Toennesen 2017.

Study characteristics
Methods	Study design: RCT parallel‐group design. 4 groups: (1) exercise group (2) diet group (3) exercise + diet group (4) control group. Results of exercise + diet group and control group have been extracted for this review. Total duration of study: 8 weeks with follow‐up at 1 year Details of any run‐in period: None stated Number of study centres and location: Single site. Denmark Study setting: Hospital Withdrawals: 37 participants randomised to exercise + diet (29 completed). 38 randomised to control (34 completed) Date of study: January 2015 to July 2016
Participants	Number recruited: Randomised to intervention (exercise + diet): 37. Randomised to control group: 38 Number completed: Completed intervention (exercise + diet): 29. Completed control: 34 Mean (SD) age: Intervention (exercise + diet) group: 43.7 (13.9) years. Control group: 38.2 (12.7) years Age range: Not stated Gender (M/F): Intervention (exercise + diet) group: 7/22. Control group: 8/26 Mean (SD) BMI: Intervention (exercise + diet) group: 26.1 (2.5) kg/m2. Control group: 25.5 (2.4) kg/m2 Severity of condition: Mild to moderate asthma. No participants received ICS in doses of greater than or equal to 1600 μg budesonide equivalents per day plus a second controller (indicating severe asthma) Diagnostic criteria: At least 1 positive diagnostic test demonstrating variable airflow obstruction (mannitol test, methacholine test, reversibility test) Baseline lung function: Mean (SD) % predicted FEV1: Intervention (exercise + diet) group: 82.6 (15.2)%. Control group: 81.9 (12.3)% Smoking history: Intervention (exercise + diet) group: non‐smoker 18 (62%); current smoker 1 (3%); former smoker 10 (34%). Control group: non‐smoker 19 (56%); current smoker 2 (6%); former smoker 13 (38%) Asthma treatment: Use of ICS: Intervention (exercise + diet) group: 17 (59%); control group: 23 (68%). Mean (SD) ICS dose (budesonide equivalents at entry in μg): Intervention (exercise + diet) group: 663 (370); control group: 739 (469) Inclusion criteria: Aged 18 to 65 years. BMI 20 to 30 kg/m2. ACQ score 1.0 or more. At least 1 positive diagnostic test demonstrating variable airflow obstruction. Either have been on stable prophylactic treatment regimen with inhaled ICS, ICS + LABA and/or leukotriene antagonist OR have had no prophylactic treatment at least 3 months before enrolment. Capable of exercising on bike. Sedentary (< 60 minutes of structured physical activity per week) Exclusion criteria: BMI > 30 kg/m2, BMI < 20.5 kg/m2. COPD. Pregnancy. Other inflammatory or metabolic diseases. Use of oral anti‐inflammatory medication or the use of antibiotic treatment during the last 8 weeks. Patients who were on oral corticosteroids or biological treatments were not included.* *Unclear if this was an exclusion criterion
Interventions	Intervention: Exercise and education: 8 weeks of high‐intensity interval training using the “10‐20‐30” concept on indoor spinning bikes 3 times per week supervised by a trained spinning instructor, with 5 group counselling sessions and 1 individual counselling session with a dietician regarding a high‐protein, low GI diet Comparison: No intervention. Encouraged to maintain usual physical activity levels and diet Concomitant medications: Participants should either have been on stable prophylactic treatment regimen with inhaled ICS, ICS + LABA and/or leukotriene antagonist OR have had no prophylactic treatment at least 3 months before enrolment. Excluded medications: Patients who were on oral corticosteroids or biological treatments were not included.* *Unclear if this was an exclusion criterion
Outcomes	Primary outcomes: Level of asthma control (ACQ) Secondary outcomes: Exercise capacity (VO2 max) on incremental cycle ergometer. Body composition (DEXA). Asthma‐related quality of life (MiniAQLQ). Lung function (spirometry). Sputum cell counts (sputum eosinophils, sputum neutrophils). Airway hyperresponsiveness (FeNO). Blood inflammatory markers (eosinophil count, serum levels of IL‐6, serum level of hs‐CRP). Urine urea excretion. Dietary GI Time points reported: Baseline, 8 weeks, 1 year Data reported as endpoint rather than as change from baseline.
Notes	Funding: None stated Notable conflicts of interest: None stated Other: Data from exercise + diet group and data from control group used for this review. Not all outcomes reported at 1 year. Email correspondence with author (A Bentzon) for VO2 data in mL/min/kg at 1 year for exercise + diet group and control group
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	“A computer‐generated block randomization method with a block size of 12, to ensure equal distribution of patients in treatment groups throughout the shifting seasons.”
Allocation concealment (selection bias)	Low risk	"The randomization was done using opaque sealed envelopes with a computer‐generated block randomization method with a block size of 12"
Blinding of participants and personnel (performance bias) Subjectively reported outcomes	High risk	Rehabilitation intervention ‐ Not blinded but inherently challenging.
Blinding of participants and personnel (performance bias) Not subjectively reported outcomes	High risk	Rehabilitation intervention ‐ Not blinded but inherently challenging.
Blinding of outcome assessment (detection bias) Subjectively reported outcomes	Low risk	“The investigators who carried out post intervention spirometry, mannitol tests, and handing out of questionnaires and all laboratory technicians were blinded to the randomization.”
Blinding of outcome assessment (detection bias) Not subjectively assessed outcomes	Low risk	“The investigators who carried out post intervention spirometry, mannitol tests, and handing out of questionnaires and all laboratory technicians were blinded to the randomization.”
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No ITT stated, but modest numbers of dropouts (<20%). Multiple imputation used but not clear what for.
Selective reporting (reporting bias)	Low risk	Almost all primary and secondary outcomes specified in clinical trial registry are reported on.
Other bias	Unclear risk	Considerable gender imbalance at baseline in one group – unclear if this may have affected outcomes, although appears to have been accounted for in analysis of primary outcome. No power calculation in the protocol, but trial registry suggests original target n = 200, and this was not achieved.