OCD and STREP

May 22, 2005
Can You Catch Obsessive-Compulsive Disorder?
By LISA BELKIN
To suffer from obsessive-compulsive disorder, many patients say, is to ''know you are crazy.'' Other forms of psychosis may envelop the sufferers until they inhabit the delusion. Part of the torture of O.C.D. is, as patients describe it, watching as if from the outside as they act out their obsessions -- knowing that they are being irrational, but not being able to stop. They describe thoughts crowding their minds, nattering at them incessantly -- anxious thoughts, sexual thoughts, violent thoughts, sometimes all at the same time. Is the front door locked? Are there germs on my hands? Am I a murderer if I step on an ant? And they describe increasingly elaborate rituals to assuage those thoughts -- checking and rechecking door locks, washing and rewashing hands, walking carefully, slowly and in bizarre patterns to avoid stepping on anything. They feel driven to do things they know make no sense.

There are researchers who believe that some of this disturbing cacophony -- specifically a subset found only in children -- is caused by something familiar and common. They call it Pediatric Autoimmune Neuropsychiatric Disorders Associated With Streptococcal Infection, or, because every disease needs an acronym, Pandas. And they are certain it is brought on by strep throat -- or more specifically, by the antibodies created to fight strep throat.

If they are right, it is a compelling breakthrough, a map of the link between bacteria and at least one subcategory of mental illness. And if bacteria can cause O.C.D., then an antibiotic might mitigate or prevent it -- a Promised Land of a concept to parents who have watched their children change overnight from exuberant, confident and familiar to doubt-ridden, fear-laden strangers.

Child psychiatrists have long known that sometimes O.C.D. in children can be like that, that it can come on fast, out of the blue, like a plague, and then last anywhere from days to months. If the typical graph of O.C.D. symptoms is a sine curve -- with episodes that ramp up slowly, peak gradually, then abate just as slowly -- the graph of rapid-onset O.C.D. is saw-toothed -- flat, then a sudden spike, followed by a relatively sharp drop, then flat again.

The patterns certainly look as if they could be two separate disorders, with similar symptoms but different causes. Across the country, many doctors are convinced of this and are putting young sudden-onset O.C.D. patients on long-term doses of antibiotics. ''If I were to place bets,'' says Judith Rapoport, the child psychiatrist who first brought O.C.D. to public attention with her book ''The Boy Who Couldn't Stop Washing,'' that bet would be on the side of those who believe in Pandas.

But as certain as some researchers are, there are others, just as smart, with just as many impressive publications and titles, who think the theory is wrong or, at best, that it is too early to tell. And this group is warning that the Pandas hypothesis is misguided, perhaps even dangerous. ''Equivocal, controversial, unproven,'' Dr. Stanford Shulman, chief of infectious disease at Children's Memorial Hospital in Chicago, says of the theory.

Pandas stands at a familiar, necessary and utterly frustrating moment in medicine -- in the gap between what doctors think and what they know. Practically every byte of scientific knowledge passes through a moment like this, on its way to being accepted as fact or dismissed as falsehood.

It has always been so, but in recent years several things about the process have changed. Science now does its thinking in public, with each incremental advance readily available online. And those waiting for answers are less patient and more involved. They don't ask their doctors; they bring their own suggestions. They don't want to wait for the results of a two-year double-blind placebo-controlled clinical trial before they act.

Which means that they often find themselves acting before all the facts are in. Can strep bacteria cause obsessive-compulsive disorder? Do these children need penicillin or Prozac? Will we look back on these questions years from now and think, How could we have believed? Or, rather, How could we have doubted?


The most vocal voice in support of Pandas is Susan E. Swedo, a pediatrician and researcher at the National Institute of Mental Health. She was the first to identify the syndrome, and the one who gave it a name. She has been studying the relationship between strep and O.C.D. for her entire career.

She began her work in the 80's, a time of discovery in the world of obsessive-compulsive disorder. Although the disease had long been known, it was not until 20 years ago that researchers began to understand how prevalent it was and not until a decade later that they came to see how often it occurred in children.

In 1989, Rapoport published her best-selling book, taking the illness into the mainstream spotlight. When the television program ''20/20'' ran a segment about her book, it prompted 250,000 calls from worried parents who thought they recognized their children. And a good number of them, Rapoport says, were right. She estimates that more than one million children in the United States suffer from O.C.D. In fact, she argues, the disorder is one that often begins in childhood, which is why doctors should start looking for it then. Half of all adult O.C.D. patients look back and remember having repetitive thoughts and rituals when they were young, which is significantly higher than the percentage of adults with other psychiatric disorders who do.

Rapoport strongly suspected that there was a medical model for at least some percentage of O.C.D. sufferers -- that the symptoms were not a result of emotional trauma (Freud's belief that it is caused by overly strict toilet training had long since fallen out of favor) but rather were caused by a biological trigger. She and her research fellows at the N.I.M.H. spent several years looking into it. Swedo was one of those fellows.

Research had already shown that O.C.D. symptoms appear when there is damage to the basal ganglia, which is a cluster of neurons in the brain that acts as a gatekeeper for movement, thought and emotion. ''So we set out to find every known condition that involved abnormalities of the basal ganglia,'' Swedo remembers.

Huntington's disease was one. Parkinson's was another. Also on the list was Sydenham's chorea -- a movement disorder known to medicine since before the Middle Ages, when it was called Saint Vitus' dance. About 70 percent of patients who develop Sydenham's also develop O.C.D. Sydenham's is caused by rheumatic fever; rheumatic fever is in turn caused by Group A beta-hemolytic streptococcal bacteria. In other words, strep throat.

The biological cascade from strep to Sydenham's starts when the body, thinking it is fighting the infection, begins to fight itself in a process known as molecular mimicry. The protein sheath that coats each invading bacterium cell is remarkably similar to the one that coats the native cells that form a particular part of the body. In this case, the protein code on the strep bacteria is a close match with the code on the cells in the basal ganglia. So the antibodies mistake the basal ganglia for strep and attack. This, of course, will not happen to every child who has strep throat, or even to most children, in the same way that every child who gets strep does not get rheumatic fever. ''It's the wrong germ in the wrong child at the wrong time,'' says Swedo, who suspects that some children are genetically predisposed toward Pandas.

By the mid-90's, Swedo had graduated to her own research laboratory at the National Institute of Mental Health. Back then the status of her research looked like this: O.C.D., she knew, could be caused by damage to the basal ganglia. Sydenham's, too, was a result of such damage. Strep, by all accounts, was the cause of the damage in Sydenham's patients. Sydenham's patients often developed O.C.D. Given all that, the next logical question seemed obvious: Can strep cause O.C.D.?

Swedo turned her attention anew to that subgroup of patients who developed their symptoms seemingly overnight. She and her collaborators hypothesized that this difference in onset could be the key to something important, a separate category, a differentiating wrinkle in a familiar pattern. It might not be the key to decoding the cause of all O.C.D., but it might explain some percentage of cases.

Swedo and her researchers put out a request among those who treat and suffer from O.C.D., looking for subjects -- children whose symptoms had come on suddenly. They received hundreds of calls and then determined that 109 of those children could accurately be described as having had a rapid onset of symptoms. The stories the parents told, while different in their particulars, were remarkably similar at their core. The symptoms came on so quickly that most parents could tell you the exact date that their children's personalities changed. All these children woke up one morning, in the words of one parent, ''full-blown somebody else.''

The exact nature of the obsessions and compulsions differed from child to child (a fact that makes all O.C.D. tricky to diagnose). Some could not stop washing their hands or insisting they needed to use the toilet or checking to make sure that doors were closed and locked. Some developed overwhelming separation anxiety or worried that they would harm someone or do something wrong.

Some had one cluster of these symptoms during their first episode and a different set of symptoms the next time around. Nearly half complained of joint pain, but not always of a sore throat. They were fidgety and moody and obstinate. They had ''bad thoughts,'' some sexual, some violent, some frightening, that they could not get out of their heads.

The children were then tested for evidence that they had recently had strep -- either via throat culture, which would find active infection, or by a blood test that measures antibodies remaining after the actual infection is gone, or, when the episode was too long ago for either test to be effective, researchers asked about a remembered history of strep. In a striking percentage of cases, the search for strep came up positive.


Disagreement is what propels all of science. Proof and disproof seems almost a requirement on the road to consensus. Copernicus's theory that the planets revolve around the sun was not fully accepted until long after his death. Pythagoras and Aristotle each suggested that the world was round, but the idea was not widely accepted for many centuries. Dr. Ignaz Semmelweis was mocked and ostracized for suggesting that by simply washing their hands, doctors could prevent women from dying during childbirth. It would be another quarter-century before Louis Pasteur and Joseph Lister confirmed that destroying germs stops the spread of disease. Much more recently, doctors were exuberant when brain surgery seemed to halt the progression of Parkinson's disease and bone-marrow transplants seemed to beat back breast cancer. But the excitement dimmed as further study found the initial data to be overly optimistic. Perhaps most significant to the discussion of Pandas, strep has been proposed as the cause of a number of conditions over the years, including Kawaski disease, but subsequent studies have repudiated the theories.

''The history of medicine is full of these examples,'' says Dr. Barron Lerner, a medical historian at Columbia University Medical Center, describing fact later shown to be quackery, flights of fancy that turn out to be fact and many ideas that bounce for decades in the shades of gray between the two. ''What looks like it's there sometimes turns out not to be there,'' Lerner says, ''and what everybody is sure of sometimes turns out not to be certain.''

Swedo and her collaborators published several small preliminary studies during the late 90's, and their first major paper claiming that Pandas was a separate syndrome appeared in 1998 in The American Journal of Psychiatry. Called ''Pediatric Autoimmune Neuropsychiatric Disorders Associated With Streptococcal Infections: Clinical Description of the First 50 Cases,'' it is exactly that, a description of children who develop O.C.D. after exposure to Type A strep.

In a way, the description is a tautology -- Pandas is classified as O.C.D. associated with strep, and therefore the only children who qualify for the diagnosis are those who have had recent strep. Swedo took the 109 rapid-onset cases and narrowed those to 50 that met her Pandas criteria, which means that 59 cases were triggered by something other than strep throat. She considers the results important, because at nearly 50 percent, the incidence of strep is far higher than would be expected in the general population and therefore statistically significant. But she agrees that her findings do not explain the cause of all O.C.D., or even all rapid-onset O.C.D.

Despite the details still up in the air, the existence of Pandas was compelling to many doctors. They saw it as inherently logical, and it gave a name to some otherwise mysterious cases that passed through their waiting rooms. ''There is no doubt in my mind,'' says Tamar Chansky, a child psychologist specializing in childhood anxiety disorders and the author of ''Freeing Your Child From Obsessive Compulsive Disorder,'' which devotes a long section to recognizing Pandas.

Not only is it real, says Chansky, who treats several patients who suffer from the disorder, but she has also noticed that each episode is often worse than the one before, creating the possibility that unless these children are treated prophylactically for strep, their O.C.D. episodes could be longer, more intense and more frequent.

''Yes, it is controversial, but I believe it is real,'' agrees Dr. Azra Sehic, a pediatrician in Kingston, Pa. One of the first times Sehic encountered Pandas was when she saw it in one of her patients, Maury Cronauer. Just before Memorial Day in 2003, when she was 6, Maury became ill with strep throat. She was treated with antibiotics and one morning soon after started acting ''odd,'' says her mother, Michelle, who is a nurse. A girl who never worried much about germs, Maury started washing her hands constantly, the most common symptom of O.C.D.

By the next day she was hysterical, saying horrid thoughts were in her head. She wasn't sure she loved her parents. She thought she was going to cheat at school or steal something. She wanted the racing thoughts to go away, and at one point her parents found her curled in a ball in the laundry room, her eyes crammed shut and her hands over her ears.

Sehic mentioned to Maury's parents that the strep might be the cause of her symptoms. She prescribed a longer course of antibiotics, to eliminate any lingering strep bacteria, which might signal the body to create more antibodies.

The O.C.D. went away. A year and a half later, Maury got strep throat again, and the O.C.D. symptoms returned. She is now taking prophylactic penicillin, an approach that is also controversial. ''It is not proven that it will help her, but it is likely that it will, so we are trying,'' Sehic says.

As Pandas was becoming widely known, and as doctors began using antibiotics as a first salvo against obsession, there was ever more research under way. Swedo was a co-author of 30 journal articles between 1998 and 2005. Across the country other lab groups took up the subject as well, and there are dozens more publications in which Swedo played no role.

Some of these merely confirmed the existence of the subgroup Swedo had described. Other studies were designed to take knowledge of Pandas to the next level -- from description to proof. What Swedo had done was identify a group in which two things were true: O.C.D. developed suddenly, and the children had evidence of recent strep. But that does not prove that the strep caused the O.C.D. Nearly all of science is a search for cause and effect -- that A made B happen, that C made B stop.

The bane of all science is coincidence. For example, a notable percentage of children develop their first signs of autism soon after a vaccination, and it is tempting to blame the shot for the symptoms. But autism as a rule tends to show itself during the years when children are also scheduled to receive fairly regular immunizations. So the odds are good that the two events will be temporally linked.

Separating correlation from causation is where every research road becomes bumpy. ''It's been more complicated to follow up on this than we ever thought it was going to be,'' Rapoport says.

There have been studies with results that were remarkably clear-cut -- the plasmapheresis trials, for instance. Plasmapheresis, also known as therapeutic plasma exchange, is essentially a cleansing of the blood, somewhat like dialysis. If strep antibodies were responsible for O.C.D. symptoms in Pandas patients, Swedo theorized, then clearing those antibodies from the bloodstream should prompt improvement.

Because the procedure is so invasive, the only subjects enrolled were those in the worst shape. Of the 29 children in the trial, 10 received plasma exchange, 9 received intravenous immunoglobulin and 10 received a placebo. According to the results published in the journal Lancet in 1999, the children receiving plasma exchange became markedly better, while those receiving placebo treatment did not.

Other studies had results that were somewhat murkier. One tested the theory that you could prevent Pandas by preventing strep. Simply treating strep does not prevent the onset of Pandas since the antibodies have already had a chance to form, which leaves prophylaxis as the most promising form of treatment. That is one way strep was first proved to cause rheumatic fever. When patients who had had rheumatic fever were given daily antibiotics, they did not get strep and they did not get a recurrence of rheumatic fever. Similarly, the hypothesis went, if strep causes Pandas, then preventing patients from getting strep would also prevent a recurrence of an episode of Pandas.

So Swedo conducted a prophylaxis study. Half of a group of Pandas patients was put on daily doses of prophylactic antibiotics, while the other half was given a placebo. After several months, the placebo and antibiotic groups were switched. If prophylaxis works, then patients should have developed more, and more intense, episodes of O.C.D. while they were taking the placebo than while taking the antibiotics.

But the antibiotic chosen for this particular study was a liquid, and unlike the case with pills, which can be counted, it was difficult for parents to keep track of whether a dose had been missed. Even one missed dose would leave a child vulnerable to strep, and some children in the antibiotic group did get sick. A percentage of those developed Pandas.

At the same time, when children in the placebo group became ill, their parents figured out that what they had been dispensing was sugar water and, fearing that the sore throat would lead to a return of Pandas, went and got a prescription for penicillin. Not nearly as many of the control group got strep or Pandas as had been predicted.

''A lot was learned about parental behavior,'' Swedo says, ''but not a lot about Pandas.''


Roger Kurlan, a professor of neurology at the University of Rochester School of Medicine and Dentistry, is not a man who minces words. ''The only thing that's a proven fact about Pandas,'' he says, ''is that children with these symptoms have been observed.'' Everything else, most specifically the role of strep in causing the symptoms, ''is nothing but speculation.''

Kurlan and his collaborator Edward L. Kaplan, an expert in strep at the University of Minnesota Medical School, have become Swedo's most vocal critics. They describe strep and O.C.D. as two things that are ''true, true and unrelated.'' Yes, it is true that some children develop rapid-onset O.C.D. And yes, it is true that a high percentage of those test positive for strep. But that does not mean that the former is caused by the latter.

''In the prior two weeks, 90 percent of these kids might also have eaten pizza,'' Kurlan says. ''Can I make an association that pizza is linked to O.C.D.?''

''If 100 kids fall out of a tree and break their arms and we test them for strep, there's going to be a very high percentage of children who have evidence of recent infection,'' echoes Stanford Shulman of Children's Memorial Hospital in Chicago. ''That doesn't mean strep is the reason they fell out of the tree.''

A more likely explanation for the presence of strep in children with Pandas, these doctors say, is that any infection, in fact any type of stress, can cause spikes in O.C.D. behavior. And they cite as an example children with Tourette's syndrome, who frequently have O.C.D. symptoms that ebb and flow with stress.

Children with neurological disorders ''are sensitive to any number of things,'' Kurlan says. ''If their dog dies. If their parents are fighting. I've seen O.C.D. get worse with a cold, with hay fever, with pneumonia. If there is anything special about strep, I don't think anyone has been able to find it.''

Yes, some children appear to develop symptoms more suddenly than others, he says, but that could be because they have hidden their earlier symptoms from their parents, which O.C.D. patients are known to do. And, yes, he agrees, patients often improve after a positive strep test and a regimen of antibiotics. But because O.C.D. is cyclical, odds are that they would have improved without the test and the medicine anyway. Add to that the fact that some children are strep carriers. They will test positive for the bacteria any time they happen to be cultured, further skewing the cause-and-effect relationship that Swedo is trying to prove.

Kurlan says that he understands why the idea of a bacterial cause for disturbing behavior is attractive to parents. A germ can be cured. A germ is not the parents' fault. ''It's a convenient link,'' he says, ''but it's very difficult to show a connection.''

Assigning blame where none exists can be dangerous, Kurlan says. Part of the harm is that of commission -- giving unnecessary medication. Patients like Maury Cronauer, he says, who take penicillin every day to prevent strep in the first place, are making themselves vulnerable to drug allergies and are promoting antibiotic resistance. And he disagrees with Swedo's view that plasmapheresis can be the answer for the most severely affected patients. The procedure leaves children vulnerable to serious infection, he says, which he considers too high a risk given that the symptoms will arguably run their course over time.

A more insidious form of harm, however, is that of omission. While turning to antibiotics to cure their child's Pandas, parents might be ignoring other treatments that could alleviate what skeptics believe the child actually has -- plain old O.C.D. It may come on slowly or gradually, in the presence of strep or not; whatever the details, a child who cannot stop washing her hands needs to be treated with one of the many drugs and behavioral-therapy regimens that are successful in battling O.C.D., he says.

''If families are distracted by a simple answer and are therefore not tackling the more serious issues, that would be a disservice,'' Kurlan says. ''Worse, that would be bad medicine.''

ndividuals are not statistics, and their stories are not proof. But as I met families and heard their tales, I came to more deeply understand why Swedo is so certain of her theory and Kurlan is so wary of it.

One 10-year-old girl in New Jersey, for instance, illustrates the hazy, sometimes illusory, difference between Pandas and O.C.D. The girl's mother (who asked that her name not be used to protect her daughter's privacy) describes two distinct times, at age 4 and age 8, when her bubbly child became riddled with disturbing thoughts: ''My mouth is full of cavities'' or ''The waiter put poison in my soda.''

The first time, the mother says, her daughter's doctors were uncertain of the cause. But the mother, after doing her own research and suspecting that it might be Pandas, called the N.I.M.H. Someone there confirmed her suspicions. Soon after, the girl took antibiotics, and, her mother says, the symptoms went away in seven months. The second time it took almost a year. The girl has had behavioral therapy but is not taking any medication for O.C.D. because her mother does not think it is necessary. The one precaution the family takes is keeping a supply of rapid strep test kits in the house and using them regularly.

Learning that her daughter had Pandas saved her own sanity, the woman says. ''It was like drowning in the middle of the ocean, and you grab onto something that will help you float.''

And yet. The second of the girl's two episodes, the mother says, was not brought on by strep but by a virus. By Swedo's definition, this would mean that the child did not have Pandas; that her parents think otherwise, Kurlan would argue, shows the danger of a bacterial scapegoat. The mother says that whatever caused the outbreaks -- strep infection, viral infection -- all that matters is that, at the moment, her daughter is fine. But when I ask the girl when she last had her bad thoughts, she tells me, ''Last week.''

Another story of another child, however, shows the damage that can be done if parents start with a psychological rather than a physical assumption. (These parents also didn't want their names used to protect their daughter's privacy.) This little girl was 6 last May, when according to her parents, she changed overnight, becoming clingy and asking the same question over and over and over and over again.

Her mother was pregnant at the time, and a psychiatrist her parents knew suggested that their daughter feared the arrival of her new sibling and was looking for attention. So first her parents reassured her. Then they began to punish her, sending her to her room so she could ''think about her behavior and change it,'' her mother says.

No one in the family, not even the girl's father, himself a doctor, linked any of this behavior to the raging strep infection she had three weeks earlier. They kept punishing her, and she kept insisting that she didn't want to act this way. ''Please stop punishing me for something I can't help,'' the mother recalls her daughter begging.

The parents took her back to the pediatrician's office (they had already been there three times), where they were given a prescription for an antidepressant. Instead of having it filled, they took her to a pediatric psychiatrist, who asked, ''Has she been sick with a sore throat?'' Blood tests showed that her level of strep antibodies was twice as high as it should have been. Two months later, after several weeks of antibiotics and several sessions with Tamar Chansky for cognitive behavioral therapy, the little girl was acting like her old self again.


From where Roger Kurlan and other doubters sit, the situation looks simple. The theory of Pandas, they say, has not been proved. Until the causal link to strep is made, these children simply have O.C.D., and anyone who thinks differently is fooling himself. From where Swedo and her supporters sit, things look equally simple. They agree that cause and effect has not yet been definitively proved. But they are adamant that what has been proved so far is too significant to be ignored and that further research is more than warranted.

In the interim, they argue, logic dictates that any child who develops full-blown O.C.D. seemingly overnight should be given a throat culture or a strep-antibody test before she is sent to a psychiatrist. ''I'm all for empirical stringency,'' Chansky says, ''but in the meantime, there's something so basic that can be done. We're talking about a throat culture and maybe a blood test. What is the downside?''

The downside, Kurlan says, is that science is not supposed to guess. ''We would be testing children as if the results had meaning for their treatment,'' he says, ''and there is insufficient evidence that it does.''

Swedo is still looking for that evidence. Her most recent publication, in the April 2005 issue of Biological Psychiatry, describes a new study of prophylactic antibiotics, one in which administration of the medication was more closely controlled. The results: Those who received the antibiotics saw ''significant decreases'' in strep infections and in ''neuropsychiatric exacerbations'' over the course of a year.

Kurlan, in turn, is conducting research of his own, a nationwide study of 80 patients -- half with a history of O.C.D. that meets the Pandas criteria and half with O.C.D. that does not. For two years, researchers have been logging the rates of strep and the episodes of O.C.D. in each group. If strep causes Pandas, then O.C.D. symptoms should be intensified in the Pandas group relative to their exposure to strep, while in the control group a variety of system-stressing triggers should cause a spike in symptoms.

When the data are compiled and made public later this year, the findings may prove that Swedo is wrong. Or they may instead prove that she is right. Most likely, this latest research will simply lead to more research, as science accumulates its evidence one bit of data at a time.


Lisa Belkin is a contributing writer for the magazine. Her last article was about Thomas Ellenson, a special-needs child in a mainstream school.



Copyright 2005 The New York Times Company