The objective of the study was to compare lipoprotein(a) [Lp(a)] concentrations in population-based samples of individuals with non-insulin-dependent diabetes mellitus (NIDDM), borderline hyperglycemia, and normoglycemia. From 2,740 male Italian Telephone Company employees aged 40 to 59 years participating in a health screening, we selected all those with NIDDM (n = 100) plus a random sample of 950 nondiabetic individuals. Diabetes was defined as fasting plasma glucose (FPG) of at least 140 mg/dL or current use of hypoglycemic drugs. Among nondiabetic individuals, 854 were defined as normoglycemic (FPG < 115 mg/dL) and 95 were defined as borderline hyperglycemic (115 < FPG < 140 mg/dL). Lp(a) level was measured on frozen plasma by enzyme-linked immunosorbent assay. Lp(a) concentrations were similar in people with NIDDM, borderline hyperglycemia, and normoglycemia: 11.2 +/- 14, 14.1 +/- 20, and 13.9 +/- 18 mg/dL, respectively (F = 1.03). Accordingly, the proportion of subjects with Lp(a) levels of at least 30 mg/dL was comparable in the three groups (12%, 15%, and 14%; chi 2 = 3.95, P = .41). Results were not confounded by differences in age, body mass index (BMI), waist to hip ratio, plasma lipids, alcohol consumption, physical activity, and use of drugs. Furthermore, within the diabetic group Lp(a) levels were not significantly different for those on diet only versus those on oral agents (10.8 +/- 14.1 v 11.7 +/- 14.7, P = .7) or for people with FPG of at least 180 as compared with people with FPG less than 180 mg/dL (9.9 +/- 12.8 v 11.5 +/- 14.8, P = .5).(ABSTRACT TRUNCATED AT 250 WORDS)