Peri-interventional antibiotic prophylaxis only vs continuous low-dose antibiotic treatment in patients with JJ stents: a prospective randomised trial analysing the effect on urinary tract infections and stent-related symptoms

Authors


  • F.M. and K.H. equal contribution.
  • Trial registry: ACTRN12612000002886.

Correspondence: Pascal Zehnder, Department of Urology, University Hospital of Bern, Inselspital, 3010 Bern, Switzerland.

e-mail: pascal.zehnder@insel.ch

Abstract

What's known on the subject? and What does the study add?

  • For urinary tract infection (UTI) rates the concept of a peri-interventional antibiotic prophylaxis during endoscopic JJ stent implantation is known to be better than no antibiotic coverage and is therefore recommended by the European Association of Urology. However, there is a lack of evidence concerning the exact antibiotic strategy for the entire stent-indwelling time. In clinical routine, it is an applied practice among urologists to continue antibiotic treatment in a low-dose fashion, even after previous uncomplicated implantations. The intention is to lower the rates of UTIs and to achieve a positive effect on stent-related symptoms (SRSs). This practice is supported by controversial recommendations from sparse publications. However, there exists neither evidence for the benefit, nor for the potential disadvantages of such empiric prevention. Moreover, increasing rates of bacterial drug resistances, growing overall healthcare costs and drug side-effects require a critical antibiotic prescription policy.
  • We analysed UTI and SRS rates in patients given a peri-interventional antibiotic prophylaxis only vs a continuous low-dose antibiotic treatment for the entire stent-indwelling time and showed that the continuous antibiotic low-dose treatment did not reduce the quantity or severity of UTIs and had no effect on SRSs, but involves undesirable disadvantages, e.g. increased drug side-effects and higher rates of resistant bacterial strains, and should therefore be avoided.

Objective

  • To evaluate the antibiotic treatment regime in patients with indwelling JJ stents, the benefits and disadvantages of a peri-interventional antibiotic prophylaxis were compared with those of a continuous low-dose antibiotic treatment in a prospective randomised trial.

Patients and Methods

  • In all, 95 patients were randomised to either receive peri-interventional antibiotic prophylaxis during stent insertion only (group A, 44 patients) or to additionally receive a continuous low-dose antibiotic treatment until stent removal (group B, 51).
  • Evaluations for urinary tract infections (UTI), stent-related symptoms (SRSs) and drug side-effects were performed before stent insertion and consecutively after 1, 2 and 4 weeks and/or at stent withdrawal.
  • All patients received a peri-interventional antibiotic prophylaxis with 1.2 g amoxicillin/clavulanic acid. Amoxicillin/clavulanic acid (625 mg) once daily was administered for continuous low-dose treatment (group B).
  • Primary endpoints were the overall rates of UTIs and SRSs. Secondary endpoints were the rates and severity of drug side-effects.

Results

  • Neither the overall UTI rates (group A: 9% vs group B: 10%), nor the rates of febrile UTIs (group A: 7% vs group B: 6%) were different between the groups.
  • Similarly, SRS rates did not differ (group A: 98% vs group B: 96%).
  • Antibiotic side-effect symptoms were to be increased in patients treated with low-dose antibiotics.

Conclusion

  • A continuous antibiotic low-dose treatment during the entire JJ stent-indwelling time does not reduce the quantity or severity of UTIs and has no effect on SRSs either compared with a peri-interventional antibiotic prophylaxis only.
Abbreviations
CFU

colony-forming units

HPF

high power field

LC

leucocytes

SRS

stent-related symptom

USSQ

Ureteric Stent Symptom Questionnaire

Introduction

The placement of JJ stents is a routine intervention in the management of patients with upper urinary tract obstruction [1]. Nevertheless, two complications are frequent, namely UTI and stent-related symptoms (SRS) [2, 3]. The reported rates for UTIs range from 2% to 34%. Risk factors are various, i.e. underlying systemic disease (e.g. diabetes mellitus, chronic renal failure), female gender or stent indwelling time [2, 4]. SRSs, e.g. dysuria, flank pain, haematuria and increased voiding frequency are the rule and can seriously effect daily activities, impair sexual function and even decrease work capacity [3, 5]. To prevent UTIs a peri-interventional antibiotic prophylaxis is recommended by the European Association of Urology [6]. However, there is a lack of evidence about the exact antibiotic strategy for the entire stent indwelling time [7]. In the clinical routine, it is an applied practice among urologists to continue a low-dose antibiotic treatment with the conception of reducing the incidence of UTIs and to possibly obtain a positive effect on SRS rates. To the best of our knowledge, there exists neither evidence for the benefit, nor for the potential disadvantages of such empiric prevention. We therefore analysed UTI and SRS rates in patients given a peri-interventional antibiotic prophylaxis only vs a continuous low-dose antibiotic treatment for the entire stent-indwelling time.

Patients and Methods

From March 2005 to April 2007, 126 patients underwent temporary JJ stent placement due to urolithiasis. Of the 126 patients, 95 met the inclusion criteria: sterile urinary culture before stent placement, no fever and no antibiotic medication within the past 2 weeks. In all, 31 patients had to be excluded due to positive urinary cultures, staghorn calculi and/or septicaemia requiring full-dose antibiotic treatment. All patients provided informed written consent. The trial was registered and allocated (ACTRN12612000002886).

Microbiological Investigation

Urine was obtained via single-use catheter in females and midstream urine samples in males. Samples were both, automatically (dipstick test) and manually (urinary sediment) analysed and consecutively cultured on agars. In case of bacterial growth, antibiotic susceptibility and resistance was determined. All analyses were conducted at the Institute of Microbiology, University Hospital Bern according to the guidelines of laboratory diagnosis.

Medical Treatment and Stent Insertion

All patients received peri-interventional antibiotic prophylaxis with 1.2 g amoxicillin/clavulanic acid given i.v. at anaesthesia induction to obtain a peak concentration at the time of highest risk during the procedure. According to the local pathogens profile and susceptibility, the antimicrobial agent of choice for continuous low-dose treatment was amoxicillin/clavulanic acid (625 mg) once daily. Patients with penicillin allergy similarly received either trimethoprim/sulfamethoxazole (2 ampules as prophylaxis, 80/400 mg once daily as continuous treatment) or ciprofloxacin (0.2 g i.v. as prophylaxis, 250 mg once daily as continuous treatment). All stent placements were performed under sterile conditions applying a retrograde Seldinger technique under epidural or general anaesthesia. Exclusively Percuflex Plus™ JJ stents with HydroPlus™ coating (Boston Scientific Corporation, Natick, MA, USA) were used. Patients were prescribed paracetamol for analgesia on demand (maximum dose 4 × 500 mg/day until stent removal), while none received α-blocking agents or anti-cholinergic drugs.

Baseline and follow-up Investigation

Evaluations for UTI were performed before stent insertion and consecutively after 1, 2 and 4 weeks and/or at stent withdrawal. Evaluations for SRSs were performed after 1, 2 and 4 weeks and/or at stent withdrawal. Simultaneously urine dipstick, urinary sediment and urine culture analysis were performed. Significant bacterial count/UTI was defined as ≥10 000 colony-forming units (CFU) per mL in the urine culture. Urinary cultures with more than three bacterial strains were considered as contaminated. Clinical symptoms caused by infections are very similar to SRSs, therefore the absence or presence of an infection was used to distinguish SRSs from UTIs. Fever was used to further stratify UTI into afebrile vs febrile infections. Consequently, the group characteristic for patients with SRSs only were: clinical symptoms positive, urinary culture negative and fever negative, while patients with afebrile UTI were defined: clinical symptoms positive, urinary culture positive and fever negative and patients with febrile UTI were: clinical symptoms positive, urinary culture positive and fever positive. Any patient who developed a UTI was given full-dose antibiotic therapy according to the resistance pattern until the control urine culture was sterile. SRSs including urinary symptoms (dysuria, haematuria, frequency: day- and night-time separately), pain (flank/suprapubic) and general health (including possible drug side-effects, e.g. fatigue, gastrointestinal symptoms and rash) were evaluated in personal interviews and quantified according to the validated Ureteric Stent Symptom Questionnaire (USSQ) [8]. Patients with positive UTIs were excluded from SRS evaluation.

Endpoints and Statistical Analysis

Primary endpoints were overall rates of UTIs. Secondary endpoints were SRS rates and severity of drug side-effects. Based on the assumption that roughly a maximum of 50% of patients will show a UTI, a sample size of 44 patients in each group was calculated to detect a difference of 30% by a two-sided Fisher's exact test with significance level of 5% (α = 0.05) and power of 80% (β = 0.2). Patients were randomised based on a computer randomisation program to either receive peri-interventional antibiotic prophylaxis during stent insertion only (group A, 44 patients) or to additionally receive a continuous daily low-dose antibiotic treatment for the entire stent-indwelling time (group B, 51). To further investigate the effect of stent-indwelling time on UTIs and SRSs, separate analyses for varying stent-indwelling times (<2, 2–4, >4 weeks) were performed in both groups. The thresholds reflect the most frequent indwelling periods in our clinical routine. A P < 0.05 was considered to indicate statistical significance. Urinary frequencies were compared also with 95% CIs for each day- and night-time at week 1 and 2, without using a correction for the multiple comparisons. The statistical analyses were performed with StatXact v.8 statistical software (Cytel, Cambridge, MA, USA) and SAS v9.1. The baseline comparisons are based on t-test (for age), Mann–Whitney test (for non-normal data, stent-indwelling time), and chi-square test or Fisher's exact tests (for categorical data). The exact CIs for the difference of two proportions are derived by inverting two one-sided tests (see Statxact). The Department of Mathematics and Statistics, University of Bern, Switzerland, performed the statistical analyses.

Results

Patient age, gender and median stent-indwelling time did not significantly differ between the groups (Table 1).

Table 1. Group specific patient profiles.
VariableGroup AGroup BP
No. of patients4451 
Median (range) age, years55 (19–82)56 (20–86)0.713
Male/female, n (%)29/15 (66/34)34/17 (67/33)0.997
Median (range) stent-indwelling time, days18 (7–90)28 (6–234)0.221
Patients with stent < 2 weeks, n (%)14 (32)14 (27)0.126
Patients with stent 2–4 weeks, n (%)17 (38)12 (23)
Patients with stent > 4 weeks, n (%)13 (30)25 (50)

UTI

The overall incidence of UTIs in the course was low (nine of 95 patients, 9.5%). The overall UTI rates and rates of febrile UTIs were similar between the groups. In group A, four of 44 patients (9%) were diagnosed with UTI and of these, three were febrile (7%), while in group B five of 51 patients (10%) had a UTI and three were febrile (6%). One patient in group A had an ascending pyelonephritis; the other five patients from both groups had febrile cystitis. In both groups, the UTI rates did not increase with progressive stent-indwelling time (Table 2). Irrespective of randomisation, we generally detected more UTIs in women (group A: three of 15 women [20%], one of 29 men [3%], group B: three of 17 women [18%], two of 34 men [6%]). UTIs caused by bacteria resistant to given antibiotic agents for prophylaxis or continuous low-dose treatment were seen in both groups. In group A, in one patient (2%) a resistant Escherichia coli species was detected. In group B, in three patients (6%) resistant bacteria, specifically one Enterococcus species and two multiresistant Staphylococcus aureus, were found, although the latter at a insignificant bacterial count (<10.000 CFU/mL).

Table 2. Patient specific details of UTIs.
VariableGroup AGroup BP
Patients with UTI/total number of patients, n/N (%)4/44 (9)5/51 (10)1.000
Patients with stent for <2 weeks, n/n (%)1/14 (7)0/14 (0)1.000
– Time of diagnosis (follow-up visit, week)1st
– Clinical appearance of infectionFever
– Antibiotic resistanceNo
Patients with stent for 2–4 weeks, n/N (%)2/17 (12)1/12 (8)1.000
– Time of diagnosis (follow-up visit, week)1st4th4th
– Clinical appearance of infectionFeverAfebrileAfebrile
– Antibiotic resistanceNoYes 
Patients with stent for >4 weeks, n/N (%)1/13 (8)4/25 (16%)0.643
– Time of diagnosis (follow-up visit, week)2nd2nd2nd4th6th
– Clinical appearance of infectionFeverAfebrileFeverFeverFever
– Antibiotic resistanceNoNoNoNoYes

The total of 250 urinary sediments were correlated with urinary cultures to evaluate their clinical value as a UTI predictor. Using a threshold of ≥5 leucocytes (LC) per high power field (HPF), 138/250 urinary sediments (55%) suggested a UTI. However, of these, 131 urinary cultures (94%) were negative. In contrast, two culture diagnosed UTI were quantified as <5 LC/HPF in the urinary sediment. Setting the threshold arbitrarily at ≥15 LC/HPF resulted in 82/250 positive sediments (33%). Similarly, 76 cultures (93%) showed no bacterial growth. Three UTIs were quantified as <15 LC/HPF, indicating the absence of a UTI. The analysis of nitrite reaction (dipstick) as a marker for a UTI showed 12 nitrite positive results within the 250 samples tested (5%). Of these, four (33%) developed a significant bacterial growth in the urinary culture, while eight did not. In five patients the nitrite reaction was false negative (Table 3a–c).

Table 3. Correlation of urinary sediment at different thresholds (5 and 15 LC/HPF) and nitrite reaction (dipstick test) with urinary culture results.
 Urinary culture negativeUrinary culture positiveTest values
  1. fn, false negative; tn, true negative; fp, false positive; tp, true positive; PPV, positive predictive value; NPV, negative predictive value.
Urinary sediment threshold of 5 LC/HPF
<5 LC/HPF110 (tn)2 (fn)

Specificity: 0.46

Sensitivity: 0.78

PPV: 0.05

NPV: 0.98

≥5 LC/HPF131 (fp)7 (tp)
Urinary sediment threshold of 15 LC/HPF
<15 LC/HPF165 (tn)3 (fn)

Specificity: 0.68

Sensitivity: 0.67

PPV: 0.07

NPV: 0.98

≥15 LC/HPF76 (fp)6 (tp)
Nitrite reaction
Negative233 (tn)5 (fn)

Specificity: 0.97

Sensitivity: 0.44

PPV: 0.33

NPV: 0.98

Positive8 (fp)4 (tp)

SRSs

Overall incidence of SRSs was high (92/95 patients, 97%). SRS rates did not differ between the groups [group A: 43 patients (98%) vs group B: 49 patients (96%)]. Only one patient (2%) in group A and two patients (4%) in group B neither complained about SRSs nor had a UTI diagnosed. Irrespective of randomisation the vast majority had at least one of the evaluated SRSs (Fig. 1). Increased day- and night-time voiding frequencies after stent insertion were seen in both groups and remained unchanged after 2 weeks (Fig. 2). Overall, low-dose antibiotic treatment did not reduce SRSs and patient stratification according to stent-indwelling time showed neither adaptation nor progression to/of SRSs (data not shown).

Figure 1.

Classification and incidence of SRSs (flank/suprapubic pain, gross haematuria and dysuria).

Figure 2.

Occurrence of increased voiding frequencies (daytime, nighttime).

Drug Side-effects

Comparable overall rates of antibiotic side-effects were seen in both groups [group A: 21 patients (48%) vs group B: 22 patients (43%)]. Most patients had a combination of rash, pruritus, nausea and/or diarrhoea. There was no difference in the incidence of fatigue (group A: 81% vs group B: 77%). However, the presence of skin and gastrointestinal symptoms was significantly increased in patients under continuous low-dose antibiotic treatment (group A: 0% and 14% vs group B: 14% and 59%; Table 4). There was no change with increasing stent-indwelling time.

Table 4. Classification and incidence of drug side-effects.
Side-effectsGroup AGroup B95% CI for group B – group A
Patients with side effects, n/N (%)21/44 (48)22/51 (43)(–0.252, 0.157)
Side-effect, n (%):   
Rash and/or pruritus03 (14)(–0.036, 0.349)
Nausea and/or diarrhoea7 (33)13 (59)(–0.051, 0.541)
Fatigue17 (81)17 (77)(–0.292, 0.224)

Discussion

For the rates of UTIs, the concept of a peri-interventional antibiotic prophylaxis during endoscopic urological procedures is known to be better than no antibiotic coverage [6, 9]. Particularly focusing on patients with indwelling JJ stents, it is a common practice to continue the antibiotic treatment in a low-dose fashion, even after previous uncomplicated implantations. The intention is to lower the rates of UTIs and to achieve a positive effect on SRSs. This practice might originate from the lack of evidence for the exact antibiotic strategy and is supported by controversial recommendations from sparse publications [9, 10]. However, increasing rates of bacterial drug resistances, growing overall healthcare costs and drug side-effects require a critical antibiotic prescription policy [11, 12]. Therefore, our aim was to evaluate the benefits and disadvantages of a peri-interventional antibiotic prophylaxis only compared with a continuous low-dose antibiotic treatment. All 95 patients received antibiotic prophylaxis during stent placement. Compared with reports using no antibiotic prophylaxis, in the present study there were lower rates for both, overall (9.5%) and febrile UTIs (6%) [7]. Interestingly, we did not detect a higher rate of UTIs in patients receiving peri-interventional antibiotic prophylaxis only compared with patients given continuous low-dose antibiotic treatment during the entire stent-indwelling time. Generally, febrile UTIs in the follow-up investigations were mainly febrile cystitis, irrespective of the antibiotic treatment strategy. This may partly be based on our trial design with frequent follow-up visits and consecutively early detection and treatment. In contrast to recent reports, in the present study the UTI rates did not increase with longer stent-indwelling times [2]. Although more UTIs were discovered in the group of patients under continuous low-dose antibiotic treatment and longer stent-indwelling times, the UTIs were diagnosed during early follow-up visits. Consistent with other reports, we generally found women to be at higher risk of developing a UTI [13]. Importantly, this finding was irrespective of the antibiotic strategy. However, the conclusion that peri-interventional prophylaxis only is sufficient for all patients is limited to the stone situation. In the present study, all patients had preoperative sterile urinary cultures. These results may not be applied to other situations, e.g. pregnancy, malignant obstruction or required endocarditis prophylaxis, where further studies are required. Uncritical widespread use of antibiotics is one of the reasons for increasing bacterial drug resistance [12, 14]. In concordance, rates of resistant bacterial strains tended to be increased in the continuous low-dose antibiotic treatment group. However, the low incidence of UTIs within the present cohort does only allow a descriptive analysis.

Another noteworthy and helpful observation for daily clinical practice was the poor value of urinary sediment and urinary dipstick results as predictive markers for significant bacterial growth in the consecutive urinary culture. Instead of effectively supporting the clinical decision-making, this lack of accuracy rather leads to unnecessary antibiotic treatments.

The vast majority of patients with JJ stents are more or less bothered by SRSs. Many different pharmaceutical and technical approaches have been investigated to reduce the symptoms [15-17]. In the present cohort, we exclusively implanted flexible HydroPlus coated stents with the shortest possible intravesical length. Despite this, the continuous low-dose antibiotic treatment had neither an effect on the rate of SRSs nor on the spectrum of symptoms. Overall, SRS rates remain high during the entire stent indwelling time and therefore the restrictive indication for stent implantation is of utmost importance. Skin and gastrointestinal symptoms were significantly increased in patients under continuous low-dose antibiotic treatment compared with patients receiving peri-interventional antibiotic prophylaxis only. This indicates that even a low-dose antibiotic treatment potentially implies individual disadvantages.

In conclusion, the present prospective randomised study showed that continuous low-dose antibiotic treatment during the entire JJ stent-indwelling time did not reduce the quantity or severity of UTIs and had no effect on SRSs compared with the peri-interventional antibiotic prophylaxis only, but involves undesirable disadvantages, e.g. increased drug side-effects and higher rates of resistant bacterial strains, and should therefore be avoided. The high SRS rates require a strict indication for stent implantation and the similarity of the symptoms to UTI symptoms makes the differentiation of the underlying cause very difficult. Concerning the use of urinary sediment and dipstick analysis as clinical screening tools, we recommend in the case of positive tests to wait for the results from the urinary culture, if the patient's condition allows for it. Only culture confirmed infections require antibiotic treatment to avoid unnecessary antibiotic prescription and antibiotic treatment should be according to resistance testing.

Conflict of Interest

None declared.

Ancillary