Home > Clinical Guides > Stable Angina: Methods, Evidence &... > Other anti anginal drugs and general...

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

National Clinical Guidelines Centre (UK). Stable Angina: Methods, Evidence & Guidance [Internet]. London: Royal College of Physicians (UK); 2011 Jul. (NICE Clinical Guidelines, No. 126.)

10Other anti anginal drugs and general drug recommendations

10.1. Introduction

Ivabradine, nicorandil, and ranolazine are anti-anginal drugs that are licensed for use in the treatment of stable angina. The GDG were interested in evidence for the use of these drugs either as monotherapy or in combination with other anti-anginal drugs, and their place in the pathway for people with stable angina.

Ivabradine is licensed for the treatment of angina in patients in sinus rhythm in combination with a BB, or when a BB is contra-indicated or not tolerated. Nicorandil has been available for longer than the other drugs considered in this chapter and although it does not have a licence to be used in combination with other antianginal drugs it is regularly used this way in practice. Ranolazine is licensed as adjunctive therapy in patients who are inadequately controlled or intolerant of first-line antianginal drugs.

The costs of drugs at standard doses are listed below and compared with the cost of long-acting nitrates.

Table 10Cost of drugs

Specific drugs used for rangeCost per day (£)Cost per year (£)
Long-acting nitratesLow = isosorbide mononitrate0.0519
High = isosorbide dinitrate0.0934
Ivabradine, 5 mg or 7.5 mg twice dailyLow = high1.39507
Ranolazine, 375 mg, 500 mg or 750 mg twice dailyLow = high1.63595
NicorandilLow = 10 mg tablets twice daily0.2799
High = 20 mg tablets twice daily0.52190

10.2. Ivabradine

Ivabradine is a heart rate-lowering agent that acts by selectively inhibiting the If current, an ionic current across the sarcolemma in cells of the sino-atrial node that is involved in pacemaker activity. Ivabradine reduces the slope of spontaneous diastolic depolarization in sino-atrial cells, and lowers heart rate at rest and during exercise. Side effects of ivabradine include bradycardia, heart block, and visual disturbances (phosphenes and blurred vision).

10.2.1. Clinical question

What is the clinical/cost effectiveness of ivabradine for the management of stable angina?

10.2.2. Clinical evidence

The “Review Protocol” for this topic can be found in Appendix C, the “Search Strategies” in Appendix D, the “List of Included and Excluded Studies” in Appendix E1, the “Clinical Evidence Tables” in Appendix E2, and the “Forest Plots” in Appendix F.

The evidence review included evidence for the use of ivabradine as monotherapy or in combination with BB to control symptoms and improve outcome in people with stable angina.

Table 10.1Ivabradine vs. Placebo

Quality assessmentSummary of findings
No of patientsEffectQuality
No of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerationsIvabradineplaceboRelative (95% CI)Absolute
Time to angina onset (sec) (trough change from baseline) (follow-up 14 days; better indicated by higher values) (g)
Borer 200345(e)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (b)None5968-MD 14.1 higher (11.73 lower to 39.93 higher)⊕⊕⊕○

MODERATE
Time to angina onset (sec) (peak change from baseline) (follow-up 14 days; better indicated by higher values) (h)
Borer45 2003randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (b)None5968-MD 43.2 higher (16.75 to 69.65 higher)⊕⊕⊕○

MODERATE
Time to 1mm ST depression (sec) (at peak of drug activity) (follow-up 14 days; better indicated by higher values)
Borer45 2003randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (b)None5968-MD 52.90 higher (26.85 to 78.95 higher)⊕⊕⊕○

MODERATE
Time to 1mm ST depression (sec) (at trough) (follow-up 14 days; better indicated by higher values)
Borer45 2003randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (b)None5968-MD 35.10 higher (9.68 to 60.52 higher)⊕⊕⊕○

MODERATE
Patients with limiting angina (j) - CV death or hospitalisation for MI or HF - (follow-up median 18 months)
Fox 200946 (BEAUTIFUL) (f)randomised trialsno serious limitation (c)no serious inconsistencyno serious indirectnessserious (d)None88/734 (12%)120/773 (15.5%)RR 0.77 (0.6 to 1)36 fewer per 1000 (from 62 fewer to 0 more)⊕⊕⊕○

MODERATE
Patients with limiting angina - all cause mortality - (follow-up median 18 months)
Fox 200946 (BEAUTIFUL)(i)randomised trialsno serious limitation (c)no serious inconsistencyno serious indirectnessserious (d)None64/734 (8.7%)77/773 (10%)RR 0.88 (0.64 to 1.2)12 fewer per 1000 (from 36 fewer to 20 more)⊕⊕⊕○

MODERATE
Patients with limiting angina - cardiac death - (follow-up median 18 months)
Fox 200946 (BEAUTIFUL)randomised trialsno serious limitation (c)no serious inconsistencyno serious indirectnessserious (d)None11/734 (1.5%)16/773 (2.1%)RR 0.72 (0.34 to1.55)6 fewer per 1000 (from 14 fewer to 11 more)⊕⊕⊕○

MODERATE
Patients with limiting angina - hospitalisation for HF - (follow-up median 18 months)
Fox 200946 (BEAUTIFUL)randomised trialsno serious limitation (c)no serious inconsistencyno serious indirectnessserious (d)None33/734 (4.5%)41/773 (5.3%)RR 0.85 (0.54 to 1.33)8 fewer per 1000 (from 24 fewer to 18 more)⊕⊕⊕○

MODERATE
Patients with limiting angina - hospitalisation for MI or unstable angina - (follow-up median 18 months)
Fox 200946 (BEAUTIFUL)randomised trialsno serious limitation (c)no serious inconsistencyno serious indirectnessserious (d)None56/734 (7.6%)65/773 (8.4%)RR 0.9 (0.64 to 1.28)8 fewer per 1000 (from 30 fewer to 24 more)⊕⊕⊕○

MODERATE
Patients without limiting angina - CV death or hospitalisation for MI or heart failure (follow-up 18 months)
Fox 200946 (BEAUTIFUL)randomised trialsno serious limitation (c)no serious inconsistencyno serious indirectnessno serious imprecisionNone756/4745 (15.9%)712/4665 (15.3%)RR 1.04 (0.95 to 1.15)6 more per 1000 (from 8 fewer to 23 more)⊕⊕⊕⊕

HIGH
Patients without limiting angina - all cause mortality (follow-up median 18 months)
Fox 200946 (BEAUTIFUL)randomised trialsno serious limitation (c)no serious inconsistencyno serious indirectnessno serious imprecisionNone508/4745 (10.7%)470/4665 (10.1%)RR 1.06 (0.94 to 1.2)6 more per 1000 (from 6 fewer to 20 more)⊕⊕⊕⊕

HIGH
Patients without limiting angina - cardiac death (follow-up median 18 months)
Fox 200946 (BEAUTIFUL)randomised trialsno serious limitation (c)no serious inconsistencyno serious indirectnessserious (d)None125/4745 (2.6%)135/4665 (2.9%)RR 0.91 (0.72 to 1.16)3 fewer per 1000 (from 8 fewer to 5 more)⊕⊕⊕○

MODERATE
Patients without limiting angina - hospitalisation for heart failue (follow-up median 18 months)
Fox 200946 (BEAUTIFUL)randomised trialsno serious limitation (c)no serious inconsistencyno serious indirectnessno serious imprecisionNone393/4745 (8.3%)386/4665 (8.3%)RR 1 (0.87 to 1.15)0 fewer per 1000 (from 11 fewer to 12 more)⊕⊕⊕⊕

HIGH
Patients without limiting angina - hospitalisation for MI or unstable angina (follow-up median 18 months)
Fox 200946 (BEAUTIFUL)randomised trialsno serious limitation (c)no serious inconsistencyno serious indirectnessno serious imprecisionNone247/4745 (5.2%)252/4665 (5.4%)RR 0.96 (0.81 to 1.14)2 fewer per 1000 (from 10 fewer to 8 more)⊕⊕⊕⊕

HIGH
All serious adverse events (follow-up median 18 months)
Fox 200946 (BEAUTIFUL)randomised trialsno serious limitation (c)no serious inconsistencyno serious indirectnessno serious imprecisionNone135/734 (18.4%)144/773 (18.6%)RR 0.99 (0.80 to 1.22)2 fewer per 1000 (from 37 fewer to 41 more)⊕⊕⊕⊕

HIGH
a

Borer 2003[44]: Randomisation, allocation concealment, double blinding and ITT reported. Baseline comparisons made, there was no clinically relevant differences in baseline characteristics were observed between the 2 groups. In Ivabradine 2.5 mg bd 3/90 (3.3%); Ivabradine 5 mg bd 4/91 (4.4%); Ivabradine 10 mg bd 3/88 (3.4%) and; Placebo 1/91 (1.1%) were lost to follow-up.

b

The upper and lower CI crosses the MID.

c

Fox 2009[45] (BEAUTIFUL): Randomisation, allocation concealment, double blinding and ITT reported. This is a post hoc analysis, and therefore the results should be considered as hypothesis generating. Sample size calculation reported.

d

95% CI around the pooled estimate of effect includes both: 1) no effect and 2) appreciable benefit or appreciable harm.

e

Ivabradine 5 mg bid vs. placebo

f

Ivabradine 7.5 mg bid vs. placebo

g

Trough = 12 hours after administration of ivabradine

h

Peak = 4 hours after administration of ivabradine

i

In this post hoc analysis, the BEAUTIFUL population was divided according to the presence of limiting angina symptoms at baseline using the New York Heart Association (NYHA) functional classification. Patients were questioned at the inclusion visit regarding the presence of symptoms limiting activity, and whether they were related to anginal pain or due to presence of heart failure (fatigue, palpitations or dyspnoea).

j

Limiting angina symptoms were identified in 13.8% of the BEAUTIFUL population at baseline (1507 out of 10917 patients). Of these, 734 were randomised to ivabradine treatment and 773 to placebo.

Table 10.2Ivabradine vs. atenolol

Quality assessmentSummary of findings
No of patientsEffectQuality
No of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerationsIvabradineAtenololRelative (95% CI)Absolute
Total exercise duration (sec) (trough change from baseline) (follow-up 16 weeks; better indicated by higher values)
Tardif 200547(d)(e)randomised trialsserious (a)no serious inconsistencyno serious indirectnessno serious imprecisionNone300286-MD 8 higher (13.26 lower to 29.26 higher)⊕⊕⊕○

MODERATE
Time to angina onset (sec) (trough change from baseline) (follow-up 16 weeks; better indicated by higher values)
Tardif 200547randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)None300286-MD 10 higher (14.96 lower to 34.96 higher)⊕⊕○○

LOW
Weekly number of angina attacks (follow-up 16 weeks; better indicated by lower values)
Tardif 200547randomised trialsserious (a)no serious inconsistencyno serious indirectnessno serious imprecisionNone307294-MD 0.5 higher (0.99 lower to 1.99 higher)⊕⊕⊕○

MODERATE
Short-acting nitrate consumption units/week (follow-up 16 weeks; better indicated by lower values)
Tardif 200547randomised trialsserious (a)no serious inconsistencyno serious indirectnessno serious imprecisionNone307294-MD 0.4 lower (1 lower to 0.2 higher)⊕⊕⊕○

MODERATE
Withdrawal due to adverse events (follow-up 16 weeks)
Tardif 200547randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (c)None28/315 (8.9%)17/307 (5.5%)RR 1.61 (0.9 to 2.87)34 more per 1000 (from 6 fewer to 104 more)⊕⊕○○

LOW
a

Tardif 2005[46]: Allocation concealment not reported. Randomisation using permutation blocks. Double blinding. Capsules of ivabradine and placebo identical. Patients, investigators, central readers of ETT data were blinded to the treatment received by the patients. Calculation of sample size reported. Baseline comparisons made no difference between the study groups. ITT reported.

b

95% CI includes no effect and the upper and lower CI crosses the MID.

c

95% CI around the pooled estimate of effect includes both: 1) no effect and 2) appreciable benefit or appreciable harm.

d

Ivabradine 5 mg bid for 4 weeks and then 7.5 bid for 12 weeks or atenolol 50 mg od for 4 weeks and then 100 mg od for 12 weeks.

e

The authors state that the study does not allow to strictly comparing the safety of ivabradine and atenolol, because about two-thirds of the patients had previously received BB and were known to tolerate these drugs; patients with known intolerance or contraindications to atenolol were specifically excluded. There was slightly higher number of deaths in the ivabradine groups (2 (0.6%) and 3 (1%) respectively) than in the atenolol group (1 (0.3%) that was not statistically significant.

Table 10.3Ivabradine + atenolol vs. atenolol+placebo

Quality assessmentSummary of findings
No of patientsEffectQuality
No of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerationsIvabradine plus atenololAtenololRelative (95% CI)Absolute
Total exercise duration (sec) (follow-up 2 months; better indicated by higher values) (e)
Tardif 200948(d)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionNone441434-MD 8.7 higher (0.98 to 16.42 higher)⊕⊕⊕⊕

HIGH
Time to angina onset (sec) (follow-up 2 months; better indicated by higher values)
Tardif 200948randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionNone441434-MD 13 higher (3.43 to 22.57 higher)⊕⊕⊕⊕

HIGH
Time to 1mm ST depression (sec) (follow-up 2 months; better indicated by higher values)
Tardif 200948randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (b)None441434-MD 27.2 higher (16.15 to 38.25 higher)⊕⊕⊕○

MODERATE
Total exercise duration (sec)(follow-up 4 months; better indicated by higher values)
Tardif 200948randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionNone441434-MD 16.6 higher (8.05 to 25.15 higher)⊕⊕⊕⊕

HIGH
Time to onset of angina(sec) (follow-up 4 months; better indicated by higher values)
Tardif 200948randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (b)None441434-MD 26.4 higher (15.64 to 37.16 higher)⊕⊕⊕○

MODERATE
Time to 1 mm ST depression (sec) (follow-up 4 months; better indicated by higher values)
Tardif 200948randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (b)None441434-MD 30.3 higher (18.4 to 42.2 higher)⊕⊕⊕○

MODERATE
angina attacks/week (follow-up 4 months; better indicated by lower values)
1 Tardif 200948randomised trialsno serious limitations(a)no serious inconsistencyno serious indirectnessno serious imprecisionnone441434-MD 0.00 higher (0.30 lower to 0.30 higher)⊕⊕⊕○

HIGH
Adverse events (follow-up 4 months)
Tardif 200948randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (c)None13/441 (2.9%)4/434 (0.9%)RR 3.2 (1.05 to 9.73)20 more per 1000 (from 0 more to 80 more)⊕⊕⊕○

MODERATE
a

Tardif 2009[47]: Randomisation, allocation concealment, double blinding and ITT reported. The random allocation was computer generated. Sample size calculation reported. 2% of patients lost to follow-up. The full analysis included 875 (98% of those randomised).

b

The upper and lower CI crosses the MID.

c

95% CI around the pooled estimate of effect includes both: 1) no effect and 2) appreciable benefit or appreciable harm.

d

Patients receiving atenolol 50 mg/day were randomised to receive ivabradine 5 mg b.i.d for 2 months, increased to 7.5 mg b.i.d for a further 2 months.

e

12 hours after last dose ivabradine, 24 hours after last dose atenolol

Table 10.4Ivabradine vs. amlodipine

Quality assessmentSummary of findings
No of patientsEffectQuality
No of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerationsIvabradineAmlodipineRelative (95% CI)Absolute
Total exercise duration (sec) (follow-up 3 months; better indicated by higher values)
Ruzyllo 200749(c)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionNone381398-MD 3.6 lower (16.5 lower to 9.3 higher)⊕⊕⊕⊕

HIGH
Time angina onset (sec) (follow-up 3 months; better indicated by higher values)
Ruzyllo 200749randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionNone381398-MD 1.9 lower (16.24 lower to 12.44 higher)⊕⊕⊕⊕

HIGH
Short-acting nitrate use (units/week) (follow-up 3 months; better indicated by lower values)
Ruzyllo 200749randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionNone389398-MD 0.8 higher (0.04 to 1.56 higher)⊕⊕⊕⊕

HIGH
Frequency of angina attacks/week - (follow-up 3 months; better indicated by lower values)
Ruzyllo 200749randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionNone389398-MD 0 higher (0.77 lower to 0.77 higher)⊕⊕⊕⊕

HIGH
Adverse events (follow-up 3 months)
Ruzyllo 200749randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (b)None181/400 (45.3%)152/404(37.6%)RR 1.2 (1.02 to 1.42)75 more per 1000 (from 8 more to 158 more)⊕⊕⊕○

MODERATE
a

Ruzyllo 2007[48] : Randomisation, allocation concealment, double blinding and ITT reported. N=1195 randomised (Ivabradine 7.5 n=400, Ivabradine 10 mg n=391, or amlodipine n=404). The ITT population consisted of 1155 patients (96.7%). Sample size calculation reported.

b

95% CI around the pooled estimate of effect includes appreciable benefit or appreciable harm.

c

Ivabradine 7.5 mg twice daily vs. amlodipine 10 mg once daily

10.2.3. Economic evidence

No economic studies were identified on this question. We calculated the daily and annual cost of ivabradine treatment based on the unit cost reported in the BNF5919.

Table 10.5Drug cost of Ivabradine

Cost per day (£)Cost per year (£)
Ivabradine, 5 mg or 7.5mg twice daily1.39507

The costs adverse effects were not estimated.

10.2.4. Evidence statements

ClinicalIvabradine versus placebo

Borer 200345: Evidence from one RCT shows that there was no significant difference between ivabradine and placebo for time to angina onset (sec) (at trough) [MD 14.1 (−11.73 to 39.93). Time to angina onset (sec) (at peak) [MD 43.2 (16.75 to 69.65)], time to 1 mm ST depression (sec) at peak [MD 52.90 (26.85 to 78.95)], time to 1mm ST depression (sec) at trough [MD 35.10 (9.68 to 60.52)] was significantly higher in the ivabradine group (5 mg) compared to placebo (follow-up 14 days).

Fox 200946: Evidence from one RCT shows that there was no statistically significant difference between ivabradine (7.5 mg) and placebo in patients with limiting angina for CV death or hospitalisation for MI or HF [RR 0.77 (0.6 to 1.0)], all cause mortality [RR 0.88 (0.64 to 1.2)], cardiac death [RR 0.72 (0.34 to 1.55)], hospitalisation for heart failure [RR 0.85 (0.54 to 1.33)], hospitalisation for heart failure or unstable angina [RR 0.9 (0.64 to 1.28)].

Evidence from RCT shows that there was no statistically significant difference between ivabradine (7.5 mg) and placebo in patients without limiting angina for CV death or hospitalisation for MI or HF [RR 1.04 (0.95 to 1.15)], all cause mortality [RR 1.06 (0.94 to 1.2),] cardiac death [RR 0.91 (0.72 to 1.16)], hospitalisation for heart failure [RR 1 (0.87 to 1.15)] and hospitalisation for heart failure or unstable angina [RR 0.96 (0.81 to 1.14)] and serious adverse events [RR 0.99 (0.80 to 1.22)] (median follow-up 18 months).

Ivabradine versus atenolol

Tardif 200547: Evidence from one RCT shows that there was no statistically significant difference between ivabradine (5 mg bid for 4 weeks and then 7.5 mg bid for 12 weeks) and atenolol (50 mg) for total exercise duration at trough (sec) [MD 8.00 (−13.26 to 29.26)], time to angina onset at trough (sec) [MD 10.00 (−14.96 to 34.96)], weekly number of angina attacks [MD 0.50 (0.99 to 1.99)], short-acting nitrate consumption (units/week) [MD −0.40 (1.00 to 0.20)], and withdrawal due to adverse events [RR1.61 (0.90 to 2.87)] (follow-up 16 weeks).

Ivabradine plus atenolol versus atenolol plus placebo

Tardif 200948 : Evidence from one RCT shows that total exercise duration at trough (sec) [MD 8.70 (0.98 to 16.42)], time to angina onset at trough (sec) [MD 13.00 (3.43 to 22.57)] and time to 1mm ST segment depression (sec) [MD 27.2 (16.15 to 38.25)] at 2 months and total exercise duration at trough (sec) [MD 16.6 (8.05 to 25.15)], time to angina onset at trough (sec) [MD 26.4 (15.64 to 37.16)] and time to 1mm ST segment depression (sec) [MD 30.3 (18.4 to 42.2)] at 4 months was significantly higher in the ivabradine plus atenolol (ivabradine 5 mg b.i.d for 2 months, increased to 7.5 mg b.i.d for a further 2 months) group compared to atenolol. There was no significant difference between Ivabradine and atenolol for angina attacks/week at 4 months [MD 0.00 (0.30 to 0.30)].

The rate of adverse events was significantly higher [(RR3.20 (1.05 to 9.73)] in the ivabradine plus atenolol group compared to atenolol alone (follow-up 2 months and 4 months).

Ivabradine versus amlodipine

Ruzyllo 200749: Evidence from one RCT shows that there were no statistically significant differences between ivabradine (7.5 mg bid) and amlodipine (10 mg/daily) for total exercise duration at trough (sec) [MD −3.60 (−16.5 to 9.3)], time to angina onset at trough (sec) [MD −1.90 (−16.24 to 12.44)], weekly number of angina attacks [MD 0.0 (−0.77 to 0.77)] or short-acting nitrate consumption (units/week) [MD 0.80 (0.04 to 1.56)]. There was significantly higher risk of adverse events with in the ivabradine group compared with amlodipine (RR1.20 (1.02 to 1.42) (follow-up 12 weeks).
EconomicNo economic evidence was found on this question. A simple cost analysis showed a significant drug cost of ivabradine.

10.2.5. Recommendations and link to evidence

RecommendationIf the person cannot tolerate beta blockers and calcium channel blockers or both are contraindicated, consider monotherapy with one of the following drugs*: Decide which drug to use based on comorbidities, contraindications, the person’s preference and drug costs.

For people on beta blocker or calcium channel blocker monotherapy whose symptoms are not controlled and the other option (calcium channel blocker or beta blocker) is contraindicated or not tolerated, consider one of the following as an additional drug*: Decide which drug to use based on comorbidities, contraindications, the person’s preference and drug costs.

* Evidence on long-acting nitrates is presented in chapter 9. Evidence on nicorandil and ranolazine is presented in sections 10.3 and 10.4 respectively of this chapter.

** When combining ivabradine with a calcium channel blocker, use a dihydropyridine calcium channel blocker, for example, slow release nifedipine, amlodipine, or felodipine.

*** At the time of publication (July 2011), nicorandil did not have UK marketing authorisation for this indication. Informed consent should be obtained and documented.
Relative values of different outcomesOutcomes of interest included long-term mortality (total and cardiovascular), rates of major adverse cardiovascular events (myocardial infarction, stroke, myocardial revascularisation), and measures of symptom severity (frequency of angina, exercise test outcomes).
Trade off between clinical benefits and harmsShort-term trials of monotherapy with ivabradine versus monotherapy with atenolol or amlodipine demonstrated similar increases in total exercise duration, and similar reductions in the frequency of angina episodes in both treatment groups.

One short-term trial reported that addition of ivabradine to monotherapy with atenolol resulted in small increases in total exercise duration (16.6s) and time to angina on the treadmill, but did not reduce the frequency of episodes of angina.

These data suggest that ivabradine is an effective anti-anginal agent with comparable short-term efficacy to atenolol and amlodipine. In addition there is a statistically significant incremental benefit of adding ivabradine to atenolol in people with angina, but this benefit is of uncertain clinical significance.

There are significantly higher rates of adverse events in ivabradine treated patients, partly due to visual disturbance (phosphenes and blurred vision).
Economic considerationsNo economic evidence on the use of ivabradine for the treatment of stable angina was available for review. The cost of ivabradine is substantially higher than the costs of other standard treatments (including BB, CCB, and long-acting nitrate).
Quality of evidenceThe trials assessing the short-term anti-anginal efficacy of ivabradine were relatively large, well-designed studies. Evidence confirming the long-term efficacy and safety of ivabradine in people with stable angina is limited.

The BEAUTIFUL trial assessed the effect of ivabradine in people with coronary artery disease and impaired left ventricular function. In a subgroup analysis of patients whose limiting baseline symptom was angina, ivabradine was associated with a reduction in the composite rate of the primary endpoint (cardiovascular death and hospitalization for myocardial infarction or heart failure) of borderline statistical significance. The rate of hospitalisation for myocardial infarction was lower in the ivabradine treated patients (RR 0.58, 95%CI 0.37–0.92, p=0.021)46. The subgroup was defined retrospectively, only includes 13.8% of the total trial population, and lacks statistical power for the primary endpoint. The GDG considered this analysis to be exploratory, rather than providing definitive evidence of benefit of ivabradine in people with stable angina and impaired left ventricular systolic function.

No economic evidence was found on this question.
Other considerationsThere is some evidence for the use of ivabradine as monotherapy or in combination with BB, but no evidence for use of ivabradine in combination with CCB was found. Concomitant use of ivabradine with heart rate reducing CCB such as verapamil or diltiazem is not recommended by the manufacturers.

Ivabradine is a relatively new drug with limited information on efficacy in stable angina. The cost of ivabradine is comparable with the costs of nicorandil and ranolazine but more than the cost of long-acting nitrate. Nevertheless the GDG considered that there was insufficient evidence to make a firm recommendation about the choice of antianginal drug as monotherapy or as an additional antianginal drug if a CCB or BB is not tolerated or is contraindicated.

The GDG concluded that monotherapy with ivabradine should not be used as an alternative to monotherapy with a BB or CCB. Monotherapy with ivabradine can be considered in people with stable angina in whom BB and CCB are contraindicated or not tolerated.

The GDG concluded that ivabradine can be considered as an additional drug for people whose symptoms are not controlled by monotherapy with a BB and the addition of CCB is contraindicated or not tolerated. Ivabradine can be combined with CCB but there is less evidence for efficacy and safety. The summary of product characteristics (SPC) for ivabradine states that ivabradine should only be combined with dihydropyridine CCB.

Ivabradine is metabolised by CYP3A4, and CYP3A4 inhibitors and inducers are liable to interact with ivabradine and influence its metabolism and pharmacokinetics to a clinically significant extent. Drug-drug interaction studies have established that CYP3A4 inhibitors increase ivabradine plasma concentrations. Specific interaction studies in healthy volunteers and patients have shown that the combination of ivabradine with the heart rate reducing agents diltiazem or verapamil resulted in an increase in ivabradine exposure (2 to 3 fold increase in AUC) and an additional heart rate reduction of 5 bpm.

10.3. Nicorandil

Nicorandil is a nitrate derivative of nicotinamide that is licensed for the prevention and long-term treatment of angina. Nicorandil is believed to have a dual mechanism of action. Specifically nicorandil provides a nitrate moiety that dilates epicardial coronary arteries and systemic venous capacitance vessels. In addition, nicorandil opens ATP-sensitive potassium channels (KATP) in vascular smooth muscle cells, thereby dilating arterial resistance vessels in the peripheral and coronary circulations. In humans nicorandil decreases ventricular filling pressure, coronary vascular resistance, and mean arterial pressure, and these combined effects increase coronary blood flow and reduce myocardial work.

KATP channels are an important mediator of ischaemic preconditioning. The molecular mechanisms have not been fully elucidated but activation of the KATP channel has a cardioprotective effect similar to ischaemic preconditioning, while KATP channel blockade prevents preconditioning. Experimental and clinical studies of myocardial ischaemia provide evidence that pretreatment with nicorandil reduces ischaemic myocardial injury. It has therefore been suggested that in addition to relieving symptoms of ischaemia nicorandil may have a clinically relevant cardioprotective effect.

10.3.1. Clinical question

What is the clinical/cost effectiveness of nicorandil for the management of stable angina?

10.3.2. Clinical evidence

The “Review Protocol” for this topic can be found in Appendix C, the “Search Strategies” in Appendix D, the “List of Included and Excluded Studies” in Appendix E1, the “Clinical Evidence Tables” in Appendix E2, and the “Forest Plots” in Appendix F.

Table 10.6Nicorandil + usual treatment versus Placebo + usual treatment

Quality assessmentSummary of findings
No of patientsEffectQuality
No of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerationsNicorandilPlaceboRelative (95% CI)Absolute
CHD death (follow-up 1.6 years)
Dargie 200250 (IONA) (d)randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)None60/2565 (2.3%)73/2561 (2.9%)RR 0.82 (0.59 to 1.15)5 fewer per 1000 (from 12 fewer to 4 more)⊕⊕○○

LOW
Non fatal MI (follow-up 1.6 years)
Dargie 200250 (IONA)randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)None56/2565 (2.2%)72/2561 (2.8%)RR 0.78 (0.55 to 1.1)6 fewer per 1000 (from 13 fewer to 3 more)⊕⊕○○

LOW
Unstable angina (follow-up 1.6 years)
Dargie 200250 (IONA)randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)None115/2565 (4.5%)127/2561 (5%)RR 0.9 (0.71 to 1.16)5 fewer per 1000 (from 14 fewer to 8 more)⊕⊕○○

LOW
All cause mortality (follow-up 1.6 years)
Dargie 200250 (IONA)randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)None111/2565 (4.3%)129/2561 (5%)RR 0.86 (0.67 to 1.1)7 fewer per 1000 (from 17 fewer to 5 more)⊕⊕○○

LOW
Worsening of angina status (follow-up 1.6 years)
Dargie 200250 (IONA)randomised trialsserious (a)no serious inconsistencyno serious indirectnessno serious imprecisionNone569/2565 (22.2%)602/2561 (23.5%)RR 0.94 (0.85 to 1.04)14 fewer per 1000 (from 35 fewer to 9 more)⊕⊕⊕○

MODERATE
GI disturbances (follow-up 1.6 years)
Dargie 200250 (IONA)randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (c)None194/2565 (7.6%)132/2561 (5.2%)RR 1.47 (1.18 to 1.82)24 more per 1000 (from 9 more to 42 more)⊕⊕○○

LOW
Combined outcome CHD death, non-fatal MI or hospital admission for chest pain (diabetes subgroup) (follow-up 1.6 years)
IONA Study Group 200451 (IONA)randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)None27/197 (13.7%)40/232 (17.2%)RR 0.79 (0.51 to 1.25)36 fewer per 1000 (from 84 fewer to 43 more)⊕⊕○○

LOW
Combined outcome CHD death, non-fatal MI or hospital admission for chest pain (age subgroup >70 yrs) (follow-up 1.6 years)
IONA Study Group 200451 (IONA)randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)None131/927 (14.1%)167/948 (17.6%)RR 0.8 (0.65 to 0.99)35 fewer per 1000 (from 2 fewer to 62 fewer)⊕⊕○○

LOW
Combined outcomes CHD death, non-fatal MI or hospital admission for chest pain (female subgroup) (follow-up 1.6 years)
IONA Study Group 200451 (IONA)randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)none86/603 (14.3%)87/613 (14.2%)RR 1 (0.76 to 1.32)0 fewer per 1000 (from 34 fewer to 45 more)⊕⊕○○

LOW
Composite (CHD death, non fatal MI or hospital admission. for chest pain) (follow-up 1.6 years)
Dargie 200250 (IONA)randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (c)None337/2565 (13.1%)398/2561 (15.5%)RR 0.85 (0.74 to 0.97)23 fewer per 1000 (from 5 fewer to 40 fewer)⊕⊕○○

LOW
Headache (follow-up 1.6 years)
Dargie 200250 (IONA)randomised trialsserious (a)no serious inconsistencyno serious indirectnessno serious imprecisionNone364/2565 (14.2%)81/2561 (3.2%)RR 4.49 (3.55 to 5.67)110 more per 1000 (from 81 more to 148 more)⊕⊕⊕○

MODERATE
a

Dargie 2002[49]: Randomisation process was reported; allocation concealment was not reported; study was double blind; Number of drop outs were reported and >20%; Intention to treat analysis was reported; the study was powered for primary outcome (CHD death, non fatal MI, or unplanned hospitalisation). In the placebo group: 2/2561 lost to follow-up, 809/2561 discontinued intervention. In the Nicorandil group: 2/2565 lost to follow-up; 1003/2565 discontinued intervention

b

95% CI around the pooled estimate of effect includes both 1) no effect 2) appreciable benefit or appreciable harm

c

95% CI around the pooled estimate of effect includes appreciable benefit or appreciable harm

d

Canadian Cardiovascular Society Functional classification of angina at the end of the study (follow-up mean 1.6 years):

  • Class I - Nicorandil 985 (43%); placebo 989 (43%)
  • Class II- Nicorandil 1159 (50%); placebo 1124 (49%)
  • Class III- Nicorandil 162 (7%); placebo 163 (7%)
  • Class IV- Nicorandil 9 (<1%); placebo 15 (1%)

Table 10.7Nicorandil versus diltiazem

Quality assessmentSummary of findings
No of patientsEffectQuality
No of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerationsNicorandilDiltiazemRelative (95% CI)Absolute
Exercise capacity (work to peak exercise) (KJ) (follow-up 90 days: better indicated by more)
Guermonprez 199352randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)None5056-MD 2.4 higher (60.15 lower to 64.95 higher)⊕⊕○○

LOW
Exercise capacity (work to onset of angina) (KJ) (follow-up 90 days: better indicated by more)
Guermonprez 199352randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)None5056-MD 3.40 higher (58.91 lower to 65.71 higher)⊕⊕○○

LOW
Adverse events (combined) (follow-up 90 days)
Guermonprez 199352randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (c)None19/60 (31.7%)19/63 (30.2%)RR 1.05 (0.62 to 1.78)15 more per 1000 (from 115 fewer to 235 more)⊕⊕○○

LOW
a

Guermonprez 1993[51]: Allocation concealment was not reported; study was double blind; intention to treat analysis was not reported. The two groups were comparable in terms of age, sex, history of myocardial infarction and severity of coronary lesions. Drop-out: Nicorandil: 3 (5%) drop outs due to insufficeint efficacy; 4(6.7%) drop outs due to other adverse events. Diltiazem: 4 (6.3%) drop outs due to insufficient efficacy; 1 (1.6%) drop outs due to other adverse events.

b

95% CI includes no effect and the upper and the lower confidence limit crosses the MID.

c

95% CI around the pooled estimate of effect includes both 1) no effect 2) appreciable benefit or appreciable harm

Table 10.8Nicorandil vs. diltiazem

*(Data for this outcome not able to analyse. Results reported as in the paper)

OutcomeNumber of studiesDesignLimitationsInconsistencyDirectnessImprecision
Frequency of anginal attacks per week1 (Guermonprez)RCT (double blind)Serious1No serious InconsistencyNo serious indirectnessNo serious imprecision
OutcomeNicorandilPlaceboRelative riskAbsolute effectQuality
Follow-up 90 days
Frequency of anginal attacks per week0.7 (mean)2--SD not reported. P=0.56 (Difference between groups not significant).MODERATE
1

Guermonprez 1993[51]: Allocation concealment was not reported; study was double blind; Number of drop-outs were reported and < 20%; Intention to treat analysis was not reported.

2

Mean value reported for both groups together. No standard deviation (SD) reported

Table 10.9Nicorandil versus amlodipine

Quality assessmentSummary of findings
No of patientsEffectQuality
No of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerationsNicorandilAmlodipineRelative (95% CI)Absolute
ETT (Total exercise duration) (min) (follow-up 8 weeks; better indicated by higher values)
Chatterjee 199953randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)none5662-MD 0.7 lower (1.69 lower to 0.29 higher)⊕⊕○○

LOW
ETT (Time to ST-segment depression) (follow-up 8 weeks; better indicated by higher values)
Chatterjee 199953randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)none5662-MD 0.6 lower (1.45 lower to 0.25 higher)⊕⊕○○

LOW
ETT (Time to onset of anginal pain) (follow-up 8 weeks; better indicated by higher values)
Chatterjee 199953randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)none5662-MD 0.9 lower (2 lower to 0.2 higher)⊕⊕○○

LOW
Sum of weekly anginal attacks (follow-up 8 weeks; better indicated by lower values)
Chatterjee 199953randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (c)none5662-MD 1.2 higher (0.54 to 1.86 higher)⊕⊕○○

LOW
Adverse events (combined) (follow-up 8 weeks)
Chatterjee 199953randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (d)none20/57 (35.1%)20/64 (31.3%)RR 1.12 (0.68 to 1.86)38 more per 1000 (from 100 fewer to 269 more)⊕⊕○○

LOW
a

Chatterjee 1999[52]: Randomised, double blind, parallel group design. Allocation concealment was not reported. The study reports of 6/121 patients drop out [5 because of adverse events (nicorandil n=3 and amlodipine n=2) and one due to compliance problems (amlodipine)]. The study reports that 118/121 were evlauted for efficacy on an ITT basis. The treatment groups were comparable for demographic and clinical characteristics within and between the two countries (Austria and Switzerland), with the exception of history of previous MI among patients recruited in Austria (nicorandil, n=2; amlodipine, n=14). The mean number of anginal attacks/week was similar in both nicorandil and amlodipine groups at baseline. However, the mean number of nitroglycerin units required for pain relief was significantly higher (p=0.04) in the nicorandil group (2.3 vs. 1.0 units/week). Baseline BP, HR and ETT variables were similar in the 2 treatment groups.

b

95% CI includes no effect and the upper and the lower confidence limit crosses the MID.

c

The upper and the lower confidence limit crosses the MID.

d

95% CI around the pooled estimate of effect includes both 1) no effect 2) appreciable benefit or appreciable harm

Table 10.10Nicorandil vs. nifedipine for stable angina

Quality assessmentSummary of findings
No of patientsEffectQuality
No of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerationsNicorandilNifedipineRelative (95% CI)Absolute
Weekly anginal attack rate (follow-up after 8 weeks of treatment; better indicated by lower values)
Ulvenstam 199254randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)none2723-MD 5.3 lower (11.48 lower to 0.88 higher)⊕⊕○○

LOW
Exercise duration (min) (follow-up after 8 weeks of treatment; better indicated by higher values)
Ulvenstam 199254randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)none2523-MD 1 higher (0.59 lower to 2.59 higher)⊕⊕○○

LOW
Time to onset of angina pectoris (min) (follow-up after 8 weeks of treatment; better indicated by higher values)
Ulvenstam 199254randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)none2322-MD 1.1 higher (0.75 lower to 2.95 higher)⊕⊕○○ LOW
Time to 1mm ST-depression (min) (follow-up after 8 weeks of treatment; better indicated by higher values)
Ulvenstam 199254randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)none2320-MD 1.6 higher (0.02 lower to 3.22 higher)⊕⊕○○

LOW
ST depression on maximal identical workload (mm) (follow-up after 8 weeks of treatment; better indicated by higher values)
Ulvenstam 199254randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (b)none2420-MD 0.2 higher (0.28 lower to 0.68 higher)⊕⊕○○

LOW
Adverse events (combined) follow-up after 8 weeks of treatment; better indicated by lower values)
Ulvenstam 199254randomised trialsserious (a)no serious inconsistencyno serious indirectnessno serious imprecisionnone25/29 (86.2%)28/29 (96.6%)RR 0.89 (0.76 to 1.05)106 fewer per 1000 (from 232 fewer to 48 more)⊕⊕⊕○

MODERATE
a

Ulvenstam 1992[53]: Double-blind, randomised, multicentre study. 55/58 completed the study. Allocation concealment not reported. ITT not reported. Baseline comparisons made-previous MI more frequent in the nicorandil group.

b

95% CI includes no effect and the upper and the lower confidence limit crosses the MID.

Table 10.11Nicorandil versus isosorbide mononitrate

Quality assessmentSummary of findings
No of patientsEffectQuality
No of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerationsNicorandilISMNRelative (95% CI)Absolute
ETT (Total exercise time) (sec) (follow-up 2 weeks; better indicated by higher values)
Zhu 200755randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious imprecision(b)None115117-MD 3.2 lower (37.26 lower to 30.86 higher)⊕⊕○○

LOW
ETT (Time to ST depression) (follow-up 2 weeks; better indicated by higher values)
Zhu 200755randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious imprecision (b)None114116-MD 2.4 higher (37.98 lower to 42.78 higher)⊕⊕○○

LOW
Adverse event (Headache) (follow-up 2 weeks)
Zhu 200755randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (c)None15/123 (12.2%)18/123 (14.6%)RR 0.83 (0.44 to 1.58)25 fewer per 1000 (from 82 fewer to 85 more)⊕⊕○○

LOW
a

Zhu 2007[54]: Randomised, allocation concealment not reported. Double blind. Power calculation used. N=232 patients completed the study (N=115 in nicorandil group and N=117 in the Isosorbide mononitrate (ISMN) group). Drop-out rate: 7% drop out (8% in the nicorandil group and 6% in the ISMN group). Intention to treat analysis was not reported. Baseline comparisons made. No significant difference between the groups.

b

95% CI includes no effect and the upper and the lower confidence limit crosses the MID.

c

95% CI around the pooled estimate of effect includes both 1) no effect 2) appreciable benefit or appreciable harm

Table 10.12Nicorandil versus propanolol

Quality assessmentSummary of findings
No of patientsEffectQuality
No of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerationsNicorandilPropanololRelative (95% CI)Absolute
Angina free in daily life (%) (follow-up 6 weeks; better indicated by higher values)
Meeter 199256randomised trialsserious (a)no serious inconsistencyno serious indirectnessno serious imprecision (b)None11/32 (34.4%)13/37 (35.1%)RR 0.98 (0.51 to 1.87)7 fewer per 1000 (from 172 fewer to 306 more)⊕⊕⊕○

MODERATE
12 hrs after medication - change in maximal work load (follow-up 3 weeks; better indicated by higher values)
Meeter 199256randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (c)None3237-MD 6 lower (14.77 lower to 2.77 higher)4⊕⊕○○

LOW
12 hrs after medication - change in maximal work load (W) (follow-up 6 weeks; better indicated by higher values)
Meeter 199256randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (e)None3237-MD 5 lower (15.72 lower to 5.72 higher)⊕⊕○○

LOW
12 hrs after treatment - change in time to angina (follow-up 3 weeks; better indicated by higher values)
Meeter 199256randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (c)None3237-MD 0.10 lower (1.05 lower to 0.85 higher)⊕⊕○○

LOW
12 hrs after treatment - change in time to angina (follow-up 6 weeks; better indicated by lower values)
Meeter 199256randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (c)None3237-MD 0.40 lower (1.35 lower to 0.55 higher)⊕⊕○○

LOW
2 hrs after treatment - change in maximal work load (follow-up 3 weeks; better indicated by higher values)
Meeter 199256randomised trialsserious (a)no serious inconsistencyno serious indirectnessno serious imprecisionNone3237-MD 5.00 lower (13.07 lower to 3.07 higher)⊕⊕⊕○

MODERATE
2 hrs after treatment - change in maximal work load (W) (follow-up 6 weeks; better indicated by higher values)
Meeter 199256randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (c)None3237-MD 5.00 lower (14.47 lower to 4.47 higher)⊕⊕○○

LOW
2 hrs after treatment - change in time to angina (follow-up 3 weeks; better indicated by lower values)
Meeter 199256randomised trialsserious (a)no serious inconsistencyno serious indirectnessserious (c)None3237-MD 0.20 higher (0.53 lower to 0.93 higher)⊕⊕○○

LOW
2 hrs after medication - change in time to angina (follow-up 6 weeks; better indicated by higher values)
Meeter 199256randomised trialsserious (a)no serious inconsistencyno serious indirectnessno serious imprecision (c)None3237-MD 0.60 higher (0.35 lower to 1.55 higher)⊕⊕⊕○

MODERATE
a

Meeter 1992[55]: Allocation concealment not reported; double blind; drop-out rate reported and < 20%; Intention to treat analysis not reported. Baseline comparisons made-Nicorandil group had shorter duration of angina at baseline compared to propranolol median 12 vs 20 months. Fewer patients had a history of MI on nicorandil vs propranolol 16/38 vs. 22/39. 5/77 were withdrawn from the trial (4 receiving nicorandil and one receiving propranolol). The patients receiving nicorandil were withdrawn because of worsening of angina complaints or headaches.

b

95% C around the pooled estimate of effect includes both 1) no effect 2) appreciable benefit or appreciable harm.

c

The upper and lower limits of 95% CI crosses the MI.

10.3.3. Economic evidence

No economic evidence was found on the use of nicorandil as monotherapy. Based on the unit cost reported in the BNF5919 the annual drug cost ranges from £99 and £190.

We found one study57 comparing the addition of nicorandil to usual care with placebo. This is summarised in the economic evidence profile below. See also Economic Evidence Tables in Appendix G.

Table 8.13Nicorandil+usual care vs. placebo+usual care – Economic study characteristics

StudyLimitationsApplicabilityOther Comments
Walker 200657Potentially serious limitations (a)Partial applicability (b)Intervention was nicorandil 20mg bd + usual care (57% BB, 55% CCB, 87% nitrates, 88% aspirin). Based on the IONA trial50 included in the clinical review.
a

Short follow-up (up to 1.6 years). Sensitivity analysis was quite limited and was applied only to the primary analysis (cost of care after discharge excluded). Morbidity associated with gastro-intestinal events is not included. Effectiveness data were reported only in the incremental analysis.

b

QALYs were not estimated.

Table 8.14Nicorandil+usual care vs. placebo+usual care – Economic summary of findings

StudyIncremental cost per patient (£)Incremental effects per patient (primary end-point averted)ICERUncertainty
Walker 200657Saves £0.12 (a)0.024 (b)DominantNicorandil is not cost-saving when:
-

cost of care after discharge is included

-

either cost of cardiology, cardiac surgery or ICU is reduced by 20%. Results were similar when the measure of effectiveness considered was the number of event-free survivors (events were cardiac death, non-fatal MI, unstable and stable angina, stroke, hospital admission for TIA) or the number of cases of definite acute coronary syndromes (coronary heart disease death, non-fatal myocardial infarction or unstable angina).

a

2002 GBP. Costs included were cost of Nicorandil (including 10% dispensing fee and two additional physician visits), adverse events related to Nicorandil, hospital admissions, surgical procedures. The cost of post-discharge care was not included in the base case analysis.

b

Primary end-points considered in the analysis were cardiac death, non-fatal MI, hospital admission for cardiac chest pain.

10.3.4. Evidence statements

ClinicalClinical Efficacy

Nicorandil + usual treatment versus Placebo + usual treatment

Dargie 200250 and IONA Study Group 200451: Evidence from one RCT shows that the composite outcomes (CHD death, non-fatal MI, or unplanned hospital admission for chest pain) for the entire group [RR 0.85 (0.74 to 0.97)] and for people aged over 70 years [RR 0.80 (0.65 to 0.99)] were significantly reduced in the nicorandil group compared to placebo (mean follow-up 1.6 years).

Dargie 200250 and IONA Study Group 200451: Evidence from one RCT shows that there were no statistically significant differences between nicorandil and placebo for CHD death [RR 0.82 (0.59 to 1.15)], non fatal MI [RR 0.78 (0.55 to 1.10)], all cause mortality [RR (0.86 (0.67 to 1.10)], unstable angina [RR 0.90 (0.71 to 1.16)], and worsening of angina status [RR 0.94 (0.85 to 1.04)]. There were no statistically significant differences between treatment groups for a composite morbidity/mortality outcome (CHD death, non-fatal MI, or unplanned hospital admission for chest pain) in subgroup analyses of results for women [RR 1.00 [0.76 to 1.31], and people with diabetes [RR 0.79 (0.51 to 1.25)] (mean follow-up 1.6 years).

Nicorandil versus diltiazem

Guermonprez 199252: Evidence from one RCT shows that there was no significant difference between nicorandil and diltiazem for exercise capacity (work to peak exercise] [MD 2.4 (−60.15 to 64.95) and exercise capacity (work required to reach onset of angina) [MD 3.40 (−58.91 to 64.95)] (follow-up 90 days).

Nicorandil versus amlodipine

Chatterjee 199953: Evidence from one RCT shows that there were no significant differences between nicorandil and amlodipine for total exercise duration (min), MD −0.70 [−1.69 to 0.29], ETT (Time to onset of anginal pain) MD −0.9 (−2 to 0.2), and ETT (Time to ST-segment depression) [MD −0.6 (−1.45 to 0.25 higher) and sum of weekly anginal attacks, [MD 1.20 [0.54 to 1.86] (follow-up 8 weeks).

Nicorandil versus nifedipine

Ulvenstam 199254: Evidence from one RCT shows that there was no statistically significant differences between nicorandil and nifedipine for weekly anginal attack rate [MD −5.3 (−11.48 to 0.88)], exercise duration (min) [MD 1 higher (−0.59 to 2.59)], time to onset of angina pectoris (min) [MD 1.1 (−0.75 to 2.95)], time to 1mm ST-depression (min) [MD 1.6 (−0.02 to 3.22)], and ST depression on maximal identical workload (mm) [MD 0.2 (−0.28 to 0.68)] (follow-up after 8 weeks of treatment).

Nicorandil versus isosorbide mononitrate

Zhu 200755: Evidence from one RCT shows that there was no significant difference between nicorandil and isosorbide mononitrate for total exercise time (sec) [MD −3.20 [−37.26 to 30.86] and ETT (time to ST-depression) [MD 2.4 (−37.98 to 42.78)] (follow-up 2 weeks).

Nicorandil versus propranolol

Meeter 199256: Evidence from one RCT shows that there was no significant difference between nicorandil and propranalol for frequency of anginal attacks [RR 0.98 (0.51 to 1.87)] (follow-up 6 weeks).

Meeter 199256: Evidence from one RCT shows that there was no significant difference between nicorandil and propanalol for change in maximal workload 12 hrs after medication at 3 weeks [MD −6 (−14.77 to 2.77)] and 6 weeks [MD −5 (−15.72 to 5.72)], change in time to angina decimal min 12 hrs after medication at 3 weeks [MD −0.10 (−1.05 to −0.85)] and 6 weeks [MD −0.40 (−1.35 to 0.55)], change in maximal workload 2 hrs after treatment at 3 weeks [MD −5.00 (−13.07 to 3.07)] and 6 weeks [MD −5.00 (−14.47 to 4.47)], change in time to angina 2 hrs after treatment at 3 weeks [MD 0.20 (−0.53 to 0.93)] and 6 weeks [MD 0.60 (−0.35 to 1.55)] (follow-up 6 weeks).

Adverse events

Nicorandil versus placebo

Dargie 200250 (IONA): Evidence from one RCT shows that there were significantly greater GI disturbances [RR 1.47 (1.18 to 1.82)] and headaches in the nicorandil compared to placebo [RR 4.49 (3.55 to 5.67)] (mean follow-up 1.6 years) (follow-up mean 1.6 years).

Nicorandil versus diltiazem

Guermonprez 199352: Evidence from one RCT shows that there were no statistically significant differences between nicorandil and diltiazem for adverse effects (combined) [RR 1.05 (0.62 to 1.78)] (follow-up 90 days).

Nicorandil versus amlodipine

Chatterjee 199953: Evidence from one RCT suggests that there were no statistically significant differences between nicorandil and amlodipine for adverse effects (combined) [RR 1.1 (0.68 to 1.86)] (follow-up 8 weeks).

Nicorandil vs. nifedipine

Ulvenstam 199254: Evidence from one RCT suggests that there were no statistically significant differences between nicorandil and nifedipine for adverse events (combined) [RR 0.89 (0.76 to 1.05)] (follow-up after 8 weeks of treatment).

Nicorandil versus isosorbide mononitrate

Zhu 200755: Evidence from one RCT suggests that there were no statistically significant difference between nicorandil and isosorbide mononitrate for adverse effects (headache) [RR 0.83 (0.44 to 1.58)] (follow-up 2 weeks).

Nicorandil versus propranolol

Meeter 199256: Adverse effects not reported (follow-up 6 weeks).
EconomicNicorandil is cost-neutral when post discharge care is not included and over a short time (1.6 years). It could be less cost effective when post-discharge care is included. This evidence has potentially serious limitations and partial applicability.

10.3.5. Recommendations and link to evidence

RecommendationIf the person cannot tolerate beta blockers and calcium channel blockers or both are contraindicated, consider monotherapy with one of the following drugs*: Decide which drug to use based on comorbidities, contraindications, the person’s preference and drug costs.

For people on beta blocker or calcium channel blocker monotherapy whose symptoms are not controlled and the other option (calcium channel blocker or beta blocker) is contraindicated or not tolerated, consider one of the following as an additional drug*: Decide which drug to use based on comorbidities, contraindications, the person’s preference and drug costs.

* Evidence on long-acting nitrates is presented in chapter 9. Evidence on ivabradine and ranolazine is presented in sections 10.2 and 10.4 respectively of this chapter.

** When combining ivabradine with a calcium channel blocker, use a dihydropyridine calcium channel blocker, for example, slow release nifedipine, amlodipine, or felodipine.

*** At the time of publication (July 2011), nicorandil did not have UK marketing authorisation for this indication. Informed consent should be obtained and documented.
Relative values of different outcomesOutcomes of interest included long-term mortality (total and cardiovascular), rates of major adverse cardiovascular events (myocardial infarction, stroke, myocardial revascularisation), and measures of symptom severity (frequency of angina, exercise test outcomes).
Trade off between clinical benefits and harmsNo evidence was found to assess the effects of monotherapy with nicorandil on long term mortality or rates of major adverse cardiovascular events in people with stable angina.

Short-term trials of monotherapy with nicorandil versus monotherapy with other anti-anginal drugs (diltiazem, amlodipine, or propranolol) demonstrated similar reductions in the frequency of episodes of angina in both treatment groups.

These trials also reported similar increases in total exercise capacity during monotherapy with nicorandil and monotherapy with diltiazem, amlodipine, propranolol, or isosorbide mononitrate.

No difference in the short-term rate of adverse effects was reported between nicorandil and diltiazem, amlodipine, or isosorbide mononitrate.
Economic considerationsNo economic evidence on the use of nicorandil monotherapy was found. The annual drug cost of nicorandil ranges from £99 to £190.
Quality of evidenceLow quality evidence from trials with small sample size and short duration of follow-up. In one trial56 no intention to treat analysis was carried out.

No economic evidence was found.
Other considerationsThe GDG concluded that there is insufficient evidence to recommend monotherapy with nicorandil in preference to monotherapy with a BB or CCB as first line treatment for angina. Nicorandil can be considered as monotherapy for the treatment of stable angina if BB and CCB are not tolerated or contraindicated.

Adverse effects of nicorandil include headache (especially on initiation of treatment), flushing, dizziness, reduction in blood pressure and/or increase in heart rate, and gastrointestinal side effects including mucosal ulceration. In the IONA trial routine treatment with nicorandil was associated with a higher risk of gastrointestinal side effects and GDG members have experience of patients who developed gastrointestinal ulceration during treatment with nicorandil.
RecommendationFor people on beta blocker or calcium channel blocker monotherapy whose symptoms are not controlled and the other option (calcium channel blocker or beta blocker) is contraindicated or not tolerated, consider one of the following as an additional drug: Decide which drug to use based on comorbidities, contraindications, the person s preference and drug costs.

* When combining ivabradine with a calcium channel blocker, use a dihydropyridine calcium channel blocker for example, slow release nifedipine, amlodipine, or felodipine.

** At the time of publication (July 2011), nicorandil did not have UK marketing authorisation for this indication. Informed consent should be obtained and documented.

NOTE: Evidence on long-acting nitrates is presented in chapter 9. Evidence on ivabradine and ranolazine is presented in sections 10.2 and 10.4 respectively of this chapter.
Relative values of different outcomesOutcomes of interest included long-term mortality (total and cardiovascular), rates of major adverse cardiovascular events (myocardial infarction, stroke, myocardial revascularisation), and measures of symptom severity (frequency of angina, exercise test outcomes).
Trade off between clinical benefits and harmsIn a large trial addition of nicorandil to standard antianginal treatment (BB 56%, CCB 55%, nitrate 87%) in people with stable angina reduced the composite of coronary heart disease death, myocardial infarction, and unplanned hospitalisation for chest pain. There were trends for lower rates of all events included in the composite primary endpoint in the nicorandil group, but these were not statistically significant. At the end of the study (1.6 years) the Canadian Cardiovascular Society angina class did not differ between the two groups. Headache, gastrointestinal disturbance, and treatment withdrawal because of adverse effects were more frequent in the nicorandil group.

The GDG concluded that the 2.4% absolute reduction in the rate of the primary composite endpoint in IONA did not justify the routine use of nicorandil as add-on therapy to standard antianginal treatment in people with stable angina, particularly as the drug is associated with an excess risk of adverse events, including headache and gastrointestinal disturbance.
Economic considerationsAn economic analysis based on IONA suggested that addition of nicorandil to standard anti-anginal treatment is cost neutral. However there is a high uncertainty over this conclusion as the cost of post-discharge care and adverse events was not considered, and the follow-up time was only 1.6 years.
Quality of evidenceModerate quality evidence from a large multicentre trial powered to detect a 20% reduction in the primary endpoint. Allocation concealment was not reported and treatment withdrawal was >20% in both groups.

The economic evidence has potentially serious limitations and partial applicability.
Other considerationsThe GDG concluded that addition of nicorandil is an option for people whose symptoms of angina are not controlled by a BB or CCB. Nicorandil is slightly cheaper than ivabradine and ranolazine but more than the cost of long-acting nitrate. Nicorandil is currently not licensed for use in combination treatment. Nevertheless the GDG concluded that there was insufficient evidence to make a firm recommendation about the choice of an additional antianginal drug if a BB or CCB is not tolerated or is contraindicated.

10.4. Ranolazine

The mechanism of action of ranolazine has not been fully elucidated, but it is believed to act by selective inhibition of late sodium influx across the sarcolemma, which attenuates the abnormalities of ventricular repolarisation and contractility associated with myocardial ischaemia. Reported side-effects include dizziness, constipation, and nausea. Ranolazine has the potential to prolong the QT interval and is contraindicated in people with pre-existing QT prolongation. Ranolazine should be avoided in severe hepatic or renal impairment. Ranolazine is available in a sustained release formulation, with an elimination half-life of about seven hours.

The summary of product characteristics (SPC) for ranolazine reports that minimal decreases in mean heart rate and mean systolic blood pressure have been observed in patients treated with ranolazine either alone or in combination with other antianginal medicinal products in controlled studies. No proarrhythmic effects were observed in 3,162 patients treated with ranolazine based on 7-day Holter monitoring in the MERLIN-TIMI 36 study58.

This section reviews evidence for the use of ranolazine as adjunctive therapy to control symptoms and improve outcome in people with stable angina.

10.4.1. Clinical question

What is the clinical/cost effectiveness of ranolazine for the management of stable angina?

10.4.2. Clinical evidence

The “Review Protocol” for this topic can be found in Appendix C, the “Search Strategies” in Appendix D, the “List of Included and Excluded Studies” in Appendix E1, the “Clinical Evidence Tables” in Appendix E2, and the “Forest Plots” in Appendix F.

Table 10.15Ranolazine (750 mg bid) + antianginal treatment vs. placebo + antianginal treatment (follow-up 12 weeks)

Quality assessmentSummary of findings
No of patientsEffectQuality
No of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerationsRanolazine (750 mg bid) + antianginalPlacebo + antianginalRelative (95% CI)Absolute
Exercise duration (sec) (trough - change from baseline) - (follow-up 12 weeks; better indicated by higher values)
Chaitman 200459 (CARISA) (c)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionnone272258-MD 23.7 higher (1.11 to 46.29 higher)⊕⊕⊕⊕

HIGH
Time to onset of angina (sec) (trough - change from baseline) - (follow-up 12 weeks; better indicated by higher values)
Chaitman 200459 (CARISA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionnone272258-MD 29.7 higher (4.62 to 54.78 higher)⊕⊕⊕⊕

HIGH
Exercise duration (sec) (peak - change from baseline) - (follow-up 12 weeks; better indicated by higher values)
Chaitman 200459 (CARISA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionnone270256-MD 34 higher (11.96 to 56.04 higher)⊕⊕⊕⊕

HIGH
Time to onset of angina (sec) (peak - change from baseline) - (follow-up 12 weeks; better indicated by higher values)
Chaitman 200459 (CARISA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionnone272256-MD 38 higher (13.91 to 62.09 higher)⊕⊕⊕⊕

HIGH
Adverse events (follow-up 12 weeks)
Chaitman 200459 (CARISA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (b)none82/279 (29.4%)71/269 (26.4%)RR 1.11 (0.85 to 1.46)29 more per 1000 (from 40 fewer to 121 more)⊕⊕⊕○

MODERATE
Angina attacks per week (follow-up 12 weeks; better indicated by lower values)
Chaitman 200459 (CARISA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionnone272258-MD 0.8 lower (1.52 to 0.08 lower)⊕⊕⊕⊕

HIGH
a

Chaitman 2004[57]: Stratified randomisation. Allocation concealment reported. Double blind. ITT reported. Baseline comparisons made, marginally fewer patients in the placebo group had undergone bypass surgery.

b

95% CI around the pooled estimate of effect includes both: 1) no effect and 2) appreciable benefit or appreciable harm.

c

Ranolazine 750 mg twice plus anti-anginal drugs including atenolol 50 mg (42%% patients), amlodipine 5 mg (31%) and diltiazem 180 mg (26%) vs. placebo plus antianginal drugs including atenolol 50 mg (44%), Amlodipine 5 mg (30%), Diltiazem 180 mg (26%)

Table 10.16Ranolazine (750 mg bid) + antianginal treatment vs. placebo+antianginal treatment – Subgroup diabetes (follow-up 12 weeks)

Quality assessmentSummary of findings
No of patientsEffectQuality
No of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerationsRanolazine (750 mg bid) + antianginal treatmentPlacebo+antianginal treatment - diabetic patientsRelative (95% CI)Absolute
Exercise duration sec (trough change from baseline) - 12 wks (follow-up 12 weeks; better indicated by higher values)
Timmis 200660 (CARISA) (c)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (b)none685d7-MD 28.7 higher (50.9 lower to 108.3 higher)⊕⊕⊕○

MODERATE
Time to onset of angina sec (trough change from baseline) - 12 wks (follow-up 12 weeks; better indicated by higher values)
Timmis 200660 (CARISA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (b)none6857-MD 50.8 higher (37.56 lower to 139.16 higher)⊕⊕⊕○

MODERATE
Angina episodes per week - 12 wks (follow-up 12 weeks; better indicated by lower values)
Timmis 200660 (CARISA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (b)none6857-MD 0.91 lower (3.25 lower to 1.43 higher)⊕⊕⊕○

MODERATE
Nitroglycerin consumption per week - 12 wks (follow-up 12 weeks; better indicated by lower values)
Timmis 200660 (CARISA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (b)none6857-MD 2.32 lower (7.18 lower to 2.54 higher)⊕⊕⊕○

MODERATE
a

Timmis 2006[58]: Randomised. Allocation concealment reported. ITT reported.

b

95% CI includes no effect and the upper and lower CI crosses the MID.

c

Ranolazine 750 mg twice plus anti-anginal drugs including atenolol 50 mg (42% patients), amlodipine 5 mg (31%) and diltiazem 180 mg (26%) vs. placebo plus antianginal drugs including atenolol 50 mg (44%), Amlodipine 5 mg (30%), Diltiazem 180 mg (26%)

Sub-group interaction between diabetic and non-diabetic patients: There was no significant treatment by subgroup interaction for exercise duration (p=0.89) and time to onset of angina (p=0.54) between diabetic and non diabetic patients. Statistical tests for interaction between diabetes status and treatment effect showed no evidence that the effects of ranolazine differed between diabetic and non-diabetic patients either in the number of angina episodes per week (p=0.81) or nitroglycerin usage (p=0.063); and therefore no evidence that the treatment effect differed between diabetic and non-diabetic patients.

Table 10.17Ranolazine (1000 mg bid) + antianginal treatment vs. placebo +antianginal treatment – Subgroup age (follow-up 6 weeks)

Quality assessmentSummary of findings
No of patientsEffectQuality
No of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerationsRanolazine (1000 mg bid) + antianginal treatmentPlacebo +antianginal treatment- ageRelative (95% CI)Absolute
Weekly angina attacks < 70 yrs (follow-up 6 weeks; better indicated by lower values)
Rich 200761 (CARISA, ERICA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionnone403409-MD 0.5 lower (1.1 lower to 0.1 higher)⊕⊕⊕⊕

HIGH
Weekly angina attacks > 71 yrs (follow-up 6 weeks; better indicated by lower values)
Rich 200761 (CARISA, ERICA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionnone135130-MD 1.13 lower (2.05 to 0.21 lower)⊕⊕⊕⊕

HIGH
Nitroglycerin consumption < 70 yrs (follow-up 6 weeks; better indicated by lower values)
Rich 200761 (CARISA, ERICA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionnone403409-MD 0.97 lower (1.64 to 0.3 lower)⊕⊕⊕⊕

HIGH
Nitroglycerin consumption > 71 yrs (follow-up 6 weeks; better indicated by lower values)
Rich 200761 (CARISA, ERICA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionnone135130-MD 0.94 lower (1.74 to 0.14 lower)⊕⊕⊕⊕

HIGH
Adverse events <70 years (follow-up 6 weeks) (c)
Rich 200761 (CARISA, ERICA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionnone194/604 (32.1%)131/420 (31.2%)RR 1.03 [0.86, 1.24]9 more per 1000 (from 44 fewer to75 more)⊕⊕⊕⊕

HIGH
Adverse events > 70 years (follow-up 6 weeks) (c)
Rich 200761 (CARISA, ERICA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (b)none102/231 (44.2%)43/132 (32.6%)RR 1.36 [1.02, 1.80]117 more per 1000 (from7 more to 261 more)⊕⊕⊕○

MODERATE
a

Rich 2007[59]: Randomised. Allocation concealment reported. Double blind. ITT reported. Baseline comparisons made between the 2 age groups, history of MI was more common among younger patients and older patients were somewhat more likely to have diabetes. Systolic blood pressure was slightly higher and diastolic blood pressure was slightly lower in older patients and there was a trend of having more women in the older subgroup.

b

95% CI around the pooled estimate of effect includes appreciable benefit or appreciable harm.

c

Adverse events- cardiac adverse events, constipation, nausea, dyspepsia, dizziness, headache, peripheral edema asthenia, serious adverse events such as MI, syncope,, transient ischemic attack. The most common events resulting in discontinuation of study drug were related to the gastrointestinal, nervous, and cardiac organ systems.

Table 10.18Ranolazine (1000 mg bid) plus amlodipine (10 mg) vs. amlodipine (10mg) (follow-up 6 weeks)

Quality assessmentSummary of findings
No of patientsEffectQuality
No of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerationsRanolazine (1000 mg bid) plus amlodipine (10mg)amlodipine (10 mg)Relative (95% CI)Absolute
Adverse events (follow-up 6 weeks)
Stone 200662 (ERICA) (b)(c)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessserious (b)none112/281 (39.9%)100/284 (35.2%)RR 1.13 (0.91 to 1.4)46 more per 1000 (from 32 fewer to 141 more)⊕⊕⊕○

MODERATE
Weekly angina frequency - (follow-up 6 weeks; better indicated by lower values)
Stone 200662 (ERICA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionnone277281-MD 0.43 lower (1 lower to 0.14 higher)⊕⊕⊕⊕

HIGH
Weekly nitroglycerin consumption - (follow-up 6 weeks; better indicated by lower values)
Stone 200662 (ERICA)randomised trialsno serious limitations (a)no serious inconsistencyno serious indirectnessno serious imprecisionnone277281-MD 0.65 lower (1.23 to 0.07 lower)⊕⊕⊕⊕

HIGH
a

Stone 2006[60]: Randomised. Randomisation wascentralised and not stratified by centre. Allocation concealment reported. Double blind. Blinding of outcome assessors reported. Ranolazine: 7/281 (2%); Placebo: 6/284 (2%) lost to follow-up. ITT reported. 95% CI around the pooled estimate of effect includes both: 1) no effect and 2) appreciable benefit or appreciable harm.

b

Ranolazine 1000 mg twice daily plus amlodipine 10 mg/daily vs. amlodipine 10 mg/daily

c

Per protocol, the patients in this study were not taking BB, and hence the authors state that this data may be especially applicable to the proportions of patients who cannot tolerate BB therapy.

10.4.3. Economic evidence

No economic studies were identified on this question. We calculated the daily and annual cost of ranolazine treatment based on the unit cost reported in the BNF5919.

Table 10.19Drug cost of Ranolazine

Cost per day (£)Cost per year (£)
Ranolazine, 375 mg, 500 mg or 750 mg twice daily1.63595

The costs of adverse effects were not estimated.

10.4.4. Evidence statements

ClinicalClinical Efficacy

Ranolazine plus antianginal treatment versus placebo plus antianginal treatment

Chaitman 200459 (CARISA): Evidence from one RCT shows that exercise duration at trough (sec) [MD 23.70 (1.11 to 46.29)], time to onset of angina at trough (sec) [MD 29.70 (4.62 to 54.76)], exercise duration at peak (sec) [MD 34 (11.96 to 56.04)] and time to onset of angina at peak(sec) [MD 38 (13.91 to 62.09)] were significantly higher in the ranolazine plus antianginal treatment compared with placebo plus antianginal treatment [follow-up 12 weeks]. Angina attacks per week was significantly lower in the ranolazine plus antianginal treatment compared with placebo plus antianginal treatment [MD 0.8 lower (1.52 to 0.08 lower)]. There were no statistically significant differences between ranolazine plus antianginal treatment and placebo plus antianginal treatment for the outcome of adverse events [RR 1.11 (0.85 to 1.46)] (follow up 12 weeks).

Timmis 200660 (CARISA): Evidence from a post-hoc sub-group analyses of one RCT shows that there were no statistically significant differences in the outcomes of exercise duration (sec) [MD 28.70 (−50.90 to 108.30)], time to onset of angina (sec) [MD 50.80 (−37.56 to 139.16)], frequency of angina attacks [MD −0.91 (−3.25 to 1.43) and nitroglycerin consumption [MD −2.32 (−7.18 to 2.54)] between ranolazine plus anti anginal treatment and placebo plus anti-anginal treatment in people with diabetes (follow-up 12 weeks).

Rich 200761 (CARISA): Evidence from one post-hoc sub-group analysis of a RCT shows that in patients younger than 70 years ranolazine plus anti anginal treatment resulted in a statistically significant reduction in nitroglycerine consumption [MD −0.97 (−1.64 to −0.30)] but no significant difference in weekly angina attacks [MD −0.50 (−1.10 to 0.10)] or adverse events [RR1.03 [0.86, 1.24] when compared with placebo plus anti-anginal treatment [follow-up 6 weeks]. In patients older than 70 years ranolazine plus anti anginal treatment resulted in statistically significantly reductions in weekly angina attacks [MD −1.13 (−2.05 to −0.21)] and nitroglycerin consumption [MD −0.94 (−1.74 to −0.14) but a statistically significant increase in adverse events [RR 1.36 [1.02, 1.80] when compared with placebo plus anti-anginal treatment (follow-up six weeks).

Ranolazine plus amlodipine versus amlodipine

Stone 200662 (ERICA): Evidence from one RCT shows that weekly nitroglycerin consumption was significantly lower with ranolazine plus amlodipine compared to amlodipine alone [MD −0.65 (−1.23 to −0.07)]. There were no statistically significant differences between ranolazine plus amlodipine and amlodipine for weekly angina frequency [MD−0.43 (1.00 to 0.14)] and adverse events (RR 1.13 (0.91 to 1.40) (follow-up 6 weeks).
EconomicNo economic evidence was found on this question. A simple cost analysis showed a significant drug cost of ranolazine.

10.4.5. Recommendations and link to evidence

RecommendationIf the person cannot tolerate beta blockers and calcium channel blockers or both are contraindicated, consider monotherapy with one of the following drugs*: Decide which drug to use based on comorbidities, contraindications, the person’s preference and drug costs.

For people on beta blocker or calcium channel blocker monotherapy whose symptoms are not controlled and the other option (calcium channel blocker or beta blocker) is contraindicated or not tolerated, consider one of the following as an additional drug*: Decide which drug to use based on comorbidities, contraindications, the person’s preference and drug costs.

* Evidence on long-acting nitrates is presented in chapter 9. Evidence on ivabradine and nicorandil is presented in sections 10.2 and 10.3 respectively of this chapter.

** When combining ivabradine with a calcium channel blocker, use a dihydropyridine calcium channel blocker, for example, slow release nifedipine, amlodipine, or felodipine.

*** At the time of publication (July 2011), nicorandil did not have UK marketing authorisation for use in this indication. Informed consent should be obtained and documented.
Relative values of different outcomesOutcomes of interest included long-term mortality (total and cardiovascular), rates of major adverse cardiovascular events (myocardial infarction, stroke, myocardial revascularisation), and measures of symptom severity (frequency of angina, exercise test outcomes).
Trade off between clinical benefits and harmsWe found no evidence on the effects of ranolazine monotherapy or ranolazine in combination with other anti- anginal drugs on long-term outcome in people with stable angina.

In one randomised trial addition of ranolazine to standard anti-anginal treatment for twelve weeks increased exercise duration (by 20 to 30 seconds) and time to angina at trough (and at peak). Ranolazine reduced the frequency of angina attacks and nitroglycerine use by about one per week. These effects were consistent in people with diabetes and in people aged over 70 years.

In one randomised trial addition of ranolazine to amlodipine reduced nitroglycerine consumption (by 0.65 doses per week) but not weekly angina frequency after six weeks follow-up.

Addition of ranolazine to amlodipine did not increase the risk of adverse events when compared with amlodipine alone. There was also no difference in risk of adverse events between ranolazine plus antianginal treatment versus placebo plus antianginal treatment for the entire group, but addition of ranolazine to antianginal treatment was associated with an increased risk of adverse events in a sub group of patients older than 70 years of age.
Economic considerationsNo economic evidence on the use of ranolazine for the treatment of stable angina was available for review. The cost of ranolazine is substantially higher than the costs of first-line anti-anginal drugs (BBs and CCBs) and long-acting nitrates.
Quality of evidenceThe available evidence for ranolazine was of moderate or high quality but were not designed to assess the long-term effects of ranolazine on mortality or other major adverse cardiac events. The improvements in exercise time and symptom severity associated with short-term ranolazine treatment are modest and of uncertain clinical significance.

No economic evidence was available on this question.
Other considerationsEvidence to support the long-term use of ranolazine as adjunctive anti-anginal therapy is very limited. The GDG concluded that there is insufficient evidence to recommend routine use of ranolazine, but ranolazine may have a role in people with stable angina who are inadequately controlled or intolerant of first-line anti-anginal therapies.

The cost of ranolazine is comparable with the costs of ivabradine but more than the cost of long-acting nitrate. Ranolazine has a licence for use in combination treatment. Nevertheless the GDG concluded that there was insufficient evidence to make a firm recommendation about the choice of an additional anti-anginal drug if a BB and/or CCB are not tolerated or are contraindicated.

10.5. General drug recommendations

RecommendationOffer people optimal drug treatment for the initial management of stable angina. Optimal drug treatment consists of one or two anti-anginal drugs as necessary plus drugs for secondary prevention of cardiovascular disease.
Other considerationsThe evidence reviews indicated benefit from secondary prevention treatment and anti-anginal treatment. The GDG considered it important to emphasise the importance for patients to receive optimal medical treatment and made a consensus recommendation for this.
RecommendationAdvise people that the aim of anti-anginal drug treatment is to prevent episodes of angina and the aim of secondary prevention treatment is to prevent cardiovascular events such as heart attack and stroke.
Other considerationsThe GDG were aware of the importance of patient adherence to secondary prevention treatment. They also considered it important that patients understand that the purpose of anti- angnal drugs is to improve symptoms. The GDG made a consensus recommendation to ensure that professionals explain these points adequately to patients.
RecommendationReview the person’s response to treatment, including any side effects, 2–4 weeks after starting or changing drug treatment.

Titrate the drug dosage against the person’s symptoms up to the maximum tolerable dosage.

Discuss how side effects of drug treatment might affect the person’s daily activities and explain why it is important to take drug treatment regularly.
Other considerationsThe GDG debated the need to review the response to treatment after starting or changing any anti-anginal medication. The GDG reached a consensus that response to treatment, including any side effects, should be reviewed 2–4 weeks after starting or changing any anti-anginal drug. If the person’s angina is not controlled, the dose of the anti-anginal drug should be titrated up to the maximum tolerable dose (within the licensed dose range) with the objective of achieving control of symptoms of stable angina. The GDG also considered it important that patients do not remain on drugs that are not providing benefit to them and healthcare professionals should stop anti-anginal drugs that are not providing symptomatic benefit.
RecommendationPatients differ in the type and amount of information they need and want. Therefore the provision of information should be individualised and is likely to include, but not be limited to:
  • what the medicine is
  • how the medicine is likely to affect their condition (that is, its benefits)
  • likely or significant adverse effects and what to do if they think they are experiencing them
  • how to use the medicine
  • what to do if they miss a dose
  • whether further courses of the medicine will be needed after the first prescription
  • how to get further supplies of medicines. [This recommendation is from ‘Medicines Adherence’ (NICE clinical guideline 76).]
Other considerationsThe GDG agreed to include this recommendation from Medicine Adherence Clinical Guideline 76 to ensure people with stable angina are given adequate information about the drugs they are prescribed.

10.6. Research recommendation

The GDG recommended the following research question:

[arrowhead]

Research question: What is the clinical and cost effectiveness of adding a newer anti-anginal drug (nicorandil, ivabradine or ranolazine) to a calcium channel blocker for treating stable angina?

[arrowhead]

Why this is important: We do not know the clinical and cost effectiveness of adding a newer anti-anginal drug to a calcium channel blocker in people with stable angina. We propose a double-blind placebo-controlled randomised trial comparing the addition of a newer anti-anginal drug to a calcium channel blocker with a calcium channel blocker alone in people with stable angina whose symptoms are not being controlled. Endpoints would include symptom severity, quality of life, long-term morbidity and mortality, and cost effectiveness. The results of the trial would influence clinical practice and inform future updates of key recommendations in this guideline.

Borer 2003[44]: Randomisation, allocation concealment, double blinding and ITT reported. Baseline comparisons made, there was no clinically relevant differences in baseline characteristics were observed between the 2 groups. In Ivabradine 2.5 mg bd 3/90 (3.3%); Ivabradine 5 mg bd 4/91 (4.4%); Ivabradine 10 mg bd 3/88 (3.4%) and; Placebo 1/91 (1.1%) were lost to follow-up.

The upper and lower CI crosses the MID.

Fox 2009[45] (BEAUTIFUL): Randomisation, allocation concealment, double blinding and ITT reported. This is a post hoc analysis, and therefore the results should be considered as hypothesis generating. Sample size calculation reported.

95% CI around the pooled estimate of effect includes both: 1) no effect and 2) appreciable benefit or appreciable harm.

Ivabradine 5 mg bid vs. placebo

Ivabradine 7.5 mg bid vs. placebo

Trough = 12 hours after administration of ivabradine

Peak = 4 hours after administration of ivabradine

In this post hoc analysis, the BEAUTIFUL population was divided according to the presence of limiting angina symptoms at baseline using the New York Heart Association (NYHA) functional classification. Patients were questioned at the inclusion visit regarding the presence of symptoms limiting activity, and whether they were related to anginal pain or due to presence of heart failure (fatigue, palpitations or dyspnoea).

Limiting angina symptoms were identified in 13.8% of the BEAUTIFUL population at baseline (1507 out of 10917 patients). Of these, 734 were randomised to ivabradine treatment and 773 to placebo.

Tardif 2005[46]: Allocation concealment not reported. Randomisation using permutation blocks. Double blinding. Capsules of ivabradine and placebo identical. Patients, investigators, central readers of ETT data were blinded to the treatment received by the patients. Calculation of sample size reported. Baseline comparisons made no difference between the study groups. ITT reported.

95% CI includes no effect and the upper and lower CI crosses the MID.

95% CI around the pooled estimate of effect includes both: 1) no effect and 2) appreciable benefit or appreciable harm.

Ivabradine 5 mg bid for 4 weeks and then 7.5 bid for 12 weeks or atenolol 50 mg od for 4 weeks and then 100 mg od for 12 weeks.

The authors state that the study does not allow to strictly comparing the safety of ivabradine and atenolol, because about two-thirds of the patients had previously received BB and were known to tolerate these drugs; patients with known intolerance or contraindications to atenolol were specifically excluded. There was slightly higher number of deaths in the ivabradine groups (2 (0.6%) and 3 (1%) respectively) than in the atenolol group (1 (0.3%) that was not statistically significant.

Tardif 2009[47]: Randomisation, allocation concealment, double blinding and ITT reported. The random allocation was computer generated. Sample size calculation reported. 2% of patients lost to follow-up. The full analysis included 875 (98% of those randomised).

The upper and lower CI crosses the MID.

95% CI around the pooled estimate of effect includes both: 1) no effect and 2) appreciable benefit or appreciable harm.

Patients receiving atenolol 50 mg/day were randomised to receive ivabradine 5 mg b.i.d for 2 months, increased to 7.5 mg b.i.d for a further 2 months.

12 hours after last dose ivabradine, 24 hours after last dose atenolol

Ruzyllo 2007[48] : Randomisation, allocation concealment, double blinding and ITT reported. N=1195 randomised (Ivabradine 7.5 n=400, Ivabradine 10 mg n=391, or amlodipine n=404). The ITT population consisted of 1155 patients (96.7%). Sample size calculation reported.

95% CI around the pooled estimate of effect includes appreciable benefit or appreciable harm.

Ivabradine 7.5 mg twice daily vs. amlodipine 10 mg once daily

Dargie 2002[49]: Randomisation process was reported; allocation concealment was not reported; study was double blind; Number of drop outs were reported and >20%; Intention to treat analysis was reported; the study was powered for primary outcome (CHD death, non fatal MI, or unplanned hospitalisation). In the placebo group: 2/2561 lost to follow-up, 809/2561 discontinued intervention. In the Nicorandil group: 2/2565 lost to follow-up; 1003/2565 discontinued intervention

95% CI around the pooled estimate of effect includes both 1) no effect 2) appreciable benefit or appreciable harm

95% CI around the pooled estimate of effect includes appreciable benefit or appreciable harm

Canadian Cardiovascular Society Functional classification of angina at the end of the study (follow-up mean 1.6 years):

  • Class I - Nicorandil 985 (43%); placebo 989 (43%)
  • Class II- Nicorandil 1159 (50%); placebo 1124 (49%)
  • Class III- Nicorandil 162 (7%); placebo 163 (7%)
  • Class IV- Nicorandil 9 (<1%); placebo 15 (1%)

Guermonprez 1993[51]: Allocation concealment was not reported; study was double blind; intention to treat analysis was not reported. The two groups were comparable in terms of age, sex, history of myocardial infarction and severity of coronary lesions. Drop-out: Nicorandil: 3 (5%) drop outs due to insufficeint efficacy; 4(6.7%) drop outs due to other adverse events. Diltiazem: 4 (6.3%) drop outs due to insufficient efficacy; 1 (1.6%) drop outs due to other adverse events.

95% CI includes no effect and the upper and the lower confidence limit crosses the MID.

95% CI around the pooled estimate of effect includes both 1) no effect 2) appreciable benefit or appreciable harm

Mean value reported for both groups together. No standard deviation (SD) reported

Chatterjee 1999[52]: Randomised, double blind, parallel group design. Allocation concealment was not reported. The study reports of 6/121 patients drop out [5 because of adverse events (nicorandil n=3 and amlodipine n=2) and one due to compliance problems (amlodipine)]. The study reports that 118/121 were evlauted for efficacy on an ITT basis. The treatment groups were comparable for demographic and clinical characteristics within and between the two countries (Austria and Switzerland), with the exception of history of previous MI among patients recruited in Austria (nicorandil, n=2; amlodipine, n=14). The mean number of anginal attacks/week was similar in both nicorandil and amlodipine groups at baseline. However, the mean number of nitroglycerin units required for pain relief was significantly higher (p=0.04) in the nicorandil group (2.3 vs. 1.0 units/week). Baseline BP, HR and ETT variables were similar in the 2 treatment groups.

95% CI includes no effect and the upper and the lower confidence limit crosses the MID.

The upper and the lower confidence limit crosses the MID.

95% CI around the pooled estimate of effect includes both 1) no effect 2) appreciable benefit or appreciable harm

Ulvenstam 1992[53]: Double-blind, randomised, multicentre study. 55/58 completed the study. Allocation concealment not reported. ITT not reported. Baseline comparisons made-previous MI more frequent in the nicorandil group.

95% CI includes no effect and the upper and the lower confidence limit crosses the MID.

Zhu 2007[54]: Randomised, allocation concealment not reported. Double blind. Power calculation used. N=232 patients completed the study (N=115 in nicorandil group and N=117 in the Isosorbide mononitrate (ISMN) group). Drop-out rate: 7% drop out (8% in the nicorandil group and 6% in the ISMN group). Intention to treat analysis was not reported. Baseline comparisons made. No significant difference between the groups.

95% CI includes no effect and the upper and the lower confidence limit crosses the MID.

95% CI around the pooled estimate of effect includes both 1) no effect 2) appreciable benefit or appreciable harm

Meeter 1992[55]: Allocation concealment not reported; double blind; drop-out rate reported and < 20%; Intention to treat analysis not reported. Baseline comparisons made-Nicorandil group had shorter duration of angina at baseline compared to propranolol median 12 vs 20 months. Fewer patients had a history of MI on nicorandil vs propranolol 16/38 vs. 22/39. 5/77 were withdrawn from the trial (4 receiving nicorandil and one receiving propranolol). The patients receiving nicorandil were withdrawn because of worsening of angina complaints or headaches.

95% C around the pooled estimate of effect includes both 1) no effect 2) appreciable benefit or appreciable harm.

The upper and lower limits of 95% CI crosses the MI.

Short follow-up (up to 1.6 years). Sensitivity analysis was quite limited and was applied only to the primary analysis (cost of care after discharge excluded). Morbidity associated with gastro-intestinal events is not included. Effectiveness data were reported only in the incremental analysis.

QALYs were not estimated.

2002 GBP. Costs included were cost of Nicorandil (including 10% dispensing fee and two additional physician visits), adverse events related to Nicorandil, hospital admissions, surgical procedures. The cost of post-discharge care was not included in the base case analysis.

Primary end-points considered in the analysis were cardiac death, non-fatal MI, hospital admission for cardiac chest pain.

Chaitman 2004[57]: Stratified randomisation. Allocation concealment reported. Double blind. ITT reported. Baseline comparisons made, marginally fewer patients in the placebo group had undergone bypass surgery.

95% CI around the pooled estimate of effect includes both: 1) no effect and 2) appreciable benefit or appreciable harm.

Ranolazine 750 mg twice plus anti-anginal drugs including atenolol 50 mg (42%% patients), amlodipine 5 mg (31%) and diltiazem 180 mg (26%) vs. placebo plus antianginal drugs including atenolol 50 mg (44%), Amlodipine 5 mg (30%), Diltiazem 180 mg (26%)

Timmis 2006[58]: Randomised. Allocation concealment reported. ITT reported.

95% CI includes no effect and the upper and lower CI crosses the MID.

Ranolazine 750 mg twice plus anti-anginal drugs including atenolol 50 mg (42% patients), amlodipine 5 mg (31%) and diltiazem 180 mg (26%) vs. placebo plus antianginal drugs including atenolol 50 mg (44%), Amlodipine 5 mg (30%), Diltiazem 180 mg (26%)

Rich 2007[59]: Randomised. Allocation concealment reported. Double blind. ITT reported. Baseline comparisons made between the 2 age groups, history of MI was more common among younger patients and older patients were somewhat more likely to have diabetes. Systolic blood pressure was slightly higher and diastolic blood pressure was slightly lower in older patients and there was a trend of having more women in the older subgroup.

95% CI around the pooled estimate of effect includes appreciable benefit or appreciable harm.

Adverse events- cardiac adverse events, constipation, nausea, dyspepsia, dizziness, headache, peripheral edema asthenia, serious adverse events such as MI, syncope,, transient ischemic attack. The most common events resulting in discontinuation of study drug were related to the gastrointestinal, nervous, and cardiac organ systems.

Stone 2006[60]: Randomised. Randomisation wascentralised and not stratified by centre. Allocation concealment reported. Double blind. Blinding of outcome assessors reported. Ranolazine: 7/281 (2%); Placebo: 6/284 (2%) lost to follow-up. ITT reported. 95% CI around the pooled estimate of effect includes both: 1) no effect and 2) appreciable benefit or appreciable harm.

Ranolazine 1000 mg twice daily plus amlodipine 10 mg/daily vs. amlodipine 10 mg/daily

Per protocol, the patients in this study were not taking BB, and hence the authors state that this data may be especially applicable to the proportions of patients who cannot tolerate BB therapy.

Copyright © 2011, National Clinical Guidelines Centre.

Apart from any fair dealing for the purposes of research or private study, criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, no part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior written permission of the publisher or, in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK. Enquiries concerning reproduction outside the terms stated here should be sent to the publisher at the UK address printed on this page.

The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore for general use.

The rights of the National Clinical Guidelines Centre to be identified as Author of this work have been asserted by them in accordance with the Copyright, Designs and Patents Act, 1988.

Cover of Stable Angina
Stable Angina: Methods, Evidence & Guidance [Internet].
NICE Clinical Guidelines, No. 126.
National Clinical Guidelines Centre (UK).

PubMed Health Blog...

read all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...