Skip Navigation

European Journal of Heart Failure 2003 5(6):717-723; doi:10.1016/S1388-9842(03)00101-6
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Zanolla, L.
Right arrow Articles by Zardini, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Zanolla, L.
Right arrow Articles by Zardini, P.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© 2003 European Society of Cardiology

Selection of endpoints for heart failure clinical trials

Luisa Zanolla* and Piero Zardini

Dipartimento di Scienze Biomediche e Chirurgiche, Sezione di Cardiologia, Università degli Studi di Verona, Ospedale Civile Maggiore Piazzale A. Stefani 1, Verona 37126, Italy

* Corresponding author. Tel.: +39-45-8072320; fax: +39-45-914727. E-mail address: luisa.zanolla{at}univr.it


   Abstract
Top
 Abstract
1. Introduction
 2. The primary endpoint...
 3. 'Hard' endpoints
 4. 'Soft' endpoints
 5. Surrogate endpoints? Not...
 6. Explanatory endpoints
 7. Conclusions
 References
 
In assessing the efficacy and the safety of a new drug, randomized clinical trials represent the standard scientific method. The selection of the best response variables in a clinical trial of a treatment in congestive heart failure patients is often not straightforward; the primary end point of a trial should be clinically relevant, directly related to the primary goal of the trial, and with favorable distributional properties. All-cause mortality is undoubtedly the most unbiased endpoint, but there is interest both in assessing cause-specific mortality and hospitalization rate and in evaluating ‘soft’ endpoints (functional status, exercise tolerance); the latter, in fact, are clinically relevant and potentially more useful in mild heart failure patients. Physiopathologic variables (e.g. left ventricular function) could provide information on drug action mechanism. In this paper, several recent large clinical trials are reviewed and the advantages and drawbacks of the response variables used, are analyzed.

Key Words: Clinical trials • Chronic heart failure • Survival • Outcomes • Mortality • Morbidity

Received November 19, 2001; Revised January 27, 2003; Accepted June 16, 2003


   1. Introduction
Top
 Abstract
 1. Introduction
2. The primary endpoint...
 3. 'Hard' endpoints
 4. 'Soft' endpoints
 5. Surrogate endpoints? Not...
 6. Explanatory endpoints
 7. Conclusions
 References
 
Patients with heart failure have high rates of mortality, hospitalization and morbid complications, along with multiple symptoms, and severe limitations in daily activities, thus resulting in a poor quality of life. Treatment of heart failure is aimed at prolongation of survival, amelioration of symptoms and of quality of life, prevention of morbidity, reduction of hospitalization and slowing down disease progression.

Randomized clinical trials represent the standard scientific method for assessing the efficacy of any treatment, and the basis for the approval of new drugs by governmental regulatory agencies; however, there is no unanimous agreement on the endpoints to be used for heart failure trials. The selection of the best response variables for the assessment of the efficacy of a treatment in congestive heart failure patients is thus still under debate [1]; the problem is relevant for the physician who is caring for the individual patient, but even more for the clinical trial investigator.


   2. The primary endpoint of a trial
Top
 Abstract
 1. Introduction
 2. The primary endpoint...
3. 'Hard' endpoints
 4. 'Soft' endpoints
 5. Surrogate endpoints? Not...
 6. Explanatory endpoints
 7. Conclusions
 References
 
The primary endpoint of a trial, i.e. the main response variable, should be a clinical event significant for the patient, clinically relevant and directly related to the primary goal of the trial.

There should generally be only one primary variable—specified in the study protocol—which is used when estimating the sample size required for the study [2]. A redefinition of the primary variable after unblinding the results, is almost always judged unacceptable [2]; in fact, a post-hoc analysis does not have protection against a type I error, with the possibility of statistically significant results occurring only by chance. There was debate on this topic following publication of the US Carvedilol Trial [3]. The paper presented the combined results of four protocols, each enrolling patients with heart failure of different clinical severity. Both total mortality and the risk of hospitalization were significantly reduced by carvedilol; however, neither of these variables had been defined as a primary or secondary endpoint in any of the four study protocols [4]. In fact, the specified endpoints were exercise distance and a quality of life score; in only one study was a composite endpoint of mortality and hospitalization specified. This last study was the only one to have a statistically significant result, however, the negative results in the other studies were not reported, and the total mortality reduction was presented as the mainstay of the trial.

Additional data-derived analyses on endpoints other than the one defined prospectively as the primary one, might be useful in suggesting hypotheses for further studies, but the conclusions should not be considered in the same manner as the principal hypothesis of the study design. This rule, however, is often violated. For instance, in the PRECISE trial [5], exercise tolerance was defined as the primary endpoint of the study. The authors pointed out that this decision had been influenced by the US Food and Drug Administration, which focused on exercise capacity as a principal variable in the assessment of new heart failure drugs. No effect was evident in the treadmill test, while in the 6-min walk test, there was a borderline significant difference favoring the carvedilol group. Nevertheless, the paper presenting the results emphasized the significant improvement in New York Heart Association (NYHA) class and in a disease severity score. The same happened with the Multicenter Oral Carvedilol Heart failure Assessment (MOCHA) trial [6], whose primary endpoint was submaximal exercise; no detectable effect was evident, but the paper emphasized the dose-related improvement in left ventricular function, and both in mortality and in hospitalization rate.

Following the protocol pre-specified analysis strictly, is the only way of avoiding a bias [7]; data driven analyses can only be reported providing they are unequivocally identified as post-hoc assessments, whose significance may only be due to chance.

A different role is played by the secondary variables when the principal one changes significantly under treatment: changes in secondary variables can be useful in the interpretation of the main results. In the MERIT-HF trial [8], for instance, there was a statistically significant difference between patients on metoprolol and patients on placebo, both in all-cause mortality and in the time to the first event, which had been defined as the principal endpoints. The authors also reported a statistically significant difference in a number of secondary variables, cardiovascular mortality, sudden death worsening heart failure; these results help in interpreting the main analysis and could suggest hypotheses for further studies.


   3. ‘Hard’ endpoints
Top
 Abstract
 1. Introduction
 2. The primary endpoint...
 3. 'Hard' endpoints
4. 'Soft' endpoints
 5. Surrogate endpoints? Not...
 6. Explanatory endpoints
 7. Conclusions
 References
 
3.1. Mortality as an endpoint
Pharmacological treatment in heart failure aims to improve patient's symptoms but mainly to prolong his life; therefore, mortality is the commonest endpoint used in heart failure clinical trials. Mortality, as an endpoint, has the advantage of being a ‘hard’ endpoint, easy to measure, not readily subject to observer bias and it clearly represents an important event for the patient himself.

One major problem in planning mortality studies is that these studies usually require a large sample size, mainly in patients with initial heart failure. The number of patients required to achieve a given statistical power for a study is a function both of the risk in the control group, and of the reduction in the risk hypothesized as an effect of the intervention; therefore, with declining mortality, the sample size required gets larger. One strategy could be to enrol high-risk patients, allowing the design of trials with a smaller sample size, due to the high event rate, but the generalizability of the results thereafter is limited.

All-cause mortality is the most unbiased endpoint. It has been used in several clinical trials [8,9].

3.2. Adjusted all-cause mortality
The use of survival regression models allows the investigator to adjust for any imbalance in prognostic variables that could be present between study groups, notwithstanding a carefully planned randomization, in particular for smaller sample size; this adjustment is methodologically acceptable, provided the analysis is pre-planned.

In the first Vasodilator Heart Failure Trial (V-HeFT) [10], there was no significant mortality difference between placebo and prazosin at the pre-planned 2-year survival analysis. The authors, however, pointed out that the log-rank statistic did not take into account important prognostic covariates, such as ejection fraction, history of coronary artery disease, heart rate and peak oxygen consumption. When these and other covariates were included in a Cox regression model, the reduction in mortality in the hydralazine–nitrate group did reach statistical significance.

It is always tempting to reanalyze the negative results of a trial when the pre-planned analysis is inconclusive, but the conclusions drawn are not statistically sound.

Only recently, a methodologically correct example was presented in the Val-HeFT trial [11] with the use of adjusted mortality to assess the size of the treatment effect, by a Cox regression model, with pre-specified clinically relevant prognostic covariates, including NYHA class, ejection fraction, cause of heart failure, age, use of ACE-inhibitor and of β-blocker.

3.3. Cause-specific mortality
Cause-specific mortality appears to be obtained relatively easily, however, the definition of the events according to the mode of death should be considered cautiously. The definition of sudden death, for instance, changes dramatically from one study protocol to another; some trials have used a time-dependent definition, such as 1 h since the onset of new symptoms, as used in the CONSENSUS [12] and ELITE [13] trials. The V-HeFT-I trial defined sudden death as either ‘observed to be instantaneous’ or ‘unwitnessed but assumed to be instantaneous on the basis of the clinical settings’. This heterogeneity of definition is shared also by death due to progressive heart failure [14], and further complicated by the inclusion of ‘intermediate’ classifications, such as death due to ‘heart failure or arrhythmias with heart failure’ [15] or ‘sudden death with some premonitory signs of worsening heart failure’ [16].

In order to standardize and verify the classification of deaths, most multi-center trials have a central adjudication committee, but its value is still under debate [7], and important trials like SOLVD [15] had the cause of a patient's death assessed only by the principal investigator at each center. However, in acute coronary syndrome trials [17], the assessment of events by endpoint committee changed the definition of events by the local investigators in a percentage sufficient to change the statistical conclusions of the studies.

A classification of the mode of death has been proposed [12], based on activity and place of death, cause of death, mode of death and events associated with death; it was used for classifying events in the MERIT-HF trial [8].

When considering cause-specific mortality, the number of events is reduced, and this reduces the statistical power of the analysis to detect any difference between the treatment groups, if present. In the absence of a difference, it is almost impossible to discriminate between a real absence of difference or insufficient statistical power of the trial to detect a possibly existing difference.

When a difference in a specific cause of death, is statistically significant, the comparison can be considered informative, providing it was a pre-planned one.

For instance, in the V-HeFT II trial [16], the comparison of all-cause mortality had a borderline significance, however, there was a significant difference in the rate of sudden death. Such post-hoc comparisons, however, should be considered cautiously and a significant difference in a specific mode of death does not offset the lack of difference in all-cause mortality. Even more complex is the interpretation of the results of the DIG trial [9] with total mortality, the primary outcome variable, unaffected by treatment. However, there was a trend to a reduction in mortality for heart failure worsening and a significant increase in death due to arrhythmias and other cardiovascular causes.

3.4. Rate of (re-)hospitalization as an endpoint
Patients with heart failure also have an impaired quality of life and increased morbidity requiring frequent hospitalizations. In the SOLVD registry approximately 35–40% of the patients had been hospitalized at least once every year [7]. Therefore, the rate of hospitalization is often considered as a secondary endpoint in clinical trials. The event of hospitalization apparently represents an objective, ‘hard’ endpoint; one should however, take into account the definition of hospitalization (either a short ward stay or a real hospitalization), different thresholds for hospitalization in centers, and different regional policies for hospitalization.

Moreover, it is difficult to define the main cause for hospitalization, especially in patients with several concomitant diseases, where one can speculate whether the patient was hospitalized with heart failure, or because of heart failure.

3.5. Composite variables
When a single primary variable cannot be selected according to the primary objective of the trial, another useful strategy is to integrate or combine multiple variables into a single or composite variable, by using a pre-defined algorithm. This approach has the advantage of increasing the number of events, and deals with the multiplicity problem without requiring an adjustment for a type I error.

In the SOLVD prevention trial [18] of asymptomatic patients with left ventricular dysfunction, only the composite endpoint (all-cause mortality plus hospitalization) showed a significant difference between patients on placebo and patients on enalapril, while the reduction in total mortality was not statistically significant; both outcomes had been defined as primary response variables. The reporting of such a result must be very clear, in order to prevent misinterpretation, otherwise it could be viewed as a positive mortality–morbidity result. While total mortality was not significantly affected by enalapril, the sample size of the study had been calculated on this endpoint, so that the study was not under-powered for detecting a difference. Also, the Val-HeFT trial [11] had both total mortality and a composite outcome variable (mortality or cardiac arrest with resuscitation or hospitalization for heart failure) as primary endpoints. The definition of two endpoints was rigorously handled by the study protocol by testing each endpoint at a significance level of 0.025, a value based on the Dunn–Sidak inequality [19]. Total mortality was unaffected by treatment, while the composite endpoint was significantly reduced by valsartan, almost exclusively due to a significant reduction in hospitalization rate; in the discussion section, however, the ‘benefit in terms of morbidity and mortality’ is emphasized.

The definition of the composite endpoint must be clearly stated. One of the US Carvedilol Studies [20], for instance, assessing mild heart failure patients, defined progression of heart failure as a composite endpoint, considering death or hospitalization for heart failure or the need for a sustained increase in heart failure medication. It was clearly stated that the events of the composite endpoint were considered in a descending hierarchical order, so that one patient could only be counted for one event.


   4. ‘Soft’ endpoints
Top
 Abstract
 1. Introduction
 2. The primary endpoint...
 3. 'Hard' endpoints
 4. 'Soft' endpoints
5. Surrogate endpoints? Not...
 6. Explanatory endpoints
 7. Conclusions
 References
 
The use of mortality as an endpoint of a clinical trial is important in the advanced stages of the disease, however, it seems inadequate to test the effect of a treatment in the earlier stages. In fact, in patients with mild symptoms, a reduction in the progression of the disease and an improvement in the functional status is the main aim. Nevertheless, mortality trials remain important in less severe degrees of heart failure. Asymptomatic class I–II patients enrolled in the SOLVD prevention trial [18] and randomized to enalapril treatment had a 15% mortality at 4 years.

4.1. Functional status
Functional status is often assessed by the NYHA classification system. It is limited to only four levels, which range from no limitation to bedridden, and can easily miss clinically important variations if the change is relatively small in magnitude, and inter-observer agreement is often poor [21]. Moreover, functional status, assessed by either NYHA classification or by any other instrument based on history, can be biased by an apparent improvement when the patient merely reduces stressful activities [22]. This may lead to an underestimate of functional disability, as evidenced when peak oxygen consumption on exercise testing is considered as a reference [23].

In addition, an improvement is often evident in the placebo arm of trials, suggesting that the increased attention paid by the physician to the patient enrolled in a clinical trial, has a positive effect on the patients’ symptoms, whatever pharmacological treatment is being administered.

4.2. Quality of life
Several scores have been proposed for assessing quality of life in heart failure patients. Their use in clinical trials implies that the score is able to change according to the effect of the treatment.

The 21-item Minnesota Living with Heart Failure (MLHF) questionnaire [24] was validated by correlating its changes with the variations in exercise tolerance and with patients’ ratings of dyspnea and fatigue induced by treatment. The correlation with changes in exercise time, although weak, was statistically significant, and associated with variations in symptom ratings. Nevertheless, in several studies the assessment of quality of life using this questionnaire, after pharmacological intervention, was inconclusive [6,25,26], and a positive response in the placebo arm was common, both with MLHF and with other scores [20,27].

4.3. Worsening heart failure
A commonly used endpoint, albeit usually a secondary one, is worsening heart failure; however, its apparent objectivity can be limited by different definitions. Some trials, like the ANZ trial [29], used a wide definition of worsening heart failure, by including not only hospital admission for worsening symptoms of heart failure, or a non-sudden death from heart failure, but also an increase in non-study treatment requirements, or an increase in NYHA functional class; such definitions obviously allow a large variability. A more ‘objective’ definition was used in the VEST trial [28] in which worsening was defined as any variation requiring treatment with intravenous inotropic or vasodilator drugs for at least 4 h. The same definition was used in the Val-HeFT trial [11] as one of the events defining morbidity in the composite endpoint. Given the imprecise definition of ‘worsening’ and the differences in functional classification of the patients enrolled and in the length of the follow-up, the rate of worsening heart failure ranges from 34.7% over more than 4 years in the DIG trial to 57.9% in the 18-month ANZ trial.

4.4. Objective evaluation of functional capacity
A more objective assessment of functional status can be obtained by the assessment of exercise capacity. If a reduction in the ability to cope with physical exercise is the main limitation in a heart failure patient, it seems quite logical to assess the effect of a new drug by evaluating changes in the tolerance to effort.

For example, the ANZ Carvedilol Trial [29] had exercise duration as a principal endpoint; as in several other carvedilol trials, notwithstanding the significant effect on survival, the drug had no effect on exercise time.

In the V-HeFT II trial [25] the peak oxygen consumption was assessed; the paper points out that the tests interrupted for reasons other than dyspnea were not included in the analysis. This represents a further source of bias in trials using functional capacity as an endpoint; patients who died before the test or who did not perform the test for a worsening heart failure or for other complications, are not included in the final analysis. If the study drug is more effective than placebo (or reference treatment) in reducing mortality and worsening of heart failure, there will be more patients who will not be able to exercise in the placebo group; these patients are likely to be the worst ones and therefore, the result in the placebo group will be overestimated due to the fact that only the ‘best’ patients remain in the trial, thus blunting any difference, which may exist, between the two study treatment groups. An ‘intention-to-treat’ analysis is obviously not possible in this situation, but Lubsen [30] proposed to consider as zero the exercise tolerance of a patient who died or is unable to exercise, and then to present the results as median value, expressing variations as 25th and 75th percentiles, then analyzing the results using non-parametric techniques. If the results of such an analysis are in agreement with the conventional one, a person can feel more confident in the conclusions of the study. Another proposal for avoiding this source of bias was applied in the PICO trial [31], which assessed exercise tolerance at 4, 12 and 24 weeks after treatment allocation: when patients had done at least one test, whatever the reason for which the following test had not been performed, the last exercise time obtained was carried forward, to replace the missing values. This kind of solution, however, implies some further potential biases.

The problem of non-completers is shared by other endpoints requiring a specific measurement at the final visit.

For the assessment of functional capacity, the 6-min walk test has been proposed as an alternative. It can be performed in the absence of stress test equipment and even by the more compromised patients. In a trial on carvedilol in severe heart failure [32], carvedilol patients significantly increased the distance walked, but the difference was below the short-term (1 day) reproducibility value [33], although it was a long-term difference. In the recently published MUSTIC trial [34], the distance walked in 6 min was the principal endpoint of the study. A significant increase was reported with active pacing, with a mean difference of 73 m, thus greater than the 55-m minimum relevant difference between two examinations at 1-day interval. The study, however, lasted 3 months and the reproducibility of long-term comparisons has not yet been assessed, to our knowledge.


   5. Surrogate endpoints? Not just semantics
Top
 Abstract
 1. Introduction
 2. The primary endpoint...
 3. 'Hard' endpoints
 4. 'Soft' endpoints
 5. Surrogate endpoints? Not...
6. Explanatory endpoints
 7. Conclusions
 References
 
Mortality has long been the standard response variable in heart failure trials. The interest in treating earlier phases of the disease and the problems in running trials with several thousand patients, has focused great interest on surrogate endpoints. A good definition of a surrogate end point is given by Temple:

A laboratory measurement or a physical sign used as a substitute for a clinically meaningful endpoint that measures directly how a patient feels, functions or survives. Changes induced by a therapy on a surrogate endpoint are expected to reflect changes in a clinically meaningful endpoint [35].

The advantage of the use of a surrogate end point is that usually only few hundred patients are required to achieve a high statistical power. A surrogate end point must obviously be a correlate of the true clinical outcome, but the correlation itself is not sufficient to identify a surrogate. The end point must have a biological relevance, and the effect of the treatment on the surrogate end point must be able to predict the effect of the treatment on the clinical outcome [36].

A pragmatic definition is provided by the Guidelines of the European Agency for the Evaluation of Medicinal Products [2]:

A variable that provides an indirect measurement of effect in situations where direct measurement of clinical effect is not feasible or practical

by mentioning surrogate endpoints, although with cautious warnings, these guidelines implicitly allow their use.

However, the use of the word ‘surrogate’ and the concept itself have been questioned [37]. If ‘surrogate’ means the replacement of an outcome variable with a physiopathologic variable, then the word itself can be misleading, and the concept is no longer acceptable. Trials using physiopathologic variables as endpoints cannot represent a ‘surrogate’ for mortality evaluation, but they aim at a better understanding of the way a treatment works.

A classification of clinical trials was proposed several years ago by Schwartz and Lellouch [38], who outlined the existence of two completely different kinds of trial. Explanatory trials, conducted in a small group of patients under conditions which could be met only in a study; a sort of laboratory environment and pragmatic trials, conducted on large groups of patients, in conditions resembling the current medical practice. Explanatory trials conducted in a small group of patients under conditions which could be met only in a study, a sort of laboratory environment, and pragmatic trials drive the choice between two modes of therapy. In this framework the so-called ‘surrogate’ variable uses an explanatory trial result to predict clinical outcomes; but an increased survival rate cannot be inferred by a favorable change in exercise capacity or cardiac function.

A definition of ‘intermediate’ endpoints has been proposed, but there is no full agreement on its meaning, ranging from variables related to the biological basis of the response to treatment (e.g. blood pressure) [39], to a combination of ‘soft’ clinical endpoints and functional measures [40]. We therefore, propose to classify as ‘explanatory’ endpoints those variables that can be useful as explanatory trial endpoints.


   6. Explanatory endpoints
Top
 Abstract
 1. Introduction
 2. The primary endpoint...
 3. 'Hard' endpoints
 4. 'Soft' endpoints
 5. Surrogate endpoints? Not...
 6. Explanatory endpoints
7. Conclusions
 References
 
Physiopathologic or ‘explanatory’ endpoints are useful in assessing the mechanism of action of a drug; they are therefore of interest in the initial drug assessment phase, as in phase II trials. They could also play a role as secondary endpoints in phase III trials.

6.1. Non-invasive measurements of left ventricular function as endpoints
The Veterans Affairs Cooperative Studies V-HeFT I and II [41] used left ventricular function measurements to evaluate the effect of therapeutic interventions: the M-mode measurement variations reported, however, were in the range of the reported intra-observer reading errors; moreover, no long-term reproducibility data are available.

Measurements of left ventricular function are strictly correlated to prognosis and in many studies their behavior parallels the changes in mortality with treatment, like the increase of ejection fraction in several carvedilol trials [6,20,29,32] and in a sub-study of the SOLVD enalapril trial [42].

However, there are trials in which a significant improvement in left ventricular function was present in association with an opposite effect on mortality and morbidity. In fact when the results of the V-HeFT II trial [25] were analyzed, the increase in the radionuclide ejection fraction was greater in the hydralazine–isosorbide dinitrate arm than in the enalapril arm, being statistically significant at the 13-week assessment; nevertheless, it was enalapril which increased the survival rate compared to hydralazine–isosorbide dinitrate.

6.2. Neurohormone levels as endpoints
Plasma hormone levels can be quite accurately measured in the laboratory. Changes in hormone levels have been demonstrated in response to pharmacological interventions [16], but a demonstration of their role in steering therapy is still missing. One example of their use in a clinical trial is the V-HeFT III Trial [43], in which plasma norepinephrine was unaffected by felodipine therapy, while atrial natriuretic peptide levels were declared to differ significantly after a logarithmic transformation, although the time trend appeared to overlap in the two treatment arms.


   7. Conclusions
Top
 Abstract
 1. Introduction
 2. The primary endpoint...
 3. 'Hard' endpoints
 4. 'Soft' endpoints
 5. Surrogate endpoints? Not...
 6. Explanatory endpoints
 7. Conclusions
References
 
In conclusion, in the study of heart failure treatment, each variable represents a different aspect of the response to a pharmacological intervention and can be relevant in different stages of the disease or in different phases of a new drug assessment. The conclusions drawn may also be apparently conflicting, according to which aspect is considered the most relevant.

The interest for the clinician caring for the individual patient, however, is for the pragmatic results [44]; the influence of a drug on cause-specific death rate may be of scientific interest, but it is almost irrelevant from the patient’ point of view.

In designing clinical trials, investigators should try to standardize response variables, keeping the attention on clinical significance, and considering that each one has methodological limitations, and should therefore, be critically assessed, as a part of the more complex picture.


   References
Top
 Abstract
 1. Introduction
 2. The primary endpoint...
 3. 'Hard' endpoints
 4. 'Soft' endpoints
 5. Surrogate endpoints? Not...
 6. Explanatory endpoints
 7. Conclusions
 References
 

  1. Packer M. How should we judge the efficacy of drug therapy in patients with chronic congestive heart failure? The insights of the six blind men. J Am Coll Cardiol (1987) 9:433–438.[Web of Science][Medline]
  2. European Agency for the Evaluation of Medicinal Products. Note for the guidance on statistical principles in Clinical Trials. ICH Topic E9, February 1998.
  3. Packer M., Bristow M.R., Cohn J.N., Colucci W.S., Fowler M.B., Gilbert E.M., et al. The effect of carvedilol on morbidity and mortality in patients with chronic heart failure. N Engl J Med (1996) 334:1349–1355.[Abstract/Free Full Text]
  4. Fisher L.D., Moyé L.A. Carvedilol and the food and drug administration approval process: an introduction. Control Clin Trials (1999) 20:1–15.[CrossRef][Web of Science][Medline]
  5. Packer M., Colucci W.S., Sacker-Bernstein S., Liang C.S., Goldscher D.A., Freeman I., et al. Double-blind, placebo-controlled study of the effects of carvedilol in patients with moderate to severe heart failure. The PRECISE trial. Circulation (1996) 94:2793–2799.[Abstract/Free Full Text]
  6. Bristow M.R., Gilbert E.M., Abraham W.T., Adams K.F., Fowler M.B., Hershberger R.E., et al. Carvedilol produces dose-related improvements in left ventricular function and survival in subjects with chronic heart failure. Circulation (1996) 94:2807–2816.[Abstract/Free Full Text]
  7. Yusuf S., Garg R. Design, results and interpretation of randomized, controlled trials in congestive heart failure and left ventricular dysfunction. Circulation (1993) 87(Suppl. VII):115–121.[Web of Science]
  8. Merit-HF Study Group. Effect of metoprolol CR/XL in chronic heart failure: Metoprolol CR/XL randomised intervention trial in congestive heart failure (MERIT-HF). Lancet (1999) 353:2001–2007.[CrossRef][Web of Science][Medline]
  9. The Digitalis Investigations Group. The effect of digoxin on mortality and morbidity in patients with heart failure. N Engl J Med (1997) 336:525–533.[Abstract/Free Full Text]
  10. Cohn J.N., Archibald D.G., Ziesche S., Franciosa J.A., Harston W.E., Tristani F.E., et al. Effect of vasodilator therapy on mortality in chronic congestive heart failure. Results of a Veterans Administration Cooperative Study. N Engl J Med (1986) 314:1547–1552.[Abstract]
  11. Cohn JN, Tognoni G, for the Valsartan Heart Failure Trial Investigators. A randomized trial of the angiotensin-receptor blocker valsartan in chronic heart failure. N Engl J Med 2001;345:1667–75.
  12. The CONSENSUS Trial Study Group. Effects of enalapril on mortality in severe congestive heart failure. N Engl J Med (1987) 316:1429–1435.[Abstract]
  13. Pitt B., Segal R., Martinez F.A., Meurers G., Cowley A.J., Thomas I., et al. Randomised trial of losartan versus captopril in patients over 65 with heart failure (Evaluation of Losartan in the Elderly Study, ELITE). Lancet (1997) 349:747–752.[CrossRef][Web of Science][Medline]
  14. Narang R., Cleland J.G.F., Erhardt L., Ball S.G., Coats A.J., Cowley A.J., et al. Mode of death in chronic heart failure. A request and proposition for more accurate classification. Eur Heart J (1996) 17:1390–1403.[Abstract/Free Full Text]
  15. The SOLVD Investigators. Effect of enalapril on survival in patients with reduced left ventricular ejection fractions and congestive heart failure. N Engl J Med (1991) 325:293–302.[Abstract]
  16. Cohn J.N., Johnson G., Ziesche S., Cobb F., Francio G., Tristani F., et al. A comparison of enalapril with hydralazine–isosorbide dinitrate in the treatment of chronic congestive heart failure. N Engl J Med (1991) 325:303–310.[Abstract]
  17. Mahaffey KW, Granger CB, Tardiff BE, Lee KL, Weaver WD, Topol EJ, et al. for the GUSTO IIb Investigators. Endpoint adjudication by a Clinical Events Committee can impact the statistical outcome of a clinical trial: results from the GUSTO II-b. J Am Coll Cardiol 1997;29:410A.
  18. The SOLVD Investigators. Effects of enalapril on mortality and the development of heart failure in asymptomatic patients with reduced left ventricular ejection fractions. N Engl J Med (1992) 327:685–691.[Abstract]
  19. Kirk R.E. Experimental design (1982) Pacific Grove: Brooks Cole. 110–111.
  20. Colucci W.S., Packer M., Bristol M.R., Gilbert E.M., Cohn J.N., Fowler M.B., et al. Carvedilol inhibits clinical progression in patients with mild symptoms of heart failure. Circulation (1996) 94:2800–2866.[Abstract/Free Full Text]
  21. Goldman L., Hashimoto B., Cook E.F., Loscalzo A. Comparative reproducibility and validity of systems for assessing cardiovascular functional class. Advantages of a new specific activity scale. Circulation (1981) 64:1227–1233.[Abstract/Free Full Text]
  22. Goldman L., Cook E.F., Mitchell N., Flatley M., Sherman H., Cohn P.F. Pitfalls in the serial assessment of cardiac functional status. How a reduction in ‘Ordinary’ activity may reduce the apparent degree of cardiac compromise and give a misleading impression of improvements. J Chronic Dis (1982) 35:763–771.[CrossRef][Web of Science][Medline]
  23. Wilson J.R., Hanamanthu S., Cjomsky D.B., Davis S.F. Relationship between exertional symptoms and functional capacity in patients with heart failure. J Am Coll Cardiol (1999) 33:1943–1947.[Abstract/Free Full Text]
  24. Rector T.S., Cohn J.N. Assessment of patients outcome with the Minnesota living with heart failure questionnaire: reliability and validity during a randomized, double-blind, placebo controlled trial of pimobendan. Am Heart J (1992) 124:1017–1025.[CrossRef][Web of Science][Medline]
  25. Rector T.S., Johnson G., Dunkman W.B., Daniels G., Farrell L., Henrick A., et al. Evaluation of patients with heart failure of the effects of enalapril compared with hydralazine plus isosorbide dinitrate on quality of life. V-HeFT II. Circulation (1993) 87(Suppl.VI):71–77.[Web of Science]
  26. Hjalmarson Å., Goldstein S., Fagerberg B., Wedel H., Waagstein F., Kjekshus J., et al. Effects of controlled-release metoprolol on total mortality, hospitalizations, and well-being in patients with heart failure. The Metoprolol CR/XL randomized intervention trial in congestive heart failure (MERIT-HF). J Am Med Assoc (2000) 283:1295–1302.[Abstract/Free Full Text]
  27. Waagstein F, Bristow MR, Swedberg K, Camerini F, Fowler MB, Silver MA, et al. for the Metoprolol in Dilated Cardiomyopathy (MDC) Trial Study Group. Beneficial effect of metoprolol in idiopathic dilated cardiomyopathy. Lancet 1993;342:1441–6.
  28. Cohn JN, Goldstein SO, Greenberg BH, Lorell BH, Bourge RC, Jaski BE, et al. for the Vesnarinone Trial Investigators. A dose-dependent increase in mortality with vesnarinone among patients with severe heart failure. N Engl J Med 1998;339:1810–6.
  29. Australia/New Zealand Heart Failure Collaborative Group. Randomised, placebo-controlled trial of carvedilol in patients with congestive heart failure due to ischaemic heart disease. Lancet (1997) 349:375–380.[CrossRef][Web of Science][Medline]
  30. Lubsen J. Exercise testing as outcome in congestive heart failure trials. Design considerations when interpreting results. Drugs (1994) 47(Suppl. 4):25–30.[Web of Science][Medline]
  31. The Pimobendan in Congestive Heart Failure (PICO) Investigators. Effect of Pimobendan on exercise capacity in patients with heart failure: main results from the Pimobendan in congestive heart failure (PICO) trial. Heart (1996) 76:223–231.[Abstract/Free Full Text]
  32. Krum H., Sackner-Bernstein, Goldsmith R.L., Kukin M.L., Schwartz B., Penn J., et al. Double-blind, placebo-controlled study of the long-term efficacy of carvedilol in patients with severe heart failure. Circulation (1995) 92:1499–1506.[Abstract/Free Full Text]
  33. Opasich C., Pinna G.D., Mazza A., Febo O., Riccardi P.G., Capomolla S., et al. Reproducibility of the six-minute walking test in patients with chronic congestive heart failure: practical implications. Am J Cardiol (1998) 81:1497–1500.[CrossRef][Web of Science][Medline]
  34. Cazeau S., Leclercq C., Lavergne T., Walker S., Varma C., Linde C., et al. Effects of multisite biventricular pacing in patients with heart failure and intraventricular conduction delay. N Engl J Med (2001) 344:873–880.[Abstract/Free Full Text]
  35. Temple R.J. A regulatory authority's opinion about surrogate endpoints. In: Clinical measurements in drug evaluation—Nimmo W.S., Tucker G.T., eds. (1995) New York: J. Wiley.
  36. Prentice R.L. Surrogate endpoints in clinical trials: definition and operational criteria. Stat Med (1989) 8:431–440.[Web of Science][Medline]
  37. Sobel B.E., Furberg C.D. Surrogates, semantics and sensible public policy. Circulation (1997) 95:1661–1663.[Free Full Text]
  38. Schwartz D., Lellouch J. Explanatory and pragmatic attitudes in therapeutic trials. J Chron Dis (1967) 20:637–648.[CrossRef][Web of Science][Medline]
  39. Roland M., Torgerson D.J. Understanding controlled trials: what are pragmatic trials? Br Med J (1998) 316:285.[Free Full Text]
  40. Temple R. Are surrogate markers adequate to assess cardiovascular disease drugs? J Am Med Assoc (1999) 282:790–795.[Free Full Text]
  41. Wong M, Johnson G, Shabetai R, Hughes V, Bhat G, Lopez B, Cohn JN for the V-HeFT VA Cooperative Studies Group. Echocardiographic variables as prognostic indicators and therapeutic monitors in chronic congestive heart failure. Veterans Affairs Cooperative Studies V-HeFT I and II. Circulation 1993;87 Suppl. VI:65–70.
  42. Konstam MA, Rousseau MF, Kronenberg MW, Udelson JE, Melin J, Stewart D, et al. for the SOLVD Investigators. Effects of the angiotensin converting enzyme inhibitor enalapril on the long-term progression of left ventricular dysfunction in patients with heart failure. Circulation 1992;86:431–8.
  43. Cohn JN, Ziesche S, Smith R, Anand I, Dunkman WB, Loeb H, et al for the Vasodilator-Heart Failure Trial (V-HeFT) Study Group. Effect of calcium antagonist felodipine as supplementary vasodilator therapy in patients with chronic heart failure treated with enalapril. V-HeFT III. Circulation 1997;96:856–863.
  44. Cleland J.G. How to assess new treatments for the management of heart failure: composite scoring systems to assess the patients’ clinical journey. Eur J Heart Fail (2002) 4(3):243–247.[Free Full Text]

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 JCOHome page
R. K.S. Wong, N. Paul, K. Ding, M. Whitehead, M. Brundage, A. Fyles, D. Wilke, A. Nabid, A. Fortin, D. Wilson, et al.
5-Hydroxytryptamine-3 Receptor Antagonist With or Without Short-Course Dexamethasone in the Prophylaxis of Radiation Induced Emesis: A Placebo-Controlled Randomized Trial of the National Cancer Institute of Canada Clinical Trials Group (SC19)
J. Clin. Oncol., July 20, 2006; 24(21): 3458 - 3464.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Zanolla, L.
Right arrow Articles by Zardini, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Zanolla, L.
Right arrow Articles by Zardini, P.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?