Short Communication: Comparison of Calculated Low-Density Lipoprotein Cholesterol (LDL-C) Values in HIV-Infected and HIV-Uninfected Men Using the Traditional Friedewald and the Novel Martin–Hopkins LDL-C Equations

Abstract

The Martin–Hopkins equation used to calculate low-density lipoprotein cholesterol (LDL-C) is more accurate than the traditional Friedewald equation, especially at higher triglyceride levels, which are more common in people with HIV (PWH). Thus, using LDL-C values calculated by the Martin–Hopkins equation may improve clinical care of PWH.

Compared with uninfected persons, people with HIV (PWH) are 1.5–2 times more likely to develop atherosclerotic cardiovascular disease (ASCVD) and to do so at a younger age.^1–3 This increased risk is due to traditional risk factors (e.g., hypertension, diabetes, smoking, and dyslipidemia), but also due to factors related to HIV infection and antiretroviral therapy (ART).³

Low-density lipoprotein cholesterol (LDL-C) is the predominant atherogenic lipid component and is often used as a target of lipid-lowering therapies, including 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitors (commonly known as statins) and other nonstatin medications. Even though statin therapy is often underprescribed in PWH,¹ it has proven effective in reducing noncalcified plaque volume in PWH with subclinical atherosclerosis⁴ as well as decreasing the risk of cardiovascular events in the general population.

LDL-C continues to be an often targeted goal of lipid management based on National Cholesterol Education Program (NCEP) goals, and the 2018 American College of Cardiology/American Heart Association (ACC/AHA) guidelines established a goal of LDL-C of <70 mg/dL in very high-risk individuals with ASCVD.⁵ The ACC/AHA also issued specific recommendations for assessing risk in those with LDL-C ≥70 mg/dL with certain ASCVD risk enhancers including HIV infection.^5–7

General population studies have shown that reducing LDL-C <70 mg/dL is associated with the lowest risk of cardiovascular events.^8,9 Because LDL-C levels and goals are well understood by both clinicians and the general population, they are routinely used to monitor efficacy of and patient adherence to treatment.^5,6 For these reasons, accurate assessment of LDL-C is critical for primary and secondary ASCVD prevention.

The gold standard for direct LDL-C measurement is ultracentrifugation, but this method is not practical for use in nonresearch settings due to cost and availability. Other methods of direct LDL-C measurement have been developed, but these have several disadvantages including lack of specificity toward abnormal lipoproteins, variability in comparison against ultracentrifugation, and are discordant with other measures of atherogenic burden, including apoB and non-high density lipoprotein cholesterol (non-HDL-C).¹⁰

Instead, LDL-C has traditionally been calculated using the Friedewald equation, which was derived in 1972 from a small population of 448 subjects.¹¹ The main limitation of this equation is inaccuracy at higher triglyceride (TG) levels, as it assumes a fixed ratio of TGs to very low-density lipoprotein cholesterol (VLDL-C). Despite this well-known shortcoming, the Friedewald equation continues to be widely used in clinical practice.¹²

A novel method of calculating LDL-C, the Martin–Hopkins equation, was developed in 2013 and uses an adjustable factor for the TG:VLDL-C ratio, which allows for more accurate estimation of LDL-C, particularly in setting of hypertriglyceridemia (150–399 mg/dL).^12,13 This equation compares favorably among men and women [mean age of 59 years (interquartile range [IQR]: 49–69)] against results derived using ultracentrifugation (correlation coefficient of 0.99; overall concordance in classification of LDL-C by guideline categories of 91.7%).^12,13 Mehta et al. showed improved concordance with LDL-C and secondary measures of cardiovascular risk, specifically apoB and non-HDL-C, using Martin–Hopkins equation compared with the Friedewald equation.¹⁴

Because hypertriglyceridemia is more common in HIV-infected versus HIV-uninfected people,¹⁵ inaccuracy of calculated LDL-C using the Friedewald equation may be an important clinical issue for PWH. Although the prevalence of hypertriglyceridemia in PWH has decreased because of improved metabolic profiles of ART, many PWH continue to experience hypertriglyceridemia.

Among the 34,163 lipid profiles tested between 2016 and 2018 in 13,848 PWH enrolled in the CFAR Network of Integrated Clinical Systems (CNICS), 41% of the lipid profiles showed hypertriglyceridemia (TGs >150 mg/dL) with 16% showing TGs >250 mg/dL, a level at which the LDL-C calculated by the Friedewald equation is inaccurate (Heidi Crane, unpublished data, personal communication).

However, to the best of our knowledge, no studies have evaluated the Martin–Hopkins equation in PWH. Therefore, we compared LDL-C values calculated using the Martin–Hopkins equation (LDL-C_MH) and the Friedewald equation (LDL-C_F) in men with or at risk for HIV infection and evaluated whether (1) the difference between the Martin–Hopkins and the Friedewald LDL-C values (LDL-C_MH − LDL-C_F) varied by HIV serostatus and (2) what influence TG levels had on these results.

For this analysis, we used fasting (>8 h since last meal) lipid measurements from 447 men enrolled in the Baltimore-Washington D.C. site of the Multicenter AIDS Cohort Study (MACS), collected at their most recent semiannual visit between October 2015 and February 2018. The study population consisted of 233 men with HIV (MWH) [median age 56.1 years (IQR: 48–64)] and 214 men without HIV [median age 64.4 years (IQR: 58–70)] (Table 1). Of the MWH, 195 (86%) had an undetectable HIV viral load (<20 copies/mL; Roche COBAS^® AmpliPrep/COBAS^® TaqMan^® HIV-1 Test, v2.0) at the time of lipid testing. Lipid testing was conducted at Heinz Nutritional Laboratory at the University of Pittsburgh (Pittsburgh, PA), using serum separated and frozen on day of phlebotomy, stored at −80° C, and shipped on dry ice to the laboratory by overnight delivery.

Table 1.

Baseline Characteristics, Lipid Values^a

Variables	HIV infected (n = 233)	HIV uninfected (n = 214)	pb
Race			<.001
Black	114 (48.9)	37 (17.3)
White	112 (48.1)	172 (80.4)
Age, median years	56 (48–64)	64 (58–70)	<.001
Total cholesterol, mg/dL	175 (149–202)	175 (154–206)	.398
Triglycerides, mg/dL	107 (77–153)	95 (69–134)	.025
HDL-C, mg/dL	54.7 (44.7–63.9)	56.3 (45.7–68.9)	.182
LDL-CF (Friedewald), mg/dL	94.6 (71.4–119.0)	93.2 (75.8–123.8)	.348
LDL-CMH (Martin–Hopkins), mg/dL	95.6 (75.3–119.7)	95.1 (77.5–126.8)	.497
HIV-specific variables
Viral load, copies/mL	60 (37–2,762)
CD4 count, cells/mL	693.5 (524.5–938.5)
Undetectable viral load	195 (83.7)
Taking protease inhibitor	48 (20.6)

^aCategorical data are presented as n (%) and continuous data as median (interquartile range).

^bp values calculated between HIV-infected and HIV-uninfected participants using a chi-square test for categorical variable and Wilcoxon rank sum test for continuous variables.

HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol.

LDL-C was calculated using both the Friedewald (LDL-C_F) and the Martin–Hopkins (LDL-C_MH) equations using the http://ldlcalculator.com spreadsheet for LDL-C_MH values. Significance of the differences in LDL-C (i.e., LDL-C_MH − LDL-C_F) between HIV serostatus groups was evaluated using the Wilcoxon rank sum test. Multivariable linear regression analyses were used to evaluate the association between these differences and TG levels, after adjusting for age, race, and HIV serostatus.

TG levels were significantly higher in the MWH versus men without HIV [median of 107 mg/dL (IQR: 77–153) vs. 95 mg/dL (IQR: 69–134); p = .025], and hypertriglyceridemia (fasting TGs >150 mg/dL) was more prevalent in MWH than men without HIV (27.0% vs. 20.1%; p = .09). There was a significant difference in LDL-C_MH − LDL-C_F at TG levels >250 mg/dL in the total cohort (p = .04). This difference was directly associated with TG levels, with an increase in LDL-C of 8.7 mg/dL for each 100 mg/dL increase in TG level (95% confidence interval: 8.4–8.9; p < .001), and there was no significant difference by serostatus in this rate of increase (p = .77) (Fig. 1).

FIG. 1.

Difference in LDL-C values as calculated by Martin–Hopkins and Friedewald equations (LDL-C_MH − LDL-C_F) in relation to triglyceride levels and HIV status. LDL-C, low-density lipoprotein cholesterol.

The difference in LDL-C_MH − LDL-C_F was significantly greater in MWH versus men without HIV (p = .01), and after adjusting for TG levels, this difference was no longer significant (p = .671), highlighting the importance of the higher TG levels in the MWH.

Among MWH in this study, the difference in LDL-C_MH − LDL-C_F was not associated with viral suppression status, although the number of MWH with detectable viral load was small (32/233). Of the 233 MWH, 48 (20.6%) were taking a protease inhibitor (PI) at time of lipid testing, and there was no significant association between the difference in LDL-C_MH − LDL-C_F and use of a PI-based regimen.

The clinical relevance of this disparity in LDL-C calculated by the two methods can be demonstrated in the case of one study participant with HIV (57-year-old African American) without diabetes whose total cholesterol was 167 mg/dL, HDL-C 30 mg/dL, and TGs 338 mg/dL with an estimated 10-year ASCVD risk score of 7.2%. The ACC/AHA guidelines suggest a moderate intensity statin if the 10-year risk is ≥7.5% and LDL-C ≥70 mg/dL. Using the Friedewald equation, the LDL-C is 69 mg/dL, which is at the goal of <70 mg/dL. However, using the Martin–Hopkins equation, the LDL-C is 92 mg/dL, more strongly favoring the need to initiate moderate intensity statin therapy based on his risk factors, including HIV, despite a 10-year risk that is <7.5%. According to the ACC/AHA guidelines, chronic HIV infection should be considered a “risk-enhancer,” and the decision to use statins may be favored.⁵

As the population of PWH ages, treatment of traditional ASCVD risk factors including dyslipidemia becomes more important. In this study, the difference between LDL-C calculated with the novel Martin–Hopkins equation and that calculated with the conventional Friedewald equation increased with higher TG levels in both MWH and men without HIV. This underestimation of LDL-C using the standard Friedewald equation in those with higher TG values may lead to underprescribing or underdosing of lipid-lowering medications, and a more accurate estimate of LDL-C will provide clinicians the information required to incorporate improved decision-making aimed at reducing ASCVD risk not only in the general population but especially in PWH who have a higher prevalence of hypertriglyceridemia.

Authors' Contributions

E.E.S. conceived the study, compiled the data for statistical analysis, conducted the literature search, performed data interpretation, and wrote the text. S.S. conducted most of the statistical analyses and interpretation and provided critical revision of the text. J.B.M. provided data analysis, expert guidance, and critical revision of the text. S.S.M. provided expert guidance and critical revision of the text. W.S.P. provided expert guidance and critical revision of the text. T.T.B. provided expert guidance in study concept and format, conducted statistical analysis and interpretation, and provided critical revision of the text.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This study was supported by the National Institute of Health (grant nos. U01 AI035042, K24 AI120834, and K12HL143957).