Pressure Support Ventilation versus  Continuous Positive Airway Pressure with the Laryngeal Mask Airway: A Randomized Crossover Study of Anesthetized Adult Patients

Forty consecutive adult patients (American Society of Anesthesiologists physical status I or II, aged 18–80 yr) undergoing peripheral musculoskeletal surgery in which the LMA was considered to be suitable were studied. Ethics committee approval and written informed consent were obtained from the Leopold-Franzens University and patients, respectively. A standard anesthesia protocol was followed and routine monitoring was applied. Patients were induced with 2 μg/kg fentanyl and 2.5 mg/kg propofol. Maintenance included 6 mg · kg⁻¹· h⁻¹propofol in oxygen and air, with the inspired oxygen concentration adjusted to 30%. One experienced LMA user (> 1,000 uses) inserted or fixed the LMA (size 5, all patients) 7according to the manufacturer’s instructions (Laryngeal Mask Co., Henley-on-Thames, UK). 8The volume of air in the cuff was adjusted so the airway sealing pressure was more than 15 cm H²O. The lungs were manually inflated until spontaneous breathing resumed. Mechanical ventilation was begun (Evita 4; Draeger Medizintechnik GmbH, Luebeck, Germany) and the patients were allocated randomly to two crossover groups. In group 1 (n = 20), patients underwent CPAP, PSV, and CPAP in sequence. In group 2 (n = 20), patients underwent PSV, CPAP, and PSV in sequence. PSV comprised positive end expiratory pressure (PEEP) set at 5 cm H²O and pressure support set at 5 cm H²O above PEEP. CPAP was set at 5 cm H²O. Pressure support was initiated when inspiratory flow reached 3 l/min. Each ventilatory mode was maintained for 10 min. No minimal respiratory rate (RR) or apnea backup was set on the ventilator. The following data were recorded by an unblinded observer once every minute for the last 5 min for each ventilatory mode and the average reading used for analysis: end tidal carbon dioxide (ET^CO2), oxygen saturation (Sp^O2), expired tidal volume (VTExp), leak fraction (LF), RR, noninvasive mean arterial pressure, and heart rate. Sp^O2was measured using the finger pulse oximeter. Spirometry was measured from the inspiratory and expiratory limb of the anesthesia circuit. In addition, epigastric auscultation was performed to detect air entering the stomach. Measurements were made before surgery. LF was determined by noting the difference between inspired and expired tidal volume. Statistical analysis was performed using the Bonferroni corrected t  test. Significance was taken as P < 0.05.

The mean (range) age, height, and weight was 36 yr (21–60 yr), 171 cm (158–188 cm), and 74 kg (50–110 kg), respectively. The male:female ratio was 19:21. There were no demographic differences between groups. Data are presented in table 1. All data were normally distributed. In both groups, PSV had lower ET^CO2(P < 0.001), higher Sp^O2(P < 0.001), and higher VTExp (P < 0.001) compared with CPAP. In both groups, PSV had similar LF, RR, mean arterial pressure, and heart rate compared with CPAP. In group 1, measurements for CPAP were similar before and after PSV. In group 2, measurements for PSV were similar before and after CPAP. Data for CPAP were similar between groups and for PSV were similar between groups. Gastric insufflation was not detected.

Our data show that PSV provides more effective gas exchange than unassisted ventilation with CPAP during anesthesia with the LMA. Minute volume increased by 50%, but ETCO²only decreased by 20% (50–40 mmHg), suggesting that 30% of the additional ventilation is wasted. This is probably related to an increase in alveolar dead space. We found that PSV improves Sp^O2by approximately 1.5%. This may be related to the increase in alveolar ventilation because the fraction of inspired oxygen was low. PSV and CPAP both cause an increase in mean airway pressure, resulting in a larger lung volume available for gas exchange. Interestingly, we found no reduction in RR with PSV. A reduction in RR with PSV has been reported in intubated critically ill patients 9and in awake healthy subjects breathing through a mouth piece. 10It may be that higher levels of PSV would have triggered a reduction in RR in our patients.

We found that LF was less than 1% and that gastric insufflation did not occur. Devitt et al.  11measured LF and gastric insufflation at peak airway pressures of 15–30 cm H²O and found that LF was 13–27% and gastric insufflation was 2–35%. Berry et al.  12reported leak fractions of less than 3% with the LMA at tidal volumes of 10 ml/kg. Peak airway pressures did not exceed 10 cm H²O and were at least 5 cm H²O below airway sealing pressure. Air leakage and gastric insufflation are highly unlikely at this level of peak pressure and airway sealing pressure.

We detected no differences in cardiovascular effects between PSV and CPAP. This is not surprising because the average intrathoracic pressure changes would have been similar between groups. No differences in cardiovascular effects were detected when comparing PPV with spontaneous breathing during LMA anesthesia. 2Finally, we did not measure work of breathing, but this is reduced during pressure support in intubated patients, 13and we speculate that this also occurs during PSV with the LMA.

We conclude that PSV provides more effective gas exchange than does unassisted ventilation with CPAP during anesthesia with the LMA while preserving LF and hemodynamic homeostasis.