Investigation of afebrile neonates with a history of fever

Pediatrics

CJEM 2004;6(5):343-348

Abstract

Résumé

Introduction

For the past decade it has been recommended that all neonates (i.e., ¾28 days of age) with a temperature >=38°C undergo diagnostic testing for bacterial infections, receive empiric intravenous antibiotics, and be admitted to the hospital.¹ This straightforward and conceptually simple approach continues to be recommended^2,3 even though more than 80% of febrile neonates do not harbour serious bacterial infections.^4-8 Given the difficulties in using the physical examination by itself to identify neonates who harbour serious bacterial infections^4,5 and the potentially devastating consequences of untreated bacterial infections in febrile neonates, a conservative approach, including laboratory testing, antibiotic administration and hospital admission seems prudent. At the present time the generally accepted approach to the management of a neonate who presents to the emergency department (ED) with a temperature >=38°C is clear and minimally controversial.

The significance of neonatal defervescence, however, remains controversial. It is unclear what investigations and disposition are indicated for neonates who have a history of fever at home, but are afebrile on arrival at the ED. A single study from 1987 suggested that the incidence of serious bacterial infections (SBIs) in neonates who defervesce prior to ED presentation is very low and might even be zero;⁹ however, at our institution, many of these neonates undergo a full septic workup, receive antibiotics, and are admitted to hospital. In addition, our ED physicians have incorporated viral testing that was not available in 1987 into the evaluation of afebrile neonates with a history of fever.

If neonates who defervesce prior to ED arrival are at extremely low risk of having an identifiable infection, it is conceivable they could be managed without laboratory testing, antimicrobials or hospitalization. This minimalist strategy would avoid exposing these neonates to iatrogenic risks such as nosocomial infections and medication errors,¹⁰ avoid unnecessary testing, treatment and hospitalization, and relieve the emergency physician of a frequently time-consuming and technically challenging laboratory evaluation.

The objective of our study was to describe the clinically significant infections identified in a cohort of afebrile neonates who underwent an ED septic workup because of the history of a measured fever at home.

Methods

Setting and patients

This retrospective medical record review was conducted at Loma Linda University Medical Center and Children's Hospital, Loma Linda, Calif., a tertiary care pediatric ED. Eligible subjects included all neonates ¾28 days of age who presented between Jan. 1, 1999, and Aug. 22, 2002, with a history of febrile illness, and who underwent lumbar puncture during their ED visit (Fig. 1). Infants were excluded if their reported home fever did not include a numeric value (e.g., the baby "just felt warm"), if their ED triage temperature was >=38°C or not recorded, or if their medical record was unavailable.

Fig. 1. Subject selection flow diagram

enlarge

Data collection

A trained data abstractor (T.S.), who was periodically monitored by the principal investigator (L.B.), reviewed the remaining medical records using a standardized data collection form.¹¹ The abstracted data included date of birth, date of visit, reported home temperature, medications administered at home (including antipyretics), ED triage temperature, bacterial culture results (including blood, urine, cerebrospinal fluid [CSF], stool and wound cultures), CSF bacterial antigen studies, CSF white and red blood cell counts, viral study results (including enterovirus, respiratory syncytial virus, rotavirus and herpes simplex virus [HSV] studies), microscopic urinalysis, the radiologist's final interpretation of a chest radiograph if performed, results of other diagnostic imaging and electroencephalograms, antimicrobial treatment administered, the condition of the patient at discharge and discharge diagnosis. Fahrenheit temperatures were converted to centigrade using the formula, C = (F - 32) * 0.5556.

Data analysis

Using a random number generator, we selected 22% of the main study group for inter-rater reliability testing. A second trained data abstractor, blinded to the results of the first, reviewed these randomly selected records for selected variables. Results of the 2 reviews were compared using a concordance rate for continuous variables and the kappa statistic for categorical variables. Other analyses were limited to descriptive statistics. Due to non-normal distributions, our data are reported as medians, ranges and interquatile ranges (IQRs). Statistical analyses were performed using Stata 8 (STATA Corporation, College Station, Tex.). This study was approved by our Institutional Review Board.

Results

We identified 206 neonates ¾28 days of age who underwent ED lumbar puncture during the study period. Of these, 108 were excluded because they did not have a home temperature documented, and 71 were excluded because they were still febrile on presentation to the ED. The main study group therefore included 27 patients who had a reported temperature at home >=38°C and who were afebrile at the time of ED arrival. Medical records, ED triage temperatures, and the home medication listing were available for all subjects.

Table 1 and Table 2 show that the study subjects were from 3 to 27 days of age (median 15 days; IQR 6-23 d) and predominantly male (63%). Reported home temperatures ranged from 38.0°C to 40.6°C (median 38.3°; IQR 38.2°-38.8°), while ED temperatures ranged from 35.9°C to 37.9°C (median 37.3°; IQR 36.9°-37.7°). Ten of the subjects (37%) had identifiable infections (Table 1) and 17 subjects did not (Table 2). No cases of bacterial meningitis, Listeria monocytogenes, Group B Streptococcus or HSV were identified. No subjects had concurrent viral and bacterial infections identified.

Two subjects demonstrated CSF pleocytosis without positive bacterial or viral testing. One of these, a 15-day-old male, underwent lumbar puncture after antimicrobials (ampicillin, cefotaxime and acyclovir) were administered. In this case, CSF culture, CSF bacterial antigen testing, and HSV type 1 and 2 tests were negative despite significant pleocytosis (1440 white blood cells/mm³ and 6 red blood cells/mm³). A brain magnetic resonance imaging revealed edema, and an electroencephalogram demonstrated right temporal lobe spikes suggestive of viral meningoencephalitis. The other subject, a 23-day-old female, underwent lumbar puncture, which demonstrated 485 white blood cells/mm³ and 495 red blood cells/mm³. Again, CSF cultures, CSF bacterial antigen testing, and HSV type 1 and 2 tests were negative.

Four subjects received home antipyretics (acetaminophen in every case). Three of these 4 had identifiable infections (2 Escherchia coli urinary tract infections [UTIs] and one rotavirus diarrhea) (Table 1). All 27 subjects received antimicrobials in the ED. Cefotaxime, ampicillin and acyclovir were the most common antimicrobials administered, given to 25, 24 and 8 subjects respectively. All 27 subjects were ultimately discharged from the hospital in good condition.

Inter-rater reliability testing revealed excellent agreement between the 2 independent data abstractors. The kappa statistic was 1.0, indicating perfect agreement for the following variables: medications administered at home, bacterial culture results, CSF bacterial antigen studies, viral study results, and the radiologist's final interpretation of chest radiography. The concordance rate was 100% for all CSF cell counts and 83% for recorded temperatures.

Discussion

In this study, identifiable infections were common in afebrile neonates who underwent an ED septic workup because of the history of a fever at home. We are aware of only 1 other study addressing this patient population. In 1987, Bonadio⁹ described 109 consecutive previously well full-term neonates who had either a tactile or rectal fever at home, were evaluated in the ED, and underwent lumbar puncture. There are important differences between studies. Bonadio excluded neonates who had received antipyretics within 4 hours and hypothermic neonates with a temperature <36°C. In the Bonadio study, 54 neonates were febrile at home and afebrile on ED presentation -- comparable to the patients described here. Bonadio found no SBIs in these 54 neonates, but did not report the results of any viral testing.

Assessing Bonadio's data using contemporary methodology,¹² an estimate of the upper limit of the 95% confidence interval (CI) for the incidence of SBIs can be calculated. Although Bonadio's incidence of SBIs was 0%, the upper limit of the 95% CI is 5.5%. In other words, the true incidence of SBIs as defined by Bonadio⁹ is most likely between 0% and 5.5%.¹³ If the true incidence is 0%, most clinicians would stop testing, treating and admitting these neonates; however, if the true incidence is over 5%, many clinicians would probably test, treat and admit these neonates to avoid a 1 in 20 chance of missing an SBI.¹⁴ This analysis suggests that Bonadio's study has too few subjects to determine if neonates who are febrile at home and afebrile on presentation to the ED can be spared a "full septic workup." In our study, we identified 3 E. coli UTIs, suggesting that the incidence of SBIs is not 0% and that testing for UTIs in this group would be prudent.

One major difference between the 2 studies relates to the use of viral testing. In 1987 it was not possible to perform viral testing in any clinically meaningful or timely manner in the ED.^15-21 As viral testing becomes more rapid, more available and less expensive, it will also become more common. Viral infections can present subtly, with only fever and nonspecific signs,^4,22,23 but effective treatments are available for some conditions (e.g., acyclovir for HSV).²⁴ There is growing evidence that viral testing influences physician behaviour,¹⁸ hence the role of specific viral tests in the ED management of neonates will need to be better defined.

Limitations

Our conclusions are not generalizable to all neonates with a history of "fever at home" because we could not identify all such cases. "Possible fever at home" is not reliably recorded in the medical record and is not computer searchable. Therefore, we chose to select subjects by age and based on the performance of a lumbar puncture (reliably recorded and searchable parameters), the latter being a marker for neonates who underwent a full septic workup. This process eliminated neonates who had no testing or limited testing -- those at the "well" end of the spectrum. Unfortunately, because of the retrospective nature of our study, we cannot describe the excluded neonates or their outcomes, and this group may be less likely to harbour clinically significant infections. It is likely that the study subjects, based on having a lumbar puncture, appeared more ill and were more likely to undergo expanded testing and hospitalization regardless of their temperature. This may not be a substantial limitation, however, as the signs and symptoms of illness in neonates can be subtle and fever may be the only unambiguous sign of serious illness at the time of presentation.^4,5,7,22,23

Our sample size is small, which limits the ability to estimate the true incidence of clinically significant infections in the population of interest; nevertheless, our study demonstrates that the incidence is not 0%. Further, our sample size precludes drawing any conclusions about the clinical utility of individual tests.

We included neonates who received antipyretics prior to ED arrival. It is possible that neonates who defervesce after antipyretic administration differ from those who defervesce spontaneously. We overcome this limitation by presenting outcomes for both groups of neonates, so readers can determine whether to "lump" or "split" the data, as they see fit.

We excluded infants whose caregivers could not provide a numeric value for temperature at home because of concerns that some of these neonates were actually afebrile. Although Graneto and Soglin reported a sensitivity of 84% and specificity of 76% for maternal detection of fever by palpation,²⁵ these authors included children up to 10 years of age, and the number of neonates studied is unclear. We are unaware of any study that reports on the parental ability to detect fever by palpation in neonates.

Future areas of study

Although our data suggest that testing for UTIs in this group would be prudent, we cannot conclude that defervescence eliminates the need to look for other serious infections, such as pneumonia, bacteremia or bacterial meningitis. Nor can we identify clinical predictors useful in guiding the use of specific testing modalities such as blood cultures and viral studies. All of these are areas for future prospective study.

Conclusion

Clinically important infections are not uncommon among afebrile neonates undergoing ED septic workup because of a measured fever at home. Some diagnostic testing is warranted in this group, although the clinical utility and indications for specific test modalities remain unclear.

References

1. Baraff LJ, Bass JW, Fleisher GR, Klein JO, McCracken GH, et al. Practice guideline for the management of infants and children 0 to 36 months of age with fever without source. Pediatrics 1993;92:1-12.
2. American College of Emergency Physicians Clinical Policies Committee and the Clinical Policies Subcommittee on Pediatric Fever. Clinical policy for children younger than three years presenting to the ED with fever. Ann Emerg Med 2003;42;530-45.
3. Steere M, Sharieff GQ, Stenklyft PH. Fever in children less than 36 months of age -- questions and strategies for management in the emergency department. J Emerg Med 2003;25:149-57.
4. Baker MD, Bell LM. Unpredictability of serious bacterial illness in febrile infants from birth to 1 month of age. Arch Pediatr Adolesc Med 1999;153:508-11.
5. Kadish HA, Loveridge B, Tobey J, Bolte RG, Corneli HM. Applying outpatient protocols in febrile infants 1-28 days of age: Can the threshold be lowered? Clin Pediatr 2000;39:81-8.
6. Chiu CH, Lin TY, Bullard MJ. Application of criteria identifying febrile outpatient neonates at low risk for bacterial infections. Pediatr Infect Dis J 1994;13:946-9.
7. Ferrera PC, Bartfield JM, Snyder HS. Neonatal fever: utility of the Rochester criteria in determining low risk for serious bacterial infections. Am J Emerg Med 1997;15:299-302.
8. Chiu CH, Lin TY, Bullard MJ. Identification of febrile neonates unlikely to have bacterial infections. Pediatr Infect Dis J 1997;16:59-63.
9. Bonadio WA. Incidence of serious infections in afebrile neonates with a history of fever. Pediatr Infect Dis J 1987;6:911-4.
10. DeAngelis C, Joffe A, Willis E, Wilson M. Hospitalization v. outpatient treatment of young, febrile infants. Am J Dis Child 1983;137:1150-2.
11. Gilbert EH, Lowenstein SR, Koziol-McLain J, Barta DC, Steiner J. Chart reviews in emergency medicine research: Where are the methods? Ann Emerg Med 1996;27:305-8.
12. Jovanovic BD, Zalenski RJ. Safety evaluation and confidence intervals when the number of observed events is small or zero. Ann Emerg Med 1997;30:301-6.
13. Young KD, Lewis RJ. What is confidence? Part 1: The use and interpretation of confidence intervals. Ann Emerg Med 1997;30:307-10.
14. Green SM, Rothrock SG. Evaluation styles for well-appearing febrile children: Are you a "risk minimizer" or a test-minimizer"? Ann Emerg Med 1999;33:211-4.
15. Mackie PL, Joannidis PA, Beattie J. Evaluation of an acute point-of-care screening for respiratory syncytial virus infection. J Hosp Infect 2001;48:66-71.
16. Titus MO, Wright SW. Prevalence of serious bacterial infections in febrile infants with respiratory syncytial virus infection. Pediatrics 2003;112:282-4.
17. Brown L, Reiley DG, Jeng A, Green SM. Bronchiolitis: Can objective criteria predict eligibility for brief hospitalization? Can J Emerg Med 2003;5(4):239-44.
18. Bonner AB, Monroe KW, Talley LI, Klasner AE, Kimberlin DW. Impact of the rapid diagnosis of influenza on physician decision-making and patient management in the pediatric emergency department: Results of a randomized, prospective, controlled trial. Pediatrics 2003;112:363-7.
19. Zbinden R, Kunz J, Schaad UB, Slongo R. Incidence and diagnosis of rotavirus infection in neonates: results of two studies. J Perinat Med 1990;18:363-8.
20. Lipson SM, Svenssen L, Goodwin L, Porti D, Danzi S, Pergolizzi R. Evaluation of two current generation enzyme immunoassays and an improved isolation-based assay for the rapid detection and isolation of rotavirus from stool. J Clin Virol 2001;21:17-27.
21. Kimberlin DW, Lin CY, Jacobs RF, Powell DA, Frenkel LM, Gruber WC, et al. Natural history of neonatal herpes simplex virus infections in the acyclovir era. Pediatrics 2001;108:223-9.
22. Filippine MM, Katz BZ. Neonatal herpes simplex virus infection presenting with fever alone. J Hum Virol 2001;4:223-5.
23. Hall CB, Kopelman AE, Douglas RG Jr, Geiman JM, Meagher MP. Neonatal respiratory syncytial virus infection. N Engl J Med 1979;300:393-6.
24. Kimberlin DW, Lin CY, Jacobs RF, Powell DA, Corey L, Gruber WC, et al. Safety and efficacy of high-dose intravenous acyclovir in the management of neonatal herpes simplex virus infections. Pediatrics 2001;108:230-8.
25. Graneto JW, Soglin DF. Maternal screening of childhood fever by palpation. Pediatr Emerg Care 1996;12:183-4.