Automated Quantitation of Hemoglobin-Based Blood Substitutes in Whole Blood Samples

Transcription

1 Hematopathology / HEMOGLOBIN-BASED BLOOD SUBSTITUTES Automated Quantitation of Hemoglobin-Based Blood Substitutes in Whole Blood Samples Jolanta Kunicka, PhD, 1 Michael Malin, PhD, 1 David Zelmanovic, PhD, 1 Marc Katzenberg, 1 William Canfield, MA, 1 Phyllis Shapiro, 1 and Narla Mohandas, DSc 2 Key Words: Blood substitutes; Hemoglobin measurement methods; RBC indices; Light scatter; Flow cytometry Abstract It is necessary to develop methods for accurate monitoring of cell-free hemoglobin in circulation. Routine monitoring of circulating cell-free hemoglobin will be useful for evaluating the efficacy of blood substitute administration and for determining the clearance rates of the blood substitute from circulation. In addition, discriminating between cell-free hemoglobin and cell-associated hemoglobin will enable accurate determination of RBC indices, mean cell hemoglobin and mean corpuscular hemoglobin concentration, in individuals receiving hemoglobinbased blood substitutes. As colorimetric methods used by hematology analyzers to quantitate the hemoglobin value of a blood sample cannot distinguish between cell-associated and cell-free hemoglobin, it is currently not feasible to quantitate the levels of hemoglobin substitutes in circulation. The advent of a technology that measures volume and hemoglobin concentration of individual RBCs provides an alternative strategy for quantitating the cell-associated hemoglobin in a blood sample. We document that the combined use of cellbased and colorimetric hemoglobin measurements provides accurate discrimination between cellassociated and cell-free hemoglobin over a wide range of hemoglobin levels. This strategy should enable rapid and accurate monitoring of the levels of cell-free hemoglobin substitutes in the circulation of recipients of these blood substitutes. Substantial progress is being made in the development of safe and effective hemoglobin-based blood substitutes. 1-6 A number of these hemoglobin-based blood substitutes are in phase 3 clinical trials for the treatment of patients with acute blood loss following severe trauma. At present there is no automated method to quantitate the circulating levels of hemoglobin-based blood substitutes in a blood sample. As the approval of these blood substitutes for clinical use is imminent, we embarked on developing an automated method to monitor the levels of hemoglobin substitutes in the circulation of recipients of these blood products. The hemoglobin value, a principal component of the CBC count, is determined by automated hematology analyzers using colorimetric methods. 7 These colorimetric methods measure total hemoglobin (Hgb total, in g/dl), which includes both cell-associated hemoglobin (Hgb cell ) and cell-free hemoglobin (Hgb free ). With the advent of a technology that simultaneously measures the volume (V, fl) and the hemoglobin concentration (HC, g/dl) of individual RBCs, it is possible to obtain an independent measure of Hgb cell as Hgb cell = RBC MCV CHCM, where RBC is the RBC count, MCV is mean cell volume, and CHCM (equivalent to mean corpuscular hemoglobin concentration [MCHC]) is the mean cellular hemoglobin concentration. In the present study, we first validated the accuracy of this cell-by-cell approach of quantitating hemoglobin levels in samples without cell-free hemoglobin by comparing Hgb cell with Hgb total values derived by colorimetric determination on blood samples. Based on measurements performed on 186 normal blood samples and 1,71 patient blood samples, we demonstrated that Hgb cell accurately reflects the circulating cell-associated hemoglobin values in these samples. As the cell-based method quantitates only hemoglobin in RBCs (Hgb cell ), while colorimetric determination of the Am J Clin Pathol 21;116: on 16 June 218

2 Kunicka et al / HEMOGLOBIN-BASED BLOOD SUBSTITUTES A Low Angle Scatter (2-3 ) B 6 12 fl g/dl C High Angle Scatter (5-15 ) 19 4 pg Figure 1 RBC cytograms and histograms. RBC cytogram (left) shows the low- and high-angle light scatter measurements from individual RBCs. Scatter cytogram map encompasses volumes between 3 and 18 fl and refractive index values between 1.38 and The measured refractive index value is used to derive cell hemoglobin concentration of individual RBCs. Histograms for cell volume (A), cell hemoglobin concentration (B), and cell hemoglobin content (C) of RBCs in the analyzed blood samples are shown (right). From these distributions the mean cell volume (MCV = 85 fl), mean cell hemoglobin concentration (CHCM = 33 g/dl), and mean cell hemoglobin content (CH = 29 pg) are derived. hemoglobin level (Hgb total ) reflects both hemoglobin in RBCs (Hgb cell ) and hemoglobin in plasma (Hgb free ), we surmised that the combined use of these 2 measurements should enable us to accurately measure cell-free hemoglobin in blood samples. By using mixtures of whole blood and 3 different hemoglobin-based blood substitutes, we were able to document that our experimental strategy can accurately quantitate the levels of hemoglobin substitutes (Hgb free ) in the circulation of recipients of these blood products. Materials and Methods Blood Samples Fresh K 3 -EDTA anticoagulated blood samples from 186 healthy volunteers and 1,71 patients were evaluated. All blood samples were analyzed within 4 hours of phlebotomy. Patient blood samples were from individuals with RBC disorders such as iron deficiency, thalassemia, and sickle cell disease. Blood Substitutes A hemoglobin-based blood substitute, PEG hemoglobin at a concentration of 5.4 g/dl (54 g/l), was obtained from Enzon (Piscataway, NJ). Hemopure (HBOC-21) and Oxyglobin blood substitutes at a concentration of 13 g/dl (13 g/l) were obtained from Biopure (Cambridge, MA). Varying proportions of blood substitute solutions and normal blood were mixed to a final volume of 5 ml to produce samples with a range of cellular and cell-free hemoglobin concentrations. The 2-Dimensional RBC Analysis Method Two-dimensional RBC analysis is an optical method in which both volume and refractive index of isotonically sphered RBCs are determined simultaneously on a cell-bycell basis by measuring 2 angles of laser light scatter. 8-1 The 2 scatter measurements are converted into volume and refractive index values using the Mie theory of light scattering for homogeneous spheres. The refractive index values subsequently are converted into cellular hemoglobin concentration values. This method is the basis for RBC analysis on the ADVIA 12 Hematology System (Bayer, Tarrytown, NY). The RBC scatter cytogram Figure 1 is a graphic representation of 2 light scatter measurements: high-angle (5-15 ) light scatter events are plotted along the x-axis and low-angle (2-3 ) light scatter events along the y-axis. The RBC scatter/scatter cytogram map resolves volumes 914 Am J Clin Pathol 21;116: on 16 June 218

3 Hematopathology / ORIGINAL ARTICLE between 3 and 18 fl and refractive index values between 1.38 and As the RBC refractive index is a monotonic function of hemoglobin concentration, the derived refractive index is converted into cell HC. This results in the resolution of HC values from 19 to 49 g/dl (19-49 g/l). The cellular hemoglobin content (CH, in pg) is derived as the product of the 2 measured values (CH = V HC). From the cell-by-cell values of V and HC and the calculated value of CH, histograms of V, HC, and CH are constructed (Figure 1), and mean values for these cell parameters (MCV, CHCM, and CH) are derived. The RBC-associated hemoglobin is calculated as Hgb cell = RBC MCV CHCM, where RBC is the RBC count. The hemoglobin concentration of the blood sample (Hgb total ) also is determined colorimetrically following lysis of RBCs by the hemoglobin reagent (Bayer). The colorimetrically determined and the cell-associated hemoglobin values, as well as the difference between these 2 hemoglobin values reflecting cell-free hemoglobin, are automatically reported on the ADVIA 12 Hematology System. Statistical Analysis The acquired data were analyzed statistically using Excel 7. for Microsoft Windows 95 (Microsoft, Redmond, WA). The parameters derived included calculations of mean values, SD, and linear regression analysis with correlation coefficient, slope, and intercept. Results Blood Hemoglobin Levels By using the automated ADVIA 12 Hematology System, the cellular hemoglobin values of blood samples (Hgb cell ) were compared with hemoglobin values derived by colorimetric determination on blood samples lysed by the hemoglobin reagent (Hgb total ). As shown in Figure 2, there is an excellent correlation between hemoglobin values derived from measured RBC indices (Hgb cell ) and colorimetric hemoglobin measurement (Hgb total ). For 186 normal blood samples, the correlation coefficient (r 2 ) was.9611, and the slope of the regression line was 1. with an intercept of.6 Figure 2A. For 1,71 patient blood samples, the correlation coefficient (r 2 ) was.9767, and the slope of the regression line was.97 with an intercept of.32 Figure 2B. Thus, cellular hemoglobin measurement (Hgb cell ) accurately reflects the cell-associated hemoglobin values of all blood samples. Three different hemoglobin measurement values are reported by the ADVIA 12 Hematology System: Hgb total reflecting both cell-associated and cell-free hemoglobin determined colorimetrically is reported as HGB ; A Cellular Hemoglobin (g/dl) B Cellular Hemoglobin (g/dl) Colorimetric Hemoglobin (g/dl) Colorimetric Hemoglobin (g/dl) Figure 2 Accuracy of cellular hemoglobin values compared with colorimetric hemoglobin values. Linear regression analysis of cellular hemoglobin and colorimetric hemoglobin for 186 normal blood samples (A) and for 1,71 patient blood samples (B). A, y = 1.16x.628; r 2 =.9611; B, y =.979x ; r 2 = Hemoglobin values are given in conventional units; to convert to Système International units (grams per liter), multiply by 1. Hgb cell reflecting only the cell-associated hemoglobin is reported as Calculated HGB ; and the difference between these 2 hemoglobin values, reflecting cell-free hemoglobin, is reported as HGB Delta. All 3 hemoglobin values appear on the display screen on the completion of the 3-second analytic cycle that follows the aspiration of the blood sample. Hemoglobin Content of RBCs The mean CH of RBCs in the blood sample was derived using the data set from a dual measurement of V and HC of individual red cells, as shown in Figure 1. The mean cell hemoglobin content (MCH) also can be derived from 2 other independent measurements, namely hemoglobin concentration and RBC count (MCH = Hgb/RBC). To verify that the CH value derived from measured volume and cell hemoglobin concentration values is equivalent to MCH derived Am J Clin Pathol 21;116: on 16 June 218

4 Kunicka et al / HEMOGLOBIN-BASED BLOOD SUBSTITUTES A CH (pg) MCH (pg) Assayed Hemoglobin (g/dl) Predicted Hemoglobin (g/dl) B CH (pg) MCH (pg) Figure 3 Relationship between cell hemoglobin content mean (CH) derived from cell by cell analysis and mean cell hemoglobin content (MCH) derived from colorimetric hemoglobin and RBC values. Linear regression analysis of CH and MCH values derived for 186 normal blood samples (A) and for 1,71 patient blood samples (B). A, y =.961x ; r 2 =.8896; B, y =.954x ; r 2 = from hemoglobin and RBC values, we compared the CH and MCH values of 186 normal and 1,71 patient blood samples. As shown in Figure 3, there was excellent concordance between CH and MCH values for normal and patient blood samples. For normal blood samples, the correlation coefficient (r 2 ) between CH and MCH was.8896, and the slope of the regression line was.96 with an intercept of.9 (Figure 3A). The mean values were as follows: MCH, 29.4 pg; and CH, 29.1 pg. For patient blood samples, the correlation coefficient (r 2 ) was.9188, and the slope of the regression line was.95 with an intercept of 1.2 (Figure 3B). The mean values were as follows: MCH, 28.9 pg; and CH, 28.7 pg. Thus, the CH value was equivalent to MCH for samples without extracellular hemoglobin. Quantitation of Cell-Free Hemoglobin From Blood Substitutes Since the cell-based method quantitates only cellular hemoglobin (Hgb cell ), while the colorimetric method Figure 4 Accurate measurement of cell-associated and cellfree hemoglobin in mixtures of normal blood and the blood substitute PEG hemoglobin (Enzon, Piscataway, NJ). Linear regression analysis of assayed (symbols) and predicted (solid line) cell-associated and cell-free hemoglobin from hemoglobin substitute in blood samples. For cell-associated and cell-free hemoglobin, the correlation coefficients (r 2 ) between assayed and predicted values were.999 and.998, respectively. The slope of the regression line and the intercept were.97 and.5 for cell-associated hemoglobin and.95 and.2 for cell-free hemoglobin. Values are given in conventional units; to convert to Système International units (grams per liter), multiply by 1. Triangles, cell-associated hemoglobin; circles, cell-free hemoglobin. (Hgb total ) quantitates the sum of both cellular and cell-free hemoglobin, we explored whether the combined use of these 2 measurements would enable the accurate quantitation of hemoglobin blood substitutes (Hgb free ) in a blood sample. A value for Hgb free is derived as the difference between Hgb total and Hgb cell. Blood samples with varying amounts of cellular and cell-free hemoglobin were prepared by mixing appropriate amounts of the hemoglobin-based blood substitute PEG hemoglobin, at a concentration of 5.4 g/dl (54 g/l), with normal blood (diluted to a hemoglobin concentration of 5.6 g/dl (56 g/l) and analyzed. As shown in Figure 4, the combined use of Hgb cell and Hgb total enabled accurate discrimination between cellular and cell-free hemoglobin over a wide range of hemoglobin substitute levels ( g/dl [5-54 g/l]) and cellular hemoglobin levels (-5.6 g/dl [-56 g/l]). Measured values of cellular and extracellular hemoglobin were within.1 g/dl (1 g/l) of the predicted values based on the ratios of PEG hemoglobin and normal blood in the prepared mixture. Blood samples with varying amounts of cellular hemoglobin and a constant amount of cell-free hemoglobin also were prepared by mixing appropriate amounts of the PEG hemoglobin with normal blood; the data from this study are shown in Table 1. Once again, measured values of both cellular and extracellular 916 Am J Clin Pathol 21;116: on 16 June 218

5 Hematopathology / ORIGINAL ARTICLE Table 1 Predicted and Measured Cellular and Extracellular Hemoglobin Values in Blood Samples Prepared by Mixing Varying Amounts of Whole Blood With a Constant Concentration of PEG Hemoglobin Substitute * Total Hemoglobin (g/dl) Cellular Hemoglobin (g/dl) Cell-Free Hemoglobin (g/dl) Predicted Assayed Predicted Assayed Predicted Assayed ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±..5.5 ± ± ±..2.3 ± ±.4 * Values are given in conventional units; to convert to Système International units (grams per liter), multiply by 1. PEG hemoglobin is from Enzon, Piscataway, NJ. Assayed Hemoglobin (g/dl) Predicted Hemoglobin (g/dl) Figure 5 Accurate measurement of cell-associated and cellfree hemoglobin in mixtures of normal blood and the blood substitute Oxyglobin (Biopure, Cambridge, MA). Linear regression analysis of assayed (symbols) and predicted (solid line) cell-associated and cell-free hemoglobin from the hemoglobin substitute Oxyglobin in blood samples. The correlation coefficients (r 2 ) between assayed and predicted values for cell-associated and cell-free hemoglobin were.999 and.999, respectively. The slope of the regression line and the intercept were 1. and.7 for cell-associated hemoglobin and 1. and.13 for cell-free hemoglobin. Data also are given in tabular form (Table 2). Values are given in conventional units; to convert to Système International units (grams per liter), multiply by 1. Triangles, cellassociated hemoglobin; circles, cell-free hemoglobin. Assayed Hemoglobin (g/dl) Predicted Hemoglobin (g/dl) Figure 6 Accurate measurement of cell-associated and cellfree hemoglobin in mixtures of normal blood and the blood substitute Hemopure (Biopure, Cambridge, MA). Linear regression analysis of assayed (symbols) and predicted (solid line) cell-associated and cell-free hemoglobin from the hemoglobin substitute Hemopure in blood samples. The correlation coefficients (r 2 ) between assayed and predicted values for cell-associated and cell-free hemoglobin were.999 and.999, respectively. The slope of the regression line and the intercept were.99 and.5 for cell-associated hemoglobin and.995 and.3 for cell-free hemoglobin. Data also are given in tabular form (Table 3). Values are given in conventional units; to convert to Système International units (grams per liter), multiply by 1. Triangles, cellassociated hemoglobin; circles, cell-free hemoglobin. hemoglobin accurately reflected the ratios of PEG hemoglobin and normal blood in the prepared mixtures. As shown in Figure 5, Figure 6, Table 2, and Table 3, the combined use of dual hemoglobin measurements, Hgb cell and Hgb total, also enabled accurate discrimination between cellular and cell-free hemoglobin for the 2 other blood substitutes, Hemopure and Oxyglobin, over a wide range of hemoglobin values (-13 g/dl [-13 g/l]). For these experiments, blood samples with varying amounts of cellular and cell-free hemoglobin were prepared by mixing appropriate amounts of the hemoglobin-based blood substitutes Oxyglobin and Hemopure, at a concentration of 13 g/dl (-13 g/l), with varying ratios of normal blood at a hemoglobin concentration of 13 g/dl (-13 g/l) and analyzed. Measured values of both cellular and cell-free hemoglobin were again within.1 g/dl (1 g/l) of the predicted values over a wide range of hemoglobin values. These data show that the experimental strategy we devised can accurately monitor the amount of various hemoglobin blood substitutes in circulation. Am J Clin Pathol 21;116: on 16 June 218

6 Kunicka et al / HEMOGLOBIN-BASED BLOOD SUBSTITUTES Table 2 Recovery of Cell-Free Hemoglobin (Oxyglobin) in Whole Blood * Composition of the Blood Sample Total Hemoglobin (g/dl) Cellular Hemoglobin (g/dl) Cell-Free Hemoglobin (g/dl) Whole Blood Oxyglobin (ml) (ml) Predicted Assayed Predicted Assayed Predicted Assayed ± ±.1.. ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±.9.. ± ±.9 * Values are given in conventional units; to convert to Système International units (grams per liter), multiply by 1. Assayed results are mean ± SD of 5 replicates. Oxyglobin is from Biopure, Cambridge, MA. Table 3 Recovery of Exogenous Hemoglobin (Hemopure) in Whole Blood * Composition of the Blood Sample Total Hemoglobin (g/dl) Cellular Hemoglobin (g/dl) Cell-Free Hemoglobin (g/dl) Whole Blood Hemopure (ml) (ml) Predicted Assayed Predicted Assayed Predicted Assayed ± ±.5.. ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±.9.. ± ±.9 * Values are given in conventional units; to convert to Système International units (grams per liter), multiply by 1. Assayed results are mean ± SD of 5 replicates. Hemopure is from Biopure, Cambridge, MA. Discussion The circulating hemoglobin value that is used routinely for assessing the degree of anemia in individual patients has long been quantitated using colorimetric methods. These colorimetric methods quantitate total hemoglobin, which includes both cell-associated and cell-free hemoglobin in plasma. The advent of technology that enables the measurement of V and HC of individual RBCs for the first time provides a cell-based strategy for quantitating exclusively cell-associated hemoglobin in a blood sample. In the present study, we documented the accuracy of this new RBC-based approach for quantitating cell-associated hemoglobin levels. Based on measurements performed on 186 normal blood samples and 1,71 patient blood samples, we showed that the cell-by-cell approach can accurately measure circulating hemoglobin values. The concordance between hemoglobin values determined by the new method and those derived by the well-established colorimetric method is excellent. More important, through the combined use of the cell-by-cell and colorimetric hemoglobin measurements, we effectively discriminated between cellular and cellfree hemoglobin over a wide range of hemoglobin levels. The range of admixed hemoglobin substitutes tested (-13 g/dl [-13 g/l]) should cover the entire range of attainable levels of these products in recipients. As there are no automated methods to discriminate between cell-associated and cell-free hemoglobin in a blood sample, the novel strategy outlined herein to discriminate between these 2 sources of hemoglobin should be particularly useful for monitoring the levels of hemoglobin-based blood substitutes in the circulation of recipients of these blood products. Another advantage of the experimental strategy we have outlined is that it will enable the accurate determination of RBC MCH and MCHC values in blood samples of recipients of hemoglobin-based blood substitutes. As the RBC MCH and MCHC values should reflect only cell-associated hemoglobin, the hematology analyzers that fail to discriminate between cell-associated and cellfree hemoglobin values will markedly overestimate the values for MCH and MCHC by failing to exclude the 918 Am J Clin Pathol 21;116: on 16 June 218

7 Hematopathology / ORIGINAL ARTICLE contribution of cell-free hemoglobin to the derived MCH and MCHC values. For example, in a blood sample from a transfusion recipient with an RBC count of /µl ( /L), an MCV of 9 µm 3 (9 fl), and a total hemoglobin value of 15 g/dl (15 g/l) composed of 1 g/dl (1 g/l) of RBC-associated hemoglobin and 5 g/dl (5 g/l) of hemoglobin-based blood substitute, the MCHC and MCH values for RBCs will be erroneously reported by standard hematology analyzers as 5.5 g/dl (55 g/l) and 45. pg respectively, instead of the correct values of 33.7 g/dl (337 g/l) and 3 pg, owing to their failure to discount the contribution of cell-free hemoglobin to the derived values. In contrast, the experimental strategy we have outlined, through its ability to discriminate between cell-associated and cell-free hemoglobin, will derive accurate values for MCHC and MCH for such a blood sample. As it is likely that during the coming years hemoglobinbased blood substitutes will be used in clinical practice, the experimental strategy we have described should have significant practical application in transfusion medicine. From the 1 Laboratory Testing Segment, Bayer, Tarrytown, NY; and 2 Life Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA. Address reprint requests to Dr Mohandas: New York Blood Center, 31 E 67th St, New York, NY 121. References 1. Winslow RM. New transfusion strategies: red cell substitutes. Annu Rev Med. 1999;5: Baron JF. Blood substitutes: haemoglobin therapeutics in clinical practice. Crit Care. 1999;3:R99-R LaMuraglia GM, O Hara PJ, Baker WH, et al. The reduction of the allogenic transfusion requirement in aortic surgery with a hemoglobin-based solution. J Vasc Surg. 2;31: Lamy ML, Daily EK, Brichant JF, et al, for the DCLHb Cardiac Surgery Trial Collaborative Group. Randomized trial of diaspirin cross-linked hemoglobin solution as an alternate to blood transfusion after cardiac surgery. Anesthesiology. 2;92: Mullon J, Giacoppe G, Clagett C, et al. Transfusion of polymerized bovine hemoglobin in a patient with severe autoimmune hemolytic anemia. N Engl J Med. 2;342: Friedman HI, Devenuto F, Kerwin A, et al. Hemoglobin solutions as blood substitutes. J Invest Surg. 2;13: Malin MJ, Sclafani LD, Wyatt JL. Evaluation of cyanidecontaining and cyanide-free methods for whole blood hemoglobin on the Technicon H*1 TM analyzer with normal and abnormal blood samples. Am J Clin Pathol. 1989;92: Mohandas N, Kim YR, Tycko DH, et al. Accurate and independent measurement of volume and hemoglobin concentration of individual red cells by laser light scattering. Blood. 1986;68: Mohandas N, Johnson A, Wyatt J, et al. Automated quantitation of cell density distribution and hyperdense cell fraction in RBC disorders. Blood. 1989;74: Lew VL, Raftos JE, Sorette M, et al. Generation of normal human red cell volume, hemoglobin content, and membrane area distributions by birth or regulation? Blood. 1995;86: Am J Clin Pathol 21;116: on 16 June 218