Objective
Uterine leiomyomata (UL) are a major source of gynecologic morbidity and the primary indication for hysterectomy. Depression can cause dysregulation of the hypothalamic-pituitary-adrenal axis, which may affect the synthesis of reproductive hormones involved in UL pathogenesis. We assessed the association between depressive symptoms and UL among 15,963 premenopausal women.
Study Design
Data were derived from the Black Women’s Health Study, a prospective cohort study. In 1999 and 2005, the Center for Epidemiologic Studies Depression Scale (CES-D) was used to ascertain depressive symptoms. On biennial follow-up questionnaires from 1999 through 2011, women reported physician-diagnosed depression, antidepressant use, and UL diagnoses. Incidence rate ratios (IRRs) and 95% confidence intervals (CIs) were estimated using multivariable Cox regression.
Results
There were 4722 incident UL cases diagnosed by ultrasound (n = 3793) or surgery (n = 929) during 131,262 person-years of follow-up. Relative to baseline CES-D scores <16, IRRs were 1.05 (95% CI, 0.98–1.13) for CES-D scores 16-24 and 1.16 (95% CI, 1.06–1.27) for CES-D scores ≥25 ( P -trend = .001). IRRs for current and past physician-diagnosed depression relative to no depression were 1.15 (95% CI, 0.98–1.34) and 1.25 (95% CI, 1.13–1.39), respectively. Results persisted after further control for antidepressant use. IRRs for current and past use of antidepressants (any indication) relative to never use were 1.11 (95% CI, 0.97–1.28) and 1.32 (95% CI, 1.14–1.52), respectively.
Conclusion
In this cohort of black women, greater depressive symptoms were associated with UL, independent of antidepressant use, supporting the hypothesis that dysregulation of the hypothalamic-pituitary-adrenal axis increases UL risk.
Uterine leiomyomata (UL) are clinically recognized in 30% of reproductive-aged women and are a major contributor to gynecologic morbidity and medical costs. African American women are 2-3 times more likely to be diagnosed with UL than white women, but established risk factors do not explain the racial disparity. Sex steroid hormones are thought to influence UL pathogenesis.
UL risk has been linked to greater exposure to psychosocial stress. A cross-sectional study found a higher prevalence of UL among women reporting more major life events and greater “stress intensity.” In the Black Women’s Health Study (BWHS), a prospective cohort study, perceived racism was positively associated with UL diagnoses. In agreement with a previous cohort study of white women, the BWHS documented a positive association between child sexual abuse and incident UL diagnoses.
Major depressive disorder is the second leading cause of years of life lost due to premature mortality in the United States, and the incidence in women is twice that of men. Depression is more common in African Americans than in other ethnic groups in some but not all studies, and African Americans with depression are less likely to take antidepressant medications. Studies limited only to those receiving medical treatment may miss an important segment of the African American population with depression.
A large body of clinical and basic research reports an association between dysregulation of hypothalamic-pituitary-adrenal (HPA) axis hormone dynamics and mood disorders such as depression. HPA-axis abnormalities observed in depressed individuals include hypercortisolemia, as reflected by elevated levels of cortisol in plasma, cerebrospinal fluid, and 24-hour urine overall and in response to psychological stress. Lower levels of dehydroepiandrosterone sulfate (androgen precursor produced by adrenal gland) have also been reported among depressed patients. Alterations in the HPA axis can cause dysregulation of the hypothalamic-pituitary-ovarian axis, which may influence the synthesis of sex steroid hormones. For instance, depressed women were found to have lower serum levels of estradiol but higher serum levels of progesterone compared with nondepressed women and progesterone may have a stimulatory effect on UL. Depressive symptoms can also deleteriously influence health-related behaviors, including sedentariness, alcohol consumption, and poor diet, all of which may increase UL risk.
In a large cohort of African American women, we evaluated prospectively the association between depressive symptoms and UL risk, and the extent to which antidepressant use mediates this association.
Materials and Methods
Study population
The BWHS is an ongoing prospective cohort study of 59,000 African American women aged 21-69 years at baseline. In 1995, subscribers of Essence –a magazine with a large African American female readership–were invited to enroll in a long-term health study by completing comprehensive self-administered baseline questionnaires. Biennial follow-up questionnaires update exposure and medical information; cohort retention was 80% through 2011. Participants reside in >17 states across the United States. The institutional review board of Boston University Medical Center approved the study protocol.
Assessment of depressive symptoms, diagnosed depression, and antidepressant use
The 1999 and 2005 follow-up questionnaires included the 20-item Center for Epidemiologic Studies Depression Scale (CES-D), which was designed to assess depressive symptoms in community samples and population-based studies. Respondents indicated the frequency of various feelings experienced during the previous week on a 4-point scale ranging from “rarely or none of the time” to “most or all of the time.” The measure included items such as “I felt sad,” “I felt lonely,” “I felt depressed,” and “I enjoyed life.” Total scores ranged from 0-60, with higher scores indicating higher levels of depressive symptoms.
Women had the opportunity to report a clinical diagnosis of depression under an open-ended question about “other serious illness” in 1995, or under questions about “depression treated with medication” on biennial follow-up questionnaires from 1997 through 2009. Because African American women are less likely to take medication for depression than other populations, we asked separately about “depression” and “depression treated with medication” in 2005 to capture clinically depressed women we may have missed in previous years. Year of first diagnosis was elicited on all questionnaires.
Users of antidepressant medications were defined as participants who reported using “medication for depression” in 1997; “antidepressants (Prozac, Zoloft, Elavil, etc.)” in 1999, 2001, 2003, or 2005; or antidepressants under “Please list all other medications or supplements that you currently take at least 3 days a week” (all years), regardless of the indication for use. The Slone Drug Dictionary was used to code medications.
Assessment of UL
Incident cases were women who reported a first diagnosis of UL confirmed by ultrasound or surgery. Ultrasound has high sensitivity (99%) and specificity (91%) relative to histologic evidence. On follow-up questionnaires, women reported whether they had been diagnosed with “uterine fibroids” in the previous 2-year interval, the calendar year in which they were first diagnosed, and whether their diagnosis was confirmed by “pelvic exam” and/or by “ultrasound/hysterectomy” (1999 or 2001) or “ultrasound” and “surgery” (2003 or later). To maximize the specificity of UL classification, cases identified by pelvic exam only (N = 350) were treated as noncases.
Assessment of covariates
On the 1995 and biennial follow-up questionnaires, we collected data on reproductive and contraceptive history, gynecologic surgeries, anthropometric factors, physical activity, smoking, alcohol consumption, geographic region of residence, socioeconomic correlates (education, marital status, occupation), medical conditions, and Pap smear frequency. Body mass index (BMI) was calculated as weight in kilograms divided by squared height in meters. We estimated total metabolic equivalents (METs) per week by summing the METs from moderate physical activity (hours/wk × 3.5) and vigorous exercise (hours/wk × 7.0). In 2003, women reported their household income. In 2007, participants reported their recency of pelvic exam (never, <5, 5-9, ≥10 years ago) and pelvic ultrasound (never, <5, 5-9, ≥10 years ago). In 2005, participants completed the 10-item Carver coping scale, with higher scores indicating higher coping skills.
Validation studies
Uterine leiomyomata
We mailed supplemental questionnaires to a random sample of 248 cases and requested permission to review their medical records. We corroborated the self-report in 122 (96%) of 126 women from whom we received medical records. Most cases (87%) reported that their condition came to clinical attention because they sought treatment for symptoms or a tumor was palpable during a routine pelvic exam. Details about the validation study may be found elsewhere.
Depressive symptoms
The validity and reliability of the CES-D have been documented in several populations, including the BWHS. In the BWHS, factor analysis indicated that the 4-factor structure of the CES-D was supported, and Cronbach alphas demonstrated high internal consistency of the CES-D responses.
Restriction criteria
Of the 22,284 BWHS premenopausal participants at baseline who had never been diagnosed with UL, we excluded women who were aged ≥50 years at baseline (n = 1576) or lost to follow-up after 1999 (n = 890); cases that did not report a year of diagnosis (n = 149) or method of confirmation (n = 110); and women with incomplete data on the CES-D (n = 3043) or covariates (n = 553), leaving 15,963 for analysis. Women with incomplete data on depressive symptoms tended to be older (mean age: 37.3 vs 36.0 years), have a higher prevalence of clinical depression (ever: 7.0% vs 6.2%), be current smokers (14.5% vs 12.7%), have fewer years of education (mean: 14.6 vs 14.9), and be parous (66% vs 59%), but were similar with respect to menarcheal age (mean, 12.4 vs 12.3 years) and BMI (28.9 vs 28.6). The age-adjusted incidence of UL among those excluded (34.4 cases per 1000 person-years) was similar to those included (36.0 cases per 1000 person-years).
Data analysis
Cox regression was used to estimate incidence rate ratios (IRRs) and 95% confidence intervals (CI). We categorized depressive symptoms as “low” (CES-D <16), “moderate” (CES-D: 16-24), and “severe” (CES-D ≥25). The reference category for all analyses was “CES-D <16.” This categorization has been used in previous studies and was informed by previous observations that a mean CES-D score of 24.4 among patients in a private psychiatric facility was strongly associated with major depression. Person-years at risk were calculated from the start of follow-up (March 1999) until first diagnosis of UL, menopause, death, loss to follow-up, or the end of follow-up (March 2011), whichever came first.
The Cox model conditioned on age (1-year intervals) and time period (2-year intervals). A second multivariable model further controlled for age at menarche (≤10, 11, 12-13, 14, ≥15 years), parity (0, 1, 2, ≥3 births), age at first birth (years), years since last birth (<5, 5-9, 10-14, ≥15), age at first oral contraceptive use (years), history of oral contraceptive use (ever, never), BMI (<20, 20-24, 25-29, ≥30 kg/m 2 ), physical activity (<10, 10-19, 20-39, ≥40 MET-hrs/wk), smoking (current, former, never), current alcohol intake (<1, 1-6, ≥7 drinks/wk), participant’s education (≤12, 13-15, 16, ≥17 years), marital status (married/living as married, divorced/separated/widowed, single), occupation (white collar, non-white collar, unemployed), annual household income (≤$25,000, $25,001-50,000, $50,001-100,000, >$100,000, missing), and region of residence (South, Northeast, Midwest, West). In sensitivity analyses, we further stratified by and adjusted for history of depression treated with medication and antidepressant use. Antidepressant use was considered a possible mediator of the association between depressive symptoms and UL, as were other lifestyle factors such as smoking and alcohol consumption. Time-varying variables (eg, antidepressant use, parity, years since last birth) were updated over time using the Anderson-Gill data structure. Control for family history of UL, maternal age at participant’s birth, abuse, victimization, and perceived racism made little difference in the effect estimates and were omitted from multivariable models.
Primary analyses assessed CES-D scores reported at baseline (1999) in relation to UL risk. Secondary analyses assessed different specifications of the CES-D variable, including: (1) updating the variable by using CES-D scores reported in 2005 (for 2005 through 2011 risk period), (2) assigning the 1999 CES-D to the 1999 through 2005 risk period and assigning the average of 1999 and 2005 CES-D measures to the 2005 through 2011 risk period (cumulative-average), and (3) examining the 2005 CES-D measure after restricting to the 2005 through 2011 risk period only.
We used stratification to examine whether associations were modified by other covariates such as coping skills, age, and smoking. P values from interaction tests were obtained using the likelihood ratio test comparing models with and without cross-product terms between depressive symptoms and selected covariates. Departures from proportional hazards were evaluated in the same manner using cross-product terms between depressive symptoms and age (<35, ≥35 years). Analyses were performed using software (SAS, version 9.2; SAS Institute Inc, Cary NC).
Results
Frequencies of CES-D scores <16, 16-24, and ≥25 were 70%, 19%, and 11%, respectively, with 6% reporting a history of physician-diagnosed depression and 4% reporting ever having used antidepressants ( Table 1 ). Higher CES-D scores were associated with younger age at baseline, higher parity, earlier ages at first birth and first oral contraceptive use, greater time since last birth, lower adult household income, higher BMI, greater tobacco and alcohol consumption, and lower coping scores. No appreciable differences were found for recent pelvic exam or ultrasound (within 5 years prior to 2007 questionnaire) across exposure categories. The correlation between CES-D scores reported in 1999 and 2005 among those with complete data for both measures was 0.54 ( P < .0001).
| Characteristic a | CES-D score in 1999 | Physician-diagnosed clinical depression (ever) | Use of antidepressants (ever) | ||||
|---|---|---|---|---|---|---|---|
| <16 | 16-24 | ≥25 | No | Yes | No | Yes | |
| No. of women | 11,165 | 3018 | 1780 | 14,959 | 1004 | 15,381 | 582 |
| Age at baseline, y (mean) | 36.3 | 35.4 | 35.1 | 35.9 | 37.4 | 35.9 | 37.9 |
| Ever use of antidepressants, % | 2.0 | 5.2 | 11.7 | 1.2 | 38.6 | 0.0 | 100.0 |
| Current use of antidepressants, % | 1.4 | 3.7 | 9.9 | 0.8 | 30.4 | 0.0 | 73.5 |
| Age at menarche, y (mean) | 12.3 | 12.3 | 12.3 | 12.4 | 12.2 | 12.3 | 12.2 |
| Parous, yes, % | 58.4 | 59.3 | 59.9 | 58.9 | 57.3 | 58.9 | 55.9 |
| Age at first birth, y (mean) b | 24.1 | 23.3 | 22.5 | 23.9 | 22.3 | 23.9 | 22.2 |
| Time since last birth, y (mean) b | 9.9 | 10.5 | 10.8 | 10.1 | 11.0 | 10.1 | 11.4 |
| Age at first OC use, y (mean) c | 19.1 | 19.1 | 18.9 | 19.1 | 18.5 | 19.1 | 18.3 |
| Education, y (mean) | 15.1 | 14.7 | 14.5 | 14.9 | 14.8 | 14.9 | 14.8 |
| Married or living with partner, % | 44.8 | 37.1 | 32.9 | 42.7 | 31.8 | 42.3 | 33.8 |
| Income ≤$25,000 in 2003, % | 6.9 | 11.4 | 16.4 | 8.4 | 13.0 | 8.6 | 11.9 |
| Body mass index, kg/m 2 (mean) | 28.2 | 29.4 | 30.2 | 28.5 | 30.2 | 28.6 | 30.9 |
| Metabolic equivalents of physical activity per week (mean) | 22.5 | 19.6 | 18.9 | 21.8 | 18.4 | 21.7 | 17.3 |
| Family history of UL, % | 35.8 | 34.6 | 35.7 | 35.4 | 37.2 | 35.6 | 35.7 |
| Smoking, current, % | 10.7 | 16.1 | 19.9 | 12.2 | 19.9 | 12.5 | 19.6 |
| Alcohol, current ≥7 drinks/wk, % | 2.9 | 4.8 | 7.0 | 3.5 | 5.2 | 3.6 | 5.1 |
| Coping score in 2005 (mean) | 29.3 | 28.3 | 27.5 | 28.9 | 28.8 | 28.9 | 28.4 |
| Pelvic exam and/or ultrasound <5 y of 2007 questionnaire, % | 80.2 | 75.4 | 71.2 | 78.3 | 77.8 | 78.3 | 78.0 |
a Characteristics are presented as means or percentages within depression categories and are age-standardized to distribution of cohort in 1999
b Restricted to parous women only
During 131,262 person-years of follow-up, there were 4722 incident UL cases confirmed by ultrasound (n = 3793) or surgery (n = 929). Higher CES-D scores were associated with higher UL risk ( Table 2 ). Relative to baseline CES-D scores <16, multivariable-adjusted IRRs for CES-D scores 16-24 and ≥25 were 1.05 (95% CI, 0.98–1.13) and 1.16 (95% CI, 1.06–1.27), respectively ( P -trend = .001). Associations were similar when we used the cumulative average of the 1999 and 2005 CES-D scores ( P -trend < .001) and when we updated scores over time ( P -trend < .001). When we restricted the incident period to 2005 through 2011, the IRRs for 2005 CES-D 16-24 and ≥25 vs <16 were 1.22 and 1.24, respectively ( P -trend = .003). The respective IRRs were 1.37 (95% CI, 1.10–1.69) and 1.36 (95% CI, 1.06–1.75) among women with low coping skills, and 1.09 (95% CI, 0.87–1.36) and 1.10 (95% CI, 0.83–1.47) among women with high coping skills, but the IRRs were not statistically different across coping strata ( P- interaction = .34). Women with CES-D scores ≥25 in 1999 and 2005 had 1.45 times the UL risk of women whose scores were <16 in both years (95% CI, 1.11–1.88). Associations were somewhat stronger among women with greater gynecologic surveillance, as indicated by a recent pelvic exam and/or ultrasound: IRRs for CES-D scores 16-24 and ≥25 vs <16 were 1.05 (95% CI, 0.97–1.14) and 1.21 (95% CI, 1.10–1.34), respectively. With regard to possible mediation by covariates, control for current smoking, alcohol consumption, BMI, and hormonal contraceptive use made little difference in the associations (data not shown).
| Characteristic | Cases | Person-years | Age-adjusted model a | Multivariable model b | ||
|---|---|---|---|---|---|---|
| IRR | 95% CI | IRR | 95% CI | |||
| CES-D score in 1999 | ||||||
| <16 | 3254 | 92,554 | 1.00 | Ref | 1.00 | Ref |
| 16-24 | 898 | 24,600 | 1.04 | 0.96–1.12 | 1.05 | 0.98–1.13 |
| ≥25 | 570 | 14,108 | 1.15 | 1.05–1.25 | 1.16 | 1.06–1.27 |
| P value, test for trend c | .004 | .001 | ||||
| CES-D score in 2005 d | ||||||
| <16 | 792 | 26,289 | 1.00 | Ref | 1.00 | Ref |
| 16-24 | 213 | 5588 | 1.20 | 1.04–1.39 | 1.22 | 1.05–1.41 |
| ≥25 | 139 | 3501 | 1.24 | 1.04–1.48 | 1.24 | 1.04–1.49 |
| P value, test for trend c | .001 | .003 | ||||
| Updated CES-D score e | ||||||
| <16 | 3237 | 93,300 | 1.00 | Ref | 1.00 | Ref |
| 16-24 | 909 | 23,753 | 1.09 | 1.02–1.18 | 1.11 | 1.03–1.20 |
| ≥25 | 576 | 14,209 | 1.17 | 1.07–1.28 | 1.19 | 1.09–1.30 |
| P value, test for trend c | < .001 | < .001 | ||||
| Cumulative average CES-D score e | ||||||
| <16 | 3257 | 93,474 | 1.00 | Ref | 1.00 | Ref |
| 16-24 | 930 | 24,603 | 1.08 | 1.00–1.16 | 1.10 | 1.02–1.18 |
| ≥25 | 535 | 13,185 | 1.16 | 1.05–1.27 | 1.18 | 1.07–1.30 |
| P value, test for trend c | < .001 | < .001 | ||||
| CES-D in 1999 vs 2005 d | ||||||
| <16 in 1999 and 2005 | 652 | 21,486 | 1.00 | Ref | 1.00 | Ref |
| <16 in 1999, 16-24 in 2005 | 115 | 2798 | 1.29 | 1.06–1.56 | 1.31 | 1.08–1.60 |
| <16 in 1999, ≥25 in 2005 | 42 | 1259 | 1.03 | 0.76–1.40 | 1.03 | 0.76–1.41 |
| 16-24 in 1999, <16 in 2005 | 105 | 3721 | 0.89 | 0.73–1.09 | 0.90 | 0.74–1.10 |
| 16-24 in 1999, 16-24 in 2005 | 61 | 1746 | 1.09 | 0.84–1.41 | 1.09 | 0.84–1.42 |
| 16-24 in 1999, ≥25 in 2005 | 36 | 969 | 1.18 | 0.85–1.41 | 1.20 | 0.86–1.67 |
| ≥25 in 1999, <16 in 2005 | 35 | 1082 | 1.01 | 0.72–1.42 | 1.00 | 0.71–1.40 |
| ≥25 in 1999, 16-24 in 2005 | 37 | 1045 | 1.10 | 0.79–1.52 | 1.10 | 0.79–1.53 |
| ≥25 in 1999 and 2005 | 61 | 1274 | 1.46 | 1.13–1.89 | 1.45 | 1.11–1.88 |
| Physician-diagnosed depression | ||||||
| Never | 4154 | 117,794 | 1.00 | Ref | 1.00 | Ref |
| Past | 405 | 9440 | 1.27 | 1.14–1.40 | 1.25 | 1.13–1.39 |
| Current | 163 | 4029 | 1.12 | 0.96–1.31 | 1.15 | 0.98–1.34 |
| Antidepressant use (for any indication) | ||||||
| Never | 4309 | 121,419 | 1.00 | Ref | 1.00 | Ref |
| Past | 201 | 4564 | 1.34 | 1.16–1.54 | 1.32 | 1.14–1.52 |
| Current | 212 | 5279 | 1.13 | 0.98–1.30 | 1.11 | 0.97–1.28 |
| Updated CES-D score and antidepressant use | ||||||
| <16, never use | 3046 | 88,674 | 1.00 | Ref | 1.00 | Ref |
| 16-24, never use | 812 | 21,166 | 1.11 | 1.02–1.19 | 1.12 | 1.0–1.21 |
| ≥25, never use | 451 | 11,577 | 1.13 | 1.03–1.25 | 1.16 | 1.05–1.28 |
| <16, past use | 100 | 2157 | 1.47 | 1.20–1.79 | 1.45 | 1.19–1.77 |
| 16-24, past use | 54 | 1190 | 1.43 | 1.09–1.88 | 1.44 | 1.10–1.89 |
| ≥25, past use | 47 | 1218 | 1.17 | 0.87–1.56 | 1.17 | 0.87–1.56 |
| <16, current use | 91 | 2470 | 1.09 | 0.89–1.35 | 1.07 | 0.87–1.32 |
| 16-24, current use | 43 | 1395 | 0.88 | 0.65–1.19 | 0.88 | 0.65–1.19 |
| ≥25, current use | 78 | 1415 | 1.57 | 1.26–1.97 | 1.56 | 1.26–1.98 |
a Adjusted for age and questionnaire cycle
b Adjusted for age, questionnaire cycle, age at menarche, parity, age at first birth, years since last birth, age at first oral contraceptive use, history of oral contraceptive use, body mass index, physical activity, smoking, current alcohol intake, participant’s education, marital status, occupation, annual household income, region of residence
c P value from Wald test of ordinal variable
d Restricted to period 2005 through 2011
e 1999 CES-D score carried forward for 470 cases and 14,266 person-years who did not complete 2005 CES-D score.
IRRs for current and past physician-diagnosed depression relative to no depression were 1.15 (95% CI, 0.98–1.34) and 1.25 (95% CI, 1.13–1.39), respectively ( Table 2 ). IRRs for the association between current and past use of antidepressants (for any indication) relative to never use were 1.11 (95% CI, 0.97–1.28) and 1.32 (95% CI, 1.14–1.52), respectively. Current users of antidepressants who scored ≥25 on the updated CES-D had 1.56 times the UL risk of never users of antidepressants with a CES-D <16 (95% CI, 1.26–1.98). Further adjustment for current and past use of antidepressants slightly attenuated the association between depressive symptoms and UL: IRRs for baseline CES-D scores 16-24 and ≥25 vs <16 were 1.04 (95% CI, 0.97–1.12) and 1.14 (95% CI, 1.04–1.25), respectively ( P -trend = .003) and IRRs for updated CES-D scores 16-24 and ≥25 vs <16 were 1.10 (95% CI, 1.02–1.19) and 1.16 (95% CI, 1.06–1.28), respectively ( P -trend = .0002).
The association between depressive symptoms and UL was uniform across strata of age, BMI, smoking status, and hormonal contraceptive use ( Table 3 ), and there was no evidence that exogenous hormone use modified the association between antidepressants and UL (data not shown). Because the requirement for UL surgery may reflect more severe disease and longer duration of subclinical disease, we stratified the data by method of confirmation: surgery vs ultrasound. The IRRs for updated CES-D scores of 16-24 and ≥25 vs <16 were 1.16 (95% CI, 0.99–1.37) and 1.15 (95% CI, 0.94–1.42) respectively ( P -trend = .064) among the 929 surgical cases, and 1.10 (95% CI, 1.01–1.20) and 1.21 (95% CI, 1.09–1.33) respectively ( P -trend < .001) among the 3793 ultrasound-confirmed cases. Results were similar using baseline CES-D scores (data not shown).
| Characteristic | Cases | Person-years | Multivariable model a | P value, interaction | |
|---|---|---|---|---|---|
| IRR | 95% CI | ||||
| Age <35 y | |||||
| Updated CES-D score | .405 | ||||
| <16 | 782 | 23,960 | 1.00 | Ref | |
| 16-24 | 271 | 7353 | 1.17 | 1.02–1.35 | |
| ≥25 | 162 | 4669 | 1.11 | 0.94–1.32 | |
| P value, test for trend b | .065 | ||||
| Age ≥35 y | |||||
| Updated CES-D score | |||||
| <16 | 2455 | 69,340 | 1.00 | Ref | |
| 16-24 | 638 | 16,400 | 1.09 | 1.00–1.19 | |
| ≥25 | 414 | 9540 | 1.22 | 1.10–1.36 | |
| P value, test for trend b | < .001 | ||||
| BMI <25 | |||||
| Updated CES-D score | .762 | ||||
| <16 | 1077 | 30,881 | 1.00 | Ref | |
| 16-24 | 229 | 6325 | 1.03 | 0.89–1.19 | |
| ≥25 | 149 | 3621 | 1.18 | 0.99–1.41 | |
| P value, test for trend b | .092 | ||||
| BMI 25-29 | |||||
| Updated CES-D score | |||||
| <16 | 1035 | 28,713 | 1.00 | Ref | |
| 16-24 | 286 | 7153 | 1.10 | 0.96–1.26 | |
| ≥25 | 166 | 3969 | 1.18 | 1.00–1.40 | |
| P value, test for trend b | .027 | ||||
| BMI ≥30 | |||||
| Updated CES-D score | |||||
| <16 | 1125 | 33,706 | 1.00 | Ref | |
| 16-24 | 394 | 10,275 | 1.16 | 1.03–1.30 | |
| ≥25 | 263 | 6620 | 1.20 | 1.05–1.38 | |
| P value, test for trend b | .002 | ||||
| Current smoking | |||||
| Updated CES-D score | .601 | ||||
| <16 | 262 | 9098 | 1.00 | Ref | |
| 16-24 | 116 | 3294 | 1.19 | 0.95–1.49 | |
| ≥25 | 85 | 2640 | 1.12 | 0.87–1.44 | |
| P value, test for trend b | .239 | ||||
| Nonsmoking | |||||
| Updated CES-D score | |||||
| <16 | 2975 | 84,203 | 1.00 | Ref | |
| 16-24 | 793 | 20,459 | 1.10 | 1.01–1.19 | |
| ≥25 | 491 | 11,570 | 1.20 | 1.09–1.32 | |
| P value, test for trend b | < .001 | ||||
| Current hormonal contraceptive use | |||||
| Updated CES-D score | .719 | ||||
| <16 | 2470 | 71,090 | 1.00 | Ref | |
| 16-24 | 714 | 18,326 | 1.12 | 1.03–1.22 | |
| ≥25 | 460 | 11,196 | 1.20 | 1.08–1.32 | |
| P value, test for trend b | < .001 | ||||
| Never or former hormonal contraceptive use | |||||
| Updated CES-D score | |||||
| <16 | 767 | 22,211 | 1.00 | Ref | |
| 16-24 | 195 | 5427 | 1.08 | 0.92–1.26 | |
| ≥25 | 116 | 3014 | 1.16 | 0.94–1.41 | |
| P value, test for trend b | .123 | ||||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
