Impeded Alveolar-Capillary Gas Transfer With Saline Infusion in Heart Failure
Abstract—The microvascular pulmonary endothelium barrier is critical in preventing interstitial fluid overflow and deterioration in gas diffusion. The role of endothelium in transporting small solutes in pathological conditions, such as congestive heart failure (CHF), has not been studied. Monitoring of pulmonary gas transfer during saline infusion in CHF was used to probe this issue. Carbon monoxide diffusion (DLCO), its membrane diffusion (DM) and capillary blood volume (VC) subcomponents, and mean right atrial (rap) and mean pulmonary wedge (wpp) pressures after saline or 5% d-glucose solution infusions were compared with baseline in 26 moderate CHF patients. Saline was also tested in 13 healthy controls. In patients, 750 mL of saline lowered DLCO (−8%, P<0.01 versus baseline), DM (−10%, P<0.01 versus baseline), aldosterone (−29%, P<0.01 versus baseline), renin (−52%, P<0.01 versus baseline), and hematocrit (−6%, P<0.05 versus baseline) and increased VC (20%, P<0.01 versus baseline), without changing rap and wpp. Saline at 150 mL produced qualitatively similar results regarding DLCO (−5%, P<0.01 versus baseline), DM (−7%, P<0.01 versus baseline), VC (9%, P<0.01 versus baseline), rap, wpp, aldosterone (−9%, P<0.05 versus baseline), and renin (−14%, P<0.05 versus baseline). Glucose solution (750 mL), on the contrary, increased DLCO (5%, P<0.01 versus 750 mL of saline) and DM (11%, P<0.01 versus 750 mL of saline) and decreased VC (−9, P<0.01 versus 750 mL of saline); aldosterone (−40%), renin (−41%), hematocrit (−3%), rap, and wpp behaved as they did after saline infusion. In controls, responses to both saline amounts were similar to responses in CHF patients regarding aldosterone, renin, hematocrit, rap, and wpp, whereas DLCO, DM, and VC values tended to rise. Hindrance to gas transfer (reduced DLCO and DM) with salt infusion in CHF, despite an increase in VC and no variations in pulmonary hydrostatic forces, indicates an upregulation in sodium transport from blood to interstitium with interstitial edema. Redistribution of blood from the lungs, facilitating interstitial fluid reabsorption, or sodium uptake from the alveolar lumen by the sodium-glucose cotransport system might underlie the improved alveolar-capillary conductance with glucose.
For the lung parenchyma to allow gas exchange between blood and gas in the alveoli, a continuous clearance is required of the excess of fluid into the interstitial space and the alveolar lumen. Excessive water accumulation in these compartments is called pulmonary and alveolar edema, respectively. The pulmonary microvascular endothelium and the alveolar epithelium constitute a barrier that is critical for gas exchange and modulation of fluid and solute passage between blood, the interstitial compartment, and alveoli.1 Despite the substantial progress that has been made in understanding the physiology of the endothelial and epithelial layers in regulating lung fluid balance, further study regarding the local and systemic regulatory factors under pathological conditions is necessary.1
An imbalance in hydrostatic forces is interpreted as the basic mechanism for volume overload and cardiogenic pulmonary edema.2 In congestive heart failure (CHF), pulmonary edema may be absent despite elevation of the hydrostatic pressures3 or may be present even if pressures are normal.4 Whether altered hemodynamics are the exclusive mechanisms for pulmonary edema in CHF and whether alterations in the capillary endothelium and/or in the alveolar epithelium barrier contribute to changes in salt, water, and gas transfer have been the subjects of only limited research.3 Cardiogenic anatomic injuries of the alveolar-capillary membrane have been reported in animals and humans.5 6
We examined the hypothesis of an alteration in the microvascular endothelium barrier by infusing saline into CHF patients and monitoring the pulmonary diffusing capacity for CO (DLCO) and its subdivisions, ie, alveolar-capillary membrane diffusing capacity (DM) and capillary blood volume available for gas exchange (VC). We reasoned that an alteration in sodium handling by the alveolar-capillary membrane would be substantiated if infusions of saline amounts devoid of hydrostatic effects were proven to impede the membrane subcomponent of the pulmonary gas transfer in CHF patients and not in healthy individuals.
Patients and Controls
We investigated male patients referred to the Institute of Cardiology, University of Milan, for evaluation of chronic heart failure. They presented with either idiopathic cardiomyopathy (cardiac enlargement, reduced ejection fraction [EF], and absence of a specific cause for cardiac failure) or ischemic heart disease (documented previous myocardial infarction). Inclusion criteria were (1) cardiac dysfunction, classified as chronic stable NYHA functional class II to III, and (2) left ventricular EF ≤40%. Exclusion criteria were (1) present or past history of smoking (>10 cigarettes per day during 1 of the past 5 years), (2) history of respiratory or renal diseases, (3) evidence of renal impairment (serum creatinine concentration ≥0.05 mmol/dL) or airway obstruction (forced expiratory volume in 1 second [FEV1] to vital capacity ratio <70%), (4) DLCO ≤75% of predicted normal value on the basis of standard nomograms incorporating age, gender, height, and weight,7 (5) mitral regurgitation exceeding grade 3 on a subjective scale from 0 to 5, and (6) angiotensin-converting enzyme (ACE) inhibitor therapy, acetylsalicylic acid, or other cyclooxygenase inhibitors within the last 2 months. ACE inhibition, in fact, ameliorates pulmonary diffusion in CHF, and cyclooxygenase blockers counteract this effect.8
Twenty-six patients (10 in NYHA functional class II and 16 in class III) took part in this investigation; none of them had participated in previous studies in our laboratory. Thirteen healthy men who were similar in age and physical characteristics to the patients and who were nonsmokers volunteered to serve as controls. They had been admitted to the hospital because of atypical chest pain and had no history of respiratory or renal disease; physical examination, serum creatinine concentration, ECG, echocardiogram, chest x-ray, and coronary angiography were all normal for this group.
The protocol was approved by the institution Ethics Committee, and written informed consent was obtained from each subject. The procedures followed were in accordance with institutional guidelines.
Pulmonary Function Testing
Measurements of FEV1, vital capacity, and total lung capacity were made with the Sensor Medics 2200 Pulmonary Function Test System. DLCO was determined twice, with washout intervals of at least 4 minutes (the average was taken as the final result), with a standard single-breath technique.9 Measured diffusing capacity was corrected for the subject’s hemoglobin (Hb) concentration. The single breath alveolar volume (VA) was derived by methane dilution. DM and VC were determined with the same equipment, according to the classic Roughton and Forster method.9 10 The high oxygen concentration in the test gas was 89.7%. Studies of reproducibility showed a high level of agreement between consecutive measurements of 1/DM, with a correlation coefficient of 0.97 and a coefficient of variation <5%. To assess any theoretical effects of CO back pressure, 4 randomly chosen patients with CHF and 5 normal subjects underwent a control study using the same study protocol but without invasive procedures and fluid infusions. We have found no evidence of significant CO back-pressure effects on serial DLCO, DM, and VC measurements under the present study protocol.
All patients were maintained on stable optimal doses of digoxin and furosemide (dosing was set at 5 pm), and none had overt signs of fluid retention. After completion of the screening tests, patients and controls were placed on a constant isocaloric diet that contained 100 mmol/L Na+, 70 mmol/L K+, and 1500 mL of water per day for the entire study. After 5 days of a controlled diet, the subjects were admitted for 6 days and 5 nights to the Heart Failure Unit, where confirmation of sodium balance was achieved, with urinary Na+, K+, and creatinine monitoring in the first 24 hours. The protocol included infusions of 0.9% NaCl and 5% d-glucose solutions, both in amounts of 150 and 750 mL. The duration of the study and the infusion sequence are depicted in Figure 1⇓. The investigators were blind to the patient clinical condition and the type of solution to be infused. Studies were begun at 8 am after an overnight fast. A triple-lumen, flow-directed, thermodilution balloon-tipped catheter was inserted into an antecubital vein and advanced to the pulmonary circulation under fluoroscopic control with the subject in the recumbent position. Then, the subject’s chest was elevated at 45°, and this comfortable position was maintained throughout the studies. For measurement of water and Na+ excretion, urine was collected in the 3 hours that preceded and in the 3 hours that followed fluid infusion. The infusion was made into the main stem of the pulmonary artery at a rate of 0.2 mL · kg−1 · min−1, and the starting solution was selected randomly. Right atrial and pulmonary wedge pressures were monitored throughout each study. The catheter was removed in the 48-hour washout interval between one type of solution and the other. DLCO, DM, and VC were determined hourly in the 2 hours preceding the infusion, shortly after the infusion, and hourly for the next 3 hours. Ten minutes before and after each infusion, mixed venous blood was withdrawn for measurements of hematocrit (Htc) and Hb, plasma protein and aldosterone concentrations, and plasma renin activity (PRA); cardiac output (average of 2 determinations) and left ventricular EF were also measured. An aliquot of the blood sample was rapidly removed for evaluation of Htc. The remainder of the blood was immediately sent for the other measurements at a central laboratory.
Left ventricular EF (Simpson’s rule) was assessed at rest by 2-dimensional echocardiography (Hewlett-Packard Sonos 1500). Standard color Doppler velocimetry was used to measure the degree of mitral regurgitation, which was graded subjectively on a scale from 0 (none) to 5 (severe). All Htc measurements were corrected for trapped plasma volume and for whole-body Htc. PRA and aldosterone plasma concentrations were determined by radioimmunoassay. Electrolyte levels in urine were measured by ion-selective electrodes.
Data are presented as mean±1 SD. χ2 analysis was applied to compare the descriptive parameters. Comparisons of the basal data were performed by unpaired t test or Wilcoxon rank test, as appropriate. To analyze temporal changes after infusion, 1-way ANOVA for repeated measurements followed by the post hoc Newman-Keuls procedure was used. Comparisons of the responses to the same solution or to different solutions were tested by unpaired and paired t test. Differences were considered to be significant at P<0.05.
Demographic characteristics, blood pressure, heart rate, renal function, and urinary output were similar in patients and controls (Table 1⇓). FEV1, total lung capacity, forced vital capacity, DLCO, DM and its ratio to effective VA (DM/VA), cardiac index (CI), and left ventricular EF were lower and plasma aldosterone and renin activity were higher in CHF patients compared with controls; mean right atrial pressure rap and mean pulmonary wedge pressure wpp, Htc, Hb, and plasma protein concentration (PPC) were similar in the 2 populations (Table 1⇓). There were no significant systematic differences in the preinfusion data between sessions. In the patient and control groups, with the initial baseline DLCO, DM, and VC used as the reference values, hourly collections of data during 2 hours of resting before each session were averaged for each subject. Mean changes in DLCO, DM, and VC during this period of observation were minimal in both populations: DLCO 2.05±0.3%, DM 1.6±0.1%, and VC 1.8±0.8% in patients, and DLCO 1.98±0.2%, DM 1.7±0.3%, and VC 1.4±0.6% in healthy subjects.
CHF Patients Versus Controls and 150 Versus 750 mL Saline
After the administration of 150 mL of saline in the CHF group, we recorded a decrease of circulating aldosterone (−9.2%) and PRA (−13.7%), without variations in Htc, Hb, PPC, left ventricular EF, and CI (Table 2⇓). There was also (Figures 2⇓ and 3⇓) a reduction of DLCO (−5%), DM (−6.6%), and DM/VA (−10%) and an increase in VC (+9%) (Table 2⇓). These changes were not associated with variations in right atrial, pulmonary arterial, and wedge capillary pressures and disappeared within <1 hour. The humoral response to 750 mL of saline in the CHF group, compared with 150 mL of saline in the same patients, consisted of a greater inhibition of aldosterone secretion (−28.8%) and PRA (−52.3%) and a reduction in Htc (−5.6%), Hb (− 5.8%), and PPC (−5.9%), without variations in left ventricular EF and CI (Table 2⇓). The infusion of 750 mL of saline produced an 8.3% decrease of DLCO, 10.3% decrease of DM, 8.9% decrease of DM/VA, and 20% increase of VC (Figures 2⇓ and 3⇓, Table 2⇓). DLCO and DM were still reduced 1 hour after infusion and reverted to baseline in the next hour. Notably, 750 mL of saline did not raise rap and wpp (Figures 2⇓ and 3⇓).
In the healthy group, effects of 150 and 750 mL of saline were similar to those in CHF patients, as far as aldosterone, PRA, Htc, Hb, PPC, left ventricular EF, CI, pulmonary arterial pressure, rap, and wpp are concerned (Figures 2⇑ and 3⇑, Table 2⇑). In spite of this, we did not observe variations in DLCO, DM, VC, and DM/VA (Figures 2⇑ and 3⇑, Table 2⇑).
Glucose Versus Saline and 150 Versus 750 mL in CHF Patients
As shown in Table 3⇓, 150 mL of glucose solution was not effective on Htc, Hb, PPC, aldosterone secretion, PRA, VA, and left ventricular EF. Variations from baseline in DLCO, DM, and DM/VA were not significant, but the differences from both the corresponding absolute values and the percent changes with a similar amount of saline were significant (Figures 4⇓ and 5⇓). Glucose caused a 13.3% reduction of VC, which was significant compared with the 9% increase with saline. rap and wpp did not vary.
Glucose and saline infusions of 750 mL similarly inhibited aldosterone secretion and PRA; Htc and Hb concentrations were reduced less by glucose, and left ventricular EF, CI, and VA were not affected (Table 3⇑). As depicted in Figures 4⇑ and 5⇑, DLCO showed a 4.4% increase that was significant compared with the 8.3% decrease with the same amount of saline. Variations of DM (10.6%), DM/VA (4.2%), and VC (−9%) with glucose were also the opposite of those elicited by saline (DM −10.3%, DM/VA −8.9%, and VC 20%); differences were highly significant. As shown in Figure 4⇑, DM values with one solution were still significantly different from those with the other solution 1 hour after infusion. Right and left ventricular filling pressures did not vary after 750 mL of glucose. Urinary output in the 3 hours after, compared with that in the 3 hours before, 150 mL of saline administration was similar in patients and controls. With the 750 mL infusions, output was raised by 64±10% with glucose and by 58±8% with saline in patients and by 68%±12% with saline in controls. Differences between patients and controls and between saline and glucose were not significant.
Several human and animal studies11 12 13 14 have investigated the transition of pulmonary edema or cardiovascular adaptations during fluid or salt loading. In these studies, liquid volumes were large enough to induce obvious hemodynamic variations. Because we aimed at avoiding these effects, we chose to infuse a volume of saline (150 mL) similar to the accepted value for pulmonary capillary blood volume in normal humans and a 5-fold greater amount.
Saline, DLCO, and Its Subdivisions
Conditions causing an increase in circulating blood volume or movement of blood from the periphery to the thorax, as during head-down tilting15 and water immersion,16 are associated with an acute increase in VC, DM, and DLCO. During sustained microgravity, DM may increase even in excess of the rise in VC17 because of capillary recruitment and uniform capillary filling. DM in these situations is primarily determined by the surface area available for gas exchange.
In the present study, saline was not effective in normal subjects, whereas in patients it reduced DLCO, DM, and DM/VA and increased the VC. These changes disappeared in <1 hour and were relatively greater with the larger volume of saline. Of surprise to us was the observation that VC was augmented with the infusion of 150 mL of saline. Because the lung is the only organ to receive the entire cardiac output, the hemodynamic effect of a fluid load is most pronounced in the pulmonary circulation12 and this may explain why small amounts of saline were effective in increasing VC. This, however, does not seem to apply to normal individuals.
An acute decrease in DM with an increase in VC implies that the thickness of the alveolar capillary membrane be acutely augmented, thus impeding the passage of the respiratory gases. Accordingly, a simple interpretation may be that saline in CHF patients causes subclinical interstitial pulmonary edema, which results in the reduction of DM. It is remarkable that there were no changes in pulmonary hydrostatic forces, left ventricular EF, and right atrial pressure. During large volume loading, central venous pressure is generally raised (Gabel et al13 have used 1500 mL of Ringer’s solution to impede lymphatic drainage in the dog), opposes lung lymph flow, and leads to excess of fluid in the lung. These considerations do not support the interpretation that in our patients changes in hydrostatic forces were the main reason for an excess of interstitial fluid and an impedance of gas transfer and suggest that CHF may be associated with an upregulation of sodium transport from blood to interstitium with interstitial edema (or perhaps swelling of the endothelium or epithelium), which reduces DM. The relatively slow response of the lymphatic system in the interstitium to a rapid increase in the net fluid filtration rate12 may facilitate accumulation of extravascular fluid in the lung and depression of DM. The relation linking interstitial fluid volume (or membrane thickness) and DM in CHF is unknown.
Our hypothesis of an upregulation in sodium handling by the endothelium disagrees with previous data in the literature. Townsley et al3 have suggested that alveolar-capillary barrier remodeling in a canine model of pacing-induced heart failure constitutes a beneficial adaptation to the pulmonary venous hyperperfusion that accompanies heart failure. Kaplan et al18 documented a normal transcapillary escape rate of transferrin in patients with decompensated heart failure. Davies et al19 have shown a reduced transferrin escape. However, it is unclear whether the mechanisms for transferrin and sodium transport are the same, ACE inhibitors had a role8 in the findings of Davies et al, and pacing-induced heart failure in dogs duplicates CHF in humans.
The reasons for an upregulated pulmonary microvascular salt transport are unclear. The simplest explanation would be that of remodeling of the alveolar-capillary membranes, which occurs in CHF.3 6 Note that epidermal growth factor upregulates alveolar epithelial sodium transport,1 tumor necrosis factor-α mediates an upregulation of sodium and fluid transfer in a rat model,20 and cardiac glycosides are potent inhibitors of alveolar fluid clearance.21
Glucose, DLCO, and Its Subdivisions
Glucose in CHF patients elicited pulmonary changes that were the opposite of those elicited by saline, ie, reduction in VC and increase in DM, with a tendency of DLCO to augment. The inhibition of renin and aldosterone secretion and the slight reduction of Htc, Hb, and PPC with glucose do not support the interpretation that VC reduction was due to systemic water displacement from the intravascular to the extravascular phase and decline of circulating blood volume. Although not proven, the explanation of a redistribution of blood from the central compartment to the periphery seems more likely. Despite the diminished surface area for gas exchange, DM rose significantly with d-glucose infusion. Redistribution of blood from the lungs may have reduced pulmonary hyperemia and interstitial fluid, thus making the alveolar-capillary membrane thinner and more conductive. This interpretation implies that improvement in the membrane conductance was such as to fully compensate (DLCO tended to be raised) for the reduced volume of blood available for gas exchange. Another hypothesis may be that in heart failure the alveolar side of the diffusing membrane is involved in the disturbances in sodium handling and that, with the infusion of glucose, sodium was taken up from the alveolar lumen through a sodium-glucose cotransport system.22 23 24
In conclusion, although part of the proposed effects are speculation, results show that (1) saline, even in small amounts, reduces alveolar-capillary membrane diffusing capacity in CHF; (2) healthy subjects do not possess such a liability; and (3) glucose improves gas diffusion across the alveolar-capillary membrane in patients with CHF.
Because DM significantly correlates with the functional status of CHF patients,25 prospective studies are required to assess whether the opposite influence of salt and glucose on the alveolar-capillary membrane function may be of any clinical relevance and produce any effect on exercise performance in cardiac failure.
This study was supported in part by a grant from the National Research Council and the Monzino Foundation, Milan, Italy.
- Received May 14, 1999.
- Revision received June 8, 1999.
- Accepted July 26, 1999.
Matthay MA, Folkesson HG, Verkman AS. Salt and water transport across alveolar and distal airway epithelia in the adult lung. Am J Physiol. 1996;270:L487–L503.
Guyton AC, Lindsey AW. Effect of elevated left atrial pressure and decreased plasma protein concentration on the development of pulmonary edema. Circ Res. 1959;7:649–657.
Townsley MI, Fu Z, Mathieu-Costello O, West JB. Pulmonary microvascular permeability: responses to high vascular pressure after induction of pacing-induced heart failure in dogs. Circ Res. 1995;77:317–325.
West JB, Tsukimoto K, Mathieu-Costello O, Prediletto R. Stress failure in pulmonary capillaries. J Appl Physiol. 1991;70:1731–1742.
Guazzi M, Marenzi GC, Alimento M, Contini M, Agostoni PG. Improvement of alveolar-capillary membrane diffusing capacity with enalapril in chronic heart failure and counteracting effect of aspirin. Circulation. 1997;95:1930–1936.
Cotes JE. Lung Function. 4th ed. Oxford, England: Blackwell Scientific Publications; 1979:239–249.
Roughton FJW, Forster FE. Relative importance of diffusion and chemical reaction rates in determining rate of exchange of gases in human lung, with special reference to true diffusing capacity of blood in the lung capillaries. J Appl Physiol. 1957;11:290–302.
Volpe M, Rao MAE, Cuocolo A, Russo R, Nappi A, Mele FA, Enea I, Trimarco B, Condorelli M. Radionuclide monitoring of cardiac adaptations to volume loading in patients with dilated cardiomyopathy and mild heart failure: effects of angiotensin-converting enzyme inhibition. Circulation. 1995;92:2511–2518.
Gee MH, Spath JA Jr. The dynamics of the lung fluid filtration system in dogs with edema. Circ Res. 1980;46:796–801.
Gabel JC, Fallon KD, Laine GA, Drake RE. Lung lymph flow during volume infusions. J Appl Physiol. 1986;60:623–629.
Miserocchi G, Negrini D, Del Fabbro M, Venturoli D. Pulmonary interstitial pressure in intact in situ lung: transition to interstitial edema. J Appl Physiol. 1993;74:1171–1177.
Nixon JW, Murray RG, Bryant C, Johnson RL Jr, Mitchell JH, Holland OB, Gomez-Sanchez C, Vergne-Marini P, Blomquist CG. Early cardiovascular adaptation to simulated zero gravity. J Appl Physiol. 1979;46:541–548.
Begin R, Epstein M, Sackner MA, Levinson R, Dougherty R, Duncan D. Effects of water immersion to the neck on pulmonary circulation and tissue volume in man. J Appl Physiol. 1976;40:293–299.
Prisk GK, Guy HJB, Elliott AR, Deutschman RA III, West JB. Pulmonary diffusing capacity, capillary blood volume, and cardiac output during sustained microgravity. J Appl Physiol. 1993;75:15–26.
Garat C, Rezaiguia S, Meignan M, D’Ortho MP, Harf A, Matthay MA, Jayr C. Alveolar endotoxin increases alveolar liquid clearance in rats. J Appl Physiol. 1995;79:2021–2028.
Sakuma T, Pittet JF, Jayr C, Matthay MA. Alveolar liquid and protein tolerance in the absence of blood flow or ventilation in sheep. J Appl Physiol. 1993;74:176–185.
Harris P, Heath D. The interstitial space of the lung. The Human Pulmonary Circulation. 3rd ed. Edinburgh, Scotland: Churchill Livingstone; 1986:359–372.
Wangensteen D, Bartlett M. D- and L-glucose transport across the pulmonary epithelium. J Appl Physiol. 1984;57:1722–1730.
Khimenko PL, Barnard JW, Moore TM, Wilson PS, Ballard ST, Taylor AE. Vascular permeability and epithelial transport effects on lung edema formation in ischemia and reperfusion. J Appl Physiol. 1994;77:1116–1121.
Puri S, Baker L, Dutka DP, Oakley CM, Hughes JMB, Cleland JGF. Reduced alveolar-capillary membrane diffusing capacity in chronic heart failure: its pathophysiological relevance and relationship to exercise performance. Circulation. 1995;91:2769–2774.