Metabolic Syndrome and Symptom Resolution in Depression: A 5-Year Follow-Up of Older Adults

This work may not be copied, distributed, displayed, published, reproduced, transmitted, modified, posted, sold, licensed, or used for commercial purposes. By downloading this file, you are agreeing to the publisher’s Terms & Conditions.

Metabolic Syndrome and Symptom Resolution in Depression:

A 5-Year Follow-Up of Older Adults

Marianna Virtanen, PhDa,*; Jane E. Ferrie, PhDb,c; Tasnime Akbaraly, PhDb,d,e,f; Adam Tabak, MDb,g; Markus Jokela, PhDh; Klaus P. Ebmeier, MDi; Archana Singh-Manoux, PhDb,j; and Mika Kivimäki, PhDb,k

Depressive symptoms are common at older ages, affecting up to 20% of adults over 65 years.1,2 The adverse impact of depression on people and society is considerable in terms of human suffering, reduced quality of life, increased health care costs, and premature mortality.3-6 Depression tends to have an episodic course characterized by the reversal of symptoms leading to full recovery.7 However, recurrent episodes are common,8 and the risk of chronicity is high.9

Metabolic syndrome, a cluster of risk factors thought to predispose an individual to cardiovascular disease,10 is characterized by elevated abdominal obesity, high level of triglycerides, low level of high-density lipoprotein (HDL) cholesterol, high blood pressure, and elevated fasting glucose or diabetes. There is some evidence to suggest that metabolic syndrome is associated with an increased risk of depression,11,12 although the relationship between components of metabolic syndrome and the onset of depression is inconsistent.13-15

Research on the biological risk factors of depression has tended to focus on first episodes, not distinguishing between episodes.8 Few previous studies have examined the contribution of metabolic syndrome or its components to the course of disease in depression.7 Small-scale studies of patient populations have analyzed lipid components of metabolic syndrome and found high HDL cholesterol16,17 and low triglyceride levels16 to be associated with better prognosis of depression, whereas low HDL cholesterol was associated with longer symptom duration18 and high triglycerides with deliberate self-harm among acute psychiatric patients.19 In 1 study20 of 168 participants with depression, metabolic syndrome was associated with chronic or repeated depressive symptoms over time.

To gain further insight into the potential prognostic role of cardiovascular and metabolic risk factors in depression, we examined the associations of metabolic syndrome and its components with symptom resolution in a large cohort of older adults.

METHODS

Participants and Procedure

Data were drawn from the Whitehall II study,21 which was established in 1985 and included all London-based civil servants, aged 35-55 years, working in 20 departments. Ethical approval for the Whitehall II study was obtained from the University College London Medical School Committee on the Ethics of Human Research; all participants provided written informed consent. The Center for Epidemiologic Studies Depression scale (CES-D)22 was first introduced to the study at phase 7 (2002-2004), which forms the baseline of the first cycle of the present study (see the sample selection procedure in Figure 1). The data were structured into 2 separate longitudinal data cycles so that the first data cycle spanned from 2002-2004 to 2007-2009 and the second data cycle spanned from 2007-2009 to 2012-2013. Data were based on a total of 965 individuals (590 men, 375 women), all of whom had depressive symptoms at the cycle baseline. Of these 965 individuals, 410 contributed only to cycle 1, 348 only to cycle 2, and 207 to both cycles. Altogether they produced 1,172 person-observations (Figure 1), among which 587 (50.1%) reported symptom resolution and 585 (49.9%) reported recurrent or chronic depressive symptoms at follow-up. Mean (SD) follow-up time was 4.63 years (0.59): 5.08 (0.33) in cycle 1 and 4.12 (0.35) in cycle 2.

Depressive Symptoms

The CES-D22 is a 20-item self-administered scale designed to measure depressive symptoms in the general population. Participants score the frequency with which specific symptoms occurred during the previous week on a 4-point scale (0 = less than 1 day, 1 = 1-2 days, 2 = 3-4 days, and 3 = 5-7 days). These were summed to yield a total score between 0 and 60. Participants scoring 16 or higher are defined as cases of CES-D depressive symptoms. Symptom resolution is denoted as being a case at baseline but not at follow-up. Recurrent or chronic cases were those who reported depressive symptoms at both baseline and follow-up. Severity of depressive symptoms at baseline was defined according to tertiles of the CES-D symptom score distribution: mild = 16-18 points, moderate = 19-23 points, and severe = greater than 23 points.

Metabolic Syndrome

Metabolic syndrome was assessed at the baseline clinical examination of the study cycle and was defined according to the American Heart Association/National Heart, Lung, and Blood Institute 2009 Joint Scientific Statement23,24 as the presence of 3 or more of the following components: (1) large waist circumference (measured at the study clinic visit): ≥ 102 cm for men (≥ 90 cm for Asian men), ≥ 88 cm for women (≥ 80 cm for Asian women); (2) high triglyceride level: ≥ 150 mg/dL (≥ 1.69 mmol/L); (3) low HDL cholesterol level: < 40 mg/dL (< 1.03 mmol/L) for men, < 50 mg/dL (< 1.29 mmol/L) for women; (4) high blood pressure: ≥ 130 mm Hg systolic blood pressure, ≥ 85 mm Hg diastolic blood pressure, or receipt of antihypertensive medication; and (5) high fasting glucose (≥ 100 mg/dL [≥ 5.56 mmol/L]) or diabetes diagnosis.

Diabetes was defined as a fasting glucose ≥ 7.0 mmol/L or a 2-hour postload glucose ≥ 11.1 mmol/L on the oral glucose tolerance test performed during the Whitehall II study clinical screening or as physician-diagnosed diabetes or use of diabetes medication.25

Covariates

Age, sex, socioeconomic status (based on occupational grade categorized as high, intermediate, and low),21 and ethnicity (white, nonwhite [South Asian, black, other]) were based on survey responses.

Information on comorbid long-standing illnesses was obtained from several data sources: prevalent coronary heart disease was based on data collected at the baseline clinical screening for each study cycle and included definite angina and nonfatal myocardial infarction. Definite angina was identified via a questionnaire and was corroborated with medical records, abnormalities in a resting electrocardiogram (ECG), an exercise ECG, or a coronary angiogram. Nonfatal myocardial infarction was defined following the World Health Organization MONICA (Multinational Monitoring of Trends and Determinants in Cardiovascular Disease) criteria26 and confirmed using data from medical examinations, hospital records of ECGs, and use of cardiac enzymes. Stroke was based on self-reported, doctor-diagnosed stroke, which was further validated by information from the National Health Service (NHS) Hospital Episode Statistics database in England, by manual searches of hospital records, and by letters from general practitioners.27 Data on prevalent long-standing illnesses other than those described above were based on a list of treated long-standing illnesses during the past 12 months reported in the survey questionnaire and transformed into ICD-10 diagnostic codes (A15-A28, A80-A99, B15-B24, B50-B60, B64, C, D00-D09, D37-D48, D50-D53, D55-D59, D60-D77, D80-D84, D86, D89, E00-E07, E20-E32, E34-E35, E40-E46, F00-F29, F30-F31, F38-F99, G, H53-H54, H90-H91, I00-I02, I05-I09, I26-I28, I30-I52, I70-I74, I77-I79, I95-I99, J30-J32, J37, J40-J47, J60-J70, J80-J81, J90-J99, K20-K21, K25-K29, K31, K50-K52, K55-K63, K65-K67, K70-K77, K80-K83, K85-K87, L10-L14, L20-L22, L24-L30, L40-L45, L50-L54, M00-M99, N00-N29, N41, N45-N46, N70-N77, Q00-Q07, Q20-Q28, R51-R52, S, T). Prevalence of these comorbidities and prevalence of diabetes (as described previously) were combined to form the variable prevalent comorbid long-standing illness (yes/no). For participants with diabetes or metabolic syndrome, comorbid long-standing illness did not include diabetes.

Body mass index was calculated from measurements of weight and height as kg/m2 and categorized as normal (< 25.0), overweight (25.0-29.9), and obese (≥ 30.0). Smoking (yes/no) and alcohol use (derived from the number of alcoholic drinks [spirits, wine, and beer] consumed in the last 7 days and classified as none; within the recommended limits [moderate and intermediate drinking], defined as up to 14/21 units for women/men; and above the recommended limits, defined as > 14/21 units)28 were based on survey responses. Antidepressant, antihypertensive, and lipid-lowering drug use (yes/no) were based on survey responses.

To examine whether the association between lipid levels and recurrent depressive symptoms was nonlinear,29 we further classified the lipid variables into 3 categories so the lower and higher cut points were the same as that used in the assessment of metabolic syndrome and the other cut point was set to the upper and lower tertiles of the distribution. Thus, the cut points were < 1.03, 1.03-1.60, and > 1.60 for HDL cholesterol among men; < 1.29, 1.29-2.0, and > 2.0 for HDL cholesterol among women; and ≤ 0.90, 0.91-1.68, and ≥ 1.69 for triglycerides in both sexes.

To examine robustness of the findings between high glucose level, diabetes, and resolution of depressive symptoms, we used an alternative formulation of prediabetes in a sensitivity analysis. This was defined as impaired fasting glucose (a fasting glucose between 6.1 and < 7.0 mmol/L and 2-hour glucose < 7.8 mmol/L), impaired glucose tolerance (a fasting glucose < 7 mmol/L and a 2-hour postload glucose between 7.8 and < 11.1 mmol/L), or both.30

Statistical Analyses

We used a repeated-measures Poisson regression analysis using generalized estimating equations (GEEs) with exchangeable correlation structure.31 The GEE takes into account the intraindividual correlation between measurements. Among persons with depressive symptoms at baseline, we first examined associations of metabolic syndrome and its components with the probability of having severe depressive symptoms or antidepressant use at baseline and then with the probability of resolution of depressive symptoms at follow-up. The GEE takes into account the nonindependence of the multiple observations from the same participants when estimating standard errors. Because the analyzed outcomes were common (probability ranging from 12% to 50%), we used log-binomial regression analysis to calculate relative risks (risk ratios) and their 95% confidence intervals (CIs). All models were adjusted for age, sex, socioeconomic status, severity of depressive symptoms, antidepressant use, and comorbid long-standing illness. Models examining metabolic syndrome, HDL cholesterol, and triglycerides were additionally adjusted for lipid-lowering drug use.

A sensitivity analysis was carried out in which participants with coronary heart disease or stroke were excluded from the analyses. In a second sensitivity analysis, we compared participants with metabolic syndrome to those without any long-standing illness to better control for the impact of other comorbid long-standing illness. In a third sensitivity analysis, we used 3-category variables of HDL cholesterol and triglycerides and tested for linearity by entering the 3-category variable into the model as a continuous variable. SAS version 9.4 (SAS, Cary, North Carolina) was used for all analyses.

RESULTS

Altogether, this study included 1,172 person-observations: 723 from men and 449 from women with depressive symptoms, with a mean ± SD age of 62.4 ± 6.6 years at cycle baselines (see Supplementary eTable 1). As expected, as participants aged, the prevalence of comorbid long-standing illness and hypertension increased from the cycle 1 baseline to the cycle 2 baseline, although a more favorable lipid profile was seen in cycle 2 than in cycle 1. Distribution of body mass index expressed as normal weight, overweight, and obese did not differ between cycles, whereas abdominal obesity tended to increase from cycle 1 to cycle 2. The prevalence of smoking was lower and the prevalence of antidepressant use was higher at the baseline for cycle 2 compared to the baseline for cycle 1.

Cross-sectional associations of metabolic syndrome and its components with severe depressive symptoms, as indexed by a CES-D symptom score > 23, and antidepressant use among participants with depression at baseline are shown in Table 1. Neither metabolic syndrome nor its components were significantly associated with severe depressive symptoms or antidepressant medication.

In the longitudinal analysis, metabolic syndrome was not associated with depressive symptom resolution at follow-up (Table 2). Of the components of metabolic syndrome, impaired fasting glucose or diabetes, abdominal obesity, and hypertension were not associated with symptom resolution. However, low HDL cholesterol (RR = 0.82 [95% CI, 0.68-1.00; P = .045] in the fully adjusted model) and high triglycerides (RR = 0.81 [95% CI, 0.70-0.95; P = .007]) were associated with a lower likelihood of symptom resolution at follow-up.

In a sensitivity analysis excluding participants with coronary heart disease or stroke (Table 3), again metabolic syndrome was not associated with symptom resolution after adjustment for all covariates. Also as before, low HDL cholesterol (RR = 0.77 [95% CI, 0.63-0.95; P = .015] after all adjustments) and high triglycerides (RR = 0.79; 95% CI, 0.67-0.94; P = .006) were associated with lower likelihood of symptom resolution at follow-up.

We also examined the association between metabolic syndrome and depressive symptom resolution after excluding participants with all other long-standing illnesses (n = 593 for the analytic sample). In this healthy subsample, again metabolic syndrome was not associated with symptom resolution at follow-up (RR = 0.93; 95% CI, 0.80-1.09; data not shown).

Supplementary eTable 2 shows that low HDL cholesterol (RR = 0.76 [95% CI, 0.62-0.94] for the lowest vs highest HDL group in the fully adjusted model) and high triglycerides (RR = 0.79 [95% CI, 0.67-0.92] for the highest vs lowest triglycerides group) were associated with lower probability of symptom resolution, and a hypothesized linear rather than nonlinear association was supported (for linear term, P = .004 [HDL cholesterol] and P = .003 [triglycerides]). These findings were replicated in a subgroup free of coronary heart disease and stroke (1,002 and 882 person-observations, respectively); the RR for the lowest versus highest HDL group was 0.73; 95% CI, 0.59-0.91; the RR for the highest versus lowest triglyceride group was 0.76; 95% CI, 0.64-0.91 (data not shown). Supplementary eTable 2 also shows that when prediabetes and diabetes were analyzed separately in comparison to those without either condition, neither was associated with symptom resolution. Of the covariates assessed at baseline, more severe depressive symptoms, antidepressant use, and age > 65 versus ≤ 65 years at baseline were associated with lower likelihood of symptom resolution.

DISCUSSION

In this large cohort of older adults, we examined the contribution of metabolic syndrome and its components to the course of depression over 5 years among participants who had depressive symptoms at baseline. We found no overall association between metabolic syndrome and resolution of depressive symptoms at follow-up. Of the components of metabolic syndrome, low levels of HDL cholesterol and high levels of triglycerides were associated with an 18% to 19% lower probability of being symptom-free at follow-up, while we found no evidence supporting an association with diabetes, elevated glucose levels, abdominal obesity, or hypertension. These findings were robust to adjustment for several confounding factors and they were replicated in a subgroup of participants free of coronary heart disease and stroke.

Previous studies on the etiology of depression have shown an association between metabolic syndrome and the onset of depression in initially nondepressed individuals.11,12 High HDL cholesterol (the favorable cholesterol)16,17 and low triglyceride levels16 were associated with better prognosis of depression, whereas low HDL cholesterol was associated with longer symptom duration18 and high triglycerides with deliberate self-harm19 in small-scale patient populations. One study20 of a nonclinical sample of 168 cases with depressive symptoms found that metabolic syndrome predicted chronic or repeated depressive symptoms over time. In general, low rather than high cholesterol levels have been associated with depression and suicide, but these finding are still inconclusive and most studies assessed only total cholesterol levels,32 although there is some evidence showing associations between low HDL cholesterol levels and suicide.32 A clear advantage of our study was the large number of observations that enabled a more accurate estimate of the association between metabolic syndrome and symptom resolution in depression. Our detailed analysis of the components of metabolic syndrome revealed that the null association was attributable to the lack of associations with glycemic status, abdominal obesity, and hypertension.

Although there is considerable evidence showing bidirectional associations between diabetes and depression, recent evidence suggests that depression-related biological changes, such as those related to the hypothalamic-pituitary-adrenal cortex and sympathetic nervous system as well as low-grade chronic inflammation, may not be consistently associated with type 2 diabetes.33 This raises the possibility that the co-occurrence of diabetes and depression is not due to a causal association. Our null finding on the prognostic role of hypertension is in line with a previous meta-analysis34 of etiologic studies showing no association between hypertension and the onset of depression in old age. However, it is noteworthy that, unlike our study, previous research on the relationship between metabolic syndrome, diabetes, hypertension, and depression has focused on the onset rather than the resolution of depressive symptoms.

Previous etiologic research has produced mixed evidence on the relationship between total cholesterol and mental disorders, with associations observed for both low and high cholesterol levels.15,29,32 Similarly, a recent meta-analysis34 found no significant association between dyslipidemia and the onset of depression. However, the meta-analysis used a composite index of dyslipidemia, which precludes direct comparison with our study. Our findings regarding the associations of HDL cholesterol and triglycerides with symptom resolution are consistent with previous prognostic studies16-19 conducted among small-scale patient populations, suggesting an association between adverse lipid levels and poor prognosis of depression. Our analyses were based on older adults within a community cohort, and the results were replicated in a subpopulation free of coronary heart disease and stroke. Our findings suggest that adverse levels of HDL cholesterol and triglycerides seem to slow recovery from depressive symptoms, even among a "healthy population," suggesting an independent association of these risk factors.

The potential pathways linking adverse lipid levels to poorer recovery from depression include cerebral arteriosclerosis through subcortical small-vessel brain lesions, endothelial dysfunction, homocysteine regulation, inflammation, and white-matter abnormalities in the brain.34-40 A specific role of HDL cholesterol is its anti-inflammatory effects in the body; it also reduces inflammation in the vascular epithelium and regresses the formation of atherosclerotic plaques.41 It is probably no coincidence that both low HDL cholesterol and high triglyceride levels were associated with reduced symptom resolution because they are highly correlated (in the present data, Pearson r = −0.39, P < .0001). Indeed, recent evidence from Mendelian randomization studies suggests that low HDL cholesterol might be a marker of raised triglyceride levels and that high triglyceride levels are causally linked to cardiovascular diseases through raised concentrations of remnant cholesterol and increased low-grade inflammation, foam-cell formation, and atherosclerotic plaques.42 Consistent with our findings, some evidence is suggestive of an association between high cholesterol levels and depression resistant to antidepressant treatment.29

There are a number of limitations to this study. As in all observational studies, causality cannot be established. More randomized controlled trials are needed to examine, for example, whether antidepressant medication augmented by lipid-lowering drugs reduces treatment-resistant depression. Second, as some unmeasured or imprecisely measured factors may lie behind the observed associations, mechanisms underlying these associations need further investigation. Third, we used the self-administered CES-D instrument to assess depressive symptoms instead of using the gold standard, structured interview-based method for the clinical diagnosis of depression. However, the CES-D has been proved to have good validity as a measure of depressive symptoms.43 In addition, we had no data on past history of depressive episodes. Finally, our participants were from white-collar occupations, which restricts the generalizability of the findings. Replication in large general population-based cohorts is needed.

Specific strengths of our study are a relatively large sample derived from a nonpatient population, coronary heart disease and diabetes measured using gold standard methods, inclusion of a large variety of confounding and mediating factors, and sensitivity analyses that tested the robustness of the findings.

In conclusion, this study adds evidence on the contribution of cardiovascular risk factors to prognosis of depressive symptoms among older adults. Individuals with low HDL cholesterol and high triglyceride levels are less likely to experience symptom resolution than individuals free of these risk factors and are therefore at an increased risk of developing recurrent or chronic depression. These data suggest that an adverse lipid profile rather than other components of metabolic syndrome may delay recovery from depression.

Submitted: September 17, 2015; accepted February 24, 2016.

Potential conflicts of interest: None.

Funding/support: The Whitehall II Study is supported by grants from the Medical Research Council (K013351); British Heart Foundation; National Heart, Lung, and Blood Institute, National Institutes of Health (NIH) (R01HL036310); and the National Institute of Aging, NIH (R01AG013196 and R01AG034454). Dr Virtanen is supported by the Academy of Finland (258598, 292824). Dr Ebmeier is supported by the Medical Research Council (G1001354), Dr Akbaraly by the National Heart, Lung, and Blood Institute (R01HL036310), and Dr Singh-Manoux by the National Institute of Aging, NIH (R01AG013196 and R01AG034454). Dr Kivimäki is supported by the Medical Research Council (K013351), the NIH (R01HL036310, R01AG034454), NordForsk (75021), and a professorial fellowship from the Economic and Social Research Council.

Role of the sponsor: The funders had no role in the conduct or publication of the study.

Acknowledgments: The authors thank all of the participating civil service departments and their welfare, personnel, and establishment officers; the British Occupational Health and Safety Agency; the British Council of Civil Service Unions; all participating civil servants in the Whitehall II study; and all members of the Whitehall II study team.

Additional information: The Whitehall II Study team comprises research scientists, statisticians, study coordinators, nurses, data managers, administrative assistants, and data entry staff, who make the study possible. Whitehall II data, protocols, and other metadata are available to bona fide researchers for research purposes. Please refer to the Whitehall II data sharing policy at http://www.ucl.ac.uk/whitehallII/data-sharing.

Supplementary material: See accompanying pages.

REFERENCES

1. Byers AL, Yaffe K, Covinsky KE, et al. High occurrence of mood and anxiety disorders among older adults: The National Comorbidity Survey Replication. Arch Gen Psychiatry. 2010;67(5):489-496. PubMed doi:10.1001/archgenpsychiatry.2010.35

2. Blazer DG. Depression in late life: review and commentary. J Gerontol A Biol Sci Med Sci. 2003;58(3):249-265. PubMed doi:10.1093/gerona/58.3.M249

3. Katon WJ, Lin E, Russo J, et al. Increased medical costs of a population-based sample of depressed elderly patients. Arch Gen Psychiatry. 2003;60(9):897-903. PubMed doi:10.1001/archpsyc.60.9.897

4. Blazer DG 2nd, Hybels CF. Origins of depression in later life. Psychol Med. 2005;35(9):1241-1252. PubMed doi:10.1017/S0033291705004411

5. Hamer M, Bates CJ, Mishra GD. Depression, physical function, and risk of mortality: National Diet and Nutrition Survey in adults older than 65 years. Am J Geriatr Psychiatry. 2011;19(1):72-78. PubMed doi:10.1097/JGP.0b013e3181df465e

6. Murray CJ, Vos T, Lozano R, et al. Disability-adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2197-2223. PubMed doi:10.1016/S0140-6736(12)61689-4

7. Richards D. Prevalence and clinical course of depression: a review. Clin Psychol Rev. 2011;31(7):1117-1125. PubMed doi:10.1016/j.cpr.2011.07.004

8. Robinson OJ, Sahakian BJ. Recurrence in major depressive disorder: a neurocognitive perspective. Psychol Med. 2008;38(3):315-318. PubMed doi:10.1017/S0033291707001249

9. Hardeveld F, Spijker J, De Graaf R, et al. Prevalence and predictors of recurrence of major depressive disorder in the adult population. Acta Psychiatr Scand. 2010;122(3):184-191. PubMed doi:10.1111/j.1600-0447.2009.01519.x

10. Mottillo S, Filion KB, Genest J, et al. The metabolic syndrome and cardiovascular risk: a systematic review and meta-analysis. J Am Coll Cardiol. 2010;56(14):1113-1132. PubMed doi:10.1016/j.jacc.2010.05.034

11. Pan A, Keum N, Okereke OI, et al. Bidirectional association between depression and metabolic syndrome: a systematic review and meta-analysis of epidemiological studies. Diabetes Care. 2012;35(5):1171-1180. PubMed doi:10.2337/dc11-2055

12. Nousen EK, Franco JG, Sullivan EL. Unraveling the mechanisms responsible for the comorbidity between metabolic syndrome and mental health disorders. Neuroendocrinology. 2013;98(4):254-266. PubMed doi:10.1159/000355632

13. Akbaraly TN, Kivimäki M, Brunner EJ, et al. Association between metabolic syndrome and depressive symptoms in middle-aged adults: results from the Whitehall II study. Diabetes Care. 2009;32(3):499-504. PubMed doi:10.2337/dc08-1358

14. Akbaraly TN, Ancelin ML, Jaussent I, et al. Metabolic syndrome and onset of depressive symptoms in the elderly: findings from the three-city study. Diabetes Care. 2011;34(4):904-909. PubMed doi:10.2337/dc10-1644

15. Almeida OP, Yeap BB, Hankey GJ, et al. HDL cholesterol and the risk of depression over 5 years. Mol Psychiatry. 2014;19(6):637-638. PubMed doi:10.1038/mp.2013.113

16. 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(7):1624-1634. PubMed doi:10.1038/npp.2014.9

17. Baldwin RC, Gallagley A, Gourlay M, et al. Prognosis of late life depression: a three-year cohort study of outcome and potential predictors. Int J Geriatr Psychiatry. 2006;21(1):57-63. PubMed doi:10.1002/gps.1424

18. Lehto SM, Niskanen L, Tolmunen T, et al. Low serum HDL-cholesterol levels are associated with long symptom duration in patients with major depressive disorder. Psychiatry Clin Neurosci. 2010;64(3):279-283. PubMed doi:10.1111/j.1440-1819.2010.02079.x

19. Roaldset JO, Linaker OM, Bjørkly S. Triglycerides as a biological marker of repeated re-hospitalization resulting from deliberate self-harm in acute psychiatry patients: a prospective observational study. BMC Psychiatry. 2014;14:54. PubMed doi:10.1186/1471-244X-14-54

20. 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(5):598-604. PubMed doi:10.4088/JCP.10m06559

21. Marmot M, Brunner E. Cohort Profile: the Whitehall II study. Int J Epidemiol. 2005;34(2):251-256. PubMed doi:10.1093/ije/dyh372

22. Radloff L. The CES-D scale: a self-report depression scale for research in the general population. Applied Psychological Measurement. 1977:1(3):385-401. doi:10.1177/014662167700100306

23. Alberti KG, Eckel RH, Grundy SM, et al; International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; International Association for the Study of Obesity. Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation. 2009;120(16):1640-1645. PubMed doi:10.1161/CIRCULATIONAHA.109.192644

24. Heiss G, Snyder ML, Teng Y, et al. Prevalence of metabolic syndrome among Hispanics/Latinos of diverse background: the Hispanic Community Health Study/Study of Latinos. Diabetes Care. 2014;37(8):2391-2399. PubMed doi:10.2337/dc13-2505

25. Tabák AG, Jokela M, Akbaraly TN, et al. Trajectories of glycaemia, insulin sensitivity, and insulin secretion before diagnosis of type 2 diabetes: an analysis from the Whitehall II study. Lancet. 2009;373(9682):2215-2221. PubMed doi:10.1016/S0140-6736(09)60619-X

26. Tunstall-Pedoe H, Kuulasmaa K, Amouyel P, et al. Myocardial infarction and coronary deaths in the World Health Organization MONICA Project: registration procedures, event rates, and case-fatality rates in 38 populations from 21 countries in four continents. Circulation. 1994;90(1):583-612. PubMed doi:10.1161/01.CIR.90.1.583

27. Britton A, Milne B, Butler T, et al. Validating self-reported strokes in a longitudinal UK cohort study (Whitehall II): extracting information from hospital medical records versus the Hospital Episode Statistics database. BMC Med Res Methodol. 2012;12:83. PubMed doi:10.1186/1471-2288-12-83

28. Heikkilä K, Nyberg ST, Fransson EI, et al; IPD-Work Consortium. Job strain and alcohol intake: a collaborative meta-analysis of individual-participant data from 140,000 men and women. PLoS One. 2012;7(7):e40101. PubMed doi:10.1371/journal.pone.0040101

29. Papakostas GI, Ongür D, Iosifescu DV, et al. Cholesterol in mood and anxiety disorders: review of the literature and new hypotheses. Eur Neuropsychopharmacol. 2004;14(2):135-142. PubMed doi:10.1016/S0924-977X(03)00099-3

30. World Health Organization, International Diabetes Foundation. Definition and Diagnosis of Diabetes Mellitus and Intermediate Hyperglycaemia: Report of a WHO/IDF Consultation. Geneva, Switzerland: World Health Organization; 2006.

31. Lipsitz SR, Kim K, Zhao L. Analysis of repeated categorical data using generalized estimating equations. Stat Med. 1994;13(11):1149-1163. PubMed doi:10.1002/sim.4780131106

32. De Berardis D, Marini S, Piersanti M, et al. The relationships between cholesterol and suicide: an update. ISRN Psychiatry. 2012;2012:387901. PubMed doi:10.5402/2012/387901

33. Tabák AG, Akbaraly TN, Batty GD, et al. Depression and type 2 diabetes: a causal association? Lancet Diabetes Endocrinol. 2014;2(3):236-245. PubMed doi:10.1016/S2213-8587(13)70139-6

34. Valkanova V, Ebmeier KP. Vascular risk factors and depression in later life: a systematic review and meta-analysis. Biol Psychiatry. 2013;73(5):406-413. PubMed doi:10.1016/j.biopsych.2012.10.028

35. Belmaker RH, Agam G. Major depressive disorder. N Engl J Med. 2008;358(1):55-68. PubMed doi:10.1056/NEJMra073096

36. Kivimäki M, Shipley MJ, Batty GD, et al. Long-term inflammation increases risk of common mental disorder: a cohort study. Mol Psychiatry. 2014;19(2):149-150. PubMed doi:10.1038/mp.2013.35

37. Virtanen M, Shipley MJ, Batty GD, et al. Interleukin-6 as a predictor of symptom resolution in psychological distress: a cohort study. Psychol Med. 2015;45(10):2137-2144. PubMed doi:10.1017/S0033291715000070

38. Taylor WD, Aizenstein HJ, Alexopoulos GS. The vascular depression hypothesis: mechanisms linking vascular disease with depression. Mol Psychiatry. 2013;18(9):963-974. PubMed doi:10.1038/mp.2013.20

39. Santos M, Kövari E, Hof PR, et al. The impact of vascular burden on late-life depression. Brain Res Rev. 2009;62(1):19-32. PubMed doi:10.1016/j.brainresrev.2009.08.003

40. Yates KF, Sweat V, Yau PL, et al. Impact of metabolic syndrome on cognition and brain: a selected review of the literature. Arterioscler Thromb Vasc Biol. 2012;32(9):2060-2067. PubMed doi:10.1161/ATVBAHA.112.252759

41. Murphy AJ, Chin-Dusting JP, Sviridov D, et al. The anti-inflammatory effects of high density lipoproteins. Curr Med Chem. 2009;16(6):667-675. PubMed doi:10.2174/092986709787458425

42. Nordestgaard BG, Varbo A. Triglycerides and cardiovascular disease. Lancet. 2014;384(9943):626-635. PubMed doi:10.1016/S0140-6736(14)61177-6

43. Head J, Stansfeld SA, Ebmeier KP, et al. Use of self-administered instruments to assess psychiatric disorders in older people: validity of the General Health Questionnaire, the Center for Epidemiologic Studies Depression Scale and the self-completion version of the revised Clinical Interview Schedule. Psychol Med. 2013;43(12):2649-2656. PubMed doi:10.1017/S0033291713000342