New modified Friedewald formulae for estimating low-density lipoprotein cholesterol according to triglyceride levels:...

The most commonly used method for estimating low-density lipoprotein cholesterol (LDLC) is Friedewald formula (FF). This study aims to extract and validate new modified Friedewald formulae for estimating LDLC according to triglyceride (TG) levels in Iranians.

Total cholesterol (TC), high-density lipoprotein cholesterol (HDLC), LDLC, and TG were measured in 5030 fasted subjects. The original FF was used for estimating LDLC. Data were divided into training and validation data sets. For extracting modified Friedewald formulae, linear regression was used with TG as independent and TC, HDLC, and directly measured LDLC (D-LDLC) as dependent variables, respectively. An overall modified formula was extracted for TG concentrations <400 mg/dL. The specific modified formulae were extracted for different TG categories. The performance of the modified formulae was assessed in the validation data set.

The overall derived formula for calculating LDLC was M-LDLC = TC-HDLC-TG/4. The coefficients of TG (specific TG terms) in modified formulae were 2.7, 3.7, 4.6, and 5 for TG <100, 100-200, 200-300, and 300-400 mg/dL, respectively. Compared to the original FF, applying a new modified formula with TG/4 to a validation data set provided a less mean difference (4.03 vs. 10.86 mg/dL), greater kappa coefficient (0.691 vs. 0.505), and better subject classification (80.8 vs. 67.1%). After applying modified formulae with specific TG terms, the difference between M-LDLC and D-LDLC decreased to 1.41 mg/dL.

Modified formulae were developed and validated for LDLC estimation that compared to the original FF, provided more accurate estimation of LDLC and better classification of subjects. These findings shall be useful for preventive, diagnostic, and therapeutic purposes.