Severinghaus, John W. MD
Karl von Vierordt (Tübingen, 1876) measured the rate of spectral changes of light penetrating tissue when circulation was interrupted. His work was ignored until Ludwig Nicolai (Göttingen, 1931) repeated that study. Nicolai’s device measured red light transmission through a hand. In 1939 Karl Matthes in Leipzig introduced ear oximetry, counterbalancing red and infrared light. J. R. Squire (London, 1940) was the first to realize that the differences of transmission of red and infrared light before and after expelling the blood from the web of the hand with a pressure cuff was a function of saturation (1). Pulse oximetry may be regarded as a sequel to Squire’s device and idea, using pulsatile changes in tissue blood volume instead of compression vascular collapse.
Oximetry development was stimulated during WW II in an effort to warn military pilots of dangerous hypoxia. Glen Millikan (1906–1947) developed a light-weight red and infrared ear oximeter in 1942 for which he coined the word “oximeter” (2).
Earl Wood (Mayo Clinic, 1949) and his PhD student, J. E. Geraci, modified the Millikan ear piece by incorporating Squire’s pneumatic cuff. Wood extended and mathematically developed the ideas of Squire, plotting the ratio of the ratio of red to infrared light optical ear density produced by compression and reperfusion as a unique function of saturation (3). After setting gains with the bloodless ear, Wood divided the decreased red signal by the decreased infrared signal to obtain saturation without calibration. However, in practice, users often set the signal to 100% while the subject breathed oxygen.
AOYAGI AND HIS PULSE OXIMETER
Takuo Aoyagi was born on February 14, 1936, in Niigata Prefecture, Japan, and graduated in 1958 from the Faculty of Engineering at Niigata University with a degree in electrical engineering. Initially he worked at Shimadzu, a scientific instrumentation company.
In February 1971 he joined the Research Division of Nihon Kohden Corporation. Initially his dream was to make a sensor of blood oxygen saturation to signal the need for artificial ventilation, and to accomplish this he studied the oximetry literature. Impressed with Wood’s plot of red and infrared light hemoglobin density (3), he obtained and studied a Japanese version (Erma) of Wood’s oximeter earpiece. He concluded that an ear oximeter could be used to record a dye dilution curve, but would require calibration with a blood sample. Because arterial pulsatile “noise” prevented accurate recording of the dye clearance, he invented a method to eliminate this noise, which led to his discovery. He wrote of this work in 2003 (4):
“These [variations due to the pulse] prevented accurate extrapolation of the down-slope of the dye curve after recirculation begins. I investigated this problem mathematically using the Lambert-Beer law. Then I conceived the idea of eliminating the pulsation by computing the ratio of optical densities of the two wavelengths. This supposition was proved workable by experiments.”
The Key Idea: The Ratio of Ratios
While testing this way of canceling out the pulsations in the ear dye densitometer, Aoyagi observed that, by holding his breath, a decrease of oxygen saturation reintroduced pulsatile waves by changing the ratio of densities at the two wavelengths. This led him to predict that this artifact might be used to measure arterial oxygen saturation. He continues:
“At this point I realized that both the pulsating portion and non-pulsating portion of optical densities of the blood in tissue must have the same information of blood color. And I imagined as follows:
(1) If the optical density of the pulsating portion was measured with two appropriate wavelengths and their ratio was obtained, the result must be same as Wood’s ratio.
(2) In this method the arterial blood is selectively measured and the venous blood does not affect the measurement. Therefore the probe site is not restricted to the ear.
(3) In this method the reference (comparable to the blanched ear reference) is set with each pulse.
Therefore a probe shift of location, or motion introduces only a brief artifact, before quickly returning to normal measurement. This was my conception of the pulse oximeter principle in December, 1972.”
He confirmed both theoretically and experimentally the validity of Wood’s plot of the density ratios. Aoyagi called the ratio of ratios φ:
Equation (Uncited)Image Tools
where AC and DC symbolize the pulsatile and nonpulsatile components of the transmitted light.
Equation (Uncited)Image Tools
Greatness in science often, as here, comes from the well-prepared mind turning a chance observation into a major discovery. “One man’s noise is another man’s signal” commented the respiratory physiologist Jere Mead half a century ago.
Development of the Oximeter
Aoyagi tested various wavelengths and methods of implementing the pulse oximetry idea. He selected the 630-nm wavelength, at which red light absorption was most sensitive to oxygen saturation, and he balanced this against a 900-nm infrared wavelength, which is not absorbed by dye. Indo-cyanine green dye was selected for cardiac output measurement because its absorption peaks at 805 nm, the isobestic point where hemoglobin and oxyhemoglobin have equal absorption, making dye dilution curves independent of saturation. Aoyagi noted that blood optical density at 900-nm decreased with desaturation, resulting in a larger signal than provided by ratios at 630 and 805 nm.
In early 1973, Aoyagi’s supervisor, Y. Sugiyama, told Dr. Susumu Nakajima, a surgeon then working at the Sapporo Minami National Sanatorium, of Aoyagi’s pulse oximetry invention. Nakajima, understanding that the method was a secret to be kept, placed an order with Nihon Kohden for the as-yet-undeveloped apparatus.
Aoyagi wrote (4):
“I assigned Mr. Michio Kishi chief of this project. For this pilot model, components of the dye densitometer were used. The light source was a small tungsten lamp. The transmitted light was divided into two and each [beam] was received with combination of an interference filter and a phototransistor. I used wavelengths of 630nm and 900nm. I selected 900 nm to avoid interference by [the dye] ICG. From the transmitted light, pulsation amplitude ‘AC’ and the total ‘DC’ were obtained and the ratio AC/DC was calculated. This AC/DC ratio was obtained at the two wavelengths and their ratio, phi (φ) was calculated. This φ was expected to correspond to Sao2.”
“For both dye densitometry and pulse oximetry, it was necessary to have a theoretical base of scattering optics.1 Dr. Kazuo Shibata of Tokyo Institute of Technology had been studying for many years methods of measurement of pigments in plants in vivo. I read his papers and consulted with him regarding the state-of-the-art. The only way to decrease the effect of error sources was to use one or two scattering plates. This method was called the “opal-glass method”. We adopted this method. By late 1973 the oximeter was ready and clinical evaluation was conducted in Sapporo.”
Disclosure of the Invention
Aoyagi reported his discovery of pulse oximetry to the Japanese Society of Medical Electronics and Biologic Engineering (MEBE) on April 26, 1974 (5) and published with his many collaborators (6). On March 29, 1974, a patent application titled “Apparatus for Photometric Blood Analysis” was submitted to the Japanese Patent Office by the Nihon Kohden Corporation, naming Aoyagi and Kishi as inventors. This patent was publicly disclosed on October 9, 1975, and published on August 2, 1978 (No. 53-26437); Patent 947714 was granted on April 20, 1979.
On April 24, 1974, two days before the MEBE meeting, a patent application also describing the use of the arterial pulse for oximetry was submitted by Masaichiro Konishi and Akio Yamanishi, named as inventors working at the Minolta Camera Company. This remarkable simultaneity of discovery led me, in reviewing this history, to inquire about the possibility that the secret was discovered in Aoyagi’s submitted abstract in advance of the meeting. Aoyagi wrote (personal communication, September, 2006):
“For presentation at MEBE, a preliminary abstract had to be submitted in the fall. In October, 1973, I submitted an application for a presentation, with a short explanation of the pulse oximeter principle. Many referees checked them and perhaps almost all of them were allowed to submit abstracts. My preliminary abstract was very short, but the [pulse oximeter] idea was written in it.2 The application was accepted.” In January 1974, Aoyagi submitted to MEBE the complete abstract describing the invention.
“Yamanishi is familiar to me because [several years later, and while employed by their different companies] we two worked together to try to make a pulse oximeter calibrator using real blood, at the request of Dr. K. Miyasaka (then head of anesthesia at National Children’s Hospital, Tokyo). Recently Yamanishi wrote a historical story of the pulse oximeter for the Japanese Society of Medical Instrumentation and it was published (7).”
In November 2006, at my request, Yamanishi shared his records and memories of these events with me [personal communication]. In the Fall of 1973, he was given a copy of Wood’s oximeter chapter (3) by his supervisor Konishi. Yamanishi had been interested in the operation of blanching and refilling the blood in the earlobe. He had been studying photo plethysmography for a year, particularly the work of Takeda et al. of Nippon Medical School (Tokyo) (8). He was aware of the effect on optical signal ratios of change of thickness of arterial blood in the tissue. In his 2005 review of events of 30 years past (7) he wrote (translation by Aoyagi): “In January, 1974, Yamanishi made up an idea of pulse oximeter and handed to a person in charge of patent saying ‘This is big invention.’”3
He thought that combining the idea of varying tissue blood volume with the larger signal of the finger would permit measurement of oxygen saturation. In April 1974 shortly before the 13th Conference of MEBE, Yamanishi noticed Aoyagi’s abstract on the pulse oximeter. He was surprised that his own group [Minolta] had only developed the theory, whereas Aoyagi et al. had already constructed an experimental model. Urged by Konishi and Yamanishi, Minolta’s patent section submitted the document to the Japanese patent office on 24th April, hoping to establish their claim before more information was disclosed in the conference. The Japanese Patent Office rejected Konishi’s patent application in 1982. In the United States, Minolta applied for and obtained patent protection with limited effect, being based on a subsequent Minolta patent in Japan.
Aoyagi’s prototype pulse oximeter was tested in conjunction with Dr. Nakajima in Sapporo on September 6 and 7, 1973 and by Nakajima and his associates on February 5 to 7, 1974 (9). It used an earpiece with incandescent light, filters, and photo transistors. However, Nihon Kohden did not continue to develop or market this instrument and made no effort to patent it abroad. Aoyagi was transferred to a post as assistant manager in the patient monitoring division of Nihon Kohden in September of 1975, and the research and development of the pulse oximeter was assigned to another worker. The pulse oximeter subsequently was marketed, but its performance was not satisfactory.
The Minolta Company developed Yamanishi’s oximeter concept using a fingertip probe to take advantage of the greater pulse amplitude. Light sources and signals were conducted through fiberoptic cables to and from the instrument, as was done in the Shaw-Hewlett Packard multiwavelength ear oximeter. Minolta’s device was marketed in 1977 as the Oximet MET-1471. Its response to hypoxia was reported to be linear and accurate to within 5% by Suzukawa et al. in 1978 (10) and Yoshiya et al. in 1980 (11). Nakajima and associates used it clinically in 1979 (12). However, when studied at Stanford in 1980 by Sarnquist et al. (13) a Minolta model 101 [identical to Oximet MET-1471] seriously underestimated the severity of hypoxia: At 50% actual Sao2 it read about 70%. Yamanishi wrote me (personal communication, 2006): “The Sarnquist data was the very important trigger for us to improve the accuracy of our pulse oximeter.” In 1984, Y. Shimada (then anesthetist of Osaka University, now professor of Nagoya University) with Minolta’s K. Hamaguri, I. Yoshiya, and N. Oka (14) published data using a Minolta Oximet MET-1471 that agreed with Sarnquist’s evidence of under-reporting the degree of desaturation. In this paper, this group developed perhaps the first theory of pulse oximetry that included scattering effects by blood cells.
Aoyagi wrote (personal communication):
“Although it was a rather too simple a theoretical formula, it encouraged me to build up my theoretical formula. There was a big difference between Minolta and Nihon Kohden. In Nihon Kohden the idea of pulse oximeter was denied by the person in charge of optical plethysmography. After my shift to another position, Nihon Kohden made no improvement in pulse oximeter technology until world-wide spread of Nellcor’s oximeters. On the contrary, Minolta made up a highly accurate instrument using their excellent optical technology, and later even made a model change before Nellcor. I appreciate Minolta. Without their recognition of idea of pulse oximeter, the idea might be buried.”
Finally, in September 1985, Aoyagi was permitted to resume research and development of his pulse oximeter. His subsequent work has focused on the mechanism causing the nonlinearity of the Beer’s law relationship of light transmission to saturation. He developed the theoretic background for both pulse oximetry and later for use of multiple wavelengths for other clinically useful purposes. His pulse spectrophotometer permits determination of plasma volume, hepatic blood flow and cardiac output after dye injection.
Aoyagi was granted the degree of Doctor of Philosophy in Engineering at Tokyo University on December 1993. He was awarded two prizes in 2002. The “Social” award was given for research other than on anesthesiology by the Japanese Society of Anesthesiologists. The Purple Ribbon Medal was given him by the Emperor of Japan for contributions to sciences and arts. Nihon Kohden was then persuaded to allow him to continue his research after retirement.
Introduction of pulse oximetry coincided with a 90% reduction in anesthesia-related fatalities. Takuo Aoyagi’s invention was serendipitous. Although he could use the infrared signal to cancel pulsatile “noise” in the dye decay optical signal, hypoxic desaturation spoiled the smooth dye curve. In that noise, he recognized a useful signal—oximetry—because his mind was well prepared to understand what he saw happen. The process of turning his insight into more accurate, convenient and inexpensive saturation monitors still continues in dozens of laboratories and firms, while he continues to innovate.
In 1985, I was asked by Professor J. Payne to present the history of pulse oximetry to a meeting in the United Kingdom, later published (15). Using only published literature, I made serious errors because the original papers in Japanese were not authored by the inventor, but by surgeons. After seeing my paper, my former collaborator Professor Yoshi Honda, MD (1926–2003) of the Department of Physiology of Chiba University investigated this discovery and introduced me to the inventor, Takuo Aoyagi, Ph.D. Honda and I published corrections of this story in 1987 (16,17). I am indebted to Dr. Aoyagi and Akio Yamanishi (Minolta) for providing additional background described here related to the invention of pulse oximetry.
1. Squire JR. Instrument for measuring quantity of blood and its degree of oxygenation in the web of the hand. Clin Sci 1940;4:331–9
2. Millikan GA. The oximeter: an instrument for measuring continuously oxygen saturation of arterial blood in man. Rev Sci Instrum 1942;13:434–44
3. Wood EH. Oximetry. In: Glasser O, ed. Medical physics, Vol 2. Chicago: Year Book Publishers, 1950:664–80
4. Aoyagi T. Pulse oximetry: its invention, theory, and future. J Anesth 2003;17:259–66
5. Aoyagi T, Kishi M, Yamaguchi K, Watanabe S. [Improvement of an earpiece oximeter]. In: Abstracts of the 13th annual meeting of the Japanese Society for Medical Electronics and Biological Engineering. Osaka, 1974:90–1
6. Aoyagi T, Kishi M, Yamaguchi K, Nakajima S, Hirai H, Takase H, Kuse A. [New pulsed-type earpiece oximeter (author’s transl)]. Kokyu To Junkan 1975;23:709–13
7. Yamanishi A. The dawn of pulse oximeter. J Jap Soc Med Instru 2005;75:852–62
8. Takeda M, Shidara E, Atarashi J. Examination into calibration of photo plethysmography. Plethysmograph 1973;4:61–3
9. Nakajima S, Hirai Y, Takase H, Kuze A, Aoyagi T, Kishi M, Yamaguchi K. [New pulsed type earpiece oximeter]. Tokyo To Junkan 1975;23:709–13
10. Suzukawa M, Fugisawa M, Matsushita F, Suwa K, Yamamura H. [Use of pulse-type finger oximeter in anesthesia (author’s transl)]. Masui 1978;27:600–5
11. Yoshiya I, Shimada Y, Tanaka K. Spectrophotometric monitoring saturation in the fingertip. Med Biol Engl Comput 1980;18:27–32
12. Nakajima S, Ikeda K, Nishioka H. Clinical application of a new (fingertip type) pulse wave oximeter. Jpn J Surg 1979;41:57–61
13. Sarnquist FH, Todd C, Whitcher C. Accuracy of a new non-invasive saturation monitor. Anesthesiology 1980;53:S180
14. Shimada Y, Yoshiya I, Oka N, Hamaguri K. Effects of multiple scattering and peripheral circulation on arterial oxygen saturation measured with a pulse-type oximeter. Med Biol Engl Comput 1984;22:475–8
15. Severinghaus JW. Historical development of oxygenation monitoring. In: Payne JP, Severinghaus JW, eds. Pulse oximetry. Berlin: Springer, 1986:1–18
16. Severinghaus JW, Honda Y. The history of blood gas analysis VII. Pulse oximetry. J Clin Monit 1987;3:135–8
17. Severinghaus JW, Astrup PB. History of blood gas analysis. Int Anesthesiol Clin 1987;25:205–12
1The experimental ratio of ratios deviates from theory due largely to scattering by tissue. Aoyagi has sought and identified ways of minimizing this error. Cited Here...
2Konishi was on the MEBE board and may have seen Aoyagi’s notes. Cited Here...
3This was, as far as I know, Yamanishi’s first mention in print (7) that he had had the idea in January 1974 of using the arterial pulse to generate the AD/DC variations. Cited Here...