Perspectives On Medical Research
Volume 4, 1993

Contents

The Process of Medical Discovery

Brandon P. Reines

Recently, certain philosophers of science have challenged the ways historians have traditionally described the process of medical discovery. Many historians have tended to provide descriptive accounts of medical discovery by focusing on particular successful experiments. Their focus, in general, has been priority disputes or biographical information about principal investigators. Consequently, most medical historians have failed to critically examine the underlying conceptual evolution of medical discoveries.(1) Philosopher Thomas Nickles' normative account of discovery in the physical sciences offers a framework for understanding discovery in the life sciences.(2,3) Nickles, seeking to describe the logic and origin of discovery, traces the evolution of physical discoveries from the initial anomalous observations to later confirmatory laboratory experiments. Nickles argues that the results of experimental manipulations (which can only approximate natural phenomena) can never be directly extrapolated from the lab to the real world. He concludes that discoveries arise from knowledge of existing theory and an ability to interpret, and perceive anomalies in, a broad range of data; significant discoveries rarely derive from an individual experimental effort.

Laboratory testing of theories in the life sciences tends to be even more problematic than in the physical sciences, because it is difficult to reproduce natural phenomena of complex, living organisms in the laboratory. Animal experiments, which involve relatively simple manipulations designed to produce a particular result, provide contrived evidence in support of a particular theory. Researchers have at their disposal a nearly infinite range of experimental manipulations that can be performed on animals of different species, and they can fashion the experimental protocol so that they obtain data that "confirm" a theory actually derived from prior clinical observation and investigation. In reality, laboratory data on other species can neither disprove a hypothesis about humans nor necessitate the formulation of a new hypothesis. Yet, historians of science routinely credit such experimentation, rather than the preceding clinical investigation and hypothesis, as the source of discovery. Consequently, medical historians generally present the animal experimenter as the discoverer and relegate the actual discoverer -- the clinician -- to anonymity or lesser importance.

Endocrinology

The Adrenal Glands

During the mid-l9th century, most physiologists contended that nervous reflexes mediate organic functions--a doctrine known as nervism. The anomalies that eventually discredited this doctrine were clinically observed.

In 1855, physician Thomas Addison described five human cases of tuberculous invasion of the adrenal glands and nearby nervous structures, noting symptoms of anemia, general debilitation, cardiac "feebleness," stomach "irritability," and skin discoloration. Addison hypothesized that the adrenal glands must secrete a substance vital to health.(4)

In an effort to test Addison's hypothesis, experimenters then removed animals' adrenal glands. No experiment reproduced the symptoms of Addison's disease,(5) but researchers were able to induce skin discoloration by experimental bilateral adrenalectomy.(6) Experimental adrenalectomy often caused rapid death of the animals used.(7) Although the researchers attributed the deaths of many of their rat and mice subjects to surgical trauma, many physiologists interpreted the deaths as evidence that the adrenal glands filter harmful toxins from the blood.(3)

In 1882, physician James F. Goodhart first firmly refuted nervism by reporting cases of Addison's disease caused by simple atrophy of the adrenal glands.(9) Addison's hypothesis then received wider acceptance.

In 1893, physician-physiologist George W. Oliver tested the hypothesis that adrenal secretions affect blood pressure, by feeding an adrenal extract to his son (who became somewhat ill).(10) Using a device of his own design, Oliver measured the diameter of the radial artery (in the forearm) and found it to be dramatically decreased. Oliver concluded that the adrenal glands must contain a substance that raises blood pressure by constricting blood vessels. Oliver never stated exactly how he derived the hypothesis, but he must have had a working hypothesis before he gave adrenal extract to his son and measured blood vessel diameter. Otherwise, his experiment on his son seems illogical. Most likely, he was inspired by the low blood pressure seen in Addison's disease.

According to the standard historical account, when Oliver arrived at experimental physiologist Edward A. Shafer's laboratory to test the adrenal extract on animals, a dog was fortuitously hooked to a mercury manometer at the time. The effects of the adrenal extract on blood pressure were noted immediately. The history of medicine literature commonly relates such dramatic, "happy accident" stories, which suggest that discoveries occur in the laboratory. Like many such stories, this one is questionable. Henry H. Dale, who related this account, heard it third-hand, and even he later wrote, "I can find no contemporary record even to suggest what the experiment might have been."(10) Oliver's writings also contradict this account.(11,12) Oliver and Shafer's experiments on animals were not lucky accidents, but well-planned manipulations that must have required forethought. For example, their experiments on dogs involved cutting the dogs' spinal cords and injecting very low doses of adrenal extract into their veins.(13) Through this highly invasive and contrived procedure, Oliver and Schafer raised the dogs' blood pressure. (Most likely, Oliver and Schafer cut the dogs' spinal cords in order to remove any inhibitory influence of the central nervous system that might have lessened the extract's effect on blood pressure.)

Although Oliver's hypothesis--derived from human investigations--provided the rationale for the experiments on dogs, medical historians have generally credited Schafer with the role of pioneer endocrinologist, highlighting the animal experiments and minimizing the contribution of the observations of humans that preceded them.

The Ovaries

In many ways, discovery of the ovaries' endocrine function paralleled discovery of the adrenal glands' function.

In the 19th century, the influential physiologist Eduard Pfluger asserted that menstruation results from the growing ovum impinging on the ovarian nerves and, by reflex, causing uterine changes.(14) However, in 1895, surgeon Robert T. Morris observed that some women menstruate and become pregnant after their ovaries have been surgically removed.(15) (Subsequent post-mortem studies of such women revealed that a small fragment of ovarian tissue had inadvertently been left behind.) In such cases the ovum cannot impinge on the ovarian nerves; therefore, Morris hypothesized that ovarian secretion, not reflex action, must be crucial in inducing menstruation. With this theory in mind, he transplanted ovarian tissue into the uteruses of women unable to menstruate or become pregnant due to prior ovariectomy or reproductive-tract pathology.(16)

According to Morris, researcher Emil Knauer quickly learned of Morris' findings by way of Morris' assistant, Morris Schlapp. Using animal experimentation, Knauer then attempted to mirror Morris' indings. In 1896, Knauer reported that rabbits whose ovaries had been excised and implanted elsewhere in their body did not experience uterine atrophy.(17) He, not Morris, became known as the discoverer of the ovaries' endocrine function.

Angiology and Reconstructive Vascular Surgery

Angiology (the science of blood vessels), which made modern reconstructive vascular surgery possible, also evolved from clinical observation.

The Vascular Endothelium

The healing capacity of the vascular endothelium (the cells comprising blood vessels' innermost layer) underlies surgical repair of blood vessels. In the process of healing, the vascular endothelium forms connective scar tissue.

In the 1880s, pathologist Ernst Ziegler observed that the endothelium of ligated arteries usually forms scar tissue and likened this capacity for repair to that of abdominal lining.(18) Surgeons began to translate their experience with abdominal surgery to vascular surgery.

In aneurysm surgery, Robert Abbe and others tested whether arteries have an inner lining capable of forming scar tissue.(19) These surgeons irritated arteries' inner lining (for example, with a surgical needle) in the hope of stimulating reparative processes, and this appeared to improve surgical results.(20)

Vascular Anastomosis

In the 1890s, as surgeons were discovering that arteries are lined by cells with a capacity for active repair, they hypothesized the possibility of performing vascular anastomosis (reuniting the two severed ends of a blood vessel).(21) Morris then attempted the procedure in rabbits. The effort failed, apparently because of resulting infections.(21) Nevertheless, Morris suggested the procedure to surgeon James B. Murphy.

In 1897, Murphy published the first report on vascular anastomosis in human patients.(22) The procedure was not entirely successful--partly because of post-surgical infection, partly because Murphy inserted one severed end of the blood vessel into the other rather than aligning them so that the two ends' inner linings could adhere to each other.(23) Nevertheless, Murphy's partially successful clinical cases influenced both clinical and laboratory investigators of vascular anastomosis, including Alexis Carrel.(24)

In 1912, Carrel won the Nobel Prize for his method of vascular anastomosis; the method, he stated, "came into existence nine years ago from the study of the technique of [physician] Payr and of Murphy."(25) Carral claimed to develop the principle using human cadavers and two living dogs.

Although Carrel and his animal research colleague Stephen Watts first demonstrated the advisability of suturing blood vessels inner lining to inner lining in nonhuman animals, their work's guiding hypothesis derived from the clinical research of surgeons, such as Abbe, who were repairing arteries with aneurysms.

Veins' Tolerance of Arterial Blood Pressure

In 1728, William Hunter first reported an arteriovenous aneurysm, an abnormal connection between an artery and a vein that introduces blood into venous circulation under unnaturally high pressure.(26)

Nearly two centuries later, surgeon Alejandro San Martin y Satrustigui analyzed cases of arteriovenous aneurysm and concluded that, because the vein abnormally connected to an artery pulsates throughout the patient's life, veins must be able to withstand the high blood pressure characteristic of arteries. San Martin hypothesized the possibility of using grafted veins as substitutes for portions of damaged arteries.(27,28)

San Martin suggested to his student José Goyanes that he attempt to connect two cut ends of an artery by means of a venous graft. In 1906, following San Martin's advice, Goyanes excised an aneurysm and replaced it with a segment of nearby vein.(27) The patient fared well. Goyanes' success encouraged other surgeons to use vein grafts to repair arterial segments.

The Bypass Principle

In 1948, French surgeon Jean Kunlin performed the first bypass operation.(29) Confronted with arterial blockage in a patient's leg, Kunlin reasoned that the chance of saving the leg might be improved if the section of blocked artery were left intact rather than excised, since excision would destroy smaller blood vessels that formed collaterals around the damaged section of artery and the collateral circulation such vessels provide. And so, Kunlin grafted one of the patient's own saphenous veins around the blockage. American surgeons quickly appreciated the principle of preserving collateral circulation,

However, the principle of grafting a patient's own vein remained largely ignored well into the 1960s. Although surgeons at Walter Reed Army Hospital successfully used autovein grafting in the early 1950s,(30) most surgeons opted instead for arterial grafts. According to vascular surgeons Rodolfo Domingo, Charles C. Fries, Philip N. Sawyer, and Sigmund A. Wesolowski, the general bias against using a patient's own vein for bypass surgery resulted from experimental findings in animals that veins engrafted in the arterial system often underwent aneurysmal dilatation (ballooning of a blood vessel due to increased pressure).(31) Nevertheless, by repeatedly performing clinical bypass surgery of leg arteries, W. Andrew Dale, Robert Linton, Gregory Mayor, and others demonstrated the success of bypass surgery using a patient's saphenous vein.(32,33)

Hematology

Red-Blood-Cell Groups

In 1900, hematologist Karl Landsteiner proposed the existence of A, B, and O human red-blood-cell groups as the explanation for transfusion reactions between human beings.(34) The validity of grouping blood types into A, B, and O was supported by in vitro assays of human sera and red blood cells and confirmed by the effectiveness of using blood typing to prevent transfusion reactions.

Rh Factor

In 1939, hematologist Philip Levine first hypothesized the existence of a third antigen (a substance that induces antibody production), in addition to A and B antigens, on the surface of human red blood cells.(35) Levine's hypothesis was prompted by the case of a woman who gave birth to a severely decomposed fetus that had been dead for several months: when the woman was subsequently transfused with her husband's ABO-matched blood, she suffered a severe reaction. Levine reasoned that the mother, due to prolonged exposure to the fetus, must have already produced antibodies to some antigen present in both the fetus's and the father's blood. That is, some previously unknown red-blood-cell antigen must have been responsible for the mother's immune response. Levine did not, however, name this newly hypothesized antigen.(36)

Landsteiner and hematologist Alexander Wiener soon dubbed the antigen "Rh (rhesus) factor."(37) When, in in vitro tests, the blood serum of humans who had shown an adverse reaction to ABO-matched blood was mixed with human red blood cells of compatible type, about 80% of samples showed clumping of red blood cells due to antibody attachment in response to an antigen. In similar tests the blood serum of rabbits who had been injected (in Landsteiner and Wiener's laboratory) with red blood cells from rhesus monkeys and had produced antibodies in reaction to rhesus red-blood-cell antigens also showed about 80% clumping when mixed with human red blood cells.(38) The misnomer "Rh factor" was based on this finding that rabbits produce an antibody to some rhesus red-blood-cell antigen that also reacts to a human red-blood-cell antigen.

Levine then attempted to confirm his own theory by replicating Landsteiner and Wiener's experimental results; having failed to do so, in 1961, Levine asserted that the rhesus-monkey antigen identified by Landsteiner and Wiener must differ from the human one.(39)

Levine's theory explained the pathogenesis of erythroblastosis fetalis ("rhesus disease of the newborn"), in which an Rh-negative mother produces antibodies against the Rh factor in the blood of an Rh-positive fetus, thereby destroying the fetus' red blood cells.(40) Nevertheless, Wiener is generally given more credit for the discovery of Rh factor and its relationship to disease than Levine, even though Landsteiner and Wiener's Rh experiments never elucidated any human disease.

White-Blood-Cell Groups

Initially white-blood-cell groups were investigated with the hope of improving the success of tissue and organ transplants. In the 1950s clinical investigators such as Jean Dausett, Jan J. van Rood, and Rose Payne began exploring human white-cell groups.(41-43) Because transfusion reactions occurred despite careful typing by all known human red-cell antigens, these investigators hypothesized the existence of previously unidentified white-cell antigens. They then used in vitro immunological tests to further characterize those antigens.

By 1948, animal experimenters such as Peter Gorer and George Snell had concluded that a specific group of genes determines which antigens are present on a mouse's cells and, therefore, which cells transplanted from another mouse will provoke an immune response to foreign antigens. They called the group of genes the "major histocompatibility complex" (MHC).(44,45) In the mouse, the MHC became known as the histocompatibility-2 (H-2) system.

While determining the human white-cell groups through studies of human patients, clinical investigators attempted to relate their findings to MHC theory--even contending, as of 1965, that they were uncovering the "human MHC"-thereby obscuring their discoveries' clinical origins.(46)

By the late 1960s, however, H-2 researchers had developed so elaborate an explanation of their laboratory findings that one of these researchers, Jan Klein, recalls: "The H-2 system was often considered an exotic curiosity completely detached from reality."(47) In contrast, clinical investigators developed a simple two-locus framework by which to interpret the white-cell data.

In 1966, Walter Bodmer, Payne, and colleagues argued that the HL-A histocompatibility complex of humans is composed of two distinct series, which they labeled A and B.(48) Snell and Erik Thorsby then maintained that the mouse H-2 framework should be simplified to reflect two main loci (K and D) analogous to the human A and B loci.(49) Thorsby wrote, "Based on serological and genetic evidence from the HL-A system, a tentative new H-2 model of two segregant series is suggested."(49) Yet, knowledge of human histocompatibility is generally attributed to the mouse research of Snell, Gorer, and their co-workers.

Conclusion

Historians have generally failed to recognize the clinical origins of medical discovery. Routinely, medical discoveries derived from observation of clinicopathologic anomalies are attributed to laboratory experimentation. Because of species differences in anatomy and physiology, extrapolation (by analogy) from laboratory to clinical phenomena is, at best, unreliable.

References

1. Reines BP. On the locus of medical discovery. The Journal of Medicine and Philosophy 1991 ;16: 183-209.

2. Nickles T. Justification and experiment, in Gooding D, Pinch T (eds). The Uses of Experiment. Cambridge, University Press, 1989, pp 299-333.

3. Nickies T. Beyond divorce: Current status of the discovery debate. Philosophy of Science 1985 ;552: 177-206.

4. Addison T. On the Constitutional and Local Effects of Disease of the Suprarenal Capsules. London, Samuel Highley, 1855.

5. Brown-Sequard CE. Recherches experimentales sur la physiologic et a! pathologie des capsules surrenales. Arch Gen Med 1856;8:385-401, 572-598.

6. Tizzoni G. Sur la physio-pathologie des capsules surrenales. Arch Ital Biol 1884;5:333-340.

7. Harley G. An experimental inquiry into the function of the supra-renal capsules and their supposed connection with bronzed skin. British and Foreign Medical-Chirurgical Review 1858;21:204-221, 498-510.

8. de Domemcis N. Pourquoi l'extirpation des capsules surrenales amene la mort chez les animaux. Arch Physiol 1894;6:810-815.

9. Goodhart J. Simple atrophy of the supra-renal capsules, accompanied by melasmasuprarenale and other symptoms of Addison's disease. Trans Path Soc London 1882;33:340-345.

10. Dale HH. Accident and opportunism in medical research. Br Med J 1948;2:451-455.

11. Oliver G. Pulse Gauging: A Clinical Study of Radial Measurement and Pulse-Pressure. London, HK Lewis, 1895.

12. Oliver G. A contribution to the study of the blood and circulation. Br Med J 1896;1:1374- 1377.

13. Oliver G, Shafer E. On the physiological action of extract of the suprarenal capsules. J Physiol 1894;17:i-v.

14. Simmer H. Pfluger's nerve reflex theory of menstruation: The product of analogy, teleology and neurophysiology. Clio Medica 1977;12:57-90.

15. Morris RT. The ovarian graft. New York Med J 1895;62:436-437.

16. Morris RT. Notes on ovarian grafting. Medical Record 1901;59:83-87.

17. Knauer E. Einige versuche uber ovanentransplantation bei kaninchen. Centrablat: fur Gynakologie 1896;20:524-528.

18. Ziegler E. A Textbook of General Pathological Anatomy and Pathogenesis. New York, Wood Pub, 1883-1887.

19. Abbe R. Aneurysmorrhaphy: Personal experience with the modern method of treating aneurysm. Ann Surg 1908;48:10-15.

20. Morris RT. The serous coat of blood vessels compared with the peritoneum. Ann Surg 1908;48: 18-21.

21. Morris RT. Fifty Years A Surgeon. New York, EP Dutton, 1935.

22. Murphy JB. Resection of arteries and veins injured in continuity. Med Record 1897:51:73-88.

23. Stewart FT. End-to-end anastomosis of the brachial artery [discussion]. Ann Surg 1908;48: 152-155.

24. Carrel A. Technique and remote results of vascular anastomoses. Surg Gynecol Obstet 1912; 14:246-250.

25. Watts S. The suture of blood vessels. Implantation and transplantation of vessels and organs. An historical and experimental study. Bull Johns Hopkins Hospital 1907;18:154-192.

26. Hunter W. XXXVI Further observations on a particular species of aneurysm. Medical Observation Sociely of the Physicians of London 1761;2:390-414.

27. Goyanes J. Substitution plastica de las artenas por las venas, o arterioplastica venosa, aplicada, como nuevo metodo, al tratamiento de los aneurismas. El Siglio Medico 1906;346.

28. Goyanes J. San Martin and his work [lecture at Athenaem of Madrid, December 1920]. Academy of the R.A.N. of Medicine, 1920.

29. Kunlin J. The treatment of arterial obstruction by vein grafting. Arch Mal Coeur 1949;42:371-375.

30. Cooke FN, Hughes CW, Jahnke EJ, Seeley SF. Homologous arterial grafts and autogenous vein grafts used to bridge large arterial defects in man: A report on fourteen cases. Surgery 1953;33:183-189.

31. Domingo RT, Fries CC, Sawyer PN, Wesolowski SA. Peripheral arterial reconstruction: Transplantation of autologous veins. Trans Am Soc Artificial Internal Organs 1963;IX:305-308.

32. Mayor G. Autogenous vein grafts, in Mackey WA (ed). Arterial Surgery. New York, MacMillan, 1964.

33. Dale WA, DeWeese JA, Merle Scott WJ. Autogenous venous shunt grafts: Rational and report of 31 for atherosclerosis. Surgery 1959;46:145-161.

34. Landsteiner K. Uber agglutinationserscheinungen normalen menschlichen blutes. Wien Klin Wschr 1901;14:1132-1135.

35. Levine P, Stetson RE. An unusual case of intra-group agglutination. JAMA 1939;113:126-127.

36. Zimmerman D. Rh: The Intimate History of a Disease and its Conquest. New York, MacMillan, 1973.

37. Landsteiner K, Wiener AS. An agglutinable factor in human blood recognized by immune sera for rhesus blood. Proc Soc Exp Biol Med 1940;43:223-224.

38. Wiener AS. Karl Landsteiner, MD: History of the Rh-Hr blood group system. NY State J Med 1969;69:2915-2935.

39. Levine P. The discovery of Rh hemolytic disease. Vox Sanguis 1984;47:187-190.

40. Levine P. The influence of the ABO system on Rh hemolytic disease. Human Biology 1958;30: 14-20.

41. Dausset J, Nenna A. Presence d'une leuco-agglutinine dans le serum d'un cas d'agranulocytose chronique. CR Society for Biology 1952:140:1534-1536.

42. Payne R. Leukocyte agglutinins in human sera. Arch Int Med 1957;99:587-591.

43. van Rood JJ, Eernisse JO, van Leeuwen A. Leucocyte antibodies in sera from pregnant women. Nature 1958;181:1735-1736.

44. Snell GD. Histocompatibility genes of the mouse IL Production and analysis of isogenic resistant lines. J Nat Canc Inst 1958;21:843-848.

45. Gorer P. The detection of antigenic differences in mouse erythocytes by the employment of immune sera. Br J Exp Biol 1936;17:42-46.

46. Dausett J, Ivanyi P, Ivanyi D. Tissue antigens in humans. Identification of a complex system (Hu-1), in Dausett J (ed). Histocompatabilily testing 1965. Copenhagen, Munksgaard, 1965.

47. Klein J. Biology of the Mouse Histocompatability-2 Complex [Chapter 1, History]. New York, Springer-Verlag, 1975.

48. Bodmer W, Bodmer J, Adler S, Payne S, Bialek J. Genetics of 4 and LA human leukocyte groups. Ann NY Acad Sci 1966;129:473-489.

49. Thorsby E. A tentative new model for the organization of the mouse H-2 histocompatability system: Two segregant series of antigens. Eur J Immunol 197 1;I:57-59.