Infection Management for Geriatrics in Long Term Care Facilities

134 Richards and Jarvis spend some time in LTCFs before death, and the majority (53%) of residents in LTCFs will spend 1 year or longer (1,2). An important distinction between LTCFs and acute care facilities is that LTCFs are residential and persons in LTCFs are generally referred to as “residents” instead of as “patients.” Even though medical care is provided to LTCF residents, other aspects of a resident’s life take on greater importance than in acute care facilities. Socialization through group activ- ities, both inside and outside the LTCF, is key to promoting good mental health for residents, although these activities may increase the risk of exposure to infec- tious agents, such as influenza. Group settings for eating and physical therapy, vi- tal to the maintenance of resident independence and functional status, may in- crease risk for foodborne outbreaks, person-to-person transmission, or exposure to potential fomites, such as physical therapy equipment. Not surprisingly, the management of infectious disease outbreaks in LTCFs is complicated because the focus of LTCF care is on the preservation of physical function and socialization, not exclusively on diagnosis and treatment of infectious diseases. The availability of clinicians to evaluate febrile residents may be limited, and diagnostic studies, including microbiologic cultures, are generally less available than in acute care fa- cilities. Consequently, nursing assistants usually perform the initial resident as- sessment, and licensed nurses relay important findings to clinicians, usually by telephone (3). In an effort to improve the evaluation of LTCF residents with fever or suspected infection, recommendations for the minimal evaluation of patients who develop fever in LTCFs were published recently (4). These guidelines spec- ify tasks appropriate for nursing assistants and licensed nurses. In LTCFs, antimicrobials for the empiric treatment of suspected infection often are prescribed without onsite clinician evaluation or diagnostic testing (5,6). When diagnostic testing is performed, only limited tests are available in most LTCFs. This, together with outsourcing of most laboratory tests, may lead to sub- optimal timeliness of reporting and, in some situations, inaccurate or misleading results. When residents are acutely ill or diagnostic testing is not available in LTCFs, residents often are transferred to the emergency departments of acute care hospitals. Not surprisingly, evaluation and management of infection accounts for approximately one-quarter of resident transfers from LTCFs to hospitals (7). III. RISK FACTORS FOR OUTBREAKS IN LTCFs Risk factors for outbreaks include both resident and institutional factors (Table 1). The typical resident in an LTCF is female, older than 80 years of age, cognitively impaired, and living with several underlying medical conditions. Individual risk factors for infection include immunologic senescence, malnutrition, multiple chronic diseases, medications (e.g., immunosuppressants, central nervous system agents that diminish cough reflex), cognitive deficits that may complicate resident

Epidemiology of Infectious Disease Outbreaks 135 Table 1 Potential Risk Factors for Infectious Disease Outbreaks in Long-Term Care Facilities Resident level Decreased immunity to infection Malnutrition Chronic disease Functional impairment including diminished cough reflex, urinary incontinence, fecal incontinence, immobility Medications, especially psychoactive medications that diminish cough reflex and consciousness Invasive devices, such as urinary catheters, enteral feeding tubes, tracheostomies, etc. Institutional level Larger size (e.g., larger number of residents) Facility design (e.g., single versus multiple resident rooms) Group activities such as meals, physical therapy, recreational activities Low immunization rates Excessive antimicrobial use Widespread colonization of residents with antimicrobial-resistant organisms compliance with basic sanitary practices (e.g., hand washing), functional impair- ments (e.g., fecal and urinary incontinence, immobility, diminished cough reflex), or invasive device use (e.g., urinary catheters, enteral feeding tubes, tra- cheostomies) (3,5,6,8). Institutional factors associated with increased risk for outbreaks are varied. Frequent group activities such as meals, physical therapy, recreational activities, or the common use of shared facilities (e.g., showers or whirlpool baths) increase the risk for outbreaks (3,5,6,8). Risk for outbreaks caused by specific pathogens (e.g., influenza, Streptococcus pneumoniae) is increased in settings where resident and healthcare worker immunization coverage is low. Finally, widespread excessive antimicrobial use and high rates of colonization with antimicrobial-resistant or- ganisms increase the risk of outbreaks from these organisms. In a study of outbreaks among New York LTCFs, institutional risk factors for respiratory or gastrointesti- nal infection outbreaks included larger LTCFs (risk ratio 1.71 per 100 bed increase), LTCFs with a single nursing unit, or LTCFs with multiple units but shared staff (9). Risk for outbreaks was lower in LTCFs with paid employee sick leave. IV. KEY ASPECTS OF INFECTIOUS DISEASE OUTBREAK INVESTIGATION The epidemiologic investigation of infectious disease outbreaks in LTCFs should be conducted in a systematic fashion. Key components of the investigation are de- scribed below and are listed in Table 2.

136 Richards and Jarvis Table 2 Key Aspects of Outbreak Investigation in Long-Term Care Facilities 1. Have an infection control plan and program 2. Ask two important questions Is this surveillance artifact? Is an epidemiologic investigation needed? 3. Develop the case definition and line listing 4. Ascertain cases 5. Determine person, place, and time Host factors (person) Geographic assessment (place) The epidemic curve (time) 6. Develop preliminary hypotheses 7. Evaluate hypotheses Cohort and case-control studies Observational studies Microbiologic studies 8. Implement intervention(s) 9. Evaluate impact of intervention(s) A. Infection Control Plan and Program Unlike hospitals, most LTCFs do not have substantial resources committed to in- fection control (10). Every infection control program should have an infection control plan outlining personnel, responsibilities, reporting relationships, and surveillance activities (see also Chapters 8 and 9). Designating a staff member as the “infection control person” is not sufficient; ideally, a trained, experienced ICP should be responsible for the program, either as a staff member at the facility or on a consulting basis. Knowing who is responsible within a facility for conduct- ing surveillance and identifying, investigating, intervening, and reporting an out- break is critical if the outbreak is to be identified and controlled as early as possi- ble. Finally, establishing an infection control committee with active participation by the LTCF administrator, medical director, ICP, and nursing staff is important not only for support and guidance during an outbreak but for continued vigilance in optimizing infection control prevention efforts to avoid outbreaks. B. Is This Surveillance Artifact? An important first question to ask in a potential outbreak situation is whether the “outbreak” actually represents surveillance artifact (11). Common causes of surveil- lance artifact include (1) introduction of new infection definitions or surveillance methods; (2) a new ICP or inexperienced infection control staff; (3) new laboratory tests or populations; and (4) change in frequency of microbial cultures. For exam-

Epidemiology of Infectious Disease Outbreaks 137 ple, surveillance data may demonstrate a markedly increased rate of pneumonia in LTCF residents. This increase may represent a true outbreak; however, broadening the definition used for pneumonia or increasing the number of residents admitted with severe neurological impairment and frequent aspiration also may increase the rate of pneumonia without representing a true infectious disease outbreak. C. Deciding When to Conduct an Investigation The decision to conduct an epidemiologic investigation may be complicated and is generally driven by three situations: (1) identification of unusual infections or organisms with high potential for morbidity or mortality (e.g., a single case of meningococcal meningitis); (2) identification of organisms or infections that, al- though relatively common, have high risk for morbidity, mortality, and transmis- sion to other residents (e.g., influenza, Norwalk virus); or (3) identification of epi- demiologically important organisms (i.e., several cases of methicillin-resistant Staphylococcus aureus [MRSA] pneumonia). The definition for what constitutes an outbreak depends on the type of in- fection and, to some extent, the facility. If, for example, a single episode of a highly contagious, potentially lethal infection, such as meningococcal meningitis, is identified in an LTCF resident, an epidemiologic investigation and early, ag- gressive infection control interventions are necessary. In contrast, knowing when to call a cluster of several LTCF residents with acute respiratory infections or gas- troenteritis an outbreak is more difficult. Generally, most authorities suggest that when the rate of infections is significantly higher than baseline endemic rates, an epidemic is occurring and an epidemiologic investigation is warranted (11,12). This definition depends on two factors to be helpful: the LTCF should have a surveillance system in place for detecting infections and calculating and compar- ing infection rates. More important, the LTCF should have a clinical staff mem- ber who understands basic principles of infection control and infectious disease epidemiology and is knowledgeable about changes in the LTCF’s resident popu- lation and infection trends in the facility. An important early consideration is whether and when to seek additional outside assistance to assist in the epidemiologic investigation. If the facility has a trained ICP, the initial investigation for most types of infectious disease outbreaks can be conducted and appropriate infection control interventions instituted by the ICP. If the facility lacks an ICP, then the clinical staff member charged with con- ducting the epidemiologic investigation may wish to seek assistance from an ICP or healthcare epidemiologist, a professional organization such as the Association of Professionals in Infection Control and Epidemiology (APIC) or the Society for Healthcare Epidemiology of America (SHEA), a state or local public health agency, or the Centers for Disease Control and Prevention (CDC). Regulations vary from state to state, but in general, LTCFs are required to report infectious dis- ease outbreaks to local or state public health agencies.

138 Richards and Jarvis Table 3 Example of a Line Listing—Influenza Outbreak Case Age Sex Ward/Room Onset Cough Fever CXR Culture Meals Physical therapy 1 87 M 4A 401 3/01/01 Yes Yes ϩ ϩ In room Yes 2 90 F 3A 304 3/02/01 Yes No ϩ ϩ On ward Yes 3 99 F 2A 208 3/02/01 Yes Yes Ϫ ϩ Main dining room Yes 4 80 F 2A 208 3/03/01 Yes No Ϫ ϩ Main dining room Yes 5 90 M 2B 240 3/05/01 Yes Yes ϩ ϩ Main dining room Yes Abbreviations: CXR, chest X-ray; M, male; F, female. D. Case Definition and Line Listing Some challenges to conducting an epidemiologic investigation in an LTCF are unique; however, the basic approach to epidemiologic investigation is the same, whether the investigation occurs in a hospital or LTCF (11,12). The initial step in an investigation is usually a case review. When clinical staff or ICPs note in- creases in rates of infection, an unusual clustering of infections, or infections with unusual agents, all residents thought to fit in the cluster should be reviewed. The easiest tool for this review is a line listing containing demographic, clinical, and exposure information for each patient (Table 3). As early as possible, investiga- tors should develop a tentative case definition. The case definition should include who, what, where, and when—a description of the infectious disease (what) along with three important parameters: person (who), place (where), time (when). For example, a case definition that could be used in a hypothetical pneumonia out- break caused by influenza might include (see Table 4): Table 4 Example of a Case Definition—Influenza Outbreak What (Disease) Respiratory illness with at least two of the following symptoms/signs Person (Population) Place New or increased cough Time New or increased sputum production Fever (temperature Ͼ 38°C) Pleuritic chest pain New or increased shortness of breath Respiratory rate Ͼ 25 breaths per minute Worsening mental or functional status X-ray compatible with pneumonia Positive respiratory tract culture for influenza A Elderly residents Ward B of LTCF A. From January 1 to January 30, 2001



































156 Rajagopalan et al. timicrobial prescribing in residents of LTCFs, with focus on age-related physio- logic, pharmacokinetic, and pharmacodynamic changes that can affect the selec- tion and dosing of such chemotherapeutic agents. B. Assessment of Antimicrobial Use in LTCFs Significant data evaluating the appropriateness of the therapeutic utility of an- timicrobial agents in LTCFs suggest that a substantial proportion of antibiotic treatments are often initiated in the absence of important diagnostic information, such as the presence of fever, leukocytosis, or culture information (3–5). One study surveying 42 nursing homes and 11 affiliated intermediate care facilities suggested that systemic antibiotics were initiated in 62.4% of cases with inade- quate initial diagnostic evaluation (6). Another study that included 3,899 residents from 52 nursing homes indicated that 22% of all antibiotics prescribed were un- necessary, that is, viral upper respiratory infection and asymptomatic bacteriuria (4). A similar study that focused on the usage pattern of a specific antibiotic, ciprofloxacin, in a long-term care setting found that only 25% of orders for that agent were appropriate and 23% were prescribed for inappropriate indications; 49% were considered inappropriate because of more effective and/or less expen- sive available alternatives (7). One study demonstrated that antibiotic prescribing for 282 elderly residents of an LTCF who received an antibiotic for a presumed urinary tract infection was inappropriate in 40% of cases; 222 (78.7%) cases, how- ever, showed clinical improvement (8). Using a Medication Appropriateness In- dex (MAI) (9) to measure appropriateness of antibiotic prescribing, 113 antibi- otics orders (39.7%) were considered inappropriate. The three antibiotics most often inappropriately prescribed were ciprofloxacin (too expensive), trimetho- prim-sulfamethoxazole (TMP-SMX) (incorrect duration), and nitrofurantoin (im- proper dosage). In addition, inappropriate prescribing accounted for an additional $560 per day in treatment costs. II. OPTIMIZING THE USE OF ANTIMICROBIAL AGENTS IN THE LTCF The optimal use of antimicrobials in LTCFs remains problematic largely because of a delay in the diagnosis and initiation of appropriate treatment of infections. Typical manifestations of infection, such as fever, may be absent or blunted in many elderly patients with serious or life-threatening infections (10,11) (see Chapter 6). Limited availability of laboratory and radiological data may, in addi- tion, preclude a precise diagnosis. General principles for initiating antimicrobial therapy are described in Table 1.

Antimicrobial Therapy 157 Table 1 Approach to Antimicrobial Prescribing in Long-Term Care Facilities 1. Not all clinical or functional changes in LTCFs residents should be attributed to infections. 2. Antibiotics should be administered only when there is potential clinical benefit. For example, studies have clearly shown that both men and woman residing in LTCFs derive no benefit from treatment of asymptomatic bacteriuria. 3. Chronic suppressive therapy with antibiotics or antimicrobial chemoprophylaxis should be restricted unless there is documented evidence of clinical efficacy and therapeutic benefit. 4. Continuation of antibiotic therapy beyond standard recommended periods should be discouraged. For example, catheter-related urosepsis should be treated until clinical sepsis is improved but should not be continued in an attempt to maintain sterile urine. 5. In circumstances in which a specific pathogen is isolated and antibiotic sensitivity studies are available, the initial broad-spectrum antibiotic should be changed to a more narrow- spectrum agent, if the organism is susceptible to such an agent. Abbreviation: LTCF, long-term care facility. A. Minimum Criteria for Initiation of Antibiotics Minimum criteria for initiating systemic antibiotics for bacterial infections have been proposed by a consensus group of physicians, geriatricians, microbiologists, and epidemiologists (12). These criteria were developed to provide guidelines for the appropriate initiation of empiric antibiotics in clinically stable LTCF resi- dents; critically ill patients with sepsis or sepsis syndrome necessitating transfer to an acute care facility were not included. Empirical regimens for common in- fections found in LTCF residents such as skin and soft-tissue infections, respira- tory infections, and urinary infections, as well as fever of unknown origin were outlined (Table 2). Other potential infections, such as intravenous catheter-related infections or infections of mucous membranes and conjunctivae; topical antibiotic use; use of antiviral and antifungal agents; prophylactic antibiotics; and chronic suppressive antibiotics were not addressed. Prospective assessment of these guidelines for appropriate antibiotic use has not been analyzed. B. Empirical Antimicrobial Therapy An empirical antimicrobial regimen should be directed against the most likely pathogens and able to achieve the desired therapeutic concentrations at the sus- pected site of infection. The choice of a specific empirical antimicrobial regimen should be based on the severity of the patient’s illness, the nature of underlying diseases, prior exposures to antimicrobials, and history of drug allergies. The So- ciety of Healthcare Epidemiology of America (SHEA) has published a position paper on antimicrobial use in LTCFs that provides recommendations for empiri-

158 Rajagopalan et al. Table 2 Minimum Criteria for Initiation of Antimicrobials in Long-Term Care Facilities Skin and soft tissue infections Either new or increasing purulent drainage at a wound, skin, or soft tissue site, or at least two of the following: 1. Fever (temperature Ͼ 37.9°C [100°F] or and increase of 1.5°C [2.4°F] above baseline temperatures taken at any site 2. Redness 3. Tenderness 4. Warmth 5. Swelling that was new or increasing at the affected site Respiratory infections 1. If the resident is febrile with a temperature Ͼ 38.9°C [102°F], at least one of the following: (1) respiratory rate Ͼ 25 breaths per minute (2) productive cough 2. If the resident has a temperature Ͼ 37.9°C [100°F] or 1.5°C [2.4°F], increase above baseline temperature, minimum criteria for initiating antibiotics requires presence of cough and at least one of the following: (1) pulse Ͼ 100/minute (2) delirium (3) rigors (shaking chills) (4) respiratory rate Ͼ 25/minute 3. For afebrile residents known to have COPD, classified as high-risk because of age Ն 65, minimum criteria for initiating antibiotics for a suspected respiratory infection include a new or increased cough with purulent sputum production. 4. For afebrile residents who do have COPD, minimum criteria for initiating antibiotics include a new cough with purulent sputum production and a least one of the following: (1) respiratory rate Ͼ 25 breaths per minute (2) delirium Urinary tract infection 1. For residents who do not have an indwelling catheter, minimum criteria for initiating antibiotics include acute dysuria alone or fever (Ͼ 37.9°C [100°F] or 1.5°C [2.4°F] increase above baseline temperature) and at least one of the following: (1) new or worsening urgency (2) frequency (3) suprapubic pain (4) gross hematuria (5) costovertebral angle tenderness (6) urinary incontinence

Antimicrobial Therapy 159 Table 2 (Continued) 2. For residents who have a chronic indwelling catheter (either an indwelling Foley catheter or a suprapubic catheter), minimum criteria for initiating antibiotics include the presence of at least one of the following: (1) fever (37.9°C [100°] or 1.5°C [2.4°F] increase above baseline temperature) (2) new costovertebral tenderness (3) rigors (shaking chills) with or without identified cause (4) new onset of delirium Fever in which the focus of infection is unknown Presence of fever (Ͼ 37.9°C [100°F] or 1.5°C [2.4°F] increase above baseline and at least one of the following: (1) new onset of delirium (2) rigors Abbreviation: COPD, chronic obstructive pulmonary disease. Source: Ref. 12. cal antimicrobial therapy for the most frequent types of infections in nursing home residents including upper and lower respiratory tract infections, urinary tract in- fections, skin and soft-tissue infections, urinary tract infection, skin and soft-tis- sue infections, diarrhea, and fever of unknown origin (13) (Table 3). (See also spe- cific chapters on each of these infections). The common cold, pharyngitis, and sinus infections are the most frequent infections of the upper respiratory tract. Because the vast majority of upper respi- ratory tract infections are viral in etiology, empirical antibiotic therapy is seldom indicated. However, if a throat culture or a reliable streptococcal screening test documents the presence of group A streptococci, penicillin would be the drug of choice. For acute bacterial sinusitis, first-line therapy includes any of the follow- ing antibiotics: trimethoprim-sulfamethoxazole, amoxicillin, and cefuroxime ax- etil. The selection of amoxicillin-clavulanic acid for acute bacterial sinusitis should be reserved for patients who respond poorly to treatment with one of the first-line antibiotics. C. Antimicrobial Utilization Review Promoting the optimal use of antimicrobials in LTCFs requires diligent antimi- crobial utilization review. Surveillance and control activities are the major foci of these programs. Antimicrobial utilization is logically within the purview of the in- fection control program. Infection control programs traditionally have advocated education, isolation techniques, and hand washing to control nosocomial infec- tions; however, they now are beginning to address problems of antimicrobial use. A recent survey found that more than one half of LTCFs had an antimicrobial uti- lization program (14).

160 Rajagopalan et al. Table 3 Empirical Antimicrobial Therapy for Common Infections Clinical syndrome Empiric antimicrobial Upper respiratory infection None Coryza/common cold None; treat only if group A Streptococcus Pharyngitis TMP-SMX ϩ amoxicillin, cefuroxime axetil, Sinus infection macrolide; second line: amoxicillin-clavulanic Lower respiratory infection acid, quinolone Acute bronchitis Acute exacerbation of chronic Most cases viral, no antibiotics indicated bronchitis Amoxicillin, TMP-SMX, doxycycline Pneumonia TMP-SMX, amoxicillin, cefuroxime axetil, Urinary tract infection macrolide, doxycycline; second line: amoxicillin-clavulanic acid, quinolone, Skin/soft tissue clindamycin (aspiration pneumonia) Cellulitis TMP-SMX, quinolone, aminoglycoside Infected pressure ulcer* (parenteral) Candidiasis Dicloxacillin: second line: cephalexin, Diarrhea clindamycin Clostridium difficile Metronidazole or clindamycin and Salmonella, Shigella TMP-SMX or quinolone; amoxicillin-clavulanic Escherichia coli O157:H7 acid Topical antifungal Metronidazole TMP-SMX, quinolone None Abbreviation: TMP-SMX, trimethoprim-sulfamethoxazole. * May require surgical debridement; if severe systemic symptoms, initial parenteral therapy should be considered. Source: Ref. 12. D. Adverse Drug Events Approximately 350,000 adverse drug events, 20,000 of which are fatal, occur each year among the 1.5 million residents of LTCFs in the United States (15). Studies that have evaluated the patterns and quality of medication prescribing in nursing homes (15,16) have found antimicrobials to be among the most frequently impli- cated drugs in causing adverse drug events. In a study of 18 community-based nursing homes located in eastern Massachusetts encompassing 28,839 nursing home resident-months, 546 adverse drug events (1.89 per 100 resident-months)

Antimicrobial Therapy 161 and 188 potential adverse drug events (0.65 per 100 resident-months) were iden- tified (15). Overall, 51% of the adverse drug events were judged to be preventable, including 171 (72%) of the 238 fatal, life-threatening, or serious events and 105 (34%) of the 308 significant events. Antibiotics were associated with 36% of non- preventable adverse drug events, but fewer than 5% of the adverse drug events were considered preventable. The majority of adverse drug events associated with antibiotics were rashes and confirmed Clostridium difficile diarrhea. Clinicians should be aware that adverse drug events occur more frequently in frail elderly and, therefore, a systematic examination of adverse events to iden- tify risk factors should be undertaken. E. Cost of Inappropriate Use of Antimicrobial Agents The consequences of inappropriate use (or overuse and misuse) of anti-infective agents and resultant financial implications include exposing patients to the poten- tial risk of adverse drug reactions, selection of resistant bacteria, and high rates of nosocomial infections—all of which will increase healthcare costs (17,18). The total cost of inappropriate prescribing of antimicrobials can only be grossly esti- mated. These agents are among the most costly of all drugs prescribed in the United States, accounting for sales between $3 to 4 billion (18). If a quarter of all prescriptions for anti-infective agents are considered inappropriate for various reasons, this could account for an additional annual cost to the healthcare system of $1 billion. Further, the increased costs of treating adverse drug reactions, in- fections, and their complications resulting from drug resistance also have to be considered. Because the geriatric population is becoming the highest user of healthcare services and the largest consumers of drugs, it is essential that a ratio- nal approach to prescribing drugs for the elderly, including antimicrobial agents, be emphasized. III. DRUG FACTORS TO CONSIDER IN PRESCRIBING ANTIMICROBIAL THERAPY Once an antimicrobial agent is selected on the basis of known or anticipated activ- ity against the pathogen(s), the goal of therapy is to deliver that drug to the site of infection in concentrations sufficient to inhibit or kill the organism(s). Most seri- ous infections require antibiotic concentrations to exceed the minimum inhibitory concentration (MIC) of the infecting organism at the site of infection. Some drugs, such as aminoglycosides and fluoroquinolones, exhibit concentration-dependent antimicrobial effects, with high drug concentrations exerting more rapid bacterici- dal action and longer post antibiotic effects (PAE) than lower concentrations. The elderly undergo age-related physiologic changes that directly influence the dispo-

162 Rajagopalan et al. sition and efficacy of various antimicrobial agents (19). The progressive decrease in the ability of vital organ systems of the elderly to maintain homeostasis can lead to alterations in drug clearance and drug receptor sensitivity. A. Pharmacokinetics Pharmacokinetics describes the fundamental mechanics of drug movement through the body over time, including factors that describe the absorption, distri- bution, metabolism, and elimination, or the overall fate of a drug in vivo (20). A summary of age-related physiological changes are shown in Table 4 (20,21). 1. Absorption Age-related changes in the gastrointestinal tract may influence drug absorption. A decrease in gastric acid secretion and an increase in gastric pH are associated with the aging process. The absorption of antimicrobials that are dependent on in- creased acidity (e.g., sulfonamides, ketoconazole) may be decreased, whereas drugs that are degraded in an acidic environment will have greater bioavailability. The significance of these changes is generally minimal and rarely affects dosing requirements. 2. Distribution Age-related changes that can affect the distribution of drugs include changes in body composition and cardiac output. In the elderly, the ratio of body fat to total Table 4 Physiological Changes Associated with Aging Pharmacokinetic parameter Physiological change with aging Absorption ↓Gastric emptying Distribution ↓Gastric acidity Elimination ↓Gastrointestinal motility ↓Absorptive surface ↓Lean body mass ↑Body fat ↑Serum 1-acid glycoprotein ↑↓Enzyme activity ↓Hepatic (liver) blood flow ↓Renal (kidney) blood flow ↓Glomerular filtration rate ↓Tubular secretion ↓ ϭ decrease activity or function. ↑ ϭ increase activity or function. ↑↓ ϭ no increase or decrease activity or function.

Antimicrobial Therapy 163 body water is increased compared with younger individuals. A decrease in lean body mass, coupled with a decrease in total body water, is associated with a de- creased volume of distribution for water-soluble drugs. Thus, with older adults, antimicrobials that are distributed primarily in body water or lean mass may have higher blood concentrations than in younger adults, which can lead to potential toxicity. Conversely, in the elderly, lipid-soluble drugs have a greater body fat dis- tribution, which may reduce blood concentrations and lead to potential subthera- peutic blood concentrations. 3. Metabolism Age appears to have no effect on the functional activity of various cytochrome P450 isoenzymes either in terms of in vitro protein content, immunohistochemical con- tent, or in vivo enzyme activity for older patients as compared to younger patients. 4. Clearance Age-related changes in renal function are probably the most significant contribu- tors to alteration in drug clearance. Reduction in kidney mass, renal blood flow, and the subsequent number of functioning nephrons, glomerular filtration rate, and the rate of tubular secretion accounts for the decreased renal excretory capac- ity observed with aging (20,21). Diminishing renal function and lack of compen- satory increases in nonrenal clearance in elderly patients have been associated with prolongation of the serum half-lives of beta-lactams, aminoglycosides, gly- copeptides, sulfonamides, and fluoroquinolones. B. Tissue Penetration Some drugs, such as aminoglycosides, macrolides, and fluoroquinolones bind ex- tensively to certain tissue components. Intracellular accumulation of aminoglyco- side is slow, however, because of its poor membrane permeability. Intracellular aminoglycoside concentrations tend to be low after initial drug exposure but are high after more sustained exposure and multiple dosing. However, the drug is mi- crobiologically inactive in an acidic environment, such as in the phagolysosome. Indeed, the high intracellular aminoglycoside concentrations achieved in the renal cortex after multiple doses may be the cause of their nephrotoxicity. C. Pharmacodynamics Pharmacodynamics refers to the action of drugs or the biological effects resulting from the interaction of a drug and its receptor site. Pharmacodynamics describes the antimicrobial effect at the site of infection as well as toxic effects in relation to the concentrations of the antimicrobial drug during the course of drug therapy.

164 Rajagopalan et al. For drugs with concentration-dependent bactericidal activity, such as aminoglycosides and fluoroquinolones, the rate and extent of bactericidal action increase with increasing drug concentrations above the minimum bactericidal concentration (MBC) up to a point of maximum effect, usually 5 to 10 times the MBC. In addition, the duration of the PAE is concentration-dependent with these drugs, with longer PAEs induced by higher drug concentrations. In contrast, the bactericidal activity of most beta-lactam antibiotics against gram-negative bacilli is relatively slow and continues as long as the concentra- tions are in excess of the MBC. It does not increase as the drug concentration is increased, that is the bactericidal action of beta-lactams is time-dependent and not concentration-dependent. For time-dependent agents that exhibit short or no post- antibiotic intervals—such as extended-spectrum beta-lactams effective against most gram-negative bacilli—multiple, small, frequent doses or continuous intra- venous infusion produces similar or superior bactericidal effects compared with infrequently administered larger doses. IV. POTENTIALLY USEFUL AND SAFE ANTIMICROBIAL AGENTS FOR LTCF RESIDENTS The vast majority of common bacterial illnesses in LTCF residents respond promptly to broad-spectrum oral antibiotics, but parenteral therapy is occasionally necessary for more severe infections. Some LTCFs have the capacity to provide parenteral therapies. There are antibiotics that may be administered via the intra- muscular route, for example select third-generation cephalosporins such as ceftri- axone that, when administered intramuscularly, demonstrate similar efficacy compared with the intravenous injection. In addition several antibiotics, such as quinolones, have oral formulations that achieve systemic concentrations compa- rable to a parenteral route (22). Such advances should mitigate the need for trans- fer of LTCF residents to an acute care facility for mild to moderate or uncompli- cated infections (23). A. Aminoglycosides Aminoglycosides in the elderly must be prescribed with caution because of the well-described risks of enhanced ototoxicity and nephrotoxicity associated with these agents and the availability of safer and less toxic drugs with comparable spectra (i.e., cephalosporins, monobactams, carbapenems, beta-lactam/beta-lacta- mase inhibitor combination antibiotics, and quinolones). However, these agents are rapidly bactericidal against staphylococci and gram-negative aerobic bacteria, including Pseudomonas sp, and often provide synergy with other agents (e.g., beta-lactams) for treatment of serious or life-threatening infections such as ente- rococcal endocarditis (24). Renal impairment (generally reversible) and ototoxic-

Antimicrobial Therapy 165 ity (generally irreversible) are the two most common and important potential ad- verse effects of these antibiotics (25). Because plasma half-life is increased in pa- tients with decreased renal function (most elderly persons), the dose should be re- duced on the basis of the creatinine clearance. Nephrotoxicity is less likely with once-daily dosing compared with the conventional every 8-hour dosing and is usually reversible (26). Nephrotoxicity, however, may lead to high serum levels of aminoglycosides, which can cause irreversible ototoxicity. Risk of ototoxicity increases with age and is highest in patients with pre-existing hearing deficiencies. Thus, aminoglycoside use in older LTCF residents should be reserved for those with serious or life-threatening infections that require hospitalization and are caused by pathogens susceptible to aminoglycosides (27). B. Beta-Lactams Select beta-lactam antibiotics (penicillins, cephalosporins, carbapenems, mono- bactams, and beta-lactamase inhibitors) may be useful in the management of in- fections in LTCFs because of their broad spectrum, favorable pharmacokinetics and favorable safety profiles. These would include parenteral cefotetan, ceftriax- one, cefoperazone, and cefipime, as well as oral agents such as penicillin, di- cloxacillin, amoxicillin, amoxicillin-clavulanate, cephalexin, cefuroxime axetil, and cefixime (27). C. Macrolides Erythromycin, clarithromycin, and azithromycin have a limited role in the man- agement of infections in the elderly in general. Clarithromycin and azithromycin have more favorable dosing regimens, improved antimicrobial activity, and lower gastrointestinal intolerance compared with erythromycin. These agents are mod- erately active against most strains of streptococci, methicillin-sensitive Staphylo- coccus aureus, anaerobes, Moraxella catarrhalis, Haemophilus influenzae, Le- gionella, Mycoplasma pneumoniae, and Chlamydia pneumoniae, as well as atypical mycobacteria such as Mycobacterium avium complex. Limited data are available regarding pharmacokinetics of these newer macrolides in elderly per- sons; decrease in drug clearance has been attributed to reduce renal clearance. The indications for the newer macrolides in elderly LTCF residents are no different than for the general population. Although macrolides have been recommended in community-acquired pneumonia treated in an ambulatory setting (28), their role as therapy for nursing home-acquired pneumonia is unclear (29). D. Clindamycin This agent is commonly used for anaerobic and staphylococcal infections and for life-threatening group A beta-hemolytic streptococcal infections (streptococ-

166 Rajagopalan et al. cal toxic shock syndrome, necrotizing fasciitis), the latter necessitating acute care facility transfer. Residents of LTCFs are particularly susceptible to antibi- otic-associated colitis caused by C. difficile; clindamycin use is a relatively com- mon association. The drug has utility in the LTCF for treating mild to moderate infections such as skin and soft tissue infections, including infected pressure ul- cers, oral and dental infections, and respiratory tract infections caused by sus- ceptible bacteria. E. Fluoroquinolones The fluoroquinolones are a group of synthetic antibiotics that have a broad spec- trum of antimicrobial activity, good absorption from the gastrointestinal tract, a unique mechanism of action (inhibition of bacterial topoisomerases), favorable pharmacokinetic properties, and a good safety profile (30). As a group, the fluoroquinolones have excellent in vitro activity against a wide range of gram-positive bacteria and many gram-negative bacteria such as Enterobacteriaceae and Aeromonas, Brucella, Campylobacter, Haemophilus, Le- gionella, Moraxella, Neisseria, and Vibrio. These agents are active against Pseu- domonas aeruginosa but are significantly less active against other pseudomonal species, including P. capacia and P. fluorescens. Ciprofloxacin is the most active quinolone against P. aeruginosa. The newer generation fluoroquinolones have ac- tivity against gram-positive bacteria including Streptococcus pneumoniae and staphylococcal species, i.e., methicillin-sensitive S. aureus (MSSA) and coagu- lase-negative species. The fluoroquinolones have less activity against streptococ- cal species and enterococci. These agents in general have very poor activity against anaerobes and No- cardia organisms. Ciprofloxacin and ofloxacin are active in vitro against Chlamy- dia trachomatis, C. pneumoniae, Mycoplasma hominis, and M. pneumoniae. Ciprofloxacin and ofloxacin are active against many species of Mycobacterium, in- cluding M. tuberculosis, M. kansasii, M. fortuitum, and M. xenopi. Cipro-floxacin initially was introduced to North America in 1987. Since its release, it has been widely used in LTCFs (22). These agents are used because they allow the conve- nience of oral therapy with an agent with good bioavailability, are easily adminis- tered by once- or twice-daily dosing, are perceived to be safe, and have wide spec- trum of activity. In the elderly, quinolones are useful in the treatment of complicated urinary tract infections, bacterial prostatitis, skin and soft tissue in- fections, pneumonia, malignant external otitis, and bacterial diarrhea caused by susceptible pathogens (22). The newer generation fluoroquinolones (e.g., levo- floxacin, gatifloxacin, moxifloxacin) with improved gram-positive (including S. pneumoniae) activity over that of the older agents in this class are now considered agents of choice for the treatment of community-acquired pneumonia in adults, in- cluding the elderly (28). Adverse effects of quinolones in the elderly occur in 5%

Antimicrobial Therapy 167 to 15% of cases, including gastrointestinal (nausea, vomiting, diarrhea) and central nervous system (dizziness, headache, insomnia) effects. Associated drug interac- tions with other medications include decreased theophylline clearance associated with increased serum levels of ciprofloxacin, but not norfloxacin or levofloxacin, and multivalent ions (e.g., calcium, iron, aluminum) contained in foods or drugs that significantly reduce absorption of quinolones from the upper gastrointestinal tract. With the intense quinolone use in many LTCFs, quinolone resistance of or- ganisms has increased. Resistance via mutations in the genes encoding topoiso- merase II and IV along with increased drug efflux is common in clinical isolates. Quinolone resistance (methicillin-resistant S. aureus (MRSA), Enterococcus fae- calis, S. pneumoniae, and P. aeruginosa) complicates management of infections by requiring parenteral therapy with other antibiotics for organisms resistant to these oral agents, as well as increasing the burden of resistant organisms (31). Hence, the appropriate use of quinolones in LTCFs must be periodically assessed. F. Trimethoprim-Sulfamethoxazole This antibiotic is commonly prescribed in the elderly, for urinary tract infections, chronic bacterial prostatitis, lower respiratory tract infections, and bacterial diar- rhea caused by susceptible pathogens. Data are limited on the pharmacokinetics of this drug in elderly persons (32). Oral drug absorption does not appear to be af- fected by age. Renal clearance of trimethoprim is decreased in older persons. The recommended doses for use in the elderly are comparable to those prescribed in younger persons: with renal impairment and a creatinine clearance of less than 30 mL/min but greater than 15 mL/min, the dosage is reduced by half. The drug should be avoided if the creatinine clearance is less than 15 mL/min. G. Miscellaneous Antibiotics Other antibiotics that could be prescribed in residents of LTCFs and deserve brief mention include vancomycin, the new FDA-approved antibiotics quinupristin ϩ dalfopristin (Synercid®) and linezolid (Zyvox®), and metronidazole. Vancomycin is a glycopeptide antibiotic used primarily for gram-positive bacterial infections. It is highly active against staphylococci (including MRSA) and streptococci (includ- ing vancomycin-sensitive enterococci). In the elderly, studies have indicated that reduced clearance of vancomycin is a consequence of reduced systemic and renal clearance as well as enhanced tissue binding of the drug. Lower parenteral doses are recommended for the frail elderly, and the dose should be adjusted according to the serum peak and trough levels as well as the creatinine clearance (33). The side effect profile in the elderly is no different from that in the general population. Quinupristin-dalfopristin, which is a streptogramin, is indicated in adults, including the elderly, for the treatment of serious and life-threatening or bac-

168 Rajagopalan et al. teremic infection with vancomycin-resistant enterococci (VRE) and complicated skin and skin structure infection with MSSA and Streptococcus pyogenes (34). The pharmacokinetics of this agent are similar to that in younger adults. Linezolid, an oxazolidinone, is active against infections caused by sensitive gram-positive bacteria as well as MRSA and VRE (35). This agent’s availability, both in par- enteral and oral formulations, as well as its relatively safe profile, is particularly advantageous in management of infections caused by such gram-positive-resistant organisms commonly encountered in elderly LTCF residents (see Section III). The indications, doses, and toxicities of metronidazole are no different in the elderly. The drug is commonly prescribed in combination with agents active against aerobic gram-positive and gram-negative bacteria to treat mixed infections in the elderly, such as infected pressure ulcers, diabetic foot ulcers, intra-abdom- inal infections, or pyogenic brain abcesses (27). Oral metronidazole is the agent of choice for the treatment of C. difficile colitis. Administered orally, this agent is absorbed well and achieves excellent tissue levels. Gastrointestinal intolerance is a common side effect. H. Antituberculous Agents Because most tuberculosis cases in the elderly are caused by isoniazid-sensitive and rifampin-sensitive Mycobacterium tuberculosis, the primary drugs for the treatment of active tuberculosis disease in this age group are isoniazid and ri- fampin. Isoniazid also should be used for the treatment of latent tuberculosis in- fection when the appropriate indications are present (see Chapter 15). I. Antifungal Agents Similar to younger adults, the commonly prescribed systemic antifungal agents in the elderly include amphotericin B, fluconazole, and itraconazole. Because of the potential toxicity of amphotericin B to renal function in the elderly, this agent must be used with caution. Fluconazole, because of its relative safety and efficacy and excellent bioavailability when administered by parenteral and oral routes, is prescribed more often in aging individuals. Itraconazole, available by parenteral and oral formulations, is an acceptable alternative, when indicated (see chapter 25). J. Antiviral Agents The antiviral agents commonly prescribed in the elderly include amantadine, ri- mantadine, acyclovir, valacyclovir, and famciclovir. Amantadine and rimantadine are recommended for influenza A infection within 48 hours of illness onset in the ambulatory elderly to reduce the duration and severity of illness; in institutional-

Antimicrobial Therapy 169 ized elderly, these drugs are recommended for prophylaxis during an influenza A outbreak within the institution (see Chapter 13). Both drugs are continued for a minimum of 2 weeks or until approximately 1 week after the end of the outbreak. The neurominidase inhibitors, zanamivir and oseltamivir, are available for use in influenza A and B infections; efficacy and safety in elderly patients have not been extensively studied. Acyclovir, valacyclovir, and famiciclovir are effective agents for the treatment of herpes simplex and herpes zoster infection. Pain from herpes zoster and chronic pain (postherpetic neuralgia) may be relatively diminished by administering these agents within the first 72 hours of the onset of illness (see Chapter 17). REFERENCES 1. Haley RW, Culver DH, White JW. The nationwide nosocomial infection rate: A new need for vital statistics. Am J Epidemiol 1985; 121:159. 2. Yoshikawa TT. VRE, MRSA, PRP, and DRGNB in LTCF: Lessons to be learned from the alphabet. J Am Geriatr Soc 1998; 46:241–243. 3. Katz PR, Beam TR Jr, Brand F, Boyce K. Antibiotic use in the nursing home physi- cian practice patterns. Arch Intern Med 1990; 150:1465–1468. 4. Warren JW, Palumbo FB, Fitterman L, Speedie SM. Incidence and characteristics of antibiotic use and aged nursing home patients. J Am Geriatr Soc 1991; 39:963–972. 5. Lee YL, Thrupp LD, Friis RH. Nosocomial infections and antibiotic utilization in geriatric patients: A pilot prospective surveillance program in skilled nursing facili- ties. Gerontology 1992; 38:223–232. 6. Zimmer JG, Bentley DW, Valenti WM, Watson NM. Systemic antibiotic use in nurs- ing homes: A quality assessment. J Am Geriatr Soc 1986; 34:703–710. 7. Pickering TD, Gurwitz JH, Zalenznik D, Noonan JP, Avorn J. The appropriateness of oral fluoroquinolone-prescribing in the long-term care setting. J Am Geriatr Soc 1994; 42:28–32. 8. Miller SW, Warnock R, Marshall LL. Appropriateness of antibiotic prescribing for urinary tract infections in long-term care facilities. Consult Pharm 1991; 14:157–177. 9. Hanlon JT, Schmader KE, Samsa GP. A method for assessing drug therapy appropri- ate. J Clin Epidemiol 1992; 45:1045–1051. 10. Norman DC. Fever and aging. Infect Dis Clin Pract 1998; 7:387–390. 11. Yoshikawa TT, Norman DC. Fever in the elderly. Infect Med 1998; 15:704–706. 12. Loeb M, Bentley DW, Bradley S, Crossley K, Garibaldi R, Gantz N, McGeer A, Muder RR, Mylotte J, Nicolle LE, Nurse B, Paton S, Simor AE, Smith P, Strausbaugh LJ. Development of minimum criteria for the initiation of antibiotics in residents of long-term-care facilities: Results of a consensus conference. Infect Control Hosp Epi- demiol 2001; 22:120–124. 13. Nicolle LE, Bentley DW, Garibaldi R, Neuhaus EG, Smith PW, the SHEA Long- Term Care Committee. Antimicrobial use in long-term-care facilities. Infect Control Hosp Epidemiol 2000; 21:537–545.

170 Rajagopalan et al. 14. Crossley K, Henry K, Irvine P, Willenbring K. Antibiotic use in nursing homes: Prevalence, cost, and utilization review. Bull NY Acad Med 1987; 63:510–518. 15. Gurwitz JH, Field TS, Avorn J, McCormack D, Jain Shailavi, Eckler M, Edmonson AC, Bates DW. Incidence and preventability of adverse drug events in nursing homes. Am J Med 2000; 109:87–94. 16. Gurwitz JH, Sanchez-Cross MT, Eckler MA, Matulis J. The epidemiology of adverse and unexpected events in the long-term care setting. J Am Geriatr Soc 1994; 42:33– 38. 17. Rho JP, Yoshikawa TT. The cost of inappropriate use of anti-infective agents in older patients. Drugs Aging 1995; 6:263–267. 18. Mylotte JM. Antimicrobial prescribing in long-term care facilities: Prospective eval- uation of potential antimicrobial use and cost indicators. Am J Infect Control 1999; 27:10–19. 19. Myers BR, Wilkinson P. Clinical pharmacokinetics of antibacterial drugs in the el- derly: Implications for selection and dosage. Clin Pharmacokinet 1989; 17:385–395. 20. Kinirons MT, Crome P. Clinical pharmacokinetic considerations in the elderly: An update. Clin Pharmacokinet 1997; 33:302–312. 21. Bennett W. Geriatric pharmacokinetics and the kidney. Am J Kidney Dis 1990; 26:283–288. 22. Guay DRP. Quinolones. In: Yoshikawa TT, Norman DC, eds. Antimicrobial Therapy in the Elderly Patient. New York, Marcel Dekker, Inc., 1994:237–310. 23. Ernest ME, Ernst EJ. Effectively treating common infection in residents of long-term care facilities. Pharmacotherapy 1999; 19:1026–1035. 24. Bouza E, Munoz P. Monotherapy versus combination therapy for bacterial infections. Med Clin North Am 2000; 84:1357–1389. 25. Zaske DE. Aminoglycides. In: Yoshikawa TT, Norman DC, eds. Antimicrobial Ther- apy in the Elderly Patient. New York, Marcel Dekker, Inc., 1994:183–235. 26. Dew RB III, Susla GM. Once-daily aminoglycide treatment. Infect Dis Clin Pract 1996; 5:12–24. 27. Rajagopalan S, Yoshikawa TT. Antimicrobial therapy in the elderly. Med Clin North Am 2001; 85:133–147. 28. Marrie TJ. Community-acquired pneumonia in the elderly. Clin Infect Dis 2000; 31:1066–1078. 29. Naughton BJ, Mylotte JM. Treatment guidelines for nursing-home acquired pneu- monia based on community practice. J Am Geriatr Soc 2000; 48:82–88. 30. Owens RC Jr, Ambrose PG. Clinical use of fluoroquinolones. Med Clin North Am 2000; 84:1447–1469. 31. Strausbaugh LJ, Crossley KB, Nurse BA, Thrupp LD, the SHEA Long-Term Care Committee. Antimicrobial resistance in long-term-care facilities. Infect Control Hosp Epidemiol 1996; 17:129–140. 32. Williams L, Bender BS. Trimethoprim-sulfamethoxazole. In: Yoshikawa TT, Nor- man DC, eds. Antimicrobial Therapy in the Elderly Patient. New York, Marcel Dekker, Inc., 1994:169–181. 33. Yoshikawa TT. Antimicrobial therapy for the elderly patient. J Am Geriatr Soc 1990; 38:1353–1372.

Antimicrobial Therapy 171 34. Nichols RL, Graham DR, Barriere SL, and the Synercid Skin and Skin Structure In- fection Group. Treatment of hospitalized patients with complicated gram-positive skin and skin structure infections: Two randomized, multicenter studies of quin- upristin/dalfopristin versus cefazolin, oxacillin or vancomycin. J Antimicrob Chemother 1999; 44:263–273. 35. Chien JW, Kucia ML, Salata RA. Use of linezolid, an oxazolidinone, in the treatment of multidrug-resistant gram-positive bacterial infections. Clin Infect Dis 2000; 30:146–151.



12 Urinary Tract Infection Lindsay E. Nicolle Health Sciences Centre, University of Manitoba, Winnipeg, Manitoba, Canada I. INTRODUCTION The most common infection that occurs in elderly residents of long-term care fa- cilities (LTCFs) is urinary tract infection (1). It is the most frequent source of bac- teremia, and a common reason for transfer of residents to acute care facilities. Uri- nary infection is also one of the most common indications for antimicrobial therapy in these facilities, and much of the antimicrobial use for urinary infection in LTCFs is inappropriate (2). Thus, an understanding of urinary infection in res- idents of LTCFs is important for optimal resident care and to promote appropriate antimicrobial use in this setting. The term urinary tract infection simply means the presence of a microbial pathogen within the normally sterile urinary tract. However, it is generally used in the context of isolation of organisms in the urine at a quantitative level that ex- cludes contamination (3). Urinary infection may be asymptomatic—also called asymptomatic bacteriuria—when microorganisms are present in the urinary tract but there are no symptoms or signs referable to urinary infection in the host (4). Individuals with asymptomatic infection usually have evidence of an inflamma- tory or immune host response in the urinary tract. The term “colonization” is sometimes used, rather than asymptomatic bacteriuria. However, this term does not have any clinical relevance and is not used in this chapter. An important group of individuals with urinary infection in LTCFs are those patients with bladder drainage who are using chronic indwelling catheters (5). The epidemiology of infection, including morbidity, differs for residents with long- term use of catheters compared with elderly individuals with urinary infection 173

174 Nicolle without long-term indwelling catheters. Thus, patients in LTCFs with long-term indwelling catheters are generally discussed separately. Observations herein should be considered relevant only for residents without indwelling catheters, un- less stated otherwise. II. EPIDEMIOLOGY AND CLINICAL RELEVANCE A. Prevalence and Incidence The prevalence of urinary infection in elderly residents of LTCFs is high (Table 1). Approximately 30% to 50% of women have positive urine cultures at any time. The prevalence in men is only slightly lower, at 20% to 40%. This remarkable prevalence has been consistently reported from different countries over many years (6–12). The incidence of both symptomatic and asymptomatic urinary infection is also high in these populations (Table 2). In prospective studies of nursing home- acquired infections, symptomatic urinary infection is reported as the first or sec- ond most frequent infection, with an incidence of 1.07 to 1.9/1,000 resident days (Table 2). The definitions used for symptomatic urinary infection, however, lack specificity and may overestimate urinary infection. Studies that have used more restrictive definitions report rates of symptomatic infection of 0.14/1,000 days (20), or 0.5/10,000 days for infection with fever (21). Similarly, symptomatic in- Table 1 Prevalence of Urinary Infection in Long-Term Care Facility Populations Female Male Population (reference) Number % positive Number % positive Nursing home, USA (6) 158 57 56 25 incontinent, mean age 85 yr 19 231 27 121 13 Nursing home, Greece (7) 37 mean age 78 yr 178 15 79 Psychiatric long-term care, 59 Denmark (8) mean age 78 yr 101 29 Veteran’s, Canada (9) mean age 80 yr 101 53 Nursing homes, USA (10) 160 18–33 incontinent, mean age 88 yr Long-term care, Canada (11) mean age 83 yr Nursing home, USA (12) mean age 83 yr

Urinary Tract Infection 175 Table 2 Incidence of Urinary Infection Reported in Long-Term Care Facility Populations Institution Asymptomatic Symptomatic Men 1.23 1.37 Veteran’s, Canada (9) 3.52 1.17 Veteran’s, Tennessee (14) 1.07 Veteran’s, Portland (17) 1.41 Women Skilled nursing facility, Canada (11) 1.9 1.26 Both/Not Stated 1.6 Veteran’s (13) 1.2–1.3 Nursing homes, San Diego (15) 2.4 Nursing homes, Wisconsin (16) Nursing homes, Maryland (18) Skilled nursing facility, NY (19) Episodes/1000 patient days. fection in women is reported to occur at a rate of 1.41/1,000 days with inclusive criteria for identification of urinary infection, but only 0.22/1,000 days with more restrictive criteria that require genitourinary symptoms (11), and in men, 1.37/1,000 days with inclusive and only 0.38/1,000 days with more restrictive cri- teria. The frequency of asymptomatic bacteriuria has also been characterized through the “turnover” of bacteriuria with repeated prevalence surveys in the same population. One study (9) reported an initial prevalence of bacteriuria of 19% for men and 27% of women in a Greek home for the aged. At 1 year, 11% of men and 23% of women with initially negative urine cultures had developed positive cul- tures; 22% and 27% with initially positive urine cultures had become negative. Another study (12) reported an initial prevalence of bacteriuria of 25% in women, with 8% of residents with negative urine cultures becoming positive every 6 months, and 31% of residents with initial positive urine cultures becoming nega- tive. A third study (22) reported an initial prevalence of bacteriuria of 15% in a group of elderly women residents in both community housing and long-term care. The monthly probability of transition from positive to negative cultures was 0.30, and 0.12 from negative to positive. In another study in elderly institutionalized men, 10% of all nonbacteriuric residents became bacteriuric in a 3-month period (9). Some residents have persistent bacteriuria, whereas others have acquisition of new organisms or clearing of bacteriuria. Thus, bacteriuria within a nursing home population is dynamic. Factors that contribute to the variation, including antimi- crobial use, are not well studied.

176 Nicolle B. Risk Factors Urinary infection in residents of LTCFs is predictably associated with certain as- pects of functional status. Residents with cognitive impairment, or who are in- continent of urine or bowel, are likely to have bacteriuria (23,24). In one study, 78% of men with persistent bacteriuria had dementia compared with 42% of those without bacteriuria, and 96% and 25%, respectively, were incontinent of urine (23). Impaired mobility and more prolonged duration of admission to the LTCF are also associated with bacteriuria in some studies (24,25). No association be- tween specific medication use and urinary infection has been reported (26). Risk factors for symptomatic and asymptomatic infection appear to be similar (27). The most important determinant of bacteriuria in the long-term care popu- lation appears to be the presence of a neurogenic bladder. Chronic neurologic dis- eases, such as Alzheimer’s, Parkinson’s, or cerebrovascular disease, are often ac- companied by a neurogenic bladder. These illnesses frequently lead to institutio- nalization and are associated with cognitive impairment and incontinence of blad- der and bowel. A neurogenic bladder results in incomplete voiding and increased likelihood of ureteral reflux, promoting both acquisition and persistence of infec- tion. Drainage devices used to manage incontinence may also increase the fre- quency of urinary infection. The use of an external condom catheter for men with incontinence is associated with at least twice the occurrence of urinary infection compared with men who are incontinent but who do not use external condom catheters (28). Previous indwelling catheter use may have initiated bacteriuria and prostatic infection. Once prostatic infection is established, it often cannot be erad- icated and may be a source for bacteriuria causing relapsing symptomatic or asymptomatic infection (29). Physiologic aging changes also have been associated with urinary infection in well, community-living elderly populations and may contribute to urinary in- fection in the nursing home setting. Women with prior genitourinary surgery or who have cystoceles are more likely to have recurrent urinary infection (30). The use of topical vaginal estrogen decreases the occurrence of both symptomatic and asymptomatic infection (31), but the extent to which estrogen deficiency inde- pendently contributes to urinary infection in this population is not established. Prostatic hypertrophy is a uniform accompaniment of aging in men. This may lead to urethral obstruction and urinary retention requiring instrumentation, and pro- motes turbulent urethral urine flow, which facilitates ascension of organisms into the bladder (29). Thus, a variety of factors contribute to the high frequency of uri- nary infection in elderly residents of LTCFs. Different influences will, of course, have different importance depending on the individual resident. C. Microbiology The diversity of infecting organisms isolated is greater in the LTCF than from uri- nary infection in community populations (Table 3). The Enterobacteriaceae are

Urinary Tract Infection 177 Table 3 Distribution of Infecting Organisms Isolated in Surveys of Bacteriuria in Nursing Home Residents Number of isolates (%) reported in studies (references) indicated Organisms 9* 11† 14* 19‡ Escherichia coli 13 (11) 30 (53) 38 (12) 67 (50) Klebsiella spp 7 (5.9) 26 (8.3) 20 (15) Enterobacter spp 2 (1.7) 14 (25) 5 (3.8) Citrobacter spp 3 (2.5) 3 (1.0) Serratia spp 11 (19) 12 (3.9) 1 (0.8) Proteus mirabilis 36 (30) 1 (1.8) 20 (6.4) Providencia stuartii 5 (4.2) 1 (1.8) 57 (18) 22 (17) Morganella morganii 3 (2.5) Providencia spp 14 (11.8) 60 (19) 5 (3.8) Pseudomonas aeruginosa 23 (19) 34 (11) Other gram-negative 6 (4.5) Enterococcuss spp 6 (5.0) 2 (0.6) 6 (4.5) Coagulase-negative staphylococci 2 (1.7) 34 (11) 1 (0.8) Staphylococcus aureus 3 (2.5) 2 (0.6) Other gram-positive 2 (1.7) 2 (0.6) Candida spp 3 (1.0) * male † female ‡ 80% female the most common infecting organisms for both symptomatic and asymptomatic infection. Escherichia coli is the most frequent bacteria isolated in women, with Proteus mirabilis usually second. For men, E. coli and P. mirabilis occur with equal frequency, or P. mirabilis is more common. Other Enterobacteriaceae iso- lated include Klebsiella pneumoniae and urease-producing organisms such as Providencia stuartii and Morganella morganii. Providencia stuartii has a unique predilection for the institutionalized population (32). Pseudomonas aeruginosa and gram-positive organisms including Entero- coccus spp and coagulase-negative staphylococci, are also common. Group B streptococci and Staphylococcus aureus are less frequent but are identified in some patients in most series. Yeast infection, principally caused by Candida albi- cans, may occur but is uncommon. The determinants of candiduria in LTCF resi- dents are not well described. Yeast infection may be more common in women with associated vulvovaginal candidiasis and in diabetic patients. The distribution of organisms isolated from symptomatic infection is similar to that for asymptomatic infection, although coagulase-negative staphylococci are uncommon in symp- tomatic infection.

178 Nicolle Polymicrobial infection is present in 10% to 25% of bacteriuric residents (9,11). Men with external condom catheters used for voiding management often have infection with more than one organism (33). Bacterial isolates from urinary infection in LTCF residents are also characterized by increased antimicrobial re- sistance (34,35). This is a result of the intense use of antimicrobials in nursing homes (2), as well as the transmission of organisms between residents in the in- stitutional setting. D. Host Response Asymptomatic bacteriuria is not simply the presence of bacteria in the bladder. In fact, at least 50% of women with asymptomatic infection have bacteria localized to the kidneys (36,37). Pyuria is present in more than 90% of bacteriuric subjects (38,39). The level of pyuria does not correlate with the presence of symptomatic infection (40). Increased urinary cytokine levels and increased local urinary or systemic antibodies to the infecting organism are further evidence of a host re- sponse with asymptomatic infection (41,42). Symptomatic subjects uniformly have pyuria. There are also elevated levels of urinary cytokines and local urinary antibody (41,42). With resolution of the symptomatic episode, urinary antibodies may decrease, particularly with E. coli infection (42). In clinical presentations with systemic manifestations such as fever, an elevated C-reactive protein is usually present, and there is an increase in systemic antibody to the infecting organism (43). E. Clinical Impact The majority of urinary infections in residents of LTCFs are clinically asymp- tomatic (4). Persistent asymptomatic bacteriuria has not been associated with neg- ative long-term outcomes, such as renal failure or hypertension. Despite the very high prevalence of urease-producing organisms, including P. mirabilis and P. stu- artii, renal or bladder stones have not been identified as a significant clinical prob- lem in LTCF residents without chronic indwelling catheters (9). Asymptomatic bacteriuria also does not affect survival (23,40). Where an association between de- creased survival and bacteriuria in residents has been observed, bacteriuria is not an independent association of mortality (24). Episodes of febrile urinary infection have been reported with a frequency of 1 to 1.5/10,000 resident days, comprising about one tenth of episodes of fever of any cause in this population (21). The urinary tract is the most common source of bacteremia in LTCF residents (44,45), although this occurs primarily in residents with an indwelling urinary catheter (46). It is also a reason for transfer of LTCF residents for acute hospitalization (47), contributing 8% of transfers at one facil-

Urinary Tract Infection 179 ity (48). Despite this, urinary infection is infrequently identified as a direct cause of death in residents (1,20). Urinary infection is one of the most common indications for antimicrobial prescriptions in LTCFs. From 20% to 60% of systemic antimicrobial courses are given for treatment of urinary infection (2). This intensive use of antimicrobials contributes to the emergence and persistence of antimicrobial resistance in the long-term care setting. F. Chronic Indwelling Catheter From 5% to 10% of residents of LTCFs have voiding managed with a long-term indwelling urethral catheter (49,50). The daily incidence of new infection for in- dividuals with a long-term indwelling catheter is similar to that reported with short-term catheters, about 3% to 7% per day (5). Thus, anyone with a long-term indwelling catheter in place for longer than 30 days will be bacteriuric, and at any time the prevalence of bacteriuria in individuals with chronic catheters approaches 100%. Residents with long-term catheters are infected with a complex bacterial flora with two to five organisms present at any time (51). Proteus mirabilis, Pro- teus spp, Morganella morganii and P. aeruginosa are the most common organ- isms isolated, although many other gram-negative and gram-positive organisms also occur. Biofilm formation occurs on the catheter, primarily on the interior sur- face (52). This material consists of bacteria, extracellular bacterial substance, Mg2ϩ, Ca2ϩ, and Tamm-Horsfall protein from urine. It may also contain struvite if infection with a urease-producing organism is present. Biofilm contributes to catheter encrustation and obstruction (53). Bacteria growing in the biofilm are in an environment where they are relatively protected from the effect of host de- fenses or antimicrobials. Residents of LTCFs with chronic indwelling catheters experience in- creased morbidity attributable to urinary infection compared with LTCFs resi- dents with bacteriuria who do not have an indwelling catheter. Febrile urinary infection is 10 times more frequent in these individuals (21,54), and bacteremia, primarily from a urinary source, is 40 times more frequent (46). A chronic in- dwelling catheter may also cause episodes of gross hematuria resulting from catheter trauma (55), and local suppurative complications including paraurethral abscesses, urethritis, epididymo-orchitis, or prostatic abscesses. Residents with indwelling catheters have a higher mortality than LTCF residents without chronic indwelling catheters, but this is attributable to underlying patient factors rather than urinary infection (50). At autopsy, residents with a long-term in- dwelling catheter have a higher frequency of histologic evidence for renal in- flammation consistent with pyelonephritis compared with bacteriuric residents without an indwelling catheter (56,57).

180 Nicolle III. CLINICAL MANIFESTATIONS Urinary infection is usually asymptomatic. However, symptomatic urinary infec- tion is an important contributor to morbidity in LTCF residents. When symp- tomatic infection occurs, the clinical presentation may vary across a spectrum of minor lower tract irritative symptoms to severe systemic symptoms and sepsis re- quiring hospitalization. Potential clinical presentations of symptomatic infection are listed in Table 4. The presentation may be similar to that in younger popula- tions. For acute lower tract infection, or acute cystitis, frequency, dysuria, ur- gency, the new onset or worsening of urinary incontinence, and suprapubic dis- comfort may occur. Acute pyelonephritis may present with costovertebral angle pain and tenderness with fever. The clinical diagnosis, however, is often not straightforward because of chronic symptoms that interfere with assessment, dif- ficulties in communication because of deafness, cognitive impairment, or dysarthria, and blunted clinical manifestations associated with aging, such as a de- creased or absent fever response (58). Sepsis, often with bacteremia, is more fre- quent with obstruction or trauma to the genitourinary tract. Epididymo-orchitis may occur in male residents. Hematuria is seldom caused by urinary infection, but gross hematuria will frequently lead to secondary sepsis in the presence of in- fected urine (55). Chronic genitourinary symptoms are common in this population. These in- clude chronic incontinence, nocturia, and frequency. A high proportion of resi- dents with chronic genitourinary symptoms have positive urine cultures, but Table 4 Presentations of Symptomatic Urinary Tract Infection in Residents of Long- Term Care Facilities Symptomatic Urinary Infection Acute cystitis (frequency, dysuria, suprapubic discomfort) Acute deterioration in continence Acute pyelonephritis (costovertebral angle pain/tenderness; often with fever) Fever with hematuria Fever with no localizing findings (10% due to urinary infection) Epididymo-orchitis Additional Presentations with Indwelling Catheters Fever with catheter obstruction Urethritis Paraurethral abscess Bladder spasms with bypassing of catheter by urine Not Symptomatic Urinary Infection Chronic genitourinary symptoms Cloudy or foul-smelling urine Clinical deterioration without fever or localizing genitourinary symptoms or signs

Urinary Tract Infection 181 chronic symptoms are not attributable to urinary infection and are not improved by antimicrobial treatment of urinary infection (59,60). Thus, chronic genitouri- nary symptoms are not a manifestation of symptomatic urinary infection. How- ever, acute deterioration in symptoms, such as in continence status, may be con- sistent with acute infection. Urinary tract infection may also present as fever without localizing findings. Fever with no apparent source is a relatively common clinical manifestation of many different infections in elderly institutionalized populations. About 50% of these episodes occur in residents with positive urine cultures because of the very high prevalence of bacteriuria in these facilities. However, only 10% of episodes of fever without localizing findings appear to be of urinary origin (20). Unfortu- nately, criteria to differentiate urinary infection from other potential sources of fever if localizing findings are not present have not been identified. Clinical dete- rioration without fever or localizing genitourinary findings is also often attributed to urinary infection. However, in the absence of fever, urinary infection is unlikely to be the cause of a nonspecific decline in resident status (61). A. Long-Term Indwelling Catheters The most common presentation of symptomatic urinary infection in the resident with a long-term indwelling catheter is fever without localizing genitourinary findings (54). Symptoms or signs localized to the genitourinary tract may include costovertebral angle pain or tenderness, suprapubic tenderness, hematuria, or catheter obstruction. Bacteremia may be present (44–46). Lower tract symptoms such as suprapubic tenderness or bypassing of the catheter because of bladder spasms may also occur, but are less common. Local suppurative complications in- cluding epididymo-orchitis, prostatic abscess, paraurethral abscesses, or urethritis occur in men. Bladder or kidney stone formation is a potential long-term compli- cation in residents with infection with urease-producing organisms such as P. mirabilis or P. stuartii. IV. DIAGNOSTIC APPROACH A. Microbiologic Diagnosis A diagnosis of urinary infection requires an appropriately collected and trans- ported urine specimen that is cultured quantitatively. For a diagnosis of asymp- tomatic bacteriuria, two consecutive urine cultures growing 105 or more colony- forming units/ml (cfu/ml) is necessary (3). Symptomatic infection is diagnosed with only one urine culture meeting this quantitative criteria. A lower quantitative count may be consistent with a microbiologic diagnosis of symptomatic infection in some clinical settings. Individuals receiving diuretics, who have renal failure,

182 Nicolle or who are infected with selected fastidious organisms may have quantitative counts lower than 105 cfu/ml. If complete obstruction is present and the infection is proximal to the obstruction, urine cultures may be negative. Urine culture will also usually be negative if antimicrobial therapy is initiated before obtaining the urine specimen. Lower quantitative counts may be consistent with infection with clinical presentations in some other populations, such as acute cystitis in young women, but lower quantitative criteria have not been validated for any presenta- tions of urinary infection in the elderly. As 10% to 25% of bacteriuric men or women in LTCFs have more than one organism isolated (9,11), urine specimens with more than one uropathogen in appropriate quantitative counts should not be dismissed as contamination. Interpretation of quantitative urine bacteriology depends on a urine speci- men collected to minimize contamination with urethral and periurethral flora. For men, a clean-catch specimen collected with voiding is usually adequate and un- likely to have substantial contamination. If a male resident cannot cooperate to provide a voided specimen, collection using a freshly applied clean condom and leg bag may provide a suitable specimen (33,62). For women, a clean-catch tech- nique also usually provides an adequate specimen (63). A clean-catch technique also has been shown to be feasible in incontinent female nursing home residents, provided the resident is cooperative and staff are properly trained (64). However, many women may not be able to cooperate with voiding for specimen collection (11). The use of pedibags, bedpans, or diapers in collecting urine specimens from women is discouraged, as these methods are subject to substantial contamination with periurethral organisms and have not been validated. When a resident is un- able to cooperate and a urine culture is indicated to assist with clinical manage- ment, in and out urethral catheterization for specimen collection should be used. Any quantitative count of a potential uropathogen is diagnostic of infection in a urine specimen collected by catheterization. This procedure, however, may intro- duce infection in as many as 5% of catheterizations and should only be used if there is a compelling clinical indication. B. Urinalysis More than 90% of elderly residents of LTCFs with bacteriuria will have pyuria, regardless of whether infection is symptomatic or asymptomatic (38,39). In ad- dition, 30% of residents without bacteriuria also have pyuria, presumably caused by periurethral contamination or other inflammatory conditions within the geni- tourinary tract. Thus, pyuria is not specific for a diagnosis of bacteriuria or for differentiating symptomatic from asymptomatic infection. However a urinalysis for determination of pyuria is useful if it is negative, as absence of pyuria has a high negative predictive value to exclude bacteriuria. The leukocyte esterase dipstick test has been evaluated for identification of pyuria in elderly institu-

Urinary Tract Infection 183 tionalized populations (6,22,39). It has a positive predictive value varying from 18% to 75% and a negative predictive value of 75% to 100% for identifying in- fection in this setting. Thus, it may be used as a rapid screening test to exclude urinary infection. C. Clinical Diagnosis When a resident presents with a new onset of irritative lower tract symptoms or with classic clinical findings of acute pyelonephritis, a diagnosis of symptomatic urinary infection may be straightforward. However, a diagnosis of symptomatic urinary infection is frequently problematic because of difficulties in communica- tion and the presence of chronic symptoms associated with comorbid disease. The fever response is less marked or may be absent (58). Acute change in chronic symptoms, such as acute deterioration in continence status or increased frequency, may support a clinical diagnosis of symptomatic urinary infection. The resident with clinical deterioration without localizing genitourinary findings may present a diagnostic problem. Because a high proportion of these in- dividuals have a positive urine culture, there is a tendency to attribute any change in clinical status for which no other explanation is apparent to urinary infection. However, urinary infection is not usually the cause for this presentation. One study reported urinary infection caused clinical deterioration without localizing findings in only 11% of such patients, and all those with urinary infection also had fever (61). In another study, fever without localizing findings was caused by uri- nary infection in only 10% of episodes (21). There were, however, no clinical or laboratory parameters to differentiate the 90% of episodes not linked to a urinary source from the 10% attributable to urinary infection. Thus, with a clinical pre- sentation of fever without localizing findings in residents with a positive urine cul- ture, some skepticism should be maintained with respect to a diagnosis of symp- tomatic urinary infection. Practitioners must recognize the uncertainty in attributing this clinical presentation to urinary infection. A negative urine speci- men is useful to exclude urinary infection, but a positive urinary culture does not confirm symptomatic urinary infection. Careful clinical evaluation to identify lo- calizing findings to support or exclude a genitourinary source should always be undertaken. The presence of “foul-smelling” or cloudy urine is also sometimes identi- fied as symptomatic urinary infection and interpreted as an indication for antimi- crobial therapy. Cloudy urine may be caused by crystals as well as pyuria, and even if pyuria is the cause, by itself it is not sufficient to diagnose symptomatic in- fection, nor is it an indication for antimicrobial therapy. An unpleasant urine odor may certainly be associated with urinary infection and is likely caused by polyamine production by infecting bacteria (65). However, not all residents with this problem have urinary infection and not all residents with urinary infection