Author + information
- Received December 4, 2007
- Revision received February 19, 2008
- Accepted March 4, 2008
- Published online April 22, 2008.
- John R. Guyton, MD⁎,⁎ (, )
- B. Greg Brown, MD, PhD†,
- Sergio Fazio, MD, PhD‡,
- Adam Polis, MA§,
- Joanne E. Tomassini, PhD§ and
- Andrew M. Tershakovec, MD, MPH§
- ↵⁎Reprint requests and correspondence:
Dr. John R. Guyton, Department of Medicine, Duke University Medical Center, Trent Drive, Durham, North Carolina 27710.
Objectives This study evaluated the safety and lipid-altering efficacy of ezetimibe/simvastatin (E/S) coadministered with extended-release niacin (N) in patients with type IIa or IIb hyperlipidemia.
Background Current guidelines recommend consideration of combination drug therapy to achieve optimal low-density lipoprotein cholesterol (LDL-C) lowering and broader lipid-altering effects when treating hypercholesterolemic patients at high risk for atherosclerotic cardiovascular events.
Methods In this 24-week multicenter, randomized, double-blind study, 1,220 type IIa or IIb hyperlipidemic patients were randomized to treatment with E/S (10/20 mg/day) + N (titrated to 2 g/day), or N (titrated to 2 g/day), or E/S (10/20 mg/day). Changes from baseline in LDL-C (primary) and other secondary variables were assessed in the completers and modified intent-to-treat populations.
Results Coadministered E/S with N resulted in significantly greater reductions in LDL-C, non–high-density lipoprotein cholesterol, triglycerides, apolipoprotein B, and lipid/lipoprotein ratios, compared with either agent alone (p < 0.001). The combination increased levels of apolipoprotein A-I and high-density lipoprotein cholesterol significantly more than E/S (p < 0.001), and reduced high-sensitivity C-reactive protein levels significantly more than N (p = 0.005). A significantly greater percentage of patients discontinued the study in the N (25.0%) and E/S + N (23.3%) groups, compared with E/S (9.6%, p < 0.001) because of clinical adverse experiences (primarily flushing). Incidences of other clinical and laboratory adverse experiences (liver-, muscle-, and gastrointestinal-related) were similar for all groups.
Conclusions Combination treatment with E/S plus N showed superior lipid-altering efficacy compared with N or E/S in type IIa or IIb hyperlipidemia patients and was generally well tolerated aside from N-associated flushing. This combination offers an effective, broad, lipid-altering therapy with improvements in lipid effects beyond LDL-C in these patients. (To Evaluate Ezetimibe/Simvastatin and Niacin [Extended Release Tablet] in Patients With High Cholesterol; NCT00271817)
In addition to high levels of low-density lipoprotein cholesterol (LDL-C), other important determinants of increased risk for coronary heart disease (CHD) include low levels of high-density lipoprotein cholesterol (HDL-C) and elevated triglyceride levels. Current guidelines for the prevention and treatment of CHD specify LDL-C lowering as the primary goal of lipid-related therapy (1–3). Therapeutic lifestyle change to reduce LDL-C and CHD risk is recommended, including dietary measures such as decreasing saturated fat and increased viscous fiber (3). However, consideration of combination therapies with broad lipid-altering effects is also recommended for treatment of high-risk patients in need of further LDL-C lowering and/or with low HDL-C and high triglyceride levels (1–3).
Recent evidence suggests that increases in HDL-C levels may be associated with reductions in the incidence of CHD, regardless of LDL-C levels (4,5). Addition of HDL-C raising to LDL-C lowering therapy may be beneficial in the treatment of patients with combined dyslipidemia, who characteristically have low HDL-C levels, elevated triglycerides, and a preponderance of small, dense LDL particles (3,4). Given the increasing prevalence of diabetes, metabolic syndrome, and obesity among patients with combined dyslipidemia, this therapeutic approach is of particular importance (4,5).
Presently, niacin is the most effective agent available for raising HDL-C levels. Niacin also decreases levels of triglycerides, LDL-C, and lipoprotein(a), and increases lipoprotein particle size, thus representing a unique therapeutic option for treatment of patients with combined dyslipidemia (4–7). Niacin has been used in clinical trials as monotherapy and combined with bile acid sequestrants, statins, clofibrate, and in triple combination with cholestyramine and gemfibrozil. These trials suggest that niacin, alone or as a component of broader treatment, can reduce cardiovascular disease–related morbidity and mortality and also can slow the progression and/or induce the regression of coronary atherosclerosis (6,15–21).
The dual cholesterol absorption and synthesis inhibitor ezetimibe/simvastatin (E/S) has been shown to be effective at lowering levels of LDL-C, non–HDL-C, and triglycerides in patients with hypercholesterolemia (22–25). The coadministration of E/S with extended-release niacin (N) (Niaspan, Kos Pharmaceuticals/Abbott, Abbott Park, Illinois) (7) might provide significant complementary LDL-C, non–HDL-C, and triglyceride lowering and HDL-C raising effects, potentially giving greater benefits than either agent alone.
This 24-week study is the initial phase of a longer study in which we assessed the safety and efficacy of E/S + N in type IIa or IIb hyperlipidemic patients. The present study compared the lipid-altering effects of E/S + N combination therapy versus N and E/S. A Completers approach was used for the primary efficacy analysis to describe the clinical effect associated with the maximum N (2 g) prescribed dose, anticipating that substantial numbers of patients would not be able to tolerate higher doses of N (6,7,26). A supportive secondary analysis was performed in the modified intent-to-treat (mITT) population, in which the last blood sample obtained before drug discontinuation is considered to reflect a patient's response to the drug.
Study population and design
This 24-week study was Part 1 of a 64-week, multicenter, randomized, double-blind study in patients with type IIa or IIb hyperlipidemia. The protocol was approved by appropriate institutional review boards, and all patients provided informed written consent.
Men and women aged 18 years to 79 years with LDL-C levels (130 to 190 mg/dl), triglyceride levels (≤500 mg/dl), and metabolic and clinical stability (e.g., euthyroid, creatinine <2 mg/dl, creatinine kinase ≤2 ×upper limit of normal [ULN], transaminases ≤1.5 × ULN) were eligible for inclusion in the study. After a 4-week washout period, eligible patients were randomized (5:2:2) via central allocation to E/S (10/20 mg) + N (titrated to 2 g), or E/S (10/20 mg), or N (titrated to 2 g). All study personnel remained blinded to treatment allocation. During the first 12 weeks of the study, N was increased by 500 mg every 4 weeks up to 2 g/day from a 500 mg/day starting dose. Patients were counseled to take N at bedtime with a low-fat snack, aspirin (325 mg), or ibuprofen (200 mg) 30 min before taking N, and to avoid alcoholic and hot beverages near the time of taking N.
Patient enrollment was monitored to achieve 30% to 50% randomization of type IIa patients and to ensure balance in gender. Patients were stratified at randomization according to baseline LDL-C (130 to 159 mg/dl; 160 to 190 mg/dl) and triglyceride (≤200 mg/dl; 201 to 500 mg/dl) levels. In Part 2, comprising an additional 40 weeks of treatment, the E/S and E/S + N groups were continued, whereas patients in the N-containing groups were randomly reassigned to those groups. Only Part 1 results are presented here; Part 2 results will be analyzed and presented separately.
The primary efficacy analysis was performed in the Completers population, which included all patients with a baseline value, and who received ≥24 weeks of active study therapy and had an on-treatment sample at the maximum titrated dose of N (2 g). This analysis, rather than a mITT-last observation carried forward (LOCF) approach, was chosen because some patients are not able to tolerate 2 g N, and it was of interest to estimate and compare the efficacy associated with N (2 g). The mITT analysis, which included all randomized patients who had baseline and at least 1 post-baseline value, and an LOCF approach to impute any missing data for the Week 24 assessment, was performed as a secondary and supportive analysis.
Safety was assessed by monitoring clinical adverse experiences (AEs) in all treated patients and laboratory AEs in all treated patients who had ≥1 post-baseline measurement. Safety variables included the incidence of serum alanine aminotransferase and aspartate aminotransferase ≥3 × ULN, serum creatine kinase (CK) ≥10 × ULN with and without muscle symptoms, discontinuation because of flushing, gallbladder-related adverse experiences, cholecystectomy, change from baseline in fasting glucose, and new onset of diabetes (patients who had an AE related to a diagnosis of diabetes, initiated an antidiabetic medication during the study, or had 2 consecutive fasting glucose measurements that increased to ≥126 mg/dl) (27).
The primary study hypothesis was that E/S + N would be superior to N with respect to percent change from baseline in LDL-C after 24 weeks of treatment. If the primary hypothesis was found to be significant, then the secondary efficacy hypotheses were tested using an ordered testing strategy to control for multiplicity in testing the hypothesized efficacy endpoints. Secondary end points, all assessed as percent change from baseline to week 24, were specified in the protocol and tested in the following order: 1) non–HDL-C (E/S + N vs. N); 2) HDL-C (E/S + N vs. E/S); 3) TG (E/S + N vs. E/S); 4) LDL-C (E/S + N vs. E/S); and 5) non–HDL-C (E/S + N vs. E/S). Other secondary variables included percent change from baseline to week 24 in total cholesterol (TC), apolipoprotein (Apo)B, ApoA-I, lipid/lipoprotein ratios, and high-sensitivity C-reactive protein (hsCRP). No strategy for multiplicity was applied to these variables.
For the primary study hypothesis, LDL-C change, it was expected that there would be 94% power to detect a 5 percentage point difference between E/S + N versus N, assuming a within-group standard deviation for an LDL-C percent change of 15% at 24 weeks. Treatment group differences were compared by an analysis of covariance model with terms for treatment, baseline LDL-C and triglyceride values, and gender for all secondary efficacy variables except triglycerides and hsCRP. For triglycerides, percent change values were transformed to ranks based on normal scores before analysis and between-treatment group differences were assessed by Hodges-Lehman location shift. The hsCRP was first transformed to the log-ratio of week 24 to baseline, and treatment differences were assessed by the delta method. Proportions of patients experiencing AEs were compared between treatment groups using the Fisher exact test. Changes from baseline in fasting glucose levels were compared by an analysis of covariance. All significance tests were 2-tailed with alpha = 0.050.
A total of 2,697 patients were screened at 106 sites in the U.S., of which 1,220 were randomized to treatment with N (n = 272) or with E/S (n = 272) or with E/S + N (n = 676). The disposition of patients is illustrated in Figure 1. The Completers population included 770 (63%) and the mITT included 1,112 (91%) of the randomized patients. Flushing was the primary reason for study discontinuation in the N-containing groups (12.1% for N; 9.9% for E/S + N; 0.4% for E/S).
There were generally no clinically meaningful differences in the distribution of baseline characteristics of all randomized patients among treatment groups (Table 1). About 28% of the patients had CHD or CHD risk equivalent status (approximately 15% diabetes, approximately 6% non-CHD forms of atherosclerosis), and half of the patients in each treatment group (slightly higher for E/S) had metabolic syndrome (3). The baseline levels of lipids and other factors were well balanced among treatment groups (Table 2). Mean baseline levels of LDL-C and HDL-C were 156 mg/dl and 50.5 mg/dl, respectively. The median baseline level of triglycerides was 156 mg/dl. Baseline levels were also similar in the mITT population (Online Appendix A).
At 24 weeks, the percent reductions from baseline in LDL-C, non–HDL-C, triglycerides, ApoB, and in ratios (TC/HDL-C, LDL-C/HDL-C, ApoB/ApoA-I, non–HDL-C/HDL-C) were significantly greater after E/S + N treatment compared with N or E/S (p < 0.001) (Table 3). Increases in HDL-C with E/S + N were significantly larger than those with E/S (p < 0.001), and were similar to those with N. The percent change in TC was significantly greater in the E/S + N group versus N (p < 0.001) and was comparable to E/S. Increases in ApoA-I levels were significantly larger in the E/S + N group compared with E/S (p < 0.001) and were comparable with N. Reduction of hsCRP with E/S + N was significantly greater than with N (p = 0.005).
Relatively similar results were obtained in the mITT analysis for all efficacy variables (Online Appendix B). In particular, the significance of treatment group differences for all reported efficacy variables was the same for both populations.
Time-related and dose-related effects
Both E/S and E/S + N treatments reduced LDL-C levels from baseline by approximately 53% at week 4 (Fig. 2A). Throughout the 16-week titration period to the maximum N dose (2 g), LDL-C levels continued to decrease for the E/S + N group, and then were maintained through 24 weeks. The LDL-C levels remained constant in the E/S group from 4 to 24 weeks. For the N group, LDL-C declined gradually from baseline levels during 24 weeks. The percent change in LDL-C was significantly greater for E/S + N compared with N (p < 0.001) for all time points, and compared with the E/S group, this difference became significant at 12 weeks.
With increasing doses of N to 2 g, levels of HDL-C increased substantially from baseline in the N and E/S + N groups throughout the 16-week titration period and then continued to increase (to 28% to 30%) from 16 to 24 weeks, although at a lesser rate (Fig. 2B). There was a small increase in HDL-C levels from baseline during 24 weeks with E/S treatment. The increase in HDL-C was significantly greater for E/S + N compared with E/S at all time points (p < 0.001), and compared with N through 12 weeks (p < 0.001), after which these differences became nonsignificant.
Levels of triglycerides declined at 4 weeks by approximately 28% for E/S and E/S + N (Fig. 2C). The levels of triglycerides were maintained for the E/S group. For the N and E/S + N groups, triglyceride levels decreased continuously during the 16-week titration period with increasing doses of niacin, and then were maintained throughout 24 weeks. The reduction in triglyceride levels was significantly greater for E/S + N compared with N over 24 weeks (p < 0.001) and was also significantly greater compared with E/S from weeks 8 to 24 (p < 0.001).
The profiles of percent change over time for these variables in the mITT population (Online Appendix C) were generally similar to those of the Completers population.
Of the 1,220 randomized patients, 1,214 received at least 1 dose of study treatment and were included in the safety analysis. The incidence of clinical AEs was higher in the N and E/S + N treatment groups than the E/S group (Table 4), mostly attributed to a greater occurrence of flushing-related AEs in the N-containing groups (Table 5). The number of patients who experienced clinical and drug-related AEs leading to study drug discontinuations was similar in the N (25.0%) and E/S + N (23.0%) groups and lower in the E/S (9.6%) group. One death due to stroke, considered by the investigator not to be study drug–related, occurred in the N group. Gallbladder-related AEs developed in 2 subjects, 1 of whom (E/S group) had a cholecystectomy.
Incidences of laboratory AEs were comparable for the N (7.0%), E/S (7.4%), and E/S + N (5.1%) groups. Similar numbers of laboratory AEs resulted in study discontinuation in the N and E/S + N groups, with no occurrences in the E/S group. No serious laboratory AEs occurred during the study. The percentages of patients with consecutive ≥3 × ULN elevations in alanine aminotransferase/aspartate aminotransferase and CK levels ≥10 × ULN were not significantly different among the treatment groups (Table 5). One patient (E/S + N group) had CK ≥10 × ULN and myalgia that were considered by the investigator to be related to strenuous exercise and not to study drug.
A total of 22 patients had AEs of increased fasting glucose, with the highest incidence occurring in the N (2.9%) versus E/S (1.8%) and E/S + N (1.3%) groups. The mean changes from baseline to week 24 in fasting glucose were 3.6 and 2.6 mg/dl, respectively, for the N and E/S + N groups, compared with −0.1 mg/dl for the E/S group (p = 0.022, E/S + N vs. E/S). Fasting glucose levels increased to a maximum of approximately 7.5% above baseline levels during the first 12 weeks in both groups receiving N, and then decreased to levels of approximately 3.0% above baseline at 24 weeks (Fig. 3).
A total of 32 patients met the criteria for new onset of diabetes in the N (2.2%), E/S (0.9%), and E/S + N (4.4%) groups (Table 5). The proportion of patients in the E/S + N versus N groups with new onset of diabetes was comparable and the difference between the E/S + N versus E/S groups was significant (p = 0.009).
This study shows that once-daily coadministration of E/S at 10/20 mg/day with N at 2 g/day provides substantial improvements in LDL-C lowering, HDL-C raising, and other aspects of the lipoprotein profile in patients with type IIa and IIb hyperlipidemia. The combination (E/S + N) significantly reduced levels of LDL-C and other lipids in comparison to either agent alone and elicited increases in HDL-C, which were significantly higher than those generated by E/S and similar to those induced by N alone. This 3-drug combination was generally well tolerated, aside from anticipated niacin-associated flushing.
In this study, this combination of E/S + N resulted in a 58.5% reduction of LDL-C levels from baseline, which was significantly greater than that of the usual recommended starting dose (10/20 mg/day) of E/S (53.5%), and comparable to that observed previously in other studies for the maximum prescribed dose of E/S (approximately 60% at 10/80 mg) (22–25). The reduction of LDL-C levels by the triple combination was also significantly greater than N alone and was more effective than, or at least comparable to, previously studied combinations of niacin and statins or bile acid sequestrants, which reduced LDL-C by 31% to 49% from baseline (4,5,10–13,15). E/S + N was also much more effective in LDL-C lowering than the triple combination of niacin, gemfibrozil, and cholestyramine, which lowered LDL-C from baseline by 22% (20).
The combination E/S has been shown to slightly increase HDL-C levels from baseline by 6% to 10% (22–25). When combined with N in this study, HDL-C levels were increased significantly more from baseline (30.2%) compared with E/S alone (8.1%) but not compared with N monotherapy (28.1%). The increase in HDL-C levels was comparable to other studies in which niacin monotherapy and combination therapy increased HDL-C levels by 24% to 43% from baseline (4,5,8–17).
The improvements in LDL-C and triglycerides after E/S + N therapy were related to the sustained effect of E/S and to increasing doses of N. These effects persisted throughout the 24 weeks of the study. In addition, HDL-C levels increased substantially during the N dose escalation period and continued to increase through 24 weeks. These results are consistent with those observed for the combination of extended-release niacin and lovastatin (Advicor, Kos Pharmaceuticals/Abbott), wherein improvements in LDL-C and triglyceride levels achieved during a similar niacin titration schedule were maintained for 1 year, and HDL-C levels continued to increase throughout the year (12,14). Results relating to the longer-term effects of the E/S + N combination will be further elucidated by Part 2 of the overall study.
Consistent with the broad lipid-altering effects of niacin demonstrated in other studies, E/S + N also reduced levels of several other lipid parameters, including non–HDL-C, TC, ApoB, and TC/HDL-C, LDL-C/HDL-C, ApoB/ApoA-I, non–HDL-C/HDL-C significantly more than E/S or N alone (10–12,14,16,20). These effects, in addition to the improvements in LDL-C and triglyceride lowering, along with the HDL-C raising effect, suggest that E/S + N may have considerable value as a treatment option for high-risk patients with combined dyslipidemia.
The primary efficacy results of this trial for E/S + N, assessed in the Completers population comprised of those individuals who completed the study at the 2 g maximum dose, were well supported by those of the mITT analysis. At 24 weeks, reductions in levels of LDL-C (−58.5% and −56.8%) and triglycerides (−42.5% and −39.9%), and increases in HDL-C (30.2% and 25.1%) in the Completers and mITT populations, respectively, were relatively comparable. Time-related and dose-related effects and changes in other efficacy variables were also comparable; moreover, the significance of treatment group differences for all efficacy variables was the same for both populations (Online Appendixes B and C). Changes in some parameters were numerically higher in the N-containing groups of the completers population, reflecting the efficacy assessment at 2 g N in this population versus a range of doses (0.5 to 2 g) in the mITT population.
Combined E/S + N treatment was generally well tolerated in this study and showed a safety profile consistent with prior experience using these agents alone or in combination indicating that E/S + N may be an option for use in patients at high risk for CHD including those with diabetes, metabolic syndrome, and obesity, in whom polypharmacy may be a concern (4–7,22–26). The rates of serious AEs were comparable for all treatment arms, and there were no statistically significant differences in muscle, liver, and gastrointestinal AEs. Study discontinuations were higher in the groups receiving N, as expected for the 2-g maximum dose, because of N-associated flushing (7,26). Clinical tolerance of niacin can be improved through patient counseling/education, bedtime dosing with a low-fat snack, prior administration of aspirin or nonsteroidal anti-inflammatory agent, and slow dose titration, all of which were recommended in this study. Even with these measures, this side effect can limit compliance with niacin therapy; however, new strategies are in development that may enable more patients to use niacin at higher doses (28,29). Although the maximum 2-g dose was assessed as a primary end point in this study, it should be noted that the average Niaspan dose used by patients in the U.S. is approximately 1 g (30). As reported in previous studies and in the extended-release niacin product circular, niacin therapy can be associated with elevations in fasting glucose levels (10,20,26,31,32). This effect was also seen in our study, although the increase in fasting glucose lessened over time, as has been the case in other studies (4,31).
Epidemiological analyses and a meta-analysis of clinical lipid trials have suggested that a 40% reduction in LDL-C and a 30% increase in HDL-C together may reduce coronary artery disease risk by approximately 70% (5). Limited placebo-controlled trials suggest that niacin as monotherapy and in combination with LDL-lowering agents improves various cardiovascular end points by 26% to 79% (15,16,19–21). Ongoing studies will provide an assessment of the effect of niacin/statin combination therapy on lipid-depletion in atherosclerotic plaques and clinical outcomes in CHD patients (33–35). Other studies are also ongoing to specifically define the effects of E/S therapy on clinical outcomes (36–38).
The significant improvements in the overall lipid and lipoprotein profile attained after 24 weeks of therapy with the triple combination E/S + N in this study are consistent with the known, complementary lipid-altering effects of the component drugs at these doses. Similarly, the observed flushing side effects, discontinuation rates of the drug groups, and the low and infrequent rates of clinical and laboratory AEs did not seem to exceed the known frequencies for the individual components of this combination. This study also documented the time-related and dose-related effects of niacin on HDL-C and glucose raising, but suggest the glucose effect is somewhat attenuated over time. The continuation of this study for an additional 40 weeks will provide a longer-term evaluation of its safety, lipid-altering efficacy, and time-related effects, and will be reported after its completion. Assessment of the full clinical impact of this triple combination therapy awaits further study in large, randomized trials with adjudicated clinical efficacy and safety outcomes.
For additional tables and an additional figure, please see the online version of this article.
Drs. Guyton, Brown, and Fazio are investigators in the study, of which Merck/Schering-Plough Pharmaceuticals, North Wales, PA, is the funding sponsor, and from whom they have also received honoraria. Drs. Tershakovec and Tomassini and Mr. Polis are employees of Merck & Co., Inc., and as such may own stock or hold other ownership interests in the company.
- Abbreviations and Acronyms
- adverse experience
- coronary heart disease
- creatine kinase
- high-density lipoprotein cholesterol
- high-sensitivity C-reactive protein
- low-density lipoprotein cholesterol
- last observation carried forward
- modified intent-to-treat
- extended-release niacin
- total cholesterol
- upper limit of normal
- Received December 4, 2007.
- Revision received February 19, 2008.
- Accepted March 4, 2008.
- American College of Cardiology Foundation
- Grundy S.M.,
- Cleeman J.I.,
- Merz C.N.B.,
- et al.
- Smith S.C. Jr..,
- Allen J.,
- Blair S.N.,
- et al.
- ↵(2002) National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report: Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). Circulation 106:3143–3421.
- Sanyal S.,
- Karas R.H.,
- Kuvin J.T.
- ↵Niaspan [product insert]. N. Chicago, IL: Abbott Laboratories, 2007.
- Taylor A.J.,
- Sullenberger L.E.,
- Lee H.J.,
- Lee J.K.,
- Grace K.A.
- Bays H.E.,
- Ose L.,
- Fraser N.,
- et al.
- (2003) Vytorin (Merck/Schering-Plough Pharmaceuticals, North Wales, PA) [product insert].
- American Diabetes Association
- O'Riordan M.
- ↵Merck/Schering Plough Pharmaceutical analysis of data from IMS Health. Whitehouse Station, NJ: NPA Monthly, 2007.
- Niacin Plus Statin to Prevent Vascular Events. http://www.clinicaltrials.gov/ct2/show/NCT00120289?term=NCT00120289&rank=1. Accessed March 20, 2008.
- Treatment of HDL to reduce the incidence of vascular events HPS2-THRIVE. http://www.clinicaltrials.gov/ct2/show/NCT00461630. Accessed March 20, 2008.
- Cannon C.P.,
- Guigliano R.P.,
- Blazing M.A.,
- et al.