Abstract
Coronary artery disease (CAD), still the main cause of mortality in the world, experienced in the 1990s major advances in treatment. Early coronary angioplasty and systematic treatment with statins and new anti-aggregators produced dramatic improvement in prognosis. Cardiac rehabilitation benefits in coronary patients demonstrated before this ‘new era’ should not be directly extrapolated to the present. The small dimension cardiac rehabilitation trials in the last century comparing cardiac rehabilitation with conservative therapy, targeting mostly male and young patients, without comorbidities, evolved more recently to larger observational studies, looking also at ethnic minorities, women, old and heart failure patients, treated with the present invasive and pharmacologic approach, becoming closer to the heterogeneous real-world scenario. To demonstrate that cardiac rehabilitation may still provide additional value to the more advanced and recent therapies, which already proved to benefit survival in coronary patients, it was mandatory to evaluate specifically these more recent studies. Without this evaluation, the doubt regarding the essential role of cardiac rehabilitation in CAD patients would still persist nowadays.
The CROS (Cardiac Rehabilitation Outcome Study) 1 is the first structured review and meta-analysis, just performed, to clarify this issue. The purpose was to evaluate the effectiveness of cardiac rehabilitation on total mortality and other clinical endpoints after an acute coronary event. Methodologically well-designed, it identified cardiac rehabilitation clinical studies performed during and after 1995, necessarily including structured and supervised exercise. Regarding the design, only one randomized controlled trial fulfilled the CROS criteria, with 1813 patients. In addition, 24 cohort clinical studies, 17 retrospective, with 206,096 patients, and seven prospective, with 12,193 patients, were selected by strict criteria, with a mean follow-up of 40 months. Regarding predefined populations, studies involved mostly mixed CAD, 158,781 patients followed by acute coronary syndromes (ACS), 46,338 patients, and lastly coronary artery bypass grafts (CABG), 14,583 patients involving a total of 219,702 patients. The CROS demonstrated that cardiac rehabilitation reduced total mortality in ACS, CABG and also in mixed CAD populations but not hospital readmissions and non-fatal cardiovascular events.
The CROS meta-analysis had some inherent limitations, as expected and pointed out by the authors. In several studies, there was scarce information on study protocols, cardiac rehabilitation content and processes of groups’ formation, different exercise volume programmes and predominantly mixed heterogeneous populations, always difficult to evaluate without separate sub-analysis. Despite the reduction in total mortality in all population groups, in the mixed CAD group heterogeneity was high, with coronary stable patients included together with acute coronary patients in different combinations. Correctly, the authors did not perform subgroup analysis in these patients because of the small number of studies and of patients in each subgroup, which would certainly have generated questionable conclusions. As the authors stated, the distribution and combination of secondary outcomes differed in every study, regarding which also conclusions should not be taken. Finally, a large variation of statistical methods used in the different studies was noticed, meaning that less objective outcomes might had been differently influenced.
One of the main aspects of the CROS meta-analysis, which at first view may be considered as a limitation, is the inclusion of only one cardiac rehabilitation trial. All other included studies were observational, nevertheless reflecting more the real-life patients as found in clinical practice. No other randomized trials regarding the benefits of cardiac rehabilitation have been recently taken, since cardiac rehabilitation is formally recommended after an acute event in CAD patients; for this reason randomization to a non-cardiac rehabilitation arm would raise an ethical problem.
It could be concerning to notice that in the CROS the only included trial, the RAMIT (Rehabilitation After Myocardial Infarction Trial), 2 had a neutral result regarding mortality; however, as already discussed in the literature, this study had several pitfalls. Statistically, it needed a greater sample (>8000 patients) in order to attain the primary endpoint of total mortality at two years, but several restraints limited recruitment. The final number (1813 patients) did not attain adequate dimension to evaluate the primary endpoint3,4 and for this reason, to increase the sample, non-randomizing centres were also considered, which was not statistically adequate. 3 In addition, more than 20% of the intervention group dropped out of the cardiac rehabilitation programme. 4 The RAMIT had no statistical power to demonstrate significant reduction in the primary endpoint. Besides the smaller sample of patients included, the diversity of centres, the heterogeneity of programs and the inferior exercise levels in the cardiac rehabilitation group limited the conclusions that could be assumed regarding cardiac rehabilitation comparatively with conventional treatment. It is well-known that the correct duration and intensity for physical activity is essential for the cardiac rehabilitation programme success, however, in the RAMIT, cardiac rehabilitation programmes were not long: 20 hours/6–8 weeks mean duration, including exercise, education and counselling. Finally, the number of patients who refused to participate is unknown, for which reason the population of RAMIT might be very dissimilar to the real-world patients. 5
Despite all the limitations of RAMIT and the inclusion of its neutral results, it is impressive to point out the unquestionable reduction on total mortality demonstrated in the CROS, 1 especially in a large number of patients after an acute cardiovascular event, ACS and CABG, most probably reflecting the clinical practice reality in the ‘new era’.
One of the strengths of the CROS is that the primary outcome, total mortality, is very objective and reliable, with no doubts regarding the benefit achieved with cardiac rehabilitation. Other endpoints, clinically very important, like quality of life improvement, would have been more difficult to compare, because of the different instruments used and the subjectivity of evaluation.
Another strength of the CROS regards the restriction in the scope of selected exercise training programmes. The dose of exercise, important to obtain the intended benefits, and the numerous protocols of exercise training need to be comparable. In most included studies, the cardiac rehabilitation setting was ‘out-patient’ and programme durations varied from 3–4 weeks to 12 months with at least two exercise sessions per week plus sessions for motivation, information, education and psychosocial interventions, with variable combinations.
The recently published 2016 Cochrane Systematic Review on exercise-based cardiac rehabilitation for coronary heart disease 6 did not obtain reduction in total mortality, even in the subgroup of studies published after 1995, although it reported the decrease in cardiovascular mortality and risk of hospitalization with further evidence supporting improvement in the quality of life. Contrarily to the CROS meta-analysis, only trials were evaluated, looking at the effectiveness of exercise-based treatments compared with no exercise, in people of all ages with CAD. It gathered 63 cardiac rehabilitation clinical trials, 14,486 CAD patients, adding 16 more recent trials (3,872 patients) to the already included studies in the Cochrane Systematic Review of 2011, 7 being the results of 2016 and 2011 identical.
The difference of results between the CROS and the Cochrane Systematic Reviews has several explanations.
One of the main problems of structured reviews and meta-analysis is the heterogeneity of the selected studies. This problem increases when prospective and retrospective observational studies are included, as happened in the CROS, although, due to the use of strict selection criteria, with just a few studies accepted from the greater initial number, this effect was at least minimized, especially in the ACS and CABG populations, which revealed low heterogeneity. Contrarily, the Cochrane review only included trials, with the purpose of improving validity and decreasing heterogeneity; however, the randomized included patients were so specific and ‘selected’ that they did not reflect the real world, as usually happens in trials. The included populations were quite different: predominantly low risk individuals, mostly male (>85%) with previous myocardial infarction or revascularization, angina or CAD diagnosis in the Cochrane review, and in the CROS, patients after an acute coronary event, including more female and older patients.
The lack of reporting of methods, though improved in recent trials, as noticed by the authors of the Cochrane Systematic Review of 2016, introduced difficulty in the assessment of overall methodological quality with risk of possible evidence bias. The CROS excluded many studies without clear methodology, with only 25 studies accepted out of the 18,534 abstracts initially identified.
Cardiac rehabilitation programmes were not equivalent, with the Cochrane review directed at exercise-based rehabilitation, also involving programmes exclusively offering the exercise component, and with the CROS including multi-component cardiac rehabilitation with structured exercise. Although exercise training is fundamental, risk factors control, education and psychological support are important core components of multidisciplinary cardiac rehabilitation, targeted in the CROS, but not necessarily contemplated in the Cochrane review of 2016. Despite this last observation, it is interesting to remark that previously the 2012 Cochrane review Patient education in the management of coronary heart disease, 8 including 13 trials (68,556 CAD patients), could not provide strong evidence that education reduced all-cause mortality, cardiac morbidity, hospitalization and revascularization compared with those who did not receive education. However, despite the difficulty to evaluate the outcomes related to the educational intervention, not so easily and objectively quantifiable as those of exercise intervention, there was some evidence that education improved quality of life. Secondary outcomes besides total mortality might be more influenced by health education.
At this point, it is important to highlight the real meaning of this study. The CROS demonstrated clearly the effectiveness of cardiac rehabilitation in mortality reduction beyond the beneficial effect of modern invasive and medical therapy, in ACS and CABG patients of real-life. After the CROS release it is easier to defend the prognostic necessity of cardiac rehabilitation after an acute coronary event in the ‘new era’. The best clinical practice for treating patients after an acute coronary event needs to include cardiac rehabilitation, otherwise intervention will not be complete.
From now on, the target for investigation should be focused on the way to achieving the greatest effectiveness in cardiac rehabilitation within the various subsets of patients. Different types of programme have been emerging with high intensity interval exercise training and combined exercise protocols. Home based programmes have just been evaluated in a recent systematic review and meta-analysis 9 and considered to be slightly more effective regarding long-term exercise capacity. Exercise programmes must always be carefully individualized and tailored. Exercise programmes must always be carefully individualized and tailored. Additionally, besides improvement in exercise training and physical activity, a greater investment in judicious education, needing adequate outcomes assessment, should be performed. There is still a real need to determine and validate each cardiac rehabilitation programme’s characteristics.
The main challenge of cardiac rehabilitation in this new millennium, knowing the value of cardiac rehabilitation in the ‘new era’, will be to find and implement in practice the right pathways to increase the uptake and the quality of effective cardiac rehabilitation programmes and make them appropriate for all groups of patients, especially minorities, keeping in mind the importance of improving long-term adherence.