JAC Advance Access originally published online on July 26, 2006
Journal of Antimicrobial Chemotherapy 2006 58(4):902-903; doi:10.1093/jac/dkl282
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Correspondence |
Comment on: Urinary tract infections in general practice patients: diagnostic tests versus bacteriological culture
Hospital Saint-Joseph Paris, France
*Tel: +33-144123650; Fax: +33-144123685; E-mail: fgoldstein{at}hopital-saint-joseph.org
Keywords: UTI , dipstick , dipslide
Sir,
The recent paper of Nys et al.1 raises several important comments. There are more than 400 papers, including ours, published in peer-reviewed journals, dealing with the evaluation of dipstick screening tests for the detection of urinary tract infections (UTIs).2–5 The results of these studies have been included in reports such as the French Consensus Conference on the diagnosis and treatment of UTI.6 The only relevant reference given by Nys et al. is a publication, in Dutch, of the ‘Guidelines of the Dutch College of General Practitioners’.
In most papers, the authors indicate in a table all the useful results for the calculation of sensitivity, specificity and predictive values according to the different chosen criteria or patient selection. One of the most important results, which is highlighted in all other studies,2–5 the negative predictive value (NPV) of the combination nitrites (NIT) + leucocyte-esterase (LE), usually 90–98%, is missing in their results.
Nys et al. indicate a positive predictive value (PPV) of 96% for NIT alone with an NPV of 30%; this is in sharp contrast with other studies where the PPV for NIT alone is 50–75% for a bacterial cut-off 
105 cfu/mL,2–5 falling to 0% for 103 cfu/mL when the NPV is 80–99%.7
There are more than 20 manufacturers of dipsticks with NIT or LE alone or in combination, yielding varying results.3 The authors do not indicate which dipsticks were used or whether the reading of results was manual or automatic.
The 103 cfu/mL cut-off is not acceptable in routine studies without careful confirmation of the UTI. I had the pleasure and honour to discuss with E. Kass, many years ago, the results of his initial study establishing the worldwide accepted criterion of a 105 cfu/mL cut-off.8 He fully agreed that a very small fraction of women have true UTI with <105 cfu/mL, but in most cases, in the absence of leucocyturia, low counts correspond to contamination. Nys et al. do not indicate how many patients were included in the 103–104, 104–105 and >105 cfu/mL bacterial range, the prevalence of LE in each group and the bacteria isolated. The authors do not indicate whether the general practitioners included diabetic patients or patients with immunodeficiency disorders, those receiving antibiotics or with indwelling catheters, usually excluded from other studies.4 It is well established that such patients may have a low bacterial and/or leucocyte count.
There are other important flaws in their study: the authors indicate that NIT and/or LE dipstick tests were performed by the general practitioners; but how many patients had both NIT + LE tests? Only 66% of the patients had an infection caused by Escherichia coli: this is 10–20% less than in most other studies with similar patients. How were the bacteria identified, since API20 was only used in case of doubt? The identification has been apparently performed on bacteria taken directly from the dipslide, without prior isolation and purity checking. Minor contaminations are extremely frequent (20–24%) in patients seen in our laboratory and such contaminations are impossible to detect on the dipslide if the bacterial count is >104 cfu/mL. Moreover, during transport by post to the central laboratory, taking at least 24 h, bacteria will easily multiply and swarm during the summer months, because of the high humidity in the dipslide container, rendering the evaluation of bacterial count and detection of contamination very hazardous. It is very likely that many bacterial counts have been overestimated. The authors apparently considered all mixed infections as contamination. It is well established that at least 5% of lower UTIs are due to two bacteria, usually two different E. coli. Symptoms of UTI can occasionally be due to another origin than infection; how many patients had a low bacterial count with a negative LE test and infection caused by enterococci?
My last comments concern the antibiotic treatment. Guidelines for the treatment of uncomplicated lower UTI are available worldwide and have been very widely accepted. Antibiotics are extremely useful for the treatment of UTI but they are completely unnecessary for patients without infections, such as 20% of these patients with a negative NIT + LE test who were treated.
Treatment duration of 7 days instead of only one simple dose of a fluoroquinolone for a lower uncomplicated UTI is not only ridiculous but certainly contributes to the emergence and spread of resistant bacteria and unnecessarily increased health costs.
Transparency declarations
None to declare.
References
1 Nys S, van Merode T, Bartelds AIM, et al. (2006) Urinary tract infections in general practice patients: diagnostic tests versus bacteriological culture. J Antimicrob Chemother 57:955–8.
2 Goldstein FW. (1991) Place actuelle des tests rapides de détection de l'infection urinaire. Med Mal Infect 21:83–8.
3 Devillé W, Yzermans JC, van Duijn NP, et al. (2004) The urine dipstick test useful to rule out infections. A meta-analysis of the accuracy. BMC Urol 4:4.[CrossRef][Medline]
4 Lammers RL, Gibson S, Kovacs D, et al. (2001) Comparison of test characteristics or urine dipstick and urinalysis at various test cutoff points. Ann Emerg Med 38:505–12.[CrossRef][Web of Science][Medline]
5 Zaman Z, Borremans A, Verhaegen J, et al. (1998) Disappointing dipstick screening for urinary tract infection in hospital inpatients. J Clin Pathol 51:471–2.[Abstract]
6 Antibiothérapie des infections urinaires. Deuxième conférence de consensus en thérapeutique anti-infectieuse de la Société de Pathologie Infectieuse de langue française Med Mal Infect (1990) 21: pp. 49–167.
7
Semeniuk H and Church D. (1999) Evaluation of the leukocyte esterase and nitrite urine dipstick screening tests for detection of bacteriuria in women with suspected uncomplicated urinary tract infections. J Clin Microbiol 37:3051–2.
8 Kass EH. (1957) Bacteriuria and the diagnosis of infections of the urinary tract. AMA Arch Intern Med 100:709–12.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||