 |
The Agency for Healthcare Research and Quality (AHRQ), through its Evidence-Based Practice Centers (EPCs), sponsors the development of evidence reports and technology assessments to assist public- and private-sector organizations in their efforts to improve the quality of health care in the United States. The reports and assessments provide organizations with comprehensive, science-based information on common, costly medical conditions and new health care technologies. The EPCs systematically review the relevant scientific literature on topics assigned to them by AHRQ and conduct additional analyses when appropriate prior to developing their reports and assessments.
To bring the broadest range of experts into the development of evidence reports and health technology assessments, AHRQ encourages the EPCs to form partnerships and enter into collaborations with other medical and research organizations. The EPCs work with these partner organizations to ensure that the evidence reports and technology assessments they produce will become building blocks for health care quality improvement projects throughout the Nation. The reports undergo peer review prior to their release.
AHRQ expects that the EPC evidence reports and technology assessments will inform individual health plans, providers, and purchasers as well as the health care system as a whole by providing important information to help improve health care quality.
We welcome written comments on this evidence report. They may be sent to: Director, Center for Practice and Technology Assessment, Agency for Healthcare Research and Quality, 540 Gaither Road, Rockville, MD 20850.
Carolyn Clancy, M.D.
Director
Agency for Healthcare Research and Quality
Jean Slutsky, P.A., M.S.P.H.
Acting Director, Center for Practice and Technology Assessment
Agency for Healthcare Research and Quality
The authors of this report are responsible for its content. Statements in the report should not be construed as endorsement by the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services of a particular drug, device, test, treatment, or other clinical service.
Objectives. This evidence-based report had two objectives. The first objective was to assess whether angiotensin-converting enzyme inhibitors (ACE inhibitors) and beta-adrenergic blocking agents (beta-blockers) are effective in patients with left ventricular systolic heart failure and whether this effectiveness differs in the following subpopulations: men, women, blacks, whites, diabetics, and nondiabetics. The second objective was to assess the cost-effectiveness of both treatment of and screening for left ventricular systolic dysfunction.
Search Strategy. We conducted a thorough computerized library search and retrieved all articles that pertained to the twelve largest placebo-controlled studies on ACE inhibitors and beta-blockers. We also contacted leading experts in cardiology for unpublished data, contacted the authors of the clinical trials for patient-level data, and obtained patient-level data from the FDA.
Selection Criteria. We selected the twelve largest randomized placebo-controlled trials of ACE inhibitors and beta-blockers.
Data Collection and Analysis. We retrieved data through published articles or patient-level data files. For each, we estimated the mortality relative risk and hazard ratio for the subgroups of interest. For example, the relative risk of mortality for women is equal to the risk of dying for women who received the drug divided by the risk of dying for women who received a placebo. We pooled these statistics across studies. We then assessed whether these risks differed statistically via a ratio statistic. For example, to assess the relative effect of the drug on the relative risk of mortality for women as compared to men, we divided the relative risk in women by the relative risk in men to produce a ratio of relative risks. We pooled these statistics and tested whether the pooled ratio estimate was significantly different from 1.
In order to assess the cost-effectiveness of screening for and treating asymptomatic left ventricular dysfunction, we created a decision model. We modeled lifetime health and economic outcomes for a hypothetical cohort of 55-year-old asymptomatic patients with ejection fraction of 35% or less but no history of heart failure (HF), using two treatment strategies and six screening strategies.
Main Results. We found evidence, with two exceptions, that treatment with ACE inhibitors or beta-blockers reduces all-cause mortality in male, female, black, white, diabetic, and nondiabetic patients. The two exceptions were the use of ACE inhibitors in women and the use of beta-blockers in black patients. Regarding the former, we found clear evidence that treating women with symptomatic heart failure with ACE inhibitors was beneficial. However, the available evidence do not support a beneficial effect in women with asymptomatic left ventricular systolic dysfunction. Regarding black patients, treatment with the beta-blocker bucindolol was associated with a nonstatistically significant increase in all-cause mortality, while treatment with other beta-blockers was associated with a nonstatistically significant reduction in mortality of similar magnitude to the statistically significant reductions observed in white patients.
In our cost-effectiveness analyses, we found that treatment of asymptomatic left ventricular dysfunction with ACE inhibitors was very cost-effective under virtually all assumptions, with typical costs per quality-adjusted life-year gained of between $5,000 and $10,000. Additional analysis showed that screening with B-type natriuretic peptide followed by echocardiography in a cohort of asymptomatic 55-year-old individuals was also cost-effective, compared with the costs of other therapies currently considered standard medical care. The number needed to screen in order to gain one year of additional life was 77. These results were only modestly sensitive to cost and were most sensitive to the prevalence of asymptomatic decreased left ventricular ejection fraction. When the prevalence falls below about 1%, a strategy of screening becomes less cost-effective than commonly accepted thresholds for cost-effective care.
Conclusions. ACE inhibitors and beta-blockers reduce mortality in a broad range of patients with left ventricular systolic dysfunction, including men and women, blacks and whites, and diabetics and nondiabetics. However, the value of ACE inhibitors in women with asymptomatic left ventricular systolic dysfunction is uncertain, and additional study is needed. In addition, based on data from a single study, the beta-blocker bucindolol may be associated with increased mortality in blacks, whereas other beta-blockers provide similar benefits in blacks and whites.
Treatment of asymptomatic left ventricular dysfunction with ACE inhibitors is very cost- effective. In addition, screening for asymptomatic left ventricular dysfunction with B-type natriuretic peptide followed by echocardiography is cost-effective in populations where the prevalence of this condition is 1% or greater.
Heart failure (HF) is associated with substantial morbidity and mortality; it is a primary or secondary cause of death for approximately 250,000 people per year in the United States. According to the 2002 Heart and Stroke Statistical Update (www.americanheart.org), HF was the first-listed diagnosis for 962,000 hospitalizations in 1999, and it is the most common diagnosis among hospital patients age 65 and older. In fact, 20 percent of all hospitalizations in this age group carry a primary or secondary diagnosis of HF. Over 3 million outpatient office visits each year are related to this illness. In 1998 alone, the estimated annual direct cost due to HF was $18.8 billion.
A series of studies has established that angiotensin-converting enzyme inhibitors (ACE inhibitors) and beta-adrenergic blocking agents (beta-blockers) provide life-saving benefits in patients with HF and left ventricular systolic dysfunction. However, most of the patients enrolled in such studies have been white males. Thus, a clinical question that is repeatedly asked is whether the mortality benefit reported in these clinical trials is also achieved for particular subpopulations, such as women, people of other races, and patients with various comorbidities such as diabetes mellitus or renal insufficiency. Since few of the randomized trials enrolled enough women, blacks, or patients with comorbidities to have sufficient statistical power to support conclusions based on subgroup analysis, this question is appropriate for meta-analysis.
In addition, because the clinical trial data support a mortality benefit for patients with asymptomatic left ventricular dysfunction, it is natural to question both the cost-effectiveness of such treatment and that of screening asymptomatic patients for left ventricular dysfunction. These clinical and policy questions form the basis for this report.
AHRQ defined the scope of work for this project to include an evidence report and quantitative analysis on the effectiveness of treatment for HF using ACE inhibitors and beta-blockers. This topic was nominated by the American College of Physicians, the American Society of Internal Medicine, and the American Academy of Family Physicians. This group submitted the following potential key questions to AHRQ:
What evidence exists on the effectiveness of nurse management programs and health food supplements?
What evidence exists on the treatment of sleep apnea in patients with HF?
What is the evidence on the treatment of specific myocardial disorders, e.g., myocarditis, sarcoidosis, and amyloidosis, in patients with HF?
What interventions are effective for patients with diastolic dysfunction?
Which patients benefit from which beta-blockers?
What are the effects of potassium levels on HF outcomes?
Do angiotensin blockers improve outcomes?
What, if any, are the differences in treatment effectiveness associated with patient gender, race, age, and income level?
After congestive heart failure was nominated as a topic, but prior to assignment of this contract to the Southern California Evidence-based Practice Center (SCEPC), the American Heart Association (AHA) and the American College of Cardiology (ACC) released practice guidelines on the management of HF. AHA/ACC graciously provided the SCEPC with a draft copy for our confidential review. On September 8, 2000, a conference call was held with our technical expert panel (TEP) to limit the key questions to be addressed in the evidence report. The purpose of the conference call was to identify topic areas for this report that would complement but not duplicate the draft guidelines, a copy of which had been made available to each TEP member. The technical experts judged that several of the original key questions posed by the nominating organizations had been answered adequately in the AHA/ACC guidelines, major studies were under way that would answer several more of the questions, and published data would be insufficient to reach meaningful conclusions for other questions. The TEP identified three areas in which they believed significant contributions could still be made:
Assessment of the effects of age over 70, gender, race, and assisted living on treatment outcomes.
Cost-effectiveness of medication combinations.
Assessment of outcomes in patients with various comorbidities, particularly diabetes mellitus, renal dysfunction, and cognitive dysfunction.
This evidence-based report addressed the following key questions regarding pharmacologic management of heart failure and left ventricular systolic dysfunction:
Are angiotensin-converting enzyme inhibitors (ACE inhibitors) and beta-adrenergic blocking agents (beta-blockers) effective in patients with HF and left ventricular systolic dysfunction and does this effectiveness differ in the following subpopulations: men, women, blacks, whites, diabetics, and nondiabetics?
What is the association between treatment with ACE inhibitors and beta-blockers and all-cause mortality for female, male, diabetic, nondiabetic, black, and white patients with HF?
Does this association vary (e.g., are there statistically significant differences) by gender (female versus male), diabetic condition (those with diabetes versus those without), and race (black versus white patients)?
What is the cost-effectiveness of both treatment of and screening for asymptomatic left ventricular systolic dysfunction?
To answer key question 1a and 1b, we first retrieved all articles that pertained to eleven large randomized placebo-controlled studies on ACE inhibitors and beta-blockers. Because the SOLVD study actually consisted of two distinct trials (one on prevention and one on treatment), we included twelve studies in total. Meta-analyses were performed separately for the ACE inhibitor and beta-blocker studies. The common outcome of interest was all-cause mortality. For some studies, both patient-level data and published summary data were available; if the two disagreed, we always chose the patient-level statistics over published group-level statistics. Among the five studies for which we had patient-level data, three datasets had minor disagreements with related publications.
All reports that presented the relevant patient sub-population data did so in the form of a two-by-two table of all-cause mortality by treatment (or placebo) group for each sub-population. Alternatively, if we were given the patient-level data, we could construct this table directly. For example, an ACE inhibitor study might provide separate two-by-two tables for men and women.
To answer key question 1a, for each sub-population (e.g., women), we estimated the log mortality relative risk, which is equal to the log of the risk of dying for women who received ACE inhibitors divided by the risk of dying for women who received placebo. The standard error for the log relative risk was also estimated, and a 95 percent confidence interval was constructed. A similar log relative risk and confidence interval were calculated for men. We then back-transformed to the unlogged scale for interpretability so that our final statistic for each sub-population in each study was the relative risk with its associated confidence interval. The analysis informed us about the association between various patient characteristics, such as gender, and mortality, with that association measured on the relative risk scale.
To answer key question 1b, that is, whether the association differed between sub-populations (e.g., female versus male), we determined whether statistical differences existed between the relative risks for two subpopulations. We did this by constructing a test statistic equal to the ratio of relative risks (RRR), which equals the female relative risk divided by the male relative risk, for example. If this test statistic differs significantly from 1, then we infer that the relative risks for the two subgroups are significantly different. As before, we performed the analysis on the log scale. The log ratio of relative risks equals the log of the relative risk for women divided by the relative risk for men, and its standard error equals the square root of the sum of the variances of the two log relative risks. We constructed a confidence interval on the log scale. We then back-transformed the estimate and its confidence interval to the unlogged scale so that our final test statistic for each study was the ratio of relative risks.
Because the followup times varied across studies and calculating the relative risk does not take this variation (or the censoring of observations) into account, we also assessed the mortality associated with ACE inhibitors and beta-blockers respectively on the hazard ratio scale. The majority of our studies presented hazard ratios and confidence intervals, and after transforming these statistics to the log scale, we extracted the log hazard ratio and its standard error for each study. We estimated the log hazard ratio for each patient subgroup of interest for each study that provided the data stratified on that dimension. We followed the same analytic strategy for the hazard ratio as for the relative risk, conducting a random-effects pooled analysis on the log scale, and back-transforming to the unlogged scale. We then constructed a ratio of hazard ratios (RHR) to compare the hazard ratios in each subgroup.
For each drug and patient comparison subgroup of studies, we assessed the possibility of publication bias by evaluating a funnel plot of the individual study log relative risks and hazard ratios. In addition, we performed a sensitivity analysis, because studies varied in their definitions of racial groups. For racial comparisons, if the study provided data separately by racial subgroup, we utilized those data. If the data were not stratified in that way, we used data for black versus nonblack patients. Our last choice was data for nonwhite versus white patients. For those studies that described the data in more than one of these ways, we compared the relative risk and hazard ratio statistics.
To address key question 2, we developed a decision model to assess the cost-effectiveness of treatment for asymptomatic left ventricular dysfunction, using EXCEL (Version 5.0, Microsoft Corporation, Redmond, WA) and DATA (Version 3.0, TreeAge Software, Boston, MA) software. Using two treatment strategies, we modeled the lifetime health and economic outcomes for a hypothetical cohort of 55-year-old asymptomatic patients with ejection fraction of 35 percent or less but no history of HF. In the first strategy, asymptomatic patients are treated with ACE inhibitors. In the second strategy, patients are not treated with ACE inhibitors until they develop symptomatic HF.
During each time period of interest (e.g., 1 month), patients with no history of HF can remain asymptomatic, develop heart failure, or die. Of those patients who developed HF, we assumed 33 percent would be hospitalized during their initial episode. Once patients develop HF, they can remain in stable heart failure, be hospitalized, or die during each time period. The model follows each patient until death.
We also developed a decision model to assess various screening options for reduced left ventricular ejection fraction. We examined six screening strategies:
Echocardiography for all patients. Patients with an ejection fraction less than 35 percent are treated (ACE inhibitors) to prevent development of HF.
Electrocardiography (ECG) first, and if abnormal, echocardiography.
Blood test for B-type natriuretic peptide (BNP) first and, if abnormal, echocardiography.
ECG only, with treatment based on the results.
BNP only, with treatment based on the results.
No screening for depressed left ventricular function.
Each screening option has one of four possible outcomes: true positive, false positive, true negative, or false negative. In our model, only true and false positives are treated. True-positive patients have a higher quality-adjusted survival than false negatives, who are treated only when HF develops. True-negative patients have a normal age-specific life expectancy. False-positive patients receive a small decrement in quality-adjusted survival to account for potential side effects of treatment.
We generated the lifetime health and economic outcomes for hypothetical cohorts of 55-year-old patients with (1) depressed ejection fraction (35 percent or less) but no history of HF treated with ACE inhibitors, (2) depressed ejection fraction but no history of HF and no treatment until HF developed, and (3) patients without depressed ejection fraction. Each month, patients with a depressed ejection fraction and without a history of HF can remain asymptomatic, develop HF, or die. Of those patients who develop HF, we assumed that 33 percent would be hospitalized during their initial episode. Once patients develop HF, they can remain in stable HF, be hospitalized, or die during each time period. The model follows each patient until death.
Effects of gender. For seven studies, we were able to obtain gender-stratified data to calculate the effect of ACE inhibitors on mortality. The data from one study could be used only in the RRR assessment, and the data from another could be used only in the RHR assessment. In aggregate, these studies included 2,898 women and 11,674 men and ranged in duration from 6 months to 42 months. The pooled random-effects estimates from the six studies with relative risk data yielded values of 0.82 for men (95% CI: 0.74, 0.90) and 0.92 for women (95% CI: 0.81, 1.04). The corresponding pooled random-effects estimates from the six studies with hazard ratio data yielded values for the men of 0.76 (95% CI: 0.66, 0.87) and for women of 0.84 (95% CI: 0.72, 0.98.) The difference in effect between men and women approached statistical significance for the ratio of relative risks (p = 0.07).
This difference between the estimates of relative risk and hazard ratios is due to the inclusion in the hazard ratio analysis of the AIRE study, which reported a slight nonsignificant mortality benefit for women compared to men treated with ramipril. In contrast, the relative risk analysis included the SAVE study, which reported a distinct but non-statistically significant higher mortality in women relative to men treated with captopril (RRR = 1.24). In a subgroup analysis, studies were divided into those that treated symptomatic HF (risk ratio analysis for CONSENSUS, SOLVD-treatment, and TRACE; hazard ratio analysis for AIRE, CONSENSUS, SOLVD-treatment, and TRACE) and those that treated for asymptomatic left ventricular systolic dysfunction (risk ratio analysis for SAVE, SOLVD-prevention, and SMILE; hazard ratio analysis for AIRE, SOLVD-prevention, and SMILE). The difference in efficacy between men and women is most pronounced for treatment of asymptomatic left ventricular dysfunction, where the evidence does not support or suggest a mortality benefit for women (relative risk = 0.96; 95% CI: 0.75, 1.22).
The evidence indicates that women with symptomatic heart failure benefit when treated with ACE inhibitors, although the benefit may be somewhat less than that seen in men. However, the evidence does not support a mortality benefit from ACE inhibitors in women with asymptomatic left ventricular systolic dysfunction.
Differences between diabetics and nondiabetics. We were able to obtain data stratified by co-occurrence of diabetes from six studies to calculate the effect of ACE inhibitors on mortality. In aggregate, these studies included 2,398 patients with diabetes and 10,188 patients without diabetes. All of these studies contributed data to our relative risk analysis; however, one study did not contain data that we could use for our hazard ratio analysis. Both analyses yielded similar results. The random-effects pooled estimate of the relative risk of mortality in patients with diabetes is 0.84 (95% CI: 0.70, 1.00) while the estimate of the relative risk in patients without diabetes is 0.85 (95% CI: 0.78, 0.92). The corresponding estimates for the hazard ratio are 0.73 (95% CI: 0.56, 0.95) for diabetics and 0.80 (95% CI: 0.69, 0.93) for nondiabetics. These results indicate that both patients with diabetes and patients without diabetes achieve reductions in mortality when treated with ACE inhibitors for HF.
Effects of race. We were able to obtain data stratified by patient race from three studies to assess the effects of ACE inhibitors on mortality. The remaining ACE inhibitor studies were conducted primarily in Scandinavian and European countries and did not enroll substantial numbers of black patients. Because one study did not present data that allowed us to calculate the hazard ratios, we had an insufficient number of studies to pool for this analysis. Therefore, only a pooled relative risk analysis was performed, which yielded an estimate in white patients of 0.89 (95% CI: 0.82, 0.97) and an estimate in black patients of 0.89 (95% CI: 0.74, 1.06). These data provide no evidence that black patients achieve lesser or greater reductions in mortality than white patients when treated with ACE inhibitors for HF. While the relative risk reduction in black patients did not achieve conventional level of statistical significance, the estimate of effect is the same as the statistically significant reduction seen in white patients. Furthermore, the two estimates of effect (for black and white patients) do not statistically differ from each other. These results are consistent with the analysis by the SOLVD investigators, who reported that there was no significant difference in mortality reduction among black and white patients in the SOLVD studies. (However, these investigators did report a difference in hospitalization rate in black patients compared to white patients.)
Effects of gender. Five studies provided gender-stratified data on the effect of beta-blocker treatment on mortality. One study contributed data only to the relative risk analysis. Our TEP determined that bucindolol, the beta-blocker evaluated in BEST, was sufficiently different in action from the other beta-blockers to justify excluding the BEST study from pooled analysis. In aggregate, the pooled studies included 2,134 women and 7,885 men. Both analyses yield similar results. The random-effects pooled estimate for the relative risk on mortality for women was 0.63 (95% CI: 0.44, 0.91), while for men the estimate was 0.66 (95% CI: 0.59, 0.75). The corresponding values for the hazard ratio analysis were 0.62 (95% CI: 0.34, 1.14) for women and 0.62 (95% CI: 0.52, 0.73) for men. Likewise, BEST reported equal effects in men and women (although in BEST, the reduction in all-cause mortality was not statistically significant). Our interpretation of these data is that both women and men with symptomatic HF have reduced mortality when treated with beta-blockers.
Differences between diabetics and nondiabetics. Three studies provided data stratified by co-occurrence of diabetes to calculate the effect of beta-blocker treatment on mortality. In aggregate, these studies included 1,883 patients with and 7,042 patients without diabetes. The only pooled estimates that were possible were the relative risks and they yielded a value of 0.65 (95% CI: 0.57, 0.74) for nondiabetic patients and a value of 0.77 (95% CI: 0.61, 0.96) for diabetic patients. This difference in relative risk was not statistically significant; however, the 95 percent confidence interval was very broad. Our interpretation of these data is that in patients with HF, with or without diabetes, beta-blocker treatment is associated with reduced mortality.
Effects of race. Four studies provided race-stratified data to assess the effects of beta-blocker treatment on mortality. As mentioned above, BEST was judged to be clinically dissimilar to the other studies and was not included in the pooled analysis. In addition, one study was conducted in Scandinavian and European countries and did not enroll appreciable numbers of black patients. In aggregate, the three studies included in the pooled analysis included 545 black patients and more than 6,000 white patients. Both the relative risk analysis and the hazard ratio analysis yielded similar results. The pooled random-effects estimate of the relative risk of the effect on mortality for blacks was 0.67 (95% CI: 0.39, 1.16), whereas for whites it was 0.63 (95% CI: 0.52, 0.77). The corresponding pooled estimates from the hazard ratio analysis were 0.64 (95% CI: 0.36, 1.16) for black patients and 0.59 (95% CI: 0.45, 0.76) for white patients.
In contrast, the BEST trial showed a statistically significant racial difference in mortality for bucindolol treatment. In fact, the relative risk and hazard ratio for mortality exceeded 1 for blacks (although this was not statistically significant). Our interpretation of these data is that black patients are likely to have the same relative risk reduction as white patients treated with the beta-blockers bisoprolol, metoprolol, or carvedilol. Bucindolol, on the other hand, was associated with worse mortality outcomes in black patients than in white patients and may actually increase mortality in blacks.
Assessing treatment of asymptomatic left ventricular dysfunction. For the base-case analysis of a 55-year-old man with an ejection fraction less than 40 percent and no history of symptomatic HF, the model predicted an average life expectancy without ACE inhibitor treatment of 8.1 years and a 5-year morbidity/mortality rate of 57 percent. These results are similar to the findings of the SOLVD prevention study. Treatment with ACE inhibitors improved survival and quality-adjusted survival by 8 months compared to no treatment. The lifetime cost of care was $3,718 greater for patients treated with ACE inhibitors than for those who received no treatment, with a cost per life-year gained of $5,802 and cost per quality-adjusted life year (QALY) gained of $5,644.
We tested the robustness of our base-case findings by varying the following assumptions: patient age, the risk of death with HF, the reduction in HF incidence, the reduction in risk of death for asymptomatic patients, the probability of hospitalization if symptomatic, cost of treatment, and quality of life. Treating asymptomatic patients with ACE inhibitors provided benefit compared to waiting for symptom development and remained economically attractive (< $20,000 per QALY gained) throughout the range of every variable tested.
Assessing screening for reduced left ventricular ejection fraction. For a population of asymptomatic 55-year-old individuals (prevalence of depressed ejection fraction 2.7 percent) we found that screening with echocardiography provided the greatest benefit but at a substantial cost. A strategy of initial screening with BNP followed by echocardiography improved outcome at a cost of only $18,300 per QALY gained compared to no screening. If quality of life is ignored, BNP screening costs $19,000 per life-year gained compared to no screening. The number needed to screen is 77 to gain 1 year of life and 70 to gain one QALY.
Because the cost-effectiveness ratio of screening with the ECG compared to no screening was greater than the ratio for BNP compared to ECG screening, the former strategy was eliminated as a possible screening option for the base-case cohort. Similarly, strategies of relying only on the ECG or BNP to determine treatment were eliminated, because they were more costly and provided fewer QALYs than the strategy using BNP followed by echocardiography.
We tested the robustness of our base-case findings by varying each of the following assumptions: prevalence of depressed left ventricular function, test characteristics of BNP, cost of testing, and impact of ACE inhibitors for patients with depressed ejection fraction. The decision to screen is influenced primarily by the prevalence of depressed ejection fraction and the accuracy of the screening tests and only slightly by the costs of screening, including echocardiography and BNP testing.
The following clinical conclusions can be reached from this evidence report. The evidence supported beneficial reductions in all-cause mortality with the use of beta-blockers in men and women, the use of ACE inhibitors in white and black patients, and the use of either drug in patients with diabetes.
We did, however, find evidence that suggests that women with asymptomatic left ventricular dysfunction may not have reduced mortality when treated with ACE inhibitors. The evidence we found does not constitute proof, and additional evidence of the effect of ACE inhibitors in women with asymptomatic left ventricular dysfunction is needed.
We also found conflicting evidence regarding the effect of beta-blocker use in black patients. Results of three of the beta-blocker studies suggested that white patients and black patients have similar reductions in all-cause mortality when treated with beta-blockers. However, the one study that assessed the beta-blocker bucindolol reported a statistically significant adverse effect on mortality in blacks relative to whites. These results suggest that not all beta-blockers have equivalent effects.
In our cost-effectiveness analyses, we found that treatment of asymptomatic left ventricular dysfunction with ACE inhibitors was cost-effective under virtually all assumptions, with typical costs of between $5,000 and $10,000 per QALY gained. Thus, this treatment is much more cost-effective than many other treatments considered standard medical practice. The demonstration of cost-effectiveness for treatment prompted an additional analysis to assess the cost-effectiveness of screening. This analysis showed that screening with BNP followed by echocardiography in a cohort of asymptomatic 55-year-old individuals was also cost-effective compared with other management strategies currently considered standard medical care. This strategy cost $19,000 per life year gained compared to a strategy without screening, with the number needed to screen equal to 77 to gain 1 year of additional life. These results were only modestly sensitive to cost and were most sensitive to the prevalence of asymptomatic depressed left ventricular ejection fraction. When the prevalence falls below about 1 percent, a strategy of screening becomes less cost-effective than commonly accepted thresholds for cost-effective care.
The findings of this evidence report suggest several important areas for future research.
Additional data are needed to support or refute the evidence that various beta-blockers may influence all-cause mortality differently in black patients. New placebo-controlled randomized clinical trials of beta-blocker therapy in black patients are likely the only way to answer this question definitively. Future studies of new or different beta-blocker drugs for heart failure need to include sufficient numbers of black patients to separately assess outcomes in this population, because a similar effect in black patients and white patients cannot be assumed.
Further assessment of the effect of ACE inhibitors is needed in women with HF, particularly the effect on women with asymptomatic left ventricular dysfunction. It may be possible to answer this question by a more complete assessment of data from existing randomized clinical trials.
Other outcomes of interest, including cardiac mortality, symptoms, and health care utilization, should be examined for all patient sub-populations. Individual patient-level data from the major randomized controlled trials may be sufficient to answer these and other original key questions regarding additional patient subpopulations (such as the aged and those with renal failure).
An additional implication of our findings is that researchers have not paid attention to ensuring that sufficient numbers of patients in important clinical subpopulations are enrolled in randomized trials. Such attention could obviate the need for future meta-analyses such as the ones on which this report is based.
If further research supports our findings of differential efficacy, additional research aimed at elucidating the cause for these findings should be undertaken. One possibility is that these findings do not represent differences in men and women or black patients and white patients, but rather reflect differing efficacy of these drugs according to the cause of heart failure (e.g. ischemic or nonischemic), which then may differ by sex or race. Alternatively, there could be a molecular basis for these results that differs by sex and race.
Given the robust evidence of benefit for ACE inhibitors and beta-blockers in reducing mortality, future work should also address how to improve the use of these therapies by focusing on potential barriers for practitioners and patients as well as empirically testing the conclusions of our cost-effectiveness analyses. Additional studies are needed to determine the true prevalence of asymptomatic left ventricular dysfunction, and to determine costs associated with making a new diagnosis of heart failure. Further research is needed to determine which patient characteristics identify a population at risk for left ventricular systolic dysfunction (prevalence greater than 1 percent). In addition, a study evaluating the health and economic outcomes of screening asymptomatic patient with BNP is warranted.
Heart failure (HF) is a clinical syndrome that can result from any cardiac disorder that impairs the ability of the ventricles to fill with and/or eject blood. The syndrome is characterized by signs and symptoms of intravascular and interstitial volume overload, which include shortness of breath and edema and/or manifestations of inadequate tissue perfusion, such as fatigue or poor exercise tolerance.
HF is a common medical condition that has a significant impact on public health. In the United States, an estimated 4.8 million individuals are affected by HF, and 400,000 to 700,000 new cases develop each year.1 The prevalence of HF increases with age: It is present among 2% of persons age 40 to 59, more than 5% of persons age 60 to 69, and 10% of persons age seventy or older. In addition, a substantial number of individuals with asymptomatic ventricular dysfunction are at risk of developing symptomatic HF. Due to the aging of the American population, the incidence and prevalence of this disease are expected to increase markedly.
HF is associated with substantial morbidity and mortality; it is a primary or secondary cause of death for approximately 250,000 people per year in the United States. According to the 2002 Heart and Stroke Statistical Update (www.americanheart.org), HF was the first-listed diagnosis for 962,000 hospitalizations in 1999, and it is the most common diagnosis among hospital patients age 65 and older. In fact, 20% of all hospitalizations in this age group carry a primary or secondary diagnosis of HF. Over three million outpatient office visits each year are related to this disease. In 1998 alone, the estimated annual direct cost due to HF was $18.8 billion.
A series of studies has established that angiotensin-converting enzyme inhibitors (ACE inhibitors) and beta-adrenergic blocking agents (beta-blockers) provide life-saving benefits in patients with HF or left ventricular systolic dysfunction. However, most of the patients enrolled in such studies have been white males. A clinical question that is consistently asked is whether or not the mortality benefit reported in these clinical trials is also achieved for other subpopulations, such as women, people of other races, and patients with particular comorbidities such as diabetes mellitus or renal insufficiency.
There are several reasons to expect that certain subpopulations might not achieve the same benefits as white males. Research evidence supports a lesser effect on blood pressure in black compared to nonblack hypertensive patients treated with ACE inhibitors,2 and one of the ACE inhibitor trials reported a lesser effect of ACE inhibitors on reducing hospitalization for black compared to nonblack patients.3 Similarly, men and women present and respond differently to particular cardiac therapies. Relevant to this topic, a preliminary analysis of one ACE inhibitor study suggested a trend toward lower mortality reduction in women than in men.4 Since few of the randomized trials enrolled enough women, blacks, or patients with comorbidities to have sufficient statistical power to support conclusions based on subgroup analysis, this question is appropriate for meta-analysis. In additional, if the clinical trial data support a mortality benefit for patients with asymptomatic left ventricular dysfunction, it is natural to question both the cost-effectiveness of such treatment and whether screening asymptomatic patients for left ventricular dysfunction is cost-effective. These clinical and policy questions form the basis for this report.
AHRQ described the scope of work as a quantitative analysis and evidence report on the effectiveness of treatment of HF using ACE inhibitors and beta-blockers. The project had five key steps:
Identify technical experts to provide input and advice to the project.
Refine the research questions.
Perform a literature search and evaluation.
Systematically synthesize the evidence.
Produce and disseminate and evidence report.
The American College of Physicians, the American Society of Internal Medicine, and the American Academy of Family Physicians nominated this topic. They submitted the following potential key questions to AHRQ.
What evidence exists on the effectiveness of nurse management programs? Health food supplements?
What evidence exists on the treatment of sleep apnea in patients with HF?
What is the evidence on the treatment of specific myocardial disorders; e.g., myocarditis, sarcoidosis, and amyloidosis, in patients with HF?
What interventions are effective for patients with diastolic dysfunction?
Which patients benefit from which beta-blockers?
What are the effects of potassium levels on HF outcomes?
Do angiotensin blockers improve outcomes?
What, if any, are the differences in treatment effectiveness associated with patient gender, race, age, and income level?
Project staff assembled a technical expert panel (TEP) that included leading cardiologists working in academic and nonacademic settings, researchers, clinicians, and health care managers. Panelists assisted the project with topic refinement, retrieval of unpublished data, and review of the final evidence report. The TEP members (and relevant affiliations) are listed here:
Michael Barrett American College of Physicians
Greg Fonarow UCLA Medical Center
Barry Greenberg UCSD Medical Center
Paul Heidenreich.Palo Alto VA Hospital
Stanford-UCSF Evidence-based Practice Center
Tom Knabel UnitedHealthcare
Marvin Konstam New England Medical Center
Michael Rich Washington University of School of Medicine
Anthony Steimle Kaiser Permanente, Northern California
Lynne Warner Stevenson Brigham and Women's Hospital
After “congestive heart failure” was nominated as a topic, but prior to assignment of this contract to the Southern California Evidence-based Practice Center (SCEPC), the American Heart Association (AHA) and the American College of Cardiology (ACC) released practice guidelines on the management of HF. AHA/ACC graciously provided the SCEPC a draft copy for confidential review. On September 8, 2000, a conference call was held with our technical expert panel to limit the key questions to be addressed in the evidence report. The purpose of the conference call was to identify topic areas for this report that would complement but not duplicate the draft guidelines, a copy of which had been made available to each TEP member. The technical experts judged that several of the original key questions posed by the nominating organizations had been adequately answered in the AHA/ACC guidelines, major studies were underway that would answer several more of the questions, and published data would be insufficient to reach meaningful conclusions for still others. The technical experts identified three areas where they believed significant contribution could still be made:
Assessment of the effects of age over 70, gender, race, and assisted living on treatment outcomes
Cost-effectiveness of medication combinations
Assessment of outcomes in patients with various comorbidities, particularly diabetes mellitus, renal dysfunction, and cognitive dysfunction.
Our TEP members determined that for clinical questions 1 and 3, only the results of placebo-controlled randomized trials (RCTs) of ACE inhibitors or beta-blockers that measured outcomes of interest to patients and policymakers (including mortality, utilization, and costs) would be accepted as evidence. The TEP judged that a formal explication of a causal pathway was not needed, because numerous randomized trials had already addressed the overarching clinical questions of the effect of the drugs on mortality and utilization. As a starting point for our research, our experts provided us with references to eight pertinent studies and the names and acronyms of the major ACE inhibitor and beta-blocker trials.
In addition to the eight reports provided by the expert panel, we searched the following databases for articles on HF treatment for the specific populations under study.
Medline, produced by the U.S. National Library of Medicine, is widely recognized as the premier source for bibliographic coverage of biomedical literature. It encompasses information form Index Medicus, Index to Dental Literature, and the Cumulative Index to Nursing and Allied Health Literature as well as other sources of literature in the areas of allied health, biological and physical sciences, humanities and information science as they relate to medicine and health care.
Healthstar, produced by the American Hospital Association, contains over one million references covering topics in hospital administration, personnel, planning, budget, accreditation, and health care delivery.
EmBase, the Excerpta Media database produced by Elsevier Science, is a major biomedical and pharmaceutical database indexing over 3,800 international journals. EMBASE currently contains over six million records, with more than 400,000 citations and abstracts added annually.
Ageline covers subjects that include aging, gerontology, health sciences, psychology, and sociology. References date from1978 to the present.
SciSearch is a database that contains all records published in Science Citation Index and additional records from about 1,000 journals listed in Current Contents. Every subject area within the board fields of science, technology, and biomedicine is included.
The Cochrane DARE (Database of Abstracts and Reviews of Effectiveness) contains structured abstracts of systematic reviews that have been critically appraised by reviewers at the Centre for Reviews and Dissemination, York, UK.
| Ace inhibitors and Beta Blockers for congestive heart failure |
| Note: A question mark after a term indicates that the term was truncated. An exclamation point after a term indicates that the term was “exploded” - I.E. all narrower terms in the subject hierarchy were included. This feature is active in the MEDLINE, HealthSTAR, and EMBASE databases. |
| SEARCH #1A |
| DATABASES SEARCHED: |
 MEDLINE |
| HealthSTAR |
| EMBASE |
| Ageline |
| Social SciSearch |
| SciSearch |
| TIME PERIOD COVERED: 1985-2000 |
| SEARCH STRATEGY: |
| angiotensin-converting enzyme inhibitors OR dipeptidyl carboxypeptidase inhibitor! OR captopril OR enalapril OR fosinopril OR lisinopril OR quinapril OR ramipril OR beta-adrenergic receptor blocking agent! OR adrenergic beta-antagonists OR carvedilol OR metoprolol OR bisoprolol OR bucindolol OR beta block? AND heart failure, congestive! OR congestive heart failure! OR heart failure AND black? OR negro? OR afr? american? OR aged! OR geriatric assessment OR aged/ti,de OR elderly care! OR geriatric? OR gerontolog? OR nursing homes! OR nursing home OR home(s) for the aged OR homes for the elderly OR institutionalized elderly OR Female? OR female! OR women! OR wom?n OR gender OR women's health OR women's health services OR kidney failure! OR kidney? fail? OR renal fail? OR kidney disease? OR renal disease? OR kidney dysfunction? OR renal dysfunction? OR diabetes mellitus! OR diabetes OR cognition disorders! OR disorders of higher cerebral function! OR thought disorder OR alzheimer disease OR alzheimer? OR senil? OR dementia! OR dementia OR (drug therapy, combination OR drug combinations OR drug? combin? OR pharmac? combin? OR polypharmac?) AND (econom? OR cost?) OR drug combination!-economic aspect! AND Document type=randomized controlled trial OR random? (COMBINED WITH ALL TOPICS EXCEPT NURSING HOMES AND DRUG COMBINATIONS) |
| TOTAL NUMBER OF TITLES RETRIEVED: 1391 |
| SEARCH #1B: |
| NOTE - THIS SEARCH WAS PERFORMED TO CORRECT AN OMISSION OF THE TERM “ACE INHIBITOR(S)” FROM THE SOCIAL SCISEARCH AND SCISEARCH DATABASES IN SEARCH 1A. |
| DATABASES SEARCHED AND TIME PERIOD COVERED: |
| Social SciSearch 1972-2001 |
| SciSearch (Archival) - 1974-1989 |
| SciSearch (Current) - 1990-2001 |
| SEARCH STRATEGY: |
| ace inhibitor? NOT (captopril OR enalapril OR fosinopril OR lisinopril OR quinapril OR ramipril OR trandolapril) AND heart failure AND black? OR negro? OR afr? american? OR aged/ti,de OR elderly OR geriatric? OR gerontolog? OR nursing home? OR home? for the aged OR home? for the elderly OR institutionali? aged OR institutionali? elderly OR Female? OR wom?n OR gender OR kidney? fail? OR renal fail? OR kidney disease? OR renal disease? OR kidney dysfunction? OR renal dysfunction? OR diabetes OR alzheimer? OR senil? OR dementia OR drug? combin? OR pharmac? combin? OR polypharmac?) |
| TOTAL NUMBER OF TITLES RETRIEVED: 256 |
Paul Shekelle, MD, and Colonel Sid Atkinson, MD, reviewed the list of retrieved titles. Of the 1,647 titles retrieved, 315 articles were deemed relevant to our undertaking and were ordered. An additional 88 articles found through mining reference lists were also ordered. Literature was tracked using Pro-Cite and Access software.
The TEP made us aware that reports of several recent studies were in press and thus would not be found through a search. Prepublication copies were provided to us.
In hopes of obtaining data on all ACE inhibitors and beta-blockers approved for HF by the Food and Drug Administration (FDA), we requested filings for each of these agents through the Freedom of Information (FOI) act. Approved ACE inhibitors included captopril, enalapril, fosinopril, lisinopril, quinipril, ramipril, and trandolapril. Approved beta-blockers included bucindolol, bisoprolol, carvedilol, and metoprolol. As discussed in the Results section, we eventually obtained data from the FDA for two studies.
Another TEP conference call was held on April 4, 2001. During this phone call, we reviewed the preliminary results of our literature search. The TEP advised us to attempt to obtain subgroup data on all RCTs that had at least 12 weeks of followup. Since most published studies did not report on our special populations of interest, project staff sent letters to original authors requesting subgroup data (see Appendix A for sample letter). Nonrespondents were sent a reminder letter on May 8, 2001. In addition, expert panel members agreed to call or email selected nonrespondents.
Our yield from this process was poor. After mailing 62 letters, we netted four agreements (all from studies with relatively small sample sizes), 12 new contacts, 10 refusals, 32 nonresponses, and four responses categorized as other.
Based on this poor response, we modified our plan to seek subgroup data more intensively from the biggest studies through personal contacts by TEP members to the authors of those studies and through attempts to obtain individual patient data on any study that had been submitted to the FDA as part of the regulatory process. Our rationale was that we had enough resources to attempt these intensive methods on only a select number of studies and the biggest studies would provide us the greatest statistical power. We calculated that the seven largest ACE inhibitor studies enrolled 14,932 patients, whereas the remaining 19 ACE inhibitor trials enrolled an aggregate of 3,033 patients. Similarly, the five largest beta-blocker studies enrolled 12,726 patients, whereas the remaining 19 beta-blocker studies enrolled 2,938 patients. Therefore, by targeting our efforts at the largest studies, we were able to make the most effective use of our resources. However, this strategy assumes that the large and small studies are measuring the same effect.
| ACE Inhibitor Trial Name | Outcome of Contact |
| AIRE | No response to letters. |
| CONSENSUS | Individual patient data obtained from Dr. Steve Snapinn at Merck. |
| SAVE | No response to letters. |
| SMILE | No response to letters. |
| SOLVD | Individual patient data obtained from Dr. Steve Snapinn at Merck. |
| TRACE | Individual patient data obtained from Dr. Christian Torp-Pederson. |
| Beta-blocker Trial Name | Outcome of Contact |
| BEST | Request refused. |
| CIBIS-II | No response to letters. |
| COPERNICUS | Study released while report was in progress. |
| Individual patient data made available to us from the FDA when we went to access MERIT-HF. | |
| MERIT-HF | Refused, “manuscript in process,” subsequently published. |
| Individual patient data obtained from FDA. | |
| U.S. Carvedilol | Agreed in principle; however, individual patient data never obtained. |
| ACE Inhibitor Trial Name | Source of Data |
| AIRE | Published data |
| CONSENSUS | Individual patient data |
| SAVE | Published data |
| SMILE | Published data |
| SOLVD | Individual patient data |
| TRACE | Individual patient data |
| Beta-blocker Trial Name | Source of Data |
| BEST | Published data |
| CIBIS-II | Published data |
| COPERNICUS | Individual patient data |
| MERIT-HF | Individual patient data |
| U.S. Carvedilol | Published data |
During the data extraction phase, it became apparent that few studies reported the relevant data stratified by age or nursing home residence. In addition, health care outcomes and health outcomes other than mortality were reported variably in the studies, making pooling less justified. For these reasons, we further restricted key questions 1 and 3 to assess only data stratified by gender, race, and diagnosis of diabetes, and to use all-cause mortality as the sole outcome of interest.
Our principal questions for meta-analysis, as determined by our TEP, were the following:
What is the association between treatments (ACE inhibitors or beta-blockers) and all-cause mortality for female patients, male patients, patients with diabetes, patients without diabetes, black patients, and white patients with HF?
Does this association vary (e.g. are there statistically significant differences) by gender (female versus male), diabetic condition (those with diabetes versus those without), and race (black versus white patients)?
Because individual studies did not enroll sufficient number of patients in the sub-populations of interest, meta-analysis is an appropriate technique to consider for these questions.
We first retrieved all articles that pertained to the eleven large placebo-controlled studies on ACE inhibitors and beta-blockers mentioned above. The SOLVD study consisted of two distinct trials on prevention and treatment respectively; thus, we considered a total of twelve studies. The same meta-analysis was done separately for the ACE inhibitor and beta-blocker sub-populations of studies, respectively.
Our outcome of interest was all-cause mortality. For studies for which both patient-level data and published statistics were available, we always chose the patient-level data over published statistics in the event of disagreement. Among the five studies for which we had patient-level data, three datasets disagreed with related publications. The differences were extremely small, never more than two patients in particular sub-populations; for example, the number of nondiabetic patients in the published article was two fewer than in the patient-level dataset. For the studies for which we had only published data, no two articles presented conflicting data about the same patient subgroup.
All published reports that included the relevant patient subgroup data presented those data in the form of a two-by-two table of all-cause mortality by treatment or placebo group for each subgroup separately. If the patient-level data were available, we could construct this table directly. For example, the report of an ACE inhibitor study that stratified data by gender would provide the data in separate two-by-two tables, one for women and one for men. For each subgroup (e.g., women), we estimated the log mortality relative risk, which is equal to the log of the risk of dying for women who received ACE inhibitors divided by the risk of dying for women who received placebo. The extraction of data from patient-level datasets is described below.
The standard error for the log relative risk was also estimated,5 and a 95% confidence interval was constructed. A similar log relative risk and confidence interval were calculated for men. We then back-transformed to the unlogged scale for interpretability so that our final statistic for each subgroup in each study was the relative risk with its associated confidence interval. The reason for conducting the estimation on the log scale is that the variance is more stable and the errors are more symmetric in this metric.
For subgroup comparisons for which we had data from more than two studies, we pooled the logs of the relative risks across studies using the DerSimonian and Laird random effects model,6 and back-transformed the pooled estimate to the unlogged scale to produce a pooled relative risk (e.g., for women) across all relevant studies. We also constructed a 95% confidence interval and provide a p-value for the test of whether the pooled relative risk is different from 1. We tested for heterogeneity using a chi-squared test.7 We note that in the case when sufficient heterogeneity across studies is not found, the DerSimonian and Laird estimate of the between-study variance is 0, and the random effects estimate is the same as a fixed effects estimate, the latter incorporating only within-study variance. Significant heterogeneity was not observed for almost all our beta-blocker pooling situations, indicating that the studies were not heterogeneous, though we acknowledge that the chi-squared test of heterogeneity has low power to detect differences across studies, and the DerSimonian and Laird estimate is only a one-step iterative method. For ACE inhibitor studies, there was substantial heterogeneity, and the random effects analysis is designed to take this into account. This meta-analysis and the ones described below were conducted in the statistical package Stata using the “meta” and associated commands.8 The analysis just described informed us about the association between various patient characteristics (such as gender) and mortality, when association is measured on the relative risk scale. Thus, this analysis answered our first question of interest.
To answer our second question, that is, whether the association differed between sub-populations (e.g., female versus male), we needed to test whether the relative risks of the two subgroups were statistically different. We did this by constructing a test statistic equal to the ratio of relative risks (RRR), which (for the example given) equals the female relative risk divided by the male relative risk. If this test statistic differs significantly from 1, then we infer that the two subgroup relative risks are significantly different. As before, we performed the analysis on the log scale. The log ratio of relative risks equals the log of the relative risk for women divided by the relative risk for men, and its standard error equals the square root of the sum of the variances of the two log relative risks. We constructed a confidence interval on the log scale. We then back-transformed the estimate and its confidence interval to the unlogged scale so that our final test statistic for each study was the RRR.
For subgroup comparisons for which we had data from more than two studies, we pooled the logs of the RRRs across studies using the DerSimonian and Laird random effects model.6 We back-transformed the pooled result to the RRR scale for interpretation, and present the pooled ratio of relative risks, its 95% confidence interval, and a p-value for the test of whether the pooled RRR is different from 1.
We note that the ratio of the pooled relative risks may not exactly equal the pooled ratio of relative risks due to the nature of the weighting. The reason for pooling of the RRRs in order to compare the relative risks, rather than pooling the relative risks separately in each subgroup and then taking the ratio, is that comparison (i.e., taking the ratio) should be done separately within each study to control for study differences.
The directions (definitions of the numerator and denominator) of the RRRs were as follows. For the effect of gender, we compared outcomes for women (numerator) versus those for men (denominator). For the effect of diabetes we compared those who had diabetes with those who did not. For the effect of race, we compared black patients to white patients if the data were stratified appropriately. If not, we compared black patients to nonblack patients, or, if necessary, we compared nonwhite patients to white patients. We conducted a sensitivity analysis as described below to assess this hierarchical approach and to determine whether the inconsistency of race classification across studies affected our conclusions.
Followup times for outcome assessment varied across studies, and the relative risk calculations do not take this variation, or the censoring of observations, into account. Thus, we also assessed the mortality associated with ACE inhibitors and beta-blockers on the hazard ratio scale. The hazard ratio accounts for the variable contribution made to followup by patients who dropped out of the study for whatever reason. We followed the strategy for data extraction and pooling as described in Parmar, Torri, and Stewart.9 The majority of the studies included in our analysis presented hazard ratios and confidence intervals, and after transforming these statistics to the log scale, we extracted the log hazard ratio and its standard error for each study.
For each patient subpopulation of interest, we estimated the log hazard ratio for each study that provided the data stratified on that dimension. We followed the same analytic strategy for the hazard ratio as for the relative risk, conducting a random-effects pooled analysis on the log scale and back-transforming to the unlogged scale. We then calculated a ratio of hazard ratios (RHR) to compare the hazard ratios in each subgroup.
We obtained data directly from patient-level datasets for five studies: CONSENSUS, COPERNICUS, MERIT-HF, SOLVD, and TRACE. For CONSENSUS, SOLVD, and TRACE, the entire patient-level files were available to us directly, and we could conduct any analyses that we wished. As described above, we constructed two-by-two tables of mortality by treatment for each subgroup of interest to estimate a relative risk and constructed a Cox proportional hazard model in SAS10 with treatment or control as the single covariate to estimate the hazard ratio for each patient subgroup of interest.
As previously mentioned, for the other two studies, COPERNICUS and MERIT-HF, we were able to analyze the patient-level data that the FDA provided. However, we were required to analyze the data at the FDA facility. The FDA allowed one of our statisticians to have access to the data at the FDA facility in Maryland. The analyst spent one day extracting and analyzing the data for both studies. The FDA provided our analyst with a computer workstation, and the data for both studies were in SAS format. The data for each study had a table of contents in a PDF file, which, along with the drug questionnaire, was used to locate the necessary variables. Once the data were compiled in a usable format for analysis, relative risks and hazard ratios were calculated for patient sub-populations. We were able to assess all-cause mortality separately from cardiac-cause mortality.
For COPERNICUS, the randomization group, gender, race, outcome status (dead or alive at the end of the trial), and time of death or dropout (i.e., censored) variables were each in separate files and had to be merged together by patient identification number. We defined the “diabetes” subgroup as any patients whose files were identified by searching the medical history text for the root “DIABET.” Two subjects who were coded as “dead” but whose files did not show dates of death were dropped from the analysis.
For MERIT-HF, an analysis file with most of our variables of interest was already available. The number of days from enrollment until death or censoring had to be calculated using either the date of death or the date of last interview.
We assessed the possibility of publication bias for the studies corresponding to each drug and patient comparison subgroup by graphically evaluating a funnel plot of the individual study log relative risk and hazard ratio for symmetry resulting from the nonpublication of small, negative studies. Because graphical evaluation can be subjective, we also conducted an adjusted rank correlation test11 and a regression asymmetry test12 as formal statistical tests for publication bias. We found no evidence of publication bias in any of the study subpopulations assessed.
As described above, studies varied in their definitions of racial groups. For the black patient versus white patient comparison, if the researchers reported data separately for blacks and whites, we utilized those data. If such data were not available, we used data reported for black versus nonblack patients, or, as a last resort, data comparing nonwhite with white patients. For those studies that provided the data for more than one of these comparisons, we compared the relative risk and hazard ratio statistics. The results of this sensitivity analysis did not differ markedly from the results of our primary hierarchical approach. We acknowledge that this sensitivity analysis cannot assess whether the potentially different race definitions (e.g., inclusion of Hispanic black patients in the Hispanic subgroup versus the black group) had an effect. However, the sensitivity analysis did permit us to evaluate some of the effects of different race definitions and stratifications across studies.
At the April 4, 2001, teleconference, Paul Heidenreich, MD, proposed to the TEP that based on his analysis of the data that were suitable for cost-effectiveness modeling, the most feasible cost-effectiveness analysis would be that of the use of ACE inhibitors for asymptomatic left ventricular systolic dysfunction, rather than an analysis of the cost-effectiveness of combinations of medications, as was originally proposed. This plan was accepted by the TEP and approved by the Task Order Officer. Later, based on the findings of this analysis, a further cost-effectiveness analysis that assessed screening for left ventricular dysfunction was proposed and approved by the Task Order Officer.
The square node represents the decision to treat asymptomatic patients with low ejection fraction with ACE inhibitors. All patients start out in the asymptomatic state. For each time period, there is a chance they can develop symptoms of heart failure or die. Once they develop symptoms, they move to the symptomatic tree, where there is a chance of remaining alive with heart failure or dying.
). In the first strategy, asymptomatic patients are treated with ACE inhibitors. In the second strategy, patients are not treated with ACE inhibitors until they develop HF.
Each time period (month), patients with no history of HF can remain asymptomatic, develop HF, or die. Of those patients who developed HF, we assumed 33% would be hospitalized during their initial episode.3 Once patients develop HF, they can remain in stable HF, be hospitalized, or die during each time period. The model follows patients until each has died (or to age 120).
Published data from the SOLVD prevention trial were used to calculate rates for the development of HF and death for asymptomatic patients with and without ACE inhibitor treatment.3 We used actual event rates during the four years of reported followup. To model outcome after four years, we used an average of the yearly event rates weighted by the number of subjects still enrolled during each year of followup. Using this method, we estimated that the yearly rate of progression to symptomatic HF would be 6.5% for patients treated with ACE inhibitors and 9.8% for those not treated. We used a similar method to determine the yearly relative risk of death (compared to the general population) for patients with asymptomatic left ventricular dysfunction who are treated (2.9) and those not treated (3.3) with ACE inhibitors.
We used data from the SOLVD treatment trial to estimate hospitalization and death rates for patients with HF treated with ACE inhibitors.3 The data consisted of actual event rates during the four years of reported followup for the SOLVD treatment trial. To model outcome following four years of living with HF, we used an average of the annual event rates weighted by the number of subjects participating during each year of the trial. This method estimated that the yearly relative risk of death (compared to the general population) for patients with symptomatic left ventricular dysfunction was 6.5 when treated with ACE inhibitors.
To determine quality-adjusted survival, we assigned a utility value of 0.71 to each year of life for patients living with HF, based on prior studies using the time-tradeoff utility of patient preferences in HF.13 Patients with asymptomatic left ventricular dysfunction were assumed to have a utility value of 0.87.13 We varied these quality-of-life assumptions in sensitivity analysis (range 0.5 to 1).
| Variable | Baseline Value (Plausible Range) | Discussion |
|---|---|---|
| Age at beginning of model | 55 | Mean age 59 in SOLVD trials |
| Probability of death during year19 (general population) | (by age) | |
| Relative risk of death compared to general population for: | ||
| Asymptomatic untreated patients3,25 | 3.3 (1.0–4.0) | |
| Asymptomatic patients treated with ACE inhibitors3,25 | 2.9 (1.0–4.0) | |
| Symptomatic patients treated with ACE inhibitors24,25 | 6.5 (2.0–8.0) | |
| Probability of transition during year from asymptomatic to symptomatic for: | Weighted average | |
| Untreated patients3 | 9.8% (5.0%–15.0%) | |
| Treated patients3 | 6.5% (5.0%–10.0%) | |
| Probability of hospitalization per year if symptomatic26 | 11% (5%–30%) | Weighted average |
| Probability of hospitalization during first episode of heart failure3 | 33% (10%–50%) | |
| QALY weights for: | Time-trade-off method, collected from general population, and visual analog scale from SOLVD study. | |
| Asymptomatic patients13,27 | 0.865 (0.68–1.0) | |
| Symptomatic patients13,27 | 0.710 (0–1.0) | |
| Cost of ACE inhibitor16,16 | $551 ($200–$1,000) | From average wholesale price of various manufacturers |
| Cost of hospitalization17 | $9,425 ($7,000–$13,000) | Baseline value for DRG 127 (Heart Failure and Shock) from MGH analysis; range includes mean of hospital charges to Medicare |
| Cost of outpatient CHF management28 | $470 ($200–$800) | Costs applied to AHA utilization survey findings |
| Cost of general health expenditures14 | (by age) | |
| Deflation rate | (by year) | Consumer Price Index, Medical Care |
| Annual discount rate for costs and utilities18 | 3% (0%–6%) | |
The square node represents the decision to screen asymptomatic patients to identify those with low ejection fraction. Each test or combination of tests will result in a true positive, false positive, true negative, or false negative finding. Only those with positive tests are treated to prevent heart failure (e.g., with ACE inhibitors).
):
Echocardiography for all patients. Patients with an ejection fraction less than 35% are treated (ACE inhibitors) to prevent development of HF.
Electrocardiogram (ECG) first, and if abnormal, echocardiography.
Blood test for B-type Natriuretic Peptide (BNP) first and, if abnormal, echocardiography.
ECG only, with treatment based on the results.
BNP only, with treatment based on the results.
No screening for depressed left ventricular function.
Each test can provide one of four possible results (true positive, false positive, true negative, false negative). Only persons who are true or false positives are referred for treatment. True-positive patients have a higher quality-adjusted survival than false negatives, who are treated only when HF develops. True-negative patients have a normal age-specific life expectancy. False-positive patients receive a small decrement in quality-adjusted survival to account for potential side effects of treatment.
A decision model was developed using EXCEL (Version 5.0, Microsoft Corporation, Redmond, WA) and DATA (Version 3.0, TreeAge Software, Boston, MA) software. We obtained the lifetime health and economic outcomes for hypothetical cohorts of 55-year-old patients with (1) depressed ejection fraction (35% or less) but no history of HF treated with ACE inhibitors, (2) depressed ejection fraction but no history of HF and no treatment until HF developed, and (3) patients with heart failure but without depressed ejection fraction.
During each time period (month), patients with a low ejection fraction and without a history of HF can remain asymptomatic, develop HF, or die. Of those patients who developed HF, we assumed 33% would be hospitalized during their initial episode.3 Once patients develop HF, they can remain in stable HF, be hospitalized, or die during each time period. The model follows each patient until death (or until age 120). Patients without depressed ejection fraction are assumed to have an average age-specific mortality based on U.S. life table data.19
| Variable | Baseline Value (Range) | Comment |
|---|---|---|
| Cost of BNP | $29 (10–100) | Biosite price of 25 tests for $725 |
| Cost of ECG | $31 (10–100) | Medicare reimbursement 2001 |
| Cost of two-dimensional echocardiography | $213 (100–1000) | Medicare reimbursement 2001 |
| Lifetime cost of care: | From Model of ACEi treatment for low ejection fraction (see appendix) | |
| Normal EF | $51,898 ($25,00–100,000) | |
| Low EF on ACE inhibitor | $40,701 ($20,000–80,000) | |
| Low EF, not on ACE inhibitor until CHF | $36,983 ($20,000–80,000) | |
| Survival: | From Model of ACEi treatment for low ejection fraction (see appendix) | |
| Normal EF | 14.96 (10–20) | |
| Low EF on ACE inhibitor | 8.76 (5–15) | |
| Low EF, not on ACE inhibitor until CHF | 8.12 (5–15) | |
| Quality-adjusted survival: | From Model of ACEi treatment for low ejection fraction (see appendix) | |
| Normal EF | 12.94 (10–20) | |
| Low EF on ACE inhibitor | 7.22 (5–15) | |
| Low EF, not on ACE inhibitor until CHF | 6.56 (3–15) | |
| Sensitivity BNP 20 | 0.89 (0.5–1.0) | Threshold of 17.9 ng/ml (European assay) |
| Specificity BNP20 | 0.71 (0.5–1.0) | |
| Sensitivity ECG21 | 0.60 (0.5–1.0) | |
| Specificity ECG21 | 0.82 (0.5–1.0) | |
| Sensitivity echocardiography22 | 0.87 (0.8–1.0) | Assumes nuclear angiography is the gold standard |
| Specificity echocardiography22 | 0.76 (0.8–1.0) | |
| Prevalence of low ejection fraction | 2.7% | Among patients >55 years of age, assumes 50% of low EF are asymptomatic |
| Deflation rate | (by year) | Consumer Price Index, Medical Care |
| Annual discount rate for costs and utilities18 | 3% (0%–6%) | |
Rates for the development of HF and death for asymptomatic patients with and without ACE inhibitor treatment were based on using published data from the SOLVD prevention trial.3 We used actual event rates during the four years of reported followup. To model outcome after four years, we used an average of the yearly event rates weighted by the number of subjects still enrolled at each year of followup. Using this method, we estimated that the yearly rate of progression to symptomatic HF would be 6.5% for patients treated with ACE inhibitors and 9.8% for those not treated. We used a similar method to determine the yearly relative risk of death (compared to the general population) for patients with asymptomatic left ventricular dysfunction who are treated (2.9) or not treated (3.3) with ACE inhibitors.
We used SOLVD treatment trial data to estimate hospitalization and death rates for patients with HF treated with ACE inhibitors.24 These data were actual event rates during the four years of reported followup for the SOLVD treatment trial. To model outcome following four years of living with HF, we used an average of the yearly event rates weighted by the number of subjects participating during each year of the trial. This method estimated that the yearly relative risk of death (compared to the general population) for patients with symptomatic left ventricular dysfunction was 6.5 when treated with ACE inhibitors.
To determine quality adjusted survival we assigned a utility value of 0.71 to each year of life for patients living with HF based on prior studies using the time-tradeoff utility of patient preferences in HF.13 Asymptomatic patients were assumed to have a utility value of 0.87.13 We varied these quality assumptions in sensitivity analysis (range 0.5 to 1).
Because of multiple strategies, a large number of comparisons were possible. For each analysis, we first ranked the strategies by increasing effectiveness. We then compared the cost-effectiveness between the most effective strategy and the strategy that had the next-highest effectiveness. Strategies that provided less effectiveness at a higher cost were eliminated (dominance). Strategies could also be eliminated by extended dominance if a combination of two other strategies provided greater outcomes at lower costs. For example, assume the order of effectiveness of strategies is no screening< ECG screening < BNP screening.18 If the cost-effectiveness ratio of electrocardiogram versus No Screening was greater than the cost-effectiveness ratio of BNP versus electrocardiogram, then electrocardiogram was eliminated by extended dominance. In our reporting, we excluded strategies that have been eliminated by dominance or extended dominance.
At the beginning of the project, we requested nominations from several organizations for technical experts to join a panel that would advise staff throughout the project. A total of eight nominations were received for the Technical Expert Panel (TEP). In addition, experts in systematic reviews and meta-analysis were selected from a pool of experts associated with the Southern California Evidence-Based Practice Center but not involved with this project. The Project Staff, in consultation with the Task Order Officer, and Dr. Michael Rich, chairman of the TEP, suggested additional prominent cardiologists to review the report.
A Sopy of the draft evidence report was mailed to each peer reviewer, along with an instruction sheet for reviewing the draft evidence report (sample letter and instruction sheet included in Appendix C). The Peer Reviewers were asked to respond within three weeks. The eight of the ten peer reviewers who responded are listed below:
Stephen Gottlieb University of Maryland Medical Center
Mariell Jessup Hospital of the University of Pennsylvania
Carl Leier The Ohio State University Medical Center
Robert McNamara Johns Hopkins University
Eric Peterson Duke Clinical Research Institute
Illeana Pina University Hospitals of Cleveland
Todd Seto The Queen's Medical Center
James Young Cleveland Clinic Foundation, Kaufman Center for Heart Failure
A copy of the draft evidence report was also mailed to the members of the Technical Expert Panel and all technical experts responded with comments. Upon receipt of all responses from the peer reviewers and technical experts, the project staff compiled a summary of the comments and changes, and revised the draft evidence report. We forwarded all comments to the Task Order Officer for review. The peer reviewers' and technical experts' comments are included in Appendix D, together with the corresponding responses or actions taken by project staff.
Figure 3
displays the results of our literature search. As noted previously, our TEP provided us references for nine studies. Our library search identified another 315 articles. By reviewing the reference lists of those articles as we received them, we identified an additional 88 articles to assess. Thus, in total, 412 articles were selected. Of these, we were able to obtain 392 through the RAND library, the UCLA library, and a consulting firm that specializes in locating hard-to-find scientific journals. Of the 392 articles screened, 174 reported the results of randomized, controlled trials (RCTs) of beta-blockers or ACE inhibitors; these progressed to the Quality Review stage (see forms, Appendix B). Of these 174, 100 were rejected because they were not placebo controlled, did not report mortality outcomes, or did not report outcomes for a minimum of 12 weeks followup. This review process left 74 articles (see Evidence Tables 1 and 2).
As mentioned in Chapter 2, many of these articles described studies that appeared to include (but did not stratify according to) our populations of interest—blacks, women, and diabetics. Thus, we attempted to correspond with the authors of all studies accepted (randomized controlled trials of beta-blockers or ACE inhibitors reporting mortality data, with a minimum of 12 week followup) in an attempt to obtain patient-level data. Of 62 authors to whom we sent letters, four agreed to send us the needed data. Ten others refused, while most others either did not reply or gave us a new contact who did not reply.
The Acute Infarction Ramipril Efficacy (AIRE) study assessed the effect of the ACE inhibitor ramipril on 1,986 patients with clinical evidence of heart failure after having an acute myocardial infarction. The average duration of followup was 15 months. The study reported a statistically significant reduction in all cause mortality with a relative risk of 0.73 for patients treated with ramipril. 29 Some subgroup analyses were also included.
The Cooperative North Scandinavian Enalapril Survival Study (CONSENSUS) assessed the effect of the ACE inhibitor enalapril in 253 patients with severe heart failure (New York Heart Association class IV). The average followup period was 188 days. The study reported that at sixth months, there was a statistically significant (40%) reduction in all-cause mortality in patients treated with enalapril.30
The Survival and Ventricular Enlargement Trial (SAVE) assessed the effect of the ACE inhibitor captopril in 2,231 patients with left ventricular dysfunction (defined as an ejection fraction of 40% or less, but without overt heart failure). The average followup time was 42 months. The study reported a statistically significant 19% reduction in all-cause mortality in patients treated with captopril. 31, 32 Subgroup analyses were also presented. 32
The Survival of Myocardial Infarction Long-Term Study (SMILE) assessed the effect of the ACE inhibitor Zofenopril in 1,556 patients who had an acute anterior myocardial infarction. The duration of followup was one year. The study reported a statistically significant 22% reduction in all-cause mortality for patients treated with Zofenopril.33 The authors also reported some subgroup analyses. Although a low left ventricular ejection fraction was not a requirement for entry into this study, our TEP judged it should be included because left ventricular dysfunction is so common following anterior myocardial infarction that the enrolled population in SMILE was sufficiently similar to the other ACE inhibitor studies to justify statistical pooling. Our test of heterogeneity supported this decision.
The Studies of Left Ventricular Dysfunction (SOLVD) contained two randomized studies of the effect of the ACE inhibitor enalapril. The first study assessed the effect in 2,569 patients with New York Heart Association Class II and III heart failure and a left ventricular ejection fraction of less than or equal to 35%.24 The average period of followup was 41.4 months. The study reported a statistically significant (16%) reduction in all-cause mortality. The second SOLVD study assessed the effect of enalapril in 4,228 patients with asymptomatic left ventricular dysfunction, defined as a left ventricular ejection fraction of 35% or less. The average followup time was 37.4 months.34 The study reported a nonstatistically significant (8%) reduction in all-cause mortality in patients treated with captopril.
The Trandolapril Cardiac Evaluation (TRACE) study assessed the effect of the ACE inhibitor trandolapril in 1,749 patients with left ventricular systolic dysfunction (defined as an ejection fraction less than or equal to 35% with or without symptoms).35 The patients were followed for 24–50 months. The study reported a statistically significant reduction in mortality (22%) for patients treated with Trandolapril.
The Beta-Blocker Survival Trial (BEST) assessed the effect of the beta-blocker bucindolol in 2,708 patients with New York Heart Association functional class III or IV and a left ventricular ejection fraction of 35% or lower.36–38 The average time of followup was two years. The study reported no overall difference in mortality between treatment and placebo groups. In a subgroup analysis, nonblack patients had a statistically significant mortality benefit with a hazard ratio of 0.82.37, 38 This benefit was counterbalanced by an unexpected nonstatistically significant higher mortality rate in black patients treated with bucindolol.
The Cardiac Insufficiency Bisoprolol II Study (CIBIS-II) assessed the effect of the beta-blocker bisoprolol in 2,647 patients with New York Heart Association class III or IV heart failure and a left ventricular ejection fraction of 35% or less.39, 40 Patients were followed up for a mean of 1.3 years. The study reported a statistically significant reduction in all-cause mortality with a hazard ratio of 0.66 for patients treated with bisoprolol. Subgroup analyses were also reported.40
The Carvedilol Prospective Randomized Cumulative Survival Study Group (COPERNICUS) assessed the effect of the beta-blocker carvedilol in 2,287 patients with severe heart failure equivalent to New York Heart Association class IV and left ventricular ejection fraction of less than 25%.41 The mean period of followup was 10.4 months. The study reported a statistically significant 35% reduction in all-cause mortality for patients treated with carvedilol. Some subgroup analyses were reported.
The Metoprolol CR/XL Randomized Intervention Trial (MERIT-HF) assessed the effect of the beta-blocker metoprolol, controlled release/extended release, in 3,991 patients with New York Heart Association functional class III to IV heart failure and left ventricular ejection fraction of 40% or less.42–44 Patients were followed for a mean of one year. The study reported a statistically significant reduction in the relative risk of mortality of 0.66 for patients treated with metoprolol. Subgroup analyses were also reported.43, 44
The United States Carvedilol Heart Failure Trials were four separate studies that assessed the effect of the beta-blocker carvedilol in patients with mild, moderate, or severe heart failure and left ventricular ejection fraction of less than 35%.45, 46 A total of 1,094 patients were studied for six months or 12 months. A pooled analysis of the four studies reported a statistically significant reduction in mortality with a relative risk of 65% for patients treated with carvedilol. Results of the subgroup analysis were also reported.46
| Study Name | Total N | N Male | N Female | Relative Risk, Male (95% CI) | Relative Risk, Female (95% CI) | Ratio of Relative Risks (95% CI) |
|---|---|---|---|---|---|---|
| CONSENSUS | 253 | 179 | 74 | 0.61 (0.44, 0.85) | 1.14 (0.68, 1.90) | 1.86 (1.01, 3.42) |
| SAVE | 2231 | 1841 | 390 | 0.80 (0.68, 0.95) | 0.99 (0.67, 1.47) | 1.24 (0.80, 1.90) |
| SMILE | 1556 | 1128 | 428 | 0.61 (0.39, 0.96) | 0.74 (0.47, 1.18) | 1.22 (0.64, 2.32) |
| SOLVD-prevention | 4228 | 3752 | 476 | 0.90 (0.77, 1.05) | 1.15 (0.74, 1.78) | 1.27 (0.80, 2.02) |
| SOLVD-treatment | 2569 | 2065 | 504 | 0.89 (0.80, 0.99) | 0.86 (0.67, 1.09) | 0.97 (0.74, 1.26) |
| TRACE | 1749 | 1248 | 501 | 0.79 (0.68, 0.91) | 0.90 (0.74, 1.11) | 1.15 (0.90, 1.48) |
| Random-effects pooled estimate | 10,213 | 2373 | 0.82 (0.74, 0.90) | 0.92 (0.81, 1.04) | 1.15 (0.99, 1.33) | |
| Study Name | Total N | N Male | N Female | Hazard Ratio, Male (95% CI) | Hazard Ratio, Female (95% CI) | Ratio of Hazard Ratios (95% CI) |
|---|---|---|---|---|---|---|
| AIRE | 1986 | 1461 | 525 | 0.76 (0.60, 0.97) | 0.69 (0.49, 0.97) | 0.91 (0.60, 1.38) |
| CONSENSUS | 253 | 179 | 74 | 0.44 (0.28, 0.68) | 1.08 (0.54, 2.15) | 2.45 (1.08, 5.59) |
| SMILE | 1556 | 1128 | 428 | 0.59 (0.36, 0.96) | 0.70 (0.40, 1.22) | 1.19 (0.57, 2.48) |
| SOLVD-prevention | 4228 | 3752 | 476 | 0.88 (0.75, 1.04) | 1.15 (0.71, 1.85) | 1.31 (0.79, 2.16) |
| SOLVD-treatment | 2569 | 2065 | 504 | 0.84 (0.73, 0.97) | 0.84 (0.62, 1.13) | 1.00 (0.72, 1.39) |
| TRACE | 1749 | 1248 | 501 | 0.75 (0.62, 0.90 | 0.86 (0.66, 1.13) | 1.15 (0.83, 1.60) |
| Random-effects pooled estimate | 9833 | 2508 | 0.76 (0.66, 0.87) | 0.84 (0.72, 0.98) | 1.11 (0.92, 1.32) | |
and 5. The corresponding pooled random effects estimates from the six studies with hazard ratio data yielded values for men of 0.76 (95% CI: 0.66, 0.87) and for women of 0.84 (95% CI: 0.72, 0.98) (Table 7 and Figures 6 and 7). The difference in effect between men and women approached statistical significance for the RRR (p = 0.07).
These differences between the estimates of relative risk and hazard ratios are due to the inclusion in the hazard ratio analysis of the AIRE study, which reported a slight nonsignificant mortality benefit for women compared to men treated with ramipril, as opposed to the relative risk analysis, which included the SAVE study. This study reported a distinct but nonstatistically significant increase in mortality in women relative to men treated with captopril (RRR=1.24).
| Analysis | Relative Risk, Male (95% CI) | Relative Risk, Female (95% CI) | Ratio of Relative Risks (95% CI) |
|---|---|---|---|
| Prevention studies | 0.83 (0.72, 0.96) | 0.96 (0.75, 1.22) | 1.25 (0.94, 1.65) |
| Treatment studies | 0.80 (0.68, 0.93) | 0.90 (0.78, 1.05) | 1.15 (0.88, 1.51) |
and 9). These results are based on a pooled analysis that included 1,079 women in the symptomatic heart failure studies and 1,294 women in the asymptomatic heart failure studies. The evidence indicates that women with symptomatic heart failure benefit when treated with ACE inhibitors, although the benefit may be somewhat less that the benefit seen in men. However, the evidence calls into question whether or not women with asymptomatic left ventricular systolic dysfunction have any mortality benefit when treated with ACE inhibitors. These results are compatible with an earlier preliminary analysis of the SOLVD data.4 Additional data are needed to answer this question. In contrast, men clearly benefit when treated with ACE inhibitors for asymptomatic left ventricular systolic dysfunction.
| Population Mortality Risk | Risk Ratio | ||||
|---|---|---|---|---|---|
| 0.5 | 0.6 | 0.7 | 0.8 | 0.9 | |
| 0.1 | 20 | 25 | 34 | 50 | 100 |
| 0.2 | 10 | 13 | 17 | 25 | 50 |
| 0.3 | 7 | 9 | 12 | 17 | 34 |
| 0.4 | 5 | 7 | 9 | 13 | 25 |
| 0.5 | 4 | 5 | 7 | 10 | 20 |
| Study Name | Total N | N Non-diabetic | N Diabetic | Relative Risk, Non-diabetic (95% CI) | Relative Risk, Diabetic (95% CI) | Ratio of Relative Risks (95% CI) |
|---|---|---|---|---|---|---|
| CONSENSUS | 253 | 197 | 56 | 0.64 (0.46, 0.88) | 1.06 (0.65, 1.74) | 1.67 (0.93, 3.01) |
| SAVE | 2231 | 1739 | 492 | 0.82 (0.68, 0.99) | 0.89 (0.68, 1.16) | 1.09 (0.79, 1.50) |
| SMILE | 1556 | 1253 | 303 | 0.79 (0.54, 1.15) | 0.44 (0.22, 0.87) | 0.56 (0.25, 1.22) |
| SOLVD-prevention | 4228 | 3581 | 647 | 0.97 (0.83, 1.15) | 0.75 (0.55, 1.02) | 0.77 (0.54, 1.09) |
| SOLVD-treatment | 2569 | 1906 | 663 | 0.84 (0.74, 0.95) | 1.01.(0.85, 1.21) | 1.21.(0.97, 1.50) |
| TRACE | 1749 | 1512 | 237 | 0.85 (0.74, 0.97) | 0.73 (0.57, 0.94) | 0.87 (0.65, 1.15) |
| Random-effects pooled estimate | 10,188 | 2398 | 0.85 (0.78, 0.92) | 0.84 (0.70, 1.00) | 1.00 (0.80, 1.25) | |
| Study Name | Total N | N Non-diabetic | N Diabetic | Hazard Ratio, Non-diabetic (95% CI) | Hazard Ratio, Diabetic (95% CI) | Ratio of Hazard Ratios (95% CI) |
|---|---|---|---|---|---|---|
| CONSENSUS | 253 | 197 | 56 | 0.50 (0.33, 0.77) | 0.68 (0.33, 1.42) | 1.35 (0.58, 3.15) |
| SMILE | 1556 | 1253 | 303 | 0.77 (0.50, 1.18) | 0.39 (0.18, 0.884) | 0.51 (0.21, 1.22) |
| SOLVD-prevention | 4228 | 3581 | 647 | 0.96 (0.81, 1.14) | 0.73 (0.52, 1.03) | 0.76 (0.52, 1.12) |
| SOLVD-treatment | 2569 | 1906 | 663 | 0.79 (0.68, 0.92) | 0.98 (0.78, 1.24) | 1.24 (0.94, 1.64) |
| TRACE | 1749 | 1512 | 237 | 0.82 (0.69, 0.97) | 0.64 (0.45, 0.91) | 0.78 (0.53, 1.16) |
| Random-effects pooled estimate | 8449 | 1906 | 0.80 (0.69, 0.93) | 0.73 (0.56, 0.95) | 0.91 (0.67, 1.23) | |
–13. We interpret these results as indicating that both patients with diabetes and patients without diabetes achieve reductions in mortality when treated with ACE inhibitors for heart failure.
| Study Name | Total N | N White | N Nonwhite | N Black | N Nonblack | Relative Risk, White (95% CI) | Relative Risk, Black (95% CI) | Ratio of Relative Risks (95% CI) |
|---|---|---|---|---|---|---|---|---|
| SAVE | 2231 | 1993 | 238 | 0.78 (0.50, 1.21) | 0.84 (0.71, 0.99) | 1.08 (0.67, 1.73) | ||
| SOLVD-prevention | 4228 | 3657 | 571 | 404 | 3824 | 0.95 (0.81, 1.12) | 0.87 (0.60, 1.25) | 0.91 (0.61, 1.36) |
| SOLVD-treatment | 2569 | 2061 | 508 | 396 | 2173 | 0.89 (0.79, 1.00) | 0.93 (0.74, 1.17) | 1.04 (0.81, 1.35) |
| Random-effects pooled estimate | 7711 | 1317 | 800 | 5997 | 0.89 (0.82, 0.97) | 0.89 (0.74, 1.06) | 1.01 (0.83, 1.24) | |
| Study Name | Total N | N White | N Nonwhite | N Black | N Nonblack | Hazard Ratio, White (95% CI) | Hazard Ratio, Black (95% CI) | Ratio of Hazard Ratios (95% CI) |
|---|---|---|---|---|---|---|---|---|
| SOLVD-prevention | 4228 | 3657 | 571 | 404 | 3824 | 0.94 (0.79, 1.111) | 0.82 (0.54, 1.25) | 0.87 (0.55, 1.37) |
| SOLVD-treatment | 2569 | 2061 | 508 | 396 | 2173 | 0.85 (0.74, 0.98) | 0.88 (0.65, 1.20) | 1.04 (0.74, 1.45) |
| Random-effects pooled estimate | 5718 | 1079 | 800 | 5997 | NC | NC | NC | |
and 15. We interpret these data as indicating that there is no evidence that black patients achieve lesser or greater reductions in mortality than white patients when treated with ACE inhibitors for heart failure. Whereas the relative risk reduction in black patients did not achieve conventional levels of statistical significance, the estimate of effect is the same as the statistically significant reduction seen in white patients. Furthermore, the two estimates of effect (for black and white patients) do not differ from each other statistically. Therefore, the most likely explanation for the lack of statistical significance in the estimate for black patients is the much smaller sample size, which increases the standard error and 95% confidence intervals. These results are consistent with the analysis by the SOLVD investigators that there was not a lesser reduction in mortality among black compared to white patients in the SOLVD studies (these investigators did, however, report a difference in hospitalization rate in black patients compared to white patients).34
| Study Name | Total N | N Male | N Female | Relative Risk, Male (95% CI) | Relative Risk, Female (95% CI) | Ratio of Relative Risks (95% CI) |
|---|---|---|---|---|---|---|
| CIBIS II | 2647 | 2132 | 515 | 0.71 (0.58, 0.87) | 0.52 (0.30, 0.89) | 0.73 (0.41, 1.30) |
| COPERNICUS | 2287 | 1822 | 465 | 0.68 (0.54, 0.86) | 0.63 (0.39, 1.04) | 0.93 (0.54, 1.59) |
| MERIT-HF | 3991 | 3093 | 898 | 0.63 (0.50, 0.78) | 0.93 (0.58, 1.49) | 1.49 (0.88, 2.51) |
| US Carvedilol HF | 1094 | 838 | 256 | 0.44 (0.24, 0.82) | 0.32 (0.11, 0.93) | 0.73 (0.21, 2.51) |
| Random-effects pooled estimate | 7885 | 2134 | 0.66 (0.59, 0.75) | 0.63 (0.44, 0.91) | 0.99 (0.70, 1.41) | |
| Study Name | Total N | N Male | N Female | Hazard Ratio, Male (95% CI) | Hazard Ratio, Female (95% CI) | Ratio of Hazard Ratios (95% CI) |
|---|---|---|---|---|---|---|
| COPERNICUS | 2287 | 1822 | 465 | 0.65 (0.51,0.83) | 0.65 (0.38, 1.10) | 1.00 (0.56, 1.78) |
| MERIT-HF | 3991 | 3093 | 898 | 0.62 (0.49, 0.78) | 0.94 (0.58, 1.53) | 1.52 (0.88, 2.60) |
| US Carvedilol HF | 1094 | 838 | 256 | 0.41 (0.22, 0.78) | 0.23 (0.07, 0.72) | 0.56 (0.15, 2.09) |
| Random-effects pooled estimate | 5753 | 1619 | 0.62 (0.52, 0.73) | 0.62 (0.34, 1.14) | 1.15 (0.75, 1.76) | |
| BEST | 2708 | 2115 | 593 | 0.91 (0.79, 1.04) | 0.91 (0.67, 1.24) | 1.00 (0.71, 1.41) |
– 19. Our interpretation of these data is that women and men with symptomatic heart failure have reduced mortality when treated with beta-blockers.
| Study Name | Total N | N Non-diabetic | N Diabetic | Relative Risk, Non-diabetic (95% CI) | Relative Risk, Diabetic (95% CI) | Ratio of Relative Risks (95% CI) |
|---|---|---|---|---|---|---|
| CIBIS II | 2647 | 2335 | 312 | 0.66 (0.54, 0.81) | 0.81 (0.52, 1.27) | 1.23 (0.75, 2.02) |
| COPERNICUS | 2287 | 1701 | 586 | 0.67 (0.52, 0.85) | 0.68 (0.47, 1.00) | 1.02 (0.65, 1.61) |
| MERIT-HF | 3991 | 3006 | 985 | 0.62 (0.48, 0.79) | 0.81 (0.57, 1.15) | 1.32 (0.86, 2.02) |
| Random-effects pooled estimate | 7042 | 1883 | 0.65 (0.57, 0.74) | 0.77 (0.61, 0.96) | 1.19 (0.91, 1.55) | |
| Study Name | Total N | N Non-diabetic | N Diabetic | Hazard Ratio, Non-diabetic (95% CI) | Hazard Ratio, Diabetic (95% CI) | Ratio of Hazard Ratios (95% CI) |
|---|---|---|---|---|---|---|
| COPERNICUS | 2287 | 1701 | 586 | 0.65 (0.50, 0.85) | 0.66 (0.43, 1.00) | 1.01 (0.62, 1.65) |
| MERIT-HF | 3991 | 3006 | 985 | 0.61 (0.48, 0.78) | 0.80 (0.55, 1.16) | 1.31 (0.84, 2.05) |
| Random-effects pooled estimate | 4707 | 1571 | NC | NC | NC | |
and 21. Our interpretation of these data is that patients with diabetes and HF have reduced mortality when treated with beta-blockers. It is possible that the relative reduction in mortality may be less for patients with diabetes than for those without diabetes, but since the absolute risk of mortality is so much greater in diabetic patients, the absolute risk reduction is almost certainly greater for diabetic than for nondiabetic HF patients treated with beta-blockers.
| Study Name | Total N | N White | N Nonwhite | N Black | N Nonblack | Relative Risk, White (95% CI) | Relative Risk, Black (95% CI) | Ratio of Relative Risks (95% CI) |
|---|---|---|---|---|---|---|---|---|
| COPERNICUS | 2287 | 2069 | 218 | 121 | 2166 | 0.66 (0.53, 0.82) | 0.62 (0.19, 2.01) | 0.94 (0.28, 3.11) |
| MERIT-HF | 3991 | 3755 | 236 | 207 | 3784 | 0.67 (0.54, 0.82) | 0.79 (0.36, 1.76) | 1.19 (0.52, 2.70) |
| US Carvedilol HF | 1094 | 217 | 877 | 0.38 (0.20, 0.70) | 0.53 (0.19, 1.48) | 1.41 (0.43, 4.68) | ||
| Random-effects pooled estimate | 5824 | 454 | 545 | 6827 | 0.63 (0.52, 0.77) | 0.67 (0.39, 1.16) | 1.17 (0.65, 2.11) | |
| BEST | 2708 | 627 | 2081 | 0.85 (0.74, 0.96) | 1.17 (0.94, 1.47) | 1.39 (1.07, 1.79) | ||
| Study Name | Total N | N White | N Nonwhite | N Black | N Nonblack | Hazard Ratio, White (95% CI) | Hazard Ratio, Black (95% CI) | Ratio of Hazard Ratios (95% CI) |
|---|---|---|---|---|---|---|---|---|
| COPERNICUS | 2287 | 2069 | 218 | 121 | 2166 | 0.63 (0.50, 0.80) | 0.60 (0.17, 20.5) | 0.95 (0.27, 3.32) |
| MERIT-HF | 3991 | 3755 | 236 | 207 | 3784 | 0.66 (0.53, 0.83) | 0.85 (0.36, 1.99) | 1.29 (0.53, 3.12) |
| US Carvedilol HF | 1094 | 217 | 877 | 0.32 (0.17, 0.62) | 0.44 (0.15, 1.28) | 1.38 (0.39, 4.81) | ||
| Random-effects pooled estimate | 5824 | 454 | 545 | 6827 | 0.59 (0.45, 0.76) | 0.64 (0.36, 1.16) | 1.21 (0.65, 2.27) | |
| BEST | 2708 | 627 | 2081 | 0.82 (0.70, 0.96) | 1.17 (0.89, 1.53) | 1.43 (1.04, 1.95) | ||
–25.
In contrast, black patients in the BEST study had a statistically significant difference in mortality compared to white patients when treated with bucindolol. In fact, the relative risk and hazard ratio for mortality exceeded 1 for black patients (although this was not statistically significant). Our interpretation of these data is that black patients are likely to have the same relative risk reduction as white patients treated with the beta-blockers bisoprolol, metoprolol, or carvedilol. Although the results for black patients were not statistically significant compared to placebo, because the point estimates of effect were similar to white patients, we judge the most likely reason for this finding to be the much smaller sample size. In contrast, bucindolol was associated with worse mortality outcomes in black patients than in white patients, and may actually increase mortality in blacks. Additional data are needed in this area.
This figure shows the expected mortality for asymptomatic patients treated with and without ACE inhibitors. The model results are similar (well calibrated) to the SOLVD prevention trial results. In that trial, the mortality at four years was 17% for the ACE inhibitor arm and 19% for the placebo arm.
This figure shows the development of symptoms or death for asymptomatic patients treated with and without ACE inhibitors. The model results are similar (well calibrated) to the SOLVD prevention trial results. In that trial, the death or heart failure at four years was 35% for the ACE inhibitor arm and 44% for the placebo arm.
) and a 57% five-year morbidity and mortality rate (Figure 27). These results are similar to the findings of the SOLVD prevention study.3
| Strategy | Total Cost | Life Expectancy | Cost per Life Year | Quality-Adjusted Life Expectancy | Cost per QALY |
|---|---|---|---|---|---|
| No ACE inhibitor | $36,983 | 8.12 | 6.56 | ||
| ACE inhibitor | $40,701 | 8.76 | $5,802 | 7.22 | $5,644 |
Patient Age. For the base-case analysis, we assumed an age of 55 years. For older age groups, both the cost and benefit of treatment with ACE inhibitors decreased. For an 80-year-old person, the marginal cost-effectiveness was $6,650 per QALY, which was only slightly higher than $4,666 per QALY for a 50-year old person.
Risk of Death with Heart Failure. Our base-case analysis assumed that the risk of death for patients with heart failure treated with ACE inhibitors was 6.5 times greater than the risk of death for the U.S. age-adjusted population.3 If we assumed a lower risk of death (relative risk 2.0), both costs and life expectancy increased, but the cost-effectiveness ratio remained favorable ($4,093 per QALY).
Reduction in Heart Failure Incidence. If the reduction in heart failure incidence with ACE inhibitor treatment was only half of the effect observed in the SOLVD trial, treatment cost remained less than $10,000 per QALY gained. Even when we assumed no reduction in mortality for asymptomatic patients treated with ACE inhibitors, treatment had to reduce the yearly probability of developing symptomatic heart failure from 9.8% (untreated) to only 9.5% (3% relative risk reduction) for the cost-effectiveness ratio to drop below $100,000 per QALY gained, and to 9.1% (7% relative risk reduction) for the cost-effectiveness ratio to drop below $50,000 per QALY gained.
The effect of varying the relative risk of death for treatment with ACE inhibitors (ACEi) versus no treatment is displayed. As the relative risk approaches 1.0, the cost-effectiveness ratio increases (treatment becomes less cost-effective). However, even when ACE inhibitors do not prevent death for asymptomatic patients, they are still inexpensive (<$15,000 per QALY gained), because they reduce the incidence of heart failure.
).
Probability of Hospitalization if Symptomatic. We assumed that 11% of patients with heart failure would he hospitalized each year. If 15% were hospitalized each year, the cost per QALY gained dropped to $5,272. Even if hospitalizations for heart failure patients were completely eliminated by ACE inhibitor treatment, preventing heart failure ($6,539 per QALY gained) still remained cost-effective because heart failure also increases outpatient costs and worsens quality of life. Our findings were also insensitive to the probability of being hospitalized with the first episode of symptomatic heart failure.
The impact of the cost of ACE inhibitors (ACEi) on the cost-effectiveness of treatment is shown. Even with a high drug cost, the cost effectiveness of treatment remains attractive (<$10,000 per QALY gained), because the increased benefits and reduced heart failure costs are substantial.
), and even if the cost of the ACE inhibitor were 0, treatment would not be cost saving because the improvement in survival (both before and after the development of symptoms) simply delays medical costs to older ages. If ACE inhibitors did not affect survival for asymptomatic patients with low ejection fraction, overall medical costs would be lower if the cost of ACE inhibitor treatment was less than $75 per year.
The cost of hospitalization had little effect on cost-effectiveness. Eliminating all hospitalizations did not raise the cost-effectiveness threshold above $7,400 per QALY gained. In addition, the cost of outpatient management did not affect our results. The cost per QALY gained ranged from $5,920 (if the annual outpatient cost was $200) to $5,306 (if the cost was $800). The discount rate also had little effect on the results. A discount rate of 0% resulted in $5,592 per QALY gained, compared to $5,776 per life year gained if the discount rate was 6%.
Quality of Life. We evaluated the effect of various utility values for living with heart failure on the cost-effectiveness of prevention with ACE inhibitors. In the base case, we assumed that quality-of-life utility would drop from 0.865 when asymptomatic to 0.71 when symptomatic (difference of 0.155), based on time-tradeoff utilities from the Beaver Dam Study.13
Similar results were found when we used the visual analog scale data from the SOLVD trial. In that study, the patients with asymptomatic low ejection fraction rated their quality of life at 0.68, compared with 0.60 for patients with symptoms. Using their values, we found the cost-effectiveness of ACE inhibitor treatment to be only $7,598 per QALY gained.
| Strategy | Cost | Marginal Cost | Life Years | Marginal Life-years | QALYs | Marginal QALYs | Marginal Cost-Effectiveness $/QALY |
|---|---|---|---|---|---|---|---|
| No screening | $43,474 | 14.7753 | 12.7677 | ||||
| ECG then Echo | $43,655 | $181 | 14.7848 | 0.0094 | 12.7775 | 0.0098 | Dominated (Extended) |
| BNP then Echo | $43,740 | $266 | 14.7893 | 0.0140 | 12.7822 | 0.0145 | $18,300 |
| Echo alone | $44,089 | $349 | 14.7906 | 0.0013 | 12.7836 | 0.0014 | $268,500 |
| ECG alone | $45,007 | $918 | 14.7831 | -0.0076 | 12.7762 | -0.0074 | Dominated |
| BNP alone | $45,918 | $1,829 | 14.7865 | -0.0042 | 12.7799 | -0.0037 | Dominated |
NOTE: Marginal costs and quality-adjusted life years (QALYs) are ordered by cost and compared to the prior nondominated strategy. The ECG then Echo strategy is eliminated by extended dominance (greater benefit and lower cost can be obtained through a combination of No screening and BNP, then echo strategies). The cost-effectiveness ratio of BNP then Echo versus ECG then Echo is $18,700 per QALY gained.
Because the cost-effectiveness ratio of screening with the ECG compared to no screening was greater than the ratio for BNP compared to ECG screening (extended dominance), this strategy was eliminated as a possible screening option for the base-case cohort. BNP screening demonstrated extended dominance over ECG screening, because the incremental cost-effectiveness ratio for BNP compared to ECG screening was less than the ratio for ECG screening compared to no screening. Willingness to pay for the benefits of ECG screening ensures a willingness to pay for the extra benefits of BNP screening. Similarly, strategies of relying only on the ECG or BNP to determine treatment were eliminated because they were more costly and provided fewer QALYs than the strategy using BNP followed by echocardiography. The ECG- and BNP-only strategies are not discussed further. All future references to BNP or ECG screening assumes that abnormal tests are followed by echocardiography.
| Population | Definition of EF Method | Prevalence (%) | Asymptomatic (%) |
|---|---|---|---|
| Primary care patients = 45 years45 | < 40% echocardiography | 1.8 | 50 |
| Community age 25-N7518 | = 30% echocardiography | 3.2 | 51 |
| Community age = 5518 | = 30% echocardiography | 5.4 | Not stated |
| Community age = 55 with ischemic heart disease18 | = 30% echocardiography | 12.1 | Not stated |
| Patients with CVA, TIA or peripheral vascular disease46 | < 40% echocardiography | 28 | 54 |
NOTE: CVA = cardiovascular accident; EF = ejection fraction; TIA = transient ischemic attack.
This figure shows the impact of prevalence of low ejection fraction (<35%) on the cost-effectiveness of screening. For the entire range listed, BNP testing followed by echocardiography is the preferred choice. At very low prevalence rates, screening is no longer cost effective.
). For the cost-effectiveness ratio with BNP screening to be less than $50,000 per QALY gained, the prevalence must be greater than 0.8%; to be under $20,000 per QALY gained, the prevalence must be 2.5%. BNP screening is never cost saving, even at 100% prevalence of disease, because treatment of asymptomatic patients with ACE inhibitors is more expensive than not treating these patients.
This figure shows the impact of BNP sensitivity for detecting depressed left ventricular systolic function (assuming a specificity of 76%). At a sensitivity greater than 85%, screening with BNP followed by echocardiography is the preferred choice (other screening strategies are dominated; thus the comparison of interest is with no screening). With a BNP sensitivity less than 85%, BNP testing is more expensive than ECG testing but also provides more benefit. The incremental cost-effectiveness of BNP versus ECG testing remains less than $50,000 per QALY gained until the BNP sensitivity drops below 70%.
This figure shows the impact of BNP specificity for detecting depressed left ventricular systolic function (assuming a sensitivity of 87%). At a specificity greater than 70%, screening with BNP followed by echocardiography is the preferred choice (other screening strategies are dominated; thus, the comparison of interest is with no screening). With a BNP specificity less than 70%, BNP testing is more expensive than ECG testing but also provides more benefit. The incremental cost-effectiveness of BNP vs. ECG testing remains less than $50,000 per QALY gained until the BNP specificity drops below 50%.
). The specificity of BNP testing for detecting depressed left ventricular function is estimated to be 71%. Even if the specificity is 50%, the cost per QALY gained would be less than $50,000, compared to screening with the ECG (Figure 32). If the specificity is at least 70%, the ECG strategy is no longer viable (eliminated by extended dominance).
Past studies have used different cut-points for an abnormal BNP test, based on the appearance of the receiver-operator characteristics curve. However, the particular cutoff chosen may not be optimal in terms of cost-effectiveness. Using various sensitivity and specificity combinations from the MONICA patient population,20 we found that both cost of care and quality-adjusted survival improve as sensitivity increases and specificity decreases. If society is willing to pay $100,000 per QALY gained, using a low BNP threshold (24ng/ml) that produces a sensitivity near 96% with specificity near 65% is optimal. However, if society will pay only $20,000 per QALY gained, then a BNP threshold slightly above 18ng/ml (sensitivity 72%, specificity 90%) is optimal.
The impact of BNP test price ($29) on the cost-effectiveness of screening is shown. At a BNP cost less than $50, screening with BNP followed by echocardiography is the preferred choice (other screening strategies are dominated; thus, the comparison of interest is with no screening). If the BNP cost is greater than $50, BNP testing is more expensive than ECG testing but also provides more benefit. The incremental cost-effectiveness of BNP versus ECG testing remains less than $50,000 per QALY gained, as long as the cost of the BNP test is less than $120.
). The cost per QALY gained with BNP screening (compared to ECG screening) remained less than $50,000 as long as the cost of the BNP test was less than $120.
The Medicare reimbursement for two-dimensional echocardiography has been dropping (without adjustment for inflation) in an attempt to better match actual costs of delivering treatment, as estimated by the Center for Medicare and Medicaid Services (formerly the Health Care Financing Administration). Significant disagreement exists between specialty societies and Medicare regarding the actual cost of an echocardiogram. However, even if echocardiography costs were as high as $1,000, BNP screening would still cost only $37,600 per QALY gained compared to ECG screening, and ECG screening would cost $34,200 per QALY gained compared to no screening.
ECG is similar in price to BNP testing. Therefore, the decision to use one over the other is based on the differences in test characteristics.
This figure shows the effect on screening cost-effectiveness of different improvements in survival following treatment of asymptomatic patients with low ejection fraction (<35%) on the cost-effectiveness of screening. For the entire range listed, BNP testing followed by echocardiography is the preferred choice. As long as treatment of asymptomatic patients with low ejection fraction improves outcome by at least 0.3 QALYs, screening with BNP followed by echocardiography will cost less than $50,000 per QALY gained.
).
ACE Inhibitor Use in Healthy Patients. We assumed a small decrement in quality-adjusted survival (0.001 years or 0.37 days) each year to account for potential side effects of ACE inhibitor treatment. Because no quality-of-life studies of ACE inhibitor use in healthy patients are available, the negative health impact of taking unneeded medication is unknown. However, our findings were similar over a wide range of quality-of-life decrements for ACE inhibitor treatment. The cost-effectiveness of BNP screening (compared to no screening; ECG screening was eliminated by extended dominance) ranged from $18,200 per QALY gained (for no decrease in quality adjusted survival) to $20,300 per QALY gained (for a three-day reduction per year in quality-adjusted survival) for normal patients taking ACE inhibitors.
The meta-analyses have several potential limitations:
An important limitation, common to many such meta-analyses, is the differential quality of the original studies. We cannot adjust for any inherent biases in the individual studies. However, to mitigate this limitation, we included studies that were double-blind randomized controlled trials and used all-cause mortality as the outcome. These design features help protect against some of the more important biases.
We restricted attention to the large trials on ACE inhibitors and beta-blockers due to resource constraints. By excluding smaller trials, we may have limited our generalizability and our ability to investigate heterogeneity in treatment effects. We did observe little to no evidence of publication bias among the large trials via visual inspection or formal testing and, given the notable nature of such trials, are fairly confident we did not miss a similar trial in our extensive search.
Between-study heterogeneity was observed, especially among the ACE inhibitor studies. Although the random-effects model is designed to take this extra variability into account in the estimation of the standard errors and generally widens the confidence interval for the pooled estimate, this model does not explain the heterogeneity. In fact, significant between-study heterogeneity should lead us to interpret a pooled result with caution.
We conducted a meta-analysis of study summary statistics (relative risks, etc.), rather than a patient-level data analysis due to the fact we could not obtain patient-level data for all trials in a timely and efficient manner. This approach limited us in two ways. First, we were unable to estimate hazard ratios for all subgroups of interest and had to rely on relative risks for some studies. The latter statistic does not adjust for differential followup intervals across studies. Second, we cannot investigate interactions between patient-level characteristics that might mitigate the treatment effect, nor can we adjust for these effects in our estimates. For example, suppose blacks are more likely to have hypertension than whites, and the treatment works less well for patients with hypertension than for those who do not have hypertension. We may conclude that the treatment works less well for blacks than whites when actually if we had controlled for hypertension status, we would not have seen differences between the two racial subgroups. By ignoring the effect of hypertension, we incorrectly attribute its association with treatment to race. A patient-level data analysis would have allowed this adjustment, had data on important confounders been collected.
Studies did differ in their definition of racial groups. A sensitivity analysis that we conducted to try to determine whether this variability affected our conclusions did not show different results using different definitions.
The cost-effectiveness analyses have several potential limitations:
We relied on the SOLVD prevention trial to determine the impact of ACE inhibitors on patients with asymptomatic left ventricular dysfunction. Although this is the best data source available, the SOLVD patients were not randomly selected from the population. It is possible that randomly selected patients with reduced left ventricular function may show less benefit with ACE inhibitors.
Our analysis did not include the potential impact of beta-blockers in the base case, because these agents have not yet been evaluated in randomized trials of asymptomatic patients. If beta-blockers have an incremental benefit over ACE inhibitors in this population, the cost-effectiveness of screening will likely improve, assuming beta-blocker treatment is cost-effective for heart failure patients.
We limited our analysis of BNP to screening asymptomatic subjects. Our study did not examine the appropriate use of BNP to adjust medications (e.g., dose of diuretics, use of digoxin) for patients with heart failure.
Our study did not specifically model other tests (biopsy, cardiac catheterization) and treatments (revascularization) that may result from the knowledge of depressed left ventricular function, because there is no accepted standard of care with additional testing or treatments (other than ACE inhibitors and possibly beta-blockers) for asymptomatic patients with reduced ejection fraction. Any additional testing such as screening for coronary artery disease should be evaluated with a separate cost-effectiveness analysis.
Our cost-effectiveness analysis did not distinguish between patient subgroups. If ACE inhibitors are more or less effective in men or women, the cost-effectiveness of treatment (and screening for depressed ejection fraction) will vary by gender. Additional studies are needed to determine which patient groups have a high enough prevalence of depressed left ventricular function (> 1%) to make screening cost-effective.
We believe several conclusions can be drawn from this evidence report. For the purposes of discussion, we divide these conclusions into those that pertain to methodological considerations, those that pertain to clinical issues, and those that pertain to the cost-effectiveness analysis.
A large enough number of placebo-controlled, randomized trials of ACE inhibitors or beta-blockers have been conducted to assess their efficacy for the prevention and treatment of heart failure.
Few of these studies have reported data stratified by patient subpopulations of interest to clinicians.
Obtaining these subgroup data by attempting to contact authors of the original studies is both time consuming and not particularly successful. Attempts on the scale used to generate this report are not within the time- and resource-constraints of typical AHRQ evidence reports.
Obtaining subpopulation data by inspecting data submitted to the FDA is a potentially fruitful area but only to the extent that the data are already in electronic form. Paper-based records are too difficult to retrieve and too voluminous to review efficiently.
Two Evidence-Based Practice Centers can successfully collaborate on the same evidence report. In this case, the cost-effectiveness analyses were performed at the Stanford-UCSF Evidence Based Practice Center.
For most of the subpopulations assessed in our meta-analysis, our results are reassuring in that we found evidence supporting beneficial reductions in all-cause mortality with the use of beta-blockers in men and women, the use of ACE inhibitors in black and white patients, and the use of either drug in patients with diabetes.
We did, however, find evidence suggesting that women with asymptomatic left ventricular dysfunction may not have reduced mortality when treated with ACE inhibitors. The evidence we found does not constitute proof, and additional evidence is needed to analyze the effect of ACE inhibitors in women with asymptomatic left ventricular dysfunction.
We also found conflicting evidence regarding the effect of beta-blocker use in black patients. For three of the beta-blocker studies, the pooled estimate of effect suggested that black patients and white patients have similar reductions in all-cause mortality when treated with beta-blockers. However, one study, which was unique in that it assessed the beta-blocker bucindolol, reported a statistically significant adverse effect on mortality in blacks relative to whites, even suggesting that use of bucindolol caused harm. These results suggest that all beta-blockers cannot be assumed to have similar effects.
We found that treatment of asymptomatic left ventricular dysfunction with ACE inhibitors was very cost-effective under virtually all assumptions, with typical costs per QALY gained of between $5,000–$10,000, which makes this treatment much more cost-effective than many other treatments considered standard medical practice.
The analysis of the cost-effectiveness of screening showed that screening with brain natriuretic peptide followed by echocardiography in a cohort of asymptomatic individuals aged 55 was also cost-effective compared with the costs of other therapies currently considered standard medical care. This strategy cost $19,000 per life year gained compared to a strategy without screening, with the number needed to screen equal to 77 in order to gain one year of additional life.
These results were only modestly sensitive to cost and were most sensitive to the prevalence of asymptomatic decreased left ventricular ejection fraction. When the prevalence falls below about 1%, a strategy of screening becomes less cost-effective than accepted thresholds for cost-effective care.
The findings of this evidence report suggest several important future research studies. The first and most important would be for additional data to support or refute the evidence that different beta-blockers may have different effects on all-cause mortality in black patients. We do not think it likely that additional analysis of existing data will be able to conclusively settle this issue. New placebo-controlled randomized clinical trials of beta-blocker therapy in black patients are likely the only way to definitively answer this question. Future studies of existing or new beta-blocker drugs for heart failure should include sufficient numbers of black patients to separately assess outcomes in this population because a similar effect in black patients and white patients cannot be assumed.
A second area for future research is further assessment of the effect of ACE inhibitors in women with heart failure, particularly the effect on women with asymptomatic left ventricular dysfunction. It may be possible to answer this question by a more complete assessment of existing data from randomized clinical trials. This would require an individual patient data meta-analysis, which in turn would require obtaining individual patient data from all of the randomized trials. While such an effort would be expensive, it may be less expensive and more ethical than mounting a new clinical trial designed to answer this question.
Additionally, other outcomes of interest, including cardiac mortality, symptoms, and health care utilization, should be examined for all patient subpopulations. Individual patient-level data from the major RCTs may be sufficient to answer these and other original key questions regarding more patient subpopulations (the aged and those with renal failure).
An additional implication of our finding is that inadequate attention has been paid to enrolling sufficient numbers of patients in important clinical subpopulations in randomized trials. Such attention would make meta-analyses like those contained in this report unnecessary.
If our findings of differential efficacy are sustained, additional research aimed at identifying the cause for these findings should be undertaken. One possibility is that these findings do not represent differences in men and women or black patients and white patients, but rather reflect differing efficacy of these drugs according to the cause of heart failure (e.g., ischemic or nonischemic), which then may differ by sex or race. Alternatively, there could be a molecular basis for these results that differs by gender and race.
Given the robust evidence of benefit for ACE inhibitors and beta-blockers in reducing mortality, future work should also address how to improve the use of these therapies by focusing on potential barriers for practitioners and patients. Yet another area for future research is empirical tests of our conclusions from our cost-effectiveness analyses.
Additional studies are needed to determine the true prevalence of asymptomatic left ventricular dysfunction and to determine the costs associated with making a new diagnosis of heart failure. Further research is needed to determine which patient characteristics identify a population at risk for left ventricular systolic dysfunction (prevalence greater than 1%). In addition, a study evaluating the health and economic outcomes of screening asymptomatic patient with BNP is warranted.
Date
REF # 0098/238/171/306
Name
Address
Address
Dear XX,,
The US Agency for Healthcare Research and Quality (AHRQ) has commissioned us to perform a meta-analysis on the treatment of systolic heart failure with ace-inhibitors in particular subgroups. The participating investigators on this project are, in addition to myself:
Dr. Michael Barrett
Dr. Greg Fonarow
Dr. Barry Greenberg
Dr. Paul Heidenreich
Dr. Tom Knabel
Dr. Marvin Konstam
Dr. Michael W. Rich
Dr. Anthony Steimle
Dr. Lynne Warner Stevenson
We read with interest your articles
“Effect of ramipril on mortality and morbidity of survivors of acute myocardial infarction with clinical evidence of heart failure. The Acute Infarction Ramipril Efficacy (AIRE) Study Investigators” in Lancet,
“Effect of ramipril on morbidity and mode of death among survivors of acute myocardial infarction with clinical evidence of heart failure. A report from the AIRE Study Investigators” in European Heart Journal,
“Angiotensin-converting enzyme-inhibitors, left-ventricular dysfunction, and early heart-failure” in American Journal of Cardiology, and
“Captopril in heart failure: a double-blind study of the effects on renal function” in Journal of Cardiovascular Pharmacology.
We would be very appreciative if you could assist us by providing data from your study specific to women, persons of African descent (Blacks), diabetics, renal insufficiency, or persons 80 years of age or older.
Name
Date
If such subgroup data are not available, we would also appreciate that information. If you have any questions regarding this request, please call me at 310-393-0411 ext. 6669 or send email to shekelle@rand.org Thank you very much for your assistance.
Sincerely,
Paul Shekelle, MD, PhD
Director
Date
Name
Address
Address
Address
Dear XX,
We are currently finishing preparation of a report on heart failure commissioned by the U.S. Agency for Healthcare Research and Quality, and are seeking peer reviewers. This report presents two analyses:
an assessment of the effect of beta-blockers or ACE inhibitors on mortality in women, blacks, and diabetics; by pooling the relevant data from the major published randomized trials; and
a cost effectiveness analysis of screening for asymptomatic left ventricular dysfunction followed with ACE inhibitor treatment.
We are hoping you will be able to be a peer reviewer of this draft report. We expect the draft report to be available in approximately two weeks, and then reviewers would have three weeks to complete their review. The Agency for Healthcare Research and Quality has agreed to have us pay an honorarium of $300 for the review.
Please fax the enclosed form to Shannon Rhodes at 310-451-6930 indicating whether or not you are willing to be a peer reviewer.
If you have any questions, please do not hesitate to contact me at 310-393-0411 ext 6669 or at Shekelle@rand.org.
Sincerely,
Paul Shekelle, MD, PhD
Director, Southern California
Evidence- based Practice Center
REVIEW QUESTIONS TO CONSIDER AND ON WHICH YOU MAY WANT TO COMMENT ARE LISTED HERE:
OVERALL EVALUATION
Is it clear what we did? You may agree or disagree with our methods, findings or conclusions, but you should be able to understand what it is we did in order to produce this report.
QUESTION FORMULATION
Are evidence report questions well formulated and easily understandable?
STUDY IDENTIFICATION
Is there a thorough search for relevant data using appropriate resources?
Are there unbiased, explicit searching strategies that are appropriately matched to the question?
STUDY SELECTION
Are appropriate inclusion and exclusion criteria used to select articles? Are selection criteria applied in a manner that limits bias? Are efforts made to identify unpublished data, if this is appropriate? Are reasons for excluding studies from the report stated? Did we miss any crucial pieces of information in our literature search?
APPRAISAL OF STUDIES
Are important parameters (e.g. setting, study population, study design) that could affect study results systematically addressed?
DATA COLLECTION
Is there a minimal amount of missing information regarding outcomes and other variables considered key to the interpretation of results? Are efforts made to reduce bias in the data collection process?
DATA SYNTHESIS
Are important parameters, such as study designs, considered in the synthesis? Are reasonable decisions made concerning whether and how to combine the data? Is precision of results reported? Are limitations and inconsistencies of studies stated? Are limitations of the review process stated?
CONCLUSIONS (stated throughout the report)
Are conclusions supported by the data reviewed? Is evidence appropriately interpreted as inconclusive (no evidence of effect) or as showing a particular strategy did not work (evidence of no effect)? Is a summary of pertinent findings provided? Are the specific issues related to the research question addressed adequately?
RESEARCH:
Are implications for research discussed? What directions for future research would you recommend based on this report that we have not covered?
| Page # | Section | Reviewers' Comments | Author's Response to Comments |
|---|---|---|---|
| General | Questions were well formulated and easily understandable. Methods were explained carefully in text. It appears as if great effort was made to exclude bias. | No response necessary | |
| General | I believe that this work is unique and valuable. | No response necessary | |
| General | I am impressed by the clarity of writing and, given the scope of the project and large amount of data/analyses, the brevity of the report. | No response necessary | |
| General | Goals were clearly stated. | No response necessary | |
| v | Abstract | Avoid the statement “additional randomized controlled evidence of the effect of ACE inhibitors in women is needed.” Most won't believe giving placebo to women with heart failure to be ethical. | This sentence has been deleted from the abstract. |
| v | Abstract | I would add a statement regarding the equivalency of both ACEI and BB in diabetics, ACEI in blacks, and BB in women. | The entire abstract has been rewritten to highlight relative risks in subgroups rather than ratio of relative risks. |
| v | Abstract | The manner in which some of the findings are reported is, in my opinion, misleading. | The manner of reporting the results has been changed. |
| 18 | Methods | My major concern centers on the low emphasis placed on the relative risks of the subgroups. Although the main stated objective is to assess whether the effect of medications differs by the subgroups (ratio of relative risks), the actual effect in each subgroup (relative risk) seems more clinically important. | See above comment. |
| 19 | Methods | The risk index as calculated is complex, as a positive value could reflect either that the therapy has an adverse effect in the subgroup compared to no effect or benefit in the larger population, OR that the subgroup has >= zero benefit, but less benefit than the larger population. This is discussed later in the methods somewhere, but could perhaps be highlighted early in the description of the index. | See above comment. |
| 19 | Methods | I find the ratio of relative risks difficult to interpret clinically. I would always include estimates of the relative risk - even in the abstract. | See above comment. |
| 19 | Methods | I had difficulty following the rationale for the initial pooling of the RRRs vs. pooling the risk ratios separately and then taking the ratio. | See above comment. |
| 19 | Methods | Confidence intervals for the RR of 0.94 for ACEI in women would be relevant. | See above comment. |
| 37 | Results | Some of the findings could be presented in a more clinically relevant and less ambitious manner. It would be helpful to present the within-subgroup pooled risks ratios and hazard ratios first, followed by the between subgroup relative risk ratios and relative hazard ratios. | See above comment. |
| 37 | Results | Within subgroup pooled risks ratios and absolute risk reductions with confidence intervals and p-values should be reported. | See above comment. |
| 37 | Results | You clearly state that a positive RRR does not necessarily exclude a mortality benefit of the drug in either subgroup. Your figures only present the RRR data, and I wonder if a table or summary figure could first show the RR for each subgroup before the RRR data is presented. | See above comment. |
| 37 | Results | I am not a fan of how the RRR was used as a summary measure. I would have rather seen the separate point estimates for treatment effect (and 95% CI) in the two comparison populations. | See above comment. |
| 37 | Results | Question is not whether a subgroup does worse than another subgroup, but whether the subgroup in question benefits from treatment. | See above comment. |
| 37 | Results | I don't think “This means that there is a 15% increase in mortality in women relative to men treated with ACEI ..” is quite accurate. For instance, women may have lower mortality on placebo than men. | See above comment. This statement actually is accurate, but confusing since it concerns relative and not absolute risk. We have completely reoriented the Results section to make it more clinically understandable. |
| 83 | Future Research | The major question is not whether women benefit as much as men or blacks as much as whites. The major questions is whether these therapies are helpful, harmful, or neither in these subgroups. | See above comment. |
| i | Title | You refer to left ventricular heart failure and left ventricular heart dysfunction. These terms are not used. You could use “heart failure and left ventricular systolic dysfunction.” | Done |
| i | Title | Possible change to “Pharmacologic management of heart failure: effects in women, black patients, and diabetics. Cost-effectiveness considerations of screening and treatment strategies.” | The title has been changed to incorporate the previous comment. This title suggestion seemed to us to overweight the Cost- Effectiveness section of the report. |
| 1 | Summary | Summary seems unnecessary in light of the following report. | This section is an AHRQ requirement. |
| 13 | Overview | Although Chapter 2 nicely describes the scope of work and original potential key questions, I found myself wondering about the background / larger context of the Evidence Report. How did the ACP and other groups get involved in nominating this topic? | Groups nominate topics for evidence reports via a mechanism that can be found on AHRQ's website (www.ahrq.gov). It is beyond the scope of the Evidence Report to explain the reasons why partners nominated topics other than the information presented in the introduction. |
| 15 | Methods | Method/rationale used to formulate the first questions was clearly defined. The only concern was why the study could not have addressed whether drug efficacy varied as a function of age as originally requested. | An analysis of efficacy by age requires individual patient data, which were not available for the majority of studies. This is stated on page 18. |
| 16 | Methods | Affiliation should be UnitedHealthcare (one word) | Done |
| 16 | Methods | Search methodologies used to identify relevant data were of high quality. Search was limited to randomized clinical trials experience, and may have omitted well performed observational studies. These are sometime used as supportive data to trial subgroup analyses. | No response necessary |
| 16 | Methods | Table 1 in the Methods section doesn't help me much. Could this be deleted? | Agreed. This table has been deleted. |
| 17 | Methods | TEP members also provided names / acronyms of the major ACE inhibitor and beta-blocker trials. | The text on page 17 has been changed. |
| 17 | Methods | Original plan for obtaining patient level results from published and unpublished studies was strong. The strategy of limiting to largest RCTs, FDA, and published subgroup data limits the scope and generalizability somewhat by not including smaller studies, unpublished studies, etc. Overall, I feel that their prioritization decisions were pragmatic and reasonable under the circumstances. | No response necessary |
| 17 | Methods | Did all of the studies have the necessary subgroup data included in the published articles? Was the rationale to pursue patient-level data only to get more reliable data? | We pursued the necessary subgroup data from all studies but were only successful for the ones listed. The rationale was to increase statistical power by increasing sample size. However, we have added new sensitivity analyses of both ACE inhibitor and beta-blocker trials by clinical condition where possible. |
| 18 | Methods | Why are we restricted to the FDA data that is available electronically? The NDA submissions always include extensive tables of subgroups. | We were advised by the FDA that retrieving the paper records would take months and that we would have to search by hand through “hundreds” of filebooks to find the data we needed. |
| 18 | Methods | Data collection is complete as to what was sought, but not enough was sought. | What we sought and obtained was all that was possible within the resource constraints of the EPC contract. |
| 18 | Methods | I am surprised that cardiac mortality data was not obtainable for all studies. Perhaps a table presenting the proportion of studies that have the outcomes of interest (resource utilization, quality of life, mortality) could be included. | Cardiac mortality was available for most studies, but subgroup data regarding mortality were available for only the studies listed. |
| 18 | Methods | I am intrigued by the lack of response of some authors and trial groups who failed to respond to requests for information. | No response necessary |
| 18 | Methods | Dr. Marion Limacher, U. Florida, debates the equal efficacy of ACEI in women compared to men in the Wenger edited book on heart disease in women, in particular related to the SOLVD trial. Did authors contact Dr. Limacher re. this database, which must have been available to her at that time? | We contacted Dr. Limacher, who sent us a copy of her book chapter, which we have now incorporated into the report. |
| 18 | Methods | Data synthesis limitations are very difficult to overcome. The solution appears elegant. | No response necessary |
| 18 | Methods | Focusing on mortality alone makes it more difficult to see effects in small subgroups. | No response necessary |
| 18 | Methods | Studies are not extensively analyzed and limitations assessed. | No response necessary |
| 18 | Methods | Need to state the known limitations of meta-analysis as compared to controlled clinical trials, along with our reasons for using this method, namely the absence of sufficient N in each subgroup for most trials. | This limitation is currently stated in the introduction. We have added it again on page 18 and in the Limitations section. |
| 18 | Methods | Some meta-analyses include a score to grade the quality of the studies. Was such an approach considered in this meta-analysis? | No. The use of quality scores has not been favored since the publication of the Juni study (Juni P. JAMA. 1999; 282(11):1054–60.) |
| 18 | Methods | If you excluded a study that was on the margin of inclusion, you might specifically mention such and the reasons for exclusion, to further illustrate your application of the criteria. | There were no studies at the margin for inclusion, we included all the RCTs with sample >1000. |
| 21 | Methods | From a clinical perspective, combining the data of the post-MI trials with the non post-MI trials is somewhat concerning. Yes, most patients had ASHD and were post-MI. However, ACE inhibitors were started within days of the MI in the post-MI trials and most did not have symptomatic HF. | A new sensitivity analysis was performed for the symptomatic and asymptomatic studies, and is presented on page 21. |
| 21 | Methods | I have concerns about the impact of combining the left ventricular dysfunction studies with the post-MI ones. Properties of the ACEIs that might have been important in the post-MI populations might not be as relevant in the LV dysfunction studies and vice-versa. | See above comment. |
| 37 | Results | You should note that your results on ACE inhibitors are based on a mix of trials in patients with heart failure and in patients with LV systolic dysfunction post -MI. | See above comment. |
| 18 | Methods | I found the use of “principle” for “principal” on numerous occasions to be distracting. | The text has been changed where appropriate. |
| 18 | Methods | The major criticism … will be the integrity of the “meta-analysis.” Perhaps, in the final version, a short commentary about the benefits and detriments of this approach could be made. | A Limitation section has been added. |
| 18 | Methods | Authors limited their analysis to studies whose primary question was specific to LV dysfunction patients. Theoretically, there is a larger body of evidence regarding these treatment efficacy available from other trial populations (I.e. in hypertension, secondary prevention trials, etc.) These other trial types would have enrolled some proportion of patients with LV dysfunction and could have supplemented their patient level analyses. | This would have required more extensive requests for individual patient data that were beyond our resources for this project. |
| 18 | Methods | Data collection appears complete | No response necessary |
| 18 | Methods | Definition of “black” varies and is not a unified population. | This acknowledges what is already explained in the text on page 19. No response necessary. |
| 19 | Methods | I am familiar with the DerSimonian and Laird random effects model. By mentioning its low power to detect differences across studies and the fact that its only a one-step iterative method, are you implying that there are other methods that are “better?” Were these considered? | The low power refers to the chi-squared test of heterogeneity and is not associated with the DerSimonian and Laird random-effects model. The low power of the chi-squared test is well-known (Hedges and Olkin 1985). Thus, to fully assess and deal with possible heterogeneity between studies, our approach is to combine the knowledge gained from this statistical test with clinical knowledge about heterogeneity, and to use a random-effects model to adjust our variance estimates for any heterogeneity that might exist. The DerSimonian and Laird random-effects model is a one-step method in terms of how it estimates the between-study variance and is equivalent to applying a method of moments approach. It is generally accepted as the most appropriate choice for a random effects estimate when one is combining a group of studies and not incorporating covariates. If one fits a multivariate model, e.g., random effects meta-regression, sometimes a restricted maximum likelihood approach is used. In our experience, the two approaches (DerSimonian and Laird and restricted maximum likelihood) produce very similar between-study variance estimates. |
| 19 | Methods | Did you use the long-term CONSENSUS data for total survival? (published in European Journal…) | We used individual patient data from the original CONSENSUS trial. The long-term CONSENSUS data showed few patients still alive, which obscure the beneficial effect of ACE inhibitors in reducing mortality up to at least 3 years of followup. |
| 19 | Methods | I think the authors should have described the trial populations more clearly in the beginning of the report, as well as tested whether treatment response varied as a function of populations studied or etiology of LV dysfunction. For example, black patients are less likely to have ischemic etiology for their LV dysfunction. Thus, the lower benefits of BB in black patients theoretically may have been confounded by disease etiology. | This level of detail requires patient level data, which was available for a minority of studies. It is plausible that the differences we saw in race and sex groups reflect differences in effectiveness of these drugs on the etiologic differences in heart failure and this has been added to the Limitations and Future Research sections. |
| 20 | Methods | Depending on your target audience, a fuller description of the hazard ratio might be helpful. | Additional explanation added on page 20 |
| 22 | Cost-effectiveness methods | The cost-effectiveness of treatment with ace inhibitors for those with LV dysfunction has been previously demonstrated. The question regarding asymptomatic screening was interesting and clinically relevant. | No response necessary |
| 22 | Cost-effectiveness methods | Overall the author did a superb job with this complex question. Hats off. | No response necessary |
| 22 | Cost-effectiveness methods | Model did not consider any therapy of LV dysfunction other than ACEI. | This is noted in the Limitations section. Currently only ACEi has been studied in a randomized trial of asymptomatic patients. |
| 22 | Cost-effectiveness methods | Model did not consider that many patients w/ LV dysfunction may need to be screened for coronary disease, which would drive up costs. | We determined “needed” treatments/tests based on randomized trial data and clinical guidelines for which only ACE inhibitors qualified. Screening for coronary disease will increase cost and likely benefits. However, the effectiveness (and cost-effectiveness) is not established and we believe such screening is not standard of care for asymptomatic patients with depressed EF. |
| 22 | Cost-effectiveness methods | Model did not consider other potential benefits of ACEI treatment on CAD, diabetes, etc (see HOPE study). | Our model applies to patients not on ACE inhibitors. The benefit observed in SOLVD is likely due in part to benefits from these groups (CAD, diabetes). |
| 22 | Cost-effectiveness methods | Rate of progression from asymptomatic LV dysfunction to symptomatic is based on SOLVD. It should be realized that patients in SOLVD had LV assessments for some reason, and are not equivalent to a totally random population. | This is an important limitation. Unfortunately there are no randomized treatment data from a totally random population. This is discussed in the Limitations section. |
| 22 | Cost-effectiveness methods | The actual annual event rates were assumed to be constant over the course of the patient's life. Is this assumption based on the SOLVD trials (at least over the first four years)? | We assumed constant a risk of death relative to the U.S. population. We determined the risk of death at year one for SOLVD, then the risk of death at year 2 conditional on surviving year one, etc. The average of these risks over the 4 year SOLVD trial (weighted by the number of patients in each years analysis) was used. |
| 22 | Cost-effectiveness methods | I'm a little surprised that you selected your baseline probability solely on the SOLVD trials, rather than meta-analyses-derived probabilities. Your sensitivity analyses mitigate this issue. | No response necessary |
| 22 | Cost-effectiveness methods | Are hospitalization rates and costs from years ago relevant to present day costs in a rapidly evolving field? | We agree that costs have changed, per- hospital day has increased while number of hospital days have decreased. Fortunately, our model was insensitive to the cost of heart failure treatment. This is noted in the Results section. |
| 22 | Cost-effectiveness methods | There have been several cost-effectiveness studies published. How does this one differ? What does it add? | Past cost-effectiveness studies have examined the treatment of symptomatic patients with ACE inhibitors. This study examines asymptomatic patients and also examines screening. |
| 22 | Cost-effectiveness methods | Why wasn't a cost-effectiveness of BB therapy considered? Was there consideration of cost-effectiveness analyses of the subgroups studies in the meta-analysis? | We limited the cost-effectiveness analyses to treatment and screening for asymptomatic patients. As yet there are no randomized trials of beta-blockers for this population. The impact of a possible additional benefit from beta-blockers on screening was evaluated with sensitivity analysis (makes screening more cost-effective). Separate cost-effectiveness analyses by race and gender was not performed. |
| 22 | Cost-effectiveness methods | Would have separated the data synthesis methods and results from that of the cost analysis. | AHRQ Evidence Report format does not allow this. |
| 22 | Cost-effectiveness methods | Unclear where the assumption that there will be a 2.7% incidence of asymptomatic LV dysfunction in asymptomatic individuals. The MONICA study found 1.5% incidence, and this was not a totally random population. | This is from reference 18 (McDonagh TA, Robb SD, Murdoch DR, Morton JJ, et al. Biochemical detection of left-ventricular systolic dysfunction. Lancet 1998;351(9095):9–13.), which describes a population screening program. |
| 22 | Cost-effectiveness methods | Data about ACEI generally are not in patients on BBs, and whether there are additive effects is unknown. | Agreed. This is noted in the Limitations section. |
| 22 | Cost-effectiveness methods | Not sure what the third hypothetical cohort is - typo? | This has been corrected. |
| 22 | Cost-effectiveness methods | The use of a single cut-point for BNP is problematic. The levels appear to go up with age and are higher in females than males. | We agree that a gender- and age-specific cut-point may improve the accuracy of BNP. However the large population based studies used a single cut-point. |
| 22 | Cost-effectiveness methods | The explanation of extended dominance was difficult for me to understand. I would try to explain it using actual base numbers. | The description of extended dominance has been revised in the Results section. |
| 27 | Methods | Literature search criteria appear strong and the selection process thorough. | No response necessary |
| 35 | Results | I'm disappointed in the poor response from individual investigators for their patient-level data. Is this response rate common for such inquiries? | This response rate is substantially worse than previous experience with obtaining additional data from original authors, where a 60% response rate is typical. |
| 37 | Results | ACEI data as it relates to the issue of CAD: As there is much data indicating the benefit of ACEI for cardiac and vascular events in patients with CAD, could some of the difference be due to lower incidence of CAD in the women? Could we look at CAD women vs. CAD men, and non-CAD women vs. non-CAD men? | Unfortunately, this is not possible without more individual patient data, since this degree of subgroup analysis is not present in published reports. Also see comment above. |
| 37 | Results | It would be useful to present absolute risk reductions in addition to relative risk reductions. Providing absolute risk reductions would allow calculation of the “number needed to treat” to save one life. | Between-study heterogeneity is generally lower on the relative scale than on the absolute scale, so the accepted approach in meta-analysis is to pool studies on the relative scale, in this case to pool relative risks or ratios of relative risks. In order to back 1 = 1 calculate a general absolute risk reduction across studies from a pooled relative risk reduction (and thereby be able to estimate a number needed to treat (NNT)), one needs to make an assumption about what the underlying risk of the outcome is in the population. This risk varies across studies, and will vary depending on the reader's experience, clinical setting, etc. Therefore, we have provided a table (see page 47) that allows the reader to determine the absolute risk reduction and associated NNT, depending on the pooled relative risk reduction and the assumption he/she wishes to make about the underlying risk in the population. |
| 37 | Results | The presence of confounding variables in the populations could have influenced the results. Perhaps this should be addressed either by further analysis of the data or at least in the discussion of the results. | This has been added to the Limitations section. |
| 37 | Results | The omission of information regarding drug dose achieved in the various sub-group is important. Could the lesser effects of ACEIs in women and beta blockers in blacks be due to dosing? | This concern has been added to the Limitations section. |
| 37 | Results | Is there a limitation to your analyses due to their univariate nature? There are likely to be a variety of characteristics associated with specific subgroups, which may influence the response to treatment. | This has been added to the Limitations section. |
| 39 | Results | I'm struck by the possible difference between your findings and those of Exner (NEJM, 2001:1351–1357). They constructed a matching white cohort and compared with blacks in SOLVD. They found no difference in effect on mortality, but a substantial difference in the effect on hospitalization. | This has been added to the text on page 39. |
| 38 | Results | As for CAD above, could the diabetes gender groups be divided up as well, as diabetes clearly changes the impact of other risks? (add both to future research if not possible at this time) | This has been added to the Future Research section as it requires more individual patient level data than we had available. |
| 39 | Results | Separate meta-analyses without the BEST trials should be done and used to draw final conclusions for BBs. Bucindilol, which was used in BEST, has intrinsic sympathomimetic activity, wile the other BBs do not. | This analysis has been done. |
| 39 | Results | I would make note of the consistency or inconsistency of the Ratio of RR in Tables 7–17. RRR in Table 7 seems very consistent, while RRs on Table 17 are very inconsistent. | See above comment. This comment reflects the difference in results in BEST and has been handled by a new primary analysis that excludes BEST. |
| 39 | Results | The grouping of beta blockers as a class might have influenced the analysis particularly in regard to race. For instance, there is evidence that bucindolol lowered plasma NE levels considerably in the BEST trials and this was likely related to the potent beta-2 blocking properties of the molecule. | See above comment. |
| 39 | Results | Most of the beta blocker black data comes from one study (BEST). | See above comment. |
| 39 | Results | It looks like the BEST data are qualitatively different from the other studies. It may be that, for whatever reasons, bucindolol is less effective than metoprolol and carvedilol in heart failure. | See above comment. |
| 39 | Results | I can't help feeling there is something odd about the BEST trial. I would like to see the effect if BEST were removed from the analysis. | See above comment. |
| 39 | Results | Finding with regard to race and beta blockers is predominantly driven by the results of BEST. Without BEST, the overall results would be close to neutral. In contrast with ACE inhibitors, the widely held view for beta blockers is that there are important pharmacologic differences from agent to agent that make extrapolation of effect from one drug to another hazardous. | See above comment. |
| 39 | Results | The analysis assumes that ACEI and BBs are all the same! I am willing to assert that this is indeed the case with ACEIs, but I am not so sure that this is the case with BBs. This is a very contentious issue at the present time. | See above comment. |
| 40 | Cost-effectiveness results | Appeared that benefits of ACEI assumed to be same in men and women. Efficacy may differ by subgroups. | This limitation is now noted in the Limitation section. |
| 40 | Cost-effectiveness results | An important part of the heart failure population involves those with diastolic heart failure. It is not clear whether BNP would effectively detect it, and an ECG certainly may not | Because there are no therapies specifically for diastolic heart failure that have been shown effective in randomized trials we have focused our analysis on those with systolic dysfunction. |
| 40 | Cost-effectiveness results | I am aware of little data on the use of BNP to screen for LV dysfunction. The report only lists one reference. | Although there have been few studies, the one by McDonagh (Reference 18) is large and well done and we believe is sufficient to base our assumptions. |
| 40 | Cost-effectiveness results | “the model predicted… These results are similar to the findings of the SOLVD prevention study.” As the SOLVD studies were used to derive the model, isn't this circular? | Yes, but it shows that we modeled what we intended to. All models should at a minimum reproduce the survival curves they were derived from. |
| 40 | Cost-effectiveness results | Cost-effectiveness analyses are very interesting and represent “new” data. | No response necessary |
| 40 | Cost-effectiveness results | The cost-effectiveness analysis results are nicely presented, although it was difficult for me to follow the Screening section. | The description of extended dominance has been revised in the Results section to make this easier to understand. |
| 40 | Cost-effectiveness results | It does not appear that sensitivity analyses were conducted to assess the importance of the proportion of patients hospitalized at the time of incident CHF diagnosis. | This was done and not reported since it had no impact on the results. This is now reported in the Results section. |
| 40 | Cost-effectiveness results | I thought the effect of using different BNP cut-offs on cost per QALY saved was fascinating and I would emphasize it more in the text- perhaps putting it into the summaries of conclusions. | We did not think this fit in the conclusions and left it in the text. |
| 40 | Cost-effectiveness results | The BNP threshold mentioned - 18 - is for a European assay. I suspect many readers will be familiar with Biosite's assay for which a comparable cutoff is around 80. Would point out which assay the 18 cut-off applies to. | This is now stated in Table 5. |
| 40 | Cost-effectiveness results | Would like to see a greater expansion on the BNP issue. BNP is being used prematurely by clinicians to apply therapy to heart failure patients. Current trials are collecting BNP in a more concerted effort to sort out its utility for prediction of events. (See John Spertus's presentation at 2002 ACC. BNP did not predict worsening of HF symptoms.) | We chose to focus on using BNP to screen asymptomatic patients for this report. We now note in the Limitations section that the use of BNP for patients to determine therapy is a separate issue. |
| 40 | Cost-effectiveness results | Do you imply that everyone over the age of 55 should have screening BNP? You don't deal directly with the impact of risk factors and history of MI on the analysis. The majority of patients with asymptomatic LV systolic dysfunction have atherosclerotic disease as the etiology. Of course, prevalence is also age-related. A point score based on age and other factors might fine-tune a cost-effective approach to screening. | Our model applies to patients not on ACE inhibitors. The benefit observed in SOLVD is likely due in part to benefits from these groups (CAD, diabetes). |
| 40 | Cost-effectiveness results | Conclusions of the model are very interesting, and gratifyingly robust. | No response necessary |
| 40 | Cost-effectiveness results | Limitations of the model might be presented more fully. Future research (rather than limitation): the difference between patients with known CAD or history of MI and no history. The population with this known history creates a concentrated one in which the benefits of screening for low EF may be even more obvious. Conversely, patients with no known history of any cardiovascular disease may have less benefit. | We agree that it is the risk of depressed ejection fraction, not age alone, that is the prime determinant of the cost-effectiveness of screening. For populations with at least 1.5% prevalence of depressed EF, screening is a reasonable value. Further work to develop such a scoring system would be helpful to determine optimal screening candidates. |
| 40 | Cost-effectiveness results | Limitation: inevitable with the current data sets is the lack of any information on truly asymptomatic patients with no history. This should be stated clearly. It is not clear how the SOLVD patients for the prevention arm were identified, but someone had already been concerned enough to obtain a measure of LV function. We would all anticipate that patients who have never come to medical attention for cardiovascular disease would have a better outcome with asymptomatic disease than those who were already under surveillance. | This is an important limitation and is discussed in the Limitations section. |
| 40 | Cost-effectiveness results | Should be cost of BNP test less than $120 (not $170). | This error has been corrected. |
| 40 | Cost-effectiveness results | The $200 price for an echo sounds way too low. | We estimated the cost of the least expensive echocardiogram that could determine LV systolic function (no Doppler needed). |
| 42 | Cost-effectiveness results | Reference to Table 21 seems incorrect. | The reference has been corrected. (Table 19). |
| 52 | Results | Table 17 (now Table 18) - should the RRR for the US Carvedilol trials be 1.39 rather than 1.15? Figure 12 appears to place it correctly. | Thank you for pointing out this problem. The table was incorrect due a transposition of numbers. The correct RRR in the table should be 1.41, not 1.14. The confidence interval of (0.43,4.68) is correct in the table. The graph is correct. |
| 52 | Results | In Figure 12 and Table 17 the ratio of relative risks for US Carvedilol seems inconsistent - about 1.35 in the figure and 1.14 in the table. | See above comment. |
| 57 | Results | Graphs of data display relative risk of benefit between groups, as opposed to relative risk of placebo vs. Rx in the groups of interest. | No response necessary |
| 71 | Cost-effectiveness results | The sensitivity analyses Figures 19–21 are difficult to interpret and would benefit from a more detailed figure legend. | These figures and their legends have been revised. |
| 81 | Conclusions | Occasional ambiguity “Neither, however is there evidence that ACE, inhibitors help women with heart failure… The results suggest but do not prove that ACE inhibitors have a beneficial effect on mortality in women with heart failure.” | This section has been rewritten. |
| 81 | Conclusions | Would recommend including a paragraph or two on how the authors view their findings being applied and to whom. For example, do the authors think that insurers will use these data to apply use of appropriate therapy as a quality measure assessment? Will CMS use for reimbursement justification? | Evidence Reports are specifically prohibited from suggesting possible practice or policy implications of the evidence. |
| 81 | Conclusions | State clearly that you are not advocating changing any treatment recommendations based on the subgroup analyses. Rather, the results should stimulate further investigation. | This section has been rewritten. |
| 81 | Conclusions | Report does not emphasize the multiple assumptions which lead to the stated conclusions. | This has been added. |
| 81 | Conclusions | There should be a Limitations section for the Cost-Effectiveness analysis and Meta-Analysis sections. | This has been added. |
| 81 | Conclusions | I am somewhat concerned about the release of some of the information such as the “not helpful” or “harmful” impression for beta-blockade in diabetic or Black patients rendered to non-scrutinizing MDs and the general public. This could have an unintended, potentially harmful effect on patients. | The Conclusions section has been rewritten to try to avoid creating this impression. |
| 83 | Future Research | The majority of beta blocker trials (except BEST) found a benefit in both blacks and whites. Is it ethical to perform further placebo controlled studies in blacks to see if this benefit is as large as in whites? | This is a question beyond the scope of the evidence report. We note that the pooled RR of mortality effect in blacks of non-BEST beta-blocker studies is not statistically significant. |
| 83 | Future Research | I am in total agreement that additional studies need to be done, in particular in the elderly and diabetic patients. | No response necessary |
| 83 | Future Research | As we move forward to screen truly asymptomatic patients, there will be some finite costs to the new diagnosis of a disease condition. This could also be mentioned as an area of future research - appropriate counseling and measurement for these costs. For patients with some other pre-existing condition, the benefit of ACEI for newly diagnosed heart failure may be diminished by those patients already on ACEI or ARB for other conditions such as HTN or diabetes. | We agree that there are unclear costs of a disease diagnosis and the need for future research is now noted. We also agree that as ACE inhibitors are more widely used, the benefits of screening will decrease. |
| 83 | Future Research | Addressing the cost-effectiveness of ACEI in women would tie in the two main methodologies of the report well. | No response necessary |
| 83 | Future Research | How would the addition of BBs to ACEI affect the cost-effectiveness of screening? Since you did not do this analysis, a statement addressing this may be helpful. | This is now stated in the Limitations section. |
| 83 | Future Research | We should also call for more controlled studies in black patients. | This clarification has been added to the Future Research section. |
| 83 | Future Research | Future work should focus on potential barriers to use of beta-blockers in patients with heart failure, including practitioner, patient, and drug-related barriers. | This has been added to the Future Research section. |
| 83 | Future Research | Future work should focus on the outcome of patients screened for heart failure with BNP and/or echocardiograms, including false positives. | This is now stated in the Future Research section. |
| 83 | Future Research | What are the prospects for answering the Original Potential Key Questions? What kind of data are required? What kind of studies? | This change has been made to the Future Research section. |
| 83 | Future Research | The implications of the findings of this project for research are understated. Perhaps add a final section “the implications, significant, and application of the findings of this project report to futures studies and trials.” | We have rewritten the Future Research section to more accurately reflect the implications of our findings. |
| 83 | Future Research | The importance of outcomes other than mortality needs to be stressed. | This has been added to this section. |
| 83 | Future Research | Mortality is probably the most appropriate end-point of use. However, information regarding the development of heart failure (in the post MI and SOLVD prevention pops) and hospitalizations might be interesting to include in the analysis. | This has been added to this section. |
| 83 | Future Research | A major point that could and should be stressed in the final document is the need to consider issues related to subgroups when studies are being designed. Under-representation of female patients and minorities in clinical trials remains a problem. | This has been added to this section. |
| 83 | Future Research | Heart failure trials have not been powered to address specific questions related to gender, race, presence or absence of ischemic heart disease, and presence or absence of diabetes. Perhaps the most important message coming from this report is that greater participation in trials by these subsets is needed. | See above comment |
| 83 | Future Research | This report should spark more basic research into molecular biodynamics that characterize race, gender, and disease-specific heart failure issues. | This has been added to this section |
| 83 | Future Research | Would like to see a better developed group of suggestions for future trials to analyze these very important and relevant subgroups. | This has been done to the extent customary in AHRQ evidence reports. |
Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]Free Full text in PMC]
