Behind every famous number lies a story. And in this season of sickness, when the thermometer frequently emerges from the bathroom cabinet, few numbers in medicine are as familiar as 98.6 - the normal temperature of the human body.
Celebrated in song and enshrined for more than a century in schoolbooks and medical texts, 98.6 degrees Fahrenheit is the benchmark many of us use to determine who goes to school or work and who stays in bed.
There's just one little problem: 98.6, it turns out, is a medical myth.
While it continues to circulate in publications ranging from this month's Parenting to the Bantam Medical Dictionary, studies in recent years have shown that 98.6 is not the normal human body temperature - and probably never was.
So how did this little scarlet line become sacred in the first place, and why has it fallen out of favor?
It's a tale that begins more than 150 years ago at a university hospital ward in Leipzig, Germany - and ends here in Baltimore, at the door of Dr. Philip Mackowiak.
Mackowiak, 59, is an infectious disease specialist and fever researcher at the Veterans Affairs Medical Center on North Greene Street. In the early 1990s, he became increasingly suspicious about the textbook value for the average oral body temperature, since it rarely reflected what he saw each day in his exam room.
"Every now and then, 98.6 would come up," he recalls. But only now and then. "It just didn't seem to make sense."
His suspicion eventually led him to conduct the first major examination of body temperature in more than a century. But he didn't stop there. An amateur medical historian who organizes an annual seminar to diagnose fatal diseases of the famous, Mackowiak decided to ferret out the number's obscure origins.
Despite its bedrock standing, "I had no idea where it came from," he says. And he suspected that most of his colleagues didn't, either.
Finding the answer would eventually require hours leafing through musty journals and century-old German medical tomes. He also wound up doing some unexpected historical sleuthing. For example, when Mackowiak heard about a 150-year-old thermometer suspected of playing a role in the 98.6 story, he gingerly toted it from Philadelphia to his Baltimore lab for testing.
In the end, Mackowiak discovered - perhaps not surprisingly - that the history of 98.6 is intimately intertwined with the thermometer's arrival in medicine itself.
Although the mercury thermometer was invented in 1714 by Gabriel Fahrenheit, it would take more than a century for physicians to apply it seriously to the measurement of body temperature and fever.
Volker Hess, a historian at Humboldt University in Berlin who has written one of the few books on the history of medical thermometry, says one problem was that early instruments were not exactly patient-friendly: A single measurement required a subject to lie quietly on his back for as long as 30 minutes.
Even then, the reading wasn't always trustworthy. According to historian John Haller Jr. of Southern Illinois University, it's not uncommon to find temperatures as high as 118 degrees Fahrenheit recorded on old patient charts. One early 19th-century physician reported treating a patient in a horse accident whose fever spiked to an incinerating 122 degrees.
But as instrument accuracy improved, so did the number of physicians investigating the link between temperature and disease. And it wasn't long before 98.6 made its appearance in medical literature.
As far as historians can tell, in 1835 French investigators Antoine Becquerel and Gilbert Breschet were the first to report that the mean, or average, temperature of a healthy adult was 98.6 degrees Fahrenheit. But, for whatever reason, their finding didn't generate much attention.
In fact, it took three more decades for 98.6 to take hold. The man generally credited with cementing its fate was a German physician, Carl Reinhold August Wunderlich.
In 1868, Wunderlich published Das Verhalten der Eigenwarme in Krankheiten. Translated three years later into English as On the Temperature in Diseases: A Manual of Medical Thermometry, it was immediately viewed as a landmark - "the first truly classical study of the thermometer in clinical practice," Haller writes.
In his book, Wunderlich reported that in 1851, he began recording the temperature of every patient who passed through his busy ward at the University of Leipzig. At first, Wunderlich ordered patient temperatures taken twice a day. Later, he increased it to four times, then six, "and in special cases even more frequently," he noted.
The result: Wunderlich's files ultimately held temperature profiles of more than 25,000 people - a pool of data so vast that even Wunderlich didn't know how many individual readings he had collected. He estimated several million.
Wunderlich's book broke new ground. The German physician was the first to point out that body temperature naturally fluctuates during the day, sinking to its lowest between 2 a.m. and 8 a.m. and reaching its peak between 4 p.m. and 9 p.m. He was also the first to set the threshold for fever, which he defined as any temperature above 100.4 Fahrenheit.
Wunderlich also noted that women have a slightly higher temperature than men, that older people are slightly cooler than younger people, and that temperature may vary by race.
But the finding that would gain the most notoriety, both in medical circles and among the public, was printed prominently on the very first page of his book: The mean temperature of the human body is 98.6 degrees Fahrenheit.
Wunderlich's findings rapidly filtered into the medical literature in Germany and the United States. Yet few scientists bothered trying to replicate them.
Wunderlich "was the Guinness Book of Records" of body temperature researchers, says Jos van der Meer, a fever researcher at Radboud University Nijmegen in the Netherlands. He says most researchers probably looked at Wunderlich's mountain of data and figured, "Why fight it?" So more than a century later, when word spread among Mackowiak's colleagues that he was planning to double check 98.6, some thought he was wasting his time. Mackowiak recalls one friend even joked, "When are you going to do one on the pulse rate? How about respiration?"
But as a fever specialist, Mackowiak was deeply interested in the boundary between normal and febrile. Flipping through medical textbooks, he found that the answer wasn't clear-cut. Some defined a fever as any reading above 98.6. Others said the threshold was 99, while still others said it was 100.4. Just as importantly, Mackowiak realized he had an easy way to put Wunderlich to the test: a trove of healthy human body temperatures right at his fingertips.
The database belonged to the University of Maryland's Center for Vaccine Development, which routinely recorded the oral temperatures of volunteers recruited to test experimental vaccines.
Mackowiak and two colleagues collected 700 temperatures from 148 healthy adults in the database and found that readings ranged from 96 degrees to 100.8 degrees. When they computed the average, they found it was just 98.2 degrees Fahrenheit. Only eight percent of the 700 temperature readings came up 98.6.
Wunderlich's famed measurement, Mackowiak and his team concluded in a 1992 article in the Journal of the American Medical Association, "has no special significance vis-M-'-vis the body temperature of healthy adults." As a result, 98.6 "should be abandoned." But, Mackowiak wondered, where had Wunderlich gone wrong? How could more than a million temperature readings have led to a different conclusion?
One reason, he realized as he pored over Wunderlich's writings, was that the German physician had placed his thermometers in his patients' armpits, not under their tongues. Although it was the customary way to take a patient's temperature then, physicians today know the armpit generally runs cooler than the mouth.
An even more suggestive clue developed from a chance encounter with the curator of the Mutter Museum in Philadelphia - home to more than 20,000 pickled fetuses, antique surgical tools and other medical oddities. After listening to Mackowiak describe his research, the woman brightened. She had an object in the museum archive he might be interested in: one of Wunderlich's original thermometers.
"I was absolutely shocked," Mackowiak recalled.
But the bigger surprise came back in his lab. When he tested it against a carefully calibrated modern digital thermometer, Mackowiak found the 19th-century instrument read as much as 3 degrees high. So even a relatively cool armpit might have been recorded as hotter than its real temperature.
Mackowiak found other reasons to be suspicious of 98.6. For starters, he says, it's hard to imagine how the German doctor, in an era before computers, could have crunched so much data. Wunderlich's research was also carried out before statistical analysis became common practice, Mackowiak notes.
In the years since Mackowiak published his research, he says many medical publishers have scrubbed 98.6 from their dictionaries and textbooks. Today, they define fever as simply a rise in temperature above normal, without specifying what normal is.
Mackowiak says that the most important conclusion to emerge from his study is not that 98.6 is wrong but that normal depends on the person, and that body temperature fluctuates based on everything from the time of day to a person's age, gender and even race.
So how do doctors diagnose fever today?
"It's a difficult number to come up with," Mackowiak concedes. In his JAMA study, he and his colleagues called fever any oral temperature higher than 99 degrees in the early morning and 100 degrees in the early evening.
But some clinicians work under different rules. "A temperature above 101.5 is when we usually start to worry," says Sherry Weinstein-Mayer, an internist in Kaiser-Permanente's Towson office.
Aside from setting the record straight, Mackowiak concedes that lowering the official norm to 98.2 would have little impact on how doctors treat patients, since temperature is just one part of the clinical picture.
But he and others say the study does teach an important lesson: "Even deeply engrained medical dogma is potentially flawed and should not be immune to periodic challenge," Mackowiak says.
And body temperature may not be the last basic measurement to come under scrutiny.
At Radboud University, van der Meer says a few researchers have begun to cast a suspicious eye at how blood pressure is measured and recorded.
But that, he says, is another story.