Although the use of bone marrow transplantation has increased greatly in recent years, the quality control procedures used in bone marrow processing laboratories remain less than ideal. Accurate marrow total nucleated cell (TNC) counts are essential for effective monitoring of bone marrow collection and processing. Aspirated marrow is variably contaminated by fat particles, resulting in overestimation of marrow TNC by automated analyzers. A recently-marketed hematological analyzer (Cobas-Helios; Roche Diagnostic Systems, Branchburg, NJ) offers the potential to correct marrow TNC counts for fat particles using available software. The authors investigated the accuracy of corrected TNC counts on 21 marrow samples, using a visual chamber count as the reference method. The correction methods studied were software correction, using the Cobas-Helios differential system, and replacement of the sample plasma with saline, Uncorrected automated marrow TNC counts (mean, 28.4 X 10(9)/L) were significantly higher than the visual reference counts (mean, 23.1 X 10(9)/L). Neither the mean corrected automated count (24.3 X 10(9)/L) nor the mean saline replaced count (24.6 X 10(9)/L) differed significantly from the mean visual reference count. For both the corrected automated and saline replaced counts, 20 of the 21 data points (95%) fell within a 95% confidence interval computed for the reference method. The authors conclude that both the corrected automated method, using the Cobas-Helios, and the saline replacement method are acceptable alternatives to the visual chamber count.
