The emerging role of high-density lipoprotein cholesterol in hypertension trials

The individual risk of cardiovascular disease (CVD) may be conditioned by the joint effect of several risk factors such as hypertension and dyslipidemia [1–3]. On the contrary, the relation between time-varying changes in the phenotypic expression of different risk factors and cause-specific outcome events is less well established. Although blood pressure (BP) reduction translates into a clearly lesser risk of both coronary and cerebrovascular events [4,5], it is not clear whether the correction of lipid abnormalities is protective from cause-specific complications different from coronary heart disease (CHD). In particular, although the increase in high-density lipoprotein cholesterol (HDL-c) translates into a less risk of coronary events, its impact on cerebrovascular disease is still unclear [6]. Observational data from Framingham [7] and the Atherosclerosis Risk in Communities study [8] showed that, at any level of low-density lipoprotein cholesterol (LDL-c), the lower the HDL-c, the higher the risk of CHD. Risk increased by almost three-fold when LDL-c was 220 mg/dl and HDL-c was 25 mg/dl, but HDL-C remained an important risk factor even in patients with low LDL-c levels [7,8]. Combining the results of four major studies (the Framingham Heart Study, the Lipid Research Clinics Prevalence Mortality Follow-up Study, the Lipid Research Clinics Coronary Primary Prevention Trial and the Multiple Risk Factor Intervention Trial), after adjustment for age, BP, smoking, BMI and LDL-c, for each 1-mg/dl increase in HDL-c, there was a 2% reduction in the risk of CHD in men and a 3% reduction in women [9]. The association between cholesterol and stroke remains controversial. A study by Wannamethee et al. [10] examined the relation between serum HDL-c and risk of stroke in 7735 men, 40–59 years of age. During a mean follow-up period of 16.8 years, 343 stroke cases were observed and higher levels of HDL-c were associated with a significantly less risk of stroke even after adjustment for potential confounders. The Northern Manhattan Stroke Study [11] reported an inverse association between HDL-c and ischemic stroke in elderly participants and among different racial or ethnic groups. More recently, the EUROSTROKE collaborative project showed a different relation of HDL-c to the risk of stroke in men as compared with women [12]. In men, there was a general trend towards a less risk of stroke with an increase in HDL-c [odds ratio per 1 mmol/l increase in HDL-c: 0.68, 95% confidence interval (CI) 0.40–1.16], whereas in women an increase in HDL-c was associated with a significantly higher risk of nonfatal stroke and of cerebral infarction [12]. However, such an increase, which means doubling of HDL-c levels (1 mmol/l = 38.6 mg/dl) is rarely seen in clinical practice. The Anglo-Scandinavian Cardiac Outcomes Trial-Blood Pressure Lowering Arm (ASCOT-BPLA) [13] was the first study to demonstrate that in-treatment HDL-c may contribute to explain the outcome benefit of a given treatment over another (in this case, amlodipine ± perindopril vs. atenolol ± bendrofluazide). In this study, the unadjusted 14% lesser risk of primary endpoint and coronary revascularization with amlodipine compared with atenolol (P = 0.005) decreased to a nonsignificant 10% (P = 0.06) after adjustment for in-treatment HDL-c [13]. In contrast, the unadjusted 13% lesser risk of stroke with amlodipine raised to 14% after correction for HDL-c [13]. Prognostic impact of high-density lipoprotein cholesterol levels in hypertension Low levels of HDL-c are frequent in hypertensive patients [3]. In the Progetto Ipertensione Umbria Monitoraggio Ambulatoriale (PIUMA) study, a high proportion (47.7%) of participants showed lipid abnormalities defined by a total cholesterol (TC) at least 6.22 mmol/l, HDL-c less than 1.05 mmol/l or LDL-c at least 4.13 mmol/l. Notably, 17 and 78% of participants included in the study had HDL-c levels below 1.05 mmol/l and below 1.55 mmol/l, respectively [3]. This study confirmed an inverse association between HDL-c and the risk of CHD, but no association emerged with the risk of cerebrovascular disease. After adjustment for several confounders, including left ventricular hypertrophy (LVH) and 24-h ambulatory pulse pressure (PP), the hazard ratio (HR) for cardiac events was 2.23 with a HDL-c less than 1.05 mmol/l (95% CI 1.06–4.70), but neither TC nor its fractions did show significant univariate or multivariate associations with cerebrovascular events [3]. An important post-hoc analysis of the Losartan Intervention For Endpoint reduction in hypertension (LIFE) study [14], published in this issue of the Journal of Hypertension, examined the prognostic impact of time-varying changes in HDL-c (determined annually during 4.8 years of follow-up) in 8611 patients with hypertension and LVH. HDL-c was higher (P < 0.001) with losartan (1.38, 1.37, 1.42, 1.47 and 1.48 mmol/l) than with atenolol (1.32, 1.30, 1.36, 1.40 and 1.42 mmol/l) at the scheduled annual visits, thereby confirming previous reports that β-blockers may exert detrimental effects on lipids, whereas angiotensin receptor blockers have neutral or favorable effects [15–18]. In both treatment groups, HDL-c decreased during the first 2 years of the study, to further increase thereafter, possibly reflecting initiation of diuretic therapy. However, as this analysis has been carried out according to an intention-to-treat approach, the potential impact of protocol violations (treatment discontinuations, etc.) should be considered when interpreting findings. When treatment allocation (losartan vs. atenolol) and in-study HDL-c were added to the same survival model, losartan was no longer predictive of a less risk of primary outcome, whereas higher levels of HDL-c were significantly associated with a lesser risk of primary outcome. These data suggest that about 36% of the protective effect of losartan over atenolol on the primary LIFE endpoint was explained by the higher HDL-c. A decrease in HDL-c by more than 0.098 mmol/l (median value) from baseline to follow-up was associated with a 67% higher risk of the primary study outcome, a 60% higher risk of death, a 70% higher risk of myocardial infarction (MI) and a 83% higher risk of stroke. Notably, baseline HDL-c above 1.43 mmol/l (median) was associated with a 51, 57, 42 and 48% reduction in the composite primary endpoint, CVD death, stroke and MI, respectively. These data are consistent with the ASCOT study [13] and suggest that antihypertensive drugs might be divided into the two broad categories of ‘lipid-hostile’ and ‘lipid-beneficial/neutral’ treatments. As the BP reduction was comparable in the LIFE study in the two treatment arms, another important message of this study is that the favorable prognostic impact of BP reduction by a given drug (in this case atenolol) may be outweighed by an adverse impact on lipids. Old and new drugs Although there is a general consensus that diuretics and nonselective β-blockers adversely affect lipid levels [15–18], disagreement still exist about the effects of other antihypertensive drug classes on lipids. A systematic overview by Kasiske et al. [15] showed that diuretics decrease HDL-c levels only in patients with diabetes and that such adverse effect does not depend on the duration of therapy. The estimated reduction in HDL-c levels (−0.10 mmol/l to approximately −3.9 mg/dl) during treatment with β-blockers is smaller than that of diuretics and this effect diminished with duration of treatment [15]. On the contrary, new antihypertensive drug classes [angiotensin-converting enzyme inhibitors (ACE-Is), angiotensin receptor blockers (ARBs) and calcium-channel blockers (CCBs)) exert little, if any, effects on lipids and HDL-c levels [15]. The effects of different drug classes on HDL-c levels were also examined in at least six recent randomized clinical trials (RCTs) [13,14,19–22], which compared ‘new’ drug classes (ACE-Is, ARBs or CCBs) with diuretics, β-blockers or both. In four RCTs [13,19–21], experimental treatment regimen was based on CCBs, whereas in the remaining two RCTs [14,22], the effect of drugs active on the renin–angiotensin system (RAS) was compared with β-blockers. Generally, the use of a ‘new’ drug was associated with a neutral or positive effect on HDL-c levels during treatment when compared with the reference treatment. Only in the National Intervention Cooperative Study (NICS) in elderly hypertensives trial [20], the use of a ‘new’ antihypertensive treatment (nicardipine) was associated with a smaller increase in HDL-c from baseline to follow-up when compared with a diuretic (0.06 vs. 0.14 mmol/l). In this trial, BP decreased more with the diuretic than with nicardipine. Therefore, a more pronounced decrease in systolic BP was associated with an improvement in HDL-c levels. Relation between blood pressure and high-density lipoprotein cholesterol The hypothesis of a relation between BP and HDL-c is not new [23–26], and at least four possible mechanistic links have been hypothesized: high BP may be associated with insulin resistance, which, in turn, may be associated with lipid abnormalities [23]; both hypertension and dyslipidemia directly affect the contraction-coupled and replication-coupled processes in the vascular smooth muscle cells, possibly through an effect on the phosphoinositide signal transduction in smooth muscle cells [24]; high levels of BP and low levels of HDL-c may trigger endothelial dysfunction through different mechanisms, including a reduced bioavailability of endothelium-derived nitric oxide [25]. Indeed, cholesterol lowering and BP lowering may improve endothelium-dependent vasodilatation [25]; high BP and dyslipidemia are associated with atherosclerotic processes, reflected by high values of PP, in the large arteries [26]. In conclusion, the strong prognostic impact of HDL-c and its modifications by treatment on the risk of coronary events in hypertensive patients justifies any effort directed at raising HDL-c through life-style measures, statins and antihypertensive drugs possibly devoid of adverse impact on lipid metabolism. After the ASCOT [13] and LIFE [14] studies, it is out of question that stable, or possibly raised, HDL-c levels during treatment of hypertension imply a highly favorable legacy. Further studies should also address the hypothesis of an association between BP reduction and increase in HDL-c. Acknowledgements This study was funded in part by the Fondazione Umbra Cuore e Ipertensione – ONLUS, Perugia, Italy. There are no conflicts of interest.