Hair analysis has been used to reflect long-term systemic exposure to exogenous drugs and toxins. Several studies have demonstrated the feasibility of measuring endogenous steroid hormones, e.g. cortisol, in hair. Recently, a study in macaques showed a significant increase in hair cortisol levels induced by stress. We explored whether hair cortisol levels may be used as a biomarker for long-term stress in humans. Patients with severe chronic pain, aged 18 years or older, receiving opioid treatment for at least one year were recruited. Controls were non-obese (body mass index, BMI < 30 mg/kg(2)) adults. The Perceived Stress Scale (PSS) questionnaire was used to assess perceived stress over the last 4 weeks. A hair sample was obtained from the vertex posterior. Cortisol was measured using an enzyme-linked immunosorbent assay. We included fifteen patients (nine females and six males) and 39 non-obese control subjects (20 females, 19 males). PSS scores (median and range) were significantly higher in chronic pain patients (24: 12-28) than in controls (12: 3-31)(P < 0.001). Hair cortisol contents (median and range) were significantly greater in chronic pain patients (83.1: 33.0-205 g/mg) than in controls (46.1: 27.2-200 pg/mg) (P < 0.01). We conclude that hair cortisol contents are increased in patients with major chronic stress. Measurement of cortisol levels in hair constitutes a novel biomarker of prolonged stress.