The Degree of Pyuria Can Help Determine Urinary Tract Infection in Elderly Women
October 1, 2023
By Richard R. Watkins, MD, MS, FACP, FIDSA, FISAC
Professor of Internal Medicine, Northeast Ohio Medical University, Rootstown, OH
SYNOPSIS: Elderly women with a urinary tract infection had a higher median number of urinary leukocytes compared to uninfected controls, including those with asymptomatic bacteriuria. For 264 leukocytes/µL, the sensitivity and specificity of microscopy were 88%. Using the standard pyuria threshold of 10 leukocytes/µL gave a specificity of 36% and a sensitivity of 100%.
SOURCE: Bilsen MP, Aantjes MJ, van Andel E, et al. Current pyuria cutoffs promote inappropriate urinary tract infection diagnosis in older women. Clin Infect Dis 2023;76:2070-2076.
Differentiating urinary tract infection (UTI) from asymptomatic bacteriuria (ASB) can be challenging in clinical practice. This is especially true in elderly patients, who may not manifest typical symptoms and can have limited communication abilities from cognitive impairment. Bilsen and colleagues sought to determine whether the degree of urinary leukocytosis could help distinguish UTI from uninfected controls, including those with ASB.
The study was conducted in the Netherlands and recruited subjects from hospitals, primary care centers, and senior living and long-term care facilities. Women aged 65 years and older were included. Exclusion criteria were an inability to express symptoms, having an indwelling urinary catheter, immunosuppressive use, antimicrobial use within 48 hours, current urolithiasis, and a UTI within the preceding month. To be included in the UTI group, subjects had to have at least two new-onset lower urinary tract symptoms (i.e., dysuria, frequency, urgency, or suprapubic pain); pyuria, defined as ≥ 10 leukocytes/µL or five or more leukocytes/high powered field or the presence of leukocyte esterase; and a monoculture (i.e., one uropathogen ≥ 104 colony forming units/mL). Fever, defined as a temperature ≥ 38°C, was used to differentiate lower from upper UTI.
Community-dwelling women and long-term care residents who did not have any lower urinary tract symptoms or fever were eligible to serve as controls. They were divided into three groups: ASB (defined as having at least two consecutive urine cultures [two to four weeks apart] with the same uropathogen with 105 CFU/mL or greater), negative culture, and mixed flora. The investigators considered Enterobacterales, enterococci, Pseudomonas aeruginosa, Staphylococcus saprophyticus, and group B streptococci as uropathogens. There were 164 subjects included in the primary analysis, 63 in the UTI group and 101 controls. Their mean age was 78.3 years. The baseline characteristics between the UTI group and controls were well matched, although the control group contained a higher proportion of long-term care residents than the UTI group (43/101 [42.6%] vs. 7/63 [11.1%], respectively). Compared to controls, subjects with a UTI had higher median urinary leukocytes (microscopy: 26 leukocytes/µL vs. 900 leukocytes/µL; flow cytometry: 23 leukocytes/µL vs. 1,575 leukocytes/µL; P < 0.001). The median leukocyte values were higher for subjects with UTI compared to those with ASB (microscopy: 900 leukocytes/µL vs. 296 leukocytes/µL [P = 0.002]; flow cytometry: 1,575 leukocytes/µL vs. 197 leukocytes/µL [P = 0.004]).
Using a cutoff of 264 leukocytes/µL, the sensitivity and specificity of microscopy were 88% (95% confidence interval [CI], 77% to 94%; positive and negative likelihood ratio [LR]: 7.2 and 0.1, respectively). Flow cytometry demonstrated a sensitivity of 91% (95% CI, 79% to 98%) and a specificity of 86% (95% CI, 78% to 92%) using a cutoff value of 231 leukocytes/µL, with a positive LR of 6.5 and a negative LR of 0.1. Furthermore, the clinically used cutoff for UTI of 10 leukocytes/µL had a sensitivity of 100% (95% CI, 94% to 100%) with a specificity of only 36% (95% CI, 28% to 48%). A cutoff of 100 cells increased specificity to 71% (95% CI, 61% to 79%) and decreased sensitivity to 93% (95% CI, 84% to 98%), while a cutoff of 400 cells increased specificity to 92% (95% CI, 86% to 96%) and dropped sensitivity to 77% (95% CI, 65% to 87%).
COMMENTARY
This was an interesting study that had two main results. First, the degree of urinary leukocytosis can help distinguish UTI from non-infection in elderly females, including those patients with ASB. Second, the currently used definition of pyuria being 10 leukocytes/µL has a very low specificity for UTI in elderly women. Thus, clinicians need to reconsider the dogma that 10 leukocytes in a urinalysis equals a UTI in an elderly female patient. In light of the findings by Bilsen and colleagues, reassessing this threshold in other patient populations at increased risk for UTIs seems warranted. Indeed, multiple pyuria thresholds may need to be developed.
Like all studies, this one has some limitations to consider. First, because the study only included elderly women, the results should not be generalized to other types of patients. Second, the control group had a higher proportion of long-term care residents than the UTI group, which may have been a confounding variable that affected the results. Third, there may have been some UTI cases that were excluded as a result of the investigators’ uropathogen definition.
This study provides good evidence that the diagnostic threshold for pyuria in elderly women needs to increase. Indeed, 300 leukocytes/µL would be more reasonable, raising specificity (88%) to avoid overtreatment while still maintaining an adequate sensitivity (84%). Of course, pyuria levels should be interpreted in the clinical context of individual patients and should never be the only determining factor for a UTI diagnosis.