Familial combined hyperlipidemia (FCH) is characterized by a variable expression of hypercholesterolemia and/or hypertriglyceridemia. We evaluated the variability in lipid phenotype expression over a 5-year period and studied factors affecting the lipid phenotype expression. A total of 32 families (299 subjects) were studied in 1994 and in 1999. Subjects were classified as having FCH when total cholesterol and/or triglyceride levels exceeded the 90th percentile adjusted for age and sex. In 1994, 93 (31%) of the 299 subjects were affected, whereas 206 (69%) of the subjects were unaffected relatives. In 1999, a diagnosis of FCH was consistent in 69 (74%) of the 93 subjects. So, 26% of the FCH subjects in 1994 showed a sporadic normolipidemic pattern (ie, total cholesterol and/or triglycerides <90th percentile) in 1999. Among the 206 unaffected relatives in 1994, 178 (86%) remained unaffected in 1999, and 28 (14%) developed an FCH lipid phenotype. Multiple regression analysis showed that sex (odds ratio 2.03, 95% CI 1.09 to 3.87; P=0.03) and body mass index (odds ratio 1.14, 95% CI 1.05 to 1.24; P<0.01) significantly contributed to the variability in lipid phenotype expression. Thus, a diagnosis of FCH, based on plasma total cholesterol and/or triglyceride levels, is consistent in only 74% of the subjects over a 5-year period. Two other major characteristics of our FCH group, compared with the unaffected relatives, included elevated apolipoprotein B (apoB) levels and the presence of small dense low density lipoprotein (LDL), as reflected by a low value of the parameter K (apoB 1461 +/- 305 versus 997 +/- 249 mg/L, respectively [P < 0.001]; K value -0.22 +/- 0.19 versus -0.02 +/- 0.19, respectively [P < 0.001]). We now report that the apoB concentration and the K value show less variability in time and are more consistently associated with FCH, inasmuch as affected FCH subjects, compared with the unaffected relatives, persistently show a higher apoB level and a lower value of parameter K, reflecting small dense LDL, even when they present a sporadic normolipidemic pattern. In conclusion, our results emphasize the need for reevaluation of the diagnostic criteria for FCH. We demonstrate that apoB and small dense LDL are attractive new candidates for defining FCH. Further studies are indicated to evaluate the role of apoB and small dense LDL as diagnostic criteria for FCH.