Respiration: Proceedings of the 28th International Congress of Physiological Sciences, Budapest, 1980

Debreczeni

Trends in respiratory research

MECHANICS OF THE CHEST WALL

Jere Mead,      Department of Physiology, Harvard School of Public Health, 665 Huntington, Avenue Boston, Massachusetts 02115, USA

Publisher Summary

This chapter discusses the mechanics of the chest wall. Diaphragmatic displacements can be estimated from abdominal volume displacements. The wall of the abdomen includes a portion that is ordinarily not regarded as abdominal at all. This is a part of the rib cage beginning at its costal margin and extending cephelad to the diaphragmatic reflection. In this region, the diaphragm is directly apposed to the rib-cage interior, and beneath lies not lung but abdominal contents. The study described in the chapter refers to this region as the area of apposition. At residual volume, it covers approximately half of the internal surface of the rib cage, decreasing to become a narrow ring at total lung capacity—as the diaphragm shortens, the area of apposition decreases, and, finally, virtually disappears. As it decreases, it expands outward and moves upward, and both of these motions need to be taken into account in estimating abdominal displacements. The total volumetric expansion of the rib cage can be estimated by the method of Konno and Mead. The fraction of its surface taken up by the area of apposition times this displacement is an estimate of the outward displacement of the area of apposition.

Rohrer (8) and later Fenn and his associates (7), started modern respiratory mechanics on its way by a process of simplification: they integrated the complexity of respiratory movements into a single variable, volume, and force into a single variable, pressure. Their respiratory system had just two parts: the lungs and chest wall, operating in series. Thus, two pressures, transthoracic and transpulmonary, and one volume, sufficed to describe the system.

Agostoni and Rahn (2) divided transthoracic pressure into transdiaphragmatic and transabdominal components, and Agostoni and Mead (1) applied these to diagrams of the respiratory system which showed the lungs in series with a two-pathway chest wall: the rib-cage providing one pathway and the diaphragm-abdomen the other, the two pathways operating parallel. But the volume scales on these diagrams were largely hypothetical and this stimulated Konno and myself (5) to seek some way to measure them. We displayed the anterior-posterior diameters of the rib-cage and abdomen X-Y, and found that when we held lung volume constant, the two diameters were tightly coupled: we could readily increase the A-P diameter of the rib-cage, if we simultaneously decreased that of the abdomen, and vice versa. When these iso-volume maneuvers were done at different volumes essentially parallel lines were developed, which, by suitable gain adjustments, could be made to have a negative slope of unity. We interpreted the tight coupling at constant volume as indicating that the chest wall had just two ways to change volume--i.e. by rib-cage or by abdominal expansion. Indeed, we found that if we were careful to maintain a constant spinal attitude it was possible to recover lung volume changes by appropriate sums of these two signals over a wide range of ventilations.

We had set out to provide Agostoni and Mead’s two pathway model of the chest wall with better measurement of the pathway volume displacements. For several years I thought we had succeeded. I thought that our estimates of the volume displacements of the rib-cage and abdomen at the body surface were all that were needed. I now appreciate that they were not sufficient. Abdominal displacements, as we measured them, did not give an adequate measure of diaphragmatic displacements.

I will show you that diaphragmatic displacements can, indeed, be estimated from abdominal volume displacements, but our idea of what constitutes the abdomen has to be extended. The wall of the abdomen includes a portion that we ordinarily do not regard as abdominal at all. This is part of the rib-cage beginning at its costal margin and extending cephelad to the diaphragmatic reflection. In this region the diaphragm is directly apposed to the rib-cage interior, and beneath lies not lung, but abdominal contents. I refer to this region as the area of apposition (6). At residual volume it covers approximately 1/2 of the internal surface of the rib-cage, decreasing to become a narrow ring at Total Lung Capacity. That is, as the diaphragm shortens the area of apposition decreases, and, finally, virtually disappears. As it decreases it expands outward and moves upward (i.e. toward the head), and both of these motions need to be taken into account in estimating abdominal displacements. The area of apposition forms a truncated cone–rather like a somewhat bent lampshade–which, as the diaphragm contracts, and the rib-cage expands, becomes shorter and wider, and tips up slightly in front. In terms of volume displacement it behaves like an expanding Krogh spirometer. The walls of the spirometer are formed by the rib-cage and hence are constrained to move with it, and this provides a way to measure it: We can estimate the total volumetric expansion of the rib-cage by the method of Konno and Mead. The fraction of its surface taken up by the area of apposition times this displacement is an estimate of the outward displacement of the area of apposition. For example, at FRC, in the upright posture, the area of apposition makes up about 1/4 of the rib-cage surface, and approximately 1/4 of its volume expansion should be reassigned to the abdomen.

The axial component can be estimated as the product of the cross-section at the level of the costal margin and a fraction (between 0.5 and 1.0 depending on the location of the hinge point) of the change in the xiphi-pubic distance. The axial component (at fixed spinal attitude) is totally dependent on the rib-cage motion and hence may also be represented as a fraction of the rib-cage excursion.

The sum of the fractions related to the upward and outward components times the rib-cage volume changes gives the total that should be assigned to rib-cage abdominal wall expansion. Next I relate this to the other abdominal displacements.

Since the abdominal contents are essentially incompressible, the sum of all abdominal wall volume displacements must be 0, and any single displacement must be equal and opposite to the sum of the others. Thus, displacement of the diaphragm dome–the pathway displacement we are seeking–is equal (and opposite) to the sum of the rib-cage and anterior-wall displacements. In terms of the Konno-Mead measurements, then, the diaphragm volume displacement is the abdominal volume displacement plus a fraction of the rib-cage volume displacement.

Let us see where we stand relative to our original objective. We set out to make better measurements of volume displacement for the Agostoni-Head model–which we did–but then found that the model needed modification. The Konno-Mead measurements were correct for the rib-cage and the anterior abdominal wall–but, as we have seen, not for the diaphragm. Only at TLC, when the area of apposition virtually disappears, is the original model approximately correct. At all other volumes, the rib-cage forms part of the abdominal wall and has the potential for sharing diaphragmatic displacements. Konno and Mead have estimated that during quiet breathing in the upright posture, the diaphragm accounted for about 1/4 of the total volume change. We think that its actual contribution is probably about twice that. Indeed, in rib-cage only inspirations–i.e. ones which the abdominal A-P diameters and cross-section remain fixed, the diaphragm shortens substantially. Indeed, for the diaphragm not to shorten during inspirations the anterior abdominaJ wall has to move inward.

This modification of the Agostoni-Mead model required rethinking the implications of pressures as well. For example, note that the rib-cage is not simply subjected to transthoracic pressure. It is subjected in part to this pressure--i.e. over the lung-apposed internal surface--and in part to something approximating transabdominal pressure. (The pressure difference across the diaphragm wall in the area of apposition is not known and needs to be measured. To the extent that its diaphragmatic tension is mainly axial and its radius of curvature in that direction large, this difference would be small. To the extent that its tension is isotropic, appositional transdiaphragmatic pressure would be about 1/2 that at the dome.)

We have developed a simple expression of the force balance for the rib-cage which neglects details but shows how these pressures act and brings up some potentially important relationships. We take total rib-cage surface area as unity and f as the fractional area of apposition. As a first approximation the force applied to the rib cage, by transthoracic pressure is (1-f) Ppl and by transabdominal pressures, fPab. These forces are equally apposed by forces developed by the rib-cage, which must arise from intrinsic passive properties, or extrinsic passive properties–i.e. forces from attached structures such as the diaphragm and accessory muscle tendons, or from their active counterparts: intercostals intrinsically and diaphragm and accessories extrinsically. Here we will lump the passive and active forces from the diaphragm together as Fdi, the passive properties of the rib-cage, intercostals and accessories as Frc, and the active properties of the intercostals and accessories, together as Fic/a. In the relaxed state Fic/a = 0; the total force balance is then

Eq. (1)

When the diaphragm is not under tension Ppl = Pab and Fdi = 0. The force balance then simplifies to Ppl = -Frc, and transthoracic pressure is then totally opposed by the intrinsic passive properties of the rib-cage. This is the classical view of the relaxation state of the rib-cage.

When the diaphragm is under tension, however, three additional influences come into play, as evidenced by the relaxation force balance solved in terms of transthoracic pressure.

Eq. (2)

Note that the capacity of each of the terms in the brackets on the right to modify transthoracic pressure is multiplied by 1/(1-f), which varies from 1 at TLC to approximately 2 at RV. It is known that in terms of transthoracic pressure, the rib-cage becomes increasingly stiff at low volume. Part of this apparent stiffness arises simply from the multiplier [(1/(1-f)]. An additional part arises from Fdi, which increases as the diaphragm is passively stretched.

Finally, the third influence is fPab. Increasing Pab, per se, makes transthoracic pressure, Ppl, increasingly negative; that is, it lowers pleural pressure. Thus, Pab drives the rib-cage--the more so the larger the area of apposition.

I would like to relate this action of abdominal pressure on the ribcage to the ideas of Goldman and Mead (3) on this subject. They were the first to demonstrate that increasing abdominal pressure exerts an inspiratory influence on the rib-cage (although this result was, in essence, anticipated by Magendie and by Duchenne more than a century before). Goldman and Mead increased transabdominal pressure by means of external compression and obtained a beguiling result: pleural pressure fell by exactly the amount that transdiaphragmatic pressure increased. They knew that when transdiaphragmatic pressure was 0, transabdominal pressure and transthoracic pressure equally well represented the pressure developed by the rib-cage. If as transdiaphragmatic pressure developed, it was directly additive to transthoracic pressure, then the underlying passive pressure developed by the rib-cage, freed from the influence of the diaphragm, would be obtained by subtraction of transdiaphragmatic pressure from transthoracic pressure, which by simple algebra is transabdominal pressure. Thus,

Eq. (3)

This was the beguiling part; if the action of the diaphragm on the rib-cage was always that observed during passive induction of transdiaphragmatic pressure, then it would increase the rib-cage volume only to the extent that it increased abdominal pressure–driving it along its hidden passive characteristic –and would drive the abdomen with exactly the same agency. In short, it would drive the entire system along its relaxation characteristic in the same way as would an external pressure applied to the entire respiratory system. The diaphragm, in other words, could do it all.

How do these ideas fit the force balance expression?

The force applied to the rib-cage by the diaphragm’s insertions must vary with diaphragmatic tension and, hence, with transdiaphragmatic pressure. Fdi can then be replaced by k(Pab - Ppl), where k includes the pertinent geometric features. k will generally be positive in sign–that is, Fdi will tend to expand the rib-cage–and less than one. For example, if the cross sections of the rib-cage at the level of insertions were equal to the surface area of the rib-cage, and if the force vectors were ideally directed to expand the rib-cage, k would be unity.

Substitution of k(Pab - Ppl) for Fdi in the force balance expression rearranges to the following:

Eq. (4)

But (Ppl - Pab) is transdiaphragmatic pressure, Pdi, and accordingly

Eq. (5)

Note that f, the fractional area of apposition, and k, are both less than one. In a particular circumstance, namely when the sum equals one, the expression becomes

Eq. (6)

which is the condition observed by Goldman and Mead during abdominal compression. Thus

Eq. (7)

A possible explanation of their observation is that abdominal compression displaced the diaphragm into the rib-cage and increased the area of apposition sufficiently so that (f + k) approached unity. But this analysis also suggests that this relationship would not be expected to hold during ordinary breathing where f is smaller, and, furthermore, decreases with inspiration.

What are realistic values for (f + k) during active contractions of the diaphragm? All the other quantities in the force balance expression are measurable. Frc, the passive pressure developed by the rib-cage, can be measured during voluntary relaxation, if the diaphragm is not under tension, when, as we have seen

Eq. (8)

Figure 1 shows preliminary results in two subjects in the upright posture. The solid lines indicate -Frc as a function of rib-cage volume. Ppl and Pab, measured respectively with esophageal and gastric balloon catheters, are shown at the end of voluntary inspirations in which the subjects consciously emphasized diaphragmatic contractions while minimizing all other respiratory muscle activity. (They were assisted in this by feedback from diaphragm and parasternal EMG signals).

Fig. 1

Graphically, (f + k) is the distance between the active Ppl and the passive one, -Frc, divided by the total distance from Ppl to Pab. values are substantially less than unity.

These admittedly preliminary results suggest that Goldman and Mead were overly hopeful in their interpretation of the diaphragm’s action. It does have an inspiratory action on the rib-cage which is greatest when it is passively stretched. But it loses its hold progressively as it shortens.

In my brief development of these ideas about pathways for volume displacement and diaphragmatic action I have left some loose threads. I could spend my remaining time picking some of them up. Instead I will extend the fabric to include some of the implications of these ideas to breathing.

The respiratory pump is a bulky object which must share the body’s facilities with a minimum of mutual embarrassment. We use respiratory muscles to walk, talk, bend, twist, strain on lifting and at stool, cough, sneeze–and give lectures. If the non-respiratory acts are brief, breathing is merely suspended, and we run on our reserves. If the nonrespiratory acts are protracted, accommodations are struck.

Clearly, there are many ways to breathe. Carrying a heavy shopping bag in your arms requires a certain degree of rib-cage fixation, but there is plenty of room for increasing ventilation by the abdominal pathways. A tight corset which does not restrict the lower rib-cage, again, leaves plenty of room for breathing.

How are these accomodations brought about? The concept of the ribcage as an element of the abdominal wall is informative on this account. At the beginning of the rib-cage force balance development, I introduced the intercostal and accessory muscles, and then dropped them out to concentrate on passive features, and, later on diaphragmatic action. I now bring them back to make a simple point. Inspiratory action of these muscles on the rib-cage lowers both abdominal and pleural pressures. In contrast, inspiratory action of the diaphragm lowers pleural pressure, but increases abdominal pressure. Clearly, then, coordinated action of the inspiratory muscles can produce inspirations which will not change abdominal pressure. As the diaphragm descends in such inspirations, the expanding rib-cage makes room for its displacements, and the abdomen is, as it were, spared. And this coin has an opposite side: changes in abdominal pressure from non-respiratory causes--such as abdominal distension in pregnancy or feeding, or abdominal muscle tensing associated with locomotion, have little influence on breathing. Increased transabdominal pressure is simultaneously expiratory to the diaphragm and inspiratory to the rib-cage. In the net, this respiratory effect is small.

This takes care of the tight corset accomodation. What about the shopping bag problem? How do we accomodate restrictions of rib-cage motion? My answer will be somewhat indirect.

We all got out of bed this morning. In this act we allowed gravity to put our diaphragms at a mechanical disadvantage; the weight of our abdominal contents pulled down on the diaphragm and shortened it. Figure 2 gives you an idea of how embarrassing this would have been to you if you hadn’t done something about it. This tracing was obtained on an individual who was powerless to do anything about it. She has a complete cord section rostral to the phrenic outflow and her diaphragm is driven by bilateral phrenic pacers, and a fixed pattern of stimulation. When she was tipped full upright, you can see that her tidal volume was cut by two-thirds.

Fig. 2

When we assume the upright posture, we automatically and completely adjust the neural drive to our respiratory muscles to accomodate what would otherwise be a severe embarrassment to respiration.

I have emphasized the favorable arrangement of the chest wall for meeting abdominal loads, and Figure 3A is an example of this.

Fig. 3

When abdominal excursions were restricted with a mechanical clamp, minute ventilation was uninfluenced, as was the intensity of diaphragm stimulation as estimated from EMG recordings. When (Figure 3B) the rib-cage was similarly restricted, again ventilation was uninfluenced, but this time the intensity of diaphragmatic stimulation was increased. With rib-cage restriction the diaphragm shortened more for a given volume change and, accordingly, required greater stimulation to achieve the same tidal exchange. Just as in the case of assuming the upright position, the cost is paid easily and completely.

How do we do it? What tells us to adjust excitation appropriately to compensate for the mechanical disadvantage incurred by changes in the operating length of our muscles? Firstly, the compensation does not appear to be chemical. Blood gases are not changed. Secondly, whatever the afferent information, apparently it need not come to consciousness. Paraplegics with cord lesions below the level of phrenic outflow, who had their diaphragm shortened by means of negative pressures applied by cuirass at the abdominal surface, showed brisk and completely adequate increases in diaphragmatic excitation. These subjects had no knowledge that their breathing was being studied and they experienced only trivial and inconsistent sensations during the application of pressure. Presumably, afferent signals by way of either the vagus or phrenic nerves were sufficient for the response. That is as far as we can go in terms of mechanisms at this time. We have elicited the response in conscious monkeys and in a cow, but we have not yet blocked the vagi in the presence of the response.

I began this journey down the pathways of the chest wall in Milan with Emilio Agostoni as we tried to write a chapter for the American Handbook for Respiratory Physiology the week following the International Congress in Leiden, and I would like in closing to list my travelling companions who form the everchanging we of this discourse. More or less in order of non-appearance these have included K. Konno, R. Kellog, D. Leith, T. Sears, M. Goldman, M. Green, J. Gillespie, E. Bruce, G. Inbar, R. Banzett, P. Koch, K. Strohl, J. Smith, and S. Loring. I am pleased that the last named now faces the task of writing a chapter on chest wall mechanics for a new edition of the American Handbook for Respiratory Physiology. I am quite sure he will find it a stimulating experience.

REFERENCES

1. Agostoni, E., Mead, J.Fenn W.O., Rahn H., eds. Statics of the respiratory system. Handbook of Respiration. American Physiological Society: Washington, D. C., 1964; 387–409.

2. Agostoni, E., Rahn, H. Abdominal and thoracic pressures at different lung volumes. J. Appl. Physiol. 1960; 15:1087–1092.

3. Goldman, M.D., Mead, J. Mechanical interaction between the diaphragm and rib cage. J. Appl. Physiol. 1973; 35:197–204.

4. Green, M., Mead, J., Sears, T.A. Muscle activity during chest wall restriction and positive pressure breathing in man. Respir. Physiol. 1978; 35:283–300.

5. Konno, K., Mead, J. Measurements of the separate volume changes of rib cage and abdomen during breathing. J. Appl. Physiol. 1967; 22:407–422.

6. Mead, J. Functional significance of the area of apposition of diaphragm to rib cage. Am. Rev. Respir. Dis. 1979; 119:31–32.

7. Rahn, H., Otis, A.B., Chadwick, L.E. The pressure-volume diagram of the thorax and lung. Am. J. Physiol. 1946; 146:161–178.

8. Rohrer, F. Der Zusammenhang der Atemkrafte und ihre Abhangigkeit vom Dehnungszustand der Atmungsorgane. Arch. f. d. ges. Physiol. 1916; 165:419.

THE PATHOPHYSIOLOGY OF AIRWAYS DISEASE

Gordon Cumming,      Midhurst Medical Research Institute, Midhurst, West Sussex, UK

Publisher Summary

This chapter focuses on the pathophysiology of airways disease. In case of bronchial asthma, the functional defects of airflow limitation and gas mixing deficit are combined and the severity of each does not correlate with the other. The extent of mixing deficit is often gross, and the size of this deficit is difficult to explain on the basis of regional distribution as produced by the large airways, although it is in these airways that flow limitation has its origin. The chapter reviews the situation seen in centrilobular emphysema where a small dilatation occurs at the level of the respiratory bronchiole and in ventilatory terms; this is identical with bronchial asthma. Pan acinar emphysema, in which there is complete destruction of the architecture of the terminal unit and which also has a mixing defect is probably explained by the long diffusive path length produced by the dilated sac, which may be several centimeters in length. The chapter also discusses the common disease of chronic bronchitis, in which cigarette smoking plays a large part in etiology.

There is a group of chest diseases in which dyspnoea is associated with limitation of expiratory airflow and it is to the treatment of this impaired airflow that therapy is mainly directed and to which the drug industry devotes much attention.

Implicit in this attitude is the assumption that the characterising feature of the group of diseases is similar and that it is only limitation of expiratory airflow. In addition there is a defect of gas mixing, and the conventional explanation is that it results from a failure of gas distribution by the airways. There is also a further assumption, which is that the process of as mixing by diffusion within the terminal ventilatory unit is complete within the time of a respiratory cycle.

So we have a unitary hypothesis which explains both airflow limitation and gas mixing by the same mechanism and this represents the conventional wisdom.

It follows from this hypothesis that the more severe the airflow, limitation, then the greater must be the defect in gas mixing so that measures of both in a series of patients would show a good correlation. Such a study carried out in my laboratory is shown in Figure 1 where FEV1.0 is used as a measure of airflow limitation and alveolar mixing efficiency for nitrogen is also plotted. As you see the correlation is poor, and the correlation coefficient calculated was 0.06.

FIG 1

This observation denies the conventional unitary hypothesis and suggests that airflow limitation and gas mixing are independant variables. Since this is so a new hypothesis for the genesis of impaired mixing is called for, and this hypothesis must not involve the function of the large airways.

It is possible therefore:-

(a) that it is concerned with the small airways and

(b) that it is concerned with the assumption made in the unitary hypothesis that gas mixing by diffusion is complete within one respiratory cycle.

Understanding the process of gas mixing is made easier by considering the situation which obtains when a subject inspires pure oxygen and the nitrogen concentration in the expirate is measured, giving a trace such as is shown in Fig. 2(a) in which expired nitrogen concentration is plotted against expired volume. From such a trace the slope of the plateau can be measured and an increase in this slope has been used for clinical diagnostic purposes and the explanation for its origin has used regional ventilation as its basis.

FIG 2 METHOD OF CALCULATION OF VOLUME OF NITROGEN

More information can be obtained from the same data if the quantity of nitrogen expired is calculated. Figure 2(b) indicates that expired flow is measured at the same time as expired concentration. The two curves are sampled by a computer as indicated in Figure 2(a) and (b), and each value of concentration is multiplied by each value of flow, the product indicating the quantity of nitrogen evolved during the short sampling period.

When all the small quantities of nitrogen are added, the pattern of nitrogen evolution with respect to expired volume may be seen in Figure 2(c). Figure 3 shows the individual data points and Figure 4 the graphic representation, and from this we can draw several conclusions - firstly the intercept of the line on the bascissa defines a dead space which we have called the series dead space and which is similar to that described by Fowler. Secondly the slope of the line indicates the mean alveolar concentration, and thirdly the total volume of nitrogen evolved in the whole breath is measured.

FIG 3 INDIVIDUAL DATA POINTS FROM A SINGLE EXPIRATE FOLLOWING AN INSPIRATE OF OXYGEN

FIG 4 APPLICATION OF LINEAR REGRESSION ANALYSIS TO DATA POINTS

It is now possible to calculate what quantity of nitrogen would be evolved from a particular lung if gas mixing within it were perfect and this calcuation, which is commonly used in physiology, is shown in Figure 5.

FIG 5 NITROGEN RECOVERY FOLLOWING A SINGLE BREATH OF OXYGEN

NITROGEN RECOVERY FOLLOWING AN INSPIRATE OF OXYGEN

The lung volume contains a mixing volume VA and a non-mixing volume VDS such that VL = VA + VDS. Following the inspirate, VA increases in volume by VT, the new volume being VL + VT. The volume of oxygen added to this volume isVT - VDS, so that the dilution ratio is:

If nitrogen concentration is C0 the new concentration C1 is

During expiration, Vn will retain its volume of nitrogen and excrete a volume VT - VDS of gas at concentration C1 or

Inserting the appropriate values.

We now have two pieces of information - the volume of nitrogen recovered from the lung by experiment and the volume predicted if gas mixing were perfect. If we divide one by the other and multiply by 100 we have the alveolar mixing efficiency expressed as a percentage.

The values for this efficiency are shown in Figure 6. Here is shown the theoretical line, the values seen in normal subjects and the values seen in disease. The last line on the figure indicates the effect on nitrogen recovery of a regional maldistribution of inspired gas using the data of West in his ten compartment model. It can be seen that distribution of ventilation can account for only a small part of nitrogen retention. In recent discussions with John West he agreed that there are other causes of impaired mixing in addition to regional distribution of ventilation.

In normals such distribution cannot explain mixing defects completely, and in disease it is even more difficult to use the parallel hypothesis to explain an efficiency of perhaps 25%.

Before embarking on an alternative explanation it is necessary to make some further remarks about the nature and the meaning of series dead space.

In the test previously described, pure oxygen flows down the bronchial tree, its linear velocity becomes progressively less as the area of cross section of the bronchial tree increases. In the trachea during quiet inspiration the velocity is about 40 cms. per second, falling in the terminal bronchiole to about 1 cm/sec and beyond this falling further as branching in the terminal ventilatory unit continues. At the same time as convective flow bears oxygen molecules forward, and with them the nitrogen molecules with which they are in contact, the nitrogen molecules are moving upwards towards the mouth due to gaseous diffusion across the concentration gradient. This upwards movement may be regarded as a velocity.

We now have two phenomena occurring at the same time - a velocity of nitrogen upwards which is fixed and determined by the diffusion characteristics of the gases, and a velocity downwards which is progressively diminishing brought about by bulk flow.

It is clear that at some point within the terminal ventilatory unit these two velocities will become equal and the bulk flow of nitrogen downwards will be exactly balanced by the diffusive flow upwards. At this anatomical site the interface between oxygen and nitrogen becomes stationary and remains so whilst inspiratory flow continues at the same rate. Beyond the stationary interface therefore, oxygen molecules gain access to the ventilatory unit only by diffusion and the rules governing gas mixing are the rules of gaseous diffusion.

Before these rules are discussed, it is important to understand the role played by bulk flow in the process of ventilation. This can best be done by considering the situation of the lung which is full of nitrogen and into which nitrogen is inspired. Since there is no concentration difference between the inspired and the resident gas there is no mixing at all, yet alveoli expand and inspire gas, which is entirely nitrogen. This volume represents the ventilation of the lung by bulk flow.

When the inspired gas is oxygen and diffusion plays its role bulk flow into alveoli occurs in exactly the same way as in the case of nitrogen. The oxygen however gains access only after diffusion and if bulk flow is large, the stationary interface moves distally and ventilation in that unit is larger but still determined by the balance between diffusive forces and bulk flow forces.

Let us now enquire about the rules which govern diffusive mixing, since events distal to the stationary interface are dominated by this phenomenon.

Three main variables are involved in diffusive mixing - (1) the diffusion coefficient between the two gases. (2) the distance over which diffusion takes place, or the diffusion path length and (3) the area of contact between the two gases - the interface area.

In normal breathing we are not concerned with the diffusion coefficient since it remains the same in health and disease so there remains the two determinants of diffusion path length and interface area. Both are determined by the geometry of the lung and by the nature of the inspiratory manouvre. When these two variables have been carefully studied using mathematical models the conclusion reached by nearly all authors is that equilibrium is almost complete within a few seconds and significant nitrogen retnetion could not be explained by a failure of diffusive mixing. Nevertheless, experiment shows that it does occur and we now see that neither the regional hypothesis nor the diffusion hypothesis appears to be capable of explaining the common observation.

When one arrives at such a position in which all known hypotheses appear to have been falsified it is useful to re-examine the original assumption upon which the disproof rests.

In the mathematical analysis of diffusion the anatomical information used has been that of a symmetrically branching structure with six generations of branching beyond the terminal bronchiole. It is now possible to look at the known anatomical information and to test the validity of the assumptions.

Figure 7 shows some generally agreed values for the branching of the airways and from which we can conclude that on each respiratory bronchiole of the third order there are about 80 alveolar ducts. The symmetrical three generation model calls for 14.

FIG 7

Figure 8 shows the agreed non-symmetrical structure distal to an RB3, a mean number of orders of six, a minimum of two and a maximum of eleven. If we assume that the structure is symmetrical and there are eleven branches then the number of ducts is 2047. Thus we see that assumptions about anatomy, particularly with regard to its symmetry or otherwise, can make an enormous difference to the assumed geometry between 14 and 2000 ducts.

FIG 8

What is required therefore is a structure for the terminal ventilatory unit which satisfies the known anatomical facts and then to draw some conclusion about its function. There are an infinite number of equivalent structures but the main features are shown in Figure 9. This structure has the right number of ducts and the correct degree of asymmetry and the first functional question to be asked is what is the position of the stationary interface.

FIG 9

Fortunately it is possible to measure this in terms of volume since during expiration the stationary interface is expelled at the lips and produces the rapid upstroke of the expired concentration curve. The volume of the series dead space therefore represents the position of the stationary interface within the lung volume. What this means in anatomical terms is rather more difficult but on average it will be formed towards the end of the third order respiratory bronchiole, in other words at the entrance of the structure depicted on this slide.

We have seen that the ventilation of the unit will be determined by bulk flow, but that mixing within it will be determined by diffusion path length and interface area. Since all pathways have similar diameters, the major determinant of mixing efficiency will be diffusion path length - in other words the distance between the stationary interface and the anatomical point considered.

Inspection of this structure will convince you that there are a great variety of path lengths within it, varying from 2 mm for the shortest to 14 mm for the longest. Consequently there will be within the terminal unit a variety of times of equilibration, and in all probability a failure of complete mixing along the longer pathways.

As we have seen the alveolar efficiency of the normal lung is high, over 85% and this small inefficiency might well be explained on this anatomical basis.

How does this hypothesis work when it is applied to disease. Those diseases which manifest mixing defect include asthma, bronchitis and emphysema - so that all possess the same functional defect of impaired mixing, in addition to sharing in common airflow limitation and dyspnoea.

To function effectively in explaining disease the anatomical hypothesis must show how increased diffusion path length and interface area are produced by the disease process.

Taking first the case of bronchial asthma in which the functional defects of airflow limitation and gas mixing deficit are combined, and the severity of each does not correlate with the other. The extent of mixing deficit is often gross, and the size of this deficit is difficult to explain on the basis of regional distribution as produced by the large airways, although it is in these airways that flow limitation has its origin.

We must turn to another aspect of asthma for the necessary clue. In asthmatic patients who are killed in accidents and who come to autopsy it is common to find small airways of about 1 mm diameter completely occluded by solid mucous secretion and such blockage is diffuse and a good proportion of airway are involved.

Taking this evidence together with the mucous spirals which are coughed up towards the end of an asthmatic attack it seems likely that blockage of small airway is a common feature in asthma.

Ventilation distal to such blocked airways is by collateral channels, as represented in Figure 10. In such a system the stationary interface is formed in the patient airway, whilst ventilation of the distal airway is (a) along a much longer diffusion pathway and (b) has a very much smaller interface area, and these conditions are as we have seen, those which produce an impairment of gas mixing.

FIG 10

Thus the gas mixing efficiency in bronchial asthma becomes a measurement of the quantity of lung in which there are occluded small airways. It is a natural corollary of this statement that treatment such as steroids which is effective in opening up closed airways, can be monitored by measuring gas mixing efficiency.

Figure 11 illustrates the situation seen in centrilobular emphysema where a small dilatation occurs at the level of the respiratory bronchiole and in ventilatory terms this is identical with bronchial asthma and the mixing defect seen in this disease is explained in the same way.

FIG 11

Pan acinar emphysema, in which there is complete destruction of the architecture of the terminal unit, and which also has a mixing defect is probably explained by the long diffusive path length produced by the dilated sac, which may be several centimetres in length.

Finally we come to the common disease of chronic bronchitis, in which cigarette smoking plays such a large part in aetiology. Here the situation is again of small airway occlusion but there is in addition to mucous plugging cicatrisation of small airways so that their occlusion is irreversible. The ratio between mucous plugging and cicatrisation is not currently known.

When the ventilatory efficiency of cigarette smokers who complain of no symptoms is measured a proportion have impaired efficiencies. Are we to conclude that there is already an attack upon their small airways, and have we identified a group which will eventually succumb to their habit? The therapeutic attack upon airways disease will become quite different if it is recognised that large airway disease and small airway disease usually co-exist and that the physiological manifestations of the two components are independant each of the other.

Whilst we all tackle large airway disease enthusiastically we give no attention at all to small airway disease, and the time has now come when this defect should be remedied.

THE PHYSIOLOGY OF CHEMORECEPTION IN THE CAROTID BODY

H. Acker and D.W. Lübbers,      Max-Planck-Institut für Systemphysiologie, Rheinlanddamm 201, 4600 Dortmund 1, FRG

Publisher Summary

This chapter focuses on the physiology of chemoreception in the carotid body. The carotid body is considered as being a peripheral chemoreceptor, which transduces changes in arterial pO2 and arterial pCO2 or pH into nervous signals. Regarding the different cell types of the carotid body tissue as a complex that collaborates in the chemoreceptive process, it can be assumed that a pO2- or pCO2-dependent transmitter release from the type-I cells excites nerve fibers, which are synaptically connected to these cells. O2 and CO2 are transported by the capillary blood flow to the chemoreceptor complex, producing specific strimuli. This chapter discusses the way by which pO2 and blood flow are involved in the chemoreceptive process or the carotid body. The chapter describes an experiment in which despite the well-known small AVpO2 difference, the mean tissue pO2 was 25 mmHg in the cat and 7 mmHg for the rabbit. These results demonstrate that the carotid body possesses mechanisms to produce low tissue pO2 values resembling a tissue hypoxia.

Since de Castro’s (7) and Heymans’ (13) fundamental findings the carotid body has been considered as being a peripheral chemoreceptor which transduces changes in arterial pO2 and arterial pCO2 or pH into nervous signals. Regarding the different cell types of the carotid body tissue as a complex which collaborates in the chemoreceptive process, _ it can be assumed that a pO2 or pCO2 dependent transmitter release from the type-I cells excites nerve fibres which are synaptically connected to these cells. O2 and CO2 are transported by the capillary blood flow to the chemoreceptor complex, producing specific stimuli. We would like to present some ideas about how pO2 and blood flow are involved in the chemoreceptive process of the carotid body. The pO2 is involved in three processes, i.e. 1) in the oxygen metabolism, and 2) in a pO2 dependent transmitter release and 3) in determining the number of active chemoreceptor complexes by producing a characteristic pO2 distribution in the tissue. The pO2 distribution determines the amount of chemoreceptor complexes which are situated in the pO2 threshold zone in which the nervous signal originates. The threshold zone varies with the degree of arterial hypoxia.

In the first figure it is shown that in spite of the well known small AVpO2 difference the mean tissue pO2 was 25 mm Hg in the cat (median 20 mm Hg) and 7 mm Hg for the rabbit (median 5 mm Hg) (23). These results demonstrate that the carotid body possesses mechanisms to produce low tissue pO2 values which resemble a tissue hypoxia. In our view this Hypoxia enables the carotid body to produce a nervous signal even if the arterial pO2 is normal and to react with an increase of the nervous activity also if the arterial pO2 decreases only by a small amount.

But before we discuss further consequences, we have to mention that Whalen and Nair (25) published tissue pO2 values of the cat carotid body with a mean value of about 60 Torr. We have no definite explanation for these different findings, but since the volume of the specific carotid body tissue is very small, we think that it is always necessary to show that the pO2 measurements are really done in the specific carotid body tissue. We tested the po2 needle electrode 1) by histological controls and 2) by a special in situ test, the O2 blowing test.

1) In the second figure an example is shown where a pO2 profile in the rabbit carotid body was correlated to the tissue compartments by histological reconstruction of the electrode tracks. It is to be seen that only a small part of the pO2 profile is measured in the zone with specific carotid body tissue, consequently, only those values can be used for a pO2 histogram of the carotid body. We could show together with Weigelt (24) that, if the pO2 values of the surrounding tissue are included, the pO2 histogram shifts to higher values.

2) The oxygen blowing test is a helpful method to decide during the experiment whether or not the pO2 measurements are done in the specific carotid body tissue. In adipose or connective tissue, the tissue pO2 increases when humidified and 37°C warm oxygen is blown over the surface of the prepared carotid body region. When the electrode reaches the specific tissue, this oxygen response cannot be seen. We have no ready explanation for this phenomenon. A similar phenomenon was described by Prühs (17) and Bingmann (5). They found that under normal conditions the nervous activity of the carotid body cannot be influenced by changing the O2 at the carotid body surface. It takes normally about 20 minutes of circulatory arrest (after the death of the animal) before this phenomenon disappears and the nervous activity can be influenced from the outside, as it is known from the carotid body in vitro described by Eyzaguirre.

Fig. 2 Correlation of tissue pO2 and depth of puncture in a carotid body of a rabbit (24)

Another reason for different results can be the difference in the vascular structure of carotid bodies as shown by E. Seidl (20). In the third figure it is shown that in the Cleveland carotid body several arterials ran along the border of the carotid bodies and side branches entered the tissue laterally. In the Dortmund carotid bodies the arteries mostly immediately enter the carotid body.

It is important in which direction the pO2 electrode enters the carotid body. We would like to stress that all measurements have to take into account the complicated anatomy of the carotid body. The pO2 electrodes which penetrate the tissue in different tracks, penetrate also different cellular elements of the tissue. These differences are responsible for the large variation in the pO2 profiles. It is not allowed to insert the electrode into the prepared carotid body and to claim without further test to measure the pO2 in the specific tissue. This should be taken into account when the contradictory results will be discussed. We think that the carotid body needs low tissue pO2 values because the pO2 threshold for the chemoreceptor complex is at low pO2 values.

Fig. 3 Location and arteries supplying carotid bodies of German and American cats (20)

This assumption is supported by experiments on fetal and newborn lambs, which we did together with Purves (19). Fig. 4 shows an example of the relationship between tissue pO2 in lamb carotid body and arterial blood in fetal and newborn state. In the fetal state the tissue pO2 and the arterial pO2 are in the same range, which means that the pO2 difference is very small. Remarkable is that at this state, the nervous activity of the carotid body is very diminished. After birth when the arterial pO2 increases to a higher level, the carotid body tissue po2 does not increase in the same way, but decreases continuously to low levels which means that now a large pO2 difference between arterial blood and tissue has developed. Concomitantly chemoreceptive nervous activity can now be registered from the sinus nerve when the arterial pO2 is decreased similarly as seen in the