Antimicrobials and UTIs: part 2- antimicrobial options (Proceedings)

Treatment of urinary tract infections offer an example of the hazards and difficulties encountered when initial response is insufficient. Treatment of bacterial UTI offers a good example of how treatment of bacterial infections might be approaches. The goal of drug therapy has been to eliminate bactiuria, but this goal should be modified to include eradication of infection while minimizing the advent of resistance. Further, the goal may need to be modified in the presence of asymptomatic bacteria (see Flow Chart).

Confirming the need for infection.

The lack of perceived toxicity to antimicrobials has contributed to our willingness to use antibiotics in the absence of convincing data supporting infection. Failure for infection to progress has been interpreted as therapeutic success...even if infection was not present in the first place. There is good reason to avoid indiscriminant antimicrobial use. Therapy can be successful in that an infection is eradicated, but resistance may still develop. However, lack of infection does not preclude the development of resistance when antimicrobials are used. Antimicrobial therapy should be used only when reasonable evidence of infection exists. Generally, a quantitative culture yielding less than 103 CFU/ml can be considered for no treatment unless mitigating circumstances indicate otherwise (ie, clinical signs consistent with UTI, immunesuppression that can lead to worsening of infection etc). In humans, the presence of bacturia is not necessarily an indication of the need for therapy. In order to avoid resistance, treatment generally is not indicated in asymptomatic bacturia except under certain conditions in which the patient is at risk, such as pregnancy, or invasive surgical procedures. On the other hand, bacterial UTI occurs much less frequently in cats than in dogs, and clinical signs indicative of cystitis in cats should not be interpreted as a need for antimicrobial therapy. Finally, culture may fail to identify infecting microbes, particularly if slow growing organisms or those requiring special media are present (eg, Mycoplasma or Ureaplasma).Likewise, for our veterinary medical patients, the risk of emerging resistance must be weighed against the risk of failing to treat. Clearly, if the decision is made to treat a UTI, then therapy must be aggressive, designed to kill invading pathogens as well as emerging mutants as rapidly as possible. In all but simple UTI, culture is indicated.

The new.

The role of empirical selection in treatment of UTI is not clear. Increasingly, scientific data supports the inappropriateness of this approach for all but the simplest of infections. An uncomplicated infection is one in which no underlying structural, neurologic, or functional abnormality can be identified. The absence of previous antimicrobial therapy is a critical criteria for classification as uncomplicated. However, in the author's retrospective study of dogs admitted to a veterinary teaching hospital and subsequently diagnosed with UTI, E .coli was the causative organism in only 50% of the cases. Although the majority of the remaining organisms were Gram negative coliforms (eg, Proteus, Klebsiella), selected Gram positive (Staphylococcus and Enterococcus being the majority) also were cultured. Indeed, Staphylococcus and other gram-positive organisms account for 25% of UTIs in dogs. Further, the rate of incidence of resistance of E coli to standard "first choice" drugs precluded empirical selection of an appropriate antibiotic. The percent of organisms resistant to first choic drugs exceeded 50% for ampicillin (eg, amoxicillin), and was 40% or more for drugs considered relatively "resistant" to beta-lactam resistance, ie, amoxicillin-clavulanic acid, and cephalothin. Forty percent of the E coli also were resistant to trimethoprim-sulfamethoxazole. Disconcertingly, 40% of organisms were resistant to fluorinated quinolones, the first choice for complicated infections, and a surprising 50% were resistant to extended spectrum penicillins (carbenicillin, piperacillin and ticarcillin). Indeed, the only drugs to which E coli was predictably susceptible were nitrofurantoin and the aminoglycosides, particularly amicakin. Factors associated with resistance included previous antimicrobial therapy and previous hospital stay.

A follow up study of 500 isolates in the United States has demonstrated regional and laboratory differences in E. coli resistance, with isolates in the south and in academic laboratories associated with higher rates of infection. This data supports the concern that the cause of an infection often is not accurately identified, and accurate identification as to genus and species does not accurately predict susceptibility.

What this data does not show, but should be implied with the advent of resistance, is that while empirical therapy may successfully result in the resolution of a UTI, resistance may have (and is likely to) have developed. In otherwise healthy patients, the advent of resistance is not likely to be problematic. But in the patient whose infection is complicated by a multitude of factors, a recurring infection is likely to emerge with a vengeance: ie, it may be characterized by a complex pattern of resistance. Failure to eradicate the subsequent infection is likely to be associated with a multi-drug resistant microbe responding only to injectable and potentially toxic drugs.

The more complicated the infection or the patient, the more important is basing therapy on susceptibility data. In human patients, diagnosis of UTI in asymptomatic patients is based on at least two clean-catch midstream urine collections. The same organism should be present in significant (see earlier) amounts in both cultures. A single culture is sufficient in the presence of symptoms. Urinary cultures should be the basis of antimicrobial selection in complicated infections (e.g., re-infection or relapse; history of antimicrobial use within the past 4 to 6 weeks) or if the infection represents a risk to the patient's health. Infection after recent urinary catheterization also should lead to culture collection. Quantitative urine culture should be used to discriminate harmless bacterial contaminations (e.g., from the urethra) from pathogenic organisms. In a properly collected urine sample, bacterial counts of more than 105 are indicative of infection; counts between 103 and 105 organisms are considered suspect and should lead to a second culture. Samples collected by catheterization or midstream catch techniques are more likely to yield falsely positive cultures than are samples collected by cystocentesis, particularly in females. Thus, cystocentesis is preferred.

Although in vitro resistance may not necessarily predict in vivo failure, evidence in human medicine suggests otherwise. As in dogs, E.coli, is the most prevalent uropathogen in humans. However, antimicrobial resistance to first choice drugs, trimethoprim- sulfonamide (TMP-SMX) and ampicillin, often exceeds 30% in humans. In women with UTI caused by E. coli characterized by in vitro resistance to TMP-SMX, and subsequently treated with the combination drug, approximately 50% were bacteriological failures and another 40% clinical failures. Previous antimicrobial therapy profoundly impacted the likelihood of resistance, with the risk greater if the antibiotic of interest has been used. Again, in women, the most important independent risk factors for TMP SMX resistance in non-hospitalized cases was use of the antibiotic within the past 3 months. Those who had taken any antibiotic were more than twice as likely to be infected with a resistant isolate; use of TMP SMX within the past 2 weeks was associated with a 16 fold greater risk of infection with a resistant isolate.

Urinary catheterization is a recognized risk factor for antimicrobial resistance. Catheterization has resulted in bacturia in previous bacteria free urine, and has been associated with changes in urine microflora, as well as increased resistance. Although aseptic techniques will reduce the risk of infection, infection is not prevented. The risk of persistent UTI in cats with experimentally –induced cystitis was increased with catheter placement, despite the use of a closed system of urine drainage. The risk of infection can be correlated with duration of catheter placement, with the risk being reduced in patients catheterized for less than three days. Previous antimicrobial therapy is likely to contribute to the risk: resistance in dogs catheterized more than 5 days was strongly associated with the advent of resistant microorganisms. Catheter type influences the risk of infection, probably because of its impact on biofilm formation and bacterial swarming. However, organisms within microcosms associated with biofilm and isolated upon urine collected from the catheter – or from the catheter tip - are not necessarily causing infection, and in the presence of infection, may not be the causative organisms. The incidence of resistance in E coli collected from catheters tips is greater than that collected by cystocentesis. Which organism is actually causing infection is not clear; it is reasonable the organisms in the catheter are not necessarily causing infection and selecting antimicrobials based on these organisms generally will require drugs generally considered second or third choice. Collection of urine via cystocentesis is likely to be more representative of infecting organisms compared to the catheter or catheter tip.

Caution is recommended when intervals shorter than 24 hours are used for aminoglycosides. Contact between drug and microbe in the urinary tract can be facilitated by administration of a drug immediately after micturition or before an anticipated micturition-free period (e.g., at night).

Drug Therapy

Attention should be paid to the pH of the urine compared with the pKa of the chosen drug. In most situations, however, even if most of a drug is ionized (e.g., a weak base in an acidic environment), because drugs are concentrated in the urine, generally sufficient un-ionized drug is available to ensure effective concentrations. In the presence of an alkaline pH, weakly basic antibiotics might be considered (aminoglycosides, fluorinated quinolones). Because urease producers may alkalinize the urine, drugs including such organisms (e.g., Proteus, Staphyloccocus, and some Klebsiella species) should be selected. In the presence of an acidic urinary pH (perhaps caused by E. coli), weakly acid drugs (e.g., penicillins, cephalosporins, poentiated sulfonamides) might be better empirical selections.

1st Tier: Beta-lactams

The beta-lactams remain excellent first choice drugs. Not only is resistance, should it emerge, likely to be reflect betalactamases and thus be limited to single drug resisitance, beta lactams also impair pili of E. coli more so than other drugs. Amoxicillin or amoxicillin clavulanic acid are recommended for first choice for empirical therapy. In general, cephalexin is a poor choice. Generic Augmentin® (human Clavamox®) is now available. However, the ratio of clavulanic acid to amoxicillin varies among the human tablets and solution, but not the small animal versions. The 400 mg human capsule has the same ratio as the veterinary ratio.The variability reflects an attempt to minimize vomiting. However, it is not clear if the ratio also impacts efficacy, although a ratio as high as 8:1 is available in the generic human product. There is no reason to believe that a similar ratio would be ineffective in dogs, although it is not clear if dogs will absorb the drug the same as in humans. An alternative approach is to dose an animal with Clavamox® but add to the regimen a dose of amoxicillin that will bring the total dose up to a target of 25 mg/kg bid to tid.

Potentiated Sulfonamides

In women with risk factors for infection with resistant bacteria, or in the setting of a high prevalence of TMP/SMX resistance, a fluoroquinolone or nitrofurantoin is recommended for empirical treatment. The goal of treatment is eradication of infection using shorter courses of therapy (ie, 3 days) with with once-a-day dosing of a selected drug or a single dose of a particularly efficacious antibiotic. The role of nitrofurantoin is increasing for treatment of UTI in women, particularly in the presence of increasing antimicrobial resistance to other urinary antimicrobials. Resistance among uropathogens to nitrofurantoin generally is consistently at a low level despite its use for five decades. An advantage to nitrofurantoin is its minimal effect on the normal gut flora. As such, selection pressure for antimicrobial resistance is reduced compared to other antimicrobials. Further, nitrofurantoin does not share cross-resistance with more commonly prescribed antimicrobials and its use is justified from a public health perspective as a fluoroquinolone-sparing agent. For example, single-dose ciprofloxacin prophylaxis increased the prevalence of ciprofloxacin-resistant faecal E. coli from 3 to 12%. After treatment with ciprofloxacin for prostatitis, 50% resulted in post-treatement faecal colonization with quinolone-resistant E. coli genetically distinct from the prostatic infection. Indeed, in humans, although flouroquinolones are effective as short-course therapy for acute cystitis, widespread empirical use is discouraged because of potential promotion of resistance. An exception is made for acute (non-obstructive) pyelonephritis, but only if culture results direct continuing therapy.

2nd Tier: Fluorinated quinolones

Assume that an isolate resistant to one FQ is resistant to all. If resistance emerges to FQ, it will be multidrug resistance. Thus, the fluorquinolones preferably are reserved as second tier drugs with use based on culture and susceptibility; use at the highest dose is preferred. As concentration dependent drugs, doses of FQs should be designed to reach at least 10 X the MIC of the infecting bug at the site of infection Ciprofloxacin oral bioavailability in dogs is 40 to 60% and in cats, 0-20%. Although ciprofloxacin is more potent toward Gram negative organisms, the dose should nonetheless be increased two fold compared to enrofloxacin, and 3 fold for Gram positive. Oral ciprofloxacin shold not be used in cats. For other fluoroquinolones in cats, marbofloxacin and apparently orbifloxacin are among the safest in regards to retinal degeneration.

3rd Tier

By the time the 3rd tier of drugs has been reached, it is critical to remove the underlying cause of infection. IT is this group in particular for which the risks of treatment need to be weighed against the risk of non-treatment; strong consideration should be given to not treat asymptomatic patients.

Aminoglycosides

The nephrotoxicity associated with aminoglycosides is exposure dependent, meaning, it can be avoided if the kidneys are granted a drug-exposure-free period during which they can excrete drug which has accumulated. Accordingly, aminoglycosides are dose once daily. Maintaining hydration (to the extent of providing sodium containing fluids at the time of dosing in at-risk patients) and dosing in the morning (perhaps the evening in cats) may reduced toxicity, as will avoiding other nephroactive drugs (eg, NSAIDS). N-acetylcysteine may help decrease the risk or extent of damage. We have actually used therapeutic drug monitoring in a patient receiving IV aminoglycoside for over 30 days as a way to follow clearance changes induced by the drugs. The risks of nephrotoxicity and difficulty in using the aminoglycosides limits therapeutic options.

Fosfomycin tromethamine (Monurol®).

Fosfomycin, a phosphonic acid which contains a carbon-phosphorous bond, is a natural antibiotic produced by Streptomyces fradiae (Fig 2). Its in vitro spectrum is broad, with potential efficacy toward isolates expressing MDR, including E coli and Gram positive isolates. Approved for human use in the USA, its indication is as a one time (or up to 3 days) treatment of E coli UTI in humans (Schito 2003). Fosfomycin is a phosophonol pyruvate analogue that irreversibly inhibits phosphoenol pyruvate transferase, an enzyme that catalyzes the first step of cell wall peptidoglycan synthesis of microbial cell walls (Fillgrove 2007). Although its mechanism of action is similar to the beta-lactams, unlike the beta-lactams, fosfomycin is not susceptible to destruction by any class of beta-lactamases. Rather, resistance to fosfomycin, which is unusual, reflects the FosX or FosA enzyme which hydrolyzes the drug. The gene for this protein is chromosomally mediated. Thus, when resistance does occur, it is usually SDR, and generally not associated with MDR (Fillgrove 2007). Further, compared to susceptible strains, fosfomycin-resistant mutants are impaired, exhibiting poorer growth rates, and reduced adherence to uroepithelial cells (particularly in the presence of -acetylcystein (Marchese 2003a) and urinary catheters (Marchese 2003b). As a cell wall inhibitor, fosfomycin is bactericidal when present at the site of infection at therapeutic concentrations. Cell wall inhibition indicates efficacy to be time dependent and facilitated if drug concentrations exceed the MIC of the infecting microbe for at least 50% of the dosing interval (T>MIC=50%). Other attributes of fosfomycin that support its use for treatment of E coli UTI include renal excretion, synergistic interaction with several other classes of antimicrobials (Olay 1974, 1989), and preparation as a 3 g sachet (granules) which is mixed with water to orally deliver approximately 40 mg/kg (in humans). Kill studies in our laboratory suggest that the efficacy of fosfomycin is not concentration dependent, suggesting, as might be expected of a cell-wall inhibitor, time-dependent effects. We anticipate that a week of treatment with Monurol® at 40 mg/kg bid will cost approximately $ 8-12/kg for 1 week of therapy. Alternative therapies are based generally on injectable drugs (aminoglycosides, carbepenems) and thus hospitalization, and for aminoglycosides, intensive monitoring. Side effects of fosfomycin appear to be limited to diarrhea.

We have demonstrated the efficacy of fosfomycin toward MDR E coli. Fecal E. coli isolates exhibiting NDR (n=50), SDR or MDR (n=25 each; isolates resistant to 14/16 drugs on the antibiogram) or non-resistant (n=50) fecal E coli (see Background) were subjected to fosfomycin E test. Isolates were obtained either from an experimental study in which multidrug resistant fecal E. coli were induced following administration of enrofloxacin (5 mg/kg/day PO for 7 days or from an ongoing surveillance study in partnership with IDEXX laboratories (1000 isolate E. coli isolates/year from dogs or cats with spontaneous disease). Thus low or high level resistance can be detected because concentrations exceed the resistant breakpoint MIC (MICBP). The fosfomycin breakpoints are < 64 mcg/ml (susceptible) and > 128 mcg/ml (resistant). The distribution MIC for fosfomycin for all of the isolates, regardless of the t ype of resistance (NDR, SDR and MDR) was well below the susceptible breakpoint (< 64 µg/m) for fosfomycin. The MIC range was 0.25 to 4 µg/ml; the MIC 50 and 90, were, respectively, 1 and 1.5µg/ml, respectively, except for MDR, for which MIC90 was 2 µg/ml. The data suggests that MDR E coli should remain susceptible to fosfomycin. We also have determined the pharmacokinetics of fosfomycin tromethamine, the commercially available drug, when administered as a single oral dose of 80 mg/kg (Figure 3).

Figure 4. Plasma drug concentration vs time curves for fosfomycin after administration of fosfomycin tromethamine (80mg/kg) po in 12 apparently healthy dogs

After oral administration, drug was detected in urine at concentrations exceeding the MIC 90 of fosfomycin for multidrug resistant (MDR) E. coli (1.5 mcg/ml) until 7 (2.5 mcg/ml) and 12 hr (9 mcg/ml) following IV and oral administration, respectively. Oral bioavailability was 88+ 32%. Gastrointestinal upset manifested as mild to moderate diarrhea was observed in four of the twelve dogs. Food increased oral bioavailability: with food, 109+31% (95% CI: 84-135) and without food, 66+16% (95%CI: 52-79%). Based on these studies, fosfomycin tromethamine appears to be a drug which may reasonably be effective for treatment of canine UTI, including that caused by MDR E. coli. However, clinical trials are now indicated to determine the most appropriate dosing regimen (bid versus once daily) and the duration of therapy.

Adjuvant Therapy

Diuresis has been advocated in the treatment of UTIs in humans. Advantages include rapid dilution of bacteria, removal of infected urine, and subsequent rapid reduction of bacterial counts. In patients with pyelonephritis, an added advantage may be enhanced host defenses: Medullary hypertonicity inhibits leukocyte migration, and high ammonia concentration inactivates complement. In the presence of vesiculoureteral reflux, however, diuresis may increase the risk of acute urinary retention. Mannose has been suggested as a preventative or adjunct to therapy; blockade of mannose receptors may preclude E coli from adhering to uroepithelial cells. However, the data supporting this approach is not clear and it is possible that the urinary bladder mounts immunodefense by detecting E. coli adherence to receptors. In contrast, limited data does exist for cranberry juice extracts which contain proranthocyaninids that block adherence to receptors. Use of drugs to modify urinary pH may facilitate the antibacterial effects of urine. The presence of ionizable organic acids (hippuric and γ-hydroxybutyric acid) in an acidic pH may enhance the antibacterial activity of the urine. Antibacterial activity may be increased by ingestion of cranberry juice (if urinary pH is acidic), which contains precursors of hippuric acid. Methenamine releases formaldehyde at a urinary pH of 5.5 or less, which also can increase antibacterial activity of urine. In human patients, urinary acidification is very difficult to achieve and can result in dissolution of crystals. Urinary acidification is recommended rarely and only with concomitant use of organic acids (or methenamine). Local urinary analgesics, such as phenazopyridine, rarely are indicated for the management of urinary tract infections. Dysuria is most likely to respond to appropriate antimicrobial therapy. These drugs cause methemoglobinemia in cats. Drugs or nutraceutical products that enhance polysulfated glycosaminoglycan synthesis (e.g., ADEQUAN, pentosan polysulfate, glucosamine, chondroitin sulfate) might be considered for patients with complicated UTI. Such materials may cover or help repair the uroepithelium, thus decreasing bacterial adherence.Probiotics might also be considered for their ability to potentially replace emerging resistant populations in the gastrointestinal tract with "good" bacteria. Note that many probiotics are characterized by poor quality; accordingly, attention should be made to stick with a brand name product. Doses should be in terms of billions in order to assure colonization of the gastrointestinal tract. IN general, target organisms should include lactobacilli, bifidobacterium, enterococci, streptococci and others.

Duration of Therapy

The duration for successful treatment of uncomplicated lower UTIs might be as short as 3 to 5 days, Such an approach is more likely to be successful if high doses and appropriate intervals are chosen. Treatment may need to be longer, however, if infection occurs anywhere other than the uroepithelium. In general, a 10- to 14-day therapeutic regimen has been recommended for the first episode of therapy.The "test for cure" can be based on a second culture 3 to 5 days into therapy. Cure should be anticipated only if the organism count at that time is less than 100 per milliliter of urine. Urine culture a second time just before discontinuation of therapy has been recommended, particularly if antimicrobial prophylaxis is to be implemented. However, increasingly, evidence is emerging that in uncomplicated cases, 3 to 5 days of therapy may be most appropriate. Drugs that have been used successfully by humans for short-term dosing include trimethoprim/sulfonamide combinations, aminoglycosides, selected cephalosporins, and fluorinated quinolones. Drugs that are metabolized to an active metabolite (e.g., enrofloxacin and ceftiofur) may be particularly conducive to single-dose therapy for animals, but these drugs apparently have not been scientifically studied for short-term use. Factors that should preclude single-dose antimicrobial therapy for a lower UTI include recurrence, historical poor response to single-dose therapy, underlying predisposing factors to a UTI (including structural abnormalities of metabolic disorders such as diabetes mellitus, and hyperadrenocorticism), and either pyelonephritis or symptoms of a UTI that have occurred for more than 7 days.

For infections that reflect a relapse, the duration of therapy should be at least 2 weeks; however, for human patients suffering from a relapse, a higher cure rate occurred with a 6-week course of therapy. For animals, a duration of 4 to 6 weeks is recommended. Because relapse is likely to occur shortly after antimicrobial therapy is discontinued, cultures should be collected 7 to 10 days after cessation of therapy. The presence of relapse should lead to a longer course of therapy, perhaps at a higher dose. A new antibiotic should be selected if infection occurs more than 10 days after cessation of therapy; as more time elapses between cessation of therapy and the presence of bacteriuria, the more likely that reinfection is the cause of recurrence.

In the event of relapse after 6 weeks of therapy, 6 months of therapy or more may be necessary. However, if the patient is asymptomatic, strong consideration should be given to no treatment unless mitigating circumstances indicate the need for therapy. Unless the underlying cause of infection can be removed, however, it is likely that resistance will emerge. An attempt should be made to assure that the infecting inoculum is killed with each therapeutic intervention such that the returning infection is characterized by lack of resistanceGreater care should be taken, however, in the selection of antibiotics for longer term therapy, with special consideration to toxicity. Drugs that are used for long-term therapy for human patients include amoxicillin, cephalexin, trimethoprim/sulfonamide combination, or a fluorinated quinolone. Cultures should be repeated monthly, and, as long as significant bacteria are not present, the drug need not be changed. Should relapse occur after a drug is discontinued, the same drug or a new drug should be administered for a longer course of therapy. Long-term therapy may be particularly important for animals in which renal parenchymal damage is a risk.

Prophylaxis

Long-term prophylaxis can be implemented for patients at risk for recurrence. Prophylaxis (by definition) can occur only after the infection has been eradicated. The use of low doses of antimicrobials in the presence of bacteriuria is likely to lead to the generation of resistant organisms and is contraindicated. Thus, prophylactic antimicrobial therapy of UTIs is indicated for reinfection but not relapse (the latter suggests that the organism was never completely eradicated). The antimicrobial chosen for long-term prophylaxis should be both safe and inexpensive. Trimethoprim/sulfonamide combinations (monitor for immune-mediated reactions) and fluoroquinolones are examples. The dose generally can be reduced to 30% to 50% of the full dose.

Despite this low dose, therapeutic concentrations of drugs are likely to be achieved in urine; in addition, subtherapeutic concentrations of drugs often are sufficiently inhibitory to prevent infection of the uroepithelium. The drug should be administered at night to maximize contact of the drug with the urinary tract. Intermittent urine cultures (monthly) are indicated to detect breakthrough infections in animals receiving long-term antimicrobial prophylaxis. Negative cultures for 6 to 9 months or more may indicate that prophylaxis is no longer necessary.

Pyelonephritis

Treatment for pyelonephritis may require hospitalization. Oral antibiotic therapy is acceptable for mild to moderate cases as long as oral therapy is tolerated well. Because renal dysfunction can be life threatening, antimicrobial selection should ultimately be based on culture and susceptibility data. Therapy can be initiated empirically; however, resistance among E. coli organisms should lead to selections other than amoxicillin and ampicillin. Trimethoprim/sulfonamide combinations, amoxicillin/clavulanic acid combinations, cephalosporins, and fluorinated quinolones remain good choices for human patients. Pyelonephritis can be associated with bacteremia, particularly gram negative. Clinical signs indicative of severe, life-threatening infection should lead to parenteral antibiotic therapy with predictably effective drugs (e.g., aminoglycosides, fluorinated quinolones, extended spectrum beta-lactams, and third-generation cephalosporins). Combination therapy also should be strongly considered. The high concentration of antibiotic that facilitates treatment of the lower urinary tract (bladder and lower) may not occur in pyelonephritis; thus attention must be closely paid to using sufficiently high doses and frequent dosing. Drugs whose efficacy is dependent on a hypotonic environment (compared with the target organism) such as beta-lactams may be less effective in the face of medullary hypertonicity. As with infection lower in the tract, bacterial numbers should decrease dramatically within the first 48 hours. For uncomplicated pyelonephritis, 14 days of therapy may be sufficient. Cultures should be repeated as previously indicated during and within 1 to 2 weeks of discontinuation of therapy. Complications such as abcessation may require surgical intervention and longer term therapy.