Search tips
Search criteria 


Logo of pchealthLink to Publisher's site
Paediatr Child Health. 1998 May-Jun; 3(3): 158–159.
PMCID: PMC2851319

Language: English | French

How accurate are leukocyte indices and C-reactive protein for diagnosis of neonatal sepsis?

Orlando da Silva, MD MSc FRCPC1 and Arne Ohlsson, MD MSc FRCPC2


Early diagnosis of neonatal sepsis is often difficult to make. Treatment on the basis of clinical suspicion and risk factors may result in overtreatment. A previous review of the usefulness of C-reactive protein and leukocyte indices concluded that these test results should be interpreted with caution. The present paper reviews and, when appropriate, revises, in light of new information, the conclusions reached in the previous systematic review of the topic.

Keywords: C-reactive protein, Leukocyte count, Infant, Newborn, Sepsis


Le diagnostic précoce d’une septicémie néonatale est souvent difficile à poser. Un traitement administré sur la base d’une suspicion clinique et des facteurs de risque peut conduire à un surtraitement. Une revue antérieure de l’utilité de la protéine C-réactive a conclu que les résultats de la protéine C-réactive doivent être interprétés prudemment. Le présent article passe en revue et, au besoin, révise, à la lumière des nouvelles données, les conclusions qui ont été tirées dans les revues systématiques antérieures concernant ce sujet.

The nonspecific signs and symptoms of neonatal sepsis make it difficult to establish an early clinical diagnosis, and antibiotic therapy is usually initiated based on risk factors and/or clinical suspicion. Given that neonatal sepsis has a low incidence (one to eight cases per 1000 live births) (1,2), treatment on clinical grounds and risk factors alone results in overtreatment. Data from St Joseph’s Health Centre, London, Ontario have shown that only 12% of infants investigated and treated for suspected sepsis had bacteremia (3).

Blood culture is considered to be the ‘gold standard’ for diagnosis of septicemia; however, its accuracy has been questioned because of spurious positive results due to contamination (4) and negative blood cultures in fatal generalized bacterial infection (5). Thus, in addition to the blood culture, various laboratory tests have been evaluated in the diagnosis of systemic infection in neonates (6,7). Complete blood count (CBC) with the various neutrophil parameters (611) and C-reactive protein (CRP) (613) are the most frequently used. CRP is an acute phase reactant synthesized by the liver that increases in the presence of inflammation caused by infection or tissue injury (14). In a systematic review of the literature of studies evaluating the usefulness of CRP and leukocyte indices for diagnosis of neonatal sepsis, da Silva et al (15) found a wide range of test sensitivity and specificity, and the investigators concluded that caution should be used when interpreting these test results.

The purpose of this paper is to review and, when appropriate, revise, in light of new information, the conclusions reached in our previous systematic review of the topic (15).


A computerized search was made of MEDLINE from 1994 to June 1997, using the following terms: ‘infant newborn’, ‘C-reactive protein’ and ‘leukocyte’, and ‘sepsis’, ‘septicemia’ or ‘infection’. Bibliographical lists of primary and review articles and personal files were also searched for any additional relevant references. The search was limited to English language only. No formalized quality assessments of the literature as previously described were performed (15).


Wagle et al (16) evaluated CRP and leukocyte parameters in 123 consecutively admitted infants of less than 30 weeks’ gestation. Thirty-six infants had proven sepsis (positive blood culture) during their hospital admission. Leukocyte indices at the time of the sepsis screen showed poor sensitivity and specificity; sensitivity and specificity were much improved 24 h later. Similarly CRP levels on day 1 of the septic episode were a poor predictor of infection; however, the accuracy of the test was significantly improved with serial measurements. The best test performance was afforded by combining CRP with left shift and white cell count results.

In a study by Buck et al (17), 11 of 222 newborns studied prospectively had proven sepsis. The sensitivity of CRP was 73%; when infants with culture-negative clinical sepsis were included, it decreased to 58%. A comparative study of CRP and various leukocyte parameters showed comparable accuracy during the first three days of life (18). In the cases of late sepsis, CRP performed better. Repeated CRP and leukocyte indices during the course of infection had high sensitivity and negative predictive values (18). Engle et al (19) evaluated leukocyte indices in a group of very low birth weight infants (less than or equal to 1500 g) with positive blood cultures during the first month of life. Using the published reference ranges of Manroe et al (20), the authors found a sensitivity of 75% for a first CBC, which increased to 96% when a second CBC taken 12 to 24 h after the first one was included in the analysis.


Sepsis in the neonatal period is a life-threatening, yet treatable condition of particular concern in the very low birth weight infant. An ideal diagnostic test for neonatal sepsis should have maximum sensitivity (not to miss any septic infant) and maximum negative predictive value (to exclude sepsis when the test is negative and consequently decrease unnecessary exposure to antibiotics).

In a recent comprehensive systematic review of the literature evaluating leukocyte ratios and CRP, the authors found a wide range of sensitivity (17% to 90%) and specificity (31% to 100%) for leukocyte count and ratios (15). Similarly CRP measurements showed variable accuracy, though better accuracy than the leukocyte parameters (15). This variability may be partially explained by the infant’s age (early versus late sepsis), laboratory expertise and the many noninfectious factors that can affect the leukocyte parameters and CRP measurements (15).

In this updated review, few articles published since 1994 that evaluated CRP and leukocyte parameters were identified (1619). The variability in the accuracy of leukocyte and CRP to predict sepsis is still remarkable. However, it is important to note that both tests are quite predictive of neonatal sepsis when measured serially (1619). Thus, these tests when repeated provide useful information when treating an infant with suspected sepsis but negative blood cultures.


Neither CRP nor leukocyte indices are accurate enough for the early diagnosis of neonatal sepsis. Serial measurements of CRP, leukocyte count and various leukocyte indices have improved accuracy in the diagnosis of sepsis and may be useful in monitoring length of antibiotic therapy. CRP is not more accurate than leukocyte parameters.


Reviewed by the Canadian Paediatric Society Board of Directors


1. Freedman RM, Ingram DL, Gross I, et al. A half century of neonatal sepsis at Yale. Am J Dis Child. 1981;135:140–4. [PubMed]
2. Placzek MM, Whitelaw A. Early and late neonatal septicemia. Arch Dis Child. 1983;58:728–31. [PMC free article] [PubMed]
3. Hammerberg O, Bialkowska-Hobranska H, Gregson D, Potters H, Gopaul D, Reid D. Comparison of blood cultures with corresponding venipuncture site cultures of specimens from hospitalized premature neonates. J Pediatr. 1992;120:120–4. [PubMed]
4. Pourcyrous M, Korones SB, Bada HS, Patterson T, Baselski V. Indwelling umbilical arterial catheter: A preferred sampling site for blood culture. Pediatrics. 1988;81:821–5. [PubMed]
5. Squire E, Favara B, Tood J. Diagnosis of neonatal bacterial infection: Haematologic and pathologic findings in fatal and nonfatal cases. Pediatrics. 1979;64:60–4. [PubMed]
6. Gerdes JS. Clinicopathologic approach to the diagnosis of neonatal sepsis. Clin Perinatol. 1991;18:361–81. [PubMed]
7. Powell KR, Marcy SM. Laboratory aids for diagnosis of neonatal sepsis. In: Remington JS, Klein JO, editors. Infectious Diseases of the Fetus and Newborn Infant. 4th edn. Philadelphia: WB Saunders Co; 1994. pp. 1223–40.
8. Philip AG. Detection of neonatal sepsis of late onset. JAMA. 1982;247:489–92. [PubMed]
9. Philip AG, Hewitt JR. Early diagnosis of neonatal sepsis. Pediatrics. 1980;65:1036–41. [PubMed]
10. Chadna A, Rao MN, Srinivas M, Shyamala S. Rapid diagnostic tests in neonatal septicemia. Indian J Pediatr. 1988;55:947–53. [PubMed]
11. Manroe BL, Rosenfeld C, Weinberg AG, Browne R. The differential leucocyte count in the assessment and outcome of early-onset neonatal group B streptococcal disease. J Pediatr. 1977;91:632–7. [PubMed]
12. Baptista-Gonzalez HA, Ibarra-Camacho A, Cedillo R, Torres-Alipi BI. Utilidad de la proteina C reactiva para el diagnostico de sepsis neonatal. Bol Med Hosp Infant Mex. 1989;46:543–6. [PubMed]
13. Pourcyrous M, Bada HS, Korones SB, Baselski V, Wong SP. Significance of serial C-reactive protein responses in neonatal infection and other disorders. Pediatrics. 1993;92:431–5. [PubMed]
14. Pepsys MB. C-reactive protein fifty years on. Lancet. 1981;i:653–6. [PubMed]
15. da Silva O, Ohlsson A, Kenyon C. Accuracy of leucocyte indices and C-reactive protein for diagnosis of neonatal sepsis: A critical review. Pediatr Infect Dis J. 1995;14:362–6. [PubMed]
16. Wagle S, Grauaug A, Kohan R, Evans SF. C-reactive protein as a diagnostic tool of sepsis in very immature babies. J Paediatr Child Health. 1994;30:40–4. [PubMed]
17. Buck C, Bundschu J, Gallati H, Bartman P, Pohlandt F. Interleukin-6: A sensitive parameter for the detection of neonatal bacterial infection. Pediatrics. 1994;93:54–8. [PubMed]
18. Berger C, Uehlinger J, Ghelfi D, Blau N, Fanconi S. Comparison of C-reactive protein and white blood cell count with differential in neonates at risk for septicemia. Eur J Pediatr. 1995;154:138–44. [PubMed]
19. Engle WD, Rosenfeld CR, Mouzinho A, Risser RC, Zeray F, Sanchez PJ. Circulating neutrophils in septic preterm neonates: Comparison of two reference ranges Pediatrics 1997. 99e10 URL: [PubMed]
20. Manroe BL, Weinberg AG, Rosenfeld CR, Browne R. The neonatal blood count in health and disease. I. Reference values for neutrophilic cells. J Pediatr. 1979;95:89–98. [PubMed]

Articles from Paediatrics & Child Health are provided here courtesy of Pulsus Group