Background The role of the neurokinin-1 (NK-1) receptor antagonists in the prevention of radiation-induced nausea and vomiting has not been established. The purpose of the GAND-emesis study was to investigate the efficacy and safety of fosaprepitant in combination with palonosetron and dexamethasone in the prevention of nausea and vomiting during 5 weeks of fractionated radiotherapy and concomitant weekly cisplatin in patients with cervical cancer. Methods This investigator initiated, multinational, randomised, double-blind, placebo-controlled phase 3 trial, included women with cervical cancer scheduled to receive fractionated radiotherapy and weekly cisplatin 40 mg/m2 for 5 weeks. Patients had to be naive to chemotherapy and radiotherapy. Patients were randomly assigned to receive either single doses of fosaprepitant 150 mg intravenously or placebo (saline) in combination with palonosetron 0·25 mg intravenously and dexamethasone 16 mg orally before cisplatin administration. Randomisation was done by the unmasked pharmacist, who used a list of six numbers (a block) provided in a sealed envelope. A web-based randomisation number generator was used to generate the full list of randomisation numbers that was split up in blocks of six numbers. All patients received oral dexamethasone 8 mg twice a day on day 2, 4 mg twice a day on day 3, and 4 mg once on day 4. The treatment was repeated for 5 weeks. The primary endpoint was the proportion of patients with sustained no emesis after 5 weeks of treatment. The modified intention-to-treat population (all patients who received study medication) was used for the statistical analyses. The study was registered with ClinicalTrials.gov, number NCT01074697. Findings Between June 15, 2010, and March 8, 2015, 246 patients from four countries consented to the study and were randomly assigned. Of these, 234 patients were eligible, having received study medication (118 received fosaprepitant, 116 received placebo). The proportion of patients with sustained no emesis at 5 weeks (competing risk analysis) was 48·7% (95% CI 25·2–72·2) for the placebo group compared with 65·7% (42·2–89·2) of patients for the fosaprepitant group. There was a significantly lower cumulative risk of emesis in the fosaprepitant group compared with the placebo group (subhazard ratio 0·58 [95% CI 0·39–0·87]; p=0·008). Treatments were generally well tolerated with few grade 3 adverse events none of which were related to the study treatment; the most common grade 3 adverse event during the 5 weeks of treatment was diarrhoea (11 [9%] of 118 patients in the fosaprepitant group vs six [5%] of 116 patients in the placebo group). There was only one report of a grade 4 adverse event (neutropenia), in the fosaprepitant group. No deaths were recorded in either group. Interpretation To our knowledge, this is the first study to investigate safety and efficacy of a NK-1 receptor antagonist during 5 weeks of radiotherapy and concomitant weekly cisplatin. Patients receiving fosaprepitant in addition to palonosetron and dexamethasone were less likely to experience emesis and nausea compared with those receiving palonosetron and dexamethasone alone. Both treatments were safe and well tolerated. Further investigations in other radiotherapy settings are warranted. Funding Private and hospital or university funding, unrestricted grants from Biovitrum and Helsinn Healthcare SA.