STUDY OBJECTIVE: A prospective study of all unplanned adult extubations was conducted for 4 months period in four intensive care units (ICUs) of a community hospital. Our objective was to document the incidence of unplanned extubations, discern possible variables predictive of occurrence and outcome, and formulate preventive measures and guidelines for reintubation. DESIGN: Prospective case series. SETTING: Hospital with 525 beds and four ICUs: medical-surgical-trauma, neurology-neurosurgery, cardiovascular, and coronary care. PATIENTS: All adult patients intubated in four ICUs for a 4-month period. INTERVENTIONS: None. MEASUREMENTS AND RESULTS: Of 460 adults intubated in the 4-month period, 13 (3 percent) unplanned extubations occurred. Six patients were reintubated because of apnea, tachypnea, hypoxia, or other factors. Seven were not reintubated. Variables examined included the ventilator settings before to self-extubation, duration of intubation, arterial blood gases after self-extubation, Glasgow Coma Scale and Acute Physiology and Chronic Health Enquiry (APACHE) scores, time and location (ICU), nasal or oral method of intubation, and the mechanism of self-extubation. CONCLUSIONS: Our data suggested that self-extubation is relatively rare in our institution and that about half of self-extubated patients were reintubated. Staff vigilance, a proper weaning period, and the nasal method of intubation were some of the factors to which we attributed this low occurrence rate. However, a larger patient study population is required to show conclusively effective preventive measures and establish guidelines for reintubation.