The Friedewald formula: another restriction?

The Friedewald formula (1) is used to calculate cholesterol in low-density lipoproteins (LDL-C) without ultracentrifugation (1). It requires measurement of total serum cholesterol (TC), triglycerides (TG), and high-density lipoprotein cholesterol (HDL-C), and its accuracy is influenced by the precision of the method used for measuring TC and TG and the effectiveness of the procedure for HDL separation. It allows determination of the atherogenic ratio, LDL-C/HDL-C, which is more sensitive than measurement of TC as predictor (or for follow-up) of coronary heart disease. This ratio usually correlates well with TC/HDL-C, another atherogenic index (2, 3). The formula is based on the assumption that the TG/chol. ratio in verylow-density lipoprotein is constant: TG/chol.= 5. There are three mAjor restrictions on itsuse: (a) it is not applicable to serum samples containing chylomicrons, which are characteristic of types I and V hyperlipoproteinemia (HPL); (b) it gives erroneously high results in type Ill HPL; and (c) it is not accurate when the serum TG concentration exceeds 4.5 mmolJL. Furthermore, some modifications previously suggested make it more accurate, particularly at high TG concentrations (4, 5), but the value for total serum cholesterol must always be >1.30 mmolfL (5). LDL-C calculated according to the Friedewald formula has been compared to LDL-C determination in several studies and it is generally agreed that its use for estimation of LDL-C is valid (6-8). We report here results of follow-up of a 42-year-old man with type I diabetes and myocardial infarction where the Friedewald formula gave a false value for LDL-C, impairing the correlation between the two atherogemc ratios, LDL-C/HDL-C and TC/HDL-C. Serum glucose (Glu), TC, and TG were determined by the Trinder enzymatic method with Boehringer Mannheim GmbH reagents (9). HDL-C was measured by precipitating all other cholesterol fractions with dextran sulfate and magnesium chloride, leaving a supernate that is assayed for cholesterol (10). LDL-C was calculated according to computational procedures of Friedewald et al. (1) [LDL= TC HDL-C TG/2.2 (mmol/L)]. Apolipoprotein B (ape B) was measured by electroimmunodiffusion according to a method modified from Laurell’s procedure (11). These variables were assayed on days one, five, and 10 after myocardial infarction. Results are sununari.zed in Table 1. LDL-C calculated by the Friedewald formula, and consequently the LDLC/HDL-C ratio, fell below the normal range on day five, in contrast with the TC/HDL-C ratio, which remained high. Interestingly, apo B and TG increased in parallel with blood glucose on day five, suggesting altered lipid and glucose metabolism. Values observed for the different lipid variables and apo B may reflect some qualitative and quantitative modifications in lipoprotein composilion. TC reportedly falls below the normal range after myocardial infarction (12), although ape B is always increased concomitantly with the VLDLC/VLDL-TG ratio (13-15). This abnormal composition of VLDL particularly affects their TG/chol ratio, invalidating the Friedewald formula even in mild hypertriglyceridemia. Thus, use of the Friedewald formula to calculate LDLC could be inaccurate when a subnormal TC value is associated with moderate hypertriglyceridemia.