CLINICAL PATHOLOGY MEANS SOLVING PROBLEMS

INTRODUCTION

Portuguese Oncology Institute patients frequently present abnormal WBC scattergrams leading to incorrect five-parameter differential count (DC). A manual leucocyte differential by microscope observation of a blood smear is in these situations necessary. We aimed to evaluate the cause of interference to DC among bilirubin (BR) and target cells (TC).

METHODS

We selected 100 samples, run on the Beckman Coulter DXH900 analyzer that yielded an optical scatter pattern with DC interference. Manual DC counts were performed by microscope observation of blood smears using Wright-Giemsa staining. All samples were reprocessed using 1⁄2 and 1⁄3  dilutions with Beckman Coulter’s DxH Diluent. The differences between the percentage of Neutrophils (NE) and Lymphocytes (LY) were calculated for pre- vs postdilution (PSD), predilution (PRD) vs manual count (MC) and PSD vs MC for all samples. Comparison among groups, PRD vs PSD and PRD vs MC , was performed using non-variable Wilcoxon test. BR values were evaluated in both pre and post dilution plasma samples. Spearman’s correlation was used to calculate correlation between BR and NE differences (PRD vs PSD) and LY differences (PRD vs PSD). BR interference was evaluated using 1:3 dilution of normal WBC scattergram samples with hyperbilirubinemia (>5 mg/dL) serum samples reprocessed on the DXH900 with no TC on blood smear. TC interference was evaluated by microscope observation of blood smears with suspicious abnormal scattergrams with no DC interference and normal BR serum levels. 

RESULTS

Hyperbilirubinemia was present in 78% of the samples whereas TC was found in 91% of the samples. Dilution level of 1/3 solved 100% of DC interference, but ½ parts dilution was only effective in 95% of the samples. Using 1/3 dilution data, no correlation was observed for BR and NE difference (PRD-PSD) or for BR and LY difference (PRD-PSD). However, we found a positive correlation between alkaline phosphatase with both NE and LY differences (p=0,007, r=0.270 and p=0.021, r=0.233, respectively). Also, Wilcoxon test evidenced statistical significance in the NE difference among PRD vs PSD (p=0.00), but no differences  for PSD vs MC (p=0,97). Similarly, Wilcoxon testing evidenced statistical significance for LY difference between PRD vs PSD (p=0.00) and between PSD vs MC (p=0.05). BR presented an 1:3 relation among differences. All 20 samples evaluated for BR interference, 1/3 dilution with hyperbilirubinemic serum showed no evidence of DC interference or debris in scattergrams. Microscopic slides observed for TC interference presented at least 5 TC for microscope field (500X) and debris in WBC scattergrams.

CONCLUSION

Our results suggest that DC interference and abnormal WBC scattergrams are more related to the presence of TC on blood samples than the BR itself. The fact that dilution solved the observed interference seems to indicate that lyses reagent is unable to effective destroy red blood cells in the presence of TC and, therefore, the sample dilution could be a solution for obtaining a correct DC, assuming that a lower erythrocyte count will favor destruction by the lyses reagent. These findings seem to bring an opportunity to introduce reflexive diluted samples to our laboratory routine.

The authors disclose no conflicts of interest.