Early studies of women with anorexia nervosa suggested that pregnancy was rare secondary to endocrinological disturbances associated with starvation coupled with the psychological and psychosocial features of the disorder.1,2 An estimated 68% to 89% of women with anorexia nervosa presentations report amenorrhea (criterion D, ie, absence of menstruation for at least 3 months) during their illness.3–6 and an additional 6% to 8% report oligomenorrhea.6 The high prevalence of menstrual disturbance has contributed to the (mis)conception that women with anorexia nervosa are unlikely to conceive.
Although results are mixed,7 several outcome studies of women with anorexia nervosa suggest that there are no differences in fertility rates in women with a history of anorexia nervosa and women in the general population.8,9 In addition, rates of fertility treatment in women with anorexia nervosa may not differ from those of the general population.8 Thus, despite high rates of menstrual dysfunction, women with anorexia nervosa are becoming pregnant.
Conception can occur in the absence of menstruation. When ovulation occurs for the first time after a period of amenorrhea and anovulation, the first egg can become fertilized. If fertilization occurs, then maintenance of the embryo will prevent shed of endometrial tissue and ovulation will not be followed by menstruation. In this rare case of fertilization on first appearance or reappearance of ovulation, a woman can become pregnant without previous menstruation. This phenomenon has been documented in the literature in premenstrual young women (on menarche), nursing mothers, older women who believe they have entered menopause, and women with amenorrhea secondary to anorexia nervosa.3
A consistent observation in population-based studies of women with anorexia nervosa is earlier age of pregnancy relative to women in the general population.10,11 Although these studies did not necessarily address age at first pregnancy, the consistent pattern was of interest and we hypothesized that this observation could be attributable to women with anorexia nervosa believing that the absence of menstruation means they are not “at risk” for pregnancy, and they are experiencing unplanned pregnancies at greater rates than women in the general population. We estimated the frequency with which women with anorexia nervosa reported unplanned pregnancies relative to a population referent group in the large Norwegian Mother and Child Cohort Study (MoBa).
MATERIALS AND METHODS
The data collection was conducted as part of MoBa at the Norwegian Institute of Public Health.12 The study has been approved by the Institutional Review Board of the University of North Carolina at Chapel Hill, appropriate regional committees for ethics in medical research, and the Norwegian National Data Inspectorate.
MoBa is a prospective pregnancy cohort study. Pregnant women are recruited through a postal invitation after registering for a routine prenatal ultrasound examination at approximately 18 weeks of gestation. Participating women sign informed consent, donate blood and urine samples, and receive a questionnaire. The present study is based on this first questionnaire (questionnaire 1). The MoBa cohort is linked to Norwegian health registries, particularly the Medical Birth Registry of Norway,13 to capture pregnancy outcome variables.
Questionnaire 1 includes items on eating disorders and behaviors derived from studies of the Norwegian Institute of Public Health Twin Panel14 that reflect Diagnostic and Statistical Manual of Psychiatric Disorders, Fourth Edition, criteria for eating disorders.15 Diagnostic algorithms captured broadly defined anorexia nervosa (Diagnostic and Statistical Manual of Psychiatric Disorders, Fourth Edition, criteria excluding amenorrhea and endorsing a body mass index less than 19.0 kg/m2 at the time of low weight). Anorexia nervosa was assessed in the period 6 months before pregnancy only because of the practical difficulties in determining low weight in the presence of pregnancy-related weight gain. Self-reported weight and height were used to calculate body mass index measures. Individuals with other eating disorders (ie, bulimia nervosa, eating disorders not otherwise specified, and binge eating disorder) were not included in the referent group and were excluded from analyses because our hypotheses concerned anorexia nervosa only.
The current study is based on version four of the quality-assured data files released in 2009. The analysis population for this report included MoBa participants who: 1) had information from both the MoBa questionnaire 1 and the Medical Birth Registry of Norway; 2) did not complete an early pilot version of questionnaire 1 (n=2,599); 3) had valid values for self-reported age, weight, and height; 4) returned questionnaire 1 before delivery; and 5) had a singleton birth. If a woman enrolled in MoBa more than once (because of additional pregnancies), then only the first pregnancy that occurred during the course of MoBa data collection was included. Of the initial 91,489 mother-child records, 74,107 (81%) met these five criteria. Overall, from 1999 to 2006, approximately 42% of invited mothers agreed to participate in MoBa.12,16 The sample used for this analysis included women with anorexia nervosa (n=62) or no eating disorder (n=62,652), as described by Bulik et al.10 Excluding women with missing values for the question regarding planned pregnancy (n=924) results in a final sample size of 62,060 (Fig. 1). Respondents from the sample completed questionnaire 1 at a median of 18.6 weeks of gestation (n=46,893, interquartile range 17.1–20.6 weeks, range, 4.1–42.3 weeks).
Unplanned pregnancy was determined by participant's answer to one question regarding whether the current pregnancy was planned. Infertility status was based on participants' self-report of receiving fertility treatment. History of induced abortion was based on a single question about past abortions. Amenorrhea was assessed by a single question, “During the last year before you became pregnant, did you lose your period for more than three months?” The response options were “No/Yes, due to another pregnancy/Yes, for other reasons.”
Descriptive statistics are means and standard errors for continuous variables and percent distribution of categorical variables. Poisson regression was used to characterize the proportion and relative risk (RR) of unplanned pregnancy by group (anorexia nervosa compared with no eating disorder).17–19 Unplanned pregnancy was a dichotomous outcome variable and group was the categorical predictor variable with and without adjustment for potential confounders of maternal age and infertility status. These two confounders were selected because of their potential association with both the subtype covariate and the response variable. In this sample, age and infertility status were inversely associated with unintended pregnancy. We used generalized estimating equations20,21 to provide robust error estimates given under-dispersion of the data.19 In post hoc analysis, we explore whether the risk of unplanned pregnancy was related to the presence of amenorrhea in the year before pregnancy. To estimate the RR of unplanned pregnancy for those with amenorrhea compared with those without in both the anorexia nervosa and referent groups, another model similar to the previous Poisson regression was used. In addition to unplanned pregnancy as the response and a predictor representing eating disorder subtype, there were two additional covariates representing a dichotomous amenorrhea status variable and an interaction term between amenorrhea status and eating disorder subtype. A χ2 statistic was used to test differences in proportion of induced abortions by eating disorder subtype. All analyses were performed using SAS/STAT for Windows. The funding source had no role in any aspect of these analyses.
Table 1 presents demographic information on the sample. The mean age of index pregnancy of women with anorexia nervosa was 26.2 years (standard deviation 4.76), and of women with no eating disorder before or during pregnancy it was 29.9 years (standard deviation 4.60). Of women with anorexia nervosa, 50.0% reported the pregnancy was unplanned compared with 18.9% of women with no eating disorder. The percentage of women who reported having used contraception in the year before pregnancy was similar across groups (94.9% anorexia nervosa; 94.7% no eating disorder). In terms of infertility treatment, 4.8% of women with anorexia nervosa and 8.3% of women with no eating disorder reported some form of infertility treatment. After adjustment for maternal age and infertility treatment, the RR of unplanned pregnancy in individuals with anorexia nervosa was 2.11 (95% confidence interval [CI] 1.64–2.72) compared with women without eating disorders before pregnancy.
In a post hoc exploratory analysis, we attempted to estimate whether unplanned pregnancies were more common in women reporting amenorrhea in the year before pregnancy. A total of 14.5% (n=9) of women in the anorexia nervosa group and 5.5% (n=3,389) women in the referent group reported amenorrhea in the year before pregnancy. The RR of unplanned pregnancy for those reporting amenorrhea compared with no amenorrhea in the anorexia nervosa group was 1.13 (95% CI 0.59–2.16) and in the referent group 1.01 (95% CI 0.94–1.08).
In separate questions dealing with reproductive history, we assessed past induced abortions and found that approximately one-quarter of the women with anorexia nervosa reported having at least one abortion (24.2%, n=15) compared with 14.6% (n=9,069) in the sample without eating disorders (χ2=4.54, degrees of freedom=1, P=.033).
The age of index pregnancy in women with anorexia nervosa in the MoBa cohort was younger than in mothers with no eating disorder. The risk of having an unplanned pregnancy was significantly increased in women with anorexia nervosa, with more than half reporting that the index pregnancy was unplanned. Further supporting our observation, women with anorexia nervosa also reported significantly higher rates of past abortion. The unplanned pregnancies did not appear to be attributable to differential contraception use, although our assessment of contraception use was cursory and consisted of one question covering the year before pregnancy. We have no information on consistency of contraceptive use. Similarly, reported amenorrhea in the year before pregnancy did not increase the RR of unplanned pregnancy, although our assessment of amenorrhea was similarly cursory. It is possible that absent or irregular menstruation and the belief that menstrual irregularities reduce the risk of conception are associated with reduced adherence to contraception guidelines or instructions, increasing the risk of unplanned pregnancy.
We present our data as suggestive and as a starting point for further, more detailed investigations. Our study has limitations. First, our assessment was not sufficiently detailed to explore the question of unplanned pregnancy in depth. More comprehensive assessments of menstrual history, contraceptive use, and pregnancy history are required to obtain a fuller picture of the mechanism of unplanned pregnancy in anorexia nervosa. Second, at the time of this version of the data files, 42% of women invited agreed to participate in MoBa. Although somewhat low, this response rate is typical for large epidemiologic studies and does not necessarily imply a biased sample.22 MoBa participants also may be somewhat more educated than the general Norwegian population.23 Third, the sample represents women invited to participate at approximately 17 or 18 weeks of gestation,12 and unplanned pregnancy estimates most likely are underestimated in this sample compared with the total cohort of women who conceived at approximately the same time. Finally, given the considerable effort required to participate in the various waves of the MoBa protocol, the women with anorexia nervosa who do choose to participate may represent the healthier end of the eating disorder severity spectrum.
In terms of clinical implications, it is important to ensure that women with anorexia nervosa recognize risk of becoming pregnant despite menstrual irregularities. Stereotypes regarding the absence of sexual activity and infertility in anorexia nervosa should not inhibit clinicians from speaking directly with patients with anorexia nervosa about sexuality, contraception, and pregnancy. Planned pregnancies in women with eating disorders increase the opportunities for appropriate nutritional and emotional support to assist them with the physical and psychological challenges of pregnancy and motherhood.
1.Weinfeld R, Dubay M, Burchell R, Mellerick J, Kennedy A. Pregnancy associated with anorexia and starvation. Am J Obst Gynecol 1977;15:698–9.
2.Strimling B. Infant of a pregnancy complicated by anorexia nervosa. Am J Disease Child 1984;138:68–9.
3.Andersen AE, Ryan GL. Eating disorders in the obstetric and gynecologic patient population. Obstet Gynecol 2009;114:1353–67.
4.Abraham SF, Pettigrew B, Boyd C, Russell J, Taylor A. Usefulness of amenorrhoea in the diagnoses of eating disorder patients. J Psychosom Obstet Gynaecol 2005;26:211–5.
5.Watson TL, Andersen AE. A critical examination of the amenorrhea and weight criteria for diagnosing anorexia nervosa. Acta Psychiatr Scand 2003;108:175–82.
6.Poyastro Pinheiro A, Thornton LM, Plotonicov KH, Tozzi F, Klump KL, Berrettini WH, et al. Patterns of menstrual disturbance in eating disorders. Int J Eat Disord 2007;40:424–34.
7.Brinch M, Isager T, Tolstrup K. Anorexia nervosa and motherhood: reproduction pattern and mothering behavior of 50 women. Acta Psych Scand 1988;77:611–7.
8.Bulik C, Sullivan P, Fear J, Pickering A, Dawn A. Fertility and reproduction in women with anorexia nervosa: a controlled study. J Clin Psychiatry 1999;2:130–5.
9.Kohmura H, Miyake A, Aono T, Tanizawa O. Recovery of reproductive function in patients with anorexia nervosa: a 10-year follow-up study. Eur J Obstet Gynecol Reprod Biol 1986;22:293–6.
10.Bulik CM, Von Holle A, Siega-Riz AM, Torgersen L, Lie KK, Hamer RM, et al. Birth outcomes in women with eating disorders in the Norwegian Mother and Child cohort study (MoBa). Int J Eat Disord 2009;42:9–18.
11.Micali N, Simonoff E, Treasure J. Risk of major adverse perinatal outcomes in women with eating disorders. Br J Psychiatry 2007;190:255–9.
12.Magnus P, Irgens LM, Haug K, Nystad W, Skjaerven R, Stoltenberg C, et al. Cohort profile: The Norwegian Mother and Child Cohort Study (MoBa). Int J Epidemiol 2006;35:1146–50.
13.Irgens LM, Bergsjø P, Lie R. The Medical Birth Registry of Norway. Epidemiological research and surveillance throughout 30 years. Acta Obstet Gynecol Scand 2000;79:435–9.
14.Reichborn-Kjennerud T, Bulik CM, Tambs K, Harris JR. Genetic and environmental influences on binge eating in the absence of compensatory behaviors: a population-based twin study. Int J Eat Disord 2004;36:307–14.
15.American Psychiatric Association. Diagnostic and statistical manual of psychiatric disorders. 4th ed. Washington, DC: American Psychiatric Press; 1994.
16.Nilsen RM, Vollset SE, Gjessing HK, Skjaerven R, Melve KK, Schreuder P, et al. Self-selection and bias in a large prospective pregnancy cohort in Norway. Paed Perinat Epidemiol 2009;23:597–608.
17.Carter R, Lipsitz S, Tilley B. Quasi-likelihood estimation for relative risk regression models. Biostat 2005;6:39–44.
18.Spiegelman D, Hertzmark E. Easy SAS calculations for risk or prevalence ratios and differences. Am J Epidemiol 2005;162:199–200.
19.Zou G. A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol 2004;159:702–6.
20.Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986;42:121–30.
21.Liang K, Zeger S. Longitudinal data analysis using generalized linear models. Biometrika 1986;73:13–22.
22.Hartge P. Participation in population studies. Epidemiology 2006;17:252–4.
23.Bulik CM, Von Holle A, Hamer R, Knoph Berg C, Torgersen L, Magnus P, et al. Patterns of remission, continuation and incidence of eating disorders during early pregnancy in the Norwegian Mother and Child Cohort Study (MoBa). Psychol Med 2007;37:1109–18.