Case-crossover studies of the triggering of disease : myocardial infarction and Ménières disease

Abstract: The general objective of this thesis is to contribute to understanding of the aetiology of myocardial infarction. It focuses on identification and risk-quantification of the factors acting late in the causal chain towards manifest myocardial infarction, so-called triggers, by using the casecrossover design. In addition, methodological aspects and sources of bias are explored. The thesis is based on three studies: the Stockholm Heart Epidemiology Program (SHEEP), a population-based case-referent study including all cases of first-time MI (n=2,246) occurring in Stockholm County from 1992 to 1994; the Onset study, a case-crossover study of all eligible nonfatal cases (n=699) in SHEEP 1993 to 1994; and, the study of triggers of attacks of Ménières disease, a case- crossover study of Ménière patients (n=46) who were recruited during regular visits to the departments of Otolaryngology and Audiology in Stockholm 2000 to 2002. The thesis identifies vigorous exertion with panting and/or overheating, episodes of anger corresponding to at least 'very angry: body tense, clenching fists or teeth', sexual activity, and work- related stressful life events as potential triggers of first-time acute myocardial infarction. The relative risk for patients without premonitory symptoms and ongoing vigorous exertion was 6.1 (95% Cl: 4.3-9.0). Beyond 45 minutes after the end of exertion no increased risk was found. Effect modification was found from physical fitness, where the risk among sedentary patients was 54.7 (95% CI: 13.7-218.4). Up to 60 minutes after an episode of anger the risk increased nine-fold. For patients without premonitory symptoms the relative risk was 15.7 (95% CI: 7.6-32.4). Patients with usually infrequent outbursts of anger (once every second week) showed a significantly higher trigger risk. There was a two-fold increase in risk up to two hours after sexual activity. The risk was substantially modified by physical fitness, which showed a relative risk of 4.4 (95% CI: 1.5-12.9) among sedentary patients. Exposure to single work-related life events was found to increase the risk of MI. Having had a high- pressure deadline at work' entailed a six-fold increase in risk during the next 24 hours; the odds ratio was 6.0 (95% Cl: 1.8-20.4). Although the increases in relative risk are often substantial the increase in absolute risk is small. Most triggers are fairly infrequent, and the periods during which there is an effect short. For the average man, 50-56 years of age, sexual activity implies only a marginal increase in the yearly risk of myocardial infarction; it changes from 0.464% without sexual activity to 0.468% with sexual activity once a week. Control-crossover analyses among a subset of healthy subjects (SHEEP referents) and within the control periods of the Ménière study were performed to establish the validity of the design. The analyses resulted in expected risk estimates of approximately 1. Analyses of the Ménière patients suggested that although the patients are well aware of their disease, many of whom have definite ideas regarding what triggers attacks, neither outcome dependent differential misclassification nor differential misclassification of exposure due to fading memory over time seems to influence the estimated risks in a crucial way.