- Email: [email protected]
Mismatch between heart failure patients in clinical trials and the real world
IJCA-15688; No of Pages 7 International Journal of Cardiology xxx (2013) xxx–xxx
Contents lists available at SciVerse ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Mismatch between heart failure patients in clinical trials and the real world ☆,☆☆ David Niederseer a, b, 1, Christoph W. Thaler c, 1, Michaela Niederseer a, b, 1, Josef Niebauer a, b, 1,⁎ a b c
University Institute of Sports Medicine, Prevention and Rehabilitation, Paracelsus Medical University, Salzburg, Austria Institute of Sports Medicine of the State of Salzburg, Sports Medicine of the Olympic Center Salzburg-Rif, Lindhofstraße 20, 5020 Salzburg, Austria Paracelsus Medical University Salzburg, Müllner Hauptstraße 48, 5020 Salzburg, Austria
a r t i c l e
i n f o
Article history: Received 12 October 2012 Received in revised form 9 December 2012 Accepted 25 December 2012 Available online xxxx Keywords: Heart failure Exercise training Rehabilitation Patient characteristics
a b s t r a c t Background: Evidence-based medicine urges physicians to translate results from clinical trials to their patients. This, however, can only work, if real world patients are represented in clinical trials. Methods: We searched the literature on chronic heart failure (1950-2/2011) for studies designed to detect effects on mortality (mortality studies, MS) and exercise training studies (ETS) as the leading non-pharmaceutical/ non-surgical treatment option in order to compare their characteristics with European (Euro Heart Survey on Heart Failure, EHSHF) and North American (Framingham Heart Study, FHS) epidemiological studies. Results: After an extensive literature search, we identified 207 ETS and 59 MS. Subjects enrolled in ETS were younger (ETS: 62.5±6.6; MS: 63.9±4.6; EHSHF: 71.0±3.5; FHS: 78.0 years), more often male (ETS: 80.9%; MS: 77.3%; EHSHF: 53.0%; FHS: 49.6%; pb 0.001), and had substantially less comorbidities such as diabetes mellitus (ETS: 13.6%; MS: 22.5%; EHSHF: 27.0%; FHS: 25.3%; pb 0.001), or hypertension (ETS: 26.3%; MS: 39.1%; EHSHF: 53.0%; FHS: 46.9%; p b 0.001). Angiotensin converting enzyme-inhibitors, beta-blockers, and angiotensin-receptorblockers were more commonly used in ETS than in EHSHF (all pb 0.001). Only 16 (10.6%) ETS and 20 (62.5%) MS reported ethnic background. Conclusion: Heart failure patients in exercise training studies and mortality studies do not represent real world patients. In order to extrapolate data to the general population future exercise training studies as well as mortality studies need to include representative patients. Otherwise, knowledge gained can only be translated to a minority of our patients. © 2013 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Evidence based medicine urges physicians to implement findings of clinical trials into everyday practice. As a consequence, patients in clinical trials need to be a representative sample of everyday patients. Van Spall et al. [1] analyzed eligibility criteria of randomized controlled trials published in high-impact general medical journals and found that women, children, elderly, and even those with common medical conditions are frequently excluded from clinical trials. It has
☆ Conflict of interest: There are no potential conflicts of interest, including related consultancies, shareholdings and funding grants. ☆☆ Part of the work was presented at EuroPRevent 2009, Stockholm on May 9th 2009, and won the Young Investigator Award of European Association of Cardiovascular Prevention and Rehabilitation of the European Society of Cardiology. ⁎ Corresponding author at: Department of Sports Medicine, Prevention and Rehabilitation, Paracelsus Medical University, Salzburg, Austria. Tel.: +43 662 4482 4270; fax: +43 662 4482 4274. E-mail address: [email protected] (J. Niebauer). 1 This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation.
been reported that multi-center trials and those involving drug interventions are most likely to have extensive and medically unjustified exclusion criteria, which may impair the generalizability of the results. Also, Masoudi et al. [2] applied the inclusion and exclusion criteria of three major chronic heart failure (CHF) trials to 20,388 representative patients in the US. Only 18%, 13%, and 25% met the enrollment criteria of the SOLVD, MERIT-HF, and RALES trials, respectively. A similar result was reported by Lenzen et al. [3] who applied the inclusion and exclusion criteria of the same three trials to the population of the Euro Heart Survey on Heart Failure (EHSHF). Of 10,701 patients, only 13% would have been eligible to participate in at least one of the three studies. Besides studies that aim to detect differences on so called “hard end points” such as mortality (mortality studies, MS), other investigations try to elucidate effects on surrogate parameters such as left ventricular ejection fraction, weight, and blood pressure. According to the recent guidelines of the ESC [4] and AHA/ACC [5,6], exercise training is the leading non-pharmacological/non surgical interventions in CHF. It was the aim of our study to compare patient characteristics of ETS and MS with epidemiological data of the EHSHF and Framingham Heart Study (FHS). We postulated that characteristics of real world
0167-5273/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijcard.2012.12.069
Please cite this article as: Niederseer D, et al, Mismatch between heart failure patients in clinical trials and the real world, Int J Cardiol (2013), http://dx.doi.org/10.1016/j.ijcard.2012.12.069
2
D. Niederseer et al. / International Journal of Cardiology xxx (2013) xxx–xxx
patients differed substantially from those enrolled in clinical trials questioning the generalizability of the study findings. 2. Methods 2.1. Literature search We searched MEDLINE, EMBASE, the Cochrane Collaboration Central Register of Controlled Clinical Trials (CENTRAL), and CINAHL (1950-2/2011) in order to locate publications on clinical heart failure trials implementing exercise training. Only adult human subjects (aged >19 years) suffering from chronic heart failure (CHF) were included. No language restrictions were applied; and clinical heart failure trials aiming at a reduction in mortality through various treatment options other than exercise training (see Table 1). For ETS we applied the search strategy used by Davies et al. [7] in their systematic Cochrane review on the same topic. Briefly, synonyms for heart failure and exercise training were combined using the operator OR; the results were then combined using the operator AND. Searches were limited to prospective clinical trials investigating the effects of any form of exercise training with the exception of electrical muscle stimulation and investigations on heart transplant recipients. We further systematically studied review articles and meta-analyses published between 2003 and 2011 and searched the references of all included exercise training studies in order to identify further publications relevant to our study. For MS we used medical subject headings (MeSH) for heart failure and mortality. Searches were limited to prospective randomized controlled CHF trials with mortality as their primary or secondary endpoint. Titles and abstracts of 5095 (for ETS) and 2057 (for MS) potentially relevant references were identified through our literature search and reviewed independently by 2 investigators (D.N., C.T.) to determine whether they met eligibility criteria for inclusion provided in Table 1. Discrepancies regarding whether or not to include a reference were resolved by consensus with another investigator (J.N.). A flow chart of the literature search is provided in Fig. 1. 2.2. Data extraction All data were extracted by one investigator (D.N.) and then checked and rechecked by another investigator (C.T.). The extracted data comprised name of first author, year of publication, journal, trial design, number of participants both in total and per group, age, gender distribution, ethnicity, socioeconomic variables, body mass index (BMI), New York Heart Association (NYHA) functional class, left ventricular ejection fraction (LVEF), primary cause of heart failure, diabetes mellitus, and hypertension. For ETS we further extracted dropouts both in total and per group, maximal oxygen consumption, and medication (beta-blockers, Angiotensin converting enzyme [ACE]-inhibitors, angiotensin receptor blocker). We did not extract medication in MS because most of the included studies investigated one of these pharmacological agents or had inclusion or exclusion criteria based on the medication of the subjects. The definition of heart failure differed among studies; however all included studies provided compelling reports on how they made sure to only include patients suffering from CHF. If a separate publication on the methods of a study had been published previously, we looked for further details in these manuscripts. No restrictions were made to whether systolic or diastolic heart failure was investigated nor did we limit the result to certain etiologies of heart failure. Only the mean values were extracted. Mostly means were provided separately for the control group and the intervention group. If the desired variable was not provided, we contacted corresponding authors of the studies, and if unsuccessful we left the space in our database blank. Epidemiological data were obtained from publications of the EHSHF [8–10] and the FHS [11,12] since they are universally accepted as the most comprehensive epidemiologic assessments of CHF in the respective continents. We extracted the same characteristics as in ETS and MS. In some ETS, baseline characteristics were only provided for those patients who completed the study; accordingly a per-protocol-analysis was employed in these studies. Thus, baseline variables of dropouts were not reported in these papers. The author(s) of this manuscript have certified that they comply with the Principles of Ethical Publishing in the International Journal of Cardiology [13].
Table 1 Inclusion criteria of exercise training studies and mortality studies of our literature search. Inclusion criteria for exercise training studies: 1) Prospective study design 2) Human adult subjects suffering from any form of chronic heart failure 3) Any form of exercise training as intervention Inclusion criteria for mortality studies: 1) Prospective randomized study 2) Human adult subjects suffering from any form of chronic heart failure 3) Any report on mortality as a primary or secondary end-point
2.3. Statistical analysis Extracted mean values of each variable were used to calculate weighted means for ETS and MS, in order to avoid misrepresentation of smaller and larger studies. All data are presented in mean ± standard deviation (SD) or percentages. We could only extract means of each included study (e.g. age) and were therefore unable to test our data on normal distribution. Because it is impossible to test properties of “studies” vs. “patients” as in EHSHF and FHS we did not test any statistical differences in continuous variables such as age, body mass index (BMI), NYHA functional class, and LVEF. In categorical variables (gender, diabetes mellitus, hypertension, and pharmacological treatment) we employed two-sided χ2-tests with Yate's continuity correction or Fisher's exact test to evaluate statistically significant differences between baseline characteristics of ETS, MS and ES. Statistical significance was defined as pb 0.05. Data were analyzed using Graph Pad InStat 3 for Windows (InStat Software, California, USA) and PASW Statistics 18, Release Version 18.0.0 (SPSS, Inc. 2009, Chicago, IL, USA).
3. Results Our abovementioned search strategy revealed 207 ETS and 59 MS studies (see Fig. 2). 3.1. Exercise training studies in heart failure Of the 207 studies analyzed, three (1.5%) studies were conducted before 1990, 53 (25.6%) between 1990 and 1999 and 152 (73.4%) between 2000 and 2011 (see Fig. 1). Publications were written in English (n=197/95.2%), Chinese (n= 3/1.5%), Polish (n= 3/1.5%), German (n=2/1.0%), French (n= 1/0.5%), or Lithuanian (n= 1/0.5%). Studies were designed as randomized controlled studies (n= 123/59.4%), controlled studies (n= 27/13.0%), non-controlled studies (n= 42/20.3%) and cross-over trials (n= 15/7.2%). 134 (64.7%) publications reported 829 drop-outs. Diabetics were excluded in 17 publications, two studies excluded uncontrolled and one decompensated diabetes. Patients suffering from hypertension were excluded in eight investigations, uncontrolled hypertension in 13 studies (see Table 2). 3.2. Mortality studies Of the 59 studies analyzed, two (3.4%) studies were conducted before 1990, 25 (42.4%) between 1990 and 1999 and 32 (54.2%) between 2000 and 2011 (see Fig. 1). All publications were written in English and designed as randomized controlled trials. Patient characteristics of subjects included in MS in heart failure are shown in Table 2. 3.3. Patient characteristics Comparison of patient characteristics of ETS, MS, and ES is shown in Table 2 and Fig. 3. Subjects enrolled in ETS were younger (ETS: 62.5 [SD 6.6]; MS: 63.9 [SD 4.6]; EHSHF: 71.0 [SD 3.5]; FHS: 78.0 years), more often male (ETS: 80.9%; MS: 77.3%; EHSHF: 53.0%; FHS: 49.6%; p b 0.001), had substantially less comorbidities such as diabetes mellitus (ETS: 13.6%; MS: 22.5%; EHSHF: 27.0%; FHS: 25.3%; p b 0.001), or hypertension (ETS: 26.3%; MS: 39.1%; EHSHF: 53.0%; FHS: 46.9%; p b 0.001), and were better treated with pharmacological agents known to reduce mortality such as Angiotensin converting enzyme-inhibitors/Angiotensin receptor blockers (p b 0.001), and beta-blockers (p b 0.001) as compared to MS and ES, respectively. Furthermore, ETS predominantly enrolled patients with a higher NYHA functional class, but a lower LVEF. Also BMI was higher in these subjects. In MS, patients differed significantly from epidemiological studies in terms of gender distribution, and comorbidities such as diabetes and hypertension (all p b 0.001). Also, age, etiology of heart failure, LVEF, NYHA functional class, and BMI differed considerably. Ethnicity was reported only in a minority of ETS, but in about half of MS, whereas socioeconomical variables were only available in 9 (4.3%) ETS, but not in MS.
Please cite this article as: Niederseer D, et al, Mismatch between heart failure patients in clinical trials and the real world, Int J Cardiol (2013), http://dx.doi.org/10.1016/j.ijcard.2012.12.069
D. Niederseer et al. / International Journal of Cardiology xxx (2013) xxx–xxx
Fig. 1. Flow chart of literature search.
Fig. 2. Publication history in the field of heart failure research: exercise training studies (ETS) and mortality studies (MS).
Please cite this article as: Niederseer D, et al, Mismatch between heart failure patients in clinical trials and the real world, Int J Cardiol (2013), http://dx.doi.org/10.1016/j.ijcard.2012.12.069
3
D. Niederseer et al. / International Journal of Cardiology xxx (2013) xxx–xxx
b0.001 n.a. b0.001 n.a.
n.a. b0.001 n.a. b0.001 b0.001 n.a. n.a. n.a. b0.001 b0.001
b0.001 n.a. b0.001 n.a.
b0.001 n.a. n.a n.a 0.011 n.a. n.a. n.a. b0.001 b0.001
b0.001 b0.001 n.a. n.a 0.094 n.a. n.a. n.a. 0.091 b0.001
MS vs EHSHF ETS vs FHS
b0.001 n.a. b0.001 n.a
MS vs FHS
4. Discussion To our knowledge we performed the first study to systematically assess whether characteristics of CHF patients of ES, ETS, and MS are comparable and how well they mirror those of everyday clinical patients. The main finding of our study is that patient characteristics of MS and ETS differ significantly from those patients included in ES with regard to age, gender distribution, ethnicity, etiology of heart failure, comorbidities, NYHA functional class, and LVEF. This questions the generalizability of the research findings and calls for an increased effort to enroll “real world” patients.
b0.001 n.a. b0.001 n.a.
n.a. b0.001 n.a. b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001
b0.001 b0.001 b0.001 0.212
b0.001 b0.001 n.a b0.001 b0.001 n.a. n.a. n.a. b0.001 b0.001 27.0 ± 5.0 No data No data No data 52.1 No data No data No data 25.3 46.9
(0.0)
(47.5)
No data 26.2 2.0 ± 1.0 No data 41.3 ± 12.0 68.0 36.9 61.8 4.5 27.0 53.0
No data
FHS (n = 655)
78.0 Yes 49.6 No
EHSHF (n = 11327)
71.0 ± 3.5 Yes 53.0 No
Mean± SD Mean± SD
27.6 ± 1.0 2.6 ± 0.4 No data 29.4 ± 6.6 55.4 n.a. n.a. n.a. 22.5 39.1 kg/m2 Class ml/kg/min % % % % % % %
Age Ethnicity reported Male Socioeconomic background reported BMI NYHA VO2peak LVEF Ischemic cause of HF Beta-blocker ACE-inhibitors ARB Diabetes Hypertension
Years n (%) of studies % n (%) of studies
29.2 ± 1.8 2.4 ± 0.3 15.9 ± 3.2 29.1 ± 5.4 46.9 54.9 74.8 2.8 13.6 26.3
(4.4)
(9.2)
63.9 ± 4.6 28 77.3 0
Mean± Mean±
SD
59
n = 134334 n = 10942
62.5 ± 6.6 19 80.9 9
2
n= 11982
Participants of interventional studies such as MS and ETS do not meet the criteria of being representatives of average patients suffering from CHF for several reasons [9]. The most common reason for declining study participation in a randomized clinical trial of behavioral therapy was “lives too far away” [18]. Also, because clinical studies are mostly conducted in university clinics, the urban population is overrepresented. Lloyd-Williams et al. [19] performed a postal survey with potential trial participants for a randomized controlled trial in heart failure patients (n= 667). The patient characteristics were known by the investigators, so that they could compare the responders with the nonresponders to the survey. Whereas no significant differences were found between responders and non-responders with respect to sociodemographic or clinical variables, it was male and younger patients as well as those on ACE-inhibitors who were significantly more likely consent to participate. Main reasons for non-participation were perceptions of being too old, too unwell, or too busy. Some of these patients were not even aware of the fact that they suffered from CHF. Corvera-Tindel et al. [20] used a study design that started with a moderate intensity of exercise training in patients with heart failure. After the first week the intensity and duration of the exercise training increased. Of 39 patients initially enrolled, 13 dropped out, whereas 26 finished the study. Dropouts significantly differed from patients who completed the study in terms of duration of CHF, comorbidities and peak VO2. 4.2. Possible reasons for misrepresentation in mortality studies
207
Total patients
Number of studies
Mortality studies (MS)
Epidemiological studies (ES)
SD
p-Value
ETS vs MS
ETS vs EHSHF
4.1. Reasons for misrepresentation in exercise training studies
Exercise training studies (ETS)
Table 2 Differences in patient characteristics between exercise training studies (ETS), mortality studies (MS) and epidemiological studies (ES) in heart failure. n.a. denotes not applicable. EHSHF denotes Euro Heart Survey on Heart Failure; FHS denotes Framingham Heart Study; HF denotes heart failure.
4
The problem of publication bias has been discussed elsewhere [21], but briefly two main aspects may contribute to our findings: Firstly, inclusion and exclusion criteria are often chosen to exclude patients with a higher potential to drop out [1]. Secondly, drugs may not be sufficiently investigated because of lack of interest of pharmaceutical companies, which sponsor about 70% of clinical trials in the US [22]. Furthermore it is also scientists who accelerate their career by publishing in high impact journals. As a consequence studies are often designed to rather lead to a positive outcome, than to enroll a representative group of patients. 4.3. Patient characteristics 4.3.1. Age, gender distribution, ethnicity, and BMI Heart failure patients still have a prognosis comparable with malignant cancer. As a matter of fact 50% of patients die within 4 years of diagnosis and 40% of patients admitted to hospital due to CHF are readmitted or dead within the following year [4]. Since the incidence of CHF increases with age, the age differences we detected are not only statistically significant, but also clinically relevant as we could show that patients are younger in ETS and MS than in ES. In some MS a special emphasis was laid on the age of the enrolled subjects. The SENIORS-study or ELITE II-study had a mean age of 76.1 and
Please cite this article as: Niederseer D, et al, Mismatch between heart failure patients in clinical trials and the real world, Int J Cardiol (2013), http://dx.doi.org/10.1016/j.ijcard.2012.12.069
D. Niederseer et al. / International Journal of Cardiology xxx (2013) xxx–xxx
5
Fig. 3. Differences in age (Panel A), gender distribution (Panel B), and comorbidities (Panel C) between exercise training studies, mortality studies, Euro Heart Survey on Heart Failure and Framingham Heart Study.
71.5 years, respectively. However, the age of these study populations was not particularly old, but rather a representative of the average age of heart failure patients [4,8–12]. Ethnicity and cultural background play an important role in pharmacogenetics. Therefore the A-HeFT study [23] was performed, showing that isosorbide dinitrate and hydralazine are also effective in African Americans. In the National Heart Care Project a retrospective
analysis showed different outcomes of heart failure patients with renal impairment in Afro-Americans and Caucasians [24], and a retrospective analysis of the SOLVD data revealed that Afro-Americans had a lower response to ACE-inhibitor therapy than Caucasians [25]. The underrepresentation of an ethnicity other than Caucasian in both ETS and MS is therefore a major methodological weakness if the results of the studies are to be extrapolated to other age groups, both genders, and
Please cite this article as: Niederseer D, et al, Mismatch between heart failure patients in clinical trials and the real world, Int J Cardiol (2013), http://dx.doi.org/10.1016/j.ijcard.2012.12.069
6
D. Niederseer et al. / International Journal of Cardiology xxx (2013) xxx–xxx
ethnicities other than Caucasians. HF-ACTION particularly addressed this issue and aimed to include a representative number of subjects from ethnicities other than Caucasians. However mean age in this largest of all ETS was 59.3 years, and only 28.4% of all randomized subjects were female. Whereas BMI directly correlates with all cause mortality in men and women, in recent studies the “reverse epidemiology” of heart failure [26] patients showed that a higher BMI is an independent predictor for longer survival in heart failure. Since in our analysis BMI differed between the ETS and MS as compared to ES, a different mortality rate would be likely in a patient population with a different BMI. 4.3.2. Socioeconomic variables In a study by Schaufelberger and Rosengren [27] performed in Sweden low socioeconomic status was found to be an independent risk factor for long-term risk of CHF in men. Even though these findings are well known, socioeconomic parameters were only addressed in 9 (4.4%) ETS. 4.3.3. Comorbidities We concentrated on diabetes mellitus and arterial hypertension, since comorbidities such as COPD, prior myocardial infarction, stroke, atrial fibrillation, orthopedic disabilities, renal disease [28], cancer or chronic infections were infrequently reported in MS and even less in ETS. In EHSHF [9] rates of these diagnoses were rather high: respiratory disease 32%; cancer 5%; atrial fibrillation 42%; renal dysfunction 17%; gout 5%; arthritis 10%; stroke 9%; and infections 24%. In many studies such comorbidities were a reason for exclusion. As a result, findings generated from ETS and MS which did not include these comorbidities cannot be applied for these subgroups of patients. Also, the NYHA class was significantly higher in ETS and MS as compared to ES. A recent review by Masoudi et al. [29] outlined and interpreted the findings of EHSHF with respect to the different etiologies of CHF in men and women, such as diastolic heart failure which is more often diagnosed in women than in men [29]. Because the majority of MS enrolled patients with low LVEF, diastolic heart failure was not assessed. As more women suffer from this kind of heart failure than men, this may be one of the reasons for the underrepresentation of women in heart failure trials. 4.3.4. Medication Because of the hypothesis that physical training has an add-on effect on the maximal pharmacological treatment, patients in training studies have to be better treated than real world patients. This was exceptionally true for the HF-ACTION trial, where 94.5%, 94.3%, 40.3%, and 18.0% of the enrolled subjects were treated with beta-blockers, ACE-inhibitors or ARBs, implantable-cardioverter-defibrillator, and biventricular pacemaker, respectively [30]. However, real world patients are not in line with current guidelines [31]. Thus, comparisons of percentages of therapy with drugs that reduce mortality have to be seen in this light. 4.3.5. The difficulty of defining “real world”, study limitations We compared the characteristics of patients in heart failure trials with those of ES, which do not necessarily represent the real world either. The FHS dataset was used to develop a risk score (the Framingham Risk Score) and has thus been considered to represent the real world patients to a large extent. However, the Framingham authors themselves cautioned about generalizing their data [14]. Nevertheless, in an epidemiological study conducted in the UK the Framingham Risk Score was found to appropriately predict the risk of cardiovascular events in this population [15]. More recent comparisons revealed reasonable agreement between Framingham predicted risk and observed risk in six US cohorts of Caucasians and Afro-Americans. This was not the case in subjects of Japanese, Hispanic, or Native American ethnic origin [16]. If the risk score is applied to different populations from Southern Europe [17],
or in studies with a more recent onset and follow-up period [17], the observed absolute risk is often substantially lower than that predicted by the Framingham algorithm. One reason for this might be the algorithm itself. But also an obvious reason could be the fact, that Framingham is not a representative US city anymore since migration has changed the demographics not only in the US but also worldwide. In addition to FHS we have chosen to also compare our data with that of the EHSHF. One of the major limitations of EHSHF was the voluntary nature of the survey. Cleland et al. [9] acknowledged that this almost certainly biased the study towards larger centers enrolling mostly patients living in urban settings. Nevertheless, they state that the EHSHF is the most representative European epidemiological study on CHF so far. Because of the lack of better alternatives we used data from the FHS and the EHSHF to depict a CHF population that mirrors real world patients as much as possible. In conclusion we compared patients' characteristics of subjects enrolled in clinical studies investigating the effects of exercise training in heart failure (ETS) as compared to clinical trials investigating the reduction in mortality through various treatments in heart failure (MS). We reported major differences in age, gender distribution, and comorbidities, and showed that patients enrolled in ETS and MS did match study populations of two epidemiological studies in Europe and Northern America. In order to extrapolate data to the general population future exercise training studies as well mortality studies need to include representative patients. Otherwise, the promising results gained can only be translated to a minority of our patients. Acknowledgments Part of this work was presented at EuroPRevent 2009, Stockholm, Sweden and won the Young Investigator Award of the European Association of Cardiovascular Prevention and Rehabilitation of the European Society of Cardiology. Appendix A. Supplementary data Supplementary data to this article can be found online at http:// dx.doi.org/10.1016/j.ijcard.2012.12.069. References [1] Van Spall HG, Toren A, Kiss A, Fowler RA. Eligibility criteria of randomized controlled trials published in high-impact general medical journals: a systematic sampling review. JAMA 2007;297(11):1233–40. [2] Masoudi FA, Havranek EP, Wolfe P, et al. Most hospitalized older persons do not meet the enrollment criteria for clinical trials in heart failure. Am Heart J 2003;146(2):250–7. [3] Lenzen MJ, Boersma E, Reimer WJ, et al. Under-utilization of evidence-based drug treatment in patients with heart failure is only partially explained by dissimilarity to patients enrolled in landmark trials: a report from the Euro Heart Survey on Heart Failure. Eur Heart J 2005;26(24):2706–13. [4] Dickstein K, Cohen-Solal A, Filippatos G, et al. ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure 2008: the Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2008 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association of the ESC (HFA) and endorsed by the European Society of Intensive Care Medicine (ESICM). Eur Heart J 2008;29(19):2388–442. [5] Hunt SA, Abraham WT, Chin MH, et al. Focused update incorporated into the ACC/AHA 2005 Guidelines for the Diagnosis and Management of Heart Failure in Adults: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines: developed in collaboration with the International Society for Heart and Lung Transplantation. Circulation 2009;119(14):e391–479. [6] Hunt SA, Abraham WT, Chin MH, et al. Focused update incorporated into the ACC/AHA 2005 Guidelines for the Diagnosis and Management of Heart Failure in Adults A Report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines Developed in Collaboration With the International Society for Heart and Lung Transplantation. J Am Coll Cardiol 2009;53(15):e1-90. [7] Davies EJ, Moxham T, Rees K, et al. Exercise based rehabilitation for heart failure. Cochrane Database Syst Rev 2010;4:CD003331. [8] Komajda M, Follath F, Swedberg K, et al. The EuroHeart Failure Survey programme—a survey on the quality of care among patients with heart failure in Europe. Part 2: treatment. Eur Heart J 2003;24(5):464–74.
Please cite this article as: Niederseer D, et al, Mismatch between heart failure patients in clinical trials and the real world, Int J Cardiol (2013), http://dx.doi.org/10.1016/j.ijcard.2012.12.069
D. Niederseer et al. / International Journal of Cardiology xxx (2013) xxx–xxx [9] Cleland JG, Swedberg K, Follath F, et al. The EuroHeart Failure survey programme—a survey on the quality of care among patients with heart failure in Europe. Part 1: patient characteristics and diagnosis. Eur Heart J 2003;24(5):442–63. [10] Komajda M, Hanon O, Hochadel M, et al. Management of octogenarians hospitalized for heart failure in Euro Heart Failure Survey I. Eur Heart J 2007;28(11): 1310–8. [11] Cheng S, Xanthakis V, Sullivan LM, et al. Correlates of echocardiographic indices of cardiac remodeling over the adult life course: longitudinal observations from the Framingham Heart Study. Circulation 2010;122(6):557–8. [12] Lee DS, Gona P, Vasan RS, et al. Relation of disease pathogenesis and risk factors to heart failure with preserved or reduced ejection fraction: insights from the Framingham heart study of the national heart, lung, and blood institute. Circulation 2009;119(24):3070–7. [13] Coats AJS, Shewan LG. Statement on authorship and publishing ethics in the International Journal of Cardiology. Int J Cardiol 2011;153:239–40. [14] Anderson KM, Wilson PW, Odell PM, Kannel WB. An updated coronary risk profile. A statement for health professionals. Circulation 1991;83(1):356–62. [15] Ramachandran S, French JM, Vanderpump MP, Croft P, Neary RH. Using the Framingham model to predict heart disease in the United Kingdom: retrospective study. BMJ 2000;320(7236):676–7. [16] D'Agostino Sr RB, Grundy S, Sullivan LM, Wilson P. Validation of the Framingham coronary heart disease prediction scores: results of a multiple ethnic groups investigation. JAMA 2001;286(2):180–7. [17] Menotti A, Puddu PE, Lanti M. Comparison of the Framingham risk function-based coronary chart with risk function from an Italian population study. Eur Heart J 2000;21(5):365–70. [18] Chang BH, Hendricks AM, Slawsky MT, Locastro JS. Patient recruitment to a randomized clinical trial of behavioral therapy for chronic heart failure. BMC Med Res Methodol 2004;4:8. [19] Lloyd-Williams F, Mair F, Shiels C, et al. Why are patients in clinical trials of heart failure not like those we see in everyday practice? J Clin Epidemiol 2003;56(12): 1157–62.
7
[20] Corvera-Tindel T, Doering LV, Gomez T, Dracup K. Predictors of noncompliance to exercise training in heart failure. J Cardiovasc Nurs 2004;19(4):269–79. [21] Easterbrook PJ, Berlin JA, Gopalan R, Matthews DR. Publication bias in clinical research. Lancet 1991;337(8746):867–72. [22] Mello MM, Clarridge BR, Studdert DM. Academic medical centers' standards for clinical-trial agreements with industry. N Engl J Med 2005;352(21):2202–10. [23] Taylor AL, Ziesche S, Yancy C, et al. Combination of isosorbide dinitrate and hydralazine in blacks with heart failure. N Engl J Med 2004;351(20):2049–57. [24] Smith GL, Shlipak MG, Havranek EP, et al. Race and renal impairment in heart failure: mortality in blacks versus whites. Circulation 2005;111(10):1270–7. [25] Exner DV, Dries DL, Domanski MJ, Cohn JN. Lesser response to angiotensinconverting-enzyme inhibitor therapy in black as compared with white patients with left ventricular dysfunction. N Engl J Med 2001;344(18):1351–7. [26] Curtis JP, Selter JG, Wang Y, et al. The obesity paradox: body mass index and outcomes in patients with heart failure. Arch Intern Med 2005;165(1):55–61. [27] Schaufelberger M, Rosengren A. Heart failure in different occupational classes in Sweden. Eur Heart J 2007;28(2):212–8. [28] Coca SG, Krumholz HM, Garg AX, Parikh CR. Underrepresentation of renal disease in randomized controlled trials of cardiovascular disease. JAMA 2006;296(11): 1377–84. [29] Masoudi FA, Havranek EP, Smith G, et al. Gender, age, and heart failure with preserved left ventricular systolic function. J Am Coll Cardiol 2003;41(2):217–23. [30] O'Connor CM, Whellan DJ, Lee KL, et al. Efficacy and safety of exercise training in patients with chronic heart failure: HF-ACTION randomized controlled trial. JAMA 2009;301(14):1439–50. [31] Drechsler K, Fikenzer S, Sechtem U, et al. The Euro Heart Survey—Germany: diabetes mellitus remains unrecognized in patients with coronary artery disease. Clin Res Cardiol 2008;97(6):364–70.
Please cite this article as: Niederseer D, et al, Mismatch between heart failure patients in clinical trials and the real world, Int J Cardiol (2013), http://dx.doi.org/10.1016/j.ijcard.2012.12.069
Contents lists available at SciVerse ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Mismatch between heart failure patients in clinical trials and the real world ☆,☆☆ David Niederseer a, b, 1, Christoph W. Thaler c, 1, Michaela Niederseer a, b, 1, Josef Niebauer a, b, 1,⁎ a b c
University Institute of Sports Medicine, Prevention and Rehabilitation, Paracelsus Medical University, Salzburg, Austria Institute of Sports Medicine of the State of Salzburg, Sports Medicine of the Olympic Center Salzburg-Rif, Lindhofstraße 20, 5020 Salzburg, Austria Paracelsus Medical University Salzburg, Müllner Hauptstraße 48, 5020 Salzburg, Austria
a r t i c l e
i n f o
Article history: Received 12 October 2012 Received in revised form 9 December 2012 Accepted 25 December 2012 Available online xxxx Keywords: Heart failure Exercise training Rehabilitation Patient characteristics
a b s t r a c t Background: Evidence-based medicine urges physicians to translate results from clinical trials to their patients. This, however, can only work, if real world patients are represented in clinical trials. Methods: We searched the literature on chronic heart failure (1950-2/2011) for studies designed to detect effects on mortality (mortality studies, MS) and exercise training studies (ETS) as the leading non-pharmaceutical/ non-surgical treatment option in order to compare their characteristics with European (Euro Heart Survey on Heart Failure, EHSHF) and North American (Framingham Heart Study, FHS) epidemiological studies. Results: After an extensive literature search, we identified 207 ETS and 59 MS. Subjects enrolled in ETS were younger (ETS: 62.5±6.6; MS: 63.9±4.6; EHSHF: 71.0±3.5; FHS: 78.0 years), more often male (ETS: 80.9%; MS: 77.3%; EHSHF: 53.0%; FHS: 49.6%; pb 0.001), and had substantially less comorbidities such as diabetes mellitus (ETS: 13.6%; MS: 22.5%; EHSHF: 27.0%; FHS: 25.3%; pb 0.001), or hypertension (ETS: 26.3%; MS: 39.1%; EHSHF: 53.0%; FHS: 46.9%; p b 0.001). Angiotensin converting enzyme-inhibitors, beta-blockers, and angiotensin-receptorblockers were more commonly used in ETS than in EHSHF (all pb 0.001). Only 16 (10.6%) ETS and 20 (62.5%) MS reported ethnic background. Conclusion: Heart failure patients in exercise training studies and mortality studies do not represent real world patients. In order to extrapolate data to the general population future exercise training studies as well as mortality studies need to include representative patients. Otherwise, knowledge gained can only be translated to a minority of our patients. © 2013 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Evidence based medicine urges physicians to implement findings of clinical trials into everyday practice. As a consequence, patients in clinical trials need to be a representative sample of everyday patients. Van Spall et al. [1] analyzed eligibility criteria of randomized controlled trials published in high-impact general medical journals and found that women, children, elderly, and even those with common medical conditions are frequently excluded from clinical trials. It has
☆ Conflict of interest: There are no potential conflicts of interest, including related consultancies, shareholdings and funding grants. ☆☆ Part of the work was presented at EuroPRevent 2009, Stockholm on May 9th 2009, and won the Young Investigator Award of European Association of Cardiovascular Prevention and Rehabilitation of the European Society of Cardiology. ⁎ Corresponding author at: Department of Sports Medicine, Prevention and Rehabilitation, Paracelsus Medical University, Salzburg, Austria. Tel.: +43 662 4482 4270; fax: +43 662 4482 4274. E-mail address: [email protected] (J. Niebauer). 1 This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation.
been reported that multi-center trials and those involving drug interventions are most likely to have extensive and medically unjustified exclusion criteria, which may impair the generalizability of the results. Also, Masoudi et al. [2] applied the inclusion and exclusion criteria of three major chronic heart failure (CHF) trials to 20,388 representative patients in the US. Only 18%, 13%, and 25% met the enrollment criteria of the SOLVD, MERIT-HF, and RALES trials, respectively. A similar result was reported by Lenzen et al. [3] who applied the inclusion and exclusion criteria of the same three trials to the population of the Euro Heart Survey on Heart Failure (EHSHF). Of 10,701 patients, only 13% would have been eligible to participate in at least one of the three studies. Besides studies that aim to detect differences on so called “hard end points” such as mortality (mortality studies, MS), other investigations try to elucidate effects on surrogate parameters such as left ventricular ejection fraction, weight, and blood pressure. According to the recent guidelines of the ESC [4] and AHA/ACC [5,6], exercise training is the leading non-pharmacological/non surgical interventions in CHF. It was the aim of our study to compare patient characteristics of ETS and MS with epidemiological data of the EHSHF and Framingham Heart Study (FHS). We postulated that characteristics of real world
0167-5273/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijcard.2012.12.069
Please cite this article as: Niederseer D, et al, Mismatch between heart failure patients in clinical trials and the real world, Int J Cardiol (2013), http://dx.doi.org/10.1016/j.ijcard.2012.12.069
2
D. Niederseer et al. / International Journal of Cardiology xxx (2013) xxx–xxx
patients differed substantially from those enrolled in clinical trials questioning the generalizability of the study findings. 2. Methods 2.1. Literature search We searched MEDLINE, EMBASE, the Cochrane Collaboration Central Register of Controlled Clinical Trials (CENTRAL), and CINAHL (1950-2/2011) in order to locate publications on clinical heart failure trials implementing exercise training. Only adult human subjects (aged >19 years) suffering from chronic heart failure (CHF) were included. No language restrictions were applied; and clinical heart failure trials aiming at a reduction in mortality through various treatment options other than exercise training (see Table 1). For ETS we applied the search strategy used by Davies et al. [7] in their systematic Cochrane review on the same topic. Briefly, synonyms for heart failure and exercise training were combined using the operator OR; the results were then combined using the operator AND. Searches were limited to prospective clinical trials investigating the effects of any form of exercise training with the exception of electrical muscle stimulation and investigations on heart transplant recipients. We further systematically studied review articles and meta-analyses published between 2003 and 2011 and searched the references of all included exercise training studies in order to identify further publications relevant to our study. For MS we used medical subject headings (MeSH) for heart failure and mortality. Searches were limited to prospective randomized controlled CHF trials with mortality as their primary or secondary endpoint. Titles and abstracts of 5095 (for ETS) and 2057 (for MS) potentially relevant references were identified through our literature search and reviewed independently by 2 investigators (D.N., C.T.) to determine whether they met eligibility criteria for inclusion provided in Table 1. Discrepancies regarding whether or not to include a reference were resolved by consensus with another investigator (J.N.). A flow chart of the literature search is provided in Fig. 1. 2.2. Data extraction All data were extracted by one investigator (D.N.) and then checked and rechecked by another investigator (C.T.). The extracted data comprised name of first author, year of publication, journal, trial design, number of participants both in total and per group, age, gender distribution, ethnicity, socioeconomic variables, body mass index (BMI), New York Heart Association (NYHA) functional class, left ventricular ejection fraction (LVEF), primary cause of heart failure, diabetes mellitus, and hypertension. For ETS we further extracted dropouts both in total and per group, maximal oxygen consumption, and medication (beta-blockers, Angiotensin converting enzyme [ACE]-inhibitors, angiotensin receptor blocker). We did not extract medication in MS because most of the included studies investigated one of these pharmacological agents or had inclusion or exclusion criteria based on the medication of the subjects. The definition of heart failure differed among studies; however all included studies provided compelling reports on how they made sure to only include patients suffering from CHF. If a separate publication on the methods of a study had been published previously, we looked for further details in these manuscripts. No restrictions were made to whether systolic or diastolic heart failure was investigated nor did we limit the result to certain etiologies of heart failure. Only the mean values were extracted. Mostly means were provided separately for the control group and the intervention group. If the desired variable was not provided, we contacted corresponding authors of the studies, and if unsuccessful we left the space in our database blank. Epidemiological data were obtained from publications of the EHSHF [8–10] and the FHS [11,12] since they are universally accepted as the most comprehensive epidemiologic assessments of CHF in the respective continents. We extracted the same characteristics as in ETS and MS. In some ETS, baseline characteristics were only provided for those patients who completed the study; accordingly a per-protocol-analysis was employed in these studies. Thus, baseline variables of dropouts were not reported in these papers. The author(s) of this manuscript have certified that they comply with the Principles of Ethical Publishing in the International Journal of Cardiology [13].
Table 1 Inclusion criteria of exercise training studies and mortality studies of our literature search. Inclusion criteria for exercise training studies: 1) Prospective study design 2) Human adult subjects suffering from any form of chronic heart failure 3) Any form of exercise training as intervention Inclusion criteria for mortality studies: 1) Prospective randomized study 2) Human adult subjects suffering from any form of chronic heart failure 3) Any report on mortality as a primary or secondary end-point
2.3. Statistical analysis Extracted mean values of each variable were used to calculate weighted means for ETS and MS, in order to avoid misrepresentation of smaller and larger studies. All data are presented in mean ± standard deviation (SD) or percentages. We could only extract means of each included study (e.g. age) and were therefore unable to test our data on normal distribution. Because it is impossible to test properties of “studies” vs. “patients” as in EHSHF and FHS we did not test any statistical differences in continuous variables such as age, body mass index (BMI), NYHA functional class, and LVEF. In categorical variables (gender, diabetes mellitus, hypertension, and pharmacological treatment) we employed two-sided χ2-tests with Yate's continuity correction or Fisher's exact test to evaluate statistically significant differences between baseline characteristics of ETS, MS and ES. Statistical significance was defined as pb 0.05. Data were analyzed using Graph Pad InStat 3 for Windows (InStat Software, California, USA) and PASW Statistics 18, Release Version 18.0.0 (SPSS, Inc. 2009, Chicago, IL, USA).
3. Results Our abovementioned search strategy revealed 207 ETS and 59 MS studies (see Fig. 2). 3.1. Exercise training studies in heart failure Of the 207 studies analyzed, three (1.5%) studies were conducted before 1990, 53 (25.6%) between 1990 and 1999 and 152 (73.4%) between 2000 and 2011 (see Fig. 1). Publications were written in English (n=197/95.2%), Chinese (n= 3/1.5%), Polish (n= 3/1.5%), German (n=2/1.0%), French (n= 1/0.5%), or Lithuanian (n= 1/0.5%). Studies were designed as randomized controlled studies (n= 123/59.4%), controlled studies (n= 27/13.0%), non-controlled studies (n= 42/20.3%) and cross-over trials (n= 15/7.2%). 134 (64.7%) publications reported 829 drop-outs. Diabetics were excluded in 17 publications, two studies excluded uncontrolled and one decompensated diabetes. Patients suffering from hypertension were excluded in eight investigations, uncontrolled hypertension in 13 studies (see Table 2). 3.2. Mortality studies Of the 59 studies analyzed, two (3.4%) studies were conducted before 1990, 25 (42.4%) between 1990 and 1999 and 32 (54.2%) between 2000 and 2011 (see Fig. 1). All publications were written in English and designed as randomized controlled trials. Patient characteristics of subjects included in MS in heart failure are shown in Table 2. 3.3. Patient characteristics Comparison of patient characteristics of ETS, MS, and ES is shown in Table 2 and Fig. 3. Subjects enrolled in ETS were younger (ETS: 62.5 [SD 6.6]; MS: 63.9 [SD 4.6]; EHSHF: 71.0 [SD 3.5]; FHS: 78.0 years), more often male (ETS: 80.9%; MS: 77.3%; EHSHF: 53.0%; FHS: 49.6%; p b 0.001), had substantially less comorbidities such as diabetes mellitus (ETS: 13.6%; MS: 22.5%; EHSHF: 27.0%; FHS: 25.3%; p b 0.001), or hypertension (ETS: 26.3%; MS: 39.1%; EHSHF: 53.0%; FHS: 46.9%; p b 0.001), and were better treated with pharmacological agents known to reduce mortality such as Angiotensin converting enzyme-inhibitors/Angiotensin receptor blockers (p b 0.001), and beta-blockers (p b 0.001) as compared to MS and ES, respectively. Furthermore, ETS predominantly enrolled patients with a higher NYHA functional class, but a lower LVEF. Also BMI was higher in these subjects. In MS, patients differed significantly from epidemiological studies in terms of gender distribution, and comorbidities such as diabetes and hypertension (all p b 0.001). Also, age, etiology of heart failure, LVEF, NYHA functional class, and BMI differed considerably. Ethnicity was reported only in a minority of ETS, but in about half of MS, whereas socioeconomical variables were only available in 9 (4.3%) ETS, but not in MS.
Please cite this article as: Niederseer D, et al, Mismatch between heart failure patients in clinical trials and the real world, Int J Cardiol (2013), http://dx.doi.org/10.1016/j.ijcard.2012.12.069
D. Niederseer et al. / International Journal of Cardiology xxx (2013) xxx–xxx
Fig. 1. Flow chart of literature search.
Fig. 2. Publication history in the field of heart failure research: exercise training studies (ETS) and mortality studies (MS).
Please cite this article as: Niederseer D, et al, Mismatch between heart failure patients in clinical trials and the real world, Int J Cardiol (2013), http://dx.doi.org/10.1016/j.ijcard.2012.12.069
3
D. Niederseer et al. / International Journal of Cardiology xxx (2013) xxx–xxx
b0.001 n.a. b0.001 n.a.
n.a. b0.001 n.a. b0.001 b0.001 n.a. n.a. n.a. b0.001 b0.001
b0.001 n.a. b0.001 n.a.
b0.001 n.a. n.a n.a 0.011 n.a. n.a. n.a. b0.001 b0.001
b0.001 b0.001 n.a. n.a 0.094 n.a. n.a. n.a. 0.091 b0.001
MS vs EHSHF ETS vs FHS
b0.001 n.a. b0.001 n.a
MS vs FHS
4. Discussion To our knowledge we performed the first study to systematically assess whether characteristics of CHF patients of ES, ETS, and MS are comparable and how well they mirror those of everyday clinical patients. The main finding of our study is that patient characteristics of MS and ETS differ significantly from those patients included in ES with regard to age, gender distribution, ethnicity, etiology of heart failure, comorbidities, NYHA functional class, and LVEF. This questions the generalizability of the research findings and calls for an increased effort to enroll “real world” patients.
b0.001 n.a. b0.001 n.a.
n.a. b0.001 n.a. b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001
b0.001 b0.001 b0.001 0.212
b0.001 b0.001 n.a b0.001 b0.001 n.a. n.a. n.a. b0.001 b0.001 27.0 ± 5.0 No data No data No data 52.1 No data No data No data 25.3 46.9
(0.0)
(47.5)
No data 26.2 2.0 ± 1.0 No data 41.3 ± 12.0 68.0 36.9 61.8 4.5 27.0 53.0
No data
FHS (n = 655)
78.0 Yes 49.6 No
EHSHF (n = 11327)
71.0 ± 3.5 Yes 53.0 No
Mean± SD Mean± SD
27.6 ± 1.0 2.6 ± 0.4 No data 29.4 ± 6.6 55.4 n.a. n.a. n.a. 22.5 39.1 kg/m2 Class ml/kg/min % % % % % % %
Age Ethnicity reported Male Socioeconomic background reported BMI NYHA VO2peak LVEF Ischemic cause of HF Beta-blocker ACE-inhibitors ARB Diabetes Hypertension
Years n (%) of studies % n (%) of studies
29.2 ± 1.8 2.4 ± 0.3 15.9 ± 3.2 29.1 ± 5.4 46.9 54.9 74.8 2.8 13.6 26.3
(4.4)
(9.2)
63.9 ± 4.6 28 77.3 0
Mean± Mean±
SD
59
n = 134334 n = 10942
62.5 ± 6.6 19 80.9 9
2
n= 11982
Participants of interventional studies such as MS and ETS do not meet the criteria of being representatives of average patients suffering from CHF for several reasons [9]. The most common reason for declining study participation in a randomized clinical trial of behavioral therapy was “lives too far away” [18]. Also, because clinical studies are mostly conducted in university clinics, the urban population is overrepresented. Lloyd-Williams et al. [19] performed a postal survey with potential trial participants for a randomized controlled trial in heart failure patients (n= 667). The patient characteristics were known by the investigators, so that they could compare the responders with the nonresponders to the survey. Whereas no significant differences were found between responders and non-responders with respect to sociodemographic or clinical variables, it was male and younger patients as well as those on ACE-inhibitors who were significantly more likely consent to participate. Main reasons for non-participation were perceptions of being too old, too unwell, or too busy. Some of these patients were not even aware of the fact that they suffered from CHF. Corvera-Tindel et al. [20] used a study design that started with a moderate intensity of exercise training in patients with heart failure. After the first week the intensity and duration of the exercise training increased. Of 39 patients initially enrolled, 13 dropped out, whereas 26 finished the study. Dropouts significantly differed from patients who completed the study in terms of duration of CHF, comorbidities and peak VO2. 4.2. Possible reasons for misrepresentation in mortality studies
207
Total patients
Number of studies
Mortality studies (MS)
Epidemiological studies (ES)
SD
p-Value
ETS vs MS
ETS vs EHSHF
4.1. Reasons for misrepresentation in exercise training studies
Exercise training studies (ETS)
Table 2 Differences in patient characteristics between exercise training studies (ETS), mortality studies (MS) and epidemiological studies (ES) in heart failure. n.a. denotes not applicable. EHSHF denotes Euro Heart Survey on Heart Failure; FHS denotes Framingham Heart Study; HF denotes heart failure.
4
The problem of publication bias has been discussed elsewhere [21], but briefly two main aspects may contribute to our findings: Firstly, inclusion and exclusion criteria are often chosen to exclude patients with a higher potential to drop out [1]. Secondly, drugs may not be sufficiently investigated because of lack of interest of pharmaceutical companies, which sponsor about 70% of clinical trials in the US [22]. Furthermore it is also scientists who accelerate their career by publishing in high impact journals. As a consequence studies are often designed to rather lead to a positive outcome, than to enroll a representative group of patients. 4.3. Patient characteristics 4.3.1. Age, gender distribution, ethnicity, and BMI Heart failure patients still have a prognosis comparable with malignant cancer. As a matter of fact 50% of patients die within 4 years of diagnosis and 40% of patients admitted to hospital due to CHF are readmitted or dead within the following year [4]. Since the incidence of CHF increases with age, the age differences we detected are not only statistically significant, but also clinically relevant as we could show that patients are younger in ETS and MS than in ES. In some MS a special emphasis was laid on the age of the enrolled subjects. The SENIORS-study or ELITE II-study had a mean age of 76.1 and
Please cite this article as: Niederseer D, et al, Mismatch between heart failure patients in clinical trials and the real world, Int J Cardiol (2013), http://dx.doi.org/10.1016/j.ijcard.2012.12.069
D. Niederseer et al. / International Journal of Cardiology xxx (2013) xxx–xxx
5
Fig. 3. Differences in age (Panel A), gender distribution (Panel B), and comorbidities (Panel C) between exercise training studies, mortality studies, Euro Heart Survey on Heart Failure and Framingham Heart Study.
71.5 years, respectively. However, the age of these study populations was not particularly old, but rather a representative of the average age of heart failure patients [4,8–12]. Ethnicity and cultural background play an important role in pharmacogenetics. Therefore the A-HeFT study [23] was performed, showing that isosorbide dinitrate and hydralazine are also effective in African Americans. In the National Heart Care Project a retrospective
analysis showed different outcomes of heart failure patients with renal impairment in Afro-Americans and Caucasians [24], and a retrospective analysis of the SOLVD data revealed that Afro-Americans had a lower response to ACE-inhibitor therapy than Caucasians [25]. The underrepresentation of an ethnicity other than Caucasian in both ETS and MS is therefore a major methodological weakness if the results of the studies are to be extrapolated to other age groups, both genders, and
Please cite this article as: Niederseer D, et al, Mismatch between heart failure patients in clinical trials and the real world, Int J Cardiol (2013), http://dx.doi.org/10.1016/j.ijcard.2012.12.069
6
D. Niederseer et al. / International Journal of Cardiology xxx (2013) xxx–xxx
ethnicities other than Caucasians. HF-ACTION particularly addressed this issue and aimed to include a representative number of subjects from ethnicities other than Caucasians. However mean age in this largest of all ETS was 59.3 years, and only 28.4% of all randomized subjects were female. Whereas BMI directly correlates with all cause mortality in men and women, in recent studies the “reverse epidemiology” of heart failure [26] patients showed that a higher BMI is an independent predictor for longer survival in heart failure. Since in our analysis BMI differed between the ETS and MS as compared to ES, a different mortality rate would be likely in a patient population with a different BMI. 4.3.2. Socioeconomic variables In a study by Schaufelberger and Rosengren [27] performed in Sweden low socioeconomic status was found to be an independent risk factor for long-term risk of CHF in men. Even though these findings are well known, socioeconomic parameters were only addressed in 9 (4.4%) ETS. 4.3.3. Comorbidities We concentrated on diabetes mellitus and arterial hypertension, since comorbidities such as COPD, prior myocardial infarction, stroke, atrial fibrillation, orthopedic disabilities, renal disease [28], cancer or chronic infections were infrequently reported in MS and even less in ETS. In EHSHF [9] rates of these diagnoses were rather high: respiratory disease 32%; cancer 5%; atrial fibrillation 42%; renal dysfunction 17%; gout 5%; arthritis 10%; stroke 9%; and infections 24%. In many studies such comorbidities were a reason for exclusion. As a result, findings generated from ETS and MS which did not include these comorbidities cannot be applied for these subgroups of patients. Also, the NYHA class was significantly higher in ETS and MS as compared to ES. A recent review by Masoudi et al. [29] outlined and interpreted the findings of EHSHF with respect to the different etiologies of CHF in men and women, such as diastolic heart failure which is more often diagnosed in women than in men [29]. Because the majority of MS enrolled patients with low LVEF, diastolic heart failure was not assessed. As more women suffer from this kind of heart failure than men, this may be one of the reasons for the underrepresentation of women in heart failure trials. 4.3.4. Medication Because of the hypothesis that physical training has an add-on effect on the maximal pharmacological treatment, patients in training studies have to be better treated than real world patients. This was exceptionally true for the HF-ACTION trial, where 94.5%, 94.3%, 40.3%, and 18.0% of the enrolled subjects were treated with beta-blockers, ACE-inhibitors or ARBs, implantable-cardioverter-defibrillator, and biventricular pacemaker, respectively [30]. However, real world patients are not in line with current guidelines [31]. Thus, comparisons of percentages of therapy with drugs that reduce mortality have to be seen in this light. 4.3.5. The difficulty of defining “real world”, study limitations We compared the characteristics of patients in heart failure trials with those of ES, which do not necessarily represent the real world either. The FHS dataset was used to develop a risk score (the Framingham Risk Score) and has thus been considered to represent the real world patients to a large extent. However, the Framingham authors themselves cautioned about generalizing their data [14]. Nevertheless, in an epidemiological study conducted in the UK the Framingham Risk Score was found to appropriately predict the risk of cardiovascular events in this population [15]. More recent comparisons revealed reasonable agreement between Framingham predicted risk and observed risk in six US cohorts of Caucasians and Afro-Americans. This was not the case in subjects of Japanese, Hispanic, or Native American ethnic origin [16]. If the risk score is applied to different populations from Southern Europe [17],
or in studies with a more recent onset and follow-up period [17], the observed absolute risk is often substantially lower than that predicted by the Framingham algorithm. One reason for this might be the algorithm itself. But also an obvious reason could be the fact, that Framingham is not a representative US city anymore since migration has changed the demographics not only in the US but also worldwide. In addition to FHS we have chosen to also compare our data with that of the EHSHF. One of the major limitations of EHSHF was the voluntary nature of the survey. Cleland et al. [9] acknowledged that this almost certainly biased the study towards larger centers enrolling mostly patients living in urban settings. Nevertheless, they state that the EHSHF is the most representative European epidemiological study on CHF so far. Because of the lack of better alternatives we used data from the FHS and the EHSHF to depict a CHF population that mirrors real world patients as much as possible. In conclusion we compared patients' characteristics of subjects enrolled in clinical studies investigating the effects of exercise training in heart failure (ETS) as compared to clinical trials investigating the reduction in mortality through various treatments in heart failure (MS). We reported major differences in age, gender distribution, and comorbidities, and showed that patients enrolled in ETS and MS did match study populations of two epidemiological studies in Europe and Northern America. In order to extrapolate data to the general population future exercise training studies as well mortality studies need to include representative patients. Otherwise, the promising results gained can only be translated to a minority of our patients. Acknowledgments Part of this work was presented at EuroPRevent 2009, Stockholm, Sweden and won the Young Investigator Award of the European Association of Cardiovascular Prevention and Rehabilitation of the European Society of Cardiology. Appendix A. Supplementary data Supplementary data to this article can be found online at http:// dx.doi.org/10.1016/j.ijcard.2012.12.069. References [1] Van Spall HG, Toren A, Kiss A, Fowler RA. Eligibility criteria of randomized controlled trials published in high-impact general medical journals: a systematic sampling review. JAMA 2007;297(11):1233–40. [2] Masoudi FA, Havranek EP, Wolfe P, et al. Most hospitalized older persons do not meet the enrollment criteria for clinical trials in heart failure. Am Heart J 2003;146(2):250–7. [3] Lenzen MJ, Boersma E, Reimer WJ, et al. Under-utilization of evidence-based drug treatment in patients with heart failure is only partially explained by dissimilarity to patients enrolled in landmark trials: a report from the Euro Heart Survey on Heart Failure. Eur Heart J 2005;26(24):2706–13. [4] Dickstein K, Cohen-Solal A, Filippatos G, et al. ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure 2008: the Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2008 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association of the ESC (HFA) and endorsed by the European Society of Intensive Care Medicine (ESICM). Eur Heart J 2008;29(19):2388–442. [5] Hunt SA, Abraham WT, Chin MH, et al. Focused update incorporated into the ACC/AHA 2005 Guidelines for the Diagnosis and Management of Heart Failure in Adults: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines: developed in collaboration with the International Society for Heart and Lung Transplantation. Circulation 2009;119(14):e391–479. [6] Hunt SA, Abraham WT, Chin MH, et al. Focused update incorporated into the ACC/AHA 2005 Guidelines for the Diagnosis and Management of Heart Failure in Adults A Report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines Developed in Collaboration With the International Society for Heart and Lung Transplantation. J Am Coll Cardiol 2009;53(15):e1-90. [7] Davies EJ, Moxham T, Rees K, et al. Exercise based rehabilitation for heart failure. Cochrane Database Syst Rev 2010;4:CD003331. [8] Komajda M, Follath F, Swedberg K, et al. The EuroHeart Failure Survey programme—a survey on the quality of care among patients with heart failure in Europe. Part 2: treatment. Eur Heart J 2003;24(5):464–74.
Please cite this article as: Niederseer D, et al, Mismatch between heart failure patients in clinical trials and the real world, Int J Cardiol (2013), http://dx.doi.org/10.1016/j.ijcard.2012.12.069
D. Niederseer et al. / International Journal of Cardiology xxx (2013) xxx–xxx [9] Cleland JG, Swedberg K, Follath F, et al. The EuroHeart Failure survey programme—a survey on the quality of care among patients with heart failure in Europe. Part 1: patient characteristics and diagnosis. Eur Heart J 2003;24(5):442–63. [10] Komajda M, Hanon O, Hochadel M, et al. Management of octogenarians hospitalized for heart failure in Euro Heart Failure Survey I. Eur Heart J 2007;28(11): 1310–8. [11] Cheng S, Xanthakis V, Sullivan LM, et al. Correlates of echocardiographic indices of cardiac remodeling over the adult life course: longitudinal observations from the Framingham Heart Study. Circulation 2010;122(6):557–8. [12] Lee DS, Gona P, Vasan RS, et al. Relation of disease pathogenesis and risk factors to heart failure with preserved or reduced ejection fraction: insights from the Framingham heart study of the national heart, lung, and blood institute. Circulation 2009;119(24):3070–7. [13] Coats AJS, Shewan LG. Statement on authorship and publishing ethics in the International Journal of Cardiology. Int J Cardiol 2011;153:239–40. [14] Anderson KM, Wilson PW, Odell PM, Kannel WB. An updated coronary risk profile. A statement for health professionals. Circulation 1991;83(1):356–62. [15] Ramachandran S, French JM, Vanderpump MP, Croft P, Neary RH. Using the Framingham model to predict heart disease in the United Kingdom: retrospective study. BMJ 2000;320(7236):676–7. [16] D'Agostino Sr RB, Grundy S, Sullivan LM, Wilson P. Validation of the Framingham coronary heart disease prediction scores: results of a multiple ethnic groups investigation. JAMA 2001;286(2):180–7. [17] Menotti A, Puddu PE, Lanti M. Comparison of the Framingham risk function-based coronary chart with risk function from an Italian population study. Eur Heart J 2000;21(5):365–70. [18] Chang BH, Hendricks AM, Slawsky MT, Locastro JS. Patient recruitment to a randomized clinical trial of behavioral therapy for chronic heart failure. BMC Med Res Methodol 2004;4:8. [19] Lloyd-Williams F, Mair F, Shiels C, et al. Why are patients in clinical trials of heart failure not like those we see in everyday practice? J Clin Epidemiol 2003;56(12): 1157–62.
7
[20] Corvera-Tindel T, Doering LV, Gomez T, Dracup K. Predictors of noncompliance to exercise training in heart failure. J Cardiovasc Nurs 2004;19(4):269–79. [21] Easterbrook PJ, Berlin JA, Gopalan R, Matthews DR. Publication bias in clinical research. Lancet 1991;337(8746):867–72. [22] Mello MM, Clarridge BR, Studdert DM. Academic medical centers' standards for clinical-trial agreements with industry. N Engl J Med 2005;352(21):2202–10. [23] Taylor AL, Ziesche S, Yancy C, et al. Combination of isosorbide dinitrate and hydralazine in blacks with heart failure. N Engl J Med 2004;351(20):2049–57. [24] Smith GL, Shlipak MG, Havranek EP, et al. Race and renal impairment in heart failure: mortality in blacks versus whites. Circulation 2005;111(10):1270–7. [25] Exner DV, Dries DL, Domanski MJ, Cohn JN. Lesser response to angiotensinconverting-enzyme inhibitor therapy in black as compared with white patients with left ventricular dysfunction. N Engl J Med 2001;344(18):1351–7. [26] Curtis JP, Selter JG, Wang Y, et al. The obesity paradox: body mass index and outcomes in patients with heart failure. Arch Intern Med 2005;165(1):55–61. [27] Schaufelberger M, Rosengren A. Heart failure in different occupational classes in Sweden. Eur Heart J 2007;28(2):212–8. [28] Coca SG, Krumholz HM, Garg AX, Parikh CR. Underrepresentation of renal disease in randomized controlled trials of cardiovascular disease. JAMA 2006;296(11): 1377–84. [29] Masoudi FA, Havranek EP, Smith G, et al. Gender, age, and heart failure with preserved left ventricular systolic function. J Am Coll Cardiol 2003;41(2):217–23. [30] O'Connor CM, Whellan DJ, Lee KL, et al. Efficacy and safety of exercise training in patients with chronic heart failure: HF-ACTION randomized controlled trial. JAMA 2009;301(14):1439–50. [31] Drechsler K, Fikenzer S, Sechtem U, et al. The Euro Heart Survey—Germany: diabetes mellitus remains unrecognized in patients with coronary artery disease. Clin Res Cardiol 2008;97(6):364–70.
Please cite this article as: Niederseer D, et al, Mismatch between heart failure patients in clinical trials and the real world, Int J Cardiol (2013), http://dx.doi.org/10.1016/j.ijcard.2012.12.069