Genetic, epidemiologic and clinical data strongly suggest that fasting or non-fasting triglycerides are independent cardiovascular risk factors

Keywords: :

• Cardiovascular disease
• Epidemiology
• Gene loss of function
• Interventional trials
• Treatment targets
• Triglycerides

Two recently published studies in the same issue of the New England Journal of Medicine analyzed the effects of rare mutations in the gene encoding for apolipoprotein C3 (Apo C3)1^,2. In the first study, 18,666 genes were evaluated in 3734 participants of European or African origin within the Exome Sequencing Project1. Four loss-of-function mutations were found, i.e. resulting in 46% lower circulating levels of Apo C3. As expected, these mutations were related to low plasma triglyceride (TG) levels, i.e. 39% lifelong lower TG levels (42% in subjects of European ancestry and 32% in subjects of African ancestry) than non-carriers of these mutations. In this cohort, 1/150 subjects was a heterozygote for at least one of these four loss-of-function mutations1. The authors then compared the incidence of overt coronary heart disease (CHD) between carriers of the loss-of-function Apo C3 mutations and the 110,472 participants who were non-carriers1. Among people having any loss-of-function Apo C3 mutation, the incidence of CHD was 40% lower than among non-carriers (odds ratio [OR] 0.60, 95% confidence interval [CI] 0.47–0.75, p = 4 × 10⁻⁶])1.

The second study assessed the relation of non-fasting TGs with CHD risk in 75,725 participants in two general population studies2. Subjects with non-fasting TG levels <90 mg/dl (1.0 mmol/l) had lower incidence of cardiovascular disease (CVD) than those with TG levels >350 mg/dl (4.0 mmol/l) (hazard ratio [HR] 0.43; 95% CI 0.35–0.54)2. This study also investigated whether the loss-of-function mutations in Apo C3 were related to CHD risk2. During follow-up, the incidence of CHD was lower in heterozygotes for Apo C3 loss-of-function mutations compared with non-carriers (HR 0.43, 95% CI 0.35–0.54, p = 0.007)2.

Collectively, these two studies showed that loss-of-function or null mutation in Apo C3 is associated with a favorable cardioprotective lipid profile, mainly characterized by lower fasting or non-fasting TG levels3. It has also been reported that a loss-of-function mutation in the Apo C3 gene is associated with markedly reduced TG and remnant cholesterol levels and reduced prevalence of coronary calcification, a surrogate marker for CHD4.

Genetic evidence that TG-rich lipoproteins are causally related to CHD was also recently presented5. These findings suggest that Apo C3 might represent a treatment target to reduce residual CVD risk1–5. Finally, a Mendelian randomization study, based on data from the Copenhagen City Heart Study with 24 years of follow-up, showed that subjects with genetically defined lower non-fasting plasma TGs have reduced all-cause mortality compared with participants with high TG levels after adjustment for various cofounders (HR 0.59, 95% CI 0.51–0.68 for individuals with non-fasting TGs <89 mg/dl [<1.0 mmol/l])6. These genetic studies suggest that lower TG levels are associated with lower risk for CVD.

There is evidence from large meta-analyses supporting the role of both high fasting or non-fasting TGs as CVD risk factors7–10. Particularly non-fasting TGs show an independent and graded relationship with CVD risk; elevated TGs measured 4 h postprandially are associated with a 4.5-fold increase in CVD risk compared with low TG levels7. In the Atherosclerosis Risk in the Communities (ARIC) study (12,339 subjects followed up for 10 years), fasting TG levels predicted CHD events in women but not in men11, while in the Reykjavik Study (18,569 subjects followed up for 20 years) fasting TG levels were associated with CHD only in men9. The European Prospective Investigation of Cancer (EPIC) – Norfolk Study (25,668 subjects followed up for 8 years) reported an association of non-fasting TG levels with increased risk for CHD in both genders9.

Although there is evidence supporting the role of fasting and non-fasting TGs as CVD risk factors this does not necessarily mean that lowering TGs with drugs will reduce the incidence of CVD events. Indeed, epidemiological and genetic data also suggested that low high density lipoprotein cholesterol (HDL-C) levels are an independent CVD risk factor12; however, substantial increases in HDL-C levels with torcetrapib, a cholesterol ester transfer protein (CETP) inhibitor, led to an increase in CVD mortality13. Recently, 2 g of extended-release niacin and 40 mg of laropiprant significantly increased HDL-C levels compared with placebo14, but had no significant effect on the incidence of major CVD events (13.2 vs 13.7%, respectively; HR 0.96, 95% CI 0.90–1.03, p = 0.29)14. Moreover, niacin increased the risk of serious adverse events (absolute increase compared with placebo, 3.7%; p < 0.001)14. These results were attributed to off-target effects of drugs but it is also possible that they suggest that the pharmacological modulation of a CVD risk factor (as established by epidemiological data) will not necessarily reduce CVD morbidity and mortality, especially when low density lipoprotein cholesterol (LDL-C) levels are very low (<70 mg/dl [1.8 mmol/l])13^,14.

In a single centre study, high TG and low HDL-C levels were associated with CHD in patients with a LDL-C <70 mg/dl (1.8 mmol/l)15. Therefore, it appears that high TG levels and to a lesser degree low HDL-C levels contribute to the residual CVD risk that is present in patients who achieve very low LDL-C levels with statin treatment15. The joint analysis of two secondary prevention trials, the Incremental Decrease in End Points through Aggressive Lipid Lowering (IDEAL, n = 8888) and the Treating to New Targets (TNT, n = 10,001), yielded similar results16. In this analysis, even slightly increased TG levels were associated with a higher risk of recurrent CVD events in statin-treated patients, many of whom had LDL-C levels <70 mg/dl (1.8 mmol/l)16. These results were also replicated by the Pravastatin or Atorvastatin Evaluation and Infection Therapy–Thrombolysis In Myocardial Infarction (PROVE IT)-TIMI 22 trial, which included 4162 patients with acute coronary syndrome (ACS)17. Patients with on-treatment TGs <150 mg/dl and LDL-C <70 mg/dl had lower CHD risk than patients with higher TG levels (HR 0.72, 95% CI 0.54–0.94, p = 0.017)17. These findings suggest that achieving low TGs should be an additional consideration after achieving low LDL-C levels in patients with ACS17. Moreover, excessive myocardial TG accumulation may lead to secondary diastolic dysfunction and heart failure with preserved ejection fraction18.

Thus, high TG levels remain an independent CVD risk factor even if LDL-C levels are very low (e.g. <70 mg/dl; 1.8 mmol/l). However, is there any effective treatment for elevated TG levels that also unequivocally reduces CVD events? In patients with type 2 diabetes mellitus (T2DM) on statins, the Pioglitazone Effect on Regression of Intravascular Sonographic Coronary Obstruction Prospective Evaluation (PERISCOPE) intravascular ultrasound (IVUS) study reported favorable effects of pioglitazone on the TG/HDL-C ratio, which was correlated with delayed atheroma progression19. This finding highlights the potential importance of targeting atherogenic dyslipidemia in diabetic patients with CHD19.

However, the findings of the PERISCOPE study19 refer to a surrogate marker of CHD. On the other hand, fibrates appear to be the treatment of choice for all kinds of hypertriglyceridemia, including mixed dyslipidemia, as monotherapy or in combination with statins without an increase in adverse effects20. In the lipid arm of the Action to Control Cardiovascular Risk in Diabetes study (ACCORD Lipid), the combination of fenofibrate and simvastatin did not reduce the primary endpoint (fatal and non-fatal CVD events) compared with simvastatin monotherapy in the overall study population (n = 5518 patients with T2DM)21. However, the subgroup of participants with atherogenic dyslipidemia (upper tertile of TGs and lower tertile of HDL-C levels) had 70% higher CVD risk than patients without atherogenic dyslipidemia and in this group the fenofibrate–simvastatin combination reduced the primary endpoint by 31% more than simvastatin monotherapy21. Nevertheless, this group of patients comprised only 17% of the ACCORD Lipid population (941 of 5518 patients) and did not affect overall outcome21. However, in everyday clinical practice, a substantial proportion of patients with T2DM have atherogenic dyslipidemia. Thus, the inclusion criteria of a randomized controlled trial might have contributed to the underestimation of the actual benefit of fibrate–statin combination in patients with T2DM in a ‘real life’ setting. In a recent post-hoc analysis of the ACCORD Lipid Trial, fenofibrate + simvastatin lowered postprandial TGs to the same extent as simvastatin monotherapy22. However, the levels of atherogenic apo-B48 particles were reduced only in participants on statin–fibrate combination with increased fasting TG levels22. Therefore, it appears that in patients with atherogenic dyslipidemia, the use of fibrates in combination with statins reduces CVD risk; on the other hand, in patients without atherogenic dyslipidemia this favorable effect is absent and fibrates may not benefit these patients23^,24.

A meta-analysis of six trials including >25,000 patients investigated the effects of fibrates on CVD risk25. Compared with placebo, the greatest benefit of fibrate treatment was seen in patients with high TGs (n = 7389), in whom fibrates reduced the risk of CVD events by 25% (95% CI 0.65–0.86, p < 0.001), as well as in patients with both high TGs and low HDL-C levels (n = 5068), in whom the risk reduction was 29% (95% CI 0.62–0.82, p < 0.001). A smaller clinical benefit was observed in patients with isolated low HDL-C (n = 15,303) (RR 0.84, 95% CI 0.77–0.91, p < 0.001)25. Among the 9872 patients with neither high TGs nor low HDL-C levels, fibrate treatment did not reduce CVD events (RR 0.96, 95% CI 0.85–1.09, p = 0.53)25. These findings support the concept that fibrates as monotherapy or in combination with statins reduce CVD risk in patients with atherogenic dyslipidemia but not in patients with isolated hypercholesterolemia23^,25.

Another important issue is whether TGs should be measured in the fasting state or postprandially and what the treatment targets should be26–29. A prospective study of 26,330 healthy women (19,983 fasting; 6347 non-fasting) showed that with the exception of TGs, other lipids (total cholesterol, LDL-C, apoB100, HDL-C, apoA-1 and total/HDL-C ratio) differed minimally (<5%) in the fasting and the non-fasting state and had similar associations with CVD risk29. In contrast, non-fasting TGs had a stronger association with CVD risk than fasting TGs29. Another analysis of 33,391 individuals (aged 20–95 years) from the Copenhagen General Population Study showed that non-fasting lipid profiles predicted increased risk of CVD events30. In the same study, 7587 women and 6394 men were followed up for 26 years; 1793 participants had a myocardial infarction (MI), 3479 developed CHD and 7818 died. Each 1 mmol/l increase in non-fasting TG levels was associated with an adjusted HR for MI, CHD and death ranging between 1.4 and 5.6 in both men and women31. This increase in risk was evident from non-fasting TGs >2.0 mmol/l (≈180 mg/dl)28. During the 31 year follow-up of the Copenhagen population study, stepwise increases in non-fasting cholesterol and TGs levels were associated to a similar degree with increasing risks for CVD32. Interestingly, non-fasting TG levels predicted CVD risk better in women and non-fasting cholesterol levels predicted risk better in men32.

Regarding the time after the meal that non-fasting TG levels better predict CVD risk, it appears that TG concentrations measured 4 h after the last meal have the strongest association with CVD events (HR 4.48, CI 95% 1.98–10.15, p for trend <0.001 for the highest tertile compared with the lowest)33.

When hypertriglyceridemic patients, especially those with low HDL-C34^,35, have chronic kidney disease (CKD) or do not want to take a fibrate, the administration of high potency statins might be a solution36^,37, especially if we use statins appropriate for patients with CKD38.

In conclusion, it would be reasonable to suggest that in future guidelines, after achieving the LDL-C target, we might probably aim to reduce non-fasting TG levels when they are >2.0 mmol/l (≈180 mg/dl). TG levels <1.6 mmol/l (142 mg/dl)36 or even lower appear to be appropriate as a secondary treatment goal, and not just as a desirable concentration. Non-HDL-C levels might serve as a tertiary target. If this suggestion is verified in prospective analyses of large interventional trials, targeting TG levels might significantly contribute to the reduction of the residual cardiovascular risk on top of statin treatment.

Transparency

Declaration of funding

This editorial was written independently. The authors did not receive financial or professional help with the preparation of the manuscript.

Declaration of financial/other relationships

Some of the authors have attended conferences, given lectures and participated in advisory boards or trials sponsored by various pharmaceutical companies. V.G.A. has disclosed that he is a member of CMRO’s International Advisory Board. D.P.M. has disclosed that he is CMRO’s Editor-in-Chief. K.T. and A.K. have disclosed that they have no significant relationships with or financial interests in any commercial companies related to this study or article.

The CMRO peer reviewer on this manuscript has no relevant financial or other relationships to disclose.