INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ventilator displays of exhaled tidal volume (VT) are traditionally used to indicate the delivered VT for critically ill patients. Many ventilators measure expired VT and airway pressure at the expiratory valve of the ventilator. Such a measurement of delivered VT does not compensate either for the compliance of the ventilator circuit or for uncontrolled variations in the circuit setup, including heaters, in-line suction devices, end-tidal carbon dioxide monitor adapters, and condensation. These confounding variables potentially invalidate the ventilator-derived VT values. Theoretically, a VT measured with a pneumotachometer positioned between the endotracheal tube and the ventilator circuit would be more reliable at indicating the VT actually delivered to the patient's lungs than would a value measured at the expiratory valve of the ventilator. Additionally, expiratory volumes may be more clinically useful than inspiratory volumes, since the inspiratory values may be falsely elevated in the presence of an endotracheal tube air leak, which is common when ventilating infants and children.

Multiple conventional mechanical ventilators, as well as respiratory mechanics monitors with pneumotachometers, are commercially available. The question that exists is not which type of device (ventilator or respiratory mechanics monitor) is inherently more accurate in measuring a volume, but rather which monitoring method better indicates the true VT delivered to the lungs on the basis of location of the relevant instrumentation in the ventilatory circuit (expiratory valve or endotracheal tube). A review of the literature reveals no prior studies comparing VT obtained at these two different locations.

To potentially achieve the goal of determining the actual VT delivered to the lungs but without requiring additional equipment, a formula can be derived to correct for the compliance of the ventilator circuit. In using this formula, the "effective" VT is defined as the ventilator-measured VT minus the volume lost because of the distensibility of the ventilator circuit. This effective VT is calculated as the ventilator-measured expired VT  (circuit compliance × [peak inspiratory pressure (PI_max)  positive end-expiratory pressure (PEEP)]) (1). The compliance of the ventilator circuit can be obtained from the manufacturer or can be calculated from the pressures and VT values measured at both ends of the circuit. However, variations in the ventilator circuit setup (i.e., heaters, humidifiers, water traps, in-line suction devices, and other devices), as well as condensation in the circuit, are not included in such a calculation. When VT is determined with a pneumotachometer positioned at the endotracheal tube, it is measured after the ventilator circuit, and variability caused by the compliance of the ventilator circuit can therefore be eliminated.

Differences among the ventilator-determined VT, the pneumotachometer-determined VT, and the calculated effective VT can be clinically significant. This is especially true for neonates, infants, and young children, owing to the relatively small VT used and the potentially significant contribution of the compliance of the ventilatory circuit as well as of variations in the circuit setup. We therefore hypothesized that neither the expiratory VT displayed by the mechanical ventilator nor the calculated effective VT would reliably represent the exhaled VT as measured with a pneumotachometer placed at the endotracheal tube during conventional mechanical ventilation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ninety-eight conventionally ventilated infants and pediatric patients were studied. Each patient was ventilated with a Servo 300 ventilator (Siemens-Elema, Solna, Sweden). The need for informed consent was waived by the Institutional Review Board of the Duke University Medical Center, where the study was conducted. Respiratory parameters were measured both with the Servo 300 ventilator and with a pneumotachometer connected to a respiratory mechanics monitor (VenTrak or CO₂SMO Plus; Novametrix Medical Systems, Wallingford, CT). The pneumotachometer was placed between the ventilator circuit and the endotracheal tube, as shown in Figure 1.

View larger version (19K): [in this window] [in a new window]   Figure 1.   Schematic diagram demonstrating pneumotachometer placement. The pneumotachometer is placed between the endotracheal tube and the ventilator circuit. The pneumotachometer is connected to the endotracheal tube by a standard "elbow" connector. The pneumo-tachometer can also be used for monitoring end-tidal carbon dioxide levels.

The respiratory mechanics monitor used in the study measures flow with a fixed-orifice differential pressure pneumotachometer located at the endotracheal tube. Respired gas flowing through the flow sensor causes a small pressure decrease across the two tubes connected to the sensor. This pressure decrease is transmitted through the tubing to a differential pressure transducer located inside the monitor, and is correlated with flow according to a factory-stored calibration. The pressure transducer is automatically "zeroed" to correct for changes in ambient temperature and electronics. Data are sampled at a rate of 100 Hz, which is significantly greater than any ventilator-based measurement rate used in conventional ventilation.

Ventilator circuit type and diagnosis were recorded for each patient. Circuit compliance specifications were 0.61 ml/cm H₂O for the infant circuit and 1.0 ml/cm H₂O for the pediatric circuit (Allegiance Healthcare Corporation, McGaw Park, IL). All ventilators, respiratory mechanics monitors, and pneumotachometers were calibrated and validated for accuracy according to the manufacturer's recommendations before and after data collection.

After mechanical ventilation had been optimized by the patient-care team, data from mechanical ventilator breaths were recorded from both the ventilator display and the respiratory mechanics monitor display. Measurements included expired VT, PI_max, and PEEP. Patients were grouped according to circuit type. Additionally, the effective VT was calculated as described earlier.

In order to compare the inherent accuracy of the VT measurements made with both the ventilator and the pneumotachometer, we performed a "bench test". For this, we utilized the SV300 along with a 15-cm ventilator circuit created from noncompliant tubing. The pneumotachometer was placed between the rigid circuit and a PMG 3000 test lung (Ingmar Medical, Pittsburgh, PA). No other circuit accessories were included in the ventilator circuit. This setup was designed to eliminate all of the confounding variables in the clinical situation. Data were collected at a range of settings in both the volume-control ventilation mode (set VT: 50 to 150 ml in 5-ml increments) and the pressure-control ventilation mode (PI_max: 10 to 50 cm H₂O in 5-cm H₂O increments). PEEP was maintained at 5 cm H₂O.

Statistical Analysis

All data were analyzed separately for the two circuit groups (infant and pediatric). Analysis of variance with repeated measures was used to compare the exhaled VT recorded with the respiratory mechanics monitor-pneumotachometer with both the ventilator-measured VT and the calculated effective VT. A regression analysis was then performed for each of the two circuit groups to describe any potential linear relationship between the different VT measurements. Similar data analysis was done for the VT data obtained during the bench-test validation. Additionally, a paired t test was used to compare the PI_max and the PEEP recorded by the respiratory mechanics monitor with the corresponding values from the ventilator display. Data are represented as mean ± SD.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Seventy patients ventilated with infant circuits and 28 patients ventilated with pediatric circuits were studied. The infant population had an age of 2.8 ± 2.3 (mean ± SD) mo (median: 2.8 mo) and a weight of 4.6 ± 3.2 kg (median: 3.8 kg). The pediatric patients had an age of 7.3 ± 5.6 yr (median: 5.7 yr) with a weight of 27.9 ± 22.9 kg (median: 17.0 kg). The largest group of patients studied were patients with congenital heart disease (n = 43; 44%). Respiratory failure from respiratory syncitial viral infection was the reason for ventilation in approximately 9% of the patient population (n = 9). Patients with respiratory failure from other infectious causes comprised 15% of the patient population (n = 15). Other diagnoses included reactive airways disease (n = 3), endocarditis (n = 1), gastroschisis (n = 1), persistent pulmonary hypertension of the newborn (n = 2), pulmonary hemorrhage (n = 2), sickle cell vasoocclusive crisis (n = 1), upper airway obstruction and/or airway surgery (n = 2), trauma (n = 2), and thymus transplantation (n = 1).

VT Comparisons

For the infants in the study (n = 70), the mean VT measured by the respiratory mechanics monitor-pneumotachometer was significantly less than the VT determined either by the ventilator (39.4 ± 21.5 ml versus 70.4 ± 31.1 ml, p < 0.0001) or the calculated effective VT (39.4 ± 21.5 ml versus 59.2 ± 28.8 ml, p < 0.0001). In the group of infants, regression analysis for comparison of the VT measured by the respiratory mechanics monitor (x axis) with the ventilator display of VT measured at the expiratory valve (y axis) revealed a correlation coefficient (r²) of 0.54 and gave an equation of y = 1.06x + 29 (Figure 2A). Regression analysis of the measured VT from the respiratory mechanics monitor (x) against the calculated effective VT (y) revealed an r² of 0.58 and gave an equation of y = 1.02x + 19 (Figure 2B).

View larger version (21K): [in this window] [in a new window]   Figure 2.   VT comparisons for neonatal circuit. (A) Regression analysis for infant ventilator circuit as ventilator VT versus respiratory mechanics-pneumotachometer VT. (B) Regression analysis for effective VT versus respiratory mechanics-pneumotachometer VT.

For the pediatric circuit (n = 28), the mean VT measured by the respiratory mechanics monitor-pneumotachometer was significantly less than the VT determined by the ventilator (135.3 ± 75.8 ml versus 185.4 ± 96.6 ml, p = 0.03). However, the VT measured by the respiratory mechanics monitor-pneumotachometer was not statistically different from the calculated effective VT (135.3 ± 75.8 ml versus 167.8 ± 94.6 ml, p = 0.16). In this group of pediatric patients, the comparison of the VT measured by the respiratory mechanics monitor (x axis) with the ventilator display of VT (y axis) revealed an r² of 0.84 and gave an equation of y = 1.16x + 28 (Figure 3A). Comparison of the VT measured by the respiratory mechanics monitor (x axis) with the calculated effective VT (y axis) revealed an r² of 0.85 and gave an equation of y = 1.15x + 12 (Figure 3B).

View larger version (20K): [in this window] [in a new window]   Figure 3.   VT comparisons for pediatric circuit. (A) Regression analysis for infant ventilator circuit as ventilator VT versus respiratory mechanics-pneumotachometer VT. (B) Regression analysis for effective VT versus respiratory mechanics-pneumotachometer VT.

Bench-Test Validation

The values of VT determined from both the ventilator and the pneumotachometer during the bench-test validation did not differ from each other. During volume-control ventilation the set VT was varied between 50 ml and 150 ml. Values of VT were recorded from both the pneumotachometer placed at the end of the noncompliant ventilator circuit and from the ventilator (96 ± 39 ml versus 98 ± 39 ml, p = 0.72). Regression analysis for this relationship resulted in an r² of 1.0. Similarly, VT values were recorded with a PI_max that varied between 10 cm H₂O and 50 cm H₂O in the pressure-control mode. The resultant VT values showed no significant differences (pneumotachometer VT 190 ± 54 ml, versus ventilator VT 205 ± 69 ml, p = 0.70). Regression analysis for this relationship resulted in an r² of 0.98.

Airway Pressure Comparisons

The PI_max as measured by both the ventilator and the respiratory mechanics monitor was similar for the infant circuit (27.0 ± 4.7 cm H₂O versus 28.5 ± 4.5 cm H₂O, p = 0.12), and this was also true for the pediatric circuit (28.3 ± 6.5 cm H₂O versus 28.1 ± 6.6 cm H₂O, p = 0.92). PEEP also did not differ between the ventilator and the respiratory mechanics monitor for either type of ventilator circuit (infant circuit: 5.9 ± 1.8 cm H₂O, versus 5.6 ± 1.8 cm H₂O, p = 0.38; and pediatric circuit: 7.6 ± 3.0 cm H₂O versus 7.1 ± 3.0 cm H₂O, p = 0.64).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Many conventional ventilators measure airway pressure and VT at the expiratory valve, and therefore do not account either for the compliance of the ventilator circuit or for uncontrolled variations in the circuit setup. The ventilator circuit compliance is a particularly relevant factor in determining the actual volume delivered to the lungs of neonates, infants, and small children, given the overall small values of VT used. In the neonate, small inaccuracies in VT may result in significant adverse consequences for the patient. If the actual volume delivered to the patient's lungs is unknown, the critically ill infant may be at increased risk for lung injury, hypoxia, and hypercapnia (2).

The need to know the exact VT delivered is essential when ventilating infants, since the volume lost because of the distensibility of the circuit may be equal to the desired VT. Our data show that in infants, the expiratory VT as measured at the endotracheal tube is on average only 56% of that measured at the ventilator. The pediatric circuit showed a somewhat better correlation, with the average VT at the endotracheal tube measuring 73% of the volume measured at the expiratory valve of the ventilator.

Theoretically, the effect of circuit compliance on the accuracy of the ventilator-determined VT can be mathematically eliminated. An effective VT can be calculated by subtracting the VT lost in the ventilator circuit from the VT displayed by the ventilator (1). However, our data show that this calculation is not sufficient. When compared with the exhaled VT measured with a pneumotachometer placed at the endotracheal tube, the correlation with the effective VT remained poor. Factors affecting the accuracy of this calculation include condensation in the ventilator circuit and additions to the circuit, such as adapters for end-tidal carbon dioxide monitors, in-line suction devices, humidifiers, and heaters. All of these factors alter the accuracy calculation of effective VT by adding uncontrolled, and variable, dead space to the circuit.

If an inappropriately small VT is utilized, atelectasis and ventilation-perfusion mismatching may occur (5). If atelectasis develops, increased airway pressures may be required to "recruit" the collapsed lung regions, potentially leading to increased barotrauma through shear injury (5). Although atelectasis can be overcome by increasing VT and/or PEEP, the VT that must be set on the ventilator to deliver the appropriate volume to the patient's lungs remains unknown.

Additionally, even before atelectasis develops, the clinician may attempt to compensate for the discrepancy in the VT measured by the ventilator by increasing the set VT or the set PI_max. However, if the clinician overcompensates, the subsequently increased delivered volume can potentially cause volutrauma and iatrogenic lung injury (2, 5, 7). Ventilation with an excessively high VT can result in disruption of the pulmonary architecture (4, 9). Rosen and coworkers demonstrated a reduction in ventilator-induced lung injury when respiratory mechanics measurements as determined at the endotracheal tube were utilized in the care of neonates (10).

With the use of a pneumotachometer and respiratory mechanics monitoring, a more reliable measurement of the delivered VT can be obtained. By utilizing a potentially more accurate determination of the delivered volume, the patient-care team may be able to minimize barotrauma and volutrauma (11, 12). Additionally, optimizing the VT actually delivered may decrease intrathoracic pressures and limit consequent adverse cardiovascular and neurologic effects (10).

Potential Limitations of the Study

A potential but unavoidable limitation of this study is that the two determinations of VT were made with different equipment. This limitation is inherent to the design of the study. All of the equipment used in this study is approved by the U.S. Food and Drug Administration and has been validated for accuracy. All equipment was calibrated according to the manufacturers' recommendations before and after data collection. The validation of VT with the test lung ensured that each of the monitoring devices used in the study was comparably accurate in measuring the volume of gas that passed through the device. This test lung validation study confirmed that when the compliance of the circuit and the potential variations in the circuit setup were removed, the values of VT displayed by the ventilator and the pneumotachometer were essentially identical.

A potential limitation in the use of a pneumotachometer is condensation, which may accumulate in the pressure ports of the device. Such condensation could potentially invalidate the instrument's readings. This potential confounding variable can be essentially eliminated by positioning the pneumotachometer in such a way that the pressure ports are directed upward. During our study, condensation did not accumulate in the pneumotachometer.

Summary

On the basis of the data in this study, it appears that in the mechanical ventilation of neonates, infants, and small children, the percentage of VT "lost" as a result of the compliance of the ventilator circuit and variations in the circuit setup may be significant. For infant ventilator circuits, there is a clinically and statistically significant difference in the ventilator-displayed VT, the calculated effective VT, and the VT measured at the endotracheal tube by a pneumotachometer-respiratory mechanics monitor. Without accurate measurements of delivered volume, infants may be subjected to unnecessary barotrauma and volutrauma, thus possibly increasing morbidity. Our results therefore suggest that in the mechanical ventilation of infants, delivered VT should be determined by a pneumotachometer placed at the airway.