StATS: Another search for evidence (June 7, 2005).

Our new EBM discussion group is asking everyone to research the following question:

In a patient (2 months 2 years of age) with a fever > 38.5 (rectal) at what CRP level should empiric antibiotic therapy be initiated?

This question was later clarified to

In an otherwise healthy pediatric patient between the ages of 2 months and 2 years) with a fever of 38.5 rectal or greater, at what level of CRP should one be concerned with a bacterial infection.

I had to confirm that CRP was C-reactive protein. Here's one possible search strategy. All searches had a limit of ages to Infant: 1-23 months.

  1. C-reactive protein (788)
  2. CRP (353)
  3. #1 OR #2 (859)
  4. bacterial infections (36,798)
  5. fever (9,414)
  6. #3 AND #5 (226)
  7. #6 AND #4 (106)
  8. #7 AND diagnosis (84)
  9. #8 AND meta-analysis [pt] (no items found)
  10. #8 AND otherwise healthy (no items found)

I went back and looked at search #8 again. There were 5 review articles among the 84, so I looked at that list:

  1. Hsiao AL, Baker MD. Fever in the new millennium: a review of recent studies of markers of serious bacterial infection in febrile children. Curr Opin Pediatr. 2005 Feb;17(1):56-61. Review. PMID: 15659965
  2. van Rossum AM, Wulkan RW, Oudesluys-Murphy AM. Procalcitonin as an early marker of infection in neonates and children. Lancet Infect Dis. 2004 Oct;4(10):620-30. Review. PMID: 15451490
  3. Kuppermann N. Occult bacteremia in young febrile children. Pediatr Clin North Am. 1999 Dec;46(6):1073-109. Review. PMID: 10629675 [PubMed - indexed for MEDLINE]
  4. Nudelman R, Kagan BM. C-reactive protein in pediatrics. Adv Pediatr. 1983;30:517-47. Review. PMID: 6369943
  5. [No authors listed] Rheumatic fever. Arthritis Rheum. 1970 Sep-Oct;13(5):461-72. Review. No abstract available. PMID: 4921514

The first article reviewed six studies, listed below:

  1. Carrol ED, Newland P, Riordan FA, Thomson AP, Curtis N, Hart CA. Procalcitonin as a diagnostic marker of meningococcal disease in children presenting with fever and a rash. Arch Dis Child. 2002 Apr;86(4):282-5. PMID: 11919107
  2. Fernandez Lopez A, Luaces Cubells C, Garcia Garcia JJ, Fernandez Pou J; Spanish Society of Pediatric Emergencies. Procalcitonin in pediatric emergency departments for the early diagnosis of invasive bacterial infections in febrile infants: results of a multicenter study and utility of a rapid qualitative test for this marker. Pediatr Infect Dis J. 2003 Oct;22(10):895-903. PMID: 14551491
  3. Galetto-Lacour A, Zamora SA, Gervaix A. Bedside procalcitonin and C-reactive protein tests in children with fever without localizing signs of infection seen in a referral center. Pediatrics. 2003 Nov;112(5):1054-60. PMID: 14595045
  4. Gendrel D, Raymond J, Coste J, Moulin F, Lorrot M, Guerin S, Ravilly S, Lefevre H, Royer C, Lacombe C, Palmer P, Bohuon C. Comparison of procalcitonin with C-reactive protein, interleukin 6 and interferon-alpha for differentiation of bacterial vs. viral infections. Pediatr Infect Dis J. 1999 Oct;18(10):875-81. PMID: 10530583
  5. Lacour AG, Gervaix A, Zamora SA, Vadas L, Lombard PR, Dayer JM, Suter S. Procalcitonin, IL-6, IL-8, IL-1 receptor antagonist and C-reactive protein as identificators of serious bacterial infections in children with fever without localising signs. Eur J Pediatr. 2001 Feb;160(2):95-100. PMID: 11271398
  6. Pulliam PN, Attia MW, Cronan KM. C-reactive protein in febrile children 1 to 36 months of age with clinically undetectable serious bacterial infection. Pediatrics. 2001 Dec;108(6):1275-9. PMID: 11731648

It turns out that all of these six articles were among the 84 studies in #8.

I have not yet read the six individual articles. The review article was a bit disappointing. They didn't discuss or use likelihood ratios, so I had to calculate them myself. They didn't mention how they identified the research studies. I suspect that they may have excluded non-English language publications. There was no explicit statement about conflict of interest.

It looks like CRP is a fair test, but not great. Excluding one extreme cutpoint, the LR+ ranged between 3.6 and 8.8 and the LR- ranged between 0.14 and 0.43.

Procalcitonin looked a bit better. One study showed likelihood ratios of 12 and 0.18.

The studies seemed to have prevalences hovering around 20. With a LR+ of 9, the post-test probability is 70% and with a LR of 12, the post-test probability is 75%. With a LR- of 0.14, the post-test probability is 3%. With a LR- of 0.18, the post-test probability is 4%.

As far as my search strategy goes, I was disappointed not to get a list down to less than 84, and if I had not found the review article, I would have had a hard time going through such a long list of studies. One thought I had after the fact was that perhaps I should have requested that C-reactive protein be in the title itself. All of the six studies listed above had either C-reactive protein of procalcitonin in the title.

We'll see what the other participants come up with.

This page was written by Steve Simon while working at Children's Mercy Hospital. Although I do not hold the copyright for this material, I am reproducing it here as a service, as it is no longer available on the Children's Mercy Hospital website. Need more information? I have a page with general help resources. You can also browse for pages similar to this one at Category: Information searching.