Cerebrovascular Risk Factors and Depression in Older Primary Care Patients: Testing a Vascular Brain Disease Model of Depression

The authors examined whether cerebrovascular risk factors (CVRFs) are associated with depressive diagnoses and symptoms in 303 primary-care patients age ≥60 years, as would be consistent with a small-vessel brain disease model of later-life depression. CVRFs were not significantly independently associated with major, minor, or subsyndromal depression, late-onset major depression, or overall depressive symptom severity. These data did not support the notion that a small-vessel brain disease model of depression might apply to the majority of older persons with depressive symptoms and syndromes in primary-care settings. Future work should include longitudinal study with larger sample sizes. The authors examined whether cerebrovascular risk factors (CVRFs) are associated with depressive diagnoses and symptoms in 303 primary-care patients age ≥60 years, as would be consistent with a small-vessel brain disease model of later-life depression. CVRFs were not significantly independently associated with major, minor, or subsyndromal depression, late-onset major depression, or overall depressive symptom severity. These data did not support the notion that a small-vessel brain disease model of depression might apply to the majority of older persons with depressive symptoms and syndromes in primary-care settings. Future work should include longitudinal study with larger sample sizes. Depression in older adults is a major public health problem of unknown etiology.1NIH Consensus Development Panel on Depression in Late Life Diagnosis and treatment of depression in late life.JAMA. 1992; 268: 1018-1024Google Scholar As reviewed elsewhere,2Alexopoulos GS Meyers BS Young RC et al.“Vascular depression” hypothesis.Arch Gen Psychiatry. 1997; 54: 915-922Google Scholar a confluence of data from risk factor, neuroimaging, and neuropsychological studies supports a theoretical model in which small-vessel cerebrovascular disease contributes to the pathogenesis of depressions of later life (especially those of later age at onset). Given this context, the term “vascular depression” has been proposed to describe patients so affected.2Alexopoulos GS Meyers BS Young RC et al.“Vascular depression” hypothesis.Arch Gen Psychiatry. 1997; 54: 915-922Google Scholar, 3Alexopoulos GS Meyers BS Young RC et al.Clinically defined vascular depression.Am J Psychiatry. 1997; 154: 562-565Google Scholar, 4Krishnan KRR Hays JC Blazer DG MRI-defined vascular depression.Am J Psychiatry. 1997; 154: 497-501Google Scholar In addition to potential implications for differential therapeutics, this model could lead to strategies for primary or secondary prevention of later-life depression. However, further empirical support for this theoretical model is needed. Supportive evidence for the model might come from demonstration of an association between systemic conditions known to be risk factors for cerebrovascular disease (e. g., hypertension, atherosclerotic heart disease, and diabetes mellitus) and depression. In nonpsychiatric populations such cerebrovascular risk factors (CVRFs) are associated with hyperintensities on T2-weighted magnetic resonance imaging (MRI) brain scans.5Kumar A Miller D Ewbank D et al.Quantitative anatomic measures and comorbid medical illness in late-life major depression.Am J Geriatr Psychiatry. 1997; 5: 15-25Google Scholar, 6Coffey CE Figiel GS Djang WT et al.White-matter hyperintensity on magnetic resonance imaging: clinical and neuroanatomic correlates in depressed elderly.J Neuropsychiatry Clin Neurosci. 1989; 1: 135-144Google Scholar These hyperintensities have an increased prevalence among older depressed patients,6Coffey CE Figiel GS Djang WT et al.White-matter hyperintensity on magnetic resonance imaging: clinical and neuroanatomic correlates in depressed elderly.J Neuropsychiatry Clin Neurosci. 1989; 1: 135-144Google Scholar, 7Rabins PV Pearlson GD Aylward E et al.Cortical magnetic resonance imaging changes in elderly inpatients with major depression.Am J Psychiatry. 1991; 148: 617-620Google Scholar, 8Zubenko GS Sullivan P Nelson JP et al.Brain imaging abnormalities in mental disorders of late life.Arch Neurol. 1990; 47: 1107-1111Google Scholar, 9Coffey CE Wilkinson WE Weiner RD et al.Quantitative cerebral anatomy in depression: a controlled magnetic resonance imaging study.Arch Gen Psychiatry. 1993; 50: 7-16Google Scholar, 10Lesser IM Miller BL Boone KB et al.Brain injury and cognitive function in late-onset psychotic depression.J Neuropsychiatry Clin Neurosci. 1991; 3: 33-40Google Scholar, 11Figiel GS Krishnan KRR Doraiswamy PM et al.Subcortical hyperintensities on brain magnetic resonance imaging: a comparison between late age of onset and early-onset elderly depressed.Neurobiol Aging. 1991; 26: 245-247Google Scholar a finding that has been a principal part of the supporting evidence for the “vascular depression” model.2Alexopoulos GS Meyers BS Young RC et al.“Vascular depression” hypothesis.Arch Gen Psychiatry. 1997; 54: 915-922Google Scholar, 4Krishnan KRR Hays JC Blazer DG MRI-defined vascular depression.Am J Psychiatry. 1997; 154: 497-501Google Scholar However, results have been mixed regarding the association between CVRFs and MRI hyperintensities in depressed patients (see Lyness et al.12Lyness JM Caine ED Cox C et al.Cerebrovascular risk factors and later-life major depression: testing a small-vessel brain disease model.Am J Geriatr Psychiatry. 1998; 6: 5-13Google Scholar for a review). Also, controlled comparisons of prevalence or severity of CVRFs between depressed and nondepressed older patients, or between patients with depression of later and younger age at onset have been relatively few,5Kumar A Miller D Ewbank D et al.Quantitative anatomic measures and comorbid medical illness in late-life major depression.Am J Geriatr Psychiatry. 1997; 5: 15-25Google Scholar, 13Greenwald BS Kramer-Ginsberg E Krishnan KRR et al.MRI signal hyperintensities in geriatric depression.Am J Psychiatry. 1996; 153: 1212-1215Google Scholar, 14Baldwin RC Tomenson B Depression in later life: a comparison of symptoms and risk factors in early- and late-onset cases.Br J Psychiatry. 1995; 167: 649-652Google Scholar and all but one failed to demonstrate an association of CVRFs with depression. Indeed, a 1995 review of investigations using a broader range of patients concluded, “Studies evaluating the risk of cardiovascular disease in the psychiatrically ill yield mixed results.”15Hayward C Psychiatric illness and cardiovascular disease risk.Epidemiol Rev. 1995; 17: 129-138Google Scholar We reported that cumulative CVRF severity was not greater in older psychiatric inpatients with major depression as compared with normal community control subjects (even when limiting the depressed group to those of later age at onset), and was not associated with greater depressive symptom severity among the depressed group;12Lyness JM Caine ED Cox C et al.Cerebrovascular risk factors and later-life major depression: testing a small-vessel brain disease model.Am J Geriatr Psychiatry. 1998; 6: 5-13Google Scholar examining individual CVRFs, and controlling for medical disability, only atrial fibrillation was associated with depression-group membership. Furthermore, most of the studies cited used psychiatric patient populations, including those with especially severe, melancholic, psychotic, or treatment-refractory major depression. Thus, the generalizability of the model to the majority of older people suffering clinically significant depressions is unknown. To our knowledge, no published work has studied the relationship of CVRFs to depression among elderly patients in primary care. It is important to extend risk factor investigations to this group for two reasons. The first is the public health imperative: Most older people suffering clinically significant depressive syndromes and symptoms are never seen in psychiatric settings, yet they do see primary care physicians.16Katon W Schulberg H Epidemiology of depression in primary care.Gen Hosp Psychiatry. 1992; 14: 237-247Google Scholar Also, disease expression from an etiologic agent may occur along a biologic gradient akin to “dose–response.”17Evans AS Causation and disease: the Henle-Koch postulates revisited.Yale J Biol Med. 1976; 49: 175-195Google Scholar A study group derived from primary care allows examination of the relationship of CVRFs to depressions of varied severity, from major to minor depression, as well as “subsyndromal” symptoms,18Judd LL Paulus MP Wells KB et al.Socioeconomic burden of subsyndromal depressive symptoms and major depression in a sample of the general population.Am J Psychiatry. 1996; 153: 1411-1417Google Scholar and allows the use of both categorical (syndromically defined) diagnoses and dimensional symptom severity scales across a broad range. Therefore, we had two hypotheses. First, we tested the hypothesis that CVRFs are associated with depressive disorder diagnosis and with greater depressive symptom severity in older patients recruited from primary care settings. Second, we tested the hypothesis that CVRFs are independently associated with depressive diagnoses and symptoms after we controlled for several other factors known to be associated with depression in older community- or psychiatric-patient samples and that might influence the association of CVRFs with depression. These included overall medical burden,19Katz IR On the inseparability of mental and physical health in aged persons: lessons from depression and medical comorbidity.Am J Geriatr Psychiatry. 1996; 4: 1-16Google Scholar functional disability,20Williamson GM Schulz R Pain, activity restriction, and symptoms of depression among community-residing elderly adults.J Gerontol. 1992; 47: P367-P372Google Scholar pain,20Williamson GM Schulz R Pain, activity restriction, and symptoms of depression among community-residing elderly adults.J Gerontol. 1992; 47: P367-P372Google Scholar personality trait neuroticism,21Hays JC Landerman LR George LK et al.Social correlates of the dimensions of depression in the elderly.J Gerontol. 1998; 53B: P31-P39Google Scholar and social support.22George LK Social and economic factors related to psychiatric disorders in late life.in: Busse EW Blazer DG The American Psychiatric Press Textbook of Geriatric Psychiatry. 2nd Edition. American Psychiatric Press, Washington, DC1996: 129-154Google Scholar We will describe our recruitment procedures briefly, because they have been described in detail elsewhere.23Lyness JM Noel TK Cox C et al.Screening for depression in elderly primary care patients: a comparison of the Center for Epidemiologic Studies-Depression Scale and the Geriatric Depression Scale.Arch Intern Med. 1997; 157: 449-454Google Scholar Subjects were engaged in the study from three internal-medicine private offices (including nine board certified internists and one physician's assistant) (n=210) and a free-standing University-affiliated family medicine clinic (n=93). After receiving a complete description of the study, all patients age ≥60 years who gave informed consent (using formal verbal or written consent procedures approved by the University of Rochester Research Subjects Review Board) were eligible to participate. Patients were screened in the offices with the Center for Epidemiologic Studies Depression Scale (CES-D).24Radloff LS The CES-D scale: a self-report depression scale for research in the general population.Applied Psychological Measurement. 1992; 7: 343-351Google Scholar Stratified sampling on the CES-D was used to oversample patients with depressive symptoms for further research assessments; specifically, all patients scoring above a cutoff of 21 on the CES-D were approached for further interview, and a random selection of subjects scoring below the cutoff were approached, up to a maximum of three total subjects enrolled per week. Thus, the final study group included patients scoring both above and below this cutoff. The in-depth assessment, obtained within 1 month of the primary-care office visit, was based on the Structured Clinical Interview for DSM-III-R (SCID),25Spitzer RL Williams JBW Gibbon M Structured Clinical Interview for DSM-III-R (SCID). New York State Psychiatric Institute, Biometrics Research, New York1986Google Scholar which was administered by trained raters with at least a master's-degree level of education. The SCID was completed from an interview with the patient, confirmation with family informant when available, and chart review. Each case was presented at a consensus conference of raters and investigators, at which final diagnoses (including age at onset) were assigned on the basis of the SCID and review of all other available records, including the primary-care chart (reviewed by a physician-investigator JML). Although not included in the SCID, the diagnosis of dementia was assigned on the basis of the available clinical data and on the patient's performance on the Mini-Mental State Exam,26Folstein 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: 185-198Google Scholar by use of DSM-III-R criteria. A total of 12 subjects had dementia: 3 in the major-depression group, 3 in the subsyndromal depression group, and 6 in the nondepressed group. Determination of depressive symptoms in patients with dementia and other medical conditions used an inclusive approach, as recommended previously by our group and others.27Lyness JM Bruce ML Koenig HG et al.Depression and medical illness in late life: report of a symposium.J Am Geriatr Soc. 1996; 44: 198-203Google Scholar Therefore, the diagnostic category “organic mood disorders” was not used. No subjects were excluded from the study on the basis of medical comorbidity because exclusion of such comorbidity, including dementia, would arbitrarily reduce the actual variability of interest.27Lyness JM Bruce ML Koenig HG et al.Depression and medical illness in late life: report of a symposium.J Am Geriatr Soc. 1996; 44: 198-203Google Scholar For purposes of analysis, subjects were assigned to one of four depression groups: current major depression (n=31); current minor depression (using criteria from the appendix of DSM-IV; n=19); subsyndromal depression (defined as a score >10 on the 24-item Hamilton Rating Scale for Depression Ham-D28Williams JBW A structured interview guide for the Hamilton depression rating scale.Arch Gen Psychiatry. 1988; 45: 742-747Google Scholar and no major or minor depression (c. f., Lyness et al.29Lyness JM, King DA, Cox C, et al: The importance of subsyndromal depression in older primary care patients. J Am Geriatr Soc (in press)Google Scholar; n=39); and nondepressed (n=214). Reasoning that CVRFs may be specifically increased in depressions with a later age at onset, we conducted separate analyses examining the variables of interest in later-onset depression (defined as any history of major depression with an age at onset of ≥60 years; n=32) vs. the nondepressed group (n=211; note that this group is smaller than the nondepressed group above because three subjects with later-onset major depression in full remission were grouped with the later-onset depressed subjects). Although depressive conditions were the most prevalent disorders, other Axis I diagnoses were present and have been described elsewhere.30Lyness JM, Caine ED, King DA, et al: Psychiatric disorders in older primary care patients. J Gen Intern Med (in press)Google Scholar Other study measures obtained at the time of the SCID included two functional measures, the Instrumental Activities of Daily Living (IADL) and Physical Self-Maintenance Scale (PSMS);31Lawton MP Brody EM Assessment of older people: self-maintaining and instrumental activities of daily living.Gerontologist. 1969; 9: 179-186Google Scholar self-reported social support, from the Duke Social Support Inventory (DSSI; comprising three subscales: social interaction, perceived social support, and instrumental support32Landerman R George LK Campbell RT et al.Alternative models of the stress-buffering hypothesis.Am J Community Psychol. 1989; 17: 625-642Google Scholar); personality trait neuroticism from the Neuroticism factor of the NEO Five-Factor Inventory;33Costa PT McCrae RR The NEO Personality Inventory: Revised Professional Manual. Psychological Assessment Resources, Odessa, FL1992Google Scholar and self-reported pain, from the Medical Outcomes Study Short Form SF-36.34Ware Jr., JE Sherbourne CD The MOS 36-Item Short-Form Health Survey (SF-36), I: conceptual framework and item selection.Med Care. 1992; 30: 473-483Google Scholar Only 260 subjects completed the NEO, DSSI, and Medical measures were completed on the basis of both patient interview and physician-investigator review. CVRF severity was with the CVRF American Scholar a of the of CVRFs cardiovascular disease, diabetes atrial and based on the American criteria for risk factors for medical illness burden was on the Rating Scale BS Rating Am Geriatr Soc. Scholar a and scale that the of in the psychiatric item from the included models for the based on the because the data were in the of with a relatively P 2nd Edition. and Scholar Each model included an for In this the for the model and for the individual are of the P 2nd Edition. and Scholar were used to with depression group all study or with later-onset depression later-onset depression vs. nondepressed In variables used in models to the first hypothesis were variables and as a for and the CVRF To the we analyses that one of the and in (e. g., and and the three of the only the models are reported because the for the CVRF did not models were used to whether were in the prevalence of CVRFs in across the depressed variables were were used because of the analyses to reduce the of a I On the basis of the of the depression and the variability of the CVRF our study would have been to a in CVRF score of to 3 with at a level of between CVRF score and for CVRF P P pain, P pain, P P P pain, P pain, P depression vs. P P pain, P pain, P CVRF cerebrovascular risk Hamilton Rating Scale for Rating Scale medical Instrumental Activities of Daily Living Physical Self-Maintenance Scale Medical Outcomes Study Short Form Duke Social Support in a CVRF cerebrovascular risk Hamilton Rating Scale for Rating Scale medical Instrumental Activities of Daily Living Physical Self-Maintenance Scale Medical Outcomes Study Short Form Duke Social Support The age was 6 years, with a range of The CVRF score had a of with a range of The study group included and clinical data by depression are in and clinical at depression depression depression are Hamilton Rating Scale for Depression symptom CVRF cerebrovascular risk factor score cumulative risk factor Rating Scale medical in a are Hamilton Rating Scale for Depression symptom CVRF cerebrovascular risk factor score cumulative risk factor Rating Scale medical the for the association of CVRF score with of the Each a separate analysis, with the and The in the at the the association of the CVRF score with the controlling for the that the CVRF score was not significantly independently associated with any of the whether or not we controlled for the in our patients assigned to the depression” or had previously of major depression that were or is such subjects might influence the of interest might or the analyses using depression group as the were after all subjects with or depression. The results did not the CVRF score did not have a significant association with depression-group the for examining the later-onset depression group was that this group should include all later-onset depressed including those with or depression, so analyses were not 3 the analyses examining the association of individual CVRFs with depression. a separate and were for of the CVRFs was significantly independently associated with depression-group or later-onset CVRFs and depression, controlled for and P P P P P P depression vs. P P P P P P CVRF cerebrovascular risk in a CVRF cerebrovascular risk We did not our hypotheses. Therefore, our data did not support the notion that a small-vessel brain disease model of depression might apply to the majority of older persons suffering depressive symptoms and syndromes in primary care settings. group the sample in studies examining the vascular model of depression among patients with depression to psychiatric treatment settings. Therefore, our results a for other derived from patient of our be by of several of our study and be in a broader of the larger regarding the “vascular depression” First, our to demonstrate a significant association of CVRFs with depression might have been to Second, we reported but should be that the use of might be our a hypotheses. we used the several of the results would have a level of and our data might have been as support for the theoretical our data are which is a because the theoretical model that CVRFs lead to depression the of small-vessel brain disease Indeed, a study such as this factors such as the The model tested would that CVRFs are associated with later-onset depression, but medical (including are known to be associated with of depression. our to age at onset of depression was by patient the of of informant and conceptual related to age at onset of depression have been reviewed JM J et at onset in late-life depression: a research report of a J Geriatr Psychiatry. Scholar Thus, our group may in have included with early-onset depression, which would the to the we note that CVRFs may be associated with overall medical disability, or the other in the analyses testing the the of finding a significant However, we the hypothesis specifically to the of any significant association with CVRFs, the first hypothesis examined the association of CVRF score with depression. be that our of cumulative CVRF severity individual CVRF to association with Although CVRFs are associated with small-vessel JC ML et vascular risk and cognitive function in a The 44: Scholar, et on magnetic resonance imaging in the I: with age and cerebrovascular risk 17: Scholar, et on magnetic resonance imaging in the 17: Scholar, et of magnetic resonance imaging and in the cerebral Neurol. 1991; Scholar as for small-vessel disease is unknown. of this we with individual CVRFs as but not a specifically of cumulative risk for small-vessel disease may have our to our hypotheses. our study used a to testing the model of We did not include data on MRI hyperintensities or other are by to be a of cerebral but one be especially when whether imaging results of hyperintensities are a of cerebrovascular such be by studies of risk factors to potential We note that in a broad patient group do not the that the small-vessel disease model might apply to patient populations, defined or Future investigations should include studies that in with and investigations will be important as including both of older patient such as and examination of the of CVRFs in life with depressive symptoms and syndromes in later These will to the patient for a small-vessel disease model might to the functional of MRI hyperintensities in and to further the between depression and medical illness with the of or implications of to patients, and