No exit says the door. A swinging gate appears to bar the stairwell, up which wafts a smell of heating plant, with just a whiff of formalin. In the basement at the bottom, Johns Hopkins Hospital performs some 350 autopsies each year.
Stainless steel, white tile, the sound of running water. Masked people in scrubs and aprons. To the unaccustomed eye, these features of the room are a blur, surrounding the all- too-real body of an adolescent boy. He is naked, stretched belly-up on a steel table, eyes closed. His hands are high above his head, as if in surrender, and they jolt up and down, just an inch or so, as the first cut is made. Soon a shallow red U loops from armpit to armpit, curving down beneath the pectoral muscles and across the breastbone.
"This is a very interesting case," says pathologist Grover Hutchins, director of the autopsy service. "We think he died of something called TTP, and mostly we can help these folks. We can treat them. So we're extremely anxious to find out why this youngster died so fast."
It is a Monday in August, 1994, and one of the less-known medical specialties, pathology, is about to swing into action. Pathology is the study of the nature and cause of disease, conducted by M.D.s, and much of its work these days is clinicalÄin the direct service of living patients, such as tending a blood bank or analyzing tissue for surgeons. The field's origin, however, and in some ways its heart, is autopsy: the scientific examination of a particular human body.
The young man on the table, a 14-year-old whom we will call Frank, had been brought to the emergency room the day before, delirious and very pale. He had swollen glands, a temperature of 102.1, and a profuse greenish diarrhea. Three and a half hours later, he died.
The illness had begun on Wednesday, said his family, with a sore throat. On Thursday, he had a cough, with a little blood in the sputum, which developed into a bad cold. However, he didn't seem all that sick till Sunday. A neighbor who is a nurse had suggested he take Tylenol.
On the table, a second incision, down the midline from chestbone to pubic bone, transforms the U into a Y, making two flaps of skin and muscle that are now turned back. It's rather like opening a book. A saw buzzes, and soon the chest plate lifts off and Frank's organs lie bare.
David Brazas, a pathology resident, leans forward to scrutinize each one in turn. As "prosector" (rhymes with "go vector"), Brazas bears the responsibility for this autopsy, from first inspection to final report. At the moment he is looking, simply, for anything unusual, anything at all that will help him determine whether Frank's illness was indeed TTP.
The answer will matter to Frank's family and his doctors, but also perhaps as a matter of public health: Of the other ailments the boy may have had, two can be epidemic.
TTP (thrombocytopenic purpura) is a rare and deadly disorder in which, for some reason, platelets clot along the small blood vessels of various organs. Thus begins a vicious cascade of effects: The clots form rough spots, which rub and injure passing red blood cells. These then self-destruct, so the patient becomes extremely anemic. Complication # l: In trying to undo the destructive clots, the body exhausts its supply of anticoagulants. And #2: At the same time, because so many platelets are tied up in clots, patients run short of them; when Frank was brought to Hopkins, his platelet count was 9,000. (Normal is 200,000.)
Since platelets help blood clot, says Brazas, with a count that low "you can bleed spontaneously. Vessels can just leak blood." Then the many microleaksÄcalled petichiae or, if larger, purpuraÄbruise and kill tiny patches of the organs, especially the pancreas and heart, sometimes the brain and kidneys. Untreated TTP is almost always fatal.
Major signs and symptoms of TTP are fever, hemolytic anemia, fluctuating mental status, a deficit in platelets, and impaired kidney function. Frank fit that picture.
But he also fit the picture for HUS (hemolytic uremic syndrome), a similar disease that is less serious but far more commonÄand that can be epidemic. HUS is often triggered by an infection, notably by the mutated E. coli that a few restaurants in the West had been serving up in tainted beef. Frank's family said he'd eaten no raw beef or barbeque, but families don't necessarily know. Teenagers eat in many places.
Perhaps Frank had TTP and HUS.
Or perhaps he had meningitis, an inflammation of the brain. "I was concerned about meningococcus," says Patricia Charache, a professor of pathology with a specialty in infectious diseases, "because of the apparent swiftness. There are very few things that cause children to die that fast." Meningitis, too, can become epidemic. If that's what Frank had, the other children in his family must be vaccinated right away.
All of which lends a certain urgency to Frank's post-mortem, which continues at about the rate that good carpenters work: they never hurry, and they never stop. At this stage, the diener (a traditional word for medical technician that is still used of autopsy assistants), Raymond Johnson, does the actual cutting.
To become a pathologist takes five years of specialized training, after medical school. Brazas, a dark-haired quick- moving fellow, is a first-year resident, and it later emerges that he has not performed many autopsies yet. This one is a pressure job because the grapevine has been buzzingÄdoctors all over the hospital are aware of the case. ("Apart from trauma," explained an OB-GYN resident, "people at Hopkins do not expect children to come into the emergency room and promptly die. It just doesn't happen.") "So I know I'm going to have a lot of help," says Brazas, dryly.
Frank's heart is startlingly enlarged, and Brazas invites Hutchins to take a look. Yes, the senior man agrees, and it is abnormally pale and mottled, showing tiny dead spots (petichiae) that support the idea of TTP or HUS.
Johnson now cauterizes the surface of each organ with a soldering iron, so the tissue samples he garners for later analysis will be sterile. Five second-year medical students observe in what looks like a slow-motion modern dance: Each time the working team changes position, the students move in silent response, heads and bodies shifting so everyone keeps a clear view. Each person in the room wears a protective mask, gloves, cap, apron, and scrub clothes.
While Brazas is busy placing each tissue sample in its own tiny case (with a label), the students edge toward the table. Encouraged by Johnson, several prod the heart, the kidney, the bowel with gloved forefingers. "They feel so different from the ones in my [preserved] cadaver," marvels a young woman. "The anatomy is much more clear in a fresh body," says Johnson.
Rejoining the group, Brazas points out more petichiae, on the lungs. The smallest class of lesions, they look like bloody pinpricks.
Then Johnson proceeds to detach the internal organs as a block, gently disengaging them from the chest wall with hand and scalpel. Later today, Brazas will separate the organs and examine each separately.
It is hard not to watch Frank's face, so utterly still and indifferent to what's going on. The feet seem poignantly human. They are not perfect, like baby feet. They have seen wear and the calves look quite muscular. Perhaps he was an athlete.
Soon the trunk is an empty shell and Johnson is making an incision in the head, from ear to ear across the crown, to remove the brain. The attending physician who had worked on Frank in the emergency room had called to ask about the brain; judging from the boy's loss of mental function, the ER team suspected cerebral hemorrhage. But the brain looks fine. "This is a beautiful brain," says Christine Brown, a second-year pathology resident who is the "charge" doctor.
It seems unbelievable that this graceful gray-white object, seen a thousand times on page or screen, could be so real, three-dimensional, and precisely like its portraits. The curving lobes are smooth and waxy-looking, with a delicate tracery of red and blue blood vessels that show no sign of bleeding.
Since this is what's called a "complete" autopsy, explains Brazas, "we usually take the eyes." In this case, though, as the family wants an open casket, they have asked for special care with the face. Frank's eyes will be left, "just to make sure."
Johnson is cutting out the spinal column, using a power tool much like a jigsaw. At one of the sinks, Brazas slits the length of bowel. First he takes a sample of its contents to be cultured for E. coli (and more), then rinses and inspects each inch. "See? This dark red patchÄthat's old bleeding. The tissue here looks a little necrotic." These lesions, again, support the diagnosis of TTP. Meningitis seems unlikely, so far, judging from the brain.
"Why does the blood in the chest cavity look so clear?" Brazas is asked. It is almost a garnet color.
"It might be diluted, and because of a lack of red blood cells," he replies. "By the time he died, this guy was running on half a tank." His blood volume would also have been low.
Because of all the little leaks in his blood vessels? "Right."
Well then...that seems incredible: How could a person be bleeding to death, yet not seem very sick? Because up to a certain point, says Brazas, "the body can compensate." Blood vessels constrict and the heart pumps faster, to keep those few red blood cells hustling.
All that does not explain, however, why Frank died so fast. Most patients with TTP are sick for weeks, even those who die. Learning more will have to wait till the slides have been prepared.
Does Brazas think of Frank as a person, or as a body? He says nothing for a minute. The water gurgles, rinsing the broad ribbon of bowel as it moves slowly across his gloved fingers. He stops to look more closely, to jot a note. "I guess I think of him as a person and a body," he says at last. "Mostly, I just think about the next patient who comes in like this. We'll do better because of what we learn today."
Learning-something-useful is the overwhelming reason to perform an autopsy, and has been since Leonardo da Vinci stealthily picked up a knife and took a look inside. One could argue, in fact, that scientific medicine began in that moment, when Aristotelian theory first gave way to direct observation. By the 19th-century medical revolution, autopsy had come out of the closet to be revered as the only reliable source of information.
How did particular diseases progress within the body? Was the diagnosis correct? Was it complete? What had the treatment accomplished, for better and for worse?
For every case, only autopsy could answer those questions. Only autopsy could help physicians improve their diagnoses and treatments. Autopsy was the only way to teach young doctors the inside of the human body, in all its variety.
The legendary Sir William Osler proclaimed his devotion to autopsy; he personally performed more than a thousand, and he often surged into the autopsy room at Hopkins with full retinue. In 1900, a future president of the American Medical Association (Abraham Jacobi) actually found it necessary to remind his academic colleagues that the practice of medicine included "not only diagnosis and autopsy, but the treatment and care of patients."
Today, though autopsy is still answering the same set of vital questions, the field is facing doubt. In 1950, half of all hospital deaths in the United States were autopsied. Now, including all the people who get a coroner's autopsy by reason of suspicious death or unknown cause of death, the Centers for Disease Control and Prevention estimate autopsies at 12 percent of deaths, and falling.
Apart from the legal mandates that govern coroners, no generally accepted protocol applies to autopsies. Physicians ask for them, or not, according to their individual judgment. No autopsy can be performed without legal permission from the next-of-kin.
When asked, U.S. families deny permission about 60 percent of the time. In a few cases, the reasons are religious; Orthodox Judaism, for instance, has an aversion to autopsy. Other families sometimes thinkÄmistakenlyÄthat the autopsy will delay the funeral, cost them something, or make an open casket impossible. ("Hospitals ordinarily absorb the costs," responds Grover Hutchins. "We can work extremely fast, if we have to. And morticians can do wonderful things. I have never seen a case where you could see any sign of the autopsy"Äand here he takes a deep breath, leans forward, and speaks with emphasisÄ"even if you knew one had taken place and knew what to look for.")
Most commonly, though, when a family denies permission, it is for emotional reasons. It is odd but true that most people say they would not object to an autopsy on their own body; what's disturbing is to imagine someone you love on that steel table. There's an illusion that life somehow lingers. "He suffered so much," people often say, when asked for permission to autopsy Grandpa. "And he's dead. What good does it do to hurt him any more?"
Physicians know families may find the idea of autopsy repellent, especially in the first burst of grief, and they don't like to ask permission unless they must. Today, given all the new diagnostic toolsÄMRI and CT scans, laparoscopy, endoscopy, blood chemistry, genetic tests, and moreÄthey less often think they must.
Elliot Fishman, a diagnostic radiologist at Hopkins, speaks for many when he says, "The detail we get of real live patients is so good, that often to get additional information at autopsy is not all that helpful. CT is almost like a living autopsy, the detail is so great." He is sure of that, after conducting research in which lungs were both scanned and autopsied. "Now the diagnosis is made while the patient is living." (Emphasis his.)
Grover Hutchins, a commanding figure behind his slab of a desk, begs to differ. "The CT is absolutely incredible," he rumbles. "It can pick up very subtle changes in the tissue. But can't tell what they are." Hutchins points out that in study after study, over many years, autopsies show that about 10 percent of patients were incorrectly or incompletely diagnosed, to the degree that treating what was missed could have lengthened life, or even cured.
Of course, the missed diagnoses change over time. For instance, leukemia used to be hard to diagnose, but today is almost never missed. Now the main stumbling block is infection: Bacteria, viruses, and fungi are on the upswing, in new, unexpected, and more virulent forms. Not only that, today's sickest patients often have not one but several medical problems, which makes it more likely that something will be missed. One disease can mask another, or make a full workup impossible, or simply get lost in the confusion. So old problems get solved, new ones arise, and the 10 percent holds steady.
"Ten percent?" says Fishman. He just can't believe it. "Well, you might say this: How often do you get autopsies now? Not that often. When do you get them? In select difficult cases. And that's how you might have a high percentage." In other words, sampling error.
Pathologists have wondered about that too. The resulting research indicates that no one, pathologist or clinician, can predict which autopsies will produce a big surprise. Being sure is not the same as being right. "Yeah, they're sure of the diagnosis," says pathologist Barbara Crain. "But they turn out to be wrong." By the same token, even when physicians are not sure, the diagnosis (and resulting treatment) can be correct.
Surprises at autopsy could be fuel for malpractice suits, or so one would think. "I have wondered sometimes," says pathologist John Yardley, "about the degree to which physicians do not seek autopsy because they know an autopsy may work against them." But of course the converse can also be true: Seeking autopsy shows confidence, and autopsy findings can form a basis for defense.
Which way it more often works is unclear, says Hutchins. "At the trial, there'll be some expert on each side who will support the argument made by that side, or they wouldn't be there." What the jury makes of it all, who can tell. "The autopsy may be irrelevant. The decision may be based on the jury's assessment of whether clinical care was adequate."
In general, from the family's point of view, Hutchins thinks it a mistake to dwell on the 10 percent. He points instead to the 90 percent, cases in which autopsies confirm that care was proper. And for that reason, he sees autopsies as therapeutic. "Families find it comforting to know that everything was done, especially with children."
He recalls the case of a 12-year-old boy who had died. The boy had severe mental handicaps all his life. They were thought to be caused by prematurity and oxygen deprivation at birth, leading to hyaline membrane disease, retardation, and seizure disorder. "At autopsy we found that he had a malformation in his brain that must have occurred, oh, four to five months into the pregnancy. I think that must have given the family more peace, knowing it was nothing the family did, nothing the doctors did. The mother doesn't have to think, What if I had been able to carry the baby to term? What if we had gone to a bigger hospital?"
"And even in the small group of cases where all did not go well," he continues, "I do not consider this an obstacle to doing autopsies. Autopsy is part of quality assurance." The Hopkins pathology chair, Fred Sanfilippo, concurs: "If you don't do autopsies, it's hard to know how to do better."
Here is a sampling of helpful information learned in recent years from autopsies:
Patricia Charache, for one, is deeply concerned about a worldwide increase in diseases from viruses, bacteria, and fungiÄsome newly discovered, some truly new, some making a comeback. "Infectious diseases are changing so swiftly," she says, "and we have an incredible amount to learn....It's happened before," she adds. "TB was the scourge of the 18th and 19th centuries. It continued till it ran out of vulnerable people. The Black Plague, it worked the same way."
In explaining the upswing of infectious diseases (her specialty), Charache first notes that doctors now see many patients who wouldn't have survived, years back: the very premature, the very frail old, AIDS patients, cancer patients. They are vulnerable, often immune-suppressed or -impaired. And microbes evolve, as they always have.
What's new is the new group of vulnerable hosts, plus the fact that 20th-century living, worldwide, is highly efficient at spreading disease. Humans routinely cluster in vast cities, travel the globe, and push out to live in new areas, where we can pick up disease from wildlife, such as Lyme disease.
To microorganisms, human technology offers splendidly dank new homes, such as reservoirs, air conditioning ducts, or tank cars full of luscious ice cream mix. And increasingly, some of these infectious agents are drug-resistant.
Charache pulls herself up short. Obviously she could talk about examples all afternoon, but she has another appointment. "Put all that together," she says, "and it's so new that we really must have tissues, and patients. There is so much more to learn!" She pauses. "We're beginning to think that latent viruses may cause certain kinds of cancer, for instance."
The difficulty is that complete autopsies are expensive. Look at Frank's case: Brazas estimates his time as 56 hours, between the initial dissection, paperwork, writing reports, viewing slides, and consultation. Also involved are the Drs. Charache, Perlman, Hutchins, Sanfilippo (as a renal pathologist), as well as a neuropathologist, hematologist, other specialists, other residents, dieners, microscopists, histotechs, admitting, and at least one typist: about 20 people, in all, most of them highly skilled.
Frank's is an exceptional case, both for its intellectual interest and its potential importance. (Brazas says, "I can tell you that most of the docs are not going to be paid, they're just involved out of interest.") The average complete autopsy is less elaborate, having no electron microscopy and little consultation. Even so, Hopkins Hospital spends about $3,600 for each one.
All that for something that is not a "reimbursable medical procedure" (meaning pathologists earn no fees)? It's hard to blame a busy pathologist in a small hospital who prefers to concentrate on clinical pathology. That would include answering questions for surgeons, while the patient is still in surgery, like "Did we get the whole tumor?" John Yardley speaks for that sort of work: "Many times it's like a detective story," he says. "And obviously it's very satisfying to work on a case where you can hope to help some living person."
It's also hard to blame a hospital administration for encouraging pathologists to take that attitude, for unless the family (not the physician) asked for the autopsy, hospitals must absorb the costs.
No wonder autopsy rates are falling.
Grover Hutchins has an idea, one that he and the autopsy committee of the College of American Pathologists hope to see as a national policy: that autopsies should be handled on a regional basis. In this plan, all autopsies would take place at research centers, where they're needed for teaching residents and medical students. He adds that only academic centers are staffed and equipped to perform today's high-tech autopsies, then maintain the necessary records and slides.
He comes back to Frank: "This young man has a disease that is rare. And in order to look at a spectrum, you need a large number of cases to review. We have a fair number of cases here, but presumably with more, we'd do better. So people who have the time, the bent, the duty to do autopsies, would do autopsies. They should not be imposed on somebody who is trying to do a clinical job."
If autopsy rates continue to fall, what will alert physicians to new diseases and the other double-whammies that changing disease will provide? For double-whammies there will be. "There will never be any such thing as 100 percent accuracy," says Ettinger.
Stainless steel, white tile, the sound of running water. That telltale whiff of formalin. But there's no body on the table, only organs; the cutting is mostly done. Frank's autopsy has entered the stage of slides, paperwork, and thoughtÄlots of thought.
A minute ago, the room was silent save for water gurgling so softly it could only be heard if you stopped to listen, and the soft footsteps of Gary March, director of the autopsy facilities and boss of the dieners. Now precisely at the dot of 2 p.m. some dozen people in white coats have arrived, all at the same time like wet snow dropping off a branch.
They stand clustered around the table, while Brazas rattles off Frank's medical history in a hail of facts and numbers: _14-year-old male brought to emergency room_diarrhea, agitated and disoriented times 3_white count of 42,800_creatinine, 1.6_elevated blood pressure, 138/97_treated for suspicion of TTP with packed red blood cells, steroids, antibiotics_Intensive efforts to resuscitate had lasted 45 minutes.
"Everything is consistent with hemolytic uremic syndrome," comments Elizabeth Jones Perlman, the attending pediatric pathologist on this case. She picks up the liver in one gloved hand to inspect it, then the left kidney. HUS and TTP are very similar, but generally in TTP the damage to the kidney is less.
In a 15-minute discussion, no conclusions emerge, nor were they intended to. The "fresh gross conference" is an inquiry, opening as many possibilities as possibleÄeven myocarditis, at the suggestion of a pediatric radiologist who heard about the case in the Doctors Dining Room.
This afternoon, Brazas will write a preliminary report to be shared with the physicians who tried to help Frank in life. Tomorrow he will view an infinitesimal speck of Frank's kidney under the electron microscope, and next week he will move on to light microscopy for the other organ systems. Then to the brain, more consultations, talking with Frank's family, and, finally, writing the final report.
"You may never see a case like this in your whole professional career," pediatric hematologist James Casella tells the pediatric residents. "I hope I never see another."
It is six weeks from the beginning of the case, Frank's autopsy is nearly finished, and the pediatricians are digesting their lesson at what's called an "M&M," a Morbidity and Mortality conference. First the attending physician who had tried to save Frank in the emergency room, Martin Pusic, tells the sad story again: "A 14-year-old male..." But this account has a detail the charts lack. An M&M considers not only what was done, but why.
When Frank entered the ER, he was "disoriented, combative, and vocalizing." His hands and feet were cool, indicating poor perfusion (that is, his body had gone into survival mode, conserving blood flow for the organs). After treatment of his anemia began, his level of consciousness continued to fluctuate, not deteriorating but remaining worrisome. Otherwise, the ER physicians thought they had him stabilized. He responded well to two transfusions of red blood cells. Perfusion recovered, glucose was normal, and breathing was adequate till near the end; at first, in fact, he did not even require oxygen.
After two hours, Frank's blood tests came back, and Casella entered the picture to make a more refined diagnosis: TTP. In the blood smears, red blood cells were severely fragmented, and platelets had been essentially absent . The teamÄwhich by now included Casella, Pusic, the senior resident, an intern, and the nursing staffÄwas about to send him up to Pediatric Intensive Care (PICU) for exchange plasmapheresis (more later) and a CT scan of the head when, suddenly, he became extremely agitated. Then his heart rate plummeted, and soon it was all over. "We didn't really know why he died," concludes Pusic.
Perlman takes the front of the room to review pathology results, with slides. The brain looked cleanÄno meningitis, no thrombi (blocked-up blood vessels).
The heart, however, was a mess. When the body was opened, the heart had been "stunning" (that word from Brazas), because it was so enlarged. Now it's clear why. Almost every blood vessel shows microthrombi (see arrows above). "We think he'd been bleeding for the whole four days," says Perlman.
When a resident asks what actually caused death, Perlman answers without hesitation: "Most likely heart failure."
The surprise lay in the kidneys, which show a few microthrombi, though nothing to compare with the heartÄand a preexisting pathology. It turns out that Frank had had membranous glomerular nephritis (GN), possibly for as long as two months. This is a condition in which deposits of antibodies build up where they don't belong: clogging the glomeruli, kidney structures that filter waste from the blood.
GN can have any number of causes: infection (syphilis, strep); toxins (gold, lead, mercury); or systemic diseases. Whatever, the body fights it by churning out immunoglobulins. Then as the kidney works to purify the blood, the immunoglobulins get pushed out of the blood. "But they're too big," had said Henry Brown, resident on call at electron microscopy. They get stuck. Mostly the cells repair themselves and the condition passes. Frank wasn't so lucky.
GN seldom has symptoms that the patient would notice, Perlman was to explain later. There'd be protein in the urine, but patients have no way to know that. Frank would have retained water, having somewhat puffy ankles and hands, "but that's not always obvious to people," says Perlman. "He may not have felt like himself."
Perlman thinks the GN and the TTP were linked in some way. Conceivably GN makes TTP more likely; one similar case is known. Or possibly both conditions were caused by lupus, an autoimmune disorder. She doubts that GN had anything to do with Frank's death. "The thrombi were in the heart," she says firmly. "He did not die of renal failure." At the same time, impaired kidney function would hardly help.
Casella then comes to the front of the room, for a discussion of Frank's case that is one part teaching TTP diagnosisÄthat's the part we'll skip hereÄand one part thinking out loud: "It might have been worth trying plasma infusions while we were waiting."
Waiting, he means, while Frank and the PICU were made ready for exchange plasmapheresis. (Among other preparations, Frank would need a large-bore catheter.) But at the time, a consultant with wide experience of TTP had doubted that simple infusion would help; he suggested it would be best to go directly to plasma exchange.
Plasma exchange is useless in HUS, Casella tells the residents, but in TTP "it makes a huge difference." With it, the majority of patients survive. And those who do, he explains, "have gained something in their blood that wasn't there when they were ill." No one is sure what. Maybe it's an enzyme, maybe an antibody. Whatever it is, it prevents platelet aggregation, and it comes in healthy plasma. Exchange plasmapheresis, taking out the patient's own blood plasma and replacing it with plasma from the blood bank, is the treatment of choice for TTP.
Frank was about to get that. Indeed, the PICU fellow was standing right there when they all ran out of time. Frank died too fast.
The hope behind every autopsy is that hindsight will help future patients, and certainly no physician who knows this tragic story could ever fail to remember that TTP can have a hidden complication, and that time may be short. For Frank himself, says Casella, "I suspect the story was already written by the time we made the diagnosis."
Elise Hancock is senior editor of the Johns Hopkins Magazine.
Send EMail to Johns Hopkins Magazine
Return to table of contents.