A-108 Evaluation of hemoglobin interference thresholds for chemical urinalysis panels

Abstract Background Chemical urinalysis is a high-volume test for most clinical laboratories that provides rapid analysis of up to 10 tests simultaneously including glucose, ketones, leukocyte esterase, nitrite, bilirubin, urobilinogen, pH, specific gravity, protein, and hemoglobin. Highly colored urines are known to impact the readability of urine dipsticks, although there is no available data on whether visible blood interferes with specific urine dipstick tests. The goal of the study was to inform an evidence-based process to handle visibly bloody urine samples submitted to clinical laboratories. Methods We hypothesized that hemoglobin would interfere with some urine dipstick results and measured hemoglobin interference with a 10-parameter urine dipstick panel. In vitro bloody urines were created by spiking graded quantities of RBCs from pools of washed, or leukoreduced RBCs. All urine samples were tested with 10SG urine dipsticks and measured using a Clinitek Advantus. The instrument was configured to report semi-quantitative levels (i.e. negative, trace, 1+, 2+, 3+) unless otherwise specified. Dipstick analysis was repeated after centrifugation at 300 x g for 10 minutes and selected in vitro urines were also reassessed after a single freeze-thaw cycle. During the study period grossly bloody patient urine samples submitted to our lab (N=12) were analyzed for chemical urinalysis using 10SG dipsticks before and after centrifugation. In all cases a single pad level change was considered remarkable. CBC analysis was performed using a Beckman Coulter DxH, and RBC pools were measured for glucose using a Nova Biomedical glucometer. Results The presence of blood in urine samples caused profound interference for leukocyte esterase, nitrite, protein, glucose, ketones, and urobilinogen. The interference was highly reproducible in vitro, and the magnitude of interference was associated with hemoglobin concentration. Urine protein was most sensitive to hemoglobin interference at >0.2 g Hgb/L when dipstick protein was trace or 1+. Clinically significant interference was observed with 10SG dipsticks at >1 g Hgb/L when leukocyte esterase and nitrite were consistently detected, and urine protein was 2+. Importantly, leukocyte esterase positivity occurred in leukoreduced RBC pools with no leukocytes. Glucose and ketones were least sensitive to hemoglobin interference in vitro at >3 g Hgb/L. Dipstick findings from bloody patient urines were remarkably consistent with in vitro urines although some clinical samples were bilirubin positive. This phenomenon was recapitulated in vitro when urines aged for several hours or after lysing samples by freeze-thaw. Centrifugation had a variable effect on clearing hemoglobin interference in patient and in vitro urines. Specific gravity and pH were unchanged in the presence of blood; and dipstick hemoglobin was 3+ in all studied urines. Conclusion Visibly bloody urines are wholly unreliable for chemical urinalysis with 10SG dipsticks when hemoglobin is >1g/L. An optimal process to handle reporting of visibly bloody urines was to test the urine, and withhold all results besides hemoglobin when dipstick hemoglobin is >2+ and protein is >1+ using 10SG dipsticks. This process allows laboratory confirmation of hematuria and limits reporting of misleading urine dipstick results.