Variability, reproducibility, and data-collection time of pulmonary bedside monitoring

Breath-by-breath pulmonary function testing at the bedside is now available both with special-purpose stand-alone equipment and with the new generation of ventilators. The authors studied the variability of, reproducibility of, and ideal length of data collection for nine indices of pulmonary function that may be useful for ventilatory management, weaning, and patient comfort. Work of breathing (as both J/L and J/min), pressure-time product, rapid shallow breathing index, respiratory time fraction, respiratory drive, change in esophageal pressure during inspiration, expiratory airway resistance, and dynamic compliance were measured in ten normal subjects and in eight patients being weaned from mechanical ventilation. All nine indices were reproducible when compared by paired t-test with two separate sets of data collected in normal subjects. Repeated measures in the normal subjects allowed calculation of 95% confidence intervals for the nine variables. There was no statistically significant difference between data collections of 5 minutes compared with those of 10 and 15 minutes. Breath-by-breath variability ranged from a coefficient of variation of 3% for the shallow breathing index in one patient to 131% for the work of breathing in J/min in another. Population variability ranged from values reported previously for other pulmonary parameters to nearly double for some parameters. The authors conclude that a 5-minute data collection time is sufficient to obtain reliable breath-by-breath data at the bedside. While taken together these indices may provide clinically useful information, their usefulness individually remains to be demonstrated because of their large variability.