ORIGINAL ARTICLE. Shidan Tosif, 1 Alice Baker, 1 Ed Oakley, 2,4,5 Susan Donath 2,3,6 and Franz E Babl 1,2,6

Transcription

1 bs_bs_banner doi: /j x ORIGINAL ARTICLE Contamination rates of different urine collection methods for the diagnosis of urinary tract infections in young children: An observational cohort studyjpc_ Shidan Tosif, 1 Alice Baker, 1 Ed Oakley, 2,4,5 Susan Donath 2,3,6 and Franz E Babl 1,2,6 1 Emergency Department, 2 Murdoch Children s Research Institute, 3 Clinical Epidemiology and Biostatistics Unit (CEBU), Royal Children s Hospital Melbourne, 4 Emergency Department, Monash Medical Centre, 5 Faculty of Medicine, Nursing and Health Science, Monash University and 6 Department of Paediatrics, The University of Melbourne, Melbourne, Victoria, Australia Aims: The optimal method for diagnostic collection of urine in children is unclear. National Institute of Health and Clinical Excellence recommend specimens taken by clean catch urine (CCU) for identification of urinary tract infection (UTI). We investigated contamination rates for CCU, suprapubic aspiration (SPA), catheter specimen urine (CSU) and bag specimen urine (BSU) collections. Method: Retrospective observational cohort study with review of microbiology data and medical records at a large tertiary children s hospital. We reviewed urine culture growth from consecutive first urine specimens of children aged <2 years, over a 3-month period in Patient demographics, collection method, location (emergency department, inpatient ward), culture growth, history of UTI, urogenital tract abnormality and antibiotic use were assessed. Contamination rates for collection methods were compared using logistic regression. Results: Urine culture specimens of 599 children (mean age 7.0 months, 54% male) were included. There were 34% CCU, 16% CSU, 14% SPA, 2% BSU and 34% with unknown sample method. Contamination rates were 26% in CCU, 12% in CSU (odds ratio (OR) 0.4, 95% confidence interval (CI) ) and 1% in SPA (OR % CI ). Concurrent antibiotics use was associated with a lower contamination rate. Contamination rates were not associated with age, sex, location, history of UTI or urogenital abnormalities. Conclusion: Contamination rates in CCU are much higher than in CSU and SPA samples. Ideally, SPA should be used for microbiological assessment of urine in young children. Collection procedures need to be optimised if CCU is used. Key words: child; clean catch urine; emergency department; suprapubic aspiration; urinary tract infection. What is already known on this topic 1 The exclusion of urinary tract infections (UTIs) is a key step in the work up of young febrile children. 2 There are a number of collection methods with variable degrees of invasiveness. Data on cross comparisons of different urine collection methods are limited. 3 Recent National Institute of Health and Clinical Excellence guidelines suggest the use of clean catch urine (CCU) samples as the recommended method of urine collection in the assessment of UTI. What this paper adds 1 In a large microbiological and clinical dataset of children less than 24 months of age, contamination rates of CCU (26%) were significantly higher than samples obtained by urinary catheter (12%) and suprapubic aspiration (SPA; 1%). 2 Ideally, SPA should be used for microbiological assessment of urine in young children. The collection procedure needs to be optimised if clean catch is to be used. 3 Documentation of collection methods for all urine specimens should be improved through education and administrative processes. Correspondence: Dr Shidan Tosif, Emergency Department, Royal Children s Hospital Melbourne, Flemington Road, Parkville, Vic. 3052, Australia. Fax: ; shidan.tosif@rch.org.au Conflict of interest: None. Ethics approval: The study was approved by the institutional ethics committee of the Royal Children s Hospital, Melbourne, Australia. No patient consent was obtained as the study was based on retrospectively obtained extant data. Accepted for publication 23 August Urinary tract infections (UTIs) are common in children under 2 years of age, and have been reported to affect between 5 and 7% of febrile children in this age group. 1,2 A diagnosis of UTI carries implications for treatment, further investigations and long-term morbidity. 3,4 A diagnosis of UTI requires a urine culture result of a sufficient bacterial colony count without evidence of contamination. Contaminated urine specimens can prevent clinicians from definitively confirming or excluding a UTI and impact on management decisions for treatment or 659

2 further investigations. This is particularly important in children less than 2 years of age in whom obvious urinary tract symptoms may be absent. While specimens can be obtained by a number of methods including bag specimen urine (BSU), clean catch urine (CCU), catheter specimen urine (CSU) and suprapubic aspiration (SPA), the most appropriate method for the diagnostic collection of urine and analysis of results is subject to ongoing debate. A recent guideline commissioned by the National Institute of Health and Clinical Excellence (NICE) 5 in the UK stated that a clean catch sample is the recommended method for urine collection. The guideline refers to a systematic review 6 that discussed five articles examining the accuracy of CCU when compared with SPA as the reference standard. The review reported good correlation between CCU and SPA in detecting cases of UTI; however, it did not examine contamination. The one article examining CCU contamination rates referred to in the guideline reported no contamination of the 23 CCU samples in their study, and 48% in BSU. 7 However, in other articles, there is considerable heterogeneity in reports of contamination. A prospective trial reported contamination rates of 15% for CCU, 27% for BSU and 29% for pad. 8 A retrospective cohort study reported contamination for CSU at 9% and BSU at 63%. 9 Three direct comparison studies were described in a recent review article, 10 which reported contamination rates between 0 and 7% for SPA and 22 and 23% for urinary catheter The same review article reported on contamination rates interpreted from original data from five studies in which children under 5 years of age had urine collection by paired clean catch and either catheter or SPA. In a secondary analysis, the reviewers reported highly variable contamination rates as ranging from 5 to 85% for CCU and 6 to 57% for CSU, and 0% contamination for SPA. To date, few studies have analysed real-life contamination rates. We set out to investigate contamination rates for CCU, SPA, CSU and BSU specimen collections at a large tertiary children s hospital. Our hypothesis was that contamination rates for urine collection in young children are highly variable based on urine collection methods. The aim was to determine the optimal collection method based on contamination rates. Methods Design and setting We conducted a retrospective cohort study of patients with urine cultures collected at a tertiary care children s hospital with inpatient admissions, outpatient visits and emergency department (ED) visits per year. The study was approved by the Institutional Ethics Committee at Royal Children s Hospital (RCH), Melbourne, Australia. Study population, inclusion and exclusion Microscopy and culture results from all urine specimens collected at RCH from 1 February to 31 April 2008 were retrieved from the hospital s microbiology database. Specimens from children less than 24 months were included for analysis. The first urine specimen collected from each child in this time period was included for analysis. Urine culture results from CCU, SPA, CSU and BSU were analysed. Urine samples were excluded if obtained by other methods such as vesicostomy or nephrostomy. Data collection Patient age, gender, location and method of collection were gathered from the microbiology database. For specimens where the method of collection was by SPA, CCU or CSU, the medical history was reviewed for documentation of a past history of UTI, urogenital abnormality and antibiotic use at time of urine specimen collection. Specimens with unknown method of collection and BSU were not included in this additional analysis. Histories were reviewed by two primary investigators using a standardised form that was piloted on 10 histories, for which there was no disagreements in data extraction. Where no comment was made in the record about past history, this was assumed to be negative. Definitions At this hospital, urine specimens are processed and reported on as per laboratory guidelines, outlined in Table 1. A urine culture was regarded as contaminated if reported as both mixed growth, which equates to two or more organisms grown, as well as a colony count above a certain threshold based on collection method used: 10 4 colony forming unit (CFU)/L for SPA or CSU, 10 8 CFU/L for CCU, indwelling catheters and unspecified samples (UK) and 10 8 CFU/L for BSU. A designation of no significant growth is given to intermediary bacteria counts in CCU and BSU, which fall between the CFU/L cut-offs for infection or no growth. Repeat collection of urine specimens are recommended in these situations. Pure growth of non-bacterial organisms (e.g. candida) were not included as contamination. These CFU cut-offs are broadly comparable with bacterial counts quoted in the NICE 5 guideline and those of the American Academy of Pediatrics Subcommittee on UTI Clinical Practice Parameters. 19 Table 1 Laboratory definition of urine culture results based on collection method at Royal Children s Hospital, Melbourne, Australia Methods of collection Urinary tract infection (pure growth of single organism) Contamination (growth of two or more organisms) SPA 10 4 CFU/L 10 4 CFU/L CSU 10 4 CFU/L 10 4 CFU/L CCU 10 8 CFU/L 10 8 CFU/L Unknown method 10 8 CFU/L 10 8 CFU/L BSU 10 8 CFU/L 10 8 CFU/L BSU, bag specimen urine; CCU, clean catch urine; CFU/L, colony forming units per litre; CSU, catheter specimen urine; SPA, suprapubic aspiration. 660

3 Table 2 Methods of urine collection, contamination rates and statistical analysis Total no. (%) Contaminated no. (%) OR (95% CI) P-value Method of collection CCU 202 (34) 52 (26) 1 CSU 97 (16) 12 (12) 0.41 (0.21, 0.81) 0.01 SPA 84 (14) 1 (1) 0.03 (0.00, 0.26) <0.005 BSU 13 (2) 6 (46) 2.47 (0.79, 7.69) 0.12 Unknown 203 (34) 40 (20) 0.71 (0.44, 1.13) 0.15 Sex Male 324 (54) 66 (20) 1 Female 275 (46) 45 (16) 0.76 (0.50, 1.16) 0.21 Age group (months) <3 218 (36) 46 (21) (20) 19 (16) 0.72 (0.40, 1.29) (22) 19 (15) 0.64 (0.36, 1.15) (22) 27 (20) 0.95 (0.56, 1.62) 0.86 Patient location Emergency department 414 (69) 73 (18) 1 Inpatient wards 145 (24) 22 (15) 0.84 (0.50, 1.40) 0.50 Other/unknown 40 (7) 16 (40) 3.11 (1.58, 6.15) <0.005 BSU, bag specimen urine; CCU, clean catch urine; CI, confidence interval; CSU, catheter specimen urine; OR, odds ratio; SPA, suprapubic aspiration. Outcome measures Primary outcome measure was the contamination rate for each urine specimen collection method. We also sought to adjust for possible confounding factors by investigating the effect of age, gender, location, past history of UTI and urogenital abnormality, and antibiotic use at time of urine specimen collection. Data analysis Logistic regression was used to compare contamination rates between collection methods, using CCU as the comparator for odds ratios (ORs) as it is the recommended method in the NICE guideline. The effect of potential confounding from the variables listed previously was also examined. Statistical calculations were performed using Stata 11.0 (Stata Corp, College Station, TX, USA) Results We identified 818 urine samples sent for urine culture to the RCH microbiology laboratory from children under 24 months during the study period. Two hundred sixteen urine samples were excluded, as we sought only the first specimen from each child. One nephrostomy and two vesicostomy urine samples were also excluded. The study sample of 599 urine results was further analysed (Table 2). Fifty-four per cent were male. The age range of children was 0 23 months with a mean age of 7 months (male 8 months, female 6 months). The largest age group (215, 36%) were infants aged less than 3 months. Overall, most urine specimens were obtained by CCU (34%), followed by CSU and SPA. Relatively few urine specimens were obtained by BSU. For a high proportion of specimens, the collection method had not been recorded (34%). Most urine collections were obtained in the ED. ED collections were by CCU (39%), CSU (15%), SPA (16%), BSU (0%) and unknown (30%), whereas inpatient collections were by CCU (19%), CSU (23%), SPA (11%), BSU (4%) and unknown (42%). Urine collections were by SPA at 19, 13, 14 and 8% in children at <3, 3 6, 7 12 and months, respectively. The contamination rates of different methods of urine collection are shown in Table 2. Contamination rates were 26% in CCU versus 12% in CSU (OR 0.4, 95% confidence interval (CI) ) and 1% in SPA (OR 0.03, 95% CI ). The few bag specimen samples showed a high rate of contamination (46%), and urine collected by an unspecified method had a contamination rate of 20%. Analysis of additional patient variables is shown in Table 3. Of the 383 patients with urine samples collected by SPA, CCU or CSU, 369 medical records were available for review and 14 were unavailable. Five per cent of children in this subset had prior UTI, 14% were receiving antibiotics when the specimen had been obtained and 11% had a known urogenital abnormality. Only antibiotic use at the time of urine collection was associated with a differential, lower contamination rate (OR 0.18, 95% CI ). Results by multivariable logistic regression when adjusted for age, gender, patient location and antibiotic use showed similar results (data not shown) as the unadjusted results shown in Table 2. Discussion This study is the first to provide real-life comparative contamination rates of the three main methods of urine collection 661

4 Table 3 Patient variables and effect on contamination, statistical analysis Total no. (%) Contaminated no. (%) OR (95% CI) P-value Past urinary tract infection No 349 (95) 59 (17) 1 Yes 20 (5) 4 (20) 1.23 (0.40, 3.81) 0.72 Antibiotic used at time No 319 (86) 61 (19) 1 Yes 50 (14) 2 (4) 0.18 (0.04, 0.75) 0.02 Urogenital abnormality No 329 (89) 57 (17) 1 Yes 40 (11) 6 (15) 0.84 (0.34, 2.10) 0.71 CI, confidence interval; OR, odds ratio. in children less than 2 years of age, and reflects the urine collection practices of a tertiary children s hospital ED and inpatient setting. Our results indicate that the method of urine collection significantly affects contamination rates. As expected, bag urine specimens were found to have an unacceptably high contamination rate of 46%. Although it is advocated as an appropriate collection method for children in the current NICE guideline, 5 clean catch specimens were contaminated at a very high rate of 26%. CSU was contaminated in 12% and SPA with a low rate of 1%. There was no statistically significant effect observed with age, gender, location and history of previous UTI or urogenital tract abnormality on contamination rates. Use of antibiotics was associated with a lower contamination rate, which is likely due to residual antimicrobial activity that kills contaminant flora. There are some limitations to this study. It did not include the clinical situation in which urine was collected, including the number of attempts required for urine collection or if one failed technique had preceded another. Because of the retrospective methodology, we were also unable to reliably extract the precise clinical indications for urine testing, which may have influenced collection method and technique. Data were not available about the use of urine dipstick tests as a method of pre-screening specimens by the treating clinician prior to being sent for culture that may have introduced bias. Children thought to be at high risk for UTI may have been more likely to have urine collected by SPA, and some urine samples may have been sent to the microbiology laboratory even though the main purpose for collection was biochemical markers, which did not require a sterile sample. The retrospective design also did not allow an assessment of the clinician response to the results of contaminated specimens nor the consequences for the children concerned. In addition, a number of data points, such as history of UTIs or urogenital abnormalities, were dependent on the recording by the treating clinicians, and although we used a number of strategies for high quality chart review as set out by Gilbert et al., 20 such as use of a piloted abstraction tool, repeat abstraction and repeat entry of a portion of records, the abstractors were not blinded to the purpose of the study. However, the strength of our data in comparison with other smaller, prospective studies under research conditions is that it is potentially more likely to reflect real-life practice. Prior studies of contamination rates have been generally small scale, mainly obtained in controlled study settings. For example, the study referred to in the NICE guideline examining CCU contamination rates was a prospective study comparing CCU and BSU in inpatients in a UK hospital. 7 Twenty-three urine samples by CCU were compared with 23 BSU. None of the CCU samples were contaminated with faecal bacteria, and 48% of BSU were contaminated. Given the controlled setting of this small study, there remain questions about contamination rates by CCU outside of a study environment. There are highly variable contamination rates published in the literature, especially for CCU where rates have ranged from 0 to 85%. 7,10,14 18 This is possibly due to variable collection methods and definitions for outcomes, making results difficult to compare between studies. While our data are based on practice by different medical and nursing practitioners of variable training, skills and experience using variable techniques at a large children s hospital, they may reflect practice in other similar settings. Notably, the contamination rates for SPA in our study are similar to the low rates in other reports ,19 There is no definitive, universally accepted definition for what constitutes contamination. In this study, we have chosen to include as contamination only high CFU growths in urine cultures with multiple organisms. While growth of multiple pathogens is consistent with previously reported definitions of contamination, 21,22 we did not include culture growths of low colony counts. Low colony counts can be associated with contamination or asymptomatic bacteriuria. 23 By excluding low colony counts, we intended to only analyse urine samples most likely to represent true contamination. Uncontaminated urine specimens are essential for the accurate diagnosis of UTI. Although non-invasive methods such as clean catch and bag specimen are often used due to clinician preferences and concerns of discomfort for the patient, urine cultures that are contaminated lead to uncertainty regarding which children to treat or investigate further. 24 While acknowledging that the NICE guideline is mainly aimed at the primary care physician, the clinical dilemma of ambiguous urine culture results exists across all care settings, and the need for accurate samples is potentially more important where the child can not be observed 662

5 over time. It has previously been demonstrated that SPA or CSU urine specimens can provide a definitive diagnosis of UTI. 25,26 Our data suggest that in children under 24 months of age, the only technique for consistently obtaining uncontaminated urine is by SPA. We suggest that in settings where SPA can be conducted, for example, in EDs and in hospital settings, it should be preferentially used. In settings where this is not possible due to skills and resources (e.g. use of ultrasound to identify the bladder), the clean catch technique should be used only after clinicians and parents are educated about the optimal techniques. There is, however, no standardised method for obtaining a CCU specimen and adherence to maintaining a sterile environment after cleaning the genital area is problematic. At the study hospital, the clinical practice guidelines states that the genitalia should be washed with water and dried and that the first few milliliters should be discarded. 27 Data regarding the adherence to this guidance or quality of cleansing and maintaining a sterile environment are not known. These results suggest that further education on how CCU specimens are obtained is required, as previous studies have demonstrated reduced bacterial growth following cleansing of the perineum prior to clean catch. 16 Repeat analysis of contamination rates with standardisation of collection techniques and training would add to this study. Our results also identified a high rate of unrecorded methods of urine specimen collection, which is consistent with past audits of the recording of the methods used to collect urine specimens. 28 This has a notable impact on laboratory culture media selection, cut-off values for contamination and diagnosis of UTI. One strategy to improve the recording of the method of collection would be that hospital laboratories refuse to process urine specimens without a specified method of collection clearly stated. Conclusion This study demonstrates an unacceptably high rate of contamination in urine culture from specimens obtained by CCU as compared with CSU and SPA in particular. While NICE UTI recommendations favour the use of CCU, our data indicate a potential for ambiguous results, need for repeat samples and potentially unnecessary treatment and investigation. Ideally, SPA should be used for microbiological assessment of urine in young children. The collection procedure needs to be optimised if clean catch is to be used. Acknowledgement We acknowledge grant support from the Murdoch Children s Research Institute, Melbourne, Australia and the Victorian Government s Operational Infrastructure Support Programme. References 1 Hoberman A, Wald ER. Urinary tract infections in young febrile children. Pediatr. Infect. Dis. J. 1997; 16: Shaikh N, Morone NE, Bost JE et al. Prevalence of urinary tract infection in childhood: a meta-analysis. Pediatr. Infect. Dis. J. 2008; 27: Ransley PG, Risdon RA. Reflux nephropathy: effects of antimicrobial therapy on the evolution of the early pyelonephritic scar. Kidney Int. 1981; 20: Becker GJ. Reflux nephropathy: the glomerular lesion and progression of renal failure. Pediatr. Nephrol. 1993; 7: National Institute for Health and Clinical Excellence. Urinary Tract Infection in Children: Diagnosis, Treatment and Long Term Management. Clinical Guideline no. 54, Available from: [accessed March 2012]. 6 Whiting P, Westwood M, Bojke L et al. Clinical and cost-effectiveness of tests for the diagnosis and evaluation of urinary tract infection (UTI) in children: a systematic review and economic model. Health Technol. Assess. 2006; 10: McKune I. Catch or bag your specimen? Nurs. Times 1989; 85: Alam M, Coulter B, Pacheco J et al. Comparison of urine contamination rates using three different methods of collection: clean-catch, cotton wool pad and urine bag. Ann. Trop. Paediatr. 2005; 25: Al-Orifi F, McGillivray D, Tange S et al. Urine culture from bag specimens in young children: are the risks too high? J. Pediatr. 2000; 137: Winckworth L, Ayat R. Question 1: urethral catheter or suprapubic aspiration to reduce contamination of urine samples in young children? Arch. Dis. Child. 2009; 94: Tobiansky R, Evans N. A randomized controlled trial of two methods for collection of sterile urine in neonates. J. Paediatr. Child Health 1998; 34: Austin BJ, Bollard C, Gunn TR. Is urethral catheterization a successful alternative to suprapubic aspiration in neonates? J. Paediatr. Child Health 1999; 35: Pryles CV, Atkin MD, Morse TS et al. Comparative bacteriologic study of urine obtained from children by percutaneous suprapubic aspiration of the bladder and by catheter. Pediatrics 1959; 24: Ramage IJ, Chapman JP, Hollman AS et al. Accuracy of clean-catch urine collection in infancy. J. Pediatr. 1999; 135: Braude H, Forfar JO, Gould JC et al. Diagnosis of urinary tract infection in childhood based on examination of paired non-catheter and catheter specimens of urine. BMJ 1967; 4: Pryles CV, Luders D, Alkan MK. A comparative study of bacterial cultures and colony counts in aired specimens of urine obtained by catheter versus voiding from normal infants and infants with urinary tract infection. Pediatrics 1961; 27: Lau AY, Wong SN, Yip KT et al. A comparative study on bacterial cultures of urine samples obtained by clean-void technique versus urethral catheterisation. Acta Paediatr. 2007; 96: Hardy JD, Furnell PM, Brumfitt W. Comparison of sterile bag, clean catch and suprapubic aspiration in the diagnosis of urinary infection in early childhood. Br. J. Urol. 1976; 48: Practice parameter: the diagnosis, treatment, and evaluation of the initial urinary tract infection in febrile infants and young children. Pediatrics 1999; 103: Gilbert EH, Lowenstein SR, Koziol-McLain J et al. Chart reviews in emergency medicine research: where are the methods? Ann. Emerg. Med. 1996; 27: Hoberman A, Wald ER, Reynolds EA, Penchansky L, Charron M. Pyuria and bacteriuria in urine specimens obtained by catheter from young children with fever. J. Pediatr. 1994; 124: Wingerter S, Bachur R. Risk factors for contamination of catheterized urine specimens in febrile children. Pediatr. Emerg. Care 2011; 27: Hoberman A, Wald ER, Reynolds EA et al. Is urine culture necessary to rule out urinary tract infection in young febrile children? Pediatr. Infect. Dis. J. 1996; 15:

6 24 Long E, Vince J. Evidence behind the WHO guidelines: hospital care for children: what are appropriate methods of urine collection in UTI? J. Trop. Pediatr. 2007; 53: Downs M. Technical report: urinary tract infections in febrile infants and young children. Pediatrics 1999; 103: e McGillivray D, Mok E, Mulrooney E et al. A head-to-head comparison: clean-void bag versus catheter urinalysis in the diagnosis of urinary tract infection in young children. J. Pediatr. 2005; 147: Clinical Guidelines Group. Urinary Tract Infection Guideline. Melbourne: Royal Children s Hospital. Available from: [accessed July 2011]. 28 Verrier-Jones K, Hockley B, Scrivener R, Pollock JI. Diagnosis and Management of Urinary Tract Infections in Children under Two Years: Assessment of Practice against Published Guidelines. London: Royal College of Paediatrics and Child Health,