An epidemiologic study of Los Angeles Marathon (LAM) applicants was conducted to investigate the relationship between self-reported infectious episodes (IE), training data, and LAM participation. Eight days before the LAM, 4926 of 12,200 applicants were randomly selected, and sent a pilot-tested four page questionnaire, which was received 7 days after the LAM. The 2311 respondents were found to be 2.0 yr older and 7.6 min faster than other LAM finishers (p less than .01). Univariate and multivariate analyses (logistic regression) were conducted to test the relationship between IE and km/wk of running (6 total categories). The final model tested controlled for age, marital status, reported sickness in other members of the runner's home, perceived feelings of stress in response to personal training regimens, and the suppressive effect of sickness on regular training. In runners training greater than or equal to 97 vs less than 32 km/wk, the odds ratio (OR) for IE during the 2 month period prior to the LAM was 2.0 (95% confidence interval (CI) 1.2-3.4). A test for trend showed an increase in OR with increase in km/wk category (p = .04) which was largely explained by the increased odds of reported sickness in the greater than or equal to 97 km/wk category. Of the 1828 LAM participants without IE before the LAM, 236 (12.9%) reported IE during the week following the LAM vs 3 of 134 (2.2%) similarly experienced runners who did not participate, OR = 5.9 (95% CI 1.9-18.8). These data suggest that runners may experience increased odds for IE during heavy training or following a marathon race.