Urinary Tract Infections Treated by Third-Generation Cephalosporins

In this issue of Pediatrics, Dasgupta-Tsinikas et al¹  present several important issues faced by clinicians in selecting antibiotic therapy for infants and children with urinary tract infection (UTI) based on laboratory-reported culture results. The laboratory report provides the clinician with 2 important pieces of data: the minimum inhibitory concentration (MIC) of each antibiotic tested against the urine isolate and an interpretation of that MIC value, given as susceptible (S), intermediate (I), or resistant (R). The MIC at which an antibiotic/organism pair goes from S to nonsusceptible (that includes I and R), is the “breakpoint,” and differentiates lower MICs for an antibiotic/pathogen pair for which antibiotic treatment should be successful at standard antibiotic doses from higher MICs at which there is an increased risk of microbiologic failure. This reported breakpoint interpretation of the MIC is under the jurisdiction of the United States Food and Drug Administration (FDA) and is generally based on the achievable concentrations of an antibiotic in the bloodstream supported by clinical data.²  Thus, the reported S, I, and R for a urine E. coli isolate is not based on antibiotic concentrations achieved in the urine, which may be >100 times greater than serum for penicillins/cephalosporins. For example, ceftazidime concentrations were documented to be 12 000 μg/mL in urine in adults during the first 2 hours after a dose compared with serum concentrations that peaked at 129 μg/mL.³  Renal parenchymal concentrations are also much higher than serum for β-lactam antibiotics, including the third-generation cephalosporins (3GC).⁴  Logically, the exposure of bacteria to an antibiotic at the site of infection is the critical determinant of whether an infection can be cured or not. The FDA and organizations who make recommendations to the FDA for breakpoints have traditionally not been asked to recognize the role that tissue-level antibiotic concentrations play in treatment outcomes; although recently, as noted by the authors, a second, lower “breakpoint for meningitis” was recently incorporated into laboratory reports. This lower breakpoint recognized that higher doses of antibiotics are required to achieve concentrations in CSF necessary to treat meningitis. Likewise, we need another, higher “breakpoint for urinary tract infections” for each antibiotic used to treat UTIs, including 3GC. Although there is recognition by FDA that 1 breakpoint reported by the laboratory may not accurately predict success at all tissue sites, the agency usually requires substantial data from well-controlled clinical trials designed to carefully assess safety and efficacy at each antibiotic dose for each infection site (or “indication”), by pathogen, before breakpoints are changed and additional drug doses can be approved as safe and effective.

Overall, as noted by the authors, <5% of current pediatric urine E. coli isolates in their dataset from all public acute care facilities in Los Angeles are third-generation cephalosporin-resistant (3GCR). For these children, previous acute health care utilization and underlying medical conditions were documented by the authors to be more common in those with 3GCR isolates compared with non-3GCR UTI controls, suggesting that these children are not previously healthy children from the community. This has important implications for determining which populations are at the highest risk of infection by resistant pathogens. For this report, the authors were not able to access data on either the site of infection within the urinary tract (separating those with community-acquired lower tract infections [cystitis] from those with upper tract infections [pyelonephritis]) or even more importantly, separating those with community-acquired infections (in previously healthy children) from those with complicated UTI, which comprises all the children with various, complex conditions that are not “simple.” For example, children with pyelonephritis complicated by renal and perinephric abscesses are expected to respond more slowly to antibiotics than those without abscesses, regardless of pathogen isolated or antibiotic used (3GCR versus non-3GCR). Fortunately, many of those without abscesses may respond to 3GC if antibiotic concentrations in the kidney exceed the MIC for a significant proportion of the dosing interval, whereas those with abscesses might not respond because of poor antibiotic concentrations within abscesses, highlighting the need to identify specific subpopulations at risk for treatment failure.

Beyond the variability of antibiotic tissue concentrations on outcomes, additional variability is present based on the resistance mechanisms in the bacteria; resistance is not always all-or-none for 3GC against E. coli and other enteric bacilli, although a laboratory report of S or R suggests that is the case. Resistance to 3GC is primarily due to the presence of β-lactamases that are capable of cleaving the β-lactam ring structure of these antibiotics. In the past, 3GCs were not degraded by the majority of β-lactamases present in E. coli, but with the continuing evolution of these β-lactamases under the pressure of 3GC use, hundreds of mutations in β-lactamase structure have now been described, all revealing some degree of increased activity against the 3GC β-lactam ring (many of these β-lactamases are generically called extended spectrum β-Lactamases, or ESBLs⁵ ). Some ESBLs have only minimal hydrolyzing activity against the 3GC, with a measured MIC that is only slightly higher than strains with no ESBL, but some β-lactamases have profound activity, completely inactivating the 3CGs. Any pathogen with an ESBL, irrespective of the MIC, is labeled as R.

It becomes instantly clear that many bacteria causing UTIs and reported as R based on blood concentrations are, in fact, susceptible when an infection is in the urinary tract. It is likely that R E. coli with ceftriaxone MICs of 2, 4, 8, 16 μg/mL, or greater may have been treated successfully with ceftriaxone; however, the authors did not have access to patient-level specific MIC data for their analyses, only the interpretation of the MIC.

Unfortunately, if the laboratory reports an isolate as R, even in a child who is responding to therapy with a 3GC, there is pressure placed on the clinician to change treatment by those who may not fully appreciate the MIC/drug exposure metrics that would actually predict success. Those clinicians bold enough to continue treatment of 3GCR E. coli in a child who is responding may get calls from antimicrobial stewards to change therapy in case we “missed” the result. We also need to consider the medical and legal consequences of treating a child with an antibiotic that the laboratory reports as R in case the outcome is not perfect.

Once a child is labeled as having a 3GCR UTI, empirical therapy for all subsequent suspected infections usually starts with more broad-spectrum agents. The authors noted this practice, which will not be easy to change until studies to assess new UTI breakpoints are conducted, subsequently reviewed, and potentially approved by the FDA.

For the past 4 decades, the 3GC (ceftriaxone, cefotaxime, ceftazidime) have been a remarkably effective therapy for UTIs caused by E. coli. We need to better understand when 3GC can continue to be used to successfully treat E. coli UTI reported as resistant before we give up their use routinely and move to antibiotics with the potential for increased toxicity and a more profound impact on the child’s microbiome. In fact, new software programs are becoming available that integrate tissue site concentrations of specific antibiotics with the MIC of the infecting pathogen to that antibiotic to allow for the suggestion of an appropriate dose. The FDA is currently evaluating these kinds of programs, and perhaps breakpoints will no longer be needed in the future. Resistance will eventually emerge to virtually every antibiotic we use; we need to be smarter about how we use them.

3GC

third-generation cephalosporin (cefotaxime, ceftriaxone, and ceftazidime)

3GCR

third-generation cephalosporin-resistant

ESBL

extended-spectrum β-lactamase

FDA

United States Food and Drug Administration

I

intermediate

MIC

minimum inhibitory concentration

NS

nonsusceptible

R

resistant

S

susceptible

UTI

urinary tract infection