Author + information
- Received May 28, 2004
- Revision received September 2, 2005
- Accepted September 12, 2005
- Published online February 7, 2006.
- Carl J. Pepine, MD, MACC⁎,⁎ (, )
- Peter R. Kowey, MD, FACC†,
- Stuart Kupfer, MD‡,
- Rainer E. Kolloch, MD§,
- Athanase Benetos, MD, PhD∥,
- Giuseppe Mancia, MD, PhD¶,
- Antonio Coca, MD#,
- Rhonda M. Cooper-DeHoff, PharmD⁎,
- Eileen Handberg, PhD⁎,
- Efrain Gaxiola, MD, FACC⁎⁎,
- Peter Sleight, MD, FACC††,
- C. Richard Conti, MD, MACC⁎,
- Ann C. Hewkin, MSc‡,
- Luigi Tavazzi, MD, FACC‡‡,
- INVEST Investigators
- ↵⁎Reprint requests and correspondence:
Dr. Carl J. Pepine, Division of Cardiovascular Medicine, University of Florida, 1600 SW Archer Road, Gainesville, Florida 32610-0277.
Objectives We sought to determine predictors for adverse outcomes in hypertensive patients with coronary artery disease (CAD).
Background Factors leading to adverse outcomes in hypertensive patients with CAD are poorly understood. The INternational VErapamil-trandolapril STudy (INVEST) compared outcomes in hypertensive patients with CAD that were assigned randomly to either a verapamil sustained-release (SR)- or an atenolol-based strategy for blood pressure (BP) control. Trandolapril and hydrochlorothiazide were used as added agents. During follow-up (61,835 patient-years), BP control and the primary outcome (death, nonfatal myocardial infarction, and nonfatal stroke) were not different between strategies.
Methods We investigated risk for adverse outcome associated with baseline factors, follow-up BP, and drug treatments using Cox modeling.
Results Previous heart failure (adjusted hazard ratio [HR] 1.96), as well as diabetes (HR 1.77), increased age (HR 1.63), U.S. residency (HR 1.61), renal impairment (HR 1.50), stroke/transient ischemic attack (HR 1.43), smoking (HR 1.41), myocardial infarction (HR 1.34), peripheral vascular disease (HR 1.27), and revascularization (HR 1.15) predicted increased risk. Follow-up systolic BP <140 mm Hg or diastolic BP <90 mm Hg (HRs 0.82 or 0.70, respectively) and trandolapril with verapamil SR (HRs 0.78 and 0.79) were associated with reduced risk.
Conclusions In hypertensive patients with CAD, increased risk for adverse outcomes was associated with conditions related to the severity of CAD and diminished left ventricular function. Lower follow-up BP and addition of trandolapril to verapamil SR each were associated with reduced risk.
The management of blood pressure (BP) in patients with coronary artery disease (CAD) remains a challenge; as the population ages, more patients survive myocardial infarction (MI), and the prevalence of diabetes, obesity, and inactivity increases (1). To better manage these patients, more prospective data are needed to understand the risks associated with clinical factors leading to adverse outcomes.
Beta-blocker/diuretic treatment is beneficial, but many trials of these drugs preceded aspirin, statins, MI reperfusion, revascularization, angiotensin-converting enzyme inhibitors, and organ protection; enrolled few patients with CAD; and many had inadequate BP control (2). Calcium antagonists and angiotensin-converting enzyme inhibitors reduce events; however, their value compared with beta-blocker/diuretic treatment has not been sufficiently tested in contemporary hypertensive patients with CAD. The foregoing provided rationale for the INternational VErapamil-trandolapril STudy (INVEST) and for this prespecified analysis to investigate factors leading to risk for adverse outcomes.
Beginning in 1997, the INVEST randomly assigned 22,576 CAD patients 50 years or older to either a verapamil sustained-release (SR)-based or an atenolol-based strategy for BP control. After an extensive cardiovascular history and physical examination, multidrug treatment strategies were titrated to BP <140/90 mm Hg, or <130/85 mm Hg for diabetes or renal impairment (3). Trandolapril and hydrochlorothiazide (HCTZ) were specified as added agents, with trandolapril primary in the verapamil SR strategy and HCTZ primary in the atenolol strategy. In both strategies, trandolapril was recommended for heart failure, diabetes, or renal impairment. The trial concluded in 2003, accumulating 61,835 patient-years follow-up. Each strategy provided excellent BP control (>70% of patients achieved BP <140/90 mm Hg) and the strategies were equivalent in preventing all-cause death, nonfatal MI, or nonfatal stroke (primary outcome). The design and results have been published (4,5).
To identify factors associated with risk for primary outcome, stepwise Cox proportional-hazards regression was used to provide hazard ratios (HRs) and 95% confidence intervals (CIs) for the following baseline covariates: age (10-year increments), gender, race/ethnicity (Caucasian, Asian, black, Hispanic, multiracial/other), U.S. residency, body mass index (BMI) in 5-kg/m2increments, MI, heart failure (functional class I to III), renal impairment, peripheral vascular disease, aspirin use, left ventricular hypertrophy, smoking (ever), coronary revascularization, stroke/transient ischemic attack, angina pectoris, unstable angina, arrhythmia, hypercholesterolemia, and diabetes. Prespecified covariates (age, gender, race/ethnicity, previous MI, and heart failure), as well as a factor for strategy were forced entries. The remaining covariates were selected by the procedure and retained when p ≤ 0.1.
Baseline systolic blood pressure (SBP) and diastolic blood pressure (DBP) were assessed by adding these terms to the aforementioned Cox model. To assess the effect of BP after randomization, we used exploratory Cox regression models excluding treatment strategy both unadjusted and stepwise with BP fit as a time-dependent variable. Here, patients were categorized by SBP <140 mm Hg or ≥140 mm Hg and DBP <90 mm Hg or ≥90 mm Hg at each visit. Additionally, a model with terms for baseline covariate, time-dependent SBP category (more than six weeks prior to primary outcome event or censoring) and the interaction term was used to assess the effect of the prognostic factors on primary outcome.
A further model used average daily dose of each strategy drug for each patient and variables indicating the proportion of verapamil SR and trandolapril, or the proportion of atenolol and HCTZ, prescribed simultaneously in addition to treatment strategy; interactions of drug-dose variables with strategy; and prespecified baseline covariates to assess the relationship between study drugs and primary outcome (6). Drug-dose variables were fit as time-dependent. Estimates are presented for selected dose combinations assuming simultaneous proportions are 100%, with atenolol 50 mg/day as the reference (HR = 1.0).
Mean and standard deviation were provided for continuous variables, or number of patients and percent (%) for categorical variables. Statistical significance was assumed if p ≤ 0.05 (two-tailed).
Conditions associated with increased risk
Baseline conditions independently associated with increased risk along with HRs and 95% CIs appear in Figure 1.Adjusted risk for the primary outcome was highest for patients with previous heart failure or diabetes. Interestingly, black race was not associated with significant increase in risk, whereas Hispanic and Asian race were associated with benefit. After adjusting for these baseline conditions, there was still no difference in risk comparing the verapamil SR and atenolol strategies (HR 0.97, 95% CI 0.89 to 1.05).
Influence of BP
Baseline SBP and DBP demonstrated no association with risk for the primary outcome, either with or without adjustment. Lower on-treatment BP was associated with benefit (SBP <140 mm Hg or DBP <90 mm Hg, unadjusted HR = 0.73, p < 0.001, adjusted HR = 0.82, p < 0.001 or unadjusted HR = 0.91, p = 0.136; adjusted HR = 0.70, p < 0.001, respectively). Similar results were observed in high-risk subgroups (Fig. 2).
Influence of antihypertensive drugs
On-treatment mean BPs and proportions of patients with SBP <140 mm Hg were very comparable between patients receiving monotherapies and those receiving multiple drugs. At month 24 for those receiving verapamil alone, those values were 127.2 mm Hg, 87.8%; atenolol alone, 127.2 mm Hg, 85.8%; and dual therapies, verapamil + trandolapril, 129.3 mm Hg, 80.1% and atenolol + HCTZ, 129.6 mm Hg, 78.9%.
At these comparable BPs, exploratory drug-dose modeling using selected doses suggested that added drugs were associated with additional benefits (Fig. 3).Although some combinations of verapamil SR and trandolapril (180 mg/2 mg/day) showed significant risk reduction, all multidrug combinations showed trends for benefit compared with atenolol 50 mg/day alone.
Previous studies have identified factors associated with adverse outcomes in populations with or at high risk for CAD, but reports that focus on hypertensive patients with CAD are limited. Thus these observations extend previous findings to older patients with CAD using contemporary drugs for hypertension. For example, in patients with CAD who had lower-extremity arterial disease, predictors of mortality included increased age, CAD severity/extent, diminished left ventricular function, hypertension, smoking, and diabetes (7). More recently, in hypertensive patients at high risk for CAD, age, left ventricular hypertrophy, diabetes, and average SBP during follow-up were independent predictors of risk (8). We found in hypertensive patients with CAD that factors related to heart failure and CAD severity (e.g., diabetes, increased age renal impairment, stroke/transient attack, smoking, MI, peripheral vascular disease, and coronary revascularization) were all independently associated with increased risk for death, MI, or stroke. Of clinical interest, exploratory analyses indicated that achieving SBP <140 mm Hg and receiving combination drug therapy each mitigate some of that risk (Figs. 2 and 3). Because CAD is both disabling and associated with subsequent morbidity and mortality, these findings identify a large group of patients for whom even closer continuing care and more intense therapy may be warranted.
Association between higher BP and increased risk is well established but has not been emphasized for elderly hypertensive patients with CAD. In general, we found in exploratory analyses that time-dependent SBP <140 mm Hg was associated with lower risk, in both the low- and high-risk subgroups defined by baseline conditions (Fig. 2).
When BP is reduced to comparable levels by different drugs, especially in patients with the high-risk conditions identified here, the choice of antihypertensive agents may be important in further reducing risk. Of current interest was the finding that trends to reduce risk appeared for addition of HCTZ to atenolol and trandolapril to verapamil SR. However, adding trandolapril to verapamil SR was associated with significant risk reduction at certain doses, with and without addition of HCTZ. This finding is consistent with results of other trials in high-risk patients (2,9,10). Beneficial trends occurred at all doses of trandolapril, and risk reduction was significant for the 2-mg dose when added to verapamil, despite the fact that trandolapril also was recommended for diabetes, heart failure, or renal impairment in the atenolol strategy. These results should be viewed as support for a multi-drug BP control strategy that includes angiotensin-converting enzyme inhibition in high-risk patients with CAD.
Patients younger than 50 years of age, those with functional class IV heart failure, with creatinine levels ≥4.0 mg/dl, or recent stroke, unstable angina, coronary revascularization, or MI were excluded. Surprisingly, increased BMI was not associated with increased risk, but these findings require additional analysis beyond the scope of this publication. Interactions with age, smoking, gender, and disease and, particularly, the increase in all cause mortality observed among those with lower BMI in large adult cohorts (11) make this analysis complex. Yet the possibility remains that when BP is well controlled and organ protection is implemented for heart failure, diabetes, and renal impairment, increased BMI could have less of an impact on outcomes. Because drugs and doses given were dependent on BP response and other factors (e.g., patient’s well-being), some confounding is unavoidable. Finally, effects of nonstrategy antihypertensive drugs were not analyzed; however, sensitivity analyses indicated they did not alter results qualitatively.
Many conditions in hypertensive patients with CAD impart excess risk. Age, as well as previous diabetes, stroke/transient ischemic attack, smoking, heart failure, renal impairment, MI, peripheral vascular disease, and revascularization are independently associated with increased risk. In general, the high risk associated with these conditions was reduced by achieving a SBP <140 mm Hg. Although both treatment strategies provided excellent BP control, with no difference in outcomes overall, the addition of trandolapril to verapamil SR was associated with reduction in risk, and this combination may offer greater protection than monotherapy. Although all patients with CAD and hypertension require aggressive risk factor modification, the subgroup with these risk conditions may benefit from particularly close surveillance for residual BP elevation and consideration for additional therapies.
The INVEST study was supported by grants from the University of Florida (Gainesville, Florida) and Abbott (Abbott Park, Illinois). Dr. Pepine received grant support for research, CME, consultant, and/or speaker’s bureau program from the following: Abbott, AstraZeneca, Aventis Pharmaceuticals Inc., Berlex Laboratories Inc., CV Therapeutics, King Pharmaceuticals Inc., Monarch Pharmaceuticals, Pfizer, and Wyeth-Ayerst Laboratories. He holds the following patents: U.S. patent no. 5,991,731 “Method and system for interactive prescription and distribution of prescriptions in conducting medical studies” and U.S. patent application no. 10/816,461 “Combination drug therapy for treating hypertension.” Dr. Kupfer was employed by Abbott. Dr. Gaxiola was part of the INVEST study as an investigator and received a small amount of money for patients enrolled. He did not receive anything for consultancies, stock ownership, or other equity interests or patent-licensing arrangements. Dr. Cooper-DeHoff received grant support from Abbott. She holds the following patents: U.S. patent no. 5,991,731 “Method and system for interactive prescription and distribution of prescriptions in conducting medical studies” and U.S. patent application no. 10/816,461 “Combination drug therapy for treating hypertension.” Dr. Handberg received grant support from Abbott and holds the following patents: U.S. patent no. 5,991,731 “Method and system for interactive prescription and distribution of prescriptions in conducting medical studies” and U.S. patent application no. 10/816,461 “Combination drug therapy for treating hypertension.” Dr. Sleight received grant support for following trials: ISIS 1-4, HPS, SEARCH, HOPE, and ONTARGET/TRANSCEND. He has received grant support from BHF, United Kingdom MRC, Canadian MRC, Ontario Heart & Stroke Foundation, AstraZeneca, Aventis, Boehringer-Ingelheim, BMS, GSK, Monarch, MSD, Nat. Vit. E Association, and Roche and has received speaker/DSMB fees from Abbott, AstraZeneca, Aventis, Bayer, Boehringer-Ingelheim, Boehringer Mannheim, BMS, Genentech, GSK, Knoll, Menarini, Merck, Monarch, MSD, Novartis, Organon, Pfizer, Pharmacia, Sanofi, Schering, and Servier. Dr. Conti was an advisor for CV therapeutics and consultant with Vasomedical. Ms. Hewkin is employed by Abbott.
- Abbreviations and Acronyms
- body mass index
- blood pressure
- coronary artery disease
- confidence interval
- diastolic blood pressure
- hazard ratio
- INternational VErapamil-trandolapril STudy
- myocardial infarction
- systolic blood pressure
- sustained release
- Received May 28, 2004.
- Revision received September 2, 2005.
- Accepted September 12, 2005.
- American College of Cardiology Foundation
- American Heart Association
- Williams B.
- Pepine C.J.,
- Handberg E.M.,
- Cooper-DeHoff R.M.,
- et al.
- Pepine C.J.,
- Handberg-Thurmond E.,
- Marks R.G.,
- et al.
- Elliott W.J.,
- Hewkin A.C.,
- Kupfer S.,
- Cooper-DeHoff R.,
- Pepine C.J.