Share this article on:

Vascular Brain Disease and Depression in the Elderly

Ikram, M Arfana; Luijendijk, Hendrika J.a,b; Vernooij, Meike W.a,c; Hofman, Alberta; Niessen, Wiro J.c,d,e; van der Lugt, Aadc; Tiemeier, Henninga; Breteler, Monique M. B.a

doi: 10.1097/EDE.0b013e3181c1fa0d
Mental Illness: Brief Report

Background: Cross-sectional studies have shown an association between vascular brain disease and depression. Longitudinal data are scarce. In a population-based study we investigated this relationship both cross-sectionally and longitudinally.

Methods: Brain MRIs were administered to 479 persons aged 60–90 years at baseline (1995–1996). Brain atrophy, white matter lesions and brain infarcts are all markers of vascular brain disease. At baseline and at follow-up examinations, we also identified persons with depressive symptoms and syndromes using the Center for Epidemiological Studies Depression Scale and psychiatric interviews. Medical records were continuously monitored to identify incident depression. Follow-up was complete until October 2005.

Results: At baseline, 36 persons had depressive symptoms. Brain atrophy, white matter lesions, and infarcts were associated with presence of depressive symptoms. During follow-up, 92 persons developed depressive symptoms, 35 of whom were categorized as having depressive syndrome. There was no association of any MRI marker with incident depressive symptoms or syndromes.

Conclusions: Markers of vascular brain disease were associated with depression cross-sectionally. However, when these markers and risk of depression were assessed longitudinally, no relationship was found.


From the aDepartment of Epidemiology, Erasmus MC University Medical Center, Rotterdam, The Netherlands; bDepartment of Geriatric Psychiatry, Parnassia BAVO Group, Institution for Mental Health Care, Rotterdam, The Netherlands; Departments of cRadiology and dMedical Informatics, Erasmus MC University Medical Center, Rotterdam, The Netherlands; and eDepartment of Applied Sciences, Delft University of Technology, Delft, The Netherlands.

Submitted 10 October 2008; accepted 20 March 2009.

The Rotterdam Scan Study was financially supported by the Health Research and Development Council and the Netherlands Organization for Scientific Research (grants 918–46–615, 904–61–096, 904–61–133).

Supplemental digital content is available through direct URL citations in the HTML and PDF versions of this article (

Correspondence: Monique M. B. Breteler, Department of Epidemiology, Erasmus MC University Medical Center, PO Box 2040, 3000 CA, Rotterdam, The Netherlands. E-mail:

Vascular disease and depression in the elderly are closely related.1,2 This has fuelled the “vascular depression” hypothesis,3 which postulates a vascular basis for late-life depression.4–6 With magnetic resonance imaging (MRI), markers of vascular brain disease can be visualized, including white matter lesions, brain infarcts, and brain atrophy. Various cross-sectional studies have reported associations of MRI markers with depression.4,7–10 Longitudinal studies have investigated only the relationship of white matter lesions with depression, with inconsistent results.9,11–13 We investigated the relationship of several MRI markers of subclinical vascular brain disease both cross-sectionally (with prevalent depression), and longitudinally (with incident depression).

Back to Top | Article Outline


Study Population

This study is based on an age-stratified (60–90 years) sample of 563 participants from the population-based Rotterdam Study,14 who underwent multisequence brain MRI in 1995–1996.15 The institutional medical-ethics committee approved the study and all participants gave written informed consent.

Of the 563 participants, 73 did not complete the MRI examination due to claustrophobia or for technical reasons, and 4 persons did not undergo psychiatric assessment at baseline. Furthermore, of the 11 persons using antidepressants at baseline, 7 had no depressive symptoms. These 7 persons were excluded because we could not determine whether the indication for using this medication was still present. A total of 479 persons were available for analysis.

Back to Top | Article Outline

MRI Measures at Baseline

Image acquisition, classification algorithm, and validation steps have been described elsewhere.15 In summary, we used the k-nearest-neighbor classifier to classify voxels (volume pixels) into cerebrospinal fluid, gray matter, normal white matter, and white matter lesions. Using nonrigid transformation, noncerebral tissues were stripped. For measurement of lobar and deep central brain volumes, we created an atlas in which the lobes were labeled according to a slightly modified version of the segmentation protocol, as described by Bokde et al16 Subsequently, we used validated nonrigid transformation to transform this atlas to each brain. Brain infarcts were rated visually as focal hyperintensities on T2-weighted images, 3 mm in size or larger and with a corresponding prominent hypointensity on T1-weighted images.15

Back to Top | Article Outline

Assessment of Depression

The assessment of depression has been described elsewhere.17 At the baseline visit and during 3 follow-up rounds (1997–1999, 1999–2000, and 2000–2001), the Center for Epidemiological Studies Depression Scale (CES-D) was used as a screening tool with a cut-off of 16. Screen-positive individuals then underwent the Present State Examination18 to diagnose major depression, dysthymia, and minor depression. The response rate was 95% (437 of the 461 eligible) at the first follow-up round, 81% (336 of 414 eligible) at the second, and 81% (309 of 381 eligible) at the third. Moreover, medical and pharmacy records of participants (eg, hospital discharge letter, specialists' reports, and notes of general practitioners) were continuously monitored for depressive episodes and for start of antidepressants during the follow-up period, by automated linkage of the general practitioners' and pharmacists' records with the database. These data ensured virtually complete follow-up among care-seeking participants.

Depressive episodes were classified as depressive symptoms for persons whose screening depression scale was positive, who had at least one core symptom of depression recorded in medical files, or who started antidepressants (without documentation of clinical symptoms). Depressive symptoms were further classified as depressive syndrome if persons were diagnosed as suffering from major depression, minor depression, or dysthymia according to the psychiatric interview or medical files. Follow-up for incident depressive symptoms and syndromes was complete to 1 October 2005.

Back to Top | Article Outline


Covariates included education, smoking, blood pressure, diabetes mellitus, body-mass index, and intima-media thickness.15

Back to Top | Article Outline

Statistical Analysis

All volumes were expressed as percentage of intracranial volume. Total white matter was the sum of normal white matter and white matter lesions. White matter lesions (WML) were analyzed as ln(WML volume), because of the skewed untransformed measures.

We used logistic regression to calculate odds ratios for presence of depressive symptoms associated with brain imaging markers. We also performed cross-sectional analyses with linear regression using the CES-D score as a continuous variable.

For longitudinal analyses, we excluded persons with depressive symptoms at baseline (n = 36). We used Cox's proportional-hazards models to calculate hazard ratios (HR) and 95% confidence intervals (CIs) for incident depressive symptoms or syndromes associated with brain imaging markers. Persons were followed until onset of depressive symptoms or syndromes, loss to follow-up, or 1 October 2005, whichever came first.

Back to Top | Article Outline


Table 1 shows the baseline characteristics of the study population. Thirty-six persons had depressive symptoms, of whom 6 had a depressive syndrome.



The smaller the brain volume, the more likely persons were to have depressive symptoms in cross-sectional analyses (Table 2). At the lobar level, parietal and temporal lobe atrophy were associated with depressive symptoms. The likelihood of having depressive symptoms increased with increasing volume of white matter lesions, especially in the frontal lobe and deep central region, and with presence of brain infarcts. Parietal lobe atrophy and deep white matter lesions were also related to CES-D in the continuous analysis (Table 2). Numbers were too small to perform separate analyses for prevalent depressive syndromes.



In the longitudinal analysis, during 3373 person-years of follow-up (mean, 7.5 years) a total of 92 persons developed depressive symptoms, of whom 35 suffered from a depressive syndrome. Neither global nor lobar brain tissue volumes were associated with incident depressive symptoms or syndromes (Table 3). Furthermore, neither white matter lesions nor brain infarcts were associated with incident depressive symptoms or depressive syndromes. Additional adjustment for cardiovascular risk factors did not change the results. Results based solely on data from the follow-up rounds show a similar lack of association (eTable,



Back to Top | Article Outline


In this population-based cohort study of elderly persons we found that structural markers of vascular brain disease were cross-sectionally related to the prevalence of depressive symptoms. However, we did not find any association between structural brain markers and incident depressive symptoms or depressive syndromes.

Strengths of this study include the population-based setting, and the cross-sectional as well as longitudinal design of the study with more than 7 years of follow-up. Moreover, unlike previous studies, we investigated various markers of vascular brain disease using automated quantification techniques. A limitation is that we lacked reliable data on depression before baseline. Therefore, some persons may already have had depression before baseline. However, given that persons with a history of depression have an increased risk of recurrent depression, together with the strong cross-sectional association between vascular brain disease and depression, longitudinal analyses would overestimate any true effect. Another limitation is possible selection bias due to differential follow-up. If persons with vascular disease were more prone to seek medical help than others, estimates from longitudinal analyses could easily overestimate effects. Conversely, if these persons are less likely to report depressive symptoms, the HR would be underestimated.

A final consideration is that we excluded persons who used antidepressants at baseline. However, post hoc analyses including these persons in either the depressed or nondepressed group did not change the results.

Cross-sectional analyses showed brain atrophy, brain infarcts and white matter lesions to be related to depressive symptoms (including depressive syndromes), consistent with various previous studies.4,8,9 Furthermore, these results concur with published data showing that atrophy in the parietal and frontal lobes, and frontal and deep white matter lesions are particularly related to depression.4,10,19

In contrast, there was no longitudinal association between MRI markers and incident depression. Although we should be cautious in interpreting these as null associations because of the relatively wide confidence intervals, the HRs do not suggest any major effect. If anything, most are in a direction opposite to what is expected based on the vascular depression hypothesis, namely that larger brain volume was associated with a decreased risk of depression (HR below 1), and white matter lesions with an increased risk of depression (HR above 1).

Several explanations for null associations in our longitudinal analyses need to be considered. First, our study design implies a time delay between baseline vascular damage and onset of depression. However, it is possible that vascular injury leads directly to depressive symptoms without any delay. A second explanation might be that depression causes vascular brain disease. Based on only limited data on progression of white matter lesions, we did not find such associations in the Rotterdam Scan Study (data not shown). However, 2 studies have reported a larger increase in white matter lesions in depressed compared with nondepressed persons.9,12 Though it is not established how depression might lead to vascular brain disease and brain atrophy, possible mechanisms include platelet dysfunction, hypotensive episodes, unhealthy lifestyle choices, and elevated cortisol levels in the brain, which in turn can cause glucocorticoid-mediated neurotoxicity.1,5,20 However, it is unclear why depression would cause vascular disease only in specific areas in the brain, eg, frontal and deep central white matter lesions. Third, it is possible that vascular brain injury does not relate to incidence of first depression but to persistent, chronic, relapsing, or recurrent depression. To test these possibilities, future studies should seek to discern more clearly first-ever depressions from possibly recurrent events, and accurately establish the duration of the depressive episode. Finally, a common etiology (eg, genetic predisposition) could link depression with vascular brain disease cross-sectionally but not necessarily longitudinally. Indeed, a twin study showed that the co-occurrence of cardiovascular disease and depression is partly explained by common genetic risk factors.21

In conclusion, we found that MRI markers of vascular brain disease were strongly associated with depression cross-sectionally. However, our study emphasizes that a cross-sectional association does not necessarily demonstrate causation, as we found no evidence for the “vascular depression” hypothesis relating these brain markers to incident depression. Even so, more longitudinal studies are needed to precisely elucidate this hypothesis. Even a noncausal relation of vascular disease with depression might point toward possibilities for prevention and treatment that thus far have remained unexplored.

Back to Top | Article Outline


1. Thomas AJ, Kalaria RN, O'Brien JT. Depression and vascular disease: what is the relationship? J Affect Disord. 2004;79:81–95.
2. Rao R. Cerebrovascular disease and late life depression: an age old association revisited. Int J Geriatr Psychiatry. 2000;15:419–433.
3. Alexopoulos GS, Meyers BS, Young RC, et al. ‘Vascular depression’ hypothesis. Arch Gen Psychiatry. 1997;54:915–922.
4. Steffens DC, Helms MJ, Krishnan KR, Burke GL. Cerebrovascular disease and depression symptoms in the cardiovascular health study. Stroke. 1999;30:2159–2166.
5. Sheline YI. Neuroimaging studies of mood disorder effects on the brain. Biol Psychiatry. 2003;54:338–352.
6. Thomas AJ, Perry R, Barber R, Kalaria RN, O'Brien JT. Pathologies and pathological mechanisms for white matter hyperintensities in depression. Ann N Y Acad Sci. 2002;977:333–339.
7. Chen PS, McQuoid DR, Payne ME, Steffens DC. White matter and subcortical gray matter lesion volume changes and late-life depression outcome: a 4-year magnetic resonance imaging study. Int Psychogeriatr. 2006;18:445–456.
8. O'Brien J, Desmond P, Ames D, et al. A magnetic resonance imaging study of white matter lesions in depression and Alzheimer's disease. Br J Psychiatry. 1996;168:477–485.
9. Steffens DC, Krishnan KR, Crump C, Burke GL. Cerebrovascular disease and evolution of depressive symptoms in the cardiovascular health study. Stroke. 2002;33:1636–1644.
10. Almeida OP, Burton EJ, Ferrier N, McKeith IG, O'Brien JT. Depression with late onset is associated with right frontal lobe atrophy. Psychol Med. 2003;33:675–681.
11. Versluis CE, van der Mast RC, van Buchem MA, et al. Progression of cerebral white matter lesions is not associated with development of depressive symptoms in elderly subjects at risk of cardiovascular disease: The PROSPER Study. Int J Geriatr Psychiatry. 2006;21:375–381.
12. Godin O, Dufouil C, Maillard P, et al. White matter lesions as a predictor of depression in the elderly: the 3C-Dijon study. Biol Psychiatry. 2008;63:663–669.
13. Teodorczuk A, O'Brien JT, Firbank MJ, et al. White matter changes and late-life depressive symptoms: longitudinal study. Br J Psychiatry. 2007;191:212–217.
14. Hofman A, Breteler MM, van Duijn CM, et al. The Rotterdam Study: objectives and design update. Eur J Epidemiol. 2007;22:819–829.
15. Ikram MA, Vrooman HA, Vernooij MW, et al. Brain tissue volumes in the general elderly population. The Rotterdam Scan Study. Neurobiol Aging. 2008;29:882–890.
16. Bokde AL, Teipel SJ, Schwarz R, et al. Reliable manual segmentation of the frontal, parietal, temporal, and occipital lobes on magnetic resonance images of healthy subjects. Brain Res Brain Res Protoc. 2005;14:135–145.
17. Luijendijk HJ, van den Berg JF, Dekker MJ, et al. Incidence and recurrence of late-life depression. Arch Gen Psychiatry. 2008;65:1394–1401.
18. WHO. SCAN Schedules for Clinical Assessment in Neuropsychiatry, Version 2.1. Distribution From Training Centers. 2 ed. Geneva: World Health Organization; 1997.
19. Sheline YI, Price JL, Vaishnavi SN, et al. Regional white matter hyperintensity burden in automated segmentation distinguishes late-life depressed subjects from comparison subjects matched for vascular risk factors. Am J Psychiatry. 2008;165:524–532.
20. Pantoni L, Garcia JH. Pathogenesis of leukoaraiosis: a review. Stroke. 1997;28:652–659.
21. Scherrer JF, Xian H, Bucholz KK, et al. A twin study of depression symptoms, hypertension, and heart disease in middle-aged men. Psychosom Med. 2003;65:548–557.

Supplemental Digital Content

Back to Top | Article Outline
© 2010 Lippincott Williams & Wilkins, Inc.