Are Low Tidal Volumes Safe? METHODS
This study was approved by the Committee on the Protection of Human Subjects in Research at the University of Massachusetts Medical Center, Worcester. Verbal informed consent was obtained from patients or responsible third parties within 24 h of entry into the trial.
All patients requiring mechanical ventilation in the surgical intensive care unit (SICU), except those recovering from cardiac surgery, were eligible for entry into the trial. Excluded after the trial was already under way were 33 victims of head trauma and patients recovering from craniotomies. Neurosurgical patients were excluded from both groups when it became apparent that efforts to maintain hypocarbia (PaC02 = 25 to 30 mm Hg) required excessively high respiratory rates (>26 breaths/min) when low Vt were employed. Patients were randomly assigned to either group 1 (Vt= 12 ml/kg) or group 2 (Vt = 6 ml/kg). A table of random numbers was used to generate the randomization schedule. The group assignments for sequential patients were enclosed in opaque envelopes that were opened in order as patients were entered into the trial.
Respiratory rate was adjusted to maintain PaCOz in the physiologic range (35 to 45 mm Hg). Positive end-expiratory pressure (PEEP) was routinely set at 5 cm H20 and increased if adequate oxygenation necessitated inspired fractional oxygen concentrations (FlOa) in excess of 0.5. Ventilator mode was not specified by the protocol and both synchronized intermittent mandatory ventilation (SIMV) and assist/control (A/C) modes were employed; SIMV was used preferentially, but in some instances, the clinical situation warranted switching to A/C mode. Puritan-Bennett (models 7200a, MA-1, and MA-2), Bear (models 1 and 2), and Siemens (Servo 900) ventilators were used and patients were randomly placed on the next disinfected machine. The inspiratory waveforms used were also not specified by protocol, but the majority of patients were started with the “down-sloping ramp” form. No inspiratory plateaus were used. Patients were sedated (usually with midazolam) as necessary. Decisions regarding weaning and extubation were guided by arterial blood gas (ABG) results, respiratory mechanics (forced vital capacity and inspiratory force), physical examination, and other pertinent clinical findings (eg, nutritional status and quantity of secretions). The decision to extubate was made by the attending physician staff of the SICU.
The presence of pulmonary infectious complications was determined on clinical grounds by a constellation of findings including the presence of otherwise unexplained fever of greater than 38.5°C, leukocytosis greater than 11,000/cu mm, and infected purulent sputum (tracheobronchitis). When the patient had roentgenographic evidence of pulmonary infiltrates in addition to the above criteria, a diagnosis of pneumonia was made. Both diagnoses were recorded and tabulated.
The following variables were recorded upon entry into the study: age, sex, principal diagnosis, and APACHE-II score. The following variables were recorded once daily between 8 and 10 am: Pa02, FIo2, peak airway pressure, and PEEP.
Durations of intubation and SICU stay were calculated by calendar days. Thus, for example, a patient admitted at 10 pm on April 3 and discharged at 11 am on April 4 was scored as having been in the SICU one day. Eight patients were excluded from the analyses of duration of intubation and duration of SICU stay, including the following: two quadriplegic patients with high cervical cord injuries requiring indefinite mechanical ventilation who were arbitrarily removed from the study after 30 days; five patients who died; and one patient in group 2 who was withdrawn from the study because she remained hypercarbic, despite respiratory rates greater than 26 breaths/min.
The analyses of the effect of Vt on pulmonary function (ie, peak airway pressure, PaO/FIOj, ratio, and PEEP requirement) were based on data recorded during the first seven days of ventilation. Between-group comparisons were performed in two ways: first, by using the extreme values for each variable recorded for each patient; and second, by using the mean value recorded for each patient. We chose to analyze data from only the first seven days because including a longer period of observation for the calculation of mean values tended to obscure the severity of acute respiratory problems.
Data were analyzed using the BMDP statistical software package on a mainframe computer (Harris 500). Where appropriate, numerical results are presented as mean ± SD, medians, and ranges. The significance of the difference in the incidence of pulmonary infectious complications was assessed using Fishers exact test. All other between-group statistical analyses were performed using the Mann-Whitney U test. Significance was declared for p<.05. A trend was considered present for .05<p<.l.