Complicated UTI: the 2025 IDSA guideline

In a fictional case, a 67-year-old man arrives at the emergency department with fever, chills and right flank pain for two days. He took ciprofloxacin four months ago for a urinary infection and has a urine culture from that time with quinolone-resistant Escherichia coli. His blood pressure is low, his respiratory rate is high and he appears confused. The attending physician needs to choose the antibiotic now, without a new culture, and then decide when to switch to a pill and how many days to treat.

The Infectious Diseases Society of America (IDSA) published in 2025 its first guideline of its own on complicated urinary tract infection, divided into four parts: classification and methods, antibiotic choice, switching from the intravenous to the oral route and treatment duration (Trautner et al., 2025). The text appeared first on the society's website, in July 2025, and later in Clinical Infectious Diseases. The recommendations answer the three questions of the attending physician in the opening.

This text summarizes the new classification, the four steps for choosing the empiric antibiotic, the conditions for switching to the oral route and the short courses. Uncomplicated cystitis is left for the text on gepotidacin; here the subject is the infection that has already left the bladder or that occurs in someone with a catheter.

Key points

• Complicated urinary tract infection is now defined by signs that the infection has left the bladder, such as fever, chills, instability or flank pain, or by the presence of a catheter, and not by anatomy (Trautner et al., 2025).
• The empiric choice follows four steps: severity, risk of resistance, patient factors and, in sepsis, the local antibiogram.
• With sepsis, the guideline suggests a third- or fourth-generation cephalosporin, a carbapenem, piperacillin-tazobactam or a fluoroquinolone before the new agents.
• In the stable patient, the suggestion is to spare the carbapenem and the new agents.
• With clinical improvement, switching to the oral route is suggested even when there is Gram-negative bacteremia.
• The suggested course is 5 to 7 days of a fluoroquinolone or 7 days of another class, including bacteremia; suspected prostatitis may require 10 to 14 days.

Summary

• Why complicated urinary tract infection is a reasoning problem
• The new classification: what makes the infection complicated
• Empiric choice: the four steps
• Sepsis or no sepsis: the prioritized classes
• Antibiogram, previous cultures and definitive therapy
• From the vein to the mouth
• How long to treat
• How NEXO helps train this reasoning
• Frequently asked questions
• Take-home messages

Why complicated urinary tract infection is a reasoning problem

Urinary tract infection is one of the most common bacterial infections, and complicated urinary tract infection is among the most frequent causes of sepsis and Gram-negative bacteremia. The previous IDSA guideline covered only cystitis and pyelonephritis in women and left out complicated infection and infection in men (Trautner et al., 2025). Since then, many randomized trials have tested new antibiotics in men and women with complicated infection, and resistance of urinary Gram-negative bacilli has gone beyond the limits the 2010 guideline used to indicate empiric treatment.

The reasoning problem lies in balancing two risks. Getting the initial antibiotic wrong in a septic patient increases the chance of death. Always using the broadest antibiotic, in everyone, selects resistant bacteria, which return in the next episode of the same patient and across the whole hospital. The global burden of antimicrobial resistance has already been estimated at more than one million attributable deaths per year, with a projected rise through 2050 (Naghavi et al., 2024). The logic of rational antibiotic use appears in the text on antimicrobial resistance and the logic of AWaRe.

What the guideline tries to solve

The guideline organizes decisions that are often made on autopilot. Instead of a single list of antibiotics, it orders the choices by severity and by the individual risk of resistance. Instead of 10 to 14 days out of habit, it proposes short courses when there is improvement. And instead of keeping the antibiotic in the vein until discharge, it suggests switching to a pill as soon as the patient improves and can swallow.

How the guideline was made

The panel brought together infectious disease specialists, microbiologists, pharmacists and physicians from primary care, emergency medicine, hospital medicine, gynecology and urology. Each question had its own systematic review, with searches up to September 2024, and the strength of the recommendations followed the GRADE method (Trautner et al., 2025). Almost all the recommendations are conditional, with certainty of evidence ranging from very low to moderate. This means the guideline points to a preferred path but leaves room for clinical judgment in each patient.

The new classification: what makes the infection complicated

The most visible change is in the definition. Previously, the infection was called complicated when there was an anatomical or functional abnormality, such as obstruction, retention, a stone or an underlying disease. IDSA replaced that criterion with something visible at the bedside: signs that the infection has gone beyond the bladder (Trautner et al., 2025).

The signs that define a complicated infection

Under the new classification, a urinary infection is complicated when there is fever, other signs of systemic illness (chills, shivering or hemodynamic instability), flank pain or costovertebral angle pain. Acute pyelonephritis falls into this group. Infection with systemic symptoms in someone using an indwelling urinary catheter, a suprapubic catheter or intermittent catheterization is also complicated, because the device favors biofilm, trauma and sediment in the bladder.

Anyone with a long-term urinary catheter, stent or nephrostomy who has symptoms linked to bacteria in the urine is also classified as having a complicated infection. The guideline recognizes one exception: a patient whose catheter was placed recently and who has no signs of infection beyond the bladder can be treated as having simple cystitis.

What was left out

Diabetes, immunosuppression, urological abnormality and recurrent infection are no longer enough to call the infection complicated. A man with cystitis and no fever or flank pain has, under the new classification, an uncomplicated infection. The guideline asks, however, that acute prostatitis be considered before classifying a man's infection as uncomplicated, because it changes the choice of antibiotic and the duration. Bacterial prostatitis, epididymitis and orchitis are outside the scope of the document.

Bacteriuria without symptoms

Every urinary infection presupposes symptoms linked to bacteria in the bladder or beyond it. Without symptoms, the patient has asymptomatic bacteriuria, which follows another IDSA guideline and, in most cases, does not require treatment. This distinction matters in people using a catheter: urethral discomfort caused by the device itself must be separated from the pain of true cystitis. Care with devices also appears in the text on phlebitis and catheter infection.

Empirical choice: the four steps

The guideline proposes a sequence of four steps for choosing the antibiotic before culture (Trautner et al., 2025). The first is to assess severity. The second is to look for individual risk factors for resistant bacteria. The third is to consider other patient factors that increase the risk of an adverse event. The fourth, reserved for the septic patient, is to consult the local antibiogram.

Step 1: severity

The first question is whether the patient is in sepsis. IDSA uses the Sepsis-3 definition: life-threatening organ dysfunction caused by a dysregulated host response to infection, identified by an increase of 2 points or more on the SOFA score (Singer et al., 2016). Screening tools such as qSOFA and SIRS can raise the suspicion in the emergency department. Septic shock is the subgroup that needs a vasopressor to maintain pressure and has lactate above 2 mmol/L despite volume replacement.

Sepsis screening in the emergency department is the topic of the text on sepsis in nursing triage, and the changes in resuscitation targets are in the text on Surviving Sepsis 2026.

Step 2: risk of resistance

The guideline makes two objective suggestions. The first is to avoid the antibiotic to which the patient has already had a resistant bacterium isolated in urine. More recent cultures are worth more than older cultures; in the studies, the median interval between paired cultures was between 3 and 6 months. The second is to avoid fluoroquinolone if the patient has used this class in the last 12 months (Trautner et al., 2025).

In the opening patient, both rules point in the same direction. He used ciprofloxacin four months ago and had quinolone-resistant E. coli. The fluoroquinolone comes off the empirical list.

Step 3: patient factors

The third step is a good practice statement: consider allergies, contraindications and drug interactions before choosing. The guideline recalls the serious adverse effects of fluoroquinolones, such as tendinitis, tendon rupture, aortic aneurysm or dissection, peripheral neuropathy, central nervous system effects, hypoglycemia, QT interval prolongation and Clostridioides difficile colitis. The older aminoglycosides, such as gentamicin, amikacin and tobramycin, are included with caution because of nephrotoxicity.

Step 4: antibiogram, only in sepsis

The antibiogram helps adjust the choice in the septic patient, provided it is local, recent and relevant. Local means from the same service. Recent means based on the previous 12 months. Relevant means done with bacteria from similar patients (Trautner et al., 2025). The guideline suggests choosing an antibiotic to which at least 90 out of every 100 relevant isolates are susceptible in septic shock, and at least 80 out of every 100 in sepsis without shock (Trautner et al., 2025). These cutoffs came from models of increased mortality linked to an inadequate initial antibiotic.

For the patient without sepsis, the guideline makes no recommendation about the antibiogram. Mortality is low in this group, and the initial error weighs little on the outcome, while routine use of broad agents fuels resistance with no clear gain.

Sepsis or no sepsis: the prioritized classes

With severity defined, the guideline separates the options into preferred and alternative. The logic is to prioritize getting it right initially in the septic patient and preserving reserve antibiotics in those who are stable.

With sepsis

For the septic patient with complicated urinary tract infection, the guideline suggests initially choosing among third- or fourth-generation cephalosporins, carbapenems, piperacillin-tazobactam and fluoroquinolones (Trautner et al., 2025). The new agents, such as the recent beta-lactam with beta-lactamase inhibitor combinations, cefiderocol and plazomicin, come later, as do the older aminoglycosides. The recommendation is conditional, with very low to moderate certainty.

Agents with coverage beyond enterobacteria, against Pseudomonas aeruginosa, enterococci or methicillin-resistant Staphylococcus aureus, may be considered when the diagnosis of urinary tract infection is unclear or when these agents are suspected.

Without sepsis

For the patient without sepsis, the suggestion is to choose among third- or fourth-generation cephalosporins, piperacillin-tazobactam and fluoroquinolones, sparing carbapenems, the new agents and the older aminoglycosides (Trautner et al., 2025). Trimethoprim-sulfamethoxazole, amoxicillin-clavulanate and first- or second-generation cephalosporins have been less studied, but may serve in selected situations for empirical oral treatment.

Nitrofurantoin and oral fosfomycin are usually inadequate in complicated infection, because they may not reach sufficient levels in the renal parenchyma and blood. This limitation explains why cystitis and complicated infection require different lists. The growing resistance of E. coli also weighs on cystitis, where the choice depends on the drug, the bacterium and the host (Cantón et al., 2026), a point also discussed in the text about gepotidacin in uncomplicated urinary tract infection.

Table: the recommendations on one page

Situation	IDSA 2025 suggestion	Strength and certainty
Sepsis from complicated infection	Third- or fourth-generation cephalosporin, carbapenem, piperacillin-tazobactam or fluoroquinolone before the new agents	Conditional; very low to moderate
Complicated infection without sepsis	Third- or fourth-generation cephalosporin, piperacillin-tazobactam or fluoroquinolone, sparing carbapenem	Conditional; very low to moderate
Resistant bacteria in a previous urine culture	Avoid the antibiotic to which it was resistant	Conditional; very low
Fluoroquinolone in the last 12 months	Avoid empirical fluoroquinolone	Conditional; very low
Susceptibility testing in sepsis	Use only if local, recent and relevant	Conditional; very low
Culture with susceptibility available	Adjust to a targeted-spectrum antibiotic	Conditional; low
Clinical improvement, oral route possible	Switch to oral route, with or without bacteremia	Conditional; low or very low
Duration with improvement	Fluoroquinolone for 5 to 7 days or another class for 7 days	Conditional; moderate or very low
Gram-negative bacteremia with improvement	7-day course instead of 14	Conditional; low

Susceptibility testing, previous cultures and definitive therapy

The empirical antibiotic is an initial bet. The guideline suggests adjusting to an effective, targeted-spectrum antibiotic as soon as the urine culture provides identification and susceptibility, instead of keeping the broad regimen until the end (Trautner et al., 2025). The recommendation is conditional, with low certainty, and gives weight to narrowing the spectrum without losing efficacy.

What an effective antibiotic is

The guideline defines an effective antibiotic as one that reaches therapeutic levels in the urine and in the relevant tissue and that is active against the causative bacterium. In bacteremia, it also needs to reach adequate levels in the blood. The count of treatment days starts on the first day of effective therapy, not on the first day of any antibiotic.

Back to the patient from the opening

The man from the opening is septic and has two risk factors for quinolone resistance. According to the guideline, the initial choice would fall among a third- or fourth-generation cephalosporin, piperacillin-tazobactam or a carbapenem. If the previous E. coli produced extended-spectrum beta-lactamase, the carbapenem gains weight in the empirical choice. When the new culture arrives, the goal is to narrow the spectrum to the narrowest effective antibiotic available. The reasoning of withdrawing what is no longer needed is the same as that of deprescribing.

Cultures before the antibiotic

The value of the previous culture depends on it existing. That is why the urine culture and, in the septic patient, the blood culture collected before the first dose of the antibiotic are part of the same reasoning. A culture properly collected today is the data that will guide tomorrow's adjustment and the empirical choice for the next episode.

From the vein to the mouth

The second part of the guideline deals with switching from the intravenous to the oral route (Trautner et al., 2025). From a pharmacological point of view, the effect of the antibiotic depends on the levels in the blood and tissues, not on the route. The switch reduces complications of venous access, nursing time, fluid and sodium volume, length of stay and costs.

The recommendation

In the patient with complicated infection, including pyelonephritis, treated initially by the intravenous route, who is improving, can take oral medication and has an effective oral option, the guideline suggests switching to the oral route instead of keeping the intravenous route until the end (conditional, low certainty). The same suggestion applies to those with associated Gram-negative bacteremia (conditional, very low certainty). In the trials with bacteremia, the switch occurred in patients without fever, stable and with the source controlled, that is, with any urinary obstruction already relieved.

The five conditions for switching

The guideline lists five points to check before the switch (Trautner et al., 2025). The patient is improving. The source has been controlled, when that applies. The intestine absorbs the tablet. There is an oral regimen active against the bacterium that reaches adequate levels where needed, including in the blood if there is bacteremia. And no psychosocial or economic factor favors the venous route.

Poor absorption, short bowel, ileus, severe diarrhea, vomiting, delayed gastric emptying and intestinal hypoperfusion due to shock can prevent the switch. Calcium, iron and magnesium supplements and sucralfate reduce the absorption of some antibiotics, such as fluoroquinolones, and call for attention to timing.

What the meta-analyses show

Two recent meta-analyses support early switching in Gram-negative bacteremia. One gathered 13 studies and about 10,000 patients and found similar or lower mortality in the group that switched to oral therapy in the first days, with no difference in clinical failure (Onorato et al., 2026). The other gathered 16 studies and found similar mortality and treatment failure between step-down oral therapy and continuous intravenous therapy, with certainty of evidence ranging from very low to moderate by GRADE (Honda et al., 2026). Both groups of authors note that most of the data come from observational studies.

How long to treat

The third recommendation addresses duration (Trautner et al., 2025). The guideline starts from an uncomfortable fact: infections are often treated for longer than guidelines suggest, and each extra day of antibiotic increases the risk of adverse effects and the selection of multidrug-resistant bacteria. The preference for the shortest effective course is also in the recommendations of antimicrobial stewardship programs (Sheppard, 2022).

The recommendation for complicated infection

In patients with complicated infection, including pyelonephritis, who improve with effective therapy, the guideline suggests a short course: 5 to 7 days of a fluoroquinolone (conditional, moderate certainty) or 7 days of an antibiotic from another class (conditional, very low certainty), instead of 10 to 14 days (Trautner et al., 2025). Most of the studies used a fluoroquinolone at a time of less resistance. With an oral beta-lactam, the evidence for short courses is more limited, and higher regimens may be needed.

The recommendation for bacteremia

In complicated infection with Gram-negative bacteremia and clinical improvement, the suggestion is 7 days instead of 14 (conditional, low certainty). A meta-analysis of six studies with 4448 patients with bacteremia of urinary origin found no difference in mortality or recurrence of bacteremia between the two courses (Luna et al., 2026). The authors conclude that 7 days may be enough for most stable patients, with the caveat that the 30-day data did not reach the formal non-inferiority limit.

Who may need more time

Most studies excluded patients with an indwelling urinary catheter, severe sepsis, immunosuppression, urinary abscess, chronic kidney disease, bacterial prostatitis, complete obstruction or urological surgery. These groups may need an individualized approach. In men with febrile infection and suspected acute prostatitis, the guideline allows 10 to 14 days, although the evidence for this subgroup is scarce. When improvement is slow, the next question is whether there is a focus that needs drainage or relief of obstruction.

Time and the transition of care

A 7-day course often ends at home. This makes the discharge prescription, the written end date and the scheduled follow-up part of the treatment. The most common failures at this point are described in the text about the transition from hospital discharge to primary care.

Haste, anchoring and cloudy urine

Urinary tract infection is a diagnosis of convenience on the ward. Cloudy urine, a positive dipstick and fever seem to close the picture, and the search stops there. The guideline itself notes that, without symptoms linked to the bladder or kidney, bacteria in the urine are bacteriuria, not infection. In an older, confused patient, labeling the fever as urinary too early can hide pneumonia, an abscess or another cause of sepsis, such as melioidosis in an endemic area. Fatigue and irritation with repeated cases also weigh in, as the text on affective bias in diagnosis shows.

How NEXO helps train this reasoning

In NEXO, the case of the day is fictional, with a real clinical presentation and up to six progressive clues, and you propose the diagnosis by searching for the ICD-10 code.

The limits on using the game are described on the Economy page of NEXO. If you research the teaching of clinical reasoning, the page for Researcher explains how to get in touch. Professionals who write or review cases find information on Curator and on Creators, and institutions can learn about the proposal on the partnerships page. Choosing the right ICD-10 code completes this training.

Frequently asked questions

What makes a urinary tract infection complicated in IDSA 2025?

Signs that the infection has left the bladder, such as fever, chills, hemodynamic instability or flank pain, and the presence of a urinary catheter, stent or nephrostomy with symptoms. Diabetes, immunosuppression and anatomical abnormalities are no longer enough on their own.

Does a man with a urinary tract infection always have a complicated infection?

The new classification allows uncomplicated cystitis in men, provided there is no fever, no systemic signs and no flank pain. The guideline asks that acute prostatitis be considered first, because it changes the choice and duration of treatment.

Which antibiotic should be used in the septic patient?

The guideline suggests choosing among a third- or fourth-generation cephalosporin, a carbapenem, piperacillin-tazobactam and a fluoroquinolone, adjusting for the previous culture, recent antibiotic use, patient factors and the local antibiogram. The new agents are kept as reserve.

Why spare the carbapenem in the stable patient?

In the stable patient, mortality from complicated infection is low and the initial error weighs little on the outcome. Routine use of a carbapenem in this group selects resistant bacteria with no clear benefit for the patient.

Does bacteremia prevent switching to an oral tablet?

The guideline suggests switching also in Gram-negative bacteremia, provided the patient is improving, afebrile, stable, with the source controlled and with an oral antibiotic that reaches good blood levels (Trautner et al., 2025).

Are seven days enough in pyelonephritis?

With clinical improvement, the guideline suggests 5 to 7 days of a fluoroquinolone or 7 days of another class, including in bacteremia. Abscess, obstruction, suspected prostatitis and immunosuppression call for an individual approach.

Take-home messages

• Complicated urinary tract infection is now defined by systemic signs or by the catheter, not by anatomy.
• The empirical choice follows four steps: severity, risk of resistance, patient factors and the local antibiogram in sepsis.
• A previous culture with resistance and a fluoroquinolone in the last 12 months change the choice.
• In the septic patient, getting the initial choice right comes first; in the stable patient, the guideline asks to spare the carbapenem and the new agents.
• The culture with susceptibility results is the moment to narrow the spectrum.
• With improvement, switching to the oral route is suggested, including in bacteremia.
• A fluoroquinolone for 5 to 7 days or another class for 7 days, with an individual approach for prostatitis, abscess and obstruction (Trautner et al., 2025).

Milestone inventory

• 2010: IDSA publishes the guideline on uncomplicated cystitis and pyelonephritis in women, without addressing complicated infection.
• 2016: Sepsis-3 defines sepsis as organ dysfunction caused by a dysregulated response to infection (Singer et al., 2016), and the FDA restricts fluoroquinolones in uncomplicated urinary tract infection.
• 2019 and 2020: randomized trials show that 7 days are not inferior to 14 in Gram-negative bacteremia, mostly of urinary origin (Luna et al., 2026).
• 2024: the global estimate of the burden of antimicrobial resistance projects a rise in deaths by 2050 (Naghavi et al., 2024).
• 2025: the BALANCE trial extends the 7-day course to bacteremias of various origins, and IDSA releases in July the first guideline on complicated urinary tract infection.
• 2026: meta-analyses of early switching to the oral route in Gram-negative bacteremia reinforce the recommendation (Onorato et al., 2026; Honda et al., 2026), and the guideline is published in Clinical Infectious Diseases.

Play the case of the day on NEXO.

References

1. Trautner BW, Cortés-Penfield NW, Gupta K, et al. Clinical Practice Guidelines by Infectious Diseases Society of America 2025 Guidelines on Management and Treatment of Complicated Urinary Tract Infections--Introduction and Methods. Clin Infect Dis. 2026. PMID 41467387. DOI 10.1093/cid/ciaf459. Read in full in the complete manuscript on the IDSA website.
2. Trautner BW, Cortés-Penfield NW, Gupta K, et al. Clinical Practice Guidelines by Infectious Diseases Society of America: 2025 Guideline on Management and Treatment of Complicated Urinary Tract Infections-Selection of Antibiotic Therapy for Complicated Urinary Tract Infections. Clin Infect Dis. 2026. PMID 41419213. DOI 10.1093/cid/ciaf460. Read in full from the complete manuscript on the IDSA website.
3. Trautner BW, Cortés-Penfield NW, Gupta K, et al. Clinical Practice Guidelines by Infectious Diseases Society of America: 2025 Guidelines on Management and Treatment of Complicated Urinary Tract Infections - Timing of Intravenous to Oral Antibiotics Transition for Complicated Urinary Tract Infections. Clin Infect Dis. 2026. PMID 41414903. DOI 10.1093/cid/ciaf461. Read in full from the complete manuscript on the IDSA website.
4. Trautner BW, Cortés-Penfield NW, Gupta K, et al. Clinical Practice Guidelines by Infectious Diseases Society of America (IDSA): 2025 Guidelines on Management and Treatment of Complicated Urinary Tract Infections--Duration of Antibiotics for Complicated UTI. Clin Infect Dis. 2026. PMID 41419448. DOI 10.1093/cid/ciaf462. Read in full from the complete manuscript on the IDSA website.
5. Singer M, Deutschman CS, Seymour CW, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA. 2016;315(8):801-810. PMID 26903338.
6. Luna G, Rodriguez A, Igbonagwam C, et al. Seven versus Fourteen Days of Antibiotics for Gram-Negative Bacteremia from a Urinary Tract Source: a Systematic Review and Meta-analysis. Infect Dis Ther. 2026;15(2):491-505. PMID 41381979.
7. Onorato L, De Luca I, Ferrante M, et al. Oral Stepdown Therapy for Gram-Negative Bloodstream Infections: A Systematic Review and Meta-Analysis. Eur J Clin Invest. 2026;56(8):e70244. PMID 42593719.
8. Honda H, Miwa T, Warren D, et al. Efficacy of a transition to oral antimicrobial therapy for gram-negative bacilli bloodstream infection: systematic review and meta-analysis. Antimicrob Steward Healthc Epidemiol. 2026;6(1):e172. PMID 42318032.
9. Naghavi M, Vollset SE, Ikuta KS, et al. Global burden of bacterial antimicrobial resistance 1990-2021: a systematic analysis with forecasts to 2050. Lancet. 2024;404(10459):1199-1226. PMID 39299261.
10. Sheppard D. Short-course antibiotic therapy: The next frontier in antimicrobial stewardship. Can Commun Dis Rep. 2022;48(11-12):496-501. PMID 38173469.
11. Cantón R, Lim AN, Cortés V, et al. Uncomplicated Urinary Tract Infections: From an Invisible Impact to a Visible Change in Complex Care. Clin Infect Dis. 2026;82(6):e1194-e1199. PMID 41738831.