We investigated the relationship between hypoxia, human papillomavirus (HPV) status and outcome in head and neck squamous cell carcinoma. Patients with stage III and IV head and neck squamous cell carcinoma treated on phase I and II chemoradiation trials with 70-Gy radiation combined with tirapazamine/cisplatin or cisplatin/fluorouracil (5FU), hypoxic imaging using [18F]-misonidazole positron emission tomography and known HPV status (by p16 immunohistochemistry) were included in this sub-study. Separate analyses were conducted to consider the impact of tirapazamine on HPV-negative tumours in the phase II trial. Both p16-positive oropharyngeal tumours and p16-negative head and neck squamous cell carcinoma tumours had a high prevalence of tumour hypoxia; 14/19 (74%) and 35/44 (80%), respectively. The distribution of hypoxia (primary, nodal) was similar. On phase II, trial patients with p16-negative hypoxic tumours had worse loco-regional control with cisplatin and 5FU compared with tirapazamine and cisplatin (P < 0.001) and worse failure-free survival (hazard ratio = 5.18; 95% confidence interval, 1.98-13.55; P = 0.001). Only 1 out of 14 p16-positive patients on the phase II trial experienced loco-regional failure. Hypoxia, as assessed by [18F]-misonidazole positron emission tomography, is frequently present in both p16-positive and negative head and neck cancer. Further research is required to determine whether hypoxic imaging can be used to predict benefit from hypoxia-targeting therapies in patients with p16-negative tumours.