The Benefits and Harms of Antibiotic Prophylaxis for Urinary Tract Infection in Older Adults (2024)

Article history

Abstract

Urinary tract infections (UTIs) are the most common bacterial infections prompting healthcare use [1] and result in 400 000 hospitalizations per year in the United States alone [2]. Urinary tract infections pose an immediate burden to patients, including physical and emotional suffering [3], and a threat of bloodstream infection and sepsis [4]. For many patients the discomfort and danger of UTIs become a recurring problem. About 1 in 4 younger patients will have a repeat UTI within 6 months of an initial episode [1], and this recurrence risk increases greatly among those with multiple previous episodes [5]. The incidence of recurrent UTI in older adults is less clear given the potential for overdiagnosis due to nonspecific symptoms [6].

Among patients with recurrent UTIs, antibiotic prophylaxis is a potentially effective means to reduce the risk of subsequent infections. A Cochrane review evaluating healthy nonpregnant women in 19 randomized controlled trials (RCTs) found that antimicrobial prophylaxis was associated with a reduction in microbiologic recurrence (relative risk [RR], .21; 95% confidence interval [CI], .13–.34; number needed to treat [NNT], 1.8), as well as a reduction in clinical recurrence (RR, .15; 95% CI, .08–.28; NNT, 1.8) [7]. The benefit of antibiotic prophylaxis is less well studied among older patients (>65 years old). A systematic review of only 3 RCTs of antibiotic prophylaxis among older women in the community identified a more modest benefit of treatment (RR, .76; 95% CI, .61–.95; NNT, 8.5) [6]. Although there is evidence that UTI prophylaxis may help to prevent UTIs, evaluation of more severe outcomes such as emergency department (ED) visit or hospitalization for urinary indications is lacking. Further, no RCTs of UTI prophylaxis have been conducted among older men or among residents of long-term-care facilities [6].

There are also potential harms of antibiotic prophylaxis, which must be weighed into the prescribing decision. The Cochrane review detected an increased risk of adverse events (RR, 1.78; 95% CI, 1.06–3.00), which consisted predominantly of mucocutaneous candidiasis and gastrointestinal symptoms [7]. A more significant patient and societal harm of antibiotic use is antibiotic resistance, which has been shown to develop more commonly among patients receiving UTI prophylaxis [8, 9]. For example, a recent RCT of UTI prophylaxis in adults using clean intermittent self-catheterization, published after the Cochrane review, confirmed a benefit of prophylaxis in preventing UTIs (incidence rate ratio, .52; 95% CI, .44–.61), but detected an increase in nitrofurantoin resistance and trimethoprim-sulfamethoxazole resistance among subsequent urinary tract isolates [9]. The harms of antibiotics, such as end-organ toxicity and antimicrobial resistance, may be more pronounced in frail elderly patients [10, 11]. Harms data tend to be underreported in RCTs [12], so a large population-based cohort study may help to better understand the balance between benefits and risks of antibiotic prophylaxis for UTIs. We sought to quantify the benefits (healthcare utilization for UTI sequelae) and risks (antibiotic resistance and adverse events) of long-term antibiotic prophylaxis among older adults with UTIs.

METHODS

Study Design

We conducted a retrospective cohort study comparing older adults (≥66 years) receiving antibiotic prophylaxis (duration ≥30 days) following a positive culture with matched patients with a positive urine culture who received antibiotics but did not receive antibiotic prophylaxis.

Setting

This study took place in Ontario, Canada’s most populous province (14.7 million in 2019). Healthcare is provided in Ontario via a universal single-payer health insurance model.

Data Sources

We obtained study data entirely from linked population-wide administrative datasets housed at ICES (formerly known as the Institute for Clinical Evaluative Sciences). ICES is an independent, nonprofit research institute whose legal status under Ontario’s health information privacy law allows it to collect and analyze healthcare and demographic data, without consent, for health system evaluation and improvement.

Urine culture and susceptibility data were captured from the Ontario Laboratory Information System (OLIS), which is a repository for hospital and outpatient laboratory results in the province of Ontario. We collected antibiotic exposure data from the Ontario Drug Benefit (ODB) database, which includes universal outpatient medication dispensing data for adults over 65 years. Previous validation of ODB has shown 99.3% accuracy compared with chart abstraction of the actual prescription [13]. Demographics and patient outcomes data were collected from the Registered Persons DataBase, the Ontario Health Insurance Plan (OHIP) Database, and Canadian Institute for Health Information (CIHI) National Ambulatory Care Recording System, CIHI Discharge Abstract Database (DAD), and Continuing Care Reporting System—Long Term Care (CCRS-LTC). These datasets are linked using unique encoded identifiers and analyzed at ICES and have been widely used for studies of antimicrobial harms and benefits [14–16].

Patient Selection Criteria

All adults living in Ontario aged 66 years (aged 65 years plus 1 year to ensure an adequate look-back) and older were eligible for inclusion. Patients in an outpatient or long-term-care setting with a first positive urine culture collected during the period 1 January 2015 and 31 December 2016 were eligible for inclusion in the cohort.

In order to capture incident use of antibiotic prophylaxis, patients with prior prescriptions for 30 or more days of consecutive antibiotics in the 90 days prior to positive urine culture were excluded. Patients who died or were not eligible for OHIP within 30 days of the positive urine culture were excluded to ensure outcomes would be available and detectable for prophylaxis recipients and nonrecipients at the beginning of the follow-up period.

Exposure

We considered the patient to have been placed on UTI prophylaxis if he/she received any antibiotic, except for doxycycline (given that this agent was not covered by drug plan insurance at the time, nor would it be indicated for UTI prophylaxis), for a prescribed duration of 30 or more days that was started within the 30-day period after the positive culture (to increase the likelihood that antibiotics were associated with a urinary indication and not for another infectious diagnosis). The day of the positive culture was assigned as the index date. We allowed for shorter durations of antibiotics to count as prophylaxis if consecutive prescriptions of the same agent sum to 30 or more days, because longer-duration prescriptions are sometimes divided into shorter intervals (eg, 7 days) for dispensing purposes. Although UTI prophylaxis is typically prescribed for longer than 30 days we opted to use a 30-day threshold to be inclusive in capturing prolonged antibiotic exposure most likely for a prophylactic indication. In general, UTI treatment courses are for 14 days or fewer (even for severe infections with bacteremia) [17] and typically 28 days or fewer for bacterial prostatitis [18], so durations exceeding 30 days ought to represent prophylaxis. Our administrative data do not provide information on patients’ symptoms for either the prophylaxis or control patients, so the presence of a positive urine culture along with a corresponding antibiotic was considered prophylaxis for study purposes.

Controls

Control patients were defined as patients with a positive urine culture who received antibiotics within 30 days after the positive culture but did not receive prophylaxis.

Covariates

Using a minimum 1-year look-back period, we collected patient characteristics that may be associated with the outcome, including patient age, sex, healthcare setting in which urine culture was collected (community vs long-term care), Deyo-Charlson comorbidity score (2-year look-back) [19], diabetes mellitus, kidney stones, genitourinary tract cancer, urinary tract surgery, number of ED visits in prior year, number of hospitalizations in past year, species of organism cultured in index urine culture, and antibiotic resistance of baseline isolate to the following agents: amoxicillin/ampicillin, amoxicillin-clavulanate, cephalexin/cefazolin/cephalothin, nitrofurantoin, trimethoprim-sulfamethoxazole, ciprofloxacin, and gentamicin.

Outcomes

Benefit (efficacy)

The primary efficacy outcome was a composite of ED visit or hospitalization for UTI, sepsis, or bloodstream infection within 30 days to 1 year after the positive urine culture (index date + 30 to index date + 365). As secondary efficacy outcomes we assessed ED visits and hospitalizations separately and assessed time to each of these events.

Harm

The primary harm outcome was detection of a repeat urinary isolate resistant to any urinary agent to which the first urine culture was susceptible. This resistance outcome was counted when the initial isolate was reported as susceptible but a subsequent isolate was reported as resistant to that same antibiotic between 30 days and 1 year after the initial positive urine culture. This outcome allowed any organism to be isolated, whether or not it was the same species as the initial isolate, as long as new resistance was reported to a specific antibacterial agent. Antibiotic susceptibility was identified based on the results reported on the culture and susceptibility report rather than the full laboratory panel, as the latter is unavailable via OLIS.

As a secondary resistance harm outcome we conducted the same analysis but limited to the susceptibility result for the agent to which the patient was treated with prophylaxis. Any newly detected resistance where the index isolate was susceptible (or susceptibility was not reported) but the subsequent isolate was resistant to the antibiotic used for prophylaxis was counted. For example, if a patient received prophylaxis with nitrofurantoin, then this outcome would only assess for new resistance to nitrofurantoin in the case and matched unexposed patient.

Another harm-related outcome was ED visit or hospitalization for any one of a composite of antibiotic-related complications: allergy, diarrhea, Clostridioides difficile, general medication adverse event, or candidiasis within 30 days and 1 year of the positive urine culture (Supplementary Table 1). This composite outcome has been used in a previous study in the Ontario long-term-care setting [15].

Statistical Analysis

The characteristics of patients undergoing UTI prophylaxis were compared with matched control patients not receiving prophylaxis (1:10 ratio). We matched on organism isolated in the index urine culture, number of positive urine cultures in the past 12 months, and the propensity score. The propensity score was based on a logistic regression model with receipt of prophylaxis as the outcome, and included all 12 other baseline variables (age, sex, location of urine culture, calendar month, Deyo-Charlson comorbidity score, diabetes, kidney stones, genitourinary tract cancer, genitourinary tract surgery, ED visits in the past 12 months, hospitalization in the past 12 months, and antibiotic susceptibility results on first urine culture based on the organism susceptibility, resistance, or nonreporting for the 7 antibiotics listed above) as covariates. Matching was conducted using a greedy algorithm, with a caliper width of 0.2 for the propensity score. We examined differences in patients who did and did not receive prophylaxis using standardized differences, both before and after matching.

In the primary analysis, we compared the 1-year risk (starting from 30 days after the index urine culture) of each benefit and harm outcome among prophylaxis patients versus matched controls. The differences were compared using the McNemar test to account for matching. The rates of each benefit and harm outcome were compared between prophylaxis patients and matched controls using a Cox proportional-hazards regression model in which the hazards of benefit and harm outcome were regressed on prophylaxis exposure. A robust variance estimator was used to account for matching. We performed subset analyses based on the following prespecified subgroups: long-term-care versus community residents, men versus women, less than 80 versus 80 years of age and above, index urinary culture organism (grouped by Escherichia coli, non–E. coli, and mixed growth not specifying a single organism), and number of positive urine cultures in the past year. In further secondary analyses, we examined each outcome based on time-to-event using Kaplan-Meier curves to illustrate the difference in outcomes between prophylaxis and non–prophylaxis patients over time. Hazard ratios (HRs) for prespecified subset analyses were displayed in forest plots. Two sensitivity analyses were performed: (1) limiting prophylaxis agents to those most likely to be used for urinary indications (nitrofurantoin, ciprofloxacin, norfloxacin, trimethoprim ± sulfamethoxazole, and fosfomycin) and (2) shifting the follow-up period start from 60 days after the urine culture collection date, to assess for immortal time bias associated with the differential length of antibiotic treatment between the prophylaxis group and controls. Analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA).

Ethics

The use of data in this project was authorized under section 45 of Ontario’s Personal Health Information Protection Act, which does not require review by a Research Ethics Board.

RESULTS

During the 2-year study period there were 5 231 404 urine cultures performed and, of those, 1 781 284 were positive. When limiting to outpatients aged 66 years and older and after exclusions, 186 589 unique patients with an incident positive urine culture who received antibiotics were eligible for inclusion (Figure 1). Of these patients, 3198 (1.0%) received antibiotic prophylaxis. Prophylaxis recipients received a median antibiotic duration of 49 days (interquartile range, 30 to 94 days). The most common agents used for prophylaxis included nitrofurantoin (n = 1401, 44.0%), trimethoprim/sulfamethoxazole (n = 542, 17.0%), ciprofloxacin (n = 500, 15.7%), cephalexin (5.9%), and trimethoprim (4.8%). The classes of antibiotics prescribed within the first 30 days for prophylaxis recipients and controls were generally similar between groups (Supplementary Table 2).

Before matching, antibiotic prophylaxis recipients were more likely to be male, live in a community setting, have a greater comorbidity score, have a higher number of ED and hospital visits, have more prior positive urine cultures, and have a greater proportion of antibiotic resistance at baseline. Matching on organism, number of prior positive urine cultures, and propensity score yielded 3190 UTI prophylaxis patients and 30 542 matched controls. Prophylaxis and control patients were well balanced, with standardized differences of less than 10% between those receiving and not receiving prophylaxis across all patient (Table 1) and pathogen (Table 2) characteristics.

Healthcare Utilization

Overall, 4.7% (n = 151) of UTI prophylaxis patients and 3.6% (n = 1092) of controls required an ED visit or hospitalization for UTI, sepsis, or bloodstream infection (HR, 1.33; 95% CI, 1.12–1.57) (Table 3, Figure 2). This increased risk of healthcare utilization with UTI prophylaxis was most evident in patients less than 80 years of age (HR, 1.66; 95% CI, 1.33–2.06), females (HR, 1.57; 95% CI, 1.25–1.94), community-dwelling patients (HR, 1.45; 95% CI, 1.21–1.73), those with no previously positive urine cultures (HR, 1.38; 95% CI, 1.07–1.74), and patients with non–E. coli-positive urine cultures (HR, 1.54; 95% CI, 1.18–1.96) (Figure 3).

Antibiotic Resistance

Acquisition of resistance to any antibiotic agent occurred in 20.9% (n = 431) of patients receiving prophylaxis and 16.4% (n = 3218) of matched controls (HR, 1.31; 95% CI, 1.18–1.44). The higher incidence of antibiotic resistance in the prophylaxis group was evident across most subgroups, except for male patients, long-term-care patients, and those with 1 or 2 prior UTIs (Figure 3).

Acquisition of resistance to the agent used for prophylaxis was more pronounced and occurred in 12.2% (n = 388) of patients on prophylaxis compared with 6.3% (n = 1911) of controls (HR, 2.01; 95% CI, 1.80–2.24) and was consistently elevated across all patient subgroups.

Antibiotic-Related Complications

Antibiotic-associated harm occurred in 7.6% (n = 243) of patients receiving prophylaxis and 7.1% (n = 2169) of non–prophylaxis controls (HR, 1.08; 95% CI, .94–1.22). Notable harms included the risk of C. difficile infection (HR, 1.56; 95% CI, 1.05–2.23) and general medication adverse events (HR, 1.62; 95% CI, 1.11–2.29), both of which were elevated in UTI prophylaxis recipients (Figure 2, Table 3).

Sensitivity Analyses

When limiting prophylaxis agents to those most likely to be used for urinary indications, the risk of an ED visit or hospitalization for UTI, sepsis, or bloodstream infection with prophylaxis persisted (HR, 1.29; 95% CI, 1.06–1.54), as did resistance to any agent (HR, 1.31; 95% CI, 1.17–1.45). The sensitivity analysis also found a similar risk of overall antibiotic-associated harm (HR, 1.10; 95% CI, .95–1.26). When limiting the follow-up window to start from 60 days after the urine culture collection, the association between UTI prophylaxis and healthcare utilization for infection remained elevated, with prophylaxis recipients at greater risk for an ED visit or hospitalization for UTI, sepsis, or bloodstream infection (HR, 1.24; 95% CI, 1.02–1.49). Similarly, the association between UTI prophylaxis and antibiotic resistance to any urinary agent (HR, 1.31; 95% CI, 1.17–1.45) and to the antibiotic used for prophylaxis (HR, 5.77; 95% CI, 5.03–6.61) was also pronounced. However, the association between UTI prophylaxis and general medication adverse events (HR, 1.22; 95% CI, .77–1.84) was attenuated.

DISCUSSION

Among older adults with a history of positive urine culture, antibiotic prophylaxis was associated with an increased risk of antibiotic-related harms, most notably antibiotic resistance to both the agent used for prophylaxis and any urinary antibiotic. Meanwhile, there was no detectable clinical benefit of prophylaxis, overall and in any subgroups, in that this strategy was associated with increased, rather than decreased, risks of an ED visit or hospitalization for UTI, sepsis, or bloodstream infection.

While there are minimal data evaluating the efficacy and safety of UTI prophylaxis with antibiotics in older adults [6], our study with over 3000 older patients receiving UTI prophylaxis confirms the results of larger studies in younger patients [8, 9] in documenting an increased risk of C. difficile and general medication adverse events.

Interestingly, our study did not find a protective benefit but instead found a risk associated with long-term antibiotics and increased ED visits or hospitalization for infection. This differs from previous literature indicating that long-term antibiotics reduced the recurrence of UTI [6, 7]. This difference may be because our study evaluated the more distal outcome of ED visit or hospitalization for infection as opposed to early symptomatic or microbiological UTI recurrence. Given the high rate of asymptomatic bacteriuria (up to 50%) in older individuals [20], our findings may reflect the harms associated with overtreatment of this condition, which does not provide benefit in this population [21]. Additionally, there is evidence from 1 prospective study in premenopausal women that using antibiotics to manage asymptomatic bacteriuria is actually associated with increased risk of infection [22], perhaps due to the disruption of protective flora that may support colonization resistance to pathogenic bacteria.

In our study, antibiotic use for UTI prophylaxis was associated with increased risk of antibiotic resistance, both to the antibiotic used for prophylaxis and to other urinary antibiotics. These findings are consistent with previous literature linking UTI prophylaxis with subsequent resistant organisms [8, 9, 23]. This is of particular concern given that antibiotic resistance will render severe infections (eg, sepsis caused by urinary pathogen) more difficult to treat and can lead to worse outcomes [24].

Limitations to this study include the possibility of unmeasured confounding-by-indication, in that patients receiving UTI prophylaxis may be at greater risk for harm due to inherent differences other than antibiotic use itself. However, using 1:10 matching on organism, prior positive urine cultures, and propensity score, our patients were well balanced on key characteristics related to UTI outcomes. The duration of prophylaxis prescribed in our cohort is generally shorter than that seen in clinical trials; however, the real-world setting may result in additional reasons for shorter durations, including adverse advents, nonadherence, intercurrent hospital admission, or death. While it is possible that the antibiotics prescribed were used for other indications besides UTI prophylaxis, the vast majority of antibiotics were those typically used for this indication. A sensitivity analysis limiting prophylaxis agents only to those most likely to be used for urinary indications was consistent with our initial findings. Additionally, since our study is based on laboratory and administrative data, we are unable to differentiate true infection from asymptomatic bacteriuria. Given the high proportion of asymptomatic bacteriuria in older populations and the fact that it is commonly unnecessarily treated with antibiotics [25], the benefit-to-harm balance may be disproportionately shifted towards harm in our cohort. However, this means that our findings may be more representative of real treatment practices in the general population and may more closely estimate these harms in comparison to RCTs in which patients are rigorously screened for urinary symptoms prior to treatment.

While most guidelines favor the use of antibiotic prophylaxis in patients with recurrent UTI, recommendations are made on the basis of efficacy: evidence for reduced risk of urinary tract infection in younger adults [26, 27]. Our findings may help highlight some of the risks of UTI prophylaxis in older adults, which could shift clinicians’ understanding of the ratio between benefits and harms in this population.

Conclusions

Among older adults, long-term antibiotic UTI prophylaxis is associated with an increased, rather than decreased, risk of hospitalization or ED visit for UTI, sepsis, or bloodstream infection. Urinary tract infection prophylaxis is also associated with the development of antibiotic resistance as well as a potential for adverse events, including C. difficile infection. These findings suggest that harms of UTI prophylaxis may outweigh benefits in older adults.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Financial support. This work was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). This work was also supported by the Canadian Institutes of Health Research (grant number 159503).

Potential conflicts of interest. The authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.

References

Her Majesty the Queen in Right of Canada, as represented by the Public Health Ontario, 2021.

Issue Section:

Major Articles and Commentaries

Comments

0 Comments

I agree to the terms and conditions. You must accept the terms and conditions.

Email alerts

Citing articles via

More from Oxford Academic