Study Suggests Infertility and Use of Fertility Drugs Increases Risk for Autism

Daniel M. Keller, PhD

May 26, 2010

May 26, 2010 (Philadelphia, Pennsylvania) — Preliminary results of a study on a cohort of women in the Nurses Health Study II suggest that a history of fertility problems and the use of ovulation-inducing drugs (OIDs) are associated with an almost 2-fold increased risk of having a child with autism spectrum disorder (ASD).

Led by Kristen Lyall, ScD, a postdoctoral research fellow at the Harvard School of Public Health in Boston, Massachusetts, the investigators compiled data from 3985 cohort participants who had their first child between 1993 and 2003 and completed a questionnaire in 2005 that included information about ASD. The women self-reported their history of infertility and use of OIDs.

Dr. Lyall told the 9th Annual International Meeting for Autism Research (IMFAR) audience here that despite 2 previous studies showing an association of a history of infertility with autism, still little is known about it, studies assessing fertility therapies have given inconclusive results, and none investigated OID use.

The Nurses Health Study II enrolled 116,608 woman aged 25 to 42 years in 1989 and mailed them questionnaires every 2 years. The nurses reported infertility and OID use (oral or injected) on each questionnaire in 1993, 1995, 1997, and 2001. "Infertility is defined as attempting to get pregnant for more than 12 months without success," Dr. Lyall said.

In 2005, they were asked if any of their children had been diagnosed as having autism, Asperger's syndrome, or other ASD. Because it was not possible to determine from the questionnaire which child was affected if the woman had multiple children, OID exposures only before the first pregnancy were considered to make sure that exposure occurred before any of the children were born.

Of the 3985 cohort participants who met the research criteria and returned the 2005 questionnaire, 111 reported a child with ASD. The women's average age at baseline was 29.1 years, the women's average age at first birth was 35.5 years, and 94% were white. Women with twin births were excluded from the analysis.

Adjusted odds ratios for the cohort demonstrated a significantly elevated risk for a history of infertility and OID use with having a child with ASD.

Table. Risk for ASD by OID Use and Infertility Status (N = 3985)

Variable No. of ASD Cases Adjusted Odds Ratio* P Value
No prior OID or INF 57 1.0  
INF only 16 1.58 .12†
INF and OID 36 1.91 .007†
INF vs no INF 52 1.81 .005
OID vs no OID 38 1.58 .04
Stratified: effect of OID among women with INF 36 1.28 .44

ASD = autism spectrum disorder; INF = infertility; OID = ovulation-inducing drug

*Adjusted for nurses' age at baseline, age at first birth, parity, race, income, marital status, spouse's education, pregnancy complications, twin births, prior miscarriages, prior induced abortions, and body mass index.

†Compared with no prior OID or INF.

Breaking out the subgroup of uniparous women (n = 1769 with 46 ASD cases) and comparing it to the entire cohort allowed for a child-specific analysis, Dr. Lyall reported that the uniparous subgroup showed results similar to the group overall.

Although the power to study the cumulative effect of OID use was limited, there was a trend toward greater risk for ASD for women who reported OID use on 2 questionnaires or more vs women with no OID use.

"Thinking about potential explanations, there may be a biological effect on fetal development perhaps through a hormonal pathway," Dr. Lyall said. "We also cannot rule out a potential confounding effect perhaps of infertility itself." She also speculated that genetic factors diminishing fertility may raise the risk for ASD or that OID use could mask effects of other forms of fertility therapy (eg, in vitro fertilization).

Some limitations of the study are the small case number, that it relied on maternal reports of OID exposure and ASD diagnoses, and that there was no child-specific information or information on possible confounding factors (eg, paternal age, maternal psychiatric history, or the birth order of an affected child) or other fertility therapies. Dr. Lyall cautioned that the results may not be generalizable to other groups of women, such as younger age at first birth, other races, or income levels.

Although these preliminary results suggest an association among a history of infertility, OID use, and the risk of having a child with ASD, she said the observations are insufficient to recommend any change in current clinical practice "and should not be used to cause undue anxiety in women who have used such treatments."

Epidemiologist Irva Hertz-Picciotto, PhD, MPH, professor of public health sciences at the University of California, Davis, MIND Institute, wondered, "If somebody reported ‘No, I never had a problem with infertility' but then they did take the ovulation-inducing drug, what would that mean and were there discrepancies?" She said she would want a better explanation of this point about the study.

Despite the limitations of the study, particularly regarding child-specific data and other variables that could not be discerned from the questionnaires, Dr. Hertz-Picciotto said, "They did what they could to try to make it as valid as possible...They seemed to have taken a lot of care in the analysis, and Donna Spiegelman, who is the statistician on that and was named as an author, is a very highly qualified biostatistician."

Dr. Lyall and Dr. Hertz-Picciotto have disclosed no relevant financial relationships.

9th Annual International Meeting for Autism Research (IMFAR): Abstract 103.002. Presented May 20, 2010.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: