Predictors of extubation success and failure in mechanically ventilated infants and children.

OBJECTIVE: To predict extubation success and failure in mechanically ventilated infants and children using bedside measures of respiratory function.

DESIGN: Prospective collection of data.

SETTING: A university-affiliated children's hospital with a 51-bed critical care unit.

PATIENTS: All infants and children who were mechanically ventilated for at least 24 hrs, except neonates < or = 37 wks gestation and patients with neuromuscular disease.

INTERVENTIONS: Bedside measurements of cardiorespiratory function were obtained immediately before extubation.

MEASUREMENTS AND MAIN RESULTS: Extubation failure was defined as reintubation within 48 hrs of extubation in the absence of upper airway obstruction. Failure rates were calculated for different ranges (selected a priori) of preextubation measures of breathing effort, ventilatory support, respiratory mechanics, central inspiratory drive, and integrated indices useful in adults. Effort of spontaneous breathing was assessed by the respiratory rate standardized to age, the presence of retractions and paradoxical breathing, inspiratory pressure, maximal negative inspiratory pressure (maximal negative inspiratory pressure), inspiratory pressure/maximal negative inspiratory pressure ratio, and tidal volume indexed to body weight of a spontaneous breath. Ventilatory support was measured by the fraction of inspired oxygen (F10(2)), mean airway pressure, oxygenation index, and the fraction of total minute ventilation provided by the ventilator. Respiratory mechanics were assessed by determination of peak ventilatory inspiratory pressure and dynamic compliance. Central inspiratory drive was assessed by mean inspiratory flow. Frequency to tidal volume ratio and the compliance, rate, oxygenation, and pressure indexed to body weight, the integrated indices useful in predicting extubation failure in adults, were also calculated. Thirty-four of the 208 patients who were studied were reintubated for an overall failure rate of 16.3% (95% confidence interval 11.3% to 21.4%). The reasons for reintubation were poor effort (n = 8), excessive effort (n = 14), altered mental status or absent airway reflexes (n = 2), cardiovascular instability (n = 3), inadequate oxygenation (n = 3), respiratory acidosis (n = 3), and undocumented (n = 1). Extubation failure increased significantly with decreasing tidal volume indexed to body weight of a spontaneous breath, increasing F10(2), increasing mean airway pressure, increasing oxygenation index, increasing fraction of total minute ventilation provided by the ventilator, increasing peak ventilatory inspiratory pressure, or decreasing mean inspiratory flow (p < .05). Dynamic compliance showed a trend of increasing failure rate with decreasing dynamic compliance but did not reach statistical significance (p = .116). Respiratory rate standardized to age, inspiratory pressure, maximal negative inspiratory pressure, inspiratory pressure/maximal negative inspiratory pressure ratio, frequency to tidal volume ratio, and compliance, rate, oxygenation, and pressure did not show any trend in failure rate with increasing or decreasing values. Threshold values that defined a low risk (< or = 10%) and a high risk (> or = 25%) of extubation failure could be determined for tidal volume indexed to body weight of a spontaneous breath, F10(2), mean airway pressure, oxygenation index, fraction of total minute ventilation provided by the ventilator, peak ventilatory inspiratory pressure, dynamic compliance, and mean inspiratory flow. Neither a low nor a high risk of failure could be defined for frequency to tidal volume ratio or the compliance, rate, oxygenation, and pressure (CROP) index.

CONCLUSIONS: Bedside measurements of respiratory function can predict extubation success and failure in infants and children. Both a low risk and a high risk of failure can be determined using these measures. Integrated indices useful in adults do not reliably predict extubation success or failure in