During their hospitalization, burn patients frequently require dressing changes that may be painful. Deep analgesia and sedation are used but carry the risk of remnant somnolence and other effects of anesthesia such as dizziness and nausea/vomiting. All these side effects may delay refeeding after the procedure, ambulation and physical therapy. Drugs from the opioid class are used to relieve pain during these procedures. Morphine with its slow onset and remnant sedation is difficult to use in these patients. Pro-emetic properties and histamine liberating effects also make this drug non optimal for iterative procedures. Fentanyl, a synthetic opioid with shorter onset and lower incidence of nausea and vomiting, is the standard drug used in dressing changes in burn patients. It is metabolized by hepatic glucoconjugation. Remifentanil, a well known novel opioid, that has a unique metabolism independent from renal or hepatic functions, is metabolized by a non specific esterase. It has a very short half-life (3.5 minutes) and should therefore be administered as a continuous infusion. The investigators hypothesized that the use of remifentanil for daily burn dressing changes is associated with less pain during procedures and faster recovery. Studied patients will be the ones requiring iterative dressing change procedures under sedation. The primary endpoint will be the maximal pain during the procedure. Secondary endpoints will be: average pain during and after the procedure; subjective sensation of comfort; total amount of opioids received; times to feeding after the procedure and ambulation after the procedure; comfort of the procedure according to the nurses; mobilisation according to the physical therapist; and safety of the analgesia technique. The study will be conducted according to the recommendations of the American Society of Anesthesiologists (ASA) that have been endorsed by the Canadian Anaesthetists' Society (CAS). All patients who consent will fast for at least 6 hours before the procedure. The two following regimens will be compared: a bolus infusion of fentanyl, starting with 1 µg/kg, followed by 0.5 µg/kg as needed every 5 to 10 minutes versus continuous infusion of remifentanil adapted to ensure analgesia. The initial dose of remifentanil will be 0.1 µg/kg/min to be adjusted between 0.05 µg/kg/min and 2 µg/kg/min. To allow blinding during the study, patients will receive a double-blinded protocol with sham (normal saline) in one arm. In other words: for each procedure, the patient will always receive boluses, either of fentanyl or saline, and a perfusion, either of remifentanil or saline. According to power calculations, 30 patients will be necessary to achieve the primary end-points. The investigators plan to enroll 40 patients in the study to allow for some drop outs and to increase their statistical power.