Chronic refractory breathlessness is common and distressing in advanced disease. Despite level I evidence to support the use of opioids for this symptom, not all patients benefit. This study aimed to discover which patient characteristics predict those most likely to gain improvement in breathlessness. This is an international, multicentre, retrospective analysis of 213 individual pooled datasets from four clinical trials of an opioid for chronic refractory breathlessness. "Response to opioid" was defined as 1) an absolute value of ≥10 mm improvement on the visual analogue scale (VAS) and 2) a relative value of ≥10% improvement from baseline VAS. We investigated baseline predictors using logistic regression. In the final model, higher baseline breathlessness intensity scores strongly predicted absolute and relative response (p<0.001). Younger age also predicted relative response (p=0.025); functional status and dominant cause of breathlessness did not. Some evidence supported the descriptor "not enough air", but was not statistically significant (p=0.052). A therapeutic trial of opioids is appropriate, irrespective of the cause of the breathlessness or functional status. Younger people or those with worse breathlessness are more likely to benefit. Opioids have a role in the management of chronic refractory breathlessness, but net benefit for individuals must be optimised.