Objective
We evaluated the relationship between antenatal depressive symptoms and preterm birth.
Study Design
Patients completed the Edinburgh Postnatal Depression Scale between 24-28 weeks of gestation. A score ≥12 (or thoughts of self-harm) indicated an at-risk woman. Symptomatic women were compared to risk-negative patients for relevant demography, historical variables, and pregnancy outcome.
Results
After screening 14,175 women we found a screen positive rate of 9.1% (n = 1298). At-risk women had a significant increase in preterm birth at <37, <34, <32, and <28 weeks of gestation. Multivariable analysis adjusting for maternal age, race/ethnicity, prior preterm delivery, and insurance status revealed a persistent association between antenatal depressive symptoms and preterm birth (adjusted odds ratio, 1.3; 95% confidence interval, 1.09–1.35), which was also observed after multiple gestations were excluded from the analysis (odds ratio, 1.7; 95% confidence interval, 1.38–1.99).
Conclusion
In this large cohort of prenatally screened women, those with depressive symptoms had an increased likelihood of preterm birth.
Preterm birth affects approximately 12% of pregnancies in the United States and approximately 10.8% of pregnancies at our institution. A number of maternal risk factors linked to preterm delivery have been described and confirmed. An association with perinatal depression has been suggested in the literature but a review of relevant studies yields conflicting results. The variability of findings to date may be attributable to methodological issues, including the suitability of the specific depression screening instrument to pregnant women, the timing and frequency of screening, lack of control for a variety of potentially confounding maternal factors associated with preterm birth, and limitations of the sample sizes available for study. Because perinatal depression is a common complication of pregnancy (occurring in up to 13% of women ), clarifying the degree to which this association exists is an important objective. Further, if a relationship can be confirmed, the possibility that treatment of this mood disorder might influence the rate of preterm birth would be appropriate to study.
Materials and Methods
A universal perinatal depression screening program was initiated in 2003 with the intention of screening all women during their pregnancy (24-28 weeks) coincident with their diabetes screening in the office setting. The Edinburgh Postnatal Depression Scale (EPDS) was chosen based on its validity during pregnancy and after delivery, its availability in multiple languages, and its relatively high sensitivity and specificity when compared to other tests. All completed screens were electronically faxed from clinical sites and the responses were entered centrally, scored automatically, and reviewed in real time by screening personnel. Screens were considered at risk if the woman had a score of ≥12 or answered any response other than “never” to the question describing thoughts of self-harm. Positive screening results were given to the program director (psychologist), medical director (psychiatrist), or coordinator (licensed social worker) for immediate patient telephone contact, evaluation, and appropriate triage. An interpretative language telephone line allowed for evaluation in each woman’s primary language. During the call, all women were also reminded about the availability of a 24/7 crisis hotline developed by our program and administered in collaboration with the Department of Health and Human Services for statewide use in Illinois.
Linkage of screening results with the patients’ obstetrics records was accomplished with our enterprisewide electronic health information system (Epic Systems Corp., Verona, WI), providing the opportunity to examine the relationship between depressive symptoms and subsequent preterm birth. To evaluate the primary outcome of preterm delivery, women were categorized as “at risk” or “not at risk” based on their prenatal screening results. Symptomatic women were compared to the risk-negative cohort for relevant demography, historical variables, and pregnancy outcomes that were readily available in Epic. These variables included maternal age, race, a history of preterm delivery, and insurance status. Other data relevant to the comparison, such as tobacco and alcohol use, the gestational age at which a prior preterm birth occurred, and prepregnancy body mass index, were not consistently available for inclusion. Duplicate screens completed over >1 pregnancy during the study time frame were excluded.
Statistical analysis was conducted with software (SPSS, version 15.0 for Windows; IBM Corp, Armonk, NY) with χ 2 and t test for categorical and continuous data, respectively. Multivariable logistic regression was performed to evaluate the impact of potential confounders on the relationship between preterm delivery and risk of depression. The Breslow-Day test was used to evaluate stratified odds ratios (ORs) for other potential confounders. A P value < .05 was considered statistically significant. For a 3% difference between preterm birth rates in at-risk women vs not-at-risk cases (using a baseline preterm birth rate of 10.8% from our center), a sample size of 958 cases and 9574 controls would be needed (95% confidence interval [CI], 90% power); as such we were sufficiently powered to detect or exclude this difference. All cases were deidentified prior to analysis and the study was provided with expedited approval by our institutional review board (EH10-366).
Results
From January 2003 through March 2011, 15,254 prenatal depression screens were completed among the 39,527 women delivering in that time frame. Reasons for failed screening included patient refusal, incomplete adherence to departmental recommendations for screening by individual practices, and late transfer of care. Unscreened subjects experienced a similar overall rate of preterm delivery (10.9%) compared to the screened cohort of women (10.6%). In all, 1079 women had multiple pregnancies screened during the study period and were excluded, leaving 14,175 unique women with prenatal EPDS screens available for analysis. Of these, 1298 women were screen positive (9.1%). This figure is slightly higher than the screen-positive rate of 7.7%, which was previously published from our center containing an original subset of 1548 women (also included in the current study), but within the range of prior observations using EPDS.
The demographic data of the women at risk for depression were significantly different from the women not at risk for depression ( Table 1 ). Symptomatic patients were, on average, older ( P < .001), overrepresented by minorities (African American and Hispanic; P < .001), more likely to be receiving public assistance ( P < .001), and less often nulliparous ( P = .03). Prior preterm birth was more common among symptomatic women compared to risk-negative cases (21.3% vs 13.7%; P < .001).
| Demographic | Not at risk for depression, n = 12,877 | At risk for depression, a n = 1298 |
|---|---|---|
| Maternal age (y), mean ± SD | 32.3 ± 5.5 | 31.3 ± 6.2 |
| Nulliparous, % | 42.2 | 39.1 |
| Prior preterm birth, % b | 9.9 | 6.9 |
| Race/ethnicity, % | ||
| Caucasian | 70.8 | 50.9 |
| African American | 7.8 | 17.0 |
| Hispanic | 13.7 | 21.3 |
| Other | 7.8 | 10.8 |
| Public assistance, % | 19.4 | 38.2 |
a All comparisons were significant ( P < .05);
Table 2 summarizes the primary outcome comparison of preterm birth between groups. The symptomatic cohort had a significant increase in preterm birth at <37 weeks (13.9% vs 10.3%) and significant differences were seen at additional gestational age breakpoints (<34, <32, and <28 weeks of gestation). The distribution between unintended (spontaneous) preterm birth and medically indicated preterm delivery, as well as cases in which the reason for preterm delivery could not be determined from the database did not significantly differ between the at-risk vs not-at-risk cohort ( P = .11). Finally, among the 1019 patients with prior preterm birth, women who screened positive were more likely to deliver preterm than were screen-negative women (OR, 1.59; 95% CI, 1.07–2.36; P = .021).
| Gestational age at delivery, wk | Not at risk for depression, n = 12,877 n (%) | At risk for depression, n = 1298 n (%) | P value |
|---|---|---|---|
| <37 | 1322 (10.3) | 181 (13.9) | < .001 |
| <34 | 349 (2.7) | 54 (4.2) | .005 |
| <32 | 156 (1.2) | 25 (1.9) | .04 |
| <28 | 29 (0.2) | 8 (0.6) | .023 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
