Urinary tract infection (UTI) is the most frequent bacterial infection in elderly populations. While urinary infection in the elderly person is usually asymptomatic, symptomatic infection occurs frequently with associated serious morbidity and, rarely, mortality. Optimal management of urinary infection in the elderly patient is challenging in the face of diagnostic uncertainty, concerns with excess antimicrobial use, and increasing antimicrobial resistance in both the community and nursing home. Current limitations in knowledge and optimal management strategies for this problem must be appreciated. In addition, the heterogeneity of the elderly populations means approaches may vary for different groups. The impact and management of urinary infection differs for women and men, and for the institutionalized and noninstitutionalized elderly person. There are also unique considerations for the subgroup of institutionalized elderly persons with chronic indwelling catheters. The discussion in this chapter is relevant to individuals without long-term indwelling catheters, unless otherwise stated.

The bladder is normally sterile. While bacteriuria is always abnormal, it is not necessarily detrimental. “Bacteriuria”, a positive urine culture without attributable signs or symptoms, is used interchangeably with “asymptomatic UTI” in this chapter. The majority of the elderly individuals with bacteriuria have evidence for a local host response. Some authors use the term “bladder colonization” in discussing asymptomatic bacteriuria. This term has not been shown to have clinical or biological relevance in elderly populations, and is not used in this discussion.

UTI encompasses a spectrum of presentations (Table 127-1). Acute uncomplicated urinary infection, also known as acute cystitis, is usually considered relevant only for premenopausal women. However, women who experience recurrent acute uncomplicated urinary infection prior to menopause often continue to experience these episodes after menopause. Complicated UTI occurs in either sex in the setting of a structurally or functionally abnormal urinary tract. A wide variety of genitourinary abnormalities are associated with infection, from prostatic hypertrophy and cystoceles to chronic renal failure and chronic indwelling catheters. A characteristic of complicated UTI is frequent and early post-therapy recurrence if the underlying genitourinary abnormality persists. These infections are also characterized by a wider variety of infecting organisms. The majority of functionally impaired older women and all men with urinary infection should be considered to have complicated infection.

Acute prostatitis is an infrequent, severe, febrile infection occurring primarily in young men. Chronic bacterial prostatitis occurs in older men. It is characterized by bacterial infection of the prostate with mild to moderate local symptoms, and is commonly associated with recurrent episodes of acute cystitis as organisms reinfect the urine from the prostate source. These symptoms respond to antimicrobial therapy. Presentations of chronic prostatitis/chronic pelvic pain syndrome in men without documentation of bacteria in the prostate are no longer thought to be attributable to infection, irrespective of the presence or absence of pyuria.

Recurrent urinary infection, either reinfection or relapse, is common for elderly persons. Reinfection is recurrent urinary infection with an organism isolated following antimicrobial therapy which differs from the pretherapy isolate. This is assumed to reflect entry of a new organism into the bladder during or following therapy. Superinfection is a new urinary infection (i.e., a reinfection) that occurs in the presence of existing bacteriuria, without an intervening episode of negative urine culture. Relapse is recurrent urinary infection when the organism-isolated post-therapy is similar to the pretherapy isolate. When relapse occurs, the organism has usually remained sequestered at some site in the urinary tract and was not eradicated with antimicrobial therapy.

Asymptomatic Infection

Elderly in the Community

There is a marked increase in the prevalence of asymptomatic bacteriuria with increasing age for women. The prevalence of bacteriuria is 2% to 3% in young women and increases to more than 10% for women older than age 65 years and 20% at 80 years or more (Table 127-2). Bacteriuria is uncommon in younger men but the prevalence of bacteriuria increases substantially with aging, particularly coincident with the development of prostatic hypertrophy. Approximately 5% of men older than age 70 years living in the community have bacteriuria. There are few studies describing the incidence of asymptomatic urinary infection in the elderly ambulatory populations. With yearly urine cultures obtained, it was reported to be 6.7/100 person years for diabetic and 3.0 for nondiabetic women aged 55 to 75 years. For 209 initially nonbacteriuric elderly ambulatory male outpatients of a veteran’s hospital, 10% had at least one episode of bacteriuria during a mean of 2.8 years follow-up. Three-quarters of these episodes resolved spontaneously.

Institutionalized Older Adults

The prevalence of asymptomatic bacteriuria in institutionalized elderly populations is remarkably high (Table 127-2). This has been consistently observed in reports over several decades from many geographic areas of the world. It is somewhat higher in women, with 25% to 50% bacteriuric, compared to 15% to 40% of men. The variation in prevalence is primarily determined by characteristics of the institutionalized population. There may be an increased prevalence with increasing age of residents, although this is not consistent among studies. In fact, bacteriuria in the very elderly, those older than age 90 years, may be somewhat lower. The incidence of asymptomatic infection amongst these populations is also very high. In a group of previously bacteriuric women screened monthly, new infections occurred at a rate of 87 infections per 100 patient years. In elderly men resident in a veteran’s hospital with urine cultures obtained every 2 weeks, an incidence of infection of 45 infections per 100 patient years was observed, with 10% of previously nonbacteriuric residents becoming bacteriuric in every 3-month period.

The remarkable occurrence of asymptomatic bacteriuria in elderly institutionalized populations has also been described in repeated prevalence surveys, with observations reported as both acquisition and loss of bacteriuria between surveys. Acquisition of bacteriuria at 1 month after initial negative screening was reported to be 11% for men and 12% for women. In another study, 8% of women acquired bacteriuria by 6 months, and 23% by 1 year. Reversion to negative urine cultures was observed in 22% of bacteriuric men at 1 year and, for women, 12% at 1 month, 31% at 6 months, and 27% at 1 year. Resolution of bacteriuria is often coincident with antimicrobial therapy.

Symptomatic Urinary Infection

Elderly in the Community

Clinically diagnosed urinary infection in an American community-living population older than 65 years of age was reported to be 13 per 100 person-years, 10.9 per 100 years in men and 14 per 100 years in women. The rate was 15 per 100 years in those aged 65 to 74 years, but 12 per 100 years in those older than age 75 years. In women 55 to 75 years enrolled in a group health plan in Washington State, symptomatic infection was 7/100 person years; 12/100 for women with diabetes and 6.7/100 for those without diabetes. Symptomatic infections were 0.17 per 1000 days in a cohort of 29 ambulatory elderly male veterans with bacteriuria. In initially bacteriuric residents of a geriatric apartment, the incidence of urinary symptoms was 0.9 per 1000 days during 6 months follow-up. Considering more severe presentations, hospitalization for acute pyelonephritis in persons older than 70 years was 10 to 15 per 10 000 population in one Canadian province, with hospitalization for women 1.3 times more frequent than men. An American study using 1997 administrative data reported pyelonephritis hospitalizations were 13.5/10 000 for women 60 to 79 years and 23.3/10 000 for women 80 years or older, and 6.3/10 000 and 12.9/10 000 for men, respectively.

Institutionalized Older Adults

The incidence of symptomatic urinary infection in long-term care facilities varies from 1.0 to 2.4 per 1000 resident days. Definitions to identify symptomatic urinary infection are, however, often imprecise and variable among studies. The reported incidence of symptomatic infection is only 0.27 to 0.4 per 1000 days for bacteriuric men or women when restrictive clinical definitions which require localizing genitourinary symptoms are used. An incidence of 0.6/1000 days was recently reported from a group of American facilities using standardized definitions, and 1.0/1000 resident days in a German nursing home. The incidence of fever from a urinary source in noncatheterized subjects in nursing homes was reported to be 0.5 to 1 per 1000 resident days using restrictive clinical or serologic criteria for identification of invasive urinary infection.

Asymptomatic Bacteriuria

Long-term adverse outcomes have not been attributed to asymptomatic bacteriuria for either ambulatory or institutionalized populations. Persistent asymptomatic urinary infection is not associated with an increased risk of development of renal failure or hypertension. In addition, despite a high prevalence of infection with urease-producing organisms such as Proteus mirabilis, complications from renal stone disease are not a common problem in institutionalized populations. Studies from Greece and Finland in the 1970s reported an association between asymptomatic bacteriuria and decreased survival for both elderly men and women. Subsequent studies in both community and institutionalized elderly populations in Finland, the United States, and Canada have not confirmed this observation. Thus, current evidence does not support an association between asymptomatic bacteriuria and decreased survival for elderly populations.

Symptomatic Urinary Infection

Symptomatic urinary infection is a common infection in the elderly population. Morbidity may occur along a continuum of limited voiding discomfort to disruption in daily activities or hospitalization for pyelonephritis or sepsis. An American study of 284 patients, older than 65 years of age, presenting to an emergency department reported 4.6% died during hospitalization; 9.5% required intensive care admission; 48.9% had greater than 2 days length of stay and 26.4% required more than 2 days intravenous antibiotics. Independent predictors of any of these adverse outcomes included mental status change, frequent past urinary infections, other non urinary infections, abnormal temperature, tachycardia, hypotension, elevated blood urea nitrogen, hyperglycemia, elevated white blood cell count and relative neutrophilia.

For the institutionalized population, deterioration in functional status may also contribute to morbidity, but this impact has not been rigorously evaluated. From 8% to 30% of transfers from long-term care to an acute care facility are necessitated by urinary infection. Urinary infection is the most common source of bacteremia in long-term care facility residents, although bacteremia is infrequent in residents without a chronic indwelling catheter. Antimicrobial pressure associated with treatment of UTIs, much of which may be asymptomatic, promotes emergence and persistence of resistant organisms in long-term care facilities. Antimicrobial therapy also contributes to morbidity through adverse effects. Urinary infection, however, is an infrequent cause of mortality, even with bacteremia from a urinary source.

Route and Site of Infection

Urinary infection occurs by the ascending route. The reservoir for infecting organisms is usually the gastrointestinal flora. Organisms colonize the periurethral area and ascend up the urethra into the bladder. Infecting organisms may subsequently reach the kidney. Renal infection is determined by virulence characteristics of the infecting organism or the presence of genitourinary abnormalities such as obstruction or reflux in the host. For men, ascending infection may also lead to prostate infection. Thus, urinary infection may be localized to the bladder, may involve the kidneys as well as the bladder and, for men, may also be localized in the prostate. Infection of the upper tract (kidney) is present in 50% of elderly women with asymptomatic infection. Renal localization is more frequent with increasing age, and in residents of nursing homes. The proportion of elderly men with a prostatic site of infection is unknown, but is likely substantial. In the institutionalized population, transfer of organisms which colonize the perineum or urinary drainage devices such as indwelling or external catheters may occur between patients. Rarely, infection may be hematogenous rather than ascending, with urinary infection secondary to bacteremia from a nonurinary source.

Microbiology

Infecting Organisms

The most frequent infecting organism isolated from urinary infection in either asymptomatic or symptomatic ambulatory elderly women is Escherichia coli (Table 127-3). For women with acute uncomplicated infection, the spectrum of virulence factors in urinary E. coli isolates is similar to younger women. Pyelonephritis strains are associated with P. fimbriae, and have an increased frequency of other virulence characteristics including capsule type, adhesins, and proteins such as iron binding proteins and hemolysins. E. coli isolated from persons with complicated urinary infection have a lower prevalence of virulence characteristics. In men, gram-positive organisms and other gram-negative organisms, particularly P. mirabilis, are more frequently isolated than E. coli.

E. coli remains the most frequent organism isolated from institutionalized women, although less common than in noninstitutionalized women. P. mirabilis is the most frequent in men (Table 127-3). E. coli may be more common in bacteremic infection than other gram-negative uropathogens. E. coli is also more likely to persist than other organisms in women with asymptomatic bacteriuria. Many other gram-negative organisms are isolated in the institutionalized population, including Enterobacteriaceae such as Klebsiella pneumoniae, Serratia spp., Citrobacter spp., Enterobacter spp., and Morganella morganii, as well as Pseudomonas aeruginosa. Providencia stuartii is an organism isolated virtually only from institutionalized subjects. When this organism is isolated, it is frequently the predominant organism in the ward or facility. Thus, it has a unique propensity for spread within the institutional setting. Gram-positive organisms include group B streptococci, which may be more common in persons with diabetes. Enterococcus spp. and coagulase-negative staphylococci are frequently isolated in men, but are seldom associated with symptomatic infection. Polymicrobial bacteriuria is common in both men and women resident in institutions, with more than one organism isolated in 10% to 25% of bacteriuric subjects.

Antimicrobial Resistance

Bacterial isolates from urinary infection in institutionalized populations are characterized by a higher frequency of antimicrobial resistance than organisms isolated in the community population. This is a consequence of intense exposure to antimicrobials together with facilitation of transmission of organisms in the institutional setting. The prevalence of resistant isolates varies widely among facilities. Resistant organisms of particular concern in urinary infection include vancomycin-resistant enterococci, extended spectrum beta lactamase producing E. coli and K. pneumoniae, and fluoroquinolone-resistant P. aeruginosa. For the elderly residents with urinary infection who present to emergency departments, prior antimicrobial use with either trimethoprim/sulfamethoxazole (TMP/SMX) or a fluoroquinolone is the most important risk factor for isolation of a resistant organism. Consistent associations for identification of resistant organisms in nursing home residents also include prior TMP/SMX exposure for TMP/SMX-resistant organisms, and prior fluoroquinolone exposure for fluoroquinolone-resistant organisms. The presence of an indwelling urethral catheter is also associated with isolation of fluoroquinolone-resistant E. coli.

Host Factors

Elderly in the Community

In the normal genitourinary tract, intermittent, complete voiding is the preeminent host defense against urinary infection. Most well-elderly women continue to empty their bladder completely. Several host factors, however, contribute to the high frequency of urinary infection in elderly populations (Table 127-4). In ambulatory postmenopausal women associations with asymptomatic bacteriuria are urinary incontinence, increased postvoid residual urine, reduced mobility and estrogen treatment. For postmenopausal women who are diabetic, an additional risk factor is the duration of diabetes. A prospective study of risk factors for symptomatic UTI in women of age 55 to 75 years reported independent predictors were a prior history of urinary infection and diabetes. An initial case control study in the same women had reported sexual activity and incontinence to be associated with symptomatic infection, but these findings were not confirmed in the prospective study.

Any structural or functional abnormality which impairs voiding will increase the likelihood of urinary infection. Genitourinary abnormalities such as urethral or ureteric strictures, bladder diverticulae, and cystoceles occur with increased frequency in older populations, and contribute to bacteriuria. Studies in ambulatory women have tended to confirm an association of increased postvoid residual being associated with UTI. An association between urinary incontinence and UTI is consistently reported from both ambulatory and institutionalized women, but is unlikely to be causative. Women with incontinence have abnormal voiding, and voiding abnormalities are a major predisposing factor for development of bacteriuria or urinary infection. Thus, urologic abnormalities which lead to incontinence are also likely to promote bacteriuria. In multivariate analyses of associations with urinary incontinence, urinary infection frequently does not persist as an independent variable. Current evidence does not support a role for immunologic changes of aging being important contributors to the high frequency of urinary infection.

The association of estrogen decline in elderly women with urinary infection is complex, and observations sometimes conflicting. Loss of the estrogen effect on the genitourinary mucosa in postmenopausal women is associated with an increased colonization of the vagina with potential uropathogens. Postmenopausal women are less likely to have lactobacilli colonizing the vagina, and more likely to have E. coli and enterococci. These changes in vaginal flora, together with the higher pH in the absence of lactobacilli, are similar to changes observed in younger women with recurrent UTI, and have been suggested to facilitate urinary infection in older women. However, both case control studies and prospective, randomized trials, have consistently reported that systemic estrogen use in postmenopausal women is associated with an increased risk of symptomatic infection. It is suggested this is because women who use estrogen are more likely to be sexually active, but this hypothesis requires confirmation. Topical vaginal estrogen maintains a vaginal environment characterized by decreased pH and predominance of lactobacilli. Prospective, randomized clinical trials have reported that vaginal estrogen may decrease the frequency of infection. However, some case control studies report an association of topical vaginal estrogen with increased risk of urinary infection. Again, these studies may be biased by inclusion of women who are receiving topical estrogen because they are experiencing recurrent urinary infection.

Prostate hypertrophy is the most important factor promoting urinary infection in elderly men. This leads to obstruction and turbulent urethral urine flow which facilitates ascension of organisms into the bladder or prostate. Once bacteria are established in the prostate they are difficult to eradicate because of poor diffusion of antimicrobials into the prostate, and the increasing presence of prostate stones with age. These stones provide a nidus from which it may be impossible to eradicate bacteria. Frequent relapse of urinary infection from a prostate source is common, although months or even years may intervene between episodes.

Institutionalized Older Adults