Intermittent Mandatory Ventilation as a Primary Means of Ventilator Support
Intermittent Mandatory Ventilation and Controlled Mechanical Ventilation
Unlike IMV, CMV imposes a fixed breathing pattern. Hence, to some extent, a comparison favors IMV, which has been claimed to be superior to CMV for the following reasons: (a) it prevents the patient from “fighting the ventilator,” reducing the need for sedation and paralysis; (b) it prevents respiratory alkalosis; (c) it improves intrapulmonary gas distribution; (d) it lowers mean airway pressure, benefiting cardiac output and preventing barotrauma; (e) it decreases oxygen consumption (); (f) it prevents muscle atrophy and discoordination; and (g) it improves renal function.
IMV prevents the patient from “fighting the ventilator,” reducing the need for sedation and paralysis. Because the patient has no autonomy to alter breathing pattern with CMV, the patient will increase inspiratory effort, which is manifested as “fighting the ventilator” when ventilatory demand rises.78 No studies, however, have compared the respective sedative doses used during CMV and IMV.
IMV prevents respiratory alkalosis. In a prospective study comparing IMV and CMV in patients with acute respiratory failure, CMV resulted in a mean pH of 7.49 and a PaCO2 of 32.3 mm Hg, whereas IMV resulted in a mean pH of 7.44 and a PaCO2 of 39.9 mm Hg.7 The CMV rate, however, was set to suppress spontaneous respiratory effort, whereas, during IMV, the rate was adjusted downward continuously as long as pH remained greater than 7.30. Hence the result in favor of IMV is to be expected. When a normal pH is achieved at a low IMV rate, subsequent data suggested that it was at the expense of increased work of breathing.42,74 Moreover, some patients on mechanical ventilation who are allowed to set their own PaCO2 do not always choose a normal level (e.g., patients with brain injury). These patients will have persistent respiratory alkalosis regardless of ventilation mode.77,96
IMV improves intrapulmonary gas distribution. This hypothesis is based on the premise that in the supine position, spontaneous breathing causes inspired gas to be distributed preferentially to dependent lung regions because the dependent diaphragm, which is displaced cephalad, operates at an improved mechanical advantage.97 Conversely, during mechanical ventilation, ventilation is distributed preferentially to nondependent lung regions because the diaphragm is not used, and the nondependent lung and chest wall are more compliant.98 Therefore, IMV’s combination of spontaneous and mechanical breaths theoretically should result in a better matching of ventilation and perfusion.99 A comparison study evaluating IMV and CMV ventilation–perfusion distribution with the inert-gas elimination technique100 was performed in stable patients recovering from major abdominal aortic surgery.101 During CMV, VT was set at 14 to 16 mL/kg with a set rate of 8 to 10 breaths/min. During SIMV, the rate was set at 50% of the CMV rate (4 to 5 breaths/min); a pressure support of 5 to 8 cm H2O was added to the intervening spontaneous breaths to overcome the ventilator circuit and endotracheal tube resistance.102 Each ventilator mode was maintained for 45 minutes. Compared with CMV, physiologic dead space increased with SIMV (22.0% vs. 26.8%, respectively; p < 0.05) and was associated with a significant increase in , resulting in a similar PaCO2. The SIMV perfusion distributions remained unaltered. This study, using the inert-gas elimination technique, shows that IMV does not improve ventilation–perfusion distributions.
IMV lowers mean airway pressure, benefiting cardiac output and preventing barotrauma. Because IMV intersperses spontaneous breaths between mechanical breaths, mean airway pressure averaged over time is lower than with CMV. The lower mean airway pressure results in maintenance of cardiac output. Several studies have shown a higher cardiac output with IMV than with CMV.90,91,103 The interactions between intrathoracic pressure associated with mechanical ventilation and cardiac performance, however, are quite complex.104 Right-ventricular and left-ventricular interaction, direct pressure on the heart, and changes in systemic and pulmonary venous return all play a role. The net effect of this interaction depends on left-ventricular filling pressure and reserve. Mathru et al105 compared the effect of CMV, IMV with 5 cm H2O PEEP (IMV–5PEEP), and IMV with 0 cm H2O PEEP (IMV–0PEEP) in two groups of patients following aortocoronary bypass surgery. VT and ventilator rate were adjusted to achieve a PaCO2 of between 35 and 40 mm Hg. One group had normal left-ventricular function. The second group had decreased left-ventricular reserve: left-ventricular diastolic pressure of greater than 16 mm Hg and ejection fraction of less than 0.6. In the first group, IMV–0PEEP resulted in an increase in cardiac output of 27% compared with CMV. In the second group, however, IMV resulted in a significant decrease in cardiac output (19%) compared with CMV. IMV–5PEEP affected cardiac output similarly to CMV. These data indicate that when compared with CMV, IMV improves cardiac output in patients with normal left-ventricular function or hypovolemia, but it may be harmful in patients with poor left-ventricular reserve.
The frequency of barotrauma with IMV and CMV was compared in a retrospective study of 292 postoperative and nonsurgical patients who received mechanical ventilation for 24 hours or more.106 The ventilator VT was set at 12 to 15 mL/kg. The IMV rate was set at 6 breaths/min or lower, provided that normocapnia was maintained. If hypercapnia developed, it was treated by increasing VT to 15 mL/kg and, if necessary, by increasing the IMV rate to 8 breaths/min. No sedation or muscle relaxants were used. In the CMV group, the rate was set at 12 breaths/min. Hypercapnia was corrected by increasing VT. The patients were sedated and paralyzed. Compared with the CMV group, patients receiving IMV had a significantly higher peak airway pressure (34 vs. 51 cm H2O, respectively) and PEEP/CPAP (15 vs. 27 cm H2O, respectively). Yet the barotrauma was 22% in the CMV group compared with 7% in the IMV group. The authors speculated that the less frequent barotrauma with IMV was related to the smaller number of mechanical breaths with large VT. Mean transpulmonary pressure, which may be responsible for the barotrauma, was not measured in either group.
IMV decreases oxygen consumption. Downs et al7 found that was lower both during mechanical ventilation and 15 minutes after its discontinuation in patients ventilated with IMV versus CMV. The authors speculated that the higher on CMV could be ascribed to respiratory alkalosis107,108 and abrupt withdrawal from mechanical ventilation rather than reflecting metabolic work of breathing. In contrast, Wolff et al103 showed that CO2 production () tended to be lower with CMV than with IMV. Because the respiratory quotient was similar with both ventilation modes, it is reasonable to assume that during CMV was lower than that during IMV. In this study, PaCO2 was 39 and 44 mm Hg on CMV and IMV, respectively, suggesting that pH was within the normal range.103 This observation suggests that, in the absence of alkalemia, is lower on CMV than on IMV.
IMV prevents muscle atrophy and discoordination. Maintaining spontaneous breathing activity during IMV has been proposed to achieve respiratory muscle conditioning and preserve respiratory muscle function.5 The periodic hyperinflation also may reinforce coordinated breathing.5,109 During CMV, muscle atrophy84,85,110,111 and myofibrillar damage,82,112,113 which account for ventilator-induced diaphragmatic dysfunction, have been demonstrated in both experimental animals and humans.84,85 Like in experimental animals, in the human diaphragm the atrophic protein kinase B–forkhead Box-O (AKT-FOXO) signaling plays an important role in activating the ubiquitin-proteasome pathway.114 Overexpressions of the atrophic genes, atrogin-1 and muscle ring finger-1 (MuRF-1) of the ubiquitin-proteasome pathway, are responsible for myofibril degradation and atrophy with CMV-induced diaphragm muscle inactivity.114 Conversely, maintaining diaphragmatic activation with ACV attenuates expression of atrogin-1 (muscle atrophy F-box [MAF-box]; see Chapter 43).83 With IMV, diaphragmatic electrical activity of both mandatory and spontaneous breaths persists. Thus, it is possible that IMV prevents respiratory muscle atrophy. This hypothesis remains to be tested.
Respiratory muscle discoordination with IMV and CMV has not been compared. IMV’s efficacy in counteracting respiratory muscle discoordination was demonstrated by Andersen et al109 in a study of twenty-eight patients during discontinuation of mechanical ventilation. The mandatory breaths were increased gradually to 75% of the patient’s . In contrast, Gibbons et al115 failed to show similar results when IMV was applied to six patients receiving prolonged mechanical ventilation. The lowest spontaneous breathing rate was 29 breaths/min at the lowest IMV rate of 10 breaths/min. Gas exchange was adequate. Five of the six patients manifested breathing discoordination in the form of either ribcage or abdominal paradox. In the patients whose diaphragmatic electrical activity was measured, an electromyographic fatigue pattern also was observed. In this study, IMV was insufficient to reduce respiratory muscle workload, as reflected by the relatively high spontaneous breathing rate. IMV’s efficacy in counteracting respiratory muscle discoordination appears to depend primarily on the extent of respiratory muscle unloading.
IMV improves renal function. Steinhoff et al116 studied the effect of CMV and IMV on renal function in patients with acute respiratory failure. With CMV, the ventilator rate was set to suppress inspiratory efforts, which averaged 10 to 16 breaths/min. IMV was set at 4 to 10 breaths/min. PEEP was maintained constant during both CMV and IMV. With CMV, urinary flow and creatinine and osmolal clearance decreased, with a net effect of water retention, in comparison with IMV. Impaired renal function during CMV was attributed to the increased intrathoracic pressure, which caused stimulation of atrial stretch receptors and release of antidiuretic hormone. A more important factor may be that increased intrathoracic pressure during CMV decreases venous return, and the consequent decrease in cardiac output produces a decrease in renal plasma flow. Hence the effects of CMV and IMV on renal function are related directly to their respective effects on cardiac function.
Intermittent Mandatory Ventilation and Assist-Control Ventilation
Comparison of IMV with ACV is more appropriate than comparison of IMV with CMV because both IMV and ACV provide partial ventilator support. Few studies have compared IMV with ACV. As discussed below, these studies primarily concern effects on cardiac output, 117,118 ,117–119 and respiratory alkalosis.77,117,120 (Because IMV’s effects on work of breathing were discussed extensively earlier, comparison of IMV with ACV is limited to its application in the neonate.121,122)
Effect on cardiac output. Groeger et al117 studied the effect of SIMV and ACV in forty patients with acute respiratory failure of various etiologies other than chronic obstructive pulmonary disease. SIMV, at a set VT of 10 to 15 mL/kg, was the initial ventilation mode in all patients, and the mandatory rate was adjusted to the minimum required to maintain a normal pH and PaCO2. When the combined mandatory and spontaneous breathing rates were greater than 35 breaths/min, the IMV rate was increased to ensure patient comfort. The ventilator mode was then switched to ACV, and VT, PEEP, inspiratory flow rate, and inspired oxygen fraction (FIO2) were held constant. After 30 minutes on SIMV or ACV, hemodynamic variables were measured. At this point, mean IMV rate was 7 breaths/min, with a total rate of 34 breaths/min, whereas the ACV rate was 15 breaths/min. Cardiac output, measured by thermodilution, was 6% higher with SIMV than with ACV. Likewise, studying twelve patients recovering from acute respiratory failure of various etiologies, Sternberg and Sahebjami118 demonstrated that cardiac index was significantly higher with SIMV than with ACV (3.6 vs. 3.3 L/min/m2). (The investigators also compared SIMV with PSV, which is discussed in the section Intermittent Mandatory Ventilation and Pressure-Support Ventilation.) The average VT with ACV was 715 mL, whereas with SIMV it was 491 mL. The SIMV mandatory breaths were set at 75% of the ACV rate. Although the cardiac output with SIMV was significantly higher than with ACV, the changes fall within the variability of the thermodilution technique.123 Despite the limited differences in cardiac output with IMV and ACV, IMV may be helpful in patients who demonstrate significant hemodynamic deterioration during ACV. In the original description of intrinsic PEEP (PEEPi),124 two patients with chronic obstructive pulmonary disease developed hemodynamic compromise secondary to significant PEEPi while receiving ACV. The institution of IMV and fluid repletion produced an improvement, although cardiac output was not measured directly during either IMV or ACV. The conflicting data concerning the effects of ACV and IMV on cardiac output underscore the complex interaction between intrathoracic pressure and cardiac function.
Effect on oxygen consumption. In the above-mentioned study by Groeger et al,117 was comparable for both pressure-triggered SIMV and ACV. When the patients were grouped according to the ratio between achieved by the mandatory breaths and total , however, the mean for patients with a ratio of less than 0.5 was significantly higher during SIMV than during ACV (320 vs. 296 mL/min/m2, respectively; p ≤ 0.05). Conversely, in the study of Sternberg and Sebahjami,118 when the mandatory breaths during SIMV were set at 75% of the ACV rate, was unaltered during SIMV despite total frequency being higher with SIMV than with ACV (20 breaths/min vs. 12 breaths/min). In healthy subjects breathing via a mouthpiece on SIMV or ACV, (measured with a metabolic cart) also was similar with the two modes.119 With SIMV, both the VT and mandatory-breath rate were equivalent to those of the ACV. These studies117–119 demonstrate that compared with ACV, the degree of machine assistance during IMV determines .
Effect on respiratory alkalosis. Three prospective studies77,117,120 comparing IMV and ACV showed a significantly lower pH and higher PaCO2 during IMV than during ACV (Table 7-1). Groeger et al117 suggested that the higher PaCO2 during IMV was related to an increased dead-space-to-tidal-volume ratio, because and were similar for both IMV and ACV. Conversely, Hudson et al77 showed that the higher PaCO2 during IMV came at the expense of a high patient workload given elevated CO2 and an unchanged alveolar ventilation level. Regardless of the mechanisms of the elevated PaCO2 during IMV, all three groups of investigators concluded that the decrease in pH and increase in PaCO2 were minimal and of questionable clinical significance. Furthermore, in studies where IMV was compared with ACV in a subgroup of patients with preexisting respiratory alkalosis, respiratory alkalosis persisted during IMV.77,120
Effects on work of breathing in the neonate. Kapasi et al122 undertook a comparison of the effects of pressure-limited ACV, SIMV, and IMV on work of breathing and inspiratory effort of clinically stable neonates with respiratory distress syndrome. The mandatory breath rate with both SIMV and IMV was set the same as that of the ACV (range: 14 to 25 breaths/min). Average total respiratory rate was not significantly different among the modes (IMV, 56.3 breaths/min; SIMV, 58.3 breaths/min; ACV, 58.8 breaths/min). The work of breathing was estimated using the esophageal pressure, calculated using the Campbell diagram.125,126 Both work of breathing and inspiratory effort were least with ACV and highest with IMV; SIMV had values between ACV and IMV. Patient–ventilator asynchrony occurred only with IMV. Jarreau et al121 reported a similar result when comparing IMV, ACV with inspiratory pressure set at 10 to 15 cm H2O, and spontaneous breathing on CPAP. Work of breathing with IMV was similar to that with CPAP: 0.81 versus 0.90 J/L. Work of breathing fell significantly only during ACV to 0.48 to 0.47 J/L at inspiratory pressures of 10 and 15 cm H2O, respectively. Thus, in neonates, patient-triggered ACV provides better patient–ventilator synchrony and unloading of workload than does continuous-flow IMV.
Table 7-1: Effect of IMV versus ACV on pH and PaCO2 |Favorite Table|Download (.pdf)
Table 7-1: Effect of IMV versus ACV on pH and PaCO2
|pH||PaCO2 (mm Hg)||Rate (Breaths Per Minute)|
|63||7.41 ± 0.06a||7.45 ± 0.06||43.0 ± 6.3a||38.0 ± 6.3||7.1 (33.6)||15.1||40|
|102||7.48 ± 0.05b||7.51 ± 0.04||29.7 ± 6.1||28.6 ± 4.9||1/2 ACV rate||NA||26|
|105||7.42 ± 0.08a||7.45 ± 0.04||40.7 ± 7.6a||37.9 ± 6.7||4 (21)||15.0||18|
|Patients with Preexisting Respiratory Alkalosis|
|63||7.49 ± 0.03||7.49 ± 0.03||27.4 ± 6.3||29.1 ± 4.7||17|
|105||7.46 ± 0.07b||7.49 ± 0.03||37.8 ± 7.4||35.7 ± 6.7||12|
In summary, the limited number of studies suggests minimal differences between the effects of high levels of IMV and ACV on cardiac output, , and respiratory alkalosis. Because of lower mean airway and intrapleural pressures, IMV should help to improve cardiovascular function in patients who exhibit hemodynamic compromise during ACV. In neonates, ACV is more effective in unloading the work of breathing and providing synchrony than IMV.
Intermittent Mandatory Ventilation and Pressure-Support Ventilation
As with IMV, PSV provides the patient with some autonomy to alter breathing patterns in response to ventilatory demand. PSV is a form of ventilatory support in which the patient’s inspiratory effort is assisted by the ventilator up to a preset inspiratory pressure level and remains at that level until a fixed127,128 or operator-adjustable129 ventilator cycle-off algorithm is activated. Unlike IMV, in which the number of mandatory breaths is fixed, PSV assists every breath, and the ventilator contribution to total workload is variable. Because the set inspiratory pressure is fixed with PSV, when patient ventilatory demand increases, inspiratory effort may exceed the ventilator contribution to total workload (see Chapter 8).130,131 PSV can be added to IMV to unload inspiratory muscle work during the spontaneous breathing cycles. The addition of a small amount of PSV (5 cm H2O) to pressure-triggered SIMV is adequate to overcome the lack of flow delivery observed with some pressure-triggered SIMV systems.41 Higher levels of PSV not only help in overcoming the ventilator circuit and endotracheal tube resistance132,133 but also augment VT and unload the elastic work of the spontaneous breaths.134,135 In one study of a small number of patients with acute respiratory failure,136 PSV application at levels of up to 30 cm H2O during SIMV (IMV rate of 6 to 10 breaths/min) and PEEP of 3 to 13 cm H2O did not result in cardiovascular compromise. In clinically stable preterm infants, application of pressure support to the intervening spontaneous breaths to deliver VT of 5 to 8 mL/kg stabilized breathing pattern,137 and prevented an increase in inspiratory effort when SIMV rate was reduced.138 Comparison between IMV and PSV is discussed pertaining to cardiac output,118 ventilation–perfusion distribution,139 and unloading of patient effort.84
Effects on cardiac output. Hemodynamics during SIMV and PSV were studied in critically ill patients by initially applying ACV to the patients.118 The VT of the mandatory breaths with SIMV then was set the same as that of ACV, at a rate of 75% of the ACV rate. With PSV, the inspiratory pressure was set to produce a VT similar to that of ACV (average pressure of 21 cm H2O). Cardiac output (measured by thermodilution), oxygen transport, and were the same for both ventilation modes. As a primary means of ventilatory support, both SIMV and PSV have comparable effects on hemodynamics.
Effects on ventilation–perfusion. Valentine et al139 studied the effect of SIMV, PSV, and airway pressure-release ventilation (see in the section Intermittent Mandatory Ventilation and Airway Pressure-Release Ventilation) on ventilation–perfusion distribution in post–cardiac surgery patients who were ready to be weaned; this section discusses only the comparison between SIMV and PSV. SIMV was the initial ventilatory support mode. The IMV rate was adjusted to maintain a pH of greater than 7.35. With PSV, pressure was titrated to achieve a mean end-tidal PCO2 of 40 mm Hg. Ventilation–perfusion distribution was assessed using the inert-gas elimination technique.100 The dispersion of ventilation–perfusion ratios, calculated as the logarithmic standard deviation of perfusion (log SD) and ventilation (log SD), was similar for SIMV and PSV. Right-to-left intrapulmonary shunt and fractional dead-space ventilation did not differ significantly. Table 7-2 shows the effects on arterial blood gases, respiratory mechanics, and . Differences in arterial blood gases were of questionable clinical significance. Peak airway pressure was significantly higher with SIMV, but mean transpulmonary pressure was comparable. As sole ventilator support, SIMV and PSV provide comparable and adequate gas exchange in postoperative patients who were ready to be weaned. No study has yet compared the efficacy of IMV and PSV as a primary means of ventilator support during acute respiratory failure.
Effects on patient effort. Leung et al43 carried out a head-to-head comparison of the efficacy of SIMV and PSV in unloading inspiratory effort at various levels of assistance. The rate of change in inspiratory effort with increasing assistance levels, estimated as pressure-time product per minute, did not differ between the two modes. Unloading efficacy, however, differed according to the level of assistance. From 0% to 60% of maximum, the decrease in pressure-time product per minute was greater with PSV than with SIMV. At a higher assistance level, the converse was observed (Fig. 7-11). Frequency decreased linearly with increase in PSV. With SIMV, frequency changed little until a high assistance level was provided (Fig. 7-11).140 Thus, when a high assistance level is needed, both SIMV and PSV provide comparable assistance. At low to medium assistance levels, however, a greater decrease in patient effort with PSV makes it more useful clinically than SIMV. Leung et al43 also assessed the patient’s wasted efforts or nontriggered attempts during both SIMV and PSV. The number of nontriggered attempts is proportional to the assistance level and highest at 100% of machine assistance (29% with SIMV and 26% with PSV). The breaths preceding the nontriggered attempts had a shorter total duration and expiratory time, higher VT, and higher dynamic PEEPi than did the breaths preceding triggered breaths. This observation suggests that nontriggered attempts resulted from an inspiratory effort that was insufficient to overcome the elevated recoil pressure associated with dynamic hyperinflation. Thus, increasing the ventilator assistance level decreases inspiratory muscle effort but also increases ineffective triggering.43,141 Ineffective triggering is associated with prolonged duration of mechanical ventilation.142
Table 7-2: Gas Exchange, Mechanics, and Oxygen Consumption during SIMV and PSV |Favorite Table|Download (.pdf)
Table 7-2: Gas Exchange, Mechanics, and Oxygen Consumption during SIMV and PSV
|FIO2||0.44 ± 0.11||0.44 ± 0.11|
|pH||7.41 ± 0.02||7.36 ± 0.02a|
|PaCO2, mm Hg||33.0 ± 2.0||39.0 ± 2.0a|
|PaCO2, mm Hg||102.0 ± 7.0||98.0 ± 8.0|
|Peak Paw, cm H2O||32.8 ± 1.3||19.4. ± 2.1a|
|Mean Paw, cm H2O||9.6 ± 1.1||8.4 ± 1.0|
|Ppl, cm H2O||3.8 ± 1.0||3.8 ± 1.1|
|Ptp, cm H2O||5.8 ± 0.6||4.6. ± 0.5|
|, liters/min||9.4 ± 0.6||9.0 ± 0.5|
|fS, breaths per minute||3.4 ± 1.8|
|fM, breaths per minute||8.4 ± 0.4||15.8 ± 0.9a|
|VTS, liters||0.08 ± 0.07|
|VTM, liters||1.03 ± 0.03||0.58 ± 0.03a|
|, ml/min||269 ± 13||268 ± 14|
Changes in PTP per minute (left panel) and frequency (right panel) as intermittent mandatory ventilation (IMV) and pressure-support ventilation (PSV) were increased progressively. PSV of 100% represents the level necessary to achieve a VT equivalent to that during ACV (10 mL/kg); IMV 100% is the same ventilator rate and VT as during ACV. (Adapted, with permission, from Leung P, Jubran A, Tobin MJ. Comparison of assisted ventilator modes on triggering, patient effort, and dyspnea. Am J Respir Crit Care Med. 1997;155:1940–1948; and Tobin MJ, Jubran A, Laghi F. Patient-ventilator interaction. Am J Respir Crit Care Med. 2001;163:1059–1063.)
Intermittent Mandatory Ventilation and Airway Pressure-Release Ventilation
Airway pressure release ventilation (APRV) consists essentially of two CPAP levels with a transient decrease or “release” of airway pressure from a higher CPAP to a lower CPAP for a set release time. Spontaneous breathing is allowed to occur between airway pressure releases.128,143–145 The pressure gradient between the two CPAP levels and the frequency of releases determine the level of ventilator support (see Chapter 11). The original version of APRV used a short release time, and in the absence of spontaneous breathing, APRV resembled pressure-controlled inverse-ratio ventilation.143 The modified version, in which the inspiration-to-expiration (I:E) ratio is adjustable, is termed biphasic intermittent positive airway pressure.145 The primary indication for APRV is the provision of ventilation during an oxygenation crisis.143 Although the indications and operating principles of IMV and APRV differ significantly, both modes allow the patient to breathe spontaneously between machine-cycled breaths.143,144 Comparison of IMV with APRV is discussed with regard to its efficacy as a primary means of ventilator support in patients with acute respiratory distress syndrome (ARDS)146 in relation to gas exchange,139,146–149 hemodynamics,146,148,150 lung recruitment,151 breathing comfort,148,152 intubation duration, and sedative use.146,153
Varpula et al146 undertook a randomized, controlled comparison of pressure-limited SIMV (with 10 cm H2O of pressure support added to the unassisted breaths) and APRV in patients with ARDS. The primary end point was the number of ventilator-free days from the time of randomization to day 28. The secondary end points were the effect on gas exchange, hemodynamics, and sedative use. PEEP was set slightly above the lower inflection point on the pressure-volume curve obtained during paralysis (or if not detected, at 10 cm H2O). The upper inflection point was never exceeded (or if not detected, the inspiratory pressure was set at less than 35 cm H2O). The pressure-release frequency with APRV and the machine assistance rate with SIMV were similar at 12 breaths/min. The I:E ratio with APRV was set at 4:1; with SIMV, it was set at 1:2 and adjusted to 2:1. Ventilator-free days, gas exchange, hemodynamics, and sedative dosage were comparable with the two modes, although average inspiratory pressure was significantly lower with APRV than with SIMV (25.9 vs. 28.6 cm H2O). Moderate hypercapnia developed during the first study day in both groups but returned toward normal after 4 days. The effects of SIMV and APRV on gas exchange and hemodynamics appear comparable in patients with lung injury147,148 or after cardiac surgery.139 In contrast to earlier studies, recently, one trial in patients with ARDS demonstrated the superiority of APRV over SIMV in terms of gas exchange and mortality.149 Mortality was 31% with APRV compared to 59% with SIMV. Unfortunately, this study was a retrospective trial conducted over 6 years with SIMV applied in the first 3 years followed thereafter with APRV; consequently, ventilator settings were not well matched.149 Application of short-term APRV also proved superior to SIMV in patients who underwent coronary artery bypass surgery.150 Intubation duration was approximately 4 hours shorter with APRV than with SIMV (mean: 10.1 vs. 14.7 hours). Analgesic and sedative dosage also was less with APRV than with SIMV. This study, however, was a nonrandomized, open clinical trial.
In a recent, prospective, randomized trial in acute lung injury using similar ventilator settings as above,146 Varpula et al151 compared the effects APRV (n = 13) and SIMV plus PSV (n = 10) in recruiting nonaerated lungs. The investigators hypothesized that a strategy employing spontaneous breathing with active diaphragmatic contractions, as with APRV, is associated with better aeration of the dependent lung region.98 The degree of lung recruitment was estimated using computer-assisted tomographic densitometry technique at the beginning of randomization, and after 7 days of the assigned ventilator mode. On day 7, the median peak Paw was significantly lower with APRV than with SIMV: 24.0 versus 30.4 cm H2O. The computer-assisted tomographic number (in Hounsfield units) was similar with both APRV and SIMV at both the carinal and dependent diaphragmatic level. Thus, APRV did not appear to improve aeration of the consolidated lung parenchyma.
Breathing comfort during SIMV and APRV were evaluated in inexperienced healthy subjects breathing via a mouthpiece and compared with PSV. SIMV was set at 8 breaths/min, VT at 5 mL/kg, and PEEP at 5 cm H2O. The low and high PEEP levels with APRV were set at 5 and 10 cm H2O, the release rate was 8 breaths/min, and the I:E ratio was 1:2. PSV was set at 10 cm H2O with PEEP of 5 cm H2O. Breathing comfort was measured with a visual analog scale (0 to 10 cm). PSV achieved the greatest comfort, SIMV the worst, and APRV fell between PSV and SIMV (2.03, 5.38, and 4.12 cm, respectively). Unfortunately, flow-limited SIMV was employed rather than pressure-limited SIMV (both APRV and PSV are pressure-limited ventilation modes).
Intermittent Mandatory Ventilation and Proportional-Assist Ventilation, Adaptive-Support Ventilation, Pressure-Regulated Volume Control Ventilation, and Volume Support Ventilation
Proportional-assist ventilation (PAV) is a mode in which the ventilator instantaneously generates pressure in proportion to the patient’s effort (see Chapter 12).154 The ventilator amplifies the patient’s inspiratory effort without any preselected target volume or pressure. A randomized crossover comparison of continuous-flow IMV versus PAV was conducted in thirty-six preterm infants.155 The inspiratory pressure with IMV was set to deliver a VT of 4 to 6 mL/kg. The IMV rate was not reported. With PAV, the volume-assist gain was adjusted to decrease lung elastance to its normal value, whereas the flow-assist gain was set at –20 cm H2O/L/s. Each mode was applied for 45 minutes. With PAV, peak and mean airway pressures and transpulmonary pressure were significantly lower, frequency was higher, and VT was comparable to that with IMV, resulting in a higher . PaCO2, however, remained the same as with IMV, PaO2 and was higher. There were no significant differences in the number of apneic or hypoxemic episodes. The lower transpulmonary pressure with PAV might help to prevent lung injury during prolonged mechanical ventilation. The results in preterm infants were similar to those in an earlier short-term trial in a few adult patients.156
Adaptive-support ventilation (ASV) is based on the work by Otis et al157 and Mead et al,158 who demonstrated that any given level of alveolar ventilation has an optimal respiratory frequency that is least costly in terms of respiratory work: the frequency at which the respiratory muscles develop the least-average force or tension.158 ASV is a mode that can alternate between pressure control and pressure support, relying on closed-loop regulation of ventilator settings that respond to changes in respiratory system mechanics and spontaneous breathing efforts (see Chapter 15). ASV adjusts inspiratory pressure, I:E ratio, and mandatory rate to maintain the target minute ventilation and respiratory rate within a frame designed to avoid both rapid, shallow breathing and excessive inflation volumes.159 Tassaux et al160 compared the short-term effects of ASV versus SIMV plus pressure support on patient–ventilator interaction in patients ready to be weaned from mechanical ventilation. ASV achieved a similar to that of SIMV. Central neural drive, however, estimated as P0·1, and sternocleidomastoid electrical activity, measured with surface electrodes, were reduced markedly. Thus, ASV provided comparable total but with a significantly decreased inspiratory load than did SIMV plus pressure support.
Pressure-Regulated Volume-Controlled Ventilation
Pressure- regulated volume-controlled ventilation (PRVCV) is a dual-control, breath-to-breath mode. PRVCV has both the benefits of pressure-controlled ventilation, with a constant VT, and automatic weaning from pressure limit as patient compliance improves and/or patient effort increases (see Chapter 15). Four prospective, randomized trials have compared the efficacy of PRVCV and IMV with different primary goals in preterm infants.161–164 Two trials were short-term in stable preterm infants during the weaning phase,161,162 and two trials during acute respiratory failure.163,164 In the stable infants, SIMV was compared with PRVCV on the effects of respiratory mechanics and gas exchange.161,162 The SIMV mandatory breaths consisted of volume guarantee or pressure-regulated volume-controlled breaths. Abubakar and Keszler161 concluded that SIMV resulted in higher work of breathing than PRVCV on the basis of a high respiratory rate, although work of breathing was not directly measured. Scopesi et al162 demonstrated that SIMV was associated with lower mean Paw and higher variability in VT than observed with PRVCV. From those studies161,162 it is not clear whether PRVCV offers advantage over SIMV because the sample size was small (ten to twelve patients), and trial duration was short (20 minutes to 2 hours).
In infants with respiratory distress syndrome, Piotrowski et al163 studied infants younger than 3 days old whose birthweight was less than 2500 g. Thirty infants received continuous-flow IMV, and twenty-seven received PRVCV; the average Paw were 18.6 and 16.2 cm H2O, respectively. The IMV rate was selected by the clinician. With PRVCV, the VT was set at 5 to 6 mL/kg. The primary end point was duration of mechanical ventilation and incidence of bronchopulmonary dysplasia. The secondary end point was complications from mechanical ventilation, consisting of air leaks, intraventricular hemorrhage, and hemodynamic instability. PRVCV did not decrease duration of mechanical ventilation or incidence of bronchopulmonary dysplasia, although it decreased the incidence of high-grade intraventricular hemorrhage (11% vs. 35%). The benefit may have occurred in part because PRVCV delivers a stable volume. In preterm newborns, large fluctuations in intrathoracic and arterial pressures cause variations in cerebral blood flow velocity that is a risk factor for intraventricular hemorrhage.165 In the second trial, D’Angio et al164 enrolled a large number of low-birthweight infants (500 to 1249 g) who were younger than 6 hours of age. Infants were assigned to pressure-limited SIMV (n = 108) or PRVCV (n = 104) until extubation, death, or meeting predetermined failure criteria on the assigned mode. With SIMV, no pressure support was applied to the spontaneous breaths. Average Paw measured at 6 hours and 12 hours after SIMV application were 15.0 and 14.0 cm H2O, respectively, and after PRVCV were 13.8 and 12.7 cm H2O, respectively. The average VT for SIMV at 6 hours and 12 hours was 18.4 and 17.8 mL/kg, respectively; whereas VT with PRVCV was 16.0 and 14.1 mL/kg, respectively. The average set ventilator rate with SIMV at 6 hours and 12 hours was 35 and 30 breaths/min, respectively; with PRVCV, the rate was 40 breaths/min at both 6 hours and 12 hours. Differences in measured ventilator variables between SIMV and PRVCV were not significant. The cumulative percentage of infants alive and extubated at 36 weeks (84%) was similar for both SIMV and PRVCV (88 out of 105 infants for SIMV; 87 out of 104 infants for PRVCV). The percentage of infants who failed the assigned mode tended to be less with PRVCV than with SIMV, 20% versus 33%. Complications associated with prematurity and mechanical ventilation such as bronchopulmonary dysplasia, intraventricular hemorrhage, or retinopathy of prematurity were similar. Interestingly, with both SIMV and PRVCV, VT was remarkably higher than that recommended with a protective ventilatory strategy.166 Yet, when ventilator settings are nearly matched, application of SIMV or PRVCV appears to have a similar impact on survival, extubation rate, or complications rate in preterm infants with acute respiratory failure.
Volume-support ventilation (VSV), or PSV combined with volume guarantee, is PSV that uses VT as feedback control for continuously adjusting the pressure-support level. All breaths are patient-triggered, pressure-limited, and flow-cycled.52 Prospective comparisons between VSV and SIMV in preterm infants were conducted during the acute167 and stable phase of respiratory failure.162,168 Nafday et al167 randomized preterm infants following surfactant treatment for respiratory distress syndrome into VSV (n = 16) or pressure-limited SIMV (n = 18) applied for 24 hours. VT with VSV was set at 5 mL/kg. Ventilator settings were adjusted to achieve predetermined arterial blood-gas values. After 24 hours, attending physicians provided ventilation management as clinically indicated. Mean airway pressure decreased over time in both VSV and SIMV groups, although the decrease was faster in the SIMV group. Both groups had similar survival rates at the time of discharge (86% for VSV, and 94% for SIMV), and complications associated with prematurity and mechanical ventilation. That is, VSV did not offer improved clinical outcomes compared with SIMV.
Two prospective, short-term trials of stable preterm infants during the weaning phase compared VSV and pressure-limited SIMV for either 30 minutes168 or 20 minutes.162 In the first trial,168 following either VSV or SIMV for 30 minutes, VSV was applied for 24 hours. The attending physician, however, had discretion to select a ventilation mode according to the infant’s condition. VSV was successfully applied for 24 hours in twenty-one of 25 infants (84%), in those infants a significantly longer duration of rhythmic breathing was observed during the 30-minute trial. VSV provided comparable fluctuations in ventilation and oxygenation, but with a lower Paw than with SIMV (15.4 vs. 19.2 cm H2O, respectively), and shorter TI (0.29 vs. 0.35 seconds, respectively). In infants with high respiratory drive, however, hyperventilation occurred with VSV, which dictated a switch to SIMV.168 Consistent with the first trial, the second trial also demonstrated lower Paw with VSV compared with SIMV (11.2 vs. 18.2 cm H2O, respectively). In summary, VSV, in the short term, did not offer ventilatory advantage over SIMV in acute or stable preterm infants with respiratory distress syndrome, except for a lower Paw in the stable infants.
Intermittent Mandatory Ventilation as a Weaning Method
Intermittent Mandatory Ventilation, Pressure Support, and T Piece
IMV was first used in adult patients as a means of discontinuing mechanical ventilation.5 This method was claimed to be more efficient, safer, and more readily accepted by the patient, and it avoided the necessity of setting up a T-piece circuit. Although preceding ventilator support may be with either ACV or IMV, IMV is applied when the patient is ready to be weaned. The number of mandatory breaths is reduced gradually (1 to 3 breaths/min) at 1- to 4-hour intervals, provided that arterial pH remains greater than 7.307 or 7.35,169 regardless of other physiologic measurements. An IMV rate of zero or close to zero is maintained for several hours, or for as long as 24 hours, before extubating the patient.
With PSV as a weaning method, the pressure level is set initially at a maximum, defined as the level that produces a VT of 10 to 12 mL/kg; then the pressure support level is reduced according to the patient’s respiratory frequency.170 When the pressure support level reaches 5 cm H2O, extubation is considered.
With a T piece, once the patient meets predefined weaning criteria, the patient is placed on a T-piece circuit.171 Progressively longer intervals of spontaneous breathing through a T piece are alternated with ACV. Extubation is considered when the patient can sustain breathing through a T piece for 1 to 2 hours. The early claim for IMV’s superiority over a T piece was not based on a controlled study, and subsequent retrospective172 and prospective studies173,174 failed to demonstrate IMV’s superiority. Studies comparing IMV and PSV showed either a significantly reduced duration175 or a tendency for a shorter weaning time176 with PSV. As PSV grew in popularity, two prospective, randomized, controlled multicenter trials177,178 simultaneously compared the three weaning modalities: SIMV, PSV, and T piece. These trials laid to rest IMV’s claim to superiority over T piece and PSV.
Brochard et al177 studied 109 patients who met three of the four defined weaning criteria and had failed a 2-hour T-piece trial. Patients were randomized to SIMV (n = 43), PSV (n = 31), and T piece (n = 35). Weaning failure was defined as continued inability to be weaned after 21 days on the same mode, the need for reintubation after 48 hours of extubation, or intercurrent events (e.g., cardiac ischemia or nosocomial pneumonia) within 72 hours in the selected mode. The initial SIMV rate was set at half the total frequency during ACV or CMV, keeping VT and flow rates constant (mean initial SIMV rate of 9.5 breaths/min). Once or twice a day, the SIMV rate was decreased by 2 to 4 breaths/min if patients did not demonstrate signs of poor tolerance. When a patient demonstrated poor tolerance, the SIMV rate was increased to its preceding level. When a patient tolerated a SIMV rate of 4 breaths/min or less over 1 day, tracheal extubation was performed. The T-piece method consisted of a gradual lengthening of the periods of disconnection from the ventilator. The initial T-piece trial was set shorter than the initial tolerance duration (mean duration: 38 minutes). The number of T-piece trials depended on the length of disconnection and on nurse availability and varied from three to eight trials per day. The T-piece periods were lengthened incrementally twice a day. Between the T-piece trials, ACV was applied. When the duration of a T-piece trial had reached 2 hours with adequate gas exchange, tracheal extubation was performed. In the patients assigned to PSV, the initial pressure was adjusted until the frequency ranged between 20 and 30 breaths/min. Twice a day, the pressure was decreased by 2 to 4 cm H2O if the patient did not show any signs of poor tolerance; if tolerance worsened, pressure was increased to its preceding level. When the patient tolerated a PSV level of 8 cm H2O or less throughout a 24-hour period, tracheal extubation was performed. During the 21-day trial, the probability of being weaned was twice as high with PSV compared with SIMV or T piece. Weaning duration did not differ between SIMV and T piece but was significantly shorter with PSV than with the other two modalities (6 vs. 9 days, respectively). The number of weaning success patients was significantly larger with PSV (77%) than with SIMV (58%) or T piece (57%) (Table 7-3).
Table 7-3: Trials Comparing Three Weaning Methods: Intermittent T Piece, Pressure-Support Ventilation (PSV), and Synchronized Intermittent Mandatory Ventilation (SIMV) |Favorite Table|Download (.pdf)
Table 7-3: Trials Comparing Three Weaning Methods: Intermittent T Piece, Pressure-Support Ventilation (PSV), and Synchronized Intermittent Mandatory Ventilation (SIMV)
|Successful Wean [n (%)]|
|Brochard et al177||20/35 (57)||24/31(77)||25/43 (58)|
|Esteban et al178||27/33 (82)||23/37 (62)||20/29 (69)|
|Risk Difference (%)a|
|T Piece vs. PSV||T Piece vs. SIMV||PSV vs. SIMV|
|Brochard et al177||−20 (−42, 2)||−1 (−23, 21)||19 (−2, 40)|
|Esteban et al178||14 (−5, 32)||7 (−13, 27)||−7 (−20, 16)|
|Time to Extubation (Days)|
|Brochard et al177||8.5 ± 8.3||5.7 ± 3.7||9.9 ± 8.2|
|Esteban et al178||3(2, 6)||4(2, 12)||5(3, 11)|
In the study by Esteban et al,178 130 patients who met two of three weaning criteria and had failed a 2-hour T-piece trial were randomly assigned to one of four methods: SIMV (n = 29), PSV (n = 37), intermittent T-piece trials (two or more per day) (n = 37), and a once-daily T-piece trial (n = 31). The initial SIMV rate was set at half the frequency during ACV (average 10 breaths/min). When the patient tolerated it, SIMV was reduced twice a day by 2 to 4 breaths/min. Tracheal extubation was performed when the patient tolerated an SIMV rate of 5 breaths/min for 2 hours without signs of distress. With the PSV, the initial pressure was adjusted to achieve a frequency of 25 breaths/min or less (average pressure of 18 cm H2O). According to patient tolerance, pressure was reduced at least twice a day by 2 to 4 cm H2O. When the patient tolerated a PSV level of 5 cm H2O for 2 hours, tracheal extubation was performed. With intermittent T-piece trials, the patient breathed through a T-piece circuit or a continuous-flow CPAP of 5 cm H2O or less at least twice a day. Trial duration was increased gradually, and when the patient tolerated a 2-hour trial, extubation was performed. Between the T-piece trials, ACV was applied. With the once-daily T-piece method, the patient breathed through a T-piece circuit, after which ACV was resumed for 24 hours. Trial duration was increased gradually. When the patient tolerated a 2-hour trial, extubation was performed. Weaning failure was defined as the need for reintubation within 48 hours after extubation or the inability to extubate the patient after 14 days. The median successful weaning duration was 5 days for SIMV, 4 days for PSV, 3 days for intermittent T-piece trials, and 3 days for the once-daily T-piece trials. The rate of successful weaning for the once-daily T-piece method was three times faster than with SIMV (rate ratio: 2.83) and two times faster than with PSV (rate ratio: 2.05). The rate of success for intermittent or once-daily T-piece trials did not differ significantly. The percentage of patients weaned successfully was 69%, 62%, 82%, and 71% for SIMV, PSV, intermittent T piece, and once-daily T piece, respectively.
These two large randomized studies showed conflicting results. In the study of Brochard et al,177 PSV was superior to T piece and SIMV. In the study of Esteban et al,178 T piece was superior to PSV and SIMV. Despite the subtle differences in methodology, both trials demonstrated that weaning time was the longest with SIMV.179 Thus, both T piece and PSV were superior to SIMV (see Table 7-3).180,181
In weaning preterm infants, Dimitriou et al182 compared pressure-limited SIMV with ACV in two separate randomized, controlled trials. With both SIMV and ACV (n = 20 each), inspiratory pressure was reduced in decrements of 2 cm H2O until a defined target pressure tailored to the infant’s body weight was reached. With SIMV, in addition to decreasing pressure, the SIMV rate was reduced in decrements of 5 breaths/min until a target rate of 20 breaths/min was reached in the first trial. In the second trial, the target was 5 breaths/min. The frequency of decrements in pressure or SIMV rate was not reported. When the infants tolerated the target pressure (for ACV) or target pressure and rate (for SIMV), ventilation was switched to CPAP for 1 hour before extubation. The end-expiratory pressure was maintained at 3 cm H2O throughout the study. Weaning failure was defined as either the failure to achieve a reduction in ventilator support within 48 hours or requirement for reintubation within 48 hours of extubation. Reintubation was indicated when respiratory acidosis developed or frequent apneas or one major apnea occurred. In the first trial, there were no significant differences in the success rate with ACV or SIMV (70% vs. 75%) or the duration of successful weaning (median: 33 vs. 30 hours). In the second trial, differences between the weaning success rates were not significant, but the median weaning duration was significantly shorter with ACV (24 hours) than with SIMV (50 hours) (P < 0.05). A reduction in inspiratory pressure alone with ACV is favored in weaning preterm infants from mechanical ventilation than with SIMV, in which both inspiratory pressure and rate are reduced.
In weaning children ages 1 month to 4 years, Moraes et al34 compared pressure-limited IMV (n = 35) and SIMV plus pressure support (n = 35). Weaning began when peak inspiratory pressure was less than 25 cm H2O and FIO2 ≤0.6 (Time 0). Respiratory rate was reduced gradually by 3 to 5 breaths/min to 10 breaths/min. Then, PEEP was gradually reduced to 7 cm H2O. These ventilator settings were maintained for 12 to 24 hours. Extubation readiness was assessed daily for 2 hours while on FIO2 ≤ 0.5 with oxygen saturation as measured using pulse oximetry (SpO2) ≥95%, and PEEP 5 cm H2O; in patients on SIMV plus pressure support, the pressure support level was tailored to the size of the endotracheal tube (pressure support of 10, 8, and 5 cm H2O for endotracheal tube size of 3.0 to 3.5; 4.0 to 4.5; and >5.0, respectively).183 The average duration of weaning to extubation was 1 day (range: 1 to 6 days) for both groups. The frequency of extubation failure of 5.7% (because of upper respiratory distress) was also similar in both groups. Unlike adults177,178 or preterm infants,182 the ventilator mode has no effects on the outcome of discontinuation from mechanical ventilation in children.34
Intermittent Mandatory Ventilation and Mandatory Minute Volume, Adaptive-Support Ventilation
Mandatory minute ventilation (MMV) allows the patient to breathe spontaneously yet ensures that a preset minute ventilation is maintained should the patient’s spontaneous ventilation decline below the set level.184 MMV was developed to overcome certain ineffective features of IMV.185 When the set mandatory IMV rate is less than required to achieve adequate ventilation, alveolar hypoventilation will ensue whenever a patient’s total minute ventilation falls below a critical level. This drawback of IMV can be circumvented with MMV, which actuates a feedback control so that the ventilator provides pressurized breaths of a fixed volume to achieve a preset total minute ventilation.
Weaning with IMV and MMV was studied prospectively in forty patients recovering from acute respiratory failure caused by parenchymal lung injury and chronic airflow obstruction.186 After meeting defined weaning criteria, the patients were randomized to IMV (n = 18) or MMV (n = 22). In the IMV group, IMV rate was decreased by 2 breaths/min at 3- to 4-hour intervals during the daytime only until the IMV rate was equal to zero. Weaning was considered complete after 4 hours of breathing on CPAP. In the MMV group, MMV was set at 75% of the total minute volume preceding the weaning trial; this was achieved by decreasing frequency while maintaining a VT of 12 mL/kg as a reference value. Weaning was considered complete after 4 hours of independent spontaneous breathing. Weaning failure was defined as an inability to complete the trial or the need for ventilator support for the same underlying disease. Successful weaning was comparable: 86% for IMV and 89% for MMV. The weaning trial was longer in the IMV group (33 hours) than in the MMV group (4.75 hours).
In neonates with healthy lungs who were intubated for medical or surgical procedures, Guthrie et al187 conducted a crossover design, short-term trial (2 hours) of MMV versus SIMV. Mandatory breaths with both MMV and SIMV were flow-limited, volume-cycled (VT 4 to 6 mL/kg), whereas spontaneous breaths were augmented with pressure support. Both modes had comparable efficacy in carbon dioxide removal, yet with lower mean Paw with MMV. Mean rate of the mandatory breaths was also significantly lower with MMV than with SIMV (4.1 vs. 24.2 breaths/min, respectively). The authors postulated that both the reduced rate of the mandatory breaths and low mean Paw with MMV potentially reduced bronchopulmonary dysplasia complications associated with mechanical ventilation. Nevertheless, a prospective long-term follow-up is required. Moreover, SIMV as a weaning method has not been compared with MMV in neonates.
SIMV and ASV were compared as weaning modalities in a prospective, randomized study in post–cardiac surgery patients.188 With both ASV (n = 18) and SIMV (n = 16), the patients underwent three ventilation phases. With ASV, in phase 1, the initial settings were the ideal body weight, the desired minute volume at the default value of 100 mL/kg of ideal body weight, and peak airway pressure of less than 25 cm H2O. Adjustment of minute volume was dictated by a PaCO2 of less than 38 or greater than 50 mm Hg. Phase 1 ended when there were no controlled breaths for 20 minutes. Phase 2 was a continuation of phase 1; it ended when pressure support was decreased to 10 cm H2O (±2 cm H2O) and maintained for 20 minutes. The patient then entered into phase 3, where pressure support was set manually at 5 cm H2O for 10 minutes. When the patient showed satisfactory tolerance, tracheal extubation was performed. The initial settings for phase 1 in the SIMV group consisted of a VT of 8 mL/kg and an SIMV rate adjusted to achieve a PaCO2 of between 38 and 50 mm Hg. The SIMV rate was then set at 12 breaths/min. When spontaneous breaths exceeded 6 breaths/min for 20 minutes, the patient was switched to PSV of 10 cm H2O (phase 2). The patient was reassessed 20 minutes later for further reduction of PSV or returned to SIMV. If the patient tolerated it, PSV was reduced to 5 cm H2O (phase 3), as in the ASV group. There was no difference in duration of tracheal intubation, and all patients except for two (one in each group) were extubated within 6 hours. In the ASV group, patients required fewer manipulations of ventilator settings and endured fewer high inspiratory pressure alarms. This study was performed in postoperative patients who had received mechanical ventilation for less than 24 hours before weaning attempts. Because mechanical ventilation duration before weaning influenced the weaning success rate,189 the response of critically ill patients to the preceding weaning methods may be different.