Skip to main content

Archived Comments for: C-reactive protein velocity to distinguish febrile bacterial infections from non-bacterial febrile illnesses in the emergency department

Back to article

  1. Detecting bacterial versus non-bacterial febrile illnesses in the emergency room

    Francisco Arnalich, Emergency Medicine Unit, La Paz University Hospital, Madrid, Spain

    1 October 2009

    Dear Sir
    I read with interest the paper by Y. Paran et al. proposing the use of C-reactive protein velocity (CRPv), defined as the ratio between CRP value on admission and the number of hours since the onset of fever, to distinguish febrile bacterial infections from non-bacterial febrile illnesses in patients admitted to the emergency room (ER). The diagnostic accuracy of CRPv expressed as AUC was better than for CRP alone (0.871 vs. 0.783), and similar to that reported by using six different biomarkers, including CRP and PCT, which yielded an AUC of 0.88 (1). Despite their study´s strengths, however, some limitations not stated by authors deserve consideration. First, the results are not valid for patients with dementia or other mental diseases which are unaware of fever or unable to provide informed consent. Second, they did not include febrile outpatients who were normothermic at the ER because of ongoing antipyretic treatment. In addition, it would be interesting to monitor changes in CRP over the period of 6 to 12 hours, which represents the average length of stay for clinical work-out in the ER. Unlike procalcitonin secretion which peaks 8 h after stimulation, CRP peaks only after 36 h and allows a time-course analysis. The magnitude of change on two measurements adds information to that obtained by the CRP velocity, especially in patients for whom the time of fever onset is doubtful. In the Emergency Department at La Paz Hospital, Madrid, Spain, a 1100-bed academic tertiary hospital serving 485.000 inhabitants, we usually measured CRP levels on admission and 6 hours afterwards. We have found that patients with proven bacterial infections had an average two-fold increase in CRP levels over the admission values compared with a mean 0.6-fold increase in patients with non-bacterial febrile illness. We think that combining information from CRP velocity and CRP increases over 6 to 12 hours is a cost-efficient approach to identify patients with bacterial infection.

    1)Kofoed K, Andersen O, Kronborg G, Tvede M, Petersen J, Eugen-Olsen J, Larsen K. Use of plasma C-reactive protein, procalcitonin, neutrophils, plasminogen activator receptor, and soluble triggering receptor expressed on myeloid cells-1 in combination to diagnose infections: a prospective study. Crit Care 2007; 11:R38.

    Competing interests

    I have no competing interests