Question 1 How common is co-existing meningitis in infants with urinary tract infection?
- 1Department of Paediatrics, The University of Melbourne, Royal Children's Hospital Melbourne, Parkville, Victoria, Australia
- 2Infectious Diseases Unit, Department of General Medicine, Royal Children's Hospital Melbourne, Parkville, Victoria, Australia
- 3Murdoch Childrens Research Institute, Royal Children's Hospital Melbourne, Parkville, Victoria, Australia
- Correspondence to Dr Marc Tebruegge, Department of Paediatrics, The University of Melbourne, Royal Children's Hospital Melbourne, Flemington Road, Parkville, VIC 3052, Australia;
- Accepted 12 March 2011
You are asked to review a febrile 2-month-old infant who presented to the accident and emergency department. The urine analysis carried out before your arrival is suggestive of urinary tract infection (UTI) (urine dipstick: positive for nitrites and white blood cells (WBCs); microscopy: 220 WBCs per high powered field). On examination the infant appears well and has no signs suggestive of meningitis. However, you recall a senior colleague stating that young infants with UTI should always have a full septic workup to rule out co-existing bacterial meningitis. You wonder if there is any evidence to support routinely performing a lumbar puncture in this setting?
Structured clinical question
In an infant with UTI [patient], is a lumbar puncture [test] indicated to rule out co-existing bacterial meningitis [outcome]?
Search strategy and outcome
Search of the Cochrane Library database using the terms ‘Urinary Tract Infection’ and ‘pyelonephritis’ retrieved 18 Cochrane reviews, none of which were relevant. Medline was searched using the PubMed interface (1950—to date/no limits set): (1) a keyword search using (Urinary Tract Infection OR pyelonephritis OR cystitis) AND (meningitis OR meningoencephalitis OR lumbar puncture) AND (neonat* OR infant) retrieved 507 publications, of which 11 were relevant,1,–,11 and (2) a search using the MeSH terms (‘Infant, Newborn’ OR ‘Infant’) AND ‘Urinary Tract Infections’ AND ‘Central Nervous System Infections’ retrieved 172 publications, but no further relevant papers were identified. Search of the ISI Web of Science (1900—to date), EMBASE (1980—to date) and Scopus (1900—to date) using the same keyword strategy as above yielded 106, 328 and 163 matches, respectively, among which no, two12 13 and one12 further relevant paper(s) were identified. Publications that described a group of fewer than 30 infants with UTI and those providing insufficient detail were excluded. All relevant publications were hand-searched for additional references, which identified a further two reports.14 15 The search date was 4 January 2011. The relevant papers are summarised in table 1.1,–,15
UTIs are common. Between 1% and 15% of young children presenting with fever in the outpatient setting have an underlying UTI.16 17 In comparison to older children, UTI in infants is more commonly associated with bacteraemia, occurring in 4–12% of cases.1 3 5 6 10 11 It has been postulated that infants with UTI are therefore at higher risk of co-existing meningitis as a result of bacterial dissemination to the central nervous system (CNS).
The detection of co-existing meningitis in children with UTI is important, as the treatment of the former differs considerably from that of UTI alone. UTI may be treated with oral antibiotics or antibiotics that have poor cerebrospinal fluid (CSF) penetration, and the treatment duration is generally shorter. Consequently, failure to detect co-existing meningitis may result in inadequate or partial treatment of the CNS infection. Although performing a lumbar puncture in all infants with UTI is arguably the safest strategy, this is a procedure associated with pain and potential complications.
Clinical bottom line
▶ Between 0% and 2% of infants under the age of 3 months with urinary tract infection have co-existing bacterial meningitis. (Grade B)
▶ There are insufficient data on the rate of co-existing bacterial meningitis in older infants (aged 3–12 months). (Grade D)
The latest National Institute for Health and Clinical Excellence clinical guideline on UTI in children (CG54) recommends starting children older than 3 months of age with ‘uncomplicated UTI’ on oral antibiotics, such as a cephalosporin or co-amoxiclav.18 For children younger than 3 months, the reader is referred to the guideline on feverish illness in children (CG47).19 However, this guideline does not specifically indicate whether infants under the age of 3 months with confirmed, uncomplicated UTI should routinely undergo lumbar puncture to rule out co-existing meningitis.
Fourteen reports identified in our literature search were retrospective studies; only one study was prospective.12 The majority of studies included only infants under 3 months of age.1 3,–,5 7 8 10 11 14 15
In all but one study,14 UTI was defined as a positive urine culture; however, different microbiological cut-offs were used to define culture positivity. Also, different methods of urine collection, including urine bag, clean catch, catheterisation and suprapubic aspirate, were used. This is important, as particularly the first two methods are prone to contamination, which may result in false-positive urine culture results. Therefore, there is the possibility that some patients with meningitis included in these studies were falsely classified as having co-existing UTI.
Only one study based the diagnosis of UTI on dipstick analysis, therefore being the only study that is directly applicable to our clinical scenario.14 Importantly, this study reflects ‘real life’ practice, as culture results are not available at the time of the initial presentation and consequently can not be used to make decisions about further investigations, including lumbar puncture, and subsequent treatment.
The frequently quoted study by Bergström et al15 found that six (19.4%) of 31 neonates with UTI had co-existing bacterial meningitis. This study has previously been criticised as likely to considerably overestimate the risk of co-existing bacterial meningitis as all specimens were obtained using urine bags, which may have resulted in a considerable overestimate of co-existing UTI.4 14 Also, fewer than half of the infants with UTI included in this report underwent lumbar puncture. Consequently, as infants with meningitis (ie, those appearing sicker) would have been more likely to be subjected to a lumbar puncture, this is likely to have skewed the estimate of co-existing meningitis even further. This important limitation also applies to some other studies included in this review in which a considerable proportion of participants did not undergo lumbar puncture.4 6 8 10 11 14
In the remaining 14 studies included in this review, the rate of co-existing bacterial meningitis in infants with UTI ranged from 0% (six studies)2 7 8 11 13 14 to 2.0% (one study).10 Notably, the latter study was relatively small and included only a single patient with co-existing meningitis. The heterogeneity of the 15 studies, particularly with respect to age of participants, urine sampling methods and microbiological definitions of UTI, precludes a formal meta-analysis, but a summary is provided in figure 1. The largest study by far (n=1609), which is likely to provide the best estimate, reported only two patients (0.1%) with confirmed bacterial meningitis, while a further three (0. 2%) were classified as ‘probable meningitis’ based on CSF findings.1
It is noteworthy that aseptic meningitis, which was addressed in only some studies,1 2 4 6 8 9 11 12 15 was identified in 5–18% of patients. Importantly, in only one study did the investigators exclude patients who had received antibiotics prior to presentation by analysing urine samples for antimicrobial activity.2 In all the other studies, the possibility that some patients had received antibiotics prior to undergoing lumbar puncture (eg, oral antibiotics not reported by the parents), which may have rendered some CSF cultures falsely negative, can not be excluded. The pathophysiology underlying aseptic meningitis in patients with UTI is not fully understood, although endotoxin-mediated phenomena resulting in CSF pleocytosis is one postulated mechanism.8
Although the available data have considerable limitations, they suggest that in infants under the age of 3 months with UTI, the ‘true’ rate of co-existing bacterial meningitis is likely to be under 1%. Whether this level of risk is sufficient to justify universal lumbar puncture is debatable. There are currently insufficient data on this issue in infants aged between 3 and 12 months. Further data from large, well-designed studies are needed to establish more precisely the risk of meningitis in infants with UTI in this age group and consequently the role of lumbar puncture.
The authors thank Professor Alison Kesson, Dr Peter Dayan, Dr Samir Shah and Dr David Schnadower for providing further information related to their original publications. MT is supported by a Fellowship award from the European Society for Paediatric Infectious Diseases (ESPID) and an International Research Scholarship from The University of Melbourne.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.