Letters to the Editor

Abstract

17 August 2009 Dear Editor, STERILE CEREBROSPINAL FLUID PLEOCYTOSIS IN INFANTS WITH URINARY TRACT INFECTION I read the paper entitled ‘Incidence of sterile cerebrospinal fluid pleocytosis in infants with urinary tract infection’1 with interest. The study is a useful addition to existing evidence (including our own2) that urinary tract infection (UTI) during early infancy may be associated with sterile cerebrospinal fluid (CSF) pleocytosis, but is rarely associated with CSF sepsis. As Yam et al. note, their study extends existing knowledge as they were able to exclude previous antibiotic therapy by measuring antimicrobial activity in the urine. However, I was surprised that the discussion regarding the clinical relevance of their findings was limited to ‘The importance of reporting CSF pleocytosis in UTI is educational so clinicians do not treat such infants for an unnecessary duration for possible bacterial meningitis’. In my view, this statement is not terribly helpful. The decision to cease intravenous antibiotic therapy in an infant with a proven UTI that has responded to therapy, and in whom the CSF cultures are sterile, is relatively straight forward. This was demonstrated by Yam et al.'s findings, as only one of the 12 infants with UTI and sterile CSF pleocytosis received a prolonged course of intravenous antibiotics. The more common and difficult clinical conundrum is: in an infant with working diagnosis of UTI (established on the basis of the full ward of an adequately collected sample), should one perform a lumbar puncture (LP) in the first place? Among the 106 infants with UTI who also had LP performed in the study by Yam et al.,1 there were no cases of bacterial meningitis identified. Similarly, among the 75 infants with UTI who also had an LP performed in our study,2 only one case of ‘probable’ bacterial meningitis was identified. There were no infants who relapsed with bacterial meningitis identified in either. Surely then, the relevant conclusion is, that in a well-looking infant with a working diagnosis of UTI, LP may not be necessary. Of course caution is required as the existing evidence is retrospective, and meningitis is a subtle and significant diagnosis in young infants. We suggested that a reasonable approach might be to limit the LP to infants at higher risk of coexisting CSF sepsis: neonates, infants with signs of septicaemia (pallor, tachycardia, inadequate perfusion), and infants with specific signs of meningitis (paradoxical irritability, a bulging fontanel, neck stiffness).