The trusted source for
healthcare information and
Abstract & Commentary
This is a study from Bordeaux, France, of immunosuppressed patients who were transferred from general medical wards to the ICU with acute hypoxic respiratory failure (AHRF). All patients had severely impaired arterial oxygenation (PaO2/FIO2 < 200 mm Hg on mask oxygen), a respiratory rate > 30/min, fever, and persistent pulmonary infiltrates on chest x-ray. Patients who were hypercapnic or hemodynamically unstable, or who had Glasgow Coma Scale scores < 8, congestive heart failure, more than 2 failing organ systems, or uncorrected bleeding diathesis were excluded, as were those who required urgent intubation on transfer.
The patients were stratified as to type of immunosuppression and were randomized to receive either standard therapy alone or that plus noninvasive positive-pressure ventilation (NPPV). The latter was delivered in the form of pressure support plus positive end-expiratory pressure via full-face mask so as to give an exhaled tidal volume of 7-10 mL/kg and respiratory rate < 25/min. NPPV was applied intermittently, but for at least 45 minutes at a time, with breaks every 3 hours. Standard criteria for intubation were used in all patients: a PaO2/FIO2 < 85 mm Hg, hypercapnia with pH < 7.30, agitation requiring sedation, seizure or other need for airway protection, hemodynamic instability, or inability to tolerate the face mask for effective delivery of NPPV.
Of 69 eligible immunosuppressed patients with AHRF, 52 (30 with hematologic malignancies and polymorphonuclear leukocytes < 100/mL; 18 were drug-induced; 4 with AIDS) were included in the study. In the NPPV group, a mean pressure support level of 15 ± 2 cm H2O was used, with PEEP 6 ± 1 cm H2O. NPPV was used 9 ± 3 h in the first 24 hours, 7 ± 3 h/d after day 1, for 1-9 days (mean, 4 ± 2 days).
The PaO2/FIO2 ratio improved from 141 to 180 mm Hg in the NPPV group as compared to a fall from 136 to 115 mm Hg in the standard care group, and this improvement was both initial and sustained (P < 0.005). Intubation was required in 12/26 (46%) of the NPPV patients as compared to 20/26 (77%) of the control patients (P = 0.03). Intubation was avoided in 47% of the patients with hematologic cancers and neutropenia who received NPPV as compared to 7% in the controls (P = 0.02). The interval to intubation and reasons for intubation were not different in the 2 patient groups. ICU mortality was 38% and 69% in the NPPV and control groups, respectively (P = 0.03).
Ventilator-associated pneumonia occurred in 2 (8%) of the NPPV patients and in 6 (23%) of the controls, although this difference was not statistically significant (P = 0.12). However, when all serious ICU-acquired complications were taken together, the difference was significant: 13 (50%) of patients vs. 21 (81%) in the 2 groups (P = 0.02). Ten NPPV patients (38%) had complications causing death in the ICU, as compared with 18 (69%) in the control patients (P = 0.03). (Hilbert G, et al. Noninvasive ventilation in immunosuppressed patients with pulmonary infiltrates, fever, and acute respiratory failure. N Engl J Med. 2001;344:481-487).
Comment by David J. Pierson, MD
This is an important study for clinicians who help to care for severely immunocompromised patients in the ICU. Its results indicate that NPPV, applied relatively early in hemodynamically stable patients with intact mental status, may prevent the need for intubation and invasive mechanical ventilation when such patients develop AHRF. Further, it suggests that by avoiding intubation and all its attendant complications, more of these patients can be brought through their episode of AHRF successfully, at least to the point of ICU discharge. Hilbert and colleagues suggest that avoidance of ventilator-associated pneumonia, an NPPV outcome that has has been suggested in studies of other patient populations, may be the fundamental reason for the mortality improvement.
Another study of NPPV in immunocompromised patients with AHRF—this one in individuals with solid organ transplants rather than with hematologic cancer—was published a few months before the Hilbert study and showed similar results. Antonelli and colleagues in Rome reported a randomized, controlled trial of 40 patients who had undergone liver (22 patients), lung (6), or kidney (12) transplants.1 Clinical presentations and criteria for NPPV management and intubation were similar to those used by Hilbert et al. Again, more patients in the NPPV group had improvement in PaO2/FIO2, and fewer of them required intubation—4 of 20 (20%) vs. 14 of 20 (70%); P = 0.002. Both the incidence of fatal complications (4 vs 10; P = 0.05) and of ICU mortality (4 vs 10; P = 0.05) were lower in the patients managed with NPPV.
Taken together, these studies suggest that patients with severe underlying immunocompromise who develop a syndrome of tachypnea, fever, persistent pulmonary infiltrates, and severe reductions in PaO2/FIO2, who have preserved mental status and who are hemodynamically stable, may benefit from a trial of NPPV. This is a patient population and clinical setting in which an extremely poor outcome is usual.2 If intubation can be avoided and the infectious or other precipitating process brought under control without the development of ventilator-associated pneumonia or other ICU complications, it may be possible to achieve a considerably better outcome in these patients.
1. Antonelli M, et al. JAMA. 2000;283:235-241.
2. Rubenfeld GD, Crawford SW. Ann Intern Med. 1996;125:625-633.