Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Is microscopy better than urine dipstick testing at identifying a UTI in children

Three Part Question

[In a child with suspected UTI] [is a positive dipstick better than microscopy at] excluding or confirming a [UTI]?

Clinical Scenario

A 6 year old child presents to the Emergency department with symptoms of UTI. You obtain an MSU and want to send it for urgent microscopy. However, the lab technician calls to tell you that there is no need for microscopy if urine dipstick has been done, as it can confirm or rule out UTI. You wonder if the lab technician is right or whether you should send the sample for microscopy after dipstick analysis.

Search Strategy

Medline database 1966 - week 2 June 2005 via OVID
EMBASE 1980 - week 2 June 2005
CINAHL 1982 - week 2 June 2005
WWW - google
[BestBETS paediatric filter] AND [urinary tract infection.mp./or exp Urinary Tract Infections] AND [diagno$.mp./or exp diagnosis] AND [urinalysis.mp./or ecp urinalysis/or exp "Sensitivity and Specificity"/or exp Reagent Strips/or dipstick.mp./or exp "Indicators and Reagents"/or exp microscopy/or microscopy.mp.]
Limit to Humans and English language
Cochrane: children AND uti AND microscopy
WWW- Rapid tests, urine sampling, urinary tract infection

Search Outcome

Medline: 592 papers. 16 relevant papers which compared microscopy and urine dipstick analysis in the investigation of childhood UTI. Of these a meta-analysis included 6 of the papers found. One was a meta-analysis which compared microscopy and dipstick and included many of the other papers found as answering the three part question. Other papers not included in the meta-analysis are also shown in the table.
EMBASE: 512 papers found, 5 of which were directly relevant to the 3 part question. Of these one was the meta-analysis found on Medline, one paper was included in the meta-analysis,one was already found on Medline and the other 2 were new papers not found on Medline. The same meta-analysis found on Medline was also found on EMBASE and included all papers identified as answering the three part question.
CINAHL: 64 papers found, none of which answered the 3 part question directly.

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Smith P, Morris A, Reller LB
2003
USA
All patients at the Duke University Medical Center, who had a urine specimin obtained by either clean-catch, MSU or single catheterisation, between 10:00 and 14:00, and only if recieved by the Clinical Microbiology Laboratory, within 10 hours of collection, over a 4 month period. Patients were aged between 1 month and 91 years, 60 of these were 14 years of age.Diagnostic studySensitivity, specificity, positive and negative predictive values of Pyuria to predict culture results: presence of a reportable pathogenSensitivity 63%, Specificity 89%, PPV 58%, NPV 91%The range of patients included in the study vary from 1 month to 91 years, so there's quite a wide range of ages of patients included, which could have an effect on the performance of the investigations under study, focusing on a particular age-range of patients may have been more approriate than analysing all ages of patients. Other weaknesses mentioned by the authors include: - Pyuria is not specific to UTI, it only suggests increased likelihood of a patient having UTI, and can be seen in other infections, e.g in common chlamydial infections - Separating contamination from normal bladder bateruria is difficult - Elderly may have low level pyuria and sterile bladder urine - Mixed voided specimens are often seen in women with proven bladder bacteruria and acute urethral syndrome.
Sensitivity, specificity, positive and negative predictive values of Haematuria to predict culture results: presence of a reportable pathogenSensitivity 18%, Specificity 89%, PPV 27% NPV 82%
Sensitivity, specificity, positive and negative predictive values of Negative microscopy to predict culture results: absence of a reportable pathogenSensitivity 76%, Specificity 74%, PPV 92%, NPV 74%
Sensitivity, specificity, positive and negative predictive values of Negative microscopy and negative dipstick to predict culture results: absence of a reportable pathogenSensitivity 83%, Specificity 76%, PPV 94%, NPV 76%
Centre for Reviews and Dissemination
1999
UK
Paper including children (aged 3 weeks-21 years) with suspected UTICritically appraised economic evaluationSensitivity, specificity, PPV, NPV of dipstick only and microscopy dipstick in combination (complete urinalysis)Dipstick:Sensitivity 92%, specificity 62%, PPV 22%, NPV 99%. Microscopy: Sensitivity 92%, specificity 49%, PPV 17%, NPV 98%Only one study analysed - Retrospective
Hospital charges$12 for dipstick, $32 for complete urinalysis
Total turn-around time16 minutes for dipstick, 28.3 minutes for complete urinalysis
AAP
1999
UK
402 articles from the literature involving young children between 2 months and 2 years of age Recommendations for the diagnosis, management, and follow-up evaluation of infants and young children with unexplained fever who are later found to have a diagnosed UTISystematic literature reviewSensitivity and specificity of components of urinalysis alone and in combinationLeukocyte esterase (LE): Sensitivity 83%, Specificity 78% Nitrite: Sensitivity 53%, Specificity 98%, LE or Nitrite: Sensitivity 93%, Specificity 72% Microscopy: WBCs Sensitivity 73% Specificity 81%, Microscopy: bacteria Sensitivity 81% Specificity 83%, LE or Nitrite or Microscopy positive: Sensitivity 99.8% Specificity 70%.Well conducted systematic literature review
Hiraoka et al
2005
Japan
Patients aged between 3 months and 94 years. 43 samples from outpatients, 282 samples from inpatients. 109 samples from patients on antibiotics and 216 samples from patients without antibioticsDiagnostic studySensitivity and specificity of quantitative unspun urine microscopySensitivity 94.6%, specificity 99.3%Did not mention if the culture (gold standard) was blinded or not. Heterogeneity of age group included. Only 37 had positive cultures, therefore not many to compare the results of positive microscopy/Gram stain, i.e. a bigger sample may have been more appropriate.
Sensitivity and specificity of Gram stainSensitivity 89.2%, specificity 98.6%
Proesmans et al
1980
Belgium
Infants and children seen in an outpatient clinic for UTI and kidney disordersDiagnostic studyWas quite a brief report with very few details. There's no mention whether the gold standard (culture) was blinded. No statistics performed in the results section, sensitivity & specificity would have been useful.
sharief et al
1998
UK
375 children with fever, admitted to an acute paediatric ward of a district general hospital. Age range 2 days – 16 years.Diagnostic studySensitivity, Specificity, PPV and NPV of Nitrite for UTISensitivity 54.6%, Specificity 96.8%, PPV 37.5%and NPV 98.4%One of the limitations, which was mentioned by the authors themselves, was that there was a low incidence of UTI (5.2%) in the group of patients studied, and this therefore increased the specificity of the results. Also, they mentioned that some children may have been taking antibiotics which they did not know about, which would have affected the results. The gold standard (suprapubic aspiration) method of urine collection was not used, and therefore their may have been contamination of urine specimens examined. Although in the Discussion section they conclude that "in the absence of strong clinical indications it is not necessary to submit a urine for microbiological examination when the LE and NIT test are both negative", they do not present the results of the sensitivity and specificity of microscopy so that we, the reader, can come up with the same conclusion.
Sensitivity, Specificity, PPV and NPV of Leukocyte esterase for UTISensitivity 100%, Specificity 78.1%, PPV 13.9%, NPV 100%
Sensitivity, Specificity, PPV and NPV of Nitrite and Leukocyte esterase for UTISensitivity 54.7%, Specificity 98.7%, PPV 60.0% and NPV 96.9%
Thysell N.
1969
Sweden
Samples from two populations: 1) 1350 women employed at the County Council hospitals, aged between 16 – 69 year, attending voluntary health control for uterine cancer. 2) 512 inpatients and outpatients at the clinicDiagnostic studySensitivity and specificity of Nitrite for detection of bacteriuria (>=10 000 bact/ml urine)Sensitivity 36.5%, Specificity highDifficult to interpret results, more tables needed.
Sensitivity and specificity of microscopy of centrifuged and uncentrifuged urine for detection of bacteriuriaSensitivity centrifuged 70.1% uncentrifuged 78.6%, specificity low
Sensitivity and specificity of microscopy of urinary sediment for the presence of puree for detection of bacteraemiaSensitivity 65.6%, Specificity 78.3%
Novak et al
2004
USA
142 children <5 years of age in the Emergency Department of Akron Children's Hospital, over a 2 month periodDiagnostic studySensitivity of unspun leukocyte count >10 microliter, positive cytocentrifuge gram stain, 2 to 5 or more leukoctyes/hpf in sediment, positive leukocyte esterase, positive blood, positive nitriteUnspun leukocyte count >10 microliter 68%, positive cytocentrifuge gram stain 60%, 2 to 5 or more leukoctyes/hpf in sediment 48%, positive leukocyte esterase 48%, positive blood 44%, positive nitrite 20%They didn't calculate the specificity and PPVs of the tests studied
Negative predictive value of unspun leukocyte count >10 microliter, positive cytocentrifuge gram stain, 2 to 5 or more leukoctyes/hpf in sediment, positive leukocyte esterase, positive blood, positive nitriteUnspun leukocyte count >10 microliter 92%, positive cytocentrifuge gram stain 92%, 2 to 5 or more leukoctyes/hpf in sediment 90%, positive leukocyte esterase 90%, positive blood 88%, positive nitrite 85%
Pfaller et al
1987
USA
340 patients seen in the emergency treatment centre of the University of Iowa Hospitals and ClinicsDiagnostic studySensitivity, specificity, PVP (predictive value of positive test) and PVN (predictive value of a negative test) of LE guided culture (>=105CFU)LE guided culture: Sensitivity 56%, Specificity 100%, PVP 100%, PVN 84.5%Small sample Non-random sampling technique No attempt to exclude patients with recent antibiotic use Urine specimens evaluated by ECT staff and not lab technicians Does not give basic details of participants (e.g. age)
Chamber count (>10WBC/mm3) guided cultureChamber count guided culture: Sensitivity 60%, specificity 100%, PVP100%, PVN 85.7%
Whiting et al
2005
UK
70 diagnostic cohort studies, including at least some children <5years with suspected UTI, index tests: mixcroscopy or dipstick tests to diagnose UTI or evaluating methods of obtaining urine specimen, with the reference standard as cultureSystematic literature review Level 1aNitrite (NIT) positive likelihood ratio (PLR) and negative likelihood ratio (NLR) and confidence intervals (CI)NIT alone: PLR 15.9, (95% CI:10.7,23.7). NLR .51, 95% CI: 0.43-0.60). Suggesting that it is useful for ruling in disease but not in ruling out disease.As mentioned by the authors one of the weaknesses of this study was publication bias in that those studies reporting higher lower estimates of performance are less likely to be published. Heterogeneity between studies. Studies included did not specify the results with respect to patient age, therefore unable to do sub-group analyses to find out how age affects test accuracy.
LE positive likelihood ratio (PLR) and negative likelihood ratio (NLR) and confidence intervals (CI)LE alone:PLR 5.5, 95% CI 4.1,7.3. NLR .26, 95% CI 0.18,0.36. Suggesting that LE alone is poor for ruling in or ruling out disease

Comment(s)

The well conducted systematic review byWhiting et al. is the most recent in a long line of research done investigating the diagnostic accuracy of dipstick analysis and microscopy for UTI in children. The overall conclusion is that negative dipstick analysis for leukocyte esterase and nitrite, or microscoopy negative for bacteriuria and pyuria can be used to rule out UTI, without the need for confirmatory culture. Positive dipstick analysis and microscopy can also be used to instigate further investigation, by sending for urine culture to confirm UTI. The value of these conclusions to clinical practice is that use of such rapid diagnmostic tests would result in saving hospital charges by not having to send for culture if tests are negative, and also result in fewer children receiving inappropiate antibiotic therapy while awaiting tests results of culture. Other papers included in the BET authenticate what has been found by the most recent sytematic literature review and meta-analysis by Whiting et al.

Clinical Bottom Line

Both dipstick analysis and microscopy are useful screening tests for the assessment of UTI in children.

References

  1. Smith P, Morris A, Reller LB. Predicting urine culture results by dipstick testing and phase contrast microscopy. Pathology. 2003 35;2:161-5
  2. Dipstick only urinalysis screen for the pediatric emergency room [Database of Abstracts of Reviews of Effects Abstract 977032]. Craver R D, Abermanis J G. Dipstick only urinalysis screen for the pediatric emergency room. Pediatric Nephrology 1997;11(3):331-333. In: The Cochrane Library, Issue 2, 2005.
  3. American Academy of Pediatrics. Committee on Quality Improvement. Practice parameter: the diagnosis, treatment, and evaluation of the initial urinary tract infection in febrile infants and young children. Pediatrics 103(4 Pt 1):843-52, 1999 Apr.
  4. M Hiraoka, Y Hida ,Y Mori, H Tsukahara,Y Ohshima, H Yoshida, M Mayumi. Quantitative unspun-urine microscopy as a quick, reliable examination for bacteriuria. Scandinavian Journal of Clinical & Laboratory Investigation. 65(2):125-132, 2005.
  5. W Proesmans, L Standaert, M Vlaeyen, R Eeckels. Urine microscopy in the diagnosis of bacteriuria. Acta Paediatrics Belgica Vol. 33(2)(pp 119-120), 1980.
  6. N Sharief, M Hameed, D Petts. Use of rapid dipstick tests to exclude urinary tract infection in children. British Journal of Biomedical Science. Vol. 55(4)(pp 242-246), 1998.
  7. Thysell N. A comparison between Albustix, Hema-Combistix, Labstix, the sulphosalicyclic-acid test, Heller's nitric-acid test, and a biuret method. Diagnosis of proteinuria. Acta Medica Scandinavica. 185(5):401-7, 1969 May.
  8. R Novak, K Powell, N Christopher. Optimal diagnostic testing for urinary tract infection in young children. Pediatric and Developmental Pathology Mar 2004 (3):226-230.
  9. Pfaller M, Ringberg B, Rames L, Hegeman J, Koontz F. The usefulness of screening tests for pyuria in combination with culture in the diagnosis of urinary tract infection. Diagnostic Microbiology and Infectious Disease March 1987 6(3):207-15.
  10. Penny Whiting, Marie Westwood, Ian Watt, Julie Cooper, and Jos Kleijnen Rapid tests and urine sampling techniques for the diagnosis of urinary tract infection (UTI) in children under five years: a systematic review [pre-publication paper ONLINE accessed 14/06/05]. Biomed Central Paediatrics 2005;5:4.