“How high was the fever?” isn’t a very useful question
© 2015 Roy Benaroch, MD
A September, 2015 study published in the Pediatric Infectious Disease Journal looked at whether fever itself – the height of a fever – is a good predictor of whether a child is likely to have a serious illness. And, once again, a common fever myth is shattered.
Researchers reviewed data from visits from children, aged 0-5 years, who presented to a pediatric hospital for fever in 2004-2006. (Yes, that was an oddly long time ago. I guess they write slowly in Australia. But in terms of the science, there haven’t been any big changes in vaccines or childhood illness since then, so I think the results are still valid.) Of the almost 16,000 fever episodes reviewed, about 1,000 children had what the authors considered a “serious infection”. That includes bacteremia (bacteria in the blood) in 64 children, pneumonia in 533, and urinary tract infection in 543. In total, only about 7% of the children had a potentially serious bacterial infection that needed to be treated. That percentage should be no surprise—we know that in the modern world, among vaccinated and healthy children, the vast majority of fevers are caused by viral infections that will get better on their own.
The authors then looked at whether the height of the children’s fever correlated well with whether or not they ended up having a serious diagnosis. They examined both the measured highest fever in the emergency department, and the parents’ self-reported “highest fever” in the last 24 hours, prior to their visit. Whichever way it was recorded, the height of the fever itself did a very poor job at discriminating between children who had and did not have a serious infection.
The statistics are complicated, and involve receiver operating curves that you may not be familiar with—but, in the author’s words, “Measured temperature at presentation to hospital is not an accurate marker of serious bacterial infection in febrile children.”
Relying on the height of a fever resulted in both false positive and negative results. For instance, evaluating only those with a temperature of 100.4 or higher would still miss 1 in three serious infections. You can’t rely on high fevers as a positive predictor, either—only 1 in 6 children with a temperature of 104 actually had a serious infection.
The authors found that a few other observations could be at least somewhat predictive. Younger babies, overall, were a little more likely to have serious infections; and children who had had fevers for more than 4 days were also somewhat more likely to be diagnosed with something more serious (though I imagine those children also had more tests done, like chest x-rays. Perhaps they just found more illness when they looked for more illness.) But even when adding in considerations of age and length of illness, the height of the fever remained a poor predictor of the seriousness of the illness.
So what’s a parent to do?
First: prevent serious illness by making sure your children are up to date on their vaccines. We can prevent most causes of meningitis, pneumonia, blood poisoning, and other serious bacterial and viral infections with vaccines. Prevention is always better than having to evaluate and treat children in an emergency department or office visit.
Then: offer comfort care to children with fever. Extra fluids are a good idea, and if fever is causing a child to be uncomfortable, treat it. If Junior perks up and acts well after fever-reducing medication, and isn’t especially at-risk for serious infection**, treat fevers at home for a few days. If the child is getting worse, acting persistently sick, or isn’t getting better, go see your doctor. Remember: the number itself is less important than how your child feels and acts.
** This advice does not include any child less than 2-3 months old, or with an immune deficiency, or who isn’t up to date on vaccines for any reason. These kids are at much higher risk of serious or deadly infections, and shouldn’t rely on advice from the internet. Call your doctor for specific fever instructions.
More about fever: