Commingling analysis of regional fat distribution measures: the Québec family study

Regional fat distribution is related to higher risk of diabetes and cardiovascular morbidity and mortality, independent of excess body mass for height. In particular, the male (android) pattern of fat deposition, which is characterized by greater truncal and abdominal fat stores relative to extremity fat levels, is associated with a higher propensity to metabolic complications. Motivated by these considerations, we have initiated a systematic investigation of several measures of regional fat distribution aimed at the detection of possible single gene effects. In this paper, we assess the evidence for commingling in the distributions of these variables in a large French-Canadian study. Two measures approximating the size of subcutaneous fat stores relative to total body fat were considered: the sum of six skinfolds (SF6 = abdominal + supra-iliac + subscapular + calf + tricep + bicep), and the sum of three trunk skinfolds (TSF3 = abdominal + supra-iliac + subscapular). In addition, two measures assessing the distributional pattern of subcutaneous fat were considered: the ratio of TSF3 to the sum of the three extremity skinfolds (TER), and a relative fat pattern index [RFPI = subscapular/(subscapular + supra-iliac)]. All four measures were assessed both prior to and after adjusting for total fat mass, which was measured using underwater weighing. Significant distributional heterogeneity was observed for some of these measures, either between generations and/or between the sexes. In general, however, fat mass adjustment tended to eliminate the heterogeneity; the exception was for RFPI, for which sex differences were noted both prior to and after the adjustment. The finding of commingling of distributions for almost all phenotypes is consistent with (but not evidence for) major gene effects. However, for some of the measures the effect of a putative major locus genotype may be mediated by covariates such as age and/or sex.