PEDIATRICS Vol. 108 No. 2 August 2001, pp. 520-521

Occult Bacteremia From a Pediatric Emergency Department

To the Editor.

Regarding the report of Alpern et al,¹ given the low rate of adverse outcome to Streptococcus pneumoniae occult bacteremia and the risks associated with false-positives and hospitalizations for positive blood cultures, one should conclude that there is no utility to blood cultures in this setting and that close follow-up without blood cultures is more valuable.

John DiTraglia, MD
Portsmouth, OH 45662

REFERENCE

1. Alpern ER, Alessandrini EA, Bell LM, Shaw KN, McGowan KL Occult bacteremia from a pediatric emergency department: current prevalence, time to detection, and outcome. Pediatrics. 2000; 106:505-511 [Abstract/Free Full Text]

In Reply.

We would like to thank Dr DiTraglia for his thoughtful comment concerning our recently published article.¹ As pointed out in his letter, the main findings of our study highlight the low prevalence of occult bacteremia and subsequent very low risk of serious adverse outcomes. We agree that the high risk of contaminated blood cultures and subsequent medical care to evaluate patients at risk for occult bacteremia is concerning. Close follow-up is incontestably agreed upon in the care of patients at risk for occult bacteremia. In the emergency department (ED) setting, however, this follow-up is often difficult. Therefore, many practitioners may elect not to eliminate obtaining blood cultures. As we concluded in our study, a continuously monitored blood culture system may allow for early differentiation between contaminated and pathogenic cultures. The early recognition (94% within 18 hours) of true pathogenic cultures with the continuously monitored system may also allow for early and important follow-up of children with "known" occult bacteremia. This "red flag" may help focus urgent and immediate follow-up in the ED setting. Each practitioner should base medical practice on a clear understanding of the blood culture system available to his or her practice. As Dr DiTraglia points out, there is no substitute for close follow-up of patients at risk for occult bacteremia.

Elizabeth R. Alpern, MD, MSCE
Evalene A. Alessandrini, MD, MSCE
Louis M. Bell, MD
Kathy N. Shaw, MD, MSCE
Karin L. McGowan, PhD
Department of Pediatrics
University of Pennsylvania
Children's Hospital of Philadelphia
Philadelphia, PA 19104-4399, USA

REFERENCE

1. Alpern ER, Alessandrini EA, Bell LM, Shaw KN, McGowan KL Occult bacteremia from a pediatric emergency department: current prevalence, time to detection, and outcome. Pediatrics. 2000; 106:505-511

To the Editor.

I read with interest the article by Dr Alpern and colleagues on occult bacteremia in the pediatric emergency department (PED).¹ Their carefully collated experience reinforces earlier reports showing that occult bacteremia, along with its complications, continues to occur, although disease attributable to Haemophilus influenzae type b (Hib) has essentially disappeared.² I agree with the authors that most pediatricians do not encounter these complications very commonly. I believe, however, that several issues in this study merit clarification and/or further comment.

1. The authors report a prevalence of non-Hib occult bacteremia of 1.9% (95% confidence interval [CI]: 1.5%, 2.3%) in young febrile children evaluated in the PED, which is similar to that of previous large prospective studies on this topic (1.6%² and 2.7%³). The prevalence of occult bacteremia reported in the current study, however, likely underestimates the true prevalence in their population. This is because patients who were already receiving antibiotics at the time of blood culture, as well as those recently immunized, were not excluded from enrollment, in contrast to earlier studies.^3,4
2. The rate of adverse outcome (meningitis or death) of children with bacteremia in this study was 2 of 104 (1.8%; 95% CI: 0, 6.8%). Although the authors report that this is an "extremely low risk" in their conclusions, the 95% CI of this rate is consistent with the rates of development of pneumococcal meningitis reported in 2 large meta-analyses on the topic (2.7%⁵ and 5.8%⁶). Many clinicians would not necessarily consider this rate of meningitis to be low given the gravity of this complication. In the current study, however, 69 (66%) of 104 patients with occult bacteremia were treated with oral antibiotics, which may have diminished the rate of complications. The authors state that "a prior randomized clinical trial demonstrated that oral antibiotics do not affect the incidence of serious focal infections." The prospective study they reference, however, was not appropriately powered to make this conclusion.⁴ Furthermore, in the meta-analysis regarding empirical oral antibiotic treatment of children with occult pneumococcal bacteremia that the authors reference, significantly fewer serious bacterial complications developed in treated than untreated children.⁵ In addition, in that study only 3 (0.8%) of 399 patients with bacteremia who received oral antibiotics developed meningitis compared with 7 (2.7%) of 257 patients who did not receive empirical antibiotics, although this difference did not achieve statistical significance (the reported odds ratio [OR] for that comparison was 0.51, with a 95% CI of 0.12-2.09).⁵ From that study, one cannot conclude that oral antibiotics do not prevent S pneumoniae meningitis, because an eightfold reduction in the odds of developing meningitis for patients receiving oral antibiotics cannot be excluded (i.e., OR: 0.12, the lower end of the 95% CI). Two other large studies on this topic report that oral antibiotics decrease the rate of focal complications in children with occult pneumococcal bacteremia⁷ and decrease the rate of development of meningitis.⁶
3. The authors conclude that, using their data, one would need to treat >3000 febrile children with empirical antibiotics to prevent one adverse outcome. The reader should be aware, however, that this estimate is based on the assumption of treating all young febrile patients with empirical antibiotics, a practice that is not condoned by many. Were one to use a screening approach using the white blood cell count² or absolute neutrophil count,⁸ however, this number needed-to-treat would be greatly reduced (to approximately 250-400 to potentially prevent 1 case of meningitis).

Serious adverse outcomes of occult bacteremia are uncommon, but should be taken seriously as these complications not infrequently lead to disability and, on occasion, death. Debate about the approach and treatment of the febrile child at risk for occult bacteremia is appropriate. Furthermore, few would argue that an overly aggressive approach to screening and treatment leads to overuse of antibiotics, while withholding antibiotics from children with bacterial pathogens in the bloodstream represents a possible failure to prevent serious complications. Dr Alpern and colleagues provide new data which will prove useful in refining the probabilities used by those who favor formal cost-effectiveness analysis. These new numbers, however, fall sufficiently close to previously published findings that I suspect that recalculating the outcomes will not change the mind of many physicians, most of whom have already taken a position based on practice style.⁹ Fortunately, the arrival and dissemination of an effective pneumococcal protein conjugate vaccine will render the arguments moot; I look forward to it.

Nathan Kuppermann, MD
Department of Pediatrics and Emergency Medicine
University of California at Davis Medical Center
Sacramento, CA 95817

REFERENCES

1. Alpern ER, Alessandrini EA, Bell LM, Shaw KN, McGowan KL Occult bacteremia from a pediatric emergency department: current prevalence, time to detection, and outcome. Pediatrics. 2000; 106:505-511
2. Lee GM, Harper MB Risk of bacteremia for febrile young children in the post-Haemophilus influenzae type b era. Arch Pediatr Adolesc Med. 1998; 152:624-628 [Abstract/Free Full Text]
3. Fleisher GR, Rosenberg N, Vinci R, Intramuscular versus oral antibiotic therapy for the prevention of meningitis and other bacterial sequellae in young, febrile children at risk for occult bacteremia. J Pediatr. 1994; 124:504-512 [CrossRef][Medline]
4. Jaffe DM, Tanz RR, Davis AT, Henretig F, Fleisher G Antibiotic administration to treat possible occult bacteremia in febrile children. N Engl J Med. 1987; 317:1175-1180 [Abstract]
5. Rothrock SG, Harper MB, Green SM, Do oral antibiotics prevent meningitis and serious bacterial infections in children with Streptococcus pneumoniae occult bacteremia? A meta-analysis. Pediatrics. 1997; 99:438-444 [Abstract/Free Full Text]
6. Baraff LJ, Oslund S, Prather M Effect of antibiotic therapy and etiologic microorganism on the risk of bacterial meningitis in children with occult bacteremia. Pediatrics. 1993; 92:140-143 [Abstract/Free Full Text]
7. Harper MB, Bachur R, Fleisher GR Effect of antibiotic therapy on the outcome of outpatients with unsuspected bacteremia. Pediatr Infect Dis. 1995; 14:760-767
8. Kuppermann N, Fleisher GR, Jaffe DM Predictors of occult pneumococcal bacteremia in young febrile children. Ann Emerg Med. 1998; 31:679-687 [CrossRef][Medline]
9. Green SM, Rothrock SG Evaluation styles for well-appearing febrile children: are you a "risk-minimizer" or a "test-minimizer"? Ann Emerg Med. 1999; 33:211-214 [CrossRef][Medline]

In Reply.

We appreciate Dr Kuppermann's letter following our article¹ and his comments concerning the very uncommon nature of complications from occult bacteremia (OB).

Dr Kuppermann correctly points out that we did not exclude patients from the study cohort due to a history of recent antibiotic use or immunization. Inclusion of these patients is similar to that of the most recent large study on this topic² and we think, by closely approximating true clinical practice, represents the most clinically relevant denominator of patients considered at risk for OB. A previous study concluded that recent vaccination does not affect the rate of OB.³ As stated in our article, only 8% of patients in our cohort were noted to have had recent antibiotic use before evaluation for OB. Subsequent analysis excluding these patients results in an overall rate of OB of 2.0% (95% CI: 1.7%, 2.5%), essentially the same as the overall risk of the entire cohort of 1.9% (95% CI: 1.5%, 2.3%).

As reported, the rate of adverse outcome (meningitis or death) in children evaluated for OB in our study was 2 of 5901 (0.03%; 95% CI: 0.004%, 0.12%). It is important to point out that the true denominator for evaluation of risk of adverse outcomes attributable to OB is the number of patients "at risk" of OB as they walk through your practice door (in our cohort that number was 5901). Only using hindsight are we able to determine the risk of adverse outcome in children with identified bacteremia. The rate of adverse outcome in children known to have bacteremia was 2 of 111 (1.8%; 95% CI: 0.2%, 6.4%). Although oral antibiotics were prescribed at the first visit for 66% of those reevaluated with repeat blood cultures, there was no statistical difference in the rate of persistent bacteremia between patients prescribed antibiotics and those not. Interestingly, both patients with "serious adverse outcomes" in our study had been prescribed oral antibiotics at initial evaluation. Kuppermann rightly points out that the study by Jaffe et al⁴ and the meta-analysis by Rothrock et al⁵ were both underpowered to show a significant effect of oral antibiotics in preventing meningitis in patients with OB. These well-designed and often-cited studies were unable to show a statistically significant difference attributable to the very rare outcome of meningitis. One of the 2 studies that is mentioned in the letter that reports a decrease in meningitis with antibiotic treatment is the meta-analysis by Baraff et al.⁶ The limitations of this study (lack of clearly defined inclusion criteria, lack of concurrent treatment and nontreatment groups, inclusion of patients not meeting accepted OB criteria, and double-counting some patients) has been delineated elsewhere.⁵ The other study cited in defense of oral antibiotic treatment included patients 3 weeks to 19 years of age, did not exclude patients with temperatures <39°C or who were ill enough at initial visit to undergo lumbar puncture, and was severely limited in power attributable to the rare outcome of meningitis.⁷

Children in our study were not routinely screened with complete blood counts because of the limitations of this test. The positive predictive value of a white blood cell count (WBC) 15 000/mm³ is 5.1.² Therefore, if we expectantly treated 100 children at risk for OB with WBC 15 000/mm³, we would expose 95 of these children to unnecessary antibiotics. Again, the limitation of this test as a screening tool is attributable to the very low incidence of OB.

We agree with Dr Kuppermann that the debate surrounding the diagnosis and treatment of OB is important as we try to improve the care provided to our patients. With the recent licensure of the heptavalent pneumococcal conjugate vaccine, we all look forward to the diminished prevalence of OB and occurrence of serious complications from this disease.

Elizabeth R. Alpern, MD, MSCE
Evalene A. Alessandrini, MD, MSCE
Louis M. Bell, MD
Kathy N. Shaw, MD, MSCE
Karin L. McGowan, PhD
Department of Pediatrics
University of Pennsylvania
Children's Hospital of Philadelphia
Philadelphia, PA 19104-4399

REFERENCES

1. Alpern ER, Alessandrini EA, Bell LM, Shaw KN, McGowan KL Occult bacteremia from a pediatric emergency department: current prevalence, time to detection, and outcome. Pediatrics. 2000; 106:505-511
2. Lee GM, Harper MB Risk of bacteremia for febrile young children in the post-Haemophilus influenzae type b era. Arch Pediatr Adolesc Med. 1998; 152:624-628
3. Burstein JL, Fleisher GR Does recent vaccination increase the risk of occult bacteremia? Pediatr Emerg Care. 1994; 10:138-140 [CrossRef][Medline]
4. Jaffe DM, Tanz RR, Davis AT, Henretig F, Fleisher G Antibiotic administration to treat possible occult bacteremia in febrile children. N Engl J Med. 1987; 317:1175-1180
5. Rothrock SG, Harper MB, Green SM, Do oral antibiotics prevent meningitis and serious bacterial infections in children with Streptococcus pneumoniae occult bacteremia? A meta-analysis. Pediatrics. 1997; 99:438-444
6. Baraff LJ, Oslund S, Prather M Effect of antibiotic therapy and etiologic microorganism on the risk of bacterial meningitis in children with occult bacteremia. Pediatrics. 1993; 92:140-143
7. Harper MB, Bachur R, Fleisher GR Effect of antibiotic therapy on the outcome of outpatients with unsuspected bacteremia. Pediatr Infect Dis J. 1995; 14:760-767 [Medline]