Patrick was a big young man who'd played the backfield before he got sick with leukemia. On a July afternoon eight years ago he was lying pale and immobile in his hospital bed. He had that unmistakable disbelieving look on his face of someone who had just realized he was dying. But it wasn't the leukemia that was killing him; it was the fungus infection clogging up his lungs. You could see the flesh between his ribs and above his collarbones being sucked in rhythmically with every noisy breath he took.
A sample of his sputum had already been cultured. It was Aspergillus, no doubt about it--a ubiquitous and usually harmless fungus that likes to set up residence in air conditioners. But when our immune defenses are down, weakened by a bone marrow transplant or chemotherapy, as was the case with Patrick, it can become a fulminant invader. A CT scan showed that clumps of the fungus were plugging up Patrick's bronchial tubes like stalactites and stalagmites inside a narrow cave. Patrick needed a bronchoscopy to free his airways and make his last days easier. As an ear, nose, and throat surgeon, I was called to do what should have been a routine procedure.
What the million-dollar CT scanner hadn't revealed was that the largest fungus ball had bored a hole clear through the wall of one of Patrick's bronchi into his pulmonary artery. The fungus was sitting there like some preposterous cork in a dam. Unsuspecting, I delicately pulled the clump away with my forceps. As it came loose, high-pressure blood gushed into the airway and flooded Patrick's lungs. There was nothing to be done-- no time to crack the chest, find the bleeder, and patch the leak. I could only cradle Patrick's head in my hands until the monitors recording his vital signs went quiet.
Minnesota law requires autopsies of all deaths occurring in the operating room, and I felt duty-bound to accompany Patrick to this last stop in the hospital. But the tiled morgue was strangely lonely when I stepped inside. The pathologist looked surprised to see me and turned off the radio when I came in. He wasn't used to company.
The pathologist was skeptical when I told him the fungus seemed to have bored a passage--called a fistula--from the bronchus to the blood vessel. But once he'd opened Patrick's chest, he found the fistula and skillfully dissected it out. Thus the autopsy confirmed what the scanner couldn't show, and we learned of a new complication of Aspergillus infection. No one had known that the fungus could cause such severe damage to a patient's airway and blood vessels. Yet only two of us were present to witness the finding. No announcement was made to the medical staff at large, informing them that an autopsy was under way. Even Patrick's team of oncologists, and the students working with them, were too busy with the living to attend.
This would not have been the case just 24 years ago, when I was a medical student. Then, attending autopsies was a required part of the curriculum. Students jammed into the tiered gallery and compulsively took notes as the pathologist carefully examined the deceased, organ by organ. The "gunners"--those with the highest grade-point averages and aspirations of getting an internship at Harvard or Stanford--usually sat in the first row and leaned over the railing to get a better look. Interns, residents, and attending physicians crowded in, too. The hospital's PA system summoned them with coded messages such as "The green light is on" or "The red door is open." I don't think many patients on the wards ever caught on.
The pathologist was teaching us fledgling doctors the singular value of following the patient all the way from diagnosis to postmortem. Medicine, we learned in the morgue, is not an exact science. Diagnoses and treatments were usually on the mark, but sometimes doctors missed something--an unusual disease complication, say, or a drug side effect--and sometimes they were just plain wrong. If the diagnosis was correct, we left the morgue confident of our hard-won clinical skills. If the diagnosis was wrong, we left with that peculiarly welcome humility that comes with knowing how things might be done better next time. Furthermore, autopsies quite often turned up surprises, conditions that weren't suspected at all, or causes of death quite unrelated to the diseases for which patients had been treated. They were eye-opening and reminded us to be always questioning and curious. Nowadays, however, fewer than half our medical schools require students to attend an autopsy, and many young doctors don't know where the morgue is anymore. Even if they find it, they'll have far less opportunity to learn anything there. American hospitals today conduct autopsies on only about 10 percent of their deceased patients. Two dozen years ago the rate varied from 30 percent in community hospitals to 50 percent in university hospitals.
Why have autopsies, one of the cornerstones of medicine, increasingly fallen out of favor? It's an odd state of affairs when you consider how much medical practice owes to the lessons learned in the morgue. After all, many diseases have been discovered solely by the autopsy: Alzheimer's disease, cystic fibrosis, most cancers, and the delayed effects of atomic bomb radiation at Hiroshima and Nagasaki, just to name a few. Without autopsies we would never understand how badly ravaged the body is by incompatible blood transfusions, Legionnaires' disease, toxic shock syndrome, or AIDS. Granted, in an age of high-tech imaging and proliferating tests, autopsies can seem messy, invasive, and totally passé. Patients today are imaged, endoscoped, even magnetized to find out what's wrong with them while they're alive. Yet a recent study showed that some 11 percent of autopsies at an elite university hospital uncovered unsuspected conditions that could have resulted in changes of diagnosis or--if treated- -prolonged survival. (For hospitals in general the rate is 10 to 30 percent.) This is just as it was 80 years ago, pre-microchip.
That point was brought home dramatically last summer when Reggie Lewis collapsed and died on the basketball court. Lewis had the benefit of all the medical expertise and technology that the Celtics could buy in Boston, one of the world's most medically sophisticated cities. Still, his doctors could not agree on a diagnosis to explain his occasional woozy spells. Were they episodes of vasovagal fainting that could easily be treated with drugs? (Vasovagal fainting results when faulty signals from the vagus nerve cause a drop in heart rate and blood pressure during exercise.) Or did Lewis have a serious cardiomyopathy--an enlarged and flabby heart due to disease--that would limit his career and his life? Unfortunately, the latter proved true. But the truth was fully revealed only postmortem. Now, alerted by the results of autopsies following Lewis's death and the death three years earlier of college star Hank Gathers, sports physicians pay serious attention to heart abnormalities in young athletes.
When death defies explanation, human intuition has always prompted us to look inside the body. Thus in the fourteenth century Pope Clement VI ordered the opening up of plague victims' bodies in an attempt to catch sight of the villain that was killing off one-third of Europe. In 1536 Jacques Cartier opened the scurvy-ridden body of one of his sailors to see why his crew were dying one by one on the St. Lawrence. Admittedly, this very intuition, in an unscientific era, gave rise to some bizarre ideas. The New World's first autopsy was done in Santo Domingo in 1533 to determine whether Siamese twins had one soul or two. But later autopsies were done to debunk fanciful ideas about disease causation, such as an imbalance of those four arcane humors that no one ever really understood.
In the nineteenth century, Rudolf Virchow, a Berlin pathologist, set out to supplant the musings of the so-called medical savants by correlating observations at the patient's bedside with scientific observations from dissection in the morgue. Virchow (who, among other things, discovered leukemia) was a great believer in the value of autopsy: he performed two a day. Karl Rokitansky, another great nineteenth-century European pathologist (who figured out pneumonia, emphysema, and liver diseases), completed 30,000 in his lifetime. The morgue, of all places, became the epicenter of learning in the hospital. The dead taught the living. Modern medicine had unquestionably begun.
In fact, not so long ago autopsies were considered so vital to the practice of U.S. medicine that they were required for hospital accreditation. Hospitals had to examine at least 20 to 25 percent of deaths to safeguard quality care, ensure the continuing education of doctors, and in general advance medical knowledge. But by the early 1970s things were starting to change. The Joint Commission on Accreditation of Healthcare Organizations dropped its autopsy requirement because, it claimed, autopsies were being done unselectively, to fulfill quotas. The implication was that the least demanding cases were being autopsied instead of the most meaningful. Of course, the least demanding autopsies were also the cheapest, which may have had something to do with the trend. By that time hospitals were beginning to feel the impact of legislative changes of the mid-1960s, changes pushing them to operate in an increasingly profit- oriented way. Perhaps it was not entirely coincidental that interest grew in performing diagnostic tests and imaging living people who were a potential source of hospital revenue. In contrast, Medicare wasn't keen to reimburse costs for services performed in the morgue, and private insurers soon followed suit. After all, the bureaucrats argued, the deceased is not really a patient and so the autopsy is not really a treatment. No treatment, no payment. If doctors wanted autopsies done, hospitals or patients' families had to foot the $1,000 to $3,000 bill.
Other changes in the last two decades have helped the decline along. In our enthusiastically litigious society, there seem to be no such things as surprises and accidents anymore. Revelations at autopsy that used to be considered purely educational are increasingly perceived as potential evidence for lawsuits. Malpractice insurance can already cost a doctor $2,000 a week; why stir up trouble by peeking inside and perhaps discovering a mistake? (There's a joke told in operating rooms about an unfortunate man who arrives at the Pearly Gates long before his appointed time. An irritated Saint Peter informs him he is about ten years early and then accusingly asks, "Who's your doctor?") Besides, doctors could rationalize, the scanners had done their job, so what else could possibly be necessary?
But there's another reason fewer autopsies are being done--and it has to do with our ambivalence toward death. Until fairly recently, doctors were used to ministering to death. Before World War II, doctors had little more than morphine, quinine, and digitalis in their black bag--palliatives to relieve pain or ease the labored breathing brought on by a failing heart, but they had precious few medicines to actually fix anything. Since the doctor's business was death, an autopsy was neither a shock nor an intrusion. And the doctor often had pressing questions to answer for himself and the patient's family--what kind of illness did this person have, was it contagious, what did it portend for the patient's children?
With the powerful therapies we now have, the doctor's business is life. Doctors have treatments that work, and often work well. Death is the enemy. It's no longer a natural denouement; it's an embarrassment, a treatment failure, or, worse, possibly somebody's fault. A doctor may be reluctant to ask the family for an autopsy because it may sound like an admission of inadequacy. As for families, they can be outright hostile to the idea. If the doctor couldn't save the patient, he or she has no business now trying to find out what happened.
These days autopsies tend to get done only in extraordinary circumstances. Medical examiners and coroners conduct forensic autopsies in cases of foul play. Some states mandate autopsies for unusual deaths in hospital patients (such as my patient Patrick's unexpected death in the operating room). When a mysterious new viral disease broke out in the Southwest last May, autopsies were done immediately to find out what was killing the victims. And no one doubted the need to figure out why Joseph Marable died last December while trying out for the basketball team at William Penn High in Philadelphia. Marable was a cousin of Hank Gathers, so heart problems might have been suspected, but in fact he died of an asthma attack.
This tendency to examine unusual deaths is reflected in higher postmortem rates for the young--most neonates who succumb in intensive care units are autopsied, for instance. Even so, the American Academy of Pediatrics announced last November that not enough autopsies were being done for children under the age of six. If we don't autopsy children, pediatricians argue, we risk missing genetic diseases in families that, if they were forewarned, could avert another tragedy. We may fail to understand the riddle of sudden infant death syndrome or to uncover the murders that masquerade as SIDS deaths.
As for adults, the more they advance in years, the more likely their deaths are to be written off as "expected." The autopsy rate drops to 8 percent in those over 65, and 5 percent in people over 85, with postmortems rarely done at all for nursing-home patients. There's a certain irony to this indifference: we're least curious about precisely the age group that receives the greatest amount of medical care. Moreover, autopsies could provide vital information about how to improve quality of life with advancing age.
What is at stake? Quality control. The autopsy is the oldest, cheapest, and safest form of quality control in medicine. It may look like an intellectual exercise with no immediate utility, but that's an extremely myopic view. The point is that the payoff may come only after months or years of amassing and analyzing the data from many autopsies. We've had 27 years' experience with heart transplants, for instance, but it was only in the last decade or so that there was enough autopsy experience to confirm a severe form of coronary artery disease in heart recipients. This artery- narrowing disease of old age may appear as soon as three months after the transplant, and in children as young as four years. In living patients the condition is diagnosed with angiograms and X-rays. But two years ago a report based on autopsies revealed that angiograms often grossly underestimate the problem and that much better techniques are needed to spot what's turning out to be the major hurdle to long-term survival in heart transplant recipients.
Without autopsies, how can we tell what we might be unleashing on ourselves with our many newfangled treatments? Did that heart attack victim die because a clot-busting drug failed to help her, or because it caused a fatal bleed in the brain? Did the kidney cancer patient succumb to malignant disease or to the toxic effects of interleukin-2? At current rates we may not be autopsying enough of us to determine whether we are subtly poisoning ourselves with Prozac, or Premarin, or Proscar. And what unforeseen things might be happening to the breast implant recipient of the 1980s, or the avant-garde, hormone-manipulated, postmenopausal mother of the 1990s?
Last October, on the shuttle Columbia, a veterinarian performed the first autopsy in space, to study the effects of weightlessness. (A rat was used as a stand-in for higher mammals like us.) There are only a handful of astronauts sailing around in space, doing what is no doubt a hazardous job. But there are a lot more of us grounded here on Earth, trying to pay the rent by extruding plastics or handling industrial chemicals. Thanks to autopsies, we've nailed down silo-filler's disease and cheesewasher's lung. We've figured out asbestosis in insulation workers, and how it leads to cancer. But what new occupational and environmental hazards might we be missing?
No one expected vinyl chloride to cause liver cancer until an alert pathologist put two and two together after autopsying a Kentucky chemical worker. No one expected beryllium to suffocate machinists--and their wives too, when they breathed in the beryllium dust from their husbands' work clothes. No one expected outbreaks of vision loss and paralysis in Japan's Minamata Bay area to be traced to mercury in the local fish.
I didn't expect Patrick to die that day in the operating room, either. Later I published the autopsy photographs of Patrick's fistula in a paper called "Acute airway obstruction due to necrotizing tracheobronchial aspergillosis in immunocompromised patients: A new clinical entity." Another Patrick will come along, because leukemia is still with us. Another fistula, too. Maybe now someone will figure out how to fix the damned thing.