Abstract
Objective
We wanted to look at the antibiotic resistance patterns in our university teaching hospital for E. coli cultured from urine to assess the resistance profiles to common antibiotics used in the hospital.
Patients and methods
Results were obtained of all positive urine cultures between January 2007 and December 2011 both from the hospital and urology department. Trends in susceptibility and resistance data for E. coli to trimethoprim, amoxicillin, ciprofloxacin, gentamicin and nitrofurantoin were examined.
Results
A total of 40,722 hospital urine specimens were culture positive in the five-year period, of which 15,311 (37.6%) grew E. coli. Across the hospital there was no evidence of change in resistance to trimethoprim over the five-year period (38.1% in 2007 compared to 36.6% in 2011; p = 0.313). The percentage resistance to amoxicillin rose slightly (52.6% to 54.4%), and this was statistically significant (p = 0.011). Overall there was a statistically significant fall in resistance to ciprofloxacin but resistance has remained stable for the last three years (15.5% to 13.5%, p = 0.013). A trend of increasing resistance to gentamicin (p = 0.002) resulted from the 2007 baseline of 3.4% to 4.9% in 2008, but resistance remained stably low thereafter. Resistance to nitrofurantoin fell significantly (from 10.4% to 1.6%; p ≤ 0.0005).
Conclusions
The number of E. coli-positive urine cultures from hospital in-patients remained broadly stable over this five-year period. Resistance to trimethoprim and amoxicillin remains high. Gentamicin resistance remains low, but the statistically significant rise over the five-year period suggests resistance rates should continue to be carefully monitored. Nitrofurantoin resistance is very low and has significantly decreased. This narrow-spectrum antibiotic should be considered the mainstay for treating uncomplicated urinary tract infections in females.
Introduction
Empiric antimicrobial guidelines aim to target the most important causative organisms of an infection and are based on local susceptibility patterns which can change over time. The purpose of these guidelines is to optimise outcomes whilst minimising excess broad-spectrum antimicrobial use, which can drive resistance selection pressures and be associated with secondary problems such as Clostridium difficile infection. It is important to continually monitor changes in antibiotic susceptibility in order to ensure that empiric treatment guidelines remain appropriate. Escherichia coli (E. coli) is the commonest cause of urinary tract infection (UTI). 1 The aim of our study was to examine the changes in trend of antibiotic resistance for urinary E. coli isolates over a five-year period in our teaching hospital and urology department.
Patient and methods
The Department of Infection of University Hospital Southampton NHS Foundation Trust (UHSFT) collected results of organisms grown from urine cultures (data collated and provided by Public Health England). The data selected for use were all E. coli-positive urine cultures from the whole hospital over a five-year period (January 2007 to December 2011). Results of urine cultures from patients throughout the hospital and within the urology department were analysed. Information on E. coli susceptibility to five antibiotics (trimethoprim, amoxicillin, ciprofloxacin, gentamicin and nitrofurantoin) was collected.
The patterns of susceptibility and resistance were analysed with Statistical Package for Social Sciences (SPSS) using the Cochran-Armitage test for trend carried out at the 5% level. The degree of trend in resistance and its significance was calculated for each antibiotic. Analysis was performed on resistance of cultures from all patients in the hospital and repeated for cultures from urology department patients.
Results
Across the hospital 40,722 urine specimens were culture positive in the five-year period 2007–2011. The organism most commonly isolated was E. coli (15,311; 37.6%). Of these E. coli isolates, 161 (1.05%) were from the urology department. Other common organisms identified in cultures from urine samples across the hospital were non-E. coli coliforms (14.39%) and enterococci (13.12%) (Table 1).
Isolates from all UHSFT urines 2007–2011.
UHSFT: University Hospital Southampton NHS Foundation Trust.
Looking at E. coli alone, results both from the hospital and urology demonstrated that the antibiotic with the highest frequency of resistance was amoxicillin (51.5%). The antibiotics with lowest frequency of resistance were gentamicin (4.8%) and nitrofurantoin (4.4%).
Over the five years, E. coli resistance to nitrofurantoin fell in the hospital and in the urology department. Within the hospital there was a statistically significant trend of decreasing resistance to nitrofurantoin (from 10.4% to 1.6%; p ≤ 0.0005) (Figure 1). Resistance to nitrofurantoin was the least frequent of all selected antibiotics analysed in 2011. A similar trend was reflected in E. coli from urines taken from urology patients (Figure 2); however, small sample size did not permit further valid statistical analysis.
Antibiotic resistance in Escherichia coli urine isolates across the hospital.
Antibiotic resistance in Escherichia coli urine isolates from the urology department. Data shown for information only – statistical interpretation not possible because of low sample numbers.
There was increasing resistance to gentamicin over the period (from 3.4% to 5.1%; p = 0.002) in the hospital, but after an initial rise occurred in 2008, the resistance rate remained stable. Resistance to gentamicin within urology patients fell (23.0% in 2007 to 9.0% in 2011) (Figure 2).
Resistance to trimethoprim and ciprofloxacin in the hospital was slightly lower in 2011 (36.6% and 13.5%) compared to 2007 (38.1% and 15.5%, respectively); tests for trend over the period showed a lack of statistical significance for trimethoprim. Ciprofloxacin resistance rates remained stable for the last three years (Table 2). The percentage resistance to amoxicillin was high and increased slightly (52.6% in 2007 to 54.4% in 2011; p = 0.011). Similar patterns were seen in urology patients (Figure 2).
Number (%) resistance of Escherichia coli in urine from UHSFT to selected antibiotics 2007–2011.
UHSFT: University Hospital Southampton NHS Foundation Trust.
Discussion
There are serious concerns that antibiotic efficacy will significantly diminish in future as increasing bacterial resistance is identified and fewer new antibiotics are developed. These risks have been highlighted by groups such as the British Society of Antimicrobial Chemotherapy Antibiotic Action forum. Prudent use of antimicrobials is important to minimise resistance selection pressures, side effects to the patient and risks of secondary infections such as Clostridium difficile. Empiric antimicrobial therapy should be selected on the basis of organism susceptibility patterns and avoid the inappropriate use of broad-spectrum antibiotics.
Similar to other studies, our study shows that E. coli is still the most common cause of UTIs.2,3 This means that attention to E. coli antibiotic susceptibility and resistance trends is particularly important in the preparation of empiric antibiotic guidelines for treatment of UTIs.
Our results demonstrated low resistance to nitrofurantoin. This has also been noted in a study in California which found 5% resistance of E. coli to nitrofurantoin in female patients. 4 Similarly, the study also demonstrated higher rates of trimethoprim resistance (18%). With regard to gentamicin resistance, our results are comparable to Ababneh et al.'s study demonstrating an increase in rate and proportion of 166% and 124% of gentamicin-resistant E. coli in the United States over seven years. 5
It is interesting to see that the trend in resistance for ciprofloxacin is plateauing and resistance remains under 14%. Ciprofloxacin continues to have a place in the treatment of UTIs in men, prophylaxis for transrectal prostate biopsies and upper renal tract infections in UHSFT because of its excellent prostate and tissue penetration. In contrast, a study in the Netherlands of 17,486 E. coli isolates from 2008 to 2010 showed a trend of increasing resistance to ciprofloxacin from 12.4% to 15.5%. 6
A strength of our work is the large number of isolates analysed (15,289). We have assumed that the majority of urine cultures will be taken from hospitalised patients prior to commencement of empiric antibiotic therapy. Trimethoprim and amoxicillin are not appropriate empiric agents for treatment of UTI. It was not possible to validate historical data for co-amoxiclav hence its non-inclusion, but of note it is a relatively broad-spectrum antibiotic with significant gastrointestinal side effects. The identification and use of an alternative narrow-spectrum antibiotics would be preferable. Nitrofurantoin is a narrow-spectrum antibiotic and resistance rates are reassuringly low, considering its local first-line use in treatment of women with uncomplicated UTIs. It is both cost effective and efficacious for empirical treatment of uncomplicated UTIs. 7 The limitations of using nitrofurantoin are its poor penetration to the prostate and that it is ineffective in the treatment of systemic sepsis or upper renal tract infections. It is contraindicated in patients with poor renal function and infants less than 3 months old and should be used with caution in patients with hepatic impairment and close to time of delivery for pregnant women. 8
The rise in gentamicin resistance over the study period is not currently clinically important as overall gentamicin resistance remains low. Gentamicin still has a place in our guidelines for the treatment of severe sepsis (including urosepsis) when it is frequently used as a single dose prior to definitive culture results being available. Therefore, continued monitoring of resistance trends is called for given the importance placed on this agent in our guidelines.
Similar patterns in resistance were found in urine cultures from our urology department though numbers were small.
It is reassuring that a narrow-spectrum agent remains highly effective in vitro even in this population. Promotion of nitrofurantoin use and avoidance of broad-spectrum antibiotics such as co-amoxiclav is particularly valuable in a hospital setting where there is a significant burden of antimicrobial use and potential for adverse consequences of broad-spectrum therapy.
Conclusions
Nitrofurantoin resistance is low and has significantly decreased and therefore should now be considered the mainstay of empiric treatment of uncomplicated UTIs in females. This can be switched to targeted therapy on the basis of definitive subsequent culture results if required. Patient-specific factors such as renal function, presence of systemic sepsis, late pregnancy, age and the site of infection should also be considered. Gentamicin resistance remains low but susceptibility trends should be monitored considering the reliance placed on it as part of the empiric treatment of severe sepsis (including urosepsis) in our hospital.
Conflict of interest
None declared.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Footnotes
Acknowledgements
The authors thank Public Health England for collating and providing the data.