Abstract

BACKGROUND:

Estimation of low-density lipoprotein cholesterol is crucial in the management of ischemic heart disease patients. Low-density lipoprotein cholesterol is routinely calculated in laboratories world over by applying Friedewald formula for logistic reasons. We derived a new formula based on multiple regression approach.

METHODS:

Lipid profiles were done on blood samples of 2008 patients. In initial 1000 patients, low-density lipoprotein cholesterol was estimated by a direct method and also by Friedewald formula. By applying linear regression methods on the data of direct estimation method, a new formula was obtained and the accuracy of this new formula was validated in the next 1008 patients.

RESULTS:

The mean low-density lipoprotein cholesterol was 116+/-41.5 mg/dl (3.02+/-1.08 mmol/l) measured by direct low-density lipoprotein cholesterol assay and that calculated by Friedewald formula was 119+/-46 mg/dl (3.09+/-1.2 mmol/l) for the initial 1000 patients. Low-density lipoprotein cholesterol measured by direct low-density lipoprotein cholesterol assay and calculated from Friedewald formula showed good correlation (r = 0.88), however, there was minimal overestimation by the Friedewald formula. The correlation improved between direct low-density lipoprotein cholesterol and calculated low-density lipoprotein cholesterol after excluding the patients with triglycerides more than 350 mg/dl (r = 0.92). The mean low-density lipoprotein cholesterol measured by the direct assay and by new formula in the next 1008 patients was 117+/-40 mg/dl (3.04+/-1.04 mmol/l) and 113.7+/-37 mg/dl (2.96+/-0.96 mmol/l), respectively with very good correlation (r = 0.97) between them.

CONCLUSIONS:

The new formula derived from multiple linear regression analysis appears to be more accurate than Friedewald formula in Indian population.