Should the upper tracts be imaged for microscopic haematuria?

• Published in: BJU international 2005-09, Vol.96 (4), p.612-617
• Main Authors: Feldstein, Murray S., Hentz, Joseph G., Gillett, Michael D., Novicki, Donald E.
• Format: Article
• Language: English
• Subjects: Adult; Biological and medical sciences; diagnostic imaging; Epidemiologic Methods; Female; haematuria; Hematuria - diagnostic imaging; Humans; Investigative techniques, diagnostic techniques (general aspects); Kidney - diagnostic imaging; kidney neoplasms; Kidney Neoplasms - diagnostic imaging; Kidney Neoplasms - urine; Kidneys; Male; Medical sciences; Nephrology. Urinary tract diseases; Patient Selection; Radiodiagnosis. Nmr imagery. Nmr spectrometry; Radiography; Tumors of the urinary system; Unnecessary Procedures; Ureter - diagnostic imaging; Urinalysis; Urinary system
• ISSN: 1464-4096; 1464-410X
• DOI: 10.1111/j.1464-410X.2005.05694.x

OBJECTIVE To determine the relative prevalence of various definitions of microscopic haematuria (MH) in patients with renal neoplasms and controls, and to predict the likely outcome of renal imaging for those definitions. PATIENTS AND METHODS In a retrospective case‐control study 278 adult men and woman seen between 1998 and 2003 with untreated renal neoplasms were compared to controls matched for age and sex. All cases and controls had renal imaging within 6 months of a urine analysis. Patients were excluded for gross haematuria or other conditions associated with MH but not relevant to upper tract imaging. Adjusted odds ratios (OR) computed for 13 definitions of MH by conditional logistic regression were the primary outcome measures. Additional outcome measures were ORs in selected subsets. Hypothetical performance characteristics of a positive urine analysis were then derived to predict the likely results of detecting renal neoplasms for each definition of MH. RESULTS The OR (95% confidence interval) for the entire series of cases and controls, both symptomatic and asymptomatic, was 2.0 (1.02–3.92, P = 0.04) for MH defined as ≥ 4 red blood cells per high‐power field (RBC/HPF) and 2.2 (1.09–4.52, P = 0.03) for ≥ 5 RBC/HPF. No significant OR was calculated for ≤ 3 RBC/HPF, nor for a subgroup of patients with MH in a routine urine analysis obtained during a periodic health examination. Symptomatic patients had an OR of 13.68 (1.6–117.1, P = 0.02) for MH defined as ≥ 5 RBC/HPF. The sensitivity of a positive test decreased from 24.8% to 5.04% as the definition for MH became more stringent. The theoretical positive predictive value (assuming a prevalence of renal cell neoplasms of 0.25%) of the most stringent definition of MH was 0.58%. CONCLUSIONS Patients with renal neoplasms have about twice the prevalence of MH with ≥ 4 or 5 RBC/HPF in a single urine sample compared with matched controls, but this difference has little impact on the hypothetical detection rate of renal cancer. Imaging the kidney for low‐grade MH in a routine urine analysis discovered at a periodic health examination in an otherwise asymptomatic patient is tantamount to screening without cause, and can be deferred for selected patients. The clinical context is as important as the degree of MH when deciding to image the kidneys.