Biological variation of serum lipids and lipoproteins in patients with clinically well controlled non insulin dependent diabetes mellitus

Abstract

To investigate how the visit-to-visit variation in serum lipids measurements affects the decision making concerning treatment according to the National Cholesterol Education Program (NCEP) guidelines in patients with clinically well controlled non-insulin-dependent diabetes mellitus (NIDDM) we have measured the biological variation (CVb) in serum total cholesterol (TC), triglycerides (TG), high density lipoprotein cholesterol (HDL-C) and low density lipoprotein cholesterol (LDL-C) in 26 patients with NIDDM. We found the CVb as follows: TC, 5.1%; TG, 17.0%; HDL-C 4.4% and LDL-C, 8.3%. Confidence intervals (95%) were determined with total intra-individual variance values around the NCEP cut-off points to evaluate how well one, two and three lipid measurements provided reliable risk classification. A single TC measurements <177 mg/dL or >263 mg/dL allowed confident classification as "desirable" or "high risk" respectively. For LDL-C, one measurement was accurate only at below 106.3 mg/dL or above 183.7 mg/dL. The average of three measurements contracted these limits to <186.7 mg/dL and >253.3 mg/dL for TC, and <116.3 mg/dL and >173.7 mg/dL for LDL-C. For HDL-C also, multiple measurements improved risk assignment in a similar fashion. There were no values which allowed assignment to the "borderline high" category with one TC measurement and with one and two LDL-C measurements. The mean of three TC and three LDL-C measurements allowed assignment to the "borderline high" category, if between 213.3 and 226.7 mg/dL for TC, 143.7 and 146.3 mg/dL for LDL-C. Seven patients (26.9%) in this risk group based on the mean of two LDL-C estimates could be placed into a different category when the mean of three estimates was taken, even though the first two LDL-C test results did not differ by more than 30 mg/dL. Our results suggest that repeated lipid measurement is important especially for the "borderline-high" risk group because big variations existed in some patients, and further that TC is the most reliable quantity.