PEDIATRICS Vol. 106 No. 2 August 2000, pp. 251-255

Outpatient Pediatric Blood Cultures: Time to Positivity

Karin L. McGowan, PhD^*, Jill A. Foster, MD, and Susan E. Coffin, MD^*

From the ^* Division of Immunologic and Infectious Diseases, Children's Hospital of Philadelphia, Department of Pediatrics, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania; and  Department of Pediatrics, St Christopher's Hospital for Children, MCP Hahnemann University School of Medicine, Philadelphia, Pennsylvania.

	ABSTRACT

Top Abstract Methods Results Discussion References

Objective.  Using a continuously monitoring blood culture system, we determined the time to positivity of blood cultures performed on immunocompetent infants and children who were not receiving antibiotics at the time of culture.

Study Design.  This study was conducted prospectively using blood cultures taken in the emergency department and outpatient clinics of an urban pediatric teaching hospital from February 1, 1993, through December 31, 1996. Cultures were excluded if obtained from patients receiving antibiotics, patients with a central line, patients with prosthetic devices, or those being followed by the oncology division. Our measures included: 1) recording the time to positive culture obtained by using a continuously monitoring blood culture instrument, 2) patient information derived from the hospital computer system concerning antibiotic use and the presence of indwelling central venous catheters and prosthetic devices, and 3) a chart review of 10% of patients from whom positive cultures were obtained.

Results.  During the 47-month study period, 10 200 single bottle blood cultures were obtained, 711 (6.97%) of which became positive. Patients ranged in age from <1 week to 24 years (mean: 2.00 years). Two hundred fifty-eight cultures (36.3%) contained only pathogens, 370 (52%) contained only skin contaminants, and 83 (11.7%) contained a mixture of contaminant and pathogen. Of the 258 cultures containing only pathogens, 14% were positive by 12 hours, 87% by 24 hours, 92% by 36 hours, 95% by 48 hours, 98% by 60 hours, and 99.7% by 72 hours. Ninety-five percent of critical pediatric pathogens including Streptococcus pneumoniae, Salmonella and other Enterobacteriaceae, Neisseria meningitidis, and groups A and B streptococci were detected in <24 hours.

Conclusion.  Because 87% of all cultures containing pathogens were detected within the first 24 hours of incubation, this study can assist emergency department, clinic, and primary care clinicians when making critical decisions concerning patients on whom blood cultures were obtained. Data on time to positivity of blood cultures can be used in conjunction with clinical status to support clinicians in making patient management decisions. Use of short stay (24 hours) or extended care units requiring less patient supervision may be easier to justify when a continuously monitoring blood culture instrument is used in the microbiology laboratory.bacteremia, sepsis.

Fever and bacteremia in children is a common problem in pediatric emergency departments and practices. A wide variety of outpatient strategies for management of febrile children have been developed almost all of which recommend taking blood cultures.^1-5 Over the past 20 years, blood culture methods, media, and systems have improved dramatically. Traditional methods involved manually reading blood culture bottles once a day for evidence of growth and physicians were instructed to wait at least 72 hours (3 readings) before discontinuing antibiotics or discharging patients who had been admitted for observation after a sepsis work-up.⁶ In the 1970s, a number of automated blood culture instruments were developed that monitored the bottles twice a day and decreased the time to detect positive cultures. Studies using such instruments have suggested that if clinically indicated, antibiotic therapy could be discontinued at 48 rather than 72 hours.⁷

Since 1990, 3 manufacturers have developed and marketed continuously monitoring blood culture systems.⁸ These systems electronically monitor blood culture bottles 24 hours a day, typically checking each bottle once every 8 to 10 minutes. To date, all comparative studies have shown that continuously monitoring blood culture instruments detect growth sooner than previous automated systems and conventional manual methods.^9-12 This study evaluates the use of a continuously monitoring blood culture system for patients seen in a large pediatric emergency department and outpatient clinic population. This study was performed to determine the time to detection of positive blood cultures in pediatric patients. In addition, the mean time for detection of common pathogens and contaminants was determined.

	METHODS

Top Abstract Methods Results Discussion References

This study was designed prospectively when the laboratory converted to the BacT/Alert Microbial Detection System (Organon Teknika, Durham, NC) for blood cultures. Questions concerning antibiotic use and the presence of an indwelling central venous catheter or prosthetic devices were incorporated into the hospital computer system such that when blood cultures were ordered, this additional patient information was required to order the test. Data on all blood cultures received by the Clinical Microbiology Laboratory at Children's Hospital of Philadelphia (CHOP) from February 1, 1993, through December 31, 1996, were collected and reviewed for this study and Institutional Review Board approval was not required. CHOP is a tertiary care hospital in an urban setting with a busy emergency department as well as numerous subspecialty and general pediatric outpatient clinics. All positive blood cultures obtained from all patients seen in the emergency department and clinics were identified. Using a standardized form, data concerning microorganisms identified, number of hours from inoculation to positive culture, source of cultured blood (peripheral versus central line), patient age at time of culture, and patient disposition were recorded. Cultures were excluded from analysis if they were obtained from a patient with an indwelling central venous catheter, from a patient with a prosthetic device, or from a patient followed by the oncology division because many of these children were routinely receiving antibiotics when cultures were obtained.

During the study period only pediatric (Pedi-BacT, Organon Teknika, Durham, NC) culture bottles containing supplemented brain heart infusion broth with .02% sodium polyethanol sulfonate were inoculated. A single bottle was used for each blood culture ordered and a maximum of 4 mL of blood was requested but not required. Physicians were not informed of the study to prevent influencing routine practices. Through a pneumatic tube delivery system from the emergency department and clinic building to the laboratories, blood cultures were routinely received in the laboratory within an hour of when they were taken and were immediately loaded into the blood culture instrument. The BacT/Alert system provides continuous agitation of blood culture bottles and monitors the carbon dioxide (CO₂) content within each bottle every 8 to 10 minutes. A positive blood culture was identified when the CO₂ concentration within an individual bottle rose from its predetermined baseline. Bottles identified as positive were immediately removed from the instrument, 24 hours a day, and an aliquot was taken for Gram stain and subculture. Physicians were notified at the time of positive culture and were given the results of the Gram stain. Bacterial isolates were identified by conventional procedures. The time from when a bottle was originally incubated to when it was noted to be positive (in hours) was automatically recorded and the information downloaded monthly for the duration of the study. Bottles were incubated for a total of 6 days before being resulted as negative.

Organisms not usually recognized as pathogens in immunocompetent patients lacking indwelling central venous catheters or prosthetic devices were considered to be probable skin contaminants. This definition of "contaminants" was made with the input and agreement of emergency department and clinic physicians. Isolates defined as contaminants were -hemolytic streptococci, coagulase-negative staphylococci, Corynebacterium species, Micrococcus species, nonpathogenic Neisseria species, and other nonpathogenic Gram-positive rods. Cultures were identified as containing multiple isolates if >1 microorganism was isolated from the same bottle or if 2 populations of the same microorganism with different antibiotic susceptibility patterns were identified. Only cultures from which a single organism was recovered were used to calculate the average time to positivity.

Mantel-Haenszel ² analysis and the Student's t test were used for analysis of categorical and continuous data. Temporal and quantitative trends were evaluated using the ² for trends analysis. Odds ratios (OR) with 95% confidence intervals (CI₉₅) were also calculated for some variables.

	RESULTS

Top Abstract Methods Results Discussion References

During the study period 10 200 blood cultures were performed at CHOP's Clinical Microbiology Laboratory from patients initially evaluated in the emergency department and clinics. Seven hundred eleven blood cultures were positive and of those 347 cultures (48.8%) were performed on patients who were hospitalized at the time of their hospital visit. Patients ranged in age from <1 week to 24 years old (mean: 2.00 years). Chart review of 10% of patients (randomly selected) from whom positive cultures were obtained showed that only 1 of 71 patients reviewed were receiving antibiotics at the time their blood culture was obtained.

Two hundred fifty-eight cultures (36.3%) contained only pathogens, 370 cultures (52.0%) contained only skin contaminants, and 83 cultures (11.7%) contained a mixture of contaminant and pathogen. Of the 258 cultures containing only pathogens, 14% were positive by 12 hours incubation, 87% by 24 hours, 92% by 36 hours, 95% by 48 hours, 98% by 60 hours, and 99.7% by 72 hours.

Two hundred thirty-nine cultures yielded pathogens that were of critical importance to specific populations of pediatric patients namely, S pneumoniae, Salmonella and other Enterobacteriaceae, N meningitidis, and groups A and B streptococci. Within 24 hours, 226 of these cultures (94.6%) had become positive (13 remained undetected [Fig 1]). By 48 hours, 98.3% of cultures containing these critical pediatric pathogens had become positive (only 4 remained undetected).

View larger version (11K): [in this window] [in a new window]   Fig. 1.   Time to detect pediatric pathogens in 239 positive blood cultures. Pathogens included: S pneumoniae, Salmonella and other Enterobacteriaceae, N meningitidis, and groups A and B streptococci.

A total of 441 contaminants were isolated. Blood cultures obtained from children <1 year of age were more likely to contain only contaminants as compared with cultures obtained from older children (60.2% vs 42.6%; OR: 2.04; CI₉₅: 1.50-2.79). Hospital admission was not associated with an increased risk of contamination. Contaminants only were isolated from 181 of 347 blood cultures (52.2%) obtained from children who were admitted.

Blood cultures containing only contaminants became positive more slowly than cultures that contained pathogens (32.8 vs 18.4 hours; P < .0001). Stratified analysis demonstrated that cultures containing only contaminants became positive more slowly from children of all ages except those 30 to 36 months old (Table 1). Most pathogens were isolated from culture more rapidly than contaminants (Table 2). Gram-positive pathogens were isolated from culture more rapidly than Gram-positive contaminants (15.5 vs 25.0 hours; P < .0001).

Of 309 cultures obtained from children <3 years old who required hospitalization, 165 (53.4%) contained only contaminants. Among young children requiring admission, cultures that contained only contaminants were significantly slower to become positive than cultures that contained pathogens (27.0 vs 19.9 hours; P < .0001).

Eight hundred nine (809) bacterial isolates were recovered from 711 blood cultures with 83 blood cultures (11.7%) yielding multiple isolates. One hundred forty-four (144) of 181 isolates (79.8%) identified from bottles containing multiple isolates were contaminants. Cultures containing multiple organisms were more likely to contain only contaminants than cultures containing a single isolate (66.3% vs 50.2%; OR: 1.95; CI₉₅: 1.17-3.27). There was no association between recovery of multiple isolates and the time to a positive culture. Multiple isolates were not recovered more frequently from blood cultures obtained from younger children (data not shown).

	DISCUSSION

Top Abstract Methods Results Discussion References

Here we report on the largest blood culture study to date involving infants and children from both emergency department and clinic settings. The major focus of this study was to determine the time required for a blood culture to become positive and the time after which a culture could safely be considered negative in this population. Such information can be used in conjunction with clinical status to assist in making patient management decisions regarding the ongoing need for antibiotic therapy or, in some cases, duration of hospitalization. Time to positivity has largely been studied in neonatal populations where most infants are receiving antibiotics and require intravenous lines for fluids and medications.^6,7,13-15 This study, in contrast, involved an immunocompetent population not receiving antibiotics and lacking indwelling central venous catheters or prosthetic devices. Our results may be more useful to clinicians who work in emergency departments, outpatient clinics, or office settings.

This study confirms results from earlier investigations in that the majority (95%) of significant positive cultures were detected by 48 hours and almost all (99.7%) were detected by 72 hours. Eighty-seven percent of all cultures containing pathogens were detected within 24 hours' incubation. In addition, 95% of cultures containing pathogens critical for specific patient populations such as S pneumoniae, Salmonella and other Enterobacteriaceae, N meningitidis, and groups A and B streptococci were positive in <24 hours. Such findings have tremendous potential for influencing patient management. The rapid time to positive culture shown in this study theoretically improves the opportunity to call back patients with positive cultures before complications might occur. Clearly the most important factor in making decisions about admission and/or discharge of patients from the hospital is the clinical assessment of the patient. However, pediatricians and other clinicians caring for febrile children may find this time to positive culture data useful in patient management. Febrile children including those at higher risk of bacteremia, such as patients <8 weeks old and those with sickle cell disease, could potentially be admitted to short stay or extended care units for a 24-hour observation period. This would help optimize emergency department efficiency and provide an alternative to hospitalizations. Equally important is the data showing that, in almost every age group, blood cultures containing likely skin contaminants became positive much more slowly than those that contained pathogens (32.8 vs 18.4 hours; P < .0001). Discharge after 24 hours of observation could be considered for patients with a benign clinical appearance and negative blood cultures. Clinically important pathogens that would have been overlooked if one assumed that only contaminants were identified after 48 hours were Neisseria gonorrhoeae (n = 1) and Campylobacter spp (n = 2).

This study also revealed some disturbing information concerning blood cultures made positive by contaminants. Our definition of a contaminant for this study was determined after consulting with emergency department and clinic physicians. The organisms they agreed to call contaminants were bacteria they did not consider pathogens for their specific outpatient populations and would not consider treating with antibiotics. During the study, 370 cultures (52.0% of all positives) contained skin contaminants. In all cases, if a child had not been initially admitted, an attempt was made to recall them for evaluation and repeat blood cultures. This represents a tremendous waste of physician time and laboratory resources. Most laboratories monitor the overall contamination rate of blood cultures performed and strive to maintain an average of 2% to 4%. The overall contamination rate for the 10 200 blood cultures taken during this study was 3.6%, but we believe such numbers are deceiving considering the cost of a repeat examination and blood culture. During this study, physicians and nurses, rather than specifically trained phlebotomists were responsible for obtaining blood cultures in outpatient settings. Other studies have shown that reduced contamination rates can be achieved by using phlebotomists in such settings.¹⁶

Our study had 2 possible limitations. No attempt was made to control the volume of blood placed in each bottle. Many studies exclude cultures that fail to meet minimum blood volume criteria. Our aim was to reflect standard daily practice at our institution. For this reason, cultures were not rejected for low blood volume despite the fact that volume of blood directly effects time to positivity. The second concern was our assumption that the patients were not on antibiotics at the time the blood culture was taken. A question in the hospital computer requiring clinicians to respond if antibiotics were present plus our exclusion of patients with indwelling central venous catheters, prosthetic devices, and those followed by oncology, in our opinion helped to limit the possibility of antibiotics being present. In addition, our random review of 71 charts revealed only 1 incident where the computer questionnaire and patient chart did not agree on this point. In that case, the patient was receiving amoxicillin for presumed otitis and the blood culture grew nontypeable Haemophilus influenzae, which was -lactamase-positive. Although less than ideal, we believe most patients were not on antibiotics at the time blood was obtained for culture.

Current practice guidelines recommend that blood cultures be incubated for a total of 7 days before they can be reported as negative.⁸ This recommendation attempts to allow for all conditions including blood cultures taken from patients receiving antibiotics and those in whom endocarditis is suspected. The only positive blood cultures requiring incubation longer than 4 days in this study yielded Corynebacterium spp that were considered contaminants by clinicians. In hospitals using continuously monitoring blood culture systems, clinicians should be made aware of the lower times necessary for detecting pathogens in blood cultures, particularly with outpatients who are not receiving antibiotics at the time of culture.

	ACKNOWLEDGMENTS

We thank the microbiology laboratory, nursing, and medical staff of Children's Hospital of Philadelphia for their assistance and support during the study period and Kathy Shaw, MD, for her insightful comments.

	FOOTNOTES

Received for publication Aug 24, 1999; accepted Dec 16, 1999.

Reprint requests to (K.L.M.) Children's Hospital of Philadelphia, 3400 Civic Center Blvd, Room 5060A, Philadelphia, PA 19104. E-mail: mcgowan{at}email.chop.edu

	ABBREVIATIONS

CHOP, Children's Hospital of Philadelphia; CO₂, carbon dioxide; OR, odds ratios; CI₉₅, 95% confidence intervals.

	REFERENCES

Top Abstract Methods Results Discussion References

1. Lieu TA, Baskin MN, Schwartz JS, Fleisher GR Clinical and cost-effectiveness of outpatient strategies for management of febrile infants. Pediatrics. 1992; 89:1135-1144 [Abstract/Free Full Text]
2. Downs SM, McNutt RA, Margolis PA Management of infants at risk for occult bacteremia: a decision analysis. J Pediatr. 1991; 118:11-20 [CrossRef][Medline]
3. Baraff LJ Management of the febrile child: a survey of pediatric and emergency medicine residency directors. Pediatr Infect Dis J. 1991; 10:795-800 [Medline]
4. Baker MD, Bell LM, Avner JR Outpatient management without antibiotics of fever in selected infants. N Engl J Med. 1993; 329:1437-1441 [Abstract/Free Full Text]
5. Kline MW Management of occult bacteremia in children. Sem Pediatr Infect Dis. 1990; 1:333-339
6. Freij B, Nelson JD. Neonatal septicemia, meningitis, and pneumonia. In: Gellis SG, Kagan BM, eds. Current Pediatric Therapy. 12th ed. Philadelphia, PA: WB Saunders Co; 1986:497-500
7. Rowley AH, Wald ER Incubation period necessary to detect bacteremia in neonates. Pediatr Infect Dis. 1986; 5:590-591 [CrossRef][Medline]
8. Reimer LG, Wilson ML, Weinstein MP Update on detection of bacteremia and fungemia. Clin Microbiol Rev. 1997; 10:444-465 [Abstract]
9. Weinstein MP Current blood culture methods and systems: clinical concepts, technology, and interpretation of results. Clin Infect Dis. 1996; 23:40-46 [Medline]
10. Wilson ML, Weinstein MP, Reimer LG, Mirrett S, Reller LB Controlled comparison of the BacT/Alert and BACTEC 660/730 nonradiometric blood culture systems. J Clin Microbiol. 1992; 30:323-329 [Abstract/Free Full Text]
11. Krishner KK, Whyburn DR, Koepnick FE Comparison of the BacT/Alert pediatric blood culture system, Pedi-BacT, with conventional culture using the 20-milliliter Becton-Dickinson supplemented peptone broth tube. J Clin Microbiol. 1993; 31:793-797 [Abstract/Free Full Text]
12. Welby PL, Keller DS, Storch GA Comparison of automated Difco ESP blood culture system with biphasic BBL Septi-Chek system for detection of bloodstream infections in pediatric patients. J Clin Microbiol. 1995; 33:1084-1088 [Abstract]
13. Kurlat I, Stoll BJ, McGowan JE Jr Time to positivity for detection of bacteremia in neonates. J Clin Microbiol. 1989; 27:1068-1071 [Abstract/Free Full Text]
14. Pichichero ME, Todd JK Detection of neonatal bacteremia. J Pediatr. 1979; 94:958-960 [CrossRef][Medline]
15. Hurst MK, Yoder BA Detection of bacteremia in young infants: is 48 hours adequate? Pediatr Infect Dis J. 1995; 14:711-712 [Medline]
16. Weinbaum FI, Lavie S, Danek M, Sixsmith D, Heinrich GF, Mills SS Doing it right the first time: quality improvement and the contaminated blood culture. J Clin Microbiol. 1997; 35:563-565 [Abstract]