Rehabilitación cardíaca con ejercicios para la cardiopatía coronaria
Resumen
Antecedentes
La cardiopatía coronaria (CC) es la causa más frecuente de muerte a nivel mundial. Sin embargo, con la disminución de las tasas de mortalidad por CC, cada vez más personas que sufren esta afección podrían necesitar apoyo para controlar sus síntomas y su pronóstico. La rehabilitación cardíaca (RC) con ejercicios tiene como objetivo mejorar la salud y los desenlaces de las personas con CC. Esta es una actualización de una revisión Cochrane publicada anteriormente en 2016.
Objetivos
Evaluar la efectividad clínica y la coste‐efectividad de la RC con ejercicios (entrenamiento con ejercicios solo o en combinación con intervenciones psicosociales o educacionales) en comparación con un control "sin ejercicios" sobre la mortalidad, la morbilidad y la calidad de vida relacionada con la salud (CdVRS) en personas con CC.
Métodos de búsqueda
Se actualizaron las búsquedas de la anterior revisión Cochrane mediante búsquedas en CENTRAL, MEDLINE, Embase y otras dos bases de datos en septiembre de 2020. También se realizaron búsquedas en dos registros de ensayos clínicos en junio de 2021.
Criterios de selección
Se incluyeron los ensayos controlados aleatorizados (ECA) de intervenciones con ejercicios con un seguimiento de al menos seis meses, en comparación con un control "sin ejercicios". La población de estudio estuvo compuesta por hombres y mujeres adultos que habían tenido un infarto agudo de miocardio (IAM), un baipás aortocoronario (BAC) o una intervención coronaria percutánea (ICP) o presentaban angina pectoris o arteriopatía coronaria.
Obtención y análisis de los datos
Se examinaron todas las referencias identificadas, se extrajeron los datos y se evaluó el riesgo de sesgo de acuerdo con la metodología de Cochrane. El metanálisis se estratificó según la duración del seguimiento: corto plazo (seis a 12 meses); medio plazo (> 12 a 36 meses); y largo plazo ( > 3 años), y se utilizó la metarregresión para explorar los posibles modificadores del efecto del tratamiento. Se utilizó el método GRADE para los desenlaces principales entre seis y 12 meses (el punto temporal de seguimiento más frecuente).
Resultados principales
Esta revisión incluyó 85 estudios que asignaron al azar a 23 430 personas con CC. En esta última actualización se identificaron 22 nuevos ensayos (7795 participantes). La población incluyó predominantemente pacientes después de un IAM y una revascularización, con una media de edad entre 47 y 77 años.
En la última década, el porcentaje medio de mujeres con cardiopatía isquémica ha aumentado del 11% al 17%, pero las mujeres siguen representando un porcentaje igualmente pequeño de los participantes reclutados en general ( < 15%). Veintiuno de los estudios incluidos se realizaron en países de ingresos bajos y medios (PIBM). En general, la información sobre los ensayos fue deficiente, aunque hubo evidencia de una mejora en la calidad en la última década. La mediana de seguimiento más larga fue de 12 meses (intervalo de seis meses a 19 años).
En el seguimiento a corto plazo (seis a 12 meses), la RC con ejercicio probablemente produce una ligera reducción de la mortalidad por todas las causas (razón de riesgos [RR] 0,87; intervalo de confianza [IC] del 95%: 0,73 a 1,04; 25 ensayos; evidencia de certeza moderada), una gran reducción de los IAM (RR 0,72; IC del 95%: 0,55 a 0,93; 22 ensayos; número necesario a tratar para un desenlace beneficioso adicional [NNTB] 75; IC del 95%: 47 a 298; evidencia de certeza alta), y una gran reducción en la hospitalización por todas las causas (RR 0,58; IC del 95%: 0,43 a 0,77; 14 ensayos; NNTB 12; IC del 95%: 9 a 21; evidencia de certeza moderada). Es probable que la RC con ejercicios conlleve poca o ninguna diferencia en el riesgo de mortalidad cardiovascular (RR 0,88; IC del 95%: 0,68 a 1,14; 15 ensayos; evidencia de certeza moderada), el BAC (RR 0,99; IC del 95%: 0,78 a 1,27; 20 ensayos; evidencia de certeza alta) y la IPC (RR 0,86; IC del 95%: 0,63 a 1,19; 13 ensayos; evidencia de certeza moderada) en un seguimiento de hasta 12 meses. Se desconocen los efectos de la RC con ejercicios sobre la hospitalización cardiovascular, con un amplio intervalo de confianza que incluye tanto efectos beneficiosos como perjudiciales considerables (RR 0,80; IC del 95%: 0,41 a 1,59; evidencia de certeza baja). Hubo evidencia de heterogeneidad considerable entre los ensayos para las hospitalizaciones cardiovasculares (I2 = 53%) y de sesgo de estudio pequeño para la hospitalización por todas las causas, que no se observó en ninguno de los otros desenlaces.
En un seguimiento a medio plazo, aunque podría haber poca o ninguna diferencia en la mortalidad por todas las causas (RR 0,90; IC del 95%: 0,80 a 1,02; 15 ensayos), el IAM (RR 1,07; IC del 95%: 0,91 a 1,27; 12 ensayos) , la IPC (RR 0,96; IC del 95%: 0,69 a 1,35; seis ensayos) el BAC (RR 0,97; IC del 95%: 0,77 a 1,23; nueve ensayos) y la hospitalización por todas las causas (RR 0,92; IC del 95%: 0,82 a 1,03; nueve ensayos), se halló una gran reducción de la mortalidad cardiovascular (RR 0,77; IC del 95%: 0,63 a 0,93; cinco ensayos). La evidencia es incierta para la diferencia en el riesgo de hospitalización cardiovascular (RR 0,92; IC del 95%: 0,76 a 1,12; tres ensayos).
En un seguimiento a largo plazo, aunque podría haber poca o ninguna diferencia en la mortalidad por todas las causas (RR 0,91; IC del 95%: 0,75 a 1,10), la RC con ejercicios podría conllevar una gran reducción de la mortalidad cardiovascular (RR 0,58; IC del 95%: 0,43 a 0,78; ocho ensayos) y del IAM (RR 0,67; IC del 95%: 0,50 a 0,90; diez ensayos). La evidencia es incierta para el BAC (RR 0,66; IC del 95%: 0,34 a 1,27; cuatro ensayos) y la IPC (RR 0,76; IC del 95%: 0,48 a 1,20; tres ensayos).
La metarregresión mostró que los beneficios en los desenlaces eran independientes de la combinación de casos de CC, el tipo de RC, la dosis de ejercicio, la duración del seguimiento, el año de publicación, el contexto de la RC, la ubicación del estudio, el tamaño muestral o el riesgo de sesgo.
Hubo evidencia de que la RC con ejercicios podría aumentar ligeramente la CdVRS en varias subescalas (SF‐36 mental component, actividad física, rendimiento físico, salud general, vitalidad, actividad social y puntuaciones de salud mental) en hasta 12 meses de seguimiento; sin embargo, podrían no ser diferencias clínicamente importantes. Los ocho estudios de evaluación económica basados en ensayos mostraron que con la RC con ejercicios el uso de los recursos es potencialmente coste‐efectivo en términos de ganancia de años de vida ajustados por la calidad (AVAC).
Conclusiones de los autores
Esta revisión Cochrane actualizada apoya las conclusiones de la versión anterior, en el sentido de que la RC con ejercicios proporciona beneficios importantes a las personas con CC, incluida la reducción del riesgo de IAM, una probable ligera reducción de la mortalidad por todas las causas y una gran reducción de la hospitalización por todas las causas, junto con los costes sanitarios asociados, y una mejora de la CdVRS hasta los 12 meses de seguimiento. Durante el seguimiento a largo plazo, los beneficios podrían incluir la reducción de la mortalidad cardiovascular y el IAM. En la última década, los ensayos fueron más propensos a incluir a mujeres y a realizarse en PIBM, lo que aumenta la generalización de los resultados. Todavía se necesitan ECA bien diseñados y que proporcionen suficiente información sobre la RC en personas con CC más representativos de la práctica clínica habitual. Estos ensayos deben informar explícitamente sobre desenlaces clínicos, incluida la mortalidad y los ingresos hospitalarios, e incluir criterios de valoración validados de CdVRS, especialmente durante un seguimiento a largo plazo y evaluar los costes y la coste‐efectividad.
PICO
Resumen en términos sencillos
Rehabilitación con ejercicios para la cardiopatía coronaria
Antecedentes
La cardiopatía coronaria (CC) es la causa única más frecuente de muerte a nivel mundial. Sin embargo, con el descenso de las tasas de mortalidad por CC, cada vez son más las personas que con esta enfermedad y que podrían necesitar ayuda para controlar sus síntomas (como la angina de pecho, la dificultad para respirar con la actividad física y la fatiga) y reducir las posibilidades de sufrir problemas en el futuro, como los ataques al corazón. La rehabilitación cardíaca con ejercicios (entrenamiento con ejercicios solo o en combinación con intervenciones psicológicas o educacionales) tiene como objetivo mejorar la salud y los desenlaces de las personas con CC.
Características de los estudios
Se buscaron en la literatura científica los ensayos controlados aleatorizados (experimentos en los que se asignan al azar a los participantes a uno de dos o más grupos de tratamiento) que examinaran la efectividad de los tratamientos con ejercicios, en comparación con ningún ejercicio, en personas de cualquier edad con CC. La evidencia está actualizada hasta septiembre de 2020.
Resultados clave
En esta última actualización se identificaron 22 ensayos (7795 participantes). Se incluyeron 85 ensayos que estudiaron a 23 430 personas con CC, predominantemente supervivientes de infartos, y a personas que se habían sometido a una cirugía de revascularización cardíaca o angioplastia (un procedimiento que ensancha las arterias o las venas estrechadas u obstruidas). En 38 (45%) ensayos se incluyeron intervenciones de solo ejercicios y en 47 (55%) se incluyeron intervenciones con ejercicio más otros componentes. El tipo de ejercicio más frecuente fue la bicicleta estática, la marcha o el entrenamiento en circuito. Veintiuna intervenciones (25%) se administraron en el domicilio del participante.
Los hallazgos de esta actualización son consistentes con la versión anterior (2016) de esta revisión Cochrane y muestran importantes efectos beneficiosos de la rehabilitación cardíaca con ejercicios, como una reducción del riesgo de muerte por cualquier causa, del infarto, del ingreso hospitalario y mejoras en la calidad de vida relacionada con la salud, en comparación con no realizar ejercicios. Se identificó un pequeño conjunto de evidencia relacionada con el tema económico que indicó que la rehabilitación cardíaca con ejercicios es coste‐efectiva. Muchos de los estudios identificados en esta actualización se llevaron a cabo en países de ingresos bajos y medios, lo que aumenta la generalizabilidad de los resultados a estos entornos en los que los niveles de CC son elevados y siguen en aumento.
Calidad de la evidencia
Aunque la manera de informar sobre la metodología ha mejorado en los ensayos recientes, la falta de información sobre aspectos metodológicos clave dificultó la evaluación de la calidad metodológica general y del riesgo de posible sesgo de la evidencia.
Authors' conclusions
Summary of findings
| Exercise‐based cardiac rehabilitation compared to 'no exercise' control for coronary heart disease | ||||||
| Patient or population: people with coronary heart disease | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with 'no exercise' control | Risk with exercise‐based cardiac rehabilitation | |||||
| All‐cause mortality | Study population | RR 0.87 | 8823 | ⊕⊕⊕⊝ | Exercise‐based cardiac rehabilitation likely results in a slight reduction in all‐cause mortality up to 12 months' follow‐up. 25 RCTs with 26 comparisons. 14 RCTs reported 0 events in both the intervention and control groups. | |
| 57 per 1000 | 50 per 1000 | |||||
| Cardiovascular mortality | Study population | RR 0.88 | 5360 | ⊕⊕⊕⊝ | Exercise‐based cardiac rehabilitation likely results in little to no difference in cardiovascular mortality up to 12 months' follow‐up. 5 RCTs reported 0 events in both the intervention and control groups. | |
| 45 per 1000 | 39 per 1000 | |||||
| Fatal and/or non‐fatal MI | Study population | RR 0.72 | 7423 | ⊕⊕⊕⊕ | Exercise‐based cardiac rehabilitation results in a large reduction in fatal and/or non‐fatal MI up to 12 months' follow‐up. 24 RCTs with 24 comparisons. 3 RCTs reported 0 events in both the intervention and control groups. NNTB 75 (95% CI 47 to 298) | |
| 48 per 1000 | 35 per 1000 | |||||
| Revascularisation ‐ CABG | Study population | RR 0.99 | 4473 | ⊕⊕⊕⊕ | Exercise‐based CR results in little to no difference in CABG revascularisation up to 12 months' follow‐up. 20 RCTs with 22 comparisons. 2 RCTs reported 0 events in both the intervention and control groups. | |
| 56 per 1000 | 56 per 1000 | |||||
| Revascularisation ‐ PCI | Study population | RR 0.86 | 3465 | ⊕⊕⊕⊝ | Exercise‐based CR likely results in little to no difference in risk of PCI revascularisation up to 12 months' follow‐up. 13 RCTs with 14 comparisons. 3 RCTs reported 0 events in both the intervention and control groups. | |
| 60 per 1000 | 52 per 1000 | |||||
| All‐cause hospital admissions | Study population | RR 0.58 | 2030 | ⊕⊕⊕⊝ | Exercise‐based cardiac rehabilitation likely results in a large reduction in all‐cause hospital admissions up to 12 months' follow‐up. 14 RCTs with 16 comparisons. One RCT reported 0 events in both the intervention and control group. NNTB 12 (95% CI 9 to 21) | |
| 214 per 1000 | 124 per 1000 | |||||
| Cardiovascular hospital admissions | Study population | RR 0.80 | 1087 | ⊕⊕⊝⊝ | We are uncertain about the effects of exercise‐based CR on cardiovascular hospitalisation, with a wide confidence interval including considerable benefit as well as harm. | |
| 78 per 1000 | 62 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| a95% CI is wide and overlaps no effect; therefore, downgraded by one level for imprecision. | ||||||
Background
Description of the condition
Coronary heart disease (CHD, see Glossary in Appendix 1) is the single most common cause of death globally, with 7.46 million deaths in 2016, accounting for one‐third of all deaths (WHO 2018). In the United Kingdom (UK), an estimated 2.3 million people live with CHD – around 1.5 million men and 830,000 women, and the condition accounts for one in seven deaths in men and one in twelve deaths in women (BHF 2020). Although remaining stubbornly constant in low‐ and middle‐income countries, the mortality rate from CHD has been falling in the UK and other high‐income settings. This is due to factors such as declines in cigarette smoking, improvements in hypertension treatment and control, widespread use of statins to lower circulating cholesterol levels, and the development and timely use of thrombolysis and stents in acute coronary syndromes (Mensah 2017). Accordingly, an increasingly large number of people live with CHD and may need support to manage their symptoms and prognosis.
Description of the intervention
Many definitions of cardiac rehabilitation (CR) have been proposed. The following definition encompasses the key concepts of CR: “The coordinated sum of activities required to influence favourably the underlying cause of cardiovascular disease, as well as to provide the best possible physical, mental and social conditions, so that the patients may, by their own efforts, preserve or resume optimal functioning in their community and through improved health behaviour, slow or reverse progression of disease” (BACPR 2017). CR is a complex intervention that may involve a variety of therapies, including exercise, risk factor education, behaviour change, psychological support, and strategies that are aimed at targeting traditional risk factors for cardiovascular disease. CR is an essential part of contemporary CHD care and is considered a priority in countries with a high prevalence of CHD. Based on evidence ‐ including from the previous version of this Cochrane Review (Anderson 2016) ‐ CR following a cardiac event is a Class I recommendation from the European Society of Cardiology, and the American Heart Association and American College of Cardiology, with exercise therapy consistently identified as a central element (Knuuti 2020; Smith 2011). However, despite these positive recommendations for exercise‐based CR, it continues to be widely underused with overall participation rates in recent decades of about 40% (Kotseva 2018). Service provision, though predominantly hospital‐based, varies markedly, and referral, enrolment and completion are sub‐optimal, especially amongst women and older people (Peters 2017; Ruano‐Ravina 2016). Home‐ and technology‐based CR programmes have been advocated to widen access and participation (Dalal 2015), and interventions aimed at improving people's uptake and adherence to CR programmes have been identified (Santiago de Araújo Pio 2019).
Exercise‐based CR appears to be a safe intervention. An observational study of more than 25,000 people undergoing CR reported one cardiac event for 50,000 hours of exercise training, equivalent to 1.3 cardiac arrests per million patient‐hours (Pavy 2006). An earlier study reported one case of ventricular fibrillation per 111,996 patient‐hours of exercise, and one myocardial infarction (MI) per 294,118 patient‐hours (Van Camp 1986). In the context of CR, higher risk CHD populations have been defined as those with severe in‐hospital complications after acute coronary syndrome (ACS), cardiac surgery, or percutaneous coronary intervention (PCI) (Pelliccia 2020; Piepoli 2010).
How the intervention might work
Exercise training has been shown to have direct benefits on the heart and coronary vasculature, including myocardial oxygen demand, endothelial function, autonomic tone, coagulation and clotting factors, inflammatory markers, and the development of coronary collateral vessels (Clausen 1976; Hambrecht 2000). However, findings of the original Cochrane Review of exercise‐based CR for CHD (Jolliffe 2001), supported the hypothesis that reductions in mortality may also be mediated via the indirect effects of exercise through improvements in atherosclerotic risk factors (i.e. lipids, smoking and blood pressure) (Taylor 2006).
Why it is important to do this review
People who have had acute MI and coronary revascularisation (along with heart failure) remain those most frequently recommended for CR referral by healthcare systems across the world (Piepoli 2010; Pelliccia 2020). Regular updates to this systematic review of randomised controlled trials (RCTs) of CR for CHD is therefore key to ensuring the contemporary nature of the evidence base in order to continue to inform healthcare policy makers and guideline producers.
The 2016 Cochrane review made the following two key recommendations for future evidence collection and clinical trials (Anderson 2016).
-
The need for further evidence from 'hard to reach' groups, including women, elderly people, and ethnic minorities.
-
The need for more consistent collection and reporting of validated health‐related quality of life (HRQoL) outcomes, costs and cost‐effectiveness.
In addition, the majority of evidence (58/63, 92%) in Anderson 2016 was collected in high‐income countries (HICs), with a need to consider trials from low‐ and middle‐income countries (LMICs) when they become available.
Objectives
To assess the clinical effectiveness and cost‐effectiveness of exercise‐based CR (exercise training alone or in combination with psychosocial or educational interventions) compared with 'no exercise' control, on mortality, morbidity and health‐related quality of life (HRQoL) in people with CHD.
Methods
Criteria for considering studies for this review
Types of studies
We included RCTs (with individual participant or cluster allocation, or cross‐over design) and quasi‐RCTs (RCTs in which treatment allocation was obtained by alternation or other predictable methods) of exercise‐based CR versus 'no exercise' control. In order to present outcome data that are meaningful and relevant for clinical and policy decision‐making, we limited our search to studies with a follow‐up period of at least six months in our 2011 update of this Cochrane Review and subsequent updates. Where a full text was not available, we contacted the study authors and attempted to collect further information. If we received no response, we placed the study into the 'awaiting classification' category.
Types of participants
We included adult (≥ 18 years) men and women, in either hospital‐based and community‐based settings, who have had a myocardial infarction (MI), or who have undergone revascularisation (CABG, PCI) or who have angina pectoris or coronary artery disease defined by angiography. We included trials with mixed indication population as long as more than 50% of the trial participants had a CHD diagnosis. Please note that the terms CHD and coronary artery disease (CAD) are (or can be) sometimes used interchangeably and terms are presented as described by trialists in the Characteristics of included studies.
We excluded studies which only included participants following heart valve surgery, with heart failure, atrial fibrillation or heart transplants, or implanted with either cardiac‐resynchronisation therapy or implantable cardiovertor defibrillators. These indications are the subject of other Cochrane reviews (Anderson 2017; Nielsen 2019; Risom 2017; Sibilitz 2016; Long 2019). We also excluded studies of participants who had completed a CR programme prior to randomisation.
Types of interventions
Exercise‐based CR is defined as a supervised or unsupervised inpatient, outpatient, community‐ or home‐based intervention which includes some form of exercise training that is applied to a cardiac patient population. The intervention could be exercise training alone or exercise training in addition to psychosocial or educational interventions, or both (i.e. "comprehensive CR").
All CR interventions were compared to a 'no exercise' control, and both the intervention and control group received usual medical care. Usual care could include standard medical care, such as drug therapy, but without any form of structured exercise training or advice.
Types of outcome measures
Studies should have intended to assess any of the following outcomes in both the CR and the control groups, but these outcomes did not form the basis of our inclusion/exclusion criteria. We collected outcome data at three follow‐up periods: short‐term (6 to 12 months), medium‐term ( > 12 to 36 months), and long‐term ( > than 36 months).
Primary outcomes
-
All‐cause mortality
-
Cardiovascular mortality
-
Fatal MI and/or non‐fatal MI
-
Revascularisation with CABG
-
Revascularisation with PCI
-
All‐cause hospitalisation
-
Cardiovascular hospitalisation
We sought data on the number of trial participants who experienced the above events.
Secondary outcomes
-
HRQoL assessed using validated instruments (e.g. SF‐36 (a 36‐item Short Form Health Survey); or EQ‐5D (a standardised measure developed by the EuroQol Group))
-
Costs and cost‐effectiveness ‐ we sought reports of total healthcare or societal costs, or both. Cost‐effectiveness analyses should have reported incremental difference in cost and outcome between CR and control (e.g. cost per quality‐adjusted life year (QALY) or cost per life year gained (LYG) analysis).
Search methods for identification of studies
Electronic searches
We updated the search from the previously published Cochrane Review (Anderson 2016), by searching the following databases on 1 September 2020.
-
Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library (Issue 9, 2020).
-
Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, MEDLINE Daily and MEDLINE (Ovid; 1946 to 1 September 2020).
-
Embase (Ovid; 1980 to 2020 week 36).
-
Cumulative Index to Nursing and Allied Health Literature (CINAHL) Plus (EBSCOHost; 1937 to 1 September 2020).
-
SCI‐Expanded and CPCI‐S on Web of Science (Clarivate Analytics; 1900 to 1 September 2020).
We designed search strategies with reference to those of the previous systematic review (Anderson 2016). We searched the databases using a strategy combining subject headings and free text terms relating to exercise‐based rehabilitation and coronary heart disease, with filters applied to limit to RCTs. The RCT filter for MEDLINE is the Cochrane sensitivity‐maximising RCT filter, and for Embase, terms as recommended in the Cochrane Handbook for Systematic Reviews of Intervenations have been applied (Lefebvre 2019, hereafter referred to as the Cochrane Handbook). For the other databases, except CENTRAL, we applied an adaptation of the Cochrane RCT filter.
We applied date limits to the previously used search terms, and we searched for the new terms without date limits. We imposed no language or other limitations. We also gave consideration to variations in terms used and spellings of terms in different countries so that studies were not missed by the search strategy because of such variations. See Appendix 2 for details of the search strategies used.
Searching other resources
We searched the following clinical trial registers on 21 June 2021, for ongoing clinical trials.
-
World Health Organization (WHO) International Clinical Trials Registry platform (ICTRP) (apps.who.int/trialsearch/).
-
ClinicalTrials.gov (www.clinicaltrials.gov).
We also handsearched reference lists of retrieved articles and systematic reviews published since the last update, for any studies not identified by the electronic searches, and we sought expert advice.
Data collection and analysis
Selection of studies
Two review authors (JF, RST) independently examined the titles and abstracts of citations identified by the electronic searches for possible inclusion, and coded them as 'retrieve' (eligible or potentially eligible/unclear) or 'irrelevant’. We retrieved full‐text publications of potentially relevant studies (and had them translated into English where required), and two review authors (JF, GD) then independently determined study eligibility using a standardised inclusion form. We resolved any disagreements about study eligibility through discussion and, if necessary, a third review author (RST) was asked to arbitrate. We identified and excluded duplicates and collated multiple reports of the same study so that each study, rather than each report, is the unit of interest in the review. We recorded the selection process in sufficient detail to complete a PRISMA flow diagram and 'Characteristics of excluded studies' table (Liberati 2009).
We re‐screened full texts excluded in previous versions of this review, where the reason for exclusion was based on reporting of outcomes. None of these studies were eligible for inclusion; we updated the reasons for exclusion.
Data extraction and management
Two review authors (GD, JF) independently extracted study characteristics of included RCTs and outcome data using a standardised data collection form which had been piloted on two RCTs included in the review. A third review author (RST) checked all extracted data for accuracy. We resolved disagreements by consensus. If data were presented numerically (in tables or text) and graphically (in figures), we used the numeric data because of possible measurement error when estimating from graphs. A third review author (RST) confirmed all numeric calculations and extractions from graphs or figures. We resolved any discrepancies by consensus. One author (GD) transferred extracted data into Review Manager 5.4.1 (Review Manager 2014), and a second author (RST) spot‐checked data for accuracy against the included study.
The following categories of data were extracted.
-
Methods: study design, total duration of study, length of follow‐up, number of centres, setting, date of study conduct.
-
Participants: number randomised, number lost to follow‐up, number analysed, age, sex, ethnicity, CHD diagnosis, and inclusion and exclusion criteria.
-
Intervention and control: mode of exercise, duration, frequency and intensity, any co‐interventions and description of comparator.
-
Outcome: primary and secondary outcomes.
-
Funding and notable conflicts of interest of authors.
If there were multiple reports of the same study, we assessed the duplicate publications for additional data. We extracted outcome results at all follow‐up points post‐randomisation. We contacted study authors where necessary to provide additional information.
Assessment of risk of bias in included studies
Two review authors (GD, JF) assessed the risk of bias in included studies using the Cochrane Collaboration's risk of bias (RoB) tool, which is a domain‐based critical evaluation of the following core risk of bias items: the quality of random sequence generation and allocation concealment, blinding of outcome assessment, incomplete outcome data, and selective reporting (Higgins 2011).
All risk of bias assessments were checked by a third review author (RST), and we resolved any discrepancies by consensus. Details of the assessments of risk of bias for each included trial are shown in the Characteristics of included studies tables.
Measures of treatment effect
We processed data in accordance with the Cochrane Handbook (Deeks 2011). Dichotomous outcomes for each comparison have been expressed as risk ratios (RR) with 95% confidence intervals (CI). For primary outcomes with an effect excluding no difference, we calculated the number needed to treat for an additional beneficial/harmful outcome (NNTB/NNTH), following methods detailed in the Cochrane Handbook (Schünemann 2021). We used the assumed risk with control from the 'summary of findings Table 1' table as the 'assumed comparator risk'.
Continuous HRQoL outcome comparisons were pooled where possible; that is, when there were more than two studies using the same HRQoL measure and reporting results on the same scale using the mean difference (MD). We interpreted these data using published minimal clinically important differences (MCIDs) where available. For the SF‐36 instrument, within‐person MCIDs which vary according to domain, have been published for people with heart disease (Wyrwich 2004; 15 for physical functioning, general health and mental health; 16.7 for emotional performance; 18.75 for physical performance and vitality; 20 for bodily pain; and 25 for social functioning). There are none available for the SF‐36 summary component scores. For the EQ‐5D, an MCID of 0.05 was used for interpretation (Briggs 2017).
Unit of analysis issues
Some trials contained two arms of CR and a single control group. In these cases, we divided the number randomised to the control group in half to obtain the denominator for data analysis; the means and standard deviation for the control group remained unchanged for both comparisons. For trials with cluster randomisation, approximately correct analyses were attempted where sufficient information (the intracluster correlation coefficient (ICC)) was available.
Given the variation in trial reporting follow‐up timings, we pooled outcome results separately at three time points; namely, short‐term (6 to 12 months); medium‐term ( > 12 to 36 months); and long‐term ( > 36 months) follow‐up.
Dealing with missing data
We contacted multiple authors to verify key study characteristics (such as randomisation), data queries and obtain missing numerical outcome data.
Assessment of heterogeneity
We explored heterogeneity amongst included studies qualitatively (through visual inspection of forest plots and by comparing the characteristics of included studies), and quantitatively (using the Chi2 test of heterogeneity and the I2 statistic). We considered the magnitude and direction of effects, and strength of evidence for heterogeneity (e.g. P value from Chi2 and number of studies) alongside a threshold of I2 greater than 50% to represent substantial heterogeneity (Deeks 2011).
Assessment of reporting biases
When 10 or more studies were included in meta‐analysis, we used the funnel plot and Egger test to examine small study bias (Egger 1997). We processed data in accordance with the Cochrane Handbook (Deeks 2011). We completed data synthesis and analyses using Review Manager 5.4.1 software (Review Manager 2014) and STATA version 16.1 (StataCorp 2020).
Data synthesis
We performed random‐effects meta‐analyses with 95% CIs where appropriate (i.e. when treatments, participants, and the underlying clinical question were similar enough for pooling to make sense). We used random‐effects meta‐analyses due to the qualitative clinical heterogeneity (types of interventions and CHD population characteristics). Compared with a fixed‐effect model, this model provides a more conservative statistical comparison of the difference between intervention and control by typically providing a wider confidence interval around the effect estimate. If a statistically significant difference was present using the random‐effects model, we also reported the fixed‐effect pooled estimate and 95% CI, because of the tendency of smaller trials ‐ which are more susceptible to publication bias ‐ to be over‐weighted with a random‐effects analysis (Heran 2008a; Heran 2008b).
Subgroup analysis and investigation of heterogeneity
We undertook univariate meta‐regression to explore heterogeneity and examine potential treatment effect modifiers. We tested ten hypotheses that there may be differences in the effect of exercise‐based CR on total mortality, cardiovascular mortality, total MI, revascularisation (CABG and PCI) and all‐cause hospitalisation across the following pre‐defined subgroups.
-
CHD case mix (% participants presenting with MI).
-
'Dose' of exercise intervention (dose (units) = number of weeks of exercise training x average number of sessions/week x average duration of session in minutes).
-
Type of CR (exercise‐only CR versus comprehensive CR).
-
Length of follow‐up period (where trial reported multiple follow‐up times, the longest follow‐up was used).
-
Year of publication (pre‐1995 versus post‐1995, where 1995 is used as proxy time to represent implementation of what might be regarded as ‘modern CHD usual care’).
-
Overall sample size (N ≤ 150 versus N > 150).
-
Setting (home‐ or centre‐based CR).
-
Risk of bias (low risk of bias in < 3 out of 5 domains).
-
Study location (continent ‐ Europe, North America, Australia/Asia or Other)
-
Studies undertaken in low‐, middle‐ or high‐income countries (according to the World Bank Group) (worldbank.org).
Given the relatively small ratio of trials to covariates, meta‐regression was limited to univariate analysis (Deeks 2011). To account for multiple testing, a Bonferroni correction was used and a P value of less than 0.005 (0.05/10 covariates) was used to define statistical significance.
Sensitivity analysis
We did not undertake sensitivity analyses.
Summary of findings and assessment of the certainty of the evidence
One author (GD) used GRADEProfiler software to assess the certainty of evidence for primary outcomes reported in the review (GRADEpro GDT). We downgraded the evidence from high certainty by one level based on the following factors: indirectness of evidence, unexplained heterogeneity, publication bias, risk of bias due to study design limitations, and imprecision of results (Balshem 2011). A second author (RST) checked the assessment. We applied a GRADE assessment to the primary outcomes at 6 to 12 months (the most commonly reported follow‐up timing across trials).
Results
Description of studies
Details of the studies included in the review are listed in the Characteristics of included studies table. Details of excluded studies are listed in the Characteristics of excluded studies table.
Results of the search
In summary, a total of 85 trials reporting data for a total of 23,430 participants have been included in this review update. This includes 30 trials (55 publications, 9552 participants) from the original Cochrane Review (Jolliffe 2001) (Andersen 1981; Bell 1998; Bengtsson 1983; Bertie 1992; Bethell 1990; Carlsson 1998; Carson 1982; DeBusk 1994; Engblom 1996; Erdman 1986; Fletcher 1994; Fridlund 1991; Haskell 1994; Heller 1993; Holmbäck 1994; Kallio 1979; Leizorovicz 1991; Lewin 1992; Miller 1984; Oldridge 1991; Ornish 1990; Schuler 1992; Shaw 1981; Sivarajan 1982; Specchia 1996; Stern 1983; Vecchio 1981; Vermeulen 1983; WHO 1983; Wilhelmsen 1975); 17 studies (26 publications, 2211 participants) identified by the second updated search (Heran 2011) (Belardinelli 2001; Bäck 2008; Dugmore 1999; Giallauria 2008; Hofman‐Bang 1999; Kovoor 2006; La Rovere 2002; Manchanda 2000; Marchionni 2003; Seki 2003; Seki 2008; Ståhle 1999; Toobert 2000; VHSG 2003; Yu 2003; Yu 2004; Zwisler 2008); an additional 16 trials (20 publications, 3872 participants) from the third updated search (Anderson 2016) (Aronov 2010; Bettencourt 2005a; Briffa 2005; Hambrecht 2004; Higgins 2001; Houle 2012; Maddison 2014; Maroto 2005; Munk 2009; Mutwalli 2012; Oerkild 2012; Reid 2012; Roman 1983; Sandström 2005; Wang 2012; West 2012), as well as one publication (Dorn 1999) which provided further follow‐up data from a study included in the original review (Shaw 1981); and 22 trials (43 publications, 7795 participants) from this 2020 updated search (Aronov 2019; Bubnova 2019; Bubnova 2020; Byrkjeland 2015; Campo 2020; Chaves 2019; Dorje 2019; Hassan 2016; Hautala 2017; He 2020; Lear 2015; Ma 2020; Pal 2013; Pomeshkina 2017; Pomeshkina 2019; Prabhakaran 2020; Santaularia 2017; Snoek 2020; Sun 2016; Uddin 2020; Xu 2017; Zhang 2018). The study selection process is summarised in the PRISMA flow diagram shown in Figure 1 (Liberati 2009).
PRISMA flow diagram of study selection process
Included studies
Study design
Seventy‐nine (93%) of the studies were two‐arm parallel RCTs. Three studies compared more than two arms (Bubnova 2020; Pomeshkina 2017; Sivarajan 1982), and one study compared three arms with a waiting‐list control design (Chaves 2019) (only outcome data at six months were used in this review as waiting‐list control participants elected which arm of the study to move into after this point). One study used quasi‐randomisation methods based on week of surgery (Uddin 2020). One study was a cluster‐randomised trial (Heller 1993), with clustered data reported at the individual level, and no report of the ICC; therefore, we were unable to attempt to approximately correct the analyses. Given that the study sample size and number of events were small, the implications are expected to be minimal.
Setting
The majority of studies (48/85, 56%) were undertaken in Europe as single centre (61/85, 72%) studies. Most trials were relatively small in sample size (median 137, range: 25 to 3959). Three large trials contributed approximately 40% (8956 participants) of all included participants (Prabhakaran 2020; WHO 1983; West 2012). The median duration of trial intervention was 6 months (range 3 weeks to 42 months) with median overall trial follow‐up of 12 months (range 6 to 228 months). Sixteen trials identified in this most recent update were undertaken in low‐ and middle‐income countries (LMICs) (Aronov 2019; Bubnova 2019; Bubnova 2020; Chaves 2019; Dorje 2019; Hassan 2016; He 2020; Ma 2020; Pal 2013; Pomeshkina 2017; Pomeshkina 2019; Prabhakaran 2020; Sun 2016; Uddin 2020; Xu 2017; Zhang 2018), although the majority of trial evidence overall remained from high‐income settings (64/85, 75%).
Participants
People with MI alone were recruited in 40 trials (47%, 17,085 participants), with one trial (He 2020) recruiting people with MI in the absence of obstructive coronary artery disease (MINOCA). The remaining trials recruited people suffering exclusively from angina (5 trials, 6%, 368 participants), post‐CABG patients (7 trials, 8%, 983 participants), post‐PCI patients (7 trials, 8%, 1035 participants) or a mixed population of people with CHD (26 trials, 31%, 3959). Two trials included a mixed indication population, where more than 50% had a CHD diagnosis: one included 4 people (2%) who received valve replacement surgery (Snoek 2020). The inclusion of these people is unlikely to have implications for the findings. Additionally, Zwisler 2008 included people with congestive heart failure (12%), and those at high risk of ischaemic heart disease (30%). However, the authors kindly provided separate outcome data for the ischaemic heart disease population only. The mean age of participants within the trials ranged from 47 to 77 years. Although over half of the trials included women (62 trials, 73%), and in the last decade the median percentage of female participants has increased from 11% to 18%, women accounted for fewer than 15% of the participants recruited overall.
Interventions
Thirty‐eight of the 85 (45%) trials involved exercise‐only interventions, and 47 (55%) trials involved interventions comprised of multiple components. Of the 47 trial interventions that included other elements, 20 (43%) were made up of exercise plus education components; 16 (34%) were made up of exercise, education and psychosocial components; 7 (15%) were made up of exercise plus psychosocial components; and four (9%) were made up of exercise plus other components such as controlled diets or dietary advice, risk factor management, smoking cessation and relaxation. One study randomised participants to receive exercise only, exercise plus education, or usual care (Chaves 2019). One study compared exercise only, or exercise plus education plus psychosocial components, to usual care control (Sivarajan 1982).
The mode of exercise training in CR programmes was most often aerobic in nature and most commonly static cycling, walking or circuit training. Twenty‐two (26%) trials specifically reported the inclusion of resistance training, most commonly in the form of weight training, callisthenics or exercises using elastic bands. The 'dose' of exercise intervention (dose (units) = number of weeks of exercise training x average number of sessions/week x average duration of session in minutes) ranged considerably across trials: overall dose (median 3540, range 450 to 32,760 units); frequency (1 to 7 sessions/week); session length (20 to 90 minutes/session); and intensity (50% to 90% of maximal heart rate, peak heart rate or heart rate reserve; 50% to 95% of maximal oxygen uptake (VO2max); Borg rating of perceived exertion 11 to 16). Due to poor and inconsistent reporting of adherence and fidelity to exercise programmes in the RCTs, we were not able to consider the actual amount of exercise that the participants received or performed in this review.
Twenty‐one studies (25%) were conducted in an exclusively home‐based setting (Bäck 2008; Belardinelli 2001; Bell 1998; DeBusk 1994; Dorje 2019; Fletcher 1994; Haskell 1994; Heller 1993; Higgins 2001; Houle 2012; Lear 2015; Lewin 1992; Ma 2020; Maddison 2014; Miller 1984; Mutwalli 2012; Oerkild 2012; Reid 2012; Snoek 2020; Uddin 2020; Wang 2012), with four of these studies randomising participants to usual care, or to an electronically‐delivered intervention designed with an element of personally tailored or structured exercise, accessed via a mobile phone or the Internet (Dorje 2019; Lear 2015; Maddison 2014; Reid 2012).
Comparators
In general, comparator groups were described as receiving usual or standard care (50/85, 59%). Twenty‐four trials (28%) reported participants in the control groups receiving usual care plus education, guidance or advice about diet, exercise, or physical activity from medical professionals or via information leaflets, but no formal exercise training. Eight trials (9%) reported participants in the control group simply received “no exercise”. One trial compared exercise training to stent angioplasty for participants with stable angina (Hambrecht 2004), while another compared exercise training to an "early return to normal activities group", where participants returned to work two weeks following a myocardial infarction, without a formal CR programme (Kovoor 2006). A third trial provided participants in the control group with blinded pedometers and instructions about how to wear the pedometer correctly during seven consecutive days from morning to bedtime (Houle 2012).
Outcomes
Eighty studies (94%) measured and reported outcomes that were used in at least one quantitative analysis (meta‐analysis or vote‐counting for HRQoL and cost‐effectiveness). One study reported clinical events as part of a composite outcome (Byrkjeland 2015). Two studies indicated that outcomes of interest were measured but did not report the results (Pomeshkina 2017; Pomeshkina 2019); triallists did not respond to our requests for data.
Funding
Fifty trials (59%) were funded by not‐for‐profit organisations, one trial (1%) was funded by industry, and six trials (7%) were funded by a combination of industry and not‐for‐profit organisations. Twenty‐eight trials (33%) did not report funding sources.
Excluded studies
We excluded 201 publications identified in the current search, for reasons listed in the Characteristics of excluded studies table. The most common reasons for exclusion were associated with study design, which included insufficient follow‐up time, or that the study was not a randomised controlled trial, or the comparator intervention included an exercise component.
We describe 15 ongoing trials which meet the inclusion criteria of this review in the Characteristics of ongoing studies table. Fourteen studies are awaiting classification, pending clarification from the authors regarding study characteristics (see Characteristics of studies awaiting classification).
Risk of bias in included studies
The overall risk of bias was low or unclear (Figure 2). A number of trials failed to give sufficient detail to assess their potential risk of bias, although the quality of reporting has generally improved over the last decade, with the percentage of studies with less than three low risk of bias domains decreased from 80% to 55% over the last decade.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study
Allocation
All the trial publications reported that the trial was 'randomised', but many provided insufficient detail to assess whether the method was appropriate. A total of 30/85 (35%) studies reported details of appropriate generation of the random sequence (Andersen 1981; Bell 1998; Bethell 1990; Briffa 2005; Bubnova 2020; Byrkjeland 2015; Campo 2020; Chaves 2019; Dorje 2019; Erdman 1986; Hambrecht 2004; Haskell 1994; He 2020; Holmbäck 1994; Houle 2012; Lear 2015; Ma 2020; Maddison 2014; Munk 2009; Oerkild 2012; Pal 2013; Pomeshkina 2017; Prabhakaran 2020; Reid 2012; Santaularia 2017; Snoek 2020; Sun 2016; Wang 2012; Wilhelmsen 1975; Zwisler 2008), and 23/85 (27%) studies reported appropriate concealment of allocation (Bell 1998; Briffa 2005; Bubnova 2020; Byrkjeland 2015; Campo 2020; Chaves 2019; Dorje 2019; Haskell 1994; Holmbäck 1994; Kovoor 2006; Lear 2015; Ma 2020; Maddison 2014; Munk 2009; Oerkild 2012; Pal 2013; Prabhakaran 2020; Reid 2012; Schuler 1992; Snoek 2020; VHSG 2003; West 2012; Zwisler 2008). One study used quasi‐randomisation methods (Uddin 2020), allocating participants to CR or usual care according to the week of surgery for participants.
Blinding
Given the nature of the exercise‐based CR intervention, it is not possible to blind participants or programme personnel.
Only 24/85 studies (28%) reported adequate details of blinding of outcome assessment (Campo 2020; Chaves 2019; Dorje 2019; Fletcher 1994; Giallauria 2008; Hambrecht 2004; He 2020; Holmbäck 1994; Lear 2015; Lewin 1992; Maddison 2014; Manchanda 2000; Marchionni 2003; Munk 2009; Ornish 1990; Prabhakaran 2020; Reid 2012; Sandström 2005; Santaularia 2017; Schuler 1992; Snoek 2020; West 2012; Wilhelmsen 1975; Zwisler 2008).
Incomplete outcome data
Although losses to follow‐up and dropout were relatively high in some studies (up to 48% in trials where losses to follow‐up were reported), follow‐up of 80% or more was achieved in 59/85 (69%) studies (Andersen 1981; Aronov 2010; Aronov 2019; Bäck 2008; Belardinelli 2001; Bell 1998; Bethell 1990; Bettencourt 2005a; Briffa 2005; Campo 2020; Carlsson 1998; Dorje 2019; Dugmore 1999; Engblom 1996; Giallauria 2008; Hambrecht 2004; Haskell 1994; Hassan 2016; He 2020; Heller 1993; Holmbäck 1994; Kallio 1979; La Rovere 2002; Lear 2015; Leizorovicz 1991; Lewin 1992; Ma 2020; Maddison 2014; Manchanda 2000; Marchionni 2003; Maroto 2005; Miller 1984; Munk 2009; Oerkild 2012; Oldridge 1991; Pomeshkina 2017; Pomeshkina 2019; Prabhakaran 2020; Roman 1983; Sandström 2005; Schuler 1992; Seki 2003; Shaw 1981; Snoek 2020; Specchia 1996; Ståhle 1999; Stern 1983; Toobert 2000; Vermeulen 1983; VHSG 2003; Wang 2012; West 2012; Wilhelmsen 1975; Yu 2003; Zhang 2018; Zwisler 2008). However, reasons for loss to follow‐up and dropout were often not reported. We judged only 38/85 (44%) studies to have adequately reported reasons for loss to follow‐up and whether there were systematic differences between groups with respect to missing data, thus having a low risk of bias. We judged 40/85 (47%) studies as having a high risk of bias, and seven studies as having an unclear risk of bias.
Selective reporting
The majority (62/85; 73%) of trials reported all outcomes listed in their methods sections, or that were prespecified in the study protocol or trial registration (Campo 2020; Chaves 2019; Dorje 2019; Fridlund 1991; Prabhakaran 2020; Santaularia 2017; Snoek 2020). Nine trials failed to report all outcomes at all time points collected (Dorje 2019; La Rovere 2002; Manchanda 2000; Oerkild 2012; Ornish 1990; Pomeshkina 2017; Pomeshkina 2019; Specchia 1996; Toobert 2000), and we judged 11 studies as having an unclear risk of bias as their methods sections did not clearly describe the outcomes to be collected (Aronov 2019; Bubnova 2019; Bubnova 2020; Byrkjeland 2015; Hassan 2016; Hautala 2017; He 2020; Ma 2020; Pal 2013; Sun 2016; Uddin 2020; Wilhelmsen 1975; Xu 2017; Zhang 2018). A number of the included studies were not designed to assess treatment group differences in morbidity and mortality (as these were not the primary outcomes of these trials) and, therefore, may not have fully reported all clinical events that occurred during the follow‐up period.
Other potential sources of bias
We did not find any other potential sources of bias amongst the studies.
Effects of interventions
Where data were available, we have presented pooled outcomes at three follow‐up timings: short‐term (6 to 12 months); medium‐term ( > 12 to 36 months); and long‐term ( > 36 months).
Primary outcomes
All‐cause mortality
Sixty‐one of the 85 included studies (72%) reported all‐cause mortality (Analysis 1.1). Four trials contributed mortality data at more than one follow‐up period (Shaw 1981; West 2012; WHO 1983; Wilhelmsen 1975). Fourteen trials reported zero events in both the intervention and control groups up to 12 months' follow‐up (Aronov 2019; Byrkjeland 2015; Chaves 2019; Hambrecht 2004; Houle 2012; Kovoor 2006; Maddison 2014; Manchanda 2000; Munk 2009; Pomeshkina 2017; Pomeshkina 2019; Santaularia 2017; Seki 2008; Zhang 2018).
Compared with 'no exercise' control, exercise‐based CR likely results in a slight reduction in all‐cause mortality up to 12 months' follow‐up (RR 0.87, 95% CI 0.73 to 1.04; I2 = 0%; 25 trials, 26 comparisons, 8823 participants). The certainty of the evidence was moderate due to imprecision, with a wide confidence interval. There was no evidence of publication bias for all‐cause mortality up to 12 months' follow‐up (Figure 3; Egger test: P = 0.50).
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: all‐cause mortality at 6 to 12 months' follow‐up
At medium‐ and long‐term follow‐up, exercise‐based CR may result in little to no difference in all‐cause mortality (medium‐term: RR 0.90, 95% CI 0.80 to 1.02; I2 = 0%; 16 trials, 11,073 participants; long‐term: RR 0.91, 95% CI 0.75 to 1.10; I2 = 35%; 11 trials, 3828 participants). There was no evidence of publication bias for all‐cause mortality at medium‐ or long‐term follow‐up (Figure 4; Egger test: P = 0.54; Figure 5; Egger test: P = 0.15, respectively).
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: all‐cause mortality at > 36 months' follow‐up
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: all‐cause mortality at > 12 to 36 months' follow‐up
Cardiovascular mortality
Thirty‐three of the 85 trials (39%) reported cardiovascular mortality (Analysis 1.2). One trial reported both short‐ and medium‐term follow‐up (WHO 1983). Up to 12 months' follow‐up, five trials reported zero events in both the intervention and control group (Byrkjeland 2015; Chaves 2019; Maddison 2014; Munk 2009; Seki 2008). At medium‐term follow‐up, one trial reported zero events in both the intervention and control groups (Belardinelli 2001).
Exercise‐based CR likely results in little to no difference in cardiovascular mortality up to 12 months' follow‐up (RR 0.88, 95% CI 0.68 to 1.14; I2 = 0%; 15 trials, 5360 participants). This result may be driven by the WHO 1983 trial which carries the majority of the weight. The certainty of the evidence was moderate due to imprecision, with a wide confidence interval. There was no evidence of publication bias for cardiovascular mortality (Figure 6; Egger test: P = 0.76).
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.2: cardiovascular mortality at 6 to 12 months' follow‐up
However, at medium‐term follow‐up, evidence suggests exercise‐based CR results in a large reduction in cardiovascular mortality (RR 0.77, 95% CI 0.63 to 0.93; I2 = 5%; 5 trials, 3614 participants), but again, this result may be driven by the WHO 1983 trial which accounts for the majority of the weight. Similarly, at long‐term follow‐up, evidence suggests a large reduction in cardiovascular mortality (RR 0.58, 95% CI 0.43 to 0.78; I2 = 0%; 8 trials, 1392 participants).
Fatal or non‐fatal myocardial infarction
Forty‐two of the 85 trials (49%) reported the risk of fatal or non‐fatal MI (Analysis 1.3). Three trials reported zero events in both the intervention and control groups up to 12 months' follow‐up (Maddison 2014; Reid 2012; Seki 2008). Five studies contributed MI data at multiple follow‐up time points (Hambrecht 2004; Haskell 1994; Hofman‐Bang 1999; West 2012; WHO 1983).
Exercise‐based CR likely results in a large reduction in fatal or non‐fatal MI up to 12 months' follow‐up (RR 0.72, 95% CI 0.55 to 0.93; I2 = 7%; 22 trials, 24 comparisons, 7423 participants). The NNTB is 75 (95% CI 47 to 298), meaning one additional MI could be prevented up to 12 months for every 75 people participating in exercise‐based CR. The certainty of the evidence was high, and there was no evidence of publication bias (Figure 7; Egger test: P = 0.12).
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.3: myocardial infarction at 6 to 12 months' follow‐up
The evidence suggests there may be little to no difference for risk of MI with exercise‐based CR at medium‐term follow‐up (RR 1.07, 95% CI 0.91 to 1.27, I2 = 0%; 12 trials, 9565 participants), which may be driven by the WHO 1983 study which carries more weight than other studies included in this analysis. There was no evidence of publication bias at medium‐term follow‐up (Figure 8; Egger test: P = 0.18).
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: myocardial infarction at > 12 to 36 months' follow‐up
At long‐term follow‐up, the evidence suggests that exercise‐based CR results in a large reduction in risk of fatal or non‐fatal MI (RR 0.67, 95% CI 0.50 to 0.90; I2 = 0%; 10 trials, 1560 participants). There was no evidence of publication bias at long‐term follow‐up (Figure 9; Egger test: P = 0.19).
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: myocardial infarction at > 36 months' follow‐up
Revascularisation ‐ CABG
Thirty‐one of the 85 included trials (36%) reported the risk of CABG (Analysis 1.4). Four studies contributed CABG data at multiple follow‐up time points (Haskell 1994; Hofman‐Bang 1999; Stahle 1999; West 2012). Two studies reported zero events in both the intervention and control groups up to 12 months' follow‐up (Maddison 2014; Seki 2008).
There was little to no difference between exercise‐based CR and 'no exercise' control for CABG up to 12 months' follow‐up (RR 0.99, 95% CI 0.78 to 1.27; I2 = 0%; 20 trials, 22 comparisons, 4473 participants). The certainty of evidence was high, and there was no evidence of publication bias for CABG at short‐term follow‐up (Figure 10; Egger test: P = 0.10).
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: CABG at 6 to 12 months' follow‐up
Similarly, at medium‐term follow‐up, evidence suggests little to no difference between exercise‐based CR and 'no exercise' control in risk of CABG (RR 0.97, 95% CI 0.77 to 1.23; I2 = 0%; 9 trials, 2826 participants), whereas across the small number of studies reporting CABG at long‐term follow‐up, evidence was uncertain about the effect of exercise‐based CR on risk of CABG, with a wide 95% CI including considerable benefit and harm (RR 0.66, 95% CI 0.34 to 1.27; I2 = 18%; 4 trials, 675 participants).
Revascularisation ‐ PCI
Twenty‐one of the 85 included trials (25%) reported the risk of PCI (Analysis 1.5). Four studies contributed PCI data at multiple follow‐up time points (Haskell 1994; Hofman‐Bang 1999; Stahle 1999; West 2012). Three studies reported zero events in both the intervention and control groups up to 12 months' follow‐up (Maddison 2014; Reid 2012; Seki 2008).
Exercise‐based CR likely results in little to no difference in PCI up to 12 months' follow‐up (RR 0.86, 95% CI 0.63 to 1.19; I2 = 7%; 13 trials, 14 comparisons, 3465 participants). The certainty of evidence was moderate due to imprecision, with wide confidence intervals. There was no evidence of publication bias for PCI at short‐term follow‐up (Figure 11; Egger test: P = 0.94).
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: PCI at 6 to 12 months' follow‐up
At medium‐term and long‐term follow‐up, the evidence is uncertain whether there is a benefit for risk of PCI with exercise‐based CR as the 95% CI is consistent with possible benefit and possible harm (medium‐term: RR 0.96, 95% CI 0.69 to 1.35; I2 = 26%; 6 trials, 1983 participants; long‐term: RR 0.76, 95% CI 0.48 to 1.20; I2 = 0%; 3 trials, 567 participants).
All‐cause hospitalisation
Twenty‐three of the 85 included studies (27%) reported all‐cause hospital admissions (Analysis 1.6). One study reported follow‐up at both short‐ and medium‐term (Hofman‐Bang 1999). One trial reported zero events in both the intervention and control groups at up to 12 months' follow‐up (Maddison 2014). No trials reported hospitalisation data at long‐term follow‐up.
Exercise‐based CR probably results in a large reduction in all‐cause hospital admissions up to 12 months' follow‐up (RR 0.58, 95% CI 0.43 to 0.77; I2 = 42%; 14 trials, 16 comparisons, 2030 participants). The NNTB is 12 (95% CI 9 to 21) meaning one additional hospital admission for any cause could be prevented up to 12 months for every 12 people participating in exercise‐based CR. The certainty of evidence was moderate, downgraded because of evidence of publication bias (Figure 12; Egger test: P = 0.003).
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: all‐cause hospitalisation at 6 to 12 months' follow‐up
At medium‐term follow‐up, evidence suggests exercise‐based CR may result in little to no difference in all‐cause hospitalisation (RR 0.92, 95% CI 0.82 to 1.03; I2 = 0%; 9 trials, 5995 participants).
Cardiovascular hospitalisation
Eight studies reported cardiovascular hospital admissions (Analysis 1.7). One study contributed cardiovascular hospital admission data over two follow‐up time points (Haskell 1994). No trials reported data at long‐term follow‐up. Definitions of cardiovascular hospitalisation differed somewhat between trials (Campo 2020: hospitalisations for a cardiovascular cause (acute coronary syndrome (ACS), cerebrovascular accident, heart failure, chronic coronary syndrome; Hambrecht 2004: hospitalisation and coronary angiography owing to worsening angina; Haskell 1994: cardiac events initiating hospitalisation (death, MI, CABG, percutaneous transluminal coronary angioplasty (PTCA)); Mutwalli 2012: elevated heart rate, deep sternal infection and heart attack; Reid 2012: rehospitalised with chest pain; Snoek 2020: hospitalisation for cardiac reasons (chronic coronary syndrome, ACS, pacemaker, PCI, endocarditis, dyspnoea); VHSG 2003: chest pain without objective evidence of ischaemia; Zwisler 2008: acute first‐time readmissions due to heart disease).
We are uncertain whether exercise‐based CR may result in reduced risk of cardiovascular hospital admissions up to 12 months' follow‐up (RR 0.80, 95% CI 0.41 to 1.59; I2 = 53%; 6 trials, 1087 participants). The certainty of evidence was low due to evidence of substantial heterogeneity and a wide confidence interval including considerable benefit as well as harm.
Similarly, evidence is uncertain in risk of cardiovascular hospitalisation in the few studies that reported medium‐term follow‐up (RR 0.92, 95% CI 0.76 to 1.12; I2 = 0%; 3 trials, 943 participants). There were insufficient studies to assess publication bias.
Secondary outcomes
Health‐related quality of life
Fifteen trials (18%) measured HRQoL at short‐term follow‐up using the same validated measure and reported outcomes on the same scale, enabling meta‐analyses using MD to be performed. For each of these validated measures, an increase in score indicates improvement in HRQoL. There were not enough data reported across trials at medium‐ and long‐term follow‐up for meta‐analysis to be performed.
Six studies (1731 participants) reported the SF‐36 summary scores (physical component score (PCS) and mental component score (MCS) at short‐term follow‐up (Analysis 1.8). Exercise‐based CR may slightly increase PCS compared to 'no exercise' control (MD 1.70, 95% CI ‐0.08 to 3.47; P = 0.06; 6 trials) and likely increases MCS (MD 2.14, 95% CI 1.07 to 3.22; 6 trials) up to 12 months' follow‐up. However, it is unclear whether these improvements are clinically meaningful. There was evidence of substantial heterogeneity for PCS (I2 = 73%, P = 0.002), but not for MCS (I2 = 21%).
Eight studies (2812 participants) reported SF‐36 individual domain scores (physical functioning, physical performance, bodily pain, general health, vitality, social functioning, emotional performance, mental health) at short‐term follow‐up (Analysis 1.9). One study did not report scores for the vitality and emotional performance domains (Belardinelli 2001). Exercise‐based CR may result in an increase in six out of eight domains: physical functioning score (MD 8.47, 95% CI 3.69 to 13.24); physical performance (MD 8.08, 95% CI 2.89 to 13.27); general health (MD 5.66, 95% CI 2.08 to 9.25); vitality (MD 5.78, 95% CI 1.89 to 9.67); social functioning (MD 1.98, 95% CI 0.26 to 3.70; I2 = 20%); and mental health (MD 5.60, 95% CI 1.21 to 9.98). There was no difference in the domains bodily pain (MD ‐0.06, 95% CI ‐8.97 to 8.84) and emotional performance (MD 0.69, 95% CI ‐1.33 to 2.71; I2 = 18%). There was evidence of substantial heterogeneity for the following domains: physical functioning (I2 = 92%, P < 0.001); physical performance (I2 = 87%, P < 0.001); bodily pain (I2 = 97%, P < 0.001); general health (I2 = 84%, P < 0.001); vitality (I2 = 85%, P < 0.001); and mental health (I2 = 93%, P < 0.001). Based on the minimally important clinical differences reported by Wyrwich 2004, the increases in each of the domains are not clinically important.
Three studies (476 participants) reported EQ‐5D visual analogue scores at short‐term follow‐up (Analysis 1.10). Exercise‐based CR may increase EQ‐5D scores up to 12 months' follow‐up (MD 0.05, 95% CI ‐0.01 to 0.10; P = 0.08). There was evidence of substantial heterogeneity (I2 = 69%, P = 0.04). The increase in EQ‐5D could potentially be clinically meaningful (Briggs 2017).
In addition to the meta‐analyses, given both the heterogeneity in HRQoL outcome measures and methods of reporting findings, a vote‐counting method was used to summarise descriptive data and direction of effect for all the studies that reported HRQoL (Table 1; Campbell 2020). Thirty‐two of the 85 trials (38%, N = 7680 participants) assessed HRQoL using a range of validated generic (e.g. SF‐36) or disease‐specific (e.g. HeartQoL) outcome measures. Thirty of these trials reported HRQoL data at short‐term follow‐up, three reported HRQoL data at medium‐term follow‐up, and only one trial reported HRQoL data at long‐term follow‐up. Although most trials demonstrated an improvement in HRQL at follow‐up compared to baseline following exercise‐based CR, a within‐group improvement was also often reported in control participants. Twenty trials reported higher levels of HRQoL in one or more subscales with exercise‐based CR compared to control at short‐term follow‐up (Belardinelli 2001; Bettencourt 2005a; Briffa 2005; Bubnova 2019; Bubnova 2020; Campo 2020; Hassan 2016; Hautala 2017; He 2020; Heller 1993; Hofman‐Bang 1999; Houle 2012; Ma 2020; Maddison 2014; Mutwalli 2012; Reid 2012; Santaularia 2017; Uddin 2020; Wang 2012; Yu 2003), with three at medium‐term follow‐up (Bell 1998; Toobert 2000; Yu 2003), and one at long‐term follow‐up (Engblom 1996). In twelve trials, there was evidence of a significantly higher level of quality of life in most (> 50%) of the subscales at short‐term follow‐up only (Belardinelli 2001; Bell 1998; Bubnova 2019; Bubnova 2020; Campo 2020; Hassan 2016; Hautala 2017; Ma 2020; Mutwalli 2012; Reid 2012; Uddin 2020; Wang 2012).
| Measure of HRQoL | Mean (SD) outcome values at follow‐up | P value | Difference between groups | |
|
| Exercise | Control |
|
|
| Aronov 2019 | ||||
| Quality of life questionnaire developed by authors (Aronov 2002) % change of mean score at 6 months | ||||
|
| Δ% | Δ% |
|
|
|
| 30.4 | “no change” | NR |
|
| Bell 1998 | ||||
| Nottingham Health Profile at 10.5 months' follow‐up: | ||||
| Energy | 17.6 (27.1) | 18.3 (29.8) | 0.87** | Exercise = Control |
| Pain | 2.8 (8.8) | 4.82 (11.9) | < 0.05 | Exercise > Control |
| Emotional reactions | 6.4 (17.0) | 12.2 (19.9) | < 0.001 | Exercise > Control |
| Sleep | 7.5 (18.4) | 20.5 (27.8) | < 0.001 | Exercise > Control |
| Social isolation | 2.3 (10.6) | 4.0 (13.3) | 0.37* | Exercise = Control |
| Physical mobility | 8.4 (11.1) | 8.9 (14.5) | 0.82** | Exercise = Control |
| Belardinelli 2001 | ||||
| SF‐36 at 6 months' follow‐up: | ||||
| Physical functioning | 78 (19) | 55 (20) | 0.001 | Exercise > Control |
| Physical performance | 75 (13) | 65 (14) | 0.01 | Exercise > Control |
| Bodily pain | 4 (9) | 22 (10) | 0.001 | Exercise > Control |
| General health | 68 (14) | 50 (19) | 0.001 | Exercise > Control |
| Vitality | NR | NR |
|
|
| Social functioning | 66 (10) | 69 (12) | 0.14* | Exercise = Control |
| Emotional performance | NR | NR |
|
|
| Mental health | 65 (12) | 48 (15) | 0.01 | Exercise > Control |
| SF‐36 at 12 months' follow‐up: | ||||
| Physical functioning | 82 (18) | 54 (20) | 0.001 | Exercise > Control |
| Physical performance | 76 (9) | 58 (14) | 0.01 | Exercise > Control |
| Bodily pain | 4 (9) | 32 (12) | 0.001 | Exercise > Control |
| General health | 70 (14) | 50 (18) | 0.001 | Exercise > Control |
| Vitality | NR | NR |
|
|
| Social functioning | 68 (11) | 68 (12) | 1.00* | Exercise = Control |
| Emotional performance | NR | NR |
|
|
| Mental health | 70 (14) | 45 (15) | 0.001 | Exercise > Control |
| Bettencourt 2005 | ||||
| SF‐36 at 1 year follow‐up: | ||||
| Physical functioning | 70 | 62 | NS* | Exercise = Control |
| Physical performance | 66 | 57 | NS* | Exercise = Control |
| Bodily pain | 73 | 65 | NS* | Exercise = Control |
| General health | 57 | 46 | < 0.02 | Exercise > Control |
| Vitality | 62 | 47 | < 0.02 | Exercise > Control |
| Social functioning | 73 | 66 | NS* | Exercise = Control |
| Emotional performance | 65 | 58 | NS* | Exercise = Control |
| Mental health | 87 | 75 | NS* | Exercise = Control |
| Mental component | 71 | 57 | 0.02 | Exercise > Control |
| Physical component | 63 | 57 | NS* | Exercise = Control |
| Briffa 2005 | ||||
| SF‐36 at 6 months' follow‐up: | ||||
|
| Δ (95% CI) | Δ (95% CI) |
|
|
| Physical functioning | 15.9 (‐8 to 23) | 7.1 (1 to 13) | NS* | Exercise = Control |
| Physical performance | 75 (0 to 100) | 75 (0 to 100) | NS* | Exercise = Control |
| Bodily pain | 26.6 (18 to 35) | 19.2 (11 to 27) | NS* | Exercise = Control |
| General health | 0.1 (‐6 to 6) | ‐0.6 (‐5 to 4) | NS* | Exercise = Control |
| Vitality | 7.1 (1 to 13) | 3.7 (‐2 to 9) | NS* | Exercise = Control |
| Social functioning | 19.6 (10 to 29) | 14.1 (7 to 21) | NS* | Exercise = Control |
| Emotional performance | 33.3 (0 to 100) | 33.3 (33 to 100) | NS* | Exercise = Control |
| Mental health | 0.5 (‐4 to 5) | 1.4 (‐3 to 5) | NS* | Exercise = Control |
| SF‐36 at 1 year follow‐up: | ||||
|
| Δ (95% CI) | Δ (95% CI) |
|
|
| Physical functioning | 17.6 (10 to 25) | 6.8 (‐1 to 14) | 0.04 | Exercise > Control |
| Physical performance | 100 (0 to 100) | 75 (12 to 30) | NS* | Exercise = Control |
| Bodily pain | 30.2 (23 to 37) | 20.9 (‐2 to 7) | NS* | Exercise = Control |
| General health | 2.7 (‐3 to 5) | 2.2 (‐2 to 7) | NS* | Exercise = Control |
| Vitality | 11.9 (6 to 18) | 6.9 (1 to 12) | NS* | Exercise = Control |
| Social functioning | 23.6 (14 to 33) | 16.4 (9 to 23) | NS* | Exercise = Control |
| Emotional performance | 33.3 (33 to 100) | 33.3 (33 to 100) | NS* | Exercise = Control |
| Mental health | 3.6 (‐1 to 9) | 3.9 (0 to 8) | NS* | Exercise = Control |
| Bubnova 2019 | ||||
| Quality of life questionnaire developed by authors (Aronov 2002) mean (SD) score after 12 months: | ||||
| Low rehabilitation potential subgroup | ‐4.9 (4.5) | ‐7.8 (3.1) | < 0.05 | Exercise > Control |
| Average rehabilitation potential subgroup | ‐5 (3.2) | ‐7.4 (4.3) | < 0.05 | Exercise > Control |
| High rehabilitation potential subgroup | ‐4.3 (3.9) | ‐5.6 (4.3) | < 0.05 | Exercise > Control |
| Bubnova 2020 | ||||
| Quality of life questionnaire developed by authors (Aronov 2002) mean (%) score change at 12 months: | ||||
|
| Δ (%) | Δ (%) |
|
|
| BMI < 30 kg/m2 group | 42 (6%) | 10 (2%) | <0.01 | Exercise > Control |
| BMI ≥ 30 kg/m2 group | 27 (5%) | 8 (2%) | <0.001 | Exercise > Control |
| Campo 2020 | ||||
| EuroQol at 6 months' follow‐up: | ||||
|
| Median (IQR) | Median (IQR) |
|
|
| VAS (visual analogue scale) | 80 (70‐90) | 70 (50‐80) | < 0.001 | Exercise > Control |
| 5 domains | N (%) | N (%) |
|
|
| Pain/discomfort: | 103 (89) | 89 (77) | 0.03 | Exercise > Control |
| Anxiety/depression: | 92 (79) | 67 (58) | 0.001 | Exercise > Control |
| Mobility: | 104 (90) | 80 (70) | < 0.001 | Exercise > Control |
| Self‐care: | 114 (98) | 87 (76) | 0.6 | Exercise = Control |
| Usual activities: | 101 (87) | 87 (76) | 0.04 | Exercise > Control |
| EuroQol at 12 months' follow‐up: | ||||
|
| Median (IQR) | Median (IQR) |
|
|
| VAS (visual analogue scale) | 75 (70‐87) | 65 (50‐80) | < 0.001 | Exercise > Control |
| 5 domains | N (%) | N (%) |
|
|
| Pain/discomfort: | 86 (77) | 72 (65) | 0.04 | Exercise > Control |
| Anxiety/depression: | 83 (74) | 58 (53) | 0.03 | Exercise > Control |
| Mobility: | 95 (85) | 74 (67) | 0.008 | Exercise > Control |
| Self‐care: | 101 (91) | 100 (91) | 0.8 | Exercise = Control |
| Usual activities: | 99 (88) | 80 (73) | 0.004 | Exercise > Control |
| Dorje 2019 | ||||
| SF‐12 at 6 months' follow‐up: | ||||
| Physical health score | 46.8 (6.9) | 45.2 (6.5) | 0.22** | Exercise = Control |
| Mental health score | 51.5 (9.3) | 50 (8.6) | 0.28** | Exercise = Control |
| Engblom 1992 | ||||
| Nottingham Health Profile at 5 years' follow‐up: | ||||
| Energy | 18 | 25 | 0.08 | Exercise = Control |
| Pain | 12 | 18 | 0.07 | Exercise = Control |
| Emotional reactions | 14 | 21 | 0.27 | Exercise = Control |
| Sleep | 24 | 29 | 0.42 | Exercise = Control |
| Social isolation | 7 | 9 | 0.42 | Exercise = Control |
| Physical mobility | 6 | 14 | 0.005 | Exercise > Control |
| Hassan 2016 | ||||
| SF‐36 8 domains at 12 months' follow‐up | ||||
| Physical functioning | 83.5 (6.5) | 76.7 (10.6) | 0.01 | Exercise > Control |
| Role limitations physical | 62.5 (23.4) | 50.8 (20.2) | 0.04 | Exercise > Control |
| Role limitations emotional | 61.1 (21.6) | 49.9 (19.1) | 0.04 | Exercise > Control |
| Energy/fatigue | 66 (11.1) | 57.7 (11.7) | 0.01 | Exercise > Control |
| Emotional well being | 69.5 (2.6) | 61.5 (7.5) | 0.000 | Exercise > Control |
| Social functioning | 67.5 (19) | 56.3 (16.3) | 0.02 | Exercise > Control |
| Pain | 79.6 (18.4) | 67.9 (15.9) | 0.01 | Exercise > Control |
| General health | 43 (7.9) | 38.5 (8.8) | 0.04 | Exercise > Control |
| Hautala 2017 | ||||
| 15D Quality of life measure at 6 months' follow‐up: | ||||
|
| 0.915 (0.07) | 0.876 (0.084) | 0.0004* | Exercise > Control |
| 15D Quality of life measure at 12 months' follow‐up: | ||||
|
| 0.922 (0.072) | 0.886 (0.088) | < 0.0015* | Exercise > Control |
| He 2020 | ||||
| SF‐36 at 12 months: | ||||
| Physical functioning | 85 (22) | 74 (19) | < 0.01 | Exercise > Control |
| Role‐physical | 80 (21) | 77 (22) | 0.362 | Exercise = Control |
| Bodily pain | 71 (32) | 68 (30) | 0.348 | Exercise = Control |
| General health | 79 (23) | 72 (19) | < 0.01 | Exercise > Control |
| Vitality | 81 (17) | 73 (25) | < 0.01 | Exercise > Control |
| Social functioning | 75 (22) | 74 (19) | 0.902 | Exercise = Control |
| Role‐emotional | 65 (34) | 65 (33) | 0.976 | Exercise = Control |
| Mental health | 72 (23) | 71 (23) | 0.825 | Exercise = Control |
| Physical health score | 79 (29) | 73 (29) | < 0.01 | Exercise > Control |
| Mental health score | 73 (28) | 71 (27) | 0.102 | Exercise = Control |
| Heller 1993 | ||||
| QLMI at 6 months' follow‐up: | ||||
| Emotional | 5.4 (1.1) | 5.2 (1.2) | 0.04 | Exercise > Control |
| Physical | 5.4 (1.2) | 5.2 (1.3) | 0.17* | Exercise = Control |
| Social | 5.9 (1.1) | 5.8 (1.1) | 0.35* | Exercise = Control |
| Hofman‐Bang 1999 | ||||
| AP‐QLQ at 12 months' follow‐up: | ||||
| Physical activity | 4.9 | 4.3 | < 0.05 | Exercise > Control |
| Somatic symptoms | NR | NR | NS | Exercise = Control |
| Emotional distress | NR | NR | NS | Exercise = Control |
| Life satisfaction | NR | NR | NS | Exercise = Control |
| Houle 2012 | ||||
| Quality of Life Index ‐ cardiac version III at 6 months' follow‐up: | ||||
| Health and functional score | 26 (5.1) | 24.5 (5.3) | 0.048 | Exercise > Control |
| Psychological/spiritual score | 25.6 (5.8) | 25.5 (3.8) | 0.383 | Exercise = Control |
| Social and economic score | 25.7 (3.8) | 25.4 (4.7) | 0.392 | Exercise = Control |
| Family score | 28.1 (2.5) | 26.7 (4.3) | 0.048 | Exercise > Control |
| Overall | 26.2 (4.3) | 25.8 (4.1) | 0.057 | Exercise = Control |
| Quality of Life Index ‐ cardiac version III at 12 months' follow‐up: | ||||
| Health and functional score | 27.8 (2.0) | 25.3 (4.6) | 0.036 | Exercise > Control |
| Psychological/spiritual score | 27.4 (2.5) | 26.2 (4.0) | 0.336 | Exercise = Control |
| Social and economic score | 27.2 (3.0) | 25.9 (5.2) | 0.638 | Exercise = Control |
| Family score | 28 (2.6) | 26.8 (5.0) | 0.092 | Exercise = Control |
| Overall | 27.7 (2.1) | 25.7 (4.2) | 0.048 | Exercise > Control |
| Ma 2020 | ||||
| SF‐12 change at 12 months' follow‐up: | ||||
|
| Δ (SD) | Δ (SD) |
|
|
| Physical component | 13.3 (6) | 9.9 (5.9) | < 0.001 | Exercise > Control |
| Mental component | 12.4 (5.4) | 9 (6.2) | < 0.001 | Exercise > Control |
| Maddison 2014 | ||||
| EQ‐5D at 24 weeks' follow‐up: | ||||
|
| 0.86 | 0.83 | 0.23 | Exercise = Control |
| SF‐36 at 24 weeks' follow‐up: |
|
|
| |
| Physical functioning | 52.9 | 51.9 | 0.20 | Exercise = Control |
| Role physical | 52.6 | 50.8 | 0.08 | Exercise = Control |
| Bodily pain | 52.4 | 51.9 | 0.71 | Exercise = Control |
| General health | 55.3 | 53.2 | 0.03 | Exercise > Control |
| Vitality | 55.7 | 55.9 | 0.79 | Exercise = Control |
| Social Functioning | 53.3 | 52.4 | 0.42 | Exercise = Control |
| Role emotional | 51.4 | 51.6 | 0.81 | Exercise = Control |
| Mental health | 54.6 | 54.0 | 0.61 | Exercise = Control |
| Mutwalli 2012 | ||||
| SF‐36 Health status score at 6 months' follow‐up: | ||||
|
| 90.14 (4.83) | 60.55 (16.21) | 0.000 | Exercise > Control |
| Oerkild 2012 | ||||
| SF‐36 at 12 months' follow‐up: | ||||
|
| Δ (95% CI) | Δ (95% CI) |
|
|
| SF 12 PCS | ‐1.1 (‐5.3 to 3.1) | ‐1.4 (‐5.2 to 2.3) | NS* | Exercise = Control |
| SF 12 MCS | ‐1.4 (‐6.1 to 3.3) | ‐0.3 (‐4.6 to 4.0) | NS* | Exercise = Control |
| Oldridge 1991 | ||||
| QLMI at 4 months' follow‐up: | ||||
| Limitations | 54 | 54 | NS | Exercise = Control |
| Emotions | 103 | 101 | NS | Exercise = Control |
| QLMI at 8 months' follow‐up: | ||||
| Limitations | 54 | 54 | NS | Exercise = Control |
| Emotions | 103 | 103 | NS | Exercise = Control |
| QLMI at 12 months' follow‐up: | ||||
| Limitations | 54 | 55 | NS | Exercise = Control |
| Emotions | 105 | 102 | NS | Exercise = Control |
| Reid 2012 | ||||
| MacNew at 6 months' follow‐up: | ||||
| Global score | 5.8 (0.6) | 5.6 (0.8) | 0.112 | Exercise = Control |
| Emotional subscale | 5.6 (0.6) | 5.4 (0.7) | 0.038 | Exercise > Control |
| Social subscale | 6.3 (0.8) | 6.0 (1.0) | 0.162 | Exercise = Control |
| Physical subscale | 6.0 (0.8) | 5.8 (1.0) | 0.031 | Exercise > Control |
| Sandstrom 2005 | ||||
| Time Trade Off (TTO) at 12 months' follow‐up: | ||||
|
| 0.86 (0.23) | 0.85 (0.21) | NS* | Exercise = Control |
| EuroQol Part one at 12 months' follow‐up: | ||||
|
| 0.87 (0.15) | 0.86 (0.16) | NS* | Exercise = Control |
| EuroQol Part two at 12 months' follow‐up: |
|
| ||
|
| 7.6 (1.46) | 7.43 (1.46) | NS* | Exercise = Control |
| Santaularia 2017 | ||||
| EuroQol‐5D at 12 months' follow‐up: | ||||
|
| N (%) | N(%) |
|
|
| Mobility | 33 (84.6) | 33 (75) | 0.019 | Exercise > Control |
| Self‐care | 38 (97.4) | 43 (97.7) | 0.172 | Exercise = Control |
| Usual activities | 32 (82) | 31 (70.5) | 0.803 | Exercise = Control |
| Pain/discomfort | 28 (71.8) | 26 (59.1) | 0.528 | Exercise = Control |
| Anxiety/depression | 22 (56.4) | 26 (59.1) | 0.429 | Exercise = Control |
| Snoek 2020 | ||||
| SF‐36 summary scores at 6 months: | ||||
| Physical | 50.2 (7.2) | 48.3 (7.5) | 0.086* | Exercise = Control |
| Mental | 54.0 (8.4) | 52.7 (9.1) | 0.322* | Exercise = Control |
| SF‐36 summary scores at 12 months: | ||||
| Physical | 50.6 (7.2) | 49 (8.2) | 0.167* | Exercise = Control |
| Mental | 53.2 (8.8) | 52.5 (9.2) | 0.604* | Exercise = Control |
| Stahle 1999 | ||||
| Karolinska Questionnaire at 12 months' follow‐up: | ||||
| Chest pain | 0.6 (1.2) | 0.4 (1.3) | NS | Exercise = Control |
| Shortness of breath | 0.4 (1.1) | 0.2 (1.0) | NS | Exercise = Control |
| Dizziness | ‐0.1 (1.1) | 0.2 (0.9) | NS | Exercise = Control |
| Palpitation | ‐0.1 (1.0) | 0.1 (0.9) | NS | Exercise = Control |
| Cognitive ability | ‐0.1 (0.6) | 0.0 (0.7) | NS | Exercise = Control |
| Alertness | 0.0 (0.9) | 0.1 (0.8) | NS | Exercise = Control |
| Quality of sleep | 0.0 (0.5) | 0.1 (0.5) | NS | Exercise = Control |
| Physical ability | 0.2 (0.7) | 0.1 (0.4) | NS | Exercise = Control |
| Daily activity | 0.3 (0.5) | 0.1 (0.5) | NS | Exercise = Control |
| Depression | 0.1 (0.3) | 0.1 (0.2) | NS | Exercise = Control |
| Self‐perceived health | 0.5 (1.3) | 0.3 (1.0) | NS | Exercise = Control |
| "Ladder of Life" present | 1.2 (1.2) | 0.9 (1.8) | NS | Exercise = Control |
| "Ladder of Life" future | 0.8 (2.7) | 0.4 (2.3) | NS | Exercise = Control |
| Fitness | 0.6 (1.4) | 0.4 (1.0) | NS | Exercise = Control |
| Physical ability | 0.7 (1.0) | 0.4 (1.1) | NS | Exercise = Control |
| Toobert 2000 | ||||
| SF‐36 at 24 months' follow‐up: | ||||
| Physical functioning | NR | NR | NS | Exercise = Control |
| Physical performance | NR | NR | NS | Exercise = Control |
| Bodily pain | NR | NR | NS | Exercise = Control |
| General health | NR | NR | < 0.05 | Exercise > Control |
| Vitality | NR | NR | NS | Exercise = Control |
| Social functioning | NR | NR | < 0.05 | Exercise > Control |
| Emotional performance | NR | NR | NS | Exercise = Control |
| Mental health | NR | NR | NS | Exercise = Control |
| Uddin 2020 | ||||
| WHOQoL‐BREF at 12 months' follow‐up | ||||
| Overall perception of HRQoL | 4.03 (0.49) | 3.2 (0.82) | < 0.01 | Exercise > Control |
| Overall perception of health | 4.06 (0.4) | 3.17 (0.38) | < 0.01 | Exercise > Control |
| Physical domain | 26.9 (2.88) | 21.17 (3.35) | < 0.01 | Exercise > Control |
| Psychological domain | 23.42 (2.84) | 17.87 (3.19) | < 0.01 | Exercise > Control |
| Social relationship domain | 11.83 (1.5) | 10.75 (0.89) | < 0.01 | Exercise > Control |
| Environmental domain | 28.8 (4.24) | 21.77 (5.31) | 0.03 | Exercise > Control |
| Wang 2012 | ||||
| SF‐36 at 6 months' follow‐up: | ||||
| Physical functioning | 80.8 (13.7) | 73.2 (13.0) | < 0.001 | Exercise > Control |
| Physical performance | 68.2 (17.3) | 56.2 (46.8) | 0.015 | Exercise > Control |
| Bodily pain | 68.2 (17.3) | 63.5 (14.6) | 0.012 | Exercise > Control |
| General health | 57.4 (20.3) | 49.0 (16.2) | 0.017 | Exercise > Control |
| Vitality | 66.3 (17.3) | 56.4 (21.7) | 0.002 | Exercise > Control |
| Social functioning | 71.3 (21.4) | 65.8 (18.0) | 0.031 | Exercise > Control |
| Emotional performance | 80.8 (37.9) | 75.9 (39.7) | 0.12 | Exercise = Control |
| Mental health | 73.5 (17.1) | 65.4 (20.7) | 0.011 | Exercise > Control |
| MIDAS at 6 months' follow‐up: | ||||
| Physical Activity | 37.7 (11.2) | 42.6 (12.3) | < 0.001 | Exercise > Control |
| Insecurity | 28.7 (9.7) | 33.4 (13.8) | < 0.001 | Exercise > Control |
| Emotional reaction | 30.4 (12.8) | 34.8 (14.4) | 0.008 | Exercise > Control |
| Dependency | 27.6 (9.4) | 31.8 (16.6) | 0.001 | Exercise > Control |
| Diet | 36.8 (15.4) | 43.6 (20.7) | 0.40 | Exercise = Control |
| Concerns over meds | 29.4 (12.6) | 37.7 (18.0) | <0.001 | Exercise > Control |
| Side Effects | 28.2 (11.1) | 30.8 (14.3) | 0.30 | Exercise > Control |
| West 2012 | ||||
| SF‐36 at 12 months' follow‐up: | ||||
| Physical function | 65 (29) | 64 (30) | NS* | Exercise = Control |
| Role physical | 69 (31) | 67 (33) | NS* | Exercise = Control |
| Role emotional | 85 (23) | 85 (25) | NS* | Exercise = Control |
| Social function | 81 (28) | 79 (29) | NS* | Exercise = Control |
| Mental health | 76 (13) | 76 (13) | NS* | Exercise = Control |
| Energy /vitality | 65 (24) | 65 (24) | NS* | Exercise = Control |
| Pain | 69 (28) | 68 (29) | NS* | Exercise = Control |
| Health Perception | 58 (25) | 57 (25) | NS* | Exercise = Control |
| Yu 2003 | ||||
| SF‐36 at 8 months' follow‐up: | ||||
| Physical functioning | 88 (12) | 82 (17) | 0.03* | Exercise > Control |
| Physical performance | 75 (33) | 66 (35) | 0.18* | Exercise = Control |
| Bodily pain | 80 (25) | 80 (25) | 1.00* | Exercise = Control |
| General health | 64 (26) | 60 (28) | 0.45* | Exercise = Control |
| Vitality | 79 (18) | 65 (17) | 0.0001 | Exercise > Control |
| Social functioning | 89 (27) | 82 (28) | 0.15 | Exercise = Control |
| Emotional performance | 93 (18) | 83 (35) | 0.05 | Exercise = Control |
| Mental health | 84 (16) | 80 (15) | 0.2 | Exercise = Control |
| SF‐36 at 24 months' follow‐up: | ||||
| Physical functioning | 88 (13) | 87 (9) | 0.67* | Exercise = Control |
| Physical performance | 80 (32) | 79 (30) | 0.87* | Exercise = Control |
| Bodily pain | 81 (21) | 85 (20) | 0.33* | Exercise = Control |
| General health | 64 (20) | 61 (18) | 0.43* | Exercise = Control |
| Vitality | 73 (21) | 73 (17) | 1.00* | Exercise = Control |
| Social functioning | 79 (30) | 90 (18) | 0.04* | Exercise > Control |
| Emotional performance | 89 (25) | 93 (25) | 0.42* | Exercise = Control |
| Mental health | 85 (14) | 85 (12) | 1.00* | Exercise = Control |
| Zwisler 2008 | ||||
| SF‐36 at 12 months' follow‐up: | ||||
| Physical Component Score | 45.2 (9.8) | 46.4 (9.8) | 0.39* | Exercise = Control |
| Mental Component Score | 50.6 (10.8) | 48.4 (11.5) | 0.16* | Exercise = Control |
AP‐QLQ: Angina Pectoris‐Quality of Life questionnaire
BMI: body mass index
EQ‐5D: five‐dimension EuroQol scale
EuroQoL: European Quality of Life Scale
IQR: interquartile range
MIDAS: Myocardial Infarction Dimensional Assessment Scale
NR: not reported
NS: not significant
QLMI: Quality of Life After Myocardial Infarction questionnaire
SD: standard deviation
SF‐36: Short Form 36‐item questionnaire
WHOQoL‐BREF: World Health Organization Quality of Life abbreviated instrument
* Calculated by authors of this report based on independent two group t test.
** Adjusted for baseline difference between groups.
Exercise = Control: no statistically significant difference (P > 0.05) between exercise and Control groups at follow up
Exercise > Control: statistically significant difference (P < 0.05) between exercise and Control groups at follow up
NS*: The authors of this review have inferred a P value of > 0.05 based either on the 95% CI, or from narrative in the paper, rather than from directly observing the P‐value.
Costs and cost‐effectiveness
Eight of the included studies reported data on costs of CR and overall healthcare costs in both groups (Briffa 2005; Hambrecht 2004; Hautala 2017; Kovoor 2006; Maddison 2014; Marchionni 2003; Oldridge 1991/Oldridge 1993; Yu 2004). These results are summarised in Table 2. While it was not possible to directly compare costs across studies due to differences in currencies and the timing of studies, it is possible to compare the within‐study costs for CR and control groups. Three studies showed no difference in total healthcare costs between groups (Briffa 2005; Kovoor 2006; Yu 2004); two studies found healthcare costs for CR to be lower (USD 2378 less per participant; EUR 1083 less per participant) compared to control (Hambrecht 2004; Hautala 2017); one study reported the healthcare costs for CR to be higher (USD 4839 more per participant) than usual care (Marchionni 2003); while two studies did not report total healthcare costs (Maddison 2014; Oldridge 1991/Oldridge 1993).
| Author/ year | Briffa 2005 | Hambrecht 2004 | Hautala 2017 | Kovoor 2006/Hall 2002 | Maddison 2014 | Marchionni 2003 | Oldridge 1991/93 | Yu 2004 |
| Follow‐up (months) | 12 | 12 | 12 | 12 | 6 | 14 | 12 | 24 |
| Year of costs (currency) | 1998 (Australian dollars ‐ AUD) | NR (US dollars ‐ USD) | NR (euros ‐ EUR) | 1999 (Australian dollars ‐ AUD) | NR (euros ‐ EUR) | 2000 (US dollars ‐ USD) | 1991 (US dollars ‐ USD) | 2003 (US dollars ‐ USD) |
| Cost of rehabilitation | ||||||||
| Mean cost/patient | AUD 694 | NR | EUR 299 | AUD 394 | EUR 127 | USD 5246 | USD 670 | NR |
| Costs considered | Details of costed elements not provided | NR | Estimated according to the average monthly fees in Finnish gyms where individual guidance in exercise training is led by a health care professional | staff, assessments, counselling, education, patient travel | NR | NR | space, equipment, staff, literature resources, operating costs, parking, patients costs | NR |
| Total healthcare costs | ||||||||
| Rehabilitation mean cost/patient | AUD 4937 | USD 3708 ± 156 | EUR 1944 | NR | NR | USD 17,272 | NR | USD 15,292 |
| Usual care mean cost/patient | AUD 4541 | USD 6086 ± 370 | EUR 3027 | NR | NR | USD 12,433 | NR | USD 15,707 |
| Absolute difference in mean cost/patient* | AUD 395 | USD ‐2378 | EUR ‐1083 | NR | NR | USD 4839 | USD 480 | USD ‐415 |
| P value for cost difference | 0.74 | P < 0.001 | NR | P > 0.05 (see below) | NR | NR | NR | P > 0.05 |
| Additional healthcare costs considered | Hospitalisations, pharmaceuticals, tests, consultations, rehabilitation, patient expenses, ambulance | Rehospitalisations, revascularisation, cycle ergometers, training facilities, and supervising staff | Primary health care costs, secondary health care costs, occupational health care service costs | Phone calls (P = 0.10); hospital admissions (P = 0.11); gated heart pool scan (P = 0.50); exercise stress test (P = 0.72); other diagnostics (P = 0.37); visits to general practitioner (P = 0.61), specialist doctor (P = 0.35), or health‐care professional (P = 0.31) | NR | NR | Service utilisation, physician costs, emergency costs, in‐patient days, allied health, other rehabilitation visits | Hospitalisations; revascularisations; private clinic visit; cardiac clinic visits; public non‐cardiac visits; casualty visits; drugs |
| Cost‐effectiveness | ||||||||
| Rehabilitation mean health care benefits | Utility‐based quality of life – | NR | Average change in 15D utility: 0.013 | NR | NR | NR | NR | NR |
| Usual care mean health care benefit | Utility 0.010 (95% CI ‐0.001 to 0.022) | NR | Average change in 15D utility: ‐0.012 | NR | NR | NR | NR | NR |
| Incremental mean health care benefit | Utility 0.013 (95% CI, NR) P = 0.38; +0.009 QALYs | NR | 0.045 QALYs (0.023‐0.077) | NR | NR | NR | 0.052 QALYs (95% CI, 0.007 to 0.1) | 0.06 QALYs |
| Incremental cost effectiveness ratio/patient | AUD +42,535 per QALY. Extensive sensitivity analyses reported. | NR | EUR ‐24,511 per QALY | NR | EUR +15,247 per QALY | NR | USD +9200 per QALY | USD ‐650 per QALY |
NR: not reported
QALY: quality‐adjusted life year
* The currency for Hambrecht 2004 is not reported, but healthcare costs are reported within the paper with $
Five studies also reported cost‐effectiveness using a cost utility approach (i.e. cost per quality‐adjusted life year (QALY)) (Briffa 2005; Hautala 2017; Oldridge 1991/Oldridge 1993; Maddison 2014; Yu 2004). Two studies showed CR (compared to control) to be economically dominant; that is, associated with more QALYs and less overall costs (Hautala 2017; Yu 2004). In the remaining three studies, the incremental cost ratio compared to control was USD 42,535 per QALY (Briffa 2005), EUR 15,247 per QALY (Maddison 2014), and USD 9200 per QALY (Oldridge 1991/Oldridge 1993). Based on these analyses, authors consistently concluded CR to be a cost‐effective use of healthcare resources compared to usual care.
Meta‐regression
We examined predictors of total mortality, cardiovascular mortality, recurrent MI, revascularisation (CABG and PCI) and all‐cause hospitalisation across the longest follow‐up of each individual study, using univariate meta‐regression. We did not perform meta‐regression where there were fewer than 10 studies included in the analysis. No statistically significant associations were seen in any of the analyses (Table 3, Table 4, Table 5, Table 6, Table 7, Table 8).
| Explanatory variable (n trials) | Exp (slope)* | 95% confidence interval, | Proportion of variance explained (adjusted R2) | Interpretation |
| Case mix (% MI patients) (n = 46) | RR = 1.00 | 1.00 to 1.00, P = 0.15 | 56.1% | No evidence that risk ratio is associated with case mix |
| Dose of exercise (number of weeks of exercise training x average number of sessions/week x average min/session) (n = 33) | RR = 1.00 | 1.00 to 1.00, P = 0.11 |
100% | No evidence that risk ratio is associated with type of CR |
| Duration of follow‐up (months) (n = 47) | RR = 1.00 | 1.00 to 1.00, P = 0.07 | 100% | No evidence that risk ratio is associated with length of follow‐up |
| Type of CR (exercise only vs comprehensive CR) (n = 47) | RR = 1.04 | 0.84 to 1.31, P = 0.70 | ‐27.1% | No evidence that risk ratio is associated with type of CR |
| Year of publication (pre‐1995 vs post‐1995) (n = 47) | RR = 0.84 | 0.70 to 0.99, P = 0.04 | 100% | No evidence that risk ratio is associated with publication year |
| CR setting (n = 47) | RR = 0.95 | 0.82 to 1.24, P = 0.95 | ‐11.3% | No evidence that risk ratio is associated with type of CR |
| Risk of bias (low risk in ≤ 3 items vs > 3 items) (n = 47) | RR = 1.02 | 0.94 to 1.09, P = 0.67 | ‐68.55% | No evidence that risk ratio is associated with risk of bias |
| Study location (continent ‐ Europe, North America, Australia/Asia or Other) (n = 47) | RR = 1.01 | 0.86 to 1.19, P = 0.93 | ‐41.24% | No evidence that risk ratio is associated with study location |
| Low‐ and middle‐income country (LMIC) vs high‐income country (n = 47) | RR = 1.02 | 0.70 to 1.48, P = 0.93 | ‐45.10% | No evidence that risk ratio is associated with LMIC |
| Sample size (≤ 150 vs > 150) (n = 47) | RR = 1.19 | 0.73 to 1.93, P = 0.47 | 16.07% | No evidence that risk ratio is associated with study sample size |
| Explanatory variable (n trials) | Exp (slope)* | 95% confidence interval, | Proportion of variance explained (adjusted R2) | Interpretation |
| Case mix (% MI patients) (n = 27) | RR = 1.00 | 0.99 to 1.01, P = 0.76 | ‐8.74% | No evidence that risk ratio is associated with case mix |
| Dose of exercise (number of weeks of exercise training x average number of sessions/week x average min/session) (n = 22) | RR = 1.00 | 1.00 to 1.00, P = 0.62 | 0% | No evidence that risk ratio is associated with dose of exercise |
| Duration of follow‐up (months) (n = 28) | RR = 0.99 | 0.99 to 1.00, P = 0.05 | 90.36% | No evidence that risk ratio is associated with length of follow‐up |
| Type of CR (exercise only vs comprehensive CR) (n = 28) | RR = 0.83 | 0.62 to 1.10, P = 0.18 | 75.69% | No evidence that risk ratio is associated with type of CR |
| Year of publication (pre‐1995 vs post‐1995) (n = 28) | RR = 1.37 | 0.89 to 2.13, P = 0.15 | 63.31% | No evidence that risk ratio is associated with publication year |
| Setting (centre vs home) (n = 28) | RR = 1.05 | 0.88 to 1.24, P = 0.61 | ‐29.66% | No evidence that risk ratio is associated with setting of CR |
| Risk of bias (low risk in ≤ 3 items vs > 3 items) (n = 28) | RR = 0.90 | 0.73 to 1.11, P = 0.30 | 85.73% | No evidence that risk ratio is associated with risk of bias |
| Study location (continent ‐ Europe, North America, Australia/Asia or Other) (n = 28) | RR = 1.02 | 0.75 to 1.39, P = 0.89 | ‐41.75% | No evidence that risk ratio is associated with study location |
| Low‐ and middle‐income country (LMIC) vs high‐income country (n = 28) | RR = 0.69 | 0.22 to 2.19, P = 0.52 | 9.36% | No evidence that risk ratio is associated with LMIC |
| Sample size (≤ 150 vs > 150) (n = 28) | RR = 1.28 | 0.69 to 2.37, P = 0.42 | 28.43% | No evidence that risk ratio is associated with study sample size |
| Explanatory variable (n trials) | Exp (slope)* | 95% confidence interval, | Proportion of variance explained (adjusted R2) | Interpretation |
| Case mix (% MI patients) (n = 41) | RR = 1.00 | 0.99 to 1.01, P = 0.93 | ‐4.57% | No evidence that risk ratio is associated with case mix |
| Dose of exercise (number of weeks of exercise training x average number of sessions/week x average min/session) (n = 33) | RR = 1.00 | 1.00 to 1.00, P = 0.68 | 0% | No evidence that risk ratio is associated with dose of exercise |
| Duration of follow‐up (months) (n = 41) | RR = 1.00 | 0.99 to 1.01, P = 0.97 | ‐12.45% | No evidence that risk ratio is associated with length of follow‐up |
| Type of CR (exercise only vs comprehensive CR) (n = 41) | RR = 0.85 | 0.58 to 1.25, P = 0.39 | 9.68% | No evidence that risk ratio is associated with type of CR |
| Year of publication (pre‐1995 vs post‐1995) (n = 41) | RR = 1.36 | 0.94 to 1.97, P = 0.11 | 25.40% | No evidence that risk ratio is associated with publication year |
| Setting (centre vs home) (n = 39) | RR = 0.80 | 0.67 to 0.95, P = 0.01 | 67.62% | No evidence that risk ratio is associated with setting of CR |
| Risk of bias (low risk in ≤ 3 items vs > 3 items) (n = 41) | RR=1.39 | 0.85 to 2.26, P = 0.18 | ‐16.70% | No evidence that risk ratio is associated with risk of bias |
| Study location (continent ‐ Europe, North America, Australia/Asia or Other) (n = 41) | RR = 0.71 | 0.49 to 1.05, P = 0.09 | 12.94% | No evidence that risk ratio is associated with study location |
| Low‐ and middle‐income country (LMIC) vs high income country (n = 41) | RR = 0.65 | 0.33 to 1.61, P = 0.20 | 0.86% | No evidence that risk ratio is associated with LMIC |
| Sample size (≤ 150 vs > 150) (n = 41) | RR = 1.69 | 1.05 to 2.72, P = 0.03 | 54.95% | No evidence that risk ratio is associated with study sample size |
| Explanatory variable (n trials) | Exp (slope)* | 95% confidence interval, | Proportion of variance explained (adjusted R2) | Interpretation |
| Case mix (% MI patients) (n = 31) | RR = 1.01 | 1.00 to 1.02, P = 0.05 | 0% | No evidence that risk ratio is associated with case mix |
| Dose of exercise (number of weeks of exercise training x average number of sessions/week x average min/session) (n = 25) | RR = 1.00 | 1.00 to 1.00, P = 0.78 | 0% | No evidence that risk ratio is associated with dose of exercise |
| Duration of follow‐up (months) (n = 31) | RR = 1.00 | 0.99 to 1.01, P = 0.75 | 0% | No evidence that risk ratio is associated with length of follow‐up |
| Type of CR (exercise only vs comprehensive CR) (n = 31) | RR = 1.04 | 0.67 to 1.61, P = 0.86 | 0% | No evidence that risk ratio is associated with type of CR |
| Year of publication (pre‐1995 vs post‐1995) (n = 31) | RR = 0.88 | 0.56 to 1.41, P = 0.59 | 0% | No evidence that risk ratio is associated with publication year |
| Setting (centre vs home) (n = 31) | RR = 1.07 | 0.87 to 1.33, P = 0.51 | 0% | No evidence that risk ratio is associated with setting of CR |
| Risk of bias (low risk in ≤ 3 items vs > 3 items) (n = 31) | RR = 0.94 | 0.64 to 1.38, P = 0.73 | 0% | No evidence that risk ratio is associated with risk of bias |
| Study location (continent ‐ Europe, North America, Australia/Asia or Other) (n = 31) | RR = 1.19 | 0.83 to 1.71, P = 0.34 | 0% | No evidence that risk ratio is associated with study location |
| Low‐ and middle‐income country (LMIC) vs high income country (n = 31) | RR = 0.51 | 0.08 to 3.18, P = 0.46 | 0% | No evidence that risk ratio is associated with LMIC |
| Sample size (≤ 150 vs > 150) (n = 31) | RR = 1.31 | 0.82 to 2.09, P = 0.25 | 0% | No evidence that risk ratio is associated with study sample size |
| Explanatory variable (n trials) | Exp (slope)* | 95% confidence interval, | Proportion of variance explained (adjusted R2) | Interpretation |
| Case mix (% MI patients) (n = 18) | RR = 1.00 | 1.00 to 1.01, P = 0.50 | 0% | No evidence that risk ratio is associated with case mix |
| Dose of exercise (number of weeks of exercise training x average number of sessions/week x average min/session) (n = 16) | RR = 1.00 | 1.00 to 1.00, P = 0.50 | 0% | No evidence that risk ratio is associated with dose of exercise |
| Duration of follow‐up (months) (n = 18) | RR = 1.00 | 0.99 to 1.01, P = 0.82 | 0% | No evidence that risk ratio is associated with length of follow‐up |
| Type of CR (exercise only vs comprehensive CR) (n = 18) | RR = 0.78 | 0.38 to 1.59, P = 0.47 | 0% | No evidence that risk ratio is associated with type of CR |
| Year of publication (pre‐1995 vs post‐1995) (n = 18) | RR = 0.95 | 0.46 to 1.95, P = 0.87 | 0% | No evidence that risk ratio is associated with publication year |
| Setting (centre vs home) (n = 18) | RR = 0.91 | 0.72 to 1.15, P = 0.41 | 0% | No evidence that risk ratio is associated with setting of CR |
| Risk of bias (low risk in ≤ 3 items vs > 3 items) (n = 18) | RR = 1.09 | 0.72 to 1.66, P = 0.67 | 0% | No evidence that risk ratio is associated with risk of bias |
| Study location (continent ‐ Europe, North America, Australia/Asia or Other) (n = 18) | RR = 0.81 | 0.53 to 1.23, P = 0.30 | 0% | No evidence that risk ratio is associated with study location |
| Low‐ and middle‐income country (LMIC) vs high income country (n = 18) | RR = 0.29 | 0.05 to 1.63, P = 0.15 | 0% | No evidence that risk ratio is associated with LMIC |
| Sample size (≤ 150 vs > 150) (n = 18) | RR = 1.19 | 0.70 to 2.01, P = 0.49 | 0% | No evidence that risk ratio is associated with study sample size |
| Explanatory variable (n trials) | Exp (slope)* | 95% confidence interval, | Proportion of variance explained (adjusted R2) | Interpretation |
| Case mix (% MI patients) (n = 23) | RR = 1.00 | 1.00 to 1.01, P = 0.71 | ‐20.91% | No evidence that risk ratio is associated with case mix |
| Dose of exercise (number of weeks of exercise training x average number of sessions/week x average min/session) (n = 19) | RR = 1.00 | 1.00 to 1.00, P = 0.44 | ‐69.78% | No evidence that risk ratio is associated with dose of exercise |
| Duration of follow‐up (months) (n = 23) | RR = 1.01 | 1.00 to 1.01, P = 0.07 | 56.52% | No evidence that risk ratio is associated with length of follow‐up |
| Type of CR (exercise only vs comprehensive CR) (n = 23) | RR = 0.93 | 0.65 to 1.33, P = 0.70 | ‐50.20% | No evidence that risk ratio is associated with type of CR |
| Year of publication (pre‐1995 vs post‐1995) (n = 23) | RR = 1.12 | 0.80 to 1.57, P = 0.48 | ‐32.69% | No evidence that risk ratio is associated with publication year |
| Setting (centre vs home) (n = 23) | RR = 0.94 | 0.83 to 1.06, P = 0.28 | ‐36.70% | No evidence that risk ratio is associated with setting of CR |
| Risk of bias (low risk in ≤ 3 items vs > 3 items) (n = 23) | RR = 1.00 | 0.71 to 1.40, P = 0.99 | ‐44.14% | No evidence that risk ratio is associated with risk of bias |
| Study location (continent ‐ Europe, North America, Australia/Asia or Other) (n = 23) | RR = 0.86 | 0.69 to 1.08, P = 0.18 | ‐137.18% | No evidence that risk ratio is associated with study location |
| Low‐ and middle‐income country (LMIC) vs high income country (n = 23) | RR = 1.06 | 0.72 to 1.55, P = 0.76 | ‐49.12% | No evidence that risk ratio is associated with LMIC |
| Sample size (≤ 150 vs > 150) (n = 19) | RR = 1.45 | 1.08 to 1.96, P = 0.02 | 100% | No evidence that risk ratio is associated with study sample size |
Discussion
Summary of main results
Exercise‐based CR provides important benefits up to 12 months' follow‐up, including a large reduction in fatal or non‐fatal MI, and likely reductions in all‐cause mortality and all‐cause hospital admissions. There was evidence that exercise‐based CR results in little to no difference in CABG, and likely results in little to no difference in cardiovascular mortality and PCI. Evidence was uncertain whether exercise‐based CR may result in reduced risk of cardiovascular hospital admissions. Imprecision (wide 95% CI), publication bias and statistical heterogeneity led to downgrading the certainty of these outcomes up to 12 months' follow‐up.
At medium‐term follow‐up ( > 12 to 36 months), although there may be little to no difference in all‐cause mortality, MI, PCI, CABG and all‐cause hospitalisation with exercise‐based CR, a large reduction in cardiovascular mortality was found. The evidence was uncertain for difference in risk of cardiovascular hospitalisation.
At long‐term follow‐up ( > 3 years), evidence suggests that exercise‐based CR may result in little to no difference in all‐cause mortality, but may result in a large reduction in risks of cardiovascular mortality and MI. The evidence was uncertain for difference in risk of CABG and PCI.
Univariate meta‐regression analysis showed that the impact of exercise‐based CR on clinical events appears to be largely consistent across trials, irrespective of case mix (% of post‐MI participants), type of rehabilitation (exercise‐only versus comprehensive), dose of exercise training (number of weeks of exercise training x average number of sessions/week x average minutes/session), duration of follow‐up (months), study location (continent ‐ Europe, North America, Australia/Asia or other, or LMIC versus HIC setting), year of publication (pre‐1995 versus post‐1995), risk of bias (low risk in ≤ 3 items versus > 3 items) or sample size ( ≤ 150 vs > 150).
We did not undertake meta‐analysis for all HRQoL outcomes, due to the range of outcome measures and methods of reporting. However, where meta‐analysis was possible, there was evidence of some small increases in HRQoL with exercise‐based CR compared with 'no exercise' control, across several SF‐36 subscales (mental component, physical functioning, physical performance, general health, vitality, social functioning and mental health scores). However, these may not be clinically important differences. These findings were supported by a vote‐counting approach to summarise HRQoL results across all studies reporting HRQoL, in which 23/32 (72%) trials reported higher levels of HRQoL in one or more subscales with exercise‐based CR compared to control at follow‐up. Whilst this method of synthesis without meta‐analysis has significant limitations, such as not taking account of the differential weights given to each trial, we believe it to be the best available method to concisely and transparently summarise the results (Campbell 2020).
The five trial‐based economic evaluation studies showed exercise‐based CR to be a potentially cost‐effective use of resources in terms of gain in QALYs.
Overall completeness and applicability of evidence
The generalisability of early versions of this review was limited, as most included studies recruited predominantly male participants (Jolliffe 2001: 9% female; current version: 16% female), following MI (Jolliffe 2001: 80% trials with MI only participants; current version: 47% trials with MI only participants). However, with the inclusion of more women in trials conducted in the last decade and with further data on the outcomes of hospitalisation and HRQoL, the findings of this updated review potentially have greater external validity. An additional 16 new studies identified and included in this current update have been undertaken in low‐ and middle‐income countries, increasing the generalisability of our results to these countries where prevalence of CHD is high and continues to increase (Prabhakaran 2018).
Quality of the evidence
In previous versions of this review, the general lack of adequate reporting of randomisation and blinding methods in the included RCT reports made it difficult to assess their methodological quality. However, the quality of reporting in studies has increased over the last decade, and reassuringly, meta‐regression showed no significant association between the effect of CR compared to control and the level of risk of bias across trials.
GRADE demonstrated that the certainty of the evidence ranged from low to high across the primary outcomes. We downgraded the certainty of the evidence for all‐cause mortality, cardiovascular mortality, PCI and cardiovascular hospitalisations by one level for imprecision, due to wide confidence intervals that overlapped the boundary for no effect (i.e. 95% CI crossed 1). We downgraded the certainty of the evidence for MI and all‐cause hospitalisations by one level due to evidence of publication bias (Egger test: P < 0.05). We downgraded the certainty of the evidence for cardiovascular hospitalisations by one level due to evidence of substantial heterogeneity (Chi2 test: P < 0.05, or I2 test for heterogeneity > 50%, or both).
Potential biases in the review process
We believe this is the most comprehensive systematic review to date of RCT‐based evidence for the impact of exercise‐based CR for people with CHD. However, it is important that we contextualise our review findings in light of some limitations.
Details of random allocation sequence generation and concealment, and blinding of outcome assessment were poorly reported (33% trials adequately reported allocation sequence generation and 29% trials adequately reported blinding of outcome assessment), and therefore may be subject to bias. Funnel plot asymmetry for the risk of MI and all‐cause hospital admission is indicative of small‐study bias and possible publication bias. There was also evidence of statistical heterogeneity for all‐cause and cardiovascular hospitalisations, and all HRQoL subscales, except SF‐12 MCS.
The number of trials reporting medium‐term ( > 12 to 36 month follow‐up) and long‐term data ( > 36 months' follow‐up) has decreased from 47% (27/57 trials) to 21% (7/33 trials) over the last decade, while sample sizes have remained relatively small over the same period (median sample size increased from 125 to 142). As a result, the number of deaths and other clinical events, including hospitalisations, reported by many trials is small, or in some cases, zero. Indeed, in many studies, we located event data in the participant flow diagrams and descriptions of losses to follow‐up and exclusions, rather than as prespecified outcomes measured, reported, or analysed within trials. In addition, cause of death was often not reported. Furthermore, in recent studies, clinical events have often been reported as a composite endpoint (e.g. major adverse cardiac events) rather than as individual events. These data reporting and evidence syntheses issues may have resulted in some of the apparently paradoxical findings of this review, such as reduced all‐cause mortality but not cardiovascular mortality in the short term.
All included studies involved a 'no formal exercise training' intervention comparator. However, a wide range of comparators were seen across the trials, including education, psychological intervention or usual medical care alone. Due to poor and inconsistent reporting of adherence and fidelity to exercise programmes in the RCTs, we were not able to consider the actual amount of exercise that participants received or performed in this review.
Agreements and disagreements with other studies or reviews
The findings of this updated review are largely in accord with the previous version of this review. Although there was a trend towards a slight reduction in all‐cause mortality with exercise‐based CR compared to 'no exercise' control, this reduction failed to reach statistical significance. This is likely explained by the inclusion of more recent trials conducted in the era of optimal medical therapy. Given the proven survival advantage of contemporary medical treatments, and the limited opportunity for mortality gain in this patient cohort, any incremental mortality benefit with exercise is likely to be small. This theory is supported by Powell and colleagues' meta‐analysis of contemporary trials (Powell 2018), demonstrating no improvement in all‐cause mortality across 19 trials (risk difference (RD) 0.00, 95% CI ‐0.02 to 0.01; P = 0.38), or 9 trials reporting cardiovascular mortality (RD ‐0.01, 95% CI ‐0.02 to 0.01; P = 0.25), published between 2000 and 2017. Our meta‐regression analysis showed a potential trend (RR 0.84, 95% CI 0.70 to 0.99; P = 0.04) suggesting all‐cause mortality could be somewhat associated with publication year. However, due to multiple testing, we cannot rule out that this finding was by chance, and did not meet the criteria for statistical significance once the Bonferroni correction was applied.
Our results are also somewhat consistent with the findings of a recently published comprehensive network meta‐analysis (Huang 2021). In this study, the authors found that comprehensive exercise‐based CR reduces the risk of all‐cause mortality, yet unlike our results, risks of PCI and CABG revascularisation were also reduced. Exercise‐only CR was found to reduce the risks of non‐fatal MI, cardiovascular mortality, and all‐cause and cardiovascular hospitalisation, but not the risk of all‐cause mortality or revascularisations compared to standard care. The authors also similarly found no strong evidence to differentiate the relative benefits of exercise‐based CR, whether delivered as an exercise‐only intervention or a comprehensive intervention.
McGregor and colleagues performed a meta‐analysis of exercise‐based CR based on HRQoL outcomes of people with CHD, including 15 short‐term (i.e. 1 to 6 months) and 9 medium‐term (i.e. 8 to 12 months) trials (McGregor 2020). Pooled HRQoL results were consistent with the present review, showing improvement with CR across a number HRQoL domain scores.
The recently updated meta‐analysis of the Cardiac Rehabilitation Outcome Study (CROSII), which included RCTs and prospective and retrospective cohort studies, reported a mortality benefit of CR in people with acute coronary syndrome and revascularisation, with an index event in 1995 or later (Salzwedel 2020). However, with inclusion of observational evidence, the prognostic benefit reported by the CROSII study is subject to selection bias and confounding.
PRISMA flow diagram of study selection process
Risk of bias summary: review authors' judgements about each risk of bias item for each included study
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: all‐cause mortality at 6 to 12 months' follow‐up
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: all‐cause mortality at > 36 months' follow‐up
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: all‐cause mortality at > 12 to 36 months' follow‐up
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.2: cardiovascular mortality at 6 to 12 months' follow‐up
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.3: myocardial infarction at 6 to 12 months' follow‐up
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: myocardial infarction at > 12 to 36 months' follow‐up
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: myocardial infarction at > 36 months' follow‐up
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: CABG at 6 to 12 months' follow‐up
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: PCI at 6 to 12 months' follow‐up
Funnel plot of comparison: exercise‐based rehabilitation versus usual care, outcome 1.1: all‐cause hospitalisation at 6 to 12 months' follow‐up
Comparison 1: Exercise‐based rehabilitation versus control, Outcome 1: All‐cause mortality
Comparison 1: Exercise‐based rehabilitation versus control, Outcome 2: Cardiovascular mortality
Comparison 1: Exercise‐based rehabilitation versus control, Outcome 3: Fatal and/or nonfatal MI
Comparison 1: Exercise‐based rehabilitation versus control, Outcome 4: CABG
Comparison 1: Exercise‐based rehabilitation versus control, Outcome 5: PCI
Comparison 1: Exercise‐based rehabilitation versus control, Outcome 6: All‐cause hospital admissions
Comparison 1: Exercise‐based rehabilitation versus control, Outcome 7: Cardiovascular hospital admissions
Comparison 1: Exercise‐based rehabilitation versus control, Outcome 8: HRQoL SF‐36 summary scores at 6 to 12 months follow up
Comparison 1: Exercise‐based rehabilitation versus control, Outcome 9: HRQoL SF‐36 8 domains at 6 to 12 months follow up
Comparison 1: Exercise‐based rehabilitation versus control, Outcome 10: HRQoL EQ‐5D at 6 to 12 months follow up
| Exercise‐based cardiac rehabilitation compared to 'no exercise' control for coronary heart disease | ||||||
| Patient or population: people with coronary heart disease | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with 'no exercise' control | Risk with exercise‐based cardiac rehabilitation | |||||
| All‐cause mortality | Study population | RR 0.87 | 8823 | ⊕⊕⊕⊝ | Exercise‐based cardiac rehabilitation likely results in a slight reduction in all‐cause mortality up to 12 months' follow‐up. 25 RCTs with 26 comparisons. 14 RCTs reported 0 events in both the intervention and control groups. | |
| 57 per 1000 | 50 per 1000 | |||||
| Cardiovascular mortality | Study population | RR 0.88 | 5360 | ⊕⊕⊕⊝ | Exercise‐based cardiac rehabilitation likely results in little to no difference in cardiovascular mortality up to 12 months' follow‐up. 5 RCTs reported 0 events in both the intervention and control groups. | |
| 45 per 1000 | 39 per 1000 | |||||
| Fatal and/or non‐fatal MI | Study population | RR 0.72 | 7423 | ⊕⊕⊕⊕ | Exercise‐based cardiac rehabilitation results in a large reduction in fatal and/or non‐fatal MI up to 12 months' follow‐up. 24 RCTs with 24 comparisons. 3 RCTs reported 0 events in both the intervention and control groups. NNTB 75 (95% CI 47 to 298) | |
| 48 per 1000 | 35 per 1000 | |||||
| Revascularisation ‐ CABG | Study population | RR 0.99 | 4473 | ⊕⊕⊕⊕ | Exercise‐based CR results in little to no difference in CABG revascularisation up to 12 months' follow‐up. 20 RCTs with 22 comparisons. 2 RCTs reported 0 events in both the intervention and control groups. | |
| 56 per 1000 | 56 per 1000 | |||||
| Revascularisation ‐ PCI | Study population | RR 0.86 | 3465 | ⊕⊕⊕⊝ | Exercise‐based CR likely results in little to no difference in risk of PCI revascularisation up to 12 months' follow‐up. 13 RCTs with 14 comparisons. 3 RCTs reported 0 events in both the intervention and control groups. | |
| 60 per 1000 | 52 per 1000 | |||||
| All‐cause hospital admissions | Study population | RR 0.58 | 2030 | ⊕⊕⊕⊝ | Exercise‐based cardiac rehabilitation likely results in a large reduction in all‐cause hospital admissions up to 12 months' follow‐up. 14 RCTs with 16 comparisons. One RCT reported 0 events in both the intervention and control group. NNTB 12 (95% CI 9 to 21) | |
| 214 per 1000 | 124 per 1000 | |||||
| Cardiovascular hospital admissions | Study population | RR 0.80 | 1087 | ⊕⊕⊝⊝ | We are uncertain about the effects of exercise‐based CR on cardiovascular hospitalisation, with a wide confidence interval including considerable benefit as well as harm. | |
| 78 per 1000 | 62 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| a95% CI is wide and overlaps no effect; therefore, downgraded by one level for imprecision. | ||||||
| Measure of HRQoL | Mean (SD) outcome values at follow‐up | P value | Difference between groups | |
|
| Exercise | Control |
|
|
| Aronov 2019 | ||||
| Quality of life questionnaire developed by authors (Aronov 2002) % change of mean score at 6 months | ||||
|
| Δ% | Δ% |
|
|
|
| 30.4 | “no change” | NR |
|
| Bell 1998 | ||||
| Nottingham Health Profile at 10.5 months' follow‐up: | ||||
| Energy | 17.6 (27.1) | 18.3 (29.8) | 0.87** | Exercise = Control |
| Pain | 2.8 (8.8) | 4.82 (11.9) | < 0.05 | Exercise > Control |
| Emotional reactions | 6.4 (17.0) | 12.2 (19.9) | < 0.001 | Exercise > Control |
| Sleep | 7.5 (18.4) | 20.5 (27.8) | < 0.001 | Exercise > Control |
| Social isolation | 2.3 (10.6) | 4.0 (13.3) | 0.37* | Exercise = Control |
| Physical mobility | 8.4 (11.1) | 8.9 (14.5) | 0.82** | Exercise = Control |
| Belardinelli 2001 | ||||
| SF‐36 at 6 months' follow‐up: | ||||
| Physical functioning | 78 (19) | 55 (20) | 0.001 | Exercise > Control |
| Physical performance | 75 (13) | 65 (14) | 0.01 | Exercise > Control |
| Bodily pain | 4 (9) | 22 (10) | 0.001 | Exercise > Control |
| General health | 68 (14) | 50 (19) | 0.001 | Exercise > Control |
| Vitality | NR | NR |
|
|
| Social functioning | 66 (10) | 69 (12) | 0.14* | Exercise = Control |
| Emotional performance | NR | NR |
|
|
| Mental health | 65 (12) | 48 (15) | 0.01 | Exercise > Control |
| SF‐36 at 12 months' follow‐up: | ||||
| Physical functioning | 82 (18) | 54 (20) | 0.001 | Exercise > Control |
| Physical performance | 76 (9) | 58 (14) | 0.01 | Exercise > Control |
| Bodily pain | 4 (9) | 32 (12) | 0.001 | Exercise > Control |
| General health | 70 (14) | 50 (18) | 0.001 | Exercise > Control |
| Vitality | NR | NR |
|
|
| Social functioning | 68 (11) | 68 (12) | 1.00* | Exercise = Control |
| Emotional performance | NR | NR |
|
|
| Mental health | 70 (14) | 45 (15) | 0.001 | Exercise > Control |
| Bettencourt 2005 | ||||
| SF‐36 at 1 year follow‐up: | ||||
| Physical functioning | 70 | 62 | NS* | Exercise = Control |
| Physical performance | 66 | 57 | NS* | Exercise = Control |
| Bodily pain | 73 | 65 | NS* | Exercise = Control |
| General health | 57 | 46 | < 0.02 | Exercise > Control |
| Vitality | 62 | 47 | < 0.02 | Exercise > Control |
| Social functioning | 73 | 66 | NS* | Exercise = Control |
| Emotional performance | 65 | 58 | NS* | Exercise = Control |
| Mental health | 87 | 75 | NS* | Exercise = Control |
| Mental component | 71 | 57 | 0.02 | Exercise > Control |
| Physical component | 63 | 57 | NS* | Exercise = Control |
| Briffa 2005 | ||||
| SF‐36 at 6 months' follow‐up: | ||||
|
| Δ (95% CI) | Δ (95% CI) |
|
|
| Physical functioning | 15.9 (‐8 to 23) | 7.1 (1 to 13) | NS* | Exercise = Control |
| Physical performance | 75 (0 to 100) | 75 (0 to 100) | NS* | Exercise = Control |
| Bodily pain | 26.6 (18 to 35) | 19.2 (11 to 27) | NS* | Exercise = Control |
| General health | 0.1 (‐6 to 6) | ‐0.6 (‐5 to 4) | NS* | Exercise = Control |
| Vitality | 7.1 (1 to 13) | 3.7 (‐2 to 9) | NS* | Exercise = Control |
| Social functioning | 19.6 (10 to 29) | 14.1 (7 to 21) | NS* | Exercise = Control |
| Emotional performance | 33.3 (0 to 100) | 33.3 (33 to 100) | NS* | Exercise = Control |
| Mental health | 0.5 (‐4 to 5) | 1.4 (‐3 to 5) | NS* | Exercise = Control |
| SF‐36 at 1 year follow‐up: | ||||
|
| Δ (95% CI) | Δ (95% CI) |
|
|
| Physical functioning | 17.6 (10 to 25) | 6.8 (‐1 to 14) | 0.04 | Exercise > Control |
| Physical performance | 100 (0 to 100) | 75 (12 to 30) | NS* | Exercise = Control |
| Bodily pain | 30.2 (23 to 37) | 20.9 (‐2 to 7) | NS* | Exercise = Control |
| General health | 2.7 (‐3 to 5) | 2.2 (‐2 to 7) | NS* | Exercise = Control |
| Vitality | 11.9 (6 to 18) | 6.9 (1 to 12) | NS* | Exercise = Control |
| Social functioning | 23.6 (14 to 33) | 16.4 (9 to 23) | NS* | Exercise = Control |
| Emotional performance | 33.3 (33 to 100) | 33.3 (33 to 100) | NS* | Exercise = Control |
| Mental health | 3.6 (‐1 to 9) | 3.9 (0 to 8) | NS* | Exercise = Control |
| Bubnova 2019 | ||||
| Quality of life questionnaire developed by authors (Aronov 2002) mean (SD) score after 12 months: | ||||
| Low rehabilitation potential subgroup | ‐4.9 (4.5) | ‐7.8 (3.1) | < 0.05 | Exercise > Control |
| Average rehabilitation potential subgroup | ‐5 (3.2) | ‐7.4 (4.3) | < 0.05 | Exercise > Control |
| High rehabilitation potential subgroup | ‐4.3 (3.9) | ‐5.6 (4.3) | < 0.05 | Exercise > Control |
| Bubnova 2020 | ||||
| Quality of life questionnaire developed by authors (Aronov 2002) mean (%) score change at 12 months: | ||||
|
| Δ (%) | Δ (%) |
|
|
| BMI < 30 kg/m2 group | 42 (6%) | 10 (2%) | <0.01 | Exercise > Control |
| BMI ≥ 30 kg/m2 group | 27 (5%) | 8 (2%) | <0.001 | Exercise > Control |
| Campo 2020 | ||||
| EuroQol at 6 months' follow‐up: | ||||
|
| Median (IQR) | Median (IQR) |
|
|
| VAS (visual analogue scale) | 80 (70‐90) | 70 (50‐80) | < 0.001 | Exercise > Control |
| 5 domains | N (%) | N (%) |
|
|
| Pain/discomfort: | 103 (89) | 89 (77) | 0.03 | Exercise > Control |
| Anxiety/depression: | 92 (79) | 67 (58) | 0.001 | Exercise > Control |
| Mobility: | 104 (90) | 80 (70) | < 0.001 | Exercise > Control |
| Self‐care: | 114 (98) | 87 (76) | 0.6 | Exercise = Control |
| Usual activities: | 101 (87) | 87 (76) | 0.04 | Exercise > Control |
| EuroQol at 12 months' follow‐up: | ||||
|
| Median (IQR) | Median (IQR) |
|
|
| VAS (visual analogue scale) | 75 (70‐87) | 65 (50‐80) | < 0.001 | Exercise > Control |
| 5 domains | N (%) | N (%) |
|
|
| Pain/discomfort: | 86 (77) | 72 (65) | 0.04 | Exercise > Control |
| Anxiety/depression: | 83 (74) | 58 (53) | 0.03 | Exercise > Control |
| Mobility: | 95 (85) | 74 (67) | 0.008 | Exercise > Control |
| Self‐care: | 101 (91) | 100 (91) | 0.8 | Exercise = Control |
| Usual activities: | 99 (88) | 80 (73) | 0.004 | Exercise > Control |
| Dorje 2019 | ||||
| SF‐12 at 6 months' follow‐up: | ||||
| Physical health score | 46.8 (6.9) | 45.2 (6.5) | 0.22** | Exercise = Control |
| Mental health score | 51.5 (9.3) | 50 (8.6) | 0.28** | Exercise = Control |
| Engblom 1992 | ||||
| Nottingham Health Profile at 5 years' follow‐up: | ||||
| Energy | 18 | 25 | 0.08 | Exercise = Control |
| Pain | 12 | 18 | 0.07 | Exercise = Control |
| Emotional reactions | 14 | 21 | 0.27 | Exercise = Control |
| Sleep | 24 | 29 | 0.42 | Exercise = Control |
| Social isolation | 7 | 9 | 0.42 | Exercise = Control |
| Physical mobility | 6 | 14 | 0.005 | Exercise > Control |
| Hassan 2016 | ||||
| SF‐36 8 domains at 12 months' follow‐up | ||||
| Physical functioning | 83.5 (6.5) | 76.7 (10.6) | 0.01 | Exercise > Control |
| Role limitations physical | 62.5 (23.4) | 50.8 (20.2) | 0.04 | Exercise > Control |
| Role limitations emotional | 61.1 (21.6) | 49.9 (19.1) | 0.04 | Exercise > Control |
| Energy/fatigue | 66 (11.1) | 57.7 (11.7) | 0.01 | Exercise > Control |
| Emotional well being | 69.5 (2.6) | 61.5 (7.5) | 0.000 | Exercise > Control |
| Social functioning | 67.5 (19) | 56.3 (16.3) | 0.02 | Exercise > Control |
| Pain | 79.6 (18.4) | 67.9 (15.9) | 0.01 | Exercise > Control |
| General health | 43 (7.9) | 38.5 (8.8) | 0.04 | Exercise > Control |
| Hautala 2017 | ||||
| 15D Quality of life measure at 6 months' follow‐up: | ||||
|
| 0.915 (0.07) | 0.876 (0.084) | 0.0004* | Exercise > Control |
| 15D Quality of life measure at 12 months' follow‐up: | ||||
|
| 0.922 (0.072) | 0.886 (0.088) | < 0.0015* | Exercise > Control |
| He 2020 | ||||
| SF‐36 at 12 months: | ||||
| Physical functioning | 85 (22) | 74 (19) | < 0.01 | Exercise > Control |
| Role‐physical | 80 (21) | 77 (22) | 0.362 | Exercise = Control |
| Bodily pain | 71 (32) | 68 (30) | 0.348 | Exercise = Control |
| General health | 79 (23) | 72 (19) | < 0.01 | Exercise > Control |
| Vitality | 81 (17) | 73 (25) | < 0.01 | Exercise > Control |
| Social functioning | 75 (22) | 74 (19) | 0.902 | Exercise = Control |
| Role‐emotional | 65 (34) | 65 (33) | 0.976 | Exercise = Control |
| Mental health | 72 (23) | 71 (23) | 0.825 | Exercise = Control |
| Physical health score | 79 (29) | 73 (29) | < 0.01 | Exercise > Control |
| Mental health score | 73 (28) | 71 (27) | 0.102 | Exercise = Control |
| Heller 1993 | ||||
| QLMI at 6 months' follow‐up: | ||||
| Emotional | 5.4 (1.1) | 5.2 (1.2) | 0.04 | Exercise > Control |
| Physical | 5.4 (1.2) | 5.2 (1.3) | 0.17* | Exercise = Control |
| Social | 5.9 (1.1) | 5.8 (1.1) | 0.35* | Exercise = Control |
| Hofman‐Bang 1999 | ||||
| AP‐QLQ at 12 months' follow‐up: | ||||
| Physical activity | 4.9 | 4.3 | < 0.05 | Exercise > Control |
| Somatic symptoms | NR | NR | NS | Exercise = Control |
| Emotional distress | NR | NR | NS | Exercise = Control |
| Life satisfaction | NR | NR | NS | Exercise = Control |
| Houle 2012 | ||||
| Quality of Life Index ‐ cardiac version III at 6 months' follow‐up: | ||||
| Health and functional score | 26 (5.1) | 24.5 (5.3) | 0.048 | Exercise > Control |
| Psychological/spiritual score | 25.6 (5.8) | 25.5 (3.8) | 0.383 | Exercise = Control |
| Social and economic score | 25.7 (3.8) | 25.4 (4.7) | 0.392 | Exercise = Control |
| Family score | 28.1 (2.5) | 26.7 (4.3) | 0.048 | Exercise > Control |
| Overall | 26.2 (4.3) | 25.8 (4.1) | 0.057 | Exercise = Control |
| Quality of Life Index ‐ cardiac version III at 12 months' follow‐up: | ||||
| Health and functional score | 27.8 (2.0) | 25.3 (4.6) | 0.036 | Exercise > Control |
| Psychological/spiritual score | 27.4 (2.5) | 26.2 (4.0) | 0.336 | Exercise = Control |
| Social and economic score | 27.2 (3.0) | 25.9 (5.2) | 0.638 | Exercise = Control |
| Family score | 28 (2.6) | 26.8 (5.0) | 0.092 | Exercise = Control |
| Overall | 27.7 (2.1) | 25.7 (4.2) | 0.048 | Exercise > Control |
| Ma 2020 | ||||
| SF‐12 change at 12 months' follow‐up: | ||||
|
| Δ (SD) | Δ (SD) |
|
|
| Physical component | 13.3 (6) | 9.9 (5.9) | < 0.001 | Exercise > Control |
| Mental component | 12.4 (5.4) | 9 (6.2) | < 0.001 | Exercise > Control |
| Maddison 2014 | ||||
| EQ‐5D at 24 weeks' follow‐up: | ||||
|
| 0.86 | 0.83 | 0.23 | Exercise = Control |
| SF‐36 at 24 weeks' follow‐up: |
|
|
| |
| Physical functioning | 52.9 | 51.9 | 0.20 | Exercise = Control |
| Role physical | 52.6 | 50.8 | 0.08 | Exercise = Control |
| Bodily pain | 52.4 | 51.9 | 0.71 | Exercise = Control |
| General health | 55.3 | 53.2 | 0.03 | Exercise > Control |
| Vitality | 55.7 | 55.9 | 0.79 | Exercise = Control |
| Social Functioning | 53.3 | 52.4 | 0.42 | Exercise = Control |
| Role emotional | 51.4 | 51.6 | 0.81 | Exercise = Control |
| Mental health | 54.6 | 54.0 | 0.61 | Exercise = Control |
| Mutwalli 2012 | ||||
| SF‐36 Health status score at 6 months' follow‐up: | ||||
|
| 90.14 (4.83) | 60.55 (16.21) | 0.000 | Exercise > Control |
| Oerkild 2012 | ||||
| SF‐36 at 12 months' follow‐up: | ||||
|
| Δ (95% CI) | Δ (95% CI) |
|
|
| SF 12 PCS | ‐1.1 (‐5.3 to 3.1) | ‐1.4 (‐5.2 to 2.3) | NS* | Exercise = Control |
| SF 12 MCS | ‐1.4 (‐6.1 to 3.3) | ‐0.3 (‐4.6 to 4.0) | NS* | Exercise = Control |
| Oldridge 1991 | ||||
| QLMI at 4 months' follow‐up: | ||||
| Limitations | 54 | 54 | NS | Exercise = Control |
| Emotions | 103 | 101 | NS | Exercise = Control |
| QLMI at 8 months' follow‐up: | ||||
| Limitations | 54 | 54 | NS | Exercise = Control |
| Emotions | 103 | 103 | NS | Exercise = Control |
| QLMI at 12 months' follow‐up: | ||||
| Limitations | 54 | 55 | NS | Exercise = Control |
| Emotions | 105 | 102 | NS | Exercise = Control |
| Reid 2012 | ||||
| MacNew at 6 months' follow‐up: | ||||
| Global score | 5.8 (0.6) | 5.6 (0.8) | 0.112 | Exercise = Control |
| Emotional subscale | 5.6 (0.6) | 5.4 (0.7) | 0.038 | Exercise > Control |
| Social subscale | 6.3 (0.8) | 6.0 (1.0) | 0.162 | Exercise = Control |
| Physical subscale | 6.0 (0.8) | 5.8 (1.0) | 0.031 | Exercise > Control |
| Sandstrom 2005 | ||||
| Time Trade Off (TTO) at 12 months' follow‐up: | ||||
|
| 0.86 (0.23) | 0.85 (0.21) | NS* | Exercise = Control |
| EuroQol Part one at 12 months' follow‐up: | ||||
|
| 0.87 (0.15) | 0.86 (0.16) | NS* | Exercise = Control |
| EuroQol Part two at 12 months' follow‐up: |
|
| ||
|
| 7.6 (1.46) | 7.43 (1.46) | NS* | Exercise = Control |
| Santaularia 2017 | ||||
| EuroQol‐5D at 12 months' follow‐up: | ||||
|
| N (%) | N(%) |
|
|
| Mobility | 33 (84.6) | 33 (75) | 0.019 | Exercise > Control |
| Self‐care | 38 (97.4) | 43 (97.7) | 0.172 | Exercise = Control |
| Usual activities | 32 (82) | 31 (70.5) | 0.803 | Exercise = Control |
| Pain/discomfort | 28 (71.8) | 26 (59.1) | 0.528 | Exercise = Control |
| Anxiety/depression | 22 (56.4) | 26 (59.1) | 0.429 | Exercise = Control |
| Snoek 2020 | ||||
| SF‐36 summary scores at 6 months: | ||||
| Physical | 50.2 (7.2) | 48.3 (7.5) | 0.086* | Exercise = Control |
| Mental | 54.0 (8.4) | 52.7 (9.1) | 0.322* | Exercise = Control |
| SF‐36 summary scores at 12 months: | ||||
| Physical | 50.6 (7.2) | 49 (8.2) | 0.167* | Exercise = Control |
| Mental | 53.2 (8.8) | 52.5 (9.2) | 0.604* | Exercise = Control |
| Stahle 1999 | ||||
| Karolinska Questionnaire at 12 months' follow‐up: | ||||
| Chest pain | 0.6 (1.2) | 0.4 (1.3) | NS | Exercise = Control |
| Shortness of breath | 0.4 (1.1) | 0.2 (1.0) | NS | Exercise = Control |
| Dizziness | ‐0.1 (1.1) | 0.2 (0.9) | NS | Exercise = Control |
| Palpitation | ‐0.1 (1.0) | 0.1 (0.9) | NS | Exercise = Control |
| Cognitive ability | ‐0.1 (0.6) | 0.0 (0.7) | NS | Exercise = Control |
| Alertness | 0.0 (0.9) | 0.1 (0.8) | NS | Exercise = Control |
| Quality of sleep | 0.0 (0.5) | 0.1 (0.5) | NS | Exercise = Control |
| Physical ability | 0.2 (0.7) | 0.1 (0.4) | NS | Exercise = Control |
| Daily activity | 0.3 (0.5) | 0.1 (0.5) | NS | Exercise = Control |
| Depression | 0.1 (0.3) | 0.1 (0.2) | NS | Exercise = Control |
| Self‐perceived health | 0.5 (1.3) | 0.3 (1.0) | NS | Exercise = Control |
| "Ladder of Life" present | 1.2 (1.2) | 0.9 (1.8) | NS | Exercise = Control |
| "Ladder of Life" future | 0.8 (2.7) | 0.4 (2.3) | NS | Exercise = Control |
| Fitness | 0.6 (1.4) | 0.4 (1.0) | NS | Exercise = Control |
| Physical ability | 0.7 (1.0) | 0.4 (1.1) | NS | Exercise = Control |
| Toobert 2000 | ||||
| SF‐36 at 24 months' follow‐up: | ||||
| Physical functioning | NR | NR | NS | Exercise = Control |
| Physical performance | NR | NR | NS | Exercise = Control |
| Bodily pain | NR | NR | NS | Exercise = Control |
| General health | NR | NR | < 0.05 | Exercise > Control |
| Vitality | NR | NR | NS | Exercise = Control |
| Social functioning | NR | NR | < 0.05 | Exercise > Control |
| Emotional performance | NR | NR | NS | Exercise = Control |
| Mental health | NR | NR | NS | Exercise = Control |
| Uddin 2020 | ||||
| WHOQoL‐BREF at 12 months' follow‐up | ||||
| Overall perception of HRQoL | 4.03 (0.49) | 3.2 (0.82) | < 0.01 | Exercise > Control |
| Overall perception of health | 4.06 (0.4) | 3.17 (0.38) | < 0.01 | Exercise > Control |
| Physical domain | 26.9 (2.88) | 21.17 (3.35) | < 0.01 | Exercise > Control |
| Psychological domain | 23.42 (2.84) | 17.87 (3.19) | < 0.01 | Exercise > Control |
| Social relationship domain | 11.83 (1.5) | 10.75 (0.89) | < 0.01 | Exercise > Control |
| Environmental domain | 28.8 (4.24) | 21.77 (5.31) | 0.03 | Exercise > Control |
| Wang 2012 | ||||
| SF‐36 at 6 months' follow‐up: | ||||
| Physical functioning | 80.8 (13.7) | 73.2 (13.0) | < 0.001 | Exercise > Control |
| Physical performance | 68.2 (17.3) | 56.2 (46.8) | 0.015 | Exercise > Control |
| Bodily pain | 68.2 (17.3) | 63.5 (14.6) | 0.012 | Exercise > Control |
| General health | 57.4 (20.3) | 49.0 (16.2) | 0.017 | Exercise > Control |
| Vitality | 66.3 (17.3) | 56.4 (21.7) | 0.002 | Exercise > Control |
| Social functioning | 71.3 (21.4) | 65.8 (18.0) | 0.031 | Exercise > Control |
| Emotional performance | 80.8 (37.9) | 75.9 (39.7) | 0.12 | Exercise = Control |
| Mental health | 73.5 (17.1) | 65.4 (20.7) | 0.011 | Exercise > Control |
| MIDAS at 6 months' follow‐up: | ||||
| Physical Activity | 37.7 (11.2) | 42.6 (12.3) | < 0.001 | Exercise > Control |
| Insecurity | 28.7 (9.7) | 33.4 (13.8) | < 0.001 | Exercise > Control |
| Emotional reaction | 30.4 (12.8) | 34.8 (14.4) | 0.008 | Exercise > Control |
| Dependency | 27.6 (9.4) | 31.8 (16.6) | 0.001 | Exercise > Control |
| Diet | 36.8 (15.4) | 43.6 (20.7) | 0.40 | Exercise = Control |
| Concerns over meds | 29.4 (12.6) | 37.7 (18.0) | <0.001 | Exercise > Control |
| Side Effects | 28.2 (11.1) | 30.8 (14.3) | 0.30 | Exercise > Control |
| West 2012 | ||||
| SF‐36 at 12 months' follow‐up: | ||||
| Physical function | 65 (29) | 64 (30) | NS* | Exercise = Control |
| Role physical | 69 (31) | 67 (33) | NS* | Exercise = Control |
| Role emotional | 85 (23) | 85 (25) | NS* | Exercise = Control |
| Social function | 81 (28) | 79 (29) | NS* | Exercise = Control |
| Mental health | 76 (13) | 76 (13) | NS* | Exercise = Control |
| Energy /vitality | 65 (24) | 65 (24) | NS* | Exercise = Control |
| Pain | 69 (28) | 68 (29) | NS* | Exercise = Control |
| Health Perception | 58 (25) | 57 (25) | NS* | Exercise = Control |
| Yu 2003 | ||||
| SF‐36 at 8 months' follow‐up: | ||||
| Physical functioning | 88 (12) | 82 (17) | 0.03* | Exercise > Control |
| Physical performance | 75 (33) | 66 (35) | 0.18* | Exercise = Control |
| Bodily pain | 80 (25) | 80 (25) | 1.00* | Exercise = Control |
| General health | 64 (26) | 60 (28) | 0.45* | Exercise = Control |
| Vitality | 79 (18) | 65 (17) | 0.0001 | Exercise > Control |
| Social functioning | 89 (27) | 82 (28) | 0.15 | Exercise = Control |
| Emotional performance | 93 (18) | 83 (35) | 0.05 | Exercise = Control |
| Mental health | 84 (16) | 80 (15) | 0.2 | Exercise = Control |
| SF‐36 at 24 months' follow‐up: | ||||
| Physical functioning | 88 (13) | 87 (9) | 0.67* | Exercise = Control |
| Physical performance | 80 (32) | 79 (30) | 0.87* | Exercise = Control |
| Bodily pain | 81 (21) | 85 (20) | 0.33* | Exercise = Control |
| General health | 64 (20) | 61 (18) | 0.43* | Exercise = Control |
| Vitality | 73 (21) | 73 (17) | 1.00* | Exercise = Control |
| Social functioning | 79 (30) | 90 (18) | 0.04* | Exercise > Control |
| Emotional performance | 89 (25) | 93 (25) | 0.42* | Exercise = Control |
| Mental health | 85 (14) | 85 (12) | 1.00* | Exercise = Control |
| Zwisler 2008 | ||||
| SF‐36 at 12 months' follow‐up: | ||||
| Physical Component Score | 45.2 (9.8) | 46.4 (9.8) | 0.39* | Exercise = Control |
| Mental Component Score | 50.6 (10.8) | 48.4 (11.5) | 0.16* | Exercise = Control |
| AP‐QLQ: Angina Pectoris‐Quality of Life questionnaire | ||||
| Author/ year | Briffa 2005 | Hambrecht 2004 | Hautala 2017 | Kovoor 2006/Hall 2002 | Maddison 2014 | Marchionni 2003 | Oldridge 1991/93 | Yu 2004 |
| Follow‐up (months) | 12 | 12 | 12 | 12 | 6 | 14 | 12 | 24 |
| Year of costs (currency) | 1998 (Australian dollars ‐ AUD) | NR (US dollars ‐ USD) | NR (euros ‐ EUR) | 1999 (Australian dollars ‐ AUD) | NR (euros ‐ EUR) | 2000 (US dollars ‐ USD) | 1991 (US dollars ‐ USD) | 2003 (US dollars ‐ USD) |
| Cost of rehabilitation | ||||||||
| Mean cost/patient | AUD 694 | NR | EUR 299 | AUD 394 | EUR 127 | USD 5246 | USD 670 | NR |
| Costs considered | Details of costed elements not provided | NR | Estimated according to the average monthly fees in Finnish gyms where individual guidance in exercise training is led by a health care professional | staff, assessments, counselling, education, patient travel | NR | NR | space, equipment, staff, literature resources, operating costs, parking, patients costs | NR |
| Total healthcare costs | ||||||||
| Rehabilitation mean cost/patient | AUD 4937 | USD 3708 ± 156 | EUR 1944 | NR | NR | USD 17,272 | NR | USD 15,292 |
| Usual care mean cost/patient | AUD 4541 | USD 6086 ± 370 | EUR 3027 | NR | NR | USD 12,433 | NR | USD 15,707 |
| Absolute difference in mean cost/patient* | AUD 395 | USD ‐2378 | EUR ‐1083 | NR | NR | USD 4839 | USD 480 | USD ‐415 |
| P value for cost difference | 0.74 | P < 0.001 | NR | P > 0.05 (see below) | NR | NR | NR | P > 0.05 |
| Additional healthcare costs considered | Hospitalisations, pharmaceuticals, tests, consultations, rehabilitation, patient expenses, ambulance | Rehospitalisations, revascularisation, cycle ergometers, training facilities, and supervising staff | Primary health care costs, secondary health care costs, occupational health care service costs | Phone calls (P = 0.10); hospital admissions (P = 0.11); gated heart pool scan (P = 0.50); exercise stress test (P = 0.72); other diagnostics (P = 0.37); visits to general practitioner (P = 0.61), specialist doctor (P = 0.35), or health‐care professional (P = 0.31) | NR | NR | Service utilisation, physician costs, emergency costs, in‐patient days, allied health, other rehabilitation visits | Hospitalisations; revascularisations; private clinic visit; cardiac clinic visits; public non‐cardiac visits; casualty visits; drugs |
| Cost‐effectiveness | ||||||||
| Rehabilitation mean health care benefits | Utility‐based quality of life – | NR | Average change in 15D utility: 0.013 | NR | NR | NR | NR | NR |
| Usual care mean health care benefit | Utility 0.010 (95% CI ‐0.001 to 0.022) | NR | Average change in 15D utility: ‐0.012 | NR | NR | NR | NR | NR |
| Incremental mean health care benefit | Utility 0.013 (95% CI, NR) P = 0.38; +0.009 QALYs | NR | 0.045 QALYs (0.023‐0.077) | NR | NR | NR | 0.052 QALYs (95% CI, 0.007 to 0.1) | 0.06 QALYs |
| Incremental cost effectiveness ratio/patient | AUD +42,535 per QALY. Extensive sensitivity analyses reported. | NR | EUR ‐24,511 per QALY | NR | EUR +15,247 per QALY | NR | USD +9200 per QALY | USD ‐650 per QALY |
| NR: not reported * The currency for Hambrecht 2004 is not reported, but healthcare costs are reported within the paper with $ | ||||||||
| Explanatory variable (n trials) | Exp (slope)* | 95% confidence interval, | Proportion of variance explained (adjusted R2) | Interpretation |
| Case mix (% MI patients) (n = 46) | RR = 1.00 | 1.00 to 1.00, P = 0.15 | 56.1% | No evidence that risk ratio is associated with case mix |
| Dose of exercise (number of weeks of exercise training x average number of sessions/week x average min/session) (n = 33) | RR = 1.00 | 1.00 to 1.00, P = 0.11 |
100% | No evidence that risk ratio is associated with type of CR |
| Duration of follow‐up (months) (n = 47) | RR = 1.00 | 1.00 to 1.00, P = 0.07 | 100% | No evidence that risk ratio is associated with length of follow‐up |
| Type of CR (exercise only vs comprehensive CR) (n = 47) | RR = 1.04 | 0.84 to 1.31, P = 0.70 | ‐27.1% | No evidence that risk ratio is associated with type of CR |
| Year of publication (pre‐1995 vs post‐1995) (n = 47) | RR = 0.84 | 0.70 to 0.99, P = 0.04 | 100% | No evidence that risk ratio is associated with publication year |
| CR setting (n = 47) | RR = 0.95 | 0.82 to 1.24, P = 0.95 | ‐11.3% | No evidence that risk ratio is associated with type of CR |
| Risk of bias (low risk in ≤ 3 items vs > 3 items) (n = 47) | RR = 1.02 | 0.94 to 1.09, P = 0.67 | ‐68.55% | No evidence that risk ratio is associated with risk of bias |
| Study location (continent ‐ Europe, North America, Australia/Asia or Other) (n = 47) | RR = 1.01 | 0.86 to 1.19, P = 0.93 | ‐41.24% | No evidence that risk ratio is associated with study location |
| Low‐ and middle‐income country (LMIC) vs high‐income country (n = 47) | RR = 1.02 | 0.70 to 1.48, P = 0.93 | ‐45.10% | No evidence that risk ratio is associated with LMIC |
| Sample size (≤ 150 vs > 150) (n = 47) | RR = 1.19 | 0.73 to 1.93, P = 0.47 | 16.07% | No evidence that risk ratio is associated with study sample size |
| Explanatory variable (n trials) | Exp (slope)* | 95% confidence interval, | Proportion of variance explained (adjusted R2) | Interpretation |
| Case mix (% MI patients) (n = 27) | RR = 1.00 | 0.99 to 1.01, P = 0.76 | ‐8.74% | No evidence that risk ratio is associated with case mix |
| Dose of exercise (number of weeks of exercise training x average number of sessions/week x average min/session) (n = 22) | RR = 1.00 | 1.00 to 1.00, P = 0.62 | 0% | No evidence that risk ratio is associated with dose of exercise |
| Duration of follow‐up (months) (n = 28) | RR = 0.99 | 0.99 to 1.00, P = 0.05 | 90.36% | No evidence that risk ratio is associated with length of follow‐up |
| Type of CR (exercise only vs comprehensive CR) (n = 28) | RR = 0.83 | 0.62 to 1.10, P = 0.18 | 75.69% | No evidence that risk ratio is associated with type of CR |
| Year of publication (pre‐1995 vs post‐1995) (n = 28) | RR = 1.37 | 0.89 to 2.13, P = 0.15 | 63.31% | No evidence that risk ratio is associated with publication year |
| Setting (centre vs home) (n = 28) | RR = 1.05 | 0.88 to 1.24, P = 0.61 | ‐29.66% | No evidence that risk ratio is associated with setting of CR |
| Risk of bias (low risk in ≤ 3 items vs > 3 items) (n = 28) | RR = 0.90 | 0.73 to 1.11, P = 0.30 | 85.73% | No evidence that risk ratio is associated with risk of bias |
| Study location (continent ‐ Europe, North America, Australia/Asia or Other) (n = 28) | RR = 1.02 | 0.75 to 1.39, P = 0.89 | ‐41.75% | No evidence that risk ratio is associated with study location |
| Low‐ and middle‐income country (LMIC) vs high‐income country (n = 28) | RR = 0.69 | 0.22 to 2.19, P = 0.52 | 9.36% | No evidence that risk ratio is associated with LMIC |
| Sample size (≤ 150 vs > 150) (n = 28) | RR = 1.28 | 0.69 to 2.37, P = 0.42 | 28.43% | No evidence that risk ratio is associated with study sample size |
| Explanatory variable (n trials) | Exp (slope)* | 95% confidence interval, | Proportion of variance explained (adjusted R2) | Interpretation |
| Case mix (% MI patients) (n = 41) | RR = 1.00 | 0.99 to 1.01, P = 0.93 | ‐4.57% | No evidence that risk ratio is associated with case mix |
| Dose of exercise (number of weeks of exercise training x average number of sessions/week x average min/session) (n = 33) | RR = 1.00 | 1.00 to 1.00, P = 0.68 | 0% | No evidence that risk ratio is associated with dose of exercise |
| Duration of follow‐up (months) (n = 41) | RR = 1.00 | 0.99 to 1.01, P = 0.97 | ‐12.45% | No evidence that risk ratio is associated with length of follow‐up |
| Type of CR (exercise only vs comprehensive CR) (n = 41) | RR = 0.85 | 0.58 to 1.25, P = 0.39 | 9.68% | No evidence that risk ratio is associated with type of CR |
| Year of publication (pre‐1995 vs post‐1995) (n = 41) | RR = 1.36 | 0.94 to 1.97, P = 0.11 | 25.40% | No evidence that risk ratio is associated with publication year |
| Setting (centre vs home) (n = 39) | RR = 0.80 | 0.67 to 0.95, P = 0.01 | 67.62% | No evidence that risk ratio is associated with setting of CR |
| Risk of bias (low risk in ≤ 3 items vs > 3 items) (n = 41) | RR=1.39 | 0.85 to 2.26, P = 0.18 | ‐16.70% | No evidence that risk ratio is associated with risk of bias |
| Study location (continent ‐ Europe, North America, Australia/Asia or Other) (n = 41) | RR = 0.71 | 0.49 to 1.05, P = 0.09 | 12.94% | No evidence that risk ratio is associated with study location |
| Low‐ and middle‐income country (LMIC) vs high income country (n = 41) | RR = 0.65 | 0.33 to 1.61, P = 0.20 | 0.86% | No evidence that risk ratio is associated with LMIC |
| Sample size (≤ 150 vs > 150) (n = 41) | RR = 1.69 | 1.05 to 2.72, P = 0.03 | 54.95% | No evidence that risk ratio is associated with study sample size |
| Explanatory variable (n trials) | Exp (slope)* | 95% confidence interval, | Proportion of variance explained (adjusted R2) | Interpretation |
| Case mix (% MI patients) (n = 31) | RR = 1.01 | 1.00 to 1.02, P = 0.05 | 0% | No evidence that risk ratio is associated with case mix |
| Dose of exercise (number of weeks of exercise training x average number of sessions/week x average min/session) (n = 25) | RR = 1.00 | 1.00 to 1.00, P = 0.78 | 0% | No evidence that risk ratio is associated with dose of exercise |
| Duration of follow‐up (months) (n = 31) | RR = 1.00 | 0.99 to 1.01, P = 0.75 | 0% | No evidence that risk ratio is associated with length of follow‐up |
| Type of CR (exercise only vs comprehensive CR) (n = 31) | RR = 1.04 | 0.67 to 1.61, P = 0.86 | 0% | No evidence that risk ratio is associated with type of CR |
| Year of publication (pre‐1995 vs post‐1995) (n = 31) | RR = 0.88 | 0.56 to 1.41, P = 0.59 | 0% | No evidence that risk ratio is associated with publication year |
| Setting (centre vs home) (n = 31) | RR = 1.07 | 0.87 to 1.33, P = 0.51 | 0% | No evidence that risk ratio is associated with setting of CR |
| Risk of bias (low risk in ≤ 3 items vs > 3 items) (n = 31) | RR = 0.94 | 0.64 to 1.38, P = 0.73 | 0% | No evidence that risk ratio is associated with risk of bias |
| Study location (continent ‐ Europe, North America, Australia/Asia or Other) (n = 31) | RR = 1.19 | 0.83 to 1.71, P = 0.34 | 0% | No evidence that risk ratio is associated with study location |
| Low‐ and middle‐income country (LMIC) vs high income country (n = 31) | RR = 0.51 | 0.08 to 3.18, P = 0.46 | 0% | No evidence that risk ratio is associated with LMIC |
| Sample size (≤ 150 vs > 150) (n = 31) | RR = 1.31 | 0.82 to 2.09, P = 0.25 | 0% | No evidence that risk ratio is associated with study sample size |
| Explanatory variable (n trials) | Exp (slope)* | 95% confidence interval, | Proportion of variance explained (adjusted R2) | Interpretation |
| Case mix (% MI patients) (n = 18) | RR = 1.00 | 1.00 to 1.01, P = 0.50 | 0% | No evidence that risk ratio is associated with case mix |
| Dose of exercise (number of weeks of exercise training x average number of sessions/week x average min/session) (n = 16) | RR = 1.00 | 1.00 to 1.00, P = 0.50 | 0% | No evidence that risk ratio is associated with dose of exercise |
| Duration of follow‐up (months) (n = 18) | RR = 1.00 | 0.99 to 1.01, P = 0.82 | 0% | No evidence that risk ratio is associated with length of follow‐up |
| Type of CR (exercise only vs comprehensive CR) (n = 18) | RR = 0.78 | 0.38 to 1.59, P = 0.47 | 0% | No evidence that risk ratio is associated with type of CR |
| Year of publication (pre‐1995 vs post‐1995) (n = 18) | RR = 0.95 | 0.46 to 1.95, P = 0.87 | 0% | No evidence that risk ratio is associated with publication year |
| Setting (centre vs home) (n = 18) | RR = 0.91 | 0.72 to 1.15, P = 0.41 | 0% | No evidence that risk ratio is associated with setting of CR |
| Risk of bias (low risk in ≤ 3 items vs > 3 items) (n = 18) | RR = 1.09 | 0.72 to 1.66, P = 0.67 | 0% | No evidence that risk ratio is associated with risk of bias |
| Study location (continent ‐ Europe, North America, Australia/Asia or Other) (n = 18) | RR = 0.81 | 0.53 to 1.23, P = 0.30 | 0% | No evidence that risk ratio is associated with study location |
| Low‐ and middle‐income country (LMIC) vs high income country (n = 18) | RR = 0.29 | 0.05 to 1.63, P = 0.15 | 0% | No evidence that risk ratio is associated with LMIC |
| Sample size (≤ 150 vs > 150) (n = 18) | RR = 1.19 | 0.70 to 2.01, P = 0.49 | 0% | No evidence that risk ratio is associated with study sample size |
| Explanatory variable (n trials) | Exp (slope)* | 95% confidence interval, | Proportion of variance explained (adjusted R2) | Interpretation |
| Case mix (% MI patients) (n = 23) | RR = 1.00 | 1.00 to 1.01, P = 0.71 | ‐20.91% | No evidence that risk ratio is associated with case mix |
| Dose of exercise (number of weeks of exercise training x average number of sessions/week x average min/session) (n = 19) | RR = 1.00 | 1.00 to 1.00, P = 0.44 | ‐69.78% | No evidence that risk ratio is associated with dose of exercise |
| Duration of follow‐up (months) (n = 23) | RR = 1.01 | 1.00 to 1.01, P = 0.07 | 56.52% | No evidence that risk ratio is associated with length of follow‐up |
| Type of CR (exercise only vs comprehensive CR) (n = 23) | RR = 0.93 | 0.65 to 1.33, P = 0.70 | ‐50.20% | No evidence that risk ratio is associated with type of CR |
| Year of publication (pre‐1995 vs post‐1995) (n = 23) | RR = 1.12 | 0.80 to 1.57, P = 0.48 | ‐32.69% | No evidence that risk ratio is associated with publication year |
| Setting (centre vs home) (n = 23) | RR = 0.94 | 0.83 to 1.06, P = 0.28 | ‐36.70% | No evidence that risk ratio is associated with setting of CR |
| Risk of bias (low risk in ≤ 3 items vs > 3 items) (n = 23) | RR = 1.00 | 0.71 to 1.40, P = 0.99 | ‐44.14% | No evidence that risk ratio is associated with risk of bias |
| Study location (continent ‐ Europe, North America, Australia/Asia or Other) (n = 23) | RR = 0.86 | 0.69 to 1.08, P = 0.18 | ‐137.18% | No evidence that risk ratio is associated with study location |
| Low‐ and middle‐income country (LMIC) vs high income country (n = 23) | RR = 1.06 | 0.72 to 1.55, P = 0.76 | ‐49.12% | No evidence that risk ratio is associated with LMIC |
| Sample size (≤ 150 vs > 150) (n = 19) | RR = 1.45 | 1.08 to 1.96, P = 0.02 | 100% | No evidence that risk ratio is associated with study sample size |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1.1 All‐cause mortality Show forest plot | 47 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 1.1.1 Follow‐up of 6 to 12 months | 25 | 8823 | Risk Ratio (M‐H, Random, 95% CI) | 0.87 [0.73, 1.04] |
| 1.1.2 Follow‐up of > 12 to 36 months | 16 | 11073 | Risk Ratio (M‐H, Random, 95% CI) | 0.90 [0.80, 1.02] |
| 1.1.3 Follow‐up longer than 3 years | 11 | 3828 | Risk Ratio (M‐H, Random, 95% CI) | 0.91 [0.75, 1.10] |
| 1.2 Cardiovascular mortality Show forest plot | 26 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 1.2.1 Follow‐up of 6 to 12 months | 15 | 5360 | Risk Ratio (M‐H, Random, 95% CI) | 0.88 [0.68, 1.14] |
| 1.2.2 Follow‐up of > 12 months to 36 months | 5 | 3614 | Risk Ratio (M‐H, Random, 95% CI) | 0.77 [0.63, 0.93] |
| 1.2.3 Follow‐up of longer than 3 years | 8 | 1392 | Risk Ratio (M‐H, Random, 95% CI) | 0.58 [0.43, 0.78] |
| 1.3 Fatal and/or nonfatal MI Show forest plot | 39 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 1.3.1 Follow‐up of 6 to 12 months | 22 | 7423 | Risk Ratio (M‐H, Random, 95% CI) | 0.72 [0.55, 0.93] |
| 1.3.2 Follow‐up of > 12 to 36 months | 12 | 9565 | Risk Ratio (M‐H, Random, 95% CI) | 1.07 [0.91, 1.27] |
| 1.3.3 Follow‐up of longer than 3 years | 10 | 1560 | Risk Ratio (M‐H, Random, 95% CI) | 0.67 [0.50, 0.90] |
| 1.4 CABG Show forest plot | 29 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 1.4.1 Follow‐up of 6 to 12 months | 20 | 4473 | Risk Ratio (M‐H, Random, 95% CI) | 0.99 [0.78, 1.27] |
| 1.4.2 Follow‐up of > 12 to 36 months | 9 | 2826 | Risk Ratio (M‐H, Random, 95% CI) | 0.97 [0.77, 1.23] |
| 1.4.3 Follow‐up of longer than 3 years | 4 | 675 | Risk Ratio (M‐H, Random, 95% CI) | 0.66 [0.34, 1.27] |
| 1.5 PCI Show forest plot | 18 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 1.5.1 Follow‐up of 6 to 12 months | 13 | 3465 | Risk Ratio (M‐H, Random, 95% CI) | 0.86 [0.63, 1.19] |
| 1.5.2 Follow‐up of > 12 to 36 months | 6 | 1983 | Risk Ratio (M‐H, Random, 95% CI) | 0.96 [0.69, 1.35] |
| 1.5.3 Follow‐up of longer than 3 years | 3 | 567 | Risk Ratio (M‐H, Random, 95% CI) | 0.76 [0.48, 1.20] |
| 1.6 All‐cause hospital admissions Show forest plot | 22 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 1.6.1 Follow‐up of 6 to 12 months | 14 | 2030 | Risk Ratio (M‐H, Random, 95% CI) | 0.58 [0.43, 0.77] |
| 1.6.2 Follow‐up of > 12 to 36 months | 9 | 5995 | Risk Ratio (M‐H, Random, 95% CI) | 0.92 [0.82, 1.03] |
| 1.7 Cardiovascular hospital admissions Show forest plot | 8 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 1.7.1 Follow‐up of 6 to 12 months | 6 | 1087 | Risk Ratio (M‐H, Random, 95% CI) | 0.80 [0.41, 1.59] |
| 1.7.2 Follow up of >12 to 36 months | 3 | 943 | Risk Ratio (M‐H, Random, 95% CI) | 0.92 [0.76, 1.12] |
| 1.8 HRQoL SF‐36 summary scores at 6 to 12 months follow up Show forest plot | 6 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 1.8.1 Physical component score | 6 | 1741 | Mean Difference (IV, Random, 95% CI) | 1.70 [‐0.08, 3.47] |
| 1.8.2 Mental component score | 6 | 1741 | Mean Difference (IV, Random, 95% CI) | 2.14 [1.07, 3.22] |
| 1.9 HRQoL SF‐36 8 domains at 6 to 12 months follow up Show forest plot | 8 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 1.9.1 Physical functioning | 8 | 2756 | Mean Difference (IV, Random, 95% CI) | 8.47 [3.69, 13.24] |
| 1.9.2 Physical performance | 8 | 2756 | Mean Difference (IV, Random, 95% CI) | 8.08 [2.89, 13.27] |
| 1.9.3 Bodily pain | 8 | 2756 | Mean Difference (IV, Random, 95% CI) | ‐0.06 [‐8.97, 8.84] |
| 1.9.4 General health | 8 | 2756 | Mean Difference (IV, Random, 95% CI) | 5.66 [2.08, 9.25] |
| 1.9.5 Vitality | 7 | 2638 | Mean Difference (IV, Random, 95% CI) | 5.78 [1.89, 9.67] |
| 1.9.6 Social functioning | 8 | 2756 | Mean Difference (IV, Random, 95% CI) | 1.98 [0.26, 3.70] |
| 1.9.7 Emotional performance | 7 | 2638 | Mean Difference (IV, Random, 95% CI) | 0.69 [‐1.33, 2.71] |
| 1.9.8 Mental health | 8 | 2756 | Mean Difference (IV, Random, 95% CI) | 5.60 [1.21, 9.98] |
| 1.10 HRQoL EQ‐5D at 6 to 12 months follow up Show forest plot | 3 | 476 | Mean Difference (IV, Random, 95% CI) | 0.05 [‐0.01, 0.10] |
