https://orcid.org/0000-0003-3825-9897
Abstract
Depression co-occurs with diabetes at twice the rate, relative to the general population without diabetes but it is unknown whether depression is longitudinally associated with diabetes control in the general population.
To characterize the longitudinal association between depressive symptoms and joint achievement of glycemic, blood pressure (BP), and cholesterol control (ABC control) among middle-aged and older adults (≥50 years) with diabetes in United States.
Data of the nationally representative Health and Retirement Study 2006-2017 were pooled across study waves conducted every 2 years. Center for Epidemiological Studies Depression (CES-D8) scale was used to assess baseline depressive symptoms (≥3 points). Joint ABC control 4 years later was ascertained using HbA1c (<7.0% [53 mmol/mol] if <65 years, <7.5% [58 mmol/mol] if ≥65 years or <8.0% [64 mmol/mol] with comorbidities), BP (systolic < 140 and diastolic < 90 mm Hg), and non-HDL cholesterol (<130 mg/dL). Survey-weighted modified Poisson regressions were used to study the association (risk ratios [RR]) of depressive symptoms with ABC control.
The study sample consisted of 3 332 observations from 2 531 individuals (mean age: 64.4 years [SD: 8.8], 55.4% women). Depressive symptoms were neither associated with the achievement of joint ABC control (RR: 0.91 [95% CI, 0.76-1.09]) nor achievement of glycemic, BP or cholesterol control after adjusting for covariates. Findings were consistent across various subgroups defined by age, gender, baseline ABC control, medication use, and duration of diabetes.
Baseline depressive symptoms do not compromise future diabetes management. Care models should focus on both conditions independently to potentially improve overall health.
Introduction
Depression co-occurs with diabetes at twice the rate, relative to the general population without diabetes,1 and is associated with lower quality of life and increased morbidity and mortality among people with diabetes.2,3 Diabetes and depression also have a bidirectional relationship with either disease increasing the risk and potentially worsening the management of the other.4–6 Joint glycemic, blood pressure, and cholesterol control, ie, ABC control, is an indicator of diabetes management and is associated with lower incidence of microvascular and macrovascular complications of diabetes.7–9 At the population level among US adults, glycemic and blood pressure (BP) control has decreased while progress in cholesterol control has plateaued between 1999 and 2018.10 There is, therefore, an urgent need to identify and manage risk factors such as depression which may be associated with poor control.
It is currently unknown whether depression is longitudinally associated with diabetes control in the general population. Potential pathways through which depression may influence diabetes control include lower adherence to medication and lifestyle recommendations.11,12 However, evidence from some longitudinal studies suggest that depression, independent of diabetes management, was not associated with long-term glycemic control.13–17 Cross-sectional studies suggest poor ABC control among people with diabetes and depression.18 Furthermore, longitudinal studies of depression and joint ABC control are few and studies such as the Pathways Epidemiologic Study 2001-2007 were based on regional samples recruited through a healthcare system.15
Using nationally representative data from the United States, we characterized whether depressive symptoms were associated with longitudinal achievement of glycemic, BP, and cholesterol control among a sample of middle-aged and older adults. We explored whether these associations differ by subgroups of gender, age, history of ABC control, medication use, and duration of diabetes.
Methods
Study design
The Health and Retirement Study (HRS), 1992-2022 is one of the longest running nationally representative studies of aging worldwide.19 The HRS consists of seven cohorts comprising individuals 50 years and older born before 1924 to those born between 1960 and 1965 living in the community or in nursing homes, and their spouses. Details of sample design, survey methodology and eligibility are available elsewhere.19,20 HRS study waves comprise follow-up visits every 2 years when anthropometry (including BP) and a standardized survey questionnaire, including assessment of depressive symptoms, are assessed for the total sample. Additionally, dried blood spots to measure biomarkers such as HbA1c, total cholesterol, and high-density lipoprotein (HDL) cholesterol are collected every alternate wave for half the community-dwelling sample (Supplementary Table 1).
For this longitudinal analysis of depression and diabetes control, we restricted the sample to those individuals who participated in Wave 8 (2006-2007), Wave 9 (2008-2009), Wave 10 (2010-2011) and Wave 11 (2012-2013) with self-reported diabetes aged 50 years and older (Supplementary Table 2). Participants were included if they completed both the assessment for depressive symptoms and participated in biomarker assessment. We excluded participants living in nursing homes and those among whom depression status was not ascertained.
Outcomes, ie, joint achievement of glycemic, BP, and cholesterol control, were ascertained in a subsequent wave 4 years later, ie, Wave 10 (2010-2011), Wave 11 (2012-2013), Wave 12 (2014-2015), and Wave 13 (2016-2017) (Supplementary Table 1). The approached sample consisted of 5 919 observations from 4 314 individuals. The analytic sample, after excluding those for whom outcome data were unavailable, consisted of 3 332 observations from 2 531 individuals. A flowchart of the analytic sample selection is provided in Figure 1.
Flowchart for analytic sample.
Data collection and variable specification
Exposure: depressive symptoms
Depressive symptoms were assessed using the short eight-item version of the Center for Epidemiological Studies Depression (CES-D8) scale.21 The CES-D8 was previously validated among older adults.22 An individual was considered to exhibit depressive symptoms if they answered “yes” to three or more questions on the scale.23,24
Outcome: glycemic, blood pressure, and cholesterol control
Hemoglobin A1c, total, and HDL cholesterol concentrations were measured using dried blood spots that were analyzed at different laboratories between 2006 and 2017 (Supplementary Table 3). To account for between lab analytical variation and ensure comparability over time, raw analytical measurements were standardized against the population distribution independent of diabetes diagnosis, assuming similar values at each percentile of the nationally representative, non-institutionalized population 50 years and older who participated in the corresponding round of the National Health and Nutrition Examination Survey (Supplementary Table 1). This implies that the average differences in HRS samples reflect secular differences in the national population over time. Non-HDL cholesterol concentration was estimated as the difference between total and HDL cholesterol. Blood pressures were measured thrice for each participant. We averaged the last two measurements to estimate the systolic (SBP) and diastolic blood pressure (DBP).
For adults with diagnosed diabetes, glycemic control was defined as HbA1c < 7.0% (53 mmol/mol) for individuals under 65 years, and <7.5% (58 mmol/mol) for individuals 65 years or older.25 We additionally incorporated current recommendations suggesting individualized targets for glycemic control among those 65 years and older, especially using liberal targets (<8.0% [64 mmol/mol]) among those with three or more comorbidities (eg, hypertension, arthritis, cancer, depression), functional dependence (in two or more instrumental activities of daily living) or cognitive impairment.25 Blood pressure control was defined as SBP < 140 mm Hg and DBP < 90 mm Hg. Cholesterol control was defined as non-HDL cholesterol <130 mg/dL. ABC control was defined as joint achievement of glycemic, BP, and cholesterol control.
Individual and household covariates
Individual socio-demographic and health characteristics associated with diabetes control were ascertained separately for each wave when depressive symptoms were measured. Individual socio-demographic characteristics include age, gender (men or women), race-ethnicity (Hispanic, non-Hispanic White, non-Hispanic Black, non-Hispanic Other), education (less than lower secondary, upper secondary or vocational training, or tertiary), and wage employment status (no wages or salaries, wage work, or retired). Health characteristics associated with depressive symptoms and diabetes control include instrumental activities of daily living (IADL) score (experiencing difficulty performing at least one of four items: managing money, taking medications, shopping for groceries, and preparing meals), body mass index (BMI; calculated as weight in kilograms divided by height in meters squared), duration of diabetes, medication use (diabetes, BP, cholesterol, depression, psychiatric disorders), moderate or vigorous physical activity, smoking, frequent alcohol consumption and comorbid hypertension. Height and weight were measured using standardized instruments and validated protocols (Supplementary Table 3). Engagement in moderate and vigorous physical activity more than once a week was self-reported by participants. Frequent alcohol consumption was defined as 5 or more days per week for men and 3 or more days per week for women when alcohol was consumed.
Household characteristics potentially associated with diabetes control included the number of household residents and household wealth. Household wealth was categorized as quintiles within each wave to indicate relative position within the overall cohort. Additional information on individual and household covariates are documented by the RAND Corporation and the Gateway to Aging Study.26,27
Statistical analyses
The objective was to characterize the association of depressive symptoms with ABC control at 4 years. First, we described individual and household characteristics using weighted means (standard deviation) and weighted frequencies. Second, we fit modified Poisson regressions to estimate risk ratios as the measure of association, with robust standard errors to account for clustering of observations within individuals.28 We fit an unadjusted model, a partially adjusted model for covariates unlikely to be mediators (gender, race-ethnicity, diabetes duration, household size, and household wealth with baseline ABC control at time of exposure assessment and fixed effects for study waves to account for unmeasured time-varying characteristics), and a model fully adjusted for all individual and household covariates associated.29 To account for differences in survey weights and pooling across study waves, we normalized biomarker survey weights across study waves. We used multiple imputation with predictive mean matching (10 datasets, 50 iterations) for missing covariate data (Supplementary Table 4) and used Rubin’s rules to pool estimates.30
We characterized effect modification of the longitudinal association of depressive symptoms and ABC control by age (≥65 vs <65 years), gender, any functional limitations (based on IADL), baseline ABC control, duration of diabetes, baseline medication use for diabetes, and depression.31–33 We did not correct for multiple hypothesis testing in this exploratory, pre-specified analysis of effect modification.34 We also estimated the association of depressive symptoms with separate achievement of glycemic, BP, and cholesterol control. Analyses were carried out with R 4.3.2. As a sensitivity analysis, among those without biomarker outcome data (Figure 1), we repeated the analysis with inverse probability weights to account for loss to follow-up due to non-response or death before outcome assessment. We also studied the relative risks for ABC control after using CES-D8 as a continuous exposure variable.
Results
The descriptive characteristics of the analytic sample of 2 531 individuals (3 332 observations) with diagnosed diabetes are reported in Table 1. The analytic sample was on average aged 64.4 years (SD: 8.8), 55.4% women, 62.5% married, 69.8% non-Hispanic White, 14.1% non-Hispanic Black, and 11.5% Hispanic. Most of the analytic sample (79.7% [95% CI, 77.3-81.8]) were taking medication to manage their diabetes, with 18.1% [95% CI, 16.0-20.3]) currently using insulin. Of the analytic sample, 27.0% (95% CI, 24.8-29.3) self-reported depressive symptoms at the first study wave when they were included in the analytic sample. The proportion with depressive symptoms was higher among women (32.6% [95% CI, 29.5-35.8]), compared to men (21.3% [95% CI, 18.2-24.8]). Relative to the analytic sample, the sample of 2 496 individuals excluded due to loss to follow-up or death were older (68.7 years), less likely to be married (54.9%), and similar in racial–ethnic composition (Supplementary Table 5). Prevalence of depressive symptoms were higher in the excluded sample among women (35.2% [95% CI, 32.2-38.4]) and men (29.9% [95% CI, 26.8-33.3]).
| Descriptive characteristics of middle-aged and older adults in Health and Retirement Study, 2006-2017.
| Total | Women | Men | |
|---|---|---|---|
| Unique individuals | 2 531 | 1 402 | 1 129 |
| Socio-demographic | |||
| Age (years) | 64.4 (63.8, 65) | 64.7 (64, 65.4) | 64.2 (63.5, 64.9) |
| 50 to 64 (%) | 57.4 (54.1, 60.6) | 56.3 (52.5, 60.1) | 58.4 (54.3, 62.4) |
| 65 to 74 (%) | 27.3 (25, 29.8) | 27.3 (24.1, 30.7) | 27.4 (24.8, 30.1) |
| 75 to 84 (%) | 13 (11.3, 14.9) | 13.6 (11.5, 16) | 12.4 (10.4, 14.7) |
| 85 and over (%) | 2.3 (1.5, 3.5) | 2.8 (1.8, 4.4) | 1.8 (1.1, 3) |
| Education level | |||
| Less than upper secondary (%) | 20.0 (17.4, 22.9) | 24.3 (21.3, 27.5) | 15.7 (12.7, 19.3) |
| Upper secondary and vocational training (%) | 58.8 (55.8, 61.8) | 60.3 (56.6, 64) | 57.3 (53.2, 61.3) |
| Tertiary (%) | 21.1 (18.6, 23.9) | 15.4 (12.5, 18.9) | 27 (23.3, 31) |
| Employment status | |||
| No wage employment (%) | 11.9 (10.2, 13.7) | 18.3 (15.7, 21.3) | 5.2 (3.5, 7.6) |
| Formal wage employment (%) | 35.1 (32.7, 37.7) | 29.6 (26.6, 32.8) | 40.9 (37, 44.8) |
| Retired (%) | 53.0 (50.2, 55.8) | 52.1 (48.8, 55.4) | 53.9 (49.6, 58.2) |
| Married (%) | 62.5 (60.2, 64.7) | 51.5 (48.2, 54.7) | 73.8 (70.2, 77.2) |
| Race-Ethnicity | |||
| Hispanic | 11.5 (8.6, 15.2) | 13 (9.9, 16.7) | 10 (6.7, 14.7) |
| Non-Hispanic White | 69.8 (65.9, 73.4) | 65.5 (62, 68.9) | 74.2 (68.9, 78.8) |
| Non-Hispanic Black | 14.1 (12.4, 16.1) | 16.5 (14.4, 18.9) | 11.7 (9.5, 14.3) |
| Non-Hispanic Other | 4.6 (3.3, 6.3) | 5 (3.6, 6.9) | 4.2 (2.7, 6.4) |
| Household | |||
| Length of current marriage (years) | 33.9 (32.7, 35.1) | 35 (33.6, 36.5) | 33 (31.7, 34.4) |
| Number of people in household (count) | 2.4 (2.3, 2.5) | 2.4 (2.3, 2.4) | 2.4 (2.3, 2.5) |
| Number of children (count) | 3.2 (3.0, 3.3) | 3.2 (3, 3.3) | 3.2 (3, 3.3) |
| Household wealth | |||
| Lowest (%) | 27.2 (24.7, 29.9) | 33.4 (30.1, 36.9) | 21.8 (18.8, 25.1) |
| Low (%) | 27.2 (23.9, 30.7) | 26.8 (23.3, 30.6) | 27.5 (23.6, 31.8) |
| Medium (%) | 16.5 (14.6, 18.6) | 15.2 (12.4, 18.5) | 17.6 (15.2, 20.4) |
| High (%) | 16.4 (14.4, 18.6) | 14.7 (12.3, 17.5) | 17.9 (15.2, 21) |
| Highest (%) | 12.7 (10.8, 15) | 10 (7.9, 12.4) | 15.2 (12.1, 18.8) |
| Health behaviors and self-reported health outcomes | |||
| Smoking | |||
| Never (%) | 43.4 (41.3, 45.5) | 52.6 (49.2, 55.9) | 34 (31.7, 36.5) |
| Former (%) | 43.5 (41.1, 46) | 34.6 (31.3, 38.1) | 52.6 (49.8, 55.4) |
| Current (%) | 13.1 (11.3, 15.1) | 12.8 (10.5, 15.6) | 13.4 (10.9, 16.2) |
| Frequent alcohol consumption (%) | 6.7 (5.5, 8.0) | 5.0 (3.6, 6.7) | 8.4 (6.6, 10.7) |
| Moderate physical activity (days/week) | 2.2 (2.1, 2.3) | 2.0 (1.9, 2.1) | 2.3 (2.2, 2.5) |
| Vigorous physical activity (days/week) | 0.8 (0.8, 0.9) | 0.6 (0.5, 0.7) | 1.0 (0.9, 1.2) |
| Diagnosis of Alzheimer’s Disease or Dementia (%) | 0.7 (0.3, 1.2) | 0.3 (0.1, 0.6) | 1.1 (0.5, 2.3) |
| Hypertension (%) | 80 (77.9, 81.9) | 80.8 (77.5, 83.7) | 79.2 (75.7, 82.3) |
| Medication for high blood pressure (%) | 71.5 (69, 73.9) | 74.5 (71.3, 77.5) | 68.4 (64.5, 72.1) |
| Medication for diabetes (%) | 79.7 (77.3, 81.8) | 79.6 (76.5, 82.4) | 79.7 (76.5, 82.6) |
| Insulin use for diabetes (%) | 18.1 (16, 20.3) | 18.6 (16.2, 21.3) | 17.6 (14.6, 21) |
| Oral medication use for diabetes (%) | 73.1 (70.8, 75.2) | 72.7 (69.9, 75.3) | 73.5 (70.1, 76.6) |
| Difficulty in instrumental activities of daily living (any of 4; proportion) | 0.2 (0.2, 0.3) | 0.3 (0.2, 0.3) | 0.2 (0.1, 0.3) |
| Vision impairment (%) | 11.2 (9.5, 13.2) | 13.2 (10.7, 16.2) | 9.1 (7.3, 11.3) |
| Balance impairment (%)a | 19.3 (17.4, 21.2) | 18.3 (16.2, 20.7) | 20.1 (17.7, 22.7) |
| Cognitive impairment (%)b | 25.1 (23.2, 27.2) | 26 (23.3, 28.9) | 24.2 (21.7, 26.9) |
| Number of conditions for which medications are used c | 3.4 (3.3, 3.4) | 3.5 (3.3, 3.6) | 3.2 (3.1, 3.3) |
| Clinical measures | |||
| BMI (kg/m2) | 32.4 (32.1, 32.6) | 33 (32.5, 33.6) | 31.6 (31.3, 32) |
| Waist circumference (cm) | 109.9 (109.1, 110.7) | 107.4 (106.3, 108.5) | 112.4 (111.2, 113.7) |
| Systolic blood pressure (mm Hg) | 131.3 (130.3, 132.3) | 129.8 (128.5, 131.2) | 132.8 (131.3, 134.2) |
| Diastolic blood pressure (mm Hg) | 79.1 (78.5, 79.6) | 78.9 (78.1, 79.7) | 79.2 (78.4, 80.1) |
| HbA1c (%) | 6.9 (6.8, 6.9) | 6.8 (6.7, 6.9) | 7 (6.8, 7.1) |
| Total cholesterol (mg/dL) | 188.5 (186.4, 190.7) | 192.5 (189.5, 195.6) | 184.4 (181.3, 187.5) |
| HDL cholesterol (mg/dL) | 49.8 (48.9, 50.7) | 52.8 (51.7, 54) | 46.6 (45.3, 47.8) |
| Non-HDL cholesterol (mg/dL) | 140.1 (138.1, 142.1) | 140.5 (137.4, 143.6) | 139.7 (136.7, 142.7) |
| Baseline ABC controld | 24.9 (22.5, 27.5) | 26.2 (22.8, 29.8) | 23.7 (20.1, 27.7) |
| Glycemic controle | 75.1 (72.8, 77.3) | 76.6 (73.5, 79.5) | 73.5 (69.5, 77.2) |
| Blood pressure control (SBP < 140 and DBP < 90 mmHg) | 67.8 (65.6, 70) | 70.6 (67.3, 73.8) | 65 (61.7, 68.2) |
| Cholesterol control (non-HDL cholesterol < 130 mg/dL) | 44.6 (41.9, 47.3) | 45.4 (41.4, 49.5) | 43.7 (39.8, 47.6) |
| Mental health | |||
| Center for Epidemiologic Studies—Depression 8 score ≥ 3 | 27.0 (24.8, 29.3) | 32.6 (29.5, 35.8) | 21.3 (18.2, 24.8) |
| Antidepressants | 22.4 (20.1, 24.9) | 28.0 (24.7, 31.5) | 16.6 (14, 19.6) |
| Antipsychotics | 17.3 (15.4, 19.5) | 22.1 (19.4, 25) | 12.4 (10.3, 14.9) |
| Follow-up observations (count) | 3332 | 1866 | 1466 |
| Number of observations achieving ABC control in prospective wave | 918 | 495 | 423 |
| Total | Women | Men | |
|---|---|---|---|
| Unique individuals | 2 531 | 1 402 | 1 129 |
| Socio-demographic | |||
| Age (years) | 64.4 (63.8, 65) | 64.7 (64, 65.4) | 64.2 (63.5, 64.9) |
| 50 to 64 (%) | 57.4 (54.1, 60.6) | 56.3 (52.5, 60.1) | 58.4 (54.3, 62.4) |
| 65 to 74 (%) | 27.3 (25, 29.8) | 27.3 (24.1, 30.7) | 27.4 (24.8, 30.1) |
| 75 to 84 (%) | 13 (11.3, 14.9) | 13.6 (11.5, 16) | 12.4 (10.4, 14.7) |
| 85 and over (%) | 2.3 (1.5, 3.5) | 2.8 (1.8, 4.4) | 1.8 (1.1, 3) |
| Education level | |||
| Less than upper secondary (%) | 20.0 (17.4, 22.9) | 24.3 (21.3, 27.5) | 15.7 (12.7, 19.3) |
| Upper secondary and vocational training (%) | 58.8 (55.8, 61.8) | 60.3 (56.6, 64) | 57.3 (53.2, 61.3) |
| Tertiary (%) | 21.1 (18.6, 23.9) | 15.4 (12.5, 18.9) | 27 (23.3, 31) |
| Employment status | |||
| No wage employment (%) | 11.9 (10.2, 13.7) | 18.3 (15.7, 21.3) | 5.2 (3.5, 7.6) |
| Formal wage employment (%) | 35.1 (32.7, 37.7) | 29.6 (26.6, 32.8) | 40.9 (37, 44.8) |
| Retired (%) | 53.0 (50.2, 55.8) | 52.1 (48.8, 55.4) | 53.9 (49.6, 58.2) |
| Married (%) | 62.5 (60.2, 64.7) | 51.5 (48.2, 54.7) | 73.8 (70.2, 77.2) |
| Race-Ethnicity | |||
| Hispanic | 11.5 (8.6, 15.2) | 13 (9.9, 16.7) | 10 (6.7, 14.7) |
| Non-Hispanic White | 69.8 (65.9, 73.4) | 65.5 (62, 68.9) | 74.2 (68.9, 78.8) |
| Non-Hispanic Black | 14.1 (12.4, 16.1) | 16.5 (14.4, 18.9) | 11.7 (9.5, 14.3) |
| Non-Hispanic Other | 4.6 (3.3, 6.3) | 5 (3.6, 6.9) | 4.2 (2.7, 6.4) |
| Household | |||
| Length of current marriage (years) | 33.9 (32.7, 35.1) | 35 (33.6, 36.5) | 33 (31.7, 34.4) |
| Number of people in household (count) | 2.4 (2.3, 2.5) | 2.4 (2.3, 2.4) | 2.4 (2.3, 2.5) |
| Number of children (count) | 3.2 (3.0, 3.3) | 3.2 (3, 3.3) | 3.2 (3, 3.3) |
| Household wealth | |||
| Lowest (%) | 27.2 (24.7, 29.9) | 33.4 (30.1, 36.9) | 21.8 (18.8, 25.1) |
| Low (%) | 27.2 (23.9, 30.7) | 26.8 (23.3, 30.6) | 27.5 (23.6, 31.8) |
| Medium (%) | 16.5 (14.6, 18.6) | 15.2 (12.4, 18.5) | 17.6 (15.2, 20.4) |
| High (%) | 16.4 (14.4, 18.6) | 14.7 (12.3, 17.5) | 17.9 (15.2, 21) |
| Highest (%) | 12.7 (10.8, 15) | 10 (7.9, 12.4) | 15.2 (12.1, 18.8) |
| Health behaviors and self-reported health outcomes | |||
| Smoking | |||
| Never (%) | 43.4 (41.3, 45.5) | 52.6 (49.2, 55.9) | 34 (31.7, 36.5) |
| Former (%) | 43.5 (41.1, 46) | 34.6 (31.3, 38.1) | 52.6 (49.8, 55.4) |
| Current (%) | 13.1 (11.3, 15.1) | 12.8 (10.5, 15.6) | 13.4 (10.9, 16.2) |
| Frequent alcohol consumption (%) | 6.7 (5.5, 8.0) | 5.0 (3.6, 6.7) | 8.4 (6.6, 10.7) |
| Moderate physical activity (days/week) | 2.2 (2.1, 2.3) | 2.0 (1.9, 2.1) | 2.3 (2.2, 2.5) |
| Vigorous physical activity (days/week) | 0.8 (0.8, 0.9) | 0.6 (0.5, 0.7) | 1.0 (0.9, 1.2) |
| Diagnosis of Alzheimer’s Disease or Dementia (%) | 0.7 (0.3, 1.2) | 0.3 (0.1, 0.6) | 1.1 (0.5, 2.3) |
| Hypertension (%) | 80 (77.9, 81.9) | 80.8 (77.5, 83.7) | 79.2 (75.7, 82.3) |
| Medication for high blood pressure (%) | 71.5 (69, 73.9) | 74.5 (71.3, 77.5) | 68.4 (64.5, 72.1) |
| Medication for diabetes (%) | 79.7 (77.3, 81.8) | 79.6 (76.5, 82.4) | 79.7 (76.5, 82.6) |
| Insulin use for diabetes (%) | 18.1 (16, 20.3) | 18.6 (16.2, 21.3) | 17.6 (14.6, 21) |
| Oral medication use for diabetes (%) | 73.1 (70.8, 75.2) | 72.7 (69.9, 75.3) | 73.5 (70.1, 76.6) |
| Difficulty in instrumental activities of daily living (any of 4; proportion) | 0.2 (0.2, 0.3) | 0.3 (0.2, 0.3) | 0.2 (0.1, 0.3) |
| Vision impairment (%) | 11.2 (9.5, 13.2) | 13.2 (10.7, 16.2) | 9.1 (7.3, 11.3) |
| Balance impairment (%)a | 19.3 (17.4, 21.2) | 18.3 (16.2, 20.7) | 20.1 (17.7, 22.7) |
| Cognitive impairment (%)b | 25.1 (23.2, 27.2) | 26 (23.3, 28.9) | 24.2 (21.7, 26.9) |
| Number of conditions for which medications are used c | 3.4 (3.3, 3.4) | 3.5 (3.3, 3.6) | 3.2 (3.1, 3.3) |
| Clinical measures | |||
| BMI (kg/m2) | 32.4 (32.1, 32.6) | 33 (32.5, 33.6) | 31.6 (31.3, 32) |
| Waist circumference (cm) | 109.9 (109.1, 110.7) | 107.4 (106.3, 108.5) | 112.4 (111.2, 113.7) |
| Systolic blood pressure (mm Hg) | 131.3 (130.3, 132.3) | 129.8 (128.5, 131.2) | 132.8 (131.3, 134.2) |
| Diastolic blood pressure (mm Hg) | 79.1 (78.5, 79.6) | 78.9 (78.1, 79.7) | 79.2 (78.4, 80.1) |
| HbA1c (%) | 6.9 (6.8, 6.9) | 6.8 (6.7, 6.9) | 7 (6.8, 7.1) |
| Total cholesterol (mg/dL) | 188.5 (186.4, 190.7) | 192.5 (189.5, 195.6) | 184.4 (181.3, 187.5) |
| HDL cholesterol (mg/dL) | 49.8 (48.9, 50.7) | 52.8 (51.7, 54) | 46.6 (45.3, 47.8) |
| Non-HDL cholesterol (mg/dL) | 140.1 (138.1, 142.1) | 140.5 (137.4, 143.6) | 139.7 (136.7, 142.7) |
| Baseline ABC controld | 24.9 (22.5, 27.5) | 26.2 (22.8, 29.8) | 23.7 (20.1, 27.7) |
| Glycemic controle | 75.1 (72.8, 77.3) | 76.6 (73.5, 79.5) | 73.5 (69.5, 77.2) |
| Blood pressure control (SBP < 140 and DBP < 90 mmHg) | 67.8 (65.6, 70) | 70.6 (67.3, 73.8) | 65 (61.7, 68.2) |
| Cholesterol control (non-HDL cholesterol < 130 mg/dL) | 44.6 (41.9, 47.3) | 45.4 (41.4, 49.5) | 43.7 (39.8, 47.6) |
| Mental health | |||
| Center for Epidemiologic Studies—Depression 8 score ≥ 3 | 27.0 (24.8, 29.3) | 32.6 (29.5, 35.8) | 21.3 (18.2, 24.8) |
| Antidepressants | 22.4 (20.1, 24.9) | 28.0 (24.7, 31.5) | 16.6 (14, 19.6) |
| Antipsychotics | 17.3 (15.4, 19.5) | 22.1 (19.4, 25) | 12.4 (10.3, 14.9) |
| Follow-up observations (count) | 3332 | 1866 | 1466 |
| Number of observations achieving ABC control in prospective wave | 918 | 495 | 423 |
Values are survey weighted means or proportions, and 95% confidence intervals for first assessment of Center for Epidemiologic Studies—Depression 8 at the wave with a biomarker measurement (2006-2007 onwards for HRS). We first normalized sample weights for an observation by dividing it by the sum of sample weights for the biomarker sample for each wave. Next, we inverse-weighted the normalized sample weight with proportion of participants in the pooled serial cross-sectional dataset corresponding to that wave to account for wave-specific sample sizes.
Difficulty in instrumental activities of daily living consisted of whether the participant had any difficulty with one of: managing money, taking medications, shopping for groceries, and preparing meals.
aThe semi-tandem test was performed first, followed by either the side-by-side if they failed the former or full tandem if they passed it. Balance impairment was failure to complete any one of the three tests.
bCognitive impairment was defined as successful completion of two or fewer tests only among those who completed all four of immediate word recall, delayed word recall, counting backwards and serial 7 subtraction tests.
cOut of 14 conditions for which medications are used: blood pressure, insulin use, oral hypoglycemic use, stroke, angina, congestive heart failure, cancer, heart attack, other heart problems, high cholesterol, breathing difficulty, other lung disease, stomach disease, depression.
dGlycemic (HbA1c), blood pressure, and cholesterol.
eHbA1c < 7.0% [53 mmol/mol] if < 65 years, < 7.5% [58 mmol/mol] if ≥ 65 years or < 8.0% [64 mmol/mol] with comorbidities.
| Descriptive characteristics of middle-aged and older adults in Health and Retirement Study, 2006-2017.
| Total | Women | Men | |
|---|---|---|---|
| Unique individuals | 2 531 | 1 402 | 1 129 |
| Socio-demographic | |||
| Age (years) | 64.4 (63.8, 65) | 64.7 (64, 65.4) | 64.2 (63.5, 64.9) |
| 50 to 64 (%) | 57.4 (54.1, 60.6) | 56.3 (52.5, 60.1) | 58.4 (54.3, 62.4) |
| 65 to 74 (%) | 27.3 (25, 29.8) | 27.3 (24.1, 30.7) | 27.4 (24.8, 30.1) |
| 75 to 84 (%) | 13 (11.3, 14.9) | 13.6 (11.5, 16) | 12.4 (10.4, 14.7) |
| 85 and over (%) | 2.3 (1.5, 3.5) | 2.8 (1.8, 4.4) | 1.8 (1.1, 3) |
| Education level | |||
| Less than upper secondary (%) | 20.0 (17.4, 22.9) | 24.3 (21.3, 27.5) | 15.7 (12.7, 19.3) |
| Upper secondary and vocational training (%) | 58.8 (55.8, 61.8) | 60.3 (56.6, 64) | 57.3 (53.2, 61.3) |
| Tertiary (%) | 21.1 (18.6, 23.9) | 15.4 (12.5, 18.9) | 27 (23.3, 31) |
| Employment status | |||
| No wage employment (%) | 11.9 (10.2, 13.7) | 18.3 (15.7, 21.3) | 5.2 (3.5, 7.6) |
| Formal wage employment (%) | 35.1 (32.7, 37.7) | 29.6 (26.6, 32.8) | 40.9 (37, 44.8) |
| Retired (%) | 53.0 (50.2, 55.8) | 52.1 (48.8, 55.4) | 53.9 (49.6, 58.2) |
| Married (%) | 62.5 (60.2, 64.7) | 51.5 (48.2, 54.7) | 73.8 (70.2, 77.2) |
| Race-Ethnicity | |||
| Hispanic | 11.5 (8.6, 15.2) | 13 (9.9, 16.7) | 10 (6.7, 14.7) |
| Non-Hispanic White | 69.8 (65.9, 73.4) | 65.5 (62, 68.9) | 74.2 (68.9, 78.8) |
| Non-Hispanic Black | 14.1 (12.4, 16.1) | 16.5 (14.4, 18.9) | 11.7 (9.5, 14.3) |
| Non-Hispanic Other | 4.6 (3.3, 6.3) | 5 (3.6, 6.9) | 4.2 (2.7, 6.4) |
| Household | |||
| Length of current marriage (years) | 33.9 (32.7, 35.1) | 35 (33.6, 36.5) | 33 (31.7, 34.4) |
| Number of people in household (count) | 2.4 (2.3, 2.5) | 2.4 (2.3, 2.4) | 2.4 (2.3, 2.5) |
| Number of children (count) | 3.2 (3.0, 3.3) | 3.2 (3, 3.3) | 3.2 (3, 3.3) |
| Household wealth | |||
| Lowest (%) | 27.2 (24.7, 29.9) | 33.4 (30.1, 36.9) | 21.8 (18.8, 25.1) |
| Low (%) | 27.2 (23.9, 30.7) | 26.8 (23.3, 30.6) | 27.5 (23.6, 31.8) |
| Medium (%) | 16.5 (14.6, 18.6) | 15.2 (12.4, 18.5) | 17.6 (15.2, 20.4) |
| High (%) | 16.4 (14.4, 18.6) | 14.7 (12.3, 17.5) | 17.9 (15.2, 21) |
| Highest (%) | 12.7 (10.8, 15) | 10 (7.9, 12.4) | 15.2 (12.1, 18.8) |
| Health behaviors and self-reported health outcomes | |||
| Smoking | |||
| Never (%) | 43.4 (41.3, 45.5) | 52.6 (49.2, 55.9) | 34 (31.7, 36.5) |
| Former (%) | 43.5 (41.1, 46) | 34.6 (31.3, 38.1) | 52.6 (49.8, 55.4) |
| Current (%) | 13.1 (11.3, 15.1) | 12.8 (10.5, 15.6) | 13.4 (10.9, 16.2) |
| Frequent alcohol consumption (%) | 6.7 (5.5, 8.0) | 5.0 (3.6, 6.7) | 8.4 (6.6, 10.7) |
| Moderate physical activity (days/week) | 2.2 (2.1, 2.3) | 2.0 (1.9, 2.1) | 2.3 (2.2, 2.5) |
| Vigorous physical activity (days/week) | 0.8 (0.8, 0.9) | 0.6 (0.5, 0.7) | 1.0 (0.9, 1.2) |
| Diagnosis of Alzheimer’s Disease or Dementia (%) | 0.7 (0.3, 1.2) | 0.3 (0.1, 0.6) | 1.1 (0.5, 2.3) |
| Hypertension (%) | 80 (77.9, 81.9) | 80.8 (77.5, 83.7) | 79.2 (75.7, 82.3) |
| Medication for high blood pressure (%) | 71.5 (69, 73.9) | 74.5 (71.3, 77.5) | 68.4 (64.5, 72.1) |
| Medication for diabetes (%) | 79.7 (77.3, 81.8) | 79.6 (76.5, 82.4) | 79.7 (76.5, 82.6) |
| Insulin use for diabetes (%) | 18.1 (16, 20.3) | 18.6 (16.2, 21.3) | 17.6 (14.6, 21) |
| Oral medication use for diabetes (%) | 73.1 (70.8, 75.2) | 72.7 (69.9, 75.3) | 73.5 (70.1, 76.6) |
| Difficulty in instrumental activities of daily living (any of 4; proportion) | 0.2 (0.2, 0.3) | 0.3 (0.2, 0.3) | 0.2 (0.1, 0.3) |
| Vision impairment (%) | 11.2 (9.5, 13.2) | 13.2 (10.7, 16.2) | 9.1 (7.3, 11.3) |
| Balance impairment (%)a | 19.3 (17.4, 21.2) | 18.3 (16.2, 20.7) | 20.1 (17.7, 22.7) |
| Cognitive impairment (%)b | 25.1 (23.2, 27.2) | 26 (23.3, 28.9) | 24.2 (21.7, 26.9) |
| Number of conditions for which medications are used c | 3.4 (3.3, 3.4) | 3.5 (3.3, 3.6) | 3.2 (3.1, 3.3) |
| Clinical measures | |||
| BMI (kg/m2) | 32.4 (32.1, 32.6) | 33 (32.5, 33.6) | 31.6 (31.3, 32) |
| Waist circumference (cm) | 109.9 (109.1, 110.7) | 107.4 (106.3, 108.5) | 112.4 (111.2, 113.7) |
| Systolic blood pressure (mm Hg) | 131.3 (130.3, 132.3) | 129.8 (128.5, 131.2) | 132.8 (131.3, 134.2) |
| Diastolic blood pressure (mm Hg) | 79.1 (78.5, 79.6) | 78.9 (78.1, 79.7) | 79.2 (78.4, 80.1) |
| HbA1c (%) | 6.9 (6.8, 6.9) | 6.8 (6.7, 6.9) | 7 (6.8, 7.1) |
| Total cholesterol (mg/dL) | 188.5 (186.4, 190.7) | 192.5 (189.5, 195.6) | 184.4 (181.3, 187.5) |
| HDL cholesterol (mg/dL) | 49.8 (48.9, 50.7) | 52.8 (51.7, 54) | 46.6 (45.3, 47.8) |
| Non-HDL cholesterol (mg/dL) | 140.1 (138.1, 142.1) | 140.5 (137.4, 143.6) | 139.7 (136.7, 142.7) |
| Baseline ABC controld | 24.9 (22.5, 27.5) | 26.2 (22.8, 29.8) | 23.7 (20.1, 27.7) |
| Glycemic controle | 75.1 (72.8, 77.3) | 76.6 (73.5, 79.5) | 73.5 (69.5, 77.2) |
| Blood pressure control (SBP < 140 and DBP < 90 mmHg) | 67.8 (65.6, 70) | 70.6 (67.3, 73.8) | 65 (61.7, 68.2) |
| Cholesterol control (non-HDL cholesterol < 130 mg/dL) | 44.6 (41.9, 47.3) | 45.4 (41.4, 49.5) | 43.7 (39.8, 47.6) |
| Mental health | |||
| Center for Epidemiologic Studies—Depression 8 score ≥ 3 | 27.0 (24.8, 29.3) | 32.6 (29.5, 35.8) | 21.3 (18.2, 24.8) |
| Antidepressants | 22.4 (20.1, 24.9) | 28.0 (24.7, 31.5) | 16.6 (14, 19.6) |
| Antipsychotics | 17.3 (15.4, 19.5) | 22.1 (19.4, 25) | 12.4 (10.3, 14.9) |
| Follow-up observations (count) | 3332 | 1866 | 1466 |
| Number of observations achieving ABC control in prospective wave | 918 | 495 | 423 |
| Total | Women | Men | |
|---|---|---|---|
| Unique individuals | 2 531 | 1 402 | 1 129 |
| Socio-demographic | |||
| Age (years) | 64.4 (63.8, 65) | 64.7 (64, 65.4) | 64.2 (63.5, 64.9) |
| 50 to 64 (%) | 57.4 (54.1, 60.6) | 56.3 (52.5, 60.1) | 58.4 (54.3, 62.4) |
| 65 to 74 (%) | 27.3 (25, 29.8) | 27.3 (24.1, 30.7) | 27.4 (24.8, 30.1) |
| 75 to 84 (%) | 13 (11.3, 14.9) | 13.6 (11.5, 16) | 12.4 (10.4, 14.7) |
| 85 and over (%) | 2.3 (1.5, 3.5) | 2.8 (1.8, 4.4) | 1.8 (1.1, 3) |
| Education level | |||
| Less than upper secondary (%) | 20.0 (17.4, 22.9) | 24.3 (21.3, 27.5) | 15.7 (12.7, 19.3) |
| Upper secondary and vocational training (%) | 58.8 (55.8, 61.8) | 60.3 (56.6, 64) | 57.3 (53.2, 61.3) |
| Tertiary (%) | 21.1 (18.6, 23.9) | 15.4 (12.5, 18.9) | 27 (23.3, 31) |
| Employment status | |||
| No wage employment (%) | 11.9 (10.2, 13.7) | 18.3 (15.7, 21.3) | 5.2 (3.5, 7.6) |
| Formal wage employment (%) | 35.1 (32.7, 37.7) | 29.6 (26.6, 32.8) | 40.9 (37, 44.8) |
| Retired (%) | 53.0 (50.2, 55.8) | 52.1 (48.8, 55.4) | 53.9 (49.6, 58.2) |
| Married (%) | 62.5 (60.2, 64.7) | 51.5 (48.2, 54.7) | 73.8 (70.2, 77.2) |
| Race-Ethnicity | |||
| Hispanic | 11.5 (8.6, 15.2) | 13 (9.9, 16.7) | 10 (6.7, 14.7) |
| Non-Hispanic White | 69.8 (65.9, 73.4) | 65.5 (62, 68.9) | 74.2 (68.9, 78.8) |
| Non-Hispanic Black | 14.1 (12.4, 16.1) | 16.5 (14.4, 18.9) | 11.7 (9.5, 14.3) |
| Non-Hispanic Other | 4.6 (3.3, 6.3) | 5 (3.6, 6.9) | 4.2 (2.7, 6.4) |
| Household | |||
| Length of current marriage (years) | 33.9 (32.7, 35.1) | 35 (33.6, 36.5) | 33 (31.7, 34.4) |
| Number of people in household (count) | 2.4 (2.3, 2.5) | 2.4 (2.3, 2.4) | 2.4 (2.3, 2.5) |
| Number of children (count) | 3.2 (3.0, 3.3) | 3.2 (3, 3.3) | 3.2 (3, 3.3) |
| Household wealth | |||
| Lowest (%) | 27.2 (24.7, 29.9) | 33.4 (30.1, 36.9) | 21.8 (18.8, 25.1) |
| Low (%) | 27.2 (23.9, 30.7) | 26.8 (23.3, 30.6) | 27.5 (23.6, 31.8) |
| Medium (%) | 16.5 (14.6, 18.6) | 15.2 (12.4, 18.5) | 17.6 (15.2, 20.4) |
| High (%) | 16.4 (14.4, 18.6) | 14.7 (12.3, 17.5) | 17.9 (15.2, 21) |
| Highest (%) | 12.7 (10.8, 15) | 10 (7.9, 12.4) | 15.2 (12.1, 18.8) |
| Health behaviors and self-reported health outcomes | |||
| Smoking | |||
| Never (%) | 43.4 (41.3, 45.5) | 52.6 (49.2, 55.9) | 34 (31.7, 36.5) |
| Former (%) | 43.5 (41.1, 46) | 34.6 (31.3, 38.1) | 52.6 (49.8, 55.4) |
| Current (%) | 13.1 (11.3, 15.1) | 12.8 (10.5, 15.6) | 13.4 (10.9, 16.2) |
| Frequent alcohol consumption (%) | 6.7 (5.5, 8.0) | 5.0 (3.6, 6.7) | 8.4 (6.6, 10.7) |
| Moderate physical activity (days/week) | 2.2 (2.1, 2.3) | 2.0 (1.9, 2.1) | 2.3 (2.2, 2.5) |
| Vigorous physical activity (days/week) | 0.8 (0.8, 0.9) | 0.6 (0.5, 0.7) | 1.0 (0.9, 1.2) |
| Diagnosis of Alzheimer’s Disease or Dementia (%) | 0.7 (0.3, 1.2) | 0.3 (0.1, 0.6) | 1.1 (0.5, 2.3) |
| Hypertension (%) | 80 (77.9, 81.9) | 80.8 (77.5, 83.7) | 79.2 (75.7, 82.3) |
| Medication for high blood pressure (%) | 71.5 (69, 73.9) | 74.5 (71.3, 77.5) | 68.4 (64.5, 72.1) |
| Medication for diabetes (%) | 79.7 (77.3, 81.8) | 79.6 (76.5, 82.4) | 79.7 (76.5, 82.6) |
| Insulin use for diabetes (%) | 18.1 (16, 20.3) | 18.6 (16.2, 21.3) | 17.6 (14.6, 21) |
| Oral medication use for diabetes (%) | 73.1 (70.8, 75.2) | 72.7 (69.9, 75.3) | 73.5 (70.1, 76.6) |
| Difficulty in instrumental activities of daily living (any of 4; proportion) | 0.2 (0.2, 0.3) | 0.3 (0.2, 0.3) | 0.2 (0.1, 0.3) |
| Vision impairment (%) | 11.2 (9.5, 13.2) | 13.2 (10.7, 16.2) | 9.1 (7.3, 11.3) |
| Balance impairment (%)a | 19.3 (17.4, 21.2) | 18.3 (16.2, 20.7) | 20.1 (17.7, 22.7) |
| Cognitive impairment (%)b | 25.1 (23.2, 27.2) | 26 (23.3, 28.9) | 24.2 (21.7, 26.9) |
| Number of conditions for which medications are used c | 3.4 (3.3, 3.4) | 3.5 (3.3, 3.6) | 3.2 (3.1, 3.3) |
| Clinical measures | |||
| BMI (kg/m2) | 32.4 (32.1, 32.6) | 33 (32.5, 33.6) | 31.6 (31.3, 32) |
| Waist circumference (cm) | 109.9 (109.1, 110.7) | 107.4 (106.3, 108.5) | 112.4 (111.2, 113.7) |
| Systolic blood pressure (mm Hg) | 131.3 (130.3, 132.3) | 129.8 (128.5, 131.2) | 132.8 (131.3, 134.2) |
| Diastolic blood pressure (mm Hg) | 79.1 (78.5, 79.6) | 78.9 (78.1, 79.7) | 79.2 (78.4, 80.1) |
| HbA1c (%) | 6.9 (6.8, 6.9) | 6.8 (6.7, 6.9) | 7 (6.8, 7.1) |
| Total cholesterol (mg/dL) | 188.5 (186.4, 190.7) | 192.5 (189.5, 195.6) | 184.4 (181.3, 187.5) |
| HDL cholesterol (mg/dL) | 49.8 (48.9, 50.7) | 52.8 (51.7, 54) | 46.6 (45.3, 47.8) |
| Non-HDL cholesterol (mg/dL) | 140.1 (138.1, 142.1) | 140.5 (137.4, 143.6) | 139.7 (136.7, 142.7) |
| Baseline ABC controld | 24.9 (22.5, 27.5) | 26.2 (22.8, 29.8) | 23.7 (20.1, 27.7) |
| Glycemic controle | 75.1 (72.8, 77.3) | 76.6 (73.5, 79.5) | 73.5 (69.5, 77.2) |
| Blood pressure control (SBP < 140 and DBP < 90 mmHg) | 67.8 (65.6, 70) | 70.6 (67.3, 73.8) | 65 (61.7, 68.2) |
| Cholesterol control (non-HDL cholesterol < 130 mg/dL) | 44.6 (41.9, 47.3) | 45.4 (41.4, 49.5) | 43.7 (39.8, 47.6) |
| Mental health | |||
| Center for Epidemiologic Studies—Depression 8 score ≥ 3 | 27.0 (24.8, 29.3) | 32.6 (29.5, 35.8) | 21.3 (18.2, 24.8) |
| Antidepressants | 22.4 (20.1, 24.9) | 28.0 (24.7, 31.5) | 16.6 (14, 19.6) |
| Antipsychotics | 17.3 (15.4, 19.5) | 22.1 (19.4, 25) | 12.4 (10.3, 14.9) |
| Follow-up observations (count) | 3332 | 1866 | 1466 |
| Number of observations achieving ABC control in prospective wave | 918 | 495 | 423 |
Values are survey weighted means or proportions, and 95% confidence intervals for first assessment of Center for Epidemiologic Studies—Depression 8 at the wave with a biomarker measurement (2006-2007 onwards for HRS). We first normalized sample weights for an observation by dividing it by the sum of sample weights for the biomarker sample for each wave. Next, we inverse-weighted the normalized sample weight with proportion of participants in the pooled serial cross-sectional dataset corresponding to that wave to account for wave-specific sample sizes.
Difficulty in instrumental activities of daily living consisted of whether the participant had any difficulty with one of: managing money, taking medications, shopping for groceries, and preparing meals.
aThe semi-tandem test was performed first, followed by either the side-by-side if they failed the former or full tandem if they passed it. Balance impairment was failure to complete any one of the three tests.
bCognitive impairment was defined as successful completion of two or fewer tests only among those who completed all four of immediate word recall, delayed word recall, counting backwards and serial 7 subtraction tests.
cOut of 14 conditions for which medications are used: blood pressure, insulin use, oral hypoglycemic use, stroke, angina, congestive heart failure, cancer, heart attack, other heart problems, high cholesterol, breathing difficulty, other lung disease, stomach disease, depression.
dGlycemic (HbA1c), blood pressure, and cholesterol.
eHbA1c < 7.0% [53 mmol/mol] if < 65 years, < 7.5% [58 mmol/mol] if ≥ 65 years or < 8.0% [64 mmol/mol] with comorbidities.
Of the 3 332 observations, 29.8% (95% CI, 27.9-31.9) achieved joint ABC control at four years, while 73.5% (95% CI, 70.8-75.9), 69.7% (95% CI, 67.6-71.7), and 54.0 (95% CI, 51.6-56.4) independently achieved glycemic, BP, and cholesterol control in the two waves following the wave when depressive symptoms were ascertained. Of the 2 531 unique individuals, 34.7% achieved joint ABC control at least once during a follow-up visit. Relative to those without depressive symptoms, individuals with depressive symptoms were less likely to achieve joint ABC control (RR: 0.82 [95% CI, 0.69-0.98]) and glycemic control (RR: 0.90 [95% CI, 0.84-0.97]) but were not different in achievement of BP control (RR: 0.99 [95% CI, 0.93-1.06]) and cholesterol control (RR: 0.96 [95% CI, 0.87-1.06]) before covariate adjustment. After adjusting for covariates unlikely to be mediators of depression and diabetes control, depressive symptoms were neither associated with achievement of joint ABC control (RR: 0.90 [95% CI, 0.76-1.07]) nor achievement of glycemic, BP, or cholesterol control (Table 2). Results from the fully adjusted model were similar to the partially adjusted model.
| Relative risk of longitudinal glycemic, blood pressure, and cholesterol control by depressive symptoms.
| Outcome | Unadjusted RR (95% CI) | Partially adjusted RR (95% CI) | Fully adjusted RR (95% CI) |
|---|---|---|---|
| Joint ABCa | 0.82 (0.69, 0.98) | 0.90 (0.76, 1.07) | 0.91 (0.76, 1.09) |
| Glycemic control | 0.90 (0.84, 0.97) | 0.96 (0.90, 1.02) | 0.95 (0.89, 1.02) |
| Blood pressure control | 0.99 (0.93, 1.06) | 1.00 (0.94, 1.07) | 0.99 (0.93, 1.06) |
| Cholesterol control | 0.96 (0.87, 1.06) | 0.99 (0.90, 1.09) | 1.01 (0.92, 1.12) |
| Outcome | Unadjusted RR (95% CI) | Partially adjusted RR (95% CI) | Fully adjusted RR (95% CI) |
|---|---|---|---|
| Joint ABCa | 0.82 (0.69, 0.98) | 0.90 (0.76, 1.07) | 0.91 (0.76, 1.09) |
| Glycemic control | 0.90 (0.84, 0.97) | 0.96 (0.90, 1.02) | 0.95 (0.89, 1.02) |
| Blood pressure control | 0.99 (0.93, 1.06) | 1.00 (0.94, 1.07) | 0.99 (0.93, 1.06) |
| Cholesterol control | 0.96 (0.87, 1.06) | 0.99 (0.90, 1.09) | 1.01 (0.92, 1.12) |
Partially adjusted associations are from separate modified multivariable Poisson regression models after adjusting for baseline glycemic, blood pressure, and cholesterol control, key socio-demographic covariates (age, gender, race-ethnicity), diabetes duration, household size and household wealth as well as wave fixed effects. Fully adjusted associations additionally adjusted for variables that may be potentially either mediators or confounders such as labor force participation status, any limitations in instrumental activities of daily living, moderate and vigorous physical activity, frequent alcohol consumption, body mass index, history of hypertension, medication use (diabetes, hypertension, high cholesterol, antidepressants, antipsychotics).
aABC: Glycemic (HbA1c), blood pressure, and cholesterol control.
| Relative risk of longitudinal glycemic, blood pressure, and cholesterol control by depressive symptoms.
| Outcome | Unadjusted RR (95% CI) | Partially adjusted RR (95% CI) | Fully adjusted RR (95% CI) |
|---|---|---|---|
| Joint ABCa | 0.82 (0.69, 0.98) | 0.90 (0.76, 1.07) | 0.91 (0.76, 1.09) |
| Glycemic control | 0.90 (0.84, 0.97) | 0.96 (0.90, 1.02) | 0.95 (0.89, 1.02) |
| Blood pressure control | 0.99 (0.93, 1.06) | 1.00 (0.94, 1.07) | 0.99 (0.93, 1.06) |
| Cholesterol control | 0.96 (0.87, 1.06) | 0.99 (0.90, 1.09) | 1.01 (0.92, 1.12) |
| Outcome | Unadjusted RR (95% CI) | Partially adjusted RR (95% CI) | Fully adjusted RR (95% CI) |
|---|---|---|---|
| Joint ABCa | 0.82 (0.69, 0.98) | 0.90 (0.76, 1.07) | 0.91 (0.76, 1.09) |
| Glycemic control | 0.90 (0.84, 0.97) | 0.96 (0.90, 1.02) | 0.95 (0.89, 1.02) |
| Blood pressure control | 0.99 (0.93, 1.06) | 1.00 (0.94, 1.07) | 0.99 (0.93, 1.06) |
| Cholesterol control | 0.96 (0.87, 1.06) | 0.99 (0.90, 1.09) | 1.01 (0.92, 1.12) |
Partially adjusted associations are from separate modified multivariable Poisson regression models after adjusting for baseline glycemic, blood pressure, and cholesterol control, key socio-demographic covariates (age, gender, race-ethnicity), diabetes duration, household size and household wealth as well as wave fixed effects. Fully adjusted associations additionally adjusted for variables that may be potentially either mediators or confounders such as labor force participation status, any limitations in instrumental activities of daily living, moderate and vigorous physical activity, frequent alcohol consumption, body mass index, history of hypertension, medication use (diabetes, hypertension, high cholesterol, antidepressants, antipsychotics).
aABC: Glycemic (HbA1c), blood pressure, and cholesterol control.
Baseline glycemic (RR: 1.60 [95% CI, 1.28-1.99]), BP (RR: 1.36 [95% CI, 1.15-1.62]), and cholesterol (RR: 1.20 [95% CI, 1.05-1.38]) control were robust predictors of future achievement of ABC control as well as achievement of the corresponding biomarker control, independent of other modifiable and non-modifiable risk factors (Supplementary Table 6). Baseline usage of diabetes (RR: 1.00 [95% CI, 0.84-1.19]) and BP (RR: 0.91 [95% CI, 0.72-1.16]) medications were not associated with joint ABC control. However, taking medication for high cholesterol (RR: 1.28 [95% CI, 1.09-1.51]) was associated with higher achievement of joint ABC control.
Association of depressive symptoms with longitudinal achievement of ABC control did not vary by subgroups of age, gender, baseline ABC control status, use of antidepressants or diabetes medication, and duration of diabetes (Figure 2). However, individuals with any functional limitations were less likely to achieve ABC control if they had depressive symptoms (RR: 0.64 [95% CI, 0.42-0.98]), while individuals without functional limitations did not differ in achievement based on depressive symptoms.
Longitudinal associations of depressive symptoms with joint glycemic, blood pressure, and cholesterol control across socio-demographic and clinical subgroups. All associations are from separate modified survey weighted multivariable Poisson regression models with statistical interaction of depressive symptoms with effect modifier after adjusting for individual and household covariates. We assessed risk of glycemic (HbA1c), blood pressure, and cholesterol control (ABC control) across subgroups of those with ABC control status at the same wave as when depressive symptoms were ascertained. Duration of diabetes was ascertained based on adjudicated self-reported data.
Associations remained null when incorporating loss to follow-up using inverse probability weights (Supplementary Table 7). Using CES-D8 as a continuous exposure variable did not change our results (unadjusted RR: 0.95 [95% CI, 0.92-0.99], adjusted RR: 0.98 [95% CI, 0.95-1.02]).
Discussion
This pooled, nationally representative analysis of middle-aged and older adults from the United States showed that depressive symptoms were not associated with longitudinal achievement of joint glycemic, BP, and cholesterol (ABC) control after adjusting for individual and household covariates. The findings were consistent across socio-demographic and clinical subgroups of age, gender, baseline ABC control, medication use, and duration of diabetes. Individuals with functional limitations were less likely to achieve ABC control if they had depressive symptoms. This study, therefore, provides robust evidence that baseline depressive symptoms are not associated with future ABC control for most middle-aged and older adults.
Our findings are consistent with a limited body of previous evidence such as the Pathways Epidemiologic Study which showed no differences in long-term glycemic, BP, and cholesterol control among patients with and without depression.14–16 Previous studies suggest that behavioral and pharmacological interventions for management of depression were associated with improved glycemic control.35,36 However consistent with this study, the Diabetes and Depression Study, a randomized controlled trial of cognitive behavioral therapy and sertraline (a selective serotonin reuptake inhibitor) reported no differences in diabetes control, although depressive symptoms improved.37 Depression was shown to be associated with an elevated risk of type 2 diabetes after adjusting for key socio-demographic factors only if the individual had pre-existing metabolic dysregulation.38 Individuals diagnosed with major depressive disorder (MDD) after diagnosis of type 2 diabetes had similar HbA1c in the years after follow-up, compared to those who did not have MDD, after adjusting for individual risk factors associated with glycemic control.39 In a systematic review of observational studies, depressive symptoms were associated with higher HbA1c.40 However, the observed effect sizes were small. Depression was also associated with a higher risk of diabetes complications in a meta-analysis of observational studies, although the studies included displayed substantial heterogeneity.41 Furthermore, the most influential study included in the meta-analysis did not adjust for baseline glycemic control.42 Another systematic review of depression and diabetes reported on the paucity of longitudinal studies and how the findings of whether these associations are bidirectional are mixed.43 Although depressive symptoms may not be associated with diabetes management as evidenced by several longitudinal studies, including ours, trials of integrated depression and diabetes management strategies report reductions in both conditions, suggesting that the observed effects may operate through independent pathways.44,45
The unadjusted inverse association between depressive symptoms and future ABC control, also reported elsewhere, and depressive symptoms and risk of diabetes complications, is likely due to the cross-sectional association of depression and diabetes management.4–6,41 For instance, glycemic control was associated with diabetes distress and may reflect subjective experiences as opposed to physiological pathways of worsened control.14 Therefore, reporting longitudinal associations between depressive symptoms and ABC control may be confounded if they do not adjust for baseline diabetes management, which remained a robust predictor of future ABC control in this study.
Although this study found that depression is not associated with future ABC control, plausible biological and behavioral pathways exist for poor diabetes outcomes among individuals with depression. For instance, some antidepressants are associated with iatrogenic weight gain and may worsen diabetes control.46 There may also be shared inflammatory pathophysiology between diabetes and depression that may explain the higher co-occurrence than expected by chance.41 Depression may also be associated with poor diet, decrease in physical activity, sleep quality, and adherence to diabetes treatments.
This study is an advance over previous studies of depressive symptoms and future diabetes control given the nationally representative data as well as adjustment for baseline diabetes management and a comprehensive set of individual and household covariates. However, there are some limitations. First, we did not have information on types of diabetes or information on medication used for the management of depression and diabetes. We also did not explore the risk of diabetes-related distress that is attributed to concerns related to maintenance of diabetes control, fear of complications, or financial burden associated with disease management.14 Second, assessment of HbA1c, total cholesterol, and HDL cholesterol concentrations were conducted using dried blood spots and in different laboratories. Although these results were calibrated, previous studies suggest that the method may have challenges in validation due to issues with size of blood spots, transport times, temperature, and humidity. Moreover, the concentration of LDL cholesterol, a more appropriate biomarker for defining cholesterol control, was not measured. Finally, there was substantial loss to follow-up between the study and the analytic samples. However, our results, using inverse probability weights for attrition, showed similar results. Furthermore, risk ratios of depressive symptoms with ABC control reflect long-term relative risks that are minimally confounded by preceding variables and reflect the total effects of all attenuating and amplifying mediator variables.
In conclusion, findings from this study suggest depressive symptoms may not be associated with future achievement of glycemic, BP, and cholesterol control, independent of baseline achievement and other individual and household characteristics. Depression among people with diabetes increased by 60% during the COVID-19 pandemic but data included in this study was pre-pandemic.6 Results from this study therefore imply that although depression may have increased among people with diabetes, it may not reflect as long-term changes in diabetes control. Care models for diabetes and depression should continue to focus on both independently to improve health. Further research is also needed to understand implications of depressive symptoms on other aspects of diabetes management, such as quality of life and medication adherence, among older adults with diabetes.
Acknowledgments
We thank the participants and enumerators of Health and Retirement Study. We thank the Gateway to Global Aging Data Team for preparing the harmonized data.
Author contributions
Namitha Mary Varghese: Conceptualization, Investigation, Writing—Original Draft, Writing—Review & Editing. Jithin Sam Varghese: Conceptualization, Investigation, Data Curation, Formal Analysis, Visualization, Writing—Review & Editing
Data Availability
All datasets used in this analysis is available for download at https://g2aging.org/downloads. The code for the analysis is available on https://github.com/jvargh7/g2aging_depression.
Funding
Research reported in this publication was partially supported by the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health under Award Number P30DK111024.
Conflicts of interest
The authors declare that they have no conflicts of interest.
Ethics approval and consent to participate
We were exempt from ethical approval by the Institutional Review Board of Emory University for this secondary data analysis study. All participants gave written informed consent before participation.
