Early complications and value of initial clinical and paraclinical observations in victims of smoke inhalation without burns

OBJECTIVE: To evaluate the incidence of early pulmonary complications and the value of initial clinical signs and paraclinical investigations in victims of smoke inhalation not suffering from burns following structural fires. DESIGN: Retrospective chart review. SETTING: Thirteen-bed ICU. PATIENTS: Sixty-four victims of smoke inhalation following household fires were admitted to the ICU between January 1987 and December 1992. Exclusion criteria from the study were patients with cutaneous burns or multiple trauma or blast injury, and patients found in cardiac arrest. METHODS: Clinical, biological, and radiologic parameters were collected over a 5-day period. RESULTS: The mortality rate in relation to progressive respiratory failure was 3.1%. Mean ICU stay was 5.8 days (range, 1 to 33 days), and was longer in the patients presenting with soot deposits in the oropharynx (p = 0.02), dysphonia (D) (p = 0.05), or ronchi (R) (p = 0.0004) at the first examination, and in those having a positive sputum bacteriologic analysis (p = 0.003) or requiring parenteral bronchodilator agents for more than 24 h (p = 0.04). Thirty-five patients underwent mechanical ventilation (MV) for a mean of 101.2 h (range, 8 to 648 h). Mean MV duration was higher in the patients presenting initially with R (p = 0.003), high carbon monoxide (but not cyanide) levels (p = 0.02), or a positive bacteriologic sample (p = 0.0001). Positive bacteriologic sampling correlated with the presence of D (p = 0.02) or R (p = 0.04) and with immediate intubation (p = 0.0003). No correlation was found with chest radiograph. CONCLUSIONS: In this selected series of fire victims without cutaneous burns, respiratory injury was frequent. The initial clinical signs may be helpful to predict pulmonary complications.