Objective: To describe our hospital's experience with noninvasive positive pressure ventilation (bilevel positive airway pressure; BiPAP) for patients with respiratory failure (RF).
Design: Retrospective, observational study.
Setting: A 300-bed community teaching hospital.
Methods: Medical records were analyzed for physiologic and outcome variables for all patients who received BiPAP for RF between January 1994 and December 1996.
Results: Eighty patients with a mean (+/- S.E.) age of 71.5+/-1.3 years and APACHE II score of 17.2+/-0.6 received BiPAP for RF during the study period. Thirty-one patients received BiPAP for hypoxemic RF, 25 for acute hypercapnic RF, 9 for chronic hypercapnic RF, 10 for postextubation RF and 5 could not be categorized. BiPAP success was defined as no need for invasive ventilation. BiPAP was successful in 47 of 75 cases that could be classified; all BiPAP successes lived whereas 18 of 28 BiPAP failures died. In the overall cohort, BiPAP success was associated with a lower ICU length of stay (5.8+/-0.9 versus 10.6+/-1.4 days, p < 0.01). The duration of BiPAP dependency in successful cases was 35.3+/-6.7 h. BiPAP was successful in 20 of 25 patients with acute hypercapnic RF and in 15 of 31 patients with hypoxemic RF. The risk of BiPAP failure was significantly greater (risk ratio = 2.6, 95% CI = 1.1-6.1) for patients with hypoxemic than for those with hypercapnic RF. BiPAP success was marked by increased PaO2/FIO2 in patients with hypoxemic RF and by increased pH and reduced PCO2 in patients with hypercapnic RF. BiPAP use was also successful in 8 of 10 patients who developed RF within 48 h of endotracheal extubation.
Conclusions: BiPAP is highly effective in selected patients with RF during routine use in a community teaching hospital. The success rate is higher amongst patients presenting with hypercapnic than amongst those with hypoxemic RF and BiPAP failure is associated with an increased likelihood of in-hospital mortality. BiPAP may also be used successfully to temporize patients who develop RF in the period following endotracheal extubation. The duration of BiPAP dependency (35 h in this study) was shorter than in previous trials, and, though this is speculative, may have been minimized by our performing a trial of unassisted breathing each day.