How Science Happened to Me

ROBERT E. FORSTER II

	ARTICLE

TOP ARTICLE

The paper I choose to discuss is "Relative importance of diffusion and chemical reaction rates in determining rate of exchange of gases in the human lung, with special reference to true diffusing capacity of pulmonary membrane and volume of blood in the lung capillaries" (J. Appl. Physiol. 1957;11:290- 302), by F. J. W. Roughton and me, the culminating article in a group of four, which described the separation of the resistance of carbon monoxide (CO) or oxygen (O₂) to diffusion in the pulmonary capillary blood from the resistance through the pulmonary alveolar membrane, as well as the measurement of pulmonary capillary blood volume, all in vivo.

I was a medical house officer at the Peter Brent Brigham Hospital in Boston under George Thorn in 1944 when Dr. (Col.) John H. Talbott, commanding officer of the Quartermaster Corps Climatic Research Laboratory, located in the environmental chambers of the Pacific Mills in Lawrence, Mass., called to tell George that Clifford Barger, then at the Climatic Research Laboratory, had contracted tuberculosis so he needed a replacement, and did George know of anyone. Thus, in December of 1944 I found myself in an old New England mill town investigating heat loss from the extremities, temperature regulation, and the effectiveness of footwear and handwear in protecting soldiers under cold conditions. We had published some papers on peripheral blood flow and heat exchange by the end of World War II when I was discharged. I then used the GI Bill to study mathematics (of the baby type), physical chemistry, and physics at Harvard for a year before returning to take up a residency at the Peter Bent Brigham Hospital. After this I spent two years supported by a Life Insurance Fund, now defunct, as a post-doctoral fellow with Eugene Landis at the Harvard Department of Physiology, still working on temperature regulation, but by the hypothalamus in cats. I looked around for a faculty position in internal medicine, interviewed at Hopkins, where they thought I was applying for admission to medical school, when surprisingly I was offered a faculty position in the Graduate School of Medicine at Penn by Robert Dripps, director of anesthesiology, then in the Department of Surgery of the School of Medicine, and Julius Comroe, in the Department of Physiology and Pharmacology. The salary was handsome for the time, $6000, and Landis thought it indecently large, but I accepted. This changed my area of research from temperature regulation to respiration, through no intellectual initiative of my own.

When I moved to Philadelphia, the Department of Physiology and Pharmacology of the Graduate School of Medicine was housed in a basement that flooded when it rained, but it had, or shortly gained, a remarkable faculty. Julius Comroe, Seymour Kety, Ward Fowler, Arthur DuBois and George Koelle were a few of the names, and there was a great excitement about the place. The philosophy was to apply the respiratory physiology that had been learned in the air and sea during World War II to the diagnosis and treatment of pulmonary diseases. Wallace Fenn had a similar philosophy for the founding of the Journal of Applied Physiology. I was assigned a small basement lab in the original School of Medicine building, with a new mass spectrometer and carbon monoxide (CO) analyzer, both of which had been ordered by Comroe with advice from Seymour Kety and John Lilly. The mass spectrometer was built to order from a low mass resolution (1/80) Instrument made by Consolidated Engineering Corporation to detect leaks in large capacity vacuum equipment. The CO meter was an infrared instrument, built to order by Max Liston, on a principle used by the Germans during World War II. Julius suggested that I repeat an experiment of Marie Krogh in 1914 that measured the diffusion capacity of the lung: the ability of the lung to exchange gases with capillary blood. In order to investigate the possibility of O₂ secretion in the lungs, she had measured the rate of disappearance of CO from the lungs during breath holding by inspiring a gas mixture containing a small amount of CO, immediately and rapidly breathing out an alveolar sample, holding the remaining lung gas for an additional 10 s, and then delivering a second alveolar sample into a rubber bag. The proportional drop in alveolar CO in 10 s plus the alveolar volume permitted her to calculate the pulmonary diffusing capacity for CO (DL_CO), a measure of the efficiency of the lungs in exchanging gas, defined as

(1)

Because of its tremendous affinity for hemoglobin, average red blood cell PCO equals zero.

This value for DL_CO could be converted to a value for DL_O2 by multiplying by a factor of 1.21, derived from the relative molecular weights of the gases and their solubilities in water. She then computed the average PO₂ difference between alveolar gas and capillary blood by dividing the oxygen uptake by DL_O2. She showed that the PO₂ difference needed to transport oxygen into the blood was reasonably small and that diffusion alone could explain O₂ uptake in exercise as well as at rest, and therefore that oxygen secretion, as proposed by her professor, Christian Bohr, was not necessary.

The major technical problem in Marie Krogh's experiment was the analysis of low concentrations of CO, which she did using a tedious combustion method. The infrared CO analyzer I had made the experiment simple enough to do, but I must have spent six months trying to improve the sensitivity and signal/noise ratio of the two new instruments before Ward Fowler came around, obviously sent by Julius Comroe, and suggested we make a measurement with the instruments as they were. If Ward had not helped me it is doubtful that I would have stopped tinkering with the instruments for many months. He also suggested including helium as a chemically inert tracer in the inspired gas mixture, as well as CO, in order to compute the CO in the first alveolar sample, rather than measuring it, thereby avoiding errors produced by variations in the distribution of inspired gas to different alveoli. From the expired alveolar and inspired helium concentrations measured by the mass spectrometer, we could calculate the dilution of inspired CO that had been in the expired gas sample at time zero, eliminating Krogh's first alveolar sample. The lesson is that one must have a clear goal, which in this case was to measure DL_CO, not to refine instrumentation. Although the outputs of the instruments were noisy, the noise was such a small proportion of the changes in CO and He concentrations during breath holding that excellent disappearance curves were obtained immediately. If physiological phenomena are of primary importance, the necessary instrumental accuracy is defined purely by the variation in the physiological variables.

We soon confirmed Marie Krogh's estimate of DL_CO and, in addition, made a new finding: that increasing alveolar PO₂ decreased DL_CO. The most reasonable explanation for this was that raising PO₂ reduced the small concentration of unliganded hemoglobin (Hb) that reacted with CO, reducing the rate of CO uptake by red cells. This meant that the rate of CO uptake by red cells in the pulmonary capillary partly limited the rate of uptake of CO in the lungs, at the time a very new idea.

I went on to derive an equation describing the rate of CO uptake in the lungs as a function of the volume of blood in the capillary bed at any instant (Vc), the rate at which a milliliter of whole blood reacted with CO at a given PCO () and the diffusing capacity of the pulmonary membrane, Dm_CO, from alveolar gas to capillary plasma outside the red cells.

However I could find no published values of these rates except for some fragmentary measurements Roughton had made on a crude instrument at the Harvard Fatigue Laboratory in 1945. In this same paper he estimated the size of the pulmonary capillary blood volume. Therefore, I wrote Professor Roughton, the world expert on the subject in Cambridge, UK, to see if he had measurements of  in human red cells at 37° C and at different PO₂ and if he would let me use them. He replied that he did not already have the necessary data but that he would be interested in coming to Philadelphia and making the measurements with us. I was delighted because not only would I obtain the rate measurements I wanted but would also learn the rapid-mixing techniques for the measurement of the velocities of the reactions of O₂ and CO with hemoglobin that had made Roughton famous. He came, of course with the encouragement of Julius Comroe, bringing a newly constructed continuous-flow rapid-mixing apparatus. In this type of instrument a suspension of red cells at a preset PO₂ is mixed in 1 ms with a solution of CO at the same PO₂, which continues to react as it flows down a glass observation tube. The HbCO concentration is analyzed spectrophotometrically at different distances along this tube, which correspond to different durations of the reaction; typically 1 cm distance equals 0.005 s. The apparatus had performed well in the UK using hemoglobin solutions and sheep red blood cells. Unfortunately in Philadelphia, using human red cells, there was so much light scattering from the flat sides of the cells as they oriented in swirls in the reaction mixture, that we could not measure HbCO accurately enough. Sheep cells had much less asymmetry. On the basis of a principle suggested by Britton Chance, Bill Briscoe and I designed an analyzer which split a white light beam into two, each passing through an interference filter at a different wavelength in the hemoglobin spectrum and thence onto a photomultiplier, the difference of whose output was recorded. The two wavelengths were chosen so that the change in absorption on forming HbCO was a maximum. The advantage of measuring the difference in two wavelengths was that the light scattering was similar so it subtracted out, providing a usable record of the relative changes in HbO₂ and HbCO in whole cell suspensions.

We obtained a large series of measurements of  on blood of normal subjects, using the rapid mixing apparatus, and of DL_CO on the same normal subjects, both at PO₂ values ranging from 100 torr to over 600 torr. Our equations permitted us to treat the diffusion path in the alveolus as two resistances in series; the reciprocal of the conductances, that is the diffusing capacities, are resistances. Thus the resistance of the pulmonary membrane, 1/DM, plus that of the pulmonary capillary blood, 1/ VC, equals the resistance of the whole path 1/DL_CO, giving us the simple equation.

(2)

By inserting values of  and DL_CO at different PO₂ values, one can easily solve for the two constant unknowns, DM and VC (which I believe is still the only method for measuring them in vivo), which was very gratifying.

As we were preparing the papers, and there were four, Poul Kruhøffer in Copenhagen published a lovely paper in 1955 in which he measured DL_CO with ¹⁴CO, a technique for the analysis of CO, and divided DL_CO into a membrane and a capillary red cell component as in Equation 2. Kruhøffer did not have any values for  at different PO₂ values, the key data, so he made the reasonable assumption, from Roughton's publications, that 1/ was proportional to PO₂. This made the capillary blood resistance zero at PO₂ equal to zero, which unfortunately is not true. However in order to defend his priority of publication, Roughton added the essence of Equation 2 on to a paper from the Department of Colloid Science at Cambridge by Gibson and colleagues, which was further along and appeared in 1955, two years before our group of papers.

At some point in writing these four papers, I realized that Roughton had applied a similar approach but with much cruder data in his 1945 paper on the transit time of red cells in pulmonary capillaries. He did not have an analytic solution for the reactions in the blood, but instead used semiquantitative arguments which few could follow. So our more recent finding that the reaction rates in blood are important in CO uptake was not so new after all.

I chose this paper because it is important, I think, in providing an experimental basis for a pulmonary function test. It emphasizes the element of chance and the influence of colleagues as contrasted with prescient planning. The diffusing capacity of the lung may not be thought as important today because many medical students and younger physicians are not taught about it.

	Footnotes

Correspondence and requests for reprints should be addressed to Robert E. Forster II, M.D., Department of Physiology, University of Pennsylvania School of Medicine, A201 Richard Building, Philadelphia, PA 19104-6085. Email: forster{at}mail.med.upenn.edu

Acknowledgments: While this story is written from my point of view as an anecdote without references, other colleagues contributed as much or more than I to this article, although they were not authors: David V. Bates, Leon Cander, Ferdinand Kreuzer, and Bernard van Lingen.