Clinical Investigations
Diabetes and Metabolism
Effectiveness of ranolazine in patients with type 2 diabetes mellitus and chronic stable angina according to baseline hemoglobin A1c Suzanne V. Arnold, MD, MHA, a,b Darren K. McGuire, MD, MHSC, c John A. Spertus, MD, MPH, a,b Yan Li, PhD, a Patrick Yue, MD, d Ori Ben-Yehuda, MD, e Luiz Belardinelli, MD, d Philip G. Jones, MS, a Ann Olmsted, PhD, d Bernard R. Chaitman, MD, f and Mikhail Kosiborod, MD a, b Kansas City, MO; Dallas, TX; Foster City, CA; New York, NY and St. Louis, MO
Background Ranolazine reduces the frequency of angina and use of sublingual nitroglycerin (SL NTG) in stable angina patients with type 2 diabetes (T2DM). Because pre-clinical data suggest that myocardial late sodium current (INaL), the target of ranolazine, is increased by hyperglycemia, we investigated whether the efficacy of ranolazine was influenced by glycemic control. Methods TERISA was a multinational, randomized, double-blind trial of ranolazine vs. placebo in patients with T2DM and stable angina. Anginal episodes and SL NTG use were recorded daily in an electronic diary. Health status was evaluated at baseline and 8 weeks post-randomization using the Seattle Angina Questionnaire (SAQ). The interaction between baseline HbA1c and treatment effect was tested across endpoints using analysis of covariance models, with HbA1c as a continuous variable with restricted cubic splines. Results
The study included 913 patients, with mean age 63.6 years, 39% women, mean T2DM duration 7.4 years, and mean HbA1c of 7.3%. Heterogeneity of efficacy by HbA1c was observed for the primary endpoint of angina frequency (Pinteraction = .027), the key secondary endpoint of SL NTG use (Pinteraction = .030), SAQ angina frequency (Pinteraction = .001), and SAQ treatment satisfaction (Pinteraction = .025) with greater efficacy of ranolazine in those with higher HbA1c values, increasing continuously from HbA1c levels N6.5%.
Conclusion Among patients with T2DM and stable angina, the therapeutic benefits of ranolazine were greater in those with higher HbA1c values. These data suggest that ranolazine is particularly beneficial in patients with stable angina who have suboptimally controlled T2DM. (Am Heart J 2014;168:457-465.e2.)
Ranolazine is an antianginal medication that exerts antiischemic actions at the cellular level, thereby reducing angina 1–4 and improving quality of life4,5 in patients with chronic angina. Recently, in the TERISA trial, ranolazine was shown to reduce the frequency of angina episodes and use of sublingual nitroglycerin (SL NTG)6 and improve the health status of patients with type 2 diabetes mellitus (T2DM) and stable angina.7 The mechanism of action by which ranolazine
From the aSaint Luke's Mid America Heart Institute, Kansas City, MO, bUniversity of Missouri-Kansas City, Kansas City, MO, cUniversity of Texas Southwestern Medical Center, Dallas, TX, dGilead Sciences, Foster City, CA, eCardiovascular Research Foundation, New York, NY, and fSt. Louis University, St. Louis, MO. RCT# NCT01425359. Submitted March 11, 2014; accepted June 2, 2014. Reprint requests: Suzanne V. Arnold, MD, MHA, 4401 Wornall Rd, Kansas City, MO 64111. E-mail: [email protected] 0002-8703 © 2014, Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.ahj.2014.06.020
is believed to reduce myocardial ischemia and angina is through inhibition of the slowly inactivating component of the cardiac sodium current (INaL), which then reduces the intracellular sodium and calcium overload that accompanies and promotes myocardial ischemia.8,9 Recent preclinical data have shown that cardiac myocytes of mice with type 1 or type 2 diabetes exhibit increased persistent sodium current. 10 We therefore hypothesized that ranolazine, which acts by inhibiting INaL,11 could potentially have a greater therapeutic effect in patients with hyperglycemia. To examine this question, we performed a post-hoc analysis using data from the TERISA trial of ranolazine vs. placebo among patients with T2DM and stable angina to assess whether the therapeutic effectiveness of ranolazine differed according to patients’ underlying baseline glycemic control.
Methods Study overview Details of the TERISA trial have been previously described. 6 Briefly, TERISA was a randomized, double-
458 Arnold et al
American Heart Journal October 2014
Figure 1
Patient flow in TERISA. Enrolment and randomization of patients with type 2 diabetes mellitus and stable angina in the HbA1c sub-study of the TERISA trial.
blind, placebo-controlled trial in which patients with T2DM, coronary artery disease, and stable angina were randomized to ranolazine or placebo for 8 weeks. The study was conducted in 104 sites in 14 countries across Asia, Europe, and North America and was approved by the national regulatory authority in each participating country. The institutional review board or local ethics committee for each site approved the study, and all participating patients provided written informed consent. Patients were required to have been treated with 1 or 2 antianginal therapies at stable doses for ≥2 weeks prior to study entry. Key exclusion criteria were New York Heart Association III-IV heart failure symptoms, acute coronary syndrome in the prior 2 months, planned coronary revascularization during the study period, and prior treatment with ranolazine.
Study design Eligible patients were provided an electronic diary to record and transmit to the coordinating center the number of angina episodes and SL NTG taken each day. After a 4-week, single-blind placebo run-in period to ensure compliance with diary entry and study medication, patients
were randomized to ranolazine (Gilead Sciences, Foster City, CA) 500 mg twice daily—increased to 1000 mg twice daily after 1 week if tolerated—versus matching placebo (patients taking verapamil or diltiazem were not uptitrated). Randomization was stratified by geographic region (Russia, Ukraine, and Belarus vs. other countries), baseline angina frequency category (b3 vs ≥3/week), and number of background antianginal medications (1 vs 2). At randomization, blood was drawn for measurement of hemoglobin A1c (HbA1c). The primary outcome of TERISA was the average weekly number of diary-reported angina episodes from weeks 2 to 8 of treatment. The key secondary endpoint was the average weekly frequency of SL NTG use by diary report. Additional outcomes included the Seattle Angina Questionnaire (SAQ), 12,13 Rose Dyspnea Questionnaire, 14 and Medical Outcomes Study 36-item Short Form, 15 which were collected at randomization and study end. The SAQ is a disease-specific health status instrument that measures clinically important dimensions of health in patients with coronary artery disease: angina frequency, angina stability, physical limitations, treatment satisfaction and disease specific quality of life. Domain scores range from 0 to
American Heart Journal Volume 168, Number 4
Arnold et al 459
Table I. Baseline characteristics of the patients, by HbA1c category
Age (y) Male White race Hypertension Dyslipidemia Current smoking Prior myocardial infarction Prior angioplasty Prior bypass graft surgery Duration of diabetes (yr) Mean ± SD Median (IQR) HbA1c (%) Mean ± SD Median (IQR) Taking diabetic medication Taking insulin Antianginal medications β-Blockers Calcium-channel blockers Long acting nitrates Statins Antiplatelet agents ACE-I/ARBs Russia/Ukraine/Belarus Average weekly angina (diary) Average weekly SL NTG (diary) SAQ angina frequency Diary compliance (%)
HbA1c b6.5%, n = 351
HbA1c 6.5%-8%, n = 311
HbA1c N8%, n = 251
P
64.2 ± 8.8 67.2% 99.7% 95.1% 77.6% 16.8% 72.2% 41.1% 18.8%
64.0 ± 8.7 60.1% 99.0% 92.9% 80.7% 16.4% 71.4% 43.5% 19.9%
62.5 ± 7.7 55.0% 98.4% 98.8% 81.1% 14.3% 80.0% 37.5% 16.3%
.050 .008 .230 .004 .510 .697 .042 .343 .542
5.5 ± 5.7 3.6 (1.4-7.2)
7.5 ± 6.8 5.5 (2.2-10.9)
10.1 ± 7.4 8.5 (4.7-12.9)
b.001
5.9 ± 0.4 6.0 (5.7-6.2) 84.9% 4.6%
7.1 ± 0.4 7.1 (6.8-7.5) 97.4% 18.0%
9.4 ± 1.0 9.1 (8.6-10.1) 98.4% 40.2%
b.001
89.2% 31.3% 31.9% 83.8% 86.0% 88.0% 74.1% 6.9 ± 4.1 4.3 ± 4.3 47.4 ± 18.6 96.0 ± 4.8
91.3% 27.3% 33.1% 82.6% 87.1% 86.8% 67.5% 6.8 ± 4.4 4.4 ± 5.8 47.4 ± 19.4 95.9 ± 5.0
90.0% 25.5% 36.7% 80.9% 92.0% 88.4% 70.5% 7.1 ± 4.4 4.5 ± 4.5 43.7 ± 18.9 95.8 ± 5.0
.651 .258 .465 .655 .068 .822 .178 .677 .790 .031 .941
100, with higher scores indicating less disease burden. For the SAQ angina frequency domain, 10-point improvement represents a clinically-important change within an individual and a 20-point change represents a moderate clinical improvement. 12 The Rose Dyspnea Questionnaire assesses the patient’s level of dyspnea with common activities. Scores range from 0 to 4, with higher scores indicating more limitation due to dyspnea. The Medical Outcomes Study 36item Short Form is a generic health status measure that provides summary scales for overall physical and mental health.15 Summary scores standardize the scores to a U.S. population mean of 50 and a standard deviation of 10, with higher scores indicating better health status. Country-specific instruments were used for each of the health status measures.
Statistical analysis Consistent with the main analysis, the pre-specified efficacy analyses included all randomized patients who took at least 14 days of study drug, completed at least 1 diary entry, and met all major eligibility criteria. 6 Average weekly angina frequency and average SL NTG use were compared between treatment groups by fitting a generalized linear model with negative binomial distribution, adjusted for log baseline angina or SL NTG frequency, number of concomitant antianginal agents, and geographic
b.001 b.001
region. Health status scores between randomization and 8 weeks were compared between treatment groups using analysis of covariance (ANCOVA), adjusted for baseline health status as a continuous variable as well as the average number of weekly angina episodes during trial run-in (≥1 and b3 vs. ≥3 and ≤28), number of concomitant antianginal agents, and geographic region. 6 We then explored the interaction of baseline HbA1c with the treatment effect of ranolazine for each outcome, treating HbA1c as a continuous variable with restricted cubic splines. For end points with significant interaction terms, the effectiveness of ranolazine vs. placebo was plotted against HbA1c. Effectiveness was displayed for diary angina frequency and SL NTG use as the incidence density ratio (IDR), which is the relative difference in the incidence rates of weekly angina frequency or weekly NTG use between ranolazine and placebo, according to a generalized linear model (ie, values b1 indicate lower rates of angina/NTG use among ranolazine patients [ranolazine is more effective]; values N1 indicate higher rates of angina/ NTG use among ranolazine patients). Effectiveness for SAQ domains was displayed as the difference in the change in SAQ scores in ranolazine vs. placebo, according to the ANCOVA model (ie, values N0 indicate greater improvement in SAQ scores among ranolazine patients [ranolazine is more effective]; values b0 indicate less improvement in
American Heart Journal October 2014
460 Arnold et al
Figure 2
Effectiveness of ranolazine vs. placebo by baseline HbA1c. A, IDR for average weekly frequency of angina by diary. B, IDR for average weekly number of SL NTG used by diary. C, Change in SAQ angina frequency. D, Change in SAQ treatment satisfaction. *IDR refers to the Incidence Density Ratio or the relative difference in the incidence rates of weekly angina frequency or weekly NTG use, according to a generalized linear model with negative binomial distribution. Dashed lines indicate 95% confidence intervals.
SAQ scores among ranolazine patients). For endpoints with significant treatment by HbA1c interactions, we examined the 3-way interactions of (1) treatment by HbA1c by insulin and (2) treatment by HbA1c by DM duration, to assure that the association between HbA1c and treatment effect was not confounded by increased insulin use or longer DM duration among patients with poorly-controlled T2DM. Finally, we examined the effect of ranolazine vs. placebo among patients within clinically relevant subgroups of HbA1c, per the most recent American Diabetes Association/European Association for the Study of Diabetes guidelines for stringent and non-stringent glycemic targets 16: well-controlled (HbA1c b6.5%), moderately controlled (HbA1c 6.5%-8%), and poorly controlled (HbA1c N8%). To further explore the health status effects of ranolazine, we also plotted cumulative response curves for SAQ angina frequency by HbA1c category. These curves plot the changes in SAQ angina frequency from baseline to 8 weeks on the x-axis (from +100 to −100)
and the percentage of patients by treatment group who achieved at least that amount of change on the y-axis. All analyses were conducted using SAS v9.2 (SAS Institute, Inc, Cary, NC) and R Version 2.11.1 (Free Software Foundation), and all tests were 2-sided with a nominal type 1 error rate of 5%. TERISA was sponsored by Gilead Sciences, Foster City, CA, USA. All statistical analyses were performed independently by Saint Luke’s Mid America Heart Institute, and the decision to submit the manuscript for publication was made by the TERISA publication committee. 6 The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper and its final contents.
Results Study population From October 2011 through July 2012, 1185 patients were screened, of whom 1142 were considered eligible
American Heart Journal Volume 168, Number 4
Arnold et al 461
Table II. Effectiveness of ranolazine by baseline HbA1c category Ranolazine
Placebo
P
Weekly angina episodes (diary) Well-controlled Moderately controlled Poorly Controlled
3.94 (3.56, 4.36) 3.77 (3.42, 4.15)
3.97 (3.60, 4.37) 4.20 (3.81, 4.64)
0.914 0.121
3.83 (3.43, 4.26)
4.82 (4.33, 5.37)
0.003
Weekly SL NTG Used Well-controlled Moderately controlled Poorly controlled
(diary) 1.86 (1.59, 2.16) 1.53 (1.30, 1.79)
1.76 (1.52, 2.04) 2.19 (1.86, 2.56)
0.624 0.002
2.04 (1.74, 2.40)
2.46 (2.09, 2.90)
0.114
14.2 (10.9, 17.4) 15.2 (11.8, 18.6)
0.869 0.897
10.0 (5.5, 14.5)
b0.001
5.6 (3.2, 7.9) 3.3 (0.9, 5.8)
0.469 0.027
3.9 (0.9, 6.9)
0.044
Change in SAQ angina frequency Well-controlled 14.5 (11.1, 17.8) Moderately 14.9 (11.6, 18.3) controlled Poorly controlled 18.7 (14.3, 23.1) Change in SAQ treatment satisfaction Well-controlled 4.6 (2.1, 7.0) Moderately 6.5 (4.1, 8.9) controlled Poorly controlled 7.1 (4.1, 10.0)
Outcomes are expressed as least squares means and 95% confidence intervals, per the generalized linear models (diary results) or ANCOVA models (SAQ results). Well-controlled diabetes (HbA1c b6.5%; n = 348); moderately controlled (HbA1c 6.5%-8%; n = 305); poorly controlled (HbA1c N8%; n = 250).
and signed informed consent (Figure 1). A total of 193 patients failed the 4-week placebo run-in period and were excluded prior to randomization, leaving 949 to be randomized to either ranolazine or placebo. We subsequently excluded 22 patients who either never initiated the study drug or discontinued it during the first 2 weeks (11 in each arm) and 14 patients with missing HbA1c measurements, leaving a final analytic sample size of 913 patients (455 in the ranolazine arm, 458 in the placebo arm). Baseline characteristics were well-matched between groups (online Appendix Supplementary Table). The mean age was 63.6 ± 8.5 years and 61% were male. Average duration of T2DM was 7.4 ± 6.8 years, and 93% of patients were treated with glucose-lowering medications, including 19% treated with insulin. Mean HbA1c across the analytic population was 7.3 ± 1.5%, with a median of 6.9% (interquartile range, 6.1%-8.2%). There were 351 patients (38%) who had well-controlled T2DM (HbA1c b6.5%), 311 (34%) who had moderately-controlled T2DM (HbA1c 6.5%-8%), and 251 (27%) who had poorly-controlled T2DM (HbA1c N8%). As expected, there were several clinical differences across baseline HbA1c categories (Table I). Patients with worse glycemic control were more likely female, had longer duration of T2DM,
were more likely to be on glucose-lowering medications, and more likely treated with insulin. Importantly, revascularization history, use of antianginal medications, concomitant guideline-recommended medical therapy, enrollment site, and baseline angina did not differ across groups.
Effect of ranolazine by baseline HbA1c Significant treatment by HbA1c interactions were observed for the primary outcome of average weekly angina frequency (P = .027), the key secondary outcome of average weekly use of sublingual nitroglycerin tablets (P = .030), and the health status outcomes of SAQ angina frequency (P = .001) and treatment satisfaction (P = .025), each interaction with greater efficacy at higher HbA1c values. There was no heterogeneity of effect of ranolazine by HbA1c for the remainder of the diseasespecific and generic health status measures (P N .1 for all). When the treatment effects of ranolazine vs. placebo were plotted against baseline HbA1c, there was a continuous relationship between greater therapeutic effectiveness of ranolazine on angina, SL NTG use, SAQ angina frequency, and SAQ treatment satisfaction and higher HbA1c (Figure 2; effectiveness of ranolazine by decile of HbA1c shown in online Appendix Supplementary Figure 1). While the effect of ranolazine by HbA1c varied slightly depending on which outcome was studied, the therapeutic benefit of ranolazine was consistently more pronounced above HbA1c levels ~6.5%. None of the spline terms were significant (P N .05 for all 4), indicating the relationship between treatment effect and HbA1c was statistically linear, with no inflection point above which the relationship changed. As patients with higher HbA1c levels had longer durations of DM and were more likely to be on insulin, we examined the 3-way interactions of treatment by HbA1c by DM duration and treatment by HbA1c by insulin. None of the interactions by DM duration were significant (P N .1 for all), indicating that treatment effect by HbA1c was not impacted by the duration of DM. The interactions by insulin use were not significant for diary angina frequency (P = .15), SAQ angina frequency (P = .70), and SAQ treatment satisfaction (P = .15), and marginally significant for diary SL NTG use (P = .045). In the plot of the treatment effect of ranolazine on SL NTG use by baseline HbA1c, stratified by insulin use, it appeared that the more pronounced effect of ranolazine on SL NTG use by increasing HbA1c was more prominent among patients not on insulin (online Appendix Supplementary Figure 2). Effect of ranolazine by baseline HbA1c category The baseline characteristics of patients within clinically relevant groups of HbA1c are shown in online Appendix Supplementary Table. Patients with poorly-controlled T2DM had a longer duration of T2DM, were more likely
462 Arnold et al
American Heart Journal October 2014
Figure 3
Cumulative response curve of change in SAQ angina frequency by treatment group, according to glycemic control. A, Well controlled T2DM (HbA1c b6.5%). B, Moderately controlled T2DM (HbA1c 6.5%-8%). C, Poorly controlled T2DM (HbA1c N8%). Left-sided panels display the entire study population within each HbA1c subset. Right-sided panels zoom in on the most common changes from baseline to 8 weeks.
to be taking glucose-lowering medications, and were more often on insulin. Neither the frequency of angina nor the pattern of antianginal treatment differed among groups, although patients with poorly-controlled T2DM had slightly lower SAQ angina frequency scores at
baseline (well vs moderate vs poor glucose control: 47.4 vs 47.4 vs 43.7, P = .031). Among patients with well-controlled T2DM (HbA1c b6.5%; n = 348), there were no significant benefits of ranolazine over placebo in terms of improvement in
American Heart Journal Volume 168, Number 4
weekly angina (P = .914), weekly SL NTG used (P = .624), SAQ angina frequency (P = 0.869) or SAQ treatment satisfaction (p = 0.469) (Table II). Among patients with moderately controlled T2DM (HbA1c 6.5%-8%; n = 305), there was modest benefit of ranolazine over placebo in terms of weekly SL NTG used (1.5 vs 2.2, P = .002), and SAQ treatment satisfaction (6.5 vs 3.3 point increase, P = .027), a non-significant trend toward a reduction in weekly angina (3.8 vs 4.2 episodes per week, P = .121) but no significant effect on SAQ angina frequency (P = .897). In patients with poorly-controlled T2DM (HbA1c N8%; n = 250), there was a marked benefit of ranolazine over placebo across 3 of the 4 outcomes, with lower weekly angina (3.8 vs 4.8 episodes per week, P = .003), higher SAQ angina frequency scores (18.7 vs 10.0 point increase, P b .001), and higher SAQ treatment satisfaction (7.1 vs 3.9 point increase, P = .044), and a non-significant trend toward lower weekly SL NTG used (2.0 vs 2.5, P = .114). Examining the cumulative response curves for SAQ angina frequency revealed no significant differences among patients with HbA1c ≤8% (Figure 3). However, among patients with poor glycemic control, 73% of patients treated with ranolazine reported a clinical improvement in angina frequency (i.e., at least a 10-point improvement) versus 46% of placebo-treated patients (P b .001; absolute difference = 27%, number needed to treat with ranolazine to have 1 patient report a clinically-relevant improvement in angina of ~4). Furthermore, 56% of ranolazine-treated patients reported at least a moderate clinical improvement in angina frequency (ie, 20-point improvement) versus 37% of placebo-treated patients (P = .002; absolute difference = 19%; number needed to treat of ~5).
Discussion In the first prospective clinical trial of antianginal therapy specifically in patients with T2DM and stable angina, we found the benefits of ranolazine in reducing angina burden varied by patients’ glycemic control. The effectiveness of ranolazine improved linearly with worsening glycemic control, with its therapeutic benefit across the broad range of patient-reported outcomes being more pronounced at HbA1c N6.5%. Patients with HbA1c N8% experienced marked improvements in angina frequency, SL NTG use, and in health status on ranolazine vs. placebo, which were both highly statistically significant and clinically meaningful. While these results need to be confirmed in future studies, they suggest that ranolazine may be more beneficial in patients with angina and suboptimally-controlled T2DM. Our observation that ranolazine is associated with greater angina relief and better health status among patients with suboptimal glycemic control is particularly relevant, as ranolazine may also produce improvements in fasting glucose and HbA1c. Post hoc analyses from the CARISA and MERLIN-TIMI 36 trials demonstrated statisti-
Arnold et al 463
cally significant and clinically-relevant reductions in HbA1c levels in patients treated with ranolazine vs. placebo, especially notable as both trials allowed openlabel use and titration of glucose-lowering medications. 17,18 While the mechanism of action for the glycometabolic effect of ranolazine remains to be fully elucidated, recent preclinical studies suggest that it may be mediated by inhibition of sodium channel activity in pancreatic alpha cells and a resultant reduction in glucagon secretion. 19–23 Several prospective randomized clinical trials are presently ongoing to more clearly establish the glucose-lowering effect of ranolazine (clinicaltrials.gov: NCT01494987, NCT01472185, NCT01555164). The mechanism(s) underlying the greater antianginal benefit of ranolazine among patients with higher HbA1c levels remain to be definitively determined, but several possible explanations exist. Preclinical studies have demonstrated that cardiomyocytes from diabetic rats exposed to high glucose concentrations have increased phosphorylation of CaMKII, 24 a kinase that increases INaL. 25 Furthermore, recent data have directly shown that cardiac myocytes of mice with type 1 or type 2 diabetes exhibit increased persistent sodium current (INaP). 10 As the antianginal effect of ranolazine is mediated by inhibition of INaL, 11 its therapeutic effectiveness is expected to be enhanced in states of hyperglycemia via this mechanism. Alternatively, hyperglycemia may worsen myocardial perfusion through a variety of mechanisms, including impaired coronary vasomotor function, greater platelet aggregation, hypercoagulability, inflammation, and oxidative stress. 26 The greater antianginal benefit of ranolazine in patients with worse glucose control may, thus, also be mediated by the purported glucose-lowering action of ranolazine, which could potentially ameliorate these detrimental effects of hyperglycemia. 27 However, because improvement in glucose control has not been shown to improve angina burden, particularly in the short time frame of the TERISA trial, this latter explanation appears to be less plausible. Finally, it is possible that other factors associated with high HbA1c levels are also associated with greater effectiveness of ranolazine. We explicitly tested the 3-way interaction between treatment, HbA1c, and insulin use and found this not significant for 3 of the outcomes and marginally significant for SL NTG. However, the direction of effect, with less ranolazine effect by HbA1c among patients on insulin, indicates the mechanism of increased antianginal is unlikely to be mediated by insulin use. As such, we believe that the increased effectiveness of ranolazine in patients with poorly-controlled T2DM likely relates directly to the increased INaP in the setting of hyperglycemia. While we observed only small differences in angina among patients treated with ranolazine vs. placebo whose glucose was well-controlled, these findings should be interpreted with caution. First, the TERISA study was designed to evaluate ranolazine in the
American Heart Journal October 2014
464 Arnold et al
population of patients with T2DM and angina and was not powered to detect differences in angina across the subgroups of patients with various HbA1c levels. Second, the confidence intervals around the estimates of the effectiveness of ranolazine were wide at lower levels of HbA1c. Third, we noted a greater placebo effect among patients with well-controlled T2DM, possibly reflecting a difference in the type of patients in this subgroup, making the detection of an effect of ranolazine above this larger placebo effect more difficult. In addition, several prior studies, including CARISA, ERICA, and MERLIN-TIMI 36, have established ranolazine as an effective antianginal agent in a broad population of patients both with and without T2DM. 3–5 Therefore, the clinical implications of this finding are uncertain but provocative in suggesting a heterogeneity of treatment benefit related to patients' glycemic control. Further studies are needed to help determine the patients most (and least) likely to benefit from ranolazine in terms of angina and quality of life. Our results should be interpreted in the context of several potential limitations. First, the analyses based on HbA1c were not pre-specified in the study protocol. Although our study was hypothesis-driven and supported by mechanistic data, given the post-hoc nature of the analysis, these results will need to be confirmed in future studies. In addition, although our observations persisted after adjustment for several key covariates and were consistent across a broad range of patient-reported outcomes, residual confounding cannot be definitively excluded. Second, the relatively short duration of followup in TERISA limits our ability to determine whether the differential antianginal effect of ranolazine in patients with various baseline HbA1c levels persists over time. Third, HbA1c levels were only assessed at baseline in TERISA. While post-hoc analyses have demonstrated ranolazine may also improve fasting glucose and HbA1c, thereby providing a dual benefit in patients with poorlycontrolled T2DM (i.e., improved glycemic control and more antianginal benefit), the glucose-lowering effect of ranolazine was not tested in TERISA and will need to be defined in on-going prospective randomized clinical trials. Finally, the majority of patients in TERISA were enrolled in European centers, and, thus, whether the observed interaction between the antianginal effect of ranolazine and baseline HbA1c applies to blacks, Hispanics and Asians—populations with growing T2DM prevalence 28—remains unclear. In summary, our study demonstrates that the antianginal effect of ranolazine is more pronounced in patients with higher vs. lower baseline HbA1c, with greater improvements observed in angina burden and health status measures. If confirmed in future studies, these findings suggest that ranolazine may be a good therapeutic option for patients with angina and suboptimallycontrolled T2DM.
References 1. Chaitman BR, Skettino SL, Parker JO, et al. Anti-ischemic effects and long-term survival during ranolazine monotherapy in patients with chronic severe angina. J Am Coll Cardiol 2004;43(8):1375-82. 2. Rousseau MF, Pouleur H, Cocco G, et al. Comparative efficacy of ranolazine versus atenolol for chronic angina pectoris. Am J Cardiol 2005;95(3):311-6. 3. Chaitman BR, Pepine CJ, Parker JO, et al. Effects of ranolazine with atenolol, amlodipine, or diltiazem on exercise tolerance and angina frequency in patients with severe chronic angina: a randomized controlled trial. JAMA 2004;291(3):309-16. 4. Stone PH, Gratsiansky NA, Blokhin A, et al. Antianginal efficacy of ranolazine when added to treatment with amlodipine: the ERICA (Efficacy of Ranolazine in Chronic Angina) trial. J Am Coll Cardiol 2006;48(3):566-75. 5. Arnold SV, Morrow DA, Wang K, et al. Effects of ranolazine on disease-specific health status and quality of life among patients with acute coronary syndromes: results from the MERLIN-TIMI 36 randomized trial. Circ Cardiovasc Qual Outcomes 2008;1(2):107-15. 6. Kosiborod M, Arnold SV, Spertus JA, et al. Evaluation of ranolazine in patients with type 2 diabetes mellitus and chronic stable angina: results from the TERISA randomized clinical trial (Type 2 Diabetes Evaluation of Ranolazine in Subjects With Chronic Stable Angina). J Am Coll Cardiol 2013;61(20):2038-45. 7. Arnold SV, Kosiborod M, McGuire DK, et al. Effects of Ranolazine on Quality of Life Among Patients With Diabetes Mellitus and Stable Angina. JAMA Intern Med 2014 Jun 2. http://dx.doi.org/10.1001/ jamainternmed.2014.2120. [Epub ahead of print]. 8. Chaitman BR. Ranolazine for the treatment of chronic angina and potential use in other cardiovascular conditions. Circulation 2006;113(20):2462-72. 9. Belardinelli L, Shryock JC, Fraser H. Inhibition of the late sodium current as a potential cardioprotective principle: effects of the late sodium current inhibitor ranolazine. Heart 2006;92(Suppl 4):iv6-iv14. 10. Lu Z, Jiang YP, Wu CY, et al. Increased persistent sodium current Due to decreased PI3K signaling contributes to QT prolongation in the diabetic heart. Diabetes 2013;62(12):4257-65. 11. Antzelevitch C, Belardinelli L, Zygmunt AC, et al. Electrophysiological effects of ranolazine, a novel antianginal agent with antiarrhythmic properties. Circulation 2004;110(8):904-10. 12. Spertus JA, Winder JA, Dewhurst TA, et al. Monitoring the quality of life in patients with coronary artery disease. Am J Cardiol 1994;74 (12):1240-4. 13. Spertus JA, Winder JA, Dewhurst TA, et al. Development and evaluation of the Seattle Angina Questionnaire: a new functional status measure for coronary artery disease. J Am Coll Cardiol 1995;25(2):333-41. 14. Rose GA, Blackburn H. Cardiovascular survey methods. Monogr Ser World Health Organ 1968;56:1-188. 15. Ware Jr JE, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 1992;30(6):473-83. 16. Inzucchi SE, Bergenstal RM, Buse JB, et al. Management of hyperglycemia in type 2 diabetes: a patient-centered approach: position statement of the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetes Care 2012;35(6):1364-79. 17. Morrow DA, Scirica BM, Chaitman BR, et al. Evaluation of the glycometabolic effects of ranolazine in patients with and without
American Heart Journal Volume 168, Number 4
18.
19.
20.
21.
22.
23.
diabetes mellitus in the MERLIN-TIMI 36 randomized controlled trial. Circulation 2009;119(15):2032-9. Timmis AD, Chaitman BR, Crager M. Effects of ranolazine on exercise tolerance and HbA1c in patients with chronic angina and diabetes. Eur Heart J 2006;27(1):42-8. Yang M, Chu R, Belardinelli L, et al. Inhibition of glucagon secretion in rat pancreatic islets via Na + channels: mechanism of anti-diabetic effect of ranolazine. Diabetes 2013;62(Suppl 1):A234. Kahlig K, Yang M, Cheng X, et al. The mechanism of the anti-diabetic effect of ranolazine: inhibition of glucagon secretion in rat pancreatic islets via Na + channel blockade. Diabetes 2013;62(Suppl 1):A272. Dhalla A, Yang M, Chu R, et al. Ranolazine inhibits glucogon secretion from human pancreatic islets via blockade of the Nav1.3 channels in alpha-cells. Diabetes 2013;62(Suppl 1):A227. Ning Y, Van Petten M, Jiang J, et al. Na channel blockers exert anti-diabetic effects via lowering glucagon levels in ZDF diabetic rats. Diabetes 2013;62(Suppl 1):A280. Ning Y, Jiang J, Van Petten M, et al. Ranolazine suppresses exaggerated postprandial glucagon response in STZ-induced diabetic rats. Diabetes 2013;62(Suppl 1):A235.
Arnold et al 465
24. Nishio S, Teshima Y, Takahashi N, et al. Activation of CaMKII as a key regulator of reactive oxygen species production in diabetic rat heart. J Mol Cell Cardiol 2012;52(5):1103-11. 25. Koval OM, Snyder JS, Wolf RM, et al. Ca2+/calmodulin-dependent protein kinase II-based regulation of voltage-gated Na + channel in cardiac disease. Circulation 2012;126(17):2084-94. 26. Paneni F, Beckman JA, Creager MA, et al. Diabetes and vascular disease: pathophysiology, clinical consequences, and medical therapy: part I. Eur Heart J 2013;34(31):2436-43. 27. Deedwania P, Kosiborod M, Barrett E, et al. Hyperglycemia and acute coronary syndrome: a scientific statement from the American Heart Association Diabetes Committee of the Council on Nutrition, Physical Activity, and Metabolism. Circulation 2008;117(12): 1610-9. 28. Department of Health and Human Services. Centers for disease control and prevention. Age-adjusted percentage of civilian, noninstitutionalized population with diagnosed diabetes, by race and sex, United States, 1980–2011. http://www.cdc.gov/diabetes/ statistics/prev/national/figraceethsex.htm. [Accessed October 22, 2013. In].
American Heart Journal Volume 168, Number 4
Arnold et al 465.e1
Appendix Supplementary Table. Baseline characteristics of the patients, by treatment group
Age (y) Male White race Hypertension Dyslipidemia Current smoking Prior myocardial infarction Prior angioplasty Prior bypass graft surgery Duration of diabetes (y) Mean ± SD Median (IQR) Absolute range HbA1c (%) Mean ± SD Median (IQR) Absolute range Taking diabetic medication Taking insulin Antianginal medications β-Blockers Calcium-channel blockers Long acting nitrates Statins Antiplatelet agents ACE-I/ARBs Baseline weekly angina by diary Russia/Ukraine/Belarus
Ranolazine n = 455
Placebo n = 458
63.1 ± 8.5⁎ 61.1% 98.9% 94.9% 79.0% 15.4% 75.2% 42.7% 18.2%
64.2 ± 8.4⁎ 61.8% 99.3% 95.8% 80.3% 16.6% 73.0% 39.2% 18.8%
7.2 ± 6.7 4.9 (2.2-10.0) 0.0-41.0
7.7 ± 7.0 6.0 (2.6-10.8) 0.0-42.6
7.3 ± 1.5 6.9 (6.2-8.3) 4.7-12.5 93.2% 17.4%
7.3 ± 1.5 6.9 (6.1-8.1) 5.0-12.9 92.6% 20.5%
90.5% 26.6% 34.9% 82.4% 89.9% 87.9% 6.8 ± 4.3 71.6%
89.7% 30.1% 32.3% 82.8% 86.2% 87.6% 6.9 ± 4.3 70.1%
⁎ P N .05 for all comparisons except age (P = .038)
465.e2 Arnold et al
American Heart Journal October 2014
Supplementary Figure 1
Effectiveness of Ranolazine Versus Placebo by Decile of Baseline HbA1c. A. IDR for average weekly frequency of angina by diary. B. IDR for average weekly number of SL NTG used by diary. C. Change in SAQ angina frequency. D. Change in SAQ treatment satisfaction.
Supplementary Figure 2
Effectiveness of Ranolazine vs. Placebo on SL NTG Use by Baseline HbA1c, Stratified by Insulin Use. IDR for average weekly number of SL NTG used by diary (IDR b1 indicates ranolazine is more effective vs. IDR N1 indicates placebo is more effective). Red line is patients on insulin. Black line is patients not on insulin.
Diabetes and Metabolism
Effectiveness of ranolazine in patients with type 2 diabetes mellitus and chronic stable angina according to baseline hemoglobin A1c Suzanne V. Arnold, MD, MHA, a,b Darren K. McGuire, MD, MHSC, c John A. Spertus, MD, MPH, a,b Yan Li, PhD, a Patrick Yue, MD, d Ori Ben-Yehuda, MD, e Luiz Belardinelli, MD, d Philip G. Jones, MS, a Ann Olmsted, PhD, d Bernard R. Chaitman, MD, f and Mikhail Kosiborod, MD a, b Kansas City, MO; Dallas, TX; Foster City, CA; New York, NY and St. Louis, MO
Background Ranolazine reduces the frequency of angina and use of sublingual nitroglycerin (SL NTG) in stable angina patients with type 2 diabetes (T2DM). Because pre-clinical data suggest that myocardial late sodium current (INaL), the target of ranolazine, is increased by hyperglycemia, we investigated whether the efficacy of ranolazine was influenced by glycemic control. Methods TERISA was a multinational, randomized, double-blind trial of ranolazine vs. placebo in patients with T2DM and stable angina. Anginal episodes and SL NTG use were recorded daily in an electronic diary. Health status was evaluated at baseline and 8 weeks post-randomization using the Seattle Angina Questionnaire (SAQ). The interaction between baseline HbA1c and treatment effect was tested across endpoints using analysis of covariance models, with HbA1c as a continuous variable with restricted cubic splines. Results
The study included 913 patients, with mean age 63.6 years, 39% women, mean T2DM duration 7.4 years, and mean HbA1c of 7.3%. Heterogeneity of efficacy by HbA1c was observed for the primary endpoint of angina frequency (Pinteraction = .027), the key secondary endpoint of SL NTG use (Pinteraction = .030), SAQ angina frequency (Pinteraction = .001), and SAQ treatment satisfaction (Pinteraction = .025) with greater efficacy of ranolazine in those with higher HbA1c values, increasing continuously from HbA1c levels N6.5%.
Conclusion Among patients with T2DM and stable angina, the therapeutic benefits of ranolazine were greater in those with higher HbA1c values. These data suggest that ranolazine is particularly beneficial in patients with stable angina who have suboptimally controlled T2DM. (Am Heart J 2014;168:457-465.e2.)
Ranolazine is an antianginal medication that exerts antiischemic actions at the cellular level, thereby reducing angina 1–4 and improving quality of life4,5 in patients with chronic angina. Recently, in the TERISA trial, ranolazine was shown to reduce the frequency of angina episodes and use of sublingual nitroglycerin (SL NTG)6 and improve the health status of patients with type 2 diabetes mellitus (T2DM) and stable angina.7 The mechanism of action by which ranolazine
From the aSaint Luke's Mid America Heart Institute, Kansas City, MO, bUniversity of Missouri-Kansas City, Kansas City, MO, cUniversity of Texas Southwestern Medical Center, Dallas, TX, dGilead Sciences, Foster City, CA, eCardiovascular Research Foundation, New York, NY, and fSt. Louis University, St. Louis, MO. RCT# NCT01425359. Submitted March 11, 2014; accepted June 2, 2014. Reprint requests: Suzanne V. Arnold, MD, MHA, 4401 Wornall Rd, Kansas City, MO 64111. E-mail: [email protected] 0002-8703 © 2014, Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.ahj.2014.06.020
is believed to reduce myocardial ischemia and angina is through inhibition of the slowly inactivating component of the cardiac sodium current (INaL), which then reduces the intracellular sodium and calcium overload that accompanies and promotes myocardial ischemia.8,9 Recent preclinical data have shown that cardiac myocytes of mice with type 1 or type 2 diabetes exhibit increased persistent sodium current. 10 We therefore hypothesized that ranolazine, which acts by inhibiting INaL,11 could potentially have a greater therapeutic effect in patients with hyperglycemia. To examine this question, we performed a post-hoc analysis using data from the TERISA trial of ranolazine vs. placebo among patients with T2DM and stable angina to assess whether the therapeutic effectiveness of ranolazine differed according to patients’ underlying baseline glycemic control.
Methods Study overview Details of the TERISA trial have been previously described. 6 Briefly, TERISA was a randomized, double-
458 Arnold et al
American Heart Journal October 2014
Figure 1
Patient flow in TERISA. Enrolment and randomization of patients with type 2 diabetes mellitus and stable angina in the HbA1c sub-study of the TERISA trial.
blind, placebo-controlled trial in which patients with T2DM, coronary artery disease, and stable angina were randomized to ranolazine or placebo for 8 weeks. The study was conducted in 104 sites in 14 countries across Asia, Europe, and North America and was approved by the national regulatory authority in each participating country. The institutional review board or local ethics committee for each site approved the study, and all participating patients provided written informed consent. Patients were required to have been treated with 1 or 2 antianginal therapies at stable doses for ≥2 weeks prior to study entry. Key exclusion criteria were New York Heart Association III-IV heart failure symptoms, acute coronary syndrome in the prior 2 months, planned coronary revascularization during the study period, and prior treatment with ranolazine.
Study design Eligible patients were provided an electronic diary to record and transmit to the coordinating center the number of angina episodes and SL NTG taken each day. After a 4-week, single-blind placebo run-in period to ensure compliance with diary entry and study medication, patients
were randomized to ranolazine (Gilead Sciences, Foster City, CA) 500 mg twice daily—increased to 1000 mg twice daily after 1 week if tolerated—versus matching placebo (patients taking verapamil or diltiazem were not uptitrated). Randomization was stratified by geographic region (Russia, Ukraine, and Belarus vs. other countries), baseline angina frequency category (b3 vs ≥3/week), and number of background antianginal medications (1 vs 2). At randomization, blood was drawn for measurement of hemoglobin A1c (HbA1c). The primary outcome of TERISA was the average weekly number of diary-reported angina episodes from weeks 2 to 8 of treatment. The key secondary endpoint was the average weekly frequency of SL NTG use by diary report. Additional outcomes included the Seattle Angina Questionnaire (SAQ), 12,13 Rose Dyspnea Questionnaire, 14 and Medical Outcomes Study 36-item Short Form, 15 which were collected at randomization and study end. The SAQ is a disease-specific health status instrument that measures clinically important dimensions of health in patients with coronary artery disease: angina frequency, angina stability, physical limitations, treatment satisfaction and disease specific quality of life. Domain scores range from 0 to
American Heart Journal Volume 168, Number 4
Arnold et al 459
Table I. Baseline characteristics of the patients, by HbA1c category
Age (y) Male White race Hypertension Dyslipidemia Current smoking Prior myocardial infarction Prior angioplasty Prior bypass graft surgery Duration of diabetes (yr) Mean ± SD Median (IQR) HbA1c (%) Mean ± SD Median (IQR) Taking diabetic medication Taking insulin Antianginal medications β-Blockers Calcium-channel blockers Long acting nitrates Statins Antiplatelet agents ACE-I/ARBs Russia/Ukraine/Belarus Average weekly angina (diary) Average weekly SL NTG (diary) SAQ angina frequency Diary compliance (%)
HbA1c b6.5%, n = 351
HbA1c 6.5%-8%, n = 311
HbA1c N8%, n = 251
P
64.2 ± 8.8 67.2% 99.7% 95.1% 77.6% 16.8% 72.2% 41.1% 18.8%
64.0 ± 8.7 60.1% 99.0% 92.9% 80.7% 16.4% 71.4% 43.5% 19.9%
62.5 ± 7.7 55.0% 98.4% 98.8% 81.1% 14.3% 80.0% 37.5% 16.3%
.050 .008 .230 .004 .510 .697 .042 .343 .542
5.5 ± 5.7 3.6 (1.4-7.2)
7.5 ± 6.8 5.5 (2.2-10.9)
10.1 ± 7.4 8.5 (4.7-12.9)
b.001
5.9 ± 0.4 6.0 (5.7-6.2) 84.9% 4.6%
7.1 ± 0.4 7.1 (6.8-7.5) 97.4% 18.0%
9.4 ± 1.0 9.1 (8.6-10.1) 98.4% 40.2%
b.001
89.2% 31.3% 31.9% 83.8% 86.0% 88.0% 74.1% 6.9 ± 4.1 4.3 ± 4.3 47.4 ± 18.6 96.0 ± 4.8
91.3% 27.3% 33.1% 82.6% 87.1% 86.8% 67.5% 6.8 ± 4.4 4.4 ± 5.8 47.4 ± 19.4 95.9 ± 5.0
90.0% 25.5% 36.7% 80.9% 92.0% 88.4% 70.5% 7.1 ± 4.4 4.5 ± 4.5 43.7 ± 18.9 95.8 ± 5.0
.651 .258 .465 .655 .068 .822 .178 .677 .790 .031 .941
100, with higher scores indicating less disease burden. For the SAQ angina frequency domain, 10-point improvement represents a clinically-important change within an individual and a 20-point change represents a moderate clinical improvement. 12 The Rose Dyspnea Questionnaire assesses the patient’s level of dyspnea with common activities. Scores range from 0 to 4, with higher scores indicating more limitation due to dyspnea. The Medical Outcomes Study 36item Short Form is a generic health status measure that provides summary scales for overall physical and mental health.15 Summary scores standardize the scores to a U.S. population mean of 50 and a standard deviation of 10, with higher scores indicating better health status. Country-specific instruments were used for each of the health status measures.
Statistical analysis Consistent with the main analysis, the pre-specified efficacy analyses included all randomized patients who took at least 14 days of study drug, completed at least 1 diary entry, and met all major eligibility criteria. 6 Average weekly angina frequency and average SL NTG use were compared between treatment groups by fitting a generalized linear model with negative binomial distribution, adjusted for log baseline angina or SL NTG frequency, number of concomitant antianginal agents, and geographic
b.001 b.001
region. Health status scores between randomization and 8 weeks were compared between treatment groups using analysis of covariance (ANCOVA), adjusted for baseline health status as a continuous variable as well as the average number of weekly angina episodes during trial run-in (≥1 and b3 vs. ≥3 and ≤28), number of concomitant antianginal agents, and geographic region. 6 We then explored the interaction of baseline HbA1c with the treatment effect of ranolazine for each outcome, treating HbA1c as a continuous variable with restricted cubic splines. For end points with significant interaction terms, the effectiveness of ranolazine vs. placebo was plotted against HbA1c. Effectiveness was displayed for diary angina frequency and SL NTG use as the incidence density ratio (IDR), which is the relative difference in the incidence rates of weekly angina frequency or weekly NTG use between ranolazine and placebo, according to a generalized linear model (ie, values b1 indicate lower rates of angina/NTG use among ranolazine patients [ranolazine is more effective]; values N1 indicate higher rates of angina/ NTG use among ranolazine patients). Effectiveness for SAQ domains was displayed as the difference in the change in SAQ scores in ranolazine vs. placebo, according to the ANCOVA model (ie, values N0 indicate greater improvement in SAQ scores among ranolazine patients [ranolazine is more effective]; values b0 indicate less improvement in
American Heart Journal October 2014
460 Arnold et al
Figure 2
Effectiveness of ranolazine vs. placebo by baseline HbA1c. A, IDR for average weekly frequency of angina by diary. B, IDR for average weekly number of SL NTG used by diary. C, Change in SAQ angina frequency. D, Change in SAQ treatment satisfaction. *IDR refers to the Incidence Density Ratio or the relative difference in the incidence rates of weekly angina frequency or weekly NTG use, according to a generalized linear model with negative binomial distribution. Dashed lines indicate 95% confidence intervals.
SAQ scores among ranolazine patients). For endpoints with significant treatment by HbA1c interactions, we examined the 3-way interactions of (1) treatment by HbA1c by insulin and (2) treatment by HbA1c by DM duration, to assure that the association between HbA1c and treatment effect was not confounded by increased insulin use or longer DM duration among patients with poorly-controlled T2DM. Finally, we examined the effect of ranolazine vs. placebo among patients within clinically relevant subgroups of HbA1c, per the most recent American Diabetes Association/European Association for the Study of Diabetes guidelines for stringent and non-stringent glycemic targets 16: well-controlled (HbA1c b6.5%), moderately controlled (HbA1c 6.5%-8%), and poorly controlled (HbA1c N8%). To further explore the health status effects of ranolazine, we also plotted cumulative response curves for SAQ angina frequency by HbA1c category. These curves plot the changes in SAQ angina frequency from baseline to 8 weeks on the x-axis (from +100 to −100)
and the percentage of patients by treatment group who achieved at least that amount of change on the y-axis. All analyses were conducted using SAS v9.2 (SAS Institute, Inc, Cary, NC) and R Version 2.11.1 (Free Software Foundation), and all tests were 2-sided with a nominal type 1 error rate of 5%. TERISA was sponsored by Gilead Sciences, Foster City, CA, USA. All statistical analyses were performed independently by Saint Luke’s Mid America Heart Institute, and the decision to submit the manuscript for publication was made by the TERISA publication committee. 6 The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper and its final contents.
Results Study population From October 2011 through July 2012, 1185 patients were screened, of whom 1142 were considered eligible
American Heart Journal Volume 168, Number 4
Arnold et al 461
Table II. Effectiveness of ranolazine by baseline HbA1c category Ranolazine
Placebo
P
Weekly angina episodes (diary) Well-controlled Moderately controlled Poorly Controlled
3.94 (3.56, 4.36) 3.77 (3.42, 4.15)
3.97 (3.60, 4.37) 4.20 (3.81, 4.64)
0.914 0.121
3.83 (3.43, 4.26)
4.82 (4.33, 5.37)
0.003
Weekly SL NTG Used Well-controlled Moderately controlled Poorly controlled
(diary) 1.86 (1.59, 2.16) 1.53 (1.30, 1.79)
1.76 (1.52, 2.04) 2.19 (1.86, 2.56)
0.624 0.002
2.04 (1.74, 2.40)
2.46 (2.09, 2.90)
0.114
14.2 (10.9, 17.4) 15.2 (11.8, 18.6)
0.869 0.897
10.0 (5.5, 14.5)
b0.001
5.6 (3.2, 7.9) 3.3 (0.9, 5.8)
0.469 0.027
3.9 (0.9, 6.9)
0.044
Change in SAQ angina frequency Well-controlled 14.5 (11.1, 17.8) Moderately 14.9 (11.6, 18.3) controlled Poorly controlled 18.7 (14.3, 23.1) Change in SAQ treatment satisfaction Well-controlled 4.6 (2.1, 7.0) Moderately 6.5 (4.1, 8.9) controlled Poorly controlled 7.1 (4.1, 10.0)
Outcomes are expressed as least squares means and 95% confidence intervals, per the generalized linear models (diary results) or ANCOVA models (SAQ results). Well-controlled diabetes (HbA1c b6.5%; n = 348); moderately controlled (HbA1c 6.5%-8%; n = 305); poorly controlled (HbA1c N8%; n = 250).
and signed informed consent (Figure 1). A total of 193 patients failed the 4-week placebo run-in period and were excluded prior to randomization, leaving 949 to be randomized to either ranolazine or placebo. We subsequently excluded 22 patients who either never initiated the study drug or discontinued it during the first 2 weeks (11 in each arm) and 14 patients with missing HbA1c measurements, leaving a final analytic sample size of 913 patients (455 in the ranolazine arm, 458 in the placebo arm). Baseline characteristics were well-matched between groups (online Appendix Supplementary Table). The mean age was 63.6 ± 8.5 years and 61% were male. Average duration of T2DM was 7.4 ± 6.8 years, and 93% of patients were treated with glucose-lowering medications, including 19% treated with insulin. Mean HbA1c across the analytic population was 7.3 ± 1.5%, with a median of 6.9% (interquartile range, 6.1%-8.2%). There were 351 patients (38%) who had well-controlled T2DM (HbA1c b6.5%), 311 (34%) who had moderately-controlled T2DM (HbA1c 6.5%-8%), and 251 (27%) who had poorly-controlled T2DM (HbA1c N8%). As expected, there were several clinical differences across baseline HbA1c categories (Table I). Patients with worse glycemic control were more likely female, had longer duration of T2DM,
were more likely to be on glucose-lowering medications, and more likely treated with insulin. Importantly, revascularization history, use of antianginal medications, concomitant guideline-recommended medical therapy, enrollment site, and baseline angina did not differ across groups.
Effect of ranolazine by baseline HbA1c Significant treatment by HbA1c interactions were observed for the primary outcome of average weekly angina frequency (P = .027), the key secondary outcome of average weekly use of sublingual nitroglycerin tablets (P = .030), and the health status outcomes of SAQ angina frequency (P = .001) and treatment satisfaction (P = .025), each interaction with greater efficacy at higher HbA1c values. There was no heterogeneity of effect of ranolazine by HbA1c for the remainder of the diseasespecific and generic health status measures (P N .1 for all). When the treatment effects of ranolazine vs. placebo were plotted against baseline HbA1c, there was a continuous relationship between greater therapeutic effectiveness of ranolazine on angina, SL NTG use, SAQ angina frequency, and SAQ treatment satisfaction and higher HbA1c (Figure 2; effectiveness of ranolazine by decile of HbA1c shown in online Appendix Supplementary Figure 1). While the effect of ranolazine by HbA1c varied slightly depending on which outcome was studied, the therapeutic benefit of ranolazine was consistently more pronounced above HbA1c levels ~6.5%. None of the spline terms were significant (P N .05 for all 4), indicating the relationship between treatment effect and HbA1c was statistically linear, with no inflection point above which the relationship changed. As patients with higher HbA1c levels had longer durations of DM and were more likely to be on insulin, we examined the 3-way interactions of treatment by HbA1c by DM duration and treatment by HbA1c by insulin. None of the interactions by DM duration were significant (P N .1 for all), indicating that treatment effect by HbA1c was not impacted by the duration of DM. The interactions by insulin use were not significant for diary angina frequency (P = .15), SAQ angina frequency (P = .70), and SAQ treatment satisfaction (P = .15), and marginally significant for diary SL NTG use (P = .045). In the plot of the treatment effect of ranolazine on SL NTG use by baseline HbA1c, stratified by insulin use, it appeared that the more pronounced effect of ranolazine on SL NTG use by increasing HbA1c was more prominent among patients not on insulin (online Appendix Supplementary Figure 2). Effect of ranolazine by baseline HbA1c category The baseline characteristics of patients within clinically relevant groups of HbA1c are shown in online Appendix Supplementary Table. Patients with poorly-controlled T2DM had a longer duration of T2DM, were more likely
462 Arnold et al
American Heart Journal October 2014
Figure 3
Cumulative response curve of change in SAQ angina frequency by treatment group, according to glycemic control. A, Well controlled T2DM (HbA1c b6.5%). B, Moderately controlled T2DM (HbA1c 6.5%-8%). C, Poorly controlled T2DM (HbA1c N8%). Left-sided panels display the entire study population within each HbA1c subset. Right-sided panels zoom in on the most common changes from baseline to 8 weeks.
to be taking glucose-lowering medications, and were more often on insulin. Neither the frequency of angina nor the pattern of antianginal treatment differed among groups, although patients with poorly-controlled T2DM had slightly lower SAQ angina frequency scores at
baseline (well vs moderate vs poor glucose control: 47.4 vs 47.4 vs 43.7, P = .031). Among patients with well-controlled T2DM (HbA1c b6.5%; n = 348), there were no significant benefits of ranolazine over placebo in terms of improvement in
American Heart Journal Volume 168, Number 4
weekly angina (P = .914), weekly SL NTG used (P = .624), SAQ angina frequency (P = 0.869) or SAQ treatment satisfaction (p = 0.469) (Table II). Among patients with moderately controlled T2DM (HbA1c 6.5%-8%; n = 305), there was modest benefit of ranolazine over placebo in terms of weekly SL NTG used (1.5 vs 2.2, P = .002), and SAQ treatment satisfaction (6.5 vs 3.3 point increase, P = .027), a non-significant trend toward a reduction in weekly angina (3.8 vs 4.2 episodes per week, P = .121) but no significant effect on SAQ angina frequency (P = .897). In patients with poorly-controlled T2DM (HbA1c N8%; n = 250), there was a marked benefit of ranolazine over placebo across 3 of the 4 outcomes, with lower weekly angina (3.8 vs 4.8 episodes per week, P = .003), higher SAQ angina frequency scores (18.7 vs 10.0 point increase, P b .001), and higher SAQ treatment satisfaction (7.1 vs 3.9 point increase, P = .044), and a non-significant trend toward lower weekly SL NTG used (2.0 vs 2.5, P = .114). Examining the cumulative response curves for SAQ angina frequency revealed no significant differences among patients with HbA1c ≤8% (Figure 3). However, among patients with poor glycemic control, 73% of patients treated with ranolazine reported a clinical improvement in angina frequency (i.e., at least a 10-point improvement) versus 46% of placebo-treated patients (P b .001; absolute difference = 27%, number needed to treat with ranolazine to have 1 patient report a clinically-relevant improvement in angina of ~4). Furthermore, 56% of ranolazine-treated patients reported at least a moderate clinical improvement in angina frequency (ie, 20-point improvement) versus 37% of placebo-treated patients (P = .002; absolute difference = 19%; number needed to treat of ~5).
Discussion In the first prospective clinical trial of antianginal therapy specifically in patients with T2DM and stable angina, we found the benefits of ranolazine in reducing angina burden varied by patients’ glycemic control. The effectiveness of ranolazine improved linearly with worsening glycemic control, with its therapeutic benefit across the broad range of patient-reported outcomes being more pronounced at HbA1c N6.5%. Patients with HbA1c N8% experienced marked improvements in angina frequency, SL NTG use, and in health status on ranolazine vs. placebo, which were both highly statistically significant and clinically meaningful. While these results need to be confirmed in future studies, they suggest that ranolazine may be more beneficial in patients with angina and suboptimally-controlled T2DM. Our observation that ranolazine is associated with greater angina relief and better health status among patients with suboptimal glycemic control is particularly relevant, as ranolazine may also produce improvements in fasting glucose and HbA1c. Post hoc analyses from the CARISA and MERLIN-TIMI 36 trials demonstrated statisti-
Arnold et al 463
cally significant and clinically-relevant reductions in HbA1c levels in patients treated with ranolazine vs. placebo, especially notable as both trials allowed openlabel use and titration of glucose-lowering medications. 17,18 While the mechanism of action for the glycometabolic effect of ranolazine remains to be fully elucidated, recent preclinical studies suggest that it may be mediated by inhibition of sodium channel activity in pancreatic alpha cells and a resultant reduction in glucagon secretion. 19–23 Several prospective randomized clinical trials are presently ongoing to more clearly establish the glucose-lowering effect of ranolazine (clinicaltrials.gov: NCT01494987, NCT01472185, NCT01555164). The mechanism(s) underlying the greater antianginal benefit of ranolazine among patients with higher HbA1c levels remain to be definitively determined, but several possible explanations exist. Preclinical studies have demonstrated that cardiomyocytes from diabetic rats exposed to high glucose concentrations have increased phosphorylation of CaMKII, 24 a kinase that increases INaL. 25 Furthermore, recent data have directly shown that cardiac myocytes of mice with type 1 or type 2 diabetes exhibit increased persistent sodium current (INaP). 10 As the antianginal effect of ranolazine is mediated by inhibition of INaL, 11 its therapeutic effectiveness is expected to be enhanced in states of hyperglycemia via this mechanism. Alternatively, hyperglycemia may worsen myocardial perfusion through a variety of mechanisms, including impaired coronary vasomotor function, greater platelet aggregation, hypercoagulability, inflammation, and oxidative stress. 26 The greater antianginal benefit of ranolazine in patients with worse glucose control may, thus, also be mediated by the purported glucose-lowering action of ranolazine, which could potentially ameliorate these detrimental effects of hyperglycemia. 27 However, because improvement in glucose control has not been shown to improve angina burden, particularly in the short time frame of the TERISA trial, this latter explanation appears to be less plausible. Finally, it is possible that other factors associated with high HbA1c levels are also associated with greater effectiveness of ranolazine. We explicitly tested the 3-way interaction between treatment, HbA1c, and insulin use and found this not significant for 3 of the outcomes and marginally significant for SL NTG. However, the direction of effect, with less ranolazine effect by HbA1c among patients on insulin, indicates the mechanism of increased antianginal is unlikely to be mediated by insulin use. As such, we believe that the increased effectiveness of ranolazine in patients with poorly-controlled T2DM likely relates directly to the increased INaP in the setting of hyperglycemia. While we observed only small differences in angina among patients treated with ranolazine vs. placebo whose glucose was well-controlled, these findings should be interpreted with caution. First, the TERISA study was designed to evaluate ranolazine in the
American Heart Journal October 2014
464 Arnold et al
population of patients with T2DM and angina and was not powered to detect differences in angina across the subgroups of patients with various HbA1c levels. Second, the confidence intervals around the estimates of the effectiveness of ranolazine were wide at lower levels of HbA1c. Third, we noted a greater placebo effect among patients with well-controlled T2DM, possibly reflecting a difference in the type of patients in this subgroup, making the detection of an effect of ranolazine above this larger placebo effect more difficult. In addition, several prior studies, including CARISA, ERICA, and MERLIN-TIMI 36, have established ranolazine as an effective antianginal agent in a broad population of patients both with and without T2DM. 3–5 Therefore, the clinical implications of this finding are uncertain but provocative in suggesting a heterogeneity of treatment benefit related to patients' glycemic control. Further studies are needed to help determine the patients most (and least) likely to benefit from ranolazine in terms of angina and quality of life. Our results should be interpreted in the context of several potential limitations. First, the analyses based on HbA1c were not pre-specified in the study protocol. Although our study was hypothesis-driven and supported by mechanistic data, given the post-hoc nature of the analysis, these results will need to be confirmed in future studies. In addition, although our observations persisted after adjustment for several key covariates and were consistent across a broad range of patient-reported outcomes, residual confounding cannot be definitively excluded. Second, the relatively short duration of followup in TERISA limits our ability to determine whether the differential antianginal effect of ranolazine in patients with various baseline HbA1c levels persists over time. Third, HbA1c levels were only assessed at baseline in TERISA. While post-hoc analyses have demonstrated ranolazine may also improve fasting glucose and HbA1c, thereby providing a dual benefit in patients with poorlycontrolled T2DM (i.e., improved glycemic control and more antianginal benefit), the glucose-lowering effect of ranolazine was not tested in TERISA and will need to be defined in on-going prospective randomized clinical trials. Finally, the majority of patients in TERISA were enrolled in European centers, and, thus, whether the observed interaction between the antianginal effect of ranolazine and baseline HbA1c applies to blacks, Hispanics and Asians—populations with growing T2DM prevalence 28—remains unclear. In summary, our study demonstrates that the antianginal effect of ranolazine is more pronounced in patients with higher vs. lower baseline HbA1c, with greater improvements observed in angina burden and health status measures. If confirmed in future studies, these findings suggest that ranolazine may be a good therapeutic option for patients with angina and suboptimallycontrolled T2DM.
References 1. Chaitman BR, Skettino SL, Parker JO, et al. Anti-ischemic effects and long-term survival during ranolazine monotherapy in patients with chronic severe angina. J Am Coll Cardiol 2004;43(8):1375-82. 2. Rousseau MF, Pouleur H, Cocco G, et al. Comparative efficacy of ranolazine versus atenolol for chronic angina pectoris. Am J Cardiol 2005;95(3):311-6. 3. Chaitman BR, Pepine CJ, Parker JO, et al. Effects of ranolazine with atenolol, amlodipine, or diltiazem on exercise tolerance and angina frequency in patients with severe chronic angina: a randomized controlled trial. JAMA 2004;291(3):309-16. 4. Stone PH, Gratsiansky NA, Blokhin A, et al. Antianginal efficacy of ranolazine when added to treatment with amlodipine: the ERICA (Efficacy of Ranolazine in Chronic Angina) trial. J Am Coll Cardiol 2006;48(3):566-75. 5. Arnold SV, Morrow DA, Wang K, et al. Effects of ranolazine on disease-specific health status and quality of life among patients with acute coronary syndromes: results from the MERLIN-TIMI 36 randomized trial. Circ Cardiovasc Qual Outcomes 2008;1(2):107-15. 6. Kosiborod M, Arnold SV, Spertus JA, et al. Evaluation of ranolazine in patients with type 2 diabetes mellitus and chronic stable angina: results from the TERISA randomized clinical trial (Type 2 Diabetes Evaluation of Ranolazine in Subjects With Chronic Stable Angina). J Am Coll Cardiol 2013;61(20):2038-45. 7. Arnold SV, Kosiborod M, McGuire DK, et al. Effects of Ranolazine on Quality of Life Among Patients With Diabetes Mellitus and Stable Angina. JAMA Intern Med 2014 Jun 2. http://dx.doi.org/10.1001/ jamainternmed.2014.2120. [Epub ahead of print]. 8. Chaitman BR. Ranolazine for the treatment of chronic angina and potential use in other cardiovascular conditions. Circulation 2006;113(20):2462-72. 9. Belardinelli L, Shryock JC, Fraser H. Inhibition of the late sodium current as a potential cardioprotective principle: effects of the late sodium current inhibitor ranolazine. Heart 2006;92(Suppl 4):iv6-iv14. 10. Lu Z, Jiang YP, Wu CY, et al. Increased persistent sodium current Due to decreased PI3K signaling contributes to QT prolongation in the diabetic heart. Diabetes 2013;62(12):4257-65. 11. Antzelevitch C, Belardinelli L, Zygmunt AC, et al. Electrophysiological effects of ranolazine, a novel antianginal agent with antiarrhythmic properties. Circulation 2004;110(8):904-10. 12. Spertus JA, Winder JA, Dewhurst TA, et al. Monitoring the quality of life in patients with coronary artery disease. Am J Cardiol 1994;74 (12):1240-4. 13. Spertus JA, Winder JA, Dewhurst TA, et al. Development and evaluation of the Seattle Angina Questionnaire: a new functional status measure for coronary artery disease. J Am Coll Cardiol 1995;25(2):333-41. 14. Rose GA, Blackburn H. Cardiovascular survey methods. Monogr Ser World Health Organ 1968;56:1-188. 15. Ware Jr JE, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 1992;30(6):473-83. 16. Inzucchi SE, Bergenstal RM, Buse JB, et al. Management of hyperglycemia in type 2 diabetes: a patient-centered approach: position statement of the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetes Care 2012;35(6):1364-79. 17. Morrow DA, Scirica BM, Chaitman BR, et al. Evaluation of the glycometabolic effects of ranolazine in patients with and without
American Heart Journal Volume 168, Number 4
18.
19.
20.
21.
22.
23.
diabetes mellitus in the MERLIN-TIMI 36 randomized controlled trial. Circulation 2009;119(15):2032-9. Timmis AD, Chaitman BR, Crager M. Effects of ranolazine on exercise tolerance and HbA1c in patients with chronic angina and diabetes. Eur Heart J 2006;27(1):42-8. Yang M, Chu R, Belardinelli L, et al. Inhibition of glucagon secretion in rat pancreatic islets via Na + channels: mechanism of anti-diabetic effect of ranolazine. Diabetes 2013;62(Suppl 1):A234. Kahlig K, Yang M, Cheng X, et al. The mechanism of the anti-diabetic effect of ranolazine: inhibition of glucagon secretion in rat pancreatic islets via Na + channel blockade. Diabetes 2013;62(Suppl 1):A272. Dhalla A, Yang M, Chu R, et al. Ranolazine inhibits glucogon secretion from human pancreatic islets via blockade of the Nav1.3 channels in alpha-cells. Diabetes 2013;62(Suppl 1):A227. Ning Y, Van Petten M, Jiang J, et al. Na channel blockers exert anti-diabetic effects via lowering glucagon levels in ZDF diabetic rats. Diabetes 2013;62(Suppl 1):A280. Ning Y, Jiang J, Van Petten M, et al. Ranolazine suppresses exaggerated postprandial glucagon response in STZ-induced diabetic rats. Diabetes 2013;62(Suppl 1):A235.
Arnold et al 465
24. Nishio S, Teshima Y, Takahashi N, et al. Activation of CaMKII as a key regulator of reactive oxygen species production in diabetic rat heart. J Mol Cell Cardiol 2012;52(5):1103-11. 25. Koval OM, Snyder JS, Wolf RM, et al. Ca2+/calmodulin-dependent protein kinase II-based regulation of voltage-gated Na + channel in cardiac disease. Circulation 2012;126(17):2084-94. 26. Paneni F, Beckman JA, Creager MA, et al. Diabetes and vascular disease: pathophysiology, clinical consequences, and medical therapy: part I. Eur Heart J 2013;34(31):2436-43. 27. Deedwania P, Kosiborod M, Barrett E, et al. Hyperglycemia and acute coronary syndrome: a scientific statement from the American Heart Association Diabetes Committee of the Council on Nutrition, Physical Activity, and Metabolism. Circulation 2008;117(12): 1610-9. 28. Department of Health and Human Services. Centers for disease control and prevention. Age-adjusted percentage of civilian, noninstitutionalized population with diagnosed diabetes, by race and sex, United States, 1980–2011. http://www.cdc.gov/diabetes/ statistics/prev/national/figraceethsex.htm. [Accessed October 22, 2013. In].
American Heart Journal Volume 168, Number 4
Arnold et al 465.e1
Appendix Supplementary Table. Baseline characteristics of the patients, by treatment group
Age (y) Male White race Hypertension Dyslipidemia Current smoking Prior myocardial infarction Prior angioplasty Prior bypass graft surgery Duration of diabetes (y) Mean ± SD Median (IQR) Absolute range HbA1c (%) Mean ± SD Median (IQR) Absolute range Taking diabetic medication Taking insulin Antianginal medications β-Blockers Calcium-channel blockers Long acting nitrates Statins Antiplatelet agents ACE-I/ARBs Baseline weekly angina by diary Russia/Ukraine/Belarus
Ranolazine n = 455
Placebo n = 458
63.1 ± 8.5⁎ 61.1% 98.9% 94.9% 79.0% 15.4% 75.2% 42.7% 18.2%
64.2 ± 8.4⁎ 61.8% 99.3% 95.8% 80.3% 16.6% 73.0% 39.2% 18.8%
7.2 ± 6.7 4.9 (2.2-10.0) 0.0-41.0
7.7 ± 7.0 6.0 (2.6-10.8) 0.0-42.6
7.3 ± 1.5 6.9 (6.2-8.3) 4.7-12.5 93.2% 17.4%
7.3 ± 1.5 6.9 (6.1-8.1) 5.0-12.9 92.6% 20.5%
90.5% 26.6% 34.9% 82.4% 89.9% 87.9% 6.8 ± 4.3 71.6%
89.7% 30.1% 32.3% 82.8% 86.2% 87.6% 6.9 ± 4.3 70.1%
⁎ P N .05 for all comparisons except age (P = .038)
465.e2 Arnold et al
American Heart Journal October 2014
Supplementary Figure 1
Effectiveness of Ranolazine Versus Placebo by Decile of Baseline HbA1c. A. IDR for average weekly frequency of angina by diary. B. IDR for average weekly number of SL NTG used by diary. C. Change in SAQ angina frequency. D. Change in SAQ treatment satisfaction.
Supplementary Figure 2
Effectiveness of Ranolazine vs. Placebo on SL NTG Use by Baseline HbA1c, Stratified by Insulin Use. IDR for average weekly number of SL NTG used by diary (IDR b1 indicates ranolazine is more effective vs. IDR N1 indicates placebo is more effective). Red line is patients on insulin. Black line is patients not on insulin.
