Exercise training for adult lung transplant recipients

Summary of findings 1. Resistance (anaerobic) exercise training versus usual care or no exercise training

Outcomes	Effect*	No. of participants (studies)	Quality of the evidence (GRADE)
Resistance (anaerobic) exercise training versus usual care or no exercise training for adult lung transplant recipients
Patient or population: adult lung transplant recipients Settings: outpatient training Intervention: resistance (anaerobic) exercise training Comparison: usual care or no exercise training
Maximal and functional exercise capacity Follow‐up: up to 6 months	Not reported	‐‐	‐‐
HRQoL Follow‐up: up to 6 months	Not reported	‐‐	‐‐
Adverse events Assessed rejection episodes of lung transplant Follow‐up: up to 6 months	One study reported a lower number of rejection episodes, and other one reported a higher number, however, there were no differences between groups	32 (2)	⊕⊝⊝⊝ very low²
Pulmonary function Follow‐up: up to 6 months	Not reported	‐‐	‐‐
Muscular strength¹ Assessed lumbar strength at seven testing positions Follow‐up: up to 6 months	One study reported greater lumbar strength in all seven positions and other one only in the three positions of test hip flexion	32 (2)	⊕⊝⊝⊝ very low³
Pathological bone fractures Estimated indirectly through BMD Follow‐up: up to 6 months	Two studies reported no differences between groups in BMD	32 (2)	⊕⊝⊝⊝ very low⁴
Death (any cause) Follow‐up: up to 6 months	Not reported	‐‐	‐‐
* For all outcomes of interest a single pooled effect estimate is not available and only a narrative synthesis of the evidence is provided * The effect of the intervention is reported based on comparison with the respective control groups in each study HRQoL: health‐related quality of life; BMD: bone mineral density
GRADE Working Group grades of evidence High certainty: Further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low certainty: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low certainty: We are very uncertain about the estimate.
¹ Increased strength from baseline ² The certainty of the evidence was downgraded one level for risk of bias (random sequence generation was assessed with a unclear risk of bias and allocation concealment, incomplete outcome data, selective reporting and other potential sources of bias were assessed with a high risk of bias),one level for imprecision (sample size was small and not calculated and 95% CI includes non‐effect value and is also broad), and one level for indirectness (there are no reported adverse effects directly related to the practice of exercise training, for example, muscle injuries) ³ The certainty of the evidence was downgraded one level for risk of bias (random sequence generation was assessed with a unclear risk of bias and allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting and other potential sources of bias were assessed with a high risk), one level for inconsistency (variability in results across studies) and one level for imprecision (sample size was small and not calculated and 95% CI includes non‐effect value and was also broad) ⁴ The certainty of the evidence was downgraded one level for risk of bias (random sequence generation was assessed with a unclear risk of bias and allocation concealment, blinding of participants and personnel, incomplete outcome data, selective reporting and other potential sources of bias were assessed with a high risk), one level for indirectness (BMD was used as an intermediate surrogate outcome of pathological bone fractures), and one level for imprecision (sample size was small and not calculated and 95% CI includes non‐effect value and was also broad)

Summary of findings 2. Resistance (anaerobic) exercise training versus another form of resistance (anaerobic) exercise training

Outcomes	Effect*	No. of participants (studies)	Quality of the evidence (GRADE)
Resistance (anaerobic) exercise training versus another form of resistance (anaerobic) exercise training for adult lung transplant recipients
Patient or population: adult lung transplant recipients Settings: outpatient training Intervention: resistance (anaerobic) exercise training (squats using WBVT (Gloeckl 2015; Gloeckl 2017); supervised upper limb exercise (Fuller 2018)) Comparison: other form of resistance (anaerobic) exercise training (squats on the floor (Gloeckl 2015; Gloeckl 2017); unsupervised upper limb exercise (Fuller 2018)
Maximal and functional exercise capacity Assessed metres walked in 6MWT Follow‐up: up to 6 months	Two studies reported a greater walking distance in the resistance (anaerobic) exercise training group, and another study reported no difference between groups	214 (3)^a	⊕⊝⊝⊝ very low¹
HRQoL Assessed by different questionnaires (CRQ and SF‐36) Follow‐up: up to 6 months	Three studies reported no differences between groups in the HRQoL	215 (3)^a	⊕⊝⊝⊝ very low²
Adverse events Assessed indirectly related to the exercise Follow‐up: up to 6 months	Two studies reported no differences in the incidence of adverse events	162 (2)	⊕⊝⊝⊝ very low³
Pulmonary function Follow‐up: up to 6 months	Not reported	‐‐	‐‐
Muscular strength Assessed upper and lower limb muscle strength Follow‐up: up to 6 months	Three studies reported no differences between groups in upper and lower limb muscle strength	245 (3)^a	⊕⊝⊝⊝ very low⁴
Pathological bone fractures Follow‐up: up to 6 months	Not reported	‐‐	‐‐
Death (any cause) Follow‐up: up to 6 months	One study reported no difference between groups in all‐cause mortality	83 (1)	⊕⊝⊝⊝ very low³
* For all outcomes of interest a single pooled effect estimate is not available and only a narrative synthesis of the evidence is provided * The effect of the intervention is reported based on comparison with the respective control groups in each study ^a In the study of Fuller 2018 the number of participants was variable because not all of them showed up for the measurements of all outcomes. The number of participants immediately after the end of the intervention was considered Gloeckl 2015 and Gloeckl 2017 included participants with mean time from transplantation of three months and 5.5 years, respectively, and were therefore not combined statistically (pooled analysis) WBVT: whole‐body vibration training; 6MWT: 6‐minute walk test; HRQoL: health‐related quality of life; CRQ: chronic respiratory questionnaire; SF‐36**: short form 36 health survey questionnaire
GRADE Working Group grades of evidence High certainty: Further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low certainty: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low certainty: We are very uncertain about the estimate.
¹ The certainty of the evidence was downgraded two levels for risk of bias (blinding of participants, personnel and outcome assessment, incomplete outcome data, selective reporting and sources of funding declared were assessed with a high risk of bias) and one level for imprecision (according to the ERS, the MID established for 6MWT is 30 metres, so the difference found in this study could be or not clinically relevant, because MID is within the CI 95% (Holland 2014). Also the lower limit is very close to 0) ² The certainty of the evidence was downgraded two levels for risk of bias (blinding of participants, personnel and outcome assessment, incomplete outcome data, selective reporting and sources of funding declared were assessed with a high risk of bias) and one level for imprecision (95% CI includes no‐effect value) ³ The certainty of the evidence was downgraded two levels for risk of bias (blinding outcome assessment, incomplete outcome data, selective reporting and sources of funding declared were assessed with a high risk of bias) and imprecision (confidence levels are very wide and there are few events), and one level for indirectness (there are no reported adverse effects directly related to the practice of exercise training, for example, muscle injuries) ⁴ The certainty of the evidence was downgraded two levels for risk of bias (blinding of participants, personnel and outcome assessment, incomplete outcome data, selective reporting and sources of funding declared were assessed with a high risk of bias) and one level for imprecision (the sample size was small and there are few events)

Summary of findings 3. Multimodal exercise training versus usual care or no exercise training

Outcomes	Effect*	No. of participants (studies)	Quality of the evidence (GRADE)
Multimodal exercise training versus usual care or no exercise training for adult lung transplant recipients
Patient or population: adult lung transplant recipients Settings: outpatient training Intervention: multimodal exercise training (continuous aerobic training + resistance training (Langer 2012); high‐intensity interval training + resistance training (Ulvestad 2020)) Comparison: usual care or no exercise training
Maximal and functional exercise capacity Assessed with different maximal and submaximal tests Follow‐up: up to 6 months	One study reported a greater increase in walking distance on the 6MWT in the multimodal exercise training group, however, this and other study reported no differences between groups in other functional and oxygen consumption tests	80 (2)	⊕⊝⊝⊝ verylow¹
HRQoL Assessed by different domains of the SF‐36 questionnaire Follow‐up: up to 6 months	Two studies reported no differences between groups in the HRQoL	80 (2)	⊕⊝⊝⊝ very low²
Adverse events Assessed the severe medical complications and exercise‐related musculoskeletal pain Follow‐up: up to 6 months	Two studies reported no differences between groups in the number of study dropouts due to severe medical complications, and other one study reported low incidence in muscle pain in the intervention group	80 (2)	⊕⊝⊝⊝ very low³
Pulmonary function Assessed by FEV₁ (absolute value and percentage of predicted value) Follow‐up: up to 6 months	Two studies reported no differences between groups in the pulmonary function	80 (2)	⊕⊝⊝⊝ very low⁴
Muscular strength Assessed hand grip, upper and lower limb muscle strength Follow‐up: up to 6 months	Two studies reported no differences in hand grip strength, but one of them did report differences in upper limb strength. The same two studies reported greater lower limb strength in the multimodal exercise training group	80 (2)	⊕⊝⊝⊝ very low⁵
Pathological bone fractures Follow‐up: up to 6 months	Not reported	‐‐	‐‐
Death (any cause) Follow‐up: up to 6 months	One study reported that one patient died due to lung rejection in the no exercise group	40 (1)	⊕⊝⊝⊝ very low⁶
* For all outcomes of interest a single pooled effect estimate is not available and only a narrative synthesis of the evidence is provided * The effect of the intervention is reported based on comparison with the respective control groups in each study 6MWT: 6‐minute walk test; FEV₁ : forced expiratory volume in 1 sec; SF‐36: short form 36 health survey questionnaire
GRADE Working Group grades of evidence High certainty: Further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low certainty: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low certainty: We are very uncertain about the estimate.
¹ The certainty of the evidence was downgraded two levels for risk of bias (allocation concealment, blinding of participants and personnel, blinding of outcome assessment and selective reporting were assessed with a high risk of bias) and one level for imprecision (the minimal important difference (MID), calculated by the rule of thumb that MID is typically 0.5 standard deviations (Norman 2003), is 5.5%, and the MID is included in the CI) ² The certainty of the evidence was downgraded two levels for risk of bias (allocation concealment, blinding of participants and personnel, blinding of outcome assessment and selective reporting were assessed with a high risk) and one level for imprecision (the minimal important difference (MID), calculated by the rule of thumb that MID is typically 0.5 standard deviations (Norman 2003), is 8.5 points in "physical functioning" and 16 points in "role functioning physical", and both MIDs are included in the respective CIs) ³ The certainty of the evidence was downgraded two levels for risk of bias (allocation concealment and selective reporting were assessed with a high risk of bias), one level by indirectness (there are no reported adverse effects directly related to the practice of exercise training, for example, muscle injuries) and one level for imprecision (CI includes non‐effect value and is also broad) ⁴ The certainty of the evidence was downgraded two levels by risk of bias (allocation concealment, blinding of participants and personnel, blinding of outcome assessment and selective reporting were assessed with a high risk of bias) and one level for imprecision (CI includes non‐effect value and is also broad) ⁵ The certainty of the evidence was downgraded two levels for risk of bias (allocation concealment, blinding of participants and personnel, blinding of outcome assessment and selective reporting were assessed with a high risk of bias) and one level for imprecision (the minimal important difference (MID), calculated by the rule of thumb that MID is typically 0.5 standard deviations (Norman 2003), is 10%, and the MID is included in the CI) ⁶ The certainty of the evidence was downgraded two levels by risk of bias (allocation concealment, blinding of participants and personnel, blinding of outcome assessment and selective reporting were assessed with a high risk of bias) and one level for imprecision (the sample size was small and there are few events)

See: Summary of findings 1 Resistance (anaerobic) exercise training versus usual care or no exercise training; Summary of findings 2 Resistance (anaerobic) exercise training versus another form of resistance (anaerobic) exercise training; Summary of findings 3 Multimodal exercise training versus usual care or no exercise training; Summary of findings 4 Fourteen‐weeks versus 7‐weeks multimodal exercise training

Summary of findings 4. Fourteen‐weeks versus 7‐weeks multimodal exercise training

Outcomes	Effect*	No. of participants (studies)	Quality of the evidence (GRADE)
Fourteen‐weeks versus 7‐weeks multimodal exercise training for adult lung transplant recipients
Patient or population: adult lung transplant recipients Settings: outpatient training Intervention: 14‐weeks multimodal exercise training Comparison: 7‐weeks multimodal exercise training
Maximal and functional exercise capacity Assessed metres walked in 6MWT Follow‐up: up to 6 months	One study reported no differences between groups in the distance walked	61 (1)	⊕⊝⊝⊝ verylow¹
HRQoL Assessed by different questionnaires (SF‐36 and AQoL) Follow‐up: up to 6 months	One study reported no differences between groups in HRQoL	59 (1)	⊕⊝⊝⊝ very low²
Adverse events Follow‐up: up to 6 months	Few participants missed some evaluations due to musculoskeletal problems or hospital readmissions, however, the group to which they belonged was not reported	66 (1)	⊕⊝⊝⊝ very low³
Pulmonary function Assessed by FEV₁ and FVC Follow‐up: up to 6 months	One study reported no differences between groups in pulmonary function	66 (1)	⊕⊝⊝⊝ very low⁴
Muscular strength Assessed by peak quadriceps muscle strength Follow‐up: up to 6 months	One study reported no differences between groups in quadriceps muscle strength	59 (1)	⊕⊝⊝⊝ very low⁵
Pathological bone fractures Follow‐up: up to 6 months	Not reported	‐‐	‐‐
Death (any cause) Follow‐up: up to 6 months	One study reported one death due to acute pneumonia in the 14‐week exercise group	66 (1)	⊕⊝⊝⊝ very low⁶
* For all outcomes of interest a single pooled effect estimate is not available and only a narrative synthesis of the evidence is provided * The effect of the intervention is reported based on comparison with the respective control groups in each study 6MWT: 6‐minute walk test; HRQoL: health‐related quality of life; FEV₁ : forced expiratory volume in 1 sec; FVC: forced vital capacity; SF‐36: short form 36 health survey questionnaire; AQoL: Australian quality of life questionnaire
GRADE Working Group grades of evidence High certainty: Further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low certainty: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low certainty: We are very uncertain about the estimate.
¹ The certainty of the evidence was downgraded two levels for risk of bias (blinding of participants and personnel and selective reporting were assessed with a high risk of bias) and one level for imprecision (according to the ERS, the MID established for 6MWT is 30 metres, so the difference found in this study could be or not clinically relevant, because MID is within the CI 95% (Holland 2014)) ² The certainty of the evidence was downgraded two levels for risk of bias (blinding of participants, personnel and outcome assessment and selective reporting were assessed with a high risk of bias) and one level for imprecision (95% CI includes non‐effect value) ³ The certainty of the evidence was downgraded by one level by risk of bias (selective reporting were assessed with a high risk of bias) and two levels by imprecision (the sample size was small and the total number of adverse events was very small) ⁴ The certainty of the evidence was downgraded one level for high risk of bias (reporting bias) and two levels for imprecision (the sample size was small and only reported P value) ⁵ The certainty of the evidence was downgraded two levels by risk of bias (blinding of participants and personnel and selective reporting were assessed with a high risk of bias) and one level by imprecision (sample size not calculated for this outcome and 95% CI includes non‐effect value and is also broad) ⁶ The certainty of the evidence was downgraded one level for high risk of bias (reporting bias) and two levels for imprecision (the sample size was small and there are few events)

Background

Description of the condition

Lung transplantation is the final treatment option for patients with chronic respiratory diseases such as chronic obstructive pulmonary disease (COPD), interstitial lung disease, pulmonary fibrosis, cystic fibrosis, and other diseases with bronchiectasis (Orens 2006; Trulock 1997). According to the records of the International Society for Heart and Lung Transplantation (ISHLT), which includes data of 241 lung transplant centres located in different countries (Yusen 2014), a total of 47,647 lung transplants have been performed in adult patients between 1985 and June 2013 (Yusen 2014). The chronic respiratory diseases most frequently undergoing lung transplantation are COPD not associated with deficiency of alpha1‐antitrypsin (A1ATD) (33% of lung transplants), interstitial lung disease (24%), cystic fibrosis (16%), and COPD associated with A1ATD (6%) (Yusen 2014).

The primary aim of transplantation is to improve patient survival and quality of life (Studer 2004). According to the ISHLT, the survival rate after lung transplantation can be 80% at one year, 53% at five years, and 32% at 10 years. Median survival time is 5.7 years independent of the underlying disease or surgical procedure (Yusen 2014).

In spite of improvements in survival (Yusen 2014) and quality of life after transplantation (Busschbach 1994; Cohen 2014; Finlen Copeland 2013; Gerbase 2008; Gross 1995; Kugler 2005; Kugler 2010; Lanuza 2000; Rodrigue 2005; Rutherford 2005; Santana 2009; Smeritschnig 2005; Studer 2004; TenVergert 1998), lung transplant recipients continue to experience certain problems, including early‐onset anaerobic threshold, persistent changes in pulmonary function, and muscle weakness which is detected at one‐year post‐transplantation (Reinsma 2006). One of the most important reasons why the positive effects of lung transplantation could be diminished in adult patients is the presence of sarcopenia (Kyle 2003; Lands 1999), a disorder clinically defined as the loss of muscle mass and muscle function (strength or performance) (Cruz‐Jentoft 2010). Another important problem is the increased risk of bone fractures secondary to osteoporosis caused by immunosuppressive therapy (Jastrzebski 2010; Kulak 2012). These problems limit a patient's functional capacity and their daily activities.

Description of the intervention

Exercise refers to all physical activity that is planned, structured, repetitive, and directed, and whose objective is to improve or maintain one or more physical fitness components (Caspersen 1985). The American Thoracic Society (ATS) and the European Respiratory Society (ERS) consider this intervention as the cornerstone of pulmonary rehabilitation (Spruit 2013). In people undergoing lung transplantation, it is recommended that exercise training includes aerobic exercises on a treadmill or a cycle ergometer, strength and flexibility exercises for upper and lower extremities, and walking up and down stairs for at least 30 minutes, 4 times a week (Downs 1996; Rochester 2014). Moreover, these guidelines make reference to the fact that intensity of such training should be adjusted individually, taking into account the maximum heart rate reached and when pulmonary function limitations related to exercise are minor (Downs 1996).

However, even though exercise training is considered to play a key role in the management of transplant patients (Rochester 2008), there is a lack of detailed guidelines on how it should be carried out. These recommendations do not give details of which specific techniques should be used either at early phases or at late phases to accomplish the objective of the treatment. These limitations, added to the variety of protocols used in observational and experimental studies, have hampered the definition of the appropriate exercise dosing regarding intensity, frequency, and appropriate duration of each session and of the full program, as well as the definition of which type of training should be carried out, namely aerobic, resistance, or interval training among other more current training modalities.

How the intervention might work

Patients undergoing lung transplantation, regardless of the underlying disease, should experience a spontaneous and gradual improvement in their overall fitness, specifically in their pulmonary function, muscle strength and exercise capacity. Improvements in the overall fitness should have a positive impact on the survival and the quality of life. However, certain limitations on physical capacity have been found up to one or two years after transplantation (Langer 2009; Reinsma 2006; Williams 1992).

The origin of these limitations may occur due to physical deconditioning of lung transplant candidates and the effect of prolonged rest and immunosuppressive treatment after the transplant. The common factors to both phases is the presence of sarcopenia (Rozenberg 2014) and a decrease in the maximal oxygen uptake (VO₂max) which is partly because the muscles use a small amount of oxygen (Evans 1997). These limitations may be resolved or minimized by incorporating exercise training, because at a cellular level it will increase the maximum rate of mitochondrial respiration and the percentage of type I fibres and their resistance to fatigue (Guerrero 2005).

Exercise training has proven to have a beneficial effect on abnormalities caused by the underlying respiratory diseases. In the case of COPD, exercise has shown to improve symptoms of dyspnoea and fatigue, and a subsequent positive impact on health‐related quality of life (HRQoL) and exercise capacity (McCarthy 2015). The same results have been described for interstitial lung disease and idiopathic pulmonary fibrosis (Dowman 2021).

In addition, exercise training post‐transplant has shown positive results in other solid organ transplantations. For example, in cardiac and kidney transplantation an improvement in muscle strength and peak oxygen uptake (VO₂peak) has been observed (Hsieh 2011; Kouidi 2013), and in the case of liver transplantation, exercise improves daily physical capacity and HRQoL (van Ginneken 2010). According to these results, exercise training may have similar results in patients undergoing lung transplantation.

Why it is important to do this review

Limitations on physical capacity in post‐lung transplant patients usually leads to physical, mental and social problems that may increase direct and indirect healthcare costs.

In the latest update of the ATS‐ERS guidelines on pulmonary rehabilitation (Spruit 2013) the importance of incorporating an exercise training program in the management of lung transplant patients is emphasised. However, the quality of the evidence (randomised controlled trials (RCTs), non‐randomised prospective cohort) supporting this recommendation is low to moderate (Wickerson 2010) and does not consider new experimental studies published recently. Additionally, the dosing and types of exercises that are most effective in this population have not been clearly defined due to the wide variability in training protocols. Considering these facts and that the most recent review on this topic only assesses critically the VO₂max in the subgroup of lung transplant recipients (Didsbury 2013), it is necessary to update, summarize and systematically review the evidence of exercise training in adult lung transplant recipients.

Objectives

To assess the benefits and harms of exercise training in adult patients that have undergone lung transplantation; more specifically, the effects on the maximal, and functional exercise capacity, HRQoL, and adverse events associated to the intervention. Second, we evaluated the effects on other variables such as patient readmission, pulmonary function, muscular strength, pathological bone fractures, return to normal activities, and death.

Methods

Criteria for considering studies for this review

Types of studies

RCTs and quasi‐RCTs (RCTs in which allocation to treatment was obtained by alternation, use of alternate medical records, date of birth or other predictable methods) looking at the effects of exercise training in patients with lung transplantation. We included studies of any publication status, including conference abstracts. Publications must be peer‐reviewed.

Types of participants

We included studies of adult patients (over 18 years) with either unilateral or bilateral lung transplantation, regardless of the underlying pulmonary disease (e.g. COPD, cystic fibrosis, interstitial lung disease). We excluded studies in which the majority of the participants needed another surgical procedure in addition to the lung transplant within the same intervention, such as heart‐lung transplantation, transplantation of another solid organ (e.g. kidney, liver) and heart surgery (e.g. valve replacement, myocardial revascularization).

Types of interventions

We included studies assessing an exercise training program. We adopted the definition of exercise provided by Caspersen 1985: all planned, structured, and repetitive physical activity that has as a final or an intermediate objective the improvement or maintenance of physical fitness. This training can consist of aerobic (continuous or interval training), resistance (anaerobic), or it could be multimodal training, which means a combination of these modalities.

Aerobic exercise is defined as carrying out activities continuously for long periods of time with the aim of conditioning the muscles of ambulation and improving cardiorespiratory fitness to increase physical activity along with the decrease in dyspnoea and fatigue (e.g. walking, running, swimming, and cycling) (Spruit 2013). Interval training involves repeated short (< 45 seconds) to long (2 to 4 minutes) bouts of exercise interspersed with recovery periods (Buchheit 2013). Resistance exercise targets specific muscle groups that are trained by repeatedly lifting relatively heavy weights (American College of Sports Medicine 2009).

All exercise training program was eligible, regardless of frequency, intensity, length of each training session or full program, whether program starts before or after discharge of hospital, where it is carried out at the gym or home, whether the program is supervised or not, and whether the patients have joined another training program before lung transplantation.

Studies including associated interventions such as drug therapy, diet or any other element involved in pulmonary rehabilitation (e.g. educational interventions, psychosocial support, and respiratory physiotherapy) were excluded if the co‐interventions were not equally implemented in both groups.

We planned to investigate the following comparison pairs from the available data:

Continuous (aerobic) exercise training versus usual care or no exercise training
Continuous (aerobic) exercise training versus resistance (anaerobic) exercise training
Continuous (aerobic) exercise training versus interval (aerobic) exercise training
Continuous (aerobic) exercise training versus multimodal exercise training
Continuous (aerobic) exercise training versus continuous (aerobic) exercise training
Resistance (anaerobic) exercise training versus usual care or no exercise training
Resistance (anaerobic) exercise training versus interval (aerobic) exercise training
Resistance (anaerobic) exercise training versus multimodal exercise training
Resistance (anaerobic) exercise training versus resistance (anaerobic) exercise training
Interval (aerobic) exercise training versus usual care or no exercise training
Interval (aerobic) exercise training versus multimodal exercise training
Interval (aerobic) exercise training versus interval (aerobic) exercise training
Multimodal exercise training versus usual care or no exercise training
Multimodal exercise training versus multimodal exercise training

Types of outcome measures

Primary outcomes

Maximal and functional exercise capacity: measured during either lab exercise tests (e.g. cardiopulmonary exercise test (CET), incremental shuttle walking test (ISWT)) or field exercise tests (e.g. six‐minute walk test (6MWT)).
Health‐related quality of life (HRQoL): measured by validated generic or specific questionnaires (e.g. SF‐36).
Adverse events: untoward haemodynamic or respiratory changes, musculoskeletal complications, postoperative complications such as wound dehiscence and thoracic instability, and any other adverse event.

Secondary outcomes

Patient readmission: defined as a subsequent hospitalisation after lung transplant hospitalisation.
Pulmonary function: measured by specific tests assessing the performance of the respiratory system (e.g. spirometry, lung plethysmography).
Muscular strength: measured by a dynamometer for assessing static strength, by calculating one‐repetition maximum (1RM) for assessing the maximal weight an individual can lift for only one repetition with correct technique or any other method reported.
Pathological bone fractures: defined as bone fractures confirmed by imaging tests (e.g. X‐ray, MRI, CT). In case, they were no reported, we used bone mineral density (BMD) measured by specific tests (e.g. dual‐energy X‐ray absorptiometry (DEXA)) as an indirect measure.
Return to normal activities: extent (in percentage) to which the patient resumes their normal activities, or median time elapsed from lung transplantation until return to normal activities or work.
Death (any cause): defined as death due to any cause or specifically due to a respiratory cause, related or not to rejection of the transplanted organ.

Search methods for identification of studies

Electronic searches

We searched the Cochrane Kidney and Transplant Register of Studies up to 6 October 2020 through contact with the Information Specialist using search terms relevant to this review. The Register contains studies identified from the following sources:

Monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL)
Weekly searches of MEDLINE OVID SP
Handsearching of kidney‐related journals and the proceedings of major kidney conferences
Searching of the current year of EMBASE OVID SP
Weekly current awareness alerts for selected kidney and transplant journals
Searches of the International Clinical Trials Register (ICTRP) Search Portal and ClinicalTrials.gov.

Studies contained in the Register are identified through searches of CENTRAL, MEDLINE, and EMBASE based on the scope of Cochrane Kidney and Transplant. Details of search strategies, as well as a list of handsearched journals, conference proceedings and current awareness alerts, are available on the Cochrane Kidney and Transplant website under CKT Register of Studies.

There were no restrictions based on language or date of publication.

See Appendix 1 for search terms used in strategies for this review.

Searching other resources

Reference lists of review articles, relevant studies and clinical practice guidelines.
Letters seeking information about unpublished or incomplete studies to investigators known to be involved in previous studies.

Data collection and analysis

Selection of studies

The described search strategy was used to obtain titles and abstracts of studies that could be relevant to the review. The titles and abstracts were screened independently by two authors, who discarded studies that were not applicable. Disagreements were solved by consensus or by a third author when consensus was not achieved. Studies and reviews that might include relevant data or information on studies were retained initially. Two authors independently assessed the full text of these studies to determine which of them met the inclusion criteria. Disagreements were solved by consensus or by a third author when consensus was not achieved.

Data extraction and management

Data extraction was carried out independently by two authors using standard data extraction forms. Disagreements were solved by consensus or a third author when consensus was not achieved. Studies reported in non‐English language journals were translated before assessment. Where more than one publication of one study existed, reports were grouped together and the publication with the most complete data was used in the analyses. Where relevant outcomes were only published in earlier versions these data were used. Any discrepancy between published versions was highlighted. If data were missing from reports, or if clarification was needed, the study authors were contacted to obtain the missing information.

Assessment of risk of bias in included studies

Two review authors, independently assessed the risk of bias of the eligible studies. Disagreements were resolved either by consensus or by a third author.

We used the risk of bias assessment tool (Higgins 2011) (see Appendix 2).

Was there adequate sequence generation (selection bias)?
Was allocation adequately concealed (selection bias)?
Was knowledge of the allocated interventions adequately prevented during the study?
- Participants and personnel (performance bias)
- Outcome assessors (detection bias)
Were incomplete outcome data adequately addressed (attrition bias)?
Are reports of the study free of suggestion of selective outcome reporting (reporting bias)?
Was the study apparently free of other problems (conflicts of interest and sources of funding declared) that could put it at a risk of bias?

In this case we determined the inclusion of one additional domain, the balance in the baseline characteristics between the groups. For this additional domain, low risk of bias was considered when major characteristics, potential prognostic factors, were reported as being balanced between groups (e.g., baseline data reported in "Table 1" without statistically significant differences); and high risk of bias when there was evidence of a clear imbalance in the main prognostic factors which was not controlled by any statistical strategy, as for example the use of change scores or control of confusing variables. Unclear risk of bias was considered when there was insufficient information to determine whether there is a low or high risk of bias.

According to the recommendations of the Cochrane Collaboration (Higgins 2011), in the case that all domains mentioned above were considered with low risk of bias, the study was considered with a low risk of global bias. In the event that at least one of the domains of one study was considered to have an unclear or high risk of bias, the risk of overall bias of that study was considered unclear or high, respectively. In addition, because of the nature of the intervention, it is not possible to achieve blinding of study participants and health personnel, so a low, unclear or high risk of bias was considered, depending on whether the different outcomes measured in each study is likely to be influenced by the lack of such blinding.

Measures of treatment effect

We planned to calculate the risk ratio (RR) with 95% confidence intervals (CI) for dichotomous outcomes (death, adverse effects, patient readmission, and pathological bone fractures) and for continuous outcomes (HRQoL, pulmonary function, maximal and functional exercise capacity, and muscular strength) the mean difference (MD) or the standardised mean difference (SMD) if different scales of a same measure were used with 95% CI.

It was not possible to carry out such analysis, however they will be considered for future updates of this systematic review.

Unit of analysis issues

The unit of analysis was the participant randomised in the included studies. A single measurement for each outcome from each participant was collected and analysed.

We planned to include only the outcome data of the first part of the study for studies with crossover design (i.e. before the crossover). In addition, if cluster‐RCTs were included, we planned to analyse the effective sample size, or the inflated standard error method based on viability and availability of the outcome data. However, the studies included in this review did not present such special features, but these will be considered in future updates of this systematic review.

Dealing with missing data

Any further information required from the original author was requested by e‐mail and any relevant information obtained in this manner was included in the review. However, satisfactory responses to the requirements were not obtained.

Evaluation of important numerical data such as screened, randomised patients as well as intention‐to‐treat, as‐treated and per‐protocol population was carefully performed. Attrition rates, for example dropouts, losses to follow‐up and withdrawals were investigated. Issues of missing data and imputation methods (for example, last‐observation‐carried‐forward) was critically appraised (Higgins 2011).

Assessment of heterogeneity

We planned to analyse heterogeneity using a Chi² test on N‐1 degrees of freedom, with an alpha of 0.05 used for statistical significance and with the I² test (Higgins 2003). We considered I² values of 25%, 50% and 75% as low, medium and high levels of heterogeneity, respectively. However, it was not possible to carry out such calculations, so it will be considered for subsequent versions of this systematic review.

Assessment of reporting biases

We planned to perform a funnel plots to assess the potential publication bias If 10 studies or more were meta‐analysed (Higgins 2011). However, due to the number of included studies and because it was not possible to perform some meta‐analysis, so this assessment will be considered in future versions of this review where possible.

Data synthesis

We assessed whether it was appropriate to pool data based on the similarity between participants’ characteristics, or the methods used to measure the outcomes. If the eligible studies had been sufficiently comparable, we planned to perform a meta‐analysis using a fixed‐effect model in the absence of clinical heterogeneity and a random‐effects model if clinical or statistical heterogeneity was likely. Potentially six studies could have been pooled in three different meta‐analysis: Braith 2007 and Mitchell 2003 were contacted, but we did not get a response, so we decided not to perform a meta‐analysis because of the risk of duplicating the information; Gloeckl 2015 and Gloeckl 2017 included participants with mean of 3 months and 5.5 years from transplant, respectively; and Langer 2012 and Ulvestad 2020 include different aerobic component of multimodal training. So, it was not possible or appropriate to carry out such analysis, so it will be considered in future updates of this systematic review.

Subgroup analysis and investigation of heterogeneity

In future versions of this review, and provided there are studies that allow subgroup analysis, these analysis will be performed according to the characteristics of the study participants and variations in the exercise training program.

Heterogeneity among participants would be assessed by analysing the following potential subgroups.

Age: (1) 18 to 40 years; (2) 41 to 55 years; (3) 56 years or over
Baseline pathology: (1) cystic fibrosis; (2) COPD; (3) idiopathic pulmonary fibrosis; (4) other.

Heterogeneity in treatments would be assessed by analysing the following potential subgroups.

Exercise intensity: (1) moderate; (2) vigorous. Defined according to the intensity of the exercise training used in the primary studies
Length of the exercise program: (1) short‐term; (2) long‐term. Defined according to the duration of the exercise training program used in the primary studies
Supervision: (1) supervised; (2) non‐supervised; (3) mixed (supervised and non‐supervised)
Hospital status: (1) inpatient; (2) outpatient; (3) mixed (inpatient and outpatient)
Beginning of the training program after lung transplantation: (1) < 1 year; (2) ≥ 1 year.

In the updates of this review, adverse effects will be tabulated and assessed with descriptive techniques, since they may differ across studies depending on the training program. Where possible, the risk difference (RD) with 95% CI will be calculated for each adverse effect, either compared to no treatment or other type of exercise training.

Sensitivity analysis

For future versions of this review, the following sensitivity analyses would be attempted to explore the influence of the following factors on the estimation of the size of the effect:

Excluding unpublished studies
Analysing only those studies with low risk of bias. We will consider low risk of bias when all domains except for blinding of participants and personnel are at low risk of bias.
Excluding studies with a total attrition rate of over 30%, or where differences in attrition between groups exceed 10%
Excluding any very large studies or long follow‐up to establish how much they dominate the results

Summary of findings and assessment of the certainty of the evidence

We presented the main results of the review in 'Summary of findings' tables in a narrative form (Murad 2017). These tables present key information concerning the quality of the evidence, the magnitude of the effects of the interventions examined, and the sum of the available data for the main outcomes (Schunemann 2011a). The 'Summary of findings' tables also include an overall grading of the evidence related to each of the main outcomes using the GRADE (Grades of Recommendation, Assessment, Development and Evaluation) approach (Guyatt 2008; Guyatt 2011a). The GRADE approach defines the quality of a body of evidence as the extent to which one can be confident that an estimate of effect or association is close to the true quantity of specific interest. The quality of a body of evidence involves consideration of within‐study risk of bias (methodological quality), directness of evidence, heterogeneity, precision of effect estimates and risk of publication bias (Schunemann 2011b).

Two authors assessed the quality of evidence of the following outcomes.

Maximal and functional exercise capacity
HRQoL
Adverse events
Pulmonary function
Muscular strength
Pathological bone fractures. In case, they were no reported, we used BMD as an indirect measure.
Death (any cause).

Results

Description of studies

Results of the search

The study selection process is shown in the PRISMA flowchart (Figure 1).

Figure 1

Study flow diagram.

Our search identified 38 records. After screening titles and abstracts 10 records were excluded. After full‐text review we included eight studies (17 records), two studies (three records) were excluded, and six ongoing studies (six records) were identified (NCT00129350; NCT00753155; NCT00808600; NCT01162148; NCT03728257; NCT03873597). These six studies will be assessed in a future update of this review.

We contacted the authors of all ongoing studies to obtain additional information, but we did not receive any responses (see Characteristics of ongoing studies).

Included studies

See Characteristics of included studies.

Study design

Eight RCTs (438 participants) met our inclusion criteria (Braith 2007; Fuller 2017; Fuller 2018; Gloeckl 2015; Gloeckl 2017; Langer 2012; Mitchell 2003; Ulvestad 2020).

Participants

The mean age of participants was 54.9 ± 9.6 years and 51.9% were men. The most common respiratory diseases that led participants to undergo lung transplant in the studies were COPD (46.2% of the total), interstitial lung disease (25.2%), cystic fibrosis (6.5%) pulmonary emphysema (5.8%), and pulmonary fibrosis (5.8%). Two studies were conducted in the USA (Braith 2007; Mitchell 2003), two in Australia (Fuller 2017; Fuller 2018), two in Germany (Gloeckl 2015; Gloeckl 2017), one in Belgium (Langer 2012), and one in Norway (Ulvestad 2020).

The median number of patients included was 60 (range from 16 to 83). Calculation of sample size was performed in six studies (Fuller 2017; Fuller 2018; Gloeckl 2015; Gloeckl 2017; Langer 2012; Ulvestad 2020).

Interventions

All included studies were parallel. Seven RCTs had two comparison groups (Fuller 2017; Fuller 2018; Gloeckl 2015; Gloeckl 2017; Langer 2012; Mitchell 2003; Ulvestad 2020) and one study had three comparison groups (Braith 2007).

Three studies compared the effects of resistance (anaerobic) exercise training and another resistance (anaerobic) exercise training (Fuller 2018; Gloeckl 2015; Gloeckl 2017). In Fuller 2018 the treatment group included a supervised upper limb program compared to unsupervised upper limb program, and in Gloeckl 2015 and Gloeckl 2017, the treatment group underwent multimodal exercise on a vibratory platform for squatting while the control group performed the same exercises on the floor.
One study compared the effects of two multimodal exercise training differing in the duration of the total program (14 weeks versus 7 weeks) (Fuller 2017).
Two studies compared multimodal exercise training with usual care or no exercise training (Langer 2012; Ulvestad 2020).
Two studies compared the effects of resistance (anaerobic) exercise training with usual care or no exercise training, where the experimental groups performed specific lumbar column strength training (Braith 2007; Mitchell 2003). In Braith 2007 (three‐armed study), the effects of a resistance (anaerobic) exercise training program was added to the consumption of alendronate (used for the fixation of calcium in people with osteoporosis or who consumed immunosuppressants), compared to a group that received alendronate only, and a group that only underwent usual care or no exercise training.

The median duration of the training program for the intervention groups was 13 weeks (range 4 to 24 weeks), and the median duration for active control groups was 4 weeks, (range 4 to 12 weeks). The training programs were all performed in hospital gyms when patients had been discharged.

Three studies included resistance training within their intervention program as a differentiating factor (Braith 2007; Fuller 2018; Mitchell 2003).
Five studies considered both aerobic (continuous or interval) and resistance training (i.e. multimodal training, within their exercise program) (Fuller 2017; Gloeckl 2015; Gloeckl 2017; Langer 2012; Ulvestad 2020). The most frequent resistance training was repeatedly lifting relatively heavy weights (Braith 2007; Fuller 2018; Gloeckl 2015; Gloeckl 2017; Langer 2012; Mitchell 2003; Ulvestad 2020) and, to a lesser extent, functional exercise training (Fuller 2017).
Aerobic training consisted of treadmill exercise with an initial velocity of 70% to 75% of the speed reached in 6MWT (Fuller 2017; Fuller 2018; Langer 2012); interval exercise with 4‐minute exercise bouts at 85% to 95% of maximum heart rate, with 2 minutes of active recovery in between (Ulvestad 2020); cycle ergometer exercise with an initial load of 60% of the basal workload (Gloeckl 2015; Langer 2012); or 20 to 25 Watts (Fuller 2018).

The median follow‐up time of the studies was six months with a range of one to 12 months.

Outcomes

For the primary outcomes, six studies reported maximal exercise capacity, functional exercise capacity, and HRQoL (Fuller 2017; Fuller 2018; Gloeckl 2015; Gloeckl 2017; Langer 2012; Ulvestad 2020). All studies reported adverse events.

For the secondary outcomes, two studies reported patient readmission (Fuller 2018; Fuller 2017), four studies reported pulmonary function (Fuller 2018; Fuller 2017; Langer 2012; Ulvestad 2020). All studies reported muscular strength. Three studies (Braith 2007; Langer 2012; Mitchell 2003) reported BMD as a indirect measure of the risk of pathological bone fractures. No data were reported on the return to normal activities and death.

Excluded studies

See Characteristics of excluded studies.

Following full‐text appraisal two studies were excluded (Ihle 2011; Warburton 2004). Ihle 2011 exercise was delivered in addition to other treatment strategies as part of a pulmonary rehabilitation program and in Warburton 2004 the control group were healthy participants.

On‐going studies

See Characteristics of ongoing studies.

We have identified six ongoing studies. Four studies included lung transplant recipients (NCT00753155; NCT01162148; NCT03728257; NCT03873597) and two studies included heart and lung transplant recipients (NCT00129350; NCT00808600). These two studies will be considered for inclusion because of the possibility of reporting separate results for lung transplant recipients.

Risk of bias in included studies

The overall risk of bias was considered high, as the information comes from studies with at least one domain considered as high risk of bias. See Figure 2 and Figure 3.

Figure 2

Risk of bias summary: review authors' judgements about each risk of bias item for each included study (blank sections indicate outcome not reported)

Figure 3

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies (blank sections indicate outcome not reported)

Allocation

Random sequence generation

Random sequence generation was considered to be at unclear risk of bias in two studies (Braith 2007; Mitchell 2003), and the remaining six studies were judged to be at low risk of bias (Fuller 2017; Fuller 2018; Gloeckl 2015; Gloeckl 2017; Langer 2012; Ulvestad 2020). Studies were stratified according to the underlying pathology (Fuller 2017), type of surgical incision (Fuller 2018), the distance walked in 6MWT (Gloeckl 2015; Gloeckl 2017) and by gender and episodes of early acute rejection (Langer 2012). Ulvestad 2020 was not stratified.

Allocation concealment

Allocation concealment was judged to be at unclear risk of bias in three studies (Braith 2007; Mitchell 2003; Ulvestad 2020) and five studies were judged to be low risk of bias (Fuller 2017; Fuller 2018; Gloeckl 2015; Gloeckl 2017; Langer 2012). Three studies used sealed opaque envelopes as a method of concealment (Fuller 2017; Fuller 2018; Langer 2012), and two studies used the telephone call method to an external centre that handled the randomisation list (Gloeckl 2015; Gloeckl 2017).

Blinding

Performance bias

Studies basing their interventions on exercise training cannot achieve blinding of participants or health personnel involved in exercise training. However, it is possible to establish with some certainty whether the lack of blinding affects the estimation of the effect depending on the characteristics of the outcomes evaluated in each study.

For our primary outcomes, maximal and functional exercise capacity and HRQoL, were considered likely to be influenced by lack of blinding, especially health personnel’s, and therefore assessed to be at high risk of bias if the interventions were not concealment. It is unlikely that the lack of blinding would affect the measurement of adverse effects and was therefore considered to have a low risk of bias. For the secondary outcomes, pathological bone fractures (measured indirectly by BMD), muscular strength and pulmonary function are variables that were probably influenced by the lack of blinding, hence their estimation was considered to be at high risk of bias, while patient readmission was considered to be at low risk of bias.

Detection bias

Three studies reported the presence of blinding for all outcomes (Fuller 2017; Fuller 2018; Langer 2012) and another only for assessing the maximal and functional exercise capacity (Gloeckl 2015). Four studies did not report blinding of outcomes assessment (Braith 2007; Gloeckl 2015; Mitchell 2003; Ulvestad 2020).

Maximal and functional exercise capacity

Four studies were judged to be at low risk of bias for maximal and functional exercise capacity (Fuller 2017; Fuller 2018; Gloeckl 2015; Langer 2012).

Fuller 2017, Fuller 2018 and Gloeckl 2015 were judged to be at low risk of bias for the distance walked in the 6MWT
Langer 2012 was judged to be at low risk of bias for VO₂max and the maximum workload (Wmax)

Two studies were judged to be at high risk of bias for maximal and functional exercise capacity (Gloeckl 2017; Ulvestad 2020).

Gloeckl 2017 was judged to be at high risk of bias for the distance walked in the 6MWT, peak work rate, and functional sit‐to‐stand test.
Ulvestad 2020 was judged to be at high risk of bias for the 15 second stair run and 30 second chair stand reported.

Health‐related quality of life

Six studies reporting HRQoL were judged to be at high risk of bias (Fuller 2017; Fuller 2018; Gloeckl 2015; Gloeckl 2017; Langer 2012; Ulvestad 2020) as the assessment instrument was self‐administered by study participants.

Adverse events

Only Fuller 2018 was judged to be at high risk of bias for adverse events as sternal instability, measured with a subjective scale, could not blind the evaluators and therefore could introduce detection bias. All other studies were judged to be at low risk of bias.

Patient readmission

Two studies reported patient readmission and were both judged to be at low risk of bias (Fuller 2017; Fuller 2018).

Pulmonary function

Two studies blinded the evaluators (Fuller 2017; Langer 2012) and were judged to be at low risk of bias, while a third study (Ulvestad 2020) did not blind the evaluators and was judged to be at high risk of bias.

Muscular strength

The estimation of muscular strength, for both limbs and respiratory muscles, was judged to be at low risk of bias in studies with blinding for three studies (Fuller 2017; Fuller 2018; Langer 2012) and high risk of bias in five studies (Braith 2007; Gloeckl 2015; Gloeckl 2017; Mitchell 2003; Ulvestad 2020).

Pathological bone fractures

Three studies reported BMD as an indirect measure of the risk of pathological bone fractures and were judged to be at low risk of bias (Braith 2007; Langer 2012; Mitchell 2003).

Incomplete outcome data

Dropout rates during follow‐up varied across the studies. Two studies were judged to be at high risk of bias. Braith 2007 had a 40% difference in the dropouts between the relevant groups for this review, and Gloeckl 2015 reported 23% of dropouts in the experimental group compared to 8% in the control group.

Two studies were judged to have unclear risk of bias. In Mitchell 2003 it was not clear whether all participants completed the study, and Fuller 2018 failed to detail the reasons for the follow‐up losses for each group.

Four studies were judged to be at low risk of bias; dropouts were ≤ 10% and the reasons were balanced across the groups (Fuller 2017; Gloeckl 2017; Langer 2012; Ulvestad 2020).

Selective reporting

Six studies registered the protocol prior to their final publication (Fuller 2017; Fuller 2018; Gloeckl 2015; Gloeckl 2017; Langer 2012; Ulvestad 2020). These studies lacked consistency between the protocol and the published results. In the published reports some changes included changing the way of reporting the measurements and showing the differences between the final and baseline values (Fuller 2017; Gloeckl 2015; Gloeckl 2017); reporting percentages of predicted values (Gloeckl 2015; Langer 2012); only reporting specific domains of HRQoL (Fuller 2017); a change in the measurement instrument of functional exercise capacity (Ulvestad 2020) and HRQoL (Gloeckl 2015; Gloeckl 2017); and inclusion of two additional non‐pre‐specified outcomes in the published protocol (Fuller 2018), however these two outcomes evaluated the effects of an exercise training program; therefore, this study was considered at low risk of bias (Fuller 2018). All other studies were judged to be at high risk of bias; two of these studies did not report relevant results, such as the functional capacity to perform exercise and the HRQoL (Braith 2007; Mitchell 2003), and one study only graphically presented muscle strength (not absolute values), which prevented us from performing a meta‐analysis (Braith 2007).

Other potential sources of bias

Initial balance between groups

No initial imbalance between groups was reported in terms of characteristics that could influence the results of the eight studies included in this review.

Declarations of interests/disclosures declared

Declarations of interest/disclosures were not reported in two studies (Braith 2007; Mitchell 2003) and were judged to be at unclear risk of bias. All other studies were judged to be at low risk of bias.

Sources of funding declared

Braith 2007 and Mitchell 2003 were the only studies that did not report the existence or not of conflicts of interest, and other hand Mitchell 2003 and Fuller 2018 did not report whether or not they had sources of funding.

Effects of interventions

We did not perform meta‐analyses due to the wide heterogeneity of the intervention and outcomes, and uncertainty of some of the data. Braith 2007 and Mitchell 2003 were contacted to clarify if they were two separate studies or if one was an extension of the other; no response was received. Gloeckl 2015 and Gloeckl 2017 included participants with mean time from transplantation of three months and 5.5 years, respectively. Langer 2012 and Ulvestad 2020 included different aerobic component of multimodal training.

For the results and outcomes of the eight individual included studies see the additional data tables in Data and analyses.

The summaries of the data are presented below.

Resistance (anaerobic) exercise training versus usual care or no exercise training

Two studies compared a resistance (anaerobic) exercise training program to usual care or no exercise training (Braith 2007 (20 participants); Mitchell 2003 (16 participants). For details of the certainty of the evidence see summary of findings Table 1.

Adverse events

Any adverse event

Braith 2007 reported 4/10 participants in the resistance (anaerobic) exercise training plus alendronate group, and no participants in the alendronate alone group experienced adverse events at 6 months (see Analysis 1.1).

Acute graft rejection

Braith 2007 reported no significant differences (P > 0.05) between the groups for acute graft rejection episodes 6 months after treatment (8‐months post‐transplant).

Mitchell 2003 reported no significant differences (P > 0.05) between the groups for acute graft rejection episodes after 6 months of treatment (8‐months post‐transplant).

We consider the certainty of evidence for adverse events as very low. The quality of the evidence was downgraded one level due to high risk of bias (selection, reporting and other bias), one level for imprecision and one level for indirectness.

Muscular strength

Braith 2007 reported significant increases (P ≤ 0.05) in lumbar extensor muscle strength at all seven testing positions at all seven testing positions (see Analysis 1.2).

Mitchell 2003 reported significant increases (P ≤ 0.05) in lumbar extensor muscle strength in the resistance training group at the final three testing positions (48, 60 and 72 degrees of lumbar flexion) compared to the control group.

We considered the certainty of evidence for muscular strength as very low. The quality of the evidence was downgraded one level due to high risk of bias (selection, performance, detection, reporting and other bias), one level for inconsistency and one level for imprecision.

Pathological bone fractures

Pathological bone fractures were measured indirectly as BMD (see Analysis 1.3).

Braith 2007 reported a significant improvement in BMD in the resistance training group but not in the control group at 8 months.

Mitchell 2003 reported a significant improvement in BMD in the resistance training group compared to the control group at 8 months (P < 0.05). Both groups lost BMD in the first 2 months post‐transplant. BMD continued to fall in the control group while the resistance exercise group regained BMD and levels returned to within 5% of pre‐transplant levels.

We consider the certainty of evidence for pathological bone fractures as very low. The quality of the evidence was downgraded one level due to high risk of bias (selection, performance, reporting and other bias), and one level for indirectness and one level for imprecision.

Other outcomes

No data were reported for maximal and functional exercise capacity, HRQoL, patient readmission, pulmonary function, return to normal activities, or death.

Resistance (anaerobic) exercise training versus another form of resistance (anaerobic) exercise training

Three studies compared a resistance (anaerobic) exercise training program to another resistance (anaerobic) exercise training (Fuller 2018 (80 participants); Gloeckl 2015 (83 participants); Gloeckl 2017 (79 participants)). Gloeckl 2015 and Gloeckl 2017 focused on different strategies for lower limb training ‐ squat exercises on a side‐alternating, whole‐body vibration platform (WBVT) versus same amount of exercise time on the floor (CON), and Fuller 2018 focused on upper limbs ‐ a structured, supervised, upper limbs program (SULP) versus no supervised upper limbs program (NULP).

For details of the certainty of the evidence see summary of findings Table 2.

Maximal and functional exercise capacity

Fuller 2018 reported the functional exercise capacity at 3 to 6 months, and Gloeckl 2015 and Gloeckl 2017 at 4 weeks. Gloeckl 2015 and Gloeckl 2017 reported the maximal exercise capacity at 4 weeks.

Gloeckl 2015 and Gloeckl 2017 both reported a significant improvement in 6MWT in both the WBVT and CON groups at 4 weeks. However, Gloeckl 2015 and Gloeckl 2017 both reported the WBVT groups walked significantly further than the CON groups: 28.4 metres (95% CI 3 to 53.7; P = 0.029) and 28.3 metres (95% CI 10.0 to 46.6; P < 0.05), respectively.

Fuller 2018 reported no significant differences in 6MWT between SULP and NULP group at 3 months (SULP group: 554.9 ± 82.6 metres; NULP group: 511.1 ± 112.8 metres; P = 0.06) but there was a significant improvement in the SULP group at 6 months (SULP group: 561.2 ± 83.6 metres; NULP group: 503.5 ± 115.2 metres; P = 0.01).

Gloeckl 2015 and Gloeckl 2017 reported significantly higher peak work rate at 4 weeks in both the WBVT groups and the CON groups at 4 weeks. However, Gloeckl 2015 and Gloeckl 2017 both reported the WBVT groups achieved a significantly higher peak work rate than the CON groups of 5.2 watts (95% CI 0.2 to 10.2; P = 0.042) and in 4.5 watts (95% CI 0.13 to 8.9; P < 0.05), respectively.

We considered the certainty of evidence for maximal and functional exercise capacity as very low. The quality of the evidence was downgraded two levels due to high risk of bias (performance, detection, attrition, reporting and other bias) and one level for imprecision.

Health‐related quality of life

Fuller 2018 reported the HRQoL at 3 to 6 months, and Gloeckl 2015 and Gloeckl 2017 at 4 weeks.

Gloeckl 2015 and Gloeckl 2017 reported significant improvement in Chronic Respiratory Questionnaire (CRQ) components for both the WBVT and CON groups at 4 weeks (see Analysis 2.2.1). However, Gloeckl 2015 and Gloeckl 2017 both reported no significant differences between the two groups in CRQ questionnaire components.

Gloeckl 2015
- CRQ dyspnoea mean difference: 0.18 points (95% CI ‐0.45 to 0.81; P = 0.569)
- CRQ fatigue mean difference: ‐0.10 points (95% CI ‐0.52 to 0.32; P = 0.644)
- CRQ emotional function mean difference: ‐0.08 points (95% CI ‐0.51 to 0.35; P = 0.701)
- CRQ emotional mastery mean difference: 0.18 points (95% CI ‐0.26 to 0.63; P = 0.412)
Gloeckl 2017
- CRQ dyspnoea mean difference: 0.03 points (95% CI ‐0.56 to 0.61 points; P > 0.05)
- CRQ fatigue mean difference: ‐0.06 points (95% CI ‐0.7 to 0.57; P > 0.05)
- CRQ emotional function mean difference: ‐0.03 points (95% CI ‐0.59 to 0.52 ; P > 0.05)
- CRQ emotional mastery mean difference: 0.19 points (95% CI ‐0.32 to 0.71; P > 0.05).

Fuller 2018 reported no significant differences in SF‐36 questionnaire for 'Physical Health', 'Mental Health' and the summary score at 3 and 6 months (see Analysis 2.2.2 and Analysis 2.2.3).

We considered the certainty of evidence for HRQoL as very low. The quality of the evidence was downgraded two levels due to high risk of bias (performance, detection, attrition, reporting and other bias), and one level for imprecision.

Adverse events

Gloeckl 2015 reported three acute infections in WBVT group (34 participants) and two acute infections in CON group (36 participants).
Gloeckl 2017 reported four acute infections in WBVT group (34 participants) and 2 acute infections in CON group (36 participants).
Fuller 2018 did not report on adverse events.

We consider the certainty of evidence for adverse events as very low. The quality of the evidence was downgraded two levels due to high risk of bias (detection, attrition, reporting and other bias) and imprecision, and one level due to indirectness.

Patient readmission

Fuller 2018 reported two participants were admitted to ICU during the 12‐week follow‐up period in the SULP group and none in NULP group.
Gloeckl 2015 and Gloeckl 2017 did not report patient readmission.

We considered the certainty of evidence for patient readmission as very low. The quality of the evidence was downgraded one level due to high risk of bias (reporting and other bias) and two levels due to imprecision.

Muscular strength

Fuller 2018 reported the muscular strength at 3 to 6 months, and Gloeckl 2015 and Gloeckl 2017 4 weeks.

Gloeckl 2015 and Gloeckl 2017 reported significant improvement in peak force for quadriceps and hamstrings in both the WBVT and CON groups at 4 weeks (see Analysis 2.4.1). Gloeckl 2015 (0.4 N, 95% CI ‐12.9 to 13.6; P = 0.956) and Gloeckl 2017 (21 N, 95% CI ‐0.75 to 42.8; P < 0.05) both reported no significant differences between the groups for peak force quadriceps however, Gloeckl 2015 (6.2 N, 95% CI ‐5.8 to 18.2; P = 0.304) reported no significant difference for peak force hamstrings, while Gloeckl 2017 (3.6 N, 95% CI 0.43 to 26.7; P < 0.05) reported a significant improvement in the WBVT group at 4 weeks.

Fuller 2018 reported no significant differences between NULP and SULP groups for average of 3 attempts of shoulder flexion force and 3 attempts of shoulder abduction force at 3 and 6 months (see Analysis 2.4.2 and Analysis 2.4.3).

We considered the certainty of evidence for muscular strength as very low. The quality of the evidence was downgraded two levels due to high risk of bias (performance, detection, attrition, reporting and other bias), and one level for imprecision.

Death (any cause)

Gloeckl 2015 reported one patient died in the WBVT group however, according to the authors the death did not appear to be associated with WBVT. Fuller 2018 and Gloeckl 2017 did not report deaths.

Other outcomes

No data were reported on pulmonary function, pathological bone fractures, and return to normal activities.

Multimodal exercise training versus usual care or no exercise training

Two studies compared a multimodal exercise training program to usual care or no exercise training (Langer 2012 (40 participants); Ulvestad 2020 (54 participants)). Langer 2012 compared continuous aerobic training and Ulvestad 2020 compared high‐intensity interval training, both in addition to resistance training. For details of the certainty of the evidence see summary of findings Table 3.

Maximal and functional exercise capacity

Langer 2012 and Ulvestad 2020 reported the maximal and functional exercise capacity at 3‐ and 12‐months post‐transplant, and 5 months of follow‐up, respectively.

Langer 2012 reported significant improvement in 6MWT in the multimodal exercise training group compared to the no exercise group at 3 months (P = 0.008) and at 12‐months post‐transplant (P = 0.002) (see Analysis 3.1.1).

Langer 2012 also reported no significant differences between groups for Wmax at 3 months (P = 0.093) and a significant improvement at 12‐months post‐transplant in the multimodal exercise training group compared to the no exercise group (P = 0.042). Langer 2012 also reported no significant differences between groups for VO₂max at 3 months (P = 0.149) and at 12‐months post‐transplant (P = 0.082) (see Analysis 3.1.2 and Analysis 3.1.3).

Ulvestad 2020 adjusted for baseline scores and reported no significant differences between groups for VO₂peak (P = 0.169), chair to stand test (P = 0.691), and stair run test (P = 0.791) at 5 months (see Analysis 3.1.4, Analysis 3.1.5 and Analysis 3.1.6).

We considered the certainty of evidence for maximal and functional exercise capacity to be very low. The quality of the evidence was downgraded two levels due to high risk of bias (selection, performance, detection and reporting bias) and one level for imprecision.

Health‐related quality of life

Langer 2012 and Ulvestad 2020 reported the HRQoL at 3‐ and 12‐months post‐transplant, and 5 months of follow‐up, respectively.

Langer 2012 reported no significant differences between multimodal exercise training and no exercise training group for SF36 questionnaire at 3 months in all domains: physical functioning (P = 0.213); role functioning physical (P = 0.448); pain (P = 0.959); general health (P = 0.212); energy/fatigue (P = 0.285); social functioning (P = 0.064); role functioning emotional (P = 0.341); and emotional well‐being (P = 0.808) (see Analysis 3.2.1).

Langer 2012 also reported no significant differences between multimodal exercise training and no exercise training group at 12‐months post‐transplant for pain (P = 0.811); general health (P = 0.632); energy/fatigue (P = 0.937); social functioning (P = 0.118); role functioning emotional (P = 0.978); and emotional well‐being (P = 0.385). However, they reported significant improvement in the multimodal exercise training group compared to the no exercise training at 12 months in the domains physical functioning (P = 0.039) and in role functioning physical (P = 0.011) (see Analysis 3.2.2).

Ulvestad 2020 reported ‐ after adjusting for baseline scores ‐ a significant improvement in the multimodal exercise group compared to the no exercise group for the mental component summary score (P = 0.02), but not for the physical component summary score (P = 0.328) (see Analysis 3.3).

We considered the certainty of evidence for HRQoL to be very low. The quality of the evidence was downgraded two levels due to high risk of bias (selection, performance, detection and reporting bias) and one level for imprecision.

Adverse events

Langer 2012 reported 3/21 participants dropped out of multimodal exercise training group and 2/19 participants in the no exercise training group due to severe medical complications without describing them or specifying any causal relationship with the exercise training.

Ulvestad 2020 reported 2 participants dropped out in the interval (aerobic) exercise training group due to health problems. They also reported no adverse events were reported during the exercise training, however 4 participants reported exercise‐related musculoskeletal pain.

We considered the certainty of evidence for adverse events to be very low. The quality of the evidence was downgraded two levels due to high risk of bias (selection and reporting bias), one level for indirectness and one level for imprecision.

Pulmonary function

Langer 2012 and Ulvestad 2020 reported the pulmonary function at 3‐ and 12‐months post‐transplant, and 5 months of follow‐up, respectively.

Langer 2012 reported no significant differences in FEV₁ between the multimodal exercise training and no exercise training' group at 3 month (P = 0.888), and at 12‐months post‐transplant (P = 0.615) (see Analysis 3.4.1).

Ulvestad 2020 reported no significant differences between the groups for FEV₁ adjusting for baseline score (P = 0.545) (see Analysis 3.5.2).

Ulvestad 2020 also reported no significant differences between groups for DL_CO adjusting for baseline score (P = 0.545) (see Analysis 3.5.3).

We considered the certainty of evidence for this outcome to be very low. The quality of the evidence was downgraded two levels due to high risk of bias (selection, performance, detection and reporting bias) and one level for imprecision.

Muscular strength

Langer 2012 reported quadriceps force, handgrip force and respiratory muscle strength by measuring maximal inspiratory pressure and maximal expiratory pressure at 3‐ and 12‐months post‐transplant, while Ulvestad 2020 reported 1 repetition maximum leg press, 1 repetition maximum arm press, and hand grip at 5 months of follow‐up.

Langer 2012 reported significant improvement at 3 months for quadriceps force in multimodal exercise training' group compared to the no exercise training group (P = 0.001) and at 12‐months post‐transplant (P = 0.001) (see Analysis 3.5.1).

Langer 2012 reported no significant differences between groups for handgrip force at 3 months (P = 0.854) and at 12‐months post‐transplant (P = 0.184). Ulvestad 2020 reported ‐ after adjusting for baseline scores ‐ no significant differences between groups for hand grip (P = 0.459) (see Analysis 3.5.2).

Langer 2012 reported no significant differences between groups for maximal inspiratory pressure at 3 months (P = 0.463) and at 12‐months post‐transplant (P = 0.246), and for maximal expiratory pressure at 3 months (P = 0.698) and at 12‐months post‐transplant (P = 0.653) (see Analysis 3.5.3 and Analysis 3.5.4).

Ulvestad 2020 reported ‐ after adjusting for baseline scores ‐ a significant improvement in the multimodal exercise group compared to the no exercise group for one repetition maximum leg press (P = 0.047), but not for the one repetition maximum arm press (P = 0.053) (see Analysis 3.5.5 and Analysis 3.5.6).

Pathological bone fractures

Langer 2012 reported BMD of the lumbar spine and femur at 12‐months post‐transplant. Ulvestad 2020 did not report pathological bone fractures.

Langer 2012 reported no significant differences between the groups for lumbar T‐score (P = 0.44) and femur T‐score (P = 0.61).

We considered the certainty of evidence for this outcome to be very low. The quality of the evidence was downgraded one level due to high risk of bias (reporting bias), one level for indirectness and one level for imprecision.

Death

Ulvestad 2020 reported one participant died by lung rejection in the no exercise training group. Langer 2012 did not report deaths.

Other outcomes

No data were reported on patient readmission or return to normal activities.

Fourteen‐ versus seven‐weeks multimodal exercise training

Fuller 2017 (66 participants) compared a multimodal exercise training program performed over 14 weeks to the same multimodal exercise training program performed over 7 weeks. For details of the certainty of the evidence see summary of findings Table 4.

Maximal and functional exercise capacity

Fuller 2017 reported the functional exercise capacity at 14 weeks and 6 month of treatment.

Fuller 2017 reported no significant differences between groups for 6MWT at 14 weeks (n = 62) and 6 months (n = 64) (P = 0.36; group time interaction on repeated‐measures analysis of variance, controlling for age and FEV₁ % predicted). The mean improvement in 7 weeks multimodal exercise training group was 202 ± 72 metres (n = 31), and 149 ± 169 metres (n = 31) in 14 weeks multimodal exercise training group. However, at 6‐months post‐transplant, the mean difference between the groups was 59.3 metres, favouring the 7 weeks multimodal exercise training group (95% CI, 12.9 to 131.6 metres; n = 64).

We considered the certainty of evidence for this outcome to be very low. The quality of the evidence was downgraded two levels for high risk of bias (performance and reporting bias) and one level for imprecision.

Health‐related quality of life

Fuller 2017 reported the HRQoL with the SF‐36 and AQoL questionnaires at 14 weeks and 6‐months post‐treatment.

Fuller 2017 reported no significant differences between the 14‐week and 7‐week multimodal exercise training groups for SF‐36 questionnaire at 14 weeks for the summary scores of physical health (P = 0.32) and emotional health (P = 0.74). Furthermore, there were no significant differences between the 14‐week and 7‐week multimodal exercise training groups at 6 months for the summary scores of physical health (P = 0.32), and emotional health (P = 0.74) (see Analysis 4.1).

Fuller 2017 reported no significant differences between the groups for the AQoL questionnaire at 14 weeks for independent living (P = 0.38); relations (P = 0.76); senses (P = 0.08); mental health P = 0.37; and AQoL 4D score (P = 0.07). Furthermore, they reported no significant differences between the groups for the AQoL questionnaire at 6 months for independent living (P = 0.38); relations (P = 0.76); senses (P = 0.08); mental health (P = 0.37); and AQoL 4D score (P = 0.07) (see Analysis 4.2).

We considered the certainty of evidence for this outcome to be very low. The quality of the evidence was downgraded two levels for high risk of bias (performance, detection a reporting bias) and one level for imprecision.

Adverse events

Five participants missed some assessments due to musculoskeletal problems or hospital readmission, however it was not reported to which group they belonged.

We considered the certainty of evidence for this outcome to be very low. The quality of the evidence was downgraded one level for high risk of bias (reporting bias) and two levels for imprecision.

Patient readmission

Fuller 2017 did not report absolute values of patient readmission. However, the authors reported most participants had spent a few days in the hospital (range 0 to 15 days), with no significant difference between groups for the number of inpatient days during the follow‐up period (P = 0.94).

We considered the certainty of evidence for this outcome to be very low, the quality of the evidence was downgraded one level for high risk of bias (reporting bias) and two levels for imprecision.

Pulmonary function

Fuller 2017 did not report absolute values of pulmonary function. However, the authors reported that respiratory function improved in both groups however there were no significant differences in FEV₁ (P = 0.98) and forced vital capacity (P = 0.7) between the treatment groups.

We consider the certainty of evidence for this outcome to be very low. The quality of the evidence was downgraded one level for high risk of bias (reporting bias) and two levels for imprecision.

Muscular strength

Fuller 2017 reported quadriceps and hamstrings strength (average of 6 repetitions and peak torque of right leg) at 14 weeks and 6‐months post‐treatment.

Fuller 2017 reported no significant differences between the two groups at 14 weeks for average quadriceps strength (P = 0.59), and peak torque quadriceps strength (P = 0.62). The study also reported no significant differences between the two groups at 6 months for average quadriceps strength (P = 0.59), and peak torque quadriceps strength (P = 0.62) (see Analysis 4.3).

Fuller 2017 reported no significant differences between the two groups at 14 weeks for average hamstring strength (P = 0.36), and peak torque hamstrings strength (P = 0.26). They also reported no significant differences between the two groups at 6 months for average hamstring strength (P = 0.36) and peak torque hamstring strength (P = 0.26) (see Analysis 4.3).

Death (any cause)

Fuller 2017 reported one acute pneumonia death in the 14‐week multimodal exercise training group.

We considered the certainty of evidence for this outcome to be very low. The quality of the evidence was downgraded two levels for high risk of bias (reporting bias) and one level for imprecision.

Other outcomes

No data were reported on pathological bone fractures and return to normal activities.

Discussion

Summary of main results

Eight studies were identified, with a total of 438 adults undergoing lung transplantation. The studies compared: resistance (anaerobic) exercise training versus usual care or no exercise training (Braith 2007; Mitchell 2003); two different resistance (anaerobic) exercise training programs differentiated by the use of a vibrating platform (Gloeckl 2015; Gloeckl 2017), and a supervised upper limb training program (Fuller 2018); multimodal exercise training versus usual care or no exercise training (Langer 2012; Ulvestad 2020); and the effect of the same multimodal exercise training but of different duration (14 weeks versus 7 weeks) (Fuller 2017).

For the primary outcomes, it is uncertain whether multimodal exercise training compared to usual care or no exercise training at 5 and 12 months improved the functional and maximal exercise capacity (Langer 2012; Ulvestad 2020). Furthermore, it is uncertain whether 14‐weeks multimodal exercise training can increase the functional exercise capacity, compared to 7‐weeks multimodal exercise training (Fuller 2017). When comparing two different resistance training, it is uncertain if the use of a vibratory platform compared to same exercise training without platform, and a supervised upper limb training compared to no supervised upper limb training achieves better performance of the functional exercise capacity (Fuller 2018; Gloeckl 2015; Gloeckl 2017).

It is also uncertain whether multimodal exercise training compared to usual care or no exercise training, improves HRQoL domains of physical functioning and physical role (Langer 2012), and mental and physical health (Ulvestad 2020), and whether there are differences in any dimension between different training programs (Fuller 2017; Fuller 2018; Gloeckl 2015; Gloeckl 2017). As for adverse effects, their incidence is probably similar between multimodal exercise training and usual care or no exercise training (Langer 2012; Ulvestad 2020), however, the existing evidence assessing the safety of exercise training is indirect, since all studies report adverse events that could not necessarily be related to exercise training, such as acute infections, medical complications and episodes of acute rejection. Only four studies refer to no lesions, cardiac events, or other serious adverse events (Fuller 2017; Fuller 2018; Gloeckl 2015; Gloeckl 2017). Two studies reported unspecified musculoskeletal problems in participants, without specifying which group they belonged to (Fuller 2017; Ulvestad 2020), and Fuller 2018 reported no adverse findings for the clamshell incision in two‐dimensional ultrasound measurement and sternal stability in supervised upper limb training and no supervised upper limb training.

For the secondary outcomes, it is uncertain whether exercise training improves muscle strength as different muscle groups were measured and at different time points. Braith 2007 and Mitchell 2003 reported improvement in lumber extensor muscle strength in the resistance exercise group and compared to no training; Braith 2007 reported improvement in all seven positions, while Mitchell 2003 only reported improvement in the last three testing positions. When comparing two resistance exercise programs only WBVT showed improvement in peak force hamstrings at 4 weeks compared to control (Gloeckl 2017). Langer 2012 reported improvement in quadriceps force at 3 months and 12‐months post‐transplant with multimodal exercise, while Ulvestad 2020 reported improvement in one repetition maximum leg and arm press at 5 months. Fuller 2017 compared the same exercise program given for different durations ‐ no differences in quadriceps and hamstring strength were reported.

Pathological bone fracture was no reported in any clinical trial. Three studies measured the BMD as an indirect measure and the certainty of the evidence is unclear (Braith 2007; Langer 2012; Mitchell 2003). Langer 2012 reported no difference in the lumbar T‐score between the multimodal versus no exercise group, while Braith 2007 and Mitchell 2003 both reported improvement in BMD in the resistance training group compared to no training.

Pulmonary function was only reported by Fuller 2017, Langer 2012 and Ulvestad 2020. Fuller 2017 reported no differences in FEV₁ and force vital capacity for different exercise duration, and Langer 2012 and Ulvestad 2020 reported no differences in FEV1 and DL_CO for multimodal exercise compared to no exercise.

Other outcomes of interest were rarely reported (adverse events, patient readmission, death) or not reported at all (return to normal activities).

The results of this review are limited mainly by the high risk of bias, imprecision, indirectness of the effect estimate, and by the lack of comparability of the studies due to differences in the population, exercise modality used, the timing of the measurement of the outcomes, and the presentation of the data.

Overall completeness and applicability of evidence

This systematic review aimed at evaluating the effects of exercise training in adult patients undergoing lung transplant. Studies that met the eligibility criteria included participants with a mean age of 54.9 years, 51.9% male, which coincides with the report from 2000 to 2013 provided by the ISHLT (Yusen 2014), making this results representative of the population usually transplanted. In addition, the distribution of the pathologies that most frequently cause a lung transplant reported by ISHLT registries is also reflected in the proportions of these conditions in the participants of the studies included in this review, with COPD being the pathology with the highest indication for lung transplantation (Yusen 2014). However, the inclusion criteria of included studies of our review did not consider people with slow progress in the post‐operative period, mainly in terms of length of hospital stay after transplantation and important medical complications, or if the patients were admitted to a special rehabilitation program (Langer 2012). Therefore, the results of our review should be carefully applied in this sub‐population of transplanted patients since these would tend to have worse performance in physical tests and a different assessment in terms of HRQoL.

As for the interventions used in the included studies, all evaluated training programs began when the patient had already been discharged, with a median duration of 13 weeks, and were carried out in rehabilitation gyms under supervision of healthcare personnel; considerations that must be taken into account when designing a training program for these types of patients. However, the modality of the intervention was heterogeneous, except for studies that applied resistance training, with lumbar extension machine (Braith 2007; Mitchell 2003), possibly because they were performed by the same research team. This training modality, with the use of lumbar extension machine, would be unlikely to be routinely applied in clinical practice. On the other hand, the use of a vibration platform, an element used in two of the studies included in this review (Gloeckl 2015; Gloeckl 2017) could be used due to its feasibility of use in terms of cost, space and as regards the possibility of being used in other populations, such as COPD, in which they have proven to be safe and have positive effects (Cardim 2016; Yang 2016).

The control groups also varied: in four studies, it was the usual care or no exercise (Braith 2007, Langer 2012; Mitchell 2003; Ulvestad 2020), in three studies, it was a different resistance (anaerobic) exercise (Fuller 2018; Gloeckl 2015; Gloeckl 2017), and in one study, it was a multimodal exercise training with a different program duration (Fuller 2017). This reflects the lack of consensus about the optimal program needed according to the condition and needs of patients and noting that in Fuller 2017 there was a strict rigor after the first part of the training program, with active monitoring of the study participants with telephone calls.

Although there is no conclusive evidence regarding the effects of a training program on patients undergoing lung transplant or regarding the best exercise plan, this intervention is currently recognized as an essential issue in the management of lung transplant recipients (Spruit 2013). Future studies are expected to contribute to determining the actual effects of exercise in this study population, and make comparisons with greater implications and potential translation to the clinical environment, such as continuous aerobic training with high intensity interval training, or supervised training with home training

Quality of the evidence

The quantity and quality of the evidence included in this review does not allow us to draw any conclusions regarding the objectives of this review. Only five of the eight included studies reported an adequate method of randomisation and allocation concealment (Fuller 2017; Fuller 2018; Gloeckl 2015; Gloeckl 2017; Langer 2012). Given the characteristics of the intervention, no study achieved blinding of participants and health personnel; however, some outcomes, such as adverse effects (Braith 2007; Fuller 2017; Fuller 2018; Gloeckl 2015; Gloeckl 2017; Langer 2012; Mitchell 2003; Ulvestad 2020) and hospital readmission (Fuller 2017; Fuller 2018), were considered as probably not influenced by the lack of blinding. On the contrary, blinding of outcome assessors was reported in four studies (Fuller 2017; Fuller 2018; Gloeckl 2015; Langer 2012), one of which did it specifically for the assessment of functional exercise capacity through 6MWT (Gloeckl 2015). The HRQoL measured in six studies, was self‐reported (Fuller 2017; Fuller 2018; Gloeckl 2015; Gloeckl 2017; Langer 2012; Ulvestad 2020), therefore, by the nature of the intervention, the evaluation presented a high risk of bias. In the case of studies that did not report the assessment blinding (Braith 2007; Mitchell 2003), we considered that regardless whether it existed or not, adverse effects, patient readmission and pathological bone fractures are unlikely to be influenced by the lack of blinding (due to the objective way of assessing them). This does not apply to BMD, muscular strength and for the assessment the adverse events by sternal instability, since it was performed with a subjective scale, so the lack of blinding of the evaluators could introduce detection bias (Fuller 2018). Regarding the incomplete result data, two studies presented a high risk of bias due to the greater loss to follow‐up in the treated group (Braith 2007; Gloeckl 2015), and unclear risk of bias, because of the non‐specific reasons for dropouts (Fuller 2018) or because the information is unclear as to whether the participants that started are the same as those who finished the study (Mitchell 2003). In terms of selective reporting of results, it appears to be the most relevant domain, since seven of the eight included studies (Braith 2007; Fuller 2017; Gloeckl 2015; Gloeckl 2017; Langer 2012; Mitchell 2003; Ulvestad 2020) were rated at high risk, mainly due to the inconsistency between data reporting and what was proposed in the different protocols, not including relevant outcomes to evaluate the effects of the type of interventions and the relative —and not absolute— numbers presenting the data with respect to baseline or reference values. Finally, when considering other types of bias, no study showed imbalance between the groups at the beginning of the studies, so all were considered at low risk of bias; two studies did not mention conflicts of interest/disclosure declared (Braith 2007; Mitchell 2003) and two did not declare if there was a source of financing (Fuller 2018; Mitchell 2003), so they were considered to be at high risk of bias.

For resistance (anaerobic) exercise training versus usual care or no exercise training, the certainty of the evidence for adverse events and BMD were downgraded to very low due to high risk of bias, indirectness and imprecision, and muscular strength were downgraded to very low due to high risk of bias, inconsistency and imprecision (summary of findings Table 1). For multimodal exercise training versus usual care or no exercise training, the certainty of the evidence for maximal and functional exercise capacity, HRQoL, pulmonary function and muscular strength were downgraded to very low due to high risk of bias and imprecision. Adverse events and pathological bone fractures were downgraded to very low due to high risk of bias, indirectness and imprecision (summary of findings Table 3).

For two different resistance exercise training programs, the certainty of the evidence for maximal and functional exercise capacity, HRQoL, patient readmission and muscular strength were downgraded to very low due to high risk of bias and imprecision; adverse events were downgraded to very low due to high risk of bias, indirectness and imprecision (summary of findings Table 2). For 14 weeks versus 7 weeks of multimodal exercise training, the certainty of the evidence for maximal and functional exercise capacity, HRQoL, adverse events, patient readmission, pulmonary function and muscular strength was downgraded to very low due to high risk of bias and imprecision (summary of findings Table 4).

Potential biases in the review process

The search strategy of this review was carried out in a sensitive manner, without limitation of time, language or publication status. In addition, in order to complement this search, we checked the references of all included studies and previous reviews addressing this issue (Didsbury 2013; Langer 2015; Wickerson 2010) so as identified studies that were not retrieved by the search strategy.

Two studies reported their results in a relative way (Gloeckl 2015; Langer 2012), either as percentages of local data reference values or as changes from baseline values, making it difficult to understand the estimate of the absolute effect. In addition, we requested unpublished data to the authors of the studies of Braith 2007 and Gloeckl 2017. Braith 2007 reported lumbar extensor muscle strength only graphically, data for the control group in the graph were not legible and therefore we did not attempt to extract these data. The author did not respond to our request. Gloeckl 2017 provided all the requested data, but these were not formally published, and were only presented as conference proceedings.

Agreements and disagreements with other studies or reviews

Two non‐Cochrane systematic reviews were published prior to the completion of this review (Didsbury 2013; Wickerson 2010). Wickerson 2010 considered non‐RCTs eligible and citing the small number of RCTs to date. It included seven studies, two of which were RCTs that also met the inclusion criteria of this review (Braith 2007; Mitchell 2003). Wickerson 2010 evaluated the methodological quality as poor to moderate, specifically rating the study by Braith 2007 with a score of 5/10 and a 6/10 for Mitchell 2003 on the PEDro scale (Maher 2003). Wickerson 2010, like our review, was cautious in its conclusions, noting but more RCTs were necessary in order to prove that lung transplant recipients clearly benefit from some exercise training program in comparison with the natural course of post lung transplant recovery.

The objective of Didsbury 2013 was similar to our review; however, the study population was broader, since it included participants who underwent heart, kidney, liver, and lung transplantation. In the case of lung transplant, Didsbury 2013 included only two studies (Langer 2012; Ihle 2011), Although the two studies were detected by our electronic search strategy, one of them was excluded from the review (Ihle 2011) because the intervention given to the treated group consisted not only of exercise training, but of a pulmonary rehabilitation program, which consists of multiple interventions that go beyond exercising.

Figure 1

Study flow diagram.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Figure 2

Risk of bias summary: review authors' judgements about each risk of bias item for each included study (blank sections indicate outcome not reported)

Ir a la figura de la revisiónAbrir en una pestaña nueva

Figure 3

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies (blank sections indicate outcome not reported)

Ir a la figura de la revisiónAbrir en una pestaña nueva

Study	Resistance (anaerobic) exercise training group	Usual care or no exercise training group	Comment
Adverse events
Any adverse event
Braith 2007	Four patients did not complete the 6‐month training phase; 3 did not adhere to the training regimen and 1 died due to transplant‐related complications	"No untoward side‐effects were documented" Two patients did not complete the 8‐month testing; 1 required a new lung graft and 1 died due to transplant‐related complications	‐‐
Acute graft rejection episodes
Braith 2007	1.8 ± 1.2	1.7 ± 1.1	"The difference in rejection episodes among groups was not significant (p ≥ 0.05)." (Data for the alendronate alone has not been reported here)
Mitchell 2003	2.5 ± 1.5	1.8 ± 1.6	"...individual subjects in the trained group experienced nearly one more rejection episode during the course of the study, compared with the control group, the difference in rejection episodes between groups was not statistically significant (P ≥ 0.05)."

Analysis 1.1

Comparison 1: Resistance (anaerobic) exercise training versus usual care or no exercise training, Outcome 1: Adverse events

Study	Resistance (anaerobic) exercise training group	Usual care or no exercise training group	Comment
Muscular strength
Braith 2007	"The alendronate training group had increased (p ≤ 0.05) maximal lumbar isometric strength at all 7 testing positions (degrees of lumbar flexion)"	"...the alendronate‐only and control groups did not have increases (p ≥ 0.05) in lumbar strength at any of the testing positions."	‐‐
Mitchell 2003	"The trained group increased (P ≤ 0.05) maximal lumbar isometric strength at all seven testing positions (degrees of lumbar flexion)."	"...the control group experienced increases (P ≤ 0.05) in lumbar strength at only 48, 60, and 72 degrees of lumbar flexion."	"The magnitude of lumbar strength gains in the trained group at 48, 60, and 72 degrees of lumbar flexion were significantly greater (twofold; P ≤ 0.05) than the control group."

Analysis 1.2

Comparison 1: Resistance (anaerobic) exercise training versus usual care or no exercise training, Outcome 2: Muscular strength

Study	Resistance (anaerobic) exercise training group	Usual care or no exercise training group	Comment
Pathological bone fractures (measured indirectly as BMD)
Braith 2007	BMD significantly improved: baseline: 0.65 ± 0.21 g/cm²; at 8 months: 0.72 ± 0.17 g/cm² hydroxyapatite; P < 0.05 (10 participants)	No significant improvement in BMD: baseline: 0.66 ± 0.19 g/cm²; at 8 months: 0.67 ± 0.24 g/cm² hydroxyapatite; P > 0.05 (10 participants)	‐‐
Mitchell 2003	BMD at baseline was 0.631 ± 0.19 and at 8 months was 0.599 ± 0.18 g/cm² hydroxyapatite (8 participants).	BMD at at baseline was 0.623 ± 0.21 and at 8 months was 0.502 ± 0.19 g/cm² hydroxyapatite (8 participants)	Both groups lost BMD in the first 2 months post‐transplant. BMD continued to fall in the control group while the resistance exercise group regained BMD and levels returned to within 5% of pre‐transplant levels (P < 0.05)

Analysis 1.3

Comparison 1: Resistance (anaerobic) exercise training versus usual care or no exercise training, Outcome 3: Pathological bone fractures (measured indirectly as BMD)

Study	WBVT group	CON group
Maximal and functional exercise capacity
Six‐minute walk test (6MWT): improvement from baseline
Gloeckl 2015	85.3 metres (95% CI 65.4 to 101.7; P < 0.001)	55.2 metres (95% CI 37.5 to 72.8; P < 0.001)
Gloeckl 2017	47.8 metres (95% CI 35.1 to 60.4; P < 0.001)	19.5 metres (95% CI 6.7 to 32.3; P < 0.01)
Peak work rate: improvement from baseline
Gloeckl 2015	16.8 watts, 95% CI 13.5 to 20.5; P < 0.001)	12.6 watts (95% CI 9 to 16.1; P < 0.001)
Gloeckl 2017	14.5 watts (95% CI 11.5 to 17.5; P < 0.001)	10 watts (95% CI 7 to 13; P < 0.001)

Analysis 2.1

Comparison 2: Resistance (anaerobic) exercise training versus other resistance (anaerobic) exercise training, Outcome 1: Maximal and functional exercise capacity

Study	CRQ dyspnoea	CRQ fatigue	CRQ emotional function	CRQ emotional mastery
Health‐related quality of life: chronic respiratory questionnaire (CRQ)
CRQ components: improvement from baseline
Gloeckl 2015	WBVT group: 1.05 points (95% CI 0.59 to 1.49) CON group: 0.86 points (95% CI 0.43 to 1.3)	WBVT group: 0.65 points (95% CI 0.35 to 0.96) CON group: 0.75 points (95% CI 0.46 to 1.04	WBVT group: 0.64 points (95% CI 0.33 to 0.95) CON group: 0.72 points (95% CI 0.43 to 1.02)	WBVT group: 0.57 points (95% CI 0.25 to 0.90) CON group: 0.39 points (95% CI 0.08 to 0.69)
Gloeckl 2017	WBVT group: 0.71 points (95% CI 0.28 to 1.14) CON group: 0.68 points (95% CI 0.34 to 1.03)	WBVT group: 0.66 points (95% CI 0.16 to 1.17) CON group: 0.72 points (95% CI 0.39 to 1.07)	WBVT group: 0.85 points (95% CI 0.38 to 1.32) CON group: 0.88 points (95% CI 0.62 to 1.14)	WBVT group: 0.59 points (95% CI 0.17 to 1.01) CON group: 0.4 points (95% CI 0.13 to 0.66)

Analysis 2.2

Comparison 2: Resistance (anaerobic) exercise training versus other resistance (anaerobic) exercise training, Outcome 2: Health‐related quality of life: chronic respiratory questionnaire (CRQ)

Study	Physical health	Mental health	Summary score
Health‐related quality of life: SF‐36 questionnaire
SF‐36 questionnaire components: 3 months
Fuller 2018	NULP group: 63.7 ± 21.6 points SULP group:69.5 ± 19.6 points P = 0.2	NULP group: 76.2 ± 17 points SULP group: 77.1 ± 17.5 points P = 0.8	NULP group: 69.8 ± 19.1 points SULP group: 74.2 ± 18.7 points P = 0.3
SF‐36 questionnaire components: 6 months
Fuller 2018	NULP group: 63.2 ± 26 points SULP group: 70.1 ± 24.9 points P = 0.2	NULP group: 78.4 ± 17.4 points SULP group: 81.4 ± 11.8 points P = 0.3	NULP group: 70.7 ± 21.4 points SULP group: 76.8 ± 19.1 points P = 0.2

Analysis 2.3

Comparison 2: Resistance (anaerobic) exercise training versus other resistance (anaerobic) exercise training, Outcome 3: Health‐related quality of life: SF‐36 questionnaire

Study	Three months	Six months
Muscular strength: peak force
Peak force: improvement from baseline
Gloeckl 2015	WBVT group: 31.1 N (95% CI 21.4 to 40.7)	WBVT group: 20 N (95% CI 11.2 to 28.8)
Gloeckl 2015	CON group: 30.7 N (95% CI 21.7 to 39.8)	CON group: 13.8 N (95% CI 5.6 to 21.9)
Gloeckl 2017	WBVT group: 39.0 N (95% CI 20.4 to 57.6)	WBVT group: 22.4 N (95% CI 12.8 to 32)
Gloeckl 2017	CON group: 18 N (95% CI 7.6 to 28.3)	CON group: 8.8 N (95% CI 0.6 to 17)

Analysis 2.4

Comparison 2: Resistance (anaerobic) exercise training versus other resistance (anaerobic) exercise training, Outcome 4: Muscular strength: peak force

Study	3 months	6 months
Muscular strength: shoulder flexion and abduction force
Shoulder flexion force
Fuller 2018	NULP group: 8.7 ± 3.4 Nm	NULP group: 9.3 ± 4.1 Nm
Fuller 2018	SULP group: 9.1 ± 4.4 Nm (P = 0.72)	SULP group: 10.2 ± 4.9 Nm P = 0.4
Shoulder abduction force
Fuller 2018	NULP group: 8.0 ± 2.8 Nm	NULP group: 8.0 ± 3.5 Nm
Fuller 2018	SULP group: 8.4 ± 3.8 Nm P = 0.59	SULP group: 8.9 ± 4.0 Nm P = 0.3

Analysis 2.5

Comparison 2: Resistance (anaerobic) exercise training versus other resistance (anaerobic) exercise training, Outcome 5: Muscular strength: shoulder flexion and abduction force

Study	Multimodal exercise group	No exercise group
Maximal and functional exercise capacity
Six‐minute walk text (6MWT)
Langer 2012	3 months: 79 ± 8% predicted (n = 18) 12‐months post‐transplant: 86 ± 7% predicted (n = 18)	3 months: 70 ± 10% predicted (n = 16) 12‐months post‐transplant: 74 ± 11% predicted (n = 16)
Wmax
Langer 2012	3 months: 63 ± 23% predicted (n = 18) 12‐months post‐transplant: 69 ± 20% predicted (n = 18)	3 months: 50 ± 22% predicted (n = 16) 12‐months post transplant: 53 ± 23% predicted (n = 16)
VO₂max
Langer 2012	3 months: 71 ± 26% predicted (n = 18) 12‐months post‐transplant: 78 ± 27% predicted (n = 18)	3 months: 56 ± 21% predicted (n = 16) 12‐months post‐transplant: 63 ± 24% predicted (n = 16)
VO₂peak (adjusted for baseline score): 5 months
Ulvestad 2020	24.1 ± 7.6 mL/kg/min (n = 23)	24.0±7.2 mL/kg/min (n = 23)
Chair to stand test (adjusting for baseline scores): 5 months
Ulvestad 2020	13.5 ± 3.1 repetitions (n = 23)	13.6 ± 2.8 repetitions (n = 23)
Stair run test (adjusted for baseline score): 5 months
Ulvestad 2020	34.4 ± 9.2 seconds (n = 23)	33.8 ± 9.4 seconds (n = 23)

Analysis 3.1

Comparison 3: Multimodal exercise training versus usual care or no exercise training, Outcome 1: Maximal and functional exercise capacity

Study	Multimodal exercise group	No exercise group	P value
Health‐related quality of life: SF‐36 components
SR‐36 components: 3 months
Langer 2012	Physical functioning: 69 ± 16	Physical functioning: 60 ± 19	P = 0.213
	Role functioning physical: 61 ± 33	Role functioning physical: 48 ± 44	P = 0.448
	Pain: 67 ± 33	Pain: 65 ± 21	P = 0.959
	General health: 62 ± 18	General health: 56 ± 16	P = 0.212
	Energy/fatigue: 70 ± 14	Energy/fatigue: 58 ± 21	P = 0.285
	Social functioning: 82 ± 20	Social functioning: 88 ± 25	P = 0.064
	Role functioning emotional: 80 ± 31	Role functioning emotional: 88 ± 34	P = 0.341
	Emotional well being: 81 ± 9	Emotional well being: 73 ± 22	P = 0.808
SF‐36 components: 12‐months post‐transplant
Langer 2012	Physical functioning: 77 ± 11	Physical functioning: 65 ± 17	P = 0.039*
	Role functioning physical: 83 ± 28	Role functioning physical: 52 ± 32	P = 0.011*
	Pain: 73 ± 21	Pain: 70 ± 28	P = 0.811
	General health: 62 ± 16	General health: 60 ± 21	P = 0.632
	Energy/fatigue: 71 ± 14	Energy/fatigue: 67±17	P = 0.937
	Social functioning: 86 ± 22	Social functioning: 74 ± 23	P = 0.118
	Role functioning emotional: 83 ± 31	Role functioning emotional: 81 ± 30	P = 0.978
	Emotional well being: 74 ± 17	Emotional well being: 74 ± 22	P = 0.385

Analysis 3.2

Comparison 3: Multimodal exercise training versus usual care or no exercise training, Outcome 2: Health‐related quality of life: SF‐36 components

Study	Multimodal exercise group	No exercise group	P value
Health‐related quality of life: SF‐36 components (adjusted for baseline scores)
SF‐36 components (adjusted for baseline scores): 5 months
Ulvestad 2020	Mental component summary score: 50.6 ± 9.2	Mental component summary score: 50.7 ± 10.9	P = 0.02*
Ulvestad 2020	Physical component summary score: 48.4 ± 8.7	Physical component summary score:52.3 ± 6.4	P = 0.328

Analysis 3.3

Comparison 3: Multimodal exercise training versus usual care or no exercise training, Outcome 3: Health‐related quality of life: SF‐36 components (adjusted for baseline scores)

Study	Multimodal exercise group	No exercise group
Pulmonary function
FEV₁
Langer 2012	3 months: 89 ± 18% predicted 12‐months post‐transplant: 92 ± 20% predicted	3 months: 80 ± 22% predicted 12‐months post‐transplant: 89 ± 25% predicted
FEV₁ (adjusted for baseline scores)
Ulvestad 2020	2.5 ± 0.6 L	2.5 ± 0.7 L
DL_CO (adjusted for baseline score)
Ulvestad 2020	6.1 ± 4.1 mmol/min/kPa)	6.4±1.4 mmol/min/kPa

Analysis 3.4

Comparison 3: Multimodal exercise training versus usual care or no exercise training, Outcome 4: Pulmonary function

Study	Multimodal exercise group	No exercise group
Muscular strength
Quadriceps force
Langer 2012	3 months: 82 ± 20% predicted 12‐months post‐transplant: 92 ± 21% predicted	3 months: 60 ± 18% predicted 12‐months post‐transplant: 71 ± 20% predicted
Handgrip force
Langer 2012	3 months: 85 ± 21% predicted 12‐months post‐transplant: 94 ± 22% predicted	3 months: 88 ± 22% predicted 12‐months post‐transplant: 91 ± 18% predicted
Ulvestad 2020	Adjusted for baseline score: 35.7 ± 12.2 kg	Adjusted for baseline score: 34.3 ± 11.9 kg
Maximal inspiratory pressure
Langer 2012	3 months: 97 ± 23% predicted 12‐months post‐transplant: 105 ± 23% predicted	3 months: 87 ± 18% predicted 12‐months post‐transplant: 92 ± 22% predicted
Maximal expiratory pressure
Langer 2012	3 months: 93 ± 25% predicted 12 months post‐transplant: 106 ± 23% predicted	3 months: 91 ± 25% predicted 12 months post‐transplant: 99 ± 23% predicted
One repetition maximum leg press
Ulvestad 2020	118.8 ± 36.0 kg	112.9 ± 30.9 kg
One repetition maximum arm press
Ulvestad 2020	49.6 ± 20.4 kg	52.0 ± 25.1 kg

Analysis 3.5

Comparison 3: Multimodal exercise training versus usual care or no exercise training, Outcome 5: Muscular strength

Study	Multimodal exercise group	No exercise group
Pathological bone fractures (measured indirectly as BMD)
Lumber T‐score at 12 months post‐transplant
Langer 2012	‐2.6 ± 1.3	‐2.2 ± 1.5
Femur T‐score at 12 months post‐transplant
Langer 2012	‐1.8 ± 0.9	‐1.9 ± 0.6

Analysis 3.6

Comparison 3: Multimodal exercise training versus usual care or no exercise training, Outcome 6: Pathological bone fractures (measured indirectly as BMD)

Study	14‐week multimodal exercise group	7‐week multimodal exercise group
Health‐related quality of life: SF‐36 components
SF‐36 questionnaire components: 14 weeks
Fuller 2017	Physical health: 71.3 ± 14.7	Physical health: 69.7 ± 17.9
Fuller 2017	Emotional health: 75.9 ± 16.5	Emotional health: 75.6 ± 16.6
SF‐36 questionnaire components: 6 months
Fuller 2017	Physical health: 71.6 ± 14.5	Physical health: 73.5 ± 18.8
Fuller 2017	Emotional health: 76.2 ± 16.8	Emotional health: 78.8 ± 18.6

Analysis 4.1

Comparison 4: 14 weeks versus 7 weeks multimodal exercise training, Outcome 1: Health‐related quality of life: SF‐36 components

Study	14‐week multimodal exercise group	7‐week multimodal exercise group
Health‐related quality of life: AQoL components
AQoL components: 14 weeks
Fuller 2017	Independent living: 0.88 ± 0.16	Independent living: 0.94 ± 0.11
	Relations: 0.87 ± 0.19	Relations: 0.88 ± 0.17
	Senses: 0.91 ± 0.12	Senses: 0.96 ± 0.06
	Mental health: 0.86 ± 0.01	Mental health: 0.90 ± 0.07
	AQol 4D score: 0.64 ± 0.23	AQol 4D score: 0.74 ± 0.19
AQoL components: 6 months
Fuller 2017	Independent living: 0.92 ± 0.19	Independent living: 0.93 ± 0.15
	Relations: 0.89 ± 0.11	Relations: 0.89 ± 0.18
	Senses: 0.93 ± 0.09	Senses: 0.95 ± 0.15
	Mental health: 0.87 ± 0.09	Mental health: 0.88 ± 0.13
	AQol 4D score: 0.70 ± 0.20	AQol 4D score: 0.74 ± 0.24

Analysis 4.2

Comparison 4: 14 weeks versus 7 weeks multimodal exercise training, Outcome 2: Health‐related quality of life: AQoL components

Study	14 weeks multimodal exercise group	7 weeks multimodal exercise group
Muscular strength
Average quadriceps strength
Fuller 2017	14 weeks: 105.41 ± 36.6 Nm	14 weeks: 111.87 ± 39.3 Nm
Fuller 2017	6 months: 114.61 ± 40.2 Nm	6 months: 115.31 ± 38.9 Nm
Peak torque quadriceps strength
Fuller 2017	14 weeks: 116.45 ± 36.5 Nm	14 weeks: 122.49 ± 41.6 Nm
Fuller 2017	6 months: 127.6 ± 42.8 Nm	6 months: 127.15 ± 44.0 Nm
Average hamstrings strength
Fuller 2017	14 weeks: 48.21 ± 16.5 Nm	14 weeks: 52.31 ± 17.3 Nm
Fuller 2017	6 months: 52.3 ± 19.0 Nm	6 months: 56.89 ± 18.5 Nm
Peak torque hamstring strength
Fuller 2017	14 weeks: 52.43.± 16.8 Nm	14 weeks: 57.27 ± 17.9 Nm
Fuller 2017	6 months: 56.72 ± 19.5 Nm	6 months: 61.98 ± 19.9 Nm

Analysis 4.3

Comparison 4: 14 weeks versus 7 weeks multimodal exercise training, Outcome 3: Muscular strength

Summary of findings 1. Resistance (anaerobic) exercise training versus usual care or no exercise training

Outcomes	Effect*	No. of participants (studies)	Quality of the evidence (GRADE)
Resistance (anaerobic) exercise training versus usual care or no exercise training for adult lung transplant recipients
Patient or population: adult lung transplant recipients Settings: outpatient training Intervention: resistance (anaerobic) exercise training Comparison: usual care or no exercise training
Maximal and functional exercise capacity Follow‐up: up to 6 months	Not reported	‐‐	‐‐
HRQoL Follow‐up: up to 6 months	Not reported	‐‐	‐‐
Adverse events Assessed rejection episodes of lung transplant Follow‐up: up to 6 months	One study reported a lower number of rejection episodes, and other one reported a higher number, however, there were no differences between groups	32 (2)	⊕⊝⊝⊝ very low²
Pulmonary function Follow‐up: up to 6 months	Not reported	‐‐	‐‐
Muscular strength¹ Assessed lumbar strength at seven testing positions Follow‐up: up to 6 months	One study reported greater lumbar strength in all seven positions and other one only in the three positions of test hip flexion	32 (2)	⊕⊝⊝⊝ very low³
Pathological bone fractures Estimated indirectly through BMD Follow‐up: up to 6 months	Two studies reported no differences between groups in BMD	32 (2)	⊕⊝⊝⊝ very low⁴
Death (any cause) Follow‐up: up to 6 months	Not reported	‐‐	‐‐
* For all outcomes of interest a single pooled effect estimate is not available and only a narrative synthesis of the evidence is provided * The effect of the intervention is reported based on comparison with the respective control groups in each study HRQoL: health‐related quality of life; BMD: bone mineral density
GRADE Working Group grades of evidence High certainty: Further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low certainty: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low certainty: We are very uncertain about the estimate.
¹ Increased strength from baseline ² The certainty of the evidence was downgraded one level for risk of bias (random sequence generation was assessed with a unclear risk of bias and allocation concealment, incomplete outcome data, selective reporting and other potential sources of bias were assessed with a high risk of bias),one level for imprecision (sample size was small and not calculated and 95% CI includes non‐effect value and is also broad), and one level for indirectness (there are no reported adverse effects directly related to the practice of exercise training, for example, muscle injuries) ³ The certainty of the evidence was downgraded one level for risk of bias (random sequence generation was assessed with a unclear risk of bias and allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting and other potential sources of bias were assessed with a high risk), one level for inconsistency (variability in results across studies) and one level for imprecision (sample size was small and not calculated and 95% CI includes non‐effect value and was also broad) ⁴ The certainty of the evidence was downgraded one level for risk of bias (random sequence generation was assessed with a unclear risk of bias and allocation concealment, blinding of participants and personnel, incomplete outcome data, selective reporting and other potential sources of bias were assessed with a high risk), one level for indirectness (BMD was used as an intermediate surrogate outcome of pathological bone fractures), and one level for imprecision (sample size was small and not calculated and 95% CI includes non‐effect value and was also broad)

Summary of findings 1. Resistance (anaerobic) exercise training versus usual care or no exercise training

Summary of findings 2. Resistance (anaerobic) exercise training versus another form of resistance (anaerobic) exercise training

Outcomes	Effect*	No. of participants (studies)	Quality of the evidence (GRADE)
Resistance (anaerobic) exercise training versus another form of resistance (anaerobic) exercise training for adult lung transplant recipients
Patient or population: adult lung transplant recipients Settings: outpatient training Intervention: resistance (anaerobic) exercise training (squats using WBVT (Gloeckl 2015; Gloeckl 2017); supervised upper limb exercise (Fuller 2018)) Comparison: other form of resistance (anaerobic) exercise training (squats on the floor (Gloeckl 2015; Gloeckl 2017); unsupervised upper limb exercise (Fuller 2018)
Maximal and functional exercise capacity Assessed metres walked in 6MWT Follow‐up: up to 6 months	Two studies reported a greater walking distance in the resistance (anaerobic) exercise training group, and another study reported no difference between groups	214 (3)^a	⊕⊝⊝⊝ very low¹
HRQoL Assessed by different questionnaires (CRQ and SF‐36) Follow‐up: up to 6 months	Three studies reported no differences between groups in the HRQoL	215 (3)^a	⊕⊝⊝⊝ very low²
Adverse events Assessed indirectly related to the exercise Follow‐up: up to 6 months	Two studies reported no differences in the incidence of adverse events	162 (2)	⊕⊝⊝⊝ very low³
Pulmonary function Follow‐up: up to 6 months	Not reported	‐‐	‐‐
Muscular strength Assessed upper and lower limb muscle strength Follow‐up: up to 6 months	Three studies reported no differences between groups in upper and lower limb muscle strength	245 (3)^a	⊕⊝⊝⊝ very low⁴
Pathological bone fractures Follow‐up: up to 6 months	Not reported	‐‐	‐‐
Death (any cause) Follow‐up: up to 6 months	One study reported no difference between groups in all‐cause mortality	83 (1)	⊕⊝⊝⊝ very low³
* For all outcomes of interest a single pooled effect estimate is not available and only a narrative synthesis of the evidence is provided * The effect of the intervention is reported based on comparison with the respective control groups in each study ^a In the study of Fuller 2018 the number of participants was variable because not all of them showed up for the measurements of all outcomes. The number of participants immediately after the end of the intervention was considered Gloeckl 2015 and Gloeckl 2017 included participants with mean time from transplantation of three months and 5.5 years, respectively, and were therefore not combined statistically (pooled analysis) WBVT: whole‐body vibration training; 6MWT: 6‐minute walk test; HRQoL: health‐related quality of life; CRQ: chronic respiratory questionnaire; SF‐36**: short form 36 health survey questionnaire
GRADE Working Group grades of evidence High certainty: Further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low certainty: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low certainty: We are very uncertain about the estimate.
¹ The certainty of the evidence was downgraded two levels for risk of bias (blinding of participants, personnel and outcome assessment, incomplete outcome data, selective reporting and sources of funding declared were assessed with a high risk of bias) and one level for imprecision (according to the ERS, the MID established for 6MWT is 30 metres, so the difference found in this study could be or not clinically relevant, because MID is within the CI 95% (Holland 2014). Also the lower limit is very close to 0) ² The certainty of the evidence was downgraded two levels for risk of bias (blinding of participants, personnel and outcome assessment, incomplete outcome data, selective reporting and sources of funding declared were assessed with a high risk of bias) and one level for imprecision (95% CI includes no‐effect value) ³ The certainty of the evidence was downgraded two levels for risk of bias (blinding outcome assessment, incomplete outcome data, selective reporting and sources of funding declared were assessed with a high risk of bias) and imprecision (confidence levels are very wide and there are few events), and one level for indirectness (there are no reported adverse effects directly related to the practice of exercise training, for example, muscle injuries) ⁴ The certainty of the evidence was downgraded two levels for risk of bias (blinding of participants, personnel and outcome assessment, incomplete outcome data, selective reporting and sources of funding declared were assessed with a high risk of bias) and one level for imprecision (the sample size was small and there are few events)

Summary of findings 2. Resistance (anaerobic) exercise training versus another form of resistance (anaerobic) exercise training

Summary of findings 3. Multimodal exercise training versus usual care or no exercise training

Outcomes	Effect*	No. of participants (studies)	Quality of the evidence (GRADE)
Multimodal exercise training versus usual care or no exercise training for adult lung transplant recipients
Patient or population: adult lung transplant recipients Settings: outpatient training Intervention: multimodal exercise training (continuous aerobic training + resistance training (Langer 2012); high‐intensity interval training + resistance training (Ulvestad 2020)) Comparison: usual care or no exercise training
Maximal and functional exercise capacity Assessed with different maximal and submaximal tests Follow‐up: up to 6 months	One study reported a greater increase in walking distance on the 6MWT in the multimodal exercise training group, however, this and other study reported no differences between groups in other functional and oxygen consumption tests	80 (2)	⊕⊝⊝⊝ verylow¹
HRQoL Assessed by different domains of the SF‐36 questionnaire Follow‐up: up to 6 months	Two studies reported no differences between groups in the HRQoL	80 (2)	⊕⊝⊝⊝ very low²
Adverse events Assessed the severe medical complications and exercise‐related musculoskeletal pain Follow‐up: up to 6 months	Two studies reported no differences between groups in the number of study dropouts due to severe medical complications, and other one study reported low incidence in muscle pain in the intervention group	80 (2)	⊕⊝⊝⊝ very low³
Pulmonary function Assessed by FEV₁ (absolute value and percentage of predicted value) Follow‐up: up to 6 months	Two studies reported no differences between groups in the pulmonary function	80 (2)	⊕⊝⊝⊝ very low⁴
Muscular strength Assessed hand grip, upper and lower limb muscle strength Follow‐up: up to 6 months	Two studies reported no differences in hand grip strength, but one of them did report differences in upper limb strength. The same two studies reported greater lower limb strength in the multimodal exercise training group	80 (2)	⊕⊝⊝⊝ very low⁵
Pathological bone fractures Follow‐up: up to 6 months	Not reported	‐‐	‐‐
Death (any cause) Follow‐up: up to 6 months	One study reported that one patient died due to lung rejection in the no exercise group	40 (1)	⊕⊝⊝⊝ very low⁶
* For all outcomes of interest a single pooled effect estimate is not available and only a narrative synthesis of the evidence is provided * The effect of the intervention is reported based on comparison with the respective control groups in each study 6MWT: 6‐minute walk test; FEV₁ : forced expiratory volume in 1 sec; SF‐36: short form 36 health survey questionnaire
GRADE Working Group grades of evidence High certainty: Further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low certainty: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low certainty: We are very uncertain about the estimate.
¹ The certainty of the evidence was downgraded two levels for risk of bias (allocation concealment, blinding of participants and personnel, blinding of outcome assessment and selective reporting were assessed with a high risk of bias) and one level for imprecision (the minimal important difference (MID), calculated by the rule of thumb that MID is typically 0.5 standard deviations (Norman 2003), is 5.5%, and the MID is included in the CI) ² The certainty of the evidence was downgraded two levels for risk of bias (allocation concealment, blinding of participants and personnel, blinding of outcome assessment and selective reporting were assessed with a high risk) and one level for imprecision (the minimal important difference (MID), calculated by the rule of thumb that MID is typically 0.5 standard deviations (Norman 2003), is 8.5 points in "physical functioning" and 16 points in "role functioning physical", and both MIDs are included in the respective CIs) ³ The certainty of the evidence was downgraded two levels for risk of bias (allocation concealment and selective reporting were assessed with a high risk of bias), one level by indirectness (there are no reported adverse effects directly related to the practice of exercise training, for example, muscle injuries) and one level for imprecision (CI includes non‐effect value and is also broad) ⁴ The certainty of the evidence was downgraded two levels by risk of bias (allocation concealment, blinding of participants and personnel, blinding of outcome assessment and selective reporting were assessed with a high risk of bias) and one level for imprecision (CI includes non‐effect value and is also broad) ⁵ The certainty of the evidence was downgraded two levels for risk of bias (allocation concealment, blinding of participants and personnel, blinding of outcome assessment and selective reporting were assessed with a high risk of bias) and one level for imprecision (the minimal important difference (MID), calculated by the rule of thumb that MID is typically 0.5 standard deviations (Norman 2003), is 10%, and the MID is included in the CI) ⁶ The certainty of the evidence was downgraded two levels by risk of bias (allocation concealment, blinding of participants and personnel, blinding of outcome assessment and selective reporting were assessed with a high risk of bias) and one level for imprecision (the sample size was small and there are few events)

Summary of findings 3. Multimodal exercise training versus usual care or no exercise training

Summary of findings 4. Fourteen‐weeks versus 7‐weeks multimodal exercise training

Outcomes	Effect*	No. of participants (studies)	Quality of the evidence (GRADE)
Fourteen‐weeks versus 7‐weeks multimodal exercise training for adult lung transplant recipients
Patient or population: adult lung transplant recipients Settings: outpatient training Intervention: 14‐weeks multimodal exercise training Comparison: 7‐weeks multimodal exercise training
Maximal and functional exercise capacity Assessed metres walked in 6MWT Follow‐up: up to 6 months	One study reported no differences between groups in the distance walked	61 (1)	⊕⊝⊝⊝ verylow¹
HRQoL Assessed by different questionnaires (SF‐36 and AQoL) Follow‐up: up to 6 months	One study reported no differences between groups in HRQoL	59 (1)	⊕⊝⊝⊝ very low²
Adverse events Follow‐up: up to 6 months	Few participants missed some evaluations due to musculoskeletal problems or hospital readmissions, however, the group to which they belonged was not reported	66 (1)	⊕⊝⊝⊝ very low³
Pulmonary function Assessed by FEV₁ and FVC Follow‐up: up to 6 months	One study reported no differences between groups in pulmonary function	66 (1)	⊕⊝⊝⊝ very low⁴
Muscular strength Assessed by peak quadriceps muscle strength Follow‐up: up to 6 months	One study reported no differences between groups in quadriceps muscle strength	59 (1)	⊕⊝⊝⊝ very low⁵
Pathological bone fractures Follow‐up: up to 6 months	Not reported	‐‐	‐‐
Death (any cause) Follow‐up: up to 6 months	One study reported one death due to acute pneumonia in the 14‐week exercise group	66 (1)	⊕⊝⊝⊝ very low⁶
* For all outcomes of interest a single pooled effect estimate is not available and only a narrative synthesis of the evidence is provided * The effect of the intervention is reported based on comparison with the respective control groups in each study 6MWT: 6‐minute walk test; HRQoL: health‐related quality of life; FEV₁ : forced expiratory volume in 1 sec; FVC: forced vital capacity; SF‐36: short form 36 health survey questionnaire; AQoL: Australian quality of life questionnaire
GRADE Working Group grades of evidence High certainty: Further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low certainty: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low certainty: We are very uncertain about the estimate.
¹ The certainty of the evidence was downgraded two levels for risk of bias (blinding of participants and personnel and selective reporting were assessed with a high risk of bias) and one level for imprecision (according to the ERS, the MID established for 6MWT is 30 metres, so the difference found in this study could be or not clinically relevant, because MID is within the CI 95% (Holland 2014)) ² The certainty of the evidence was downgraded two levels for risk of bias (blinding of participants, personnel and outcome assessment and selective reporting were assessed with a high risk of bias) and one level for imprecision (95% CI includes non‐effect value) ³ The certainty of the evidence was downgraded by one level by risk of bias (selective reporting were assessed with a high risk of bias) and two levels by imprecision (the sample size was small and the total number of adverse events was very small) ⁴ The certainty of the evidence was downgraded one level for high risk of bias (reporting bias) and two levels for imprecision (the sample size was small and only reported P value) ⁵ The certainty of the evidence was downgraded two levels by risk of bias (blinding of participants and personnel and selective reporting were assessed with a high risk of bias) and one level by imprecision (sample size not calculated for this outcome and 95% CI includes non‐effect value and is also broad) ⁶ The certainty of the evidence was downgraded one level for high risk of bias (reporting bias) and two levels for imprecision (the sample size was small and there are few events)

Summary of findings 4. Fourteen‐weeks versus 7‐weeks multimodal exercise training

Comparison 1. Resistance (anaerobic) exercise training versus usual care or no exercise training

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Adverse events Show forest plot	2		Other data	No numeric data

1.1.1 Any adverse event	1		Other data	No numeric data
1.1.2 Acute graft rejection episodes	2		Other data	No numeric data
1.2 Muscular strength Show forest plot	2		Other data	No numeric data

1.3 Pathological bone fractures (measured indirectly as BMD) Show forest plot	2		Other data	No numeric data

Comparison 1. Resistance (anaerobic) exercise training versus usual care or no exercise training

Comparison 2. Resistance (anaerobic) exercise training versus other resistance (anaerobic) exercise training

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Maximal and functional exercise capacity Show forest plot	2		Other data	No numeric data

2.1.1 Six‐minute walk test (6MWT): improvement from baseline	2		Other data	No numeric data
2.1.2 Peak work rate: improvement from baseline	2		Other data	No numeric data
2.2 Health‐related quality of life: chronic respiratory questionnaire (CRQ) Show forest plot	2		Other data	No numeric data

2.2.1 CRQ components: improvement from baseline	2		Other data	No numeric data
2.3 Health‐related quality of life: SF‐36 questionnaire Show forest plot	1		Other data	No numeric data

2.3.2 SF‐36 questionnaire components: 3 months	1		Other data	No numeric data
2.3.3 SF‐36 questionnaire components: 6 months	1		Other data	No numeric data
2.4 Muscular strength: peak force Show forest plot	2		Other data	No numeric data

2.4.1 Peak force: improvement from baseline	2		Other data	No numeric data
2.5 Muscular strength: shoulder flexion and abduction force Show forest plot	1		Other data	No numeric data

2.5.2 Shoulder flexion force	1		Other data	No numeric data
2.5.3 Shoulder abduction force	1		Other data	No numeric data

Comparison 2. Resistance (anaerobic) exercise training versus other resistance (anaerobic) exercise training

Comparison 3. Multimodal exercise training versus usual care or no exercise training

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Maximal and functional exercise capacity Show forest plot	2		Other data	No numeric data

3.1.1 Six‐minute walk text (6MWT)	1		Other data	No numeric data
3.1.2 Wmax	1		Other data	No numeric data
3.1.3 VO2max	1		Other data	No numeric data
3.1.4 VO2peak (adjusted for baseline score): 5 months	1		Other data	No numeric data
3.1.5 Chair to stand test (adjusting for baseline scores): 5 months	1		Other data	No numeric data
3.1.6 Stair run test (adjusted for baseline score): 5 months	1		Other data	No numeric data
3.2 Health‐related quality of life: SF‐36 components Show forest plot	1		Other data	No numeric data

3.2.1 SR‐36 components: 3 months	1		Other data	No numeric data
3.2.2 SF‐36 components: 12‐months post‐transplant	1		Other data	No numeric data
3.3 Health‐related quality of life: SF‐36 components (adjusted for baseline scores) Show forest plot	1		Other data	No numeric data

3.3.3 SF‐36 components (adjusted for baseline scores): 5 months	1		Other data	No numeric data
3.4 Pulmonary function Show forest plot	2		Other data	No numeric data

3.4.1 FEV1	1		Other data	No numeric data
3.4.2 FEV1 (adjusted for baseline scores)	1		Other data	No numeric data
3.4.3 DLCO (adjusted for baseline score)	1		Other data	No numeric data
3.5 Muscular strength Show forest plot	2		Other data	No numeric data

3.5.1 Quadriceps force	1		Other data	No numeric data
3.5.2 Handgrip force	2		Other data	No numeric data
3.5.3 Maximal inspiratory pressure	1		Other data	No numeric data
3.5.4 Maximal expiratory pressure	1		Other data	No numeric data
3.5.5 One repetition maximum leg press	1		Other data	No numeric data
3.5.6 One repetition maximum arm press	1		Other data	No numeric data
3.6 Pathological bone fractures (measured indirectly as BMD) Show forest plot	1		Other data	No numeric data

3.6.1 Lumber T‐score at 12 months post‐transplant	1		Other data	No numeric data
3.6.2 Femur T‐score at 12 months post‐transplant	1		Other data	No numeric data

Comparison 3. Multimodal exercise training versus usual care or no exercise training

Comparison 4. 14 weeks versus 7 weeks multimodal exercise training

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Health‐related quality of life: SF‐36 components Show forest plot	1		Other data	No numeric data

4.1.1 SF‐36 questionnaire components: 14 weeks	1		Other data	No numeric data
4.1.2 SF‐36 questionnaire components: 6 months	1		Other data	No numeric data
4.2 Health‐related quality of life: AQoL components Show forest plot	1		Other data	No numeric data

4.2.1 AQoL components: 14 weeks	1		Other data	No numeric data
4.2.2 AQoL components: 6 months	1		Other data	No numeric data
4.3 Muscular strength Show forest plot	1		Other data	No numeric data

4.3.1 Average quadriceps strength	1		Other data	No numeric data
4.3.2 Peak torque quadriceps strength	1		Other data	No numeric data
4.3.3 Average hamstrings strength	1		Other data	No numeric data
4.3.4 Peak torque hamstring strength	1		Other data	No numeric data

Comparison 4. 14 weeks versus 7 weeks multimodal exercise training