Abstract
Background
Metabolic syndrome (MetSyn) is the co-occurrence of obesity with metabolic derangements. Prior research implicates MetSyn in prolonging the course of depression in older adults, but MetSyn’s effect on antidepressant response is unknown in this population.
Objectives
To determine if MetSyn and related metabolic dyscrasias are associated with decreased remission from depression among older adults treated pharmacologically for depression.
Design
Secondary analysis of a randomized controlled trial.
Participants
Adults 60 years and older with major depressive disorder (MDD, n=435) recruited from three academic medical centers in North America.
Intervention
Open-label, protocolized treatment with venlafaxine extended-release for twelve or more weeks.
Main Measures
Time-to-remission from depression, with remission defined as Montgomery-Åsberg Depression Rating Scale (MADRS) score ≤ 10 at last two visits.
Key Results
Of 435 participants (mean age 69.1 years), 222 (51%) met criteria for MetSyn at baseline; MetSyn was associated with greater severity (MADRS score) and chronicity of depression at baseline. Remission was achieved in 182 participants (42%). In the unadjusted analysis, MetSyn was associated with prolonged time-to-remission (hazard ratio for remission 0.71, 95% CI 0.52 – 0.95) but this relationship was not significant in the adjusted model; greater number of MetSyn components and lower HDL cholesterol had similar effects. Only diastolic blood pressure (DBP) was a significant predictor of time-to-remission before and after adjustment, with higher DBP predicting prolonged time-to-remission. Insulin sensitivity did not predict time-to-remission.
Conclusion
The presence of MetSyn in depressed, older adults is associated with greater symptom severity and chronicity of depression in our sample, which appear to account for the decreased antidepressant response observed among those with MetSyn. Additionally, our preliminary finding relating higher DBP with decreased antidepressant response bears further examination and replication.
Keywords: late-life depression, major depressive disorder, metabolic syndrome, elderly, venlafaxine
INTRODUCTION
The prevalence of MDD is estimated at 4.4% worldwide (Ferrari 2013), high rates of treatment resistance: fewer than 40% will achieve remission during an initial course of treatment (Kemp 2008, Rush 2006). In older adults, depression’s dire consequences, including cognitive decline (Zahodne 2014), increased health care utilization (Shao 2017), functional decline (Callahan 1998, Barry 2009), and increased overall mortality (Laursen 2016) have prompted an aggressive search for predictors of therapeutic response, such as cerebrovascular disease (Taylor 2013, Aizenstein 2016), genetic factors (Marshe 2017), peripheral biomarkers and symptomatic phenotypes (Rutherford 2016, Nelson 2009). Gallo et al (2016) recently demonstrated that management of depression reduces mortality among depressed older adults with multimorbidity.
Metabolic Syndrome (MetSyn) is defined as a combination of obesity and two or more of the following components: hypertension, high triglycerides (TG), low high-density lipoprotein (HDL) cholesterol, and fasting hyperglycemia (Alberti 2006). MetSyn is postulated to impact prevalence and severity of depressive disorders (Pan 2012, Mansur 2015). The association is bidirectional, with depression predicting MetSyn and vice versa (Pan 2012). Recent inquiries into treatment relevance of comorbid MetSyn on depression in adults of mixed ages have yielded varied results: Sagud and colleagues (2013) found no association between MetSyn and treatment resistance, while Vogelzangs et al (2014) and Woo et al (2016) associated treatment resistance with metabolic dysfunction but did not evaluate MetSyn specifically.
Treatment response in late-life depression (LLD) might be particularly linked to MetSyn, as both conditions are highly prevalent among older adults, share a variety of mechanistic links (Hryhorczuk 2013), and result in shared downstream effects, including cardiovascular and cerebrovascular sequelae (Almas 2015, Daskalopoulou 2016). Only two studies have examined the course of LLD in the context of MetSyn. Vogelzangs et al (2011) found MetSyn (but no individual MetSyn components) to predict depression chronicity. Marijnissen et al (2017) also found MetSyn to predict a more chronic course of depression but not after multivariate adjustment; number of MetSyn components, waist circumference and HDL were predictive of depression chronicity as well. A limitation of both studies is that the treatment of LLD was not standardized or well-described, thus these studies cannot evaluate MetSyn’s association with treatment resistance.
The Incomplete Response in Late Life Depression: Getting to Remission (IRL GRey) study was carried out to better inform therapy for treatment-resistant depression in older adults (Lenze 2015). Participants ≥ 60 years with MDD were treated with open-label venlafaxine extended-release (XR) 150–300 mg/day for 12 weeks; non-remitters were then randomized 1:1 to receive either aripiprazole or placebo. When the study was initiatially conceived, the concept of metabolic depression had not been fully explored in the literature. However, during the course of study conduct, it became clear that metabolic parameters and conditions may impact depression treatment response. Data from the initial phase of treatment with venlafaxine offered a prime opportunity to test the hypothesis that, among older adults in North America (where MetSyn is quite prevalent) comorbid MetSyn may predict poorer response to treatment of MDD. Based on the aforementioned associations between MetSyn and depression, we hypothesized the following: (1) presence of MetSyn would predict a prolonged time-to-remission from depression (primary hypothesis); (2) participants with a higher number of components of MetSyn would have a prolonged time-to-remission; (3) given the the central role of insulin resistance in MetSyn and the multiple pathophysiologic links between diabetes mellitus and MDD (Moulton 2015), we hypothesized that a surrogate measure of insulin resistance, the Homeostasis Model Assessment for Insulin Resistance (HOMA-IR) value, would also be independently associated with a prolonged time-to-remission.
METHODS
Description of the parent study
Inclusion criteria for IRL GRey were as follows: adults 60 years and older with non-psychotic MDD and MADRS score 15 or more (Lenze 2015). Exclusion criteria included dementia or other cognitive impairment evidenced by score ≤ 20 on the Mini-Mental Status Examination (MMSE, Folstein 1975), bipolar disorder, schizophrenia, schizoaffective disorder, psychosis, alcohol or substance misuse or dependence in the preceding 6 months, and unstable medical illness. Participants were recruited at the University of Pittsburgh, the Centre for Addiction and Mental Health in Toronto, and Washington University in St Louis, from July 2009 to December 2013. The study protocol was approved by respective site institutional review boards, and all participants provided written, informed consent. Further details on sample-size determination, recruitment, enrollment, randomization, assessments, procedures, and safety review are available elsewhere (Lenze 2015).
Assessments
The MADRS assesses the severity of 10 depressive symptoms with total score ranging from 0 (absence of any depressive symptoms) to 60: apparent sadness, reported sadness, inner tension, reduced sleep, reduced appetite, concentration difficulties, lassitude, inability to feel, pessimistic thoughts, and suicidal thoughts (Mongtomery and Åsberg 1979). Presence of MetSyn was determined at baseline using the International Diabetes Federation’s (IDF) worldwide consensus definition, which requires central obesity based on body mass index (BMI) or waist circumference (WC) (see below), and two or more of the following components: elevated triglycerides (TG), low HDL cholesterol, elevated blood pressure, and elevated fasting plasma glucose (Alberti 2006). Criteria for obesity are met in the presence of BMI ≥ 30 kg/m2 or by reaching a sex- and ethnicity-specific waist circumference cutoff (≥ 80 cm in women, ≥ 90 or ≥ 94 cm in men). Elevated TG is defined as ≥ 150 mg/dL (1.7 mmol/L) or specific treatment for this lipid abnormality. Low HDL cholesterol is defined as < 40 mg/dL (1.03 mmol/L) in men or < 50 mg/dL (1.29 mmol/L) in women, or specific treatment for this lipid abnormality. Elevated blood pressure is defined as a systolic pressure ≥ 130 mm Hg or diastolic ≥ 85 mm Hg, or treatment for hypertension. Elevated fasting glucose is defined as a fasting plasma glucose ≥ 100 mg/dL (5.6 mmol/L) or previously diagnosed diabetes mellitus type 2. Waist circumference is determined by measuring circumference halfway between the inferior margin of the ribs and the iliac crest. Weight and height are determined with light clothing on, shoes removed, and pockets emptied. Blood pressure is measured seated after five minutes of rest. In cases where multiple baseline readings of these parameters were taken, data that most closely preceded initiation of treatment was used. History of hypertension, diabetes mellitus, low HDL, or hypertriglyceridemia and active pharmacotherapy for the same were assessed by participant report. As per IDF recommendations, individuals meeting criteria for these MetSyn components by history and/or medication use were included in our analyses. Blood samples were drawn from participants in the fasting state. HOMA-IR was calculated as follows: [fasting insulin (in μU/mL)] x [fasting glucose (in mg/dL)]/405 (Matthews et al, 1985).
Procedures
Open-label venlafaxine XR was initiated at 37.5 mg daily in all participants and uptitrated in increments of 37.5 mg/d after a minimum of three days, targeting 150 mg/d. After 6 weeks of treatment, if MADRS remained greater than 10, then venlafaxine dosage was further increased by increments of 37.5–75 mg/d over at least three days based on clinical response and as symptomatically tolerated up to a maximum dose of 300 mg/d. Those experiencing side effects underwent slower titration or reduced dosage. MADRS score was re-assessed at in-person follow-up visits every 1–2 weeks. After 12 weeks (or longer, if necessary, to allow for four weeks on the highest-tolerated dose of venlafaxine or to clarify remission status), participants were categorized as having remitted or not, with remission defined as reaching MADRS 10 or less for the last two follow-up visits. Participants were allowed to continue outside psychotherapy and medications for anxiety or sleep.
Statistical Analysis
Differences in baseline characteristics between those with and without MetSyn were evaluated using a two-tailed t-test or Wilcoxon test for continuous variables and a chi-square or Fisher’s Exact test for categorical variables. The association between MetSyn status and remission from depression was assessed with a chi-square test. For our primary hypothesis, Cox proportional hazards regression was employed to test for MetSyn status’ effect on time-to- remission. In this model, a greater time-to-remission corresponds to a decreased hazard ratio for remission. Univariate proportional hazards regression models were similarly created for each metabolic variable listed in Table 1 to investigate effect on time-to-remission. Then, MetSyn status, individual values for metabolic paramters and treatment with antihypertensive, lipid- and glucose-lowering agents were included separately in an adjusted model. The adjusted model included those demographic and depression-related variables found in stepwise proportional hazards regression to predict time-to-remission at a significance level of 0.05. The variables evaluated for predictive effect were age, sex, education, living status (alone or with others), study site, referral source, outside psychotherapy, baseline total MADRS score, duration of current depressive episode (dichotomized to > 2 years or ≤ 2 years), recurrent depression or single episode, MMSE, Repeatable Battery for the Assessment of Neuropsychological Status (RBANS, Randolph et al 1998), and prior antidepressant treatment assessed with the Antidepressant Treatment History Form (ATHF, Sackeim 2001) as dichotomized to values 0–2 (no prior adequate antidepressant trial in this depressive episode) or 3–4 (≥1 adequate trials). As obstructive sleep apnea has been previously associated with treatment resistance in this sample (Waterman et al 2016), this analysis was not repeated here. Analyses were performed on SAS 9.4 (SAS Institute, Cary, NC).
Table 1.
Baseline participant characteristics by MetSyn status.
| All Patients* n=435 |
MetSyn Absent n=211 |
MetSyn Present n=222 |
Chi square/t/Wilcoxon value | p-value | |
|---|---|---|---|---|---|
|
| |||||
| Age, years (SD) | 69.1 (7.2) | 69.1 (7.4) | 69.1 (7.1) | 0.05 | 0.96 |
|
| |||||
| Female sex, n (%) | 281 (65%) | 136 (64%) | 143 (64%) | 0.0001 | 0.99 |
|
| |||||
| Race, n (%) | Fisher’s Exact† | 0.67 0.56 |
|||
|
| |||||
| Caucasian | 381 (88%) | 187 (89%) | 192 (86%) | ||
|
| |||||
| African American | 45 (10%) | 20 (9%) | 25 (11%) | ||
|
| |||||
| Asian or Pacific Islander | 8 (2%) | 3 (1%) | 5 (2%) | ||
|
| |||||
| Native American | 1 (0.2%) | 1 (0.5%) | 0 (0%) | ||
|
| |||||
| Site, n (%) | 18.13 | <0.001 | |||
|
| |||||
| Pittsburgh | 189 (43%) | 72 (34%) | 116 (52%) | ||
|
| |||||
| Toronto | 113 (26%) | 71 (34%) | 41 (18%) | ||
|
| |||||
| St. Louis | 133 (31%) | 68 (32%) | 65 (29%) | ||
|
| |||||
| Education, years (SD) | 14.4 (2.8) | 14.5 (2.9) | 14.3 (2.8) | 0.88 | 0.38 |
|
| |||||
| Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) (SD) | 94.4 (15.9) | 93.9 (17.9) | 95.0 (13.9) | 42932 | 0.67 |
|
| |||||
| Mini Mental Status Examiation (MMSE)‡ (SD) | 28.7 (1.5) | 28.7 (1.5) | 28.6 (1.6) | 0.67 | 0.50 |
|
| |||||
| Montgomery Asberg Depression Rating Scale (MADRS) (SD) | 26.6 (5.8) | 25.8 (6.1) | 27.4 (5.4) | 42124 | 0.005 |
| Apparent sadness (SD) | 2.7 (1.0) | 2.6 (0.9) | 2.8 (1.0) | 43452 | 0.06 |
| Reported sadness (SD) | 3.7 (0.7) | 3.6 (0.8) | 3.7 (0.7) | 43202 | 0.03 |
| Inner tension (SD) | 2.9 (1.1) | 2.8 (1.1) | 2.9 (1.1) | 44432 | 0.27 |
| Reduced sleep (SD) | 3.4 (1.6) | 3.2 (1.7) | 3.6 (1.6) | 42531 | 0.01 |
| Reduced appetite (SD) | 1.7 (1.4) | 1.7 (1.5) | 1.7 (1.4) | 46015 | 0.86 |
| Concentration difficulties (SD) | 2.7 (1.4) | 2.6 (1.4) | 2.8 (1.4) | 43577 | 0.08 |
| Lassitude (SD) | 3.2 (1.1) | 3.1 (1.1) | 3.4 (1.0) | 41754 | 0.001 |
| Inability to feel (SD) | 3.1 (1.2) | 3.0 (1.2) | 3.1 (1.2) | 44503 | 0.31 |
| Pessimistic thoughts (SD) | 2.7 (1.1) | 2.6 (1.2) | 2.7 (1.1) | 45126 | 0.60 |
| Suicidal thoughts (SD) | 0.7 (0.9) | 0.7 (1.0) | 0.7 (0.9) | 44551 | 0.30 |
|
| |||||
| Hamilton Rating Scale for Depression (HRSD) (SD) | 19.9 (5.0) | 19.3 (5.4) | 20.5 (4.5) | −2.55 | 0.01 |
|
| |||||
| Age of first depressive episode, years (SD) | 42.7 (21.5) | 42.8 (22.1) | 42.6 (21.1) | 0.09 | 0.93 |
|
| |||||
| Duration of current depressive episode, weeks (SD) | 286 (600) | 201 (401) | 363 (730) | 42113 | 0.008 |
|
| |||||
| Suicidal Ideation Scale§ (SD) | 2.25 (4.4) | 2.07 (4.2) | 2.41 (4.6) | 44185 | 0.26 |
|
| |||||
| Antidepressant Treatment History Form (ATHF) (SD) | 2.4 (1.6) | 2.3 (1.6) | 2.6 (1.5) | −1.96 | 0.051 |
|
| |||||
| Body mass index, kg/m2 (SD) | 29.7 (6.8) | 26.5 (5.6) | 32.7 (6.5) | −10.51 | <0.001 |
|
| |||||
| Waist circumference, cm (SD) | 99.1 (16.9) | 90.5 (14.0) | 106.9 (15.5) | −11.10 | <0.001 |
|
| |||||
| Systolic blood pressure, mmHg (SD) | 132.3 (19.0) | 128.8 (19.1) | 135.5 (18.5) | −4.10 | <0.001 |
|
| |||||
| Diastolic blood pressure, mmHg (SD) | 76.4 (11.4) | 74.7 (10.7) | 78.1 (11.6) | −3.23 | 0.001 |
|
| |||||
| Pulse pressure||, mmHg (SD) | 55.9 (15.7) | 53.9 (15.5) | 57.7 (15.7) | −2.56 | 0.01 |
|
| |||||
| Total cholesterol, mg/dL (SD) | 192.0 (44.1) | 196.9 (44.4) | 187.2 (43.4) | 2.30 | 0.02 |
|
| |||||
| High-density lipoprotein, mg/dL (SD) | 53.4 (17.4) | 61.4 (17.7) | 45.6 (13.3) | 57910 | <0.001 |
|
| |||||
| Low-density lipoprotein, mg/dL (SD) | 109.6 (37.9) | 113.5 (38.5) | 105.9 (37.2) | 2.09 | 0.04 |
|
| |||||
| Triglycerides, mg/dL (SD) | 145.9 (84.8) | 110.4 (54.9) | 180.0 (94.0) | 33108 | <0.001 |
|
| |||||
| Glucose, mg/dL (SD) | 109.7 (40.6) | 95.0 (20.6) | 123.7 (49.4) | 33114 | <0.001 |
|
| |||||
| Insulin¶, mIU/L (SD) | 10.5 (8.5) | 8.5 (7.0) | 13.3 (9.7) | 5273 | <0.001 |
|
| |||||
| HOMA-IR# (SD) | 2.7 (2.5) | 2.1 (2.0) | 3.6 (2.8) | 5552 | <0.001 |
|
| |||||
| Number of MetSyn Components (SD)** | 2.7 (1.4) | 1.6 (0.9) | 3.8 (0.8) | 23642 | <0.001 |
|
| |||||
| Cumulative Illness Rating Scale for Geriatrics (CIRS-G) †† (SD) | 9.8 (4.4) | 8.4 (4.1) | 11.1 (4.3) | −6.65 | <0.001 |
|
| |||||
| Co-treatment with anti-hypertensive, n (SD)& | 224 (55) | 72 (37) | 152 (72) | 50.52 | <0.0001 |
|
| |||||
| Co-treatment with lipid-lowering agent, n (SD)∋ | 208 (51) | 83 (43) | 124 (59) | 10.65 | <0.0001 |
|
| |||||
| Co-treatment with glucose-lowering agent, n (SD)Ψ | 76 (19) | 11 (6) | 65 (31) | 42.18 | <0.0001 |
|
| |||||
| History of cardiovascular disease‡‡, n (%) | 133 (31%) | 44 (21%) | 89 (40%) | 18.81 | <0.001 |
|
| |||||
| History of cerebrovascular disease, n (%) | 25 (5.8%) | 9 (4.3%) | 16 (7.2%) | 1.72 | 0.19 |
|
| |||||
| History of smoking, n (%) | 137 (31%) | 68 (32%) | 69 (31%) | 0.07 | 0.80 |
|
| |||||
| History of obstructive sleep apnea, n (%) | 80 (18.4%) | 20 (9.5%) | 60 (27.0%) | 22.1 | <0.001 |
Two participants could not be assigned a MetSyn status due to missing data, but available metabolic data was included in secondary analyses;
Top p-value for all 4 categories, bottom p-value for Caucasian and other 3 categories combined;
Of note, 54 participants’ data was missing for this assessment;
Pulse pressure is calculated by subtracting diastolic pressure from systolic pressure;
Total sample sizes for insulin and HOMA-IR were reduced to 142 due to availability of usable blood samples;
Homeostasis Model Assessment for Insulin Resistance;
These include obesity, hypertension, high TG, low HDL, and hyperglycemia as defined respectively by the International Diabetes Federation’s worldwide consensus definition of MetSyn (Alberti 2006);
includes calcium channel blockers, ACE inhibitors, beta blockers, angiotensin II receptor blockers, diuretics, alpha receptor antagonists and nitrates;
includes statins, niacin, fish oil, cholesterol absorption inhibitors, bile acid binding resins, fibrates, omega-3 fatty acids;
includes biguanides, sulfonylureas, meglitinide derivatives, alpha-glucosidase inhibitors, thiazolidinediones, glucagonlike peptide-1 agonists, dipeptidyl peptidase inhibitors, selective sodium-glucose transporter-2 inhibitors, insulin, amylinomimetics and dopamine agonists;
Please see appendix for further details of smoking, cardiovascular disease, and cerebrovascular disease designations.
RESULTS
Participant characteristics
Of 435 participants with MDD beginning venlafaxine (Table 1), 211 did not meet criteria for MetS, 222 did meet criteria for MetSyn, and two could not be classified due to missing data. Notably, the MetSyn group had greater baseline symptom severity as measured with total MADRS and Hamilton Rating Scale for Depression (HRSD, Hamilton 1960) scores, higher MADRS symptom scores for reported sadness, reduced sleep, and lassitude, and greater duration of depressive episode; site distribution varied by MetSyn status, as well. The majority of our sample had been previously treated for their current depressive episode, with 59.1% having received an adequate course of antidepressant treatment corresponding to ATHF score of 3 or more, with nearly statistically greater ATHF score observed among those with MetSyn. Measures of most metabolic parameters differed by MetSyn status as expected, except LDL cholesterol and total cholesterol which were higher in the group without MetSyn. Those with MetSyn also had greater medical illness burden as measured by the Cumulative Illness Rating Scale for Geriatrics (CIRS-G, Miller 1992) and greater prevalence of cardiovascular disease. Otherwise, groups were similar in all other baseline characteristics (Table 1).
Relationship of MetSyn with remission
Of the participants above, 182 (41.8%) reached a sustained MADRS ≤ 10 (defined as remission), 160 did not remit, and 93 left the study before completing 12 weeks of therapy (Fig. 1). Those with MetSyn were less likely to remit (36.9%) than those without MetSyn (46.5%, Chi-Square = 4.03, d.f., = 1, p = 0.045). There was no difference in proportion of participants withdrawing from the study by MetSyn status (21.6% vs 21.3%, Chi-Square = 0.0056, d.f., = 1, p =0.94).
Figure 1. Kaplan Meier curve depicting incidence of remission by MetSyn status.
Those withdrawing from study and those who did not remit were censored from this graphical representation of remission status versus time, with censoring events represented with, “+”.
Univariate Cox proportional hazards regression demonstrated greater time-to-remission among those with MetSyn (HR for remission 0.71, 95% CI = 0.52–0.95, p = 0.02; Table 2) compared to those without MetSyn. Stepwise proportional hazards regression identified female sex (HR 1.55, 95% CI 1.21–1.88) and ATHF score ≤ 2 (HR 1.52, 95% CI 1.23–1.82) as predicting decreased time-to-remission, while greater duration of depressive episode (HR 0.68, 95% CI 0.36–0.99), and greater baseline MADRS score (HR 0.93, 95% CI 0.90–0.96) predicted increased time-to-remission. When these variables were combined to construct the fully adjusted model, MetSyn loses significance as a predictor of time-to-remission (HR 0.86, 95% CI 0.64–1.16).
Table 2.
Effects of metabolic and atherosclerotic variables on time-to-remission.
| Univariate | Adjusted | |||
|---|---|---|---|---|
| HR (95% CI) | p-value | HR (95% CI) | p-value | |
| Metabolic syndrome | 0.71 (0.52–0.95) | 0.02 | 0.86 (0.64–1.16) | 0.32 |
| Number of metabolic syndrome components, per component | 0.89 (0.80–0.99) | 0.03 | 0.95 (0.86–1.06) | 0.37 |
| Body mass index, per kg/m2 | 0.99 (0.97–1.02) | 0.54 | 1.01 (0.98–1.03) | 0.55 |
| Waist circumference, per cm | 0.99 (0.99–1.00) | 0.23 | 1.00 (0.99–1.01) | 0.46 |
| Systolic blood pressure, per 10 mmHg | 0.94 (0.87–1.01) | 0.10 | 0.94 (0.87–1.02) | 0.15 |
| Diastolic blood pressure, per 10 mmHg | 0.88 (0.77–0.995) | 0.04 | 0.87 (0.77–0.99) | 0.04 |
| Pulse pressure, per 10 mmHg | 0.98 (0.89–1.07) | 0.65 | 0.99 (0.90–1.08) | 0.83 |
| Total cholesterol, per 10 mg/dL | 1.00 (0.97–1.04) | 0.82 | 1.00 (0.96–1.03) | 0.79 |
| High-density lipoprotein, per 10 mg/dL | 1.11 (1.02–1.21) | 0.01 | 1.03 (0.95–1.13) | 0.46 |
| Low-density lipoprotein, per 10 mg/dL | 1.00 (0.97–1.04) | 0.91 | 1.00 (0.96–1.04) | 0.83 |
| Triglycerides, per 10 mg/dL | 0.98 (0.97–1.00) | 0.10 | 0.99 (0.97–1.01) | 0.43 |
| Glucose, per 10 mg/dL | 0.98 (0.95–1.02) | 0.40 | 1.00 (0.97–1.04) | 0.79 |
| Insulin, per 10 μU/mL | 0.85 (0.59–1.21) | 0.37 | 0.97 (0.71–1.34) | 0.88 |
| HOMA-IR, per unit | 0.92 (0.81–1.04) | 0.18 | 0.96 (0.85–1.07) | 0.44 |
| History of cardiovascular disease | 0.86 (0.63–1.18) | 0.36 | 0.99 (0.72–1.36) | 0.95 |
| History of cerebrovascular disease | 0.88 (0.46–1.67) | 0.69 | 0.98 (0.51–1.86) | 0.95 |
| History of smoking | 1.01 (0.74–1.38) | 0.96 | 0.96 (0.70–1.32) | 0.81 |
For continuous and numerical variables, hazard ratios reflect increasing number of the stated variable, with a lower hazard ratio corresponding to decreased incidence of remission and thus greater time-to-remission. For binary variables (metabolic syndrome, cardiovascular disease, cerebrovascular disease, and smoking), the hazard ratio presented is that for presence of the stated entity versus absence of the stated entity.
Individual metabolic variables as predictors of time-to-remission
Univariate proportional hazard regression models found increasing HDL cholesterol (HR 1.11 per 10 mg/dL, 95% CI 1.02–1.21) to predict shorter time-to-remission, while increasing DBP (HR 0.88 per 10 mmHg, 95% CI 0.77–0.995), and greater number of components of MetSyn (HR 0.89 per component, 95% CI 0.80–0.99) predicted greater time-to-remission. Upon introduction to the adjusted model, of these only DBP (HR 0.87 per 10 mmHg increase, 95% CI 0.77–0.99) remained a significant predictor. HOMA-IR was not a significant predictor of time-to-remission in either univariate or adjusted models (Table 2). Cerebrovascular disease, cardiovascular disease, and smoking were not predictive of time-to-remission.
We also examined whether medications targeting blood pressure, lipids, and glucose would confound analyses of these metabolic parameters’ respective effects on time-to-remission by inserting each as an individual variable in the univariate model, without adjustments for disease-related and demographic variables, and found no such effects (data not shown). Use of co-prescribed beta-blockers, or any antihypertensives, had no effect on time-to-remission for DBP (which remained significant) nor on SBP or pulse pressure (which remained non-significant). HDL remained significant after adjusting for use of lipid-lowering drugs, and LDL, total cholesterol, and TG remained non-significant. Finally, no moderating effect was found between use of glucose-lowering drugs and insulin, glucose, or HOMA-IR on time-to-remission, with all remaining non-significant (data not shown).
DISCUSSION
In this study of older adults with MDD, we evaluated MetSyn’s effect on remission with antidepressant treatment. To our knowledge, our study is the first to examine this relationship in older adults, as well as the first to do so in a North American population with a high prevalence of MetSyn. We had three key findings. First, MetSyn was associated with more severe depression including higher symptomatic severity and greater duration of depression. Second, MetSyn predicted greater time-to-remission with antidepressant treatment; however, significance was lost after controlling for these severity markers. In other words, reduced treatment response among those with MetSyn appears to be accounted for by the greater severity of depression observed in this group. Third, a greater number of MetSyn components and lower HDL cholesterol predicted greater time–to-remission, but not after multivariate adjustment; only higher DBP predicted greater time-to- remission with and without inclusion of covariates.
While Vogelzangs et al (2014) and Marijnissen et al (2017) both found number of MetSyn components and low HDL cholesterol to predict chronicity of depression, to our knowledge, such an effect of DBP on chronicity or treatment response in depression has not been previously reported. Moreover, DBP was the only variable to remain significant in our adjusted analyses. These results tentatively link elevated DBP with treatment-resistant depression. If present, cerebrovascular disease burden could be one mediator of such an effect. Hypertension is an established risk factor for vascular depression (Krishnan 2004), but its relationship with LLD more broadly has not been proven (Long 2015). Vascular depression lacks universally accepted diagnostic criteria, but is generally characterized by later onset in the setting of atherosclerotic risk factors and radiographic evidence of cerebrovascular disease (Aizenstein 2016, Viscogliosi 2014). Clinical features include psychomotor slowing, apathy, and greater impairment of cognition than is found in other forms of LLD, with increased risk of treatment failure, progression to dementia, and mortality (Aizenstein 2016). Elevated DBP could also denote hypothalamic-pituitary-adrenal (HPA) axis hyperactivity, which serves as a central common mechanism linking depression with MetSyn, though its relavence to antidepressant response is not well established (Jacobson 2014).Limitations and Strengths
A primary limitation of our study concerns the associations observed between MetSyn and greater symptom severity, longer duration of depressive episode, and borderline-significant greater ATHF score. Given the cross-sectional nature of the baseline data, causality cannot be assessed. Literature on MetSyn’s association with depressive symptom severity has been equivocal, with some (van Reedt Dortland 2010, Sekita 2013), but not all (Marijnissen 2017), previous authors reporting a positive relationship. Other collinear effects could also partly explain MetSyn’s effect on time-to-remission: for example, obstructive sleep apnea (OSA), a potent predictor of resistance to antidepressant treatment in this same study population (Waterman et al 2016), is highly correlated with MetSyn status in our sample.
We were unable to detect an effect of insulin concentration or HOMA-IR on remission, despite well-known epidemiological and mechanistic links between diabetes/insulin resistance and depression (Moulton et al 2015). HOMA-IR is a well-validated surrogate measure of insulin resistance in large populations, and also has reasonable correlations with more sensitive measures in younger populations (Lorenzo et al, 2010). However, HOMA may have limited utility to detect age-related insulin resistance in smaller populations, particularly in older adults (Chang et al 2006). This, in addition to missing insulin samples for a portion of study participants (Table 1), likely limited the power to detect effects of insulin resistance on treatment response in late-life depression.
Racial and ethnic minorities were not well represented in the study sample, decreasing generalizability of our findings to non-white populations. These results should be interpreted with caution, as race has been shown to modulate obesity’s effects on prevalence of depression (Hicken 2013, Xiang 2015). Inclusion of individuals in the MetSyn group who achieved the requisite number of MetSyn components due to medical history and treatment of low HDL, high TG, hypertension or diabetes, in accordance with IDF criteria for MetSyn, into our analytic model could have increased the likelihood of type 2 error. Additionally, while higher DBP could conceivably prolong time-to-remission via increased cerebrovascular disease burden, no cerebral imaging was available to correlate cerebrovascular disease markers with metabolic derangement and treatment response. Finally, use of venlafaxine in a highly standardized dosing algorithm by all participants is both a weakness and strength. Though the generalizability of these results to other antidepressants is unknown, more reliable comparisons between participants can be drawn. Indeed, a recent meta-analysis of treatment studies for MDD in older adults demonstrated greater efficacy for sertraline, paroxetine, and duloxetine relative to placebo, which was not the case for venlafaxine in the sole included study featuring venlafaxine; metabolic variables were not discussed or adjusted for (Thorlund 2015).
The sample is metabolically well characterized and may offer useful insights into the relationship between MetSyn, other metabolic variables, and depression treatment response. Other study strengths include a fairly large sample of older adults with MDD and a high prevalence of MetSyn from three sites in North America, lending greater generalizability to the study’s findings.
In conclusion, we report an association between MetSyn and reduced antidepressant response among elderly persons with MDD, which appears to be accounted for by greater symptom severity and duration of depression at baseline. Randomized, prospective depression treatment studies measuring both metabolic and depressive outcomes are indicated to better understand the causal relationship between MetSyn, depression severity, and treatment response. Among individual MetSyn components, DBP remained a significant predictor of treatment outcome after multivariate adjustment. The results of this study should be interpreted with caution. We do not recommend that clinicians treat older adults with depression differently in the face of co-occuring MetSyn. Further verification of these findings is indicated using more reliable markers of insulin resistance, with other classes of antidepressants in racially diverse populations in the hopes of better informing treatment decisions in LLD.
Supplementary Material
Impact Statement.
We certify that this work is novel or confirmatory of recent novel clinical research. The potential impact of this research on clinical care or health policy includes the following: In a large sample of older adults, the metabolic syndrome was associated with greater disease severity and reduced treatment response in late-life depression. Further study in larger and more diverse populations, using sensitive metabolic testing – particularly for insulin resistance – is needed to validate these results, which could lead to changes in treatment guidelines for late-life depression.
Acknowledgments
Funding Sources: The parent study is registered with ClinicalTrials.gov, number NCT00892047. R01 MH083648 to Washington University, and R01 MH083643 to University of Toronto) with additional funding provided by the UPMC Endowment in Geriatric Psychiatry, the Taylor Family Institute for Innovative Psychiatric Research (at Washington University), the Washington University Institute of Clinical and Translational Sciences grant (UL1 TR000448) from the National Center for Advancing Translational Sciences (NCATS), and the Campbell Family Mental Health Research Institute at the Centre for Addiction and Mental Health, Toronto. Bristol-Myers Squibb contributed aripiprazole and placebo tablets, and Pfizer contributed venlafaxine extended-release capsules for the parent study.
The authors would like to thank the participants, clinical research staff, and data and safety monitoring board members of the original IRL GRey study. We also thank Phil Miller for his assistance with data analysis techniques. The parent study was supported primarily by the National Institute of Mental Health (R01 MH083660 and P30 MH90333 to University of Pittsburgh.
Footnotes
Additional baseline participant characteristics by MetSyn status
DATA AVAILABILITY
The dataset analyzed for purposes of this study is available from the corresponding author on reasonable request.
AUTHOR CONTRIBUTIONS
JSM was the primary author of this manuscript, and was responsible for the study design, including oversight and interpretation of statistical analyses. GEN provided assistance in the developing the study design, in the interpretation of results and in developing the visual representation of the data, as well as contributing substantially to all aspects of manuscript content. DD provided statistical support, including expertise in the use of SAS to conduct analyses and in the creation of tables and figures. EJL was the principal investigator of the parent study, and provided overall guidance and intellectual support in developing the study hypotheses, analytic stragtegy, interpretation of results, and contributing to all aspects of manuscript development. JFK, CFR and DMB were all primary investigators of the original IRL-Grey study, and provided expertise in conducting the original study, providing background information regarding study conduct for the methods section of the manuscript, assistance in the interpretation of results and in crafting the introduction and discussion sections of the manuscript.
CONFLICT OF INTEREST
JSM has no conflicts of interest to report. GEN receives research support from Otsuka America, Inc. for an investigator-initiated study. DD has no conflicts of interest to report. EJL has received support from NIH, FDA, Takeda, Lundbeck, Janssen, Alkermes, Taylor Family Institute for Innovative Psychiatric Research, McKnight Brain Research Foundation, Barnes/Jewish Foundation, and the Patient-Centered Outcomes Research Institute (PCORI). JFK has received medication supplies from Indivior and Pfizer within the past five years for investigator-initiated studies. CFR has no conflicts of interest to report. DMB receives research support from the Canadian Institutes of Health Research (CIHR), Brain Canada, Weston Brain Institute, National Institutes of Health (NIH), Temerty Family through the Centre for Addiction and Mental Health (CAMH) Foundation and the Campbell Family Research Institute. He received non-salary operating funds and in-kind equipment support from Brainsway Ltd. for an investigator-initiated study. He is the site-principal investigator for three sponsor-initiated clinical trials from Brainsway Ltd. He received in-kind equipment support from Tonika/Magventure for an investigator-initiated study. He received medication supplies from Indivior for an investigator-initiated trial. BHM currently receives research support from Brain Canada, the Canadian Institutes of Health Research, the CAMH Foundation, the Patient-Centered Outcomes Research Institute (PCORI), the US National Institute of Health (NIH), Eli Lilly (medications for a NIH-funded clinical trial), Pfizer (medications for a NIH-funded clinical trial), Capital Solution Design LLC (software used in a study founded by CAMH Foundation), and HAPPYneuron (software used in a study founded by Brain Canada). Within the past three years he has also received research support from Bristol-Myers Squibb (medications for a NIH-funded clinical trial) and Pfizer/Wyeth (medications for a NIH-funded clinical trial). He directly own stocks of General Electric (less than $5,000).
References
- 1.Ferrari AJ, Somerville AJ, Baxter AJ, et al. Global variation in the prevalence and incidence of major depressive disorder: A systematic review of the epidemiological literature. Psychol Med. 2013;43:471–81. doi: 10.1017/S0033291712001511. [DOI] [PubMed] [Google Scholar]
- 2.Kemp AH, Gordon E, Rush AJ, Williams LM. Improving the prediction of treatment response in depression: Integration of clinical, cognitive, psychophysiological, neuroimaging, and genetic measures. CNS spectr. 2008;13:1066–86. doi: 10.1017/s1092852900017120. [DOI] [PubMed] [Google Scholar]
- 3.Rush AJ, Trivedi MH, Wisniewski SR, et al. Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: A STAR* D report. Am J Psychiatry. 2006;163:1905–17. doi: 10.1176/ajp.2006.163.11.1905. [DOI] [PubMed] [Google Scholar]
- 4.Zahodne LB, Stern Y, Manly JJ. Depressive symptoms precede memory decline, but not vice versa, in non-demented older adults. J Am Geriatr Soc. 2014;62:130–4. doi: 10.1111/jgs.12600. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Shao H, Mohammed MU, Thomas N, et al. Evaluating excessive burden of depression on health status and health care utilization among patients with hypertension in a nationally representative sample from the Medial Expenditure Panel Survey (MEPS 2012) J Nerv Ment Dis. 2017;205:397–404. doi: 10.1097/NMD.0000000000000618. [DOI] [PubMed] [Google Scholar]
- 6.Callahan CM, Wolinsky FD, Stump TE, Nienaber NA, Hui SL, Tierney WM. Mortality, Symptoms, and Functional Impairment in Late-Life Depression. J Gen Intern Med. 1998;13:746–52. doi: 10.1046/j.1525-1497.1998.00226.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Barry LC, Allore HG, Bruce ML, Gill TM. Longitudinal association between depressive symptoms and disability burden among older persons. J Gerontol A Biol Sci Med Sci. 2009;64:1325–32. doi: 10.1093/gerona/glp135. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Laursen TM, Musliner KL, Benros ME, Vestergaard M, Munk-Olsen T. Mortality and life expectancy in persons with severe unipolar depression. J Affect Disord. 2016;193:203–7. doi: 10.1016/j.jad.2015.12.067. [DOI] [PubMed] [Google Scholar]
- 9.Taylor WD, Aizenstein HJ, Alexopoulos GS. The vascular depression hypothesis: mechanisms linking vascular disease with depression. Mol Psychiatry. 2013;18:963–74. doi: 10.1038/mp.2013.20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Aizenstein HJ, Baskys A, Boldrini M, et al. Vascular depression consensus report – a critical update. BMC Med. 2016;14:161. doi: 10.1186/s12916-016-0720-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Marshe VS, Maciukiewicz M, Rej S, et al. Norepinephrine transporter gene variants and remission from depression with venlafaxine treatment in older adults. Am J Psychiatry. 2017;174:468–75. doi: 10.1176/appi.ajp.2016.16050617. [DOI] [PubMed] [Google Scholar]
- 12.Rutherford BR, Taylor WD, Brown PJ, Sneed JR, Roose S. Biological aging and the future of geriatric psychiatry. J Gerontol A Biol Sci Med Sci. 2016;72:343–52. doi: 10.1093/gerona/glw241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Nelson JC, Delucchi K, Schneider LS. Anxiety does not predict response to antidepressant treatment in late life depression: results of a meta-analysis. Int J Geriatr Psychiatry. 2009;24:539–44. doi: 10.1002/gps.2233. [DOI] [PubMed] [Google Scholar]
- 14.Gallo JJ, Hwang S, Joo JH, et al. Multimorbidity, depression, and mortality in primary care: Randomized clinical trial of an evidence-based depression care management program on mortality risk. J Gen Intern Med. 2016;31:380–6. doi: 10.1007/s11606-015-3524-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Alberti KG, Zimmet P, Shaw J. Metabolic syndrome—a new world-wide definition. A Consensus Statement from the International Diabetes Federation. Diabet Med. 2006;23:469–80. doi: 10.1111/j.1464-5491.2006.01858.x. [DOI] [PubMed] [Google Scholar]
- 16.Pan A, Keum N, Okereke OI, et al. Bidirectional association between depression and metabolic syndrome. Diabetes care. 2012;35:1171–80. doi: 10.2337/dc11-2055. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Mansur RB, Brietzke E, McIntyre RS. Is there a “metabolic-mood syndrome”? A review of the relationship between obesity and mood disorders. Neurosci Biobehav Rev. 2015;52:89–104. doi: 10.1016/j.neubiorev.2014.12.017. [DOI] [PubMed] [Google Scholar]
- 18.Sagud M, Mihaljevic-Peles A, Uzun S, et al. The lack of association between components of metabolic syndrome and treatment resistance in depression. Psychopharmacology (Berl) 2013;230:15–21. doi: 10.1007/s00213-013-3085-x. [DOI] [PubMed] [Google Scholar]
- 19.Vogelzangs N, Beekman AT, van Reedt Dortland AK, et al. Inflammatory and metabolic dysregulation and the 2-year course of depressive disorders in antidepressant users. Neuropsychopharmacology. 2014;39:1624–34. doi: 10.1038/npp.2014.9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Woo YS, McIntyre RS, Kim JB, et al. Association of treatment response with obesity and other metabolic risk factors in adults with depressive disorders: Results from a National Depression Cohort study in Korea (the CRESCEND study) J Affect Disord. 2016;203:190–8. doi: 10.1016/j.jad.2016.06.018. [DOI] [PubMed] [Google Scholar]
- 21.Hryhorczuk C, Sharma S, Fulton SE. Metabolic disturbances connecting obesity and depression. Front Neurosci. 2013;7:177. doi: 10.3389/fnins.2013.00177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Almas A, Forsell Y, Iqbal R, Janszky I, Moller J. Severity of depression, anxious distress and the risk of cardiovascular disease in a Swedish population-based cohort. PloS One. 2015;10:e0140742. doi: 10.1371/journal.pone.0140742. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Daskalopoulou M, George J, Walters K, et al. Depression as a risk factor for the initial presentation of twelve cardiac, cerebrovascular, and peripheral arterial diseases: data linkage study of 1.9 million women and men. PloS One. 2016;11:e0153838. doi: 10.1371/journal.pone.0153838. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Vogelzangs N, Beekman AT, Boelhouwer IG, et al. Metabolic depression: a chronic depressive subtype? Findings from the InCHIANTI study of older persons. J Clin Psychiatry. 2011;72:598–604. doi: 10.4088/JCP.10m06559. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Marijnissen RM, Vogelzangs N, Mulder ME, van den Brink RH, Comijs HC, Voshaar RO. Metabolic dysregulation and late-life depression: a prospective study. Psychol Med. 2017;47:1041–52. doi: 10.1017/S0033291716003196. [DOI] [PubMed] [Google Scholar]
- 26.Lenze EJ, Mulsant BH, Blumberger DM, et al. Efficacy, safety, and tolerability of augmentation pharmacotherapy with aripiprazole for treatment-resistant depression in late life: a randomised, double-blind, placebo-controlled trial. Lancet. 2015;386:2404–12. doi: 10.1016/S0140-6736(15)00308-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Moulton CD, Pickup JC, Ismail K. The link between depression and diabetes: the search for shared mechanisms. Lancet Diabetes Endocrinol. 2015;3:461–71. doi: 10.1016/S2213-8587(15)00134-5. [DOI] [PubMed] [Google Scholar]
- 28.Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12:189–98. doi: 10.1016/0022-3956(75)90026-6. [DOI] [PubMed] [Google Scholar]
- 29.Montgomery SA, Åsberg MA. A new depression scale designed to be sensitive to change. Br J Psychiatry. 1979;134:382–9. doi: 10.1192/bjp.134.4.382. [DOI] [PubMed] [Google Scholar]
- 30.Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and β-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28:412–9. doi: 10.1007/BF00280883. [DOI] [PubMed] [Google Scholar]
- 31.Randolph C, Tierney MC, Mohr E, Chase TN. The Repeatable Battery for the Assessment of Neuropsychological Status (RBANS): Preliminary clinical validity. J Clin Exp Neuropsychol. 1998;20:310–9. doi: 10.1076/jcen.20.3.310.823. [DOI] [PubMed] [Google Scholar]
- 32.Sackeim HA. The definition and meaning of treatment-resistant depression. J Clin Psychiatry. 2001;62:10–7. [PubMed] [Google Scholar]
- 33.Waterman L, Stahl ST, Buysse DJ, et al. Self-reported obstructive sleep apnea is associated with nonresponse to antidepressant pharmacotherapy in late-life depression. Depress Anxiety. 2016;33:1107–13. doi: 10.1002/da.22555. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry. 1960;23:56–62. doi: 10.1136/jnnp.23.1.56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Miller MD, Paradis CF, Houck PR, et al. Rating chronic medical illness burden in geropsychiatric practice and research: Application of the Cumulative Illness Rating Scale. Psychiatry Res. 1992;41:237–48. doi: 10.1016/0165-1781(92)90005-n. [DOI] [PubMed] [Google Scholar]
- 36.Krishnan KR, Taylor WD, McQuoid DR, et al. Clinical characteristics of magnetic resonance imaging-defined subcortical ischemic depression. Biol Psychiatry. 2004;55:390–7. doi: 10.1016/j.biopsych.2003.08.014. [DOI] [PubMed] [Google Scholar]
- 37.Long J, Duan G, Tian W, et al. Hypertension and risk of depression in the elderly: A meta-analysis of prospective cohort studies. J Hum Hypertens. 2015;29:478–82. doi: 10.1038/jhh.2014.112. [DOI] [PubMed] [Google Scholar]
- 38.Viscogliosi G, Andreozzi P, Manzon L, Ettorre E, Cacciafesta M. Metabolic syndrome, executive dysfunction, and late-onset depression: Just a matter of white matter? J Am Geriatr Soc. 2014;62:1607–9. doi: 10.1111/jgs.12960. [DOI] [PubMed] [Google Scholar]
- 39.Jacobson L. Hypothalamic-pituitary-adrenocortical axis: Neuropsychiatric aspects. Compr Physiol. 2014;4:715–38. doi: 10.1002/cphy.c130036. [DOI] [PubMed] [Google Scholar]
- 40.van Reedt Dortland AK, Giltay EJ, van Veen T, Zitman FG, Penninx BW. Metabolic syndrome abnormalities are associated with severity of anxiety and depression and with tricyclic antidepressant use. Acta Psychiatr Scand. 2010;122:30–9. doi: 10.1111/j.1600-0447.2010.01565.x. [DOI] [PubMed] [Google Scholar]
- 41.Sekita A, Arima H, Ninomiya T, et al. Elevated depressive symptoms in metabolic syndrome in a general population of Japanese men: A cross-sectional study. BMC Public Health. 2013;13:862. doi: 10.1186/1471-2458-13-862. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Hicken MT, Lee H, Mezuk B, Kershaw KN, Rafferty J, Jackson JS. Racial and ethnic differences in the association between obesity and depression in women. J Womens Health (Larchmt) 2013;22:445–52. doi: 10.1089/jwh.2012.4111. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Xiang X, An R. Obesity and onset of depression among U.S. middle-aged and older adults. J Psychosom Res. 2015;78:242–8. doi: 10.1016/j.jpsychores.2014.12.008. [DOI] [PubMed] [Google Scholar]
- 44.Thorlund K, Druyts E, Wu P, Balijepalli C, Keohane D, Mills E. Comparative efficacy and safety of selective serotonin reuptake inhibitors and serotonin-norepinephrine reuptake inhibitors in older adults: A network meta-analysis. J Am Geriatr Soc. 2015;63:1002–9. doi: 10.1111/jgs.13395. [DOI] [PubMed] [Google Scholar]
- 45.Beck AT, Kovacs M, Weissman A. Assessment of suicidal intention: The Scale for Suicide Ideation. J Consult Clin Psychol. 1979;47:343–52. doi: 10.1037//0022-006x.47.2.343. [DOI] [PubMed] [Google Scholar]
- 46.Lorenzo C, Haffner SM, Stancakova A, Laasko M. Relation of Direct and Surrogate Measures of Insulin Resistance to Cardiovascular Risk Factors in Nondiabetic Finnish Offspring of Type 2 Diabetic Individuals. J Clin Endocrinol Metab. 2010;95:5082–5090. doi: 10.1210/jc.2010-1144. [DOI] [PubMed] [Google Scholar]
- 47.Chang AM, Smith MJ, Bloen CJ, Galecki AT, Halter JB, Supiano MA. Limitation of the Homeostatis Model Assessment to PredictInsulin Resistance and B-Cell Dysfunction in Older People. J Clin Endocrinol Metab. 2006;91:629–634. doi: 10.1210/jc.2005-1803. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.