Study	Reason for exclusion
Bose 1987	Children had clear cut asthma defined as history of recurrent wheeze (> 5 episodes). RCT in children with nocturnal wheeze or cough.
Cazzola 1993	RCT (double blind cross over design) in 10 adults with angiotensin converting enzyme inhibitor induced cough. After 2 weeks of theophylline complete remission was seen in 8 of the 10 adults.
Chyrek 1988	Adult study. Participants had clear cut asthma defined with spirometry change.
Cloutier 1981	Non randomised controlled trial. 15 children with chronic cough treated with theophylline (15‐20 mg/kg/day). Cough resolved within 3‐5 days in all children.
Hannaway 1982	Non randomised controlled trial. 10 of 32 children had abnormal chest auscultation findings and hence some children in cohort did not have non‐specific cough. All responded to theophylline (20‐24 mg/kg/day). Authors reproted "dramatic response" in some children.
Konig 1981	Non randomised controlled trial. 6 of the 11 children in this cohort study had dyspnoea on exertion and 2 had chest pain. Thus at least 55% had recognisable clinical asthma. Theophylline (16‐20mg/kg/day) and metaproterenol given to all for 1‐2 weeks. Authors reported "disappearance of, or great reduction" in cough.
Perez 1994	Cross over RCT that examined the effect of a xanthine (acebrophylline for a week) compared to placebo (one week) using outcomes of cough (in addition to dyspnoea and wheeze). No washout period was described. The entry criteria (asthma and acute bronchitis) was undefined and mean baseline FEV1 was 72%. Thus some children clearly had asthma rather than non‐specific cough. The authors reported significantly less cough patient‐days in the acebrophylline group (2%) than in the placebo group (5.3%).
Rachelefsky 1980	RCT but clear cut asthma; defined as presence of airway reversibility on spirometry. Study compared metoproterenol to theophylline.
Selby 1997	RCT in adults comparing salbutamol and theophylline. No placebo.
Usmani 2005	RCT in well healthy adults (and animals) and not in adults with chronic cough. Authors assessed effect of theobromine on cough sensitivity .
Yahav 1982	Non randomised controlled trial. 4 of the 15 children had reversibility on spirometry and thus only 11 had non specific cough. 10 children treated with theophylline, 5 with inhaled salbutamol and 2 received additional corticosteroids. Authors reported that cough cleared in all children within 2‐3 days.