Tan 2016.

Study characteristics
Methods	Design: randomised, placebo‐controlled, parallel‐group design Duration: 52 weeks Location: outpatient department at a medical university hospital
Participants	Population: 49 adults with COPD (GOLD stages 2 to 4) were randomly assigned to erythromycin (n = 36) or control (n = 18) Baseline characteristics: age (mean years): 68.4 (SD 6.6); % male: 89; % FEV₁ predicted (mean): 44.4 (SD 13.8); pack‐years (mean): 41.3 (SD 25.8); exacerbation history: not reported Inclusion criteria: stable COPD diagnosis according to GOLD guidelines, FEV₁ < 80% predicted, FEV₁‐to‐FVC < 70%; no acute exacerbation, no change in therapeutic treatment, no antibiotic treatment in the last 4 weeks Exclusion criteria: bronchial asthma; primary bronchiectasis; diffuse pan bronchiolitis; active tuberculosis; lung cancer; pneumoconiosis; other lung disease with restrictive ventilatory impairment; cardiovascular, nervous system, or endocrine disease; blood, hepatic, or kidney disease; malignant tumour, inability to communicate, serious adverse reaction to erythromycin
Interventions	Group A: erythromycin (125 mg erythromycin 3 times a day for 6 months) Group B: erythromycin (125 mg erythromycin 3 times a day for 6 months followed by 6 months' follow‐up) Control group Allowed co‐medications: supplemental oxygen, theophylline, inhaled bronchodilators and corticosteroids. Other macrolides, histamine antagonists, non‐steroidal anti‐inflammatory drugs, and oral glucocorticosteroids were not allowed
Outcomes	Concentrations of IL‐17 and IL‐23 in peripheral blood and induced sputum, 6‐minute walk distance, serious adverse events
Notes	Funding: National Nature Science Foundation of China (81460009) and Guangxi Natural Science Foundation Identifier: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"Randomly divided"; no other details were provided
Allocation concealment (selection bias)	Unclear risk	No further information was provided
Blinding of participants and personnel (performance bias) All outcomes	High risk	No blinding of participants or personnel was described. Study was assumed to be open‐label (although abstract states double‐blind). Study authors have been contacted; we await clarification response
Blinding of outcome assessment (detection bias) All outcomes	High risk	No blinding of outcome assessors was described. Study was assumed to be open‐label (although abstract states double‐blind). Study authors have been contacted; we await clarification response
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Dropout was low and balanced, but details of how many people were analysed at each time point were not given
Selective reporting (reporting bias)	Unclear risk	No prospective trial registration or protocol was identified, so it is not clear if outcomes of interest for this review may have been collected but not reported (e.g. serious adverse events, exacerbations, quality of life)
Other bias	Low risk	No other bias was identified