The oxygen that we obligate aerobes need for survival is transported from the lungs to peripheral tissues by the hemoglobin that is densely packed in our red blood cells (erythrocytes). Hemoglobin is the most intensively studied protein in the world, and its structure is known in intimate detail. It is made up of four protein subunits. Normal human adult hemoglobin A is a tetramer comprised of a pair of alpha-globin chains and a pair of beta-globin chains. Nestled deep in each of these protein's folds, is a planar structure called a porphyrin, which binds in its center a single atom of iron, most commonly in the 2+ valency state (see diagram below). The iron-porphyrin group is called heme. In the lung, oxygen diffuses across the alveolar membrane, and then the red cell membrane in lung capillaries. When it encounters a molecule of hemoglobin, it wedges itself between the iron atom and a nitrogen attached to the globin chain, which helps to hold the heme group in place in the protein. One molecule of hemoglobin with its four heme groups is capable of binding four molecules of diatomic oxygen, O2. The loaded pigment is called oxyhemoglobin, and it is a brilliant red color as in arterial blood. Pressure from dissolved oxygen in plasma and in the surroundings in the red cell helps to keep the oxygen on its binding site.
As the blood circulates to the periphery, the small amount of dissolved oxygen is consumed first by cells in organs and tissues. This release in pressure makes available the much larger reservoir of heme-bound oxygen, which begins a sequential unloading of its four oxygen molecules. At the most, under normal circumstances only 3 molecules of oxygen are unloaded. Partially or fully unloaded hemoglobin is called deoxyhemoglobin. It is a dark blue to purplish color as in venous blood.
During oxygen unloading, the hemoglobin tetramer undergoes subtle intramolecular conformational changes called cooperativity. As a result of cooperativity, once the first oxygen has been unloaded, the unloading of the second oxygen is facilitated. The second oxygen can dissociate after a much smaller change in oxygen pressure than was needed to unload the first. Another conformational change facilitates dissociation of the third oxygen. Cooperativity is an important phenomenon that permits the loading and unloading of large amounts of oxygen at physiologically relevant oxygen pressures. Chemicals that interfere with oxygen transport and/or cooperativity are potentially lethal.
Oxygen lack is known as hypoxia; the complete absence of oxygen is called anoxia. There are four basic types of hypoxia. Stagnant or hypokinetic hypoxia is characterized by a decreased rate of blood flow, but a normal arterial oxygen tension (pressure). The arterial oxygen tension is related to the amount of oxygen dissolved in blood plasma, not the much larger pool that is bound to hemoglobin. This condition could result from an overdose of a vasodilator drug, such as nitroglycerin. Arterial or anoxic hypoxia occurs when there is a problem in oxygenating the blood, and the arterial oxygen tension is abnormally low. Anoxic hypoxia can be the result of exposure to an irritant gas such as phosgene or various oxides of nitrogen, including nitric oxide. These gases penetrate deeply into the lung, and their irritant effects provoke local inflammatory changes resulting in increased capillary permeability. The outpouring of fluids into the alveoli impairs gas exchange resulting in increased hypoxia, which triggers further capillary vasodilatation in a vicious cycle. The end result is pulmonary edema (adult respiratory distress syndrome). A frothy, sometimes bloody exudate may be seen from the nose and mouth. The low arterial oxygen tension is sensed by chemoreceptors in the carotid bodies, epithelioid structures located just above the bifurcation of the common carotid artery on each side of the neck. Activation of the chemoreceptors results in the labored breathing pattern known as dyspnea, as the body tries to compensate for the hypoxia.
The third type of hypoxia is known as anemic, in which there is a decrease in the oxygen content of the blood, but the arterial oxygen tension may be normal. Anemic hypoxia might result from blood loss, or chemicals that prevent oxygen binding to hemoglobin or affect cooperativity. Purists would not include histotoxic hypoxia as a fourth type, because both the arterial oxygen tension and the oxygen content of blood are normal. The lesion here is caused by poisons such as cyanide or hydrogen sulfide, which block the utilization of oxygen at the cellular level.
The story, "A Woman with a Headache" can be found in Roueché, B., The Man Who Grew Two Breasts, Truman Talley Books/Plume, New York, New York, 1995.
Also found in Rouché, B., The Orange Man and Other Narratives of Medical Detection, Little Brown and Company, 1971. (Out of Print)
"A Woman with a Headache" Addendum
Carbon monoxide ranks among the leading causes of morbidity and mortality due to poisonings world-wide, but even the most common of poisonings is a rare event in the career of a given physician. A headache and florid complexion are seen far more commonly in alcoholic inebriation, and in this case this erroneous impression was further reinforced by the unconventional behavior of the Andersons. Hypoxia, or a lack of oxygen, has some of the initial effects of central nervous system depressant drugs. This loss of inhibitions is usually explained as a selective, early depression of higher inhibitory control mechanisms, a loosening of the brakes, as it were, on more primitive emotions and behavior. The mild hypoxia was also responsible for the headache, perhaps secondary to dilation of cerebral blood vessels. It was easy to be misled by Mrs. Anderson's pre-existing kyphoscolosis.
As indicated in the story, carbon monoxide and oxygen bind reversibly and competitively to the same heme sites on the hemoglobin tetramer. Actually the affinity of carbon monoxide is about 250 times greater than that of oxygen. That works out to a dangerous concentration of 50% carboxyhemoglobin, when the blood is in equilibrium with a gas mixture containing only 0.1% carbon monoxide. Any increase in the carbon monoxide binding translates into an exactly parallel decrease in oxygen content. It is not quite true, however, that carbon monoxide inhibits the release of oxygen bound to hemoglobin. The measurements make it look that way, but that is an experimental artifact, which makes look as if, "It inhibits the release of whatever oxygen has managed to combine with hemoglobin" (see diagram below). It seems probable today that there are hybrid species of tetramers which have bound both oxygen and carbon monoxide, instead of tetramers that have bound only oxygen or only carbon monoxide. Let us take as an example of such a hybrid species a tetramer that has bound two oxygens and two carbon monoxides. If an oxygen comes off first, cooperativity can facilitate the release of the second oxygen, but then there is no more oxygen left to take advantage of cooperativity. Thus, carbon monoxide poisoning is compounded by a loss in the number of opportunities for cooperativity to facilitate oxygen unloading. Victims are more severely compromised than in an equivalent anemia where there is less than the normal complement of hemoglobin, but the hemoglobin that is present is able to function normally in terms of cooperativity.
For many years it was thought that carbon monoxide poisoning represented a simple asphyxia due only to exclusion of oxygen. There were numerous unconvincing experiments purporting to show that carbon monoxide had adverse effects on cellular metabolism. It is only in recent years that there have been convincing studies in situ to show that carbon monoxide has direct, inhibitory effects on brain aerobic metabolism. These brain effects persist in experimental animals even after the blood concentration of carboxyhemoglobin has returned to normal. This has led to the recommendation that severe poisonings be treated in hyperbaric chambers. Although the experimental evidence indicates that hyperbaric oxygen is, indeed, superior to pure oxygen at atmospheric pressure, there is the practical consideration of how close the victim is to an available chamber. The logistics of transportation may be so formidable that conventional treatment is the only viable alternative.
Carboxyhemoglobin is indistinguishable from oxyhemoglobin to the naked eye. Since carbon monoxide is not consumed in peripheral tissue, as is oxygen, it continues its circulation in veins back to the lungs where some may be excreted or the carbon monoxide load may be increased. The presence of this red pigment in the venous return is responsible for the abnormally red color of the skin and mucus membranes.
Carbon monoxide poisoning is an example of anemic hypoxia where the arterial oxygen tension remains near normal until substantial blood concentrations of carboxyhemoglobin are reached. Thus, there are no signs of respiratory distress, such as dyspnea. Instead, in response to a slowly developing hypoxia, the body initiates widespread peripheral vasodilation. In order to compensate for the increased volume of the vascular space, there must be an increased cardiac output. The cardiac output cannot increase indefinitely, and eventually the victim may faint, as did both of the Andersons, because of brain hypoxia, and remain unconscious for long periods before death.
It was appropriately pointed out in the story that carbon monoxide is not a cumulative poison. Recovery from recurrent acute exposures is usually complete. However in any hypoxic insult of sufficient severity or duration, such as strangulation or drowning, there is a possibility of long-lasting or permanent neurologic sequelae. Such cases of carbon monoxide poisoning have been documented in the medical literature.
About one per cent of the total circulating hemoglobin in non-smoking, non-occupationally exposed humans is in the form of carboxyhemoglobin. For many years it was assumed that this represented ambient, environmental exposure. It has since been learned that much or all of this carbon monoxide is actually generated in the body. Paradoxically, it originates during the breakdown of hemoglobin heme. These endogenous levels rise when there is a demand for extra hemoglobin, the consequences of which are that eventually more hemoglobin is broken down as in pregnancy.
FiornalR for headache is an empirical mixture of aspirin, caffeine and butalbital, a barbiturate. It is unlikely to be recommended by modern physicians. NembutalR is another barbiturate, which was prescribed by a different physician. Codeine was prescribed for the headache pain, and later Mrs. Anderson was given DemerolR, which is also an opioid analgesic. This illustrates a common problem in dealing with more than one physician at a time, when there is faulty communication between or among them.
In the early 1990's, a man was found dead in his automobile in a Burlington, VT parking lot. The ignition and air conditioning were on, but the vehicle had run out of gas. Analysis of a blood sample showed clearly lethal levels of carboxyhemoglobin. A reputable industrial hygienist was hired to perform tests on the car. The results of the tests were unsatisfactory. Under no circumstances could the hygienist get the levels of carbon monoxide inside the vehicle high enough to account for the lethal blood level of carboxyhemoglobin. Was this an accident, or was it a clever and still unsolved murder? The victim could have been exposed to carbon monoxide at some other site, and the body placed in the car, although that scenario might have resulted in signs of a struggle. Alternatively, a small canister of carbon monoxide could have been opened and placed in the car to create an overwhelming concentration of the gas. On entering the car, the victim quickly loses consciousness perhaps having only time to adjust his seat belt and turn on the ignition. The air conditioner clears the vehicle of carbon monoxide, and the empty canister is retrieved later for disposal.
Smith, R. P. Chapter 11, Toxic responses of the blood. In Casarett and Doull's Toxicology The Basic Science of Poisons, 5th ed., C. D. Klaassen, Ed., McGraw-Hill, New York, 1996.
The story, "Eleven Blue Men" can be found in Roueché, B., The Medical Detectives, Truman Talley Books/Plume, New York, New York, 1991.
Also found in Rouché, B., Eleven Blue Men and Other Narratives of Medical Detection, Little Brown and Company, Boston, 1953. (Out of Print)
Also found in Rouché, B., Annals of Epidemiology, Little Brown and Company, Boston, 1967. (Out of Print)
"Eleven Blue Men" Addendum
After having read "A Woman with a Headache", one wonders how the initial misdiagnosis of carbon monoxide poisoning by the examining physician could have been entertained in this story? These victims were blue, not red. The condition is called cyanosis, but cyanosis is much more commonly due to abnormally high concentrations of bluish or purple deoxyhemoglobin in arterial blood. It indicates a problem in hemoglobin oxygenation, which may have any of a number of causes. Asphyxia as in strangulation or drowning is one. Fluid in the lungs (pulmonary edema or adult respiratory distress syndrome) is another. Methemoglobinemia is an unusual cause of cyanosis. Methemoglobin ("changed hemoglobin") is a greenish to chocolate brown to almost black pigment depending on its concentration in blood. It occurs as a result of the oxidation of the heme iron from 2+ to 3+. In effect, the iron has rusted. In that state it can no longer combine reversibly with oxygen or carbon monoxide, and oxygen transport is compromised.
Given a sufficiently high concentration of nitrite in blood all of the heme irons can be oxidized. At lower concentrations of nitrite, where oxidation is incomplete, so-called valency hybrid species are generated. Two such species are known, in one the two alpha-chain hemes are oxidized and the two beta chain are reduced; the converse is true in the other species. The reduced subunits function normally in that they continue to bind oxygen. In non-fatal methemoglobinemia, four species of hemoglobin exist in blood: the fully reduced, the fully oxidized and the two valency hybrid species. The existence of hybrid species in the case of carbon monoxide was only a logical hypothesis, but the existence of the hybrid species of methemoglobin can be shown unequivocally by electrophoretic separation. The hybrid species have the same effect on oxygen dissociation as do the postulated species in carbon monoxide poisoning, and for the same reason. They make it appear that the residual oxygen is bound more tightly than normal, but the true effect is that, again, there is simply a loss in the number of opportunities for cooperativity to facilitate oxygen dissociation. In the hybrid species there is only one opportunity to utilize cooperativity. Thus, a person with a fifty percent methemoglobinemia is also more severely compromised than a person with a simple fifty per cent anemia.
Finally, nitrite has a third effect in that it relaxes directly most smooth muscles including the smooth muscle in blood vessels. That results in vasodilation, and the volume in the vascular space becomes greater than the blood volume. Blood pressure drops precipitously. The brain is no longer perfused adequately, and the victim faints (hypovolemic shock). Fainting, in this case is autoprotective, since it usually improves circulation to the brain. Elevating the feet has a further salutatory effect. The clinical observations in the story did not include hypotension, although blood pressure measurements must have been made. Since it is very unlikely that they would have required the patients to stand for the measurements, we must assume that the blood pressures were taken in the supine position, which would tend to normalize them.
It is more difficult to explain the retching, abdominal cramps, diarrhea and rigidity. Vomiting is perhaps the most common and nonspecific sign of poisoning. The well-known effects of nitrite on smooth muscle would suggest, however, that gut relaxation should occur instead of the contractions that resulted in cramps and diarrhea. Perhaps some or all of these signs could be ascribed to alcohol withdrawal.
It is now believed that the hemoglobin oxidation and the smooth muscle relaxation produced by nitrite are really due to its spontaneous conversion to nitric oxide, a toxic, free radical gas found, for example, in automobile exhaust. Do not confuse nitric oxide with the anesthetic, nitrous oxide, or "laughing gas". Nitric oxide is a contributor to air pollution and acid rain. Incredibly, this noxious material is a natural mediator in the body, where it is synthesized from an amino acid to perform a variety of important physiological functions, including contributions to the regulation of blood pressure by relaxing vascular smooth muscle. The story indicates that nitrite has been used to manage heart conditions and high blood pressure. Perhaps that was a short-lived practice many years ago. Today the drug that might be employed for the treatment of angina is nitroglycerin, the same chemical as the explosive. Nitroglycerin has been in clinical use for three quarters of a century, and no one knew how it worked until nitric oxide was discovered. Nitroglycerin and nitrite belong to a small family of drugs known as the nitric oxide vasodilators, because they are converted in the body to nitric oxide.
Much is made in the story about the rarity of sodium nitrite poisoning, and, indeed, it is very unusual. However, methemoglobinemia can be caused by many chemicals other than nitrites, so the resulting pathophysiological condition is not as rare. All of the nitric oxide vasodilators, in toxic doses, generate methemoglobin to some degree, and also produce hypovolemic shock. A wide variety of organic chemicals such as aniline and nitrobenzene and hundreds of derivatives of them, including some aniline dyes, generate methemoglobin without vasodilation. Unlike nitrite, these organic compounds must be metabolized to active forms, so they are not reactive per se in shed blood.
Ten of the men in the story recovered spontaneously, because there is an enzyme, methemoglobin reductase, in red blood cells, which is responsible for maintaining hemoglobin heme iron in its functional 2+ valency state (see diagram below). Because the iron is in close contact with its reactive load of molecular oxygen, "accidental" oxidation of the iron occurs continuously. That damage is reversed continuously by methemoglobin reductase.
There are, however, rare individuals born with a congenital deficiency of methemoglobin reductase. There is a relatively high incidence of this trait among Alaskan Eskimos and among some families in Appalachia. These individuals may go through life with as much as half of their total hemoglobin in the form of methemoglobin. As it turns out, however, they are more blue than sick. They can compensate for the defect by making more red blood cells than normal individuals (polycythemia). Even though these extra red cells are defective, the functional fraction of hemoglobin is increased. These phenotypes are exquisitely sensitive to methemoglobin-generating chemicals.
A redox dye, methylene blue, can be used to treat acquired methemoglobinemia, even in people deficient in methemoglobin reductase. The dye is reduced to its colorless ("leuco") form by an enzyme in red blood cells that seems to have no physiological function. The leuco-dye can then nonenzymatically reduce methemoglobin to hemoglobin. Thus, the intravenous injection of methylene blue, while invasive, can be life-saving. A physician in Appalachia amazed some of his blue patients by giving them a blue dye that turned them pink, at least temporarily. Lifetime intravenous injections of methylene blue, however, are not a satisfactory solution for a problem that is primarily cosmetic. Large oral doses of vitamin C are somewhat less effective, but much safer.
In contrast to sodium nitrite, sodium nitrate is biologically inert. Many common bacteria, however, including some that are normally found colonized in the human gastrointestinal tract are capable of reducing nitrate to nitrite. During the decade of the 1940's a miniepidemic of a condition known variously as well water methemoglobinemia, or the blue baby syndrome, appeared in the midwestern United States. This was a disease of infants under one year of age, who lived in rural areas, and consumed water from shallow dug wells. Other members of the family, also consuming well water, were typically not affected. Because of the intensive use of nitrate fertilizers and close proximity to feed lots heavily contaminated with animal wastes, the well water often contained high concentrations of nitrate. In a typical case, an apparently normal infant would be brought home from the hospital. After a variable period of time, cyanosis would be noted. The infant would be rushed to the hospital, and placed under observation (this was before methylene blue). There would be a slow spontaneous recovery due to the activity of methemoglobin reductase. The infant would be discharged, and, not uncommonly, the same senario would be repeated.
The only other regular feature of the syndrome was gastrointestinal upset with nausea, vomiting and diarrhea. At least some of the infants seemed to have a high gastric pH, so that the environment of the stomach and the short adjoining portions of the small intestine were more alkaline than usual. An acidic gastric pH is a deterrent to the growth of most bacteria. The hypothesis was formulated, that in the case of affected infants, the normal intestinal microflora were able to establish colonies much higher in the intestinal tract and closer to the stomach than they were normally found, because of the more favorable pH. In older children and adults, nitrate is rapidly absorbed high in the intestine before it reaches the area containing microflora. Thus, nitrate and bacteria do not come into intimate contact. With the bacteria much closer to the stomach in affected infants, there is opportunity for contact, reduction of nitrate to nitrite, and absorption of nitrite into the blood stream. In theory, the problem would be self-solving after one year of age, when infant gastric acidity falls to adult levels. The hypothesis was supported by the observations of physicians who deliberately introduced nitrate into the gastrointestinal tract through a colostomy in an infant. Sure enough, the infant became methemoglobinemic. Needless-to-say, this was in the days before human subject review boards became popular. It also fit with the well-known sensitivity of ruminants to nitrate. Many plants accumulate nitrate above their nutritional needs and store the excess. The lower chambers of the ruminant stomach contain bacterial microflora that aid in digestion. So, in the ruminant stomach, plants high in nitrate come in contact with reducing bacteria, and there have been epidemics of methemoglobinemia among cattle herds.
Without any conscious or deliberate public health intervention, the cases of well water methemoglobinemia reached a peak in the mid-1940's, and began to decline. Only sporadic cases are described today. This strange episode constitutes the basis for the setting of nitrate levels in potable water supplies by the Environmental Protection Agency. Nitrate in drinking water became a key issue for environmental activists. Many of the potable water supplies in the midwest have nitrate levels above the EPA limits during the growing season. However, there is a growing group of investigators, who believe that nitrate in drinking water may be only a red herring, and they have begun to challenge the hypothesis about the etiology of the blue baby syndrome.
For example, it is very clear that the major source of nitrate exposure for human adults is food, not water. For most of the population in the United States only three per cent of the total nitrate consumed comes from water, and ninety seven per cent from food. Thus, the incremental contribution of nitrate in drinking water to the total adult intake is trivial. Infants, however, constitute a special case. Breast-fed infants are exposed to very little nitrate, but infants fed formula are exposed to nitrate in the water used for its preparation.
It is very difficult to study cases of well water methemoglobinemia, when the number of them has declined so precipitously. You can't study an epidemic that is nearly over. However, a recent small scale study in Israel generated some startling results. The investigators found that infantile methemoglobinemia is more frequent among hospitalized infants than previously thought. The hospitalizations were usually a result of acute diarrhea. There was no correlation between methemoglobinemia and the consumption of nitrate/nitrite in food or water. Despite this, the infants had unusually high blood levels of nitrates. These findings cast some doubt on the long-standing hypothesis about the etiology of infantile methemoglobinemia. They point away from nitrate intake and toward an abnormal increase in the synthesis of nitric oxide triggered in an as yet unknown way by diarrhea. Some infections are known to stimulate endogenous nitric oxide production by white blood cells. The nitric oxide could react with oxyhemoglobin to generate methemoglobin and nitrate, which would increase blood nitrate concentrations in a way that would not correlate with nitrate intake. Again, this problem needs further study before communities invest heavily in equipment to remove nitrate from drinking water, which may have nothing to do with the blue baby syndrome.
Like botulinum toxin, sodium nitrite has found a tiny niche in modern medicine as part of an overall regimen to treat cyanide poisoning. It is injected intravenously in carefully controlled doses to deliberately induce a tolerable level of circulating methemoglobin (see diagram below). Methemoglobin avidly binds cyanide as cyanmethemoglobin, a biologically inactive complex. Although the cyanide is bound very tightly, cyanmethemoglobin is a fully reversible complex that can dissociate free cyanide as the blood concentration of cyanide falls. This might cause a relapse in the course of the recovery. Another chemical, sodium thiosulfate, is also given to permanently convert cyanide to thiocyanate, a relatively nontoxic form that is excreted in the urine. These treatment kits are found in most emergency rooms, but they were designed for adults. On one occasion when a toddler was given the full adult dose of nitrite, he died of an overwhelming methemoglobinemia.
The narrator in this story is never identified, but at the end he is saying to himself, "The toxic dose of sodium nitrite is three grains" (about 200 mg). There are several troubling aspects about such statements. "Toxicity" is not defined, although it presumably means the dose that produces the distinctive blue color (cyanosis) of the skin and mucous membranes. Much is made in the story about the rarity of sodium nitrite poisoning. Since it is a rare condition, it is unlikely that the "toxic" dose in humans would be known with such precision. If it were as common as aspirin poisoning, it would be clear that the toxic dose is not a single number, but a range of numbers. People vary widely in sensitivity to the toxic effects of chemicals, as pointed out in the addendum to the first story. A 200 pound person is likely to be less sensitive to the same dose of a chemical than is a 100 pound person. Thus, it is common to express toxic or lethal doses in terms of the average dose or mean dose to produce a given effect that must be clearly defined. Variation around the mean or average lethal dose in a large human population might be as much as 10%. The additional nitrite contributed by salting the oatmeal, however, represents a 20% increment over the "average toxic dose". It would be entirely reasonable to conclude that such an increase might convert a borderline dose into a toxic dose.
The laboratory played an important role in making and confirming the diagnosis in these cases. The positive findings, however, were extremely fortuitous. The enzyme, methemoglobin reductase, which is contained within red blood cells is just as active in shed blood as it is in circulating blood. Moreover, it remains active even at refrigerator temperatures, albeit somewhat slower than at body temperature. There have been cases where samples known to contain methemoglobin showed negative results after storage overnight in a refrigerator. The positive finding of nitrite in blood samples is even more suspicious. Since nitrite reacts avidly with hemoglobin, they should have found the product of that reaction, nitrate, but not nitrite. Had the plasma been separated from the red cells immediately after the sample was drawn, the chances for a positive finding of nitrite in plasma would be much improved.
Smith, R. P. Chapter 11, Toxic responses of the blood. In Casarett and Doull's Toxicology The Basic Science of Poisons, 5th ed., C. D. Klaassen, Ed., McGraw-Hill, New York, 1996.