Journal Logo



HEAT Trial Moves Treatment for Fever Closer to a Resolution

Morgenstern, Justin MD

Author Information
doi: 10.1097/01.EEM.0000484510.18123.66

    You have just finished seeing a 72-year-old woman with a fever, cough, and altered mental status. She is tachycardic, tachypneic, and has decreased urine output. You diagnose her with sepsis, and before confirming pneumonia on chest x-ray, you rapidly draw cultures, give empiric broad-spectrum antibiotics, and start fluid resuscitation. You already spoke to the intensivist, who has a bed available and is ready for the patient to be transferred upstairs. You were ready to get back to the long line of ankle sprains and runny noses when the nursing student grabs your arm, and with a slightly worried look on his face points to the 103°F temperature.

    The value of fever in infection is a common medical debate. Some say it helps; others say it hurts. Some say it must be treated; others say it should be left alone. Everyone seems to have an opinion, including my patients after their inevitable consult with Dr. Google.

    Two arguments prevail about the physiologic effects of fever in infection. The increased metabolic demands required to mount a fever might be detrimental to sick patients with limited reserves. (Am J Respir Crit Care Med 1995;151[1]:10.) But, says the counterpoint, a fever is a natural immune response and is evolutionarily designed to fight infection, and bacterial growth is inhibited at higher temperatures. (Infect Immun 1986;52[2]:484.)

    As a quick historical aside, the 1927 Nobel Prize for medicine was awarded to Julius Wagner-Jauregg, MD, for his discovery of the therapeutic value of malaria inoculation in treating dementia paralytica. Neurosyphilis was a terminal disease in this pre-antibiotic era, but Dr. Wagner-Jauregg injected patients with Plasmodium vivax, provoking a fever that cured them. The resultant malaria was then treated with quinine. (Science 2013;342[6159]):686.)

    After years of debating the value of fever based on animal data and observational trials, we finally have some high-quality evidence from a new study: the HEAT trial. (New Engl J Med 2015;373[23]:2215.) This randomized, double-blinded, controlled trial in ICU patients 16 and older with a temperature of 38°C or higher being treated for a known or suspected infection compared acetaminophen 1 gram IV every six hours with a placebo.

    The results? For the primary outcome — ICU-free days at day 28 (where death counted as no days free from the ICU) — no difference was seen between the two groups (23 ICU free days in the acetaminophen group vs 22 with placebo). No differences were found in the secondary outcomes of all-cause mortality at 28 or 90 days, ICU length of stay, and hospital length of stay.

    Shorter ICU Stay

    Acetaminophen was associated with a shorter ICU length of stay among patients who survived, according to a pre-specified subgroup analysis, but a longer ICU length of stay among nonsurvivors. It's not clear how you would use those data, however, because you won't know ahead of time which patients are going to survive unless you have a clinical crystal ball.

    The authors deserve a lot of credit. This was an incredibly well run, large, multicenter trial. They published their protocol prior to enrolling patients and then published the planned statistical analysis before looking at the data. They used an intention-to-treat analysis, and had complete follow-up data on all patients except 10 (1.4%) who withdrew their consent after being enrolled in the study.

    Like any trial, though, a few weaknesses could have potentially influenced the results. The study did not enroll consecutive patients, with more than 1,000 patients who were eligible for the study not being enrolled. And a large number of protocol violations were made, with almost one in five patients either missing a dose or being given an extra dose of the study medication. Acetaminophen was used open-label after the study period, and was given to about 30 percent of patients in both groups. It's not clear how or if these weaknesses affected the trial's results. The bottom line is that, in the highest quality study to date, no benefit was found in routinely treating febrile patients with acetaminophen.

    How should we apply this information clinically? I think this reinforces the general consensus in emergency medicine that a fever does not need to be treated. It is unlikely, based on these results that treating a fever is going to change clinically important outcomes.

    Does that mean that febrile patients should never be given acetaminophen? Of course not. Think back to the last time you had a flu-like illness. You probably felt horrible, and acetaminophen has long been our first choice for symptom control. As I write this, I am recovering from a “man cold,” and I took acetaminophen. This trial did not demonstrate a benefit from routinely providing acetaminophen, but it also did not indicate any harm.

    So what do I do? I don't routinely treat fevers. I reassure my patients and my emergency department team that the best evidence we have indicates that fever is not dangerous and that we don't need to treat it with acetaminophen. We all know, though, that fever can cause discomfort, so I order acetaminophen for any febrile patient who wants it for symptom control.

    Share this article on Twitter and Facebook.

    Access the links in EMN by reading this on our website or in our free iPad app, both available at

    Comments? Write to us at [email protected].

    Wolters Kluwer Health, Inc. All rights reserved.