Comparison of Ciprofloxacin (7 Days) and Trimethoprim-Sulfamethoxazole (14 Days) for Acute Uncomplicated Pyelonephritis in Women a Randomized Trial

ORIGINAL CONTRIBUTION

Comparison of Ciprofloxacin (7 Days) and Trimethoprim-Sulfamethoxazole (14 Days) for Acute Uncomplicated Pyelonephritis in Women A Randomized Trial

David A. Talan, MD Context The optimal antimicrobial regimen and treatment duration for acute un- Walter E. Stamm, MD complicated pyelonephritis are unknown. Thomas M. Hooton, MD Objective To compare the efficacy and safety of a 7-day ciprofloxacin regimen and a 14-day trimethoprim-sulfamethoxazole regimen for the treatment of acute pyelo- Gregory J. Moran, MD nephritis in women. Thomas Burke, MD Design Randomized, double-blind comparative trial conducted from October 1994 Abdollah Iravani, MD through January 1997. Jonathan Reuning-Scherer, PhD Setting Twenty-five outpatient centers in the United States. Deborah A. Church, MD Patients Of 378 enrolled premenopausal women aged at least 18 years with clinical diagnosis of acute uncomplicated pyelonephritis, 255 were included in the analysis. N THE UNITED STATES AT LEAST Other individuals were excluded for no baseline causative organism, inadequate re- 250 000 episodes of acute pyelo- ceipt of study drug, loss to follow-up, no appropriate cultures, and other reasons. nephritis occur annually among Interventions Patients were randomized to oral ciprofloxacin, 500 mg twice per day adult women, resulting in as many for 7 days (with or without an initial 400-mg intravenous dose) followed by placebo Ias 100 000 hospitalizations.1,2 How- for 7 days (n = 128 included in analysis) vs trimethoprim-sulfamethoxazole, 160/800 ever, the optimal antimicrobial regi- mg twice per day for 14 days (with or without intravenous ceftriaxone, 1 g) (n = 127 men and duration of therapy for this in- included in the analysis). fection have not been established. Main Outcome Measure Continued bacteriologic and clinical cure, such that al- In contrast to the paucity of con- ternative antimicrobial drugs were not required, among evaluable patients through trolled treatment trials in women with the 4- to 11-day posttherapy visit, compared by treatment group. pyelonephritis, recent studies of women Results At 4 to 11 days posttherapy, bacteriologic cure rates were 99% (112 of 113) with uncomplicated lower urinary tract for the ciprofloxacin regimen and 89% (90 of 101) for the trimethoprim- infections have established that these sulfamethoxazole regimen (95% confidence interval [CI] for difference, 0.04-0.16; infections can be treated effectively with P = .004). Clinical cure rates were 96% (109 of 113) for the ciprofloxacin regimen antimicrobials for less than 7 to 14 and 83% (92 of 111) for the trimethoprim-sulfamethoxazole regimen (95% CI, 0.06- days.3-5 Three-day regimens of trimeth- 0.22; P = .002). Escherichia coli, which caused more than 90% of infections, was more frequently resistant to trimethoprim-sulfamethoxazole (18%) than to ciprofloxacin (0%; oprim-sulfamethoxazole or fluoroqui- PϽ.001). Among trimethoprim-sulfamethoxazole–treated patients, drug resistance was nolones appear to provide adequate associated with greater bacteriologic and clinical failure rates (PϽ.001 for both). Drug- cure rates, while being associated with related adverse events occurred in 24% of 191 ciprofloxacin-treated patients and in lower cost and fewer adverse effects 33% of 187 trimethoprim-sulfamethoxazole–treated patients, respectively (95% CI, than longer regimens. The current stan- −0.001 to 0.2). dard duration of therapy for acute un- Conclusions In our study of outpatient treatment of acute uncomplicated pyelo- complicated pyelonephritis is 14 days, nephritis in women, a 7-day ciprofloxacin regimen was associated with greater bac- but few studies of shorter-course teriologic and clinical cure rates than a 14-day trimethoprim-sulfamethoxazole regi- therapy for this infection have been un- men, especially in patients infected with trimethoprim-sulfamethoxazole–resistant strains. dertaken. Increasing antimicrobial re- JAMA. 2000;283:1583-1590 www.jama.com Author Affiliations and Financial Disclosures are listed MD, Olive View-UCLA Medical Center, 14445 Olive See also Patient Page. at the end of this article. View Dr, North Annex, Sylmar, CA 91342 (e-mail: Corresponding Author and Reprints: David A. Talan, [email protected]).

Downloaded From: https://jamanetwork.com/ on 09/29/2021 SHORT-COURSE ANTIMICROBIAL REGIMENS FOR PYELONEPHRITIS

sistance and concomitant reduced clini- of more than 229 µmol/L (2.6 mg/dL) or at days 3 to 5 during therapy, and at days cal efficacy of amoxicillin and ampicillin creatinine clearance level of less than 30 4 to 11 and 22 to 48 posttherapy (ie, fol- for urinary tract infections have been mL/min (0.50 mL/s); persistent vomit- lowing the end of the initial 14-day regi- observed for the last 2 decades, greatly ing; hospital admission; fluoroquino- men for each treatment group). Organ- limiting the use of these drugs for treat- lone, sulfaonamide, penicillin, or ceph- isms present in a concentration of 103 ment of cystitis and pyelonephritis.4 alosporin allergy; pregnancy or lactation; colony-forming units/mL or more were Recently, increasing in vitro resis- a systemic antimicrobial within the pre- identified using standard microbiologi- tance of uropathogens to trimethoprim- vious 72 hours; investigational drug ad- cal techniques. Antimicrobial suscepti- sulfamethoxazole has been reported in ministration within 30 days; and previ- bility testing was performed according the United States.6 However, the ex- ous enrollment in this study. to National Committee for Clinical Labo- tent of this resistance and its clinical im- ratory Standards laboratory proce- plications are unclear. Furthermore, the Study Procedures dures using the modified Kirby-Bauer current shift to outpatient manage- Subjects had a medical history and procedure or by microdilution.7,8 Pre- ment and emphasis on cost-saving strat- physical examination performed, in- therapy blood cultures, if positive, were egies make it timely to evaluate criti- cluding a pelvic examination. Those who repeated within 72 hours. cally treatment regimens for acute, met entry criteria and provided written uncomplicated pyelonephritis. informed consent had pretherapy blood Main Outcome Measures We conducted a multicenter, random- cultures (2 sets) and urine specimens Primary study outcomes were study drug ized,double-blind,outpatienttrialtocom- (collected by midstream clean catch or efficacy based on continued bacterio- pare the efficacy, safety, and relative costs bladder catheterization) obtained for cul- logic and clinical cure (such that no alter- of oral ciprofloxacin, 500 mg twice per ture and susceptibility testing. native antimicrobial therapy for uri- day for 7 days, and oral trimethoprim- Initial stratification to decide whether nary tract infection was prescribed) sulfamethoxazole, 160/800 mg twice per the initial dose of antimicrobial would through the 4- to 11-day posttherapy visit day for 14 days (with or without an ini- be given intravenously or orally was among all enrolled patients who were tial intravenous dose of ciprofloxacin or made by the treating physician. Sub- evaluable for efficacy analyses (efficacy- ceftriaxoneineachgroup)inwomenwith jects were then randomly assigned to re- valid group). Criteria for efficacy evalu- acute uncomplicated pyelonephritis. We ceive 500-mg tablets of ciprofloxacin ation were: (1) enrollment criteria for hypothesized that the 7-day ciproflox- (Cipro, Bayer Corporation Pharmaceu- acute uncomplicated pyelonephritis, (2) acin regimen would have equivalent tical Division, West Haven, Conn) or pretherapy growth of uropathogens of efficacy to the 14-day trimethoprim- 160/800-mg tablets of trimethoprim- 104 colony-forming units/mL or more sulfamethoxazole regimen and in vitro sulfamethoxazole (Bactrim, Roche Labo- (midstream clean catch) or of 103 colony- resistanceoftheinfectingpathogenwould ratories, Nutley, NJ). Subjects receiv- forming units/mL or more (catheteriza- be associated with decreased efficacy. ing an initial intravenous dose were tion); (3) study drug received 5 full days administered 400 mg of ciprofloxacin or more (unless clinical or bacterio- METHODS (ciprofloxacin group) or1gofceftriax- logic failure occurred requiring another Study Population one (trimethoprim-sulfamethoxazole antimicrobial); (4) no other systemic Premenopausal women aged at least 18 group, Rocephin, Roche Laboratories) antimicrobial use; and (5) 1 or more years with a clinical diagnosis of acute over 60 minutes. Antimicrobial in- posttherapy follow-up visits including uncomplicated pyelonephritis were en- fusions were identical and oral medi- repeat urine culture and measurement rolled at 25 outpatient centers (ie, emer- cations were encapsulated in opaque of pyuria. Patients remained in the study gency departments, clinics, and offices) gelatin capsules (#000) for blinding pur- whether the pretherapy blood or urine throughout the United States between poses. Patients who were unable to tol- isolate was subsequently found to be October 1994 and January 1997. The erate oral medications after up to 12 resistant to either or both study drugs. study was approved by each institu- hours of observation were discontin- Continued bacteriologic cure was de- tional review board. Eligible patients had ued from the study. All patients re- fined as pathogen growth of less than 104 flank pain and/or costovertebral angle ceived 1 capsule twice per day for 14 (clean catch) or less than 103 (catheter- tenderness; a temperature higher than days; ciprofloxacin-treated subjects re- ized) colony-forming units/mL through 38.0°C, orally, or higher than 38.6°C, rec- ceived placebo twice per day for the last the posttherapy follow-up visits. Bacte- tally; and pyuria (Ͼ8 leukocytes per mi- 7 days. Compliance was evaluated by riologic failure through the 4- to 11- croliter by hemocytometer or Ͼ5 leu- conducting pill counts from blister pack- day posttherapy visit was classified as su- kocytes per high-power field by the ets at each follow-up evaluation. perinfection with a new uropathogen or sediment examination method). Exclu- persistence of the original uropatho- sion criteria were: severe sepsis; immu- Microbiological Methods gen (based on standard microbiologi- nocompromised condition; diabetes; uro- All patients had a quantitative urine cul- cal identification techniques). Contin- logic abnormality; serum creatinine level ture collected before the start of therapy, ued clinical cure was defined as absence

Downloaded From: https://jamanetwork.com/ on 09/29/2021 SHORT-COURSE ANTIMICROBIAL REGIMENS FOR PYELONEPHRITIS

of all signs and symptoms of illness clinical or bacteriologic failure, or adverse Table 1. Examples of Cost Assumptions through the posttherapy follow-up vis- drug events. The perspective of the cost Used in Health Economics Analysis its. Clinical failure was defined as dete- analysis was that of the third-party payer. Variable Cost, $ rioration, persistence of signs and symp- Direct medical costs were assigned ret- Study drugs toms of the initial infection after 3 days rospectively and were reported in 1997 Oral ciprofloxacin, 500 mg9 3.48 9,10 Intravenous ciprofloxacin, 400 mg9 30.01 of antimicrobial treatment, or recur- dollars. Total costs were calculated Oral trimethoprim- 0.34 rence of signs and symptoms after ini- based on the sum of the costs associated sulfamethoxazole, tial cure or improvement. According to with the following: the initial visit, includ- 160/800 mg9 Intravenous ceftriaxone, 1 g9 36.77 the study protocol, patients with bacte- ing site of care (ie, office or emergency Medical care service riologic or clinical failure discontinued department), antimicrobial, urinalysis, Outpatient clinic or office visit12,13 55.39 Emergency department visit12,13 91.14 taking the study antimicrobial and were and urine culture; a single follow-up visit Hospitalization (per day)10,11 1041.61 prescribed alternative antimicrobial (if there was at least 1 follow-up visit) Medical test Urinalysis14,15 6.68 therapy. with urinalysis and urine culture; and Urine culture14,15 33.38 Secondary study outcomes included additional resources used associated with Diagnostic and imaging test Ultrasound (renal)12,13 90.55 bacteriologic and clinical responses failures and adverse drug events includ- Excretory urogram12,13 86.79 through the 22- to 48-day posttherapy ing subsequent nonstudy-required medi- visit for the efficacy-valid group. For all cal visits, laboratory and radiology tests, patients receiving at least 1 dose of study hospitalizations, therapeutic adjuncts, than anticipated, in part because of vari- drug, intent-to-treat analyses were done. and other antimicrobials, through the able times of patient follow-up. Conse- Adverse drug events, evaluated by ques- patient’s last visit. Costs of hospitaliza- quently, prior to data analysis and tioning patients at each follow-up visit, tions, office visits, laboratory tests, and unblinding, follow-up evaluation inter- were calculated for all enrolled patients. other procedures were estimated by mul- vals were extended to 4 to 11 days and Continued bacteriologic and clinical cure tiplying relative value units by an esti- 22 to 48 days, resulting in 27 (15 cipro- rates were calculated for all enrolled mate of the average cost per relative value floxacin, 12 trimethoprim-sulfamethoxa- patients who had a pretherapy and at unit (TABLE 1).9-15 Cost per cure was cal- zole) and 12 (7 ciprofloxacin, 5 trimeth- least 1 posttherapy evaluation, using the culated based on the total costs for all oprim-sulfamethoxazole) additional last posttherapy visit. Health resource use patients divided by the number of cured efficacy-valid patients, respectively. and cost-per-cure analyses were done for patients in each treatment group. Baseline medical variables were ana- all enrolled patients with cure deter- lyzed using contingency table analysis, mined by clinical status at the last visit, Statistical Methods adjusting for a possible stratum effect. For regardless of whether the patient received It was originally estimated that 180 effi- continuous variables, a 2-way analysis of alternative antimicrobial therapy. cacy-valid patients needed to be enrolled variance model was fitted with stratum Patients without follow-up or with to have an 80% power to demonstrate and treatment as factors to compare the inconclusive follow-up were classified equivalence in cure rates at 5- to 9-day 2 treatment groups. The 95% CIs around as having clinical failure. and 25- to 45-day posttherapy visits the differences in rates for secondary vari- Patients were withdrawn from the using a 1-sided 95% confidence inter- ables were constructed using a normal study if they were either classified as a val (CI) around the difference in bacte- approximation of the binomial distribu- clinical or bacteriologic failure, or if they riologic and clinical responses for the effi- tion with continuity correction. experienced a serious adverse drug cacy-valid group. During the study, new event. At this time, patients were also US Food and Drug Administration RESULTS evaluated for bacteriologic and/or clini- guidelines were issued recommending Study Population cal outcome. Unblinding of study drug 2-sided 95% CIs.16 Thus, for the pri- A total of 378 patients were enrolled and assignment was discouraged, occurred mary efficacy variables, 2-sided 95% CIs followed up for a median of 45 days af- after study withdrawal, and excluded weighted by stratum using a Cochran- ter study entry (range, 0-98 days). subjects from the efficacy-valid group. Mantel-Haenszel procedure were con- Numbers of patients assigned to treat- An attempt was made to follow pa- structed around the difference in success ment groups, who were evaluable for tients (including those who discontin- rates. The 2 treatments were consid- study drug efficacy and assessed at fol- ued taking the study drug) through the ered equivalent at the 5% significance low-up visits, and reasons for nonevalu- 22- to 48-day posttherapy visit. level if the lower bound of the 95% CI ability are summarized in FIGURE 1. around the difference in success rates was Evaluable subjects in the 2 treatment Health Resource Use −10% or more. Based on these changes, groups were well matched with re- and Cost Analyses the estimated sample size was increased spect to demographic characteristics All health care resources used were pro- (ie, 250 efficacy-valid patients) and and symptom severity on study entry spectively collected. Additional resources enrollment was extended, but the num- (TABLE 2). No evaluable patient missed were defined as those associated with ber of efficacy-valid patients was lower more than 2 doses of her prescription.

Downloaded From: https://jamanetwork.com/ on 09/29/2021 SHORT-COURSE ANTIMICROBIAL REGIMENS FOR PYELONEPHRITIS

Evaluation of Bacteriologic Figure 1. Patient Assignment and Follow-up Assessments and Clinical Outcomes Causative Uropathogens and Their Anti- 378 Patients Enrolled and Randomized microbial Resistance. For the efficacy- valid group, 117 (91%) ciprofloxacin and

191 Assigned to Ciprofloxacin Group 187 Assigned to Trimethoprim-Sulfamethoxazole Group 119 (94%) trimethoprim-sulfamethoxa- 66 Patients Received Initial Dose Intravenously 61 Patients Received Initial Dose Intravenously zole–treated subjects had Escherichia coli 125 Patients Received Initial Dose Orally 126 Patients Received Initial Dose Orally isolated from their baseline urine specimens. Other less-frequently isolated uro-

63 Discontinued Study 60 Discontinued Study pathogens included: Enterobacter aero- 17 Did Not Have Baseline Causative Organism 23 Did Not Have Baseline Causative Organism genes (5 ciprofloxacin, 4 trimethoprim- 13 Inadequately Received Study Drug 16 Inadequately Received Study Drug sulfamethoxazole); Staphylococcus 6 Were Lost to Follow-up 13 Were Lost to Follow-up saprophyticus (2 ciprofloxacin, 3 trimeth- 7 Did Not Have Appropriate Cultures 8 Did Not Have Appropriate Cultures oprim-sulfamethoxazole); Klebsiella 7 Had Other Protocol Violations pneumoniae (1 ciprofloxacin, 2 trimeth- 7 Had Other Protocol Violations 2 Had Blind Broken 2 Had Blind Broken 2 Were Given Another Antibiotic for Another Infection oprim-sulfamethoxazole); Citrobacter (1 ciprofloxacin, 1 trimethoprim-sulfamethoxazole); Proteus mirabilis (2 cipro- 128 Were Evaluable for Efficacy Analyses 127 Were Evaluable for Efficacy Analyses floxacin, 1 trimethoprim-sulfamethoxazole); and Enterobacter agglomerans (1 ciprofloxacin). Four patients had 2 uro- Bacteriologic Outcome, Clinical Outcome, Bacteriologic Outcome, Clinical Outcome, pathogens isolated. Forty-seven (18.4%) 4-11 Days After Therapy 4-11 Days After Therapy 4-11 Days After Therapy 4-11 Days After Therapy of all uropathogens (44 E coli,2E aero- 113 Were Evaluable 113 Were Evaluable 101 Were Evaluable 111 Were Evaluable 15 Did Not Have a 15 Did Not Have 26 Did Not Have a 16 Did Not Have genes, and 1 P mirabilis) were resistant Bacteriologic a Clinical Bacteriologic a Clinical to trimethoprim-sulfamethoxazole, while Evaluation Evaluation Evaluation Evaluation 1 strain (P mirabilis) was resistant to ciprofloxacin (0.4%; PϽ.001). Among all Bacteriologic Outcome, Clinical Outcome, Bacteriologic Outcome, Clinical Outcome, E coli strains, trimethoprim-sulfa- 22-48 Days After Therapy 22-48 Days After Therapy 22-48 Days After Therapy 22-48 Days After Therapy 111 Were Evaluable 106 Were Evaluable 108 Were Evaluable 106 Were Evaluable methoxazole resistance was 7% at East- 17 Did Not Have a 22 Did Not Have 19 Did Not Have a 21 Did Not Have ern US centers, 14% in the Midwest, and Bacteriologic a Clinical Bacteriologic a Clinical 32% in Western states (range [per cen- Evaluation Evaluation Evaluation Evaluation ter], 0%-46%). Ceftriaxone resistance was

The ciprofloxacin group was given oral ciprofloxacin, 500 mg twice per day for 7 days (with or without initial reported for 1 E coli isolate. 400-mg intravenous dose) followed by placebo for 7 days. The trimethoprim-sulfamethoxazole group was Fourteen subjects (5.5%) had bacter- given oral trimethoprim-sulfamethoxazole, 160/800 mg twice per day for 14 days (with or without initial in- emia pretherapy, all with E coli. All E coli travenous ceftriaxone, 1 g). blood isolates were susceptible to ciprofloxacin, whereas 4 (29%) were resistant to trimethoprim-sulfamethoxazole. Bacteriologic Outcome. For the ef- Table 2. Demographic and Clinical Characteristics of Women in the Efficacy-Valid Group* ficacy-valid group, 112 (99%) of 113 Ciprofloxacin Trimethoprim-Sulfamethoxazole subjects given oral ciprofloxacin for 7 Group† Group‡ Variable (n = 128) (n = 127) days with or without an initial intrave- Age, median (range), y 25 (18-53) 23 (18-58) nous dose of ciprofloxacin and 90 (89%) White 74 (58) 61 (48) of 101 subjects given oral trimethoprim- Excellent or good health status 126 (99) 124 (98) sulfamethoxazole for 14 days with or Pyelonephritis in last year 13 (10) 12 (10) without an initial intravenous dose of Duration of illness, median (range), d 3 (1-28) 3 (1-30) ceftriaxone had continued bacterio- Severe flank pain§ 45 (35) 48 (38) logic cure through the 4- to 11-day post- Vomiting 38 (30) 38 (30) therapy visit (95% CI, 0.04-0.16 for the Temperature Ͼ39.0°C 30 (23) 30 (24) difference; P = .004; FIGURE 2). In this Bacteremia 4 (3) 10 (8) interval, all failures were due to persis- *Values are expressed as number (percentage) unless otherwise indicated. tence except for 2 trimethoprim-sul- †Given oral ciprofloxacin, 500 mg twice per day for 7 days (with or without an initial 400-mg intravenous dose). ‡Given trimethoprim-sulfamethoxazole, 160/800 mg twice per day for 14 days (with or without intravenous ceftriax- famethoxazole–treated subjects who had one, 1 g). superinfection caused by Staphylococ- §Defined as intense and debilitating. cus aureus and Citrobacter freundii, re-

Downloaded From: https://jamanetwork.com/ on 09/29/2021 SHORT-COURSE ANTIMICROBIAL REGIMENS FOR PYELONEPHRITIS

spectively. Continued bacteriologic cure was given intravenously (Table 3). ued bacteriologic cure rates through the rates through the 22- to 48-day post- Intent-to-treat analysis of continued 4- to 11-day posttherapy visit among tri- therapy visit were 94 (85%) of 111 cipro- clinical cure rates for the ciprofloxacin methoprim-sulfamethoxazole–treated floxacin-treated patients and 80 (74%) group were 82% (137 of 167 patients) subjects were significantly higher among of 108 trimethoprim-sulfamethoxazole– compared with 72% (124 of 172 those infected with trimethoprim- treated patients (95% CI, 0.00-0.21; patients) for the trimethoprim-sulfa- sulfamethoxazole–susceptible as op- P = .08; Figure 2). methoxazole group (95% CI, 0.01-0.19). posed to resistant E coli, 96% (73 of 76) Among women receiving ciprofloxa- Of the 14 subjects with E coli vs 50% (7 of 14), respectively (95% CI, cin, bacteriologic cure rates were simi- bacteremia, 1 of 10 trimethoprim- 0.15-0.77; PϽ.001). Corresponding con- lar regardless of whether the initial dose sulfamethoxazole–treated subjects was tinued clinical cure rates through the 4- was given intravenously (TABLE 3). considered a clinical failure; the re- to 11-day posttherapy visit among tri- Among women receiving trimethoprim- maining subjects were clinical cures. methoprim-sulfamethoxazole–treated sulfamethoxazole, an initial intrave- Effect of Trimethoprim-Sulfa- subjects infected with trimethoprim- nous dose of ceftriaxone was associated methoxazole Resistance on Bacterio- sulfamethoxazole–susceptible vs resis- with a greater bacterial cure rate at days logic and Clinical Outcomes. Contin- tant E coli were also significantly differ- 4 to 11, but not at 22 to 48 days posttherapy, compared with that seen for Figure 2. Continued Bacteriologic and Clinical Cure Rates Through the 4- to 11-Day and 22- women receiving only oral trimetho- to 48-Day Posttherapy Visits for Women with Acute Uncomplicated Pyelonephritis prim-sulfamethoxazole (Table 3). Intent- to-treat analysis of continued bacterio- Ciprofloxacin Trimethoprim-Sulfamethoxazole logic cure for the ciprofloxacin group was Bacterial Eradication Clinical Cure P = .004 84% (128 of 153 patients) compared with P = .002 100 P = .02 74% (112 of 152 patients) for the tri- P = .08 methoprim-sulfamethoxazole regimen (95% CI, 0.01-0.19). Of the 14 patients 80

with E coli bacteremia, 2 (20%) of 10 tri- 60 methoprim-sulfamethoxazole–treated subjects had urinary bacterologic per- 40 Patients, % sistence. All 4 ciprofloxacin-treated subjects achieved bacterologic cure. 20 Clinical Outcome. For the efficacy- 0 valid group, 109 (96%) of the 113 sub- n = 113 n = 101 n = 111 n = 108 n = 113 n = 111 n = 106 n = 106 jects given oral ciprofloxacin for 7 days 4-11 Days 22-48 Days 4-11 Days 22-48 Days with or without an initial intravenous After Therapy After Therapy After Therapy After Therapy dose of ciprofloxacin and 92 (83%) of The ciprofloxacin group was given an oral dose of ciprofloxacin, 500 mg twice per day for 7 days (with or 111 subjects given oral trimethoprim- without initial 400-mg intravenous dose) followed by placebo for 7 days. The trimethoprim-sulfamethoxazole group was given oral trimethoprim-sulfamethoxazole, 160/800 mg twice per day for 14 days (with or without sulfamethoxazole for 14 days with or initial intravenous ceftriaxone, 1 g). without an initial intravenous dose of ceftriaxone had continued clinical cure Table 3. Continued Bacteriologic and Clinical Cure Rates Stratified by Oral or Intravenous through the 4- to 11-day posttherapy visit Administration of the First Dose of Antimicrobial* (95% CI, 0.06-0.22; P = .002; Figure 2). Trimethoprim- Sulfamethoxazole Of the clinical failures, 1 (25%) of the Ciprofloxacin Group† Group‡ ciprofloxacin-treated patients and 8 (42%) of the trimethoprim-sulfa- Response Oral Intravenous Oral Intravenous methoxazole–treated patients failed Bacteriologic cure through 4-11 75 (100) 37 (97) 61 (85) 29 (100) days after therapy within the first 4 days of treatment. Con- Bacteriologic cure through 22-48 60 (83) 34 (87) 54 (72) 26 (63) tinued clinical cure rates through the 22- days after therapy to 48-day posttherapy visit were 96 Clinical cure through 4-11 days 72 (96) 37 (98) 66 (85) 26 (79) (91%) of 106 ciprofloxacin-treated pa- after therapy tients and 82 (77%) of 106 trimethoprim- Clinical cure through 22-48 days 63 (89) 33 (94) 58 (78) 24 (75) after therapy sulfamethoxazole–treated patients (95% *Includes patients with eradication of the initially infecting strain and absence of recurrent infection, or continued clini- CI, 0.03-0.23; P = .02; Figure 2). cal cure, not requiring alternative antimicrobial therapy through the visits after therapy. Values are expressed as number (percentage). For both treatment groups, clinical †Given oral ciprofloxacin, 500 mg twice per day for 7 days (with or without an initial 400-mg intravenous dose). cure rates were similar regardless of ‡Given trimethoprim-sulfamethoxazole, 160/800 mg twice per day for 14 days (with or without intravenous ceftriaxone, 1 g). whether an initial dose of antimicrobial

Downloaded From: https://jamanetwork.com/ on 09/29/2021 SHORT-COURSE ANTIMICROBIAL REGIMENS FOR PYELONEPHRITIS

ent: 92% (76 of 83) and 35% (6 of 17), patients vs 2 of 9 patients, respectively). events causing study drug discontinu- respectively(95%CI,0.29-0.83;PϽ.001). However, continued clinical cure rates ationoccurredin11(6%)oftheciproflox- Oftrimethoprim-sulfamethoxazole–treat- were similar (4 of 7 patients vs 2 of 10 acin-treated patients and 21 (11%) tri- ed women with resistant strains, 6 of 7 patients, respectively). The 1 patient with methoprim-sulfamethoxazole–treated who had bacteriologic failure also had a ciprofloxacin-resistant P mirabilis strain patients. clinical failure. Among ciprofloxacin- was treated with ciprofloxacin and Additional health resource use (ie, ex- treated women who were infected with achieved bacteriologic and clinical cure. cluding resources required for initial trimethoprim-sulfamethoxazole–resis- management, and posttherapy fol- tantstrains,bacteriologicandclinicalcure Evaluation of Safety, low-up visits for patients without ear- ratesthroughthe4-to11-dayposttherapy Health Resource Use, and Costs lier treatment failure) was higher in all visit were 100% (18 of 18) and 95% (18 Safety,healthresourceuse,andcostswere categories for patients treated with tri- of 19), respectively. evaluatedforall378enrolledpatients.Ad- methoprim-sulfamethoxazole than Among women infected with trimeth- verse drug events occurred in 24% of 191 ciprofloxacin, with the exception of ra- oprim-sulfamethoxazole–resistantE coli, ciprofloxacin-treated patients and in 33% diological procedures (TABLE 5). Since continued bacteriologic cure through the of 187 trimethoprim-sulfamethoxazole– more trimethoprim-sulfamethoxazole– 4- to 11-day posttherapy visit occurred treated patients, respectively (95% CI, treated patients had clinical failures and more frequently among those receiving −0.001 to 0.2) (TABLE 4). Gastrointesti- accrued high costs, the mean total cost an initial intravenous dose of ceftriax- nal events, headache, and rash tended per patient was greater for trimethoprim- one than those only treated with oral tri- to occur more frequently in the trimeth- sulfamethoxazole–treated patients methoprim-sulfamethoxazole (5 of 5 oprim-sulfamethoxazole group. Adverse ($687) than for ciprofloxacin-treated patients ($531) by 29% or $156 (95% CI, Table 4. Rates of Adverse Drug Events Among Women* −$118 to $443). By the last follow-up Ciprofloxacin Trimethoprim-Sulfamethoxazole visit, clinical cure occurred in 86.4% and Group† Group‡ 89.3% of ciprofloxacin and trimetho- Body System (n = 191) (n = 187) prim-sulfamethoxazole–treated pa- Any event 46 (24) 62 (33) tients, respectively (95% CI, −0.096 to Events causing study antimicrobial discontinuation 11 (6) 21 (11) 0.037). Seventeen ciprofloxacin and 10 Any digestive 17 (9) 29 (16) trimethoprim-sulfamethoxazole– Nausea 10 (5) 14 (7) treated patients had inconclusive fol- Vomiting 8 (4) 10 (5) low-up and were classified as failures. Diarrhea 6 (3) 6 (3) Since trimethoprim-sulfamethoxazole– Any central nervous system 12 (6) 25 (13) treated patients ultimately required more Headache 9 (5) 19 (10) Dizziness 3 (2) 6 (3) interventions and new antimicrobial pre- Rash 4 (2) 14 (7) scriptions to achieve cure, the mean cost *Values are expressed as number (percentage). per cure was higher for trimethoprim- †Given oral ciprofloxacin, 500 mg twice per day for 7 days (with or without an initial 400-mg intravenous dose). sulfamethoxazole ($770) than cipro- ‡Given trimethoprim-sulfamethoxazole, 160/800 mg twice per day for 14 days (with or without intravenous ceftriaxone, 1 g). floxacin-treated patients ($615) by 25% or $155 (range of difference based on up- Table 5. Additional Medical Resource Use for Women by Treatment Group* per and lower bounds of the 95% CI in cure rates, $102-$207). Ciprofloxacin Trimethroprim-Sulfamethoxazole Group† Group‡ Resource (n = 191) (n = 187) COMMENT Hospital stay, d 20.9 38.0 Although acute uncomplicated pyelo- Medical visits/telephone consultations 23.6 30.5 nephritis is a relatively common and po- Radiological procedures 13.1 9.1 tentially serious infection in young Other therapeutic procedures§ 3.7 6.4 women, few controlled treatment trials Laboratory tests࿣ 33.0 72.7 have been conducted to define optimal Cultures 18.8 28.9 therapy.17 Previous studies suggest that Antimicrobial prescriptions¶ 47.1 57.2 ampicillin should no longer be used since *Values are rates per 100 intent-to-treat subjects for additional resources associated with failures and adverse drug events (ie, excluding resources required for initial management and follow-up visits after therapy for patients without at least 30% of causative E coli strains are earlier treatment failure). ampicillin-resistant.4,18,19 In addition, †Given oral ciprofloxacin, 500 mg twice per day for 7 days (with or without an initial 400-mg intravenous dose). ‡Given trimethoprim-sulfamethoxazole, 160/800 mg twice per day for 14 days (with or without intravenous ceftriax- these and other studies demonstrate that one, 1 g). §Such as intravenous fluids, bladder catheterization, or cystoscopy. treatment of acute uncomplicated py- ࿣Such as urinalysis, complete blood cell count, or serum chemistry. elonephritis in the outpatient setting is ¶Includes all topical and systemic antimicrobials (including antibacterial and antifungal agents). safe and effective in appropriately se-

Downloaded From: https://jamanetwork.com/ on 09/29/2021 SHORT-COURSE ANTIMICROBIAL REGIMENS FOR PYELONEPHRITIS

lected women.20-22 Unfortunately, most acute uncomplicated pyelonephritis were not generally appear to enhance cure previous trials have not assessed pa- generally less than 10% at investigative rates compared with those observed in tients with causative organisms resis- sites. However, increasing rates of tri- patients receiving only oral regimens. tant to the study antimicrobials, thus pre- methoprim-sulfamethoxazole resis- However, patients were not random- cluding evaluation of the effects of in tance have been reported among uri- ized to these treatment strategies and, vitro resistance on clinical outcome. nary E coli isolates.6,27-30 Significantly, in therefore, no conclusions can be made In general, existing studies of treat- vitro resistance to trimethoprim-sulfa- as to their relative therapeutic benefit. ment duration in acute uncomplicated methoxazole was strongly associated with We also evaluated the use of health pyelonephritis have been uncontrolled, bacteriologic and clinical failure in our care resources by women presump- small, and underpowered, or have patients. In vitro resistance, which is cor- tively treated for acute pyelonephritis included patients with complicated infec- related with achievable serum antimi- with the 2 antimicrobial regimens. This tions.17,23-26 In 1 randomized trial com- crobial levels, would not necessarily be trial was not powered to show statisti- paring 2 vs 6 weeks of oral therapy with expected to predict clinical outcome in cal significance in the economic differ- trimethoprim-sulfamethoxazole or ampi- urinary tract infections since urine anti- ences between treatment groups for cillin for women with acute uncompli- microbial levels are many times greater health care resource use. However, re- cated pyelonephritis, 2-week therapy was than serum levels for most antimicrobial source use did appear to be higher in as effective as 6-week therapy with either agents. Indeed, for cystitis, bacteriologic all categories among trimethoprim- drug and resulted in fewer adverse effects, eradicationcanfrequentlybeachievedde- sulfamethoxazole–treated patients, with less frequent selection of resistant strains, spite in vitro resistance of the uropatho- the exception of radiological proce- and lower costs.19 Whether a duration of gen to the treatment antimicrobial.5 Py- dures. Although the prescription cost of therapy even less than 14 days is effec- elonephritis appears to be a deep tissue the ciprofloxacin regimen was greater tive is unclear. In a small trial of hospi- infection for which adequate serum an- than that of the trimethoprim-sulfa- talized men and women with compli- timicrobial levels are important. methoxazole regimen, patients treated cated and uncomplicated pyelonephritis, We also observed that trimethoprim- with the trimethoprim-sulfamethoxa- who were given parenteral netilmicin or sulfamethoxazole resistance was not uni- zole regimen tended to have greater over- ciprofloxacin, Bailey et al23 demon- form among E coli causing acute uncom- all costs, particularly those related to sub- strated that 5 days of treatment resulted plicated pyelonephritis in the United sequent hospitalizations, office visits, and in short-term bacteriologic cure rates of States, with much higher rates seen in the laboratory tests. It should be noted that, approximately 90%. However, late recur- Western states. These data suggest that because these results were derived from rence was observed in some patients.23 trimethoprim-sulfamethoxazole may no a clinical trial, they may not accurately A trial among men and women with acute longer be appropriate as an empiric reflect actual or optimal practice. For ex- complicated and uncomplicated pyelo- therapy in certain geographic areas. In ample, whereas some physicians may nephritis, conducted by Jernelius et al,24 these areas, initial empiric outpatient routinely change antimicrobial drug demonstrated that a 7-day oral regimen therapy with a fluoroquinolone should therapy for an improved patient whose of pivampicillin-pivmecillinam resulted be strongly considered. Patients treated urine isolate is subsequently found to in significantly fewer bacteriologic cures with trimethoprim-sulfamethoxazole demonstrate in vitro resistance to their compared with a 14-day regimen (9 should be followed up carefully until sus- treatment drug, this was not mandated [28%] of 32 patients vs 20 [69%] of 29 ceptibility results are available. in our protocol. Also, since the cost patients, respectively). In contrast, our A frequent practice in emergency de- analysis was conducted from the per- data demonstrate bacteriologic and clini- partments is to administer the first an- spective of the third-party payer, our re- cal cure rates exceeding 90% with a 7-day timicrobial dose intravenously to sicker sults did not capture indirect or intan- course of ciprofloxacin for women with patients with pyelonephritis, particu- gible costs (eg, missed work or school, acute uncomplicated pyelonephritis. The larly those with nausea and vomiting.22 diminished health-related quality of life). difference in bacteriologic outcomes in To mimic clinical practice and include The current investigation was designed these studies may be due to multiple fac- patients with greater illness severity in to address many of the limitations of pre- tors, including differences in the drugs the current trial, physicians were al- vious studies and to conform to the Infec- used and patient inclusion criteria.24 lowed to give the first dose of ciprofloxa- tious Diseases Society of America’s rec- Our data demonstrate an unexpect- cin intravenously prior to an oral cipro- ommendations for conducting clinical edly high prevalence of in vitro resis- floxacin regimen, or ceftriaxone prior to trials by limiting the study population to tance to trimethoprim-sulfamethoxa- an oral trimethoprim-sulfamethoxa- a homogeneous group of young women zole among E coli strains causing acute zole regimen. Ceftriaxone was chosen as with acute uncomplicated pyelonephri- uncomplicated pyelonephritis. In 1994, most representative of emergency de- tis; by employing a randomized blinded prior to beginning the study, resistance partment practice at that time based on design; by treating all women, includ- rates to trimethoprim-sulfamethoxa- an informal survey. Administration of the ing those later found to have resistant zole among uropathogens associated with first antimicrobial dose intravenously did strains with the assigned regimen; by fol-

Downloaded From: https://jamanetwork.com/ on 09/29/2021 SHORT-COURSE ANTIMICROBIAL REGIMENS FOR PYELONEPHRITIS

lowing patients for approximately 4 to Olympia, Wash (Dr Burke); Central Florida Medical timicrobial Disk Susceptibility Tests. 4th ed. Villa- Research Center, Orlando (Dr Iravani); and Bayer Cor- nova, Pa: National Committee for Clinical Labora- 6 weeks; and by enrolling an adequate poration Pharmaceutical Division, West Haven, Conn tory Standards; 1993. sample size to ensure statistical power.31 (Drs Reuning-Scherer and Church). 8. Approved Standard: Methods for Dilution in An- Investigators: J. Plouffe, Ohio State University Hospi- timicrobial Susceptibility Tests for Bacteria That Grow However, this trial does have method- tal, Columbus; J. Johnson, University of Minnesota, Min- Aerobically. 2nd ed. Villanova, Pa: National Commit- ological limitations as well. We intended neapolis; C. Cox, University of Tennessee, Memphis; tee for Clinical Laboratory Standards; 1993. the clinical response to be judged inde- S. Traub, Mercy Hospital, Springfield, Mass; G. Rich- 9. 1997 Drug Topics Redbook. Montvale, NJ: Medi- ard, University of Florida, Gainesville; R. Bessette, Fal- cal Economics Co Inc; 1997. pendently of bacteriologic outcome by lon Clinic, Worcester, Mass; J. Dalovisio, Ochsner Clinic, 10. American Hospital Association. Hospital Statis- the treating physician, but the physi- New Orleans, La; R. Cooper, Madigan Army Medical tics. Chicago, Ill: American Hospital Association; 1995. Center, Tacoma, Wash; L. Payne, United States Air Force 11. Consumer Price Index for Medical Care Services. cians did have access to culture and sus- Medical Center, Lackland AFB, Texas; I. Klimberg, The Washington, DC: Bureau of Labor Statistics; 1997. ceptibility results. Therefore, assess- Urology Center of Florida, Ocala; G. Wells, Alabama 12. Physicians’ Current Procedural Terminology CPT Urology Associations, Birmingham; R. Feldman, Urol- 1998: Professional Edition. Chicago, Ill: American ment of the clinical outcome, and ogy Specialists, Waterbury, Conn; N. Ruiz, Henry M. Medical Association; 1997. subsequent management, could have Jackson Foundation, National Naval Medical Center, 13. Segal MJ, Heald RB, eds. Medicare RBRVS: The Bethesda, Md; F. Wenger, Jr, Albert Einstein Medical Physician’s Guide. Chicago, Ill: American Medical As- been influenced by this knowledge. Center, Philadelphia, Pa; P. Vinson, Norwood Clinical sociation; 1995. In conclusion, for the outcomes of bac- Research Center, Birmingham, Ala; C. Terregino, Coo- 14. Appendix D: revised laboratory RVU list. In: Ac- teriologic and clinical cure, we have dem- per Hospital, University Medical Center, Camden, NJ; counting and Budget Manual for Fiscal and Operating M. Cancio, St Joseph’s Comprehensive Research Insti- Management, Revision #23. Baltimore: State of Mary- onstrated that the 7-day ciprofloxacin tute, Tampa, Fla; R. Shesser, George Washington Clinic, land Health Services Cost Review Commission; 1996. regimen is at least as efficacious as the Washington, DC; M. Sperling, Edinger Medical Group, 15. Current Rates Report. Baltimore: State of Mary- Fountain Valley, Calif; L. A. Walker III, Vanderbilt Uni- land Health Services Cost Review Commission; 1998. 14-day trimethoprim-sulfamethoxa- versity Medical Center, Nashville, Tenn; F. Harchelroad, 16. Points to Consider. Rockville, Md: US Food and zole regimen, and statistically, the cipro- Allegheny General Hospital, Pittsburgh, Pa; J. Linden, Drug Administration; 1992:20. Boston City Hospital, Boston, Mass. 17. Warren JW, Abrutyn E, Hebel JR, et al. Guide- floxacin regimen was superior. In prac- Financial Disclosures: Drs Talan and Moran have re- lines for antimicrobial therapy of uncomplicated acute tice, the shorter duration ciprofloxacin ceived research support and lecture honoraria from Bayer bacterial cystitis and acute pyelonephritis in women. regimen, which was associated with Inc and lecture honoraria from Roche Laboratories. Drs Clin Infect Dis. 1999;29:745-758. Talan and Moran also have received research support 18. Johnson JR, Lyons MF II, Pearce W, et al. Therapy fewer adverse effects, would be ex- from Pfizer Inc, Ortho-McNeil Pharmaceutical, and Eli for women hospitalized with acute pyelonephritis. pected to lead to higher patient accep- Lilly Co, and have been paid speakers for Pfizer Inc and J Infect Dis. 1991;163:325-330. Ortho-McNeil Pharmaceutical. Dr Stamm has re- 19. Stamm WE, McKevitt M, Counts GW. Acute renal tance and medication compliance rates. ceived honoraria and research support from Bayer Inc, infection in women. Ann Intern Med. 1987;106:341- The higher cost of the ciprofloxacin pre- Ortho-McNeil Pharmaceutical, and Proctor & Gamble. 345. Dr Reuning-Scherer was an employee of Bayer Inc and 20. Safrin S, Siegel D, Black D. Pyelonephritis in adult scription would have to be weighed is a statistical consultant for Bayer Inc. Dr Hooton has women. Am J Med. 1988;85:793-798. against the potential additional costs re- received research support and lecture honoraria from 21. Abraham E, Baraff LJ. Oral versus parenteral Bayer Inc. Dr Burke has received research support from therapy of pyelonephritis. Curr Ther Res. 1982;31: lated to more treatment failures with tri- Bayer Inc and Merck & Co Inc. 536-542. methoprim-sulfamethoxazole, espe- Funding/Support: This work was supported by a re- 22. Ward G, Jorden RC, Severance HW. Treatment cially in geographic areas with high rates search grant from Bayer Corporation, Pharmaceuti- of pyelonephritis in an observation unit. Ann Emerg cal Division, West Haven, Conn. Med. 1991;20:258-261. of E coli resistance to trimethoprim- Previous Presentation: Presented in part at the Eighth 23. Bailey RR, Lynn KL, Robson RA, Peddie BA, Smith sulfamethoxazole. These findings should International Congress on Infectious Diseases, Bos- A. Comparison of ciprofloxacin with netilmicin for the ton, Mass, May 15-18, 1998, and the 38th Inter- treatment of acute pyelonephritis. N Z Med J. 1992; not be extrapolated to men, patients with science Conference on Antimicrobial Agents and Che- 105:102-103. complicated infections, or those with se- motherapy, San Diego, Calif, September 24-27, 1998. 24. Jernelius H, Zbornik J, Bauer C-A. One or three Acknowledgment: We are indebted to Adrienne L. week’s treatment of acute pyelonephritis? Acta Med vere sepsis. Patients with acute uncom- Block, PhD; Teresa Tartaglione, PharmD; Roberta Scand. 1988;233:469-477. plicated pyelonephritis should have urine Amore, RN; and Blair Robertson, PhD, for editorial and 25. Gleckman R, Bradley P, Roth R, Hibert D, Pel- scientific contributions. letier C. Therapy of symptomatic pyelonephritis in cultures obtained to both confirm the di- women. J Urol. 1985;133:176-178. agnosis and assess the antimicrobial sus- 26. Ode B, Bro¨ ms M, Walder M, Cronberg S. Failure REFERENCES ceptibility of the infecting uropatho- of excessive doses of ampicillin to prevent bacterial 1. Stamm WE, Hooton TM, Johnson JR, et al. Uri- relapse in the treatment of acute pyelonephritis. Acta gen, so as to better predict bacteriologic nary tract infections. J Infect Dis. 1989;15:400-406. Med Scand. 1980;207:305-307. and clinical outcome. Continued local 2. McCarthy E. Inpatient Utilization of Short-Stay Hos- 27. Beunders AJ. Development of antibacterial resis- pitals, by Diagnosis: United States-1980. Hyattsville, tance: the Dutch experience. J Antimicrob Che- and national surveillance of uropatho- Md: National Center for Health Statistics; 1989. DHHS mother. 1994;33(suppl A):S17-S22. gen antimicrobial-resistance patterns is publication 83-1735. 28. Alon U, Davidai G, Berant M, Merzbach D. 3. Norrby SR. Short-term treatment of uncompli- Five-year survey of changing patterns of suscepti- essential to provide optimal care for cated lower urinary tract infections in women. Rev In- bility of bacterial uropathogens to trimethoprim- women with acute uncomplicated py- fect Dis. 1990;12:458-467. sulfamethoxazole and other antimicrobial agents. 4. Stamm WE, Hooton TM. Management of urinary Antimicrob Agents Chemother. 1987;31:126-128. elonephritis in an era of increasing an- tract infections in adults. N Engl J Med. 1993;329: 29. Winstanley TG, Limb DI, Eggington R, Hancock timicrobial resistance. 1328-1334. F. A 10 year survey of the antimicrobial susceptibility 5. Hooton TM, Winter C, Tiu F, Stamm WE. Ran- of urinary tract isolates in the UK: the Microbe Base Author Affiliations: Divisions of Emergency Medi- domized comparative trial and cost analysis of 3-day project. J Antimicrob Chemother. 1997;40:591-594. cine and Infectious Diseases, Department of Medi- antimicrobial regimens for treatment of acute cystitis 30. Hooton TM, Stamm WE. Diagnosis and treat- cine, Olive View-UCLA Medical Center, University of in women. JAMA. 1995;273:41-45. ment of uncomplicated urinary tract infection. Infect California, Los Angeles (Drs Talan and Moran); Divi- 6. Gupta K, Scholes D, Stamm WE. Increasing preva- Dis Clin North Am. 1997;11:551-581. sion of Infectious Diseases, Department of Medicine lence of antimicrobial resistance among uropatho- 31. Rubin RR, Shapiro ED, Andriole VT, Davis RJ, (Drs Stamm, Hooton, and Burke), Harborview Medi- gens causing acute uncomplicated cystitis in women. Stamm WE. Evaluation of new anti-infective drugs for cal Center, University of Washington, Seattle (Drs JAMA. 1999;281:736-738. the treatment of urinary tract infection. Clin Infect Dis. Stamm and Hooton); Providence St Peter’s Hospital, 7. Approved Standard: Performance Standards for An- 1992;15(suppl 1):S216-S227.

Downloaded From: https://jamanetwork.com/ on 09/29/2021