American University of Beirut, Faculty of Health Sciences, Department of Epidemiology and Biostatistics, Beirut, Lebenon. e-mail: ansibai{at}aub.edu.lb
SirWe read with interest the article of Moore et al. about the relationship between psychological stress and ischaemic heart disease (IHD) in middle-aged Caucasian men.1 Several dimensions of psychological stress, including occupational and those originating from lifestyle and life events, had little or no impact on the incidence of IHD. As their study pooled IHD into one category and was limited to men, they could not discuss the potential relationship for the different categories of heart diseases or among women.
In our previous two studies, one using the case-control2 and the other cohort approach,3 we had different results. Several measures of war-related stressful events associated significantly with heart disease morbidity and mortality, in both men and women. War-related stressors were drawn from events experienced by individuals during the war in Lebanon, similar to those originating from life events. These included acute events such as deaths, injuries and kidnappings in the family, and chronic strains such as economic pressures, frequent displacements, and crossing green lines. In both studies, long-term stress has shown more conclusive results. For example, in our case-control study, crossing the green-lines that separated the two belligerent sides during wartime, considered an attribute of chronic stress, associated significantly with arteriographically determined coronary disease. In the more recent cohort study, while specific types of acute and chronic events associated with cardiovascular disease mortality, the largest contribution to risk of mortality in women appeared to ensue from stressors due to kidnappings where the individual remained missing. In both studies, there was a dose-response relationship between exposure to cumulative score of war events and outcome. Moore et al correctly raise several issues and limitations in studies of psychological stress and heart disease. They also indicate that their negative findings may be explained by information bias, because of the potential of frequent change in exposure during follow-up. They suggest that, in future cohort studies, stress should be measured at regular intervals. In our cohort study,3 stressful events were assessed for the 10-year follow-up period, and were modelled in the analysis as time-dependent variables. This allowed for an examination of impact of long-term psychological stress on risk of cardiovascular disease mortality.
Although the role of psychosocial stress factors in the natural history of heart disease has been widely investigated, several issues remain not yet as clear. Studies vary in their definition of psychological stress, recall period, study population and outcome considered, allowing only very general comparisons to be made. Researchers in the field need to differentiate between the immediate impact of events and the role of accumulation of stressful experiences at different stages of the natural history of heart disease.
References
1 Moore L, Meyer F, Perusse M et al. Psychological stress and incidence of ischaemic heart disease. Int J Epidemiol 1999;28:65258.[Abstract]
2 Sibai AM, Armenian HK, Alam S. Wartime determinants of arteriographically confirmed coronary artery disease in Beirut. Am J Epidemiol 1989;130:62331.[Abstract]
3 Sibai AM. Wartime determinants of cardiovascular and all-cause mortality among middle-aged and older populations in Beirut. Ph.D. thesis. London: University London, 1996.
Cancer Research Centre, Pavillion Hotel-Dieu de Quebec, Quebec City, Canada.
SirWe were made aware of the interesting research conducted by Sibai et al. during our literature review. In their case-control study they investigated the relation of war stressors to coronary artery disease (CAD) as determined by coronary angiography and in their cohort study war exposures were studied with reference to CVD mortality.
We find it difficult to compare their research with ours due to the different nature of the stressors studied. Lebanon is a country that emerged in 1991 from 16 years of conflict including multiple invasions and civil disturbances. It is interesting to hypothesize that the extreme level of stress over long periods of time experienced by their study population compared to the relatively variable sources and degrees of stress experienced by the subjects of our study population could explain the difference in results. It may be that only a high level of stress experienced over a long period of time can affect cardiovascular health. Authors found CAD to be related to the number of war events and frequency of crossing green lines. It is interesting to note that the OR associated with one war-related event was below one and an increased odds ratio was only observed after two or more events. As authors have acknowledged in their article concerning the case-control study, the risk of referral bias, misclassification of exposure, over-reporting and recall bias may at least in part explain their results.
As noted by the authors, the heterogeneity of studies in this field does not allow for valid comparison. In order to correctly evaluate the differing aspects of psychological stress, a validated instrument should be used to measure exposure several times during follow-up in a large cohort. The use of regression models with time-dependant covariates could be a very useful statistical tool for analysing such data.