Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Feb 12.
Published in final edited form as: Eur J Prev Cardiol. 2018 Dec 6;26(5):533–543. doi: 10.1177/2047487318816495

Effects of high-density lipoprotein targeting treatments on cardiovascular outcomes: A systematic review and meta-analysis

Haris Riaz 1,#, Safi U Khan 2,#, Hammad Rahman 3, Nishant P Shah 1, Edo Kaluski 4, A Michael Lincoff 1, Steven E Nissen 1
PMCID: PMC7879587  NIHMSID: NIHMS1621088  PMID: 30861690

Abstract

Background:

The effects of increasing high-density lipoprotein cholesterol on cardiovascular outcomes remain uncertain.

Design:

We conducted a meta-analysis to investigate the effects of high-density lipoprotein cholesterol modifiers (niacin, fibrates and cholesteryl ester transfer protein inhibitors) on cardiovascular outcomes.

Methods:

Thirty-one randomized controlled trials (154,601 patients) with a follow-up of 6 months or more and a sample size of 100 or more patients were selected using MEDLINE, EMBASE and CENTRAL database (inception January 2018).

Results:

High-density lipoprotein cholesterol modifiers had no statistically significant effect on cardiovascular mortality in terms of relative risk (RR) (RR 0.94, 95% confidence interval (CI) 0.89–1.00, P = 0.05, I2 = 13%) or absolute risk (risk difference −0.0001, 95% CI −0.0014, 0.0011, P = 0.84, I2 = 28%). High-density lipoprotein cholesterol modifiers reduced the RR of myocardial infarction (RR 0.87, 95% CI 0.82–0.93, P < 0.001, I2 = 37%). This significant effect was derived by the use of fibrates (RR 0.80, 95% CI 0.73–0.87, P < 0.001, I2 = 22%) and meta-regression analysis showed that this benefit was consistent with an absolute reduction in low-density lipoprotein cholesterol. High-density lipoprotein cholesterol modifiers had no effect on stroke (RR 1.00, 95% CI 0.93–1.09, P = 0.94, I2 = 25%) or all-cause mortality (RR 1.02, 95% CI 0.97–1.08, P = 0.48, I2 = 49%). Meta-regression analyses failed to demonstrate a significant association of pharmacologically increased high-density lipoprotein cholesterol with key endpoints. In studies with background statin therapy, high-density lipoprotein cholesterol modifiers had no statistically significant impact on cardiovascular mortality, myocardial infarction, stroke or all-cause mortality (P > 0.05).

Conclusion:

The use of high-density lipoprotein cholesterol modifying treatments had no significant effect on cardiovascular mortality, stroke or all-cause mortality. The beneficial effect on myocardial infarction was lost when drugs were used with statin therapy.

Keywords: Niacin, fibrates, CETP inhibitors, meta-analysis

Introduction

The association of high-density lipoprotein (HDL) cholesterol and the risk of coronary heart disease (CHD) remain uncertain. A number of observational studies have demonstrated reduced coronary events with increased levels of HDL cholesterol;1–3 while recent population-based data suggested that patients having extremely elevated HDL cholesterol paradoxically had a higher incidence of all-cause mortality.4 Numerous clinical trials have explored the potential beneficial effects of pharmacological therapies to raise HDL cholesterol. Initial trials of drugs that increase HDL cholesterol such as niacin or fibrates reported a reduced incidence of major cardiovascular events.5–7 However, in the contemporary statin era, similar trials on niacin and fibrates were neutral or showed harm. Despite these outcomes, the scientific enthusiasm on HDL cholesterol increasing therapies has persisted, resulting in the development of cholesteryl ester transfer protein (CETP) inhibitors. The CETP inhibitor trials so far have also shown mixed results, ranging from negative to neutral to marginally positive. In the current study, we performed a systematic review and meta-analysis to assess the impact of HDL cholesterol increasing therapies on the risk of cardiovascular outcomes.

Methods

This meta-analysis was performed and reported in accordance with Cochrane collaboration guidelines8 and preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines.9

Data sources and searches

Two authors (SUK and HR) developed the search strategy. We searched MEDLINE, EMBASE and the Cochrane Central Register of Controlled Trials (CENTRAL) databases from inception to 15 January 2018. We used a modified broad search strategy by using relevant keywords (‘fibrates’, ‘niacin’, ‘cholesteryl ester transfer protein inhibitor’, ‘CETP’, ‘clofibrate’, ‘fenofibrate’, ‘gemfibrozil’, ‘bezafibrate’, ‘torcetrapib’, ‘dalcetrapib’, ‘evacetrapib’, ‘anacetrapib’, ‘high density lipoprotein cholesterol’, ‘HDL-C’ and ‘cardiovascular disease’). We applied restrictions on randomized controlled trials (RCTs) and humans. No restrictions were applied on language, year of publication or text availability. Additional sources included online libraries of www.clinicaltrialresults.com, www.esccardio.org, www.cardiosource.org, www.clinicaltrials.gov, proceedings of major cardiovascular meetings and bibliographies of relevant articles. The citations were downloaded in Endnote X 7 (Thompson ISI ResearchSoft, Philadelphia, PA, USA) and duplicates were identified and removed through EndNote and manually. Two authors (SUK and HR) independently screened studies based on prespecified inclusion criteria. Any disagreements were resolved by mutual consensus or third party review.

Study selection

The prespecified inclusion criteria were: (a) RCTs comparing HDL cholesterol targeting treatments (niacin, fibrates and CETP inhibitors); (b) follow-up of 6 months or greater (minimum duration required by a drug to demonstrate a clinical benefit);10 (c) sample size 100 or more patients (to avoid small study effects);11 and (d) studies had to report at least one clinical event for outcomes of interest (see below). We excluded non-RCTs and studies comparing treatments which predominantly act on low-density lipoprotein (LDL) cholesterol, i.e. statins, ezetimibe, proprotein convertase subtilisin/kexin type 9 inhibitors and bile acid sequestrants.

Quality assessment and data extraction

The data abstraction was done by two authors (SUK and HR) independently on a prespecified data collection form. The following information was extracted: baseline characteristics of participants, treatment groups, events, sample size, change in HDL cholesterol levels, LDL cholesterol levels and triglycerides in each group and difference between the groups, and follow-up duration of each trial. In older studies, in which LDL cholesterol was not reported, LDL cholesterol was estimated from total cholesterol (TC) using the regression equation: LDL cholesterol = (TC) *[(TC) * 0.0012 + 0.3793].12

The absolute change in lipid parameters was calculated as the mean difference averaged over the course of follow-up between two interventions. If not available, we then used the mean achieved lipoprotein value at the point close to 50% of the median follow-up.12 When possible, we focused on estimates adjusted for baseline measurements.13 When available, we extracted data for intention to treat analysis. Disagreements related to data were resolved by discussion, referring back to the original article or opinion of the third author. The quality of the trials was assessed using the Cochrane bias risk assessment tool14 (Supplementary Table 1).

The primary outcome was cardiovascular mortality. The secondary endpoints were myocardial infarction (MI), stroke, all-cause mortality, change in HDL cholesterol, LDL cholesterol and triglyceride levels. The definitions of the endpoints were taken as reported in the individual trial.

Data synthesis and analysis

Outcomes were combined using the DerSimonian and Laird random effects model.15 Dichotomous estimates were reported as risk ratio (RR) and risk difference (RD), and continuous variables were calculated as mean difference (MD) with 95% confidence interval (CI). Heterogeneity was assessed using Cochrane Q statistics and was quantified by I2 with values greater than 25%, 50% and 75% consistent with a low, moderate and high degree of heterogeneity, respectively.16 Publication bias was assessed using funnel plot and Egger’s regression test.17 All analyses were conducted at a 5% significance level.

Subgroup analyses were performed to explore the effect of background statin therapy on all outcomes. For estimating the expected change in effect size per change in HDL cholesterol levels, meta-regression analyses were conducted for absolute and percentage increase and log RR of each outcome using random effects models with the restricted maximum likelihood estimation. The Knapp and Hartung adjustment was applied for estimation of standard errors of the estimated coefficients to calculate summary effect estimates.18 The index R2 value (defined as the ratio of explained variance to total variance) was used to report the proportion of variance accounted for by the change in lipoprotein values. For significant outcomes, we performed sensitivity analyses by conducting meta-regression between an absolute reduction in LDL cholesterol and the clinical endpoint to assess variation in outcome due to change in LDL cholesterol. Comprehensive meta-analysis software version 3.0 (Biostat, Englewood, NJ, USA) was used for all analyses.

Results

Initial electronic search yielded 1131citations, 158 citations were removed as duplicates, and out of the remaining 973 articles 565 records were excluded at title and abstract level screening. A further 377 full text articles were excluded based on priori inclusion/exclusion criteria. Ultimately, 31 RCTs (154,601 patients) were included in this meta-analysis (Figure 1). The Coronary Drug Project19 was a three arm study comparing fibrates, niacin and placebo and thus it contributed the data for both fibrates and niacin. Overall, 10 trials (78,542 patients) assessed CTEP inhibitors,20–29 16 trials (41,765 patients) assessed fibrates5,6,19,30–42 and six trials(34,294 patients) assessed niacin.7,19,43–46 Sixteen trials used background statin therapy20–29,40,42–46 and 15 had no statin therapy.5–7,19,30–39,41 The mean age of study participants was 58.4 ± 6.7 years and 78% were men. The pooled mean baseline HDL cholesterol was 42.7 ± 4.9 mg/dL. The mean follow-up duration was 3.2 ± 1.6 years (Table 1). The funnel plot (Supplementary Figure 9) and Egger’s regression test did not detect publication bias (P (two-tailed) > 0.05) (Supplementary Table 2).

Figure 1.

Figure 1.

Open in a new tab

Search strategy based on PRISMA guidelines.

Table 1.

Clinical characteristics of the study participants, number of patients (%) and mean (SD) or median values (IQR) where available.

Studies (year) Arms N Age (years) Men (%) CHD (%) HTN (%) DM (%) Baseline HDL-C (mg/dL) Statin therapy (%) HDL-C increase from baseline (%) Achieved LDL-C (mmol/L) Follow-up (months)
ILLUMINATE (2007)28 Torcetrapib 7533 61.3±7.6 5854 (77.7) 4481 (61.1) 5423 (72.3) 3271 (43.5) 48.6±12.0 100 72 0.53 12
Placebo 7534 61.3±7.6 5861 (77.8) 4497 (60.0) 5554 (73.9) 3390 (45.2) 48.5±12.2
ILLUSTRATE (2007)25 Torcetrapib 591 56.9±9.1 416 (70.4) 591 (100.0) 440 (74.5) 119 (20.1) 46.0±12.8 100 61 0.44 24
Placebo 597 57.0±9.2 421 (70.5) 597 (100.0) 463 (77.6) 133 (22.3) 45.2±11.2
RADIANCE 1 (2007)26 Torcetrapib 450 46.8±12.0 214 (47.6) – 110 (24.4) 12 (2.7) 52.9±12.7 100 52 0.70 24
Placebo 454 45.2±12.9 232 (51.1) – 114 (25.1) 19 (4.2) 51.8±12.8
RADIANCE 2 (2007)27 Torcetrapib 377 57.9±8.1 237 (63.0) – 193 (51.0) 68 (18.0) 47.6±11.2 100 63 0.23 24
Placebo 375 56.5±8.2 245 (65) – 185 (49.0) 92 (25.0) 47.6±9.7
DEFINE (2010)24 Anacetrapib 811 62.5±8.7 629 (77.6) 447 (55.1) 560 (69.1) 430 (53.0) 40.5±9.3 99 139 0.83 18
Placebo 812 62.9±9.0 618 (76.1) 441 (54.3) 541 (66.6) 432 (53.2) 40.4±9.1
Dal-PLAQUE (2011)22 Dalcetrapib 64 62.6±8.2 51 (80) 57 (89) 47 (73) 19 (30) 46.2±11.5 87 31 0.00 24
Placebo 66 64.6±7.8 55 (83) 54 (82) 48 (73) 20 (30) 46.2±15.4
Dal-OUTCOMES (2012)21 Dalcetrapib 7938 60.3±9.1 6365 (80.2) 7,938 (100.0) 5336 (67.0) 1930 (24.0) 42.5±11.7 97 40 0.00 31
Placebo 7933 60.1±9.1 6436 (81.1) 7933 (100.0) 5419 (68.0) 1952 (25.0) 42.2±11.5
Dal-VESSEL (2012)23 Dalcetrapib 232 62.3±7.0 211 (91.0) 147 (63.0) 171 (74.0) 108 (47.0) 39.2±7.3 95 31 1.38 9
Placebo 234 61.9±7.9 211 (90.0) 155 (66.0) 175 (75.0) 102 (44.0) 38.5±7.2
ACCELERATE (2017)20 Evacetrapib 6038 64.8±9.4 4648 (77.0) 3620 (60.0) 5272 (87.3) 4127 (68.4) 45.3±11.7 97 132 0.75 28
Placebo 6054 65.0±9.5 4660 (77.0) 3637 (60.1) 5301 (87.6) 4109 (67.9) 45.3±11.7
REVEAL (2017)29 Anacetrapib 15,225 67.0±8.0 12,769 (83.9) 13,325 (87.5) – 5654 (37.1) 40.0±10.0 100 104 0.67 48
Placebo 15,224 67.0±8.0 12,765 (83.8) 13,354 (87.7) – 5666 (37.2) 40.0±10.0
Newcastle (1971)6 Clofibrate 244 52.0±8.7 192 (78.7) 244 (100) – – – 0 – 0.00 48
Placebo 253 53.0±8.4 208 (82.2) 253 (100) – – –
Scottish (1971)5 Clofibrate 350 52.9±7.7 288 (82.3) 350 (100) – – – 0 – 0.00 48
Placebo 367 51.2±7.4 305 (83.1) 367 (100) – – –
Cooperative Study Clofibrate 272 – 272 (100) 76 (28.0) 163 (60.0) 63 (23.0) – 0 – 0.00 54
Group (1973)30 Placebo 269 – 269 (100) 56 (21.0) 186 (69.0) 65 (24.0) –
The Coronary Drug Fibrates 1103 – 1103 (100) 1103 (100) NA NA – 0 – 0.10 58
Project (1975)19 Niacin 1115 – 1115 (100) 1115 (100) – – –
Placebo 2789 – 2789 (100) 2789 (100) – – –
WHO COOP (1978)31 Clofibrate 5331 45.9 5331 (100) NA – – – 0 – 0.41 60
Placebo 5296 45.8 5296 (100) NA – – –
Carlson et al (1988)32 Clofibrate 106 60.0 83 (78.5) 48 (45.2) – 3 (3.0) – 0 – 0.00 60
Placebo 105 59.6 85 (80.8) 48 (45.7) – 4 (3.6) –
HHS (1987)33 Gemfibrozil 2051 47.2±4.6 2051 (100) – 297 (14.5) 49 (2.4) 47.1±10.5 0 10 0.54 60
Placebo 2030 47.4±4.6 2030 (100) – 272 (13.4) 59 (2.9) 47.1±11.0
DIS (1991)34 Clofibrate 379 45.8±8.8 198 (52.2) – 119 (31.4) 379 (100) – 0 10 0.27 60
Placebo 382 46.2±7.0 231 (60.5) – 109 (28.5) 382 (100) –
SENDCAP (1998)35 Bezafibrate 81 50.8±8.0 61 (75.3) – – 81 (100) 39.4±4.6 0 6 0.32 60
Placebo 83 50.9±8.1 56 (67.5) – – 83 (100) 36.3±5.0
VA-HIT (1999)36 Gemfibrozil 1264 64±7 1264 (100) 771 (61.0) 720 (57.0) 303 (24.0) 32.0±5.0 0 6 0.00 61
Placebo 1267 63±7 1267 (100) 773 (61.0) 722 (57.0) 317 (25.0) 32.0±5.0
BIP (2000)37 Bezafibrate 1548 60.1±6.8 1412 (91.2) 1548 (100) 482 (31.2) 155 (10.0) 34.6±5.5 0 15 0.16 75
Placebo 1542 60.1±6.7 1413 (91.6) 1542 (100) 518 (33.6) 154 (10.0) 34.6±5.5
DAIS (2001)38 Fenofibrate 207 57.4±5.7 149 (72.0) 100 (48.0) 113 (55.0) 207 (100) 39.1±7.0 0 8 0.27 6
Placebo 211 56.3±6.2 156 (74.0) 100 (47.0) 102 (48.0) 211 (100) 40.6±7.7
LEADER (2002)39 Bezafibrate 783 68.4±8.9 783 (100) 151 (20.5) – 137 (17.5) 42.9 0 8 0.31 55
Placebo 785 68.0±8.8 785 (100) 160 (21.7) – 131 (16.7) 43.7
FIELD (2005)40 Fenofibrate 4895 62.2±6.8 3071 (63.0) 1068 (22.0) 2776 (57.0) 4895 (100) 42.5±10.1 26 4 0.36 60
Placebo 4900 62.2±6.9 3067 (63.0) 1,063 (22.0) 2768 (56.0) 4900 (100) 42.5±10.1
AFREGS (2005)41 Fibrates 71 63.3±7.5 64 (90.1) 71 (100) – 0.0 34.8±7.7 0 – 0.56 30
Placebo 72 63.1±6.8 68 (94.4) 72 (100) – 0.0 34.9±3.9
ACCORD (2010)42 Fenofibrate 2765 62.2±6.7 1914 (69.2) 1008 (36.5) – 2765 (100) 38.0±7.0 100 8 0.00 60
Placebo 2753 62.3±6.8 1910 (69.4) 1008 (36.6) – 2753 (100) 38.2±7.2
CLAS (1987)7 Niacin 94 53.9±0.5 94 (100) 94 (100) 0.0 0.0 44.6±1.15 0 37 0.68 24
Placebo 94 54.5±0.5 94 (100) 94 (100) 0.0 0.0 43.7±1.13
ARBITER 2 (2004)43 Niacin 87 67±10 78 (89.7) 87 (100) 64 (73.6) 24 (27.6) 39.0±7.0 100 21 0.25 12
Placebo 80 68±10 74 (92.5) 80 (100) 61 (76.3) 22 (27.5) 40.0±7.0
Guyton et al. (2007)44 Niacin 676 56.9±10.9 324 (47.9) 34 (5.0) 431 (63.8) 105 (15.5) 50.5±13.2 100 30 0.08 6
Placebo 272 57.5±10.3 152 (55.9) 18 (6.6) 181 (66.5) 43 (15.8) 49.9 (13.9)
AIM-HIGH (2011)45 Niacin 1718 63.7±8.8 1465 (85.3) 968 (56.3) 1250 (72.8) 588 (34.2) 34.5±5.6 100 25 0.16 36
Placebo 1696 63.7±8.7 1445 (85.2) 955 (56.3) 1189 (70.1) 570 (33.6) 34.9±5.6
HPS2-THRIVE (2014)46 Niacin 12,838 64.5±7.5 10,614 (82.7) 8685 (67.7) – 4134 (32.2) 43.9±11.2 100 17 0.26 48
Placebo 12,835 64.5±7.5 10,615 (82.7) 8660 (67.5) – 4165 (32.5) 44.0±11.2
Open in a new tab

IQR: interquartile range; CHD: coronary heart disease; DM: diabetes mellitus; HTN: hypertension; HDL-C: high-density lipoprotein cholesterol.

Effect on lipid profile

HDL cholesterol modifiers significantly increased the HDL cholesterol (mg/dL) levels compared to control (MD 4.54, 95% CI 2.89, 6.19, P < 0.001, I2 = 98%). CETP inhibitors had the greatest impact on HDL cholesterol (mg/dL) (MD 56.30, 95% CI 24.92, 87.68, P < 0.001, I2 = 98%), followed by niacin (MD 7.65, 95% CI 4.36, 10.94, P < 0.001, I2 = 68%) and fibrates (MD 3.30, 95% CI 1.40, 5.21, P < 0.001, I2 = 98%, Supplementary Figure 1). Similarly, HDL cholesterol modifiers significantly reduced LDL cholesterol levels (mg/dL) (MD −10.58, 95% CI −13.61, −8.00, P < 0.001, I2 = 98%), with CETP inhibitors (MD −12.52, 95% CI −17.20, −7.84, P < 0.001, I2 = 98%), fibrates (MD −9.78, 95% CI −13.77, −5.78, P < 0.001, I2 = 90%), and niacin (MD −9.67, 95% CI −14.56, −4.78, P < 0.001, I2 = 79%) showing significant reductions compared to control (Supplementary Figure 2). HDL cholesterol modifiers were associated with a significant reduction in triglycerides (mg/dL) (MD −10.94, 95% CI −13.25, −8.62, P < 0.001, I2 = 79%), with maximum reduction provided by fibrates (MD −44.50, 95% CI −54.73, −34.26, P < 0.001, I2 = 40%), followed by niacin (MD −31.41, 95% CI, −49.38, −13.44, P < 0.001, I2 = 80%) and CETP inhibitors (MD −8.72, 95% CI, −11.12, −6.32, P < 0.001, I2 = 44%, Supplementary Figure 3).

Cardiovascular mortality

Twenty-five trials (151,900 patients) reported 5190 events of cardiovascular mortality. HDL cholesterol modifiers had no statistically significant effect on cardiovascular mortality (RR 0.94, 95% CI 0.89–1.00, P = 0.05, I2 = 13%; Figure 2). There was no benefit in terms of absolute risk (RD −0.0001, 95% CI −0.0014, 0.0011, P = 0.84, I2 = 28%). Class wise, none of the individual drug classes had a statistically significant effect on cardiovascular mortality. Meta-regression analysis showed that there was no association between an absolute increase in HDL cholesterol levels and cardiovascular mortality (slope −0.00002, P = 0.99; R2 = 0, Supplementary Figure 4).

Figure 2.

Figure 2.

Open in a new tab

Forest plot showing cardiovascular mortality.

Myocardial infarction

Thirty trials (154,332 patients) reported 7687 events of MI. HDL cholesterol modifiers reduced the relative risk of MI by 13% (RR 0.87, 95% CI 0.82–0.93, P < 0.001, I2 = 37%, Supplementary Figure 6). This risk reduction was achieved with fibrates only (RR 0.80, 95% CI 0.73–0.87, P < 0.001, I2 = 22%), and neither CTEP inhibitors nor niacin had significant effects on MI. In terms of absolute risk, the use of HDL cholesterol modifiers prevented four MIs per 1000 patients (RD −0.004. 95% CI −0.006, −0.001, P = 0.02, I2 = 52%). Meta regression analysis did not show a significant association between the risk of MI and absolute increase in HDL cholesterol (slope 0.061, P = 0.13, R2 = 4%, Supplementary Figure 4). However, every 1 mmol (38.5 mg) reduction in LDL cholesterol generated a 39% relative risk reduction in MI in fibrates trials (RR 0.61, 95% CI 0.39–0.94, P = 0.03). Absolute LDL cholesterol accounted for 95% of the variability (R2) in outcome for fibrates. In the case of niacin (RR 0.58, 95% CI 0.32–1.1, P = 0.06, R2 = 97%) and CETP inhibitors (RR 0.49, 95% CI 0.10–2.63, P = 0.34, R2 = 5%) no significant benefit per unit decrease in LDL cholesterol was observed (Pinteraction = 0.01) (Figure 3).

Figure 3.

Figure 3.

Open in a new tab

Relationship between absolute low-density lipoprotein (LDL) cholesterol reduction (mmol) and the log relative risk (RR) of myocardial infarction in cholesteryl ester transfer protein (CETP) trials, fibrates trials, and niacin trials. Each trial is represented by one circle, the size of which is proportional to the weight in the meta-regression. The meta-regression slope (predicted risk for degree of LDL cholesterol reduction) is represented by the linear regression line. Blue represents the regression line for CETP inhibitor, yellow for Niacin and green for fibrate.

Stroke

Twenty-two trials (137,069 patients) reported 3980 events of stroke. The use of HDL cholesterol had a neutral effect on the relative risk of stroke compared to control (RR 1.00, 95% CI 0.93–1.09, P = 0.94, I2 = 25%, Supplementary Figure 7). This equitable effect was consistent across all treatments. Similarly, HDL cholesterol modifiers had no significant effect on the absolute risk of stroke (RD 0.0003, 95% CI −0.0011, 0.0016, P = 0.72, I2 = 29%). Meta regression analysis did not show a significant association between the risk of stroke and absolute change in HDL cholesterol (slop: 0.002, P = 0.97; R2 = 0, Supplementary Figure 4).

All-cause mortality

Twenty-five trials (150,196 patients) reported 10,014 events of all-cause mortality. HDL cholesterol modifiers did not reduce the relative risk of all-cause mortality compared to control (RR 1.02, 95% CI 0.97–1.08, P = 0.48, I2 = 49%, Supplementary Figure 8). Individually, a consistent effect was shown by each drug class. HDL cholesterol modifiers had no effect on the absolute risk of all-cause mortality (RD 0.0011, 95% CI −0.0019, 0.0042, P = 0.46, I2 48%). Meta-regression analysis did not demonstrate a significant association between the risk of all-cause mortality and an absolute increase in HDL cholesterol (slope −0.046, P = 0.26, R2 = 0.19, Supplementary Figure 4).

Subgroup analysis showed that in trials which did not use statins as background therapy, HDL cholesterol modifiers were associated with a significant risk reduction in cardiovascular mortality (RR 0.88, 95% CI 0.79–0.98, P = 0.02, I2 = 31%) and MI (RR 0.78, 95% CI 0.71–0.85, P < 0.001, I2 = 9%), while HDL cholesterol modifiers had no effect on stroke or all-cause mortality (Figure 4). Conversely, in studies in which statins were used as background therapy, HDL cholesterol modifiers had no statistically significant impact on cardiovascular mortality (RR 0.97, 95% CI 0.90–1.05, P = 0.46, I2 = 0%) or MI (RR 0.95, 95% CI 0.88–1.01, P = 0.12, I2 = 21%) or stroke or all-cause mortality (Figure 4).

Figure 4.

Figure 4.

Open in a new tab

Forest plot showing effect of high-density lipoprotein (HDL) cholesterol modifiers on all endpoints in subgroups of patients with concomitant background statin therapy or no statin therapy.

Discussion

In this systematic review and meta-analysis of 31 RCTs involving 154,601 patients and 26,871 cardiovascular events, we report that over a mean follow-up duration of 3.2 years, none of the key HDL increasing drugs (fibrates, niacin or CETP inhibitors) were associated with a statistically significant reduction of cardiovascular or all-cause mortality. There was a significant risk reduction in MI by HDL cholesterol modifying therapies, which was mainly attributable to fibrates and was linked to an absolute reduction in LDL cholesterol rather than HDL cholesterol. However, this benefit was not observed in the contemporary studies that included background statin therapy. There was no significant impact of these agents on the risk of stroke. Meta-regression analyses did not reveal any significant differences in the key endpoints with respect to HDL cholesterol levels. These results suggest that despite the long-standing interest in pharmacological interventions aimed at increasing HDL cholesterol, such therapies are unlikely to generate meaningful clinical benefits.

Older studies on fibrates demonstrated a reduced incidence of cardiovascular events with the use of these therapies. In the Newcastle6 and Scottish5 studies, the use of clofibrate was associated with a substantial reduction in the cardiovascular risk. However, the beneficial effect of these interventions was not observed when they were added to the statin therapy. For example, in the Action to Control Cardiovascular Risk in Diabetes (ACCORD) study,42 which enrolled patients with type 2 diabetes treated with background statin therapy, the use of fenofibrate compared with placebo was not associated with any significant reduction of major adverse cardiovascular events. Therefore, the beneficial effect seen in the initial studies may reflect effects on other lipoproteins such as LDL cholesterol or triglyceride reduction.

The relatively new pharmacological class of CETP inhibitors has been shown to increase the HDL cholesterol level by as much as 133%;20 however, the development of these drugs has been challenged since three compounds failed to reduce cardiovascular outcomes in phase III clinical trials. Recently, the Randomized Evaluation of the Effects of Anacetrapib through Lipid Modification (REVEAL)29 trial demonstrated a 9% risk reduction (RR 0.91, P < 0.004) in the composite primary endpoint of coronary death, MI or coronary revascularization for anacetrapib on a background statin therapy. However, the mechanism which resulted in a cardiovascular risk reduction in this trial likely reflects the18% mean reduction in LDL cholesterol which fell close to the regression line from a large meta-analysis comparing reduction in cardiovascular events to the absolute decrease in LDL cholesterol levels.47 These findings strongly suggest that the reduction in the primary endpoint was largely driven by a reduction in atherogenic lipoproteins, rather than the 104% increase in HDL cholesterol generated by the drug.29

We compare our results with prior reports on the same topic. Birjmohun and colleagues48 investigated the safety and efficacy of niacin and fibrates. This was a pooled analysis of 53 trials with 16,802 patients. The authors suggested that the use of fibrates was associated with a significant reduction in major coronary events while concluding that the data on niacin were limited. Their study included trials with small sample size with limited follow-up duration and missed some important clinical endpoints. These issues can generate small study effects and underestimate the treatment effect due to the short duration of drug exposure. Moreover, the majority of these studies were from the pre-statin era that may have resulted in overestimation of the effect sizes of the fibrates and niacin. Another study by Keene et al.49 was a much larger analysis (39 trials, 117,411 patients), which also suggested that none of these therapies were associated with any meaningful improvements in the cardiovascular endpoints. Besides the inclusion of small studies, the other main limitations of that analysis were that major CETP trials had not been published and thus there was a lack of assessment of these datasets; moreover, no attempt of meta-regression was done to explore the relationship between HDL cholesterol and key endpoints. We have attempted to address these limitations in our review. First, we have updated the evidence by pooling all the major trials including contemporary CETP trials. Second, we applied strict study selection criteria to generate the most robust possible evidence; and third, we utilized trial-level meta-regression analysis to investigate the effect of HDL cholesterol cardiovascular outcomes. Furthermore, by looking into three distinct therapies, all of which increase HDL cholesterol, we provide strength to the notion that increasing HDL cholesterol on its own is unlikely to be of mechanistic relevance, unless coupled with an alternative pathway such as lowering LDL cholesterol. Finally, the genetic data also support the concept that increasing HDL cholesterol may not confer any significant clinical benefit. For instance, a large Mendelian randomization study (20,913 MI cases, 95,407 controls) identified 14 single nucleotide polymorphisms that increase HDL cholesterol and found that such polymorphisms had no significant association with the risk of MI (OR 0.99, 95% CI 0.88–1.11, P = 0.85) despite significantly increasing HDL levels.50

The major limitations of this analysis are that this is a pooled analysis of individual trials as we did not have access to the individual patient data. However, these studies have been published over decades and it seemed logistically implausible to obtain access to all the data from the individual trials. We relied on the endpoint definitions of the individual studies and the possibility of heterogeneity in the outcome definitions cannot be excluded. Similarly, it is also likely that the adjudication of these endpoints varied across the trials. Finally, the background therapies have varied over the duration of time. The new studies on CETP inhibition almost uniformly have background statin therapy, whereas such therapy was inconsistent with the older studies of fibrates and niacin.

In conclusion, we report that three classes of drugs that raise HDL cholesterol by distinct mechanisms failed to reduce cardiovascular mortality, stroke or all-cause mortality significantly. The apparent clinical benefits of these therapies are driven by the studies of the pre-statin era, and the use of effective statins as background therapy had nullified the influence of HDL-raising therapies. Meta-regression analyses confirmed that the observed MI benefit by fibrates was due to LDL cholesterol reduction and there is a lack of association between HDL cholesterol and a reduction in the outcomes of interest. Our findings refute the efficacy of HDL cholesterol modifiers on hard clinical endpoints. Studies on HDL infusion agents are underway and the results of those will be further illuminating.

These findings once again raise an important clinical lesson in that the observational studies on surrogate endpoints can be considered hypothesis generating only and the benefit of a therapy can be only proved by virtue of large, adequately powered outcome trials.

Supplementary Material

Supplement

Acknowledgments

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Footnotes

Declaration of conflicting interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: AML is consultant for Amgen, Novo, Nordisk, Sanofi, Abbott, Sarpeta, Sermonix; and received the following research grants: Pfizer, Astra Zeneca, Esperion, AbbVie, Eli Lilly, Roche. SEN receives grant support from Pfizer, the Medicines Company, Amgen, Cerenis, AstraZeneca and Esperion Therapeutics. EK is consultant for Bristol-Myers Squibb, Pfizer, Janssen and Daiichi-Saknyo.

References

  • 1.Castelli WP, Anderson K, Wilson PW, et al. Lipids and risk of coronary heart disease. The Framingham Study. Ann Epidemiol 1992; 2: 23–28. [DOI] [PubMed] [Google Scholar]
  • 2.Sharrett AR, Ballantyne CM, Coady SA, et al. Coronary heart disease prediction from lipoprotein cholesterol levels, triglycerides, lipoprotein(a), apolipoproteins A-I and B, and HDL density subfractions: the Atherosclerosis Risk in Communities (ARIC) Study. Circulation 2001; 104: 1108–1113. [DOI] [PubMed] [Google Scholar]
  • 3.Gordon DJ, Probstfield JL, Garrison RJ, et al. High-density lipoprotein cholesterol and cardiovascular disease. Four prospective American studies. Circulation 1989; 79: 8–15. [DOI] [PubMed] [Google Scholar]
  • 4.Madsen CM, Varbo A and Nordestgaard BG. Extreme high high-density lipoprotein cholesterol is paradoxically associated with high mortality in men and women: two prospective cohort studies. Eur Heart J 2017; 38: 2478–2486. [DOI] [PubMed] [Google Scholar]
  • 5.Research Committee of the Scottish Society of Physicians. Ischaemic heart disease: a secondary prevention trial using clofibrate. BMJ 1971; 4: 775–784. [PMC free article] [PubMed] [Google Scholar]
  • 6.Trial of clofibrate in the treatment of ischaemic heart disease. Five-year study by a group of physicians of the Newcastle upon Tyne region. BMJ 1971; 4: 767–775. [PMC free article] [PubMed] [Google Scholar]
  • 7.Blankenhorn DH, Nessim SA, Johnson RL, et al. Beneficial effects of combined colestipol–niacin therapy on coronary atherosclerosis and coronary venous bypass grafts. JAMA 1987; 257: 3233–3240. [PubMed] [Google Scholar]
  • 8.Higgins JPT and Green S (eds). Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. The Cochrane Collaboration, 2011. Available at http://handbook.cochrane.org. [Google Scholar]
  • 9.Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ (Clinical Research ed) 2009; 339: b2535. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Heart Protection Study Collaborative Group. Effects on 11-year mortality and morbidity of lowering LDL cholesterol with simvastatin for about 5 years in 20,536 high-risk individuals: a randomised controlled trial. Lancet (London, England) 2011; 378: 2013–2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Schwarzer G, Carpenter JR and Rücker G. Small-study effects in meta-analysis In: Meta-Analysis with R. Cham: Springer International Publishing, 2015, pp.107–141. [Google Scholar]
  • 12.Silverman MG, Ference BA, Im K, et al. Association between lowering LDL-C and cardiovascular risk reduction among different therapeutic interventions: a systematic review and meta-analysis. JAMA 2016; 316: 1289–1297. [DOI] [PubMed] [Google Scholar]
  • 13.Vickers AJ and Altman DG. Statistics notes: analysing controlled trials with baseline and follow up measurements. BMJ (Clinical Research ed) 2001; 323: 1123–1124. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Higgins JPT, Altman DG, Gøtzsche PC, et al. The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials. BMJ (Clinical Research ed) 2011; 343: d5928. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.DerSimonian R and Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986; 7: 177–188. [DOI] [PubMed] [Google Scholar]
  • 16.Higgins JPT, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ 2003; 327: 557–560. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Egger M, Davey Smith G, Schneider M, et al. Bias in meta-analysis detected by a simple, graphical test. BMJ (Clinical Research ed) 1997; 315: 629–634. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Knapp G and Hartung J. Improved tests for a random effects meta-regression with a single covariate. Stat Med 2003; 22: 2693–2710. [DOI] [PubMed] [Google Scholar]
  • 19.Clofibrate and niacin in coronary heart disease. JAMA 1975; 231: 360–381. [PubMed] [Google Scholar]
  • 20.Lincoff AM, Nicholls SJ, Riesmeyer JS, et al. Evacetrapib and cardiovascular outcomes in high-risk vascular disease. N Engl J Med 2017; 376: 1933–1942. [DOI] [PubMed] [Google Scholar]
  • 21.Schwartz GG, Olsson AG, Abt M, et al. Effects of dalcetrapib in patients with a recent acute coronary syndrome. N Engl J Med 2012; 367: 2089–2099. [DOI] [PubMed] [Google Scholar]
  • 22.Fayad ZA, Mani V, Woodward M, et al. Safety and efficacy of dalcetrapib on atherosclerotic disease using novel non-invasive multimodality imaging (dal-PLAQUE): a randomised clinical trial. Lancet (London, England) 2011; 378: 1547–1559. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Luscher TF, Taddei S, Kaski JC, et al. Vascular effects and safety of dalcetrapib in patients with or at risk of coronary heart disease: the dal-VESSEL randomized clinical trial. Eur Heart J 2012; 33: 857–865. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Cannon CP, Shah S, Dansky HM, et al. Safety of anacetrapib in patients with or at high risk for coronary heart disease. N Engl J Med 2010; 363: 2406–2415. [DOI] [PubMed] [Google Scholar]
  • 25.Nissen SE, Tardif JC, Nicholls SJ, et al. Effect of torcetrapib on the progression of coronary atherosclerosis. N Engl J Med 2007; 356: 1304–1316. [DOI] [PubMed] [Google Scholar]
  • 26.Kastelein JJ, van Leuven SI, Burgess L, et al. Effect of torcetrapib on carotid atherosclerosis in familial hypercholesterolemia. N Engl J Med 2007; 356: 1620–1630. [DOI] [PubMed] [Google Scholar]
  • 27.Bots ML, Visseren FL, Evans GW, et al. Torcetrapib and carotid intima-media thickness in mixed dyslipidaemia (RADIANCE 2 study): a randomised, double-blind trial. Lancet (London, England) 2007; 370: 153–160. [DOI] [PubMed] [Google Scholar]
  • 28.Barter PJ, Caulfield M, Eriksson M, et al. Effects of torcetrapib in patients at high risk for coronary events. N Engl J Med 2007; 357: 2109–2122. [DOI] [PubMed] [Google Scholar]
  • 29.Bowman L, Hopewell JC, Chen F, et al. Effects of anacetrapib in patients with atherosclerotic vascular disease. N Engl J Med 2017; 377: 1217–1227. [DOI] [PubMed] [Google Scholar]
  • 30.Veterans Administration Cooperative Study Group. The treatment of cerebrovascular disease with clofibrate. Final report of the Veterans Administration Cooperative Study of Atherosclerosis, Neurology Section. Stroke 1973; 4: 684–693. [DOI] [PubMed] [Google Scholar]
  • 31.Oliver MF, Heady JA, Morris JN, et al. A co-operative trial in the primary prevention of ischaemic heart disease using clofibrate. Report from the Committee of Principal Investigators. Br Heart J 1978; 40: 1069–1118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Carlson LA and Rosenhamer G. Reduction of mortality in the Stockholm Ischaemic Heart Disease Secondary Prevention Study by combined treatment with clofibrate and nicotinic acid. Acta Medica Scand 1988; 223: 405–418. [DOI] [PubMed] [Google Scholar]
  • 33.Frick MH, Elo O, Haapa K, et al. Helsinki Heart Study: primary-prevention trial with gemfibrozil in middle-aged men with dyslipidemia. Safety of treatment, changes in risk factors, and incidence of coronary heart disease. N Engl J Med 1987; 317: 1237–1245. [DOI] [PubMed] [Google Scholar]
  • 34.Hanefeld M, Fischer S, Schmechel H, et al. Diabetes Intervention Study. Multi-intervention trial in newly diagnosed NIDDM. Diabetes Care 1991; 14: 308–317. [DOI] [PubMed] [Google Scholar]
  • 35.Elkeles RS, Diamond JR, Poulter C, et al. Cardiovascular outcomes in type 2 diabetes. A double-blind placebo-controlled study of bezafibrate: the St. Mary’s, Ealing, Northwick Park Diabetes Cardiovascular Disease Prevention (SENDCAP) Study. Diabetes Care 1998; 21: 641–648. [DOI] [PubMed] [Google Scholar]
  • 36.Rubins HB, Robins SJ, Collins D, et al. Gemfibrozil for the secondary prevention of coronary heart disease in men with low levels of high-density lipoprotein cholesterol. Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial Study Group. N Engl J Med 1999; 341: 410–418. [DOI] [PubMed] [Google Scholar]
  • 37.BIP Study Group. Secondary prevention by raising HDL cholesterol and reducing triglycerides in patients with coronary artery disease. Circulation 2000; 102: 21–27. [DOI] [PubMed] [Google Scholar]
  • 38.Diabetes Atherosclerosis Intervention Study Investigators. Effect of fenofibrate on progression of coronary-artery disease in type 2 diabetes: the Diabetes Atherosclerosis Intervention Study, a randomised study. Lancet (London, England) 2001; 357: 905–910. [PubMed] [Google Scholar]
  • 39.Meade T, Zuhrie R, Cook C, et al. Bezafibrate in men with lower extremity arterial disease: randomised controlled trial. BMJ (Clinical Research ed) 2002; 325: 1139. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Keech A, Simes RJ, Barter P, et al. Effects of long-term fenofibrate therapy on cardiovascular events in 9795 people with type 2 diabetes mellitus (the FIELD study): randomised controlled trial. Lancet (London, England) 2005; 366: 1849–1861. [DOI] [PubMed] [Google Scholar]
  • 41.Whitney EJ, Krasuski RA, Personius BE, et al. A randomized trial of a strategy for increasing high-density lipoprotein cholesterol levels: effects on progression of coronary heart disease and clinical events. Ann Intern Med 2005; 142: 95–104. [DOI] [PubMed] [Google Scholar]
  • 42.Ginsberg HN, Elam MB, Lovato LC, et al. Effects of combination lipid therapy in type 2 diabetes mellitus. N Engl J Med 2010; 362: 1563–1574. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Taylor AJ, Sullenberger LE, Lee HJ, et al. Arterial Biology for the Investigation of the Treatment Effects of Reducing Cholesterol (ARBITER) 2: a double-blind, placebo-controlled study of extended-release niacin on atherosclerosis progression in secondary prevention patients treated with statins. Circulation 2004; 110: 3512–3517. [DOI] [PubMed] [Google Scholar]
  • 44.Guyton JR, Brown BG, Fazio S, et al. Lipid-altering efficacy and safety of ezetimibe/simvastatin coadministered with extended-release niacin in patients with type IIa or type IIb hyperlipidemia. J Am Coll Cardiol 2008; 51: 1564–1572. [DOI] [PubMed] [Google Scholar]
  • 45.The AIM-HIGH Investigators. Niacin in patients with low HDL cholesterol levels receiving intensive statin therapy. N Engl J Med 2011; 365: 2255–2267. [DOI] [PubMed] [Google Scholar]
  • 46.Landray MJ, Haynes R, Hopewell JC, et al. Effects of extended-release niacin with laropiprant in high-risk patients. N Engl J Med 2014; 371: 203–212. [DOI] [PubMed] [Google Scholar]
  • 47.Baigent C, Keech A, Kearney PM, et al. Efficacy and safety of cholesterol-lowering treatment: prospective meta-analysis of data from 90,056 participants in 14 randomised trials of statins. Lancet (London, England) 2005; 366: 1267–1278. [DOI] [PubMed] [Google Scholar]
  • 48.Birjmohun RS, Hutten BA, Kastelein JJ, et al. Efficacy and safety of high-density lipoprotein cholesterol-increasing compounds: a meta-analysis of randomized controlled trials. J Am Coll Cardiol 2005; 45: 185–197. [DOI] [PubMed] [Google Scholar]
  • 49.Keene D, Price C, Shun-Shin MJ, et al. Effect on cardiovascular risk of high density lipoprotein targeted drug treatments niacin, fibrates, and CETP inhibitors: meta-analysis of randomised controlled trials including 117 411 patients. BMJ 2014; 349: g4379. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Voight BF, Peloso GM, Orho-Melander M, et al. Plasma HDL cholesterol and risk of myocardial infarction: a Mendelian randomisation study. Lancet (London, England) 2012; 380: 572–580. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement
NIHMS1621088-supplement-Supplement.pdf (493.8KB, pdf)

RESOURCES