• 
• 

Three Part Question

In [non-toilet-trained children] is [urethral catheterisation as good as suprapubic aspiration] at [obtaining uncontaminated urine samples and so detecting true urinary tract infections]?

Clinical Scenario

You are the paediatric senior house office running the prolonged jaundice clinic. A urine sample from a baby was collected via the "clean catch" method and has a mixed (contaminated) growth on culture. You need to recall the baby for repeat urine culture and wonder whether it is best to perform a suprapubic aspiration to minimise the chance of another contaminated sample or whether an in–out urethral catheter sample would be as good.

Search Strategy

Medline via PubMed was the primary source of articles. Search terms: catheter* AND suprapubic AND (urine OR urinary). Limits were: human, English language and ages 0–5 years. Dates included 1966 to September 2008. An initial search for direct comparisons was undertaken, and because of the low numbers of studies, all papers which compared suprapubic aspiration or urethral catheterisation with clean catch urine samples were also included.
Secondary searches on the Cochrane database, Clinical Evidence and SUMsearch were performed using the same search terms.

Search Outcome

A total of 123 papers were found via PubMed. Studies not comparing suprapubic aspiration and urethral catheter samples were excluded, as were any that did not have culture result as an outcome. This left two relevant direct comparison papers. The references of the above papers were checked, along with the linked articles, and one further article was found (which dated from before 1966).

An additional six articles were found comparing suprapubic aspiration or urethral catheterisation with clean catch sampling. Clean catch was chosen as the comparator as it is the first line collection method recommended by the National Institute for Health and Clinical Excellence (NICE).

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Tobiansky et al, 1998	Preterm babies with suspected sepsis n = 33 UC: n = 16 SPA: n = 17	RCT (level 1b)	Primary: success in obtaining urine Secondary: urine culture results, amount of urine collected, complications	14/17 SPAs successful with 10/17 getting >0.5 ml urine. 1/14 contaminated (7.1%) 13/16 UC got urine, 5/16>0.5 ml. 3/13 contaminated (23.1%) No significant difference in contamination between UC and SPA More success in getting adequate urine volumes in babies over 34 weeks	If the child was better by the time the result was available then it was considered as contaminant but we assumed all patients were started on empirical antibiotics (not specified in paper) which could have changed clinical outcome Babies were cleaned with aqueous chlorhexidine before catheter and SPA samples were taken Study terminated early as little benefit of UC was shown and there were more practical problems with catheters, so potential for type 2 error
Austin et al, 1999	Babies in NICU with suspected sepsis 24–44 weeks n = 90 SPA: 65 attempts UC: 43 attempts (18 after failed SPA)	Observational cohort (level 2b)	Primary: success in obtaining urine Secondary: culture result	33/43 UC successful (one sample lost before culture), 7/32 contaminated (22%) 32/65 SPA successful, 0/32 contaminated	UC successful in more patients than SPA Babies were cleaned with normal saline before catheter samples and with an alcohol swab before SPA No power calculation had been done and clinicians could choose which method they attempted, so no randomisation – most chose SPA as first line unless contra-indicated
Pryles et al, 1959	Children with no clinical evidence of UTI n = 42 All children had both methods simultaneously Urine sample from catheter was split into two aliquots and evaluated separately	Crossover (level 2b)	Primary: bacteriuria Secondary: culture result	1/42 first half UC contaminated (2.4%) 0/42 second half UC contaminated 0/42 SPA contaminated	This was a healthy population, so is it reasonable to extrapolate to babies with potential sepsis? Data were evaluated using the cut-off for significant bacteriuria as 104 and 105 CFU/ml – the results reported here are for 105 (standard NICE definition of UTI) All patients were free from infection at 4 months so any bacteria introduced during the procedures did not lead to UTI

Comment(s)

Up to 40% of urinary tract infections in infants can be asymptomatic and opinion over whether to treat such asymptomatic bacteriuria differs: NICE guidelines say no to treatment (NICE), whereas American practice guidelines recommend treating any culture positive urine. Symptomatic urinary tract infections (UTIs), however, always need treatment, but as infants rarely present with specific signs or symptoms urine samples are often taken as part of wider screens (eg, septic workup or prolonged jaundice screen). It is important not to misdiagnose a contaminated urine sample as a true UTI as children will need appropriate treatment to prevent complications such as renal scarring or urosepsis and may need further investigations, which may be invasive.

Current NICE guidelines (2007) recommend clean catch urine samples, but if this is contaminated what should your next step be? A Cochrane review into urine collection methods is currently ongoing, but no data have yet been published and a diagnostic review in 2006 did not address this question (Whiting).

Suprapubic aspiration (SPA) is the gold standard for uncontaminated urine collection in babies under 1 year old but is the most technically demanding. Although studies have shown that SPA is more painful than urethral catheterisation (UC)(Kozer) and often causes increased parental anxiety, it is significantly quicker to perform (Pollack) (mean time for SPA 16.73 s, for UC 80.7 s, p<0.001) so any discomfort is over faster. Although the Austin et al study showed SPAs were significantly less likely to be successful at obtaining sufficient urine for culture, there are certain strategies to improve success rate including adequate training, ensuring the nappy is dry beforehand and the baby is well hydrated, and using an ultrasound machine to visualise the bladder pre-procedure (which can increase the success rate to 96%) (Kiernan). However, infants with sepsis are more likely to be dehydrated with a smaller bladder volume and so the above strategies may not be enough: one study using blind SPA in infants presenting to A&E with presumed sepsis had only a 46% success rate (Pollack).

In–out catheters are thought to be a less traumatic method of urine collection and can be used by nursing staff. It is harder to maintain sterile conditions during insertion though, leading to contaminated samples. Therefore, samples taken from a catheter should be interpreted against the pre-test probability of urinary tract infection (Cheng). Catheters can also be hard to pass, especially in premature babies, resulting in more urine leakage and a smaller final urine volume collected. One study (Dayan) showed a small but significant increase in contamination in the first portion of the catheterised urine stream when compared to the later portion. However, due to the small volume of urine collected in young infants it may not be possible to separate it into early and late samples. In–out catheters are useful in cases where attempt at SPA has failed or is contra-indicated.

From the three studies which directly compared SPA and UC samples, two (Tobiansky, Pryles)showed no significant increase in sample contamination from catheters compared to SPA, whilst the third study (Austin) had 22% contamination in UC samples and 0% in SPA. Simply grouping the three studies together increased the numbers in each group and, using Fisher’s exact test, shows catheter samples are significantly more likely to be contaminated than SPA samples (p = 0.002). However, differences in methodology between the studies make such inter-study comparisons difficult. For example, the studies varied in their use of cleaning agent. Austin et al cleaned with saline before catheter samples but alcohol before SPA, whereas Tobiansky et al used the same cleaning method before both methods (aqueous chlorhexidine and iodine). It is known that cleaning with soap and water before collecting midstream clean catch samples reduces the contamination rate in toilet-trained populations (Vaillancourt), but there have been no studies examining the effect of cleaning with saline or sterile water versus chlorhexidine or iodine. Therefore, we are unable to comment if this difference could have influenced results and so the studies employing different cleaning methods should be compared with caution.

A further difference was the number of colony forming units (CFU/ml) used to distinguish a true infection from a contaminated sample. The cut-off value used historically is >105 CFU/ml, which is the NICE definition of UTI, but this was from a study with adult patients using catheter samples (Kass) and should only be a guide, depending on method of collection and the clinical state of the patient. This value of pure growth in catheter samples correlates with 95% probability of infection, whereas in an SPA sample, any number of gram negative bacilli or >103 CFU/ml gram positive cocci on culture is more than 99% likely to indicate a true infection (American Academy of Pediatrics). Two of the reviewed papers used >105 CFU/ml as the cut-off for samples taken via SPA, so based on this, the results may have been affected by under-diagnosing the number of infections or contaminations in SPA samples (as any cultures with <105 CFU/ml were counted as culture negative) and hence over-diagnosing the proportional rate of catheter contamination.

In order to gain more evidence to add to the above, results of the six indirect studies were analysed. Any true UTI samples were excluded and then the total number of contaminated samples by each method was totalled as a proportion of the number of samples taken by that method in each study. By combining the results of the studies the contamination ratios could be determined. Only children under 5 years old were included in this table as our original question was about non-toilet-trained infants, and the cut-off for true UTI in the presence of symptoms was any growth in SPA samples and more than 105 CFU/ml of pure growth in catheter or clean catch samples. Any other growth or any mixed growth was considered as contaminant.

The odds of contaminations were compared using the Peto method, with weighted Peto odds ratio for contamination of 0.04 for SPA samples and 0.17 from urethral catheters. This difference in contamination ratios is statistically significant, with increased contamination in catheter samples (p<0.01). By analysing proportions of contamination, rather than absolute numbers, it was hoped to minimise the effects of differences between studies, such as cleaning methods. Although this provides supportive evidence that contamination rates are higher in catheter samples, these numbers can only be used as a guide to trend rather than as absolutes.

Editor Comment

NICE, National Institute for Health and Clinical Excellence; NICU, neonatal intensive care unit; RCT, randomised controlled trial; SPA, suprapubic aspiration; UC, urethral catheterisation; UTI, urinary tract infection.

Clinical Bottom Line

Combined evidence indicates that urethral catheter samples are more likely to be contaminated than samples obtained by suprapubic aspiration (Grade C), but not all evidence is of high quality.

Catheter sample contamination can be reduced by culturing the second half of the sample only. (Grade B)

Catheter samples may still be more appropriate than suprapubic aspiration (number needed to harm (NNH) 9) where it is felt there is a higher chance of success, influenced by clinical factors such as age, the size of the baby, other co-morbidities and potentially parental preference.

References

1. Tobiansky R, Evans N. A randomized controlled trial of two methods for collection of sterile urine in neonates. J Paediatr Child Health 1998;34:460–2.
2. Austin BJ, Bollard C, Gunn TR. Is urethral catheterization a successful alternative to suprapubic aspiration in neonates? J Paediatr Child Health 1999;35:34–6.
3. 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;24:983–91.
4. Conn NK. A study of some of the methods of urinary collection in children. J Clin Pathol 1970;23:81–4.
5. 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;48:279–83.
6. Ramage IJ, Chapman JP, Hollman AS, et al. Accuracy of clean-catch urine collection in infancy. J Pediatr 1999;135:765–7.
7. 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;4:702–5
8. Pryles CV, Luders D, Alkan MK. A comparative study of bacterial cultures and colony counts in paired specimens of urine obtained by catheter versus voiding from normal infants and infants with urinary tract infection. Pediatrics 1961;27:17–28.
9. 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;96:432–6.
10. NICE. Urinary tract infection: diagnosis, treatment and long-term management of urinary tract infection in children . Clinical guideline CG54. London: NICE, August 2007.
11. American Academy of Pediatrics Committee on Quality Improvement Subcommittee on Urinary Tract Infection. Practice parameter: the diagnosis, treatment, and evaluation of the initial urinary tract infection in febrile infants and young children. Pediatrics 1999;103:843–52.
12. Thomas J, Kurien A, Philipraj SJ. Methods for collecting urine samples in adults and children with suspected urinary tract infection (protocol). Cochrane Database Syst Rev 2008;(2):CD006025.
13. Whiting P, Westwood M, Bojke L, et al. Clinical effectiveness and cost-effectiveness of tests for the diagnosis and investigation of urinary tract infection in children: a systematic review and economic model. Health Technol Assess 2006;10:iii–iv, xi–xiii, 1–154.
14. Kozer E, Rosenbloom E, Goldman D, et al. Pain in infants who are younger than 2 months during suprapubic aspiration and transurethral bladder catheterisation: a randomised, controlled study. Pediatrics 2006;118:e51–6.
15. Pollack CV Jnr, Pollack ES, Andrew ME. Suprapubic bladder aspiration versus urethral catheterization in ill infants: success, efficiency and complication rates. Ann Emerg Med 1994;23:225–30.
16. Kiernan SC, Pinckert TL, Keszler M. Ultrasound guidance of suprapubic bladder aspiration in neonates. J Pediatr 1993;123:789–91.
17. Cheng YW, Wong SN. Diagnosing symptomatic urinary tract infections in infants by catheter urine culture. J Paediatr Child Health 2005;41:437–40.
18. Dayan PS, Chamberlain JM, Boenning D, et al. A comparison of the initial to the later stream urine in children catheterized to evaluate for a urinary tract infection. Pediatr Emerg Care 2000;16:88–90.
19. Vaillancourt S, McGillivray D, Zhang X, et al. To clean or not to clean: effect on contamination rates in midstream urine collections in toilet-trained children. Pediatrics 2007;119:e1288–93.
20. Kass EH. Pyelonephritis and bacteriuria. A major problem in preventive medicine. Ann Intern Med 1962;56:46–53.