A 27-year-old woman is triaged with a complaint of suprapubic pain and provides a urine sample. The clinical picture does not suggest cystitis, but her urinalysis shows the presence of leukocyte esterase and nitrites. The patient tells you she urinated directly into the cup. How do you interpret this urinalysis? Should she receive antibiotics?
Inaccurate interpretation of urinalyses (UAs) can lead to overdiagnosis of infection, unnecessary treatment, and antibiotic resistance. (Clin Infect Dis 2017;65:1199.) Accurate results are particularly important in special populations, including men, pregnant women, geriatric patients, and patients with diabetes mellitus. (Dtsch Arztebl Int 2010;107:361.)
For an accurate interpretation of a UA, a physician must decide if the presence of leukocytes or nitrites represents a pathogenic bacterial infection of the urinary tract. Many clinicians rely on the presence or absence of squamous epithelial cells on UA to determine if a urine sample is contaminated. (Nephrol Dial Transplant 1999;14:2746.) Evaluation for the presence of squamous epithelial cells requires microscopy, and is more time-consuming than the rapid bedside urine dipstick routinely used in many emergency departments.
A contaminated urine sample could be a urine culture with the growth of vaginal or skin contaminants such as Lactobacilli, Corynebacteria species, Gardnerella, and alpha-hemolytic Streptococcus. (Nephrol Dial Transplant 1999;14:2746.) Polymicrobial urinary tract infections are rare, and samples demonstrating polymicrobial growth are considered contaminated unless the patient has an ileal conduit, neurogenic bladder, vesicocolic fistula, UTI with stones, chronic renal abscess, or indwelling urinary catheter. (Nephrol Dial Transplant 1999;14:2746.)
Urine cultures can take up to three days for pathogen growth, so this method of distinguishing UTI from contamination is impractical for EPs, who must decide whether to treat with antibiotics before culture results are available. (Clin Infect Dis 2017;65:1199; N Engl J Med 1993;328:289.)
The best urine samples are those collected early in the morning or when a patient has not passed urine in the preceding four hours, providing a higher concentration and more time for bacteria to multiply. (Nephrol Dial Transplant 1999;14:2746; Evidence-Based Diagnosis in Primary Care: Practical Solutions to Common Problems. Philadelphia: Elsevier Health Sciences; 2012.) The timing of collection, however, can rarely be controlled in the ED, so midstream catch or single catheterization has been recommended. (Nephrol Dial Transplant 1999;14:2746.)
To decrease the risk of contaminated urine samples, women are instructed to spread their labia and vaginal opening, clean with water, and catch urine midstream, while uncircumcised men are instructed to pull back the foreskin prior to urination. Water is preferred over soaps and antiseptics, which can be bactericidal and lead to misleadingly low bacterial counts. (Nephrol Dial Transplant 1999;14:2746.)
Beyond Collection Technique
Contrary to popular belief, multiple studies have shown no difference in bacterial contamination rates when comparing clean-catch samples against non-clean-catch samples. (N Engl J Med 1993;328:289; J Emerg Med 2015;48:706; J Hosp Infect 1991;18:71; Arch Intern Med 2000;160:2537.) The method by which patients were educated on how to provide a clean-catch sample like verbal instructions and posters in patient bathrooms failed to decrease contamination rates. (J Emerg Med 2017;52:639; Am J Public Health 1977;67:640.) One study found that 45 percent of patients who received verbal instruction on urine collection technique actually collected a midstream sample, and only 15 percent of women parted their labia during urine collection. (West J Emerg Med 2012;13:401.)
Considering contamination is common and difficult to eliminate in ED samples, how can EPs interpret urinalyses? The presence of nitrites, a metabolic product of pathogens, increases the positive likelihood ratio of UTI by 2.6 to 10.6 times, but is an insensitive marker for infection. (Dtsch Arztebl Int 2010;107:361.) The presence of leukocyte esterase alone is also a poor predictor, with a likelihood ratio between 1.0 and 2.6. (Dtsch Arztebl Int 2010;107:361.) The presence of nitrites and moderate leukocytes increases the likelihood of infection more than sevenfold. (Fam Pract 2003;20:103.) The combined presence of nitrites, leukocytes, and blood increases the likelihood by more than 15 times. (Eur J Emerg Med 2011;18:221.)
These findings must be interpreted in the context of the patient's symptoms and the clinician's pre-test suspicion for infection. The clinician can ask about common symptoms of UTI: painful voiding, urgency, urinary frequency, and tenesmus (LR 1.16 to 1.31). (Fam Pract 2003;20:103.) The characteristics of the urine, however, are more useful predictors of infection; cloudy urine has a positive likelihood ratio of 2.1 and specificity of 60%, while foul odor has a positive likelihood ratio of 5.1 and a specificity of 96% for infection. (Evidence-Based Diagnosis in Primary Care: Practical Solutions to Common Problems. Philadelphia: Elsevier Health Sciences; 2012.) A careful history and inspection of the urine specimen can greatly assist in diagnosing UTI when the UA is equivocal. Women with typical UTI symptoms may be treated empirically without a need for a UA.
Urine contamination commonly occurs in UAs, and we frequently treat based on the results of these contaminated tests. None of the efforts to decrease contamination rates in collection methods has been effective. A history from the patient and key indicators of infection in these imperfect urinalyses will help EPs make the right decisions regarding treatment.