Why We Get Sick

by Randolph M. Nesse

  Now imagine a woodland songbird about to lay a clutch of eggs that she and her mate will incubate. Her reproductive success for this breeding season will depend entirely on those eggs. How many should she lay? Remember, she is not trying to assure the survival of the species, she is trying to maximize her own lifetime reproductive success. Laying too few eggs would obviously be foolish, but laying too many can also decrease her total lifetime reproduction if there is not enough food and some of the chicks die, or if she exhausts her energy reserves in caring for her brood and thus jeopardizes her chances of living until the next breeding season. These considerations apply equally to every individual in the woodland, but different birds reach different decisions on how many eggs to lay. If the average for a species is four eggs per pair, some pairs may have five and some only three. Do we conclude that all are trying for four but some can’t count? Or do we perhaps conclude that egg numbers are not subject to optimization by natural selection?

  An adaptationist forgoes such explanations until after considering the possibility that the birds deserve more credit. Could it be that, as a general rule, three eggs is best for those that lay only three, four for those that lay four, and so on? A simple sort of experiment provides the answer. If there are thirty nests with four eggs, leave ten randomly selected nests alone. From ten other nests remove an egg (the owners are now down to three) and add them to the ten remaining nests (four-egg birds now have five eggs). Now measure the average success of the three groups of birds: those allowed to choose their own egg number and those with one more or one less than they originally laid.

  If all relevant factors are carefully considered, the results of such studies usually vindicate the conclusion reached fifty years ago by Oxford ornithologist David Lack: birds adjust the number of eggs they lay to maximize their individual reproductive success. To do this requires an accurate assessment of their own individual health and capabilities and experience. Having to provide food for four nestlings is more difficult and hazardous than providing for only three. Nestlings in more crowded nests may weigh less at fledging and be less likely to survive the following winter. Conditions vary unpredictably from year to year, and worse-than-normal years are especially dangerous for the more crowded broods. Surely such knowledge enhances a naturalist’s pleasure in watching a pair of wild birds feed their young. Those birds are doing it right—not just right in general or on average, but right for them as unique individuals.

  In this discussion of clutch size we considered the optimal number of offspring. We ignored the fact that there are two kinds of offspring, male and female. Should our birds ideally produce one or the other or both in some ideal proportion? In the natural selection of sex ratio one overwhelmingly important strategy maximizes fitness: producing offspring of whichever sex is in short supply. Any frequenter of singles bars knows that the minority sex has a mating advantage. In nature, individuals that produce male offspring when females are scarce will be selected against because many of those males will never have offspring. If males are scarce, individuals that produce females will not have as many grand-offspring as individuals who produce males. The operation of this process of selection explains why there are equal numbers of males and females. This simple, elegant evolutionary explanation was first recognized by the great evolutionary geneticist R. A. Fisher in 1930. If you are thinking that an equal sex ratio arises because an individual has an equal chance of getting an X or a Y chromosome from its father, you are right, but this is a proximate explanation. The insufficiency of a proximate explanation is demonstrated by the many special cases such as ants and fig wasps, which are too complex to describe here but in which grossly unequal sex ratios turn out to match the more complex predictions

  Does natural selection in fact produce populations with exactly the same number of males and females? No, it does not, as would be expected by detailed reflection on factors such as the two sexes reaching maturity at different ages, differing death rates, differing costs to male and female parents, and other factors. Careful calculations support the conclusion that, for organisms with sex-determination and reproductive processes like ours, the sex ratio will stabilize when the parents collectively spend equal resources on rearing sons and rearing daughters. The demography of human and many other populations conforms closely to these expectations.

  We hope to convince you in the coming chapters that the modern theory of natural selection can be just as helpful in making medically important discoveries as it is for predicting the foraging patterns of beavers, the effects of altered clutch sizes of birds, and the sex ratios of mammals. The reasoning will always start with some prior information about health or disease and a question about evolved adaptation: Is this feature of the human body a part of some adaptive machinery? If so, what must the rest of the machinery be like? How can we test our predictions for unknown aspects of the machinery? If any feature of human biology seems functionally undesirable, how can natural selection have permitted it to arise? Is an undesirable trait the price of a positive feature? Could it be a trait that was adaptive in the Stone Age but that now causes disease? What are the medical consequences of natural selection acting to improve adaptation in our pathogens and parasites? These are just a few of the sorts of questions now routinely asked by evolutionary biologists, and efforts at answering them have been enormously fruitful.

  We must temper our enthusiasm with a note of caution. A question about function can have more than one right answer. For instance, the tongue is important both for chewing and for speech; the eyebrows, both for keeping the sweat out of the eyes and for communication. Second, the evolutionary history of a species or a disease is like any other kind of history. There is no experiment, in the usual sense, that we can do now to decide how long ago our ancestors first started to use fires for cooking or other purposes and what subsequent evolutionary effects that change may have had. History can be investigated only by examining the records it has left. Charred bones or even carbon deposits from an ancient campfire can be informative documents to people who know how to read them. Likewise, the chemical structure of proteins and DNA may be read to reveal relationships among now strikingly different organisms. Until a time machine is invented, we will not be able to go back and watch the evolution of major traits, but we can nonetheless reconstruct prehistoric events by the records they left in fossils, carbon traces, structures, and behavioral tendencies, as well as protein and DNA structures. Even when we cannot reconstruct the history of a trait, we can often still be confident that it was shaped by natural selection. This can be supported by evidence for its function in other species and by the match between the trait’s characteristics and its functions.

  So hypotheses about the evolutionary origins and functions of a trait, just like hypotheses about proximate aspects of a trait, need testing and are often testable. Special difficulties attend the testing of evolutionary hypotheses, but these are no reason to give up—they just make the work more challenging and interesting. Do we claim to test evolutionary hypotheses in this book? Not really. While we will try to separate speculation from fact, and will cite evidence for most of our examples, hardly any of them can be considered proven by the evidence we present. Some of the examples are based on many studies, each with different data bearing on a different aspect of the problem, but even this is often insufficient.

  Our goal is not to prove any specific hypothesis but to show that evolutionary questions are interesting, important, and testable. We want people to start asking new questions. So, without apology, we ask questions about the possible evolutionary significance of diverse aspects of disease and offer answers that are often speculative. Some people will, despite our warnings, insist on taking these speculations as facts. Perhaps in a few years Darwinian medicine will have enough confirmed findings to fill a book. For now, our goal is not to exhaustively test a few hypotheses but to encourage patients, doctors, and researchers to ask new questions about why disease exists. As Gertrude Stein said on her de
athbed, “The answer, the answer, the answer. What is the answer?… In that case, what is the question?”

  3

  SIGNS AND SYMPTOMS OF INFECTIOUS DISEASE

  Suppose you are on the side of the mice in cat-mouse conflicts. The mice say they hate the smell of a cat. It makes them jittery and unable to concentrate on important matters, such as food and courtship and babies. You know of a drug that will dull the sense of smell so that the mice will no longer be bothered by the odor of cats. Do you prescribe the drug? Probably not. The ability to detect cat odor, however unpleasant it may be, is a valuable asset for mice. The presence of the cat’s smell may signal the imminent arrival of its claws and teeth, and avoiding these is far more important than the stress of an unpleasant odor.

  More realistically, suppose you are a pediatrician treating children with colds. Colds bring many symptoms that children dislike—runny nose, headache, fever, and malaise. Acetaminophen (e.g., Tylenol) can reduce or eliminate some of these symptoms. Do you tell the parents of cold-stricken children to give them acetaminophen? If you are a traditional physician or are in the habit of using acetaminophen yourself to relieve similar symptoms, you probably do. Is this wise? Consider the analogy between acetaminophen and the drug we were considering for the mice. Like the smell of a cat, fever is unpleasant but useful. It is an adaptation shaped by natural selection specifically to fight infection.

  FEVER AS DEFENSE AGAINST INFECTION

  Matt Kluger, a physiologist at the Lovelace Institute, believes that “there is overwhelming evidence in favor of fever being an adaptive host response to infection that has persisted throughout the animal kingdom for hundreds of millions of years.” He believes that using drugs to suppress fever may sometimes make people sicker—and even kill them. Some of the best evidence comes from his laboratory. In one experiment, he showed that even cold-blooded lizards benefit from fever. When infected, they seek out a place warm enough to raise their body temperature about two degrees Celsius. If they cannot move to a warm place, they are more likely to die. Baby rabbits also cannot generate a fever, so when they are sick they too seek out a warm place to raise their body temperature. Adult rabbits do get fever when infected, but if the fever is blocked with a fever-lowering drug, they are more likely to die.

  Fever results not from any mistake in temperature regulation but from the activation of a sophisticated evolved mechanism. If you put a rat with a two-degree fever into a very hot room, the rat activates its cooling mechanisms to keep its body temperature two degrees above normal. If you put it into a cooler room, it activates heat-conservation mechanisms to maintain that two-degree fever. Body temperature is carefully regulated even during fever; the thermostat is just set a bit higher.

  Perhaps the most dramatic human evidence for the value of fever comes from studies by Julius Wagner-Jauregg in the early decades of this century. After noting that some syphilis patients improved after getting malaria and that syphilis was rare in areas where malaria was common, he intentionally infected thousands of syphilis patients with malaria. In an era when fewer than one in a hundred syphilis patients recovered, this treatment achieved remission rates of 30 percent, an advance that made Wagner-Jauregg worthy of his 1927 Nobel Prize in Physiology or Medicine. At that time, the value of fever was much more widely recognized than it is now.

  Doctors still say, as the joke goes, “Take two aspirin and call me in the morning.” This isn’t so surprising, given that only a few human studies have tried to evaluate fever as an adaptation to combat infection. In one study, children with chicken pox who were given acetaminophen took on average about a day longer to recover than those who took a placebo (sugar pill). In another study, fifty-six volunteers got colds on purpose, from an infectious nasal spray. Some then took aspirin or acetaminophen, others a placebo. The placebo group had a significantly higher antibody response and less nasal stuffiness. They also had a slightly shorter period of infectious dispersal of viruses. The paucity of detailed studies of this sort, given that so many drugs are used to relieve the symptoms of so many infectious diseases in so many patients, shows the reluctance to study the adaptive aspects of unpleasant symptoms.

  This may be about to change. Dr. Dennis Stevens, professor of medicine at the University of Washington, cites “evidence that treating a fever in certain circumstances actually may make it more likely the patient will develop septic shock.” Medications that block fever apparently interfere with the normal mechanisms that regulate the body’s response to infection, with results that may be fatal.

  Before going on to other defenses, we should emphasize that a given expression of a defense need not be adaptive, and that even when it is, it may not be essential. We would not dream of recommending that people never take drugs to reduce fever. Even if many studies were to establish decisively that fever is usually important for combating infection, that would not justify an unbending policy of encouraging fever or even of routinely letting it rise to its natural level. An evolutionary perspective draws attention to the costs as well as the benefits of an adaptation like fever. If there were no compensating disadvantage in having the human body operate at 40° C. (103° F.), it ought to stay at that temperature all the time, so as to prevent infections from ever getting started. But even this moderate fever has costs; it depletes nutrient reserves 20 percent faster and causes temporary male sterility. Still higher fevers can cause delirium and perhaps seizures and lasting tissue damage. It should also be realized that no regulation mechanism can perfectly anticipate all situations. We would expect temperature to rise, on average, to a level close to an optimum to fight infection, but because regulatory precision is limited, fever will sometimes rise too much and at other times not enough.

  Even if we knew that it would prolong an infection, we would still sometimes want to block fever. Maintaining and improving health are, after all, not the only goals of medicine. If she is about to sing Nanetta in a Metropolitan Opera performance of Falstaff, soprano Barbara Bonney might well decide to take a medication to relieve a touch of laryngitis, even if she knew it might delay her complete recovery. The rest of us may choose to take drugs just to feel better during a cold, even though our recovery might be slower.

  The important point, with respect to the adaptive significance of fever, is that we need to know what we are doing before we interfere with it. At present we don’t. If discomfort were the whole story, we could always choose to reduce or eliminate it. But if reducing fever will often delay recovery or increase the likelihood of secondary infection, we should interfere only when the expected gain is worth the risk. We hope that medical research will soon produce the evidence to help doctors and patients decide when fever is and is not useful.

  IRON WITHHOLDING

  Our bodies have a related defense mechanism, of which most people are unaware and which physicians sometimes unwittingly attempt to frustrate. Here are some clues about how it works. A patient with chronic tuberculosis is found to have a low level of iron in his blood. A physician concludes that correcting the anemia may increase the patient’s resistance, so she gives him an iron supplement. The patient’s infection gets worse. Another clue: Zulu men often drink beer made in iron pots and often get serious liver infections caused by an amoeba. In contrast, less than 10 percent of Masai tribesmen have amoebic infections. They are herdsmen and drink large amounts of milk. When a group of Masai were given iron supplements, 88 percent soon got an amoebic infection. In another study, well-meaning investigators gave iron to supplement the low levels found in Somali nomads. At the end of one month, 38 percent had infections versus 8 percent of those who had not taken the supplements.

  Yet another clue: eggs are a rich source of nutrients, but their porous shells can be readily penetrated by bacteria. So how can eggs stay fresh so long? They contain lots of iron, but it is all in the yolk, none in the surrounding white. Egg white protein is 12 percent conalbumin, a molecule whose structure tightly binds iron and thereby withholds it from any bacteria th
at might get in. Prior to the antibiotic era, egg whites were used to treat infections.

  The protein in human milk is 20 percent lactoferrin, another molecule designed to bind iron. Cow’s milk has only about 2 percent lactoferrin, and breast-fed babies consequently have fewer infections than those fed from bottles. Lactoferrin is also concentrated in tears and saliva and especially at wounds, where an elevated acidity makes it especially efficient in binding iron. The researchers who discovered conalbumin predicted that there should be a similar molecule to bind iron within the body. This led to the discovery of transferrin, another protein that binds iron tightly. Transferrin releases iron only to cells that carry special recognition markers. Bacteria lack the needed code and can’t get the iron. People suffering from protein deprivation may have levels of transferrin less than 10 percent of normal. If they receive iron supplements before the body has time to rebuild its supply of transferrin, free iron in the blood makes fatal infections likely—as has been a tragic outcome of some attempts to relieve victims of famine.

  By now the nature of this defense is surely obvious. Iron is a crucial and scarce resource for bacteria, and their hosts have evolved a wide variety of mechanisms to keep them from getting it. In the presence of infection, the body releases a chemical called leukocyte endogenous mediator (LEM), which both raises body temperature and greatly decreases the availability of iron in the blood. Iron absorption by the gut is also decreased during infection. Even our food preferences change. In the midst of a bout of influenza, such iron-rich foods as ham and eggs suddenly seem disgusting; we prefer tea and toast. This is just the ticket for keeping iron away from pathogens. We tend now to think of bloodletting as an example of early medical ignorance, but perhaps, as Kluger has suggested, it did help some patients by lowering their iron levels.