The genetic basis of familial combined hyperlipidemia (FCHL) has eluded investigators for 20 years, despite the apparent segregation of FCHL as an autosomal dominant disorder affecting 1% to 2% of individuals. Etiologic heterogeneity and additive effects of traits controlled by other genetic loci have been suggested. Two traits have been implicated in FCHL. The first is the predominance of a small, dense low-density lipoprotein (LDL), LDL subclass phenotype B, which segregates as a mendelian trait. The second is a mendelian locus with large effects on apolipoprotein (apo) B levels that is defined by complex segregation analysis (predicted apoB level genotype). This study shows that these factors appear to be separate genetic effects, both of which aid in the prediction of FCHL in four large pedigrees. The results suggest that FCHL may be best predicted by a threshold model in which apoB level genotype and LDL subclass phenotype each act to increase the risk of FCHL. Heterogeneity in the transmission of apoB levels among families is suggested, supporting the etiologic heterogeneity of FCHL. These results emphasize the advantages inherent in the study of large pedigrees when disease heterogeneity is suspected.