Chemiluminescent measurements of nitric oxide pulmonary diffusing capacity and alveolar production in humans.

Measurements of nitric oxide (NO) pulmonary diffusing capacity (DL(NO)) multiplied by alveolar NO partial pressure (PA(NO)) provide values for alveolar NO production (VA(NO)). We evaluated applying a rapidly responding chemiluminescent NO analyzer to measure DL(NO) during a single, constant exhalation (Dex(NO)) or by rebreathing (Drb(NO)). With the use of an initial inspiration of 5-10 parts/million of NO with a correction for the measured NO back pressure, Dex(NO) in nine healthy subjects equaled 125 +/- 29 (SD) ml x min(-1) x mmHg(-1) and Drb(NO) equaled 122 +/- 26 ml x min(-1) x mmHg(-1). These values were 4.7 +/- 0.6 and 4.6 +/- 0.6 times greater, respectively, than the subject's single-breath carbon monoxide diffusing capacity (Dsb(CO)). Coefficients of variation were similar to previously reported breath-holding, single-breath measurements of Dsb(CO). PA(NO) measured in seven of the subjects equaled 1.8 +/- 0.7 mmHg x 10(-6) and resulted in VA(NO) of 0.21 +/- 0.06 microl/min using Dex(NO) and 0.20 +/- 0.6 microl/min with Drb(NO). Dex(NO) remained constant at end-expiratory oxygen tensions varied from 42 to 682 Torr. Decreases in lung volume resulted in falls of Dex(NO) and Drb(NO) similar to the reported effect of volume changes on Dsb(CO). These data show that rapidly responding chemiluminescent NO analyzers provide reproducible measurements of DL(NO) using single exhalations or rebreathing suitable for measuring VA(NO).

Simultaneous measurement of nitric oxide production by conducting and alveolar airways of humans.

Human airways produce nitric oxide (NO), and exhaled NO increases as expiratory flow rates fall. We show that mixing during exhalation between the NO produced by the lower, alveolar airways (VL(NO)) and the upper conducting airways (VU(NO)) explains this phenomenon and permits measurement of VL(NO), VU(NO), and the NO diffusing capacity of the conducting airways (DU(NO)). After breath holding for 10-15 s the partial pressure of alveolar NO (PA) becomes constant, and during a subsequent exhalation at a constant expiratory flow rate the alveoli will deliver a stable amount of NO to the conducting airways. The conducting airways secrete NO into the lumen (VU(NO)), which mixes with PA during exhalation, resulting in the observed expiratory concentration of NO (PE). At fast exhalations, PA makes a large contribution to PE, and, at slow exhalations, NO from the conducting airways predominates. Simple equations describing this mixing, combined with measurements of PE at several different expiratory flow rates, permit calculation of PA, VU(NO), and DU(NO). VL(NO) is the product of PA and the alveolar airway diffusion capacity for NO. In seven normal subjects, PA = 1.6 +/- 0.7 x 10(-6) (SD) Torr, VL(NO) = 0.19 +/- 0.07 microl/min, VU(NO) = 0.08 +/- 0.05 microl/min, and DU(NO) = 0.4 +/- 0.4 ml. min(-1). Torr(-1). These quantitative measurements of VL(NO) and VU(NO) are suitable for exploring alterations in NO production at these sites by diseases and physiological stresses.

Rate of nitric oxide production by lower alveolar airways of human lungs.

This report describes methods for measuring nitric oxide production by the lungs' lower alveolar airways (VNO), defined as those alveoli and bronchioles well perfused by the pulmonary circulation. Breath holding or vigorous rebreathing for 15-20 s minimizes removal of NO from the lower airways and results in a constant partial pressure of NO in the lower airways (PL). Then the amount of NO diffusing into the perfusing blood will be the pulmonary diffusing capacity for NO (DNO) multiplied by PL and by mass balance equals VNO, or VNO = DNO(PL). To measure PL, 10 normal subjects breath held for 20 s followed by exhalation at a constant flow rate of 0.83 +/- 0.14 (SD) l/s or rebreathed at 59 +/- 15 l/min for 20 s while NO was continuously measured at the mouth. DNO was estimated to equal five times the single-breath carbon monoxide diffusing capacity. By using breath holding, PL equaled 2.9 +/- 0.8 mmHg x 10(-6) and VNO equaled 0.39 +/- 0.12 microl/min. During rebreathing PL equaled 2.3 +/- 0.6 mmHg x 10(-6) and VNO equaled 0.29 +/- 0.11 microl/min. Measurements of NO at the mouth during rapid, constant exhalation after breath holding for 20 s or during rebreathing provide reproducible methods for measuring VNO in humans.

Determination of production of nitric oxide by lower airways of humans--theory.

Exercise and inflammatory lung disorders such as asthma and acute lung injury increase exhaled nitric oxide (NO). This finding is interpreted as a rise in production of NO by the lungs (VNO) but fails to take into account the diffusing capacity for NO (DNO) that carries NO into the pulmonary capillary blood. We have derived equations to measure VNO from the following rates, which determine NO tension in the lungs (PL) at any moment from 1) production (VNO); 2) diffusion, where DNO(PL) = rate of removal by lung capillary blood; and 3) ventilation, where V A(PL)/(PB - 47) = the rate of NO removal by alveolar ventilation (V A) and PB is barometric pressure. During open-circuit breathing when PL is not in equilibrium, d/dt PL[V(L)/ (PB - 47)] (where V(L) is volume of NO in the lower airways) = VNO - DNO(PL) - V A(PL)/(PB - 47). When PL reaches a steady state so that d/dt = 0 and V A is eliminated by rebreathing or breath holding, then PL = VNO/DNO. PL can be interpreted as NO production per unit of DNO. This equation predicts that diseases that diminish DNO but do not alter VNO will increase expired NO levels. These equations permit precise measurements of VNO that can be applied to determining factors controlling NO production by the lungs.

Stabilization of lung surfactant particles against conversion by a cycling interface.

The large active particles of pulmonary surfactant are depleted in patients with acute respiratory distress syndrome and in animal models of this disorder. We studied in vitro conversion of large to small particles, separated by differential sedimentation, to determine how factors lavaged from rabbits injured by intravenous oleic acid would affect conversion. In half-filled test tubes rotated end over end, samples from injured animals increased the recovery of large particles from 40 +/- 6% of uncycled samples for controls to 62 +/- 21%. We hypothesized that proteins in the injured samples, and perhaps also the proteinase inhibitors used previously to block conversion (N. J. Gross and R. M. Schultz. Biochim. Biophys. Acta 1044: 222-230, 1990), stabilized surfactant particles by limiting access to the cycling interface. Hemoglobin, neutrophil elastase, and alpha1-antiproteinase (alpha1-PI) oxidized to eliminate its antiproteinase activity all stabilized large particles against conversion. Hemoglobin was most effective, increasing recovery from 18 +/- 5% for controls to 86 +/- 5% with 0.4 mg/ml hemoglobin. Native alpha1-PI had no effect on conversion. Our results suggest that acceleration of normal conversion is unlikely to explain the depletion of large particles in injured lungs. They also suggest that conversion of surfactant particles separated by differential sedimentation requires no proteinase susceptible to inhibition by alpha1-PI. They provide an alternate hypothesis related to interfacial effects rather than proteinase inhibition for the previously reported effect of alpha1-PI on conversion of particles separated according to density.

Changes in subphase aggregates in rabbits injured by free fatty acid.

Rabbits treated with intravenous free fatty acid suffer an acute lung injury. Material obtained from these lungs by bronchoalveolar lavage (BAL) has dramatically impaired ability to lower surface tension in vitro despite normal levels of surfactant phospholipids. Although large quantities of surface-active inhibitors are present in BAL, their effects are not sufficient to explain the magnitude of surfactant inactivation. This study determines if alterations in the surfactant aggregates can explain the loss of surfactant function in fatty acid lung injury. In injured animals, the larger, most active surfactant particles recovered by centrifugal pelleting were found to be decreased in amount. The remaining large particles had reduced surface activity compared with control aggregates. In addition, large particles in injured animals had a higher density than control animals on sucrose gradients following equilibrium centrifugation. Interaction with serum components present in the injured BAL could explain these higher densities. The ability of the injured BAL to lower surface tension was improved by supplementation with normal levels of particles from injured lungs. Supplementation of injured BAL with control large aggregates improved activity further and restored the ability to lower surface tension to < 1 mN/m. Therefore both the decreased amount and activity of large surfactant aggregates in injured animals contributed significantly to the observed inactivation of surfactant. Diminished surfactant function from alteration in surfactant aggregates is a mechanism common to other forms of acute lung injury, and the design of therapies with exogenous surfactants in injured lungs will need to consider strategies that restore surfactant function towards normal.

Inhibition of pulmonary surfactant by oleic acid: mechanisms and characteristics.

The inhibitory effects of oleic acid (OA) on the surface activity of pulmonary surfactant were characterized by use of the oscillating bubble surfactometer, the Wilhelmy balance, and excised rat lungs. Oscillating bubble studies showed that OA prevented lavaged calf surfactant [0.5 mM phospholipid (PL)] from lowering surface tension below 15 mN/m at or above a molar ratio of OA/PL = 0.5. In contrast to inhibition of surfactant by plasma proteins, increasing the surfactant concentration did not eliminate inhibition by oleic acid, which occurred at OA/PL greater than 0.67 on the oscillating bubble even at surfactant concentrations of 1.5 and 12 mM PL. Studies of surfactant adsorption showed that preformed films of OA had little effect on the adsorption of pulmonary surfactant. Wilhelmy balance studies showed that OA did interfere with the ability of spread films of surfactant to reach low surface tensions during dynamic compression. Further balance experiments with binary films of OA and dipalmitoyl phosphatidylcholine showed that these compounds were miscible in surface films. Together these findings suggested that OA inhibited pulmonary surfactant activity by disrupting the rigid interfacial film responsible for the generation of very low surface tension during dynamic compression. Mechanical studies in excised rat lungs showed that instillation of OA gave altered deflation pressure-volume characteristics with decreased quasi-static compliance, indicating disruption of pulmonary surfactant function in situ. This alteration of mechanics occurred without major changes in the composition of lavaged PLs or in the tissue compliance of the lungs defined by mechanical measurements during inflation-deflation with saline.(ABSTRACT TRUNCATED AT 250 WORDS)

Effect of pattern of aerosol inhalation on clearance of technetium-99m-labeled diethylenetriamine pentaacetic acid from the lungs of normal humans.

The clearance rate of inhaled aerosols of technetium-99m-labeled diethylenetriamine pentaacetic acid (99mTc-DTPA) from the lungs provides a rapid, clinically useful, noninvasive index of pulmonary epithelial permeability. In order to identify a method that minimizes intrasubject and intersubject variability and thereby provides a reliable means to identify patients with abnormal values, we administered a submicronic aerosol of 99mTc-DTPA to 10 healthy, nonsmoking male subjects with either tidal breathing (Vtidal) or multiple vital capacity maneuvers (VVC). Subjects then spontaneously breathed room air while counting continued for 30 min. Monoexponential clearance rates over 7, 15, and 30 min were compared with a two-compartment, biexponential analysis over 30 min. Intrasubject reproducibility was evaluated by repeating clearance 2 to 156 days later. Monoexponential clearance following VVC at 30 min equaled 1.36 +/- 0.55%/min compared with 0.83 +/- 0.25%/min for Vtidal (p less than 0.025). VVC inhalations resulted in a larger fast compartment of 16 +/- 12% compared with 3 +/- 2% with tidal breathing (p less than 0.01). The least intrasubject variability with coefficient of variation (CV) of +/- 18% was obtained with monoexponential analyses after Vtidal during 15 min of scanning and with either breathing maneuver over 30 min. Monoexponential clearance for 30 min with Vtidal gave the least scatter between subjects, with CV of +/- 30%. These data show that simple tidal inhalations of 99mTc-DTPA followed by a monoexponential analysis of the 30-min time-activity curve from both lungs minimize the degree of variability between and among subjects and provide a predicted normal value of clearance of 0.83 +/- 0.25%/min. The development of a more rapid curvilinear clearance followed by delivery VVC suggests that several deep breaths transiently increase epithelial permeability or reduce the volume of liquid in the alveolar subphase in some regions. Resting for 20 min prior to inhaling the aerosol of 99mTc-DTPA is recommended to avoid alterations in clearance rates from deep breathing.

Altered function of pulmonary surfactant in fatty acid lung injury.

To determine whether acute fatty acid lung injury impairs pulmonary surfactant function, we studied anesthetized ventilated rabbits given oleic acid (55 mg/kg iv, n = 11) or an equivalent volume of saline (n = 8). Measurements of pulmonary mechanics indicated a decrease in dynamic compliance within 5 min of injury and a decrease in lung volume that was disproportionately large at low pressures, consistent with diminished surfactant activity in vivo. Bronchoalveolar lavage fluid obtained 1 h after injury had significantly increased erythrocytes and total leukocytes, largely polymorphonuclear cells. The phospholipid content and composition of the cell-free fraction had only minor changes from those of controls, but the protein content was increased 35-fold. Measurements of lavage surface activity in vitro showed an increase in average minimum surface tension from 1.3 +/- 0.4 (SE) dyn/cm in controls to 20.2 +/- 3.9 dyn/cm in injured animals. The alterations in static pressure-volume curves and decrease in lavage surface activity suggest a severe alteration of surfactant function in this form of lung injury that occurs despite the presence of normal amounts of surfactant phospholipids.

Assessment of rebreathing O2 consumption in humans with normal and diseased lungs.

We investigated sources of error in estimating steady-state O2 consumption (VO2ss) by calculating O2 uptake from an anesthesia bag containing O2, He, and N2 during 10-20 s of rebreathing (VO2rb). In 11 normal resting subjects, VO2rb calculated with end-tidal sampling overestimated VO2ss by 16 +/- 15% (SD) (P less than 0.003). This error was proportional to the increase in pulse rate during rebreathing, so that pulse-corrected VO2rb slightly underestimated VO2ss by 2.1 +/- 12.2% (P = 0.66) in the six subjects who rebreathed 28% O2 in the rebreathing bag but significantly underestimated VO2ss by 7.5 +/- 6.7% (P less than 0.04) in the six subjects who rebreathed 21% O2 in the rebreathing bag. During exercise, VO2rb underestimated VO2ss by 4 +/- 12% (P less than 0.001) and by 7 +/- 6% at O2 consumptions greater than 2,000 ml/min if O2 in the rebreathing bag was kept above 20% throughout rebreathing. We found that VO2rb calculated with end-tidal gas concentrations underestimated VO2ss by 1-43% in patients with moderate-to-severe obstructive lung disease, with even greater errors when mixed expired samples were used. The magnitude of the discrepancy correlated poorly with abnormalities in standard pulmonary function tests. Based on these data, VO2rb closely approximates VO2ss in normal subjects, provided hypoxia during rebreathing is avoided and cardiac acceleration from rebreathing is taken into account during resting measurement.

Lung weight in vivo measured with computed tomography and rebreathing of soluble gases.

In nine anesthetized dogs, accuracy of noninvasive measurements of lung weight (W) and gas volume in vivo was determined from volume and density determined by computed tomography (CT) and by rebreathing helium and the soluble gases dimethyl ether (WDME) and acetylene (WC2H2). Reference standards were obtained from the postmortem scale weight of the frozen lungs (Wscale) and compared with the CT lung weights measured in the living dog (WCT-38) and the frozen carcass (WCT-cold). WCT-cold did not significantly differ from Wscale [-2 +/- 9% (SD), P = 0.7]. WCT-cold was 10% greater than WCT-38 (0.10 greater than P greater than 0.05), suggesting an increase in lung weight despite immediately commencing freezing after death. WDME measured 64 +/- 6% and WC2H2 56 +/- 12% of WCT-38. Serial multiple measurements in three dogs over 14 wk showed a coefficient of variation (CV) of 10 +/- 2% for WDME, 18 +/- 2% for WC2H2, 4.1 +/- 0.9% for WCT, 2.6 +/- 0.8% for CT density, and 3.5 +/- 1.6% for functional residual capacity (FRC) by CT. FRC calculated from CT consistently underestimated FRC measured by rebreathing helium by 18 +/- 8% (P less than 0.005). This error, despite good agreement between WCT and Wscale, was explained by underestimation of CT total lung volume and overestimation of lung density by factors known to affect CT readings, such as partial volume effects, beam hardening, and limited number of input signals. These data show that CT scanning can provide serial measurement of the mass, density, and volume of the lungs with a CV in the order of 5%, but the rebreathing of soluble gases gives more than double this variability. Measurements of WDME performed on the same day had a CV of 3 +/- 1%, so that WDME provides a precise noninvasive means to measure lung weight in acute studies.

Effects of posture on stimulated ventilation in quadriplegia.

Quadriplegics are able to compensate for alterations of operational length of the diaphragm by reflexly increasing neural drive to the diaphragm. This increase in neural drive is adequate to maintain required tidal volume and minute ventilation during quiet breathing in these patients with limited inspiratory muscle function. It is not known, however, if this neural compensation is sufficient to preserve ventilation when the diaphragm is stressed by simultaneously changing its operational length and increasing ventilatory demands. This issue was explored in 7 quadriplegics whose vital capacity was reduced to 15 to 53% of predicted. The diaphragm was stressed by shortening its length from the supine to a 60 degree tilted position, and also by inducing hyperventilation by having the subjects rebreathe 7% CO2. Response to this stress was recorded by monitoring the ventilatory response to rebreathing CO2 (delta VE/delta PCO2), and also by measuring mouth pressure 0.1 s after occluding the airway at the start of inspiration during CO2 rebreathing (delta P0.1/delta PCO2). A change from the supine to the tilted position caused an increase in resting end-expiratory volume of 0.8 +/- 0.2 L (SD) and therefore shortened the diaphragm. Despite this shortening of diaphragm length and the stress of CO2 rebreathing, there was no significant change in delta VE/delta PCO2 and delta P0.1/delta PCO2 with changes in posture. The delta VE/delta PCO2 was 0.82 +/- 0.42 L/min/mm Hg supine versus 0.95 +/- 0.65 L/min/mm Hg when tilted. The delta P0.1/delta PCO2 was 0.18 +/- 0.08 cm H2O/mm Hg supine versus 0.20 +/- 0.10 cm H2O/mm Hg tilted.(ABSTRACT TRUNCATED AT 250 WORDS)

Cardiac output by rebreathing in patients with cardiopulmonary diseases.

Noninvasive estimates of cardiac output by rebreathing soluble gases (Qc) can be unreliable in patients with cardiopulmonary diseases because of uneven distribution of ventilation to lung gas volume and pulmonary blood flow. To evaluate this source of error, we compared rebreathing Qc with invasive measurements of cardiac output performed by indicator-dilution methods (COID) in 39 patients with cardiac or pulmonary diseases. In 16 patients with normal lung volumes and 1-s forced expiratory volumes (FEV1), Qc measured with acetylene [Qc(C2H2)] overestimated COID insignificantly by 2 +/- 9% (SD). In subjects with mild to moderate obstructive lung disease, Qc(C2H2) slightly overestimated COID by 6 +/- 15% (P = 0.11). In patients with restrictive disease or combined obstructive and restrictive disease, Qc(C2H2) underestimated COID significantly by 9 +/- 14% (P less than 0.04). The magnitude of the discrepancy between Qc and COID correlated with size of the volume rebreathed and an index of uneven ventilation calculated from helium mixing during rebreathing that determined a dead space to inspired volume ratio (VRD/VI). Rebreathing volumes less than 40% of the predicted FEV or VRD/VI of 0.4 or greater identified all subjects with a discrepancy between Qc(C2H2) and COID of 20% or greater.

Chemical breakdown of technetium-99m DTPA during nebulization.

Aerosols of 99mTc diethylenetriaminepentaacetic acid ([99mTc]DTPA) used for measuring lung permeability and lung ventilation require a radioaerosol delivery system to produce an aerosol with reproducible size and radiochemical purity. To test how well nebulizers meet this requirement, radiochemical purity of aerosols produced with a jet and an ultrasonic nebulizer was evaluated. The activity median aerodynamic diameter (AMAD) and geometric standard deviation (sigma g) of radioaerosols were 0.46 micron (sigma g = 1.6) for the jet nebulizer and 0.70 micron (sigma g = 1.7) for the ultrasonic nebulizer. Paper and liquid chromatographic assays were obtained on the [99mTc]DTPA aerosol solute produced with each nebulizer. The results of these tests showed major differences in radiochemical purity. Aerosols produced in the jet nebulizer consistently showed greater than 90% of the radioactivity bound to the DTPA ligand whereas aerosols produced in the ultrasonic nebulizer showed less than 10% of the radioactivity bound to DTPA. The results support the need to test radiochemical purity of aerosols before using an aerosol nebulizer for pulmonary imaging and clearance studies.

Bronchial and alveolar absorption of inhaled 99mTc-DTPA.

The clearance rate of 99mTc-DTPA deposited in the lung by inhalation has been used as an index to measure lung epithelial permeability. To determine if differences exist between alveolar and bronchial absorption of 99mTc-DTPA we measured regional clearance rates in 4 beagle dogs for 30 min after preferential bronchial and alveolar deposition. Alveolar deposition was maximized by inhalation for 2 min of small 99mTc-DTPA particles (activity median aerodynamic diameter, AMAD = 0.5 micron; geometric standard deviation, GSD = 1.6) with deep slow ventilation (VT = 350 ml; f = 9 min-1), and bronchial deposition was increased by inhalation of large particles (AMAD = 4.1 microns, GSD = 2.3) with rapid shallow ventilation (VT = 50 ml; f = 65 min-1). Respective clearance rates from basal regions, which represent mainly alveolar absorption, were: for small particles, 2.29% min-1; for large particles, 1.57% min-1 (p = 0.10). Apical regions, which contain relatively more bronchial surface than do the basal regions, showed the following clearance rates: for small particles, 1.76% min-1; for large particles, 1.31% min-1 (p less than 0.05). These results indicate that in vivo alveolar absorption of 99mTc-DTPA is more rapid than bronchial absorption. Control or verification of the site of deposition of the tracer in the lung is of importance for the interpretation of the results of the 99mTc-DTPA lung permeability assay.

Intermittent positive pressure breathing in patients with respiratory muscle weakness. Alterations in total respiratory system compliance.

Intermittent positive pressure ventilation (IPPB) is reported to improve lung compliance and decrease the work of breathing in subjects with kyphoscoliosis. These results suggest that IPPB may improve chest wall and lung compliance in patients with neuromuscular disease. We studied the short-term effects of IPPB on total respiratory system compliance in 14 subjects with neuromuscular disease. Seven were quadriplegics, and seven had muscular dystrophy. Vital capacity was reduced to 38 +/- 14 percent of the predicted normal values. Baseline measurements of total respiratory system compliance were 57 +/- 18 percent when compared to normal control values. After a 20 minute treatment of IPPB delivered with inspiratory pressures of 20 to 25 cm H2O that more than tripled resting tidal volume, there was no significant change in total respiratory system compliance in either group of patients. These findings indicate that patients with quadriplegia or muscular dystrophy do not derive immediate improvement in ventilatory mechanics from IPPB treatments.

Lung fluid balance during acute renal failure in sheep.

To determine whether uremia changes lung vascular permeability, we measured the flow of lymph and proteins from the lungs of acutely uremic sheep. Acute renal failure was induced by either bilateral nephrectomy or by reinfusing urine. Both models of renal failure increased the plasma creatinine from 0.8 +/- 0.3 to 11 +/- 1 mg/dl in 3 days but caused no significant change in the flow of lymph from the lungs. To determine whether uremia increased the protein clearance response to elevated pulmonary microvascular pressures, we inflated a balloon in the left atrium for 2 h before and 3 days after inducing acute renal failure. In seven sheep, before removing the kidneys, the 20 cmH2O elevation of left atrial pressure increased the protein clearance 3.9 +/- 3.0 ml/h (from 9.5 +/- 4.9 to 13.4 +/- 5.4 ml/h). Three days after the bilateral nephrectomy the same increase in left atrial pressure increased the protein clearance 6.4 +/- 3.6 ml/h (from 6.1 +/- 2.1 to 12.5 +/- 5.2 ml/h), which was a significantly larger increase than that measured before the nephrectomy (P less than 0.05). Sham nephrectomy in seven sheep caused the protein clearance response to the elevated left atrial pressure to fall from 4.7 +/- 1.9 ml/h before the sham nephrectomy to 2.6 +/- 1.4 ml/h 3 days later (P less than 0.05). Uremia due to reinfusion of urine in five sheep did not affect the protein clearance response to elevations in left atrial pressure. Neither model of acute uremia increased the postmortem extravascular lung water volume.(ABSTRACT TRUNCATED AT 250 WORDS)

Measurement of lung gas volume and regional density by computed tomography in dogs.

To determine if computed tomography (CT) can accurately measure lung volume, we compared lung gas volume measured by helium dilution with the equivalent volume calculated from CT total lung volume and density in 13 supine dogs. CT lung gas volume underestimated helium volume by 34% (range: -63 to 0%). Studies of wooden lung phantoms varying in density from 0.082g/cc to 0.776g/cc showed that only 15% of this error could be mimicked by the phantoms. The rest of the discrepancy is attributed to the lung's irregular borders, and the sharp density gradients surrounding and within the lung that result in x-ray beam hardening, sampling limitations, and partial volume measurement errors. Serial biweekly measurements in three dogs for 14 weeks showed CT gas volume to be highly reproducible with less scatter than seen in the helium measurements. Density in the lungs of all dogs showed a uniform gradual decrease from approximately 0.60g/cc at the dependent surface to 0.20g/cc at the superior surface with relatively constant density at any horizontal level. These studies show that whereas CT underestimates gas volume in the lungs, serial measurements are highly reproducible in experimental studies and are a promising technique to monitor diseases or response to therapy. Density gradients in the lungs were sufficiently uniform so that disruption of the normal gradient may be an indicator of early lung disease.