BACKGROUND: At present, there is no established standard treatment for frail older patients with recurrent or metastatic head and neck squamous cell carcinoma. We aimed to compare the efficacy and safety of cetuximab to those of methotrexate (the reference regimen) in this population. METHODS: This randomised, open-label, phase 3 trial was done at 20 hospitals in France. Patients aged 70 years or older, assessed as frail by the ELAN Geriatric Evaluation, with recurrent or metastatic head and neck squamous cell carcinoma in the first-line setting and with an Eastern Cooperative Oncology Group (ECOG) performance status of 0-2 were eligible for inclusion. Patients were randomly assigned (1:1) to receive cetuximab 500 mg/m(2) intravenously every 2 weeks or methotrexate 40 mg/m(2) intravenously every week, with minimisation by ECOG performance status, type of disease evolution, Charlson Comorbidity Index score, serum albumin concentration, and geriatrician consultation. To avoid deterministic minimisation and assure allocation concealment, patients were allocated with a probability of 0·80 to the treatment that most reduced the imbalance. Treatment was continued until disease progression or unacceptable toxicity, whichever occurred first. The primary endpoint was failure-free survival (defined as the time from randomisation to disease progression, death, discontinuation of treatment, or loss of 2 or more points on the Activities in Daily Living scale, whichever occurred first) and was analysed in the intention-to-treat population. 151 failures expected out of 164 patients were required to detect a hazard ratio (HR) of 0·625 with 0·05 alpha error, with 80% power. A futility interim analysis was planned when approximately 80 failures were observed, based on failure-free survival. Safety analyses included all patients who received at least one dose of the study drug. This study is registered on ClinicalTrials.gov (NCT01884623) and was stopped for futility after the interim analysis. FINDINGS: Between Nov 7, 2013, and April 23, 2018, 82 patients were enrolled (41 to the cetuximab group and 41 to the methotrexate group); 60 (73%) were male, 37 (45%) were aged 80 years or older, 35 (43%) had an ECOG performance status of 2, and 36 (44%) had metastatic disease. Enrolment was stopped for futility at the interim analysis. At the final analysis, median follow-up was 43·3 months (IQR 30·8-52·1). At data cutoff, all 82 patients had failure; failure-free survival did not differ significantly between the groups (median 1·4 months [95% CI 1·0-2·1] in the cetuximab group vs 1·9 months [1·1-2·6] in the methotrexate group; adjusted HR 1·03 [95% CI 0·66-1·61], p=0·89). The frequency of patients who had grade 3 or worse adverse events was 63% (26 of 41) in the cetuximab group and 73% (30 of 41) in the methotrexate group. The most common grade 3-4 adverse events in the cetuximab group were fatigue (four [10%] of 41 patients), lung infection (four [10%]), and rash acneiform (four [10%]), and those in the methotrexate group were fatigue (nine [22%] of 41), increased gamma-glutamyltransferase (seven [17%]), natraemia disorder (four [10%]), anaemia (four [10%]), leukopenia (four [10%]), and neutropenia (four [10%]). The frequency of patients who had serious adverse events was 44% (18 of 41) in the cetuximab group and 39% (16 of 41) in the methotrexate group. Four patients presented with a fatal adverse event in the cetuximab group (sepsis, decreased level of consciousness, pulmonary oedema, and death of unknown cause) as did two patients in the methotrexate group (dyspnoea and death of unknown cause). INTERPRETATION: The study showed no improvement in failure-free survival with cetuximab versus methotrexate. Patients with an ECOG performance status of 2 did not benefit from these systemic therapies. New treatment options including immunotherapy should be explored in frail older patients with recurrent or metastatic head and neck squamous cell carcinoma, after an initial geriatric evaluation, such as the ELAN Geriatric Evaluation. FUNDING: French programme PAIR-VADS 2011 (sponsored by the National Cancer Institute, the Fondation ARC and the Ligue Contre le Cancer), GEMLUC, GEFLUC, and Merck Santé. TRANSLATION: For the French translation of the abstract see Supplementary Materials section.