ALTE indicates acute life-threatening event; UA, urinalysis; and UTI, urinary tract infection.
eAppendix 1. Search Strategy
eAppendix 2. Quality Assessment
eTable. MOOSE Checklist: Impact of Diagnostic Criteria on UTI Prevalence in Bronchiolitis: a Meta-Analysis
Customize your JAMA Network experience by selecting one or more topics from the list below.
McDaniel CE, Ralston S, Lucas B, Schroeder AR. Association of Diagnostic Criteria With Urinary Tract Infection Prevalence in Bronchiolitis: A Systematic Review and Meta-analysis. JAMA Pediatr. 2019;173(3):269–277. doi:10.1001/jamapediatrics.2018.5091
How are diagnostic criteria associated with the reported prevalence of urinary tract infection (UTI) in bronchiolitis?
In this systematic review and meta-analysis, the estimated prevalence of UTI in bronchiolitis was 3.1%, using the individual study definitions of UTI. With the inclusion of a positive urinalysis result (pyuria or nitrites) as a diagnostic criterion, the UTI prevalence was 0.8%.
When a positive urinalysis result is added as a diagnostic criterion for UTI, the estimated prevalence of concomitant UTI in bronchiolitis is less than the threshold for testing of 1% to 3% referenced in the 2011 American Academy of Pediatrics UTI guidelines.
Concomitant urinary tract infection (UTI) is a frequent concern in febrile infants with bronchiolitis, with a prior meta-analysis suggesting a prevalence of 3.3%. However, the definition of UTI in these studies has generally not incorporated urinalysis (UA) results.
To conduct a systematic review and meta-analysis examining the prevalence of UTI in infants with bronchiolitis when positive UA results are incorporated into the UTI definition.
Medline (1946-2017) and Ovid EMBASE (1976-2017) through August 2017 and bibliographies of retrieved articles.
Studies reporting UTI prevalence in bronchiolitis.
Data were extracted in accordance with meta-analysis of observational studies in epidemiology guidelines via independent abstraction by multiple investigators. Random-effects models generated a weighted pooled event rate with corresponding 95% confidence intervals.
Main Outcomes and Measures
Prevalence of UTI.
We screened 477 unique articles by abstract, with full-text review of 30 studies. Eighteen bronchiolitis studies reported a UTI prevalence and 7 of these reported UA data for inclusion in the meta-analysis. The overall reported prevalence of UTI in bronchiolitis from these 18 studies was 3.1% (95% CI, 1.8%-4.6%). With the addition of positive UA results (defined as the presence of pyuria or nitrites) as a diagnostic criterion, the prevalence of UTI as reported in the 7 studies in bronchiolitis was 0.8% (95% CI, 0.3%-1.4%). Sensitivity analyses yielded similar results, including for infants younger than 90 days. Heterogeneous definitions of UTI and UA criteria introduced uncertainty into prevalence estimates.
Conclusions and Relevance
When a positive UA result is added as a diagnostic criterion, the estimated prevalence of concomitant UTI is less than recommended testing thresholds for bronchiolitis.
Multiple studies have reported an association between bronchiolitis and urinary tract infections (UTIs), with some studies suggesting that as many as 7% to 12% of infants with bronchiolitis have a concomitant UTI.1-3 A meta-analysis4 of UTI in infants younger than 3 months with bronchiolitis reported a prevalence of 3.3%.
However, because the sensitivity of the urinalysis (UA) historically has been considered suboptimal,5-7 the diagnosis of UTI in these studies and in most studies of febrile infants has been defined by a positive urine culture alone regardless of the UA results.8,9 However, there is increasing recognition that a positive urine culture in the absence of pyuria on a UA may reflect asymptomatic bacteriuria (AB) or contamination rather than a true UTI.10-13 In response to this concern, in 2011 the American Academy of Pediatrics (AAP) revised the diagnostic criteria for UTIs to include both an abnormal UA result (pyuria or bacteriuria) and a positive urine culture (>50 000 cfu/mL).
For a condition such as bronchiolitis, where the annual disease burden globally is 20 to 50 million children,14 the accurate diagnosis of a concomitant UTI is important given the associated exposure to antibiotic therapy, imaging, potential for hospital admission or prolonged courses of intravenous antibiotics in young infants,15 prophylactic antibiotics, and additional follow-up appointments. In the 2011 AAP UTI guidelines, pretest probability thresholds referenced a testing range from 1% to 3%,11 derived from a survey of directors of outpatient pediatric clinics and emergency departments.16 Consistent with this threshold range, a UTI risk calculator published in 201817 established a threshold for testing of 2%.
We hypothesized that by adding a positive UA result, defined as the presence of pyuria or nitrites, as a diagnostic criterion along with a positive urine culture for the diagnosis of UTI, the prevalence of concomitant UTI in bronchiolitis would be less than these recommended thresholds for testing. To evaluate this, we undertook a systematic review and meta-analysis to assess the scope of current definitions for the diagnosis of UTI within the bronchiolitis literature and to determine the prevalence of UTI in infants with bronchiolitis when positive UA results are included as a diagnostic criterion for UTI.
This study was conducted in accordance with the consensus statement on reporting Meta-Analysis of Observational Studies in Epidemiology (MOOSE)18 (eTable in the Supplement) and is registered with PROSPERO International Prospective Register of Systematic Reviews (CRD42018088978).19 The initial search was conducted by a medical librarian in September 2017. The search included Medline and Medline-in-process from 1946 to August 2017 and Ovid EMBASE from 1976 to August 2017. Search terms were determined by diagnoses and Medical Subject Headings, limiting to articles in English, including infants from birth through age 23 months, and excluding variations of the word pregnancy to filter out maternal articles on UTI. The full search strategy is available in eAppendix 1 in the Supplement. Additional studies were identified through hand searching and a manual review of the bibliographies of qualifying studies and relevant other systematic reviews.
Studies were considered for inclusion if they reported a UTI prevalence rate in infants having a primary diagnosis of bronchiolitis and for subanalysis if they provided both urine culture and UA data. Studies were then excluded if they included patients with active immunodeficiency or oncologic processes or specifically included patients with known anatomically abnormal urinary tracts. Additionally, review articles, case reports or series, editorials, meta-analyses, and guidelines were excluded.
Studies identified from the search results underwent title and abstract review, conducted by 2 independent reviewers for inclusion (C.E.M. and S.R. or C.E.M. and A.R.S.). Any discrepancy regarding inclusion was discussed by the entire study team and consensus reached. For studies reporting a UTI prevalence in bronchiolitis but not reporting UA data, the corresponding authors were contacted via email to determine whether additional data were available.
For full-text articles meeting inclusion criteria, data abstraction occurred independently by 2 study investigators (C.E.M. and S.R. and C.E.M. and A.R.S.). Using a standardized data extraction form piloted by 3 team members (C.E.M., S.R., and A.R.S.), the extraction was conducted by 2 reviewers, compared, and referred to the third for any disagreements. With discrepancies resolved by consensus, final data were extracted from 18 bronchiolitis studies (Figure 1).
We defined the diagnosis of a UTI and the definition of a positive UA result using the individual study definition. If the study did not provide a formal definition, we defined a positive UA result based on reporting a positive leukocyte esterase, positive nitrite, or more than 5 white blood cells per high-powered field. We did not include bacteriuria in the definition of a positive UA result. Owing to suggestions from prior investigations that the sensitivity of the UA is lower for non–Escherichia coli organisms,20 the proportion of E coli UTIs were recorded for studies reporting urinary pathogen data.
To assess for study quality, we used a critical appraisal tool for prevalence studies, with modifications tailored to our study question.21 Specifically, 2 investigators independently evaluated the quality of all included full-text studies. The team modified the tool to address critical issues of validity for prevalence studies relevant to bronchiolitis, including ensuring a representative sample, appropriate recruitment, objective study definitions for UTI and bronchiolitis, appropriate accounting for confounding factors, and appropriate statistical analyses. Individual components for each study were rated as 0 (failing the measure), 1 (passing the measure), or not reported. The maximum score for a given study was 7.
We used a κ statistic to report interobserver agreement on study inclusion and exclusion. To conduct a meta-analysis, we first transformed proportions with the Freeman-Tukey double arcsine transformation.22 For proportions of zero, we used a continuity correction. Point estimates and 95% confidence intervals were then computed using inverse-variance weighted random effects. For the 18 studies providing prevalence data, a baseline point estimate of UTI prevalence was calculated. From the 7 studies providing UA data, point estimates were calculated using the study definition for a UTI and then a second estimate based on the inclusion of positive UA criteria in addition to positive urine culture. Heterogeneity was assessed using the I2 statistic. Sensitivity analyses were performed removing studies identified as outliers on the study quality assessment tool. Lastly, a prevalence rate for E coli UTI with and without positive UA was calculated using the same analysis as discussed previously. All analyses were conducted with Stata, version 15.1 (StataCorp).
A total of 463 studies were retrieved in our initial search and 5 were eliminated as duplicates. We identified an additional 20 studies through reference list review; thus, we reviewed 478 unique study abstracts and identified 31 studies for full-text review. Interobserver agreement on study inclusion/exclusion was high, with a κ of 0.89 (95% CI, 0.83-0.95). After full-text review, 13 studies were further excluded, with the most frequent reason for exclusion being studies not reporting UTI prevalence (10 of 12) (Figure 1). One study was excluded owing to using the same cohort of patients for 2 articles.23 Although corresponding authors with available contact information were emailed, no studies were able to provide additional UA data. Of the remaining 18 studies,1-3,24-38 11 did not report UA data, leaving 7 for meta-analysis by reporting UA data and urine culture positivity (Tables 1 and 2).1,24,26,29,30,32,33
Of the 7 included studies with UA data, 3 defined bronchiolitis by a positive respiratory syncytial virus test result only,26,32,33 while the other 4 included children with clinical bronchiolitis as well.1,24,29,30 All of the studies excepting 2 required fever as part of the inclusion criteria.1,26,29,30,33 Three of the studies exclusively included infants younger than 90 days.24,29,30 Overall quality scores ranged from 5 to 7 on a 7-point scale. Scores within the individual categories are available in eAppendix 2 in the Supplement.
The criteria for diagnosing a UTI and the definition of a positive UA result varied between studies (Tables 1 and 2). Definitions of a positive urine culture ranged from any growth of a known pathogen35 to, most commonly, at least 10 000 cfu/mL on a catheterized specimen.2,24,27,30,38 Several studies used varying standards of colony-forming unit per milliliter for positivity based on method of obtaining the urine.3,26,28,37 Most studies included only cultures that grew a single organism, and 5 studies did not provide a definition for a positive culture.29,31,32,34,36 For the 7 studies that provided UA data, 2 of the studies used positive leukocyte esterase, positive nitrite, or more than 5 white blood cells per high-powered field.1,29 Several of the included studies presented UA data but did not define a positive UA result.24,30,32 There was no consistency in the definition of a UTI when stratified based on studies that provided UA data vs those that did not. There also was no consistency in the definition of UTI or of a positive UA data when stratified by year of manuscript publication.
The overall reported prevalence from the 18 included studies of UTI in bronchiolitis was 3.1% (95% CI, 1.8-4.6; I2 = 89%) (Figure 2). When narrowed to the 7 studies that provided UA data, the overall prevalence of UTI without inclusion of this data was 2.2% (95% CI, 0.8-4.4; I2 = 85%). However, when a positive UA result was included as a diagnostic criterion, the prevalence of UTI was 0.8% (95% CI, 0.33-1.40; I2 = 39%) (Figure 3). From the 7 studies providing UA data, a sensitivity analysis removing 3 studies that defined bronchiolitis only by respiratory syncytial virus testing26,32,33 did not substantially alter the summary statistics, producing a point estimate for UTI of 1.0% (95% CI, 0.1%-2.7%; I2 = 61%) when a positive UA result was required. Similarly, for the studies exclusively including infants younger than 90 days,24,29,30 the point estimate for concomitant UTI with positive UA was 0.5% (95% CI, 0.1%-1.2%; I2 = 0%).
In the 6 bronchiolitis studies with at least 1 positive urine culture (1 study had zero positive cultures29), all reported the specific causative UTI pathogens. In total, the weighted proportion of E coli UTIs was 63.9% in bronchiolitis.1,24,26,30,32,33 When applying the diagnostic criterion of a positive UA result, the weighted prevalence of E coli was 80.3%.1,24,26,30,33
In studies examining the prevalence of UTI in bronchiolitis, inclusion of a positive UA result as a diagnostic criterion was significantly associated with the estimated prevalence of concomitant UTI. The overall prevalence of UA–positive UTI is 0.8%, less than the pretest probability threshold of 1% to 3% discussed in the 2011 AAP UTI guidelines.11 These results held consistent across sensitivity analyses.
Similar to the previously published meta-analysis of UTI in bronchiolitis,4 our study found an overall reported prevalence of 3.1%, with the addition of several newer studies published after the previous analysis. We also noted a low proportion (67%) of urine culture pathogens to be E coli compared with the proportion described in most febrile infant studies of 80% to 90%.39-42 Interestingly, when a positive UA result was required for diagnosis of a UTI, the E coli proportion increased to 80% in UA-positive UTIs. Non–E coli organisms have been associated with a lower rate of pyuria in several studies published in the last 5 years and may indicate that these organisms incite less of an inflammatory response when causing a UTI.20,43 Alternatively, these organisms may be more likely to associated with contamination or asymptomatic bacteriuria. Eliacik et al44 found that follow-up suprapubic aspirates performed on patients with untreated catheterized urine cultures positive for non–E coli organisms (most of with negative UA results) were predominantly negative.44 These findings suggest that most urine cultures that are positive for non–E coli organisms without pyuria either do not reflect true UTI or represent UTIs that spontaneously resolve and that the negative UA results found in these patients therefore reflect true negatives rather than false-negatives.
Within our study cohort, the definitions of UTI, a positive urine culture, and a positive UA result were highly variable, pointing to the need for more adoption of standardized definitions. In the 1999 AAP UTI Practice Parameter, the definition of UTI was based on urine culture results alone and varied by culture method (suprapubic aspiration, catheterization, or bagged specimen) and quantity of bacterial growth.45 The recommended diagnostic criteria were subsequently revised in 2011 to include both an abnormal UA result and a positive urine culture for infants older than 2 months.11 All 3 of the bronchiolitis studies1,26,33 published after the 2011 guidelines included subanalyses comparing diagnosis based on urine culture alone with urine culture plus a positive UA result; however, the definitions that they used to define a positive urine culture and UA were disparate.1,26,33
The revised AAP recommendations exclude infants younger than 2 months, the highest risk population for serious bacterial infections and where questions regarding the reliability of the UA still linger. Accordingly, none of the 3 studies in our meta-analysis that exclusively included infants younger than 90 days required a positive UA result for the primary reporting of a concomitant diagnosis of a UTI, regardless of year of study enrollment. Interestingly, both the baseline reported pooled prevalence (1.4%) and the pooled prevalence of UA-positive UTI (0.5%) was lower in this younger age group.
Since the publication of the revised diagnostic criteria for UTI, the utility of an abnormal UA in young infants as a necessary component for the diagnosis of a UTI has been re-evaluated. Two investigations46,47 published within the past 5 years have reported UA sensitivities in bacteremic UTI in young infants nearing or reaching 100%. These findings are relevant given that simultaneously positive urine and blood cultures with the same organism cannot be explained by asymptomatic bacteriuria or contamination and therefore, in contrast to a positive urine culture alone, bacteremic UTI represents true infection. Additionally, Hoberman et al10 found that evidence of acute pyelonephritis was not present on dimercaptosuccinic acid renal scan in any patients with positive urine cultures and less than 10 leukocytes/mm3 on UA,48 and Wettergren et al13 demonstrated that asymptomatic bacteriuria (detected by suprapubic aspirate) is surprisingly common in healthy infants, with a period prevalence of 0.9% in girls and 2.5% in boys during the first year of life and a point prevalence as high as 1.5% in boys younger than 2 months. These studies suggest that the paired finding of a positive urine culture without pyuria is unlikely to reflect true UTI.
Misdiagnosis of UTI in infants with bronchiolitis has significant implications for patient care. Early exposure to antibiotics has been associated with obesity, asthma, anaphylaxis, allergies, and development of inflammatory bowel disease49-52 and may lead to unnecessary laboratory workup, imaging and radiation, antibiotic prophylaxis, and hospitalization. Additionally, inflated prevalence estimates are likely to drive initial urine testing, which is often invasive. While we are unable to conclude that all reported UTIs with unremarkable UA results were misdiagnosed, our findings do suggest that prior estimates of the prevalence of UTI in bronchiolitis may have been exaggerated. Given that most UTI investigations have not required a positive UA to define UTI, investigations similar to this that reassess UTI risk using UA criteria in other clinical conditions, such as fever without source, jaundice, influenza, or brief resolved unexplained events, would be informative.
Under the premise that UA results should indeed be incorporated into the definition of UTI, the estimated prevalence of UTI in bronchiolitis of 0.8% translates to 125 infants being tested to detect 1 UTI. This number needed to test is high enough that routine urine testing in infants with bronchiolitis and fever seems unjustified. However, even within bronchiolitis, the probability of UTI is likely affected by known risk factors described for UTI in general,17 such that a 2-month-old uncircumcised boy with bronchiolitis may have a higher probability of UTI than a 14-month-old circumcised boy with bronchiolitis. Ultimately, clinicians can incorporate pretest probability, severity of illness, and parent preferences to determine when urine testing may be appropriate. For those infants in whom testing may be desired, a screening UA by a urine bag may help to further reduce the trauma of urethral catheterization.53
In this cohort, the definitions of UTI and positive UA result were highly heterogeneous between studies. Owing to these variations, we were forced to define a UTI based on the individual study’s definition rather than a standard applied to all included studies. Because many of the studies include urine culture results for growth less than 50 000 cfu/mL and the significant variability within the UA positivity definition, the exact prevalence of concomitant UTI in bronchiolitis remains elusive. Several studies in their definition of a positive UA included nitrites, which are indicators of a positive urine culture but may not help distinguish a UTI from asymptomatic bacteriuria. As such, the inclusion of nitrites in these studies may further inflate the calculated prevalence of UTI greater than the actual prevalence. There was a difference in overall prevalence of UTI between the calculations based on all 18 included studies vs the 7 that provided UA data (3.1% vs 2.2%). Qualitatively, there are no apparent differences in inclusion or exclusion criteria or consistency in UTI definitions between the 7 studies and the larger cohort that would explain this difference, so the reason for this small difference remains unclear. Lastly, there were 2 studies that did not require fever as an inclusion criteria,24,32 which introduces potential threats to generalizability given that practitioners are less apt to worry about UTI in afebrile patients with bronchiolitis. In one of these studies, 72% of infants had a documented fever,24 and in the other, presence of fever was not documented.32 However, the prevalence of UTI in these 2 studies (1.2% and 1.2%, respectively) was similar to the studies that mandated fever.
The prevalence of UTI in infants with bronchiolitis is affected by UTI diagnostic criteria. With the incorporation of a positive UA result into the disease definition, the estimated prevalence of concomitant UTI is less than proposed thresholds for testing. Applying a positive UA result as a diagnostic criterion for UTI may have the potential to decrease the misdiagnosis of UTI in infants with other recognizable clinical syndromes as well.
Accepted for Publication: November 19, 2018.
Published Online: January 28, 2019. doi:10.1001/jamapediatrics.2018.5091
Corresponding Author: Corrie E. McDaniel, DO, M/S FA.2.115, PO Box 5371, Seattle, WA 98105(email@example.com).
Author Contributions: Dr McDaniel had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: McDaniel, Ralston, Schroeder.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: McDaniel, Ralston, Lucas.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Ralston, Lucas.
Administrative, technical, or material support: McDaniel.
Conflict of Interest Disclosures: Dr Schroeder has received occasional honoraria from universities/children's hospitals for invited talks on urinary tract infection.
Create a personal account or sign in to: