ATLAS - Rivaroxaban in patients with a recent ACS

ATLAS ACS 2-TIMI 51: Mega JL, et al. Rivaroxaban in patients with a recent acute coronary syndrome. NEJM 2012;366:9-19.

Bottom line:

• Over ~13 months of follow-up, the addition of rivaroxaban 2.5 mg BID to ASA+clopidogrel/ticlopidine reduced the risk of CV events, mainly driven by fatal events, which led to lower all-cause mortality (NNT 63) in patients post-ACS. Adding rivaroxaban also reduced stent thrombosis (NNT 143). This came at the cost of a greater risk of major bleeds (NNH 84), including intracranial hemorrhages (NNH 500).

• Adding rivaroxaban 5 mg BID did not reduce mortality, & further increased the risk of major bleeds.

• It remains unclear which of the following regimens would have the best balance between efficacy & safety: Ticagrelor-based DAPT, clopidogrel-based DAPT + low-dose rivaroxaban, or ticagrelor-based DAPT + low-dose rivaroxaban.

Patients (n=15,526)

• Included
  • ACS (within 7 days of admission, after revascularization if planned)
  • Plus either
    • Age >55 y
    • Previous MI
    • Diabetes
  • Excluded
    • Previous intracranial hemorrhage
    • Clinically-significant GI bleed within 12 months
    • CrCl <30 mL/min
    • Hb <100 g/L
    • Platelet count <90
• Baseline characteristics
  • Age 61.5 y (9.6% 75 y or greater)
  • Female 25%
  • STEMI (~50%), NSTEMI (~25%), UA (~25%)
  • Previous MI 27%
  • CV risk factors: HTN ~67%, diabetes ~32%, dyslipidemia ~50% 
  • CrCl 86 mL/min
  • Meds
    • ASA 99%, P2Y12 inhibitor ~93%
    • ACEI/ARB ~40%
    • Beta-blocker ~66%
    • Statin ~15%

Interventions

• Intervention1: Rivaroxaban 2.5 mg BID
• Intervention2: Rivaroxaban 5 mg BID
• Control: Placebo
• All groups: ASA + clopidogrel or ticlopidine

Results @ 13.1 months

View fullsize

Considerations

• Low risk of bias (allocation, performance, detection & attrition)
  • Allocation concealed by central computer/phone allocation
  • Participants, clinicians & investigators blinded to study drug
  • Modified ITT & full ITT analysis
  • 0.3% lost to follow-up, though ~28% in all groups discontinued study drug before end of study
• Generalizability
  • Patients represented a group of post-ACS patients at very high risk of recurrent ASCVD with very low use of proven secondary prevention therapies
  • Additionally, this trial was performed before approval of more potent P2Y12 inhibitors (prasugrel & ticagrelor), so background DAPT included ASA + clopidogrel/ticlopidine (how many specifically received clopidogrel not reported)
  • Based on the preliminary results from ATLAS ACS-TIMI 46 (see below), the investigators selected 2.5 and 5 mg BID doses of rivaroxaban for this trial; this is pharmacokinetically rational given the 5-13h half-life of rivaroxaban (apixaban & dabigatran have similar half-lives and are generally dosed BID), however, it is a very different regimen than that used for AF/VTE. As a result, it's unclear if this regimen preserves the efficacy of rivaroxaban for these other conditions.

Summary of ATLAS ACS-TIMI 46

• Dose-ranging study with identical enrolment criteria as ATLAS ACS 2-TIMI 51

View fullsize

RE-DUAL PCI - Dabigatran-based dual antithrombotic regimen in patients with AF after PCI

Cannon CP, et al. Dual antithrombotic therapy with dabigatran after PCI in atrial fibrillation. NEJM 2017

Bottom line:

• RE-DUAL PCI provides further evidence supporting dual therapy with an oral anticoagulant + P2Y12 inhibitor in patients with AF post-PCI instead of triple therapy.

• Dual therapy with dabigatran 150 mg BID reduced clinically relevant bleeds (NNT 19) as well as major bleeds (NNT 36-56 depending on definition), & was non-inferior in terms of thromboembolic events over approximately 1 year.

• Dual therapy with dabigatran 110 mg BID showed a possible increase in death, MI & definite stent thrombosis, & was potentially inferior for the composite efficacy outcome. Despite reducing major & clinically-relevant bleeds, this regimen can't be recommended due to inadequate evidence of efficacy in this setting.

  • Notably, the PIONEER trial did not assess for non-inferiority of rivaroxaban-based dual therapy regimens to triple therapy, but if it had, it would not have met the non-inferiority criteria from RE-DUAL PCI. However, the PIONEER results did not show the concerning numerical trends seen here with dabigatran 110 mg BID.

Patients (n=2725)

• Included

  • Non-valvular AFib (paroxysmal, persistent or permanent)

  • PCI (bare-metal or drug-eluting stent) within 120h for stable CAD or ACS

• Exclusion

  • Bioprosthetic or mechanical heart valve

  • CrCl <30 mL/min

  • "Other major coexisting conditions"

• Baseline characteristics

  • ~72 y

  • Female ~24%

  • AF characteristics

    • Paroxysmal (~50%), persistent (~17%), permanent (33%)

    • CHA2DS2-VASc score ~4, HAS-BLED score ~3

    • Previous stroke 10%

  • PCI characteristics

    • Previous: MI ~25%, PCI ~33%, CABG ~10%

    • Indication for PCI: ACS (~50%), stable angina/+ stress test (44%), other (~6%)

    • Drug eluting stent 85%

Interventions

• Intervention1: Dabigatran 150 mg BID + clopidogrel/ticagrelor

  • Note: Elderly outside the US not eligible for this group due to dabigatran labeling

• Intervention2: Dabigatran 110 mg BID + clopidogrel/ticagrelor

• Control ("triple therapy"): Warfarin (INR 2-3) + clopidogrel/ticagrelor + low-dose ASA

  • ASA D/Ced after 1 month with bare-metal stent or 3 months with drug-eluting stent

  • Time in therapeutic INR: 64%

• In all groups

  • P2Y12 inhibitor was continued for at least 12 months

  • P2Y12 inhibitor chosen: Clopidogrel 88%, ticagrelor 12%

  • Mean duration of trial anticoagulant: 12.3 months

Results @ mean 14 months

Dabigatran 150 mg BID vs triple therapy (control group excludes elderly outside US not eligible for higher dabigatran dose)

• Primary outcome (major or clinically relevant non-major bleed, ISTH definition):

  • Dabi150 20.2%, control 25.7%, NNT 19

  • HR 0.72 (0.58-0.88), p<0.001 for non-inferiority

• Major bleed:

  • ISTH definition: Dabi150 5.6%, control 8.4%, HR 0.64 (0.43-0.94), NNT 36

  • TIMI definition: Dabi150 2.1%, control 3.9%, HR 0.51 (0.28-0.93), NNT 56

  • Intracranial hemorrhage: Dabi150 0.1%, control 1.0%, p=0.047

• Composite efficacy outcome (death, MI, stroke, or systemic embolism, or unplanned PCI/CABG): Dabi150 11.8%, control 12.8%, HR 0.89 (0.67-1.19)

Dabigatran 110 mg BID vs triple therapy

• Primary outcome:

  • Dabi110 15.4%, control 26.9%, NNT 9

  • HR 0.52 (0.42-0.63), p<0.001 for non-inferiority

• Major bleed:

  • ISTH definition: Dabi110 5.0%, control 9.2%, HR 0.52 (0.37-0.74), NNT 24

  • TIMI definition: Dabi110 1.4%, control 3.8%, HR 0.37 (0.20-0.68), NNT 42

  • Intracranial hemorrhage: Dabi110 0.3%, control 1.0%, p=0.06

• Composite efficacy outcome: Dabi110 15.2%, control 13.4%, HR 1.13 (0.90-1.43) - did not meet non-inferiority criteria

  • Thromboembolic events or death: Dabi110 11.0%, control 8.5%, HR 1.30 (0.98-1.73)

    • Death Dabi110 5.6%, control 4.9%

    • MI: Dabi110 4.5%, control 3.0%

    • Definite stent thrombosis: Dabi110 1.5%, control 0.8%

Considerations

• Low to unclear risk of bias

  • Unclear randomization & allocation concealment (not adequately reported)

  • Open-label design - low risk of performance bias, but high risk of detection bias for softer outcomes (ie clinically significant non-major bleeds)

  • Low risk of attrition bias (ITT analysis that included all randomized patients regardless of receipt of study intervention; 0.2% lost to follow-up, <4% withdrew consent with no vital status available at end of study)

• Non-inferiority trial

  • Non-inferiority margin 1.38 for HR upper end of 95% confidence interval for both efficacy & safety outcomes

  • Primary analysis using ITT population with sensitivity on-treatment analysis

  • Dual therapy with dabi150 met non-inferiority for both bleeding & thromboembolic outcomes, but dabi110 only met non-inferiority criteria for bleeding

• Excellent generalizability due to broad eligibility criteria & enrolment of a representative, relatively elderly population

• Unadjusted bleed rates from a subgroup analysis by choice of P2Y12 inhibitor (clopidogrel vs ticagrelor) suggest that risk of bleeding increases gradually with number of antithrombotic agents as well as potency of the inhibitor, e.g. ISTH major bleed over mean 14 months from Figure 2:

  • Clopidogrel + dabigatran: ~5%

  • Clopidogrel + ASA + warfarin OR ticagrelor + dabigatran: ~8%

  • Ticagrelor + ASA + warfarin: ~16.5%

CANTOS - Canakinumab for patients with previous MI

Ridker PM, et al. Antiinflammatory therapy with canakinumab for atherosclerotic disease. NEJM 2017

Bottom line:

• In patients with prior MI & hsCRP >2 mg/L, canakinumab reduced the risk of non-fatal coronary events (NNT ~160/year), but increased the risk of fatal infections (NNH ~1000/year).

• Due to unresolved issues of cost & feasibility of use of this therapy, CANTOS has limited direct applicability to real-world practice, but serves as a proof of concept for anti-inflammatory therapy to reduce the risk of ASCVD.

Patients (n=10,061 after 17,482 were screened)

• Included
  • Hx of MI
  • High-sensitivity CRP 2 mg/L or higher
• Exclusion
  • Hx of chronic/recurrent infection
  • Previous CA
  • Known/suspected immunocompromised
  • Hx/high risk of TB or HIV-related disease
  • Ongoing use of other systemic anti-inflammatory tx
• Baseline characteristics
  • 61 y
  • Female 26%
  • STEMI 54%, NSTEMI 34%, unknown 12%
  • Previous PCI 66%, CABG 14%
  • HF 22%
  • CV risk factors: Smoker 23%, HTN ~80%, diabetes ~40%
  • Labs
    • Median hsCRP 4.2 mg/L
    • LDL-C 2.1 mmol/L
  • Meds
    • ACEI/ARB ~80%
    • Anti-ischemic therapy 92%
    • Statin ~90%

Interventions

• Interventions: Canakinumab 50 mg, 150 mg or 300 mg subcutaneously q3 months
  • 300 mg dose group: 300 mg q2 weeks x2 doses, then q3 months
• Control: Matching placebo
   

Results @ median 3.7 years

Labs

• hsCRP reduction: 50 mg (26%), 150 mg (37%), 300 mg (41%)
• LDL, HDL: No effect

Efficacy

• Primary outcome (CV death, MI or stroke):
  • Canakinumab 150 mg: 14.0% over follow-up (incidence rate: 3.86 per 100 patient-years)
  • Placebo: 16.0% over follow-up (incidence rate: 4.50 per 100 patient-years)
  • HR 0.85 (0.74-0.98), NNT 50 (NNT ~160/y)
• Secondary efficacy outcome (primary + unstable angina hospitalization leading to unplanned revascularization):
  • Canakinumab 150 mg: 4.29 per 100 patient-years
  • Placebo: 5.13 per 100 patient-years
  • HR 0.83 (0.73-0.95)
• Death: Canakinumab 150 mg 2.73/100 pt-y, placebo 2.97/100 pt-y, HR 0.92 (0.78-1.09)
• MI: Canakinumab 150 mg 1.90/100 pt-y, placebo 2.43/100 pt-y, HR 0.76 (0.62-0.92)
• Stroke: HR 0.98 (0.71-1.35)
• Revascularization: Canakinumab 150 2.49/100 pt-y, placebo 3.61/100 pt-y, HR 0.68 (0.58-0.81)
• Note: For brevity, I only include the efficacy data for the 150-mg dose here (50 mg generally ineffective/less effective, & 300 mg similar to 150 mg)

Safety/tolerance

• Discontinued study drug: Canakinumab 18.7%, placebo 18.1%
• Serious adverse events: Canakinumab 11.8/100 pt-y, placebo 12/100 pt-y, p=0.79
• Serious adverse event from infection: Canakinumab 3.1/100 pt-y, placebo 2.9/100 pt-y, p=0.14
• Fatal infection or sepsis: Canakinumab 0.3/1000 pt-y, placebo 0.2/100 pt-y, p=0.02

Considerations

• Low risk of bias
  • Central, computerized randomization (allocation concealed)
  • Participants, clinicians & investigators blinded
  • 0.3% lost to follow-up
• Generalizability
  • Eligibility criteria for this trial are broad (any prior MI + hsCRP >2 mg/L) & primarily excluded patients at high risk of adverse effects of immunosuppressive therapy, though enrolled patients had a relatively high risk of ASCVD (~4%/year in the placebo group)
• Currently, canakinumab is approved under the organ drug status for rare diseases, & is priced at $200,000/year (US price in USD)

Heparin, enoxaparin & fondaparinux in ACS

NSTE-ACS

Heparin: Systematic review of 6 RCTs (3 were double-blind) published 1988-1995

• P: NSTE-ACS patients treated with ASA (n=1353)
• I: Heparin bolus + infusion x2-7 days
• C: Placebo/no treatment
• O:
  • Death/MI: Heparin 7.9%, control 10.4% (NNT 40), OR 0.67 (0.45-0.99)
  • Recurrent angina: OR 0.94 (0.58-1.54)
  • Revascularization: OR 1.25 (0.76-2.06)
  • Major bleed: OR 1.88 (0.60-5.87)

Enoxaparin (A to Z, ESSENCE, SYNERGY, TIMI 11B)

View fullsize

Fondaparinux: OASIS 5 - Double-blind non-inferiority RCT

• P: NSTE-ACS (n=20,078)
  • Clopidogrel given to 2/3
• I: Fondaparinux 2.5 mg  once daily up to 8 days or until discharge
• C: Enoxaparin 1 mg/kg q12h (if CrCl <30: 1 mg/kg q24h) x2-8 days or until "clinically stable"
• O @ day 9:
  • Primary outcome (death/MI/refractory ischemia): Fondaparinux 5.8%, enoxaparin 5.7%, HR 1.01 (0.90-1.13)
    • Death: 1.8% vs 1.9%
    • MI: 2.6% vs 2.7%
    • Refractory ischemia: 1.9% for both
  • Major bleed: 2.2% vs 4.1% (NNT 53), HR 0.52 (0.44-0.61)
• O @ 180 days:
  • Primary outcome: 12.3% vs 13.2% (NNT 112), HR 0.93 (0.86-1.00)
    • Death: 5.8% vs 6.5%
  • Major bleed: 4.3% vs 5.8%, HR 0.72 (0.64-0.82)
• Benefit & safety preserved in 1/3 of patients who underwent PCI during initial ACS hospitalization (in fondaparinux group, heparin given during PCI)

 

STEMI

Enoxaparin: (ASSENT 3, EXTRACT TIMI 25)

• ASSENT 3
  • P: STEMI undergoing fibrinolysis with TNK (n6095)
  • I: Enoxaparin 30 mg IV bolus, then 1 mg/kg subcutaneously q12h continued for max 7 days
  • C: Heparin bolus, then infusion for at least 48h (also 3rd group given abciximab & no anticoagulant)
  • O @ 30 days:
    • Primary outcome (death, in-hospital re-infarction or refractory ischemia): Enoxaparin 11.4%, heparin 15.4% (NNT 25)
      • Death: 5.4% vs 6.0% (p=0.25)
      • Re-infarction: 2.7% vs 4.2%
      • Refractory ischemia: 4.6% vs 6.5%
    • ICH: 0.9% in both groups
    • Major bleed, non-ICH: 3.0% vs 2.2% (NNH 125)
• EXTRACT TIMI 25
  • P: STEMI undergoing fibrinolysis (n=20,506)
  • I: Enoxaparin 30 mg IV bolus, then 1 mg/kg subcutaneously q12h continued for max 8 days or until discharge
    • If >75 y/o: No bolus, 0.75 mg/kg q12h
    • If CrCl <30 mL/min: 1 mg/kg q24h
  • C: Heparin infusion to aPTT 1.5-2.0x above normal, given for at least 48h
  • O @ 30 days:
    • Primary outcome (death/MI): Enoxaparin 9.9%, heparin 12.0% (NNT 48), RR 0.83 (0.77-0.90)
      • Death: 6.9% vs 7.5%
      • MI: 3.0% vs 4.5%
    • Urgent revascularization: 2.1% vs 2.8%
    • Major bleed: 2.1% vs 1.4% (NNH 143), RR 1.53 (1.23-1.89)

Fondaparinux: OASIS 6 - Double-blind RCT

• P: STEMI (n=12,092)
• I: Fondaparinux 2.5 mg daily (first dose IV if lytic/PCI) for up to 8 days or until discharge
• C: Stratified by indication for placebo
  • No indication for heparin: Placebo
  • Indication for heparin (fibrin-specific thrombolytic or scheduled for 1o PCI): Heparin x1-2 days
• O @ 30 days (only presenting data for those with indication for heparin)
  • Primary outcome (death/MI): Fondaparinux 8.3%, heparin 8.7%, HR 0.96 (0.81-1.13)
  • Major bleed: Fondaparinux 2.1%, heparin 2.3%, HR 0.93 (0.67-1.30)

Beta-blockers post-MI or in stable CAD

Summary:

High-quality but outdated evidence (with likely overestimated benefit) from RCTs, corroborated by contemporary observational studies, supports the use of beta-blockers in patients post-MI without HF or LV dysfunction to reduce the risk of death;

• Acutely post-MI, beta-blockers reduce deaths due to arrhythmias & re-infarction. In the long-term where patients with normal LV function & low risk of ventricular arrhythmias, the main mechanism for mortality reduction of beta-blockers would be by reducing re-infarctions;
• Much uncertainty remains due to the indirectness of old evidence and high risk of bias of newer observational studies; only a contemporary, adequately-powered RCT of patients without HFrEF post-MI will provide clarity.

The evidence for post-MI beta-blocker use (in those without HF or LV dysfunction) is limited to an average of 3 years, after which the benefit of continued use is unclear;

• After 3 years, clinicians should re-assess the benefit/risk of continuing beta-blockers based on presence/control of angina, arrhythmias and risk factors for re-infarction, as well as tolerability and patient willingness to continue taking the beta-blocker.

Beta-blockers do not appear to reduce CV events in patients with uncomplicated stable CAD (no prior MI, LV dysfunction or HFrEF), so they should only be used in the presence of a compelling indication (such as angina, for which a calcium-channel blocker could also be used as first-line therapy).

 

Current guideline recommendations

Acute coronary syndrome (ACS; AHA 2013 STEMI guidelines, AHA 2014 NSTE-ACS guidelines, AHA 2011 secondary prevention guidelines)

• Start an oral beta-blocker on the first day of no contraindications (STEMI Class I recommendation, Level of evidence B; NSTE-ACS I, A)
• Use a beta-blocker in all patients with prior-MI & EF 40% or less unless contraindicated" (I, A)
• Continue during & after hospitalization in all patients with STEMI and with no contraindications" (I, B); also reasonable to continue in patients with NSTE-ACS with normal LV function" (IIa, C)
• Continued for 3 years after an ACS in all patients with normal LV function (I, A)
  • It is also reasonable to continue beyond (I, B)

For stable coronary artery disease (CAD)/ischemic heart disease (IHD; CCS 2014 stable IHD guidelines)

• Use a beta-blocker in all patients with stable IHD & LV dysfunction (strong recommendation, high-quality evidence) or prior MI (conditional recommendation, moderate-quality evidence)
• Use either a beta-blocker or calcium-channel blocker for stable angina if none of the above (conditional recommendation, moderate-quality evidence)
• Consider a beta-blocker for all other patients with coronary or other vascular disease" (AHA 2011 secondary prevention guidelines; IIb, C)

The focus of this article will be patients with CAD without HF/LV dysfunction. We have covered beta-blocker use for post-MI LV dysfunction & HFrEF elsewhere.

Early/short-term use during ACS

• COMMIT provides the best-available evidence in a contemporary population
  • Double-blind RCT of 45,852 patients with suspected MI (87% with STEMI, mean 10h from symptom onset) with no planned PCI
  • Randomized to metoprolol (5 mg IV x3 over 15 min, then 200 mg/d until discharge or up to 4 weeks) or placebo
  • There was no difference in the co-primary outcomes
    • Death, re-MI, VF, or other arrest) in hospital: Metoprolol 9.4%, placebo 9.9% (odds ratio [OR] 0.96, 0.90-1.01)
    • Death: Metoprolol 7.7%, placebo 7.8% (OR 0.99, 0.92-1.05)
  • Increased risk of cardiogenic shock (metoprolol 5.0% vs placebo 3.9%, NNH 91, OR 1.30), but
  • Decreased risk of re-MI (2.0% vs 2.5%, NNT 200, OR 0.82) & VF (2.5% vs 3.0%, NNT 200, OR 0.83)

Should beta-blockers be used post-MI in patients with normal LV function?

• A 1999 systematic review with meta-analysis remains the best-available evidence on this topic
  • Major caveats:
    • Included trials were published between 1966-1991, which precedes widespread use of many ACS therapies, including PCI & statins (most of the trials also preceded use of fibrinolytics, ASA)
    • Patients were not systematically assessed for HF or LV dysfunction, so it is unclear how many of these patients had normal LV function
    • Maximum average follow-up of 3 years
  • Over 2 years, use of a beta-blocker decreased the risk of death (NNT 42, OR 0.77, 0.69-0.85)
• A newer systematic review, which attempted to determine the efficacy of statins in the modern era, had numerous issues limiting clinical utility:
  • Arbitrarily classified trials as being in the "reperfusion era" if >50% of patients underwent revascularization with PCI/CABG, reperfusion with a fibrinolytic, or received ASA+statin
  • Results of the "reperfusion era" analysis dominated by COMMIT, which was a trial of short-term metoprolol use
  • There were no "reperfusion era" trials with beta-blocker duration >1 year
• Observational studies show conflicting results on beta-blocker use after MI
  • A 2017 cohort of 179,810 patients with MI found a reduction in 1-year mortality in unadjusted comparisons (4.9% versus 11.2% without beta-blockers), but not in adjusted analyses using propensity score matching or instrumental variables
    • Notably, 95% of participants in this study received a beta-blocker on discharge, leading to very high risk of selection bias
  • A 2015 systematic review of 10 cohort studies with 40,973 patients who underwent PCI for MI found a reduction in death with beta-blocker use (relative risk (RR) 0.58, 0.48-0.79)
    • The relative risk reduction was numerically greater for those with reduced EF (RR 0.60, 0.36-1.00) compared to those with EF >40% (RR 0.79, 0.59-1.07)
  • A 2015 cohort study (that did not exclude patients with HFrEF or LV dysfunction) not included in the above review  found that beta-blocker use after an MI reduced the relative risk of death (HR ~0.6) at a median 2.1 years 
  • Notably, a 2012 cohort study using the REACH registry that initially led to questioning the utility of beta-blockers post-MI was underpowered to identify a clinically meaningful difference

How long should beta-blockers be continued post-MI?

• Average duration of beta-blocker use in the 1999 systematic review was 2-3 years
• A 2016 cohort study of 2679 patients with MI without HF or LV dysfunction demonstrated a reduction in the risk of death at 30 days (hazard ratio (HR) 0.46, 0.26-0.82)
  • This study was underpowered to identify a clinically-important difference in death at 1 year (HR 0.77, 0.46-1.30) or 5 years (HR 1.19, 0.65-2.18)
• There is no evidence that discontinuing beta-blockers after a certain duration post-MI is safe or beneficial

What about patients with stable CAD (without prior MI, or HFrEF/LV dysfunction)?

• A 2016 systematic review with meta-analysis of cohort studies that included 17,397 patients with angiographically-proven CAD without MI or LV dysfunction found no difference in all-cause death (OR 0.91, 0.79-1.04) at 3-5.4 years
• A 2014 cohort study of 26,793 patients with newly-diagnosed CAD found that that the effect of beta-blockers on the risk of death/MI differed on MI history (p=0.005 for interaction)
  • Prior MI: HR 0.87 (0.82-0.93)
  • PCI or CABG but no prior MI: HR 1.03 (0.93-1.13)
• INVEST trial (see previous nerdcat summary): A beta-blocker-based regimen was not superior to a verapamil-based regimen over 2.7 years in patients with CAD+HTN without prior MI