OVERDUE FINES ARE 25¢ PER DAY PER ITEM Return to book drop to remove this checkout from your record. THE EFFECT OF PULMONARY VENOUS PRESSURE ON COLLATERAL VENTILATION IN DOGS By Steven D. Fuller A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Physiology 1979 ABSTRACT THE EFFECT OF PULMONARY VENOUS PRESSURE ON COLLATERAL VENTILATION IN DOGS By Steven D. Fuller Steady state collateral resistance (Rcoll), residual volume (RV) and lung compliance (C) was measured in an iso- lated perfused dog lung (IPL) perfused with blood from a donor dog. A double lumen catheter was inserted through the mainstem bronchus and wedged in a peripheral airway, isolat- ing a segment of lung. Rcoll was calculated as: Rcoll = —————P§ - PL Vcoll where PS is pressure at the catheter tip as measured by the inner lumen, PL is lung inflation pressure and Vcoll is the flow of 95% O2 and 5% CO2 (95/5), room air (21/0) or 100% O2 (100/0) through the outer lumen and into the isolated segment. When compared to 21/0, 95/5 decreased Rcoll while Rcoll with 100/0 was not consistently different. Elevating pulmonary venous pressure when both the lung and segment were ventilated with 95/5 decreased Rcoll, increased RV, did not affect C at 50% total lung capacity (TLC) and decreased C at 90% TLC. ACKNOWLEDGEMENTS To be a major professor and advisor is a duty. To be a friend is an act of kindness which extends far beyond duty. During the past two and one-half years Dr. N. E. Robinson has served in both capacities, and it is to him with his qualities of insight, understanding and cheerful leadership that I express my unending gratitude. I would like to thank Dr. J. B. Scott and Dr. Robinson for their guidance and assistance in my work with the iso- lated perfused lung preparation. Deepest appreciation is extended to these gentlemen as well as to Dr. J. R. Hoffert for their valuable criticism during the preparation of this thesis. A special thanks to Bobbi Milar, whose interminable tolerance of a struggling graduate student shall not be forgotten. ii TABLE OF CONTENTS LIST OF TABLES ...................................... LIST OF FIGURES ..................................... LIST OF ABBREVIATIONS ............................... I. INTRODUCTION AND LITERATURE REVIEW .............. A. Pathways for Collateral Ventilation ....... ... 1. Pores of Kohn ............................. 2. Canals of Lambert (bronchiolealveolar communications) ........................... 3. Martin's Channels (interbronchiolar respiratory bronchioles) ........ . ......... B. Some Factors Affecting Collateral Ventilation 1. Lung Volume and Volume History ............ 2. Pharmacologic Agents ...................... a. Histamine .............. ..... ........... b. Acetylcholine .......................... c. Epinephrine and ISOproterenol .......... 3. Alveolar and Blood Gas Tensions ........... a. Carbon Dioxide .............. . .......... b. Oxygen ................................ 4. Pulmonary Vessels and Vascular Pressures. 11. MATERIALS AND METHODS ........... . ............... A. Isolated Perfused Lung Preparation (IPL) ..... B. Measurement of Steady State Collateral Resistance ............... . ............. . ..... . Response of the Isolated Perfused Lung to Hypocapnea and Hyperoxia ..................... iii Page vi vii 36 38 TABLE OF CONTENTS--continued - Page D. The Effect of Pulmonary Venous Pressure on Steady State Collateral Resistance, Residual Volume and Lung Compliance .................. 39 E. Analysis of Data ............................ 40 1. Response of the Isolated Perfused Lung to Inspired Hypocapnea and Hyperoxia.. ...... 40 2. The Effect of Pulmonary Venous Pressure on Steady State Collateral Resistance.... 41 3. The Effect of Pulmonary Venous Pressure on Residual Volume ....................... 42 4. The Effect of Pulmonary Venous Pressure on Lung Compliance ....................... 42 S. The Effect of Pulmonary Venous Pressure on Pulmonary Vascular Resistance ......... 43 III. RESULTS ...................... . ....... ... ........ 44 A. Response of the Isolated Perfused Lung to Hypocapnea and Hyperoxia ..... . ........ . ...... 44 B. The Effect of Pulmonary Venous Pressure on Steady State Collateral Resistance ........... 50 C. The Effect of Pulmonary Venous Pressure on Residual Volume .............................. 61 D. The Effect of Pulmonary Venous Pressure on Lung Compliance .............................. 61 E. The Effect of Pulmonary Venous Pressure on Pulmonary Vascular Resistance ................ 61 IV. DISCUSSION ...................................... 67 V. SUMMARY AND CONCLUSIONS ......................... 82 LIST OF REFERENCES ............... . .................. 83 APPENDICES A. TYPICAL DATA SHEETS .......................... 89 B. ANOVA TABLES AND RESULTS OF COMPARISON OF MEANS OF ALL DATA ............................ 92 iv LIST OF TABLES TABLE 1. Mean pulmonary arterial and'pulmonary venous CO2 tensions for Group I in three of the four experiments ................. . ................... 2. Average isolated perfused lung weight before and after the experiments, weight gain and flow rate ............................................ 3. Results of a_paired t test showing1 the average difference (D) in Rcoll (mm Hg ml'1 min x 10 at adjacent va' 5 (mm Hg), the standard devia- tion of the difference (SD), the calculated t value (P‘<. 05, * indicateD significant differ- ences) and the degrees of freedom (df) for the grouped PL's ........... ... .. ..... .............. 4. The change in Rvas (mm Hg-ml-1-min) between control va's (va = 5 mm Hg) and elevated va's at PL = 0, 2 and 4 cm H20 ....................... Page 49 SO 54 66 FIGURE 1. LIST OF FIGURES A diagram of the isolated perfused lung prep- aration (IPL) ................................. . The effect of 95% Oz and 5% CO (95/5), room air (21/0) and 100% O (100/0) on steady state collateral resistance (Rcoll) at two lung inflation pressures (PL = 2 and 4 cm H20) ..... Data from Figure 2 is regrouped to show the effect of time on steady state collateral resistance (Rcoll) in the isolated perfused lung (IPL) .................................... . The effect of pulmonary venous pressure (va) on steady state collateral resistance (Rcoll) in two representative dogs (each graph desig- nates a single dog) ........................... Steady state collateral resistance (Rcoll) at elevated pulmonary venous pressures (va) is plotted against Rcoll at control va's (va = 5 mm Hg) ............................... . The effect of pulmonary venous pressure (va) on residual volume (RV) in the isolated per- fused lung (IPL)... ........................... . The effect of pulmonary venous pressure (va) on the compliance of the isolated perfused lung (IPL) at two different lung volumes ...... . A schematic diagram showing a possible mechan- ism by which elevated pulmonary venous pres- sure (va) decreases steady state collateral resistance (Rcoll) ............................ Schematic curves showing two different rela- tionships between steady state collateral resistance (Rcoll) and lung volume ............ vi Page 35 46 48 52-53 56-60 63 65 74 79 ESGV FRC VC RV TLC PaCO ACO IPL 95/5 21/0 100/0 LIST OF ABBREVIATIONS Exhaled segmental gas volume Functional residual capacity Vital capacity Residual volume Total lung capacity Arterial CO2 tension Alveolar CO2 tension Arterial O2 tension Alveolar OZ tension Isolated perfused lung Pulmonary venous pressure Pulmonary arterial pressure Systemic arterial pressure Positive end expiratory pressure Steady state collateral resistance Pressure in an isolated lung segment measured at the tip of the wedged catheter Lung inflation pressure measured at the mainstem bronchus Flow of gas through the wedged catheter and into the isolated segment 95% O and 5% O 2 room air 2 100% O2 vii LIST OF ABBREVIATIONS-~continued CL Lung compliance between forty and fifty percent vital capacity CH Lung compliance between ninety and one hundred percent vital capacity 0 Blood perfusion rate of the isolated perfused lung Rvas Pulmonary vascular resistance viii I. INTRODUCTION AND LITERATURE REVIEW Prior to 1930 the lung was thought to consist of a series of dichotomously branching airways terminating in alveoli with no anastomoses between adjacent lung lobules. This anatomical arrangement was considered to explain atelec- tasis produced by bronchial obstruction (61). Beginning in 1930 VanAllen, Lindskog and associates demonstrated anatomic connections between adjacent lung lobules within the same lobe of human, cat, dog and rabbit lungs. Their conclusions were based on three observations: 1) atelectasis occurred following lobar but not lobular obstruction (62), 2) an atelectatic lobe could be reinflated via a cannula wedged in a lobular airway (59, 63, 64, 65, 66) and 3) emphysema could be produced on the lobar but not the lobular scale by allow- ing gas flow into the lobe on inflation but not out on deflation (60). The findings of these investigators caused rejection of the older theory of the lung as a series of dead-end pathways. Their experiments are the foundation of today's research involving collateral ventilation. A. Pathways for Collateral Ventilation l. Pores of Kohn Pores of Kohn occur in alveolar walls, have a diameter of 2-10 u and have been described in virtually all mammalian species (32). In early life they are uniformly distributed over the lung. In middle and later life they are present in largest numbers in the borders and apices of the upper and lower lobes (35). Interalveolar pores were described as early as 1847 (l), but their existence in the normal lung was debated for the next 100 years. Miller (1, 41) denied their presence in the normal lung and believed the shedding of alveolar epithelium explained the findings of Kohn (1) who traced fibrin threads through alveolar pores in cases of pneumonia. Martin (36) suggested that pores might be sites in alveolar walls previously occupied by inactive phagocytes which migrate during disease, leaving a pore. There are three points against this argument: 1) phagocytes may simply be passing through a pore rather than causing its formation. 2) In the manatee the pores are approximately 50 p in diameter (33), much larger than an alveolar phago- cyte which is approximately 10 p in diameter. 3) Loosli (32) infected monkeys and dogs intratracheally with type I pneomococci and found that lung septa of recovered animals have the same appearance as in non—infected lungs. Caradonna, according to Loosli (32), obtained two groups of guinea pigs: the first group was kept in a quiet environment while the second group was violently agitated several times per day. Pores appeared in the former group at 14-15 months of age while in the latter group pores appeared at 3-4 months of age and increased in number there- after. Caradonna concluded that pores are a result of the "excessive effort and dilatation of alveolar walls during the act of respiration." Because of the regularity of pore shape, Ogawa (44) believed pores are normally found in the lung and are not the result of histological fixation. Loosli (32) further suggests the pores to be avenues through which disease organisms may pass between adjacent lung lobules. VanAllen, Lindskog and colleagues were the first to demonstrate some type of anatomical connection between adjacent lung lobules. Because the Pores of Kohn were the only pathways of this type known at that time, they were implicated as being the sole route for collateral ventila— tion. It is likely, however, these investigators were measuring gas flow through other collateral channels more so than through the pores because the resistance of the pores would require much greater driving pressures than those measured to initiate and maintain gas flow. In their early experiments, VanAllen et a1. (64) placed a freshly excised lung lobe from a dog in an airtight chamber and connected the lobar bronchus and the first side branch of this bronchus to the outside with tubes. Air was withdrawn from the chamber just to the point where the lung was uni- formly expanded. The tip of one tube was submerged in water while air was run into the other tube. When pressures in the latter reached 1.0 cm H20, there was bubbling in the water at the tip of the submerged tube. When the lung was completely deflated, 6.0 cm H20 was required to initiate air passage and 2.0 cm H20 to maintain it. It is quite likely in these experiments that the bubbling which occurred at the tip of the submerged tube was from air which passed through other collateral channels and very little, if any, from air passing through the pores. This is suggested by the the following evidence: the proximal tip of the sub- merged catheter was 0.5 cm beneath the surface of the water (31), so the effective driving pressure needed to maintain the bubbling was really not 1.0 cm H20 but rather 0.5 cm H20: likewise, when the lung was deflated the effective driving pressure would actually be 1.5 cm H20. This sug- gests the pores provide a very minimal resistance to air flow. The following evidence, however, indicates this may not be true: 1) the pores are probably closed at very low lung volumes. This may be the reason why some of the early investigators were not able to find pores in the normal lung since their fixations were performed with the lungs in the collapsed state (28, 35). Also, Kuno and Staub (28) found that artificial microholes (25 and 51 u in diameter) pro- duced in cat lungs are closed at zero transpulmonary pres- sure. 2) Martin (37) calculated the Opening pressure of an alveolar pore with a surface tension of 47 dynes/cm to be extremely high. The driving pressures in VanAllen's experi- ments (64) were much too low to be attributed solely to the Pores of Kohn. Since, according to LaPlace's law (P = ZST/r, where ST is the surface tension and r is the cross-sectional radius of the pore), the calculated Opening pressures become much smaller with a larger radius, larger collateral path- ways must have been involved. 2. Canals of Lambert (bronchiole- alveolar communications) In 1953 Lambert (29) and later Duguid and Lambert (14) described occasional interruptions in the walls of terminal and respiratory bronchioles in which the lumen of the bron- chiole communicated directly with the surrounding alveoli. These bronchiole-alveolar communications are up to 30 u in diameter and have been found in cats, rats, humans and sheep (27). They become more prominent in certain pathologic con- ditions and may not be affected by lung volume (29). It is difficult from Lambert's original paper to ascer- tain how frequently these pathways occur. However, she made a tracing of several bronchioles taken from 500 serial sec- tions in human lung and found 32 canals in a 1.0 x 1.5 mm area. 3. Martin's Channels (interbronchiolar respiratory bronchioles) The third collateral channel was described by Martin (37) in 1966. He excised the left upper lobe from four dog lungs and cannulated the first two bronchial branches. By perfusing polystyrene spheres 60-710 u in diameter suspended in saline into one of the cannulae and collecting the out- flow through the other cannula, it was possible to pass spheres as large as 120 u in diameter through collateral pathways. Next, he passed India ink aerosol through the catheter system and noted ink deposits on respiratory bron- chioles which connected two terminal bronchioles. From these studies he suggested that collateral flow may occur through these respiratory bronchioles. Martin's results raise the following question: are the channels through which the spheres pass the same connecting channels in which the ink was deposited? The respective sizes of the respiratory bronchioles and the spheres suggest the answer is no. The diameter of the respiratory bron- chioles is normally in the range of 470—540 u (in human lungs fixed at about 75% total lung capacity) (10), so why were spheres no larger than 120 u found in the outflow of Martin's experiments? Raskin and Herman (48) describe interacinar communications 200 u in diameter in the human lung fixed at a transpulmonary pressure of 25 cm HZO' If these lungs were fixed at a somewhat lower transpulmonary pressure, the diameter Of the channels would decrease and may approach the size of the spheres in Martin's outflow. The distribution Of Martin's interbronchiolar respira- tory bronchioles throughout the lung is unknown since he only studied upper lobes. But in this region, using serially sectioned lung samples, he traced a collateral channel from a terminal bronchiole in one segment to an adjacent segment and found it 1.56 mm in length. In the 1600 serial sections which he examined, he found twenty-two collateral respiratory bronchioles. B. Some Factors Affecting Collateral Ventilation 1. Lung Volume and Volume History The effect Of lung volume on collateral flow was first indicated by VanAllen et a1. (64) in which a driving pres- sure Of 1.0 cm H20 was needed to pass air collaterally in slightly inflated lungs whereas a driving pressure of 2.0 cm H20 was needed in lungs which were collapsed. Lindskog and Bradshaw (31) passed a cannula with a dilatable tip through a tracheostomy of an anesthetized dog and wedged it in a bronchus Of a lower lobe. The proximal end of the cannula was attached to one of two openings Of a water bottle. The other end of this Opening was attached to a tube which Opened 0.5 cm beneath the surface of the water. The second Opening of the bottle was attached to a delicately balanced Krogh spirometer. When the dog breathed, gas passed from the surrounding lung through collateral channels and into the isolated segment. During exhalation, the vol- ume Of gas exhaled from the segment passed into the Krogh spirometer and was recorded on a kymograph. Exhaled gas from the rest of the lung was collected in a Roth-Benedict spirometer. Measurements were made under three conditions: 1) initial measurements were taken during normal tidal breathing at functional residual capacity (FRC). 2) Lindskog then placed several small weights on top Of the Roth- Benedict spirometer which raised FRC and caused the dog to breathe at a higher end-eXpired lung volume. This increased exhaled segmental gas volume (ESGV). 3) When the dog in— creased its tidal volume (but maintained a normal FRC) by rebreathing its own expired gas, ESGV again increased. In more recent experiments using slightly different methods, Hogg et al. (20), Woolcock and Macklem (67), Menkes et al. (39), Inners et al. (22) and Robinson and Sorenson (49) have all reported collateral flow resistance to be inversely prOportional to lung volume, thus confirming Lindskog's original findings. An issue On which there is no agreement is the effect of volume history on collateral resistance. Some labora- tories report that collateral resistance is greater on inflation than on deflation (20, 49) while others report the converse (67). Hogg et al. (20) isolated a lower lobe segment in excised human lungs by wedging a catheter in a bronchus beyond the first branch to the superior segment. Polyethylene catheters (3 mm O.d.) were inserted through the pleural surface on either side Of the interlobar fissure Of the segment. When flow was introduced into the segment, collateral resistance was calculated as: Palv - Palv (LL) V Rcoll = (UL) (1) where Rcoll is collateral resistance, Palv(LL) is the alveo- lar pressure in the isolated segment, Palv(UL) is the alveo- lar pressure in the apical segment Of the same lobe and V is the flow of gas into the segment. After suspending the lungs in a volume displacement plethysmograph, the cannu- lated segment was quasistatically inflated to a distending pressure Of 30 cm H20 and deflated to 0 cm H20. When flow was plotted against the pressure drop across the collateral channels, Rcoll was greater on inflation than on deflation at the same lung volume. They suggest this is due to the effect Of surface tension on collateral channels since tissue forces would be the same during both inflation and deflation. Robinson and Sorenson (49), using Hilpert's technique (33), inserted a double lumen catheter (2.5 mm O.d.) through the lobar bronchus of a lower lobe Of excised dog lungs and wedged it in a peripheral airway, thus isolating a lung 10 segment. Steady state collateral resistance (Rcoll) was calculated as: Rcoll = ——————— (2) where V is the flow Of air introduced into the outer lumen of the catheter, PS is the pressure at the catheter tip as measured by the inner lumen and PL is the lobar inflation pressure. These workers found inflation Rcoll to be signifi- cantly greater than deflation Rcoll at equal transpulmonary pressures. Using pressure-volume curves Of the lobe, they plotted Rcoll against percent vital capacity of the lobe. At a given lobar volume, inflation Rcoll was less than deflation Rcoll below forty percent total lobe capacity, and nO significant difference existed between the two values above forty percent total lobe capacity. The former result is easily explained on the basis that at equal transpulmonary pressures lung hysteresis caUses deflation lobar volume tO be greater than inflation lobar volume. The collateral channels have a greater diameter at higher volumes and thus have less resistance tO flow Of gas. The latter finding is difficult to explain. Both Hogg et al. (20) and Robinson and Sorenson (49) calculated lobar volume from pressure- volume curves. Hogg et a1. (20), however, Obtained their pressure-volume curves at the same time they made their measurements of Rcoll, whereas Robinson and Sorenson (49) 11 Obtained their pressure-volume curves before or after the measurements Of Rcoll. Perhaps in the latter case, lobar compliance changed over time so that a transpulmonary pres- sure Of 10 cm H20, for example, did not produce the same lobar volume when plotting the pressure-volume curve as it did when making Rcoll measurements. In contrast with Hogg et al. (20) but in agreement with Robinson and Sorenson (49), Woolcock and Macklem (67) found inflation Rcoll to be less than deflation Rcoll in experi- ments in which collateral ventilation was studied using two different methods. In the first method a segment Of an excised lobe of lung was isolated with a double lumen cathether inserted through the lobar bronchus. One lumen measured pressure at the catheter tip when a small quantity Of air was rapidly injected with a syringe through the other lumen. The effective compliance Of the segment (Ceff) was calculated as the change in volume Of the segment divided by the change in the pressure difference (PC) between the catheter tip and the pleural surface. The second method involved the same experimental prep- aration as the first method but only a single lumen catheter was used. After subjecting the lobe to a particular volume history, it was held at a given lung volume and oscillated by applying a quasi-sinusoidally varying pressure to the bronchial or tracheal cannula at frequencies Of 0.3, 0.5, 1.0 and 1.5 cycles per second using a loud speaker powered 12 by a variable frequency sine wave generator. Lung volume (V1) was plotted against PC on a storage oscilloscope, and the following calculations were made: 9 = tan-12wfRC (3) where O is the phase angle by which PC lags V and is 1 determined by the time constant (T) for ventilation of the isolated segment. R was assumed to be the resistance through the collateral channels since resistance through larger airways would probably be much smaller in comparison. C is the compliance of the lung (assumed to be equal to the compliance Of the isolated segment), and f is the frequency Of oscillations. T could be solved for by rearranging the equation: RC = T = %%%Q . (4) In this manner, collateral resistance (Rcoll) could be calculated by dividing T by the effective compliance (determined as described previously): T Rcoll = Ceff (5) There are several points of criticism to the methods Of Woolcock and Macklem. Concerning their first method: 1) they state that "... volumes Of air were rapidly injected until peak pressure registered by the catheter was about 30 cm H20 and not more than 40 cm H20. The quantity Of 13 injected air depended on the position Of the catheter. When it was wedged in a small bronchus, less than 0.5 ml was needed to distend the lung distal to the catheter to a PC of 30 cm H20, but when the catheter was wedged in a large bronchus as much as 15 ml was needed." They calculated their segment volumes (based on lobe volume fixed at 30 cm H20) to be 14, 34, 60, 140 and 225 ml. In all cases, the volume of air required to produce a peak pressure of 30 cm H20 was much less than 1/10th the segment volume. It is generally accepted that a static distending pressure Of 30 cm H20 causes a lung inflation tO total lung capacity, and yet these investigators reached such a distending pressure in the segment with an injection of air less than 1/10th the segment capacity. Because of this and because the air was injected rapidly, it is possible that most of the pressure recorded after the injection may have been needed to over- come the resistance of small airways distal to the tip of the catheter while only a small portion of the pressure was applied toward overcoming the elastic and surface tension forces in distending the segment. Even if the segment was considered as a closed compartment without collateral chan- nels, placing a volume of gas in the segment equivalent to less than l/lOth its total capacity would cause only a mdnimal change in pressure. 2) Some of the air injected ‘would leak out through collateral channels resulting in an 11nderestimation of Pc' It is known that Rcoll varies 14 inversely with lung volume (20, 22, 39, 49, 67). Because their determinations were made at different lung volumes, Rcoll would vary and thus the amount of air leaking out dur- ing rapid injection would increase as lung volume increased. This causes a greater experimental error at higher lung volumes; that is, PC is underestimated to a greater extent at higher lung volumes causing an overestimation Of Ceff and an underestimation Of Rcoll. Concerning the second method, accurate measurement of the phase angle is possible only when the tracing is a per- fect ellipse. Any deviation from this shape would mean arbitrary placement of the axes and a corresponding error. Also, the calculation Of Rcoll contains the error introduced by inaccurate measurement Of Ceff‘ 2. Pharmacologic Agents Airways respond to many pharmacologic agents, and gen- erally one would predict collateral channels to respond in qualitatively the same manner since both are surrounded by smooth muscle. Only four drugs have been tested for their effect on collateral ventilation: histamine, acetylcholine, epinephrine and iSOproterenol. a. Histamine Histamine constricts airways (9, 12, 13, 46). This effect may be size and species dependent since Persson and Eckman (46) have reported that in vitro preparations of cat 15 bronchioles (1 mm) were constricted in a histamine bath (0.08-10.0 ug/ml) but larger airways (> 3 mm) were unaffect- ed, whereas both large and small airways from dog, guinea pig and human constricted in response to histamine. Lung volume history may also affect the response to histamine. Drazen et al. (13) report that histamine-induced increases in peripheral airway resistance were mostly re- versed by an inflation to total lung capacity. Likewise, Colebatch et al. (9) found that all changes in lung com- pliance and resistance caused by the intravenous administra- tion of histamine to cats were completely reversed by a temporary inflation of the lungs to a transpulmonary pres- sure Of 20 cm H20. Constriction Of collateral channels by histamine was first demonstrated by Ally and Lindskog (2) who wedged a cannula with a dilatable tip in a subsegmental bronchus Of a lower lobe in anesthetized dogs. As the dog inhaled through a tracheostomy, air which entered the surrounding lung passed through collateral channels and into the iso- lated segment. During exhalation, a portion Of the gas volume passed through the wedged cannula and was recorded with a Krogh spirometer while the remaining gas may have passed back through collateral channels and into the sur- rounding lung. Histamine given intravenously in doses of 0.02-2.0 mg resulted in a marked diminution or elimination of ESGV. There was a gradual return to the preinjection 16 rate in a period of minutes. Benadryl, an antihistamine, completely eliminated this effect. In more recent experi- ments with a similar method, Johnson and Lindskog (25) also noted ESGV is sharply reduced or obliterated in anesthetized dogs given 0.01 mg/kg histamine intravenously. b. Acetylcholine Acetylcholine given intravenously (8-800 ug) increases pulmonary resistance and decreases lung compliance in anes- thetized cats (9). Isolated airways Of cats, dogs, guinea pigs and man all constrict in response to acetylcholine (0.05-10.0 ug/ml media) regardless of their size (46). Shon and Batra (52) noted an increase in collateral resistance in anesthetized dogs caused by Urecholine. Their results as well as those Of Colebatch et al. (9) were re- versed by atropine. Chen et al. (8) used a method similar to that of Johnson and Lindskog (25) in anesthetized dogs and found that ESGV completely stopped in response to 0.1 mg/kg mecholyl iv. c. Epinephrine and ISOproterenOl ESGV increased 66-160 percent Of the control level within 2 minutes after iv injection of epinephrine (0.01 mg/kg) in anesthetized dogs (8). In similar experiments, ESGV increased 14.4-64.0 percent in anesthetized dogs given 1.0-13.4 ug/kg iSOprOterenol iv (25). 17 3. Alveolar and Blood Gas Tensions a. Carbon Dioxide Peters (47) ventilated anesthetized dogs with 5.0, 7.5, 10.0, 15.0, 20.0 and 25.0 percent CO2 in 02. Bronchocon- striction occurred with arterial COz tensions (PaCOZ) from 56-84 mm Hg and was reversed when PaCOZ dropped to 38-40 mm Hg. These effects Were abolished by either atropine or vagotomy. Green and Widdicombe (19) demonstrated a decrease in dead space, lung compliance, tracheal volume and an in- crease in lung resistance in anesthetized dogs ventilated with 4% and 8% C02. Similar effects, but to a less magni- tude, were seen with vagal stimulation. In the latter studies, control end-expired CO2 tension ranged from 20-30 mm Hg, and administration of 4% CO2 increased end-expired COz tension to a mean of 42 mm Hg, and 8% CO2 increased the mean value to 56 mm Hg. Bilateral cervical vagotomy reduced the changes in all parameters. It appears, therefore, that the constricting effect Of systemic hypercapnea may be caused by a vagally mediated mechanism, since: 1) elevated PaCOZ is needed for the changes to occur, and 2) the effects are reduced or abolished by vagotomy. Ingram (21) performed experiments with anesthetized dogs which were tracheotomized and cannulated with a Carlens tracheal divider so that each lung could be ventilated independently. A balloon tip catheter was passed through the femoral vein and into the left pulmonary artery so that 18 the artery was occluded by baloon inflation. The right lung was used for gas exchange while measurements of lung compli- ance and resistance were made in the left lung. This experi- mental design allowed independent variations in both sys- temic and alveolar gas tensions and a determination Of their effects on the airways. Increasing PaCOZ to 70 mm Hg while holding alveolar CO2 tension (PACOZ) to 35 mm Hg decreased lung compliance and increased lung resistance. This effect was abolished by vagotomy, thus demonstrating its reflex origin. Increasing PACOZ to 72 mm Hg while holding PaCOZ to 37 mm Hg had no effect. It may be that the direct dilat- ing effect Of CO2 on the airway was not strong enough to overcome the constricting effect of tonic vagal tone. When, however, PACOZ was lowered to 2 mm Hg and PaCOZ was kept at 35 mm Hg, compliance decreased and resistance increased. This shows the direct constrictor effect of hypocapnea on the airways. The greatest changes in compliance and resist- ance were seen when arterial hypercapnea (PaCOZ = 68 mm Hg) was coupled with alveolar hypocapnea (PACO2 = 2 mm Hg). This response was only attenuated, not abolished, by vagotomy, probably because alveolar hypocapnea caused con- striction while the reflex constriction from arterial hyper- capnea was removed. Severinghaus et al. (51) produced unilateral pulmonary artery occlusion in close-chested anesthetized dogs by inserting a balloon tip catheter into either the left or 19 right pulmonary artery. When either pulmonary artery was occluded by balloon inflation, there was a shift Of ventila- tion away from the unperfused lung due to bronchoconstric- tion. This effect was prevented by inhalation Of 6% CO2 into that side, but not by atropine or vagotomy. They attribute their results to hypocapnea directly constricting airway smooth muscle. Daly et a1. (11) reported evidence Of the possible role Of the central chemoreceptors and vagus nerve in the regula- tion Of airway caliber. The brain of anesthetized dogs were perfused with blood containing various mixtures of O2 and CO2 while these gas tensions were held to within a normal range in the blood perfusing the rest Of the body. Passing anoxic blood through the brain caused bronchoconstriction, and hypocapnic blood caused bronchodilation. These effects were abolished by either atrOpine or vagotomy. In denervated lung preparations, CO2 directly dilates the airways. Nisell (43) found that isolated perfused cat lungs previously constricted with carbaminoylcholine, muscarine or histamine dilated when the lungs were venti- lated with 10-14 percent C02. This correlated with his studies in which a decrease in the pH Of the perfusing blood caused dilation whereas an increase in the pH of the per- fusate caused constriction of the airways. Since collateral channels may be small airways sur- rounded by smooth muscle, they should be affected by C02 in 20 qualitatively the same manner as small airways. Most stud- ies (8, 25, 50, 57) involve administration of inspired hypercapnic gas mixtures to anesthetized dogs, and the results indicate that CO2 dilates collateral channels. Chen et al. (8) found that 15% CO2 caused an increase in ESGV compared to room air or 100% 02' PaCOZ was usually in the range of 35-43 mm Hg, considerably lower than that need- ed tO induce reflex bronchoconstriction. Johnson and Lindskog (25) noted varying increases in ESGV from 31-198 percent caused by adding 5% CO2 to the inspired gas. In this case, control arterial pH was approximately 7.54 and decreased to 7.50 with 5% C02. Sealy and Seaber (50) varied inspired C02 concentrations from 2.5-15 percent and noted that the peak increase in ESGV occurred with 6% C02. NO fur- ther increase was seen with 15% C02, perhaps because the smooth muscle Of the airways and collateral channels was maximally dilated. In excised lungs, 15% CO2 increased ESGV by as much as 150%. Traystman et al. (57), using Hilpert's method (33), isolated a segment of lung by wedging a double lumen catheter in a peripheral airway of an anes- thetized dog. The dog was ventilated with room air with a tidal volume adjusted to give an end-expired CO2 concentra- tion Of 4%. Collateral resistance was calculated as: Ps Rcoll = (6) V 21 where Rcoll is collateral resistance, P8 is pressure at the catheter tip measured by the inner lumen and V is the flow Of air, 5% CO2 or 10% CO2 into the segment through the outer lumen. When air flowing through the segment was replaced by 5% C02, Rcoll fell by 46%. When CO2 concentration was increased from 5% to 10%, Rcoll fell an additional 9%. These studies indicate that CO2 may act directly on smooth muscle to produce a decrease in collateral resistance. b. Oxygen Nisell (43) reports that previously constricted airways of isolated perfused cat lungs dilate when ventilated with pure nitrogen and constrict when exposed to 100% 02‘ Pulmonary vessels respond in the reverse manner. Since a hypoxic condition would cause bronchodilation and pulmonary vascular constriction (while the Opposite would be true for a hyperoxic condition), he suggests 02 plays a major role in the regulation Of ventilation and perfusion Of the lung. In agreement with Nisell are the results of Severinghaus et a1. (51) who reported that unilateral bronchoconstriction caused by occlusion Of the pulmonary artery to the same lung was prevented by the inhalation Of 100% nitrogen into that lung. Nisell did not consider, however, the reflex effect of altered arterial O2 tension on the lung. Green and Widdicombe (19) reported that 10% O2 in N2 decreased dead space, lung compliance and tracheal volume, and increased 22 total lung resistance in anesthetized dogs. Small changes occurred with 100% 02 generally in the Opposite direction. They report that hypoxia increases tracheobronchial smooth muscle tone by a reflex effect from the carotid body chemo- receptor. Also, Nadel and Widdicombe (42) report that inhalation Of 10-15 percent 02 in N2 resulted in constric- tion of both upper and lower airways which was prevented by tying the glossopharyngeal nerves or cooling the vagus. This demonstrates the reflex effect of arterial hypoxemia (causing constriction) was stronger than the local effect of airway hypoxia (causing dilation). In regards to collateral channels, Traystman et al.(57) reported that collateral resistance rose 37% when air venti- lating an isolated segment was replaced by 5% O2 in N2. However, 5% O2 and 5% CO in N2 caused a slight reduction 2 in Rcoll, suggesting that CO2 had the dominant effect. Arterial blood gases were not measured, so it is not known if central or peripheral chemoreceptors were activated. 4. Pulmonary Vessels and Vascular Pressures Recent anatomical studies (4, 16, 24) indicate that capillaries and other small vessels are directly attached tO the terminal airways and can influence airway behavior. Kapanci et al. (26) and Assimacopolous et al. (4) using rats, lambs, humans and monkeys found interstitial cells and fibers which are possibly attached to the capillary and 23 alveolar basement membranes as well as being found in the pillars Of alveolar capillaries (53). In the human lung, Gehr et al. (16) have shown that capillaries are embedded in a network Of connective tissue fibers intercalated between endothelium and epithelium. They have also found that in the thin portion Of the alveolar septum the endo- thelial and epithelial basement membranes are fused. These studies indicate that the terminal airways and the vessels which surround them are connected to each other and there- fore are interdependent with each other. Thus, pulmonary vessels potentially influence the mechanical behavior of the lung. A possible role for the vascular influence on terminal airways may be seen in the lung Of the newborn. Jaykka (23, 24) studied the effect Of vascular distention in fetal and newborn lambs. Unperfused lungs inflated to an airway pressure of at least 20 cm H20 were not uniformly expanded, whereas when India ink was injected into the pulmonary artery under a pressure Of 80 mm Hg, the microscopic picture resembled that Of a normally aerated lung. He suggests the capillary system was rendered rigid by the liquid and formed a framework that supported the terminal airspaces, so that in the fetus capillary erection would cause amniotic fluid to enter the potential airspaces. The results Of airway distention caused by pulmonary vascular congestion may depend on whether the lungs are 24 excised or intact within a close-chested animal. Avery et al. (5) performed studies on the gas-free excised lungs Of cats, dogs and human stillborn infants. After cannulat- ing the pulmonary artery and left auricle, vascular pres- sures were elevated by infusing Dextran in Ringer's-lactate into the system. An intratracheal cannula was used to record simultaneous pressure changes in the airways. At vascular pressures Of 20-30 cm H20, the pressure in the airways was 1.0-1.5 cm H20 subatmospheric. At vascular pres- sures Of 70 cm H20, the pressure in the airways was 5.0 cm H20 subatmospheric. They concluded, however, that these airway pressures resulting from capillary distention alone would not approximate the pressures required for inflation Of the lung from the degassed state. Frank et al. (15) reported that the effect of pulmonary vascular congestion on the elastic recoil Of excised cats' lungs may be dependent on lung volume. At large lung vol- umes (airway pressures of 5 cm H20 or more) recoiling force was increased by pulmonary vascular congestion; at inter- mediate lung volumes (airway pressures of 2.5-3.0 cm H20, approximately equal to FRC) the change was negligible and, at smaller volumes recoiling force was reduced. Thus, at low lung volumes pulmonary vascular congestion exerts an erectile effect on the airways, whereas at high lung volumes vascular congestion has a compressing effect by increasing lung recoil. These studies indicate that pulmonary vascular 25 congestion tends to return the lung to a resting volume which is approximately FRC. In this condition, a change in lung volume in either direction will require a greater force than if the pulmonary vessels are not filled. Similar effects have been shown in the intact lungs of normal human subjects (17). In contrast, other studies (18, 30) of intact lungs of close-chested animals have indicated a decreased FRC, decreased compliance and a decreased end- expiratory transpulmonary pressure in response to pulmonary vascular engorgement. At this point there is no agreement regarding the effect of pulmonary vascular pressures on collateral venti- lation. Ankeney et al. (3) found that ESGV decreased in anesthetized dogs in response to increased vascular pres- sures caused by pulmonary venous obstruction, myocardial infarction or pulmonary venous stasis. Close examination of their methods, however, would allow the reader to support the Opposite conclusion. They obstructed a secondary bron- chus in the right lower lobe with a cannula similar to the one used by VanAllen (59). The proximal end Of the cannula was attached to glass tubing, the other end of which was placed just below the water level in a trap bottle. Another piece of glass tubing whose tip was above the water level in the trap bottle was connected to a volume recorder. Gas from the obstructed segment bubbled through the water, and the volume was recorded. With this method, gas passes from 26 the surrounding lung into the isolated segment during inspiration. During expiration, however, the gas may leave the segment via two routes: 1) through the cannula to the trap bottle, or 2) back out through collateral channels and into the surrounding lung. Their conclusion that increased pulmonary vascular pressures decreased collateral ventila- tion was made on the reduced volume of gas collected, and they assumed that elevated pressures distended alveolar capillaries which then encroached upon the Pores of Kohn and thus decreased the volume of gas passing through them; less gas would be allowed to enter the segment, causing a smaller collateral volume to be recorded. This assumption may be erroneous for three reasons: 1) these workers did not consider that there may be pathways other than alveolar pores which provide for collateral flow. 2) It has been shown that capillary distention does not have a measureable effect on pulmonary blood volume (38). This indicates that capillaries change their size only minimally in response to high vascular pressures and therefore should not occlude the pores any more than they would at low vascular pressures. 3) The capillaries are interdependent with the terminal air- ways they surround (4, 16, 53) and may act as a supportive structure to them (23, 24). If this is true, vascular engorgement would have the effect of making capillaries and small vessels more rigid. This would add increased support to the small airways and collateral channels and tend to 27 prevent their closure during expiration. The decreased ESGV seen by these workers may thus be explained as follows: at low vascular pressures, most of the air leaving the seg- ment during expiration passes through the obstructing cannu- la because its resistance is much lower than the resistance through collateral channels. High vascular pressures, however, stabilize collateral channels and causes them to remain more dilated. The resulting decrease in resistance through these pathways causes a greater prOportion of the expired gas to leave the segment via this route, causing less collateral volume to be recorded. The conclusions of Ankeney et al. may therefore be in exact contradiction to their data. Other experiments dealing with the effect of vascular pressure on collateral ventilation were performed by Johnson and Lindskog (25) who clamped the left pulmonary artery in dogs and noted a 16.6-53.0 percent decrease in ESGV which then rose after the clamp was released. They were not able to explain this finding. Menkes et al. (40) and Traytsman et al. (58) used a stop-flow technique in which pulmonary artery pressures of anesthetized dogs on right heart bypass were elevated by as much as 20 mm Hg. Lung perfusion was then abruptly stopped so that pulmonary vascular pressures declined to 0 mm Hg within 12-15 seconds. Rcoll increased as pulmonary vascular pressures decreased, and this was attributed to the effect 28 of local hypocapnea, since: 1) when air was continually introduced into the segment under conditions of no blood flow, the resultant local hypocapnea caused constriction of collateral channels. 2) The effect of vascular pressure on Rcoll was greatly attenuated when 10% CO2 was introduced into the segment instead of room air. They conclude that the change in Rcoll is not directly related to the changes in intravascular pressure since the fall in vascular pres- sures preceded the increase in Rcoll. This conclusion may be in error, however, because Rcoll was not measured until three seconds after lung perfusion was stopped. Their figures indicate that the greatest changes in pulmonary vascular pressures occurred during this time, so the mechan- ical effect of the pulmonary vessels on the airways may have occurred prior to the time their measurements were taken. Also, they did not measure PaCOZ’ so it is unknown if reflexes influenced their results. The local hypocapnea produced in this method may have influenced these workers' results but, in contrast to their conclusion, the change in vascular pressures may be an important factor. Their data indicate that the increase in Rcoll occurred after pulmonary vascular pressures had de- clined to low levels. This may be explained as follows: high vascular pressures may tend to hold collateral pathways Open. When vascular volume is suddenly and rapidly depleted (as with stop-flow), these pathways will tend to narrow but 29 at a rate which may be slower than the rate of vascular volume depletion. This is because the smooth muscle sur- rounding the collateral pathways as well as other elastic elements attached to them may continue to contract several seconds after the distending force has disappeared. In fact, this is a major cause of the time-dependent increase in Rcoll in unventilated lungs statically held at a given transpulmonary pressure (49). Because there is no conclusive evidence which demon- strates the mechanical effect of pulmonary vessels on col- lateral pathways, the following studies measure the effect of pulmonary venous pressure on steady state collateral resistance. If pulmonary vascular congestion exerts an erectile effect on the lung, the resultant changes in lung volume may account for the change in collateral resistance. To determine this, residual volume is also measured. Since measurements were made at different lung inflation pres- sures, changes in lung compliance may affect the lung volume at which measurements were made. Pressure-volume curves are therefore plotted at various pulmonary venous pressures. An isolated perfused lung was chosen as the model for the experiments so that alveolar and blood gas tensions could be controlled during alterations in pulmonary venous pres- sure. The viability of the isolated perfused lung was tested by measuring its response to inspired hypocapnic and hyperoxic gas mixtures. II. MATERIALS AND METHODS A. Isolated Perfused Lung Preparation (IPL) Two mongrel dogs were used in each experiment. The larger dog (16-18 kg, subsequently referred to as the donor dog) was anesthetized with sodium pentobarbitol 33 mm/kg (Abbott Laboratories, North Chicago, IL) and intubated. Ventilation was controlled with a Harvard pump (Model 607, Harvard Apparatus, Dover, MA) to maintain an end-expired CO2 concentration of 5% as measured by an infra-red CO2 analyzer (LB-2, Beckman Instruments, Schiller Park, IL) which was calibrated before each experiment with a known concentration of C02. After injecting heparin (Sigma Chem- ical Co., St. Louis, M0) for anticoagulation (10,000 units intravenously), the left femoral artery was cannulated and connected to a pressure transducer (4710, Bell and Howell, Pasedena, CA) in order to monitor systemic arterial pressure (P ). A fall in P may cause the release of broncho- syst syst active substances (catecholamines, etc.) from the donor dog which affects the mechanical behavior of the isolated lung. Therefore, care was taken (such as the regulation of depth Of anesthesia) to prevent this from occurring. The left cephalic vein was cannulated for infusion of supplemental 30 31 anesthetic. The right femoral artery and vein were cannu- lated to provide and receive blood for subsequent perfusion of the isolated lung. The animal was kept warm with a heat- ing pad. A second dog (9-12 kg) was euthanized with 20 ml of 5% sodium pentobarbitol containing 5000 units of heparin. A thoracotomy was performed at the sixth intercostal space on the left side. After carefully wrapping the left lung with a gauze pad, it was gently held to prevent damage while the left mainstem bronchus was cut at the bifurcation with the trachea. In order to expose the heart, the pericardium was opened and cut around the heart at the level of the atrioventricular valves. After ligation, all pulmonary veins of the right lung as well as the posterior vena cava were sectioned, and the right lung was discarded. The heart was divided in the lower third midway between the apex and the atrioventricular valves, and the apex was discarded. The left pulmonary artery was severed at the bifurcation with the right pulmonary artery. After cutting all remain- ing tissue attachments, the left lung and remaining portion of the heart were removed as a unit and weighed. At the termination of the experiment, the heart and lung unit were again weighed to determine total weight gain as an estima- tion of fluid gain of the preparation. At this time, the airways were checked for gross edema by dissecting several airways as distally as possible with a scissors. 32 A "T-tube” was tied into the left mainstem bronchus of the isolated lung and attached via a sidearm to a tripod so that the heart and lung unit were suspended. The outer sur- face of the lung was kept moist with a saline spray. Lung inflation pressure (PL) was measured at the mainstem bron- chus with a cannula attached to a transducer (PM 131, Statham Instruments, Hato Rey, PR), and under static condi- tions was assumed to be equal to alveolar pressure in the surrounding lung. The lung was inflated several times with a blower to PL = 30 cm H20 in order to remove atelectasis. In the studies which measured the effect Of gas composition on collateral resistance, the lung was ventilated with three different gas mixtures: room air (21/0), 95% O2 and 5% CO2 (95/5) and 100% O2 (100/0). In the studies which measured the effect of pulmonary venous pressure on collateral resist- ance, the lung was ventilated with 95/5 only. The tidal volume was 250-300 ml with a respiratory rate of 10 breaths per minute, and the lung was regularly given deep breaths to PL = 30 cm H20. A cannula (9 mm O.d.) was inserted into the left pulmonary artery so that blood from the donor dog's right femoral artery could be pumped into the lung at a con- stant flow rate using a roller pump (designated as the in- flow pump) (Masterflex, Cole Palmer, Chicago, IL). Another cannula attached 10 cm upstream from the pulmonary artery was connected to a transducer (P23 Db, Statham Instruments, Hato Rey, PR) and measured pulmonary arterial pressure (Ppa). 33 Blood flow (0), as measured with a stopwatch and graduated cylinder, was initially adjusted to produce a mean Ppa in the low to normal range (approximately 10 mm Hg), and O was maintained constant thereafter. The left atrium was cannulated with a rigid large bore tube with a flared end (29 mm O.d.). Pulmonary venous pres- sure (va) was measured with a transducer (P23 Gb, Statham Instruments, Hato Rey, PR) connected to a cannula whose tip rested at the opening of the large bore tube. Ppa and va were recorded with a multipen recorder (Model 60297, Soltec Corp., Sun Valley, CA). Blood leaving the lung passed through the left atrium and was pumped with a second roller pump (designated as the outflow pump) (Masterflex, Cole Palmer, Chicago, IL) into the right femoral vein of the donor dog. Blood therefore passed through the following circuit: donor dog, inflow pump, excised lung, left atrium, outflow pump, donor dog (Figure l). The total time of ischemia of the isolated lung was approximately 25 minutes. A feedback control system regulated va. A propor- tional gain current output controller (Leeds and Northrup, North Wales, PA) received a signal from the va amplifier, compared this to a predetermined va set point and adjusted the speed of the outflow pump appropriately to maintain a constant va. The response time of the system was such that only small changes in va (1 8 mm Hg) occurred with lung inflation and deflation. 34 Figure l A diagram of the isolated perfused lung preparation (IPL). PEEP = positive and expiratory pressure. PL = lung inflation pressure measured at the mainstem bronchus. PS = pressure in the isolated segment as measured by the inner lumen of the wedged catheter. Vcoll is the flow of gas through the outer lumen of the wedged catheter into the isolated segment. Ppa pulmonary arterial pressure. va pulmonary venous pressure. Psyst system1c arterial pressure. 3S [-— vcumnon (—- PE E P VEOLI““" LEE? ATRIUH ' ' va fi 0mm! RECORDER ourvur (5 H001 CONTROLLER cutout INFLOW OUTFLOU 'UHP '0'? F O hLI“FROM FEIORJL IRTERY YO FEIORAL VII“ 0 Psyst DONOR DOG L I no EXPIRED co; . 51. Figure l 36 Pulmonary vascular resistance (Rvas) of the IPL was calculated as: Rvas = P29 - va (7) Q where Ppa is pulmonary arterial pressure, va is pulmonary venous pressure and Q is the rate of blood flow through the IPL. B. Measurement of Steady State Collateral Resistance Steady state collateral resistance (Rcoll) was measured according to the method of Hilpert (described by Macklem (33)). With the lung inflated to 30 cm H O, a double lumen 2 catheter with a flared tip (4 mm O.d.) was inserted into the mainstem bronchus and wedged in a peripheral airway of the lower lobe, thus isolating a segment of lung. The inner lumen was connected to a transducer which measured pressure at the catheter tip (PS). Gas flowed through the outer lumen and into the isolated segment from a compressed gas source and through a rotameter (Model 7431T, Matheson Gas Products, Joliet, IL) which measured collateral flow (Vcoll), and it leaked out of the segment through collateral channels and into the surrounding lung. To measure the pressure gradient between the segment and the lung, PS and PL were 37 recorded on opposing ends of a differential transducer (PM 131, Statham Instruments, Hato Rey, PR). Rcoll was measured at P = 0, 2 and 4 cm H20 while P L L was maintained by adjusting a blower attached to the exhaust port of the ventilator. When Vcoll was adjusted with the rotameter to maintain PS - PL = 3 cm H20, Rcoll was calcu- lated as: P - P Rcoll = S L (8) Vcoll The tracing of PL and PS - PL were recorded on a photo recording oscillosc0pe (Electronic for Medicine, White Plains, NY). It may be possible that some of Vcoll leaks around the tip of the wedged catheter and leaves the segment through more proximal airways rather than through collateral chan- nels. However, the following evidence suggests that this probably does not occur: 1) the catheter is wedged in the segment while the left lung is inflated to total lung capa- city. The lung is then deflated so that the airway col- lapses around the tip of the catheter such that removal of the catheter with the lung collapsed may only be done at the expense of damaging the lung. 2) The tip of the cathe- ter is flared, contributing to a more secure wedge. 3) Rcoll measurements were made at low lung volumes (resid- ual volume to functional residual capacity). The airways 38 are thus more constricted than when the catheter is inserted and tend to retain the catheter in its place. 4) In an animal with poor collateral ventilation (such as the horse), air injected through the outer lumen of the wedged catheter distends the segment and remains trapped when the catheter is clamped (unpublished observations, Robinson and Sorenson, 1975). 5) When recording PS and PL on two different chan- nels of a recorder, during early inflation of the IPL (when collateral channels have not yet opened) the two pressures are out of phase with each other, such that as the lung expands around the isolated segment, PL becomes positive and PS becomes negative. C. Response of the Isolated Perfused Lung to Hypocapnea and Hyperoxia Since the IPL model clearly differs from an intact preparation, it was decided to verify its use for measure- ments of Rcoll and its viability by determining the response to inspired hypocapnea and hyperoxia. Lungs were ventilated with 21/0 (hypocapnea), 95/5 (normocapnea) and 100/0 (hyperoxia). Blood samples were taken from the pulmonary arterial and pulmonary venous sides of the IPL and analyzed for CO2 tensions with a blood gas machine which was cali- brated before each measurement (Model BM3, Mark 2, Radio- meter Co., Copenhagen, Denmark). A comparison of P paCO2 39 and P was an indication of the gas exchange function va02 Of the IPL. Because the sequence in which the gases are administered may influence the lung's response (8), each gas mixture was administered three times in random order (latin square design) to determine the effect of sequence on the repeatability of measurements. Four experiments were performed in which both the iso- lated segment and surrounding lung were ventilated with the same gas mixture. When changing gas mixtures, 5 minutes were allowed for equilibration. A minimum of three Rcoll measurements were taken at PL = 2 and PL = 4 cm H20 in each experiment. A typical data sheet is shown in Appendix A. D. The Effect of Pulmonary Venous Pressure on Steady State Collateral Resistance, Re§idual Volume and Lfipg Compliance Rcoll, lung pressure-volume curves and residual volume (RV) were measured as a function of va. Both the lung and the isolated segment were ventilated with 95/5. In each of eleven experiments, va was varied in the following sequence: va = 5(1), 10, 5(2), 15, 5(3), 20, 5(4), 25, 5(5), 30 5(6) mm Hg*. A minimum of three Rcoll measurements were made at P = 4, P L L cm H20 in 6 experiments so that the mean values between = 2 cm H20 in eleven experiments and at PL= 0 * va = 5 1 , 5( , etc. are the first, second, etc. times that va 6f 5 £1 Hg appeared in the protocol sequence. 40 adjacent va's could be compared. A typical data sheet is shown in Appendix A. At each va, RV was measured in nine of eleven experi- ments by equilibrating the gas in the IPL at PL = 0 cm H20 with a known volume of room air and analyzing the nitrogen concentration of the mixture with a rapidly responding nitrogen analyzer (Model 410, Cardio-pulmonary Instruments Corp., Houston, TX). Known volumes of gas were injected into the lung through the mainstem bronchus in stepwise fashion while PL was recorded, and pressure-volume curves were plotted. Total lung capacity (TLC) was defined as the volume of gas contained in the IPL at P = 20 cm H20. L E. Analysis of Data Unless otherwise stated, all data were analyzed with a randomized complete block analysis of variance (55) and the means compared using the indicated tests with a type I error probability set at 5%. 1. Response of the Isolated Perfused Lung t6 Inspired Hypocapnea and Hyperoxia At each P the treatments were divided into three L groups with each gas administered once per group. This is shown as follows: 41 GROUP I GROUP II GROUP III 95/5 21/0 100/0 21/0 100/0 95/5 100/0 95/5 21/0 1 l l 2 2 2 3 3 3 At least three Rcoll measurements were made per treatment for each dog; when more than three treatments were made, the excess values furthest from the mean were discarded. In this manner, each treatment contained three measurements of Rcoll at each PL. The groups were analyzed with a ran- domized complete block analysis of variance with replicates (54) and the means compared using Tukey's w-procedure. To analyze the effect of time on Rcoll, the gas mix- tures of Groups I, II and III were regrouped as follows: GROUP A GROUP B GROUP C 21/0 21/0 21/0 95/5 95/5 95/5 lOO/O 100/0 100/0 1 2 3 l 2 3 l 2 3 The means Of each group were compared using Tukey's w-procedure or the least significant difference (lsd) test. 2. The Effect of Pulmonary Venous Pfessure on Steidy‘State COlIateral Resistance The effect of va on Rcoll was analyzed in two ways: 1) the mean Rcoll at each PL of adjacent va's were compared using the lsd test, and 2) Rcoll values at the different PL's were grouped at each va, and the resulting average 42 Rcoll at adjacent va's were compared using the paired Student t test. To analyze the effect of time and sequence on Rcoll, the means of the grouped PL's at all va's = 5 mm Hg were compared using the lsd test. 3. The Effect of Pulmonary Venous Pressure on ResiduaI Volume The effect of va on RV was analyzed by comparing the means of adjacent va's using the lsd test. One experiment terminated prematurely because the IPL developed pulmonary edema indicated by large amounts of frothy edema fluid in the airways. Therefore, the data were divided into two groups: GROUP I GROUP II va 5C1) 10 5C2) 15 5(3) 20 5(4) 25 5(5) 30 5(6) A separate analysis of variance was performed on each group. 4. The Effect of Pulmonary Venous Pressure on Lung Compliance Pressure-volume curves were plotted in ten experiments. The curves were analyzed by calculating the compliance on two different portions of the curve. The compliance on the low portion of the curve, CL, was calculated as: _ 50% - 40% TLC CL ‘ APL (9) 43 and the compliance on the upper portion of the curve was calculated as: _ 100% - 90% TLC CH - APL (10) The effect of va on CL and CH were analyzed by comparing the means at adjacent va's using Tukey's w-procedure. In two of the ten dogs, pressure-volume curves were not plotted at every va. One IPL developed pulmonary edema and the other IPL developed a tear in the parenchyma, both at va = 20 mm Hg. Therefore, the data were divided into two groups: GROUP 1 GROUP 11 va 5(1) 10 5(2) 15 5(3) 20 5(4) 25 5(5) 30 5(6) n = 10 n = 8 A separate analysis of variance was performed on each group. 5. The Effect of Pulmonagy Venous Pressure on PuImonary Vascular Resistance The change in Rvas between the following adjacent va's were compared: P -1 5 - 5 - pv 5(1) 0, (2) 1,5 20,5 (3) - 25, 5 - 30 (5) mm Hg. (4) The means were compared using a Student-Newman-Keuls' (SNK) test. III. RESULTS A. Response of the Isolated Perfused Lung to Hypocapnea and Hyperoxia Figure 2 shows the effect Of changing alveolar gas composition on Rcoll during repetitions of the various gas mixtures. In all Groups, Rcoll with 21/0 was significantly higher than with 95/5. The following differences occurred between Groups II and III and Group I: 1) in Groups II and III, Rcoll with 21/0 was not significantly different than Rcoll with 100/0, but in Group I this difference was sig- nificant. 2) In Groups II and 111 there was a significant difference in Rcoll with 95/5 and 100/0, but in Group I this difference was not significant at PL = 2 cm H20. 3) In Group I at PL = 4 cm H20, there were no significant differ- ences between any of the means when using Tukey's w-pro- cedure (the same statistical test used for comparison of the other means). However, when using a more liberal test (the lsd test), all three means were significantly different. The effect of time on the IPL is indicated in Figure 3. The data in Figure 2 have been regrouped so that each sec- tion of Figure 3 shows the repetition of the same gas mix- ture with each mixture administered once near the beginning, middle and end of the experiment, and each experiment 44 4S .QSOHU some you v u : .mo._vm um acmofimMcmflm Ohm aungHmMawwm «ocv mz uoxume no: moo=0hommww HH< .xuwuoflo mo oxmm one how 65mm onu mm nsouw sumo :wsuwz oonoscom one .ho>ozo: .ouawwm one an .OAOHXwE some mo .kao>muooanh .mewumnu -mwcwsvm wuwnu can vnooom .umuwm on» campcoo HHH one HH .H mnsouu page Om mcwwmov Ohmscm :«umav oonoacom Eovcmn cw moEwu oops» vopoumficfisum was ousuxfis mow summ .aoN: Eu e was N a any monsmmoun OOwumHmcw wcza o:« no nHHoomv oucwummeh Honoumnfioo oumum Sumoum :o fio\ooav No wooa use RO\H~U new soap .fim\mmv Noo em was No 5mm mo ougemo one N ousmfim 46 n O\OO. n Ohm n N m\mm O\OO~ N ouswwm N DEN N D\mm O\OO. . _ _ Dim m\mm wz m2 q H “Bozo 0 m2 m2 1 NH “Some J undo euro 1 l J l H .328 L mQm mcfium>ofio umau OONH may :o wcflafimm munwom .m.>gm woam>ofio :mnu uoumoem m“ m.>mm Houucou um Haooz Hana mafia onu soaon wcwafimm muafloa away On :SOHO ma mo mafia < .Awm as m u >nm. m.>nm HOHunoo oumomvcw um HHoum oumowwcw Nuwucmefi we “Hoax umcfimwm wouuoHO mw m>mmv mousmmoug msoco> zumcoeflam woum>oao um fiHHoomv oocmumwmou Hahoumaaou oumum hpmoum m ohsmwm 56 «m muswflm :..E.....E .0»... Eu .9 x float OOm OOV OOM OON OO. O - q a - . DI SE 30 u >ll 4 OO. 1 OON 1 OOM JOOV POOm ,o: x 11003 ugw - Hm - 03H we 57 mm «semen 55.15%: .8 .2 x 58m com 00.. com OON oo. o 1 q W _ q s: eesmuid . a m ..OO. .n X m. -OON 3 w room 2H nu ....a w A I. H loov .: 9 w w m. w H Loom D S8 Um ousmfim ENE . .LE . ON: EU QO. X 4400K OOm OOV OOm OON OO. O — - u q q a L: OIEE mu 1 o o ma 0 O .00. .n X m. -OON 3 w -OOm H 7v o M. w .. .oov .... m w w U w a H 000 .D am ouzwwm :.E..-.E.O~I Eu .2 x 38¢ 59 000 OOV OI EEAQW u >QQ OOM OON OO. O a lhfld 6H um: 93 OO. OON OOM OOV 00m 90! X T7098 ww - c-IW-OZH we 60 mm museum SE . .-.E . of EO OOm OOV OOm OON '0. X 44001 OO. a... EEofi q u>mQ 6H mm 02 : Add O OO. OON 00m 009. 00m ,0: x 11°38 l-lw ° 03H “13 UHJJ ' 61 C. The Effect of Pulmonary Venous Pressure on ResidualiVolume Figure 6 shows the effect of va on RV. Elevated va's were generally associated with larger RV's while lower va's resulted in smaller RV's. Statistically significant changes begin between va = 10 and 5(2) mm Hg. D. The Effect of Pulmonary Venous Pressure on Lung Compliance Pressure-volume curves were plotted in ten experiments. Figure 7 shows the effect of va on the average compliance Of the lung at the low and high portion of the pressure- volume curve (CL and CH’ respectively). The only signifi- cant effect Of va on CL exists between va = 30 and 5(5) mm Hg, while elevating va is associated with decreasing CH (and vice versa) beginning between va = 5(2) and 15 mm Hg and continuing throughout the protocol (see ANOVA table in Appendix B). E. The Effect of Pulmonary Venous Pressure on Pulmonary Vascular Resistance Table 4 indicates differences in Rvas between control va's (va 5 mm Hg) and elevated va's. In each case Rvas decreases at the higher va. There is little or no further decrease once va has reached 15 mm Hg (see ANOVA table and SNK test in Appendix B). 62 Figure 6 The effect of pulmonary venous pressure (va) on residual volume (RV) in the isolated perfused lung (IPL). Elevating va increases RV and lowering va decreases RV. Stars indicate significant changes (P< .05). n = 9 RV ml sor 88- 86- 84- 82- 80- 78- 76 74 72 70- 68- 66 64 62- 60- 63 P l l I l l l l l l I 5m '0 5(2fl5 5&20 5:4125 5:51:50 5(6) PPV mm Hg Figure 6 64 Figure 7 The effect of pulmonary venous pressure (va) on the compliance of the isolated perfused lung (IPL) at two different lung volumes. C = compliance between 40 and L 50% TLC, CH = compliance between 90 and 100% TLC. Stars indicate significant changes (P< .05). n = 10. COMPLIANCE %TLC-cm H20" ~ ~ (3 .— 00 (N 45 tn (D ‘\l 00 (O 65 L I I I I I I I I I I 501,0 5(2) '5 5(3) 20 5(4) 25 5(5130 5(6) PPV mm Hg Figure 7 Table 4. 66 The change in Rvas (mm Hg-ml'lomin) between con- trol va's (va = 5 mm Hg) and elevated va's at PL = 0, 2 and 4 cm H20. va (mm Hg) p 0 -10 5 -15 -2 - -30 L (cm “2 ) 5(1) (2) S(3) 0 5(4) 25 5(5) PL = 4 -0.017 -0.031 -0.035 -0.037 -0.038 1 SE 0.002 0.004 0 006 0.006 0.007 pL = 2 —0.016 -0.026 -0.031 -0.032 -0.037 1 SE 0.002 0.005 0.005 0.005 0.006 pL = 0 -0.016 -0.032 -0.039 -0.040 -0.039 : SE 0.006 0 007 0.007 0.008 0.009 n 11 11 11 10 10 IV. DISCUSSION The results of the present study indicate that hypo- capnia increases Rcoll in an IPL while the effects of 100% 02 are not consistently different from 21% 02 (21/0). Since the preparation lacks extrinsic neural supply, it is probable that CO2 acts directly on airway smooth muscle so that hypocapnia causes constriction of airways and smooth muscle-containing collateral channels. The similar effects Of 100% O2 and 21/0 suggest the following possibilities: I) hyperoxia has no effect on the airways, and the IPL responds only to the resultant hypocapnea. 2) Hyperoxia constricts airways in the IPL, but the effect is not seen because the airways may be maximally constricted by the hypo- capnia. With the present methods, it is not possible to discern which mechanism occurs. In most studies in which various gas mixtures were administered to either an isolated or intact lung, the gases were given only once. However, there is evidence that the quantitative response of the lung differs if the same gas is given more than once. Chen et al. (8) ventilated anesthe- tized dogs with several different mixtures of O2 and CO2 and found that if any one gas mixture was given a second time, 67 68 the amount of exhaled segmental gas volume (ESGV) was much greater than when the gas was first given. In the present studies, therefore, the three gas mixtures were each random- ly given three times. As shown in Figure 3, Rcoll generally did not change with repeated exposures of the IPL to either 21/0 or 100/0. Rcoll progressively decreased, however, when measured during the second and third exposures of the IPL to 95/5. Chen et al. (8) suggest that once the lung is exposed to C02, the collateral channels either become larger or new channels Open. This is not supported in the present study since Rcoll with either 21/0 or 100/0 changes little or not at all even after exposure to CO2 (Figure 2). The cause of the progressive decrease in Rcoll with repeated exposures to 95/5 remains unsolved. Sealy and Seaber (50) measured Rcoll in unperfused excised dog lungs ventilated with room air. ESGV increased only slightly dur- ing the initial 40 minutes after excision but then increased by 300% of its initial value after 4 hours. In this case, the lung may have been dying due to lack of nutrient deliv- ery, but this is unlikely in the present studies since the lung is perfused with blood from a living animal. In specu- lation, it may be that CO2 causes the release of a broncho- active substance from the lung which produces bronchodila- tion. Upon repeated exposure to C02, airway smooth muscle may become hypersensitized to the substance and respond by 69 dilating in greater magnitude. Further studies are needed to clarify the mechanism. Because the increased Rcoll in response to hypocapnea in the IPL is in agreement with the findings of other inves- tigators (19, 21, 43, 47, 50, 51, 57) who demonstated that hypocapnea causes constriction of small airways and collat- eral channels, it is concluded that the IPL remains viable during the experiments and provides a suitable model for studying Rcoll. A major advantage of this preparation is lung denervation and absence of reflex changesimlbronchiolar tone. In the intact preparation, lung inflation decreases vagal output and lung deflation increases vagal output to the airways, resulting in airway dilation on inflation and airway constriction on deflation (10). Measurements for Rcoll are thus affected not only by the volume history pre- ceding the measurement but also by the lung volume at which the measurement is made. Similarly, there is no reflex effect from changes in lung vascular volume. Increased left atrial or pulmonary venous pressure in the intact preparation activates low pressure receptors in this area and may alter the reflex vagal and/or synpathetic output to the lung (6, 7), result- ing in changes in airway caliber. Also, any procedure hav- ing a chronotropic or inotropic effect on the heart influences the behavior of these receptors and reflexes by modifying atrial filling pressure. 70 Other advantages of the IPL preparation lie in the ability to precisely control alveolar gas composition by controlling the gas tensions in the inspired gas and per- fusate, and in the ability to alter any of the pulmonary vascular parameters (Ppa, va, 0) without concern for main- taining a normal cardiac output as in the intact preparation. A major difference when comparing an excised lung to an intact preparation is that in the former, Rcoll values are generally lower. In the IPL, control Rcoll at FRC ranges 4 to 200 x 10"4 cm H 0-m1'1-min., whereas in 2 intact dog lungs at FRC Rcoll ranges from 150 x 10'4 to 2500 x 10.4 cm HZO-ml-l'min. (57). Woolcock and Macklem (67) report Rcoll in excised unperfused human lungs near FRC to be approximately 233 x 10'4 cm HZO-ml'lomin., whereas 4 from 75 x 10- Terry et al. (56) report an average Rcoll of 510 x 10- cm HBO-ml-lomin. in human volunteers at FRC. There are four possible explanations for these findings: 1) in excised lungs, FRC is defined as the lung volume when trans- pulmonary pressure equals 5 cm H O, and this may exceed FRC 2 in close-chested animals which is the lung volume at the end of expiration. Since Rcoll is inversely proportional to lung volume, this would cause excised lungs to have a lower Rcoll. 2) Disruption of the normal blood supply to the IPL may cause death of airway smooth muscle. This is not believed to be a major factor, however, since: a) there is no edema in the airways at the end of the experiment, 71 indicating the alveolar epithelial and capillary endothelial membranes are still intact; b) gas exchange in the IPL remains satisfactory, as shown in Table l; c) the airways change their caliber in response to variations in inspired gas mixtures, and this reaction is consistent, predictable and correlates with data from 1p yiyp preparations. 3) Since the IPL lacks neural connections, absence of vagal tone to airway smooth muscle may cause dilation and a decrease in Rcoll. Olson (45), however, was unable to demonstrate any change in Rcoll with vagotomy. 4) The IPL is suspended from the mainstem bronchus and surrounded by atmospheric pressure, whereas intact lungs are suspended from the hilum, rest on the diaphragm and are surrounded by a negative intrapleural pressure. It is possible that these gravitational effects in the IPL cause the small airways and collateral channels to be relatively dilated, resulting in a decreased Rcoll. Another difference between the IPL and intact lungs is that the blood perfusion rate in the IPL averaged only 16% of the ip yiyp rate. This is not unusual, however, since it is not possible to achieve more than 20% of the physio- logical flow rate without inordinately high vascular pres- sures and resultant pulmonary edema (N. C. Staub, personal communication). In the present studies it is felt that since va is controlled, this low perfusion rate will not 72 affect pulmonary vascular volume and is therefore of minor consequence. The result of the studies on the effect of va on Rcoll indicate that pulmonary vascular pressures influence steady state collateral resistance when va equals or exceeds 15 mm Hg, such that elevated va's are associated with low Rcoll while low va's are associated with high Rcoll. Residual volume is increased by high va's, and lung compliance is not affected by these vascular pressures at the PL's at which the measurements were made. The rela- tionship between lung volume and pulmonary vascular pressure was first described by vonBasch (according to Gray et al. (18)) who suggested that pulmonary vascular engorgement pro- duced an increase in resting lung volume or "Lungenschwel- lung." This observation is supported by data from excised cat and dog lungs (5, 15) where airway pressures were reported to decrease in response to elevated pulmonary vascular pressure, and also in the histological appearance of the human fetus and newborn lung (23, 24). The evidence suggests, therefore, that pulmonary vascular engorgement may exert an erectile force on the lung which increases lung volume and thereby decreases Rcoll. One possible mechanism in which this occurs is shown schematically in Figure 8. The airways are surrounded by, attached to and interdepend- ent with alveoll, large blood vessels and their associated pulmonary capillary bed. Anything which tends to stiffen 73 Figure 8 A schematic diagram showing a possible mechanism by which elevated pulmonary venous pressure (va) decreases steady state collateral resistance (Rcoll). When va is low, the alveolar capillaries (C) and pulmonary arterioles (PA) contain only a small blood volume, are very distensible and do not support the small airways (TB = terminal bron- chiole) or the alveoli (A). When va is high, the alveolar capillaries and pulmonary arterioles are engorged with blood and are expanded to their elastic limit. They are therefore more rigid and act to splint the small air- ways and alveoli and prevent them from collapsing. ‘ awn“ \ "v o \ u. {"15 O ‘ s V ' i‘h’ ‘n ‘I’obfifilb‘:,o.‘---‘. fit‘o.q¢n°u;ggo€fi'£gg’ ‘ L ‘ \':}".v "a. ‘ 5‘,» Figure 8 75 the parenchyma acts to support the compliant, noncartilag- enous small airways and tends to prevent collapse. On the other hand, surface tension forces and creeping of smooth muscle tend to constrict small airways which do not have adequate support from the surrounding parenchyma. At low vascular pressures, the alveolar capillaries and larger vessels contain only a small volume of blood; they are therefore distensible and not supportive to the airways. At very high vascular pressures the vessels expand to their elastic limit, becoming much stiffer and tending to support and stabilize the airways. This stabilization then in- creases resting (residual) volume and decreases Rcoll. The erectile effect exerted by the pulmonary vessels on the small airways may not occur at all lung volumes. Frank (15) measured airway pressure in excised cat lungs while simultaneously increasing left atrial pressure to 20-30 cm H20. At low lung volumes (airway pressures of less than 3 cm H20) airway pressure decreased in response to high vascular pressures, while at high lung volumes (airway pressures of 5 cm H20 or more) airway pressure increased. At intermediate lung volumes (airway pressures of 3-5 cm H20) airway pressure did not change significantly. Similar results were also demonstrated by Goldberg et al. (17) who Obtained pressure-volume curves from normal human subjects when mouthpiece pressure was at atmospheric pres- sure and also when mouthpiece pressure was 30 cm H20 below 76 atmospheric pressure. They found in the latter case that lung compliance decreased above FRC and increased below FRC in comparison to the former and attributed its effect to the increased pulmonary vascular volume when mouthpiece pressure was -30 cm H20. This suggests that vascular engorgement tends to move the vessels and the airways they surround and the matrix of tissue in which they are embedded to a "resting volume" which is approximately FRC. The vasculature then resists changes in lung volume in either direction: that is, it tends to expand lungs held at lower volumes and compress lungs held at higher volumes. This may explain why in the present studies elevated va's had the smallest effect at P = 4 cm H20 and the greatest effect L at P = 0 cm H O. L 2 It may be argued that the increased RV at high vascular pressures is the cause of the decreased Rcoll since Rcoll is inversely proportional to lung volume (20, 22, 39, 49, 67). But there is some evidence which suggests this may not be true: 1) the changes in RV were small in comparison to the TLC of the left lung (ARV/TLC x 100s 5%). If this resulted in a subtle change in Rcoll, the change may be too small to be measured by the present methods. 2) In two experiments there was no change in RV throughout the entire protocol, and yet Rcoll still changed in response to changes in va. 3) In portions Of nearly all the experiments there were changes in RV without corresponding changes in Rcoll. 77 An hypothesis is that the increased RV caused by elevated va's is confined mainly to the small airways (i.e., alveolar ducts) while the decreased Rcoll is due to a great- er cross-sectional area of the collateral channels. If the longitudinal dimension of the channels decreased as the cross-sectional area increased, the increased RV would thus occupy only the small airways since the volume within the channels would be unchanged. It is still possible, however, that changes in RV influence Rcoll in some dogs but not in others. The graphs in Figure 9 schematically show the relationship between Rcoll and lung volume taken from two different dogs (unpub- lished observations, Robinson and Sorenson, 1975). In Figure 9A a slight increase in RV (RV1 to RVZ) has no effect on Rcoll, whereas in Figure 9B a slight increase in RV has a marked effect in Rcoll. These curves were not plotted in the present studies, so it is possible that this effect of residual volume on Rcoll was present in some of the experiments. In consideration of the relationship between residual volume and Rcoll, the following question is raised: Where, within the isolated segment, does the change in Rcoll occur? If the increased residual volume at high va's causes the decreased Rcoll, then perhaps the major site of Rcoll is in the small airways between the tip of the wedged catheter and the beginning of the collateral channels. These airways may 78 Figure 9 Schematic curves showing two different relationships between steady state collateral resistance (Rcoll) and lung volume. In 9A, a small change in residual volume (RV) has little effect on Rcoll. In 9B, a small change in RV has a marked effect on Rcoll. 79 ‘ “ “ VOLUME meg VOLUME Figure 9 80 be dilated to contain the additional volume, and the resist- ance to gas flow may be decreased. Conversely, if there is no relationship between residual volume and Rcoll, then perhaps the major site of Rcoll is within the collateral channels, explained as follows: suppose the total resist- ance to gas flow through the isolated segment (Rcoll(T)) is composed of two resistances in series, resistance through small airways (Rcoll(1)) and resistance through collateral channels (Rcoll(2)), and suppose further that Rcoll(2)>> Rcoll(1). A small change in RcollCl) due to a change in residual volume would not effect Rcoll(T) since Rcoll(2) is the rate-limiting factor. Rcoll(2) would thus be the major site of resistance. It is not yet possible with the present methods to determine which mechanism occurs. An additional consideration is the question of why sig- nificant decreases in Rcoll did not occur until va had reached 15 mm Hg. A possibility is that the compliance of the pulmonary vessels remains unchanged until va nears 15 mm Hg. At this point, the increased blood volume has expanded the vessels to their elastic limit, and they become stiffer and supportive to the airways. This is sug- gested by the data in Table 4 which shows that there are no further changes in Rvas when va is raised above 15 mm Hg. The effect of va on Rcoll becomes Obvious only when va equals or exceeds 15 mm Hg. Perhaps recruitment of pulmon- ary vessels occurs pari ppssu as va is raised. 81 At va = 15 mm Hg, the critical Opening pressure of all the vessels may have been exceeded so that the entire popula- tion of vessels participate in supporting the airways. It is possible that both recruitment and distension of vessels occurs simultaneously and that vascular support of the airways depends on vessel compliance as well as the total number Of Open vessels surrounding the airways. V. SUMMARY AND CONCLUSIONS 1) Elevated va's decrease Rcoll and low va's increase Rcoll. 2) Elevated va has no effect on lung compliance at 50 percent TLC. 3) Elevated va decreases lung compliance at 90-100 percent TLC. 4) Elevated va increases RV. 5) The increased RV with elevated va's may be associ- ated with the decreased Rcoll. 6) Inspired hypocapnea increases Rcoll, probably by a direct effect on airway smooth muscle. 7) The effect of inspired hyperoxia on Rcoll is not consistently different than that of inspired normoxia. 8) The IPL remains responsive to changes in inspired gas composition during the course Of the experiments. 9) The IPL provides a suitable model for studying Rcoll. 82 LI ST OF REFERENCES (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) LIST OF REFERENCES Adriani, quoted from Miller, W. S., The Lung. Springfield, 111.: Thomas, 1937, pp. 64-68. Alley, R. D. and G. E. Lindskog. Pharmacologic factors influencing collateral respiration: possible relation to the etiOlOgy of pulmonary complications. Ann. Surg, 128: 497-508, 1948. Ankeney, J. L., C. A. Hubay and F. W. Tillotson. The effect of changes in the pulmonary circulation on collateral ventilation. Surg. Forum. 1: 25-33, 1950. Assimacoloulus, A. R. Guggenheim and Y. Kapanci. Changes in alveolar capillary configuration at different levels of lung inflation in the rat. An ultrastructural and morphologic study. Lab. Invest. 34: 10-22, 1976. Avery, M. E., N. R. Frank and I. Gribetz. The effects of vascular distention on inflation of the newborn lung. A.M.A. J. Diseases Child. 96: 505-506, 1958. Aviado, D. M., Jr., T. H. Li, W. Kalow, C. F. Schmidt, G. L. Turnbull, G. W. Peskin, M. E. Hess and A. J. Weiss. Respiratory and circulatory reflexes from the perfused heart and pulmonary circulation of the dog. Am. J. Physiol. 165: 261—277, 1951. Aviado, D. M., Jr. and C. F. Schmidt. Reflexes from stretch receptors in blood vessels, heart and lungs. Physiol. Rev. 35: 247-300, 1955. Chen, C., W. C. Sealy and A. V. Seaber. The dynamic nature of collateral ventilation. J. Thoracic Supg. 59: 518-529, 1970. Colebatch, H. J. H., C. R. Olsen and J. A. Nadel. Effect of histamine, serotonin and acetylcholine on the peripheral airways. J. Appl. Physiol. 21: 217-226, 1966. Comroe, J. H. Physiology Of Respiration. Year Book Medical Publishers, Inc. Chicago, 1974. 83 (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) 84 Daly, M. de Burgh, C. J. Lambertsen and A. Schweitzer. The effects upon the bronchial musculature of altering the oxygen and carbon dioxide tensions of the blood perfusing the brain. J. Physiol. 119: 292-341, 1953. Drazen, J. M. and F. Austin. Atropine modification of the pulmonary effects of chemical mediators in the guinea pig. J. Appl. Physiol. 38: 834-838, 1975. Drazen, J. M., S. H. Loring, A. Jackson, J. R. Snapper and R. H. Ingram. Differential effects on fowler dead- space and pulmonary resistance of histamine aerosol and vagal bronchoconstriction in dogs (Abst). Fed. Proc. 37: 221, 1978. Dyguid, J. B. and M. W. Lambert. The pathogenesis of coal miners' pneumoconiosis. J. Path. Bacteriol. 88: 389-403, 1964. Frank, N. R. Influence of acute pulmonary vascular congestion on recoiling force of excised cats' lungs. J. Appl. Physiol. 14: 905-908, 1959. Gehr, P. M. Bachofen and E. R. Weibel. The normal human lung: ultrastructure and morphometric estimation of diffusion capacity. Resp. Physiol. 32: 121-140, 1977. Goldberg, H. S., W. M. Mitzner, K. Adams, H. Menkes, S. Lichtenstein and S. Permutt. Effect of intra- thoracic pressure on pressure-volume characteristics of the lung of man. J. Appl. Physiol. 38: 411-417, 1975. Gray, B. A., D. R. McCaffree, E. D. Sivak and H. T. McCurdy. Effect of pulmonary vascular engorgement on respiratory mechanics in the dog. J. Appl. Physiol. 45: 119-127, 1978. Green, M. and J. G. Widdicombe. The effects of venti- lation of dogs with different gas mixtures on airway caliber and lung mechanics. J. Physiol. 186: 363-381, 1966. Hogg, J. C., P. T. Macklem and W. M. Thurlbeck. The resistance of collateral channels in excised human lungs. J. Clin. Invest. 48: 421-431, 1969. Ingram, R. H., Jr. Effects of airway versus arterial CO changes on lung mechanics in dogs. J. Appl. Physiol. 38: 603-607, 1975. (22) (23) (24) (25) (26) (27) (28) (29) (30) (31) (32) (33) (34) 85 Inners, C. R., P. B. Terry, R. J. Traystman and H. A. Menkes. Lung volume and resistance to collateral venti- lation in man. Fed. Proc. 37: 927, 1978. Jaykka, S. Capillary erection and lung expansion. Acta Paediat. 45(suppl. 112): 1-91, 1957. Jaykka, S. Capillary erection and the structural appearance of fetal and neonatal lungs. Acta Paediat. 47: 484-500, 1958. Johnson, R. M. and G. E. Lindskog. Further studies on factors influencing collateral ventilation. J. Thoracic Surg. 62: 321-329, 1971. Kapanci, Y., A. Assimacopoulus, C. Irle, A. Zwahlen and G. Gabbiani. "Contractile interstitial cells" in pulmonary alveolar septa: a possible regulator of ventilation/perfusion ratio? Ultrastructural, immuno- fluorescence and in vitro studies. J. Cell Biol. 60: 375-392, 1974. Krahl, V. S. Microscopic anatomy Of the lungs. Am. Rev. Resp. Diseases. 80(1, part II): 24-44, 1959. Kuno, K. and N. C. Staub. Acute mechanical effects of lung volume changes on artificial microholes in alveolar walls. J. Appl. Physiol. 24: 83-92, 1968. Lambert, M. W. Accessory bronchiole-alveolar communi- cations. J. Path. Bact. 70: 311-314, 1955. Levine, O. R., R. B. Mellins and A. P. Fishman. Quantitative assessment of pulmonary edema. Circ. 17: 414-426, 1965. Res. Lindskog, G. E. and H. H. Bradshaw. Collateral respira- tion. The chemical composition and volume of the collaterally respired gases. Am. J. Physiol. 108: 592, 1934. 581- Loosli, C. G. Interalveolar communications in normal and in pathologic mammalian lungs. Arch. Path. 24: 743-776, 1937. Macklem, P. T. ventilation. Airway obstruction and collateral Ppysiol. Rev. 51: 368-436, 1971. Macklem, P. T., A. J. Woolcock, J. C. Hogg, J. A. Nadel and N. J. Wilson. Partitioning of pulmonary resistance in the dog. J. Appl. Physiol. 26: 798-805, 1969. (35) (36) (37) (38) (39) (40) (41) (42) (43) (44) (45) (46) 86 Macklin, C. C. Alveolar pores and their significance in the human lung. Arch. Path. 21: 202-216, 1936. Martin, H. B. The effect of aging on the alveolar pores of Kohn in the dog. Am. Rev. Resp. Diseases. 88: 773-778, 1963. Martin, H. B. Respiratory bronchioles as the pathway for collateral ventilation. J. Appl. Physiol. 21: 1443-1447, 1966. Maseri, A. P. Caldini, P. Harward, R. Joshi, S. Permutt and K. Zierler. Determinants of pulmonary vascular volume. Recruitment versus distensibility. Circ. Res. 31: 218-228, 1972. Menkes, H., D. Lindsay, G. Gamsu, L. Wood, A. Muir and P. T. Macklem. Measurement of sublobar lung volume and collateral flow resistance in dogs. J. Appl. Physiol. 35: 917-921, 1973. Menkes, H. A., R. J. Traystman, P. B. Terry, G. K. Batra, B. R. Pitt and J. T. O'Neill. The effects of changes of pulmonary vascular pressure and flow and collateral resistance (Abst). Fed. Proc.-35: 395, 1976. Miller, W. S. The alveolar pores of pneumonia. J. Exp. Med. 42: 779-783, 1925. Nadel, J. A. and J. G. Widdicombe. Effect of changes in the blood gas tensions and carotid sinus pressure on tracheal volume and total lung resistance to airflow. J. Physiol. 163: 13-33, 1962. Nisell, 0. L. The action of oxygen and carbon dioxide on the bronchioles and vessels of the isolated perfused lungs. Acta. Physiol. Scand. 21(suppl. 73): 1-62, 1950. Ogawa, C. Contributions to the histology of the respiratory spaces of the vertebrates lungs. Am. J. Anat. 27: 333-393, 1920. Olson, L. E. Vagal effects on collateral flow resist- ance in dog lungs. M. S. Thesis, Michigan State University, 1978. Persson, C. G. A. and M. Eckman. Contractile effects of histamine in large and small respiratory airways. Agents and Actions. 6: 389-393, 1976. (47) (48) (49) (50) (51) (52) (53) (54) (55) (56) (57) (58) (59) 87 Peters, R. M. Relationship between percent carbon dioxide in inspired air and degree of bronchoconstric- tion. Ann. Surg. 142: 461-468, 1955. Raskin, S. P. and P. G. Herman. Interacinar pathways in the human lung. Am. Rev. Resp. Diseases. 111: 489- 495, 1975. Robinson, N. E. and P. R. Sorenson. Collateral flow resistance and time constants in dog and horse lungs. J. Appl. Physiol. 44: 63-68, 1978. Sealy, W. C. and A. V. Seaber. The action of carbon dioxide on the collateral pathways of pulmonary venti- lation. J. Thoracic Surg. 49: 533-538, 1975. Severinghaus, J. W., E. W. Swenson, T. N. Finley, M. T. Lategola and J. Williams. Unilateral hypoventilation produced in dogs by occluding one pulmonary artery. J. Appl. Physiol. 16: 53-60, 1961. Shon, B. and G. K. Batra. Effects of urecholine on mechanics of collateral ventilation (Abst). Fed. Proc. 37: 871, 1978. Sobin, S. S., H. M. Tremer and Y. C. Fung. Morpho- metric basis of the sheet-flow concept of the pulmonary alveolar microcirculation in the cat. Circ. Res. 26: 397-414, 1970. Sokal, R. R. and Rohlf, F. J. Biometry. W. H. Freeman and Company. San Francisco, 1969. Steel, R. G. D. and Torrie, J. H. Principles and Pro- cedures of Statistics. McGraw-Hill Book Company, Inc. New York, 1960. Terry, P. B., R. J. Traystman, H. H. Newball, G. B. Batra and H. A. Menkes. Collateral ventilation in man. N. Eng. J. Med. 298: 10-15, 1978. Traystman, R. J., G. K. Batra and H. A. Menkes. Local regulation of collateral ventilation by oxygen and carbon dioxide. J. Appl. Physiol. 40: 819-823, 1976. Traystman, R. J., P. B. Terry and H. A. Menkes. Carbon dioxide--a major determinant of collateral ventilation. J. Appl. Physiol. 45: 69-74, 1978. VanAllen, C. M. A dilatable bronchial cannula. Yale J. Bio. Med. 2: 295-296, 1930. (60) (61) (62) (63) (64) (65) (66) (67) (68) 88 VanAllen, C. M. Obstructive pulmonary emphysema and collateral respiration. Surg: Gynecol., Obst. 55: 303- 307, 1932. VanAllen, C. M. and W. F. Adams. The mechanism of obstructive pulmonary atelectasis. Surg. Gynecol., Obst. 50: 385-396, 1930. VanAllen, C. M. and G. E. Lindskog. Obstructive pulmon- ary atelectasis. Problems of pathogenesis and clinical management. Arch. Surg. 21: 1195-1213, 1930. VanAllen, C. M. and G. E. Lindskog. Collateral respir- ation in the lung. Role in bronchial obstruction to prevent atelectasis and to restore patency. Surg. Gynecol., Obst. 53: 16-21, 1931. VanAllen, C. M., G. E. Lindskog and H. G. Richter. Gaseous interchange between adjacent lung lobules. Yale J. Bio. Med. 2: 297-300, 1930. VanAllen, C. M., G. E. Lindskog and H. G. Richter. Collateral respiration. Transfer of air collaterally between pulmonary lobules. J. Clin. Invest. 10: 559- 590, 1941. VanAllen, C. M. and Y. C. Soo. Collateral respiration. Spontaneous reinflation of an atelectatic pulmonary lobule by collateral respiration. J. Clin. Invest. 12: 171-179, 1933. Woolcock, A. J. and P. T. Macklem. Mechanical factors influencing collateral ventilation in human, dog and pig lungs. J. Appl. Physiol. 30: 99-115, 1971. Woolcock, A. J., P. T. Macklem, J. C. Hogg and N. J. Wilson. Influence of the autonomic nervous system on airway resistance and elastic recoil. J. Appl. Physiol. 26: 814-818, 1969. APPENDICES APPENDIX A TYPICAL DATA SHEETS 89 coco. eooo. coco. Hooo. “coo. eooo. Nooo. coco. Hooo. mm a HNNO. NHNO. ammo. ammo. mmmo. acme. mmNO. some. meNO. m «ONO. memo. Heme. ammo. w6~o. “mmo. Home. some. News. NON: sue meNo. wHNo. emNO. GNNO. Hmmo. NNNO. cowo. oeNo. Name. A memo. MHNC. ONNQ. mNNo. eemo. acme. emmo. emwo. eemo. N u a mooo. moco. Nooo. ouoo. Hooo. mooo. Hooo. Nooo. Hooo. mm H mmNo. ease. NmNo. Hmfio. mmNO. HmNO. eNNO. NmNO. hwfio. m HmNO. mmfio. mmNo. moss. meO. NeNo. ommo. mmmc. mafia. .oN: soc emNo. mesa. mmNo. mNHo. oeNo. MHNO. ewmo. Heme. case. 3 HeNo. eofio. mmNO. OONO. mmNo. mmNo. ammo. emNo. mafia. e u a bkwmlnnmnmelnprpH M\mm oNbOH o\H~ O\ooH oNHN m\mm mass. -He.w: see HHH some, HH msoeo H macho H Hfloue moo won mzo mom Hmmzm H < xHszmm< m Dz< m OH AQH NIH mo mmzommmm 91 Hooo. Hooo. Hooo. Hooo. Hooo. Hooo. Hooo. Hooo. Hooo. Nooo. mooo. mm « ooHo. oHHo. ooHo. NNHo. omHo. HNHo. omHo. omHo. HmHo. ooHo. mmHo. m Homo. HNHo. mNHo. HNHo. hmHo. oNHo. NmHo. NmHo. ooHo. ooHo. omHo. HON: so. ooHo. oHHo. HHHo. HNHo. mmHo. NNHo. HoHo. omHo. HoHo. oHHo. mmHo. H moHo. HHHo. mmHo. HNHo. HmHo. mNHo. NmHo. HHHo. hmHo. NHHo. NoHo. o u H Hooo. Hooo. Hooo. Hooo. Hooo. Hooo. oooo. mooo. oooo. Nooo. Nooo. mm a oooo. Nooo. Hooo. oooo. Hooo. oooo. oooo. oooo. Hooo. mHoo. Hooo. m ooHo. Hooo. Hooo. Hooo. Hooo. oooo. Hooo. mHoo. Hooo. mHoo. Hooo. Hon soo Hooo. Hooo. Hooo. Hooo. Hooo. onoo. ouoo. oooo. Hooo. oHoo. Hooo. H mooo. Hooo. ohoo. oooo. Hooo. oooo. HHoo. mooo. oooo. muoo. Ahoo. N u o Hooo. Hooo. mooo. Hooo. Hooo. oooo. Hooo. Hooo. Hooo. Hooo. Hooo. mm a oooo. Nooo. oooo. Hooo. Nooo. Hooo. Hooo. Hooo. Hooo. omoo. mooo. m Hooo. Nooo. Hooo. oooo. Hooo. oooo. Hooo. oooo. Hooo. Hooo. oooo. HON: aoo oooo. Hooo. oooo. Nooo. Hooo. omoo. oooo. omoo. Hooo. omoo. oooo. H Hooo. omoo. omoo. Hooo. Hooo. Hooo. Nooo. oooo. omoo. omoo. Hooo. o u o m Hoom om Hmom mo Hoom om H om mH Hmom oH HHom HoHe.H-Hs.oz ego mm: ES. >mm HHoom HHoom zo >oo Ho HOHHHH mmo ooo mzo moo ommmm H < xHozmoo< APPENDIX B ANOVA TABLES AND RESULTS OF COMPARISON OF MEANS OF ALL DATA 92 93 APPENDIX B ANOVA COMPARISON OF THE EFFECTS OF DIFFERENT GAS MIXTURES ON Rcoll P = 4 cm H O L 2 Source of Variation df SS MS F Group I Dogs 3 .00145 .00048 1542.9* Treatment 2 .00003 .000017 53.7* Interaction 6 .00008 .000013 40.7* Error 24 .000008 .0000003 Total 35 .00157 w = .0012 There is 21/0 and Group 11 Dogs Treatment Interaction Error Total w = .0023 There is 21/0 and 95/5,and 95/5 and 100/0. Group III Dogs 3 .001216 .000405 1206. Treatment 2 .000098 .000049 145. Interaction 6 .000044 .000007 22. Error 24 .000008 .00000034 Total 35 .001366 w = .0003 There is a significant difference between a significant 95/5,and 21/0 .0013 .000049 .000029 .000026 .0014 a significant difference between and 100/0. .000448 408. .000024 22. .000005 4. .000001 difference between 21/0 and 95/5,and 95/5 and 100/0. 94 ANOVA COMPARISON OF THE EFFECTS OF DIFFERENT GAS MIXTURES ON Rcoll P = 2 cm H O L 2 Source of Variation df SS MS F Group 1 Dogs 3 .001502 .000501 934.7* Treatment 2 .000023 .000012 21.9* Interaction 6 .000093 .000015 28.8* Error 24 .000013 .000001 Total 35 .001632 w = .0018 There are no significant treatment differ- ences lsd(.05)= .0012 There is a significant difference between 21/0 and 95/5,and 21/0 and 100/0. Group II Dogs 3 .001502 .000501 33l.3* Treatment 2 .000073 .000036 24.0* Interaction 6 .000037 .000006 4.1* Error 24 .000036 .000002 Total 35 .001648 w = 0025 There is a significant difference between 21/0 and 95/5,and 95/5 and 100/0. Group III Dogs 3 .001370 .000457 493.5* Treatment 2 .000126 .000063 68.0* Interaction 6 .000038 .000006 6.9* Error 24 .000022 .000001 Total 35 .001556 w = .0018 There is a significant difference between 21/0 and 95/5,and 95/5 and 100/0. 95 ANOVA COMPARISON OF THE EFFECTS OF LIKE GAS MIXTURES ON Rcoll PL = 4 cm H20 Source of Variation df SS MS F Group A Dogs 3 .00157 .00052 868.3* Treatment 2 .00002 .00001 16.5* Interaction 6 .00008 .00001 23.2* Error 24 .00001 .000001 Total 35 .00169 w = .0018 There are no significant treatment differ- ences. lsd(.05) = .0012 There is a significant difference between 21 d 2 d 2 d 21/g(1) an l/Ocz),an 1/0(1) an (3) Group B Dogs 3 .00093 .00031 404.9* Treatment 2 .00005 .00003 34.2* Interaction 6 .00006 .00001 12.3* Error 24 .00002 .000001 Total 35 .00105 w = 0018 There is a significant difference between 5 . 9 /5(1) and 95/5(3) Group C Dogs 3 .00152 .00051 1321.2* Treatment 2 .000001 .0000005 1.5 Interaction 6 .000011 .000002 4.8* Error 24 .000009 .0000004 Total 35 .00154 96 ANOVA COMPARISON OF THE EFFECTS OF LIKE GAS MIXTURES ON Rcoll PL = 2 cm H20 Source of Variation df SS MS F Group A Dogs 3 .00164 .00055 274.7* Treatment 2 .00001 .000003 1.7 Interaction 6 .00004 .000007 3.6* Error 24 .00005 .000002 Total 35 .00174 Group B Dogs 3 .00118 .00039 l421.2* Treatment 2 .00008 .00004 152.3* Interaction 6 .00007 .00001 41.2* Error 24 .00001 .0000003 Total 35 .00133 w = .0010 There is a significant difference between 95/5C1) and 95/5(2), 95/5(1) and 95/5(3), and 95/5(2) and 95/5(3). Group C Dogs 3 .00152 .00051 723.5* Treatment 2 .00001 .000005 7.8* Interaction 6 .00009 .00001 21.0* Error 24 .00002 .000001 Total 35 .00164 w = .0018 There are no significant treatment differ- ences. lsd(.05) = .0012 There is a significant difference between loo/0(2) and 100/0(3). 97 ANOVA THE EFFECT OF va ON Rcoll P = 4 cm H O L 2 Source of Variation df SS MS F va = 5(1) through va = 20 mm Hg Dogs 10 .00227 .00023 71.5* Treatment 5 .00003 .00001 2.12 Error 50 .00016 .000003 Total 65 .00246 va = 5(4) through va = 5(6) mm Hg Dogs 9 .00242 .00027 52.6* Treatment 4 .00004 .00001 2.19 Error 36 .00018 .00001 Total 49 .00265 98 ANOVA THE EFFECT OF va ON Rcoll P = 2 cm H O L 2 Source of Variation df SS MS F va = 5(1) through va = 20 mm Hg Dogs 10 .00570 .00057 68.0* Treatment 5 .00007 .00001 1.7 Error 50 .00042 .00001 Total 65 .00619 va = 5(4) through va = 5(6) mm Hg Dogs 9 .00622 .00069 41.6* Treatment 4 .00016 .00004 2.44 Error 36 .00060 .00002 Total 49 .00698 99 ANOVA THE EFFECT OF va ON Rcoll PL = 0 cm H20 Source of Variation df SS MS F va = 5(1) through va = 20 mm Hg Dogs 5 .00606 .00121 178.9* Treatment 5 .00006 .00001 1.7 Error 25 .00017 .00001 Total 35 .00628 va = 5(4) through va = 5(6) mm Hg Dogs 3 .00251 .00084 312.9* Treatment 4 .00007 .00002 6.5* Error 12 .00003 .000003 Total 19 .00262 lsd(.05) = .0023 There is a significant difference between 5(5) and 25 mm Hg, va = 5(5) and 30 mm Hg and va = 5(6) and 30 mm Hg. 100 ANOVA THE EFFECT OF TIME ON THE IPL ALL va's = 5 mm Hg Source of Variation df SS MS F va = 5(1), 5(2), 5(3) mm Hg Dogs 27 870426.0 32238.0 43.0* Treatment 2 3496.1 1748.1 2.3 Error 54 40517.2 750.3 Total 83 914438.7 va = 5(4), 5(5), 5(6) mm Hg Dogs 23 997252.0 43358.8 41.6* Treatment 2 4637.5 2318.8 2.22 Error 46 47947.8 1042.34 Total 71 1049837.3 101 ANOVA THE EFFECT OF va ON RV Source of Variation df SS MS F Group I Dogs 8 22456.6 2807.1 69.6* Treatment 5 865.9 173.2 4.3* Error 40 1612.4 40.3 Total 53 24934.9 lsd(.OS) = 5.0 Group 11 Dogs 7 9899.9 1414.3 11.8* Treatment 4 1738.1 434.5 3.6 Error 28 3362.3 120.1 Total 39 15000.3 lsd(.05) = 9.4 102 ANOVA THE EFFECT OF va ON CL Source of Variation df SS MS F Group I Dogs 9 36.3 .0 2.8* Treatment 5 19.7 3.9 2.7* Error 45 65.0 1 4 Total 59 121.0 w = 1.6 Group 11 Dogs 7 55.8 .0 26.6* Treatment 4 .7 .2 3.9* Error 28 8.4 3 Total 39 68.9 w = 0.8 Note: In both ANOVA's, the significant treatment F indi- cates there is a significant difference between means within the Groups. However, in Group I there are no significant differences between adjacent va's, and in Group II the only significant difference between adjacent va's is between va = 5(5) and 30 mm Hg. 103 ANOVA THE EFFECT OF va on CH Source of Variation df SS MS F Group 1 Dogs 9 4.5 0.5 16.7* Treatment 5 0.5 0.1 3.3* Error 45 1.4 0.03 Total 59 6.5 w = 0.23 There is a significant difference between va = 5(3) and 20 mm Hg. Group 11 Dogs 7 1.9 0.3 19.3* Treatment 4 1.4 0.4 25.0* Error 28 0.4 -.014 Total 39 w = 0.16 There is a significant difference between the means Of all adjacent va's. 104 ANOVA THE EFFECT OF va ON Rvas Source of Variation df SS MS F PL = 4 cm H20 Dogs 9 .009 .00100 12.5* Treatment 4 .003 .00075 9.5* Error 36 .003 .00008 Total 49 .015 P = 2 cm H O L 2 Dogs 9 .008 .00089 16.8* Treatment 4 .003 .00075 12.0* Error 36 .002 .00006 Total 49 .0013 P = 0 cm H O L 2 Dogs 3 .003 .00100 20.2* Treatment 4 .002 .00050 8.6* Error 12 .001 .00008 Total 19 .005 105 Results of SNK Test Comparing Rvas Means (P< .05) P H 5 -10 5 -15 5 -20 5 -25 5 -30 P" ("m g) (1) (2) (3) (4) (5) 3(- (PL = 4) -0.017 -0.031 -0.035 -0.037 -0.038 3; (PL = 2) -0.016 -0.026 -0.031 -0.032 -0.037 R (P = O) -0.016 -0.032 -0.039 -0.040 -0.039