[Classification exposure error in studies on tobacco smoke in pregnancy and birth weight of newborns]

The purpose of the paper was to assess the validity of the self-reported tobacco smoke exposure in pregnancy against plasma cotinine measurements. A total of 158 patients from obstetrical wards was included in the cotinine study. Biochemical smokers were defined as persons with serum cotinine levels greater than 25 ng/ml. The data showed that the exposure classification based on self-reported smoking habit status confronted with cotinine values was of low sensitivity (52%) but of high specificity (98%). To assess the effect of this exposure classification error on the association between the low birth weight (LBW) and smoking in pregnancy, the data from the recent survey in Cracow children have been used. It was shown that RR estimates for smoking and LBW after adjustment for misclassification error were substantially higher than that not adjusted (crude RR = 2.9, corrected RR = 5.1). Due to the exposure misclassification error the attributable fraction (AF) of LBW due to mother's smoking was heavily biased as well. Estimated attributable fraction AF(pop) based on crude RR amounted to 22%, however, after adjustment reached 50%. The corresponding values for attributable fraction in exposed group AF(exp) were 66% and 80%.