Mickleborough 2005.

Methods	Randomised placebo‐controlled cross‐over study.
Participants	24 participants (mean age 24 years) were recruited from a university population and local community. 14 participants were on inhaled short beta 2 agonists and 12 participants on inhaled corticosteroids. All had documented exercise‐induced asthma ‐ wheezing, shortness of breath and chest tightness after exercise and atopic asthma. Exclusion criteria: pregnancy, hypertension, hyperlipidaemia, diabetes, bleeding disorders, delayed clotting time or taking aspirin.
Interventions	All participants had a run‐in period of 1 week on a normal sodium diet; after this all participants consumed a low sodium diet (65 mmol of sodium a day as meal plan). Participants were then randomised to either high sodium limb (sodium tablets 174 mmol of sodium a day) or low sodium limb (placebo tablets). After 2 weeks there was a wash‐out period on a normal sodium diet for one week. Finally participants had two weeks on the different limb.
Outcomes	Pre‐ and post‐exercise FEV1, FVC, FEV1/FVC, FEF25‐50%, DLCO, KCO, Va, DMC0, VL, VC/VA. Post‐exercise measures were 1, 5, 20, 45, 75, 90, 105, 120 minutes. In addition inhaled sputum was collected for total cell count, neutrophils, lymphocytes, macrophages, epithelial cells, interleukin 8, mean leukotriene b4, cysteinyl leukotriene, PGD2‐methoxine. This was collected pre‐exercise, 1 hour, 6 hours and 24 hours post‐exercise.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Independent investigator had no contact with the subjects and no involvement in data collection or analysis used a computerised random number generator to create the randomisation sequence.
Allocation concealment (selection bias)	Low risk	Randomised study with independent investigator generating the sequence.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Described as blinded; no other information available.
Incomplete outcome data (attrition bias) All outcomes	Low risk
Selective reporting (reporting bias)	Unclear risk	Not clear from information given.
Other bias	Unclear risk	Subjects had to stop maintenance medication during the study. The study was controlled for cross‐over effect. Power calculation was not fully described.