INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In adults breathing normally, under the influence of gravity both blood flow and ventilation are distributed preferentially to the dependent zones of the lungs (1). Body position may affect gas exchange by altering the matching of ventilation to perfusion within the lungs (2). However, arterial blood gas levels remain relatively constant in normal subjects during changes in body position (3). In contrast, body position can have significant effects on gas exchange in patients with pulmonary disease. It is well known that in adult patients with unilateral lung disease, oxygenation improves when the good lung is dependent and the diseased lung is uppermost (3, 4). Improved right-to-left shunt or improved ventilation-perfusion matching was reported when the unaffected lung is dependent (5).

However, previous studies have shown that improved oxygenation with the good lung dependent is not always the case for adult (3, 6) and infant (10) patients with unilateral lung disease. To our knowledge, the mechanism of this different postural effect on oxygenation in unilateral lung disease is not yet understood.

In conditions in which closing volume is increased because of old age or smoking, the dependent airways would be closed during tidal breathing (11), so that reduced ventilation-perfusion ratio and hypoxia would result and ventilation would be preferentially distributed to the uppermost lung.

We reasoned that oxygenation in the lateral decubitus position in adults with unilateral lung disease might be affected by an increase in closing volume through a consequent alteration in the distribution of ventilation to the dependent lung.

We therefore undertook an observational study of the effects of the lateral decubitus position on arterial oxygenation in a group of patients with unilateral lung disease. We also tested the hypothesis that oxygenation in lateral recumbency in adults with unilateral lung disease would be influenced by the distribution of ventilation to the dependent lung and that the distribution of ventilation to the dependent lung would be influenced by the closing volume.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

The study group consisted of 44 randomly selected hospitalized patients who had unilateral parenchymal lung diseases such as pneumonia or pulmonary tuberculosis, in which at least one pulmonary lobe was involved. There were 20 men and 24 women 16 to 82 yr of age (mean ± SD: 49.9 ± 18.7 yr). Twenty-three cases of pneumonia and 21 cases of pulmonary tuberculosis were included. In these patients, arterial blood gases were analyzed and the alveolar-arterial PO₂ difference (AaPa_O2) was calculated.

In 24 randomly chosen patients, who volunteered to be the subjects of this study, ventilation lung scans and closing volumes were measured, in addition to arterial blood gas analyses. From the results of these tests, only 16 patients (10 men and 6 women 49.2 ± 18.2 yr of age) were selected to be included in the study, as eight of the patients failed to meet the acceptability criteria. Ten cases of pneumonia and six cases of pulmonary tuberculosis were studied.

Study Protocol

Samples of arterial blood were drawn anaerobically by radial artery puncture in all patients while they were breathing room air spontaneously, 20 min after assumption of the supine, right lateral, and left lateral decubitus positions. The positions were adopted in random order. The blood samples were stored on ice and analyzed within 10 min with a blood-gas analyzer (Corning 288; Ciba-Corning Diagnostics Corp., Norwood, MA).

The patients were divided into two groups based on whether individual Pa_O2 with the normal lung in the dependent position was higher than that with the diseased lung in the dependent position (control group) or opposite condition (reversed group). PA_O2 (alveolar oxygen tension) was calculated by the following equation: PA_O2 = (barometric pressure-47) × FI_O2  Pa_CO2/R. In this study, R was assumed to be 0.8. The AaPO₂ was calculated by subtraction of Pa_O2 from PA_O2. The values of Pa_O2 and AaPO₂ with the normal lung in the dependent position were compared with those with the diseased lung in the dependent position in both groups.

In the 16 patients who underwent complete studies, the difference in Pa_O2 between the normal and the diseased lung in the dependent position (Pa_O2 with the normal lung in the dependent position minus Pa_O2 with the diseased lung in the dependent position) was computed as an index of postural effect on oxygenation in both lateral decubitus positions in unilateral lung disease.

The effect of body position on the distribution of ventilation to each lung was measured by radionuclide ¹³³Xe lung scans. Ventilation scanning was performed with a closed xenon delivery system (Pulmonex Xenon System; Atomic Products Co., Shirley, NY) and a gamma camera (Diacam; Siemens, Hoffman Estates, IL) connected to a dedicated on-line computer. The subject was instructed to take a deep breath from functional residual capacity to total lung capacity just after a ¹³³Xe gas bolus (1.5 mCi) was injected into the delivery system in the supine position and then to breathhold for 20 s for the counting of radioactivity. At the same time, a posterior image was obtained. After sufficient time was allowed to wash out of the gas, reventilation studies were performed in both lateral decubitus positions. The radiation dose was equivalent to that of a standard chest radiograph. To examine postural changes in fractional ventilation (the proportion of ventilation distributed to each lung), ¹³³Xe activity from the normal lung was expressed as a percentage of the activity from both lungs. The difference in the fractional ventilation going to the normal lung between the dependent and supine position (fractional ventilation to the normal lung in the dependent position minus fractional ventilation to the normal lung in the supine position) was computed as an index of postural effect on distribution of ventilation to the normal lung in the dependent position compared with the supine position.

Closing volume (CV) by the single breath nitrogen test method (12) was measured in the seated position using SensorMedics System 2100 (SensorMedics Corporation, Yorba Linda, CA). Criteria for acceptability of closing volume curves had to be met (12), and three acceptable tracings were obtained. The mean of the three values was taken as the final value. Values are expressed as percent predicted using the formula of Lee and colleagues (13). All test results were analyzed after complete data were obtained.

The study protocol was approved by the Institutional Review Board, and informed consent was obtained from all patients.

Statistical Analysis

The all data are expressed as the mean ± SD. Pa_O2 and AaPO₂ in both lateral decubitus positions were compared using a paired t test. A nonparametric paired t test (Wilcoxon's signed rank test) was used for the data that did not show a normal distribution.

Linear regression analysis was used to determine how the distribution of ventilation influences oxygenation and also how the closing volume influences the distribution of ventilation. All statistical analyses were performed using custom software (SPSS for Windows, V7.5, 1997; SPSS Inc., Chicago, IL). A value of p < 0.05 was considered to be significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Of the 44 randomly selected patients with unilateral lung diseases, 26 (59%) were classified as the control group and 18 (41%) as the reversed group. For the control group, Pa_O2 with the normal lung in the dependent position was significantly higher than that with the diseased lung in the dependent position (86.2 ± 10.2 versus 78.8 ± 9.4 mm Hg, p < 0.001) (Figure 1). In the reversed group, the corresponding value for Pa_O2 with the normal lung in the dependent position was significantly lower ( 76.0 ± 8.3 versus 83.4 ± 8.4 mm Hg, p < 0.001). In the control group, the AaPO₂ with the normal lung in the dependent position was also significantly lower than that with the diseased lung in the dependent position (18.6 ± 11.0 versus 25.7 ± 11.0 mm Hg, p < 0.001); it was significantly higher (27.5 ± 11.4 versus 21.2 ± 11.9 mm Hg, p < 0.002) in the reversed group.

View larger version (24K): [in this window] [in a new window]   Figure 1.   PaO2 with the normal lung in the dependent position (NLD) compared with that with the diseased lung in the dependent position (DLD) for the two groups. The difference between the two dependent positions was statistically significant (p < 0.001 for each) in both groups. Horizontal bars represent mean values.

The data on the 16 patients subjected to the complete study are shown in Table 1. Coincidentally, higher Pa_O2 values were found in eight patients with the normal lung in the dependent position, as well as in eight patients with the diseased lung in the dependent position.

Relationship between Oxygenation and the Distribution of Ventilation

The difference in Pa_O2 between the normal and the diseased lung in the dependent position was related significantly to the difference in the fractional ventilation going to the normal lung between the dependent and the supine position (r = 0.642, p = 0.007) (Figure 2).

View larger version (19K): [in this window] [in a new window]   Figure 2.   Relationship between the difference in PaO2 between the normal and the diseased lung in the dependent position and the difference in the fractional ventilation to the normal lung between the dependent and supine position in 16 patients with unilateral lung disease. A significant positive relationship was observed (r = 0.642, p = 0.007).

Relationship between the Distribution of Ventilation and the Closing Volume

The difference in the fractional ventilation going to the normal lung between the dependent and the supine position was related significantly to the CV/VC (% predicted) (r = 0.597, p = 0.015) (Figure 3).

View larger version (16K): [in this window] [in a new window]   Figure 3.   Relationship between the difference in the fractional ventilation to the normal lung between the dependent and supine position and the closing volume (CV/VC, % predicted) in 16 patients with unilateral lung disease. A significant negative relationship was observed (r = 0.597, p = 0.015).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study has shown that for patients lying in the lateral decubitus position, a significantly increased gas exchange was present with the normal lung down in about 60% of randomly selected patients with unilateral parenchymal lung disease, and the opposite was true in about 40% of patients.

In our 16 patients in whom a full study was performed, we found that the differences in Pa_O2 between the normal and the diseased lung in the dependent position were related significantly to the differences in the fractional ventilation going to the normal lung between the dependent and the supine position. Furthermore, we found that the latter was related significantly to the closing volume. Thus, this result indicates that the higher the fractional ventilation going to the normal lung in the dependent position, the better oxygenation with the normal lung in the dependent position. Conversely, the lower the fractional ventilation to the normal lung in the dependent position, the worse the oxygenation with the normal lung in the dependent position. It was also shown that the lower the closing volume, the higher the fractional ventilation to the normal lung in the dependent position and, conversely, the higher the closing volume, the lower the fractional ventilation to the normal lung in the dependent position. Therefore, a normal pattern of postural effect on the distribution of ventilation to the dependent lung was reversed in patients with increased closing volume. Thus, oxygenation in lateral recumbency in patients with unilateral lung disease might be influenced by the fractional ventilation to the dependent lung and the latter, in turn, might be influenced by the closing volume.

In the normal upright lung, because of the hydrostatic pressure differences within the lung, blood flow per unit volume increases from top to bottom reaching the highest values at the bases of the lungs (1). This pattern is affected by changes of posture. In the lateral decubitus position, the dependent regions are best perfused.

The distribution of ventilation is also gravity-dependent, so that normally the ventilation to the base exceeds that to the apex when breathing at functional residual capacity (FRC). Any change in posture affects the distribution of ventilation, as it does blood flow. The dependent lung is better ventilated than the uppermost lung in the lateral decubitus position (2).

In normal adults, there are no significant differences in gas exchange between the left and right lateral decubitus positions (3). However, it has been reported that in unilateral lung disease, a significantly higher Pa_O2 occurs when the patient is lying with the normal lung down (3, 4). The increased Pa_O2 values are probably the result of an improved right-to-left shunt or an improved matching of ventilation to perfusion when the unaffected lung is dependent, compared with the situation when the affected lung is dependent (5).

In contrast to the situation in adults, oxygenation in infants can be improved if the child with unilateral lung disease is positioned with the normal lung uppermost (10). In infants and very young children, it has been shown that ventilation is distributed preferentially to the uppermost lung (14). An infant's chest wall is floppier than an adult's, and the resting pleural pressure in an infant is closer to atmospheric pressure. Thus, in the lateral decubitus posture, closure of peripheral airways is more likely to occur in the dependent lung regions and ventilation to be distributed toward the uppermost lung (15). A further explanation is based on a much reduced transabdominal hydrostatic pressure difference in the infant, thereby reducing the functional advantage of the dependent diaphragm.

Even in adults a reversal of gravitational distribution of ventilation at FRC has been observed when closing volume is increased. Then, the dependent airways will tend to close during tidal breathing, as they do in elderly subjects (16). The closing capacity (CC) is the sum of the closing volume and the residual volume. When the closing volume increases until it encroaches a normal breathing range (i.e., when CC > FRC), some of the pulmonary blood flow will be distributed to alveoli with closed airways, usually in the dependent parts of the lungs. This will constitute a shunt or reduced ventilation-perfusion ratio and will decrease the arterial oxygenation. Any physiologic decrease in FRC will enhance this difference. The most common physiological influence is postural. The FRC in the supine posture is considerably less than that in the upright posture, so that even in normal subjects 44 yr of age or older, CC may equal or exceed FRC when supine. On the other hand, in the erect posture, CC does not exceed FRC until about 65 yr of age (11). Thus, the encroachment of closing volume on FRC is a function of aging. In this study, arterial blood gas analysis was performed in the recumbent position, so that any difference between CC and FRC might have been enhanced. In turn, this would have made the problem more apparent in patients who already had an increased closing volume.

Closing volume also increases in cigarette smokers (17) and in patients with congestive heart failure (18). It could be assumed that unilateral pneumonia or pulmonary tuberculosis itself might produce an increased closing volume. However, to our knowledge, there have been no published reports on this subject. Moreover, despite the fact that these two groups of patients had different responses to posture, they had in common unilateral lung disease. Therefore, unilateral lung disease alone could not be the direct cause of the increased closing volume and the different response to posture.

In the elderly subjects, the upper parts of the lungs are better perfused than in the young subjects (16). Moreover, in a parenchymal lung disease such as pneumonia or tuberculosis, alveolar filling may cause gas exchange abnormalities, and hypoxic pulmonary vasoconstriction may direct pulmonary blood flow away from poorly ventilated alveoli.

Therefore, we consider that the postural effect on gas exchange in patients with increased closing volume and unilateral lung disease might be different from the situation in normal adults, but similar to that in infants.

In adult patients with normal closing volume and unilateral lung disease, gravitationally increased perfusion to the dependent lung would be matched by enhanced ventilation with the normal lung in the dependent position; resulting oxygenation would be good. When the diseased lung is dependent, regional ventilation to this dependent abnormal lung could not be increased appropriately (4) and oxygenation would be poor. Thus, in this study we computed the difference in Pa_O2 between the normal and the diseased lung in the dependent position as an index of the postural effect on oxygenation, for both lateral decubitus positions in patients with unilateral lung disease. We also computed the difference in the fractional ventilation going to the normal lung between the dependent and the supine position as an index of the postural effect on the distribution of ventilation to the normal lung in the dependent position compared with the supine position.

In adult patients with increased closing volume and unilateral lung disease, we reasoned that, with the normal lung in the dependent position,ventilation would be distributed preferentially to the uppermost, poorly perfused diseased lung because of increased closing volume; resulting oxygenation would be poor. Furthermore, with the diseased lung in the dependent position, ventilation would be distributed preferentially to the uppermost, well-perfused normal lung, which is attributable to a substantially reduced dependent blood flow by the hypoxic vasoconstriction of the diseased lung; resulting oxygenation would be good. Thus the normal pattern of the postural effect on oxygenation would be reversed in patients with increased closing volume. Therefore, we hypothesized that the improvement in Pa_O2 with the normal lung in the dependent position in patients with unilateral lung disease may depend on the magnitude of the closing volume.

The question remains as to how much of the lung must be diseased in order for the perfusion relative to ventilation to be sufficiently reduced to produce this effect. The degree of flow diversion away from the hypoxic site with a localized pathologic process depends on the size of the hypoxic segment; the smaller the hypoxic region, the greater the percentage diversion of blood flow away from that segment (19, 20). As the hypoxic test segment becomes larger, its activity becomes manifest as more substantial increases in transmural pulmonary arterial pressure and perfusion pressure; the larger hypoxic segment becomes less effective in diverting blood flow (20). We therefore consider that a minimum amount of lung involvement, in the sense of a threshold, may not be an essential requirement for blood flow to be diverted from a diseased lung. In humans, hypoxia in a single lobe of the lung results in a rapid decline in perfusion of the lobe such that regional blood flow is half that during normoxia (21). In this study, we purposely selected subjects with unilateral lung disease, in which at least one pulmonary lobe was involved.

In order to estimate roughly what proportion of general patient population with unilateral lung disease may obtain improved oxygenation by placing the diseased lung in the dependent position, we performed arterial blood gas analysis in 44 randomly selected patients with unilateral pneumonia or pulmonary tuberculosis. We found that in 44 patients with unilateral lung disease, about 60% showed improved oxygenation with the normal lung in the dependent position, but also that about 40% showed improved oxygenation with the diseased lung in the dependent position. Furthermore, this improvement was statistically significant, suggesting that the magnitude of the postural effect is not trivial and that the effect could be clinically meaningful. In addition, between the two positions, both groups showed a significantly different AaPO₂. Because AaPO₂ is a measure of the magnitude of ventilation-perfusion ratio, this finding implies that a significantly different ventilation-perfusion ratio might exist between the two positions in both groups.

This finding of a different postural effect is not consistent with a previous report, which suggested that lying on the normal lung resulted in maximal arterial oxygen pressure in every patient (4). However, other studies have demonstrated that a favorable gas exchange was not always present when the patients lay in the lateral decubitus position with the normal lung down (3, 6).

Despite these inconsistent results in previous reports, no systemic study has been carried out as to why different postural effects had been observed. Therefore, in order to understand this phenomenon better, we have studied the influence of distribution of ventilation on oxygenation in lateral recumbency and the influence of closing volume on the distribution of ventilation in a small subpopulation of patients with unilateral lung disease. The results of this study have shown that oxygenation in lateral recumbency in unilateral lung disease may be influenced by the closing volume through a consequent alteration in the distribution of ventilation to the dependent lung. This finding seems to be consistent with previous reports that an improved gas exchange was observed when the sick lung was placed in the dependent position in infant patients with unilateral lung disease (10) whose dependent airways tend to close during tidal breathing, and also in adult patients with unilateral pulmonary embolism who were receiving mechanical ventilation (8). The reasons for the similarity between the latter case and our findings could be that the nondependent lung receives higher regional ventilation in mechanically ventilated patients, and that blood flow is directed to the healthy lung in patients with pulmonary embolism (8). Further study in a larger number of patients with unilateral lung disease is needed to confirm our present findings.

In this age of assisted ventilation and respiratory aids, it is certainly encouraging to establish the possibility of improving arterial oxygenation simply by changing body position rather than by resorting to expensive breathing devices (22). It is well known that changes in position are necessary for the prevention of decubitus ulcer in immobilized patients and it is generally recommended that bedridden patients should be turned from side to side every 2 h. However, consideration of the different response in oxygenation with different body positions may be required for optimal oxygenation during respiratory care in patients with unilateral lung disease.The difference in Pa_O2 between the normal and the diseased lung in the dependent position was closely correlated with the difference in Pa_O2 between the normal lung in the dependent and the supine position (Figure 4). If Pa_O2 with the normal lung in the dependent position is higher than that with the diseased lung in the dependent position, it tends to be higher than the Pa_O2 in the supine position. On the other hand, if Pa_O2 with the normal lung in the dependent position is lower than that with the diseased lung in the dependent position, it tends to be lower than the Pa_O2 in the supine position. Therefore, for positioning patients with normal closing volume and unilateral lung disease, placing the normal lung in the dependent position might be preferable to placing the diseased lung in the dependent position or the supine position. Conversely, for positioning patients with increased closing volume, placing the diseased lung in the dependent position or supine position might be preferable to the normal lung in the dependent position. Therefore, we recommend that position of the patient for maximal oxygenation should be chosen on an individual basis according to documentation of the patient's best oxygenation by arterial blood gas analysis or pulse oximetry, and not as a matter of routine protocol.

View larger version (22K): [in this window] [in a new window]   Figure 4.   Correlation between the difference in PaO2 between the normal and diseased lung in the dependent position and the difference in PaO2 between the normal lung in the dependent and supine position. A significant positive correlation was observed (r = 0.812, p < 0.001). The line of identity is indicated.

In conclusion, this study has shown that for patients with unilateral parenchymal lung disease, lying with the normal lung in the dependent position will not always result in an optimal gas exchange and a maximal arterial oxygen pressure. Not only posture, but also the closing volume, plays a role in determining the degree of oxygenation in lateral recumbency. As closing volume increases with the diseased lung dependent, less of the total ventilation is distributed to the dependent lung and oxygenation progressively improves. This observation suggests that blood flow, as well as ventilation, is shifted away from the dependent abnormal lung in patients with unilateral lung disease and increased closing volume.