Counting of cells in cerebrospinal fluid is currently performed manually. Because of the inherent analytical and economical disadvantages, we attempted to introduce a fully automated method. Therefore, we validated the Abbott CellDyn-4000 haematology analyser for counting cells in cerebrospinal fluid. The analyser was used in its standard configuration with the simple precaution of a preceding blank sample. As for leukocyte counting the analyser yielded high precision (CV approximately 5% above the upper reference limit), good linearity, low limit of detection (2/microl) and excellent correlation (r > 0.99) with the counting chamber method. The differential leukocyte count was equally accurate and precise, even in the low concentration range. Performance of the erythrocyte count was impaired by its high limit of detection (6/nl) and it appeared satisfactory only for detecting blood admixture due to traumatic puncture. The specificity of the analyser is excellent, since it correctly classified non-viable leukocytes and excluded yeast cells from the leukocyte count in a patient with cryptococcal meningitis. We conclude that the CellDyn-4000 is well suited for quickly and reliably counting leukocytes in cerebrospinal fluid. Developing some software modifications might make the analyser useful also for performing erythrocyte counting in cerebrospinal fluid.

To establish the optimal flow cytometric method for HLA-B27 phenotyping, we compared several strategies, using three monoclonal anti-B27 antibodies (from the HLA-ABC-m3, GS145.2, and FD705 clones). We used a triple-color direct immunofluorescence assay, including a T-lymphocyte-specific antibody as an internal control and an anti-HLA-Bw4 antibody. Blood samples from >400 subjects were tested. From ROC curve analysis none of the three antibodies appeared to be suitable for use as a single typing reagent. The efficiency of the test was affected by cross-reactions with other HLA antigens, notably the HLA-B7 antigen. Preincubation with anti-B7 serum efficiently inhibited this cross-reaction and raised the test efficiency considerably. We concluded that none of the anti-B27 antibodies investigated is suitable for use as a single typing reagent. Additional typing of Bw4 is not valuable, whereas inhibition of cross-reactions due to the B7 antigen will considerably improve the performance of the test. We recommend that two different monoclonal anti-B27 antibodies be used for accurate and reliable HLA-B27 phenotyping with flow cytometry.

In pregnant women subject to abdominal trauma or other foetomaternal haemorrhage, foetal red blood cells containing haemoglobin-F (HbF) can be found in the circulation. Recently, a monoclonal antibody to HbF has become commercially available, enabling application of a flow cytometric immunofluorescence method for accurately determining the concentration of HbF+ red blood cells. We demonstrate that white blood cells are included in the cluster selected as red blood cells and that these white blood cells exhibit a level of autofluorescence that coincides with the fluorescence signal from HbF+ red blood cells. However, these white blood cells can be excluded from the analysis, thus preventing spuriously increased HbF+ red blood cell counts. We present the results of patient samples containing HbF+ red cells as illustrations of the technique and as a potential interference by HbF-containing cells of nonfoetal origin. Using samples spiked with cord blood, the method is exactly linear with a high coefficient of correlation (r=0.997). Furthermore, the assay has excellent precision (CV < 2.4%), a low limit of detection (0.12% HbF+ RBC), is independent of Rhesus D and can be completed within 1.5 h. This method is suitable for accurate determination of foetomaternal haemorrhage.

Paraffin-embedded lymph nodes from patients with Hodgkin's disease were examined for flow cytometric DNA content. In order to increase the sensitivity of the assay we tried to enrich for the neoplastic cells by bivariate analysis using a polyclonal anti-nucleolar antibody (AN-AB) and the forward scatter (FSC). DNA aneuploidy was found to be present in 67 of all 137 cases (49%), in 24 cases only demonstrable by dual-parameter analysis. The DNA index varied from 0.69 to 1.89 with a total of 22 hypo-diploid cases. The number of aneuploid nuclei exceeded the expected frequency of Reed-Sternberg (RS) and Hodgkin (H) cells in most of the analysed specimens. In conclusion, flow cytometry in Hodgkin's disease appears to give useful information regarding the ploidy status and evidence has been provided that the malignant cell population in Hodgkin's disease is not limited to the classical RS/H cells.

No abstract