Comparing risk of agranulocytosis estimated with case-control and case-population methods: ticlopidine as an example

INTRODUCTION The case-population approach aims to estimate a risk by comparing the incidence of a disease among the exposed population to the drug of interest with the incidence among the nonexposed.' Using the estimation of the risk of agranulocytosis induced by ticlopidine as an example, the results obtained in a case-control study are compared with those of casepopulation approach, and the effects of limiting the population at risk and of controlling by potential confounding factors are showed. METHODS Cases of agranulocytosis were identified through a case-control study which methods have been published elsewhere.2 The exposed population-time was estimated from consumption data within National Health Service and the defined daily dose. RESULTS The adjusted OR with the case-control approach was 160.75 (95% CI, 35.01 - 738.05). The non-adjusted RR with the case-population approach was 358.36 (95% CI, 210.54 - 587.75), while adjusting by the population at risk and by other risk exposures it was 159.07 (95% CI 84.20 - 292.62). CONCLUSIONS The case-population approach could be an alternative method for a good estimation of risk. This example shows how similar the estimated risks of agranulocytosis by ticlopidine using casecontrol and case-population approaches are when appropriate adjustments are applied. REFERENCES.