The IOU Consensus Recommendations for Empirical Therapy of Cystitis in Nursing Home Residents

Abstract

Objectives:

To establish consensus guidelines for empiric treatment of uncomplicated cystitis with anti-infectives in non-catheterized older nursing home residents.

Design:

Two-round modified Delphi Survey.

Participants:

Expert panel of 19 clinical pharmacists.

Measurements:

Comprehensive literature search, and development/review/edit of draft survey by the investigative group (i.e., 1 geriatric clinical pharmacist, 2 physicians from geriatric and 1 from infectious disease internal medicine). The expert panel rated their agreement with each of 31 recommendations for drugs of choice, dosing medications at various levels of renal function, drug-drug interactions to avoid, and duration of therapy by gender on a 5-point Likert scale (1= strongly disagree to 5=strongly agree). Consensus agreement was defined as a lower 95% confidence limit of ≥4.0 for the recommendation–specific mean score.

Results:

The response rate was 95% for the first round and 3 recommendations achieved consensus (i.e., dosing for nitrofurantoin and trimethoprim/sulfamethoxazole in those without chronic kidney disease, and drug-drug interaction between trimethoprim/sulfamethoxazole and warfarin). In the second round, 90% responded and reached consensus on an additional 8 recommendations (2 for nitrofurantoin or trimethoprim/sulfamethoxazole as initial drugs of choice, 3 for dosing ciprofloxacin, nitrofurantoin, and trimethoprim/sulfamethoxazole at various levels of chronic kidney disease, and 3 drug-drug interactions to avoid: trimethoprim/sulfamethoxazole with phenytoin and ciprofloxacin with theophylline, or tizanidine)

Conclusion:

An expert panel of clinical pharmacists was largely able to reach consensus agreement on a number of recommendations for the optimal empiric treatment of cystitis with oral anti-infective medications in older nursing home residents. The recommendations were incorporated into a treatment algorithm for uncomplicated cystitis in non-catheterized nursing home residents and used in educational materials for health professionals in an ongoing controlled intervention study.

INTRODUCTION

Suspected urinary tract infection (UTI) is the most common reason that antibiotics are prescribed to older nursing home residents.^1-3 While UTI may include a number of clinical entities, uncomplicated cystitis accounts for most cases of UTI.^4-7 Patients with cystitis are typically not severely ill and present with lower urinary symptoms such as dysuria, urinary urgency / frequency, or suprapubic pain.⁸ Warning signs or symptoms such as fever, rigor, flank pain, hypotension, or prostatic pain suggest more complicated upper tract disease or prostatitis.⁸

The appropriateness of antibiotics is important as they can have considerable unintended consequences such as adverse drug events and promoting development of antibiotic resistant organisms.^1,9 Moreover, antibiotic resistant urinary tract organisms are associated with increased transfers to acute care facilities, and mortality.^1,10 In addition, suboptimal antibiotic use is the leading cause of Clostridium difficile infections which can result in serious and life threatening complications.^1,11

Unfortunately, much of the antibiotic use for cystitis among nursing home residents is potentially suboptimal. For example, Rotjanapan and colleagues conducted a six-month study in two Rhode Island nursing homes (total of 270 beds) and found that 72% of residents with cystitis did not receive a drug of choice.¹² In addition, 46% of antibiotics given had a suboptimal dosage and nearly 70% were given an antibiotic for an excessive duration. Another study by Miller and colleagues reported on the appropriateness of antibiotic prescribing for 282 residents aged 65 and older who received an antibiotic for cystitis in 18 U.S. nursing homes over a 12-month period.¹³ Overall, 40% of antibiotics were deemed to be potentially suboptimal. The most common problem was excessive doses of primarily renally cleared antibiotics in 19% of residents. Other problems seen included the use of higher cost antibiotics, therapeutic duplication, impractical directions, prolonged duration and drug-interactions.¹³

While there are clinical practice guidelines for defining, evaluating and monitoring infections in nursing home residents, they are not specific to optimal use of antibiotics in treating cystitis.^14-18 Given this background, the study objective was to establish consensus guidelines for the optimal prescribing of empiric anti-infectives for uncomplicated cystitis in non-catheterized older nursing home residents via a modified Delphi survey of an expert panel of clinical pharmacists.^19,20

METHODS

To develop the modified Delphi survey, the investigative team worked with two medical librarians at the University of Pittsburgh to conduct a comprehensive literature review restricted to English language articles in PubMed and Embase using a combination of terms including urinary tract infections, cystitis, antibacterial agents, nursing homes, and aged. Abstracts of these articles were reviewed. Studies deemed relevant underwent full article review. In addition, reference lists from review and retrieved articles were scanned for any additional relevant studies.

Using the identified literature and considering four important aspects of medication appropriateness (i.e., drugs of choice, dosing medications at various levels of renal function, drug-drug interactions to avoid, and duration of therapy by gender), a clinical pharmacist researcher developed the initial draft of the Delphi survey.^19,20 This draft survey was further refined by three internal medicine physician investigators (2 from geriatric medicine and 1 from infectious disease). The final survey included 31 recommendations and supporting references to be considered by an expert panel. The expert panel consisted of 19 clinical pharmacists (see Acknowledgment section for details) and they were sent the modified Delphi survey via email and were blinded to the identity of other panel members. They were asked to rate their agreement with each of 31 recommendations on a 5-point Likert scale (1= strongly disagree to 5=strongly agree). After receipt of the completed surveys, means and 95% confidence intervals were estimated for each recommendation. Consensus on agreeing to a recommendation was achieved when the lower 95% confidence limit calculated from participant responses was greater than or equal to 4.0. Consensus on disagreeing to a recommendation was achieved when the upper 95% confidence limit calculated from participant responses was less than or equal to 3.0, and the remainder was treated as failure to reach consensus.

Recommendations that did not achieve consensus in the first round were included in a second round during which the same expert panel was asked to further consider the recommendation after providing the group rating mean for each item from the first round, and additional targeted or suggested items from the expert panel to better glean consensus. This approach is consistent with previous published work.^21,22 For all statistical analyses, SAS^® version 9.3 (SAS Institute, Inc., Cary, NC) was used. The University of Pittsburgh Institutional Review Board approved the study as exempt. The results were summarized into simple treatment recommendations to facilitate clinical use.

RESULTS

The literature search yielded 253 treatment studies in PubMed and 332 in Embase for a total of 462 unique articles, of which 53 were deemed relevant after abstract review. No additional studies were identified from reference lists. Three quarters (75%) of expert panel members were female. Additionally, 68% were board certified geriatric pharmacists, whereas an additional 26% were board certified in another speciality.

The first-round response rate was 95%. After the first round of the survey, consensus was reached for 3 recommendations (using nitrofurantoin 100 mg twice daily and trimethoprim/sulfamethoxazole 160/800mg twice daily in those without chronic kidney disease, and avoiding a drug-drug interaction between trimethoprim/sulfamethoxazole and warfarin (Table 1). The panel also reached consensus against considering fosfomycin as a drug of choice, avoiding nitrofurantoin in those with a creatinine clearance<60 ml/min, avoiding trimethoprim/sulfamethoxazole in those with a creatinine clearance<30 ml/min, avoiding ciprofloxacin in patients concurrently receiving glipizide/glyburide, 5 days as the minimum treatment duration for men, and fosfomycin as the cheapest alternative (Table 1).

Table 1.

Clinical Pharmacist Expert Panel Round 1 Survey Responses for Antibiotic Treatment of Cystitis in Older Nursing Home Residents

Appropriateness Domain	Recommendations	Mean (95% CI)
Drugs of Choice for Empiric Treatment	With adequate renal function or sulfonamide allergy, nitrofurantoin	4.22 (3.86-4.59)
	With adequate renal function but without sulfonamide allergy, trimethoprim/sulfamethoxazole	4.28 (3.90-4.65)
	Fosfomycin is always the first choiceb	1.61 (1.26-1.96)
Renal Dosing	When CrClr<60, avoid nitrofurantoin b	1.89 (1.65-2.12)
	When CrClr<40, avoid nitrofurantoin	3.17 (2.50-3.83)
	When CrClr<30, avoid nitrofurantoin	4.44 (3.96-4.83)
	When CrClr<30, avoid trimethoprim/sulfamethoxazoleb	2.22 (1.75-2.69)
	When CrClr<15, avoid trimethoprim/sulfamethoxazole	4.33 (3.91-4.75)
	With adequate renal function, nitrofurantoin 100mg twice dailya	4.56 (4.30-4.81)
	With adequate renal function, trimethoprim-sulfamethoxazole 160/800mg twice daily a	4.61 (4.36-4.86)
	With reduced renal function but not to the level it should be avoided, trimethoprim-sulfamethoxazole 160/800mg daily	4.00 (3.58-4.42)
	When CrCl<30, ciprofloxacin 250mg daily	3.56 (3.04-4.07)
	With adequate renal function, ciprofloxacin 250mg twice daily	4.17 (3.68-4.66)
	Regardless of renal function, fosfomycin 3 gm single dose	4.17 (3.71-4.63)
Clinically Significant Drug-Drug Interactions	Ciprofloxacin and theophylline	4.28 (3.90-4.65)
	Ciprofloxacin and corticosteroids	3.22 (2.79-3.66)
	Ciprofloxacin and warfarin	4.06 (3.69-4.42)
	Ciprofloxacin and glipizide/glyburide b	2.33 (1.77-2.90)
	Trimethoprim/sulfamethoxazole and procainamide	3.83 (3.31-4.35)
	Trimethoprim/sulfamethoxazole and phenytoin	4.33 (3.95-4.71)
	Trimethoprim/sulfamethoxazole and warfarin a	4.89 (4.73-5.05)
	Trimethoprim/sulfamethoxazole and methotrexate	3.94 (3.45-4.44)
	Trimethoprim/sulfamethoxazole and angiotensin converting enzyme inhibitors	3.72 (3.35-4.10)
	Trimethoprim/sulfamethoxazole and angiotensin receptor blockers	3.72 (3.35-4.10)
	Trimethoprim/sulfamethoxazole and sulfonylureas	3.50 (3.11-3.89)
Minimum Duration of Anti-Infective Treatment	In women, 3 days	3.83 (3.34-4.32)
	In women, 5 days	2.89 (2.35-3.43)
	In men, 5 days b	2.06 (1.69-2.42)
	In men, 7 days	4.11 (3.73-4.49)
	In men, 10 days	2.67 (2.15-3.18)
Cost of Treatment	Fosfomycin is the least expensive choice b	1.50 (1.15-1.85)

^aConsensus agreement

^bConsensus disagreement

Abbreviations: CI =confidence interval; CrClr= creatinine clearance in milliliters per minute; gm=gram; mg=milligram

The second-round response rate was 90%. Four new recommendations were added. The panel reached consensus agreement on: nitrofurantoin as the drug of choice for empiric treatment of cystitis among individuals with adequate renal function and a sulfonamide allergy; trimethoprim/sulfamethoxazole as the drug of choice in those with adequate renal function and without a sulfonamide allergy; avoiding nitrofurantoin in those with a creatinine clearance<30 ml/min; avoiding trimethoprim/sulfamethoxazole in those with a creatinine clearance<15ml/min; for those with adequate renal function, 250mg twice daily as the appropriate dose of ciprofloxacin; significance of the drug-drug interaction between ciprofloxacin and theophylline, between ciprofloxacin and tizanidine, and that between trimethoprim/sulfamethoxazole and phenytoin (Table 2). The panel disagreed that ciprofloxacin and sulfonylureas constitute a significant drug-drug interaction, and men should be treated for a minimum of 10 days (Table 2). Consensus was not reached for the remaining items (Table 2). Figure 1 operationally summarizes the survey results as treatment guidelines.

Table 2.

Clinical Pharmacist Expert Panel Round 2 Survey Responses for Antibiotic Treatment of Cystitis in Older Nursing Home Residents

Appropriateness Domain	Recommendations	Mean (95% CI)
Drugs of Choice for Empiric Treatment	With adequate renal function or sulfonamide allergy, nitrofurantoin a	4.41 (4.15-4.67)
	With adequate renal function but without sulfonamide allergy, trimethoprim/sulfamethoxazole a	4.35 (4.04-4.66)
Renal Dosing	When CrClr<60, avoid nitrofurantoin	3.35 (2.72-3.98)
	When CrClr<30, avoid nitrofurantoin a	4.82 (4.62-5.03)
	When CrClr<15, avoid trimethoprim/sulfamethoxazole a	4.41 (4.00-4.82)
	With reduced renal function but not to the level it should be avoided, trimethoprim-sulfamethoxazole 160/800mg daily	4.12 (3.72-4.52)
	When CrCl<30, ciprofloxacin 250mg daily	3.76 (3.34-4.19)
	When CrCl<30, ciprofloxacin 500mg daily c	3.06 (2.44-3.67)
	With adequate renal function, ciprofloxacin 250mg twice dailya	4.47 (4.21-4.74)
	With adequate renal function, ciprofloxacin 500mg twice daily c	2.53 (1.90-3.16)
	Regardless of renal function, fosfomycin 3 gm single dose	4.24 (3.74-4.73)
Clinically Significant Drug-Drug Interactions	Ciprofloxacin and theophylline a	4.41 (4.09-4.73)
	Ciprofloxacin and corticosteroids	3.06 (2.72-3.40)
	Ciprofloxacin and warfarin	4.06 (3.72-4.40)
	Ciprofloxacin and glipizide/glyburide b	2.35 (1.91-2.80)
	Ciprofloxacin and tizanidine a,c	4.41 (4.05-4.78)
	Trimethoprim/sulfamethoxazole and procainamide	4.06 (3.72-4.40)
	Trimethoprim/sulfamethoxazole and phenytoin a	4.41 (4.15-4.67)
	Trimethoprim/sulfamethoxazole and methotrexate	4.06 (3.67-4.44)
	Trimethoprim/sulfamethoxazole and angiotensin converting enzyme inhibitors	3.71 (3.31-4.10)
	Trimethoprim/sulfamethoxazole and angiotensin receptor blockers (or potassium supplements) c	3.71 (3.31-4.10)
	Trimethoprim/sulfamethoxazole and sulfonylureas	3.59 (3.22-3.95)
Minimum Duration of Anti-Infective Treatment	In women, 3 days	4.00 (3.69-4.31)
	In women, 5 days	2.71 (2.27-3.14)
	In men, 7 days	4.00 (3.64-4.36)
	In men, 10 days b	2.47 (2.06-2.88)

^aConsensus agreement

^bConsensus disagreement

^cNew recommendation added in round 2

Abbreviations: CI =confidence interval; CrClr=estimated creatinine clearance in milliliters per minute; gm=gram; mg=milligram

Figure 1:

Empiric Treatment of Uncomplicated Bladder Infection (Cystitis) In Nursing Home Residents ≥ 65 Years Without a Urinary Catheter

DISCUSSION

This study found that an expert panel of clinical pharmacists could achieve consensus agreement on the optimal prescribing of empiric oral anti-infectives in uncomplicated cystitis in older nursing home residents. Regarding drugs of choice, the panel reached agreement that preferred drugs were nitrofurantoin and trimethoprim/sulfamethoxazole. This is consistent with recent guidelines and reviews and also is clinically sensible as they are capable of successfully treating the three most common cystitis organisms, Escherichia coli (E. coli), Klebsiella spp. and Proteus spp. for uncomplicated cystitis in older nursing home.^17,23-30 In addition, nitrofurantoin may also be effective in treating Enterococcus that maybe resistant to other choices and is also a good choice in those with a sulfa allergy.²³ Where resistance to E. coli is less than 20%, trimethoprim/sulfamethoxazole may be a reasonable alternative as its average wholesale price is less than nitrofurantoin and allows one to avoid the small risk of pulmonary reactions that can be seen with nitrofurantoin.^{23-24, 29,31-32}

The expert panel agreement also provides guidance regarding the renal dosing recommendations for several anti-infectives consistent with a previous Delphi Survey.²² Although controversial, nitrofurantoin should not be used in residents with estimated creatinine clearance <30ml/min because urinary concentrations may be insufficient for it to be effective.³³ In those whose creatinine clearance is >15ml/min, trimethoprim/sulfamethoxazole is an acceptable alternative. Below this level of renal function, the risk of hyperkalemia with trimethoprim/sulfamethoxazole is greater and alternatives such as ciprofloxacin or fosfomycin should be considered.^24,31,34

The panel also identified four clinically important drug-drug interactions, three of which involved narrow therapeutic range drugs (i.e., phenytoin, theophylline, and warfarin). A case control study from Canada found that trimethoprim/sulfamethoxazole in combination with warfarin increased the risk of upper gastrointestinal hemorrhage by nearly 4-fold (adjusted odds ratio [aOR]= 3.84).³⁵ This finding was confirmed by another case-control study from the U.S. in which trimethoprim/sulfamethoxazole increased the risk of bleeding with warfarin by nearly 3-fold (aOR=2.70).³⁶ This same anti-infective was found in another case-control study to be the culprit interacting with phenytoin in which the risk of toxicity requiring hospitalization was increased more than two-fold (aOR=2.11).³⁷ In addition, ciprofloxacin increases the risk of theophylline toxicity by nearly two-fold (aOR=1.86).³⁸

This study had several potential limitations. First, though the panel agreed that anti-infective treatment in men more than 5 days but less than 10 days in duration was acceptable for uncomplicated cystitis, we were unable to reach consensus regarding optimal duration of treatment. For women, most sources are clear that the duration of use for anti-infectives for cystitis should range between 3-7 days.^23-24,28-29 Therefore, we operationalized the duration of use as 7 days for men and 3 days for women in the final recommendations (Figure 1). Second, we had used a convenience sample of clinical pharmacists for this modified Delphi survey. However, these individuals came from 10 states and included clinical pharmacists from academia, community base care and managed care. Third, face-to-face meetings of the panelists were not held due to the prohibitive cost and other practical considerations. However, non-face-to-face participation is felt to be advantageous in maintaining blinding and limits confounding arising from dominant personalities.³⁹ Finally, as new research contributes additional information about emerging resistance patterns, and safe and effective medication use in the elderly, the list of anti-infective drugs and associated prescribing information will likely need to be updated.

Like with all treatment guidelines, clinical judgment is important. It is important to note that active patient monitoring without empiric antibiotic use may also be acceptable while awaiting culture results. These guidelines are meant for empiric treatment pending culture results. Clinicians should reassess a patient’s clinical progress along with antibiotic appropriateness once culture results return. Referred to as an antibiotic time out, this reassessment offers the clinician an opportunity to refine antibiotic prescribing by narrowing the antimicrobial spectrum, reducing treatment duration, and potential discontinuing unnecessary additional treatment. When using this treatment guideline, clinicians should also be mindful of local resistance patterns in their facility as well as disease severity, recent antibiotic use, and/or prior history of antimicrobial resistance in an individual patient as these factors may influence the choice of empiric antimicrobial selected.

In conclusion, an expert panel of clinical pharmacists was largely able to reach consensus agreement on a number of recommendations for the optimal empiric treatment of cystitis with oral anti-infective medications in older nursing home residents. These recommendations were incorporated into a treatment algorithm for uncomplicated cystitis in non-catheterized older nursing home residents and used in educational materials for health professionals in an ongoing controlled intervention study (AHRQ R18 HS023779).

Impact Statement:

We certify that this work is NOVEL clinical research. While clinical practice guidelines exist for the diagnosis of urinary tract infections in nursing home residents, there are no clinical practice guidelines focusing specifically on the optimal use of antibiotics for the treatment of uncomplicated cystitis in non-catheterized nursing home residents.