- *Golisano Children’s Hospital at Strong Memorial Hospital, Department of Pediatrics, University of Rochester Medical Center, Rochester, NY.
- †Director of Hospital Medicine at Children's Hospital at Dartmouth, Hanover, NH.
Drs Biondi, Murzycki, Ralston, and Gigliotti have disclosed no financial relationships relevant to this article. This commentary does not contain a discussion of an unapproved/investigative use of a commercial product/device.
Is it necessary to monitor blood cultures for a full 48 hours to consider a febrile infant to be free of serious bacterial infection?
Fever that could be due to bacteremia in infants is a common and serious problem faced by pediatricians. Because of their immature immune systems and nonspecific signs, infants are more susceptible to serious bacterial infections (SBIs) than adults. A wide variety of strategies have been developed for the assessment of febrile but well-appearing children younger than age 90 days, known collectively as the “rule out sepsis” evaluation. This clinical presentation is a common reason for infant hospitalization. However, it remains unclear how long blood cultures from this population need to be monitored and how long infants need to be hospitalized. The literature is inconclusive, and there is variability in current practice.
A review of currently published data did not identify a clear-cut observational period. Randhawa et al suggest a 96-hour observational period, although it was not obvious from this study what type of culture detection system was used or whether the patients were symptomatic. (1) In 2011, Guerti et al demonstrated support for an evaluation period of more than 36 hours. (2) They found that 56% of blood cultures were positive for bacteremia after an incubation time of 24 hours, 89% after 48 hours, and 97% after 72 hours. Another large study by Kumar et al demonstrated similar results. (3) Interestingly, if all coagulase-negative Staphylococcus are removed in these studies, the results become 79%, 93%, and 96% and 77%, 93%, and 97%, respectively.
An older study by Garcia-Prats et al using previous versions of the automated blood culture detection systems demonstrated that 77% of blood cultures will turn positive within 24 hours. (4) Jardine et al concluded that 25% of pathogenic bacteria will grow after 24 hours in children older than 2 days. (5) However, because of the rarity of positive blood cultures in infants, the vast majority of positive culture data in these studies came from infants admitted to intensive care units. This cohort of patients is vastly different from that of the patient population in which the typical “rule out” is performed and is at risk for different pathogens that often require longer incubation times, such as coagulase-negative Staphylococcus and Candida.
Those who support shorter observation times may cite the Byington et al article that states, “approximately 90% of bacterial pathogens are identified within the first 24 hours of incubation.”(6) These data, however, are actually derived from three other studies. The first is the McGowan et al study that examined blood cultures from outpatient pediatric patients aged 0 to 18 years, with a mean age of 2 years, at a single center from 1993 to 1996. (7) These patients are not our target population, and these data conflict with the aforementioned studies done in infants. The second study is the Kumar et al article mentioned above.
The third study, by Kaplan et al, did not include infants younger than age 28 days and used a now-obsolete culture detection system for a majority of the study cultures. (8) Although these studies are certainly of high quality, they demonstrate not only that there is a lack of consensus within the literature, but also that the available literature does not assess our specific target population: the nontoxic, febrile infant.
Commonly used medical textbooks generally support a 48-hour observational period, but none was found to be firm in its conclusion. The most recent edition of Comprehensive Pediatric Hospital Medicine reads “…no clear consensus exists regarding the optimal approach to febrile infants younger than 90 days.” It recommends 48-hour inpatient evaluations for children younger than age 28 days, while stating that earlier discharge may be indicated for older infants. This source states further that clinicians must determine the degree of diagnostic uncertainty they are willing to tolerate, understand their institutional or regional practice variations, and know that risk minimization may influence practice patterns. (9)
Infectious Diseases of the Fetus and Newborn Infant, published in 2011, also recommends a 48-hour period of evaluation. (10) The lack of guidance from the most recent editions of many popular textbooks such as Red Book, (11) The Nelson Textbook of Pediatrics, (12) and The Harriett-Lane Handbook (13) is notable. Although these texts do discuss at length the details of various evaluation techniques and approaches to the febrile infant, our reviews did not identify specific recommendations regarding length of inpatient observation. We feel that this finding speaks to the paucity of conclusive data on the subject.
The inconclusive nature of the primary and secondary literature regarding the topic of length of admission for our patient population has hindered a nationally standardized approach. Local institutions are likely to be operating independently on this issue. Informal discussions with colleagues at academic centers and community hospitals indicate that most hospitals use a 48-hour observation period while awaiting blood culture results. However, there is significant variability depending on the institutional and regional preferences. As is often said, “All politics and microbiology is local.”
We recently performed an institutional review board–approved pilot study that examined all positive blood cultures since January 2009 of infants less than age 91 days at the University of Rochester. The microbiology laboratory identified 515 positive blood cultures identified via the BacT/ALERT detection system® (bioMérieux, Inc, Durham, NC). After filtering out critically ill and postoperative cardiac patients and any culture that was determined to be a contaminant, we were left with only 14 cultures from 10 infants. Of these 14 cultures, none took longer than 15 hours to become positive. Notably, Escherichia coli was identified as the most common pathogen, and we did not uncover a single case of Listeria infection.
These results suggest that the current practice of inpatient observation for 48 hours pending blood culture results may be unnecessary. However, even a much larger study done by a single center, with its own local microbiology, may prove insufficient to change physician behavior at the national level.
In July 2012, a large, observational study done within the Intermountain Healthcare System reported outcomes from the use of an evidence-based clinical pathway for febrile infants. (14) For admitted infants, the pathway bases the duration of length of stay on results of bacterial and viral testing at 24 hours. Admitted culture-negative infants who are classified as high risk for SBI by the Rochester criteria (15) and who test positive for a viral pathogen or who are at low risk for SBI are eligible for discharge at 24 hours. All other culture-negative infants are eligible for discharge at 36 hours. The study reported no cases of missed SBI, and the readmission rate remained stable after implementation of the pathway. (14) This is the largest study to date examining the evaluation of febrile infants, and these data suggest that nontoxic, febrile infants can be discharged safely by 36 hours.
However, although this was a well-powered study, it remains isolated to one region of the country and was designed to report outcomes of an evidence-based care plan, not to determine time to positivity of infantile blood cultures. Although the pathway appears to be safe, it may not operate at the optimal length of stay. In fact, because this protocol is based on evidence that does not specifically examine the nontoxic infant, one could argue hypothetically that it might be safe to discharge patients even earlier. Of course, the current literature does not support discharge before 24 hours; but a well-powered, national study examining time-to-positivity of blood cultures in this unique target population could prove more definitive with regard to the determination of optimal length of stay.
A recent institutional review board–approved survey of 57 pediatric hospitalists performed by the authors (E.B., S.R.) shows that 48 of these clinicians (84%) continue to favor a standard inpatient observation period longer than 36 hours for infants younger than age 30 days, with widely varying practices for infants who are 30 to 90 days old. At the present time, there is no strong evidence indicating a need to change the conventional 48-hour waiting period.
- Guerti K,
- Devos H,
- Ieven MM,
- Mahieu LM
- Kumar Y,
- Qunibi M,
- Neal TJ,
- Yoxall CW
- Garcia-Prats JA,
- Cooper TR,
- Schneider VF,
- Stager CE,
- Hansen TN
- Byington CL,
- Enriquez FR,
- Hoff C,
- et al
- McGowan KL,
- Foster JA,
- Coffin SE
- Zaoutis LB,
- Chiang VW
- Remington JS
- Pickering LK
- Kliegman RM,
- Stanton BMD,
- St. Geme J,
- Schor N,
- Behrman RE
- 13.↵Johns Hopkins Hospital, Arcara K, Tschudy M. The Harriet Lane Handbook. 19th ed. Philadelphia, PA: Elsevier; 2012
- Byington CL,
- Reynolds CC,
- Korgenski K,
- et al
- Jaskiewicz JA,
- McCarthy CA,
- Richardson AC,
- et al
- © American Academy of Pediatrics, 2013. All rights reserved.