Rapid intravenous infusion of 20 ml/kg saline does not impair resting pulmonary gas exchange in the healthy human lung

Abstract

Rapid infusion of intravenous saline, a model of pulmonary interstitial edema, alters the distribution of pulmonary perfusion, raises pulmonary capillary blood volume, and increases bronchial wall thickness in humans. We hypothesized that infusion would disrupt pulmonary gas exchange by increasing ventilation/perfusion (V̇˙a/Q̇˙) inequality as opposed to a diffusive impairment in O₂ exchange. Seven males (26 ± 3 yr; FEV1: 110 ± 16% predicted.) performed spirometry and had V̇˙a/Q̇˙ mismatch measured using the multiple inert gas elimination technique, before and after 20 ml/kg iv of normal saline delivered in ∼30 min. Infusion increased thoracic fluid content from transthoracic impedance by 12% (P < 0.0001) and left FVC unchanged but reduced expiratory flows (FEF_25–75 falling from 5.1 ± 0.4 to 4.2 ± 0.4 l/s, P < 0.05). However, V̇˙a/Q̇˙ mismatch as measured by the log standard deviation of the ventilation (LogSDV̇˙) and perfusion (LogSDQ̇˙) distributions remained unchanged; LogSDV̇˙: 0.40 ± 0.03 pre, 0.38 ± 0.04 post, NS; LogSDQ̇˙: 0.38 ± 0.03 pre, 0.37 ± 0.03 post, NS. There was no significant change in arterial Po₂ (99 ± 2 pre, 99 ± 3 mmHg post, NS) but arterial Pco₂ was decreased (38.7 ± 0.6 pre, 36.8 ± 1.2 mmHg post, P < 0.05). Thus, infusion compressed small airways and caused a mild degree of hyperventilation. There was no evidence for a diffusive limitation to O₂ exchange, with the measured-predicted alveolar-arterial oxygen partial pressure difference being unaltered by infusion at Fi_O2 = 0.125 (4.3 ± 1.0 pre, 5.2 ± 1.0 post, NS). After infusion, the fraction of perfusion going to areas with V̇˙a/Q̇˙ < 1 was increased when a subject breathed a hyperoxic gas mixture [0.72 ± 0.06 (Fi_O2 = 0.21), 0.80 ± 0.06 (Fi_O2 = 0.30), P < 0.05] with similar effects on ventilation in the face of unchanged V̇˙a and Q̇˙. These results suggest active control of blood flow to regions of decreased ventilation during air breathing, thus minimizing the gas exchange consequences.

to function as an efficient gas exchange organ, the lung has an elaborate fluid balance system. Any increase in pulmonary vascular pressure has been shown to increase the lymphatic flow rate as a result of increased fluid transudation through the pulmonary capillaries (23). Alveolar edema (accumulation of fluid in the alveoli) alters gas exchange, with the development of shunt and significant gas exchange impairment, often with serious consequences (21). The development of alveolar edema is, however, generally preceded by interstitial edema in which fluid accumulation occurs not in the alveolar spaces, but in the pulmonary interstitium (22, 24, 25).

As the lung transitions from a normal “dry” state to one with alveolar edema, one would expect to see a progressive impairment of gas exchange. Volume overload edema (hydraulic edema) has been used in the past to examine this transition from the healthy “dry” lung to that in which edema is present. It has been shown that in the early stages of the development of interstitial edema, fluid accumulation occurs not in the gas exchange region of the thin alveolar-capillary membrane, but rather in locations around the pulmonary bronchioles and arterioles, with a common finding of peribronchial and perivascular cuffing observed on radiographs (13, 25). The development of such cuffing brings with it changes in lung mechanics (17) and the potential to alter the distribution of airflow, blood flow, or both, each with the possibility to alter the matching of ventilation to perfusion (V̇˙a/Q̇˙) and thus pulmonary gas exchange. Such studies show that rapid (within ∼30 min) intravenous infusion of 20 to 30 ml/kg of normal saline results in increased pulmonary capillary blood volume (6, 18) and cardiac output. However, in normal subjects the lung appears more able to handle this volume overload than in the case of patients with congestive heart failure (8).

We hypothesized that rapid saline infusion in an otherwise healthy lung would result in the development of a pulmonary gas exchange defect by increasing V̇˙a/Q̇˙ mismatch, as opposed to a diffusive impairment to O₂ exchange (as might be expected if alveolar-capillary membrane thickening were to occur) as seen in isolated lung models (12, 30). To test this hypothesis, seven normal young males had pulmonary gas exchange measured using the multiple inert gas elimination technique (MIGET) at baseline and between 15 and 90 min following rapid (30 min) intravenous infusion of 20 ml/kg of normal saline. To better assess any degree of diffusive limitation to O₂ transport, measurements were also made under hypoxic conditions (Fi_O2 = 0.125). Measurements were also made at an elevated Fi_O2 (0.30) to enhance detection of any hypoxic pulmonary vasoconstriction that may have developed in the saline-loaded lung. Contrary to our primary hypothesis, despite a clear effect of saline infusion on spirometric variables, there was no significant increase in V̇˙a/Q̇˙ heterogeneity, suggesting that during the early stages of hydraulic pulmonary interstitial edema, active matching of blood flow, and ventilation may be operative. Furthermore, and as expected, there was no evidence for a diffusive limitation to O₂ transport as a result of saline infusion.

METHODS

Subject population.

We studied seven males (26 ± 3 yr; FEV1: 110 ± 16% predicted) all of whom had no history of pulmonary or cardiac disease, were nonsmokers, and had a normal physical examination. The study was approved by the University of California, San Diego Human Research Subject Protection Program, and written, informed consent was obtained from each subject.

Study protocol.

Subjects were instrumented as described below, and the inert gas mixture was infused (see below) for 30 min to first establish steady-state gas exchange. Subjects remained supine throughout the study. Data collection blocks (described below) were performed at baseline breathing air and then at two values of Fi_O2 (0.125, 0.30) cycled in a block design fashion between subjects to eliminate any ordering effect. With the subjects breathing air, 10 ml/kg of saline was infused rapidly (over an ∼10-min period), and a data collection block performed. A second infusion of 10 ml/kg was performed resulting in the complete (20 ml/kg) infusion being performed in a 30- to 40-min period. Another data collection block was performed with the subject breathing air immediately following infusion. Following this, three data collection blocks were performed at three values of Fi_O2 (0.125, 0.21, 0.30) cycled in a block design fashion between subjects to eliminate ordering effects.

Instrumentation and measurements.

Subjects performed forced spirometry (model VRS 2000; S&D Instrumentation, Doylestown, PA) in the standing posture prior to the start of the study. They were then placed in the supine posture and remained in that posture until the end of the study (∼4 h later), at which time forced spirometry was again performed immediately after assuming the standing posture.

A radial artery cannula (20 gauge) was inserted under local anesthesia in the nondominant arm for collection of arterial samples. A thermistor for monitoring blood temperature was coupled to the radial arterial line. A peripheral venous cannula (20 gauge) was inserted in a forearm vein of the contralateral arm for infusion of the inert gas mixture (described below). A second peripheral venous cannula (∼18 gauge) was inserted into a large forearm vein on the same arm as the arterial line for the rapid saline infusion.

Subjects were instrumented with a tetrapolar impedance cardiograph (model BioZ ICG; CardioDynamics, San Diego, CA) providing both ECG monitoring and measurement of cardiac output and thoracic fluid content from the impedance signal. This device has been previously validated against direct Fick measures of cardiac output (31). A Hans Rudolph (Shawnee, KS) nonrebreathing valve (model 2700) connected the heated expiratory line to a heated mixing box for sampling of expired inert gases. Expiratory flow and oxygen consumption was measured with a metabolic cart (Parvomedics TrueMax; Parvo, UT) connected distal to the mixing box. The inspiratory line was connected to air or to bags containing an hypoxic gas mixture (Fi_O2 0.125) or hyperoxic gas mixture (Fi_O2 0.30) using large-bore valves.

Subjects breathed through the mouthpiece for ∼10 min prior to inert gas sampling and for all periods of breathing hypoxic or hyperoxic mixtures (∼20 min each). At other times, subjects were disconnected from the mouthpiece and breathed room air.

A data collection block consisted of duplicate inert gas samples (see below for details), chest impedance measures, and measurement of closing volume and expired ventilation. A single data collection block lasted ∼20 min.

MIGET.

The multiple inert gas technique was applied in the usual manner (7). Briefly, an inert gas solution containing dissolved sulfur hexafluoride, ethane, cyclopropane, enflurane, ether, and acetone was prepared in sterile 0.9% sodium chloride solution and infused in intravenously for ∼30 min before collection of the first samples. The rate of inert gas infusion was 3 ml/min. The total volume of fluid infused from the inert gas solution during the course of the study was ∼500 ml over ∼3–4 h, while the total volume of blood drawn from each subject was ∼200 ml. Duplicate samples of mixed expired gas (15 ml) and arterial blood (6 ml) were obtained in gas-tight glass syringes. Concentrations of six inert gases in mixed expired gas and arterial blood were measured using gas chromatography (model 5890A; Hewlett-Packard, Wilmington, DE) (27). Mixed venous concentrations of inert gases were calculated from the cardiac output (determined from the impedance cardiography measurements) and the measured arterial and mixed expired concentrations by using the Fick principle. V̇˙a/Q̇˙ distributions were obtained using a least-squares, best-fit regression analysis with enforced smoothing (5). We report the first moment (i.e., mean of the distribution) and second moment (i.e., log standard deviation of the distribution) of the V̇˙a/Q̇˙ distributions obtained. The residual sum of squares was used as an indicator of adequacy of fit of the data to the 50-compartment model.

Arterial sampling and blood gas measurements.

Following each collection of 6 ml of arterial blood for inert gas analysis, 2 ml of arterial blood was drawn into a separate, preheparinized syringe for arterial blood gas assessment. Blood gas samples were stored on ice until analyzed for Po₂, Pco₂, and pH using an IL Synthesis-45 blood gas analyzer (Instrumentation Laboratories, Lexington, MA). Samples were measured no more than 30 min after being taken. All blood gas values were corrected to the blood temperature measured during blood sampling.

Statistical analysis.

Results from duplicate inert gas measurements were averaged. Values of cardiac output, stroke volume, thoracic fluid content, and heart rate were averaged over stable 2-min periods just before the inert gas sampling. Resulting values were analyzed using repeated-measures ANOVA (StatView version V5.1, SAS Institute, Cary, NC). In cases where significant main effects were observed, post hoc testing was performed using a two-tailed Student's t-test. Significance was accepted at P < 0.05.

RESULTS

All subjects tolerated the procedures well, and no subjects reported any adverse symptoms as a result of infusion.

Effects of infusion while breathing 21% O₂.

Spirometry performed before and after the supine portion of the study showed significant changes as a consequence of infusion (Table 1). Forced vital capacity was unchanged by infusion, but there were significant reductions in FEV₁/FVC (by 3%), and in flows at midlung volumes (FEF_25–75 by 19% and FEF₅₀ by 13%). However, closing volume was not able to be reliably measured in this subject population.

Table 1.

Spirometry following infusion

	Preinfusion	Postinfusion
FVC, liters	6.31 ± 0.48	6.19 ± 0.46
FEV1/FVC, %	80.7 ± 1.4	78.1 ± 1.3*
FEF25–75, l/s	5.13 ± 0.45	4.16 ± 0.42*
FEF50, l/s	5.78 ± 0.52	4.98 ± 0.51*

Values are means ± SE.

^*P < 0.05 compared with preinfusion.

Infusion resulted in a significant increase in thoracic fluid content as measured by impedance cardiography (Fig. 1). There was a consistent, progressive, and significant increase in thoracic fluid content after both 10 and 20 ml/kg infused saline. The overall increase in thoracic fluid content after saline infusion was 12%, which was maintained through the postinfusion measurement period.

Fig. 1.

Thoracic fluid content (impedance cardiography). Values are those measured at an Fi_O2 of 0.21. Preinfusion is the average of 2 measurement blocks. a.u., Arbitrary units. *P < 0.05 compared with preinfusion.

Other cardiovascular changes measured by impedance cardiography were very much smaller than the change in thoracic fluid content and are documented in Table 2. In brief, cardiac output was unaltered by infusion, although there was a slight decrease in stroke volume and a concomitant increase in heart rate. Systolic arterial blood pressure showed a slight increase of 7 mmHg as a result of infusion (P = 0.003), but there were no significant increases in mean or diastolic blood pressure (Table 2).

Table 2.

Effects of infusion

	Preinfusion	Midinfusion	End Infusion	Postinfusion	P Values from ANOVA
HR, beats/min	57 ± 2	59 ± 2	65 ± 5	65 ± 4	0.02
SV, ml	106 ± 7	103 ± 7	100 ± 7	97 ± 7	<0.0001
Q̇˙c, l/min	5.7 ± 0.3	5.8 ± 0.4	5.9 ± 0.5	5.7 ± 0.4	0.70
Systolic BP, mmHg	137 ± 5	144 ± 5	145 ± 6	144 ± 5	0.003
Mean BP, mmHg	88 ± 2	90 ± 2	90 ± 3	91 ± 3	0.53
Diastolic BP, mmHg	69 ± 2	71 ± 1	71 ± 2	72 ± 2	0.51
V̇˙e, l/min, BTPS	7.9 ± 0.4	8.1 ± 0.3	8.2 ± 0.4	8.3 ± 0.4	0.72
V̇˙D/V̇˙T	0.40 ± 0.03	0.40 ± 0.03	0.38 ± 0.04	0.38 ± 0.04	0.295
V̇˙a, l/min, BTPS	4.7 ± 0.2	5.0 ± 0.2	5.0 ± 0.2	5.1 ± 0.2	0.51
V̇˙o2, l/min	0.25 ± 0.01	0.25 ± 0.01	0.25 ± 0.01	0.26 ± 0.01	0.82
V̇˙co2, l/min	0.21 ± 0.01	0.22 ± 0.01	0.22 ± 0.01	0.22 ± 0.01	0.75
PaO2, mmHg	99 ± 2	99 ± 3	100 ± 2	99 ± 3	0.998
PaCO2, mmHg	38.7 ± 0.6	38.3 ± 0.8	37.5 ± 0.8	36.8 ± 1.2	0.042
A-aDO2, mmHg	3.1 ± 2.0	5.3 ± 3.3	5.6 ± 3.1	5.7 ± 2.7	0.63
Q̇˙s/Q̇˙t	0.000 ± 0.000	0.000 ± 0.000	0.000 ± 0.000	0.000 ± 0.000	1.0
LogSDQ̇˙	0.38 ± 0.03	0.36 ± 0.03	0.36 ± 0.03	0.37 ± 0.03	0.68
LogSDV̇˙	0.40 ± 0.03	0.37 ± 0.04	0.38 ± 0.04	0.38 ± 0.04	0.94
Disp(R-E)	2.21 ± 0.32	1.93 ± 0.30	2.21 ± 0.40	2.10 ± 0.35	0.86
A-aDO2, measured-predicted, mmHg	−3.6 ± 1.7	−0.9 ± 2.6	−0.2 ± 2.7	−1.0 ± 1.7	0.41

Values are means ± SE. All data were collected at a Fi_O2 of 0.21. Preinfusion, average of data prior to infusion (2 points); midinfusion, following intravenous saline infusion of 10 ml/kg; end infusion, following intravenous saline infusion of 20 ml/kg; postinfusion, collected between 30 and 90 min after the termination of infusion (see text for details); HR, heart rate; SV, stroke volume; Q̇˙c, cardiac output; V̇˙e, expired minute ventilation (l[BTPS]/min); BTPS, body temperature pressure saturated; V̇˙o₂, oxygen consumption (l[STPD]/min); STPD, standard temperature pressure dry; V̇˙co2, carbon dioxide production (l[STPD]/min); Pa_O2, arterial Po₂; Pa_CO2, arterial Pco₂; A-aDO₂, alveolar-arterial oxygen partial pressure difference; V̇˙D/V̇˙T, fractional dead space ventilation from multiple inert gas elimination technique (MIGET); V̇˙a, alveolar ventilation (l[BTPS]/min); Q̇˙s/Q̇˙t, fractional shunt from MIGET; LogSDQ̇˙: log standard deviation (SD) of the perfusion distribution from MIGET. LogSDV̇˙: log SD of the ventilation distribution from MIGET; Disp(R-E), dispersion of the retention and excretion data from MIGET. A-aDO₂ measured less that predicted from MIGET, see text for details.

Pulmonary ventilation and gas exchange were only slightly altered by infusion with a significant decrease in the arterial Pco₂ (by 1.9 mmHg, Table 2), although changes in expired minute ventilation and alveolar ventilation were not statistically significant. There was no evidence for impairment in pulmonary gas exchange as evidenced by the lack of change in both Pa_O2, and the alveolar-arterial oxygen partial pressure difference (A-aDO₂).

No measures of heterogeneity of pulmonary gas exchange from the multiple inert gas data were significantly altered by infusion (Table 2). In particular, there was no alteration in V̇˙a/Q̇˙ heterogeneity and no evidence for diffusive limitation to O₂ exchange (i.e., no difference between measured arterial Po₂ and that predicted by MIGET from V̇˙a/Q̇˙ inequality) since A-aDO₂ (measured-predicted) was low and not significantly different from zero (all cases P > 0.05).

Effects of hypoxia.

There was a reduction in Pa_CO2 due to increased ventilation at reduced Fi_O2. However, there were no significant effects on any of the variables pertaining to matching of ventilation to perfusion resulting from hypoxia. While there were changes from the normoxic condition resulting from the hypoxia itself, these were unaltered by infusion (Table 3). The increase in LogSDQ̇˙ in hypoxia seen in this study is consistent with previous studies at high altitude (29) The A-aDO₂ (measured-predicted), indicative of diffusion limitation, was not increased by hypoxia, nor did infusion alter this.

Table 3.

Effects of infusion at different Fi_O2

	Hypoxia, 12.5%		Normoxia, 21%		Hyperoxia, 30%		P Values from ANOVA
	Preinfusion	Postinfusion	Preinfusion	Postinfusion	Preinfusion	Postinfusion	Effect of Infusion	Effect of FiO2	Effect of Infusion × FiO2
TFC, a.u.	38.7 ± 3.1	43.9 ± 3.5	38.9 ± 3.2	43.7 ± 3.5	39.1 ± 3.0	43.9 ± 3.7	0.0001	0.34	0.50
Q̇˙c, l/min	6.3 ± 0.4	6.4 ± 0.5	5.7 ± 0.3	5.7 ± 0.4	5.7 ± 0.3	5.7 ± 0.4	0.62	<0.001	0.77
V̇˙e, l/min, BTPS	8.6 ± 0.3	9.5 ± 0.2	7.9 ± 0.4	8.3 ± 0.4	8.0 ± 0.6	6.9 ± 0.4	0.79	0.002	0.006
V̇˙D/V̇˙T	0.34 ± 0.03	0.31 ± 0.05	0.40 ± 0.04	0.38 ± 0.04	0.39 ± 0.04	0.36 ± 0.05	0.32	0.006	0.99
V̇˙a, l/min, BTPS	5.9 ± 0.4	6.1 ± 0.3	4.7 ± 0.2	5.1 ± 0.2	5.0 ± 0.4	4.7 ± 0.2	0.57	0.0005	0.32
PaO 2, mmHg	42 ± 1	42 ± 1	99 ± 2	99 ± 3	155 ± 4	151 ± 4	0.54	<0.0001	0.38
PaCO2, mmHg	36.3 ± 0.7	34.2 ± 0.8	38.7 ± 0.5	36.8 ± 1.2	38.3 ± 1.1	38.1 ± 1.0	0.006	0.0004	0.12
A-aDO2, mmHg	6.4 ± 0.9	8.7 ± 0.8	3.1 ± 2.0	5.8 ± 2.7	8.4 ± 3.1	10.0 ± 3.6	0.22	0.14	0.85
A-aDO2, measured-predicted, mmHg	4.3 ± 1.0	5.2 ± 1.0	−3.6 ± 1.6	−1.0 ± 1.7	0.0 ± 3.5	2.4 ± 3.1	0.14	0.06	0.82
Mean of Q̇˙ Dist	0.82 ± 0.05	0.80 ± 0.06	0.77 ± 0.04	0.82 ± 0.06	0.80 ± 0.08	0.90 ± 0.06	0.19	0.14	0.20
LogSDQ̇˙	0.43 ± 0.02	0.48 ± 0.04	0.38 ± 0.03	0.37 ± 0.03	0.44 ± 0.04	0.44 ± 0.04	0.57	0.003	0.30
Mean of V̇˙ Dist	0.85 ± 0.07	0.90 ± 0.06	0.71 ± 0.09	1.00 ± 0.12	1.07 ± 0.08	0.91 ± 0.07	0.16	0.04	0.005
LogSDV̇˙	0.47 ± 0.03	0.53 ± 0.04	0.40 ± 0.03	0.38 ± 0.04	0.43 ± 0.02	0.58 ± 0.09	0.06	0.04	0.19
Disp(R-E)	2.87 ± 0.32	3.79 ± 0.53	2.21 ± 0.32	2.10 ± 0.35	2.58 ± 0.26	3.53 ± 0.57	0.03	0.002	0.21

Values are mean ± SE. Q̇˙ Dist, distribution of perfusion; V̇˙ Dist, distribution of ventilation; other variables as in Table 2.

Effects of hyperoxia.

Hyperoxia resulted in a reduction in expired ventilation following infusion (Table 3) in contrast to normoxia and hypoxia. There was, however, no change in alveolar ventilation.

There was a small but significant effect of altered Fi_O2 on LogSDQ̇˙, although there was no effect of saline infusion. There was a nonsignificant trend for an increase in LogSDV̇˙ postinfusion, which rose by 26% as a result of infusion when breathing the hyperoxic gas mixture. This effect is shown in Fig. 2. Dispersion of the combined retention and excretion curves [Disp(R-E)] was significantly increased by both infusion and by an alteration in Fi_O2 (Fig. 2C, Table 3).

Fig. 2.

Changes in log standard deviation of the ventilation (LogSDV̇˙) (A) and perfusion (LogSDQ̇˙) (B) and dispersion of the combined retention and excretion curves (R-E) (C) following infusion at 3 different values of Fi_O2. Fi_O2, fraction of inspired oxygen.

Prior to infusion, the fraction of V̇˙ or Q̇˙ going to areas of low V̇˙a/Q̇˙ fell when Fi_O2 was increased to 0.30 (Fig. 3). However, after infusion, the fraction of ventilation going to areas with V̇˙a/Q̇˙ < 1 was increased when breathing a hyperoxic gas mixture [0.36 ± 0.04 (Fi_O2 = 0.21), 0.41 ± 0.03 (Fi_O2 = 0.30), P < 0.05] with similar effects on perfusion [0.72 ± 0.06 (Fi_O2 = 0.21), 0.80 ± 0.06 (Fi_O2 = 0.30), P < 0.05].

Fig. 3.

Fractional ventilation (A) and perfusion (B) to areas of the lung with ventilation/perfusion (V̇˙a/Q̇˙) between 0.1 and 1.0. Note that there was a significant interaction between infusion and a change in Fi_O2 as evidenced by the increase in ventilation to areas of low V̇˙a/Q̇˙ when Fi_O2 was 0.30.

DISCUSSION

Overall effect of infusion.

Our results show that rapid infusion of 20 ml/kg of normal saline resulted in significant effects on forced expiratory flows as evidenced by a drop in the FEV₁/FVC ratio and large reductions in forced expiratory flows at midlung volumes. Electrical impedance measurements showed a substantial rise in the fluid content of the thorax (Fig. 1), which was maintained over the course of the postinfusion period. This change was accompanied by a slight but significant drop in arterial Pco₂ of ∼2 mmHg (Table 2), likely as a result of a dilutional acidosis from the infusion (14, 18). Infusion also resulted in a small increase in heart rate (by 8 beats/min), although there was no change in cardiac output due to a concomitant reduction in cardiac stroke volume and no change in arterial blood pressure (Table 2). While impedance provides an indirect measurement of cardiac output, it has been well validated against direct Fick (31). Furthermore, errors in the measurement of cardiac output have only minimal influence on heterogeneity measures (LogSDQ̇˙, LogSDV̇˙) derived from MIGET (28). The errors induced in these variables are an order of magnitude smaller than those in cardiac output and thus are physiologically insignificant.

In these respects, our results agree with and support those of other investigators. Muir et al. (14) showed an increase in airway closure and a redistribution of pulmonary blood flow measured using ¹³³Xe as a result of infusion of 2 liters of saline in men in an upright position. Interestingly, they showed a short-term improvement in gas exchange as evidenced by a transient increase in Pa_O2; however, Pa_O2 in two of four subjects subsequently fell below control values in the period after completion of the infusion. These changes were attributed to increases in lung water (which could not be directly measured) altering the distribution of both ventilation and perfusion. Subsequent studies show a marked increase in pulmonary capillary blood volume following 2 liters of saline infusion with little or no change in membrane diffusing capacity or lung water measured using soluble gas uptake (6). More recent studies have included patients with congestive heart failure and show differences between these subjects and normals (8, 9, 16), which is likely due to elevated left atrial pressures in these patients. The principal difference appears to be that the normal lung presents much more resistance to the development of edema within the gas exchange region than that in patients with congestive heart failure, and so changes, such as a reduction in membrane diffusing capacity with increased lung water is much smaller or absent in normal subjects. In isolated dog lungs, interstitial edema fluid accumulation results in the alteration of the caliber of extra-alveolar vessels and in diminished patency of small airways (12, 30). A recent study using high resolution computed tomography showed a marked increase in the thickness of the airway walls in normal subjects who received 3 liters of infused saline (13), consistent with the previous studies in isolated lungs.

However, despite the presence of an obvious effect of infusion based on changes in ventilation, heart rate, and poststudy spirometry, there was no detrimental effect on the efficiency of pulmonary gas exchange (there was no increase in A-aDO₂). Consistent with this, MIGET showed only minimal changes in any of the key measures of heterogeneity of matching of ventilation to perfusion. Indeed, what was remarkable was the degree to which heterogeneity measures (LogSDV̇˙ and LogSDQ̇˙) remained unchanged following infusion of 20 ml/kg (see Table 2).

No evidence for diffusive limitation to O₂ transport from infusion.

Previous studies have shown that infusion, and the subsequent development of pulmonary interstitial edema, occurs not in the gas exchange region of the lung, but in regions proximal to gas exchange, namely in regions surrounding the bronchioles and arterioles (13). Our data support this.

If interstitial edema were to develop in the gas exchange region of the lung in the form of fluid accumulation in the alveolar-capillary membrane, one would expect that this would be reflected in a limitation to O₂ diffusion, especially in hypoxia (26). For this reason, we performed the MIGET studies at an Fi_O2 of 0.125. If diffusive limitation to O₂ transport were present, we would expect to see a discrepancy between the measured A-aDO₂ and that predicted on the basis of the observed distribution of V̇˙a/Q̇˙ from MIGET (10).

However, our data show that there was no effect of infusion in hypoxia (there was no significant interaction between infusion and Fi_O2; see Table 3), suggesting that any fluid accumulation in the lungs did not occur in the gas exchange region or was insufficient to result in diffusive limitation. This functional measurement is consistent with the previous radiological and functional observations (6, 13, 20) that fluid accumulation in the early stages of pulmonary interstitial edema is proximal to the gas exchange region and does not influence alveolar gas exchange (25).

Possible contribution of hypoxic pulmonary vasoconstriction.

If hypoxic pulmonary vasoconstriction (HPV) were present in the lung as a result of saline infusion, then breathing 30% O₂ would be expected to alleviate the constrictive effect and increase blood flow to poorly ventilated regions of the lung. This would be expected to increase the V̇˙a/Q̇˙ inequality, in particular LogSDQ̇˙. We chose 30% O₂ to preclude any confounding effect of alveolar instability at high O₂ fractions (4) that might occur if regions of low V̇˙a/Q̇˙ resulted from infusion. The absence of any development of shunt at high Fi_O2 (Table 3) shows that absorption atelectasis did not occur.

Before infusion, LogSDV̇˙ was 0.43 ± 0.04 at an Fi_O2 of 0.30, a value typically seen in healthy subjects (Fig. 2A, Table 3). Following infusion, this was 0.58 ± 0.09 (P = 0.06 for effects of infusion), which on the basis of previous studies is above that normally seen and close to that seen in previous studies of V̇˙a/Q̇˙ heterogeneity following heavy exercise (11, 15). There was a similar and significant increase in Disp(R-E), [a global measure of V̇˙a/Q̇˙ mismatch, which is not dependent on the 50-compartment model used to calculate other parameters of V̇˙a/Q̇˙ mismatch (19)] (Fig. 2C), but there was no effect on LogSDQ̇˙ (Fig. 2B).

Figure 3 shows the effects of a raised Fi_O2 on regions of low V̇˙a/Q̇˙ (those areas with a low Pa_O2). Prior to infusion, raising the Fi_O2 did not result in an increase in ventilation or perfusion to areas of low V̇˙a/Q̇˙ (● Fig. 3). However, following saline infusion, the fractional ventilation and perfusion to these areas was increased when Fi_O2 was increased (○, Fig. 3; both effects P < 0.05), consistent with the release of vasoconstriction in areas of the lung with low V̇˙/Q̇˙ ratios. Similarly, there was a fall in the mean of the ventilation distribution following saline infusion only at an Fi_O2 of 30% (Table 3) (the mean of the ventilation distribution showed a significant interaction between saline infusion and changing Fi_O2). Taken together with a lack of change in Q̇˙, or in V̇˙ (as evidenced by a lack of change in Pa_CO2 at Fi_O2 = 0.30, Table 3), these effects show a modest broadening of the distribution of ventilation with a move toward lower values of V̇˙a/Q̇˙, consistent with an increase in perfusion to areas of low V̇˙a/Q̇˙. More perfusion to areas of low V̇˙a/Q̇˙ results in a leftward shift of the ventilation and perfusion distributions as shown in previous studies of HPV using MIGET which have shown similar effects (1–3).

Thus our results suggest that a mild degree of HPV was operative in these subjects. That the effect is seen only following saline infusion suggests that these areas have restricted ventilation as a result of infusion, likely from peribronchial cuffing, and that HPV has acted to maintain a more uniform distribution of V̇˙a/Q̇˙ postinfusion than would otherwise be the case.

Conclusions.

Infusion resulted in a clearly demonstrable effect on forced expiratory flow limitation, and an increase in thoracic fluid content based on impedance cardiography and in increase in heart rate. These effects, taken with previous studies, suggest that our intervention of 20 ml/kg iv saline in ∼30 min likely resulted in the development of pulmonary interstitial edema, consistent with prior studies in the field (6, 13). However, despite these changes, infusion did not result in a impaired pulmonary gas exchange as measured by either arterial blood gases, or by the multiple inert elimination technique. Our data obtained breathing at an elevated Fi_O2 provide limited support for interstitial edema resulting in the development of regions of low ventilation that under normoxic conditions are subject to HPV, normalizing the distribution of V̇˙a/Q̇˙ and, thus, gas exchange.

GRANTS

This study was supported by National Heart, Lung, and Blood Institute Grants HL-080203, HL-081171, and HL-084281, and American Heart Association Grant 00540002N. I. M. Olfert received funding support as a Parker B. Francis fellow.

DISCLOSURES

No conflicts of interest are declared by the author(s).