The clinical significance of asymptomatic bacteriuria (AB) in the elderly is not clear. Nevertheless, elderly women have the highest prevalence of this entity which increases with increasing age. In the absence of signs or symptoms, bacteriuria ?10⁵ colony forming units per milliliter (CFU/ml) has come to be synonymous with urinary tract infection (UTI). Few studies of the inflammatory response of the urinary tract in the elderly population have emphasized its relation with bacteriuria. However, the absence of inflammatory response could explain the difficulty in controlling asymptomatic UTI in elderly patients. The measurement of pyuria is available by means of establishing the presence of host injury, thus AB associated with pyuria represents UTI, and AB without pyuria may be a transient colonization state. In the present work, the association between AB and pyuria in elderly ambulatory women was determined.

Materials and Methods

During the 12-month period from January–December 1996, a cross-sectional study of elderly ambulatory women residing in seven nursing homes was undertaken. A woman was considered elderly if she was 65 years or older. All elderly women who were eligible for participation in the study fulfilled the following criteria: be ambulatory; have no symptoms of UTI (incontinence, dysuria, frequency, urgency, suprapubic pain, flank pain); have no indwelling bladder catheter, and be of normal mental state. Normal mental state was evaluated for orientation to time, place, person and recent memory. Those who had received any antimicrobial therapy within the last 30 days and/or if they had a urinary catheter were not included in this study.

Before urine samples were obtained, all of the elderly women were questioned and records from the nursing home were reviewed. Information concerning exposure to antibiotics, immunosuppressive agents, underlying disease (active condition), history of UTI, vaginal symptoms (vaginal discharge, itching, swelling or vulvar burning), and other genitourinary abnormalities were recorded. Two investigators were instructed to obtain specimens. The procedure was carried out using the following method. The urethral meatus was washed with soap and rinsed with sterile water before specimen collection. The first morning voided urine specimens were collected in sterile containers, using a midstream clean-catch technique. Urine specimens were refrigerated immediately after collection. Specimens were transported on ice to the laboratory for urinalysis and culture no longer than 1 h after collection. The urinalyses were performed in a blinded fashion, without knowledge of the urine culture results. The absolute leukocyte count in uncentrifuged urine specimens was determined by using a Neubauer counting chamber (Sigma, St. Louis, MO, USA), and pyuria was defined as ?10 leukocytes in a cubic millimeter of uncentrifuged urine.

Microscopic examination of urinary sediment was made after the centrifugation of urine at 400 × g for 3 min and leucocytary casts were researched. Specimens were cultured by inoculating 0.01 ml of urine on blood agar and eosin-methylene blue agar plates and were incubated at 37°C. These plates were evaluated at 24, 48, and 72 h; the amount of colonies was determined, and organisms were identified using a microplate method (MicroScan, Baxter, West Sacramento, CA, USA). A urine specimen with more than one organism was considered to be contaminated, and a second specimen was obtained. In elderly women with bacteriuria, a positive urine culture was confirmed by a second positive culture with the same organism within 1 week. When two organisms were present in three specimens, the elderly women were considered bacteriuric with two bacterial species.

Asymptomatic bacteriuria in the elderly woman was defined as the absence of symptoms of UTI and bacteriuria. The elderly women with bacteriuria were considered bacteriuric in two categories: with bacteriuria ?10⁵ CFU/ml of urine and with bacteriuria <10⁵ CFU/ml of urine. Urine culture results were subsequently analyzed in conjunction with absolute leukocytes count in urine.

Student t test and chi-square test were calculated to assess significant differences between elderly women with and without bacteriuria, and bacteriuria with and without pyuria. p <0.05 was accepted as statistically significant. To assess the presence and absence of pyuria with and without bacteriuria, sensitivity, specificity, positive predictive value, negative predictive value, relative risk (RR) and 95% confidence intervals (CI) were calculated.

Results

Of 188 elderly women, 178 were enrolled for participation in the study. AB was detected in 44 (24.7%) of 178 women. The ages of the women ranged from 65–104 years. The mean age of those with AB was 81 ± 8 years, whereas the mean age of those without bacteriuria was 78 ± 9 years, p <0.050. An underlying active disease was present in 19 women with AB (43%), and 59 without bacteriuria (44%), p <0.921. Systemic arterial hypertension was the most frequent disease in both groups. A previous UTI was reported in eight elderly women with AB (18.1%), of which three occurred within the previous 4 months (37.5%). No evidence of vaginal symptoms was found at the time the urine specimens were obtained.

Of the 44 women with AB, 35 were found to have bacteriuria ?10⁵ CFU/ml (79.5%), and nine (20.4%) were found to have bacteriuria of 10⁴ CFU/ml. Twenty-seven (77.1%) of the 35 women with bacteriuria ?10⁵ CFU/ml were found to have pyuria in the first specimen, while in women with bacteriuria of 10⁴ CFU/ml, pyuria was rarely present, being found in only one woman (11.1%); p <0.001 (RR = 6.94; 95% CI 1.08–44.45). Two (7.4%) of the 27 women with bacteriuria ?10⁵ CFU/ml and pyuria were found not to have pyuria on the second specimen. The presence of pyuria had a sensitivity of 63.6% for bacteriuria, with a specificity of 91%. The positive predictive value for the presence of pyuria predicting those with bacteriuria was 70% (28 of 40), and the negative predictive value for the absence of pyuria predicting those without bacteriuria was 88.4% (122 of 138).

The positive predictive value for the presence of pyuria predicting those with bacteriuria ?10⁵ CFU/ml was 96.4% (27 of 28), and the negative predictive value for the absence of pyuria predicting those with bacteriuria of 10⁴ CFU/ml was 50% (8 of 16). Pyuria was found to be present in 12 women without bacteriuria (8.9%). Leukocytary casts were not present in elderly women with or without bacteriuria.

Enterobacteriaceae were isolated from 100% of the bacteriuric women, and Escherichia coli was the most common organism isolated, being found in 36 women (81.8%). Two bacterial species were found in three women. Resistance of Escherichia coli to ampicillin (68%), amoxicillin (36%), trimethoprim-sulfamethoxazole, norfloxacin, and ciprofloxacin (28.5%) was present in elderly women with bacteriuria ?10⁵ CFU/ml.