INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Currently, the FRC is the only lung volume that can be accurately, and repeatedly measured in ventilated patients of all ages. Other lung volumes such as total lung capacity (TLC), vital capacity (VC), and residual volume (RV) can also be measured, but these techniques are rarely employed. Nevertheless, there is a strong rationale for the assessment of the relative distribution of lung volumes because such measurements could add significantly to our understanding of pulmonary hyperinflation and its consequences in ventilated patients. Infants in particular and children are prone to complete blockage of small airways and atelectasis from bronchial edema and peribronchial inflammation, as commonly seen in bronchiolitis. As a result of this combination of obstructive and restrictive changes, FRC may remain (in absolute numbers) within normal limits, and pulmonary hyperinflation can be verified only by an increase in the FRC/TLC ratio (1, 2).

Several studies have questioned the accuracy of plethysmographic measurements of thoracic gas volume (Vtg or FRCpleth), particularly in patients with obstructive lung disease because of discrepancies between the FRCpleth determined after occlusion at end-exhalation and the corrected FRCpleth measured from occlusions at higher lung volumes (3). These discrepancies were not predictable either in direction or in magnitude. The hypothesis was advanced that airway closure during tidal breathing, uneven distribution of pleural pressure, and compliance of extrathoracic airways were responsible for overestimation or underestimation of FRCpleth determined at end-exhalation (6).

In contrast to the Vtg determined by body plethysmography, FRC measured by gas dilution techniques represents only volume that is in connection with the mouth and can be diluted or washed out. Thus, these techniques may tend to underestimate lung volume in patients with gas trapping. It is reasonable to assume that discrepancies between measurements made at low and high lung volumes will be more pronounced when lung volume is measured by gas dilution since closed airways will reopen and recruit lung areas when positive pressure and volume are applied, especially in patients with obstructive lung disease.

We have previously reported the accuracy of the N₂ washout technique for measurements of lung volumes above FRC in intubated rhesus monkeys with normal lungs (7). Nevertheless, it remains unclear if gas dilution techniques can be used to measure higher lung volumes (e.g., TLC) accurately in infants with lung disease.

The aims of the present study were: (1) to determine whether the N₂ washout technique can be used to measure TLC accurately and reproducibly in intubated infants with and without lung disease, and (2) to determine whether TLC measured by N₂ washout (TLCN₂) differs from the TLC (TLCFRC+IC) calculated from the sum of FRC measured by N₂ washout (FRCN₂) and the inspiratory capacity (IC) measured by pneumotachograph from a passive exhalation from +40 cm H₂O inspiratory pressure to FRC.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Fifty infants and children (mean age [SEM], 19.9 [4.6] mo; mean weight, 9.1 [0.9] kg) requiring mechanically assisted ventilation for various clinical reasons were recruited for the study from the Pediatric Intensive Care Unit at Children's Hospital Los Angeles. Patients were eligible for enrollment if they were intubated with a cuffed endotracheal tube. The cuff was inflated to prevent any air leak detectable by auscultation when the lungs were inflated to +40 cm H₂O pressure. The infants were placed supine, and stomach contents were drained via a pediatric nasogastric tube. Sedation was provided by diazepam (0.1 mg/kg) given intravenously and/or fentanyl (0.1 mg/kg). All infants received vecuronium (0.1 mg/kg) for neuromuscular blockade during the testing procedure. Each infant was suctioned for secretions at the beginning of testing and subsequently as necessary. This study was approved by the Committee on Clinical Investigations of our institution, and informed consent was obtained on behalf of the patients in this study.

The single-breath occlusion technique was used to assess total respiratory system compliance (Crs) and resistance (Rrs) (8). Testing was performed by interrupting a breath close to end-inspiration during mechanical ventilation (SensorMedics 2600; SensorMedics, Yorba Linda, CA). The resulting passive exhalation was analyzed for flow, relaxation airway pressure, and volume to calculate Crs and Rrs. At least eight curves free of artifact were analyzed for each patient, the values averaged, and the data recorded.

Lung volumes were measured by a modified automated N₂ washout technique using a computer-based data acquisition system (SensorMedics 2600) incorporating a two-point calibration method developed and validated by Sivan and colleagues (9). The technique has recently been summarized and reviewed in detail (10, 11). In brief, the system consists of a second intermittent flow ventilator (Siemens Servo 900C; Siemens, Solna, Sweden) with settings matched to the patient's ventilator of the same type, in volume-controlled ventilation mode but delivering 100% O₂. Both ventilators are connected to the proximal end of the endotracheal tube through a slider valve that activates and switches the patient to the washout ventilator. The gas leaving the latter ventilator via the exhalation port passes through a mixing chamber and the N₂ concentration is measured continuously by an analyzer within the testing cart. The system was calibrated by using a syringe filled with known amounts of room air attached to the washout ventilator using the same minute ventilation (E) for calibration and patient washout. A computer-based data acquisition system integrates the N₂% signal electronically, provides a real-time display of the N₂ washout curve, and automatically calculates the lung volume from which the washout started at the end of the procedure according to the following equation:

(1)

Because E during calibration and testing are the same when using an intermittent flow washout ventilator in volume-controlled mode, there is no need to measure E by argon dilution or any other technique because the ratio E_test/E_cal is equal to 1.

The N₂ washout to measure FRCN₂ was initiated after allowing the patient to exhale to barometric pressure at zero PEEP. The N₂ washout to measure TLCN₂ was initiated from +40 cm H₂O inspiratory pressure. For the latter measurement, the inflation pressure was held static for at least 3 s using the inspiratory hold button of the ventilator to standardize volume history. At least two measurements were obtained for each lung volume and the values averaged. The measurements were accepted when they were within 10% of each other (11). None of our patients required a FI_O2 > 0.7 at the time of the testing.

Quadruplicate passive expiratory maneuvers were performed from +40 cm H₂O inspiratory pressure down to barometric pressure at zero PEEP and the volume measured by means of a pneumotachograph (Model 4700; Hans Rudolph, Kansas City, MO). The inflation pressure was again held static for at least 3 s. This volume was called inspiratory capacity (IC). These data were averaged and the mean added to FRCN₂ to calculate TLCFRC+IC. Similarly, FRCTLCIC was calculated by subtracting IC from TLCN₂. Patients were allowed to exhale to barometric pressure for the purpose of these measurements, even if they were receiving PEEP. In between the measurements and during the nitrogen washout, the patients were ventilated with the PEEP settings chosen by the pediatric intensivist in charge. There were no side effects or complications associated with disconnecting the PEEP for the exhalation or with inflating the patients to 40 cm H₂O for 3 s. The use of neuromuscular blockade during the testing procedure allowed full control over the breathing cycle.

The study patients were classified on the basis of pulmonary function testing using our previously published reference values established in intubated infants into: (1) infants without lung disease, (2) infants with purely restrictive lung disease, and (3), infants with obstructive lung disease (2, 12). Infants without lung disease were required to have clear chest radiographs, a normal ratio of arterial-to- alveolar oxygen tension (> 0.8), and normal respiratory mechanics. Reference values for normal respiratory mechanics are  0.8 ml/cm H₂O/kg for total respiratory system compliance (Crs) and in the range of 0.04 to 0.08 cm H₂O/ml/s (as much as 0.1 cm H₂O/ml/s for endotracheal tubes < 3.5 mm I.D.) for total respiratory system resistance (Rrs) (2, 12). The classification using pulmonary function tests was also consistent with the clinical diagnosis and physical examination of the patients.

Statistical Analysis

Values are reported as means ± SEM for the entire sample and for each clinical subgroup when appropriate. A one-way analysis of variance, the Bonferroni adjustment, and the nonparamtetric Kruskal-Wallis test were used for overall group comparisons, with a significance level of p < 0.05. For paired data analysis, Wilcoxon's signed-rank test and the paired two-tailed t tests were used to determine a significant difference between the means of TLCN₂ and TLCFRC+IC in each subgroup. Agreement between TLCN₂ and TLCFRC+IC was determined using the analysis of Bland and Altman (13) for the entire sample and for each subgroup. The degree of agreement was expressed by two components: (1) the relative bias (d), as estimated by the mean difference, and (2) the random variation (s), as estimated by the standard deviation of differences. The "limits of agreement," or 95% range, were expressed by d ± 2s.

In addition, the individual mean TLCFRC+IC were considered significantly different when lying outside 10% of the individual mean of TLCN₂. This cutoff point of 10% was arbitrarily chosen on the basis of being approximately double our previously published mean coefficients of variation for repeated FRC and TLC measurements in intubated infants, which are in the range of 4 to 6% (1, 9).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The clinical diagnoses of the study patients in each subgroup are summarized in Table 1 and their lung function results are shown in Table 2. There was no statistically significant difference in age between groups because of the scatter of the data. However, the restrictive patients were significantly heavier than those in the other two groups (Kruskal-Wallis test statistics, 6.6; p = 0.04). The mean ± SEM (range) of the endotracheal tube internal diameters used in the study patients was 3.8 ± 0.2 (3 to 4.5) mm for the infants without lung disease, 3.5 ± 0.1 (3 to 4.5) mm for the obstructive patients, and 4.4 ± 0.3 (3.5 to 7) for the restrictive patients. The tube sizes used in the restrictive group were significantly larger than those used in the other two groups (Kruskal-Wallis test statistics, 11.4; p = 0.003). There were significant differences in lung function between the three groups. The group with no lung disease had significantly higher Crs and TLC than did the restrictive and the obstructive groups. The restrictive group was distinguished by smaller FRC and TLC than both of the other groups, and the obstructive patients demonstrated significantly higher Rrs and FRC/TLC ratios than did both other groups. The mean (SEM) coefficients of variation for the repeated FRCN₂, IC, and TLCN₂ measurements in this study were 3.4 (0.6), 3.6 (0.4), and 1.5% (0.3), respectively. However, we did reject a few N₂ washout measurements that were uncertain to have commenced at the correct lung volume (either passive end- exhalation or end-inhalation at +40 cm H₂O) because they were more than 10% off from previous and succeeding measurements.

The results of TLC calculated from the N₂ washouts at end-exhalation plotted against those from the washouts made at +40 cm H₂O inspiratory pressure are shown in Figure 1. For grouped mean data, there were significant differences between TLCN₂ and TLCFRC+IC, and between FRCN₂ and FRCTLCIC, respectively, only in the group with obstructive lung disease. This is also reflected in the analysis of Bland and Altman displayed in Table 3, where the obstructive group had a strikingly higher mean difference and limits of agreement between the two techniques than either of the other groups. The differences between TLCN₂ and TLCFRC+IC were unrelated to the size of the measurements, as shown in Figure 2. The analysis of Bland and Altman of the agreement between FRCN₂ and FRCTLCIC yields identical results since the absolute differences are the same as between TLCN₂ and TLCFRC+IC.

View larger version (15K): [in this window] [in a new window]   Figure 1.   Relationship between TLCFRC+IC (FRC measured by N2 washout [FRCN2] + inspiratory capacity [IC] measured by pneumotachograph from +40 cm H2O inspiratory pressure) and TLCN2 (measured by N2 washout from +40 cm H2O inspiratory pressure). Each point represents the mean of two to three measurements. The line of identity is also shown.

View larger version (17K): [in this window] [in a new window]   Figure 2.   Scatter plot, according to the method of Bland and Altman (14), of the differences between TLCN2 (TLC measured by N2 washout) and TLCFRC+IC (calculated by adding FRC, as determined by N2 washout, and inspiratory capacity, as measured by pneumotachograph, from a passive exhalation from +40 cm H2O inspiratory pressure) against their respective means. The figure demonstrates considerable lack of agreement for infant patients with obstructive disease.

The percent difference between TLCN₂ and TLCFRC+IC with respect to the different groups is shown in Figure 3. The individual percent differences ranged from 7.1% to 6.4% in the infants with no lung disease, from 20 to +22.4% in the obstructive patients, and from 9.7 to 12.6% in the restrictive infants. Ten infants (nine obstructive and one restrictive) had differences between TLCN₂ and TLCFRC+IC that were greater than 10%. Applying the same mathematics to FRC yields results that are slightly less favorable with respect to the percent difference since the absolute differences between TLCN₂ and TLCFRC+IC are the same as between FRCN₂ and FRCTLCIC. These data are shown in Figure 4.

View larger version (11K): [in this window] [in a new window]   Figure 3.   Percent difference between TLCN2 (measured by N2 washout from +40 cm H2O inspiratory pressure) and TLCFRC+IC (calculated by adding FRC and inspiratory capacity). Differences > 10% were observed in a number of patients with obstructive disease.

View larger version (12K): [in this window] [in a new window]   Figure 4.   Percent difference between FRCN2 (measured by N2 washout) and FRCTLCIC (calculated by subtracting inspiratory capacity from TLCN2 [measured by N2 washout]). Differences > 10% were observed in a number of patients with obstructive and restrictive disease.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The reports that suggest conventional methods for estimating end-expiratory lung volume (FRC or Vtg) in wheezy infants might be inaccurate and provide potentially misleading information raise questions about the usefulness of these variables for the clinical management of such patients. Several groups reported discrepancies between plethysmographic measurements of Vtg determined after occlusion at end exhalation and those measured from occlusion at higher lung volumes (3). Several factors such as gas trapping caused by small airway closure at low lung volumes and recruitment of obstructed units at higher lung volumes, uneven distribution of pleural pressure, and compliance of central airways were implicated as responsible for those differences that were not predictable in magnitude and direction.

We have previously demonstrated in intubated rhesus monkeys that lung volumes above FRC can be reliably measured by the N₂ washout technique (7, 9). Lung volumes directly measured did not differ from preset lung volumes achieved by artificial inflation of the monkey's lungs above FRC with known volumes of air. However, these animals had normal lungs and airways and the findings are, therefore, not directly applicable to infants with lung disease. In the present study, we found that in intubated infants and children without obstructive lung disease, the measurements of TLC or FRC by gas dilution are little different whether measured from low or high lung volumes. However, there was less agreement in infants with obstructive lung disease. These differences were most likely caused by recruitment of atelectatic lung units behind obstructed or collapsed airways after a deep inspiration to TLC and changes in the distribution of secretions and airway blockage in these patients over the time these measurements were obtained. Nevertheless, except for a few patients, these differences were within a range of 20%, which, according to Hedenstierna (14), is accurate enough for most clinical purposes.

With all techniques there is controversy regarding how many measurements should be made in each patient, and the way in which results should be expressed. Whereas it is relatively simple and quick to obtain three to five repeat measurements of Vtg in the body plethysmograph, this is less feasible with gas dilution techniques not only because of the duration of rebreathing or washout required but because of the necessary interval between tests. Results based on a single recording are unlikely to be reliable. It was recently suggested that the absolute minimum is the mean of two measurements within 10% of each other, although greater confidence would be obtained if three such measurements were available (11). The time-consuming nature of the N₂ washout maneuver prevented us from obtaining more than two or three repeated measurements at each lung volume in these critically ill patients.

The determination of lung volumes such as TLC and FRC, together with the measurements of respiratory mechanics and maximal expiratory maneuvers, allow a thorough assessment of lung function in intubated infants analogous to adult pulmonary function laboratory standards. We have previously reported such complete pulmonary function data in intubated infants with severe RSV infection and demonstrated that the FRC/TLC ratio is a useful value for assessment of pulmonary hyperinflation and its response to therapy (1). After albuterol inhalation, the FRC/TLC ratio decreased in all those patients who demonstrated a reduction in airway resistance and an improvement in maximal expiratory flows. In this study, the lung volumes and their ratio clearly discriminated the patients with respect to the underlying pulmonary pathophysiology. In addition, pulmonary hyperinflation may increase the risk of barotrauma and interfere with ventilator weaning because, during assisted ventilation, it constitutes an extra burden on the inspiratory muscles that must be counterbalanced before inspiratory flow can commence (15). Hence, the combined measurement of FRC and TLC could be of assistance in the ventilatory management of restrictive and obstructive patients.

Reference values for TLC in intubated, healthy infants and children are very limited. Strictly speaking, the values for TLC reported here on infants without lung disease may not be regarded as true normal data because these ventilated infants did not fulfill the stringent criteria that were recently recommended to define health (18). Nevertheless, we believe our study group is as close to normal as one can possibly find in the pediatric intensive care environment. Besides, the measurements of FRC and respiratory mechanics are in agreement with published normal data in this age group (19). In this respect, we believe that our values may serve as a useful reference for clinical practice. Thorsteinsson and colleagues (20) reported the only other TLC data for intubated, healthy infants under neuromuscular blockade and also used a gas dilution technique, but they defined TLC at +30 cm H₂O inspiratory pressure. In their study, mean ± SD TLC was 52 ± 13 ml/ kg in infants and 87 ± 11 ml/kg in older children up to 16 yr of age. Measuring TLC at +30 cm H₂O, however, is impractical for critically ill patients with lung disease as many of these patients are ventilated with positive inspiratory pressures greater than +30 cm H₂O. This is one reason why +40 cm H₂O is a better choice to define TLC in intubated patients. In addition, we have previously demonstrated that the flat portion of the pressure-volume curve has not been reached with +30 cm H₂O inspiratory pressure under conditions of neuromuscular blockade in intubated infants (21). This can be explained by the use of neuromuscular blockade, which may result in an upward shift of the flat portion of the pressure-volume curve by increasing chest wall compliance compared with nonparalyzed infants. Therefore, we and others defined TLC at +40 cm H₂O inspiratory pressure in intubated infants and children (22, 23). This may explain why our reference values for TLC are higher than those reported by Thorsteinsson and colleagues (20) for infants of similar age and weight (66.4 compared with 52.0 ml/kg).

We conclude that TLC can be reliably measured by N₂ washout alone or in combination with pneumotachography in intubated infants, but with limited accuracy in those with obstructive disease. The measurement of the subdivisions of lung volume and the calculation of the FRC/TLC ratio will help to estimate the degree of pulmonary hyperinflation in intubated and ventilated infants. Although it seems rational that lung volume measurements should be helpful in the management of ventilated patients, the impact of this knowledge on morbidity and mortality has yet to be evaluated.