Value of Diagnostic Tests for Bacterial Infections from Birth to 90 Days


Synopsis: A meta analysis of 52 scientific studies of the value of laboratory studies (WBC, bandemia, CRP) to identify 0 to 90-day-old infants with possible bacterial infection showed that the results of these tests did not reliably identify infants with proven bacterial infections.

Source: Fowlie PW, Schmidt B. Diagnostic tests for bacterial infection from birth to 90 days-a systematic review. Arch Dis Child Fetal Neonatal Ed 1998;78:F92-F98.

Fowlie and schmidt performed a medline search of literature published in a 30-year period, 1966-1995, to determine the clinical value of common diagnostic tests (e.g., total white blood cell count, immature to total white cell ratio, and the C-reactive protein) for identifying bacterial infection in infants younger than or 90 days of age. Fifty-two of 572 papers that were identified met three methodologic criteria of quality. The likelihood ratios, the measure of efficacy of the tests used in the 52 studies, generally fell within the indeterminate range leading Fowlie and Schmidt to state that any single test or combination of tests was "unlikely to change the pre-test probability of a given child either being infected or not being infected." Fowlie and Schmidt conclude that "tests are of limited value in the diagnosis of infection in this population."


Fowlie and Schmidt from Dundee, Scotland, should be applauded for a significant effort at reviewing a large number of articles and for focusing on this important topic. In order to interpret the results of this study in the context of other recent work in this area, the following should be considered.

First, several large studies have focused on the evaluation and management of febrile infants and implicitly assume that screening tests such as the white blood cell count and differential counts, while helpful in defining risk for serious bacterial illnesses, are insufficiently predictive of and/or sensitive for these infections. Those should not be used alone; rather the screening tests should be combined with observation, history, and physical exam data, as well as results of urinalysis and CSF analysis.1,2 Thus, the findings in the current study of the limited value of screening laboratory tests, such as the total white blood cell count reflect what is current practice, that is, non-reliance on these tests alone to diagnose bacterial illnesses.

Second, Fowlie and Schmidt combine newborns and infants up to 90 days of age. The bacterial pathogens that cause disease vary in infants depending on age. Moreover, recent studies reported before the Ambulatory Pediatric Association draw distinctions between the ability of screening protocols to diagnose bacterial illness in infants younger than 29 days vs. infants 30-89 days of age.3,4 Yet, these age groups are combined in the current study.

For these reasons, I believe the results reported in this article, while of interest, do not offer insights beyond those that have been integrated into current clinical practice. (Dr. McCarthy is Professor of Pediatrics at Yale University School of Medicine and developed the Yale Scoring System for evaluation of the degree of illness in infants and children.)5


    1. Baker MD, et al. N Engl J Med 1993;329:1437-1441.

    2. Jaskiewicz JA, et al. Pediatrics 1994;94:390-396.

    3. Baker MD, et al. Program and Abstracts: Ambulatory Pediatric Association (38th Annual Meetings, New Orleans); 1998:114.

    4. Kadish HA, et al. Programs and Abstracts: Ambulatory Pediatric Association (38th annual Meeting, New Orleans); 1998:115.

    5. McCarthy PL. Pediatrics 1982;70:802-809.