Despite multiple medical and interventional technologies to reduce myocardial ischemia, chronic angina still affects nearly 8 million people in the United States (1). Although it is associated with worse health-related quality of life ((2),3), repeat hospitalizations, and increased healthcare costs (4), angina remains frequently undertreated ((5),6). Patients with angina and concomitant type 2 diabetes mellitus (T2DM) represent a particularly challenging group, as they often have more diffuse and extensive coronary artery disease (CAD) as compared with those without T2DM ((7),8). Furthermore, patients with CAD and T2DM may also have a greater burden of angina than those without diabetes ((9),10). Targeted approaches to reduce the burden of angina specifically among patients with T2DM could have a substantial impact on quality of life.

Ranolazine, a selective inhibitor of the late sodium current (late I_Na) (11), has been proven effective in treating chronic angina both as a monotherapy ((12),13) and in combination with other commonly prescribed antianginal medications ((14),15). Furthermore, among patients with poorly controlled T2DM, ranolazine may lower fasting glucose and HbA1c ((10),(16),17). Although post hoc analyses of prior trials have suggested an antianginal benefit among patients with T2DM ((10),17), this hypothesis has not been prospectively tested. Accordingly, we sought to test the efficacy of ranolazine in reducing angina among patients with T2DM, CAD, and chronic angina who remain symptomatic despite treatment with other agents.

TERISA (Type 2 Diabetes Evaluation of Ranolazine in Subjects With Chronic Stable Angina) was a randomized, double-blind, placebo-controlled trial in which subjects with stable angina and T2DM were randomized to twice daily placebo or ranolazine for 8 weeks (Figure 11_gr1). The study was conducted in 14 countries across Asia, Europe, and North America, and was approved by the national regulatory authority in each participating country and by the institutional review board or local ethics committee for each site. All participating subjects gave written informed consent. The full study protocol and the list of participating sites and investigators can be found in the Online Appendix ((6), 6). The primary aim of the trial was to examine the efficacy of ranolazine versus placebo on weekly angina frequency in subjects with T2DM, CAD, and chronic stable angina who remain symptomatic despite treatment with 1 or 2 antianginal agents.

Full inclusion and exclusion criteria may be found in the Online Appendix (6). Briefly, to be eligible for randomization in TERISA, subjects had to have a documented history of both T2DM and CAD, and at least a 3-month history of chronic stable angina. Subjects were further required to be treated with 1 or 2 antianginal therapies (beta-blockers, calcium-channel blockers, long-acting nitrates) at a stable dose for at least 2 weeks prior to study entry. Key exclusion criteria were New York Heart Association functional class III to IV heart failure symptoms, acute coronary syndrome in the prior 2 months, planned coronary revascularization during the study period, stroke or transient ischemic attack within 6 months prior to screening, uncontrolled hypertension, clinically significant hepatic impairment, prior treatment with ranolazine, and dialysis.

Eligible subjects entered a 4-week, single-blind placebo run-in period and were provided a novel, handheld electronic diary (LogPad LV diary, PHT Corporation, Boston, Massachusetts) (6), with built-in electronic prompts for daily entry. Subjects were instructed to record and transmit the data to the coordinating center every evening, including the number of angina episodes and number of sublingual nitroglycerin (SL NTG) doses taken since the previous evening. Subjects taking >2 antianginal medications at screening were allowed to washout additional antianginal therapies 2 weeks prior to the run-in period. To be randomized at the end of the run-in period, subjects were required to meet the following criteria: 1) ≥85% adherence to daily electronic diary data entry (including angina frequency and SL NTG use) with no week with <5 days of diary use; 2) an average weekly angina frequency between 1 and 28 and at least 1 angina episode during each week; and 3) ≥80% adherence with single-blind placebo. Subjects were then randomized in a double-blind fashion to either ranolazine or placebo for 8 weeks (6). Ranolazine (Gilead Sciences, Foster City, California) or matching placebo was initiated at 500 mg twice daily (bid) for 1 week and, if tolerated, increased to 1,000 mg bid (subjects taking verapamil or diltiazem were maintained on 500 mg bid of ranolazine or matching placebo). Randomization was stratified by the following: 1) average number of weekly angina episodes during trial run-in (≥1 and <3 vs. ≥3 and ≤28); 2) number of concomitant antianginal agents (1 vs. 2); 3) geographic region (Russia, Ukraine, and Belarus vs. Other). The proportion of subjects being treated with 2 concomitant antianginal medications was capped at 50%, and the proportion of subjects with baseline average weekly angina frequency <3 was capped at 30%. Number and doses of concomitant antianginal medications were kept stable during the study per protocol.

The primary outcome was the average number of angina episodes per week from weeks 2 to 8 of treatment. Pre-specified secondary endpoints included average weekly frequency of SL NTG use, number of angina-free days, proportion of subjects with ≥50% reduction in average weekly angina frequency, and health-related quality of life, as assessed by the Medical Outcomes Short Form-36 (SF-36) (18) and the Patient's Global Impression of Change scale score (19).

Full details of the statistical analysis plan may be found in the Online Appendix (6). Assuming an average weekly angina frequency of 2.0 episodes on placebo, we estimated that a sample size of 900 subjects, randomized in a 1:1 ratio to ranolazine and placebo, would provide 90% power to show a relative reduction of 20% in weekly angina frequency with ranolazine as compared with placebo.

The pre-specified efficacy analyses included all randomized subjects who took at least 14 days of the study drug, completed at least 1 diary entry, and met all major eligibility criteria. The primary efficacy analysis comparing weekly angina frequency was performed by fitting a generalized linear model with negative binomial distribution, as was the analysis of average weekly SL NTG use. Pre-specified subset analyses included the following: 1) average number of weekly angina episodes at baseline (<3 vs. ≥3); 2) number of concomitant antianginal agents (1 vs. 2); 3) geographic region (Russia, Ukraine, and Belarus vs. Other); 4) age (<65 vs. ≥65 years); and 5) sex (male vs. female). In addition, a number of exploratory stratified analyses were performed, including history of percutaneous coronary intervention (yes vs. no); history of coronary artery bypass graft surgery (yes vs. no); and HbA1c at various thresholds.

All analyses were conducted using SAS version 9.2 (SAS Institute, Inc., Cary, North Carolina), and statistical significance was determined by a 2-sided p value of <0.05. A pre-specified testing sequence incorporating Hochberg adjustments was used to preserve the type I error rate.

The study was sponsored by Gilead Sciences (Foster City, California). Statistical analysis was performed by Saint Luke's Mid America Heart Institute independent of the sponsor. The decision to submit the manuscript for publication was made by the publication committee (6).

From October 5, 2011, to July 20, 2012, 1,185 patients were screened (Figure 11_gr1). Of these, 1,142 patients were considered trial-eligible, signed informed consent, and entered into a 4-week single-blind placebo run-in period. A total of 193 patients failed the run-in period and were excluded prior to randomization; 949 patients were randomized (473 to ranolazine, 476 to placebo). Twenty-two patients that either never initiated or discontinued the study drug during the first 2 weeks (11 in the ranolazine arm, 11 in the placebo arm) were excluded from the analysis, leaving the final analytic sample size of 927 patients (462 in the ranolazine arm, 465 in the placebo arm). The target dose of ranolazine or matching placebo (1,000 mg bid or 500 mg bid in the presence of diltiazem or verapamil) was achieved in >95% of patients.

Baseline characteristics of the randomized patients are listed in (Table 1), and were well matched between groups. The mean age was 64 ± 8.5 years, 61% were male, and 99% were white. Seventy percent of subjects were enrolled in Russia, Ukraine, and Belarus, and 30% were enrolled in other countries (enrollment by country, 6). CAD risk factors were common, including hypertension (96%), dyslipidemia (80%), and smoking (16%); and the majority of patients had a previous myocardial infarction (74%) and prior revascularization (51%). Average diabetes duration was 7.5 ± 6.8 years, mean HbA1c was 7.3 ± 1.5%, and 93% of patients were treated with glucose-lowering medications, including 19% treated with insulin.

A total of 56% of patients were treated with 1 antianginal agent, with the remainder receiving 2 agents at baseline. Most patients were treated with beta-blockers (90%), with lower rates of calcium channel blocker (29%) and long-acting nitrate use (34%). The use of concomitant cardiovascular medications was high, including statins (82%), antiplatelet agents (88%), and angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (88%).

Compliance with electronic diary submission of patient-reported angina frequency and SL NTG was high (98% of all patient-days during weeks 2 to 8 of treatment had a diary entry, both in ranolazine and placebo arms). During the 4-week run-in phase, average weekly angina frequency was similar between the ranolazine and placebo groups (6.6 [95% confidence interval (CI): 6.27 to 7.02] episodes vs. 6.8 [95% CI: 6.42 to 7.19] episodes, p = 0.54). Though patients treated with placebo had a substantial decrease in angina frequency, weekly angina frequency was significantly lower in the ranolazine group than in the placebo group during weeks 2 to 8 after randomization (3.8 [95% CI: 3.57 to 4.05] episodes vs. 4.3 [95% CI: 4.01 to 4.52] episodes, p = 0.008) ((Figure 11_gr2)A, Table 2).

At baseline, there was no statistical difference in average weekly SL NTG use between the ranolazine and placebo groups (4.1 [3.74 to 4.60] vs. 4.5 [4.05 to 4.98] doses, p = 0.27). During weeks 2 to 8 after randomization, the average weekly number of SL NTG doses was significantly lower in the ranolazine group (Figure 11_gr2B), and was significantly lower in the ranolazine group than in the placebo group (1.7 [95% CI: 1.58 to 1.92] vs. 2.1 [95% CI: 1.92 to 2.31] doses, p = 0.003) ((Figure 11_gr2)B, Table 2).

During weeks 2 to 8 after randomization, the proportion of angina-free days did not differ between the ranolazine and placebo groups (67% vs. 64%, p = 0.068) (Table 2). The proportion of patients achieving at least 50% reduction in weekly angina frequency was higher in the ranolazine than placebo group (47% vs. 42%, p = 0.034), and the increase from baseline to end of treatment in SF-36 Physical Component Summary Score was also greater in the ranolazine than placebo group (2.9 [95% CI: 2.3 to 3.5] points vs. 1.9 [95% CI: 1.3 to 2.5] points, p = 0.005). However, these latter 2 differences were not considered statistically significant (despite p values <0.05) based on the pre-specified multiple testing procedure. There was no difference between the ranolazine and placebo groups in the change in SF-36 Mental Component Score (1.0 [95% CI: 0.18 to 1.82] points vs. 1.1 [95% CI: 0.28 to 1.92] points, p = 0.77) or in the Patient's Global Impression of Change scale score at end of treatment (4.0 [95% CI: 3.82 to 4.19] vs. 3.9 [95% CI: 3.74 to 4.10], p = 0.41).

The results of subgroup analyses are presented in (Figure 11_gr3)A and (Figure 11_gr3)B. The superior efficacy of ranolazine versus placebo on the primary endpoint was consistent in the pre-specified subgroups of baseline average weekly angina episodes (<3 vs. ≥3; p_interaction = 0.85), number of concomitant antianginal medications (1 vs. 2; p_interaction = 0.89), age (<65 vs. ≥65; p_interaction = 0.97), and sex (p_interaction = 0.46). There was a significant interaction in the effect of ranolazine versus placebo on the primary endpoint by the geographic region of enrollment (Russia, Ukraine and Belarus vs. Other; p_interaction = 0.016). Baseline characteristics by enrollment region stratified by randomized treatment assignment are presented in (6). The average number of weekly angina episodes during the treatment phase decreased significantly in both treatment groups (6), but was not statistically different between the ranolazine and placebo arms among patients enrolled in Russia, Ukraine, and Belarus (4.1 [95% CI: 3.9 to 4.4] vs. 4.3 [95% CI: 4.1 to 4.6], p = 0.31). However, among patients enrolled in other countries, those treated with ranolazine experienced a significant reduction in average weekly angina episodes as compared with placebo-treated patients (3.1 [95% CI: 2.8 to 3.5] vs. 4.1 [95% CI: 3.7 to 4.6], p = 0.002) (6).

In exploratory analyses, the greater therapeutic benefit of ranolazine versus placebo was consistent across revascularization subgroups (Figure 11_gr3A). However, the therapeutic superiority of ranolazine versus placebo on reducing weekly angina frequency was more pronounced in patients with higher baseline HbA1c, regardless of the HbA1c threshold used (p_interaction = 0.027) (Figure 11_gr3B).

There was no difference in the incidence of serious adverse events between groups (Table 2). Of the 20 subjects discontinued from the study due to adverse events, 9 were in the ranolazine group and 11 were in the placebo group. Five subjects died during the study, 3 in the ranolazine group (2 myocardial infarctions and 1 sudden cardiac death) and 2 in the placebo group (1 acute cardiac failure and 1 pulmonary embolism). Nonserious adverse events of nausea, dizziness, and constipation occurred more frequently with ranolazine compared with placebo.

In this trial of patients with T2DM, established CAD, and stable angina, ranolazine was more effective than placebo in reducing the primary outcome of average weekly angina episodes, as well as average weekly sublingual nitroglycerin use. These results were consistent across the subgroups of baseline average weekly angina episodes, number of concomitant antianginal medications, age, and sex. The therapeutic benefit of ranolazine versus placebo was greater among patients enrolled outside of Russia, Ukraine, and Belarus, and among those with higher baseline HbA1c. In addition, ranolazine was safe and well tolerated in this patient population.

While patients with T2DM and CAD have more extensive disease ((7),8) and worse outcomes ((20),21) than those without DM, the data on whether they experience more angina are conflicting. Several older studies suggested that patients with DM have less angina than their non-DM counterparts due to an increased likelihood of “silent” ischemia related to diabetic autonomic neuropathy ((22),(23),24). However, both a recent observational study and a multinational clinical trial demonstrated that patients with DM experienced a higher burden of angina compared with patients without DM after an acute coronary syndrome ((9),10). Given the high prevalence of angina in patients with DM, as well as its relationship with worse health-related quality of life ((2),3), greater risk of repeat hospitalizations, and increased healthcare costs (4), effective therapeutic strategies for angina management in this patient group are clinically important.

To our knowledge, the TERISA trial is the first randomized clinical trial to study antianginal medical therapy (ranolazine or otherwise) specifically in patients with T2DM—a high-risk and therapeutically challenging group. A post hoc analysis of the CARISA (Combination Assessment of Ranolazine in Stable Angina) randomized trial that evaluated ranolazine versus placebo among patients with stable angina showed that ranolazine decreased angina frequency to a similar degree in patients with and without T2DM (17). Moreover, a subgroup analysis of the MERLIN−TIMI 36 (Metabolic Efficiency With Ranolazine for Less Ischemia in Non-ST Elevation Acute Coronary Syndromes−Thrombolysis In Myocardial Infarction 36) study of subjects following acute coronary syndromes suggested a trend toward greater therapeutic benefit of ranolazine versus placebo in reducing recurrent ischemia among patients with T2DM than those without T2DM (10). Although these observational analyses suggested an antianginal benefit of ranolazine in patients with T2DM, this hypothesis has not been prospectively tested until now. By specifically focusing on patients with T2DM in a prospective, randomized clinical trial, the TERISA trial establishes, for the first time, the therapeutic effectiveness of ranolazine in this patient population.

The subgroup analyses of the TERISA trial produced intriguing findings, particularly the greater efficacy of ranolazine versus placebo among patients with higher HbA1c. The mechanisms underlying this finding remain to be determined. There is, however, evidence that cardiomyocytes exposed ex vivo to high glucose, or isolated from hearts of diabetic animals have up-regulated phosphorylated CaMKII (25), a kinase known to phosphorylate the cardiac sodium channel resulting in an increased late I_Na (26), thereby leading to increases in intracellular sodium and calcium (27). Ranolazine, an inhibitor of late I_Na (11), would then be expected to have a greater therapeutic effect in patients with poor glycemic control. Alternative explanations for this observation may also exist. Beyond its antianginal benefit, ranolazine may also lower HbA1c, particularly among patients with suboptimal glucose control ((16),17). Hyperglycemia may worsen myocardial perfusion through microvascular and endothelial dysfunction, enhanced platelet aggregation and prothrombotic state, and increased oxidative stress—detrimental effects that may be reduced with better glucose control (28). Ranolazine's antianginal effect mediated through inhibition of late I_Na may, therefore, be potentiated by its glucose-lowering properties through some of these mechanisms. However, because glucose-lowering per se has never been shown to improve angina, this mechanism is a less plausible one. Regardless of the underlying physiology, these findings, if confirmed in future clinical trials, would suggest that T2DM patients with angina and suboptimal glucose control may receive greater benefit from ranolazine therapy—both in terms of its antianginal and glucose-lowering effects.

We also observed a difference in the therapeutic effectiveness of ranolazine among patients enrolled in Russia, Ukraine, and Belarus versus other countries. Preliminary analyses have determined that this observation was driven primarily by several sites located in Russia and was not explained by differences in baseline characteristics between geographic regions. The underlying reasons behind this finding remain uncertain and will be investigated in future detailed analyses.

The TERISA study has several strengths. The design of the study, including the single-blind placebo run-in period, ensured high angina frequency at baseline and compliance with the study medication and self-reporting of outcomes among the study patients. It also allowed appropriate accounting for the marked placebo effect, which is of critical importance in clinical trials of antianginal agents. Prior antianginal trials often compared on-treatment angina frequency to a baseline value without a parallel placebo group ((29),(30),31). The primary endpoint of the TERISA trial (patient-reported angina frequency) was both patient-centered and clinically meaningful. Other studies, even when showing a reduction in intermediate measures of angina (e.g., time to limiting angina on exercise testing), failed to demonstrate a reduction in angina frequency ((32),33). Additionally, we used a novel electronic diary instrument (6) to record patient-reported angina and SL NTG use, with outstanding compliance. Studies using traditional paper diaries are often limited by inappropriate input of multiple entries in a single sitting (so-called “hoarding”) and low subject compliance (as low as 11% in 1 study) (34). Our successful use of the electronic diary should inform the design of future clinical studies that evaluate patient-reported outcomes. Finally, the use of concomitant medical therapy for CAD, such as antiplatelet agents, antihypertensive medications, and statins, was high in both treatment groups.

The results of this study should be interpreted in the context of several potential limitations. Most of the subjects were enrolled in European centers, thereby limiting inclusion of a racially diverse study population. More data are needed on the effect of ranolazine among blacks, Hispanics, and Asians. In addition, several of the subgroup analyses were not pre-specified and should be regarded as exploratory and hypothesis-generating. Specifically, while the findings of greater ranolazine efficacy in patients with higher HbA1c were adjusted for several key factors (including geographic region), a possibility of residual confounding cannot be definitively ruled out. Finally, the relatively short duration of treatment in the TERISA trial limits our ability to know the durability of the observed antianginal effect of ranolazine; however, prior studies have demonstrated this effect to persist for at least 12 months (35).

Among patients with T2DM, CAD, and persistent chronic angina, despite treatment with up to 2 antianginal agents, ranolazine significantly reduced angina frequency and SL NTG use and was safe and well tolerated.