Objective
The purpose of this study was to assess pregnancy, obstetric, and perinatal health outcomes and complications in women with lifetime eating disorders.
Study Design
Female patients (n = 2257) who were treated at the Eating Disorder Clinic of Helsinki University Central Hospital from 1995-2010 were compared with unexposed women from the population (n = 9028). Register-based information on pregnancy, obstetric, and perinatal health outcomes and complications were acquired for all singleton births during the follow-up period among women with broad anorexia nervosa (AN; n = 302 births), broad bulimia nervosa (BN; n = 724), binge eating disorder (BED; n = 52), and unexposed women (n = 6319).
Results
Women with AN and BN gave birth to babies with lower birthweight compared with unexposed women, but the opposite was observed in women with BED. Maternal AN was related to anemia, slow fetal growth, premature contractions, short duration of the first stage of labor, very premature birth, small for gestational age, low birthweight, and perinatal death. Increased odds of premature contractions, resuscitation of the neonate, and very low Apgar score at 1 minute were observed in mothers with BN. BED was associated positively with maternal hypertension, long duration of the first and second stage of labor, and birth of large-for-gestational-age infants.
Conclusion
Eating disorders appear to be associated with several adverse perinatal outcomes, particularly in offspring. We recommend close monitoring of pregnant women with either a past or current eating disorder. Attention should be paid to children who are born to these mothers.
Eating disorders (EDs) are common psychiatric disorders among women at childbearing age. According to epidemiologic studies, at least 1 in 20 women experiences some form of ED during pregnancy. Salient symptoms of EDs include disturbed eating behavior, pronounced fear of weight gain, and dissatisfaction with one’s body. Accompanying dysfunctional behaviors, such as very restrictive eating, binge and purge episodes, excessive exercise, and laxative abuse depend on the ED subtype and often severely compromise the homeostatic balance of the body. Anxiety and depressive symptoms are also ubiquitous in these disorders. Residual symptoms are common even years after the recovery from ED.
On the other hand, many otherwise healthy women experience worry related to weight gain and their changing body during pregnancy and the postpartum period, and food cravings and fluctuations in eating patterns are physiologic during these periods. With this background, it is evident that pregnancy and the postpartum period represent extra challenges for women with EDs.
Adequate nutrition and weight gain are crucial for fetal development, and maternal stress has potentially detrimental effects on offspring. Current evidence on the effects of maternal EDs on pregnancy, delivery, and perinatal outcomes suggests increased risks for several complications, although studies are not fully conclusive. This is the reason that we aimed to examine the risk of pregnancy and obstetric complications and adverse perinatal health outcomes in a large patient cohort who was treated for EDs. Based on current evidence, we hypothesized that women with anorexia nervosa (AN) would have higher risk of pregnancy, obstetric, and perinatal complications related to undernourishment (eg, anemia, low birthweight, small-for-gestational-age [SGA]), that women with binge eating disorder (BED) would have an elevated risk of complications related to binge-eating and obesity (eg, hypertension, gestational diabetes mellitus, preterm birth, large-sized infants), and that complications of women with bulimia nervosa (BN) would be a mixture of these. In addition, we hypothesized that complications related to stress, anxiety, and depressive symptoms (eg, premature contractions, preterm birth) would be present in patients from all 3 ED categories.
Materials and Methods
From hospital records, we manually identified all patients who had been treated in the ED clinic at the Helsinki University Central Hospital from 1995-2010. Matched unexposed control women were selected randomly from the Central Population Register as described previously. A register search on pregnancy, obstetric, and perinatal outcomes was conducted on 2257 patients and 9028 unexposed women for the follow-up period (extending from admission to Dec. 31, 2010/death/moving abroad/reaching age 50 years). We focus hereby on pregnancies that led to childbirth. All births during the follow-up period were included; however, multiple births were excluded (n = 104).
The ED diagnoses were set by attending physicians at the clinic with the use of International Statistical Classification of Diseases and Related Health Problems, 10th revision (ICD-10) criteria, where F50.0, F50.1, F50.2, and F50.3 indicate AN, atypical AN, BN, and atypical BN, respectively. We used broad criteria for AN and BN, with atypical forms combined with full disorders. In the clinic, diagnosis of BED was set with the use of Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV), research criteria.
Outcome measures
Data on outcome measures were obtained from the Medical Birth Register, which covers all delivery hospitals in Finland (live births and stillbirths with ≥22 weeks’ gestation or birthweight ≥500 g). Data quality studies indicate that most of the register content corresponds well/satisfactorily with hospital records.
Pregnancy complications
The following pregnancy complications (recorded since 2004) were included: gestational diabetes mellitus (pathologic oral glucose tolerance test), initiation of insulin treatment during pregnancy, anemia, antenatal corticosteroid treatment, and pregnancy-related ICD-10 diagnoses of the mother (since 2004). ICD-10 diagnoses included preeclampsia (O14), hypertension (O13, O16), slow fetal growth (O36.5, P05.0, P05.1, P05.9), fast fetal growth (O36.6), oligohydramnios (O41.0), infection of amniotic fluid (O41.1), premature rupture of membrane (O42), any placental disorder (O43, O44, O45, O73, and a separate check-box on placenta previa in the Medical Birth Register data collection form), fear of childbirth (O99.80), premature contractions (O47), proteinuria (O12.1), hyperemesis gravidarum (O21.0, O21.1, O21.2, O21.9), any vein complication (O22), urogenital infection (O23), hepatogestosis (O26.6), exhaustion (O26.82), symphyseolysis (O26.7), cervix insufficiency (O34.3), suspected fetal injury because of alcohol/drugs (O35.4, O35.5), and suspected fetal hypoxia (O36.3). Information on eclampsia was available for the whole duration of the follow-up period.
Obstetric complications
The studied obstetric complications included induction of labor, asphyxia, breech presentation, epidural anesthesia, use of forceps/vacuum, delivery by cesarean section, delivery by elective cesarean section, duration of the first and the second stage of labor (minutes, since 2004), and delivery related ICD-10 diagnoses (since 2004). The studied diagnoses included fetal distress (O68), rupture of perineum (O70), and maternal distress (O75.0).
Perinatal health outcomes and complications
The following perinatal outcomes were included: perinatal death, gestational age (by fetal ultrasound examination at the first maternity care visit), premature birth (<37 weeks’ gestation), very premature birth (<28 weeks’ gestation), birthweight (grams), low birthweight (<2500 g), very low birthweight (<1500 g), SGA (using cut-points for the Finnish population ), large for gestational age (LGA), low Apgar score at 1 minute (≤6), very low Apgar score at 1 minute (≤3), assisted ventilation, resuscitation, and neonatal monitoring.
Covariates
We used maternal age, parity, marital status (dichotomous variable single vs married or cohabitation), and smoking status (yes/no) as covariates in the adjusted models.
Statistical analyses
We used Stata statistical software (version 12.1; StataCorp, College Station, TX) for the data analysis. Analyses were done in 2 phases, both unadjusted and adjusted for covariates described earlier. We used linear regression to analyze continuous variables and logistic regression for the analyses of categoric variables. χ 2 test was used for demographic variables and t test for prenatal care indicators. Comparisons were done between diagnostic groups and the joint group of unexposed women. In analyses, we took into account the clustering of births within mothers.
Ethical considerations
The Ethics committee of National Institute of Health and Welfare has reviewed the study with a positive statement (DnroTHL/184/6.02.00/2011). Data handling was performed according to the Finnish data protection legislation and the rules of National Institute of Health and Welfare. All institutions gave their permission to use their register data in this study. The authors did not have access to the personal identification data; only research codes were used in all analyses.
Results
We identified 1078 singleton births among patients and 6319 among unexposed women during the follow-up period. Mothers with broad AN (n = 182) delivered 302 babies; mothers with broad BN (n = 436) and BED (n = 39) delivered 724 and 52 babies, respectively. Unexposed mothers (n = 3642) delivered 6319 babies. Only singleton childbirths are reported here.
Demographic characteristics
The mean age at childbirth was 29.4 ± 5.0 (SD) years in women with AN, 30.4 ± 1.2 years in women with BN, 30.2 ± 1.0 years in women with BED, and 29.1 ± 4.8 years in unexposed women. Being married was less common among women with AN compared with unexposed women ( P < .001), and being divorced was more common in women with AN and BED ( P < .001 and = .005, respectively). There were no differences across the groups in being single. Among those who gave birth, parity was distributed equally in women with AN and BED compared with unexposed women, whereas the number of previous births was lower among women with BN ( P = .005). Smoking during pregnancy was less common among women with an ED compared with unexposed women ( P = .04 for BN; not significant for AN and BED).
Prenatal care
The total number of prenatal care visits was higher among women with BN and BED compared with unexposed women, and all patient groups had an elevated number of hospital outpatient visits before delivery. The mean number of prenatal care visits was 16.7 ± 0.3 ( P = not significant) in women with AN, of which 3.6 ± 0.2 ( P < .001) were hospital outpatient visits, 17.2 ± 0.2 ( P < .001) in women with BN, of which 3.7 ± 0.1 ( P < .001) were in the hospital, 19.6 ± 0.9 ( P < .001) in women with BED, of which 4.7 ± 0.5 ( P < .001) were in the hospital, and 16.4 ± 0.06 in unexposed women, of which 2.9 ± 0.03 were in the hospital.
Pregnancy complications
Most of the pregnancy complications occurred in similar percentages across the exposure groups ( Table 1 ).
| Pregnancy complication | Anorexia nervosa | Bulimia nervosa | Binge eating disorder | Unexposed |
|---|---|---|---|---|
| Gestational diabetes mellitus | ||||
| n (%) | 5 (1.66) | 27 (3.73) | 2 (3.85) | 257 (4.07) |
| Odds ratio (95% CI) | ||||
| Crude | 0.40 (0.14–1.13) | 0.91 (0.57–1.46) | 0.94 (0.24–3.75) | 1 |
| Adjusted | 0.38 (0.13–1.10) | 0.81 (0.51–1.31) | 0.85 (0.22–3.23) | 1 |
| Anemia | ||||
| n (%) | 12 (3.97) | 12 (1.66) | 0 | 97 (1.54) |
| Odds ratio (95% CI) | ||||
| Crude | 2.65 (1.38–5.11) a | 1.08 (0.57–2.06) | 1 | |
| Adjusted | 2.39 (1.20–4.76) a | 1.05 (0.54–2.03) | 1 | |
| Hypertension b | ||||
| n (%) | 3 (1.4) | 6 (1.22) | 4 (22.22) | 87 (2.24) |
| Odds ratio (95% CI) | ||||
| Crude | 0.62 (0.19–1.98) | 0.54 (0.21–1.4) | 12.48 (3.82–40.82) a | 1 |
| Adjusted | 0.63 (0.20–2.00) | 0.51 (0.20–1.33) | 13.29 (4.03–43.81) a | 1 |
| Slow fetal growth b | ||||
| n (%) | 14 (4.64) | 22 (3.04) | 0 | 122 (1.93) |
| Odds ratio (95% CI) | ||||
| Crude | 2.47 (1.36–4.48) a | 1.59 (0.99–2.55) | 1 | |
| Adjusted | 2.59 (1.43–4.71) a | 1.53 (0.94–2.48) | 1 | |
| Fast fetal growth b | ||||
| n (%) | 0 | 7 (1.42) | 1 (5.56) | 36 (0.93) |
| Odds ratio (95% CI) | ||||
| Crude | 1.54 (0.68–3.51) | 6.29 (0.80–49.67) | 1 | |
| Adjusted | 1.54 (0.69–3.47) | 6.06 (0.72–50.99) | 1 | |
| Premature contractions a | ||||
| n (%) | 7 (3.26) | 16 (3.25) | 1 (5.56) | 59 (1.52) |
| Odds ratio (95% CI) | ||||
| Crude | 2.18 (0.99–4.83) | 2.18 (1.18–4.00) a | 3.82 (0.49–29.77) | 1 |
| Adjusted | 2.31 (1.05–5.11) a | 2.20 (1.17–4.14) a | 3.96 (0.51–30.95) | 1 |
a Statistically significant findings
Anemia was more frequent among women with AN compared with unexposed women. The risk of maternal hypertension was elevated in women with BED. Furthermore, slow fetal growth was observed more frequently in women with AN compared with unexposed women. Women with AN and BN had increased odds of premature contractions compared with unexposed women.
Obstetric complications
The rate of induction of labor was 14.8% in women with AN, 18.2% in women with BN, 26.4% in women with BED, and 15.5% in unexposed women ( P = not significant). The elective cesarean section rate was 7.1% in women with AN, 7.3% in women with BN, 11.3% in women with BED, and 5.9% in unexposed women ( P = not significant).
Duration of the first stage of labor was the shortest among women with AN (mean, 733 ± 401 minutes, adjusted P = .031) and the most lengthy among women with BED (mean, 1249 ± 309 minutes; adjusted P < .001) relative to unexposed women (mean, 811 ± 503 minutes). Similarly, the second stage of labor was prolonged among women with BED (mean, 110 ± 73 minutes) comparing with unexposed women (mean, 43 ± 55 minutes; adjusted P = .018). There were no statistically significant differences between the exposure groups in terms of other obstetric outcomes.
Perinatal health outcomes and complications
Women with AN and BN gave birth to babies with lower birthweight compared with unexposed women (mean, 3302 ± 562 g; adjusted P < .001 in women with AN; mean, 3464 ± 563 g; adjusted P = .037 in women with BN; mean, 3520 ± 539 g in unexposed women), whereas birthweight was higher among babies of women with BED (mean, 3812 ± 519 g; adjusted P < .001). Similarly, women with AN had an increased odds for SGA infants and infants with low birthweight ( Table 2 ), whereas odds for LGA infants was increased among women with BED. Gestational age was the lowest among women with AN and the highest among women with BED (AN: mean, 39.6 ± 2.1 weeks; adjusted P = .032; BN: mean 39.7 ± 1.9 weeks; adjusted P = .026; BED: mean, 40.1 ± 1.4 weeks; adjusted P = .27; unexposed: mean, 39.9 ± 1.8 weeks). Women with AN had an increased risk of very premature birth. All 3 cases of very premature birth among women with AN were spontaneous in nature. Assisted ventilation and monitoring of the neonate occurred in similar percentages across the groups, whereas resuscitation and very low Apgar score at 1 minute after the birth were more common among infants born to women with BN compared with unexposed women. Babies of women with AN had a 4-fold risk of perinatal death (adjusted odds ratio, 4.06; 95% confidence interval, 1.15–14.35). All of these 3 babies were born prematurely; 2 of them were born very prematurely at <28 weeks’ gestation.
| Perinatal health complication | Anorexia nervosa | Bulimia nervosa | Binge eating disorder | Unexposed |
|---|---|---|---|---|
| Resuscitation | ||||
| n (%) | 3 (0.99) | 15 (2.07) | 0 | 58 (0.92) |
| Odds ratio (95% CI) | ||||
| Crude | 1.08 (0.34–3.45) | 2.28 (1.29–4.03) a | 1 | |
| Adjusted | 1.06 (0.33–3.37) | 2.12 (1.18–3.79) a | ||
| Perinatal death | ||||
| n (%) | 3 (0.99) | 3 (0.41) | 1 (1.92) | 21 (0.33) |
| Odds ratio (95% CI) | ||||
| Crude | 3.00 (0.89–10.2) | 1.25 (0.37–4.20) | 5.88 (0.79–43.57) | 1 |
| Adjusted | 4.06 (1.15–14.35) a | 1.78 (0.51–6.19) | 9.51 (1.33–68.26) a | |
| Small for gestational age | ||||
| n (%) | 13 (4.30) | 23 (3.18) | 0 | 133 (2.10) |
| Odds ratio (95% CI) | ||||
| Crude | 2.09 (1.17–3.73) a | 1.53 (0.94–2.47) | 1 | |
| Adjusted | 2.20 (1.23–3.93) a | 1.51 (0.92–2.48) | ||
| Large for gestational age | ||||
| n (%) | 1 (0.33) | 19 (2.62) | 5 (9.62) | 155 (2.45) |
| Odds ratio (95% CI) | ||||
| Crude | 0.13 (0.02–0.94) a | 1.07 (0.65–1.78) | 4.23 (1.64–10.92) a | 1 |
| Adjusted | 0.13 (0.02–0.91) a | 1.10 (0.66–1.84) | 4.32 (1.64–11.36) a | |
| Premature birth | ||||
| n (%) | 15 (4.98) | 36 (5.01) | 0 | 259 (4.11) |
| Odds ratio (95% CI) | ||||
| Crude | 1.22 (0.67–2.22) | 1.23 (0.82–1.84) | 1 | |
| Adjusted | 1.28 (0.71–2.33) | 1.28 (0.85–1.91) | ||
| Very premature birth | ||||
| n (%) | 3 (0.99) | 3 (0.42) | 0 | 18 (0.29) |
| Odds ratio (95% CI) | ||||
| Crude | 3.51 (1.02–12.09) a | 1.46 (0.43–5.01) | 1 | |
| Adjusted | 4.59 (1.25–16.87) a | 1.84 (0.51–6.62) | ||
| Low birthweight | ||||
| n (%) | 19 (6.31) | 30 (4.16) | 0 | 201 (3.19) |
| Odds ratio (95% CI) | ||||
| Crude | 2.05 (1.23–3.40) a | 1.32 (0.88–1.98) | 1 | |
| Adjusted | 2.16 (1.30–3.58) a | 1.37 (0.90–2.07) | ||
| Very low birthweight | ||||
| n (%) | 3 (0.99) | 6 (0.83) | 0 | 34 (0.54) |
| Odds ratio (95% CI) | ||||
| Crude | 1.86 (0.56–6.11) | 1.55 (0.65–3.70) | 1 | |
| Adjusted | 2.14 (0.64–7.20) | 1.81 (0.72–4.57) | ||
| Low Apgar score at 1 min (<7) | ||||
| n (%) | 16 (5.39) | 39 (5.41) | 5 (9.62) | 299 (4.75) |
| Odds ratio (95% CI) | ||||
| Crude | 1.14 (0.67–1.96) | 1.15 (0.82–1.61) | 2.13 (0.72–6.35) | 1 |
| Adjusted | 1.17 (0.68–2.01) | 1.13 (0.79–1.60) | 2.23 (0.74–6.68) | |
| Very low Apgar score at 1 min (<3) | ||||
| n (%) | 4 (1.35) | 19 (2.64) | 1 (1.92) | 80 (1.27) |
| Odds ratio (95% CI) | ||||
| Crude | 1.06 (0.39–2.89) | 2.10 (1.25–3.54) a | 1.52 (0.21–10.87) | 1 |
| Adjusted | 1.16 (0.42–3.20) | 2.31 (1.34–3.98) a | 1.74 (0.25–12.1) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
