Skip to main content

Benefits of cardiac rehabilitation following acute coronary syndrome for patients with and without diabetes: a systematic review and meta-analysis

Abstract

Aim

The benefits of cardiac rehabilitation (CR) after acute coronary syndrome (ACS) are well established. However, the relative benefit of CR in those with comorbidities, including diabetes, is not well understood. This systematic review and meta-analysis examined the benefit of CR on exercise capacity and secondary outcomes in ACS patients with a co-diagnosis of diabetes compared to those without.

Methods

Five databases were searched in May 2021 for randomised controlled trials (RCTs) and observational studies reporting CR outcomes in ACS patients with and without diabetes. The primary outcome of this study was exercise capacity expressed as metabolic equivalents (METs) at the end of CR and ≥ 12-month follow-up. Secondary outcomes included health-related quality of life, cardiovascular- and diabetes-related outcomes, lifestyle-related outcomes, psychological wellbeing, and return to work. If relevant/possible, studies were pooled using random-effects meta-analysis.

Results

A total of 28 studies were included, of which 20 reported exercise capacity and 18 reported secondary outcomes. Overall, the studies were judged to have a high risk of bias. Meta-analysis of exercise capacity was undertaken based on 18 studies (no RCTs) including 15,288 patients, of whom 3369 had diabetes. This analysis showed a statistically significant smaller difference in the change in METs in ACS patients with diabetes (standardised mean difference (SMD) from baseline to end of CR: − 0.15 (95% CI: − 0.24 to − 0.06); SMD at the ≥ 12-month follow-up: − 0.16 (95% CI: − 0.23 to − 0.10, four studies)).

Conclusion

The benefit of CR on exercise capacity in ACS patients was lower in those with diabetes than in those without diabetes. Given the small magnitude of this difference and the substantial heterogeneity in the results of the study caused by diverse study designs and methodologies, further research is needed to confirm our findings. Future work should seek to eliminate bias in observational studies and evaluate CR based on comprehensive outcomes.

Peer Review reports

Introduction

Cardiac rehabilitation (CR) is highly recommended after acute coronary syndrome (ACS) due to its beneficial effects on cardiac mortality, hospitalisation, and health-related quality of life (HRQoL) [1]. However, ACS patients with multimorbidity are underrepresented in studies evaluating CR [1]. Thus, less is known about the effectiveness of CR and the management of ACS patients living with multiple diseases.

Diabetes is a known risk factor for ACS and more than doubles the risk for cardiovascular disease [2]. The conditions have similar risk factors and are closely related in aetiology [3]. Hence, diabetes is one of the most prevalent comorbidities in CR patients; notably, up to one-third of CR patients have been estimated to have diabetes [4, 5]. Compared to ACS patients without diabetes, those with a combination of ACS and diabetes exhibit a higher mortality, accelerated loss of physical function, and a poorer HRQoL [6,7,8]. The adverse prognoses for ACS patients with diabetes call for CR interventions adapted to the needs of this high-risk group to ensure effective CR irrespective of having a co-diagnosis [9, 10].

Evidence suggests that intensified, multidisciplinary CR targeting lifestyle and medication is achievable for ACS patients with diabetes and improves their prognosis [11, 12]. Patients with a comorbidity of diabetes should be able to attend CR safely and the fundamental CR recommendations for exercise and healthy lifestyle are considered compatible with diabetic treatment irrespectively of type of diabetes [3, 12, 13]. However, safety precautions as frequent self-monitoring of blood glucose concentration before, during and after exercise are recommended [13]. Despite concordant treatment recommendations, management of patients with diabetes remains suboptimal in CR, and generally, the growing literature on multimorbidity suggests that traditional disease-specific rehabilitation potentially overlooks interactions of multiple diseases and their management [4, 10]. The insufficient management in relation to a co-diagnosis of diabetes could be explained by inherent precautions that might cause differential needs at entry to CR such as diabetes-related comorbidities, glucose-lowering medication use, dietary patterns, self-management and psychosocial wellbeing [12]. These factors might prevent the realisation of the recommended treatment and require a person-centred and multidisciplinary approach [12]. It is therefore important to examine whether these needs are adequately addressed in CR traditionally developed from a disease-specific model and how outcomes are affected [10, 14]. Knowledge in this field may contribute to evolving CR to best address the comprehensive needs of patients with co-diagnoses.

Exercise capacity is a key outcome in CR due to its ability to reduce mortality and morbidity in the general CR population as well as in patients with diabetes [1, 15]. The primary objective of the current review was therefore to examine the benefit of CR on exercise capacity in ACS patients with a co-diagnosis of diabetes compared to those without. Second, the review aimed to examine the benefit of CR on HRQoL, cardiovascular- and diabetes-related outcomes, lifestyle-related outcomes, psychological wellbeing, and return to work in ACS patients with a co-diagnosis of diabetes compared to those without.

Methods

This systematic review was reported according to the PRISMA statement [16]. The study protocol has been registered in the PROSPERO database (CRD42019151055).

Study eligibility criteria

Studies published in 2000 or later were included to reflect the current guideline-recommended management of ACS (e.g., up-to-date surgical and medical procedures and secondary prevention) [17]. The study eligibility criteria are presented in Table 1.

Table 1 Study selection criteria

The population comprised two groups: ACS patients with a co-diagnosis of diabetes (exposure) compared to those without (comparison group). Structured exercise training (Table 1) was an inclusion criterion, and other core components for CR could be included in accordance with the British Association for Cardiovascular Prevention and Rehabilitation (BACPR) [18]. Only studies published in 2000 or later were included to reflect the current guideline-recommended management of ACS (e.g., up-to-date surgical and medical procedures and secondary prevention) [17].

Outcomes

The primary outcome, cardiorespiratory fitness (CRF), referred to as exercise capacity in this paper, was measured directly using a physical test with four possible end points (i.e., VO2 max, VO2 Peak, sub maximum or symptom-limited). All exercise test results were unified through the use of metabolic equivalents (METs), which were assessed directly by a maximal test (using facial mask monitoring gas exchange) or estimated based on the workload associated with a submaximal test. All MET values were extracted as reported, and VO2 reported values were converted into METs assuming 1 MET equals 3.5 ml/kg VO2 [19]. Secondary outcomes are outlined in Table 1.

Search strategy

The search strategy was developed with support from a specialist librarian. Searches in the databases PubMed (U.S. National Library of Medicine, NCBI), EMBASE by Elsevier, Cochrane Central Register of Controlled Trials (CENTRAL), Web of Science (WoS), and CINAHL (via EBSCO-HOST) were conducted on May 24, 2021, using a strategy combining selected MeSH terms or descriptors and free text terms relating to four blocks: (1) ACS, (2) diabetes, (3) CR and (4) study design. Search strategies and search terms are documented in the additional file 1. In addition to the structured search, Cochrane reviews matching the topic "Myocardial ischaemia/coronary disease" in the Cochrane Database of Systematic Reviews were hand searched for eligible studies. The included randomised controlled trials (RCTs) from the most recent Cochrane Review on exercise-based CR were examined, and an updated search was performed in CENTRAL from 2014 to2020 for eligible studies [1]. Furthermore, reference lists of key literature [1, 12, 14, 15] were examined, and ClinicalTrials.gov was searched to identify ongoing studies (see search terms in additional file 1).

Study selection

The study selection process was conducted using Covidence software (www.covidence.org) [20]. The titles and abstracts were screened independently by at least two of three reviewers (KKWP, MBR, BBG). Next, all full-text articles marked with “yes” or “maybe” were retrieved, and the eligibility of each study was assessed by at least two of three reviewers (BBG, MBR, TM). The primary reason for exclusion of each study was recorded. Any conflicts between the two reviewers were discussed with the third reviewer until consensus was reached.

Data extraction

A predefined data extraction form was designed and used. Details are outlined in Table 2. Data extraction was performed by the first author consulted by PD, AH or JC. CR interventions in the selected studies were quality checked according to the six core components for cardiovascular disease prevention and rehabilitation outlined by BACPR (see Additional file 2) [18]. For the primary outcome, exercise capacity (METs) at baseline, end of CR and ≥ 12-month follow-up was extracted along with number of patients (n) and standard deviations (SDs) for the two groups, namely, ACS patients with a co-diagnosis of diabetes versus those without.

Table 2 Table of characteristics

Risk of bias assessment

The risk of bias judgements were assessed independently by two authors (BBG and MBR). Individual assessments were compared, and consensus was reached in discussion with a third author (TM). The Cochrane risk-of-bias tool for randomised trials, version 2 (RoB 2.0), was used to assess the risk of bias in the RCTs [21]. A modified version of the Risk Of Bias In Non-randomized Studies of Exposures (The ROBINS-E) was used to assess the risk of bias in the observational studies [22]. The modification of the ROBINS-E included leaving out domain 2 (selection of participants into the study) and domain 4 (departures from intended exposures) from the assessment. Domain 2 seemed irrelevant, as the exposure (diabetes) is a chronic condition. Instead, the definition of diabetes was extracted for all studies (Additional file 4). Signalling questions for domain 4 were found to be non-applicable for the aim of this study, e.g., "Was selection of participants into the study (or into the analysis) based on variables measured after the start of the exposure?". Instead, loss to follow-up from the study populations was noted. The studies were assessed individually in the remaining domains. Each domain was judged as low, moderate, serious, or critical. Finally, an overall risk of bias judgement was made for each study. The ROBINS-E assessment was visualised by a traffic light plot adapted from the visualisation tool robvis provided in the web app [23].

Statistical analysis

For the primary outcome, the MET change scores for each group were extracted or generated by subtracting the end of CR and 12-month METs from the baseline METs. The baseline and 12-month MET SDs were obtained from the standard error of the mean (SEM) when missing [25]. Regarding the change score SDs, imputation of these SDs was calculated in case of incomplete statistical information using a correlation coefficient or by using summary statistic level imputation [24, 25]. To evaluate the impact of the imputation strategy, a sensitivity analysis was applied based on the median observed SD from studies using an estimated cardiopulmonary exercise test (serving as the worst-case scenario) and studies using a direct cardiopulmonary exercise test (serving as the best-case scenario). The difference in change scores between the groups was calculated by a random-effects model adjusting to Hedges’ g, using change scores and change score SDs, and reported as the standardised mean difference (SMD) with 95% confidence intervals (CI) [25]. The SMD was interpreted according to the Cochrane Handbook guiding rules for interpreting SMDs [26]. Statistical heterogeneity was examined using the Cochrane Q test, quantified with the I2 statistic and interpreted according to the thresholds for the interpretation of the I2 statistic in the Cochrane Handbook [27]. Publication bias was assessed by Egger’s test and visually by a funnel plot [25]. A number of subgroup analyses were planned, and a detailed description can be found in the PROSPERO protocol (CRD42019151055). Subgroup analyses were performed by random-effects models as described above using meta-regression analyses. If planned subgroup analyses were not possible, reasons for this were addressed.

Results

The search yielded a total of 5,205 unique studies. The full text of 117 of these studies was assessed for eligibility, with 28 studies eligible for inclusion (Fig. 1). In total, 20 studies reported on the primary outcome, exercise capacity [28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47]. Of these, one RCT was eligible for inclusion [29]; however, only observational data were extracted for the purpose of this review. Ten of the studies reporting on exercise capacity also included reporting on one or more of the secondary outcomes used in this systematic review, and an additional eight studies from the literature search were identified reporting on secondary outcomes; thus, in total, 18 studies were used to assessed secondary outcomes.Additional file 3 contains references and results on secondary outcomes. Hence, in total, 28 studies were included in the current review.

Fig. 1
figure 1

PRISMA flow diagram

Study characteristics

Additional file 3 presents the study characteristics and reporting on secondary outcomes. A total of 16,661 ACS patients were included from the 20 studies reporting on exercise capacity. For the meta-analysis, two studies were subsequently excluded due to insufficient reporting of the test protocol [30] and results only being presented graphically [29]. Thus, n = 15,288 patients were analysed at the end of CR in the 18 studies included in the meta-analysis evaluating exercise capacity [28, 31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47]. Table 2 presents detailed information on the included studies.

Across the studies reporting on exercise capacity, 19–48% of the patients were diagnosed with diabetes. The total number of ACS patients with a co-diagnosis of diabetes was 3,369 (22.0%]. ACS patients with type 2 diabetes were exclusively included in 11 studies [31, 34,35,36, 38, 39, 41,42,43,44, 46]. Four studies included ACS patients with type 1 or type 2 diabetes [29, 32, 40, 47], and five studies did not account for the type of diabetes [28, 30, 33, 37, 45]. A diagnosis of diabetes was classified from a fasting blood glucose test or from hospital records in 11 of the studies [34,35,36,37,38, 40,41,42, 44, 46, 47]. In seven studies, diabetes was classified from a self-reported history, taking diabetes medication, or a lack of information on classification [28, 31,32,33, 39, 43, 45]. Additional file 4 presents specific classification procedures.

The CR programmes described in the studies reporting the primary outcome were provided as outpatient services lasting from 22 days to two years and were provided in a hospital, medical centre or community-based centre. Home-based interventions with outpatient consultations were reported in three studies [36, 39, 43]. The number of weekly sessions was 1–5, and each session lasted from 30–90 min. In addition to exercise sessions, CR components compromised educational sessions (risk factor management, psychological management and nutritional counselling). In four studies, the intervention was only reported as exercise [29, 31, 36, 39]. However, when providing a quality check of all the interventions according to the BACPR core components (Additional file 2), all of the studies were assessed as comprising elements of "lifestyle risk factor" and "audit and evaluation". Thirteen studies reported elements related to "health behaviour change and education" [28, 32,33,34,35, 37, 38, 40, 42,43,44, 46, 47]. However, less reported were the elements of "psychosocial health" (seven studies) [28, 32, 33, 38, 40, 41, 44], "medical risk management" (seven studies) [28, 33,34,35, 38, 42, 46], and "long-term strategies" (three studies) [32, 42, 44].

Adherence or compliance to the CR intervention was missing or inconsistently addressed in the majority of the studies. Four studies [28, 32, 40, 47] reported lower measures of adherence or compliance among ACS patients with a co-diagnosis of diabetes, whereas one study oppositely reported higher adherence [44].

Risk of bias

Risk of bias assessments were performed on all 20 studies reporting on exercise capacity, and the assessments are summarised in Fig. 2. For the studies reporting on exercise capacity, two were assessed as having a serious or moderate bias [46, 47], and the rest were assessed as having a critical risk of bias. Limitations were mainly related to bias due to confounding, classification of exposure and outcome as well as risk of bias due to missing data.

Fig. 2
figure 2

Risk of bias judgement of the included studies

Test procedures for measuring exercise capacity

All 20 studies measuring exercise capacity applied the same cardiopulmonary exercise test procedure for the baseline test as for the follow-up test. Exercise capacity estimated from the maximal work rate achieved was performed in eleven of the studies [28, 30, 32, 33, 38, 40, 41, 43,44,45,46], while direct measurement of V̇O2 was performed in nine studies [29, 31, 34,35,36,37, 39, 42, 47]. A ramp loading of gradual resistance was applied in six studies [28, 29, 35, 37, 41, 47], whereas two studies [34, 36] reported incremental loading. In 12 studies [30,31,32,33, 38,39,40, 42,43,44,45,46], the loading procedure was not specified. A treadmill was used in 12 studies [28, 32, 33, 35, 38, 40,41,42,43,44,45,46], and seven studies used a bicycle ergometer [29, 31, 34, 36, 37, 39, 47]. In one study, the test device was not clear [30]. Exercise capacity was reported as metabolic equivalents (METs), VO2peak (ml O2/kg per minute) or both. Follow-up was performed after the final CR session in all 20 studies. In four studies [32, 40, 42, 47], follow-up was also performed at 12 months from baseline. Additional file 5 presents the specific test methods. Two studies were excluded from the meta-analysis due to results only being presented graphically [29] and insufficient reporting of the test protocol [30].

Comparison of changes in exercise capacity from the start to the end of the intervention

After including n = 15,288 patients from 18 studies [28, 31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47], the comparison showed a significantly smaller change in exercise capacity (METs) in ACS patients with a co-diagnosis of diabetes than in those without (-0.15 (95% CI: -0.24; -0.06) I2 = 74%, p < 0.01) (Fig. 3). However, the effect size was considered small (SMD < 0.40) [26]. The sensitivity analysis to evaluate the impact of the SD imputation strategy did not give rise to concern regarding the primary imputation strategy (results not shown). Because only half of the studies used a cardiopulmonary exercise test with direct measures of VO2, which is considered the gold standard for measuring exercise capacity [48], a post hoc sensitivity analysis on the exercise test (direct versus estimated test protocol) was applied and did not show a significant difference in the estimate (p = 0.34).

Fig. 3
figure 3

Forest plot: Meta-analysis of changes in exercise capacity (expressed in METs) from the start to the end of CR intervention in ACS patients with a co-diagnosis of diabetes compared to those without

Narrative synthesis of the two studies excluded for meta-analysis reported comparable benefits of exercise capacity in ACS patients with a co-diagnosis of diabetes compared to those without in one study including n = 28 participants (estimates not reported) [29]. The study with an insufficient test protocol including n = 1,312 participants reported significantly less benefit in exercise capacity in ACS patients with a co-diagnosis of diabetes compared with those without (change in METs: 1.70 (95% CI: 1.50–1.90) vs. 2.50 (95% CI: 2.40–2.70) p < 0.05) [30].

Comparison of long-term (> 12 months) changes in exercise capacity

After including n = 5,909 patients from four studies [32, 40, 42, 47], the comparison showed a significantly smaller change in exercise capacity (METs) in ACS patients with a co-diagnosis of diabetes compared to those without (-0.16 (95% CI: -0.23; -0.10) I2 = 0%, p ≤ 0.01 (Fig. 4)). However, the effect size was considered small (SMD < 0.40) [26].

Fig. 4
figure 4

Forest plot: Meta-analysis of changes in exercise capacity (METs) from start of CR intervention to ≥ 12 months follow-up in ACS patients with a co-diagnosis of diabetes compared to those without

Assessment of publication bias

No funnel plot asymmetry (Egger’s test (p = 0.39)) was present for studies reporting on exercise capacity at the end of intervention; hence, this is interpreted as the results not being affected by small study bias (see Additional file 6, Fig. 6.4).

Subgroup analyses

In the protocolised univariate subgroup analyses, no statistical difference in METs change between groups were found on age (p = 0.39), BACPR score of CR interventions (p = 0.96), type of diabetes (p = 0.48), type of intervention (p = 0.35), and length of follow-up (p = 0.96) (Figures AD 1–3, Additional file 6). It was not possible to conduct subgroup analyses for study design, risk of bias and sex.

Secondary outcome results

From eight studies, it was possible to conduct a meta-analysis on cardiac mortality, reinfarction, revascularisation, weight and BMI. The three studies [47, 49, 50] reporting on cardiac mortality showed an increased risk of cardiac mortality at the ≥ 12-month follow-up in ACS patients with a co-diagnosis of diabetes compared to those without (OR, 2.16 [95% CI: 1.49–3.13], I2 = 49% p < 0.01). Three studies [47, 49, 50] reporting on reinfarction and revascularisation events showed a comparable risk of reinfarction at the ≥ 12-month follow-up (reinfarction: OR, 0.94 95% CI [0.617, 1.445], I2 = 3%, p = 0.79, revascularisation: OR, 1.07 95% CI [0.86,1.45], I2 = 19%, p = 0.54). Four studies on weight [30,31,32, 44] and six studies on BMI [30, 31], 33, 38, 44, 46] showed comparable changes in ACS patients with a co-diagnosis of diabetes compared to those without at the end of CR (weight: 0.20 (95% CI: 0.04; 0.37) I2 = 48%, p = 0.10; BMI: 0.19 (95% CI: 0.13; 0.26) I2 = 10%, p = 0.27). Additional file 3 provides a narrative description of the secondary outcome results that could not be analysed using meta-analysis.

Discussion

This systematic review aimed to compare the benefit of CR on exercise capacity and secondary outcomes between ACS patients with a co-diagnosis of diabetes and those without. From 18 observational studies, our findings suggest that compared to ACS patients without diabetes, those with a co-diagnosis of diabetes showed a reduction in benefit on exercise capacity. The magnitude of this difference is, however, considered small. As we found substantial heterogeneity and high levels of risk of bias among the included studies, the results should thus be interpreted with caution. For a more definite conclusion, consistency in methodologies are need with special attention to correct classification of diabetes diagnosis and confounding factors. Exploration of the subgroup analyses including clinical factors (age, type of intervention, type of diabetes), indicated that the observed heterogeneity on the primary outcome was more likely to be explained by methodological heterogeneity rather than clinical heterogeneity.

Our findings on secondary outcomes based on the results from 18 observational studies yielded diverse results; therefore, we cannot determine a definite conclusion as to whether there is evidence for differential benefits of CR on secondary outcomes for ACS patients with a co-diagnosis of diabetes in comparison to those without.

A clinically significant improvement in exercise capacity has been suggested at one MET (with each MET reducing mortality by 12%) [51]. The results from Fig. 3 show that in 11 of the 20 included study populations in the meta-analysis, improvements in exercise capacity reached or exceeded one MET at the end of the study in ACS patients with diabetes. This suggests that although we did identify a statistically significant difference in benefit after CR between patients with and without diabetes, clinically meaningful improvements can be reached for ACS patients with diabetes at the end of intervention. More studies are needed to draw conclusions on a long-term basis.

For the secondary outcomes, synthesising evidence was challenged due to variation, e.g., in choice of outcome, interventions and follow-up time across studies (Additional file 3). We found an increased risk of cardiac mortality for ACS patients with a comorbidity of diabetes compared to those without at the ≥ 12-month follow-up. Regarding reinfarction, revascularisation, weight and BMI changes seemed comparable between the ACS patients with and without diabetes. The results on blood glucose levels were not judged eligible for meta-analysis; however, improvements were not maintained in the long term for ACS patients with diabetes in one study [47]. Assessment of glycaemic control is recommended as a crucial element for optimised CR for ACS patients with diabetes and should be provided as an add-on to CR for these patients combined with strategies to improve long-term adherence to medication and healthy lifestyle to maintain decreases in blood glucose levels from a life-long perspective [12]. Future studies in ACS patients with a co-diagnosis of diabetes should strive to evaluate CR on comprehensive and standardised outcomes reflecting the biopsychosocial nature of CR.

The prognosis for ACS patients with diabetes is reported to be remarkably poor when compared to that for ACS patients without diabetes [6,7,8]. CR programmes have been reported to be underused, which is a plausible explanation for the insufficient management of ACS patients with diabetes [52]. This possibility is also supported by Jiménez-Navarro et al., who showed that although CR reduced mortality after percutaneous coronary intervention (PCI) for patients with diabetes, CR participation was paradoxically lower in patients with diabetes [53]. Furthermore, a recent study suggests that having diabetes is a strong factor affecting CR uptake [5]. Challenges regarding non-participation in CR for patients with diabetes should be a subject for future studies to identify risk factors for non-attendance to target uptake and intervention to ensure delivery of CR for ACS patients with diabetes.

Strengths and limitations

This study presents the most comprehensive systematic overview of existing evidence on differences in exercise capacity and secondary outcomes in ACS patients with and without diabetes involved in CR. Several limitations including bias from study designs and diverse methodologies in included studies however, need to be addressed as this might contribute to the vast heterogeneity observed on the primary outcome. Most importantly, included studies failed to control for confounding elements such as differential patients characteristics at baseline. Demographic and clinical covariates such as age, sex, baseline exercise capacity and surgical intervention have been identified as predictors of suboptimal gain in exercise capacity and would be relevant parameters to take into account [54, 55]. In addition to controlling for confounding elements, retrospectively formed study populations made it difficult to assess bias for the selection of participants into the study. Criteria for these study populations were, e.g., exclusion of patients registered with no follow-up exercise test [31, 33, 38, 42] or exclusion of patients who were not able to complete the CR programme [31, 37, 38, 46]. Exclusion of these groups limits the generalisability of the results to ACS patients attending and completing CR. Furthermore, limited information on patients lost to follow-up made it difficult to assess the impact of missing outcomes [35, 36, 47]. In this regard, Pischke et al. [32] reported that patients with diabetes who were lost to follow-up were significantly older and less educated than those with complete follow-up. In this case, patients lost to follow-up might have affected the results of this review and potentially diminished the difference between patients with and without diabetes.

For a pooled effect estimate in the meta-analysis, VO2 were converted into METs in five studies [31, 34, 36, 37, 44]. This does not seem to bias the result to a better or worse result, but might give a higher variation in these studies and thus a potential limitation 56].

Several studies did not report systematically screening for diabetes at the beginning of CR [28, 31,32,33, 38, 40, 43, 45]. As the prevalence of diabetes has previously been found to be considerably underestimated among patients with coronary disease [4], it is likely that misclassification of diabetes diagnosis has occurred. Additionally, diagnostic criteria of diabetes varied across the included studies. This might have contributed to the observed heterogeneity in the results on the primary outcome.

Despite our research question addressing effectiveness, the global implementation of CR as standard care [57] makes it impossible to address this with an RCT design due to ethical issues. Hence, the question naturally calls for observational studies, as confirmed by the included observational studies. The general lack of control groups not receiving CR prevents us from comparing results to the natural disease progression in patients with ACS and diabetes. However, from Kenttä et al. [39], it is indicated that CR itself prevents loss of physical function in patients with diabetes, as a control group not receiving CR was found to have greater loss in physical function [39].

Regarding the risk of bias assessment, we did not find a suitable tool to evaluate the effect of an intervention among different subgroups (ACS patients with a co-diagnosis of diabetes versus those without). The applicability of the ROBINS-E tool for our research question was challenged, as the tool originally was developed for studies examining the effects of environmental exposures on health outcomes [58]. Additionally, ROBINS-E fails to discriminate between studies with a single risk of bias or multiple risks of bias. ROBINS-E is severely limited at determining whether confounders will bias study outcomes [58]. An alternative tool, such the checklist by Wells and colleagues [59], were considered, but the focus on intervention effects was not appropriate for the aim of this review. Nevertheless, we believe that the risk of bias assessment from ROBINS-E (Fig. 2) addressed relevant methodological issues. Until a more suitable risk of bias tool is available, we did not find it relevant to define the quality of evidence according to the Grading of Recommendation, Assessment, Development and Evaluation (GRADE) approach as described in the protocol [60].

Implications for practice and further research

The findings from this systematic review highlight the need for further high-quality research into the content and effects of CR for patients with diabetes as well as participation over the course of CR for patients with diabetes. Most importantly, future studies should make efforts to eliminate potential confounding parameters such as demographic, behavioural and clinical factors that differ between ACS patients with diabetes and those without. Additionally, when a suitable checklist is available, a formal risk of bias assessment of secondary outcomes should be carried out, and clinical practice should continue to ensure the inclusion of ACS patients with diabetes in CR, as clinically meaningful benefits regarding exercise capacity seem to be reached.

Conclusion

The benefit of CR on exercise capacity in ACS patients was lower in patients with a co-diagnosis of diabetes than in those without. Given the small magnitude of this difference in exercise capacity together with substantial heterogeneity in the results of the study, further research is needed. Future work should seek to eliminate bias in observational studies, evaluate CR on comprehensive outcomes and investigate participation in CR for patients with diabetes.

Availability of data and materials

The data used and/or analysed during the current study are available from the corresponding author on reasonable request.

Abbreviations

CR:

Cardiac rehabilitation

ACS:

Acute coronary syndrome

METs:

Metabolic equivalents

HRQoL:

Health-related quality of life

SMD:

Standardised mean difference

BACPR:

British Association for cardiovascular prevention and rehabilitation

CRF:

Cardiorespiratory fitness

OR:

Odds ratio

CI:

Confidence interval

SD:

Standard deviation

References

  1. Dibben G, Faulkner J, Oldridge N, Rees K, Thompson DR, Zwisler AD, Taylor RS. Exercise-based cardiac rehabilitation for coronary heart disease. Cochrane Database Syst Rev. 2021. https://doi.org/10.1002/14651858.CD001800.pub4.

    Article  PubMed  Google Scholar 

  2. Einarson TR, Acs A, Ludwig C, Panton UH. Prevalence of cardiovascular disease in type 2 diabetes: a systematic literature review of scientific evidence from across the world in 2007–2017. Cardiovasc Diabetol. 2018;17(1):83.

    Article  Google Scholar 

  3. Squires RW, Stewart KJ. Cardiac Rehabilitation for Patients with Diabetes Mellitus. In: Reusch MDJEB, Regensteiner PMABAJG, Stewart EDFMFKJ, Veves MDDA, editors. Diabetes and Exercise: From Pathophysiology to Clinical Implementation. Cham: Springer International Publishing; 2018. p. 319–27.

  4. Gyberg V, de Bacquer D, de Backer G, Jennings C, Kotseva K, Mellbin L, et al. Patients with coronary artery disease and diabetes need improved management: a report from the EUROASPIRE IV survey: a registry from the Euro observational research programme of the European society of cardiology. Cardiovasc Diabetol. 2015;14(1):133.

    Article  Google Scholar 

  5. Harrison AS, Doherty P, Phillips A. An analysis of barriers to entry of cardiac rehabilitation in patients with diabetes: Using data from the national audit of cardiac rehabilitation. Diab Vasc Dis Res. 2018;15(2):145–9.

    Article  Google Scholar 

  6. Peterson PN, Spertus JA, Magid DJ, Masoudi FA, Reid K, Hamman RF, et al. The impact of diabetes on one-year health status outcomes following acute coronary syndromes. BMC Cardiovasc Disord. 2006. https://doi.org/10.1186/1471-2261-6-41.

    Article  PubMed  PubMed Central  Google Scholar 

  7. McGuire DK, Svensson A-M, Sattar N, Rawshani A, Mifraraj M, Franzen S, et al. Mortality and cardiovascular disease in type 1 and type 2 diabetes. N Engl J Med. 2017;376(15):1407.

    Article  Google Scholar 

  8. De Smedt D, Clays E, Annemans L, Doyle F, Kotseva K, Pająk A, et al. Health related quality of life in coronary patients and its association with their cardiovascular risk profile: Results from the EUROASPIRE III survey. Int J Cardiol. 2013;168(2):898–903. https://doi.org/10.1016/j.ijcard.2012.10.053.

    Article  PubMed  Google Scholar 

  9. Salman A, Doherty P. To what extent is the variation in cardiac rehabilitation quality associated with patient characteristics? BMC Health Serv Res. 2019. https://doi.org/10.1186/s12913-018-3831-1.

    Article  PubMed  PubMed Central  Google Scholar 

  10. Taylor RS, Singh S. Personalised rehabilitation for cardiac and pulmonary patients with multimorbidity: Time for implementation? Eur J Prev Cardiol. 2021;28(16):e19–23. https://doi.org/10.1177/2047487320926058.

    Article  Google Scholar 

  11. Soja AMB, Zwisler ADO, Frederiksen M, Melchior T, Hommel E, Torp-Pedersen C, et al. Am Heart J. 2007;153(4):621–8. https://doi.org/10.1016/j.ahj.2007.01.030.

    CAS  Article  PubMed  Google Scholar 

  12. Hansen D, Kraenkel N, Kemps H, Wilhelm M, Abreu A, Pfeiffer AFH, et al. Management of patients with type 2 diabetes in cardiovascular rehabilitation. Eur J Prev Cardiol. 2019;26(2_suppl):133–44. https://doi.org/10.1177/2047487319882820.

    Article  PubMed  Google Scholar 

  13. Hansen D, Niebauer J, Cornelissen V, Barna O, Neunhaeuserer D, Stettler C, et al. Exercise prescription in patients with different combinations of cardiovascular disease risk factors: a consensus statement from the EXPERT working group. Sports med (Auckland). 2018;48(8):1781–97.

    Article  Google Scholar 

  14. Miketic JK, Hravnak M, Stilley CS, Robertson RJ, Sereika SM. Factors influencing the outcomes of patients with both coronary artery disease and diabetes enrolled in standard cardiac rehabilitation programs: a literature review. J Cardiovasc Nurs. 2011;26(3):210–7.

    Article  Google Scholar 

  15. Kemps H, Kränkel N, Dörr M, Moholdt T, Wilhelm M, Paneni F, et al. Exercise training for patients with type 2 diabetes and cardiovascular disease: what to pursue and how to do it. A position paper of the European association of preventive cardiology (EAPC). Eur J Prev Cardiol. 2019;26(7):709–27. https://doi.org/10.1177/2047487318820420.

    Article  PubMed  Google Scholar 

  16. Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.

    Article  Google Scholar 

  17. Puymirat E, Simon T, Cayla G, Cottin Y, Elbaz M, Coste P, et al. Acute myocardial infarction. Circulation. 2017;136(20):1908–19.

    Article  Google Scholar 

  18. The BACPR Standards and Core Components for Cardiovascular Disease Prevention and Rehabilitation 2017 2017.

  19. Kokkinos P, Myers J. Exercise and physical activity: clinical outcomes and applications. Circulation. 2010;122(16):1637–48.

    Article  Google Scholar 

  20. Covidence systematic review software, Melbourne, Australia. : Veritas Health Innovation; [Available from: www.covidence.org.

  21. Sterne J A C, Savović Jelena, Page Matthew J, Elbers Roy G, Blencowe Natalie S, Boutron Isabelle, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019. https://doi.org/10.1136/bmj.l4898.

    Article  PubMed  PubMed Central  Google Scholar 

  22. Preliminary risk of bias for exposures tool template, July 2017 version 2017 [Available from: https://www.bristol.ac.uk/population-health-sciences/centres/cresyda/barr/riskofbias/robins-e/.

  23. McGuinness LA. robvis: An R package and web application for visualising risk-of-bias assessments. 2019.

  24. Ma J, Liu W, Hunter A, Zhang W. Performing meta-analysis with incomplete statistical information in clinical trials. BMC Med Res Methodol. 2008;8(1):56.

    Article  Google Scholar 

  25. Higgins JPT TJ, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (editors). Cochrane Handbook for Systematic Reviews of Interventions version 6.0 2019 [updated July 2019; cited 2020. Available from: www.training.cochrane.org/handbook.

  26. Schünemann HJ VG, Higgins JPT, Santesso N, Deeks JJ, Glasziou P, Akl EA, Guyatt GH. . Chapter 15: Interpreting results and drawing conclusions. (updated February 2021). . 2021. In: Cochrane Handbook for Systematic Reviews of Interventions version 62 [Internet]. Cochrane. Available from: Available from www.training.cochrane.org/handbook.

  27. Deeks JJ HJ, Altman DG (editors). Chapter 10: Analysing data and undertaking meta-analyses. In: Cochrane Handbook for Systematic Reviews of Interventions version 62 (updated February 2021) [Internet]. Cochrane, 2021. Available from: Available from www.training.cochrane.org/handbook.

  28. Banzer JA, Maguire TE, Kennedy CM, O’Malley CJ, Balady GJ. Results of cardiac rehabilitation in patients with diabetes mellitus. Am J Cardiol. 2004;93(1):81–4.

    Article  Google Scholar 

  29. Wu YT, Wu YW, Hwang CL, Wang SS. Changes in diastolic function after exercise training in patients with and without diabetes mellitus after coronary artery bypass surgery. A randomized controlled trial. Eur J Phys Rehabilit Med. 2012;48(3):351–60.

    Google Scholar 

  30. St M, Clair HM, Sacrinty M, Johnson D, Robinson K. Effects of cardiac rehabilitation in diabetic patients: both cardiac and Noncardiac factors determine improvement in exercise capacity: effects of CR in diabetic patients. Clin Cardiol. 2014;37(4):233–8. https://doi.org/10.1002/clc.22245.

    Article  Google Scholar 

  31. Svacinová Hana, Nováková Marie, et al. Benefit of combined cardiac rehabilitation on exercise capacity and cardiovascular parameters in patients with Type 2 diabetes. Tohoku J Exp Med. 2008;215(1):103–11. https://doi.org/10.1620/tjem.215.103.

    Article  PubMed  Google Scholar 

  32. Pischke CR, Weidner G, Elliott-Eller M, Scherwitz L, Merritt-Worden TA, Marlin R, et al. Comparison of coronary risk factors and quality of life in coronary artery disease patients with versus without diabetes mellitus. Am J Cardiol. 2006;97(9):1267–73.

    Article  Google Scholar 

  33. Hindman L, Falko JM, LaLonde M, Snow R, Caulin-Glaser T. Clinical profile and outcomes of diabetic and nondiabetic patients in cardiac rehabilitation. Am Heart J. 2005;150(5):1046–51.

    CAS  Article  Google Scholar 

  34. Vergès B, Patois-Vergès B, Cohen M, Lucas B, Galland-Jos C, Casillas JM. Effects of cardiac rehabilitation on exercise capacity in Type 2 diabetic patients with coronary artery disease. Diabet Med. 2004;21(8):889–95.

    Article  Google Scholar 

  35. Mourot L, Boussuges A, Maunier S, Chopra S, Rivière F, Debussche X, et al. Cardiovascular rehabilitation in patients with diabetes. J Cardiopulm Rehabil Prev. 2010;30(3):157–64.

    Article  Google Scholar 

  36. Karjalainen JJ, Kiviniemi AM, Hautala AJ, Niva J, Lepojarvi S, Makikallio TH, et al. Effects of exercise prescription on daily physical activity and maximal exercise capacity in coronary artery disease patients with and without type 2 diabetes. Clin Physiol Funct Imaging. 2012;32(6):445–54.

    Article  Google Scholar 

  37. Nishitani M, Shimada K, Masaki M, Sunayama S, Kume A, Fukao K, et al. Effect of cardiac rehabilitation on muscle mass, muscle strength, and exercise tolerance in diabetic patients after coronary artery bypass grafting. J Cardiol. 2013;61(3):216–21.

    Article  Google Scholar 

  38. Toste S, Viamonte S, Barreira A, Fernandes P, Gomes JL, Torres S. Cardiac rehabilitation in patients with type 2 diabetes mellitus and coronary disease: a comparative study. Rev Port Cardiol. 2014;33(10):599–608.

    Article  Google Scholar 

  39. Kenttä T, Tulppo MP, Nearing BD, Karjalainen JJ, Hautala AJ, Kiviniemi AM, et al. Effects of exercise rehabilitation on cardiac electrical instability assessed by T-wave alternans during ambulatory electrocardiogram monitoring in coronary artery disease patients without and with diabetes mellitus. Am J Cardiol. 2014;114(6):832–7.

    Article  Google Scholar 

  40. Armstrong MJ, Martin B-J, Arena R, Hauer TL, Austford LD, Stone JA, et al. Patients with diabetes in cardiac rehabilitation: attendance and exercise capacity. Med Sci Sports Exerc. 2014;46(5):845–50.

    Article  Google Scholar 

  41. Boukhris M, Tomasello SD, Khanfir R, Ibn Elhadj Z, Terra AW, Marzà F, et al. Impacts of cardiac rehabilitation on ventricular repolarization indexes and ventricular arrhythmias in patients affected by coronary artery disease and type 2 diabetes. Heart & Lung J Acute Critical Care. 2015;44(3):199–204.

    Article  Google Scholar 

  42. Kim HJ, Joo MC, Noh SE, Kim JH. Long-term outcomes of cardiac rehabilitation in diabetic and non-diabetic patients with myocardial infarction. Ann Rehabil Med. 2015;39(6):853–62.

    Article  Google Scholar 

  43. Szalewska D, Tomaszewski J, Kusiak-Kaczmarek M, Niedoszytko P, Gierat-Haponiuk K, Haponiuk I, et al. Influence of a hybrid form of cardiac rehabilitation on exercise tolerance in coronary artery disease patients with and without diabetes. Kardiol Pol. 2015;73(9):753–60.

    Article  Google Scholar 

  44. Khadanga DS, Savage AP, Ades AP. Insulin resistance and diabetes mellitus in contemporary cardiac rehabilitation. J Cardiopulm Rehabil Prev. 2016;36(5):331–8.

    Article  Google Scholar 

  45. Kasperowicz A, Cymerys M, Kasperowicz T. Effectiveness of cardiac rehabilitation in exercise capacity increase in patients with st-segment elevation myocardial infarction. Int J Environ Res Public Health. 2019;16(21):4085.

    Article  Google Scholar 

  46. Laddu DR, Ozemek C, Hauer TL, Rouleau CR, Campbell TS, Wilton SB, et al. Cardiometabolic responses to cardiac rehabilitation in people with and without diabetes. Int J Cardiol. 2020;301:156–62.

    Article  Google Scholar 

  47. Eser P, Marcin T, Prescott E, Prins LF, Kolkman E, Bruins W, et al. Clinical outcomes after cardiac rehabilitation in elderly patients with and without diabetes mellitus: The EU-CaRE multicenter cohort study. Cardiovascular Diabetology. 2020;19(1).

  48. Mezzani A, Hamm LF, Jones AM, McBride PE, Moholdt T, Stone JA, et al. Aerobic exercise intensity assessment and prescription in cardiac rehabilitation: a joint position statement of the European association for cardiovascular prevention and rehabilitation, the American association of cardiovascular and pulmonary rehabilitation and the Canadian association of cardiac rehabilitation. Eur J Prev Cardiol. 2013;20(3):442–67.

    Article  Google Scholar 

  49. Suresh V, Harrison RA, Houghton P, Naqvi N. Standard cardiac rehabilitation is less effective for diabetics. Int J Clin Pract. 2001;55(7):445–8.

    CAS  PubMed  Google Scholar 

  50. Reibis R, Treszl A, Bestehorn K, Karoff M, Schwaab B, Wirth A, et al. Comparable short-term prognosis in diabetic and non-diabetic patients with acute coronary syndrome after cardiac rehabilitation. Eur J Prev Cardiol. 2012;19(1):15–22.

    CAS  Article  Google Scholar 

  51. Myers J, Prakash M, Froelicher V, Do D, Partington S, Atwood JE. Exercise capacity and mortality among men referred for exercise testing. N Engl J Med. 2002;346(11):793–801.

    Article  Google Scholar 

  52. Ruano-Ravina A, Pena-Gil C, Abu-Assi E, Raposeiras S, Hof Avt, Meindersma EP, et al. Participation and adherence to cardiac rehabilitation programs. A systematic review. International journal of cardiology. 2016;223:436-43

  53. Jimenez-Navarro MF, Lopez-Jimenez F, Perez-Belmonte LM, Lennon RJ, Diaz-Melean C, Rodriguez-Escudero JP, et al. Benefits of Cardiac Rehabilitation on Cardiovascular Outcomes in Patients With Diabetes Mellitus After Percutaneous Coronary Intervention. Journal of the American Heart Association. 2017;6(10):e006404-e.

  54. Bargehr J, Thomas CS, Oken KR, Thomas RJ, Lopez-Jimenez F, Trejo-Gutierrez JF. Predictors of suboptimal gain in exercise capacity after cardiac rehabilitation. Am J Cardiol. 2017;119(5):687–91.

    Article  Google Scholar 

  55. Laddu D, Ozemek C, Lamb B, Hauer T, Aggarwal S, Stone JA, et al. Factors associated with cardiorespiratory fitness at completion of cardiac rehabilitation: identification of specific patient features requiring attention. Can J Cardiol. 2018;34(7):925–32.

    Article  Google Scholar 

  56. Dibben GO, Gandhi MM, Taylor RS, Dalal HM, Metcalf B, Doherty P, et al. Physical activity assessment by accelerometry in people with heart failure. BMC sports science, medicine & rehabilitation. 2020;12:47-.

  57. Piepoli MF. 2016 European guidelines on cardiovascular disease prevention in clinical practice : the sixth joint task force of the european society of cardiology and other societies on cardiovascular disease prevention in clinical practice (constituted by representatives of 10 societies and by invited experts). Int J Behav Med. 2017;24(3):321.

    Article  Google Scholar 

  58. Bero L, Chartres N, Diong J, Fabbri A, Ghersi D, Lam J, et al. The risk of bias in observational studies of exposures (ROBINS-E) tool: concerns arising from application to observational studies of exposures. Syst Rev. 2018;7(1):242.

    Article  Google Scholar 

  59. Wells GA, Shea B, Higgins JP, Sterne J, Tugwell P, Reeves BC. Checklists of methodological issues for review authors to consider when including non-randomized studies in systematic reviews. Res Synth Methods. 2013;4(1):63–77.

    Article  Google Scholar 

  60. Schünemann H BJ, Guyatt G, Oxman A, editors. GRADE handbook for grading quality of evidence and strength of recommendations Updated October 2013 [Available from: guidelinedevelopment.org/handbook. .

Download references

Acknowledgements

We would like to thank Kathrine Kam Wium Pedersen (KKWP) for assistance with screening titles and abstracts. We would also like to thank all the authors for providing additional information about their studies.

Funding

The project was funded by Public Health in the Central Denmark Region, a joint effort counting both municipalities and the Region, grant no. A1960, and a research training supplement from Aarhus University and Centre for Rehabilitation Research, Department of Public Health, Aarhus University, Aarhus, Denmark. LHT is currently funded by a grant from the Danish Regions and The Danish Health Confederation through the Development and Research Fund for financial support (project no. 2703) and a grant from Region Zealand (Exercise First).

Author information

Authors and Affiliations

Authors

Contributions

BBG, LHT, TM, ADZ, PD, and RT designed this systematic review. BBG performed the systematic literature search with assistance from a librarian. BBG, MBR and TM selected and evaluated the retrieved studies. BBG extracted the data consulted by PD, AH and JC. BBG, MBR and TM assessed the risk of bias. BBG analysed the extracted data with assistance from JC. BBG, LTH and TM drafted the manuscript. All authors read and approval the final maniscript.

Corresponding author

Correspondence to Thomas Maribo.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1.

Structure of search strategy.

Additional file 2.

BACPR Standards and Core Components.

Additional file 3.

. Secondary outcomes.

Additional file 4.

Exposure measurement methods, classification of diabetes status.

Additional file 5.

Outcome measurement methods, exercise capacity.

Additional file 6.

Subgroup analysis.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Gadager, B.B., Tang, L.H., Ravn, M.B. et al. Benefits of cardiac rehabilitation following acute coronary syndrome for patients with and without diabetes: a systematic review and meta-analysis. BMC Cardiovasc Disord 22, 295 (2022). https://doi.org/10.1186/s12872-022-02723-5

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12872-022-02723-5

Keywords

  • Acute coronary syndrome
  • Diabetes
  • Secondary prevention
  • Cardiac rehabilitation
  • Multimorbidity