 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Gérard Bréart, MD
Institut National de la Santé et de la Recherche Médicale U149
75014 Paris, France
Martin Chalumeau, MD, PhD
Department of Pediatrics
Clinical Epidemiology Unit
Saint-Vincent-de-Paul Hospital
75014 Paris, France
Institut National de la Santé et de la Recherche Médicale U149
75014 Paris, France
We greatly appreciate Chevalier and Gauthier's interest in our article.1 They suggest that the use of sterile bags for urine collection biased our results. They base their comments largely on a technical report2 that support the American Academy of Pediatrics practice parameter on urinary tract infection (UTI).3 As stated in the discussion of our article,1 we agree that the use of sterile bags for urine collection is not as specific as suprapubic aspiration or transurethral catheterization.
However, data concerning the specificity of sterile bags in non–toilet-trained children older than 1 month is scarce. The only valid method for ascertaining the specificity of sterile bags requires comparing urine specimens collected by this technique and by suprapubic aspiration or transurethral catheterization within a very close time frame in patients suspected of having UTI. What constitutes a positive bacterial culture of urine collected by sterile bags needs to be defined in advance for this comparison. In his 1999 technical report, Downs2 reported that the specificity of sterile bags varied between 14% and 84% based on 4 published studies.4–7 Unfortunately, a careful analysis of 2 of these studies reveals that they do not provide data about the specificity of sterile bags.4,5 The specificity reported in the third study was based on pooled results obtained from 2 different urine-collection techniques (sterile bags and clean-voided midstream); it is not possible to calculate the specificity of sterile bags alone.6 Data from the fourth study did not permit calculation of the specificity of sterile bags either. Indeed, there were no available data about children with a negative culture of urine collected by sterile bags.7 As mentioned but not discussed in Downs' technical report,2 Pylkkanen et al8 reported a specificity of sterile bags of 88% (95% confidence interval [CI]: 78–93) in a study of 272 children. The authors defined UTI as a positive bacterial monoculture of 
105 colony-forming units per mL for sterile-bag specimens. This threshold was also the one used in our study1 (because it is recommended by the French Society of Pediatrics9), and it also was used in a recent study by the Pediatric Research in Office Settings' group.10 Other studies11,12 published before Downs' technical report support the results reported by Pylkkanen et al. For example, Benito Fernandez et al12 reported in 1996 the comparison of sterile-bag collection with suprapubic aspiration in 48 children <1 year of age. The specificity of sterile bags was 89% (95% CI: 74–95) when using the same previously defined urine-culture threshold for bag specimens. It is unclear to us why Downs' technical report2 failed to cite and comment on those key studies. Thus, the level of evidence for the range of specificity of sterile-bag collection that he reported (and used in Chevalier and Gauthier's letter) is questionable. It may indicate room for a more systematic review.
Furthermore, if the relationship between vesicoureteral reflux (VUR) and high procalcitonin (PCT) was due only to the selection bias related to the use of sterile bags as Chevalier and Gauthier suggest, then (1) the prevalence of VUR would be higher in children with UTI diagnosed by suprapubic aspiration or transurethral catheterization than sterile bags, and (2) the external validation of this relationship would fail in other populations of children with UTI diagnosed by suprapubic aspiration or transurethral catheterization. As a matter of fact, we recently conducted a large multicenter validation study of 398 patients in 8 centers (5 using suprapubic aspiration or transurethral catheterization) in 7 European countries.13 The prevalence of VUR did not differ statistically (P = .9) according to the urine-collection technique used (26% in the centers using suprapubic aspiration or transurethral catheterization vs 25% in those using only sterile bags or clean-voided midstream). The relationship between VUR and high PCT remained strong and independent (adjusted odds ratio: 2.4; 95% CI: 1.4–4.1; P = 10–3). High PCT sensitivity and specificity were very close to those found in our single-center study1: the sensitivity was 75% for all-grade VUR and 89% for high-grade VUR, with a specificity of 43% in both cases.
We believe that the specificity of sterile bags as a urine-collection technique is not that poor, especially if febrile UTI is defined as the association of fever, positive urine monoculture (105 colony-forming units per mL), and positive biological inflammation, as it was in our study.1 That is probably one reason, together with the adverse effects of suprapubic aspiration or transurethral catheterization (including pain and trauma), why many pediatricians in the United States,10 as well as in numerous European countries,9,14–16 use sterile bags. Our data also provide strong evidence that PCT is a powerful predictor of VUR in children with a first febrile UTI regardless of the urine-collection technique used for the diagnosis of UTI.
REFERENCES
Related articles in Pediatrics:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2005 American Academy of Pediatrics. All rights reserved. |
||
|
||
