Comparison of tolerance of 4 interfaces for preventive non invasive ventilation after abdominal surgery in intensive care units assessed by patients and caregivers: A prospective randomized cross-over study.

INTRODUCTION

Interface selection is one of the major determinants of success for noninvasive ventilation (NIV). The aim of this study was to compare the tolerance of 4 interfaces (auto and hetero-evaluation) used during preventive NIV in Intensive Care Units (ICU).

METHODS

ICU patients receiving preventive NIV post-extubation after abdominal surgery were included in a prospective, single-center, randomized, crossover study. Four interfaces: two full-face (Helmet®, Bacou®) and two oro-nasal (Respironics®, Intersurgical®) interfaces were evaluated. An auto-evaluation (patients) and an hetero-evaluation (caregivers) were performed at the end of each NIV trial for each interface. Tolerance was evaluated with a visual numeric scale including: comfort (0 = maximum discomfort, 10 = perfect comfort), leaks (0 = maximum leaks, 10 = no leak), and communication (0 = no communication, 10 = optimal communication). A p-value <0.05/6 = 0.008 (Bonferroni correction) was considered significant.

RESULTS

Twenty-six consecutive patients were included. For auto-evaluation of comfort and leaks, no significant difference was observed between the interfaces. For hetero-evaluation of comfort, significantly higher scores were observed for Helmet (9 (8-10)) compared to Respironics and Intersurgical (respectively 9 (7-9) p = 0.0073 and 8 (7-8) p = 0.0046), whereas no difference was observed for hetero-evaluation of leaks. Concerning the auto-evaluation of communication, higher scores were observed for Helmet (9 (6-10)), in comparison to the other interfaces (5 (3-7) p = 0.003, 5 (3-8) (p = 0.0017, 2 (0-5) p < 0.0001) for Bacou, Respironics and Intersurgical, respectively). Similar results were observed for hetero-evaluation. The caregivers overestimated comfort scores and communication scores for each interface (p < 0.008), except for Helmet (p = 0.05).

CONCLUSION

The results suggest that none of the interfaces is universally better than the others, with no differences in comfort scores. The choice of interface in NIV should be personalized, and the patient asked for the preferred interface. Auto-evaluation differed from hetero-evaluation.