Patients who do not complete cardiac rehabilitation have an increased risk of cardiovascular events during long-term follow-up

Cardiac Rehabilitation (CR) is a class I recommended intervention in patients with coronary artery disease and has beneficial effects on physical fitness, quality of life, cardiovascular risk factors and clinical outcome, including mortality [1‐5]. Despite the evidence of these beneficial effects, CR programmes are still largely underutilised [6]. What’s more, a substantial number of patients that do participate in CR attend only a few sessions and then drop out prematurely. It might reasonably be expected that such suboptimal CR participation will in less favourable results as well. Suaya and Beauchamp were the first to describe a possible relation between the number of sessions attended and mortality [7, 8]. In more recent studies such a relation was also found in specific populations, including diabetic patients and women with coronary artery disease [9, 10]. However, contradictory findings of CR results were also reported [11], which may be caused by the lack of a unanimous definition of CR ‘participation’ and ‘completion’. For example, in several studies the attendance of at least one session was already regarded as CR participation [12, 13], whereas others used a more stringent definition [7, 8].

Recently, we presented the OPTImal CArdiac REhabilitation (OPTICARE) trial; a randomised controlled CR trial that enrolled 914 patients post acute coronary syndrome (ACS) [14, 15]. All patients started a ‘standard’ 12-week CR programme, and we aimed to assess an association between completion of the programme and incidence of adverse cardiac events during prolonged follow-up. Furthermore, we looked for possible determinants of CR completion. For the current analysis, CR is defined as ‘complete’ if the participant attended at least 75% of the physical programme, and ‘incomplete’ if otherwise.

We found that post-ACS patients who did not complete CR—defined as participation of at least 75% of the sessions—had an almost threefold higher incidence of MACE during prolonged follow-up. We further found that smoking behaviour was associated with the degree of CR completion—smoking cessation was a success factor. Still, the reasons why patients did not complete the 12-week ‘standard’ CR programme remained largely unknown.

Duration and intensity of CR programmes are highly variable [18]. In a systematic review of Rauch et al. [19] no comparison could be made between beneficial effects and the type or length of CR offered, given the wide heterogeneity of CR programmes which varied both in duration (3 weeks to 12 months) and intensity (2 to 5 sessions per week). Not surprisingly therefore, a definition of ‘complete CR’ is absent in the guidelines of the European Association for Cardiovascular Prevention & Rehabilitation [20] both in terms of length of a standard programme and in particular in terms of the number of attended sessions. Unfortunately, completion is not defined at all in most studies [12], whereas in other studies it ranged from minimal attendance of one session [13], to 50% in EUROASPIRE IV [21], and to 75% in the study of Beauchamp et al. [8]. We used the most stringent Beauchamp criterion, because in our opinion CR will only be effective if the absenteeism is minimal. This position is supported by our data, as incomplete CR according to Beauchamp was a strong and independent predictor of increased risk for cardiovascular events. It should be noticed that the majority of patients in the incomplete CR group quitted CR very early in the programme, resulting in an important difference in attended sessions: on average 6 versus 23 sessions. Patients who attended a number of sessions in-between were a minority. Therefore, in our opinion, the minimal number of sessions should be the chosen >75% of the physical programme, even though this is an arbitrary choice. In contrast to what one might expect, the occurrence of MACE during CR was not the reason for incomplete CR, as all 5 patients with MACE during CR ultimately completed CR after a new coronary intervention.

Benefits of complete versus incomplete CR were shown in a few retrospective studies. In 2009 Suaya et al. [7] was the first to show a relation between mortality and less than 25 attended CR sessions in elderly with coronary artery disease. Subsequently, these results were confirmed by a study of Beauchamp et al. in which patients who suffered from an acute myocardial infarction or who underwent PCI and attended less than 25% of the CR sessions had a more than two-fold increased mortality risk during 14-year follow-up, compared with those attended >75% of the sessions [8]. More recently, Armstrong et al. reported in a large study in almost 3000 diabetic patients, included between 1996 to 2010, that diabetic patients were less likely to start and to complete CR [9]. Although complete CR was not well defined, patients who fully participated in the 12-week CR programme had reduced mortality and hospitalisation in comparison with CR participants who did not complete CR. Finally, in a large study by Colbert et al. in over 6000 women with at least one-vessel coronary artery disease who participated in CR in 1996, complete CR, defined as at least 12 of 24 CR sessions (50%) including a 12-week post-CR assessment, had the lowest mortality during long-term follow-up [10].

In a recent retrospective study of our group we already demonstrated a reduction in 10-year mortality in ACS patients. We studied 1159 ACS patients who had primary PCI. We found that patients who attended a CR programme had a significantly lower 10-year mortality than their no-CR counterparts (14.7% versus 23.5%), and that patients who completed CR had a lower 10-year mortality compared with patients who started CR but did not complete the programme (13.6% versus 18.9%) [22]. The current study differs from the ones mentioned, because: 1) our patients constituted a more homogeneous population with ACS in the large majority treated with primary PCI; 2) medical therapy was more intense; and 3) medical therapy did not differ between the patients with complete and incomplete CR.

Since incompleteness of CR is associated with clinical outcome, it seems important that we identify patients at risk for drop-out. The single most important predictor for incompleteness of CR was smoking persistence. Other authors have also shown that smoking is an important predictor for incompletion of CR [23]. We can only speculate why. Patients are motivated but may simply not be able to quit for whatever reason. However, patients who are not motivated to quit smoking may also not be motivated to work on a healthier lifestyle in general either. It could also be that smoking is just part of other patient characteristics, such as a certain type of personality, limited education, and lower socio-economic status [24]. Patients do not always see the relevance of quitting smoking, which suggests that our efforts to educate society about the negative influence of smoking is still not optimal. Another explanation might be that smokers feel stigmatised and isolated during CR as suggested by Beauchamp et al. [8]. Importantly, many authors warn that we should interpret the results of studies on smoking with caution, since in most studies ‘smoking’ was poorly defined [24]. Studies often relied on hospital records or self-reporting, without biochemical verification as we did in our current study. Since stop smoking rates are even lower when verified by objective biochemical verification we plea to continue this measurement in the future.

Since these were patients included in the OPTICARE trial, a potential bias from patients’ willingness to participate in a trial may exist. Certainly, the number of patients that drop-out may be higher in routine clinical care settings. Whether these patients have a different profile and/or cardiac outcome is not known. Also, our study population was at a relatively ‘low risk’ for drop-out, with relatively young patients without heart failure or renal impairment. Another limitation is that we do not know the reasons behind incomplete CR. From our own daily experience we know that this could be a wide variety of reasons, both medical and non-medical.

Finally, it is difficult to define what exactly constitutes ‘complete CR’ as it comprises a multidisciplinary programme, including exercise sessions, diet, smoking cessation and stress management courses. Each individual patient has different needs of attention, for instance attending dietary advise sessions may be more important for an obese diabetic patient than for a non-obese non-diabetic patient. All studies up to now focus on the number of exercise sessions: the effects of the other multidisciplinary sessions are still unknown and were therefore not incorporated in the current definition of complete CR. Future studies should address the importance of attending the non-exercise sessions.

Post-ACS patients who did not complete a ‘standard’ 12-week CR programme had a higher incidence of adverse cardiac events during long-term follow-up than their counterparts who completed the programme. Persistent smokers are at risk of non-completion, which is, indirectly, an extra argument to motivate patients to quit smoking. Still, patients could not satisfactorily be identified as potential non-completers on the basis of their smoking behaviour alone. As CR is proven to be beneficial, further research is needed to understand the reasons why patients terminate prematurely.