This study examined predictors of prolonged hospitalisation (>30 days) and significant disability (modified Rankin Scale >2) in 257 patients with acute ischaemic stroke. These patients were assessed prospectively regarding stroke severity, comorbidities and complications in hospital. Multivariate logistic regression was used to select variables that best predicted prolonged hospital stay and significant disability on discharge. Four factors significantly predicted prolonged hospital stay: older age (>65); diabetes mellitus; in-hospital infection; and significant disability on discharge. Significant disability on discharge was in turn associated with diabetes, infection, premorbid disability, stroke in progression and atrial fibrillation. Diabetes and in-hospital infection, together with other factors, can significantly predict prolonged hospital stay and disability in stroke patients. These two potentially modifiable factors are possible targets for interventions to reduce the burden of illness and healthcare costs of stroke.