Abstract
Introduction
Early life abuse has been linked to later Type 2 diabetes, but its association with gestational diabetes has not been examined. The aim of this study was to examine the association between childhood and adolescent abuse victimization and risk of gestational diabetes in the Nurses’ Health Study II.
Methods
Participants were asked about experiences of physical and sexual abuse in childhood or adolescence in 2001 and about history of pregnancy complications in 2009. Mothers of singleton live births who provided information on their abuse history comprised the study sample. Modified Poisson regression was used to estimate risk ratios and 95% CIs for gestational diabetes as a function of physical and sexual abuse victimization. Analyses were conducted in 2014–2015.
Results
Of 45,550 women in the analysis, 8% reported severe physical abuse and 11% reported forced sexual activity in childhood or adolescence. Approximately 3% (n=3,181) of pregnancies were complicated by gestational diabetes. In adjusted models, severe physical abuse was associated with a 42% greater gestational diabetes risk (risk ratio=1.42, 95% CI=1.21, 1.66) relative to no physical abuse and forced sexual activity was associated with a 30% greater risk (95% CI=1.14, 1.49). Women with histories of both physical and sexual abuse were at higher risk than women exposed to a single type of abuse. These associations were not explained by overweight status in early adulthood or prior to pregnancy.
Conclusions
Childhood and adolescent victimization is associated with increased risk of gestational diabetes in adult women.
Introduction
Gestational diabetes mellitus (GDM) is a metabolic condition in which impaired glucose tolerance emerges or is first recognized during pregnancy.1 GDM increases incidence of perinatal complications such as caesarean section and macrosomia,2,3 thus its prevention is an important public health goal. GDM risk factors include overweight/obesity4 and sociodemographic factors such as minority race and low SES.5 Psychosocial factors, such as stress6 and prenatal depression,7,8 may also contribute to GDM.
Abuse victimization in early life (childhood or adolescence) is a prevalent psychosocial exposure affecting more than 20% of U.S. women.9,10 There are several reasons that early abuse may be a GDM risk factor. First, abuse victimization is associated with Type 2 diabetes in midlife,11 which is strongly correlated with GDM12; it is possible that the metabolic risk associated with abuse emerges first in pregnancy. Second, women with abuse histories are more likely to be obese in young adulthood,13–19 and obesity is a major contributor to GDM risk.4 Third, other plausible biological mechanisms could link abuse to GDM, including abuse-related elevations in stress hormones that may trigger insulin resistance.20–22 To the authors’ knowledge, the association between early abuse and GDM has not been investigated.
Understanding the relationship between early life abuse and GDM can inform theories of GDM etiology and guide prevention approaches. This study therefore examined the hypotheses that:
Early life physical or sexual abuse exposure is associated with increased GDM risk.
This association is explained by early adult and prepregnancy weight status. Analyses were conducted in the Nurses’ Health Study II (NHSII) longitudinal cohort in 2014–2015.
Methods
Study Population
The NHSII is a longitudinal cohort of 116,430 female registered nurses recruited from 14 states at age 25–42 years in 1989. Biennial questionnaires ascertain sociodemographic, behavioral, and medical information. In 2001, a supplemental Violence Questionnaire asking about physical and sexual abuse in childhood or adolescence was sent to 91,297 NHSII participants who had responded to the previous biennial questionnaire within three mailings. Questionnaires were returned by 68,376 (75%) of recipients. The 2009 biennial questionnaire included a comprehensive reproductive history, and was returned by 57,580 Violence Questionnaire responders; these women were similar to 2001 questionnaire responders overall (data not shown). Women were excluded if they had had no live births (n=10,236), did not report the year of the birth (n=4), had their first birth prior to age 18 years (n=1,126) owing to overlap with the window of abuse exposure, or had a twin or triplet first birth (n=664) given potentially unique GDM etiologies in these pregnancies; women with singleton first births who had a subsequent twin or triplet pregnancy contributed singleton pregnancy observations until the twin or triplet birth, when their subsequent pregnancies were excluded. This left 45,550 women with 103,370 pregnancies for analyses. Of these, women missing physical abuse data were excluded from physical abuse analyses (n=135), and women missing sexual abuse data were excluded from sexual abuse analyses (n=278). Analyses of combined physical and sexual abuse excluded women missing either physical or sexual abuse (n=319). This study was approved by the IRB of Brigham and Women’s Hospital.
Measures
Child and adolescent physical abuse were assessed using questions from the Revised Conflict Tactics Scales,23 which ascertained the frequency in childhood (age 0–10 years) and adolescence (age 11–17 years) with which a parent, step parent, or adult guardian: pushed, grabbed, or shoved; kicked, bit, or punched; hit with something that hurt; choked or burned; or physically attacked the participant. As in previous analyses,11,24 this study used the most severe event in either childhood or adolescence to categorize physical abuse as:
none;
mild (pushed/grabbed/shoved at any frequency or kicked/bitten/punched once or hit with something once);
moderate (hit with something more than once or physically attacked once); and
severe (kicked/bitten/punched more than once or physically attacked more than once or choked/burned at any frequency).
In analyses of the mediating role of prepregnancy overweight, which were conducted in the subset of the sample with a pregnancy after the 1989 baseline, moderate and severe abuse were combined into a single category to maintain adequate cell sizes. Preliminary analyses indicated that associations were similar for moderate and severe abuse.
Child and adolescent sexual abuse was ascertained by asking participants whether and how often as a child (age 0–10 years) or adolescent (age 11–17 years) the following occurred:
They had been touched in a sexual way by an adult or an older child or were forced to touch an adult or an older child in a sexual way when they did not want to.
An adult or older child had forced or attempted to force them into any sexual activity by “threatening you, holding you down, or hurting you in some way when you did not want to?”25
Sexual abuse was categorized into the following three categories based on the most severe event24: none, sexual touching only, and forced sexual activity.
To assess whether exposure to both physical and sexual abuse conferred additional risk, two composite variables were also examined: an eight-category variable for physical abuse severity with and without any sexual abuse, and a six-category variable for sexual abuse severity with and without any physical abuse. The ref for both composite variables was no physical or sexual abuse.
The 2009 questionnaire included a reproductive history that asked participants to report diagnosis of GDM in each pregnancy. A previous validation study of self-reported GDM on earlier biennial questionnaires confirmed 94% of self-reported GDM cases against medical record of GDM diagnosis, suggesting that participants tend to report GDM accurately.26
The following shared risk factors for abuse and GDM were included as potential confounders in adjusted models: age at pregnancy (continuous), nurse participant’s year of birth (continuous), race/ethnicity (indicator for nonwhite, with white as the ref), participant’s mother’s and father’s educational attainment (indicators for <9 years, 9–11 years, 12 years, 13–15 years, with ≥16 years as the ref), participant’s mother in professional occupation (yes/no), participant’s father in professional occupation (yes/no), participant’s parents’ home ownership (yes/no), recalled body size at age 5 years (each participant indicated the image of the female figure that best approximated her body type at age 5 years, ranging from 1=very lean to 9=very obese27; ordinal), and participant’s parents’ history of diabetes (yes/no).
Because abuse is associated with weight status,15,17,18 which predicts GDM,4 overweight status was examined at two time points as potential mediators of the abuse–GDM relationship. First, the study examined mediation by overweight at age 18 years, defined as BMI >25 kg/m2, based on weight at age 18 years reported by nurse participants at NHSII baseline; comparisons to recorded weights on nursing school entry health exams indicate reasonable validity of this measure.28 Second, prepregnancy overweight was examined as a mediator. Prepregnancy overweight was defined as BMI>25 kg/m2 on the biennial questionnaire prior to each pregnancy. This analysis was restricted to the subset of the analytic sample that had at least one pregnancy after the 1989 NHSII baseline (i.e., those for whom prepregnancy overweight was measured). The potential mediating role of weight trajectories was examined in a supplemental analysis examining weight change from age 18 years to prepregnancy. Finally, the authors ran a supplemental analysis of mediation by prepregnancy smoking, which is common after abuse29 and may be associated with GDM30,31; prepregnancy smoking was defined as current smoking reported on the biennial questionnaire prior to the pregnancy.
Statistical Analysis
Modified Poisson regression32 was used to estimate risk ratios (RRs) for GDM as a function of abuse, using generalized estimating equations to handle repeated outcomes.33 For each abuse exposure, the authors ran a crude model and a model adjusted for potential confounders. Missing covariate data were handled with complete case analyses; fewer than 10% of participants were missing covariates. Because the main analyses relied on GDM reports on the 2009 cumulative pregnancy questionnaire, which have not been validated, the authors ran a sensitivity analysis using GDM reported on the previous biennial questionnaires, which performed well in a validation study.26 They also ran supplemental analyses of GDM as a function of:
timing of abuse (childhood only, adolescence only, or both); and
sexual abuse by frequency.
All analyses were conducted in SAS, version 9.3.
To examine BMI at age 18 years as a potential mediator of the abuse–GDM association, associations were estimated between:
abuse and risk of overweight at age 18 years;
overweight status at age 18 years and GDM risk; and
abuse and GDM risk adjusted for overweight at age 18 years.
Models adjusting for a mediator can provide an estimate of the direct effect of the exposure on the outcome, independent of the mediator, under several assumptions including no interaction between the exposure and mediator.34 The authors therefore tested for an interaction between categorical abuse and overweight at age 18 years, using a Wald test with a cutoff of p<0.05; a significant interaction term implies different direct effects of abuse on GDM depending on overweight status.34
For the subset of the analytic sample that had at least one pregnancy after the 1989 NHSII baseline (n=14,643), the authors also assessed the mediating role of prepregnancy overweight (approximated by overweight status at the biennial questionnaire prior to each pregnancy). The association was estimated between abuse and prepregnancy overweight, and between prepregnancy overweight and GDM. The authors then ran abuse–GDM models adjusted for overweight status, after testing for an abuse–overweight interaction. They ran a supplemental analysis adjusting for weight change from age 18 years to prepregnancy.
Results
Approximately one third (34%) of the 45,550 women included in the analysis reported moderate or severe physical abuse in childhood or adolescence, with 8% (n=3,684) reporting severe abuse. Thirty-three percent reported sexual abuse, with approximately 11% (n=4,806) reporting forced sexual activity (Table 1). Approximately 3% of the 103,370 pregnancies were complicated by GDM and 5% of women had experienced a GDM pregnancy. The incidence of GDM ranged from 5% among women with no history of physical abuse to 7% for women with a history of severe physical abuse. Women with histories of physical or sexual abuse were slightly younger at pregnancy than their non-abused counterparts, were less likely to have fathers who worked professional jobs, and were more likely to have a parental history of diabetes (Table 1).
Table 1.
Distribution of Outcome and Covariates by Physical and Sexual Abuse in Childhood or Adolescence
| Childhood and adolescent physical abuse |
Childhood and adolescent sexual abuse |
Missing (%) |
||||||
|---|---|---|---|---|---|---|---|---|
| Analysis variables |
None (n=21,21 6) |
Mild (n=8,5 32) |
Moderate (n=11,9 83) |
Severe (n=3,6 84) |
None (n=30,3 72) |
Touch only (n=10,0 94) |
Forced sex (n=4,8 06) |
|
|
Gestational diabetes in any pregnancy n (%) |
1,053(5.0) | 465(5.5) | 685(5.7) | 258(7.0) | 1,561(5.1) | 574(5.7) | 318(6.6) | 0.0 |
|
Continuous covariates (Mean, SD) | ||||||||
| Maternal year of birth |
1954.4(4.8) | 1954.4(4.6) | 1954.3(4.6) | 1953.9(4.6) | 1954.4(4.7) | 1954.0(4.6) | 1954.1(4.5) | 0.0 |
| Age at pregnanc y (years) |
28.5(4.8) | 28.4(4.9) | 28.1(4.9) | 27.7(5.2) | 28.5(4.8) | 28.1(4.9) | 27.6(5.1) | 0.0 |
| Parity in 2009 |
2.3(0.9) | 2.3(0.9) | 2.3(0.9) | 2.2(0.9) | 2.3(0.9) | 2.3(0.9) | 2.3(0.9) | 0.0 |
| Mother’s educatio n (years) |
12.5(2.0) | 12.4(1.9) | 12.2(1.9) | 12.1(2.0) | 12.4(1.9) | 12.2(2.0) | 12.1(2.0) | 6.7 |
| Father’s educatio n (years) |
12.6(2.5) | 12.5(2.4) | 12.3(2.4) | 12.2(2.5) | 12.6(2.4) | 12.3(2.5) | 12.2(2.5) | 7.9 |
| Age 5 somatogr ama |
2.5(1.2) | 2.5(1.2) | 2.5(1.2) | 2.5(1.3) | 2.5(1.2) | 2.5(1.2) | 2.5(1.3) | 1.8 |
|
Categorical covariates (column %) | ||||||||
| Physical abuse |
0.1 | |||||||
| None | -- | -- | -- | -- | 52.0 | 40.5 | 26.8 | |
| Mild | -- | -- | -- | -- | 18.7 | 19.7 | 17.7 | |
| Mode rate |
-- | -- | -- | -- | 24.1 | 30.8 | 31.6 | |
| Sever e |
-- | -- | -- | -- | 5.3 | 9.0 | 23.9 | |
| Sexual abuse |
0.4 | |||||||
| None | 74.6 | 66.6 | 61.3 | 43.7 | -- | -- | -- | |
| Touch only |
19.3 | 23.4 | 26.0 | 24.9 | -- | -- | -- | |
| Forced sex once |
6.1 | 10.0 | 12.7 | 31.4 | -- | -- | -- | |
| Nonwhite race | 5.5 | 5.4 | 7.6 | 7.5 | 5.7 | 6.9 | 7.4 | 0.0 |
| Mother professional |
11.8 | 11.1 | 10.0 | 10.1 | 11.3 | 10.6 | 10.3 | 0.0 |
| Father professional |
30.0 | 27.6 | 23.9 | 22.1 | 28.9 | 24.8 | 22.7 | 0.0 |
| Parents owned home |
53.2 | 50.3 | 48.6 | 45.0 | 51.7 | 49.5 | 48.2 | 0.0 |
| Parental diabetes history |
24.5 | 25.5 | 27.2 | 29.9 | 24.9 | 27.3 | 29.1 | 0.0 |
| Overweight (BMI>25 kg/m2) at age 18 yearsb |
7.4 | 7.8 | 8.7 | 9.8 | 7.6 | 8.7 | 9.4 | 0.9 |
Participants chose the image of a female figure that best approximated their body type at age 5, ranging from 1 (very lean) to 9 (obese)
Nurse participants were asked to self-report their weight at age 18 on the 1989 baseline questionnaire; comparisons to recorded weights on nursing school entry health exams indicate good validity of these reports.29
In crude analyses, physical abuse severity was associated with elevated GDM risk in a dose–response manner (Table 2). Mild, moderate, and severe physical abuse were associated with GDM RRs, relative to no physical abuse, of 1.13 (95% CI=1.01, 1.27), 1.16 (95% CI=1.05, 1.29), and 1.50 (95% CI=1.30, 1.73), respectively. Adjustment for covariates attenuated these RRs to 1.08 (95% CI=0.96, 1.22), 1.16 (95% CI=1.04, 1.29), and 1.42 (95% CI=1.21, 1.66), respectively. The adjusted RR for forced sex was 1.30 (95% CI=1.14, 1.49, Table 2). Exposure to unwanted sexual touching was unassociated with GDM risk (RR=1.08, 95% CI=0.96, 1.22, Table 2), except when experienced more than once (data not shown); forced sex had similar associations with GDM regardless of frequency. Sensitivity analyses using reports of GDM on prior biennial questionnaires, rather than reports from the 2009 pregnancy survey, generally found similar results (data not shown), with a slightly stronger association for unwanted sexual touching.
Table 2.
Early Life Physical and Sexual Abuse Associations With Gestational Diabetes
| Physical abusea |
Crudeb |
Adjustedcd |
By sexual abuse exposurece |
||||
|---|---|---|---|---|---|---|---|
| No sexual abuse exposure |
With sexual abuse Exposuref |
||||||
| N(cases) | RR (95%CI) |
RR (95 %CI) | N(cases) | RR (95%CI) | N(cases) | RR (95%CI) | |
| None | 21,216(1,053) | 1-- | 1 -- | 15,774(751) | 1 -- | 5,371(297) | (0.98,1.14 1.32) |
| Mild | 8,532(465) | 1.13 (1.01, 1.27) | (0.96,1.221.08 ) | 5,662(300) | (0.92,1.07 1.24) | 2,834(165) | (1.03,1.24 1.49) |
| Moderate | 11,983(685) | 1.16 (1.05, 1.29) |
(1.04,1.291.16 ) | 7,313(408) | (1.04,1.18 1.35) | 4619(274) | (1.04,1.21 1.42) |
| Severe | 3,684(258) | 1.50 (1.30, 1.73) | (1.21,1.661.42 ) | 1,597(100) | (0.98,1.24 1.57) | 2,061(156) | (1.38,1.68 2.04) |
| Sexual abuse |
Crudeg |
Adjustedch |
By physical a buse exposureci |
||||
| No physical abuse exposure |
With physical abuse exposurej |
||||||
| N(cases) | RR (95% CI) |
RR (95% CI) | N(cases) | RR (95% CI) | N(cases | RR (95%CI) ) |
|
| None | 30,372(1,561) | 1 -- | 1 -- | 15,774 (751) | 1 -- | 14,572(808) | (1.03,1.14 1.28) |
| Touch only |
10,094(574) | 1.1 (1.00,1 1.23) | 1.1 (0.99,0 1.22) | 4,083(218) | 1. (0.91, 08 1.27) | 5,999(356) | (1.09,1.25 1.44) |
| Forced sex |
4,806(318) | 1.(1.15,3 1.49) | 1.(1.14,3 1.49) | 1,288(79) | 1.(1.04, 34 1.72) | 3,51 | (1.20,51.42 1.67) |
| 1 | 0 | (239) | |||||
Mild physical abuse was defined as being pushed, grabbed, or shoved at any frequency or being kicked, bitten, or punched once or hit with something once; moderate physical abuse was defined as being hit with something more than once or physically attacked once, and severe physical abuse was defined as being kicked, bitten, punched, or physically attacked more than once or ever choked or burned.
n=45,415 women with 103,059 pregnancies
Adjusted for maternal age at pregnancy, maternal year of birth, race, mother’s educational attainment, father’s educational attainment, mother in professional occupation, father in professional occupation, parental home ownership, age 5 body size (somatogram), parental history of diabetes.
Complete case analysis included 41,540 women with 94,340 pregnancies
Complete case analysis included 41,373 women with 93,974 pregnancies
Any sexual abuse, including sexual touching only, forced sexual activity once, and forced sexual activity more than once.
n=45,272 women with 102,750 pregnancies
Complete case analysis included 41,408 women with 94,058 pregnancies
Complete case analysis included 41,373 women with 93,974 pregnancies
Any physical abuse, including mild, moderate, or severe
When modeled as a composite variable, severe physical abuse in combination with sexual abuse was associated with a 68% higher GDM risk (RR=1.68, 95% CI=1.38, 2.04) relative to no physical or sexual abuse. When combined with physical abuse, forced sexual activity was associated with a 40% higher GDM risk (RR=1.42, 95% CI=1.20, 1.67) compared with no sexual or physical abuse (Table 2). Supplemental analyses of the timing of abuse suggested that physical abuse was more strongly associated with GDM if it occurred in adolescence (versus childhood), whereas the opposite was true for sexual abuse (data not shown).
Moderate and severe physical abuse and forced sex were associated with modestly higher risks of overweight at age 18 years (moderate physical abuse, RR=1.15, 95% CI=1.06, 1.24; severe physical abuse, RR=1.20, 95% CI=1.07, 1.35; forced sex, RR=1.10, 95% CI=1.00, 1.22, data not shown). Being overweight at age 18 years was associated with a 43% higher risk of GDM (RR=1.43, 95% CI=1.24, 1.66) relative to not being overweight (data not shown). However, adjustment for overweight at age 18 years made no difference to the estimated association between either physical or sexual abuse and GDM (Table 3). A physical abuse–age 18 years overweight interaction term was statistically significant (Wald p<0.05), with severe physical abuse associated with an RR for GDM of 1.74 (95% CI=1.09, 2.78) among the small number (8%) of women who were overweight at age 18 years versus 1.40 (95% CI=1.19, 1.66) among those who were not overweight. Stratified analyses for sexual abuse are presented to parallel the physical abuse results, but the sexual abuse–overweight interaction was not significant.
Table 3.
Early Life Abuse Associations With Gestational Diabetes by Overweight Status (BMI>25 kg/m2) at Age 18 Years
| Exposure | Adjusted for overweight statusa at age 18 |
Stratified by weight status at age 18b |
|||||||
|---|---|---|---|---|---|---|---|---|---|
| Not overweight at age 18 |
Overweight at age 18 |
||||||||
| N(cases) | RRc | (95%CI) | N(cases) | RRc | (95% CI) | N(cases) | RRc | (95% CI) | |
| Physical abused | |||||||||
| None | 21,216(1,053) | 1 | -- | 19,468(948) | 1 | -- | 1,554(89) | 1 | -- |
| Mild | 8,532(465) | 1.09 | (0.97,1.23) | 7,802(403) | 1.05 | (0.93,1.20) | 659(59) | 1.52 | (1.04,2.21) |
| Moderate | 11,983(685) | 1.16 | (1.04,1.29) | 10,843(589) | 1.10 | (0.98,1.24) | 1,034(89) | 1.74 | (1.26,2.40) |
| Severe | 3,684(258) | 1.43 | (1.22,1.67) | 3,293(228) | 1.40 | (1.19,1.66) | 358(29) | 1.74 | (1.09,2.78) |
| Sexual abuse | |||||||||
| None | 30,372(1,561) | 1 | -- | 27,834(1,396) | 1 | -- | 2,279(147) | 1 | -- |
| Touch only |
10,094(574) | 1.10 | (0.99,1.22) | 9,130(492) | 1.06 | (0.95,1.19) | 867(74) | 1.40 | (1.03,1.90) |
| Forced sex | 4,806(318) | 1.31 | (1.15,1.50) | 4,312(272) | 1.30 | (1.12,1.50) | 449(45) | 1.40 | (0.97,2.03) |
BMI>25kg/m2
Wald p for interaction with physical abuse=0.045 for physical abuse; Wald p for interaction with sexual abuse=0.248
Adjusted for maternal age at pregnancy, maternal year of birth, race, mother’s educational attainment, father’s educational attainment, mother in professional occupation, father in professional occupation, parental home ownership, age 5 body size (somatogram), parental history of diabetes.
Mild physical abuse was defined as being pushed, grabbed, or shoved at any frequency or being kicked, bitten, or punched once or hit with something once; moderate physical abuse was defined as being hit with something more than once or physically attacked once, and severe physical abuse was defined as being kicked, bitten, punched, or physically attacked more than once or ever choked or burned.
In the subsample of women with at least one pregnancy after study baseline (n=14,643), moderate to severe physical abuse (combined for this smaller subsample) was associated with an adjusted 15% higher risk of being overweight in prepregnancy (RR=1.15, 95% CI=1.08, 1.23), and forced sex was associated with a 26% higher risk (RR=1.26, 95% CI=1.16, 1.38, data not shown). Prepregnancy overweight was in turn associated with a doubling of GDM risk (RR=2.25, 95% CI=1.98, 2.55, data not shown). Abuse–GDM associations in this subsample were similar to those in the main analysis (moderate to severe physical abuse, RR=1.30, 95% CI=1.12, 1.51; forced sex, RR=1.34, 95% CI=1.09, 1.65, Table 4). Adjustment for prepregnancy overweight attenuated these associations slightly, to 1.25 (95% CI=1.08, 1.45) for moderate to severe physical abuse and 1.25 (95% CI=1.01, 1.53) for forced sex. Adjustment for continuous prepregnancy BMI and change in BMI from age 18 years to prepregnancy produced similar results. Interactions between abuse and prepregnancy overweight were not significant. Adjustment for prepregnancy smoking did not have an important impact on effect estimates (data not shown).
Table 4.
Early Life Abuse Associations With Gestational Diabetes, Adjusted for Pre-Pregnancy Overweight Status (BMI >25kg/m2) a
| Adjustedb |
Additionally adjusted for pre- pregnancy overweightc |
||||
|---|---|---|---|---|---|
| Exposure | N (cases) | RR | (95% CI) | RR | (95% CI) |
| Physical abuse | |||||
| None | 6,035 (458) | 1 | -- | 1 | -- |
| Mild | 2,413 (208) | 1.20 | (1.00, 1.43) | 1.18 | (0.99, 1.41) |
| Moderate or severe | 3,921 (394) | 1.30 | (1.12, 1.51) | 1.25 | (1.08, 1.45) |
| Sexual abuse | |||||
| None | 8,608 (687) | 1 | -- | 1 | -- |
| Touch only | 2,611 (252) | 1.13 | (0.96, 1.32) | 1.11 | (0.95, 1.30) |
| Forced sex | 1,126 (118) | 1.34 | (1.09, 1.65) | 1.25 | (1.01, 1.53) |
Among women with ≥1 pregnancy after 1989, n=14,643
Adjusted for maternal age at pregnancy, maternal year of birth, race, mother’s educational attainment, father’s educational attainment, mother in professional occupation, father in professional occupation, parental home ownership, age 5 body size (somatogram), and parental history of diabetes.
BMI >25kg/m2 reported on questionnaire prior to pregnancy; in cases where women were pregnant during the questionnaire cycle prior to their first birth (e.g., a woman could have had her first birth in 1992 and been pregnant at the time she reported her BMI on the 1991 questionnaire), we used the reported BMI from the questionnaire prior to the one in which she reported being pregnant (e.g., for woman with her first birth in 1992, who was pregnant during report of BMI in 1991, we used BMI in 1989 as her pre-pregnancy BMI).
Discussion
Early life physical or sexual abuse was associated with elevated risk of GDM. Severity of physical abuse was related to GDM in a dose–response manner, with moderate physical abuse associated with a 16% greater GDM risk and severe physical abuse associated with a 42% greater risk, compared with no physical abuse. Forced sex was associated with a 30% greater risk compared with no sexual abuse. Experiencing sexual abuse in addition to physical abuse elevated the risks further, with the approximately 5% of women who experienced severe physical abuse in addition to sexual abuse having an almost 70% greater risk of GDM than their non-abused counterparts.
Only a handful of studies to date have examined early life abuse exposure as a risk factor for poor pregnancy health; these have found early life abuse to predict prenatal depression,35 preterm birth,36,37 low birth weight,37 and cesarean section.38 To the authors’ knowledge, this is the first study to examine GDM risk in relation to early life abuse.
These findings build upon previous results in this cohort showing that early life abuse exposure predicts Type 2 diabetes in middle age—a metabolic condition that is strongly correlated with GDM.12 Current findings suggest that metabolic risk in abused and non-abused women diverges by reproductive age, consistent with previous prospective work documenting abuse-related divergence in weight status by late adolescence.15
As abuse has been associated with heavier adult weight status in several studies,13–18,39 the authors hypothesized that weight status might explain the elevated risk of GDM in abused women. However, they did not find that adjustment for either being overweight at age 18 years or at prepregnancy attenuated the associations between physical and sexual abuse and GDM risk, indicating that weight at these time points is not a driving factor. This suggests that other abuse–GDM pathways may be important; for example, chronic elevation in stress hormones in response to abuse may contribute to insulin resistance even in the absence of observable weight status changes.20–22
Limitations
Limitations of this observational study include retrospective reports of both early life abuse and pregnancy outcomes. Because most abuse is not reported to authorities, there is no gold standard against which to validate the retrospective abuse reports in this cohort. Reassuringly, when abuse–obesity associations have been compared across self-reported and externally substantiated abuse, results are similar.39 GDM data were also retrospectively collected in this cohort, but validation against medical records of GDM reported on previous biennial questionnaires suggests that women in this cohort report GDM with a high degree of accuracy.26 Weight status may be also misclassified, although validation studies indicate good agreement with measured weight.28 Prepregnancy or prenatal depression may play a role in the association between abuse and GDM, but the authors were unable to examine this owing to lack of data on these variables in the cohort. As with any observational study, the results may be influenced by residual confounding. Finally, the NHSII is a non-representative cohort of mostly white women; however, the observed associations are likely driven by mechanisms that are generalizable across population groups.
Findings indicate that highly stressful events in early life should be further explored as determinants of poor pregnancy health and might be used to more accurately identify women at increased risk for GDM. Those designing interventions to prevent GDM might explore expanding the scope of intervention strategies, which currently focus largely on nutrition, exercise, and weight control.40 If replicated, these findings suggest that development and testing of tailored interventions that address psychological and social sequelae of early abuse exposures might be worthwhile.
Conclusions
This study finds observational evidence that abuse victimization elevates women’s risk of GDM in pregnancy, and that this is not explained by previously observed links between abuse victimization and young adult weight status. Additional work is needed to further explore mechanisms and to develop and test interventions that can address the unique risks faced by women with abuse histories.
Acknowledgements
Dr. Mason’s time is supported by the Building Interdisciplinary Research Careers in Women’s Health Grant (K12HD055887) from the Eunice Kennedy Shriver National Institutes of Child Health and Human Development (NICHD), the Office of Research on Women’s Health, and the National Institute on Aging, NIH, administered by the University of Minnesota Deborah E. Powell Center for Women’s Health. The content is solely the responsibility of the authors and does not necessarily represent the office views of the NICHD or NIH. Dr. Clark is supported by grant R03HD068045 from NICHD. Additional support was provided by grant RO1HL081557 from NIH. Nurses’ Health Study II infrastructure is supported by grant UM1CA176726 from NIH.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
No financial disclosures were reported by the authors of this paper.
References
- 1.American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2011;34(Suppl 1):62–69. doi: 10.2337/dc11-S062. http://dx.doi.org/10.2337/dc11-S062. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Alberico S, Montico M, Barresi V, et al. The role of gestational diabetes, pre-pregnancy body mass index and gestational weight gain on the risk of newborn macrosomia: results from a prospective multicentre study. BMC Pregnancy Childbirth. 2014;14(1):23–30. doi: 10.1186/1471-2393-14-23. http://dx.doi.org/10.1186/1471-2393-14-23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Gorgal R, Gonçalves E, Barros M, et al. Gestational diabetes mellitus: a risk factor for non-elective cesarean section. J Obstet Gynaecol Res. 2012;38(1):154–159. doi: 10.1111/j.1447-0756.2011.01659.x. http://dx.doi.org/10.1111/j.1447-0756.2011.01659.x. [DOI] [PubMed] [Google Scholar]
- 4.Saldana TM, Siega-Riz AM, Adair LS, Suchindran C. The relationship between pregnancy weight gain and glucose tolerance status among black and white women in central North Carolina. Am J Obstet Gynecol. 2006;195(6):1629–1635. doi: 10.1016/j.ajog.2006.05.017. http://dx.doi.org/10.1016/j.ajog.2006.05.017. [DOI] [PubMed] [Google Scholar]
- 5.Link CL, McKinlay JB. Disparities in the prevalence of diabetes: Is it race/ethnicity or socioeconomic status? Results from the Boston Area Community Health (BACH) survey. Ethn Dis. 2009;19(3):288–292. [PMC free article] [PubMed] [Google Scholar]
- 6.Hosler AS, Nayak SG, Radigan AM. Stressful events, smoking exposure and other maternal risk factors associated with gestational diabetes mellitus. Paediatr Perinat Epidemiol. 2011;25(6):566–574. doi: 10.1111/j.1365-3016.2011.01221.x. http://dx.doi.org/10.1111/j.1365-3016.2011.01221.x. [DOI] [PubMed] [Google Scholar]
- 7.Ertel KA, Silveira M, Pekow P, et al. Prenatal depressive symptoms and abnormalities of glucose tolerance during pregnancy among Hispanic women. Arch Womens Ment Health. 2014;17(1):65–72. doi: 10.1007/s00737-013-0379-2. http://dx.doi.org/10.1007/s00737-013-0379-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Bowers K, Laughon SK, Kim S, et al. The association between a medical history of depression and gestational diabetes in a large multi-ethnic cohort in the United States. Paediatr Perinat Epidemiol. 2013;27(4):323–328. doi: 10.1111/ppe.12057. http://dx.doi.org/10.1111/ppe.12057. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Tjaden P, Thoennes N. Prevalence, incidence, and consequences of violence against women: findings from the National Violence Against Women Survey. Washington, DC: U.S. Department of Justice; 2000. [Google Scholar]
- 10.Black M, Basile K, Breiding M, et al. The National Intimate Partner and Sexual Violence Survey (NISVS): 2010 Summary Report. Atlanta, GA: National Center for Injury Prevention and Control, CDC; 2011. [Google Scholar]
- 11.Rich-Edwards JW, Spiegelman D, Lividoti Hibert EN, et al. Abuse in childhood and adolescence as a predictor of type 2 diabetes in adult women. Am J Prev Med. 2010;39(6):529–536. doi: 10.1016/j.amepre.2010.09.007. http://dx.doi.org/10.1016/j.amepre.2010.09.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Malcolm J. Through the looking glass: Gestational diabetes as a predictor of maternal and offspring long-term health. Diabetes Metab Res Rev. 2012;28(4):307–311. doi: 10.1002/dmrr.2275. http://dx.doi.org/10.1002/dmrr.2275. [DOI] [PubMed] [Google Scholar]
- 13.Felitti VJ. Childhood sexual abuse, depression, and family dysfunction in adult obese patients: a case control study. South Med J. 1993;86:732–736. doi: 10.1097/00007611-199307000-00002. http://dx.doi.org/10.1097/00007611-199307000-00002. [DOI] [PubMed] [Google Scholar]
- 14.Johnson JG, Cohen P, Kasen S, Brook JS. Childhood adversities associated with risk for eating disorders or weight problems during adolescence or early adulthood. Am J Psychiatry. 2002;159(3):394–400. doi: 10.1176/appi.ajp.159.3.394. http://dx.doi.org/10.1176/appi.ajp.159.3.394. [DOI] [PubMed] [Google Scholar]
- 15.Noll JG, Zeller MH, Trickett PK, Putnam FW. Obesity risk for female victims of childhood sexual abuse: a prospective study. Pediatrics. 2007;120(1):e61–e67. doi: 10.1542/peds.2006-3058. http://dx.doi.org/10.1542/peds.2006-3058. [DOI] [PubMed] [Google Scholar]
- 16.Boynton-Jarrett R, Rosenberg L, Palmer JR, Boggs DA, Wise LA. Child and Adolescent Abuse in Relation to Obesity in Adulthood: The Black Women’s Health Study. Pediatrics. 2012;130(2):245–253. doi: 10.1542/peds.2011-1554. http://dx.doi.org/10.1542/peds.2011-1554. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Shin SH, Miller DP. A longitudinal examination of childhood maltreatment and adolescent obesity: Results from the National Longitudinal Study of Adolescent Health (AddHealth) Study. Child Abuse Negl. 2012;36(2):84–94. doi: 10.1016/j.chiabu.2011.08.007. http://dx.doi.org/10.1016/j.chiabu.2011.08.007. [DOI] [PubMed] [Google Scholar]
- 18.Mamun AA, Lawlor DA, O’Callaghan MJ, Bor W, Williams GM, Najman JM. Does childhood sexual abuse predict young adult’s BMI? A birth cohort study. Obesity (Silver Spring) 2007;15(8):2103–2110. doi: 10.1038/oby.2007.250. http://dx.doi.org/10.1038/oby.2007.250. [DOI] [PubMed] [Google Scholar]
- 19.Hollingsworth K, Callaway L, Duhig M, Matheson S, Scott J. The Association between Maltreatment in Childhood and Pre-Pregnancy Obesity in Women Attending an Antenatal Clinic in Australia. PLoS One. 2012;7(12):51868. doi: 10.1371/journal.pone.0051868. http://dx.doi.org/10.1371/journal.pone.0051868. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Pervanidou P, Chrousos GP. Metabolic consequences of stress during childhood and adolescence. Metabolism. 2012;61(5):611–619. doi: 10.1016/j.metabol.2011.10.005. http://dx.doi.org/10.1016/j.metabol.2011.10.005. [DOI] [PubMed] [Google Scholar]
- 21.Gragnoli C. Hypothesis of the neuroendocrine cortisol pathway gene role in the comorbidity of depression, type 2 diabetes, and metabolic syndrome. Appl Clin Genet. 2014;7:43–53. doi: 10.2147/TACG.S39993. http://dx.doi.org/10.2147/TACG.S39993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Kyrou I, Tsigos C. Stress hormones: physiological stress and regulation of metabolism. Curr Opin Pharmacol. 2009;9(6):787–793. doi: 10.1016/j.coph.2009.08.007. http://dx.doi.org/10.1016/j.coph.2009.08.007. [DOI] [PubMed] [Google Scholar]
- 23.Strauss MA, Gelles RJ. Physical Violence in American Families: Risk Factors and Adaptations to Violence in 8,145 Families. New Brunswick: Transaction Publishers; 1990. [Google Scholar]
- 24.Rich-Edwards JW, Mason S, Rexrode K, et al. Physical and Sexual Abuse in Childhood as Predictors of Early-Onset Cardiovascular Events in Women. Circulation. 2012;126(8):920–927. doi: 10.1161/CIRCULATIONAHA.111.076877. http://dx.doi.org/10.1161/CIRCULATIONAHA.111.076877. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Finkelhor D, Moore D, Hamby SL, Straus MA. Sexually abused children in a national survey of parents: methodological issues. Child Abuse Negl. 1997;21(1):1–9. doi: 10.1016/s0145-2134(96)00127-5. http://dx.doi.org/10.1016/S0145-2134(96)00127-5. [DOI] [PubMed] [Google Scholar]
- 26.Solomon CG, Willett WC, Rich-Edwards J, et al. Variability in diagnostic evaluation and criteria for gestational diabetes. Diabetes Care. 1996;19(1):12–16. doi: 10.2337/diacare.19.1.12. http://dx.doi.org/10.2337/diacare.19.1.12. [DOI] [PubMed] [Google Scholar]
- 27.Stunkard A, Sorenson T, Schulsinger F. The Genetics of Neurological and Psychiatric Disorders. New York: Raven Press; 1983. [Google Scholar]
- 28.Troy LM, Hunter DJ, Manson JE, Colditz GA, Stampfer MJ, Willett WC. The validity of recalled weight among younger women. Int J Obes Relat Metab Disord. 1995;19(8):570–572. [PubMed] [Google Scholar]
- 29.Jun HJ, Rich-Edwards JW, Boynton-Jarrett R, Austin SB, Frazier AL, Wright RJ. Child Abuse and Smoking Among Young Women: The Importance of Severity, Accumulation, and Timing. J Adolesc Heal. 2008;43(1):55–63. doi: 10.1016/j.jadohealth.2007.12.003. http://dx.doi.org/10.1016/j.jadohealth.2007.12.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Moore Simas TA, Szegda K, Liao X, Markenson G, Chasan-Taber L. Cigarette smoking and risk of gestational diabetes mellitus in a cohort of hispanic women. Diabetes Res Clin Pr. 2014;105(1):126–134. doi: 10.1016/j.diabres.2014.04.026. http://dx.doi.org/10.1016/j.diabres.2014.04.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Wendland EMDR, Duncan BB, Belizán JM, Vigo A, Schmidt MI. Gestational diabetes and pre-eclampsia: common antecedents? Arq Bras Endocrinol Metabol. 2008;52(6):975–984. doi: 10.1590/s0004-27302008000600008. http://dx.doi.org/10.1590/S0004-27302008000600008. [DOI] [PubMed] [Google Scholar]
- 32.Zou G. A Modified Poisson Regression Approach to Prospective Studies with Binary Data. Am J Epidemiol. 2004;159(7):702–706. doi: 10.1093/aje/kwh090. http://dx.doi.org/10.1093/aje/kwh090. [DOI] [PubMed] [Google Scholar]
- 33.Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics. 1986;42(1):121–130. http://dx.doi.org/10.2307/2531248. [PubMed] [Google Scholar]
- 34.Vanderweele TJ, Vansteelandt S. Odds ratios for mediation analysis for a dichotomous outcome. Am J Epidemiol. 2010;172(12):1339–1348. doi: 10.1093/aje/kwq332. http://dx.doi.org/10.1093/aje/kwq332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Rich-Edwards JW, James-Todd T, Mohllajee A, et al. Lifetime maternal experiences of abuse and risk of pre-natal depression in two demographically distinct populations in Boston. Int J Epidemiol. 2011;40(2):375–384. doi: 10.1093/ije/dyq247. http://dx.doi.org/10.1093/ije/dyq247. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Seng JS, Low LK, Sperlich M, Ronis DL, Liberzon I. Post-traumatic stress disorder, child abuse history, birthweight and gestational age: A prospective cohort study. BJOG An Int J Obstet Gynaecol. 2011;118(11):1329–1339. doi: 10.1111/j.1471-0528.2011.03071.x. http://dx.doi.org/10.1111/j.1471-0528.2011.03071.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Gavin AR, Hill KG, Hawkins JD, Maas C. The role of maternal early-life and later-life risk factors on offspring low birth weight: Findings from a three-generational study. J Adolesc Heal. 2011;49(2):166–171. doi: 10.1016/j.jadohealth.2010.11.246. http://dx.doi.org/10.1016/j.jadohealth.2010.11.246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Lukasse M, Vangen S, Øian P, Schei B. Childhood abuse and caesarean section among primiparous women in the Norwegian mother and child cohort study. BJOG An Int J Obstet Gynaecol. 2010;117(9):1153–1157. doi: 10.1111/j.1471-0528.2010.02627.x. http://dx.doi.org/10.1111/j.1471-0528.2010.02627.x. [DOI] [PubMed] [Google Scholar]
- 39.Danese A, Tan M. Childhood maltreatment and obesity: systematic review and meta-analysis. Mol Psychiatry. 2014;19(5):544–554. doi: 10.1038/mp.2013.54. http://dx.doi.org/10.1038/mp.2013.54. [DOI] [PubMed] [Google Scholar]
- 40.Muktabhant B, Lumbiganon P, Ngamjarus C, Dowswell T. Interventions for preventing excessive weight gain during pregnancy (Review) Cochrane database Syst Rev Online. 2012;4:CD007145. doi: 10.1002/14651858.CD007145.pub2. http://dx.doi.org/10.1002/14651858.CD007145.pub2. [DOI] [PMC free article] [PubMed] [Google Scholar]