Fever at home, no fever in the ED


How to handle vanishing symptoms once you see the patient.

Your next patient is a four-week-old girl brought in by her parents due to fever. At home this evening, she felt warm and had a rectal temperature of 100.9 (38.3). The parents brought her immediately to the ED. They gave no antipyretic medications prior to arrival.


The baby was born via repeat C-section at term after an uncomplicated pregnancy. She has had no problems up to this point. There has been no change in her feeding, voiding or stools. She has had no cough or congestion.

In the ED, the baby is afebrile. She looks good and has no source for fever based on her physical examination. So, does this baby still require a work-up? Should you base this decision on her current vitals or does the fever at home still count?

Three fairly recent studies have looked at this very question, specifically, what is the likelihood of a serious bacterial infection in a young, non-toxic appearing infant with a history of fever documented at home, but no fever on presentation to the ED?


Yarden-Bilavsky and colleagues did an observational, prospective study in Israel, looking at infants 28 days or younger presenting to the ED with fever without a source.[1] They defined fever as 38 degrees Celsius or higher. Infants who felt warm, but did not have a temperature measured by their parents were not included if they did not have a fever in the ED. Babies born prematurely, with underlying medical conditions or pretreated with antibiotics were also excluded.

They looked at 399 infants of which 143 (35.8%) had fever by history only and 256 (64.2%) had fever documented in the ED. In the fever by history group, 12 (8.4%) had serious bacterial infections, all of which were urinary tract infections (UTIs).

In the group that had fever in the ED, 46 (18%) had serious bacterial infections, including 33 UTIs, 4 UTIs with bacteremia and 4 pneumonias. The authors stated that the small number of infections in the group with no fever on arrival to the ED made them question the need to work-up these babies.

Ramgopal and colleagues looked at a larger number of patients.[2] They performed a secondary analysis of the PECARN (Pediatric Emergency Care Applied Research Network) data set. This data was collected as a prospective, cross sectional cohort study of patients from 26 EDs. In that study 3,835 infants were included. Excluded were former premature infants, those with focal infection (except otitis media), septic infants, those with recent antibiotic use, serious underlying illness or indwelling catheters or shunts.


Of the 3,825, 1,233 (32.2%) had fever by history and 2,592 (67.8%) had fever on arrival to the ED. There were 108 serious bacterial infections in the fever by history group, for a rate of 8.8%. There were 331 serious bacterial infections in the fever on presentation group, for a rate of 12.8%. The bulk of the infections were UTIs, but there were also some cases of bacteremia and meningitis. The authors concluded that the slightly smaller risk in the group without fever on presentation should not change the decision to perform a fever evaluation.

It is interesting to see that both this study and the Yarden-Bilavsky study found almost the same number of infections in the group that was afebrile on arrival, 8.4 versus 8.8 %. However their conclusion, based on almost the same percentage, was the complete opposite!

Expanding the age range to 90 days, there is one more study to consider. Mintegi and colleagues looked at infants in Spain presenting to the ED with fever by history. [3] Their study was a prospective look at infants less than 90 days of age presenting to the ED with no source for fever on exam.

Like the prior two studies, their definition of fever was 38 Celsius or above and they excluded tactile fevers. They included term infants who were well-appearing, with no chronic disease or antibiotic exposure. They looked at 2,470 infants. 678 had fever by history only and 1,792 had fever in the ED. Their rates of serious bacterial infection were much lower than the prior two studies, which could be expected due to the older age range. The rate of serious bacterial infection was 2.4 % in both groups, although none of the babies that were afebrile had meningitis (out of 678) whereas there were seven in the febrile group that had meningitis (out of 1,792).

So what is the overlying message? Young infants with measured fever at home, but no fever when they come into the ED had a lower likelihood of serious bacterial infection than their counterparts with fever on presentation in two out of the three studies, but still had a significant risk of infection, particularly UTI.

I agree with the authors of the PECARN analysis that you should approach fever in the young infant in the same manner regardless of whether the temperature was documented at home or in the ED.

But what should you do with subjective fever, the infant who “felt warm” at home? None of these studies included that group. But one study from 1996[4] and another from 2003[5] found that parents’ assessment of fever by touch had a sensitivity greater than 80%.  When in doubt, you should probably err on the side of treating this the same as a documented fever at home.

So our young patient gets a full sepsis work-up. On to the next patient.

[1] Yarden-Bilavsky H, Bilavsky E, Amir J, et al. Serious bacterial infections in neonates with fever by history only versus documented fever. Scand J Infect Dis. 2010;42: 812-816,

[2] Ramgopal S, Janofsky S, Zuckerbraun NS, et al. Risk of serious bacterial infection in infants age < 60 days presenting to emergency departments with a history of fever only. J Pediatr 2019;204:191-195.

[3] Mintegi S, Gomez B, Carro A, et al. Invasive bacterial infections in young afebrile infants with a history of fever. Arch Dis Child. 2018;103:665-669.

[4] Hooker EA, Smith SW, Miles T, and King L. Subjective assessment of fever by parents: comparison with measurement by noncontact tympanic thermometer and calibrated rectal glass mercury thermometer. Ann Emerg Med. 1996:28:313-317.

[5] Callanan D. Detecting fever in young infants: reliability of perceived pacifier and temporal artery temperatures in infants younger then 3 months of age. Pediatr Emerg Care. 2003;19:240-243.


Dr. Levine is a professor of pediatrics in the Division of Pediatric Emergency Medicine at the University of North Carolina.

Mackenzie Esch, MD specializes in Pediatric Emergency Medicine in Chapel Hill, NC and Chapel Hill, NC.


  1. Christopher Russo on

    “Full sepsis” workup includes a lumbar puncture–this is presented in the article as an obvious thing to do but this is not necessarily indicated in this well-appearing four-week old.

    • Bedjinie Bayard on

      Hi Christopher Russo, does the disappearance of the fever in those children with a history of fever indicate a low level of white blood cells or does it happen because they lose the white blood cells in the urine (positive leukoesterase)?

Leave A Reply