Estimating familial aggregation while adjusting for covariates. Application to pulmonary function data from black and white sibships

Although crude correlations are useful in family studies, some adjustment for effects of risk factors that vary both within and among families if often needed. A linear model for estimating sibship correlations while simultaneously considering height, age, race, sex, ascertainment, and smoking status was used on pulmonary function data on 1-second forced expiratory volume (FEV1) and the natural logarithm of the ratio of FEV1 to forced vital capacity (lnFEV%) from 402 adults in 152 white sibships and 172 adults in 59 black sibships. Crude correlations of .271 +/- .048 (FEV1) and .342 +/- .047 (lnFEV%) decreased significantly to .206 +/- .048 and .231 +/- .048, respectively, after adjustment. For black and white sibs, adjusted intraclass correlations, although not statistically different, were .153 +/- .089 and .225 +/- .055 (FEV1), respectively, and were .103 +/- .088 and .275 +/- .054 (lnFEV%), respectively, suggesting that pulmonary function may aggregate more strongly among whites. This analysis illustrates how risk factor adjustment can be readily incorporated into familial correlation studies.