Drug use and acute leukemia

To study the occurrence of acute leukemia in relation to preceding use of drugs a case-control study has been carried out in Rome, Italy.

Two hundred and two patients (age >15 years) with a diagnosis of acute leukemia during the period July 1992-June 1994 were enrolled. For each patient, 10 controls matched by age and gender were randomly drawn from the source population. Through the individual beneficiary code the information relevant to the drugs received within the National Health Service during the period January 1989-December 1992 was retrieved. The use of drugs was considered etiologically related to leukemia if the prescription occurred before the 12 months preceding the diagnosis. Exposure was categorized as 'any use' (at least one prescription during the etiologic period), 'high use' (duration of use greater than the median in the control group) and, for NSAIDs, 'very high use' (duration of use greater than 180 days).

Among drugs suspected to cause leukemia, users of high doses of chloramphenicol presented an OR of 1.8 (95% CI: 0.6-5.3). Among other categories of drugs with an increase in the ORs, though not statistically significant, we found tricyclic antidepressants (OR=1.7; 0.8-3.4) and oral contraceptives (OR=1.8; 0.8-4.0). No excess risk was observed for users of calcium-channel blockers (OR=0.9; 0.5-1.7). Use of very high doses of NSAIDs appeared to decrease the occurrence of acute leukemia (OR=0.4; 0.1-1.5).

Even with several limitations, this study provides an initial frame of reference for the potential causal role of drugs in acute leukemia.