OBJECTIVE: The objective of this study was to evaluate the ratio of dead space to tidal volume (VD/VT) as a predictor of extubation failure of children from mechanical ventilation. METHODS: From September 2001 to January 2003 we studied a cohort consisting of all children (1 day-15 years) submitted to mechanical ventilation at a pediatric intensive care unit who were extubated and for whom pre-extubation ventilometry data were available, including the VD/VT ratio. Extubation success was defined as no need for any type of ventilatory support, invasive or otherwise, within 48 hours. Patients who tolerated extubation, with or without noninvasive support, were defined as success-R and compared with those who were reintubated. Statistic analysis was based on a VD/VT cutoff point of 0.65. RESULTS: During the study period 250 children received mechanical ventilation at the pediatric intensive care unit. Eighty-six of these children comprised the study sample. Twenty-one children (24.4%) met the criteria for extubation failure, with 11 (12.8%) of these requiring non-invasive support and 10 (11.6%) reintubation. Their mean age was 16.8 (+/-30.1) months (median = 5.5 months). The mean VD/VT ratio for all cases was 0.62 (+/-0.18). Mean VD/VT ratios for patients with successful and failed extubations were 0.62 (+/-0.17) and 0.65 (+/-0.21) (p = 0.472), respectively. Logistic regression failed to reveal any statistically significant correlation between VD/VT ratio and success or failure of extubation (p = 0.8458), even for patients who were reintubated (p = 0.5576). CONCLUSIONS: In a pediatric population receiving mechanical ventilation due to a variety of etiologies, the VD/VT ratio was unable to predict the populations at risk of extubation failure or of reintubation.