Effect of external PEEP in patients under controlled mechanical ventilation with an auto-PEEP of 5 cmH2O or higher

Giuseppe Natalini, Daniele Tuzzo, Antonio Rosano, Marco Testa, Michele Grazioli, Vincenzo Pennestrì, Guido Amodeo, Francesco Berruto, Marialinda Fiorillo, Alberto Peratoner, Andrea Tinnirello, Matteo Filippini, Paolo F. Marsilia, Cosetta Minelli, Achille Bernardini, for the VENTILAB group

In some patients with auto-positive end-expiratory pressure (auto-PEEP), application of PEEP lower than auto-PEEP maintains a constant total PEEP, therefore reducing the inspiratory threshold load without detrimental cardiovascular or respiratory effects. We refer to these patients as "complete PEEP-absorbers." Conversely, adverse effects of PEEP application could occur in patients with auto-PEEP when the total PEEP rises as a consequence. From a pathophysiological perspective, all subjects with flow limitation are expected to be "complete PEEP-absorbers," whereas PEEP should increase total PEEP in all other patients. This study aimed to empirically assess the extent to which flow limitation alone explains a "complete PEEP-absorber" behavior (i.e., absence of further hyperinflation with PEEP), and to identify other factors associated with it. One hundred patients with auto-PEEP of at least 5 cmH2O at zero end-expiratory pressure (ZEEP) during controlled mechanical ventilation were enrolled. Total PEEP (i.e., end-expiratory plateau pressure) was measured both at ZEEP and after applied PEEP equal to 80 % of auto-PEEP measured at ZEEP. All measurements were repeated three times, and the average value was used for analysis. Forty-seven percent of the patients suffered from chronic pulmonary disease and 52 % from acute pulmonary disease; 61 % showed flow limitation at ZEEP, assessed by manual compression of the abdomen. The mean total PEEP was 7 ± 2 cmH2O at ZEEP and 9 ± 2 cmH2O after the application of PEEP (p < 0.001). Thirty-three percent of the patients were "complete PEEP-absorbers." Multiple logistic regression was used to predict the behavior of "complete PEEP-absorber." The best model included a respiratory rate lower than 20 breaths/min and the presence of flow limitation. The predictive ability of the model was excellent, with an overoptimism-corrected area under the receiver operating characteristics curve of 0.89 (95 % CI 0.80-0.97). Expiratory flow limitation was associated with both high and complete "PEEP-absorber" behavior, but setting a relatively high respiratory rate on the ventilator can prevent from observing complete "PEEP-absorption." Therefore, the effect of PEEP application in patients with auto-PEEP can be accurately predicted at the bedside by measuring the respiratory rate and observing the flow-volume loop during manual compression of the abdomen.