provided an age at acne, voice break, regular shaving and first nocturnal emission were 64.2%, 60.4%, 75.7% and 32.9%, respectively (Table I). In girls, mean age at menarche was 13.3 years (SD 1.4). In boys, mean age at acne, voice break, regular shaving and first nocturnal emission were 14.5 (SD 1.5), 14.5 (SD 1.4), 17.2 (SD 1.4) and 14.7 (SD 1.9) years, respectively.

Regardless of adjustment, we saw no significant differences in age at menarche among girls who were born of fertile couples, untreated infertile couples, infertile couples who sought medical help or couples who did not plan the pregnancy. We also saw no significant differences in age at acne, voice break, regular shaving and first nocturnal emission among different groups of boys, except that boys born of couples who had not planned their pregnancy showed slightly earlier age at acne, voice break and first nocturnal emission (Table II). Furthermore, we found no association between age at each event and TTP (Table III).

Table II.

Hazard ratio (HR) of occurrence of the event according to pregnancy planning status and TTP

	TTP >12 months		Examined or treated		Unplanned
	Crude HR	Adjusted HRa (95% CI)	Crude HR	Adjusted HRa (95% CI)	Crude HR	Adjusted HRa (95% CI)
Girls
Menarche	1.08	1.06 (0.93–1.21)	1.09	1.08 (0.94–1.24)	1.03	1.01 (0.93–1.10)
Boys
Acne	0.97	0.99 (0.82–1.19)	1.09	1.13 (0.95–1.35)	1.12	1.16 (1.03–1.31)
Voice break	0.84	0.85 (0.71–1.02)	1.07	1.04 (0.85–1.27)	1.09	1.12 (0.98–1.27)
Regular shaving	1.03	1.03 (0.86–1.23)	1.00	0.97 (0.82–1.15)	0.95	0.97 (0.87–1.08)
First nocturnal emission	0.91	0.92 (0.69–1.24)	1.07	1.08 (0.84–1.40)	1.09	1.14 (0.96–1.36)

Fertile couples with TTP of 0–12 months used as reference. Cox regression: data included those who had experienced the event and provided the age at which she/he had the event and those who had not experienced the event (censored at the age at the time of interview). CI, confidence interval.

^aAdjusted for maternal age, smoking, pre-pregnancy BMI, municipality and family’s socioeconomic status, all measured during pregnancy.

Table III.

HR of occurrence of the event among children born of fertile and untreated infertile couples, according to TTP

	TTP 7-12 months		TTP >12 months
	Crude HR	Adjusted HRa (95% CI)	Crude HR	Adjusted HRa (95% CI)
Girls
Menarche	1.00	1.00 (0.88–1.14)	1.08	1.07 (0.93–1.22)
Boys
Acne	1.08	1.08 (0.91–1.30)	0.98	0.99 (0.82–1.20)
Voice break	1.05	1.05 (0.88–1.25)	0.84	0.86 (0.71–1.03)
Regular shaving	1.10	1.10 (0.93–1.30)	1.04	1.04 (0.87–1.24)
First nocturnal emission	1.07	1.07 (0.84–1.38)	0.92	0.93 (0.69–1.25)

TTP of 0–6 months used as reference. Cox regression: data included those who had experienced the event and provided the age at which she/he had the event and those who had not experienced the event (censored at the age at the time of interview).

^aAdjusted for maternal age, smoking, pre-pregnancy BMI, municipality and family’s socioeconomic status, all measured during pregnancy.

Categorizing age at each event by using a cut-off of ∼10% at the low or high end of the age distribution (in integer year) revealed similar results (data not shown). Neither restricting the analysis to the age determined only by year (i.e. ignoring month) nor restricting the analysis to those who provided complete information on age (i.e. both year and month) changed the results substantially (data not shown).

In a subgroup of children who provided information on weight and height, which were measured at the same time or <6 months apart during 2003–2005 (n = 1303 girls and n = 1691 boys), further adjustment for BMI did not change the estimates, except that the slightly earlier age at acne, voice break and first nocturnal emission seen in boys born of couples who had not planned the pregnancy disappeared (data not shown).

Discussion

In this analysis, we saw no association of parental infertility or infertility treatment with age at appearance of markers of sexual maturation, which is in line with results from a Belgian study in 8 year olds following ICSI (Belva et al., 2007). We observed that boys born of couples who had not planned their pregnancy may have a slightly earlier sexual maturation, which could reflect an association with parental fecundity, a result of uncontrolled confounding, or simply a chance finding.

The participants in this study were 18–21 years of age. The birth cohort was established two decades ago with a high participation rate (87%), whereas the response rate at follow-up was relatively low (63%). If participation rates depended on both parental infertility and altered sexual maturation, our estimates could be biased. Participation among adolescents born of untreated infertile couples (60.2%) or born of couples with unplanned pregnancy (61.6%) was slightly lower than those of the other groups (63.8–65.4%) (P = 0.041). We have no information on the maturation pattern of the non-participants, but we checked whether participation in girls differed as a function of size at birth, which predicts age at menarche (Adair, 2001). Girls who were relatively long [>52 cm (median birth length)] and thin at birth [<3480 g (median birthweight)] participated in the follow-up in a similar proportion as the rest (70.3% versus 72.7%, P = 0.33). Participation of long and thin girls was also similar across the exposure groups of interest (66.7–80.0%, P = 0.63). Owing to recruitment of mothers around the 36th week of pregnancy, our study population had few children born preterm (2.5% born before 37 weeks of gestation), which limits our ability to assess a possible effect in this group.

We collected information on TTP in three categories to facilitate women’s recall, which did in turn led to a loss in precision of the estimates that is, however, likely to be modest. We had no pregnancy-specific information on infertility examination and treatment for the women and no information on the male partner. Use of estrogen to stimulate ovulation was the treatment of choice for female infertility in Denmark in the 1980s. Thus, our assessment about impact of infertility treatment is limited.

More girls responded (and provided more complete information) than boys but, in both sexes, lack of completeness in reporting did not depend on parental infertility status. The large number of missing data suggests that, after 5–10 years, boys do not accurately remember when the investigated events happened, which may in part be due to the fact that these events have a less clear time definition than age at menarche. Recording of puberty in real time is needed to get data with better precision.

Age at menarche has been extensively studied as a marker of puberty for girls, and the mean age of 13.3 years in our study was comparable with the age of 13.0–13.4 years reported by studies among adolescents born in the 1970s or 1980s in Denmark (Helm and Helm, 1984; Boldsen et al., 1993; Helm and Grolund, 1998; Juul et al., 2006). Sexual maturation in boys has been much less studied, and the mean age of 14.5 years at voice break in our study is slightly higher than 14.0 years reported for Danish boys admitted to the Copenhagen Municipal Choir School during 1990–1999 (Juul et al., 2007). As a daily practice in the choir school, the conductor of the choir records voice break when it happens. The difference from our data could, therefore, be due to information bias (self-perception and recall). Age at menarche appears to be heritable (Herbison, 2007; Ong et al., 2007), and timing of puberty in boys may also be under genetic control. However, we had no information on the age of parental sexual maturation, and we were thus unable to take this aspect into consideration.

The increasing number of infertile couples who become parents leads to concerns about risk of reproductive problems in their offspring. Jensen et al. reported that men of mothers who received infertility treatment had poor semen quality (Jensen et al., 2007), and similar results have been found in the Aalborg–Odense Birth Cohort (Ramlau-Hansen et al., 2007, 2008). However, significant changes in reproductive hormones in men of mothers who were examined or treated for infertility have not been reported (Jensen et al., 2007; Ramlau-Hansen et al., 2008).

Our data suggest no association between parental infertility or hormonal treatment and sexual maturation in children. Denmark has the highest proportion of ART usage in the world (Andersen and Erb, 2006), which may reflect major fertility problems. Our findings should, thus, be interpreted with caution, as they may not apply to other populations. Moreover, other reproductive events, including pregnancy outcomes among offspring of infertile couples, should be evaluated as well as sexual maturation in offspring born after ART.

Author’s Role

Contributions to conception and design, data interpretation, paper revision and final approval of the submitted version: J.L.Z., O.B., C.O., B.H.B., E.A.N., A.S., J.O.

Acquisition of data: J.L.Z., B.H.B., E.A.N., J.O.

Data analysis: J.L.Z., B.H.B., E.A.N., J.O.

Drafting the article: J.L.Z.

Funding

This work was supported by a grant from the Danish Medical Research Council (271-05-0115) and, in part, by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences (Z01 ES044003).

Conflict of interest: none.