Infectious Madness: The Surprising Science of How We "Catch" Mental Illness


Growing Pains: “Catching” Anorexia, Obsessive-Compulsive Disorder, and Tourette’s

Moods and thoughts are just as biological as digestion and respiration.


In March of 2013, I visited the offices of Susan Swedo, MD, in Building 10 of the sprawling Bethesda, Maryland, National Institutes of Health complex. This city of science houses everything from an army of investigators led by various public-health czars to support personnel to, of course, patients and subjects. I had always imagined this insular research kingdom as being nestled in some remote outpost, but the Washington, DC, Metro delivered me to its door.

After navigating the administration building’s screening machines and humorless security guards requesting identification, I was photographed and permitted onto one of the shuttle buses that endlessly ply the NIH streets and parking lots. It reminded me of a border crossing.

Swedo graciously ushers me into the inner office of her Maryland clinic’s suite. Although we’ve never met, we spoke in 1998 by telephone when I called to discuss her work for a Psychology Today article.

Then I had been struck by the forthright precision of her speech. Saved from severity by the faintest Midwestern twang, her hyperfluency was punctuated by gentle irony and sprinkled with easy laughter. Afterward, I looked for her image online out of curiosity—though I called it research—and was rewarded with a frisson of recognition: Sissy Spacek, I thought. All-American girl.

Sporting a chin-length reddish-blond bob, she looked directly at the camera through glasses, model-pretty but with a wide, guileless smile rather than the melancholic scowl of the fashionable. If her image has changed a bit over the decades, Swedo remains the perennial girl next door—if that girl were a medical prodigy who rose to become chief of the Pediatrics and Developmental Neuroscience Branch at NIMH.

Clad in a knee-length skirt and an impeccable white coat, Swedo, with her smilingly direct manner, appears conventional enough to play a doctor on television. Yet after speaking with me for a quarter hour in her office, she leans forward in her seat, looks directly at me, and crisply pronounces, “I’m not like most researchers.”

As I’m sure my gelid smile conveys, I don’t quite know what she means.

But over the course of the afternoon, a pattern emerges that sheds light on her claim. For one thing, I quickly discover that if you ask Swedo a question about herself, she’ll end up telling you about a patient.

“How did you move from being a busy pediatrician who treated mostly underprivileged adolescents in Chicago mental-health clinics to heading a research wing at NIMH?” I ask, and in response she begins describing her odyssey from ambitious twenty-one-year-old medical student to newly minted MD juggling simultaneous positions in several Illinois adolescent mental-health clinics. But she quickly veers off into recalling the challenges faced by her first patients, including a sixteen-year-old from a well-to-do family whose fashionable mother couldn’t be troubled to come to the hospital after her daughter’s suicide attempt.1

When Swedo remembers this neglected child of thirty years ago, her eyes are bright with empathy, but what’s audible is anger and impatience. “We are failing these children. How can you not do something?”

Instead of being festooned with the obligatory framed diplomas, testimonial letters, and conference posters that I’ve come to expect in doctors’ offices, her walls are home to three mesmerizing thirty-by-forty-five-inch photographs taken by her husband, all images that glorify natural wonders. The sun-dappled florescence of lush, verdant woods enchants in one photo, while a nearly hyper-real shot that looks like Colorado’s Pine Tree Arch radiates copper-hued glory and dominates a second wall. On the facing wall is a large frameless vertical triptych of angelically pretty, deeply saturated redheads—“my daughters.” Atop a file cabinet against an adjacent wall, a seated persimmon-haired tot clad in an immaculate tennis sweater and shorts smiles into the camera, so beautiful and poised that at first I assumed that the image came with the frame. But this is her grandson. When I praise the red-haired beauties, she smiles broadly, then wistfully. “No one believes I was a redhead,” she murmurs.

What really distinguishes Swedo’s office, however, are the pandas. Two of the amply stuffed bears armed with hypodermic needles crown a massive bookcase; a grinning panda statue perches on a table edge exhorting well-child care; panda cartoons lampoon managed care; panda postcards litter a corkboard; and a panda poster announces a European conference on acquired mental illness after infections in adolescence.

It’s fitting that pandas should dominate this otherwise lightly adorned space, because when Swedo is not speaking of her patients, she is speaking of PANDAS, which is, in a sense, another child that needs her protection and the key to helping her patients. To understand what PANDAS is, it’s helpful to first hear the stories of some children who have been affected by it, like Seth.

A child’s sea change

Flushed with irritation, Jane emerged from Seth’s bedroom.2 Until recently, her son had been a quiet ten-year-old whose small rebellions rarely went beyond balking at bedtime or objecting to limits on his Internet use. But this past week, life had become an endless series of complaints and arguments. Tonight he had refused to eat his dinner, complaining that the food “looked funny,” and what if it were poisoned? He could die. First patiently, then angrily, Jane had sought to reassure him, and she finally gave up after a two-hour standoff during which he had resentfully pushed food around his plate without even pretending to eat it.

Now he wouldn’t go to sleep. When she came in to turn off his light an hour after bedtime, this too became a tense debate. Seth had seen a documentary that included video images of rats running amok in the New York City subway and he whimpered that he was afraid of rats attacking him in the night.

Why was he behaving so childishly? she wondered. It was as if he were regressing, she thought, and she hated to admit it, but she was burned out on his nonstop whining and arguing. “There are no rats here. Go to sleep!” she snapped, turning out the light and slamming the door. As she headed down the hallway, she heard the light flick on again, and she stalked back, furious, and flung the door open.

“Rats don’t like the light! They might not attack me if it’s on. Please don’t turn it off,” Seth begged.

No more Discovery Channel for you, she thought wearily, and then, as the room came into focus, her heart sank. Seth was cowering on the far edge of his bed, his eyes continuously scanning the floor. She suddenly realized that behind his cranky recalcitrance and nonstop complaints, he was terrified.

She knelt by his bed. “Sweetheart, there’s nothing to worry about. Would you like to sleep in my room; would you feel safer?” Nodding gratefully, Seth hugged her waist, and soon he was sleeping beside her. She too fell into a deep slumber, but later, she awoke to the empty depression where Seth had lain and to a strange, insistent sound. The clock read 5:15 a.m. As she padded out of the room, she realized that she was hearing running water and that she had been hearing it in her dreams for a long time, maybe all night. As the word drown flashed into her consciousness, she broke into a run.

But Seth stood before the sink, fiercely washing his raw, reddened hands in the running water with a worn shard of soap—it had been a new bar yesterday—a washcloth, and a coarse nailbrush. “Seth, honey, what are you doing?” she asked gently. “Please stop. Please.” He didn’t seem to hear her, and she knew what she had to do. “Come on, honey, you have to get dressed; we’re going to the hospital.”

When they arrived at the emergency department, she noticed that Seth’s lower lip was twitching. He sat down but then leaped from his seat to pick up every piece of paper from the filthy floor with his raw, reddened hands, his head bobbing like a strange overgrown bird’s.

Suddenly he stopped, transfixed, and then ran over and pulled her arm with all his might.

“Mommy, Mommy, they’re coming to kill us! Let’s go! Now! We have to go, now!” Jane tried to calm him, but she too was beginning to panic. Then the nurse called Seth’s name.

Dr. Vogel, the pediatrician, told Jane that Seth had obsessive-compulsive disorder, or OCD. He explained that children with OCD feel great anxiety and can’t stop worrying. Repeating certain behaviors, like turning lights out in a certain order, tapping a certain number of times, or compulsively washing their hands, helps to allay their fears, and it’s very difficult for them to stop these comforting rituals and behaviors.

Jane had heard of OCD, and she was relieved to have a name for Seth’s bewildering behavior and learn that it was treatable. But she was also surprised that his symptoms had come on so quickly, virtually overnight. “Is this normal?” she asked Vogel, who responded that OCD was one of the most common childhood psychological disorders and a variety of things could trigger it, but Seth had probably had these inclinations for a long time; she had just been too close to him to notice the progression.

Jane didn’t believe this. Seth had always been a placid child who smiled easily and who took skinned knees, reprimands, and playground frictions pretty much in stride. But this had changed recently, and the only unusual event in his life she could think of was several bouts with a sore throat. Seth had really suffered, spending whole days on the couch complaining bitterly about being sick, which was unlike him, unable to play, eat, or swallow without pain. As soon as one sore throat ended, another seemed to begin. After he’d been laid up with three, she learned that two of his classmates had recently recovered from strep throat. She belatedly realized that Seth, too, might have had this more serious strep infection, and she decided to take him to the doctor if he suffered another one, but he did not. He just became an anxious complainer.

Jane was racked with guilt to think that the two might be related: Would Seth have OCD now if she had taken him to the doctor for an antibiotic to halt his illness? But Dr. Vogel smiled indulgently at her fears and reassured her that Seth’s problems were psychological. They had nothing whatever to do with a sore throat, strep or otherwise.

In 1994, when Seth was diagnosed, virtually all doctors would have agreed. Psychiatry recognized that children could fall prey to adult syndromes, from schizophrenia-type psychoses to anxiety disorders like OCD. Some psychiatric diseases, such as anorexia, affected children and adolescents disproportionately.

And in many ways, Seth fit the description of a typical child with OCD. The disorder typically strikes children around age ten, some of whom stop eating or fall into the grip of irrepressible tics, ceaselessly flexing their fingers, waving their hands, or jerking their heads arrhythmically while others, like Seth, begin to wash their hands over and over, even after the skin was cracked and bleeding.

Pediatricians ascribed diseases like OCD to psychosocial forces, and there was some evidence of a genetic predisposition; it ran in families. Even Tourette’s syndrome, which plagues people with involuntary movements and utterances and is considered a genetic rather than a psychological disorder, is treated with talk therapy as well as antianxiety medication because it is so often accompanied by other psychological disorders.

Jane left the hospital with a prescription for an antianxiety drug and a suggestion that she take Seth to his pediatrician for follow-up.

Sitting in the pediatrician’s Maryland office, Jane told the doctor that she just couldn’t shake the idea of a connection between Seth’s sore throat and the sea change. She knew her son, and this sudden transformation just felt, well, biological, to her; it felt like something that had happened to him, not something that he was. Or did every parent of an OCD child feel that way? She explained to the doctor that things had escalated very quickly: Seth had become more anxious as he recovered from these sore throats, suddenly developing nameless fears that kept him from eating or sleeping. Her messy son, Jane belatedly realized, had recently acquired a zeal for organization as well, categorizing and boxing his Legos instead of leaving them spilled across the floor of his room, alphabetizing his books on the shelf, and neatly hanging up the clothes that he’d once left strewn in piles. It hadn’t occurred to Jane to regard this newfound neatness as a problem, but now, she thought of it as a symptom.

Jane half expected Seth’s pediatrician to shrug off her worries as the ER doctor had done, but fortunately, Seth’s pediatrician was Susan Swedo, who listened, intrigued, because Seth’s story was unusual in ways that sounded familiar to her.

Swedo had been investigating Sydenham’s chorea, a movement and emotional disorder that often arises after streptococci infections like a sore throat.3 Sydenham’s mostly affects children between ages five and fifteen. It is characterized by rapid, involuntary, spasmodic movements, mostly of the face, feet, and hands. The chorea, from the Greek word for “dance,” refers to these movements, which are often accompanied by muscle weakness as well as emotional and behavioral problems. Seventeenth-century English physician Thomas Sydenham described the condition in the medical literature, but by then the disease already had a long history as St. Vitus’s Dance, named after the patron saint of dancers. Our forebears knew of this illness as a compulsive danse macabre that they regarded as satanic in nature—it was a component of the devil’s rites described during the Salem witch trials.

Sydenham’s is now associated not with satanism but with rheumatic fever, or RF, which causes muscle aches, swollen and painful joints, a rash, and difficulty in concentration and writing. As many as 30 percent of children who contract RF develop Sydenham’s, which, as Swedo knew, was the legacy of an untreated streptococcal infection. Although antibiotic use has rendered RF rare in developed countries like the United States, where it affects only one in every two hundred thousand children, it has recently made a comeback among the nation’s undertreated, such as poor children in inner-city neighborhoods.4

Not only does Sydenham’s follow streptococcal infections, but it is also seasonal, striking most frequently, like schizophrenia, during the winter and early spring. In the U.S., it is most common in the northern states.5

Swedo understood that the diagnosis of Sydenham’s was fraught with opportunity for confusion, beginning with the specific nature of a child’s involuntary movements, which are not simple tics and are not repetitive like the behavior of hyperactive children. They are truly random, small, contained “piano-playing” tics rather than the wilder gesticulations of children with other disorders. Also, Sydenham’s is sometimes confused with cerebral palsy, which by definition is caused by some traumatic events during pregnancy or a baby’s first year of life. By contrast, Sydenham’s has a later onset, after an infection. Two clinical tests enable pediatricians to diagnose Sydenham’s. In one, the doctor asks the child to stick out her tongue and keep it in that position; a Sydenham’s patient will have trouble holding her mouth open and her tongue out for more than a second or two. In the other test, the doctor asks the child to squeeze her hand; a Sydenham’s patient is unable to maintain a grip with steady pressure, so the child’s hand will erratically tighten and relax, creating what doctors call the milking sign. Twice as many girls as boys are diagnosed with Sydenham’s.

Because the rheumatic fever that results in Sydenham’s is a rare complication of a strep infection, like the ones that Seth had weathered, Swedo wondered whether such infections were closely associated with other psychiatric symptoms—like Seth’s OCD. She suspected that a syndrome might connect Group A streptococci, or GAS, infections that cause strep throat to a variety of childhood mental disorders.

She knew that for some children, psychiatric symptoms were the first harbingers of Sydenham’s, as they became unusually restless, aggressive, or hyperemotional even before the physical symptoms of chorea, or dancing tic movements, appeared. Other symptoms were frequent mood changes, episodes of uncontrollable crying, behavioral regression—that is, acting like much younger children—mental confusion, general irritability, difficulty concentrating, and impulsive behavior. In the most common childhood psychiatric syndrome, OCD, intrusive thoughts, images, or impulses recurred, and children seemed powerless to abandon their compulsive behaviors. Often, affected children were seized by fears of harm coming to a family member or of intruders. They sometimes felt compelled to count silently, wash their hands over and over, organize items, or check repeatedly to ensure that a door was locked.

The rheumatic fever is itself a rare complication of a strep infection, and Swedo came to realize that such strep infections were closely associated with a repertoire of symptoms in OCD, tics and Tourette’s syndrome, anorexia, and other psychiatric illnesses. Were the GAS infections really triggering mental disorders? “It was like a mystery or detective novel,” recalls Swedo. “I had to find out.”


Flemish painter Pieter Brueghel the Younger (1564–1636) painted this representation of the dancing mania, also known as choreomania or St. Vitus’s Dance, seizing a pilgrimage of epileptics en route to the church at Molenbeek. Such compulsive dancing was originally ascribed to satanic influence, and later to a collective hysterical disorder, but now seems due to infection of rye and other grains by the fungus Claviceps purpurea. When people ate the tainted bread their symptoms included compulsive dancing. Streptoccocal infections also produced some cases.

Swedo set about finding other children to whom this had happened and came up with a cache of children who had also suddenly become mentally ill, acquiring OCD symptoms or tic disorders shortly after a bout with strep throat or another GAS infection. As word spread that she was investigating the link, dozens of parents from the surrounding communities in the District of Columbia, Virginia, Maryland, and even as far away as Illinois and Michigan made pilgrimages to the NIMH complex to bring their anxious, OCD, anorexic, or tic-plagued children to her.

In 1995 Swedo’s team at the National Institute of Mental Health studied a group of fifty children with OCD, both with and without tics. All their symptoms had been preceded by strep throat or a similar infection. When they tested these children, they found high levels of an antigen6 (a substance that stimulates an immune response against a pathogen) that suggested a genetic susceptibility to rheumatic fever and to Sydenham’s chorea.7 Swedo found that these antigen levels were also high in children with autism.8

In 1998 Swedo published the landmark paper that laid out her theory of pediatric autoimmune neuropsychiatric disorders associated with streptococcal infections,9 or PANDAS, that was afflicting normal children whose behavior exploded into madness within days, and sometimes overnight. First, they were paralyzed by an unfathomable anxiety. Without an apparent cause, this heart-stopping, unfocused fear of the sort that seized Seth was a harbinger of the full force of the psychiatric illness to come.

PANDAS is a syndrome, which means that it encompasses a number of disorders—OCD, Tourette’s, anorexia, and others—that share a cause. Swedo and other scientists estimate that PANDAS accounts for perhaps three of every twenty cases of such diseases. She cautioned researchers that PANDAS was not a default diagnosis and should be considered only in cases where the conventional model of illness did not explain a child’s symptoms.

Such signs and symptoms include a rapid onset. In the case of PANDAS, symptoms arise a few days after infection; Sydenham’s lag six to nine months behind. PANDAS symptoms show a gender disparity, with males more likely to have tics and females more likely to have obsessive-compulsive symptoms. Moreover, PANDAS children regress in ways that other Sydenham’s, OCD, and Tourette’s patients do not. PANDAS children suffer a rapid deterioration of fine motor control, as shown by loss of handwriting and drawing skills, whereas the decline is much more gradual in garden-variety, non-PANDAS-caused Sydenham’s. A picture drawn by a sixteen-year-old PANDAS sufferer looks like the work of a six-year-old. This instant infantilism extends to other behaviors. Out of the blue, humiliated twelve-and thirteen-year-olds resume wetting the bed; some find that they cannot stem the flow of their urine even during the daytime. Preteens begin throwing tantrums, refusing to speak or eat, although the latter is often triggered by an unshakable conviction that their food is tainted or poisoned.

In 1998 I had felt excited as I read Swedo’s PANDAS paper in the American Journal of Psychiatry; she had provided contemporary evidence for a broader role of infection in mental illness than I had imagined.

I already knew, at that time, that some mental disorders could be triggered by infection; the discovery—and rediscovery—that general paresis is one of the final devastations of syphilis is now common knowledge. Links between influenza and schizophrenia and between bornavirus and schizophrenia in Europe had also been discovered numerous times and just as consistently forgotten. I knew that contemporary psychoneuroimmunologists were wondering how much of the dementia and suicidal behavior in their HIV-infected patients could be attributed to the direct action of the virus on the nervous systems of the infected, rather than to the social pressures, despair, and medication side effects that were usually blamed. But in the late 1990s, human research into the microbial roots of madness seemed frustratingly sparse.

I wanted to know more, and Swedo was running studies in children, not petri dishes. She detailed how the body’s response to an infection could go haywire in a young person with an inexperienced immune system, generating a vigorous but inaccurate friendly fire that damaged the body’s own tissues instead of wiping out the invaders. In the model proposed by Swedo, antibodies that linger long enough interfere with functioning of the brain’s basal ganglia.

How, I ask Swedo, did she gain this insight into the infectious nature of these childhood mental disorders? “It was a mom who first made the connection, not us!” she declared gleefully. “I always give her the credit because she, like other parents, know their children better than we ever can; if doctors will just listen to them, they can give us the answers.”

The distinction between patient and subject is an ethically important one, but it is clear that Swedo’s research subjects never cease to be patients in her eyes. The eagerness most researchers radiate when speaking of their theories is audible when Swedo speaks of the children and parents in her care and in her studies.

The fact that she puts her work as a pediatrician first has served her research well; Swedo pays careful attention to her charges, and her hypotheses grow from her experiences with them. Because some children with PANDAS suffered from tics after an infection—including grunts, vocal utterances, and even sometimes curses—Swedo began to think they might have a form of Tourette’s as well.

About two hundred thousand Americans have the most severe form of Tourette’s syndrome, or TS. It is named after Georges Gilles de la Tourette, the French neurologist who first described it in 1885. Usually diagnosed in children between three and nine years old, the neurological disorder is characterized by repetitive, stereotyped, involuntary movements and shouts, eye blinks, grunts, and curses or other vocalizations, even barking, that are collectively called simple tics. Some experience more complex motor tics that include facial grimacing combined with a head twist and a shoulder shrug. Tics are often worse during periods of excitement or anxiety and better during calm, focused activities. For such a rare disorder, TS has received a surfeit of media attention since the 1970s, and as many as one in every one hundred U.S. residents now report milder symptoms of TS such as tics, or involuntary sudden, brief, repetitive movements that involve a limited number of muscle groups.

There are no blood, laboratory, or imaging tests for a TS diagnosis. Instead, children are diagnosed when they have suffered both motor and vocal tics for at least one year. TS is chronic in 10 to 15 percent of affected people but most children who are diagnosed exhibit the worst symptoms in their early teens, and the tics gradually subside as they enter adulthood.10 This provided solid evidence of a connection to PANDAS in Swedo’s eyes, but more studies were necessary to prove the causal relationship and to characterize the mechanism by which GAS caused mental disease. Importantly, Swedo sought to discover whether treatments for GAS infections, such as filtering antibodies from the children’s blood, would reliably alleviate the children’s PANDAS symptoms.

As she recruited more children who had had experiences like Seth’s, the word spread through pediatricians’ offices, support groups, and mommy blogs, the theory resonating with many parents who felt that insidious infections, not genetics or family tensions, were behind their children’s OCD, anorexia, or Tourette’s.

On one such blog, a mother from Flint, Michigan,11 shared the story of her daughter’s sojourn in OCD hell.

One July day, Bertha, her “outgoing, friendly, and spunky” nine-year-old daughter, “woke up transformed” into a toddler, erupting in screams, tantrums, and whining at the slightest frustration. Bertha reverted to bed-wetting and baby talk and seemed tortured by a compulsion to repeatedly touch surfaces and door handles, crying, “Mommy, Mommy, help me, I can’t stop doing this!” Even her handwriting and drawing reverted to that of a three-year-old.

“It was as though she was possessed,” wrote Bertha’s mother. Her daughter was diagnosed with OCD at ten years old, the typical age of onset.

But Bertha’s overnight descent into illness seemed unnatural, and her mother was convinced that something physiologic was afoot. While Bertha took medication and saw a behavioral therapist, her mother read everything she could on the subject, and one day she stumbled upon the PANDAS theory and Swedo’s NIMH studies. She drove her daughter to Maryland, where Bertha joined a study of twenty-seven children with obsessive-compulsive disorder. The treatment involved filtering the offending antibodies from the children’s blood. Swedo used immunomodulatory interventions, including steroids, intravenous immunoglobulins, and plasma exchange, to treat the underlying infections in carefully controlled clinical trials.

Like most of the study’s subjects’, Bertha’s symptoms abated. Almost immediately she was able to resist the compulsions, and as her antibody levels fell, her verbal expression and drawings drew near the age-appropriate level. Within a month, her normal speech and bubbly demeanor resurfaced, and she was restored to her family. Of the eighteen children diagnosed with PANDAS who were treated similarly, all but two improved. Seth, whose mother had first glimpsed the connection between his sore throat and a mental disorder, was among them.

This improvement is important because it helps bridge the gap between correlation and causation. Not only are the high antibodies to the infection associated with the mental-illness symptoms, but as the antibodies are banished, the symptoms abate, which suggests a causal relationship between the madness and the antibodies and, therefore, between the madness and the infection.

A mania for thinness

Another PANDA illness that strikes mostly girls is the queen of childhood psychological disorders.

Ten-year-old Greta sprang out of bed, grateful for the flood of early-morning sunshine through her window. The cold climate and usually gray skies of Rochester, New York, rarely gave way to such glorious weather, even in mid-May. She swallowed tentatively, then smiled; her throat still felt completely better. April had been cold and rainy, and she’d kept getting sore throats, one after another. Until a few days ago, she’d also had sharp stomach pains that came and went without warning. Luckily, they’d faded just as her mom began speaking of taking her to the doctor. At least her appetite hadn’t returned, a good thing, because Greta was seriously dieting. She was determined to be a size 6 by September, when she would turn eleven. Lately, she could think of nothing else.

Donning her plaid shorts, the “lucky pants” she’d taken to wearing every day, she was gratified to see how loose they were. She drank her daily cup of nonfat milk, or half of it, and quickly jumped on her bike to burn it off. She soared through the streets of her small city, going past the university, and eventually reached the village green of a suburb nine miles away. Slowly wheeling her bike down Grand Street to rest, she peered into the windows of the upscale boutiques and dreamed of the day she would be able to wear such clothes. But first she had to lose weight. The thought made her stomach clench for a minute, but then she remembered how loose her shorts were; she was on her way.

Suddenly she longed to weigh herself, so she pedaled home as fast as she could, drank the rest of her milk, iced, and called it lunch. Only then did she allow herself to get on the scale—she saw she’d lost a pound since yesterday. She frowned; not enough. She’d skip the cup of spinach she’d taken to eating for dinner every day and go for another ride in an hour or so. Greta then began her regimen of daily sit-ups.

By September, Greta was a size 4! On her relatively tall five-six frame, this looked thin, but she didn’t think so; compared to the models in magazines, she was still too fat. She did think she looked like a different person, though, and she felt all eyes on her. When a few people, including her favorite teacher, took her aside to warn her not to lose too much weight, she wanted to laugh; she still had far to go. She’d set her sights on becoming a size 2 before Christmas. Every time she thought of her new goal, she rapped twice, softly, on her desk; somehow, this helped reassure her that she would achieve it.

Greta didn’t know it, but she suffered from anorexia nervosa, often referred to as simply anorexia or AN. She was so obsessed with weight control that she ate only very small quantities of certain foods, which resulted in an abnormally low body weight. Like other eating disorders, anorexia is a disease of young people; 95 percent of those who develop eating disorders are between the ages of twelve and twenty-six,12 and anorexia is the third most common chronic illness among adolescents.

The anorectic’s distorted body image makes her see herself as overweight no matter how thin she becomes. Her relentless pursuit of thinness is accompanied by obsessive thoughts about food, calories, and weight. To allay this obsession she frequently engages in self-weighing, compulsive exercise, or even binge eating followed by extreme methods of purging the food, such as vomiting, enemas, or laxatives. Unusual eating behaviors, such as eating only raw green vegetables or only even numbers of grapes, are common.

Although some with anorexia nervosa recover after one treatment session, others go on to develop chronic illness. Their health declines, their menstrual periods stop, their hair and nails become dry and brittle, and their internal body temperature drops, causing them to constantly feel cold. These symptoms are followed by weakness, anemia, muscle wasting, and low blood pressure, and finally the victim suffers heart, brain, and other organ damage, which can be irreversible and result in death.

OCD and anorexia are related not only by their compulsive symptoms but also by the way neurotransmitters malfunction in both.13 What’s more, as in OCD and Tourette’s, some cases of childhood anorexia are triggered or dramatically worsened by GAS infection, placing this autoimmune species of anorexia under the PANDAS umbrella of disorders.14

Anorexia is usually treated by medical monitoring and nutritional counseling with individualized psychotherapy, including cognitive and behavioral approaches that are tailored to the disease. The FDA has also approved antidepressant medications as part of the treatment.

Anorectics can be completely cured,15 yet in 2009 the American Journal of Psychiatry reported that one of every twenty-five people treated for anorexia nervosa dies of the disease,16 and the number of deaths due to anorexia may be even higher, because relatively few sufferers are formally diagnosed.17

As many as 90 percent of U.S. anorectics are girls, and male anorectics are far less likely to seek treatment because of the perception that it is a woman’s disease. Because eating disorders have the highest mortality rate of any mental illness,18 we need more and better treatments for anorexia. Researchers have discovered that anorexia is caused by a complicated mixture in which genetics, psychology, and social factors interact. But GAS infection may be an underacknowledged biological risk factor for anorexia, and addressing it might save people who are not helped by psychotherapies.

Brain imaging and genetic studies may provide clues to how each person develops the disorder. Such knowledge may allow researchers to create specific treatments for preventing and curing infection-driven medical anorexia.19For this to work, doctors would need a means of identifying the PANDAS anorectics.

In 2000 Swedo reported in the Journal of Child and Adolescent Psychopharmacology that she’d tested four children who showed the clinical signs of having PANDAS anorexia and found the same antigens that were elevated in the other PANDAS disorders, indicating the telltale GAS infections.20

Mistaken identity?

PANDAS doesn’t cause every case of OCD, or even most of them. In fact, current research ascribes just one in ten cases of OCD and Tourette’s to PANDAS.21

Or maybe, some critics scoff, no cases at all. Swedo’s article was met by an initial acceptance and a flurry of corroborative studies. Most of the 173 articles focusing on human PANDAS studies I found on PubMed cite researchers convinced that PANDAS is real—convinced enough that they continue to refine, quantify, and augment the diagnosis and mechanism.

But some scientists were skeptical from the first, and soon enough the PANDAS theory was roundly assailed. Some questioned whether the connection was really causal, noting that the frequent sore throats characteristic of PANDAS cases are too common to constitute a distinguishing feature of the syndrome, especially because some children are never diagnosed with strep throat.

Perhaps, the naysayers suggested, Swedo was confusing garden-variety movement disorders—such as Sydenham’s and Tourette’s—with the PANDAS movement disorders. Might they not be the same disease?

This question seems illogical because it ignores Swedo’s admonition, even in her earliest PANDAS writings, that doctors must first rule out the normal varieties of these illnesses before deciding that a child may have PANDAS.

And she scoffs at the suggestion that she has confused common movement and psychiatric disorders with PANDAS when in fact she has painstakingly tracked the differences.

For example, the onset of typical OCD is gradual, even insidious, taking months or years to manifest, while the dramatic OCD symptoms of PANDAS spring up literally overnight. PANDAS is frequently preceded by an incapacitating fear and anxiety that persists through the illness. Parents often report that a child can remain relatively symptom-free at school, only to explode in a fit of anxiety and aggression when he arrives home, immediately consumed by frightening rituals and tics. Typical OCD strikes children around age ten, but PANDAS sufferers can be half that age. And in Tourette’s and Sydenham’s, Swedo says, “the movements are very different. The choreoatheoid movements of Sydenham’s chorea are random, purposeless, snake-like writhing movements or quick muscle jerks and jumps that interrupt a volitional movement,22 while the choreiform movements that characterize PANDAS are small ‘piano-playing’ movements of the fingers that are seen only in certain postures.”23

The theory that GAS causes many of the intractable mental disorders of adolescence—OCD, Tourette’s, anorexia, autism, and others—has suffered a backlash that has forced Swedo and others who treat and research PANDAS to address questions about study design, the suitability of animal models, and the very nature of proof.

“Several excuses are always less convincing than one,” noted Aldous Huxley, and some critics ask how GAS can cause so many different mental disorders. Might it not be more logical to regard these very common bacteria not as causative agents but loiterers at the scene of a crime committed by some more conventional trigger of madness—stress, trauma, or genetics?

The belief that PANDAS needlessly complicates the diagnoses of garden-variety anorexia and Tourette’s invokes Occam’s razor, a scientific concept stating that even far-fetched theories are unnecessary when a simple explanation will do. As thirteenth-century philosopher William of Occam insisted, when theories compete, the simplest is preferred. The medical shorthand is “If you hear hoofbeats in Central Park, don’t think of zebras.”

The versatility shown by PANDAS, however, implies a survival strategy that should not be surprising in pathogens. Microbes, more rapidly driven by evolution than humans, display an impressive adaptive range of effects on their hosts, but their hosts are players in the evolutionary game too, so not all strategies end up improving the microbes’ lot. Strains of the human papillomavirus, or HPV, promote warts and cancer of the cervix, penis, and anus, as well as cancers of the head and neck. Epstein-Barr virus causes mononucleosis, Burkitt’s lymphoma, and Hodgkin’s disease. Helicobacter pylori causes both ulcers and heart attacks when it is not protecting against obesity; Clostridium botulinum brings on rapid death by paralysis, but it also seems to alleviate depression, as noted in chapter 6; Porphyromonas gingivaliscauses heart disease and gingivitis; Chlamydophila pneumoniae triggers heart disease and pneumonia; and Streptococcus mutans causes not only dental decay but heart disease.24

Moreover, PANDAS relies not on a direct infection but on collateral damage; the maladaptive response to infection can trigger different neurological effects depending on which structures of the basal ganglia it harms.

These ganglia are interconnected areas of the forebrain that cooperate to control voluntary movement, learning, some habits, emotions, and thinking. The basal ganglia control motor neurons that generate movement and are also thought to control action selection, or intentional actions. When these ganglia are hampered in moderating the body’s motor systems, the person is racked by involuntary uncoordinated motions.25

Biological… and benign?

As I read the blogs, support-group posts, and online personal videos by parents struggling with their children’s sudden, mysterious symptoms—and even by children who suspected that they might have PANDAS—I was astonished to discover how many wrote of hoping that they would be diagnosed with the syndrome, which would neatly explain their woes and present a course of action. Many, in fact, wrote of feeling dismissed after receiving other diagnoses, such as mass hysteria or malingering. Even worse, some received no diagnosis at all, just a baffled therapeutic silence. An actual diagnosis with a tangible cause—PANDAS—was, for many of these families, devoutly to be wished.

It’s easy to understand why people are eager to identify their illnesses with PANDAS. For many, a biological cause is far easier to sympathize with than a mental-illness diagnosis, which still carries a stigma for not only the sufferer but his or her entire family.

In biblical times, people asked as a matter of course what sins had been committed by the mentally ill person or his parents to explain his symptoms, and as I’ve noted, Sydenham’s chorea was known as a satanic danse macabre. Now we know that in many cases it is the fruit of a Group A beta-hemolytic streptococcus infection.

But a stubborn moral taint lingers. As chapter 2 explains, as late as the 1980s, psychosis was still being blamed on “schizophrenogenic mothers,” and autism was still being ascribed to poor parenting; someonehad to be to blame. By contrast, mental disorders with a biological basis, such as influenza, bornavirus, and GAS infection, seem morally neutral. Children get infections as a matter of course; no one blames the parents or accuses the victims of intellectual or moral weakness.

“When I practiced medicine at Children’s Memorial Hospital in Chicago, I saw parents suffer horribly when they lost their children to leukemia,” Swedo recalled. “When I came to the NIMH, I began to see parents lose their children to OCD and schizophrenia. These parents’ grief is so much more profound. The fact that their children’s illnesses are socially unacceptable makes their pain almost unbearable.”26

The desire for the morally neutral refuge of a biological cause may explain why people with symptoms who identify themselves as having PANDAS but who do not meet the criteria express impatience or frustration with doctors. Parents often insist that doctors miss the diagnosis because they are unaware of PANDAS, not because the ill person does not meet the diagnostic criteria. And when children with symptoms of OCD, anorexia, Tourette’s, or anxiety do not meet the criteria for PANDAS, they and their parents often resist hearing this, perhaps because a PANDAS diagnosis can represent an escape for the whole family from the stigmatized label of mentally ill.

In short, people with PANDAS have begun to claim new identities as victims of a biomedical brain disease as that explanation for their madness gains greater sympathy from the public.

PANDAS offers an alternative identity in the manner described by philosopher Ian Hacking, who has written of how new labels of mental disorders are embraced as redefinitions that allow people to escape the confines of labeling or loosen the shackles of stigma.27

There are precedents. In 1968, for example, around the time that the Diagnostic and Statistical Manual of Mental Disorders deemed homosexuality a mental disease, the creation of the homosexual as a specific kind of person was often traced to a paper by Mary MacIntosh entitled “The Homosexual Role,”28 published in Social Problems, a journal that Hacking says “was much devoted to ‘labeling theory.’” Hacking’s article asserts that social reality is “conditioned, stabilized, or even created by the labels we apply to people, actions, and communities.”29 Similarly, multiple personality disorder, describing a syndrome in which a person is plagued with several identities, was invented around 1875, after which people flocked to become diagnosed with the disorder. Theories abound as to why people sought out the diagnosis, and the motivations probably differ from person to person. But people often find a diagnosis, almost any diagnosis, more comforting than grappling with bewildering mental symptoms that make them fear for their sanity. The role of clinicians is also important because psychiatry has its trends and fads, leading many symptom-ridden patients to receive whatever diagnoses are currently in fashion.

As we’ve seen from the stubborn bias against homosexuals and their former characterization as “mentally diseased,” a redefinition does not always banish stigma or mistreatment. Recasting mental illness as a form of infection can also backfire, especially in the case of especially dreaded or sexually transmitted diseases. When syphilis was demonstrated to cause paresis, judgment rained down on paretics, and the STD stigma may even explain the willingness of researchers of the time to engage in malaria therapy, infecting paresis sufferers with a chronic and debilitating disease. Tuberculosis, by contrast, was morally rehabilitated after it was discovered to be infectious. Before that, TB was referred to as consumption, which, says Hacking, “was not only a sickness but also a moral failing, caused by defects of character. That is an important nineteenth-century social fact about TB. We discovered in due course, however, that the disease is transmitted by bacilli that divide very slowly and that we can kill.” The idea of the consumptive as “a particular kind of person” with certain character traits, rather than simply a person suffering from illness, was, Hacking says, “an artifact of the nineteenth century.”30

But several studies have determined that, even when the infections themselves carry no social taint, “emphasizing the biological aspects of mental illness does not reduce stigma and discrimination among the general public.”31Although ascribing mental disorders to frankly physical causes like brain-chemistry imbalance or infection is an approach that discourages the assigning of blame for illness, this also promotes the belief that the illness, being biological, is intractable. As an article in Schizophrenia Bulletin explains, such biological underpinnings foster the belief that the sick person is impervious to treatment and therefore may be dangerous.32

Medical professionals harbor their own brand of prejudice, and their discrimination against mental disease is a product of the very mind-body dualism that has prevented the profession from recognizing the infectious contributions to mental illnesses in the first place. Toronto mental-health commissioners Thomas Unger and Stephanie Knaack explain why:

When presented with a symptom or set of symptoms, for example, physicians will start by using the fundamental schematic categorisation of “Is it functional or is it organic?” If categorised as organic (i.e. in the body) it is assumed to be real, legitimate and material. From the physician’s point of view, this means it is something that can be observed, studied, treated and corrected. Arguably, this reduces stigma and discrimination. However, if categorised as functional (i.e. a problem of the mind, with no physiological correlates), the physician will consider it less real and the patient may be more likely to be stigmatised and discriminated against.33

For those with mental diseases caused by infection, better treatment may be in the offing, because the infection and its damage present a discrete medical target, unlike the murky but widespread theories of brain-chemistry imbalance that have not always held up well to researcher scrutiny.

A contested diagnosis

Sydenham’s is known to be caused by childhood infection with Group A beta-hemolytic streptococcus, and it affects 30 percent of children who suffer acute rheumatic fever. But unlike Sydenham’s, causation in PANDAS remains contested.

It’s not that anyone questions whether the affected children were infected with GAS; Swedo and others have rigorously documented the presence of antibodies to the bacteria, although less sensitive tests by doctors who are not PANDAS specialists may fail to detect them. Skeptics, however, attack every other tenet of the theory.

They ask, “Is PANDAS really distinct from garden-variety Tourette’s and OCD?” and point out that the discovery of PANDAS was made from case-finding among sick children who fit the general profile rather than from forward-looking studies of large numbers of children gathered at random, and some wonder whether this method creates an illusion of causality.

The ubiquity of GAS also works against the PANDAS theory in the eyes of some. One can see for oneself that strep throat and related infections are everywhere but Tourette’s and OCD are not. Does this mean that GAS infection is a cofactor, insufficient to cause disease on its own but exacerbating the damage from genetics, stress, trauma, or even poor parenting? Or is GAS just a near-ubiquitous innocent bystander? To those convinced of PANDAS’s disease status, it is clear that not everyone with GAS becomes mentally ill because many factors affect vulnerability. Genetics, immunological vigor, general state of health, and perhaps environmental insults may all determine who develops PANDAS and who is able to avoid antibody damage to the basal ganglia.

Correlation and cause are two different things; for strep infections and PANDAS, the former has been demonstrated, but the latter is proving far more elusive. Correlation is the Achilles’ heel of research into disease causation. Heart disease, for example, is strongly associated with stress. But do people suffer from heart disease because they are stressed, or are they stressed because they suffer from heart disease? Or is there some more complicated explanation for why stress and heart disease are frequent traveling partners?

Here is one example of the difficulty of teasing cause from correlation: Studies done before 1992 revealed that hypochondriacs were much less likely than their peers to develop atherosclerosis, or hardening of the arteries. Leaping to a narrow causative explanation, you may conclude that excessive worrying about your health is good for your heart. But in 1999, a Journal of the American Medical Association study showed that tetracycline use is associated with a lower incidence of heart attacks.34 People open to broader analyses may reason that because hypochondriacs are more likely than others to take antibiotics to ward off infections, and because the atherosclerosis associated with heart disease is caused by various bacteria, it may be the antibiotic, not the worrying, that is protective. And they’re right. Hypochondriacs are less likely to develop atherosclerosis because they are more likely to take the antibiotics that knock out heart pathogens such as Porphyromonas gingivalis and Chlamydophila pneumoniae.35

A lack of consensus

Swedo seems dismayed by the vigor with which critics like Harvey Singer, the Haller Professor of Pediatric Neurology and director of Child Neurology at Johns Hopkins, chip away at PANDAS. How do you prove an infection causes an illness?

“We need consensus,” she explains as we sit together in her office, “and we had a meeting at NIH in July 2010 to reach agreement on the clinical picture of the acute-onset cases. Unfortunately,” she snaps, tapping a paper impatiently, “three of the forty-one attendees elected to publish a ‘minority report’ entitled ‘Moving from PANDAS to CANS’ (childhood, rather than pediatric, acute-onset neuropsychiatric syndrome) and the debates intensified, rather than being resolved by the meeting.”

This is yet another paper criticizing the evidence behind the PANDAS model and proposing an utterly different model and acronym, CANS, that removes any reference to an infectious agent.36 Swedo has circled all the paper’s points that, in her opinion, do not accurately reflect the evidence; coarse black circling fills the pages.

But compelling evidence must accompany consensus, and attaining traditional proof is hampered by research constraints. Ethically, you can’t infect people with GAS and wait for symptoms to develop. You can’t remove people from every other known risk factor for insanity—genetics, stress, inflammation, brain damage—to see whether GAS alone triggers it. The usual methods of proving medical theories seem infeasible.37

Further questioning PANDAS’s disease status, some pointedly note that neither PANDAS nor its newest incarnation, the very similar pediatric acute-onset neuropsychiatric syndrome, or PANS,38 is a disease entry in either the International Statistical Classification of Diseases and Related Health Problems (ICD) or the psychiatrist’s bible, the Diagnostic and Statistical Manual of Mental Disorders.39 The current edition, the DSM-5, defines a mental disorder only as “a clinically significant behavioral or psychological syndrome or pattern that occurs in an individual [which] is associated with present distress… or disability… or with a significant increased risk of suffering.” In the case of PANDAS, the DSM-5 refused to commit itself, noting that it remains a controversial diagnosis, citing both the PANS (Swedo’s) and CANS (Singer’s) revisions to the clinical picture.

Straddling the fence

But this failure to endorse PANDAS/PANS means little, because the wheels of psychiatric epidemiology turn glacially slowly. The DSM in particular is quite malleable and tends to reflect the sociopolitical climate as much as the medical one, so a disorder’s inclusion or banishment from the manual follows closely on the heels of its political fortunes.

Witness the removal of homosexuality as a mental illness, which took place only after gay-rights activists demonstrated at the 1970 American Psychiatric Association meeting in San Francisco. By 1973, homosexuality was removed and replaced by sexual orientation disturbance.

Today a political furor swirls about the DSM-5’s consolidation of autism, Asperger’s syndrome, and similar conditions within the overarching category of autism spectrum disorder, or ASD, which affects one in eighty-eight U.S. children. “The change signals how symptoms of these disorders represent a continuum from mild to severe, rather than being distinct disorders,”40 notes APA literature, but its new definitions also reduce the number of people who are diagnosed with autism-like ailments such as Asperger’s syndrome by nearly one-third, according to a Columbia University study in the Journal of Autism and Developmental Disorders.41 There’s much dissension from those who fear this “may leave thousands of developmentally delayed children each year without the ASD diagnosis they need to qualify for social services, medical benefits and educational support,” as the Columbia researchers predict, and although the DSM editors wrote that this consideration did not figure in their decisions, they had to be aware of it. The change is also problematic for adults who identify with their Asperger’s diagnosis. They stand to lose not only material resources and employment protections but the psychological and social benefits of a diagnosis that runs social interference for them. An awkward or standoffish person with a diagnosis of Asperger’s is likely to be met with more respect and understanding than someone who exhibits the same behavior but is unprotected by a diagnosis; people may think he is simply unfriendly or judgmental.

The DSM has its medical flaws as well, according to critics who decry its invalid and inconsistent “cookbook” symptomatology, its arbitrary dividing lines between normalcy and pathology, and its cultural bias. African Americans, for example, are much more likely to earn the label of psychotic for the same behavior that elicits a milder neurotic label in whites,42 but this had been the case long before the DSM-5hit the shelves.43 The manual has also been accused of medicalizing human experiences that are not necessarily pathological. For example, under the Diagnostic and Statistical Manual of Mental Disorders IV, clinicians were advised not to diagnose major depression in people who had suffered the death of a loved one within the previous two months. But the DSM-5 (the roman numerals were dropped after the fourth edition) has abandoned this bereavement exclusion, and many people take issue with treating the grief of the bereaved as a pathological condition. Some also criticize the book for reflecting the opinions of a closed circle of influential psychiatrists, and many are uncomfortable with the fact that it is used much more often in the United States than abroad.

There is little attention paid to addressing how profoundly culture affects the way mental disease appears within the DSM’s pages, and, as if all this weren’t enough to hobble it as a tool, the text is also beholden to corporate and other financial interests, including pharmaceutical companies and the APA, which has raked in $100 million from its sales and licensing.44

However, the chief flaw of the DSM, from the perspective of this discussion of mental diseases caused by infection, is a key error of omission: despite the typically sluggish fourteen years of deliberation and voluminous documentation in which the authors of the newest revision indulged, the manual has maintained a perfect silence on what Ferris Jabr’s 2013 Scientific American essay called “the biological underpinnings of mental disorders.”45

Thus, although PANDAS is not a valid DSM-5 disease category,46 this signifies little. So the question remains: How do we determine whether the evidence that correlates infection with PANDAS rises to proof of causation? This issue applies not only to PANDAS but to all the possible links between infection and mental states that this book discusses.

Interrogating proof

The discovery of syphilis spirochetes in the brains of paretics struck the blinders from the eyes of the nineteenth-century doctors who treated those patients. The physicians produced detailed charts to document how often the disease and the paresis were associated and whether the syphilis seemed to precede the madness. What they emerged with was a correlation between the two disorders.

Yet approximately a century intervened between this correlation and the routine curing of paresis with antibiotics, which proved it was infectious. The change in treatment was the definitive indication that medicine had finally accepted the association between spirochetes and madness as a proven fact, but the slowness to translate that realization into practice is a drearily familiar scenario. Typically a very long lag time elapses between Eureka! and the acceptance of a new cause and treatment.

But how do we know when we have reached the eureka stage?

Despite the public’s discomfort with infection as a cause of mental illness, it is a simple extension of the accepted nineteenth-century germ theory that posits infectious causes for a wide array of physical diseases. Should we not have to apply the standards set up by the architects of germ theory to establish proof? Can we?

In a word, no.

Basic standards for the proof of infectious-disease causation were laid down in 1883 by the German bacteriologists Robert Koch and Friedrich Loeffler, whose criteria came to be known as Koch’s postulates. According to these, a suspected pathogen can be said to cause a disease only when (1) the germ is consistently associated with the disease; (2) it can be isolated from the sick organism and cultured; and (3) inoculating an organism with the pathogen should cause symptoms of the disease to appear. In 1905, another requirement was appended: The pathogen must be isolated again from the experimentally infected host.47 However, at least one prominent researcher claims that only the first two postulates are Koch’s and only they matter.48

Critics have invoked Koch’s postulates to question the etiology of PANDAS and other madness caused by infection. But even in the nineteenth century, scientists realized the criteria’s limitations; some microbes that caused disease failed to fulfill the postulates. Mary Mallon, dubbed Typhoid Mary, comes to mind; she was an asymptomatic carrier—that is, she carried the typhoid pathogen without suffering signs or symptoms of the disease herself—and similar carriers are found in cholera. This carrier scenario is so common in infectious disease, especially viral diseases such as polio, herpes simplex, and hepatitis C, as to invalidate Koch’s first postulate. Polio causes paralysis in only a small number of infected people, yet we know polio is caused by the virus because the vaccine against poliomyelitis successfully prevents it.

Koch’s second postulate rests on equally sandy ground, because some disease-causing microorganisms, such as prions, infectious proteins that many think responsible for Creutzfeldt-Jakob disease, cannot be grown in culture.

Koch himself knew that the third postulate was flawed; ever since the establishment of germ theory, it’s been known that not all organisms exposed to a pathogen will fall ill. Immunological resistance, genetics, and variations of general health happen to them all. Noninfection may be due to such factors as having acquired immunity from previous exposures or vaccination.

Then, too, genetic immunity protects some; having an allele for Tay-Sachs confers some degree of immunity to tuberculosis, for example, and having sickle-cell trait does the same for some strains of malaria.49 Perhaps this is why the third postulate specifies that the pathogen should cause symptoms, rather than that it must.

In short, the evidence tells us that Koch’s postulates are sufficient—but not necessary—to establish causation.

Today, quite a few infectious agents are accepted as the cause of disease even though they do not fulfill Koch’s postulates.50 “We have to be ready to think of all sorts of new, clever ways to identify pathogens,” says evolutionary biologist Paul Ewald, author of The Evolution of Infectious Disease and Plague Time. “We will have to abandon Koch’s postulates in some cases.”51

Arrowsmith in the twenty-first century

What, I wonder, does epidemiologist Ian Lipkin think? Like Dr. Martin Arrowsmith, the intrepid protagonist of Sinclair Lewis’s 1925 novel Arrowsmith, Dr. Ian Lipkin is a peerless microbe hunter. He has identified hundreds of viruses, tracked pathogens from the Bronx to Beijing to Burundi, fingered West Nile virus as the cause of a mysterious 1999 encephalitis epidemic in New York City,52 and advised the makers of the film Contagion. He knows a thing or two about linking infection to disease.

He also directs Columbia University’s Center for Infection and Immunity, whose mission statement reads,

We are committed to assembling a “global immune system” that will enable scientists and clinicians to manage potential threats before they can affect the health of communities worldwide.

The first step toward achieving this goal is being able to quickly identify the pathogens that cause disease.

I arrive at Columbia’s Mailman School of Public Health, just blocks from my former Harlem home, to ask Lipkin how his institute definitively fingers pathogens.

“It’s nice to be back in Harlem,” I volunteer to the affable security guard at the glassed-in front desk as I proffer my Columbia ID. I am quickly corrected. Peering closely at my card, he says, smiling, “This is not Harlem; it’s Hamilton Heights.” I return his smile, but nomenclature doesn’t change the fact that I could throw rocks from here and hit no one but Harlemites; the university’s medical enclave is nestled within it.

When I reach Lipkin’s institute, however, I better understand the guard’s distinction. The heart of the globe’s immune system generates an ambiance that is light-years from the colorful urbanity surrounding it.

Everything looks gray, beige, or black. When the elevator reaches the seventeenth floor, the doors slide back smoothly, in silence, to reveal a capacious modern beehive housing workers in banks of identical desks; this is the anteroom of the center. As I step forward, their heads swivel briefly in response to this stranger among them; immediately, a neatly dressed young man walks over and politely questions me in hushed tones before ushering me past the first flank of podlike gray workspaces. There are no cubicle walls, and each desk is graced with a black phone, a charcoal-gray monitor, and a seated employee. I’m invited into a black ergonomic chair, where I sit alone and unregarded within the glassed-in conference room. Before me, the workers busily attend to their tasks in eerie silence; the conference room must be soundproofed.

A fifty-inch black monitor and a bone-colored spherical microphone depend from the ceiling, and a grayish keyboard and a few pages from an autism-study protocol are the only items on the table, which easily seats ten. It is made from the delicately varnished cross-section of a mammoth tree—rings, knots, and all—and it’s the only object in sight that is visibly organic and that looks as if it could have been designed in the previous century. After twelve minutes, a young woman clad in dark gray and white opens the door, admitting a subdued hum of background activity. She utters my name, then wordlessly guides me down a corridor toward Ian Lipkin’s inner office. It’s adjacent to its own conference room, into which she motions me. I take a seat, and as she glides out of sight, I hear her announce to Lipkin: “Your ten o’clock.”

I’ve been allotted twenty minutes. After a few of them pass, Lipkin enters, a slim man of average height who’s fiftyish but looks a decade younger. He is wearing a pointed-collar cotton shirt of vaguely institutional green, neat, belted brown pants, rectangular bronze-rimmed glasses, and a moue of impatience. As we make eye contact, this changes to a small but pleasant smile that vanishes when he sees my extended hand. He demurs. “I don’t shake hands, especially in winter.”

“I understand,” I say, because it’s only logical behavior for a microbe hunter. He pauses and adds, awkwardly, “It’s nothing personal,” then slumps into his seat and stares at me with a dour expression.

There’s little time, so I get right to it. “How do you prove causality in cases where you cannot apply Koch’s postulates?”

Lipkin minces no words. “Koch’s postulates are obsolete.”

I point out that journal articles often invoke them as criteria.

“Well, they sound good, don’t they?” he counters, raising an eyebrow and grinning. “But that’s not the way you prove causation. Proof falls into three categories—the possible, the probable, and the definitive.”

Lipkin deals in the definitive. “We’ve discovered more than five hundred viruses since I arrived at Columbia in 2002,” including West Nile virus, which he identified as the cause of a North American encephalitis outbreak in 1999. Lipkin was the first to use high-throughput sequencing for pathogen discovery and he uses MassTag PCR and GreeneChip technology, two multiplex assays53 that have identified and characterized his hundreds of viruses.

Moreover, Lipkin’s work isn’t limited to viruses. His empire of pathogen hunters investigates protozoa and fungi as well. He’s indicted inflammatory neuropathy in some ailments and shown that it can be treated with plasmapheresis. He’s also shown that an infant’s exposure to viral infections early in life changes the way his or her neurotransmitters function, suggesting a role for infections in schizophrenia and possibly autism.54

As he summarizes his team’s protean achievements, I am pleasantly surprised by the witty and genial conversationalist who emerges. He even takes care to speak in accessible language; for example, he veers midphrase from “in vitro results” to “laboratory results” for clarity’s sake. This isn’t necessary, but it’s thoughtful. I tell myself that his initial abruptness probably came of his being badly pressed for time; it must be hard to tear himself away from his well-oiled machinery of epidemiology, even for twenty minutes.

“I think that there are many examples where you cannot fulfill Koch’s postulates,” Lipkin continues. In those cases, what determines when something qualifies as proof? “There are the original Bradford Hill criteria, and other criteria people have talked about for years,” Lipkin says, referring to Austin Bradford Hill, the English epidemiologist who suggested nine criteria for proof. They include strength of association (the larger the association, the greater the chance it is causal), consistency of association, and biological gradient, the idea that a greater number of exposures leads to a higher incidence of the effect. He also introduced the criterion of plausibility, stipulating that a believable mechanism must be proposed to tie cause to effect.

Plausibility. Believable. These words give me pause. Couldn’t such a criterion exclude a true cause just because it isn’t within one’s habits of thought—that is, within the current paradigm that explains such illnesses? Dismissing a theory as unbelievable—scientists who were loath to admit that paresis was caused by the bacteria of syphilis did just this, as did physicians who refused to accept that pellagra was caused by a nutritional deficiency, not by an infectious disorder that was limited to blacks. Everyone “knew” that paresis was madness born of psychological causes and that pellagra was a racially bound disease of filth, so even in the face of definitive evidence to the contrary, some clung to these theories decades after they had been disproven. In other words, isn’t the criterion of plausibility, in some cases, an invitation to cling to dogma?

Another Bradford Hill criterion seems even more problematic: temporality, which suggests that the effect must follow the putative cause closely in time. The infectious transmission of mental illness often transpires over a long interval that can obscure the cause-and-effect relationship. Paresis appears as long as three decades after an infection by T. pallidum, and influenza can trigger schizophrenia after twenty years. Rabies, however, can trigger madness within weeks or even days of infection, clarifying its connection to a precipitating animal bite. This makes the madness of rabies much easier to associate with infection, yet rabies, too, goes undiagnosed when months have intervened between an exposure and the onset of symptoms.

Koch’s theory is also riddled with limitations. “Koch’s postulates require that you grow something, put it in an animal model, and replicate disease,” says Lipkin. “But there are agents that you can’t cultivate in laboratories. You have infectious agents for which there is no animal model because you have to have a receptor for the virus,” he explains. “Or you have to be able to grow the bacterium; all these things are difficult.55 That’s why we use what we describe as possible, probable, and definitive evidence of disease.”

Possible means you have found an association. Probable incorporates such factors as location of the agent in the target tissues, levels of the pathogen or of antibodies to it, and biological plausibility. The ability to create animal models of diseases you are studying is key in establishing causality, Lipkin adds. “Do you have an analogous situation, and can you come up with a way of explaining it using an animal model that holds true?” he asks rhetorically.

Definitive proof, which you may not get to for a while, means you have, one, satisfied Koch’s postulates, or, two, demonstrated that introducing the vaccine reduces the incidence of disease or eliminates it completely, or that you have a specific drug that can improve the situation and that reduces the presence of the pathogen or antibodies,” Lipkin summarizes. Just as the polio vaccine sharply reduced polio, researchers demonstrate the viral cause of an illness when they show that the vaccine against it lowers the frequency of disease.

“So there are three ways that you definitively prove something,” continues Lipkin. “You prevent it with a vaccine; you treat it with a drug; or you go through Koch’s postulates to identify the virus, grow it, then re-introduce it to replicate disease. But Koch’s is an obsolete proof in an era of molecular markers.”

Creatures of scientific habit

Two decades ago I was a visiting fellow at the Harvard School of Public Health. As a writer in residence at the Longwood medical complex in Boston, I learned a great deal about immunology, toxicology, psychiatric epidemiology, and medical writing. But intriguing, tacit elements of my education transpired outside of the classrooms and amphitheaters as I observed details ruling the social dynamics among researchers and physicians. I learned that the brilliant scientists and dedicated healers I observed were not always immune to illogic and could sometimes succumb to the same sort of biases as the rest of us.

I also learned that when all you have is a hammer, everything looks like a nail. Whether your tool of choice is an exhaustive command of genetics or an encyclopedic intimacy with Freudian theory, it can became your preferred or even your default approach to medical problems, even after its limitations emerge.

Schizophrenia, for example, remains characterized as primarily genetic even though it has never been adequately explained by genetics. The mere 30 to 40 percent concordance of the disease in identical twins is testimony to this, as discussed in the previous chapter. Yet many cling to a purely genetic theory, which has become a dominant physiological paradigm for explaining disease rates and disparities, only recently tempered by epigenetics of the sort described in chapter 2. I recall that ascribing the disease to toxic family dynamics persisted long after theories of schizophrenogenic mothers and absent fathers had lost their credibility.

Scientists had fallen into a habit of paradigmatic thought, focusing on genetics, psychological trauma, and brain injury, all of which are implicated in schizophrenia but none of which completely explain the disease’s prevalence. This habit made physical causes, like brain chemistry or even infection, less believable candidates.

In The Structure of Scientific Revolutions, Thomas Kuhn illuminates how introducing new knowledge that entails a shift in the prevailing paradigm evokes resistance and hostility to the new ideas. The hurdles are not always scientific, because all too often the mantle of science shrouds politics, social biases, and even petty jealousies that hamper the understanding of physical as well as mental diseases.

The concern, of course, is that science has sometimes been suppressed for nonscientific reasons—usually because it is politically inexpedient or it violates dogma or the tenets of the dominant belief systems.

I recall Kuhn’s admonition as I notice that the word controversial in a journal article about PANDAS on Swedo’s desk has been heavily circled in black. When I ask her about this, she shows she is quite alive to its semantic implications. Raising her eyebrows, she questions the rhetorical strategies of some critics: “They call PANDAS ‘controversial,’ but it is so only because certain people question its elements without understanding them, sometimes in contravention to the evidence. It’s not aboveboard, but the very word conjures up a suggestion of… marginality, or of questionable science, without offering any specific criticism. It’s just not logical.”

I’ve heard the term controversial used in informal discussions and debates about the nature of PANDAS infections more often than in peer-reviewed journals. But it is also used in more accessible publications or sites such as Wikipedia, where laypeople and patients are likely to get their first exposure to PANDAS, an introduction that may be colored by such a loaded term. It implies skepticism in the same manner that the verb claim signals that a statement may be untrustworthy.

In 2005, Joanna Kempner of Rutgers University and her colleagues interviewed forty-one researchers about science that had been suppressed, and her team found that the manner in which such knowledge became marginalized and dismissed was “self-imposed, reflecting social, political, and cultural pressures on what is studied, how studies are performed, how data are interpreted, and how results are disseminated…. We were surprised that respondents felt most affected by what we characterize as ‘informal constraints.’” In other words, scientists who ventured into forbidden knowledge were often warned away from the topics by oblique disapproval from their colleagues that did not invoke scientific criteria and sometimes even descended to ad hominem attacks: “He’s crazy,” “He’s not well thought of around here,” and “That’s controversial” are comments I heard when I asked academics about novel contested theories, from the role of the enteric nervous system in psychological disease to the flawed history of genetic theories of intelligence dissected by Stephen Jay Gould in The Mismeasure of Man. Such dismissals say nothing about the theory itself and everything about the theory’s status as an acceptable area of study.

Spreading informally, forbidden knowledge demarcates some scientific territory as off-limits, Kempner wrote. “Researchers sometimes only know that they have encountered forbidden knowledge when their research breaches an unspoken rule and is identified as problematic by legislators, news agencies, activists, editors, or peers.”56 Scientists often choose to abandon the verboten topic or theory.

As researchers entertain the idea of infection as another root of mental disorders, we realize how often diseases, physical and mental, that have long been ascribed to genetics, diet, and behavior are actually infectious in nature. Acceptance has been hard and slow, and it’s sometimes hampered not by scientific hurdles, but by social ones.

Hard-to-swallow proofs

Consider that until 1994, doctors told patients that their stress levels and diets—particularly their fondness for coffee and spicy foods—were causing their ulcers. Most physicians treated patients with acid control, prescribing not only Tagamet, which reduces stomach acid, but also bland diets and milk. (Except for a few mavericks, like John Lykoudis, a Greek practitioner who prescribed antibiotics in the 1950s and 1960s.) The milk and bland diets were only partially helpful, however, and so were supplemented by biofeedback and even psychological counseling, which did little to reduce the rates of peptic ulcer disease, or PUD.

However, in 1982 Australian physicians Robin Warren and Barry Marshall proved that Helicobacter pylori, a familiar bacteria living in our intestinal tracts, accounts for 90 percent of stomach and duodenal ulcers as well as for some other gastric diseases, including stomach cancer.

But wait: In 1982, H. pylori was discovered to cause most ulcers, so why did doctors continue to prescribe Tagamet and milk until 1994? And why does the medical literature show that Dr. John Lykoudis of Missolonghi, Greece, was curing ulcer patients with antibiotics as far back as 1958?

To find out, I perused the PubMed site for gastroenterology journals tracing ulcer treatment. I learned that during the century before Marshall and Warren’s breakthrough, an infectious cause of ulcers had been “discovered” on least fifteen separate occasions by different clinicians.

More than a hundred and ten years ago, Professor Walery Jaworski of Krakow’s Jagiellonian University described “spiral-shaped microbes in the human stomach” of his ulcer-plagued patients;57 Giulio Bizzozero, a nineteenth-century Italian anatomist, had been the first to describe these troublemaking “corkscrew” bacteria in the stomach; and in the 1960s, Iranian surgeon Emami-Ahari saw evidence of bacterial infection in ulcers suffered by patients in his Tehran private clinic. He also used antibiotic therapy to cure them. These were just a few of the physicians who glimpsed the connection between gut pathogens and ulcers, although they didn’t have the tools or intimacy with microbiology to specify H. pylori as the culprit. Nonplussed, I thought of Goethe, who wrote, “Everything has been thought of before, but the problem is to think of it again.”

Lykoudis, however, took therapeutics a step further when he formulated a safe, effective antibiotic cocktail consisting of two quinolines, which are aromatic compounds with antibiotic qualities—the quinine used against malaria is one—and streptomycin, taken with oral vitamin A. In the heroic if not altogether wise tradition of physicians, he tested it on himself in 1958 and cured his own ulcers.58 Calling his patented remedy Elgaco,59 formed from the words elkos (Greek for “ulcer”), gastritis, and colitis, he reported curing thirty thousand patients. Word spread quickly and new patients thronged his office, but the Athens Medical Association responded by denouncing Lykoudis as a huckster and fining him four thousand drachmas. The Greek government quickly followed suit, censuring and fining him for using Elgaco without the proper testing.60

The government’s accusation was true, Lykoudis retorted, but he hadn’t tested it because pharmaceutical companies had flatly refused to run studies of his controversial treatment. Lykoudis did not have the advanced training and expertise that would allow him to identify the organism that caused ulcers, but testing by others who did might have helped him amass an airtight case for infection. No university or company he approached was willing to do so, however.61 He felt unfairly marginalized and persecuted by the medical establishment and the government alike. Was he? In retrospect, it seems likely. “No science is immune to the infection of politics and the corruption of power,” wrote Jacob Bronowski, a Polish biologist, historian of science, and poet. Lykoudis and his shunned cure could have been Bronowski’s poster child. In 1966, JAMA refused to publish the Greek physician’s paper entitled “Ulcer of the Stomach and Duodenum,” which outlined the link between infection and ulcers. With the exception of a self-published pamphlet, none of his writings on the subject ever found a publisher, ensuring that physicians the world over would cling to their prescriptions of Tagamet and milk.62

A century of resistance to accepting or even discussing and testing the role of bacteria in ulcers63 made that relationship forbidden knowledge. The point is not that the dissenters were correct and the conventional practitioners were wrong; many would-be scientific innovators blame medical intolerance for their obscurity when that is not the case. Some are simply misguided, and still others are outright quacks and charlatans who decry the hostility of the medical establishment and compare themselves to Galileo and Semmelweis as they line their pockets with the fees of the gullible.

The real point is the nature of the refutation offered by the medical establishment. Rejection of Lykoudis’s theory should have been based on science, data, evidence, and logic. When resistance to change is instead based on factors such as personality, bias, academic snobbery, political considerations, and a conspiracy of silence that ignores uncomfortable theories, the question becomes, as Lykoudis plaintively wrote, “Why the refusal even to test it?”

Without scientific publication or rigorous testing, the infection theory of ulcers was consigned to the forgotten annals of forbidden knowledge. Lykoudis’s notebooks detail a life blighted by professional frustration; he died in 1980, just two years before Warren and Marshall validated his life’s work.

The duo could prove the connection because they had access to tools unknown to Lykoudis, including the flexible fiber-optic endoscope developed in the late 1970s, which provided a safe technique to view the stomach and collect specimens from the gastric mucosa of live patients for more accurate diagnosis. Modern nutrient media and incubation techniques also allowed Warren and Marshall to grow the organisms in culture, as Koch and his scientific progeny dictate.

In 1985, having satisfied Koch’s postulates, Warren and Marshall triumphantly published their findings that H. pylori, not stress and spices, causes ulcers. But once again, the H. pylori hypothesis failed to change physicians’ behavior. According to the CDC, at that time, most physicians knew of the H. pylori association, but half of primary care doctors did not test their ulcer patients for H. pylori.64 They ignored it while doling out treatments like acid suppressants, which were ineffective against the root cause of the disease. Despite their superior tools and access to publication, Warren and Marshall were about to share Lykoudis’s professional fate: studied indifference and obscurity.

“Everyone knew that bacteria couldn’t survive in the stomach’s acid environment,” Marshall told the Sydney Morning Herald to explain the widespread resistance to the discovery. “They’d been taught so at medical school.”65 In the end, it took more than scientific evidence to get the medical world’s attention; it took showmanship. The same self-experimentation that Lykoudis had conducted privately, Marshall shrugged off his lab coat and his gravitas and performed publicly.

To illustrate his claim, Marshall drank a beaker of H. pylori in culture66 in 1984, and within days, he was rewarded by nausea and vomiting. An endoscopy revealed the appearance of both H. pylori and gastritis, which Marshall was then able to banish with antibiotics. The fading of his symptoms in two weeks demonstrated, for at least the fifteenth time, that a microbe can cause gastric woes and that antibiotics can cure them.

In 1985 the victorious pair published their results in a wildly popular Australian Medical Journal paper, and in 1994, JAMA followed up with a National Institutes of Health consensus opinion that most duodenal and gastric ulcers were caused by H. pylori and that antibiotics were now the recommended treatment.67 It turned out that milk and Tagamet had enjoyed their limited success only because lowering the stomach’s acidity changed the stomach milieu sufficiently to discourage H. pylori infection.

The infectious cause of ulcers finally entered the medical canon when, in 1997, the Centers for Disease Control and Prevention spearheaded a public-health campaign to spread the word that ulcers were a curable infection, and in 2005 Marshall and Warren scored the ultimate validation: they were awarded the Nobel Prize in Physiology or Medicine for their “discovery of the bacterium Helicobacter pylori and its role in gastritis and peptic ulcer disease.” Today we understand that ulcers are caused by H. pylori, spread by contaminated food and groundwater and through human saliva via kissing.68

Such mischaracterizations of microbial disease abound in history.69 In nineteenth-century Florence, for example, Italian physician Domenico Antonio Rigoni-Stern noticed that cervical cancer affected married women and prostitutes70 but spared nuns. His conclusion: Cervical cancer was caused by tight corsets.71

By the end of the Victorian era, women’s doctors tied cervical cancer to early and frequent sexual contact with multiple partners—and to poor male hygiene. In the mid-nineteenth century, Alabama women’s doctor James Marion Sims infamously declared that due to black women’s lasciviousness, 60 percent of the black female patients seen in hospitals had cervical cancer. Cardiologist Daniel Hale Williams challenged his evidence, or rather the lack thereof, wondering in print how Sims could have determined this when he could produce no records, and black women were rarely permitted into the white hospitals of the time.72 Still, Sims’s views were widely adopted. By the 1970s, cervical cancer was laid at the doorstep of the sexual villain of the day: herpes infection. It was not until the 1980s that Harald zur Hausen finally discerned that strains 16 and 18 of the human papillomavirus, HPV, accompanied most cervical cancers and theorized that HPV was the cause in 70 percent of cases.

Only then could a vaccine against a major source of cervical cancer be crafted. It worked, demonstrating that HPV causes cervical cancer, for which zur Hausen won the 2008 Nobel Prize in Physiology or Medicine. For HPV, cancer is a strategy. Because sexual transmission of HPV is inefficient except in the most active and promiscuous individuals, the virus would have few chances to spread to another host if HPV strains did not trigger the cells they infect to continuously and recklessly divide—the definition of cancer—allowing HPV to divide and proliferate right along with them. The virus remains shrouded from surveillance by the immune system. Cancer is thus a means to an end, and once again the health of the human host becomes collateral damage.

In yet another example, heart disease has been especially riddled with stubborn mythology, including psychological determinism, a subtle form of blaming the victim, as the hostility and aggressive behavior of people with type A personalities was broadly indicted in heart attacks. We now know that infectious agents such as C. pneumoniae are key players that triple the risk of coronary artery disease. Stanley Prusiner’s theory that infectious proteins caused both mad cow disease and its human variant, CJD, was met with hostility and ad hominem attacks. He was reviled as a self-promoting huckster for decades before winning the Nobel Prize for discovering the prion.

Resistance to discussing and testing the role of microbes in ulcers, heart disease, and cervical cancer transformed those relationships into forbidden knowledge. The pertinent question is not whether the theory is correct, but why the nature of the criticism leveled at it is informal and nonscientific or, in other cases, why the theory is ignored, guaranteeing it will be forgotten. Why does science treat some theories as taboo and forbid objective discussion and testing?

Certainly, compelling evidence is necessary before we change our thinking about the infectious transmission of diseases that were once “known” to have noninfectious causes. But definitive data is not always enough, because often, once it is presented, the evidence is ignored.

Even proven theories and demonstrated facts sometimes become forbidden knowledge, for social and political reasons as well as scientific ones. Such knowledge can be unacceptable for many reasons—it may call the theories or work of other scientists into question, or it may sabotage established hierarchies or an entrenched scheme of thinking about disease.

Sometimes the new paradigm exposes treatments as illogical and ineffective, and always, it does little to bolster the egos of scientists whose careers depend on subscribing to established disease paradigms. Although science is shaped by formal regulations and policies, researchers who traced its genesis found that, according to respondents, most of its constraints are informal or self-imposed, reflecting social, political, and cultural pressures on what is studied, how studies are performed, how data are interpreted, and how results are disseminated.

Discovering the roots of Tourette’s and anorexia in sore throats is far from the strangest paradigm shift along the infection-and-mental-disease spectrum. There is also the case of mental disorders such as depression and autism that are cued by the microbes of the “second brain”—the one that resides in your gut, as the next chapter explains.

If you find an error or have any questions, please email us at Thank you!