Article, Urology

Performance characteristics of urinalyses for the diagnosis of pediatric urinary tract infection

a b s t r a c t

Purpose: The purpose of this study is to determine whether point-of-care (POCT) urinalysis (UA) is as accurate as laboratory-performed UA in diagnosing Urinary tract infections in the pediatric emergency department (PED). Basic procedures: This was a retrospective series of children (0-18 years old) seen at a tertiary care PED from July 2008 to December 2012 in whom UA and urine culture were obtained. Urinalyses were considered positive if Leukocyte esterase and/or nitrites were positive. Performance characteristics for the 2 types of UAs were calculated using urine culture as the reference standard.

Main findings: A total of 43 452 specimens were sent for laboratory-performed UA and culture, and 6492, for POCT UA and culture (in 2908 specimens, both UAs were performed). Sixty-four percent of specimens were from girls, 51% were catheterized, and 7.5% had UTIs. The sensitivity of POCT UAs and laboratory-performed UAs was 82.5% (confidence interval [CI], 79.4%-85.3%) and 89.1% (CI, 86.4%-88.8%), respectively. The superior performance of laboratory-performed UAs was driven by the sensitivity of microscopy. Laboratory-performed UAs were more sensitive than the POCT in girls (90.6% [CI, 89.4%-91.8%] vs 82.8% [79.4%-85.8%]).

Principal conclusions: Although POCT UAs offer more rapid turnaround times, the sensitivity is greater for laboratory- performed UAs. Given the difficulty in following up PED patients after discharge as well as the potential morbidity from untreated UTIs, the rapidity of the POCT UA must be balanced against the lower sensitivity of this assay. The benefit of more accurate diagnosis may outweigh the potentially longer PED length of stay associated with a laboratory-performed UA.

(C) 2013

Introduction

Urinary tract infections are the most common bacterial infection in febrile infants, seen in up to 10% [1]. The nonspecific symptoms of UTI in the young child, coupled with inability to discern dysuria symptoms in the young child, lead to widespread testing for UTI using urinalysis (UA). However, UAs are not perfect. False positives may occur due to asymptomatic colonization with non- pathogenic bacteria [2], eosinophilia, Trichomonas, and concomitant use of certain antibiotics (carbapenems, clavulanate) [3] as well as with inappropriate cleaning before voiding. False-negative UAs are seen in young infants who may not be able to respond to UTIs with pyuria [4]. One meta-analysis showed that Urine dipsticks had a sensitivity of 88% if either leukocyte esterase or nitrite was positive, and sensitivity of enhanced urinalysis (which includes Gram stain of fresh unspun urine) was 93% [5].

? Sources of support: This study was supported by internal funding.

* Corresponding author. Tel.: +1 832 824 5582; fax: +1 832 825 1182.

E-mail address: [email protected] (A.T. Cruz).

Since 2001, there has been a shift toward performing Point-of-care tests (POCTs) as a means to rapid diagnosis of several conditions [6]. Perhaps the area in which this has been most useful is the emergency department (ED), where both promptness of diagnosis and accuracy may impact a clinician’s decision to use a given test. Point-of-care test UAs may decrease length of stay [7,8], and certain enzymes may be better measured in the real-time fashion of POCTs. However, the decision to use POCT vs laboratory-based UAs should not be driven solely by rapid diagnosis; the sensitivity and specificity of these tests in the clinical setting should also be carefully considered. The goal of this investigation was to delineate the performance characteristics of POCT and laboratory-based UAs for UTI in a pediatric population.

Methods

This was a retrospective series of children seen at Texas Children’s Hospital (Houston, TX) from July 2008 to December 2012 in whom both a UA and urine culture were obtained. Urinalysis was defined as either a urine dipstick as a POCT or UA in the laboratory (termed laboratory-performed UA). During the study

0735-6757/$ – see front matter (C) 2013 http://dx.doi.org/10.1016/j.ajem.2013.06.037

1406 B.A. Kazi et al. / American Journal of Emergency Medicine 31 (2013) 14051407

period, 2 urine POCTs were available: Urisys 1000 Urine Analyzer (Roche, Indianapolis, IN) from January 1, 2008, to October 31, 2011, and Clinitek Status Analyzer (Siemens, Tarrytown, NY) from November 1, 2011, to December 31, 2012. The laboratory-performed UA used Iris 9EB sticks (Iris Diagnostics, Chatsworth, CA). Urinalyses were considered positive if leukocyte esterase and/or nitrites were positive. Laboratory-performed UAs also were considered positive if microscopy showed more than 10 white blood cells per high-power field. Gram stains of unspun urine were not routinely performed during the study period.

urine cultures were considered positive if voided or catheterized specimens grew at least 50000 colony-forming units per milliliter of a single organism or suprapubic aspirates grew at least 1000 colony- forming units per milliliter [9]. Urine cultures were held for up to 3 days after being plated on blood agar and MacConkey agar plates. Specimens received more than 2 hours or more than 24 hours refrigerated after collection were not processed. The following were not considered UTIs: specimens growing more than 1 organism, Lactobacillus, Corynebacterium species, or coagulase-negative staphy- lococci in children 2 to 24 months old. Urine culture specimens collected via a bag, Foley catheter, indwelling stent, or urinary tract fistula were excluded.

The performance of UAs was compared to the reference standard of urine culture. Sensitivity, specificity, and positive/negative predictive values were calculated using Stata 11 (Stata, Inc, College Station, TX). Institutional review board approval was obtained before study initiation.

Results

During the study period, 43 452 specimens were sent for laboratory-performed UA, and 6492, for POCT UA (in 2908 specimens, both UAs were performed). Sixty-four percent were female. Specimen sources included 24146 (51.4%) catheterized, 22751 (48.4%) voided, and 99 (0.2%) suprapubic aspirations. Urine cultures were consistent with UTI in 3503 (7.5%). Urinary tract infections were found most frequently in children 6 to 11 months old (8.4%), followed by 2 to 5

months old (8.1%), older than 24 months (7.6%), 12 to 24 months old

(6.5%), and younger than 2 months (6%). The organisms were 2729

Escherichia coli (77.9%), 255 Klebsiella species (7.3%), 183 Proteus species (5.2%), 252 other Gram-negative bacteria (7.2%), and 50 gram- positive organisms (1.4%); 34 cultures (1%) grew candidal species.

The sensitivity of POCT UAs and laboratory-performed UAs were 82.5% (confidence interval [CI], 79.4%-85.3%) and 89.1% (CI, 86.4%- 88.8%), respectively (Table 1). There was a trend toward increasing sensitivity with increasing patient age, and the laboratory-performed UA was more sensitive in girls than boys. The POCT UA was much more specific in boys than girls (Table 2).

Discussion

This large retrospective series of children evaluated for UTI demonstrated several findings. First, the sensitivity of the laborato-

ry-performed UA was higher than for the POCT UA. Second, there were twice as many false positives as true positives for both assays, resulting in low positive predictive values. Third, microscopy was the most sensitive parameter for the laboratory-performed UAs. Finally, UA performance varied in patients of different demographics.

The laboratory-performed UA was significantly more sensitive than the POCT UA. This has been noted in prior studies when dipsticks are compared with enhanced UA [5,10]. One study of children younger than 2 years evaluated for UTI in the ED showed sensitivity of enhanced UA to be 94%, compared with a dipstick sensitivity of 79% [10]. However, enhanced UA is not available in all centers, whereas the dipstick and laboratory-performed UA routinely are available to clinicians. Hence, evaluating performance of the latter assays is important. Often in the pediatric ED (PED), children may provide a Urine sample while in the waiting room, and the specimen may sit for prolonged periods before a POCT-UA being performed. This can lead to degradation in leukocyte esterase and red blood cell lysis, either of which may result in false negatives. White blood cell clumping can prevent dispersion of leukocyte esterase and cause false-negative results on POCT UA. However, leukocyte clumping would be evident on microscopy, making it a more sensitive test. Although the rapidity of POCT-UA is attractive for the emergency physician due to the high volume and turnover of patients, sources of false-negative testing (specimen stasis, white blood cell clumping, and high concentrations of substances such as Ascorbic acid, glucose, macrolides, and cephalosporins) must be considered and weighed [3].

The percentage of false-positive results on POCT-UA was found to

be high. In the PED, the aim is often to optimize sensitivity, given that patient follow-up is unpredictable at best. The low positive predictive value of POCT UA may result in a large number of children being unnecessarily treated with antibiotics. This overtreatment can result in at least 2 outcomes. One is the cost to families for purchasing third- generation cephalosporins often used for empiric UTI treatment. The difficulty of interpreting UA results was previously demonstrated in 1 systematic review, which was unable to determine the most cost- effective manner of diagnosing UTIs in children [11]. A second concern surrounds the injudicious use of antibiotics selecting for drug- resistant organisms. In recent years, the spread of drug-resistant organisms such as extended-spectrum ? lactamase (ESBL)-producing organisms has been well documented. Extended-spectrum ? lacta- mase-encoding genes are most common in E coli and Klebsiella species, confer high-grade resistance to cephalosporins, and have higher mortality rates than in patients with susceptible organisms. This has a tremendous impact on the well-being of children with UTIs, with a national surveillance system recently reporting an increase in ESBL-producing bacteremic strains from 27% in 2007 to 44% in 2010 [12]. Over a 1-year period at our hospital, 94 isolates were identified as having at least 1 ESBL gene [13]. The main risk factor shown to be associated with ESBL-producing strains has been the use of broad- spectrum cephalosporins [14].

The positive predictive value of POCT-UA increased with increasing age and with the female sex, thus minimizing the number of false-

Table 1

Performance characteristics of POCT UA and laboratory-based UA compared with urine culture in pediatric patients

Test

Assay type

n

Sensi, % (95% CI)

Speci, % (95% CI)

PPV, % (95% CI)

NPV, % (95% CI)

Overall performancea

Lab-performed UA

43452

89.1 (87.9-90.2)

76.1 (75.7-76.5)

22.7 (22-23.5)

98.9 (98.8-99)

POCT-UA

6492

82.5 (79.4-85.2)

81.3 (80.3-82.3)

33.9 (31.6-36.2)

97.6 (97.1-98)

Leukocyte esterase

Lab-performed UA

43452

79.1 (77.6-80.5)

84.8 (84.4-85.1)

29.1 (28.1-30)

98.1 (97.9-98.2)

POCT-UA

6492

71.7 (68.1-75)

83.1 (82.1-84.1)

33 (30.6-35.5)

96.2 (95.6-96.7)

Nitrites

Lab-performed UA

43452

39.5 (37.8-41.3)

98.1 (98-98.3)

62.7 (60.6-64.9)

95.4 (95.2-95.6)

POCT-UA

6492

47.3 (43.5-51.2)

96.7 (96.2-97.1)

62.4 (58.1-66.6)

94.1 (93.4-94.6)

Microscopy

Lab-performed UA

43452

84.8 (83.5-86)

79.3 (78.9-79.7)

24.4 (23.6-25.2)

98.5 (98.4-98.6)

Abbreviations: NPV, negative predictive value; PPV, positive predictive value; Sensi, sensitivity; Speci, specificity.

a Either leukocyte esterase or nitrites positive for POCT UAs and any of the following positive for laboratory-based UAs: leukocyte esterase, nitrites, or microscopy.

B.A. Kazi et al. / American Journal of Emergency Medicine 31 (2013) 14051407 1407

Table 2

Age- and sex-based performance characteristics of urinalyses*

Population

Assay type

n

Sensi, % (95% CI)

Speci, % (95% CI)

PPV, % (95% CI)

NPV, % (95% CI)

Female

Lab-performed UA

27394

90.6 (89.4-91.8)

70.8 (70.3-71.4)

22.8 (21.9-23.6)

98.8 (98.6-98.9)

POCT-UA

5049

82.8 (79.4-85.8)

78.1 (76.8-79.3)

32.5 (30.1-35)

97.3 (96.7-97.8)

Male

Lab-performed UA

16058

84.5 (81.8-86.9)

84.8 (84.2-85.4)

22.6 (21.2-24.2)

99.1 (98.9-99.2)

POCT-UA

1443

80.6 (71.4-87.5)

92.2 (90.6-93.6)

44.4 (37.1-51.8)

98.4 (97.5-99)

b 2 mo

Lab-performed UA

4033

86.1 (80.9-90.1)

89.3 (88.2-90.2)

33.5 (29.8-37.4)

99 (98.6-99.3)

POCT-UA

39

75 (35.6-95.5)

90.3 (73.1-97.5)

66.7 (30.9-91)

93.3 (76.5-98.8)

2-5 mo

Lab-performed UA

4616

87.2 (83.3-90.3)

87 (85.9-88)

37.7 (34.5-41)

98.7 (98.5-99)

POCT-UA

237

74.1 (53.4-88.1)

93.8 (89.4-96.5)

60.6 (42.2-76.6)

96.6 (92.8-98.5)

6-11 mo

Lab-performed UA

5059

88.1 (84.6-90.9)

85.3 (84.3-86.3)

36.2 (33.3-39.2)

98.7 (98.3-99)

POCT-UA

802

79.2 (67.7-87.5)

96 (94.3-97.3)

66.3 (55.2-75.9)

97.9 (96.5-98.8)

12-23 mo

Lab-performed UA

4813

89.9 (85.9-92.9)

86.8 (85.8-87.8)

32.4 (29.4-35.6)

99.2 (98.8-99.4)

POCT-UA

837

82.5 (70.5-90.6)

96.6 (95-97.8)

66.7 (55-76.7)

98.6 (97.3-99.2)

>=2 y

Lab-performed UA

24931

90 (88.5-91.3)

68.1 (67.4-68.7)

18 (17.2-18.8)

98.9 (98.7-99)

POCT-UA

4577

83.5 (79.9-86.6)

75.1 (73.7-76.4)

29.3 (27-31.7)

97.4 (96.7-97.9)

* Either leukocyte esterase or nitrites positive for POCT UAs and any of the following positive for laboratory-based UAs: leukocyte esterase, nitrites, or microscopy.

positive results obtained with those demographic categories. The reason for this may be that the POCT-UA was developed for use in those patients symptomatic of UTI with dysuria, making the pretest probability of a true-positive result much higher in older, develop- mentally appropriate children who are able to communicate the presence of dysuria [15]. In addition, the virtual inability of infants to mount an inflammatory response to UTIs contributes to a sensitivity of the POCT-UA in younger ages [16]. The PED physician may consider these points when deciding whether to use POCT UA or laboratory- performed UA for certain patient populations.

Microscopy was the most sensitive parameter for laboratory- performed UAs. This may be due to direct visualization of cells contained in urine samples that do not degrade over time, in contrast to enzymes such as leukocyte esterase. Microscopy is not dependent on the age of the urine specimen or on the absence of substances that can falsely alter POCT UA results. It also confers the ability to evaluate for contamination of the specimen by squamous epithelial cells or by protozoa such as Trichomonas. This point has the most relevance for clinicians practicing in centers that do have the capability to conduct laboratory-performed UAs as well as POCT UAs. As the microscopy part of laboratory-performed UA was found to be more sensitive than any part of the POCT UA, the physician may consider using laboratory- performed UAs rather than using a tiered approach, such as first obtaining a POCT UA, then ordering microscopy only for samples with abnormal POCT UAs.

There were limitations to this retrospective study. First, clinical indication for UA testing in individual patients was not known. Second, it is possible that specimens were not received by the laboratory in a timely manner, thus decreasing the sensitivity of laboratory-performed UA. The time from specimen collection to performance of UA was also not known. It is possible that the colder temperature maintained in the laboratory could have affected enzymatic activity. Third, clinical variables impacting UA parameters (eg, neutropenia, chronic steroid use, immunosuppressant therapy) were not evaluated. These could result in decreased rates of pyuria, causing false-negative results. Next, enhanced UA was not available in our center. Finally, results may not be generalizable to laboratories using alternative urinalysis machines or strips.

In conclusion, although POCT-UAs offer more rapid turnaround times, the sensitivity is greater for laboratory-performed UAs. Given the difficulty in following up PED patients after discharge as well as the potential morbidity from untreated UTIs, the rapidity of the POCT-UA must be balanced against the lower sensitivity of this

assay. Future studies could evaluate the benefit of more accurate diagnosis potentially outweighing the longer PED length of stay associated with a laboratory-performed UA, from both clinical and cost perspectives.

References

  1. Zorc JJ, Levine DA, Platt SL, et al. Clinical and demographic factors associated with urinary tract infection in young febrile infants. Pediatrics 2005;116:644-8.
  2. Whiting P, Westwood M, Watt I, Cooper J, Kleijnen J. Rapid tests and urine sampling techniques for the diagnosis of urinary tract infection in children under 5 years: a systematic review. BMC Pediatr 2005;5:4.
  3. Plapp FV. Urinalysis testing and urinary tract infections. Allied Healthcare Group Publication; 2013.
  4. Lin DS, Huang SH, Lin CC, et al. Urinary tract infection in febrile infants younger than eight weeks of age. Pediatrics 2000;105:e20.
  5. Gorelick MH, Shaw KN. Screening tests for urinary tract infection in children: a meta-analysis. Pediatrics 1999;104:e54.
  6. Jang JY, Shin SD, Lee EJ, Park CB, Song KJ, Singer AJ. Use of a comprehensive metabolic panel point-of-care test to reduce length of stay in the emergency department: a randomized controlled trial. Ann Emerg Med 2013;61:145-51.
  7. Fermann GJ, Suyama J. Point of care testing in the emergency department. J Emerg Med 2002;22:393-404.
  8. Ryan RJ, Lindsell CJ, Hollander JE, et al. A multicenter randomized controlled trial comparing central laboratory and point-of-care cardiac marker testing strategies: the Disposition Impacted by Serial Point of Care Markers in Acute Coronary Syndromes (DISPO-ACS) trial. Ann Emerg Med 2009;53:321-8.
  9. Subcommittee on Urinary Tract Infection, Steering Committee on Quality Improvement and Management, Roberts KB. Urinary tract infection: clinical practice guideline for the diagnosis and management of the initial UTI in febrile infants and children 2 to 24 months. Pediatrics 2011;128:595-610.
  10. Shaw KN, McGowan KL, Gorelick MH, Schwartz JS. Screening for urinary tract infection in infants in the emergency department: which test is best? Pediatrics 1998;101:e1-5.
  11. 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.
  12. Castanheira M, Farrell SE, Deshpande LM, Mendes RE, Jones RN. Prevalence of ?- lactamase encoding genes among Enterobacteriaceae bacteremia isolates collect- ed in 26 USA hospitals: report from the SENTRY Antimicrobial Surveillance Program (2010). Antmicrob Agents Chemother 2013;57:3012-20.
  13. Chandramohan L, Revell PA. Prevalence and molecular characterization of extended-spectrum-beta-lactamase-producing Enterobacteriaceae in a pediatric patient population. Antmicrob Agents Chemother 2012;56:4765-70.
  14. Muro S, Garza-Gonzalez E, Camacho-Ortiz A, et al. Risk factors associated with extended-spectrum ?-lactamase-producing Enterobacteriaceae nosocomial Bloodstream infections in a tertiary care hospital: a clinical and molecular analysis. Chemotherapy 2012;58:217-24.
  15. Hamill T. Quality assurance and point of care testing procedures. Siemens Product Insert 2012:7-11.
  16. Crain EF, Gershel JC. Urinary tract infections in febrile infants younger than 8 weeks of age. Pediatrics 1990;86:363-7.