Discriminatory performance of positive urine hemoglobin for detection of significant hemolysis in patients with continuous-flow left ventricular assist devices.

Overview

abstract

BACKGROUND: Serum lactate dehydrogenase (LDH) is the standard measure for detection of hemolysis and thus surveillance for device thrombosis in patients on continuous-flow left ventricular assist device (CF-LVAD) support. Significant hemolysis has been defined as LDH ≥600 IU/L. However, LDH testing requires phlebotomy, precluding frequent home monitoring. Simple dipstick urinalysis (UA) for urine hemoglobin (U-Hb) overcomes this limitation. This study correlated U-Hb and LDH levels and evaluated the performance of UA for detection of significant hemolysis in patients with CF-LVADs. METHODS: U-Hb and LDH were measured concurrently 956 times in 221 patients with CF-LVADs. Statistics were computed to determine accuracy of UA in detecting LDH ≥600 IU/L, with a positive result being any detected U-Hb. All analyses were performed with and without excluding for 1) conditions associated with tissue damage, which are known to increase LDH, and 2) suspected or confirmed urinary tract infections or hematuria, which are known to cause hemoglobinuria for reasons other than hemolysis. RESULTS: Mean LDH for absent/mild/severe U-Hb was 360 IU/L/467 IU/L IU/L/777 IU/L without exclusions, 354 IU/L/444 IU/L IU/L/651 IU/L after excluding non-hemolytic LDH elevations, 370 IU/L/513 IU/L IU/L/1,357 IU/L after excluding urinary tract infections and hematuria, and 367 IU/L/470 IU/L IU/L/1,217 IU/L when both exclusions applied (all p < 0.001). Absent U-Hb had a negative predictive value for LDH ≥600 IU/L of >90% for all analyses. CONCLUSIONS: Serum LDH is significantly associated with U-Hb levels. Absence of U-Hb appears to efficiently exclude significant hemolysis in patients with CF-LVADs. Because it can be performed by patients at home, hemoglobinuria monitoring may enable more intense surveillance and earlier diagnosis of device thrombosis.