During a 12-week period, 204 consecutive patients admitted to the multidisciplinary intensive-care unit of a children's hospital were prospectively studied for complications of mechanical ventilation. METHODS A respiratory therapist completed a standardized data form at the end of each shift for each ventilated patient. Patient age, sex, length of ventilation, diagnosis, and complications were noted. Complications were classified as relating to the endotracheal tube (ETT), the ventilator, or the patient's medical management, and were analyzed according to incidence (number per 100 patients or per 100 ventilator days) and to associated mortality. RESULTS Patients ranged in age from newborn to 24 years. Sixty-three percent were male. Twenty-one percent of patients were managed by the medical staff, 11% by the general surgical staff, and 68% by the cardiac surgical staff. Average length of ventilation was 5.2 days. Overall survival rate was 91.7%. ETT complications reported as number per 100 patients were: pre-necrosis (13.0 [4/57 orally intubated patients and 23/147 nasally intubated patients]), ETT retaping complications (6.0), ETT plugging (1.0), and self-extubation (3.0). Ventilator complications reported as number per 100 ventilator days were: alarm failures (6.5), ventilator failures (0.7), and circuit problems (7.0). Medical complications reported as number per 100 patients were: massive gastric distension (8.8), right-upper-lobe collapse (4.4), pneumothorax (4.4), subcutaneous air (1.5), and pneumoperitoneum (1.0). ETT and ventilator complications showed no association with mortality. The large number of cardiac infants less than 24 months of age (n = 101) led us to further analyze this group for survival rate. We found that the survival rate was 93% for those requiring less than 7 days mechanical ventilation and 89.3% for those requiring greater than or equal to 7 days. As the study progressed, the respiratory therapists independently noted that their attentiveness to both patient and machine increased as did their awareness of complications. The incidence of alarm failure, circuit problems, and pre-necrosis was higher among the first 103 patients compared to the 101 patients entered into the study subsequently.