[Collection of familial data in pharmacogenetics. Methodological problems]

The purpose of this paper is to show the problems with family data collection in a pharmacogenetic study, the aim of which was to study the genetic polymorphism of inducibility of cytochromes P4501A by polycyclic aromatic hydrocarbon (PAH). Data were collected from 76 smoker nuclear families (315 volunteers). A caffeine test, a blood sample and answers to a questionnaire were obtained from each healthy volunteer. It was a crucial problem to recruit nuclear families with healthy smoker father and/or mother and 2 smoker children. On 127 answers, 22 families were not eligible, 27 refused and 10 had a single child, that meant secondary refusals. Problems differed with origin of the recruitment. Included were 40 families obtained from 3 antismoking outpatient departments, 29 from general practitioners but 6 only from students. The family rates with 2 parents/2 children were significantly higher with general practitioners (p < 0.01). This emphasized their part in epidemiologic studies. Problems with the use of methodology were bound to acceptability of the tests which were better in females (p < 0.05), and a change in caffeine form could improve this acceptability. Lastly, difficulties with laboratories constraints required a very good coordination between families, nurses and laboratories.