It is 3AM and you are called about Mr. Smith, a 74-year-old man who presented to the hospital tonight with a cough and was found to have pneumonia and sepsis. You have started him on antibiotics, but he now has a temperature of 38.3°C (100.9°F). His nurse asks, “Should we give him something for that fever?”
Some of us, concerned that fevers may put excess stress on the critically ill, might order acetaminophen as an antipyretic. Others, acknowledging studies that suggest elevated temperatures may enhance immune cell function and inhibition of pathogen growth, might recommend against lowering Mr. Smith’s fever. The HEAT trial (Permissive Hyperthermia through Avoidance of Acetaminophen in Known or Suspected Infection in the Intensive Care Unit), published OLF last month in NEJM, aims to provide some clarity for those who have struggled with this same question: does treatment with acetaminophen improve or worsen outcomes in the critically ill and febrile?
The HEAT authors, Young et al., conducted a blinded, randomized controlled trial at 23 medical- and surgical-intensive care units in New Zealand and Australia, comparing scheduled 1 gram IV acetaminophen to placebo (5% dextrose in water) every 6 hours. HEAT enrolled patients ≥16 years of age with a temperature of at least 38°C (100.4°F) who were on antibiotics for a known or suspected infection. Study treatment was to be continued for 28 days, or until one of the following occurred: the patient was discharged from the ICU, the patient’s fever resolved, antibiotics were stopped, the patient developed a contraindication to ongoing therapy, or the patient died. Study outcomes included ICU-free days, mortality, length of stay, fever characteristics and specific laboratory values.
Following analysis of the 346 patients in the acetaminophen group and the 344 patients in the placebo group, the authors found that the acetaminophen group had significantly lower mean daily peak temperature (38.4° vs 38.6°) and mean daily average temperature (37.0° vs 37.3°). Despite this, the authors did not detect a significant difference between the acetaminophen and placebo groups when it came to ICU-free days to day 28 (23 days vs 22 days, p=0.07) and 90-day mortality (15.9% vs 16.6%, p=0.84), as well as ICU length of stay, hospital length of stay, and key laboratory results. Additionally, no significant differences in outcomes were noted between four pre-specified subgroups: those with septic shock vs those without, those taking aspirin vs those not taking aspirin, those with “high fever” (≥39°C/102.2°F) versus those without, and those identified as having community- versus hospital- versus ICU-acquired infections.
However, study authors warn about two key limitations. First, duration of treatment was short (median number of doses given was 8 in the acetaminophen group, and 9 in the placebo group). Second, there was frequent use of open-label acetaminophen in both groups after completion of the study drug course (30.0% of the patients in the acetaminophen group, 29.4% of patients in the placebo group). As such, this study may only apply to the consideration of acetaminophen early in an ICU course.
Keeping these limitations in mind, HEAT still provides us with a reasonably powered randomized controlled trial that finds neither a significant benefit or harm from scheduled IV acetaminophen use in patients who are critically ill and febrile from a suspected infection.
You assess Mr. Smith who remains febrile and appears uncomfortable. His wife and nurse are at the bedside and agree. You discuss that while acetaminophen may reduce his fever, it is unlikely to result in longer-term clinical benefit or harm. Given his discomfort, you all agree on a trial of acetaminophen with reassessment of his fever and discomfort in the morning.