These data in an overall younger population than that covered by the 2011 guidelines adds evidence that pyuria (but not microscopic bacteriuria) is critical to diagnosing UTI. Pyuria plus positive culture has been a combination for the pediatric infectious diseases practitioner’s toolkit for decades. Likewise, it seems to me that primary care pediatric clinicians also often decide whether to undertake the expense of culture based on UA results. For example, a completely normal UA may obviate need for culture except in selected unusual cases.
Requiring UA evidence of inflammation to diagnose UTI (per the 2011 guidelines and the recommendations of the authors of the recent 2015 study) makes sense because most UTIs in otherwise healthy children are caused by gram-negative organisms (> 90% from Escherichia coli) (J. Pediatr. 1994;124:513-9). Why are UA results so important?
A positive nitrite test strongly suggests UTI because nitrites in the urine indicate viable gram-negative organisms also are present in the urine. Nitrates in the urine are converted to nitrites by metabolic activity of gram-negative pathogens. For WBCs or LE in the urine, their presence indicates inflammation in the urinary tract, Consider that lipopolysaccharide (LPS), also known as gram-negative endotoxin, is a major component of the cell membrane of > 90% of uropathogens like E. coli. Moreover, LPS elicits about the strongest innate immune response via toll-like receptor 4 (TLR4) from monocytes/macrophages, inducing a large pro-inflammatory and chemotactic response – interleukin-6, interleukin-8, tetrahydrofuran-alpha. Remember that LPS is also a major cause of fever and of shock during gram-negative sepsis.
So a UTI diagnosis based on a “positive” culture without evidence of metabolic products of gram negatives (nitrites) or without inflammation (no pyuria or negative LE) should be questioned. The combination of > 50,000 CFU/mL with no detectable LE or < 3-5 WBC/HPF in a febrile child is most likely evidence for AB in a child with the fever caused by some non-UTI process.
In contrast, selected SBIs may occur when the culture is “positive” without inflammation or nitrites. The first of three examples is a renal parenchymal abscess, where bacteria enter the urine sporadically in only small numbers, and do not actually infect the urinary tract mucosa. The scenario of no inflammation but “positive” culture also may occur when a large bacteremic load causes results in organisms filtering through the kidney into the urine, again without urinary mucosal infection, such as Staphylococcus aureus, group A streptococcus, or group B streptococcus bacteremia/sepsis. The third scenario with a “positive” culture and no pyuria can be with organisms that have blunted abilities to induce inflammation, such as enterococcus. Enterococcal cell components have weak inflammatory and chemotactic capability. So a urinary mucosal infection in the collecting system or bladder may occur without much if any pyuria. In fact, the patients from the recent study with insufficient evidence of pyuria/inflammation were those who had either gram-positive organisms or considerably less than 50,000 CFU/mL of gram-negative organisms.
The sensitivity and specificity of the LE or pyuria was higher in the recent study (Pediatrics 2015;135:965-71) than any prior study. The authors comment that they had not expected such a high sensitivity of 97.6% (94.5-99.2) for LE in confirmed bacteremic UTI, nor did they expect the high specificity of 93.9% (87.9-97.5). The presence of microscopic pyuria defined as > 3 WBC/HPF was nearly as sensitive, 96%, and specific, 91.3%. Disappointingly, positive nitrite testing was only 39.5% sensitive, but it was 100% specific. This likely reflects the short time that urine resides in the bladder of infants < 90 days of age, so there is insufficient time for the pathogens to metabolically convert the nitrates to nitrites.
So how would the UA help with our example case? There is microscopic bacteriuria, pyuria, and positive LE, but negative nitrites. Using the suggestions of the authors of the recent report (Pediatrics 2015;135:965-71) and those of another report on the utility of UA results (Acta Paediatr. 2010;99:581-4), the UA in our case indicates that we should be highly suspicious of a UTI in this child < 2 months old for whom the 2011 guidelines do not directly apply. But remember that these impressive sensitivity and specificity values relate to bacteremic UTI. Whether they apply to nonbacteremic UTI is not known. Likewise, the authors caution that their study design did not allow calculation of positive or negative predictive values – aspects that would clarify things even further.
So we still cannot be more than highly suspicious. Without a positive predictive value, we do not know the odds of this case having a UTI with mathematical precision. The authors do point out that only one of their subjects had a completely normal UA and actually had a bacteremic UTI. If you guessed that it was a gram-positive pathogen, you win the prize. So it seems reasonable to predict that a normal UA has a high specificity for not being a UTI (87.8%), but a positive UA remains only highly suggestive. It is still not clear if a negative UA statistically justifies not submitting the culture of the sterilely obtained urine because we still don’t have a negative predictive value.