from long-term antithrombotic protection. Fragmin in Un-stable Coronary Artery Disease (FRISC) Study Group. J AmColl Cardiol 1997; 29: 43–8.

[5] Hamm CW, Heeschen C, Goldmann B et al. Benefit ofabciximab in patients with refractory unstable angina inrelation to serum troponin T levels. N Engl J Med 1999; 340:1623–29.

[6] Fragmin and Fast Revascularisation during InStability inCoronary artery disease Investigators. Long-term low-molecular-weight heparin in unstable coronary-artery disease:FRISC II prospective randomised multicentre study. Lancet1999; 354: 701–7.

[7] Cohen M, Demers C, Gurfinkel EP et al. A comparison oflow-molecular-weight heparin with unfractionated heparin forunstable coronary artery disease. N Engl J Med 1997; 337:447–52.

[8] The CAPTURE investigators. Randomized placebo-controlled trial of abciximab before and during coronaryintervention in refractory unstable angina: the CAPTUREstudy. Lancet 1997; 349: 1429–35.

[9] Abrahamsson P, Anderson K, Eriksson P, Dellborg M. Prog-nostic value of maximum ST-vector magnitude during the first24-hours of vectocardiographic monitoring in patients withunstable angina pectoris. Eur Heart J 1999; 20: 1166–74.

[10] Lindahl B. More is worse — ST-segment deviation in unstablecoronary artery disease. Eur Heart J 1999; 20: 1611–12.

[11] Lindahl B, Andren B, Ohisson J, Venge P, Wallentin L andthe FRISK study group. Risk stratification in unstable

coronary artery disease. Additive value of troponin T deter-minations and pre-discharge exercise tests. Eur Heart J 1997;18: 762–70.

[12] Lin SS, Lauer MS, Marwick TH. Risk stratification ofpatients with medically treated unstable angina using exerciseechocardiography. Am J Cardiol 1998; 82: 720–4.

[13] Brown KA. Prognostic value of thallium-201 myocardialperfusion imaging in patients with unstable angina whorespond to medical treatment. J Am Coll Cardiol 1991; 17:1053–7.

[14] Sitges M, Pare C, Azqueta M et al. Feasibility and prognosticvalue of dobutamine-atropine stress echocardiography earlyin unstable angina. Eur Heart J 2000; 21: 1063–71.

[15] Ferro G, Duilio C, Spinelli L, Liucci GA, Mazza F, Indolfi C.Relation between diastolic perfusion time and coronary arterystenosis during stress-induced myocardial ischemia. Circu-lation 1995; 92: 342–7.

[16] Beleslin BD, Ostojic M, Djordjevic-Dikic A et al. Integratedevaluation of relation between coronary lesion featuresand stress echocardiography results: the importance ofcoronary lesion morphology. J Am Coll Cardiol 1999; 33:717–26.

[17] Lu C, Picano E, Pingitore A et al. Complex coronary arterylesion morphology influences results of stress echocardiogra-phy. Circulation 1995; 91: 1669–75.

[18] Little WC. Angiographic assessment of the culprit coronaryartery lesion before acute myocardial infarction. Am J Cardiol1990; 66: 44–47.

European Heart Journal (2000) 21, 1043–1045doi:10.1053/euhj.2000.2138, available online at http://www.idealibrary.com on

Editorials 1043

Coronary artery disease, depression and social support:only the beginning

See page 1072 for the article to which this Editorialrefers

In this issue, Horsten and colleagues[1] report thatsymptoms of depression and lack of social integration3–6 months after discharge for an acute coronarysyndrome have independent negative impacts on5-year prognosis in women 65 years of age andyounger. Although finding increased risks associatedwith depression or with low social support are notnew, studies examining the joint impact of depressionand measures of social support in cardiac patients arerare.

Of all the negative emotions, including depression,anxiety and hostility, that have been hypothesized toinfluence outcomes in patients with established cor-onary artery disease, the epidemiological findings aremost consistent for depression[2,3]. Research in post-myocardial infarction samples has documented thatdepression increases the risk of mortality from two toseven times, depending on the sample characteristics

and the duration of follow-up[4–7]. We recentlyobserved similar increases in the combined risk ofcardiac mortality and myocardial infarction in un-stable angina patients[8]. Post-hoc analyses of severallarge longitudinal data sets also suggest that depres-sion is associated with the incidence of coronaryartery disease in initially healthy individuals[9–12].

What kind of depression increases the risk of car-diac events, and what type of events are involved? Thestudy by Horsten and colleagues[1] suggests an in-crease in the combined risk of revascularization, car-diac mortality, and myocardial infarction in womenreporting two or more out of nine depressionsymptoms. Most previous research highlightingdepression-related risks for mortality and myocardialinfarction in coronary artery disease patients hasfound the risk to be concentrated in the approxi-mately 25 to 30% of patients with the highest level ofdepression. However, in this study the risk did notincrease beyond two symptoms, the level reported by72% of the women. Two symptoms of depression is

� 2000 The European Society of Cardiology

1044 Editorials

not the equivalent of major or even minor depression,and is not a condition for which current psychiatricguidelines recommend treatment. The question maynot be what a normal level of depressive symptoms is,but at what threshold medical attention is needed.The answer is dependent both on the magnitude of theincreased risk, and the costs and benefits of treatment.

There were too few deaths and infarctions over 5years of follow-up in this sample to assess the linkbetween depression and hard cardiac events. Fromthe data provided, it appears that there were, at most,26 deaths or myocardial infarctions among the 81patients with one of the combined events. The risk inthis study was primarily a risk of revascularization.What is the clinical significance of these findings? Weneed to consider what revascularization represents:evidence of myocardial ischaemia with a coronaryanatomy amenable to angioplasty or bypass surgery.After hospital discharge what is the process thatgenerates this evidence? In the usual situation, patientcomplaints of angina or routine procedures to detectresidual ischaemia, including exercise stress tests ornuclear medicine evaluations, lead to angiographicassessment. However, in the study by Horsten et al.there was another potential pathway. Some 131 of the292 patients took part in a quantitative angiographystudy within 3 months following their baseline inter-view. According to an earlier report of the studypublished in this journal in 1998 ‘all angiographieswere carried out based on the study protocol i.e.primarily because the patient was a participant in theFemCorRisk study.’[13]. Thus, the procedures weredone on an ‘availability basis’ rather than based onclinical factors. Although the authors do not describewhat happened to patients with significant stenoses,the 1998 report indicates that 53 patients had at leastone coronary artery obstructed by 50% or more. It islikely that many of them became the revascularizedpatients in the present analyses. In this context, itseems that middle-aged women who report less thantwo depression symptoms, 3–6 months after anadmission for an acute coronary syndrome areunlikely to have significant stenoses requiring revas-cularization. However, it is unclear whether this is achicken or egg phenomenon. By this point, how manyof these women knew or suspected that they were inline for a procedure because of symptoms of residualischaemia? The authors found that the severity ofangina symptoms was highly linked to depression(P=0·0003). However, control for angina class didnot alter the relationship between depression andtheir combined end-point. In summary, these resultssuggest that if a woman coronary artery diseasepatient aged 65 or younger reports more than onedepression symptom, thorough evaluation for revas-

Eur Heart J, Vol. 21, issue 13, July 2000

cularization should be undertaken regardless of herangina level.

In addition to their results for depression, theauthors also observed a statistically significantincrease in risk with decreasing levels of social inte-gration, the degree to which an individual is inter-connected with others and has multiple contacts inthe family and community. Although women withtwo or more depression symptoms reported signifi-cantly lower levels of this sort of social networkcontacts, the risk associated with low social inte-gration was present for both depression groups, andso was independent of depression. Previous work bythe investigators with middle-aged men free fromcardiac disease at baseline demonstrated that socialintegration predicted myocardial infarctions and cor-onary deaths during the subsequent 6-year period[14].So social integration predicts the incidence of cardiacdisease in middle-aged men, and revascularizationand cardiac events in middle-aged women with al-ready established coronary artery disease. However,in their 1998 quantitative angiography study, involv-ing a subsample of the current study patients, theauthors also found that major coronary stenoses weresignificantly more common in women with low socialintegration. Thus, it makes sense that the combinedend-point of revascularization, myocardial infarctionand cardiac deaths would also be linked to low socialintegration in the full sample.

In comparison, the authors did not find muchimpact of decreasing levels of what they term ‘attach-ment’ on combined outcomes in women with cor-onary artery disease (P=0·14). Again this parallelstheir 1998 results with coronary blockages. However,their previous work with men initially free of cor-onary disease found that attachment acted much likesocial integration to predict hard cardiac events.Their measure of attachment assessed the availabilityof a close special relationship with one or morepeople. As the authors suggest, it seems plausible thatthe lack of stronger findings is due to the scale’s verylimited range (only 6 points) with the data skewedtowards higher scores, possibly because of socialdesirability response bias.

By what mechanisms might social integration orattachment influence coronary stenoses and via themaffect cardiac events and revascularization? The lit-erature suggests two major models to explain theimpact of social support: the main effect and thestress-buffering models[15]. In the main effect model,social support has an impact on disease outcomesregardless of other physical or psychosocial risks.This is what the authors observed with depression:integration had an impact independent of depression.In fact, most of the evidence supporting this type of

Editorials 1045

model has come from studies that assess social inte-gration. There are several plausible explanations.First, people with more inter-connections with thecommunity tend to have healthier lifestyles. In thecurrent study, smoking was less common with higherintegration. Second, when people begin to become illthey may reduce their contacts with others because oflower energy levels so that lower social integrationmay be a proxy measure of greater illness severity.However, control for measures of cardiac risk did notexplain the link observed by the authors. It has alsobeen hypothesized that social integration has its im-pact on cardiac outcomes because integrated patientsmay be more likely to receive medical treatment, orbe more motivated to alter risk factors. In contrast tothe main effect model of social support, the stress-buffering model suggests that social support isbeneficial only in the presence of ‘high stress’ or highpsychological risk. This model is primarily supportedby measures of how individuals view their closesocial support, essentially what was not adequatelymeasured by the attachment scale in the currentstudy. If we consider that depression is a ‘high stress’situation, the stress buffering model suggests thatperceived social support should act to buffer theimpact of depression on cardiac events, that is, ifproperly measured, depression and attachmentshould interact to predict cardiac outcomes.Although the authors did not report testing thispossibility, we recently examined this relationship in asample of 887 men and women following myocardialinfarction[16]. We found that while perceived socialsupport was not directly related to 1-year survival,very high levels of support buffered the impact ofdepression on mortality. Depressed individuals withhigh support did not experience a depression-relatedincrease in risk. Further, high levels of support pre-dicted improvements in depression symptoms overthe first post-myocardial infarction year in depressedpatients.

Although depression and poor support have noweach been associated with poorer prognosis in severalstudies of coronary artery disease patients, the searchfor the mechanisms has just begun. While there is asyet no scientific proof that intervening on thesefactors will improve prognosis, the epidemiologicaldata should finally be sufficiently strong to intrigue allof us and trigger a surge of new research.

N. FRASURE-SMITHF. LESPEuRANCE

Montreal Heart Institute,McGill University and University of Montreal,

Montreal, Quebec, Canada

References

[1] Horsten M, Mittleman MA, Wamala SP, Schenck-GustafssonK, Orth-Gomer K. Depressive symptoms and lack of socialintegration in relation to prognosis of CHD in middle-agedwomen; the Stockholm Female Coronary Risk Study. EurHeart 2000; 21: 1072–80.

[2] Glassman AH, Shapiro PA. Depression and the course ofcoronary artery disease. Am J Psychiatry 1998; 155: 4–11.

[3] Rozanski A, Blumenthal JA, Kaplan J. Impact of psychologi-cal factors on the pathogenesis of cardiovascular disease andimplications for therapy. Circulation 1999; 99: 2192–217.

[4] Ladwig KH, Kieser M, Konig M, Breithardt G, Borggrefe M.Affective disorders and survival after acute myocardial infarc-tion: Results from the post-infarction late potential study. EurHeart J 1991; 12: 959–64.

[5] Frasure-Smith N, Lesperance F, Talajic M. Depression fol-lowing myocardial infarction: Impact on 6-month survival.JAMA 1993; 270: 1819–25.

[6] Frasure-Smith N, Lesperance F, Talajic M. Depression and18-month prognosis after myocardial infarction. Circulation1995; 91: 999–1005.

[7] Frasure-Smith N, Lesperance F, Juneau M, Talajic M,Bourassa MG. Gender, depression and one year prognosisafter myocardial infarction. Psychosom Med 1999; 61: 26–37.

[8] Lesperance F, Frasure-Smith N, Juneau M, Theroux P.Depression and 1-year prognosis in unstable angina. Arch IntMed 2000; 60 (in press).

[9] Barefoot JC, Schroll M. Symptoms of depression, acutemyocardial infarction, and total mortality in a communitysample. Circulation 1996; 93: 1976–80.

[10] Aromaa A, Raitasalo R, Reunanen A, Impivaara O,Heliovaara M, Knekt P et al. Depression and cardiovas-cular diseases. Acta Psychiatr Scand 1994; 377 (Suppl): 77–82.

[11] Anda R, Williamson D, Jones D, Macera C, Eaker E,Glassman A et al. Depressed affect, hopelessness, and the riskof ischemic heart disease in a cohort of U.S. adults. Epidemi-ology 1993; 4: 285–94.

[12] Ford DE, Mead LA, Chang PP, Cooper-Patrick L, WangN-Y, Klag MJ. Depression is a risk factor for coronary arterydisease in men. Arch Intern Med 1998; 158: 1422–6.

[13] Orth-Gomer K, Horsten M, Wamala SP et al. Social relationsand extent and severity of coronary artery disease. Eur HeartJ 1998; 19: 1648–56.

[14] Orth-Gomer K, Rosengren A, Wilhelmsen L. Lack of socialsupport and incidence of coronary heart disease in middle-aged Swedish men. Psychosom Med 1993; 55: 37–43.

[15] Cohen S, Wills TA. Stress, social support, and the bufferinghypothesis. Psychol Bull 1985; 98: 310–57.

[16] Frasure-Smith N, Lesperance F, Gravel G, Masson A, JuneauM, Talajic M, Bourassa MG. Social support, depression andmortality during the first year following myocardial infarc-tion. Circulation 2000; 101: 1919–24.

Eur Heart J, Vol. 21, issue 13, July 2000