I'd called my cousin one evening, asking if he and his fiancé would like to come to lunch that weekend.
"So can you come?"
"I don't think so," my cousin replied placidly. "We're busy, but even if we weren't, Ann says you're a ditz."
"She says you're not serious enough; you're too enthusiastic."
IT TOOK ME A FULL WEEK TO REALIZE that this conversation was a result of a less-than-perfect prefrontal cortex, the part of the brain that metes out sensitivity and restraint. A few years back, my cousin, age 48, had had a major stroke. While he was in the shower, lathered in Lifebuoy, a blood clot lodged in the middle cerebral artery that delivers blood to the front of one side of his brain. He felt at once weak and odd. For two days, he lay in bed, convinced he was getting the flu. But the progressive uselessness of half of his body made it clear this was no virus. He went to his doctor, then ended up at the Johns Hopkins Bayview Medical Center, where truly heroic measures--no hype--kept him alive. It was close. He went in and out of a coma; his brain swelled; his head assumed the color of a ripe peach. Ultimately, however, rehab helped him regain essentially all of his speech and learn to walk again. There wasn't much they could do, though, about his lapses of social nicety.
My cousin isn't alone. Each year, more than 550,000 people in this country experience a "cerebrovascular accident," or "brain attack," and thousands of the survivors have losses that range from inability to speak or walk to relatively minor lapses like my cousin's. Stroke is our third leading cause of death, after heart attacks and cancer, and the leading cause of disability in adults. But that could change. If Hopkins is any forerunner of what's about to happen, then a rock has just been thrown into the previously quiet pool of stroke treatment in this country.
Had my cousin's "event" been this year, he might have been one of the fortunate few at Hopkins to get an enzyme called alteplase. Better known as tPA, for tissue plasmogen activator, it could have spared his prefrontal cortex, and perhaps he would still be the sensitive, tactful person I once knew.
Hopkins has become one of a number of major medical centers quick to embrace tPA for strokes; that's important but not special. What's interesting, however, is how tPA is making waves in the way the disease is managed here, sort of Hopkins-as-the-country's-microcosm. Still more interesting is how a handful of Hopkins researchers--a recent assemblage of the hottest-shots in the field--are using tPA as a base to leapfrog into a whole new way of viewing stroke, both literally and figuratively. What they're doing now should not only optimize use of tPA, but will also give a hearty push to what many hail as the coming revolution in stroke treatment.
"UNTIL RECENTLY," says assistant professor of neurology Robert Wityk, who heads the neurological end of stroke treatment at Hopkins Hospital, "a tremendous nihilism existed among those who do stroke work, especially interns. That's because there was never much you could do in the first 24 hours that would make much difference in how patients came out in the end.
"You'd put people in a quiet room and tell them not to get excited. You could regulate blood pressure and try to prevent further damage. You could do something--locate the source of a stroke to see if you should try to prevent another one with blood thinners--but it was always secondary." Says neurologist Christopher Earley, Wityk's counterpart at Hopkins Bayview, "The whole feeling has been, We'll support you, see you through this acute phase and then see how nature takes its course." That bottom line, however---holding on tight while nature calls the shots--has been so frustrating, Earley says, that few interns have been taking up stroke as a specialty. One result is that research that could change things has lagged.
But now, there's tPA. The drug digests blood clots by accelerating a preexisting, natural clot-breaking process we share with other mammals. Most people have heard of tPA in connection with heart attacks, the drug's original target. On the market since 1987, tPA, along with the famed beta-blockers, has revolutionized the treatment of heart attack patients. It can restore blood flow to the heart by dissolving the clots that close coronary arteries. Studies show that roughly 35 percent more of those in the hospital with a heart attack survive than would without the drug. Knowing that has changed people's reaction to a heart attack. The idea's gotten out that the more quickly you reach a hospital, with tPA and support treatment, the greater your chances.
Will the same happen for stroke? tPA was approved only last year by the FDA for the most common type of brain attack, called ischemic stroke. That's the sort that occurs when an artery to the brain gets blocked, and tissues essentially suffocate. Roughly 80 percent of all strokes are ischemic, like my cousin's. Symptoms such as dizziness and loss of feeling may come on slowly and without pain (for additional symptoms, see page 21). The remaining cases of stroke are hemorrhagic; they occur when an artery in the brain ruptures, often bringing "the worst headache of your life" followed by a loss of consciousness. Then, the last drug a sufferer would want is a clot-breaker like tPA.
"The thing about tPA is that it's the first crack in the wall of nihilism," says Wityk. Agrees Earley: "tPA's changed things from black to white, from closed to open," with the result that both Hopkins Hospital and Bayview have set up acute stroke teams of neurologists, radiologists, nurses, and special emergency care staff skilled in the drug's use. Wityk says he himself was probably hired--a lily-pure clinician among Hopkins's hybrid clinician/researchers--on a wave of tPA enthusiasm.
Something, however, waxes quite different in the way tPA is used for heart attacks as opposed to strokes. The key, says Wityk quietly, lies in its limited applicability. "We get dozens of stroke admissions every month," he says, "but of that number, only one or two qualify for tPA." At Bayview, they've treated, maybe, 10 patients with tPA in the last year. A far higher percentage of heart attack victims receive the drug. Why?
The reason for the disparity has to do with the biology of the brain and human nature. "The course of a stroke varies considerably from person to person," says Earley in a pithy understatement. "Three hours into a stroke, it's hard to predict who will have a good outcome and who won't. You can see people who are hemiplegic--completely paralyzed on one side--and the next day, they're up and moving around just fine, and we didn't give them anything! Other people with what we thought were really small events can be incapacitated months later, from depression or a variety of reasons.
"A lot depends on a person's individual pattern of brain
circulation," adds Wityk. He says he's seen people with both of
their carotid arteries blocked (these large vessels in the neck
normally carry 80 percent of the blood to the brain) and the only
reason they tootled off to the doctor was, perhaps, a slight
dizziness. As arteries get blocked, Wityk explains, alternate or
"collateral" vessels take on their job. It's the people who have
a clot form suddenly, who didn't have the gradual strangling of
circulation that prompts collaterals to form, the doctors say,
who really suffer when a blood clot or other obstruction hits.
"So, for stroke, unlike heart attacks, we had to see if tPA could
make improvements against a background that's basically hazy,"
In late 1995, the National Institute of Neurological Disorders and Stroke (NINDS) published the results of its major study looking at tPA outcomes. The researchers concluded, says Earley, that "no more people died or ended up as vegetables than those without the drug. That was the real worry--that you'd keep more people alive, but that a lot would be severely disabled, bedridden. And more important was that all categories of disabilities improved. "What we saw was an even shifting of people (from 12 to 15 percent) from each category into the next milder one. You can't say tPA's a cure, but it's certainly good to have," says Earley. "Some are able to talk where they wouldn't have otherwise, or walk. Sometimes the improvement's dramatic."
I think of my cousin and what a difference a single category would make. He lost his ability to think in a certain analytical way and had to give up his job as a speechwriter in Washington. If I'd only known he didn't have the flu; if only they'd had tPA then, I would've dragged him to my car myself in the dead of night in a snowstorm and driven to Hopkins.
tPA for ischemic strokes comes, however, with a catch. The NINDS and earlier studies revealed that taking the drug for ischemic strokes heightens chances of a brain hemorrhage. The tendency was 6 percent in those with the drug versus 0.6 percent in those without. "Bleeds happen naturally after strokes," says Earley, "but the numbers here were significant and caused concern."
The problem stems from changes that take place in the blood vessels in the brain after a clot lodges. Without oxygen, blood vessel cells begin to die, and cell structure starts breaking down. The tiny molecular tethers that anchor one cell to another are among the first to go, making blood vessels both frail and leaky. The cells "serviced" by the vessel also begin to die, in a telltale patch of death--an infarction. What happens when you add tPA? It dissolves the clot and blood once again flows through artery walls. But under returning blood pressure, the now-weakened vessels can give way like wet tissue paper.
The trick, it seems, is for patients to get tPA within about
three hours after suffering a stroke--before broad areas of
vessel cells are significantly weakened and infarcts appear.
Then, surprisingly, even if a few small hemorrhages spout,
patients are relatively unaffected.
This risk-benefit balancing has introduced a time-consciousness as never before in stroke treatment. It's why the acute stroke teams at Hopkins are on call 24 hours a day and why Wityk and Earley can't go to a barbecue without their beepers. "Nerve cells in the brain have almost no reserves and, minute by minute, require blood flow," says Earley. "Lost time is lost brain."
Every piece of literature that comes with Activase, the gene-engineered form of tPA manufactured by Genentech, has a time caveat printed in bold at least twice: treatment should only be initiated within three hours of onset of stroke symptoms. "We're not going to cut hairs over five minutes," says Earley, "but the reality is that this drug carries a certain risk." Go beyond the three-hour window, and those tissuey, weakened vessels may well give way to hemorrhage.
But hitting the three-hour window can be a challenge, because many people have a hard time recognizing and pinning down the onset of symptoms. "I had a patient a few months ago," says Wityk, "who said as far as she knew, her stroke began a half-hour before she came to the ER. Then a neighbor showed up and said, 'No, she had trouble walking two hours ago.' A family member told me, 'She's been having trouble walking all week.' I hadn't started anything; I tend to be very conservative. If time of onset is unclear, I won't treat a patient with tPA."
Unfortunately, the number of people who come into the hospital within three hours is small, just 4 to 6 percent of those with stroke, Earley says. "Then a third more, I suspect, waltz in a day or two later. Our biggest uphill battle has nothing to do with tPA--it has to do with education. It's hard to convince people that when an arm goes limp, they should do something about it immediately. I know that seems hard to believe! But people invariably say the same thing: 'It didn't hurt. I thought I'd go to bed and sleep on it; when I woke up, it was still there."
Continues Earley, "tPA has done more than offer a new way to treat people; it's changed our perspective even on how we should communicate the problem of stroke, on the semantics. That's why the National Stroke Association has started promoting the phrase 'brain attack' instead of 'cardiovascular accident.' The American Heart Association did the same thing years ago with heart attack; they knew 'myocardial infarction' just wouldn't stay in everyone's mind. 'Brain attack' implies a certain urgency, a 911 feeling."
One other way to make tPA an option for more people, Hopkins neurologists told me, is to streamline diagnosis of those who do reach the ER in time. "Our goal," says Wityk, "is to have all the tests done and the decision on whether or not to treat with tPA in 40 minutes."
Many things masquerade as ischemic strokes, says Wityk, and probably a fifth of the time, what looks to be one, isn't. So whenever someone symptomatic reaches the ER, critical care staff order tests, such as an ECG to detect heart attacks, a known source of clots. Blood tests rule out severe anemia, for example, or diabetes. Patients' histories can help eliminate a severe migraine or seizure. Also, reports of earlier fleeting, strokelike spells called transient ischemic attacks (TIAs) signal a greater likelihood that a stroke has occurred (or another TIA). And nothing substitutes for a physical exam: the neurologist's careful prodding, watching, and comparing to evaluate balance, gait, strength, memory, and mind. It helps focus the site of the problem. A weak leg, for example, probably means defects in a motor area of the brain; weakness plus numbness adds an affected sensory area. Knowing symptoms is also a first step in separating ischemic strokes from hemorrhagic.
But sometimes the symptoms of the two strokes overlap, or a
patient may come in unconscious, and because a tPA drip could be
fatal for someone with a hemorrhagic stroke, something even more
definitive needs to be done to rule out brain bleeding. The CT
scan is currently the test of choice. The sequential brain X-ray
quickly picks up a hemorrhage. So CTs have become routine for
stroke: wheeling a patient to radiology and back and doing the
scan takes less than 30 minutes. Earley reports a 13-minute
record. A gurney with wings.
CT scans, however, aren't as good at identifying ischemic stroke. Damage--infarctions--wreaked by even a large ischemic stroke may not show up on a CT scan until hours or days later, and evidence of small strokes, especially those deep in the brain, may never be visible, says Wityk.
CT fails to pick up damage right away because the imaging technique is based on differences in density--between water and brain, say, or bone. With stroke injury, density changes as injured tissues begin to swell with water. But it takes a fair amount of water for CT to detect that density change. Patients can be in a bad way tissue-wise before something shows up. "I would feel much more enthusiastic about giving someone tPA if I knew for sure what was happening right then in the arteries and surrounding tissues," says Wityk.
IN A WARREN OF BASEMENT ROOMS at a Hopkins magnetic resonance imaging (MRI) facility, neuroradiologist Norman Beauchamp Jr. leans against a wall while he draws me a graph on a blackboard. It's late, and the poor man's eyes twitch with fatigue. He's been overseeing the MRI scanner, checking images, working with technicians until early afternoon, then been holed up in a dark film-reading room, confiding scan results into a small tape recorder for the next few hours. Out of curiosity, I ask him something theoretical about MRI, which I know I'll never put into this article. I tell him that, but he answers anyway. He sheds his fatigue like a coat and talks to me because he's driven. Beauchamp is Aladdin; he's been in the cave, seen the treasure, and now is trying to figure out how best to get it to the light of day. Not many have seen what he has, and it's hard to sit on it.
"We need to extend tPA's therapeutic window," Beauchamp declares. An extension from three hours to, say, 12, could almost quadruple the number of potential tPA patients, he says, based on numbers from a large stroke trial conducted by a consortium of European medical centers. "But such an extension means we'd unfailingly have to be able to pick out those 'at risk' [of bleeding]. So what do we really need? A test that makes a person's infarction conspicuous. MRI can do that." MRI, he says, will soon routinely be able to tell if infarctions exist and you can't give tPA, or if they don't and you can.
Beauchamp and a coterie of fellow neuroradiologists and
physicists--several recently lured from other institutions to
Hopkins--have found ways to tweak the conventional MRI scanner so
it does more than the one at your local community hospital.
(Conventional MRI takes very nice, sharp pictures of soft tissues
noninvasively and without radiation.) What they're doing is so
cutting edge they all should carry Band-Aids. They've pushed
the scanner so it can tell, ostensibly within a half-hour, if a
stroke is ischemic or hemorrhagic. It can find a blocked vessel
of all but the smallest size; it can locate infarctions, can tell
if a stroke is new or old, and if someone has a lot of
salvageable brain or if the stroke damage is already too great to
Unlike CT, says Beauchamp, "MRI is sensitive directly to hydrogen in water, so it responds even to small changes pretty darn quickly."
Beauchamp warms to the subject, offers "good MRI" scenarios: "Say Mr. Jones comes in with symptoms. He gets MRI right away and we see a vessel's blocked. We also see no infarction yet. You give him tPA.That's great.
"Say Jones has what looks like stroke. CT shows no bleeding. You give him tPA and everything's fine. But suppose Jones's blood clot had cleared on its own? They do that sometimes, and patients can still have all the symptoms of a stroke until their tissues 'perk up.' Then they're back to normal. But you couldn't tell that with CT. You could with MRI. And you'd save the cost of an expensive drug as well as the cost of personnel to monitor you over the next day or so.
"What if Jones has a seizure, but you don't know that, and he has a kind of paralysis, called Todd's paralysis, afterward? CT might say it's OK to treat. But he wouldn't have had a stroke at all!"
A REAL JONES, though that's not his name, sits in front of me in a wheelchair, waiting for an MRI scan. He's 76, with wispy white hair and a long history of heart disease, high blood pressure, and migraines. He was admitted a few days earlier for a coronary bypass, which went well. Later, however, his right hand went weak and his speech slurred. He veered to the right when he walked. A stroke? His first CT was normal, as was a follow-up at 12 hours. So his attending physician thought MRI might clear things up. The patient wasn't a case for the acute stroke team-- having recently undergone surgery, he was hardly a tPA candidate. But, still, you need to know what's wrong. Small blood vessels deep in a motor area of the brain, called the internal capsule, were suspect.
I like the man. Even though he's short of breath, he talks to me. "I've never had an MRI before," he wheezes. "They tell me it won't hurt, though. I hear it's noisy."
Technician Cindy Maranto confirms this. There is a loud banging but the patient can have earphones with music piped in. "And I can talk to you with a small speaker," she adds. He looks anxious and his breathing is faster as he's wheeled to the vault-thick door of the scanning room. "I have a problem breathing!" he tells Maranto. She offers him something to help him relax, but he refuses. He'll be on oxygen while he's in the machine.
From the doorway, he must walk to the scanner. "No metal wheelchairs in here!" Maranto warns. "We've already had one stick to the machine and it was a pain to pry it away." The strong magnetic field precludes things metallic in the room. I stow my tape recorder.
"I'd like Tommy Dorsey," the patient says. The scanner's a bonanza for masochistic claustrophobics. Only the patient's shins and feet stick out the end. We leave, shutting the heavy door. Once in the control booth nearby, Maranto gives a quick look through the window at the patient, talks to him via the mike, turns on Tommy Dorsey, and types instructions at a keyboard.
This MRI session consists of a series of five scans, each sensitive to a different aspect of brain tissue. At the end, the patient's condition will be summarized from more than 1,000 separate shots. MRI works by placing the patient's brain in a strong magnetic field, so strong that positively charged hydrogen protons in brain atoms align in the direction of the field. Then pulses of radio frequency energy kick the protons out of position momentarily. They "relax" back into alignment and the energy given off during the episode reflects the chemical makeup of tissues.
Tightly bound protons in bone, for example, give off less energy- -translated as a weak signal. Loosely held protons in water give a stronger signal. The brain scans are mostly attuned to water.
Maranto can tap some keys and increase sensitivity to various things that influence "relaxation"--such as having lots of other protons (read: water) nearby. The effect, which seems to me like pulling a rabbit out of a hat, is that different scans are programmed to highlight different tissue characteristics. One scan spots dead tissue, another tells if tissue is mildly injured and water's beginning to leak in or if it's so shot that the cells are already waterlogged; a third senses the rate of local circulation to a tissue. Taken together, the scans form the patient's individual "signature." This signature, and the ability to vary the scans that go into it, makes Hopkins unique. Nobody anywhere else is doing it.
Beauchamp looks at one image, and an infarct shows up as three bright spots on his brain. Another type of scan shows the same spot as dark. "See?" he says, "This pattern tells us it's a new stroke. He's just had this." Another scan shows the surrounding tissue, however, is not "at risk," which means the man's stroke is finished and enough circulation is present that no further damage should appear.
Was this a speedy process? No. The actual scanning time was indeed short, perhaps 20 minutes, but programming the scanner-- often a three-person effort--took time. At one point, the patient moved so much that everyone in the control booth feared a "wigglegram," a fuzzy image; they discussed pulling him out of the scanner and replacing him later, under sedation, but then he calmed. Another time, the scanner repeatedly dumped instructions, forcing reprogramming. "Any complex machinery has gremlins," said one research fellow. All in all the patient, whose results also contribute to an ongoing Hopkins study on MRI and stroke, was in the scanner for more than an hour.
"Boy, that wasn't for everybody," Jones said, back in his wheelchair, but he looked pleased with himself. "Don't I get a cookie? You do when you give blood."
BEAUCHAMP KNOWS there are bugs to work out, not only during the actual scanning but later. Some of the scans aren't real time: they require a physicist to cajole data into a more-useful enhanced image. That can take several hours. But, he says, the physicists are busy putting themselves out of a job; they're working with the scanner manufacturer to enhance images while the patient's still there. "We haven't yet used any of the scans to justify use of tPA. We can't do that until our study's complete and we show that MRI's useful."
Soon, Hopkins will get two more scanners, one of them dedicated to analyzing heart and brain attacks.That should increase the number of patients steered into Beauchamp's study. So far the number stands at 15, though he and co-workers have done more previously, outside the study. "Then we can convince the folks who make the clinical decisions for patients." Within a year or so, Beauchamp predicts, MRI will be eased into the tPA routine. Then, things should take off. They'll share the techniques, he says, with hospitals worldwide--anywhere there's an MRI scanner.
And in five years or so, Earley says, people will be using MRI to decide which is better, tPA or some competing drug. Genentech is already readying a new tPA variant. With a better way to tell who could use drugs like tPA and when, competition for the optimal clot-breaker should become fierce.
And then what?
Ten years ago, we used to view stroke like this: something that happens to an artery and BOOM-you're-dead, or BOOM-you're-somehow-damaged. Now, thanks in a measure to work here, we're finding that stroke is, as one neurologist puts it, "an unfolding upset in whole brain metabolism" that merely begins with a problem in a blood vessel. Animal and human studies at Hopkins show that the process of irrevocable brain damage can take, perhaps, a half a day in some people, under some conditions.
The Big One for stroke won't be in agents like tPA that clear clogged vessels--what some physicians refer to as "plumbing solutions." The future for stroke sufferers, the Hopkins experts agree, lies in neuroprotection: drugs that will shield brain tissue from these unfolding, damaging chemical changes until blood flow can be restored (see story at left). Such drugs are already in early study at Hopkins. Soon, their progress will undoubtedly be monitored by the new techniques coming from our MRI labs. In the future, then, only massive vessel blocks or hemorrhages will do much harm. Smaller events may not cause a stroke at all. And wiser cousins will hold their tongues and come to lunch.
Marjorie Centofanti is a freelance writer living in Severna Park, Maryland.
RETURN TO NOVEMBER 1997 TABLE OF CONTENTS.