Rethinking fever: New study redefines body temperature

Researchers at Boston CHildren's Innovation & Digital Health Accelerator take new look at body temperature.
PHOTO ILLUSTRATION: PATRICK BIBBINS/BOSTON CHILDREN’S HOSPITAL

Your first patient of the day presents with a sore throat and a temperature of 99.5 degrees. Although a little higher than normal, it’s not technically a fever, right? Jonathan Hausmann, MD, a rheumatologist at Boston Children’s Hospital and Beth Israel Deaconess Medical Center, might disagree.

Hausmann, Fatma Dedeoglu, MD, and their colleagues from the Boston Children’s Hospital Innovation & Digital Health Accelerator, recently published a study in the Journal of General Internal Medicine that they hope will begin a larger dialogue among physicians and others about normal body temperature and the definition of fever.

Body temperature, fever lower than current standard

“The assumption is that normal temperature is 98.6 degrees and fever begins at 100.4. But we found that the average temperature is more like 97.7,” he says. “And by defining a ‘fever’ as temperatures above the 99th percentile of normal, we found fever typically begins at about 99.5 degrees.”

His interest in the topic stems from his work treating patients for rare periodic fever syndromes, also called autoinflammatory diseases. “Before studying patients with these rare syndromes that cause fever, I wanted to first understand normal and febrile temperatures in otherwise healthy patients,” Hausmann explains.

He says this project came out of speaking with parents and patients, many of whom told him they believe their normal body temperature is lower than 98.6 degrees.

“They would tell their doctors they had a fever when their thermometer hit 99 degrees, but because this didn’t meet the traditional definition of fever, their concerns were dismissed” he says. “So, we wanted to find out what a ‘normal’ temperature really is and also to take a new look at what constitutes a fever.”

Taking advantage of new technology

To conduct the study, Hausmann and his colleagues took a novel approach — smartphone crowdsourcing.

“Apple recently came out with a platform called ResearchKit that allows us to conduct a research study by developing our own app,” he says. “Patients can then download the app, read more about the study, and sign the consent form on their phones. Then they answer questions about medical history and log their temperatures on the app.”

Hausmann and colleagues created the app, Feverprints, and opened their study to both children and adults after being approved by the Boston Children’s Hospital Institutional Review Board.

“Participants could take their temperature any way they wanted, at any time of day, whether they thought they had a fever or not. We also asked them to record their symptoms at the time the temperature was taken, and whether they had taken medications to help lower fever.”

For the published study, Hausmann included only temperatures taken orally in asymptomatic adults who were not taking antipyretics.

On average, women had slightly higher temperatures than men, and temperature decreased slightly with age. They also found that body temperature seems to run on a circadian rhythm, with the lowest readings in the early morning, peaking in the late afternoon.

“While a temperature of 99.5 first thing in the morning is probably a fever, it may not be so at 4 p.m.,” he says.

Hausmann says the results also show the value of crowdsourcing research.

“There was a lot of skepticism when we started this study. Many people wondered whether we could get reliable data from people reporting their own information,” he says. “But we found that our data matched perfectly the data from smaller, published studies that were conducted under controlled conditions, showing that this way of conducting research can be valid. It’s also a fairly inexpensive way of recruiting a large number of patients for a study over a short period of time.”

Changing how we view fever

He hopes this study will help doctors and patients realize that body temperature should not be seen as a binary variable, but should be interpreted in light of who the patient is and the time of day the temperature was taken.

“Even though we have known the circadian fluctuation of temperature for many years, most doctors and patients aren’t aware of it,” he says. “It hasn’t yet changed medical care or how we view fevers.”

With this study under his belt, Hausmann is interested in pursuing further research on fevers. Two areas of interest include learning the effect of antipyretics on length of illness, as there is some research that suggests lowering fever may prolong illness, and determining a “fever print” for various illnesses, using fever patterns as a way to diagnose illnesses faster and more accurately.

Learn more about the Innovation & Digital Health Accelerator.