---

What is non-HDL-C?

Non HDL-C represents the sum of the cholesterol content of all the lipid particles in blood, except HDL (Figure 1). Since all lipid particles other than HDL [i.e. VLDL, IDL, LDL, lipoprotein (a)] are atherogenic, non-HDL-C actually represents cholesterol content of all the atherogenic lipid particles in blood. Measurement of non-HDL-C does not require any additional test as it can be easily calculated by subtracting HDL-C from total cholesterol, available as part of a routine lipid profile (non HDL-C= total cholesterol – HDL-C).

Non HDL-C is a robust lipid measurement and has distinct advantages over LDL-C (Table 1). First, unlike LDL-C, measurement of non-HDL-C does not necessarily require fasting sample as both total cholesterol and HDL-C are not acutely affected by feeding. In contrast, LDL-C, which is often calculated from Freidwald equation instead of direct measurement, requires a fasting sample. As per Freidwald equation, LDL-C is estimated as-

LDL-C = Total cholesterol – HDL-C – (serumtriglycerides/5)

Since serum triglyceride level is acutely elevated by feeding, a non-fasting sample will underestimate LDL-C by this method.

Second, as discussed below, non-HDL-C, by including all the different atherogenic lipid molecules in blood, has been shown to be a more complete measure of atherogenicity and a much stronger predictor of CVD risk. For these reasons, non-HDL-C has already been included as a secondary target of therapy in several guidelines on lipid management [9]. It is recommended that the non-HDL-C level should not exceed LDL-C by more than 30 mg/dl at any point.

The rationale behind this recommendation is that non-HDL-C primarily reflects the sum total of LDL-C and VLDL-C and 30 mg/ dl cholesterol in VLDL-C would correspond to serum triglyceride level of 150 mg/dl, which is the acceptable upper limit of normal for triglycerides [9].

Dynamic relationship between atherogenic lipoproteinsand apolipoproteins

Each lipid molecule in the blood contains at least one molecule of an apolipoprotein. The apolipoprotein in LDL, IDL, VLDL, lipoprotein (a) and chylomicrons is apolipoprotein-B (Apo-B) whereas it is Apo-A in HDL molecules. The Apo-B concentration in blood is therefore directly proportional to the number of atherogenic lipid particles present and thus reflects the total atherogenic potential. For this reason, Apo-B levels have been shown to be an excellent predictor of CV risk, far superior to LDL-C [19-21]. As non-HDL-C measures cholesterol content of all the Apo-B containing particles, it correlates closely with Apo-B levels. In contrast, the relationship between LDL-C and Apo-B is quite variable. Under normal conditions, almost 90% of all the Apo-B in the blood is carried by LDL particles and therefore LDL-C levels tend to correlate with Apo-B levels. However, this concordant relationship between Apo-B and LDL-C is diminished in patients who have elevated triglycerides. When the levels of IDL and VLDL increase as in hypertriglyceridemia, the contribution of these particles to total Apo-B pool increases and the relationship between LDL-C and Apo-B weakens [22,23]. In addition, higher VLDL levels are also associated with greater proportion of small-dense LDL particles, which further weakens this relationship. Excess triglyceride in VLDL particles is exchanged for the cholesterol present within the LDL particles. Subsequent hydrolysis of triglycerides transferred to LDL renders the LDL particles smaller, denser and more atherogenic. As the proportion of small-dense LDL particles increases, more number of LDL particles are now available for the same amount of cholesterol. As a result, LDL-C alone will underestimate the total atherogenic burden whereas non-HDL-C, by incorporating information about VLDL and IDL, still remains more accurate and shows better correlation with serum Apo-B levels. The clinical recognition of this phenomenon has become increasingly important now with the rising prevalence of obesity, metabolic syndrome and diabetes mellitus. These are the conditions characterized by disproportionately elevated triglycerides, low HDL-C, increased proportion of small-dense LDL and normal or near-normal LDL-C levels [24]- the typical setting in which non- HDL-C becomes a superior risk marker than LDL-C.

Non-HDL-C as a measure of atherosclerosis risk

Non HDL-C and CV risk prediction: As stated earlier, the atherogenic lipoprotein paradigm suggests that for determination of CV risk, total number of atherogenic particles is more important than the conventional lipid parameters such as LDL-C. Being an aggregate measure of all atherogenic lipoproteins, non-HDL-C is therefore expected to be a more accurate predictor of CV risk as compared to LDL-C. This hypothesis has been proven in a number of studies that have shown better prediction of CV risk by non HDL-C, than LDL-C, with an accuracy comparable to that of Apo-B [14,18,25-27].

In the Lipid Research Clinics Program, 4462 middle agedindividuals who were free from CVD were followed up for an average of 19 years. It was found that non-HDL-C was a much stronger predictor of all cause and CV mortality as compared to LDL-C (chi square 24.3 for non-HDL-C and 5.0 for LDL-C) [14]. A 30 mg/dl increase in non-HDL-C resulted in 19% increase in mortality in men and 11% increase in women compared to 15% and 8% respectively for LDL-C. In 2003, Lu et al studied 4549 diabetic patients in the Strong Heart Study and compared the ability of various lipid measures for prediction of CV risk [22]. In both men and women, the hazard ratio for the upper tertile of non-HDL-C was found to be higher than the same for LDL-C and triglycerides. The greater predictive accuracy ofnon-HDL-C than LDL-C in diabetics was further established in a posthoc analysis of 4 large prospective studies-The Framingham Cohort Study, The Framingham off spring study, The Lipid Research Clinics Program follow up study and Multiple Risk factor intervention trial25. A total of 19381 individuals who were above 30 years of age and were free from CVD were included in this post-hoc analysis. This analysis demonstrated that compared to non-diabetics, the diabetic subjects had significantly higher non-HDL-C levels (176.7 and 194.1mg/dl respectively, p< 0.001) but almost identical LDL-C levels (148.0 and 148.6 mg/dl respectively, p=0.68). On multivariate analysis, CVD risk in diabetics increased with increase in non-HDL-C but not LDL-C.

The predictive accuracy of various lipid measures in a secondary prevention setting was evaluated in the BARI trial (Bypass Angioplasty Revascularization Investigation) [26], which was primarily conducted to compare coronary bypass surgery with angioplasty in patients with multi-vessel coronary artery disease. A total of 1514 patients were included and followed up for 5 years. Once again, non-HDL-C was found to be a significant independent predictor of non-fatal myocardial infarction and angina whereas LDL-C did not predict either of these end-points.

Non-HDL-C and CV risk in relation to serum triglycerides levels: Since elevated non-HDL-C, when out of proportion to LDL-C levels, suggests increased concentration of triglyceride rich lipoproteins such as VLDL, it becomes imperative to determine whether the increased CV risk associated with non-HDL-C is attributed to increased serum triglyceride levels. Several studies have demonstrated that non- HDL-C is a strong predictor of CV risk independent of triglyceride levels [23,28]. Non-HDL-C was found to maintain its predictive value for CV events in patients with triglycerides levels< 200mg/dl as in the EPIC-Norfolk (European Prospective Investigation into Cancer and nutrition-Norfolk) study [28], and also at triglyceride levels > 400mg/dl as in the SHEP (Systolic Hypertension in the Elderly Program) study [23]. In contrast, LDL-C lost its predictive value when triglyceride levels exceeded 400mg/dl [23].

Non-HDL-C in comparison to Apo-B: Since non-HDL-C servesas a surrogate for Apo-B, the two have also been compared for their ability to predict CV risk. In the Health Professionals Follow-up study involving 18225 men, 40 to 75 years of age, non-HDL-C was found to be an inferior predictor of CV events as compared to Apo-B [21]. However, in another study done in 15632 healthy women aged 45 years and above (The Women Heart study), the highest quintile of non-HDL-C had similar relative risk for major CV events as the highest quintile of Apo-B [27]. It is noteworthy that in both these studies, non-HDL-C performed much better than LDL-C.

Non-HDL-C and correlation with measure of subclinicalatherosclerosis: Orakzai et al studied asymptomatic individualsundergoing coronary artery calcium (CAC) screening and foundstatistically significant association between non HDL-C and presence of CAC [29]. In yet another study, Kowamoto et al evaluated the relationship between non-HDL-C and carotid atherosclerosis as assessed on carotid duplex imaging. The study showed similar results, with non HDL-C being a better predictor of carotid atherosclerosis than LDL-C [30]. The PDAY (Pathological Determinance of Atherosclerosis in Youth) study [31] provided further evidence about the relationship between non-HDL-C and subclinical atherosclerosis. The study correlated autopsy findings with various lipid parameters in young individuals who had died of external causes. It found that non-HDL-C was significantly associated with presence of fatty streaks at the site of examination (i.e. thoracic and abdominal aorta and right coronary artery) but no significant association was found between Apo-B and fatty streaks.

Lipid lowering therapy and residual CV risk: Several patients who are on statins and who succeed in achieving the “target” LDL-C levels may still suffer from CV events [11]. These patients bear the brunt of having residual CV risk, not identified by traditional lipid markers. It has been shown that ‘on treatment’ LDL-C levels in patients on lipid lowering therapy do not very accurately reflect the residual CV risk whereas non-HDL-C and Apo-B continue to remain relatively accurate predictors of CV events even in this setting [12,32]. A combined analysis of the TNT (Treating to new Targets) and IDEAL (Incremental Decrease in Endpoints through Aggressive Lipid Lowering) studies involving 18018 patients clearly demonstrated a much stronger relationship between the risk of CV events and the‘on-treatment’ levels of non-HDL-C and Apo-B, as compared to LDL-C [12]. Yet another combined analysis of 8 trials showed that in statin-treated patients, though the levels of LDL-C, non–HDL-C, and Apo B were each associated with the risk of future major adverse CV events, the strength of this association was greater for non–HDL-C than for LDL-C (p=0.002) and Apo B (p=0.02) [32].

Treating elevated non-HDL-C

Since LDL-C is the main component of non-HDL-C, the initial treatment approach for elevated non-HDL-C is same as that for LDL-C [9]. Thus, aggressive life-style measures and statins form the mainstay of treatment. If non-HDL-C remains elevated despite adequate LDL-C lowering, further intensification of life-style measures and increase in the statin dose, if not already on maximum tolerated dose, should be tried. Weight reduction, smoking cessation and abstinence from alcohol are all quite effective in lowering triglycerides (and thus non-HDL-C) and should be actively pursued.In addition, at the same time, one should also diligently look for and correct any secondary causes of hypertriglyceridemia if present, such as uncontrolled diabetes, nephrotic syndrome, chronic renal failure, certain drugs (e.g. corticosteroids, protease inhibitors for human immunodeficiency virus infection, beta blockers, estrogens) etc [9].

If the non-HDL-C level still remains high in spite of abovemeasures, addition of a fibrate is a useful option. Although the most recent American College of Cardiology/ American Heart Association guidelines have discounted the role of non-statin drugs in the current management of dyslipidemia [10], there is reasonable evidence to suggest that fibrates may provide incremental CV risk reduction in patients with elevated triglyceride levels [33-35]. For example, in the ACCORD (Action to Control Cardiovascular Risk in Diabetes) Lipid study, when fenofibrate was added to the background statin therapy, it significantly reduced the incidence of CV events in patients who had atherogenic dyslipidemia [33]. This is consistent with the data from older studies comparing gemfibrozil and bezafibrate with placebo, which showed significant CV risk reduction with these agents, with more marked effects seen in those with elevated triglyceride levels [34,36-38].

Implications for clinical practice

Enough evidence has accumulated over the last decade to suggest that non-HDL-C may be a better predictor of CV risk than LDL-C in different patient populations. Basis these studies, many investigators have suggested to elevate the role of non-HDL-C as the primary target of lipid lowering therapy instead of LDL-C. The value of non-HDL-C in routine clinical practice was underscored by the recent consensus report from the American Diabetes Association and the American Heart Association that stated “routine calculation and use of non- HDL-C constitute a better index than LDL-C for identifying high risk patients and that calculation of non-HDL-C should be provided in all laboratory reports’’[24].

An additional advantage of using non-HDL-C as the primarytarget for therapy is that it does not involve any deviation from the current practice and does not attempt to undermine the significance of LDL-C [9]. As already mentioned above, the initial treatment strategy for elevated non-HDL-C is same as that for elevated LDL-C, irrespective of whether LDL-C is the primary target for therapy or non-HDL-C. However, using non-HDL-C as the main target of therapy ensures appropriate corrective measures in patients who continue to have elevated non-HDL-C levels despite having achievedLDL-C target. Although this too is included in the current guidelines as the secondary target for therapy, it often gets overlooked because of continued overemphasis on LDL-C alone.

Though Apo-B has been shown to be a better predictor of CVrisk than non HDL-C or LDL-C, its application in routine clinical practice is difficult. Measurement of Apo-B requires a special assay which is expensive, time-consuming, not readily available and is still not standardized. Moreover, there is yet to be a consensus on optimum target levels of Apo-B in different patient populations [24,39,40]. Given these practical limitations, non-HDL-C remains the best alternative for regular clinical use.