Treatment and prophylaxis of UTIs: Commentary on current controversies, part 2

For infants and children with a suspect or documented UTI, selection of antibiotics is relatively straightforward. The most important factor in selection is the knowledge of local antibiotic resistance in Escherichia coli (far and away the most frequent urinary tract pathogen) to allow an informed decision on the likelihood of success when one empirically selects a particular antibiotic. Virtually all hospitals keep track of antimicrobial susceptibility data for E. coli and other gram-negative enteric bacilli in the community and in hospital units, and they report the data in the hospital’s annual antibiogram. This allows clinicians to track the local resistance to the most commonly used agents: beta-lactams (penicillins, cephalosporins, carbapenems), trimethoprim-sulfamethoxazole, aminoglycosides and fluoroquinolones.

The second most important factor is the clinician’s need to “get it right” from the first dose. For example, looking at the antibiogram from “Anytown” General Hospital, if TMP-SMX susceptibility of local E. coli isolates from the community is about 80%, then one is likely to fail in about 20% of children treated empirically with TMP-SMX, before knowledge of the culture and susceptible results.

If your patient is a 2-year-old girl with a possible UTI based on symptoms of urethritis of frequency and dysuria, with no fever, no abdominal pain or flank pain, and no previous history of UTI, then it is reasonable to take a 20% chance of failing initial treatment, pending further information on the susceptibility her E. coli (if she indeed has a UTI).

This risk is reasonable because it is unlikely that she will suffer any long-term injury to her kidneys based on a clinical diagnosis of cystitis, and knowing that she has medical follow-up to make sure that she does not subsequently develop signs and symptoms of upper tract disease while on TMP-SMX. Once the susceptibility results are available within a day or two, the clinician will be reassured if the organisms are susceptible to TMP-SMX or will be able to switch therapy to an effective agent, if resistance is demonstrated.

On the other hand, for the 1-year-old boy with fever to 40^· C, irritability, vomiting, no focus of infection, but cloudy urine with an abundance of leukocytes and perhaps leukocyte casts, one cannot take a risk of failing empiric therapy for a presumed pyelonephritis. In 2011, for most hospitals in the United States, third-generation cephalosporins are active against 95% to 98% of community isolates of E. coli. For this child, parenteral therapy is preferred, given the concerns for vomiting, to assure that the antibiotic will achieve an effective exposure in the kidney. Ceftriaxone (intramuscular or IV) or cefotaxime have been used effectively for decades in this context.

For oral therapy, any of the third-generation cephalosporin agents — cefixime (Suprax, Lupin Pharmaceuticals), cefdinir (Omnicef, Abbott Laboratories), ceftibuten (Cedax, Pernix Therapeutics), cefpodoxime (Vantin, Pharmacia and Upjohn) — and most second-generation cephalosporin agents should be effective because they are all concentrated in the kidney during renal elimination. More than a decade ago, Hoberman and Wald demonstrated prospectively that for infants aged 1 to 24 months, with no history of previous UTI, a documented febrile UTI could be treated as effectively with oral cefixime as with ceftriaxone, demonstrating that it is the exposure of the antibiotic at the site of infection that counts. Although gentamicin is also effective in vitro against 95% to 98% of community E. coli isolates, aminoglycosides are more nephrotoxic than cephalosporins, and we usually select the safest among equally effective agents. The same applies to ciprofloxacin, which is effective but is still under evaluation for cartilage injury in children, and it should not be used as first-line therapy if other, safer alternatives exist.

The duration of treatment has also not been prospectively evaluated in risk-stratified children. Those with simple cystitis and no urinary tract anomalies should be effectively treated with 3 days of therapy. Those with pyelonephritis probably need longer therapy because small abscesses are likely to be present within the renal parenchyma, suggesting a 7- to 10-day course of therapy. Those with lobar nephronia (renal parenchymal cellulitis) or renal abscess may need at least 10 to 14 days. Because we have no way to routinely and accurately describe the extent of infection in each child without imaging, most recommendations — by necessity — simply suggest a range of days of therapy. However, rather than simply suggesting a standard number of days, it would make far more sense to treat each child after the clinical response and an objective laboratory marker for acute inflammation (such as the C-reactive protein). However, prospective studies to evaluate these markers to define duration of therapy and link duration with outcomes for children with a range of severity of infection have not yet been performed.

For children with recurrent UTIs, particularly those with anatomic abnormalities or previous urinary tract surgery, no single recommendation for therapy can be made. Information from the child’s past infections must be reviewed to better understand the susceptibilities of previous urinary tract pathogens, which are likely to include other pathogens such as Pseudomonas aeruginosa and Enterococcus faecalis, in addition to E. coli. The more resistant and more recently isolated the previous pathogen, the more aggressive initial empiric therapy should be — awaiting culture results to see if the spectrum of therapy can be narrowed; for example, meropenem stepped down to ceftriaxone, or even to ampicillin.

Antibiotic prophylaxis

Prophylaxis is designed to prevent infection, assuming that recurrent infection leads to cumulative renal damage. Therefore, if we prevent infection, we can prevent damage. However, as noted above, we do not accurately know how much injury is caused with each infection, as we have no standard, accepted definitions of severity of disease (including associated comorbid conditions, such as obstruction) that are uniformly applied by investigators, so we don’t actually know how much injury we can prevent with each infection.

Adding to the confusion regarding prophylaxis is the fact that E. coli and other enteric bacilli will develop resistance in time to the antibiotics commonly used for prophylaxis. TMP-SMX is a favorite antibiotic for prophylaxis, based on long-term safety data in children. However, for the child receiving prophylaxis, colonizing TMP-SMX-susceptible strains of E. coli will eventually be replaced by TMP-SMX-resistant strains, with subsequent infections caused by these resistant strains.

The interval of prophylaxis before infection by a resistant strain is variable and unpredictable, from weeks to years. Certainly, any TMP-SMX susceptible strains that may enter the urethra and bladder will be killed after the daily dose of TMP-SMX. However, to say that antibiotic prophylaxis does not work for a particular child because that child will ultimately develop a UTI caused by an antibiotic-resistant pathogen is to undervalue another metric that some others have used to judge effectiveness: time-to-event analysis (eg, time to development of the next UTI).

The questions that arise with this include: 1) the duration of prophylaxis needed to document improved outcomes with respect to renal damage; and 2) which prospectively evaluated subpopulations may benefit most if prophylaxis leads to a prolonged time-to-event. These questions are not well-addressed in the literature, even in prospective studies. A recent publication in The New England Journal of Medicine from Australia by Craig and colleagues documented a modest but statistically significant decrease of 6% in the number of children with UTIs after 12 months of TMP-SMX prophylaxis, but long-term renal injury was not assessed. A recent publication in Pediatrics from Italy by Pennesi and colleagues using 2 years of TMP-SMX prophylaxis in children with VUR did not show a decrease in infections or renal scars in the TMP-SMX group. But they failed to take into account the expected development of resistance while on long-term therapy and subsequent infection caused by TMP-SMX resistant E. coli.

Choosing the right antibiotics

With respect to selection of antibiotics for prophylaxis, besides TMP-SMX, nitrofurantoin has been used for prophylaxis based on adequate concentrations of antibiotic in the urine after oral administration. Data suggest that resistance emerges less quickly with nitrofurantoin compared with TMP-SMX, but resistance will ultimately develop to nitrofurantoin as well. However, for the child who may benefit from prophylaxis, you now may have TMP-SMX and, subsequently nitrofurantoin, two sequential antibiotics that can prolong the time to UTI.

Similarly, resistance is currently uncommon to the second- and third-generation oral cephalosporin antibiotics. These can be used as well, until resistance develops to those agents, too. Once antibiotic pressure from prophylaxis is removed, susceptible strains will often re-colonize the child. Unfortunately, no prospective evaluation of prophylaxis that assesses the development of resistance and the use of a clinical trial design of sequential, rotating antibiotics (based on susceptibility data from isolated strains) as a means to maximize the time-to-UTI interval has been conducted. And of course, there is a real risk for developing multiresistant strains. Fortunately, for the child with a break-through infection caused by a multidrug-resistant strain, parenteral therapy still exists in the aminoglycoside and carbapenem classes of antimicrobials, as neither is available in an oral formulation for prophylaxis.

We clearly must better define the benefits and risks of prophylaxis for various subpopulations of children at risk of long-term renal injury.

Lastly, an NIH/NIDDK-funded prospective study is under way: the Randomized Intervention for Children with Vesicoureteral Reflux (RIVUR) study. In this study, TMP-SMX vs. placebo is provided to children with VUR (all grades) in a randomized, double blind trial design, and the primary endpoint is prevention of UTI with 2 years of prophylaxis. Secondary endpoints are renal scarring and the development of TMP-SMX resistance in E. coli that may cause infection or colonization. A total of 600 children will be enrolled through 2013. However, not addressed in the study are the issues regarding development of resistance and changing antimicrobial prophylaxis strategies to ensure that active agents are used. We can be assured that TMP-SMX resistance will develop in children who are given TMP-SMX prophylaxis for 2 years.

John S. Bradley, MD, is director of the division of infectious diseases at Rady Children’s Hospital in San Diego and associate clinical professor of pediatrics at the University of California San Diego. He is also a member of the Infectious Diseases in Children Editorial Board.

Disclosure: Dr. Bradley reports no relevant financial disclosures.

For more information:

• Craig JC. N Engl J Med. 2009;361:1748-1759.
• Hoberman A. Pediatrics. 1999;104(1 Pt1):79-86.
• Hoberman A. N Engl J Med. 2009;361:1804-1806.
• Pennesi M. Pediatrics. 2008;121:e1489-1494.
• RIVUR Study Online Posting. Available at: http://clinicaltrials.gov/ct2/show/NCT00405704?term=rivur&rank=1.
• Shaikh N. Pediatrics. 2010;126:1084-1091.
• Williams G. Cochrane Database Syst Rev. 2011;3:CD001534.

Follow the PediatricSuperSite.com on Twitter.