Introduction

Neurogenic bladder is one of the risk factors for urinary tract infection (UTI) because these patients often have complications such as disturbances in urination, catheterization, urinary tract obstruction, and residual urine lead to the emergence of bacteriuria or pyuria. During UTI management, antibiotic use needs to be monitored since uncontrolled use may lead to the emergence of antibiotic-resistant strains and also involves unnecessarily high institutional costs [1]. Antimicrobial stewardship (ASW) programs have become more common as a way to address these problems [2]. ASW programs have emphasized the prevention of antibiotic-resistant strains from unnecessary broad-spectrum antibiotic use, long duration use, insufficient doses, or inappropriate spectrum coverage [3]. The related researchers have spread this concept and increased the number of specialists managing these issues [4, 5], including the UTI field. In particular, outpatients with NB require special attention because most patients with NB have urination disturbances and a high risk of catheter-associated UTI (CAUTI) leading to the increased likelihood of antibiotic-resistant bacterial infection owing to frequent antibiotic exposure for chronic pyuria or bacteriuria. Antibiotic treatments are contraindicated for patients without UTI in many clinical settings [6]. Outpatients with NB specifically need to be studied for antibiotic consumption, antibiotic stewardship intervention, and changes in the emergence of antibiotic-resistant bacteria.

In this study, we retrospectively examined outpatients with NB at risk for CAUTI after intervention for appropriate use of antibiotics, and analyzed the amounts of antibiotics prescribed and the incidence rate of antibiotic-resistant bacteria over a 6-year period in a rehabilitation hospital to investigate the effect of our ASW program [7] in decreasing the use of oral third generation cephalosporins and quinolones, and to reduce the occurrence of antibiotic-resistant bacteria.

Methods

Study setting

We retrospectively investigated the amount of oral antibiotics used for outpatients in Hyogo Prefectural Rehabilitation Central Hospital, Hyogo, Japan from 2012 to 2017. This is a rehabilitation hospital with 7 inpatient wards (two rehabilitation wards, one internal medicine and neurology ward, two orthopedic wards, one spinal cord injury ward, and one pediatric rehabilitation ward) with 330 beds, 70 days of average hospitalization, an average 630 cases of yearly surgery, and 30 full-time doctors (orthopedics, urology, internal medicine, rehabilitation, neurology, rheumatology, pediatric orthopedics, and pediatric sleep disorders) [8]. This hospital is one of the hospitals that had been referred in our previous study [8]. We had found that there had been significant correlations between antibiotic resistance of UTI-causing bacteria such as Escherichia coli and Klebsiella pneumoniae and both annual number of days of antibiotic use and days of therapy in the hospital. There were 28,973 drug prescriptions for outpatients and 31,874 prescriptions for inpatients in 2017, all of which were prepared in the hospital. Two hundred five prescriptions were provided per working day. Our infection control team (ICT) consists of a pharmacologist (team leader), physician, medical technologist, and nurses based on Japanese Ministry of Health, Labor, and Welfare standards for intervention initiated in 2016. The ICT suggests that physicians limit or stop the use of third generation cephalosporins and fluoroquinolones. Penicillin or first generation cephalosporins or sulfamethoxazole–trimethoprim (ST) is often recommended for use as de-escalation. We began intervening in Urology Department use of third generation cephalosporins and fluoroquinolones at the initiation of the study. Most of the patients with NB and UTI in this study were caused by spinal cord injury. The number of patients with indwelling catheter or intermittent catheterization was 100 cases and 350 cases, respectively. In total, 103 patients were routinely changed with urethral catheter or cystostomy catheter. The frequency of outpatient clinical attendance and the indication for attendance was once per 1–3 months.

Antibiotic-resistant bacteria

In this study, we particularly focused on extended-spectrum β-lactamase (ESBL) producing-E. coli, ESBL-producing-K. pneumoniae, and multi-drug resistant Pseudomonas aeruginosa, since these are representative UTI-causative bacteria with a high likelihood of antibiotic resistance and are likely to well represent the whole spectrum of antibiotic resistance. ESBL-producing and multidrug resistant P. aeruginosa were primarily screened by the cefotaxime and ceftazidime disk diffusion method. Isolates, which showed an inhibition zone of 27 mm for cefotaxime and 22 mm for ceftazidime were investigated for ESBL enzyme production. The suspected isolates were further assessed by the double disc method [9, 10]. A multidrug-resistant P. aeruginosa phenotype is defined as resistance to three or more antipseudomonal antimicrobial classes: carbapenems, fluoroquinolones, penicillins/cephalosporins, and aminoglycosides [10].

Particularly, we focused on the quantity of cefdinir (CFDN) and levofloxacin (LVFX), two antibiotics that were especially heavily used in the Urology Department and whose use has been remarkably reduced over the 6-year study period. We also investigated the incidence rate of antibiotic-resistant bacteria in urine cultures from outpatients over the 6-year study period. This study was approved by the ethics board of Hyogo Prefectural Rehabilitation Central Hospital.

Statistical analysis

Data were analyzed by Statcel 3 (OMS publishing Inc., Tokyo, Japan) to compare the changes in parameters between the first three years and the last three years. Welch’s t test was used for comparison because of non-normal distribution. The Spearman’s correlation coefficient by the rank test was conducted to determine the correlation between years and the following parameters for CFDN and LVFX: number of prescriptions, dosing periods, dosages, total number of antibiotics used, number of prescriptions in the Urology Department, and dosages in the Urology Department. The proportion of resistant bacteria was analyzed using the χ2 test to compare 2012 and 2017. Statistical differences among means were considered significant when p < 0.05.

Results

Antibiotic use

We investigated the number of prescriptions, dosing periods, and amount prescribed for outpatients with NB in the Urology Department. The averages and medians are shown in Table 1. The total number of prescriptions of CFDN significantly decreased from 463 cases to 130 cases over 6 years (p = 0.012). The average dosing periods and dosages in each year slightly decreased from 7.7 days to 6.5 days after 2012, but the median of those factors did not change substantially (Table 1). Mean dosages of CFDN significantly decreased from 2205 mg in 2012 to 1930 mg in 2017 (p = 0.047). We also found that the proportion of prescription and total dosages significantly decreased from 2012 to 2017 (p = 0.047 and p = 0.025, respectively).

Table 1 Change in prescriptions of cefdinir in the urological outpatient setting.
Full size table

The average and median of LVFX prescriptions, dosing periods, and amount prescribed for outpatients in the Urology Department and for total hospital outpatients are shown in Table 2. The number of prescriptions of LVFX significantly decreased from 640 cases in 2012 to 171 cases in 2017 in the Urology Department (p = 0.025). The total amount of LVFX used for outpatients in 2017 significantly decreased from 2334 g in 2012 to 610 g in 2017 (p = 0.035). We also found that the proportion of prescription and total dosages significantly decreased from 2012 to 2017 (both p = 0.048) (Table 2).

Table 2 Changes in levofloxacin prescriptions in the urological outpatient setting.
Full size table

Antibiotic-resistant bacteria

We investigated the incidence rate of antibiotic-resistant E. coli, K. pneumoniae, and P. aeruginosa from urine culture. ESBL-producing E. coli and K. pneumoniae were the most commonly detected antibiotic-resistant Enterobacteriaceae. Antibiotic-resistant P. aeruginosa showed resistance to three or more antibiotics. As seen in Fig. 1, the proportion of ESBL-producing E. coli to total E. coli slightly, but not significantly, decreased from 26% in 2012 to 24% in 2017 (p = 0.744). The incidence of ESBL-producing K. pneumoniae showed a remarkable decrease from 25% in 2012 to 7% in 2017 (p < 0.001). The incidence of P. aeruginosa decreased from 21% in 2012 to 14% in 2017 (p = 0.193) (Fig. 1). LVFX-resistant E. coli and K. pneumoniae were also seen and the changes in these over the 6-year study period are shown in Fig. 2. The incidence of LVFX-resistant E. coli significantly decreased from 46% in 2012 to 31% in 2017 (p = 0.03). The incidence of LVFX-resistant K. pneumoniae decreased from 12% in 2012 to 3% in 2017 (p = 0.016) (Fig. 2).

Fig. 1: Antibiotic resistance rate of E. coli, K. pneumoniae, and P. aeruginosa.
figure 1

The proportion of extended-spectrum β-lactamase-producing E. coli and Klebsiella versus total E. coli and K. pneumoniae is shown. Antibiotic-resistant P. aeruginosa showed resistance to three or more antibiotics.

Full size image
Fig. 2: Quinolone resistance rate of E. coli and K. pneumoniae.
figure 2

The proportion of quinolone-resistant E. coli and K. pneumoniae is shown.

Full size image

Discussion

ASW programs have been encouraged globally, and the number of related studies has increased over the last decade [11]. In this study, we found a significant reduction in the prescription of third generation cephalosporins and quinolones between 2012 and 2017 due to the implementation of an ASW program, associated with a decrease of ESBL-producing K. pneumoniae, antibiotic-resistant P. aeruginosa, LVFX-resistant E. coli, and LVFX-resistant K. pneumoniae in outpatients with NB, where antibiotic exposure was commonly seen in the presence of pyuria or bacteriuria and high risk for UTI [12].

It is important to reduce antibiotic use and thus prevent the emergence of antibiotic-resistant bacteria even in this high-risk population. Our study showed reduced use of antibiotics, especially broad-spectrum third generation cephalosporins, CFDN and fluoroquinolones like LVFX for urological outpatients with NB over the 6 years of the study. We found that the incidence of antibiotic-resistant bacteria in urine cultures did not increase. The decrease in CFDN and LVFX prescriptions seems to have been one of the key factors, since the dosing periods and median amount of CFDN and LVFX for each patient did not show a dramatic change or decrease from 2012 to 2017.

De-escalation of antibiotics and switching to narrower-spectrum coverage is one modality by which ASW programs can help us to avoid unnecessary use of broad-spectrum antibiotics [13]. Carbo et al. showed that ASW programs including prescription review and feedback, restricted formularies, guideline development, and education significantly shortened the duration of antibiotic therapy and improved clinical outcomes of inpatients and outpatients [14]. Hicks et al. found that comprehensive oversight of outpatient antibiotic prescriptions was an essential strategy for reducing the spread of antibiotic resistance and a critical step in identifying where appropriate-use interventions can have the most impact [15]. In the present study, we found that dosing periods for CFDN and LVFX were slightly shortened and total consumption was reduced to achieve de-escalation over the study period. The prescribed amounts of CFDN and LVFX decreased after 2012, while the amount of ST and cefaclor tended to increase over the study period (data not shown), indicated that ASW stewardship led to less use of broad-spectrum and more focused use of narrower-spectrum antibiotics, decreasing the chance of emergence of antibiotic-resistant strains. The use of narrow-spectrum antibiotics such as ST or first generation cephalosporin was reported to yield low antibiotic resistance rates in outpatients with UTI [16], supporting our results.

The total amount of LVFX used for outpatients gradually decreased until 2016, but slightly increased from 2016 to 2017. In contrast, LVFX use in the Urology Department continuously decreased from 2012 to 2017. If this pattern of use spreads to other departments, such as internal medicine which also regularly prescribes antibiotics, it might lead to a decreasing chance of emergent antibiotic-resistant strains in the whole hospital.

Several studies have correlated the establishment of ASW programs with lower antimicrobial therapy costs. Fukuda et al. prospectively monitored ASW management of meropenem, ciprofloxacin, and amikacin for 251 cases in a hospital with 465 beds over 2 years [17]. They showed that antimicrobial therapy costs significantly decreased by 26% after an ASW program was instituted. They also found that consumption of aminoglycosides decreased by 80% and the detection rate of methicillin-resistant staphylococcus aureus decreased by 48%. In this study, ASW reduced almost 80% of the CFDN cost and 90% of the LVFX cost in the Urology Department over the 6-year observation period. The Urology Department showed the greatest cost reduction in this study (data not shown).

Regarding antibiotic-resistant bacteria, the total incidence of antibiotic-resistant bacteria tended to decrease from 2012 to 2016, but slightly increased in 2017. In particular, we had an outbreak caused by ESBL-producing K. pneumoniae in an inpatient ward [18]. The increased incidence of antibiotic-resistant K. pneumoniae in outpatients was possibly affected by this outbreak. In addition, urine culture was done for discharged patients carrying antibiotic-resistant bacteria as outpatients after the outbreak was controlled. The main purpose of ASW programs is to inhibit or prevent the emergence of antibiotic-resistant bacterial infections. Many other studies have shown that ASW intervention reduced antibiotic use and improved patient outcomes, resulting in lowered incidence of antibiotic resistance [19]. Our results did not show a significant change (decrease) of antibiotic-resistant bacteria and further follow-up needs to be performed for definitive evaluation.

ASW programs have spread to many health care institutions [11]. We suggest that ASW should be done in cooperation with an ICT and linked hospital staff to carry out the following: (1) Early monitoring and intervention in treatments for infectious diseases after the initiation of antibiotic treatment and/or detection of positive blood culture. (2) Approval for intended antibiotic use in cooperation with the Pharmaceutical Department. (3) Management of appropriate antibiotic use to monitor dosing and duration along with therapeutic effects after initiation of antibiotic therapy. When necessary, the ASW can send feedback to the doctors in charge regarding the appropriate use of antibiotics. (4) The ASW can order microbiological tests required for early diagnosis, and educate the staffs to collect appropriate specimens. (5) The ASW can monitor the incidence of antibiotic-resistant bacteria in the whole hospital. (6) The ASW can educate medical staff to promote ASW activities using antibiograms and local data. (7) The ASW should have a meeting round once a week if possible.

We would like to emphasize the limitations of this study. First, this study is retrospective and was carried out in one institution. We have missed a longitudinal cohort study consisting of a group exposed to the risk factor (in this case treatment with the extended-spectrum antibiotics) and another control not exposed to this risk factor. Other variables influencing on bacterial resistance such as a previous history of UTI, previous exposure to antibiotics, gender, ASIA impairment scale, history of urological manipulation, presence of urinary catheters, patient histories, and inpatient status was not investigated. Second, patient history and outcome data by treatment were not assessed. Third, the cost issue was not assessed in this study because some antibiotics were changed to generic brands while other new antibiotics were introduced, complicating the possibility of relevant calculation. Fourth, we did not verify the relationship between the use of extended-spectrum antibiotic and bacterial resistance by mechanistic studies. Further prospective multicenter studies need to be performed to draw definitive conclusions. Our study was not designed to address these limitations, but to conduct a retrospective study to assess the effectiveness of our ASW program in terms of NB patients and antibiotic-resistant bacterial strains.

In conclusion, we showed that the use of CFDN and LVFX decreased for outpatients with NB in our Urology Department from 2012 to 2017. At the same time, the incidence of antibiotic-resistant bacteria from urine cultures did not significantly increase. Even though no direct correlation was shown, the results are consistent with the fact that the decrease in the use of CFDN and LVFX in urological outpatients contributed to the decrease in the incidence of antibiotic-resistant and ESBL-producing bacteria. Although a prospective controlled study is necessary to confirm this hypothesis.