Patients with borderline resectable pancreatic ductal adenocarcinoma have relatively low resection rates and poor survival despite the use of adjuvant chemotherapy. The aim of our study was to establish the feasibility and efficacy of three different types of short-course neoadjuvant therapy compared with immediate surgery. ESPAC5 (formerly known as ESPAC-5f) was a multicentre, open label, randomised controlled trial done in 16 pancreatic centres in two countries (UK and Germany). Eligible patients were aged 18 years or older, with a WHO performance status of 0 or 1, biopsy proven pancreatic ductal adenocarcinoma in the pancreatic head, and were staged as having a borderline resectable tumour by contrast-enhanced CT criteria following central review. Participants were randomly assigned by means of minimisation to one of four groups: immediate surgery; neoadjuvant gemcitabine and capecitabine (gemcitabine 1000 mg/m2 on days 1, 8, and 15, and oral capecitabine 830 mg/m2 twice a day on days 1-21 of a 28-day cycle for two cycles); neoadjuvant FOLFIRINOX (oxaliplatin 85 mg/m2, irinotecan 180 mg/m2, folinic acid given according to local practice, and fluorouracil 400 mg/m2 bolus injection on days 1 and 15 followed by 2400 mg/m2 46 h intravenous infusion given on days 1 and 15, repeated every 2 weeks for four cycles); or neoadjuvant capecitabine-based chemoradiation (total dose 50·4 Gy in 28 daily fractions over 5·5 weeks [1·8 Gy per fraction, Monday to Friday] with capecitabine 830 mg/m2 twice daily [Monday to Friday] throughout radiotherapy). Patients underwent restaging contrast-enhanced CT at 4-6 weeks after neoadjuvant therapy and underwent surgical exploration if the tumour was still at least borderline resectable. All patients who had their tumour resected received adjuvant therapy at the oncologist's discretion. Primary endpoints were recruitment rate and resection rate. Analyses were done on an intention-to-treat basis. This trial is registered with ISRCTN, 89500674, and is complete. Between Sept 3, 2014, and Dec 20, 2018, from 478 patients screened, 90 were randomly assigned to a group (33 to immediate surgery, 20 to gemcitabine plus capecitabine, 20 to FOLFIRINOX, and 17 to capecitabine-based chemoradiation); four patients were excluded from the intention-to-treat analysis (one in the capecitabine-based chemoradiotherapy withdrew consent before starting therapy and three [two in the immediate surgery group and one in the gemcitabine plus capecitabine group] were found to be ineligible after randomisation). 44 (80%) of 55 patients completed neoadjuvant therapy. The recruitment rate was 25·92 patients per year from 16 sites; 21 (68%) of 31 patients in the immediate surgery and 30 (55%) of 55 patients in the combined neoadjuvant therapy groups underwent resection (p=0·33). R0 resection was achieved in three (14%) of 21 patients in the immediate surgery group and seven (23%) of 30 in the neoadjuvant therapy groups combined (p=0·49). Surgical complications were observed in 29 (43%) of 68 patients who underwent surgery; no patients died within 30 days. 46 (84%) of 55 patients receiving neoadjuvant therapy were available for restaging. Six (13%) of 46 had a partial response. Median follow-up time was 12·2 months (95% CI 12·0-12·4). 1-year overall survival was 39% (95% CI 24-61) for immediate surgery, 78% (60-100) for gemcitabine plus capecitabine, 84% (70-100) for FOLFIRINOX, and 60% (37-97) for capecitabine-based chemoradiotherapy (p=0·0028). 1-year disease-free survival from surgery was 33% (95% CI 19-58) for immediate surgery and 59% (46-74) for the combined neoadjuvant therapies (hazard ratio 0·53 [95% CI 0·28-0·98], p=0·016). Three patients reported local disease recurrence (two in the immediate surgery group and one in the FOLFIRINOX group). 78 (91%) patients were included in the safety set and assessed for toxicity events. 19 (24%) of 78 patients reported a grade 3 or worse adverse event (two [7%] of 28 patients in the immediate surgery group and 17 [34%] of 50 patients in the neoadjuvant therapy groups combined), the most common of which were neutropenia, infection, and hyperglycaemia. Recruitment was challenging. There was no significant difference in resection rates between patients who underwent immediate surgery and those who underwent neoadjuvant therapy. Short-course (8 week) neoadjuvant therapy had a significant survival benefit compared with immediate surgery. Neoadjuvant chemotherapy with either gemcitabine plus capecitabine or FOLFIRINOX had the best survival compared with immediate surgery. These findings support the use of short-course neoadjuvant chemotherapy in patients with borderline resectable pancreatic ductal adenocarcinoma. Cancer Research UK.