Objectives Loss of appetite is prevalent in palliative care and distressing for patients and families. Therapies include corticosteroids or progestogens. This study explores the net effect of dexamethasone on anorexia.

Methods Prospective data were collected when dexamethasone was started for anorexia as part of routine care. The National Cancer Institute's Common Toxicity Criteria for Adverse Events (NCICTCAE) Likert scales assessed severity of anorexia and immediate and short-term harms at 2 time points: baseline and 7 days.

Results This study (41 sites, 8 countries) collected data (July 2013 to July 2014) from 114 patients (mean age 71 (SD 11), 96% with cancer). Median Australian-modified Karnofsky Performance Scale was 50% (range 20–70). Mean baseline NCICTCAE anorexia score was 2.7 (SD 0.6; median 3). 6 patients died by day 7. Of 108 evaluable patients, 74 (68.5%; 95% CI 59.0% to 76.7%) reported ≥1 reduction anorexia scores by day 7, of whom 30 were 0. Mean dexamethasone dose on day 7 was 4.1 mg/day (SD 3.4; median 4; range 0–46 mg). 24 patients reported ≥1 harms (32.4% CI 22.6% to 44.1%; insomnia n=10, depression n=7, euphoria n=7 and hyperglycaemia n=7). Of 24 patients with no benefit, 10 reported ≥1 harms.

Conclusions This study shows positive and negative effects of 7 days of dexamethasone as an appetite stimulant in patients with advanced life-limiting illnesses. Identifying clinicodemographic characteristics of people most at risk of harms with no benefit is a crucial next step. Longer term follow-up will help to understand longer term and cumulative harms.