Detection and quantitation of hemoglobin (Hb) fractions

Transcription

1 Evaluation of a Dual Hemoglobin A 2 /A 1c Quantitation Kit on the Bio-Rad Variant II Automated Hemoglobin Analyzer John D. Lafferty, MLT, ART; Andrew G. McFarlane, MLT, ART; David H. K. Chui, MD, FRCP Context. Quantitation of hemoglobin (Hb) A 1c and investigation of hemoglobinopathy on the Bio-Rad Variant analyzers require a switch between 2 separate kits that is time consuming and causes errors. Objective. Evaluation of a new Variant II HbA 2 /HbA 1c Dual kit capable of both Hb A 1c quantitation and hemoglobinopathy investigation on a single kit. Design. We evaluated Hb A 1c,HbA 2, and Hb F quantitation for precision, linearity, and correlation with current methodology. We also evaluated detection of Hb variants and correlation of Hb Barts quantitation. Setting. Hamilton Regional Laboratory Medicine Program, Provincial Hemoglobinopathy Laboratory, St Joseph s Healthcare Site, Hamilton, Ontario. Patients. Patient blood samples submitted for Hb A 1c quantitation or hemoglobinopathy investigation. Main Outcome Measures. Precision, linearity, linear regression, and reference interval validation. Results. We provide tables and figures illustrating precision, linearity, linear regression, and quantitation of Hb variants. We validated reference intervals for Hb A 1c,Hb A 2, and Hb F. Conclusions. The dual kit provides precise Hb A 1c,Hb A 2, and Hb F quantitation. The results show good linearity and correlate well with the results of current methods. We detected all clinically important Hb variants and a wide variety of rare variants. The dual kit has several advantages: it eliminates the need for extensive kit switch over; improves utility for newborn screening because of its quantification of Hb Barts; permits quantification of Hb A 1c using the -Thal method; and eliminates the need for separate Hb A 2 reference intervals for patients with Hb S because of its accurate quantitation of Hb A 2 in the presence of Hb S. (Arch Pathol Lab Med. 2002;126: ) Detection and quantitation of hemoglobin (Hb) fractions is clinically relevant in 2 situations. First, the quantitation of Hb A 1c, the major glycated form of Hb A, is useful in the clinical management of diabetes. Second, the quantitation of Hb A 2 and Hb F, and the detection and quantitation of various Hb variants is an essential tool in the diagnosis of hemoglobinopathies (eg, thalassemia syndromes or sickle cell syndromes). Methods of Hb A 1c analysis traditionally include electrophoresis, immunoassay, or chromatography. Methods of hemoglobinopathy investigation traditionally include alkaline and acid Hb electrophoresis in combination with Hb A 2 quantitation by electrophoresis or chromatography and Hb F quantitation by alkali denaturation or radial immunodiffusion. Hemoglobin fraction analysis by cation exchange high-performance liquid chromatography (HPLC) has the advantage of providing Hb A 1c quantitation as well as hemoglobinopathy investigation using a single technology. 1 3 Automated analysis of Hb fractions by cation exchange Accepted for publication June 14, From the Hamilton Regional Laboratory Medicine Program, Hamilton, Ontario (Mr Lafferty, Mr McFarlane, and Dr Chui) and the Departments of Medicine (Mr Lafferty) and Pathology and Molecular Medicine (Dr Chiu), McMaster University, Hamilton, Ontario. Reprints: John Lafferty MLT, ART, Core Laboratory, Hamilton General Site, 237 Barton St E, Hamilton, Ontario, Canada LL 2X2 ( lafferty@hhsc.ca). HPLC has been available through Bio-Rad Laboratories Inc (Hercules, Calif) since ,5 The first-generation instrument, the Bio-Rad Variant, provides automated hemolysate injection, chromatography, and chromatogram integration, as well as Hb fraction identification and quantitation. The second-generation instrument, the Bio-Rad Variant II, performs the same functions with the addition of cap piercing, automated hemolysate preparation, bar code reading, and a Windows NT compatible Clinical Data Management (CDM) system capable of interfacing with a laboratory information system. Until recently, performance of both Hb A 1c quantitation and hemoglobinopathy investigation on the Variant or Variant II has required the use of 2 separate methods or kits. Quantitation of Hb A 1c required a Hb A 1c kit (current Hb A 1c kit) consisting of a specific chromatography column (cartridge) and reagents used in conjunction with an Hb A 1c program (method) optimized for Hb A 1c quantitation. Hemoglobinopathy investigation required the use of a thalassemia Short kit (current -Thal kit) with a different cartridge and reagent set used in conjunction with a - Thal method optimized for Hb A 2 and Hb F quantitation as well as the detection, identification, and quantitation of Hb variants. The use of the 2 kits on a single instrument requires changing the cartridge and converting the instrument from one reagent set to another, depending on which analyte is being measured. In our experience, this conversion is time consuming, requires extensive manual 1494 Arch Pathol Lab Med Vol 126, December 2002 Dual Hb A 2 /A 1c Kit Lafferty et al

2 n Table 1. Precision Analysis of Hb A 1c Quantitation by Hb A 1c Method* Hb A 1c Normal HB A 1c Increased Hb A 1c Normal HB A 1c Slightly Increased HB A 1c Increased purging of the pumping mechanism, and is a common cause of quality control failures and instrument malfunction. A new Variant II HbA 2 /HbA 1c Dual kit (dual kit) has been developed by Bio-Rad Laboratories (Munich, Germany) and is in use outside North America. This kit is capable of both Hb A 1c quantitation and hemoglobinopathy investigation using a single cartridge and reagent set. The only conversion required is selection of the Hb A 1c or -Thal method on the CDM system and calibration with specific calibrator sets. The purpose of this study was to evaluate the new dual kit for the quantitation of Hb A 1c,HbA 2, and Hb F, as well as the detection and quantitation of Hb variants compared with the current Hb A 1c and -Thal kits. MATERIALS A METHODS We used routine patient blood samples submitted for Hb A 1c quantitation or hemoglobinopathy investigation in the study. Blood samples were drawn into tubes containing K 2 EDTA (potassium ethylenediaminetetraacetic acid) (Becton Dickinson Vacutainer Systems, Franklin Lakes, NJ) using standard venipuncture techniques, shipped at room temperature, and analyzed within 4 hours of receipt in the laboratory. We used the following kits in the analysis: the Variant II Hemoglobin A 1c Reorder Pack, the Variant II -Thalassemia Short Program Reorder Pack, and the Variant II HbA 2 /HbA 1c Dual Program Reorder Pack. All 3 kits use cation exchange chromatography to separate Hb fractions. In all methods, the samples are mixed by the Variant II sampling station, diluted with the kitspecific hemolyzing/wash buffer, and injected onto the kit-specific analytic cartridge. The Variant II dual pumps deliver a programmed buffer gradient of increasing ionic strength to the cartridge, where the Hb fractions are separated based on their ionic interactions with the cartridge material. Buffer systems used are Bis-Tris/sodium phosphate buffer (current Hb A 1c kit), sodium phosphate buffer (current -Thal kit), and Bis-Tris/sodium phosphate buffer (dual kit). The separated Hb fractions pass through a flow cell where absorbance is measured at 415 nm. Calibration with method-specific calibrators is used to adjust the quantitation of the various Hb fractions being analyzed. The current Hb A 1c kit and dual kit use a 2-level calibration system, and the current -Thal kit uses a single-level calibrator. Raw data are integrated by the CDM system, and a chromatogram/sample report is generated. In all methods, a kit-specific CDM program is used depending on the analysis required (ie, Hb A 1c quantitation or hemoglobinopathy investigation). 1,6,7 We used the Hb A 1c Calibrator/Diluent set to calibrate Hb A 1c analysis for both the current and dual Hb A 1c methods. The calibration protocol is traceable to the Diabetes Complications Control Trial and the National Glycohemoglobin Standardization Program. We used the Hb A 2 /F Calibrator/Diluent set to calibrate Hb A 2 and Hb F quantitation for the current -Thal method, and we used the -Thal Calibrator set to calibrate Hb A 2,HbF,and Hb A 1c quantitation for the dual kit -Thal method. All methods were performed as described in the respective instruction manuals for each kit. 1,6,7 All analyses were performed on the Bio-Rad Variant II automated Hb analyzers using CDM version 3.11 integration software. All analyses were controlled with commercial Hb A 1c,HbA 2, and Hb F controls purchased from Bio-Rad Laboratories. Hemoglobin A 1c,HbA 2, and Hb F quantitation were each evaluated for within-run precision, between-run precision, linearity, and correlation with current methodology by linear regression. We conducted reference interval validation for Hb A 1c,HbA 2, and Hb F quantitation by analyzing samples from normal volunteer donors. Detection of common Hb variants Hb S, Hb C, Hb E, Hb Constant Spring, Hb Barts, and Hb H was evaluated on EDTA-anticoagulated whole blood specimens encountered during the evaluation period. Detection of rarer Hb variants (eg, Hb Lepore, Hb D Punjab, Hb G, and Hb J) was evaluated on frozen hemolysate specimens from the laboratory s Hb variant library. Correlation of Hb Barts quantitation was evaluated by linear regression between the laboratory s in-house cation exchange HPLC system and the dual kit using the -Thal method. We determined all lines of best-fit estimates by the least squares method. All data analysis was performed using Microsoft Excel 97 (Microsoft Corporation, Redmond, Wash). Reference intervals were validated in accordance with the guideline approved by the National Committee for Clinical Laboratory Standards. RESULTS Hb A 1c Quantitation (Hb A 1c Method) Precision results for Hb A 1c quantitation are shown in Table 1. Linearity results are illustrated in Figure 1. Linear regression results and line of best-fit estimate are shown in Figure 2. Reference interval validation values ranged from % to 6.2% (0.044 to 0.062). One individual with a Hb A 1c concentration of 6.2% (0.062) exceeded the present laboratory and manufacturer s recommended reference interval of 4.0% to 6.0% (0.040 to 0.060). Hb A 2,HbA 1c, Hb F, and Hb Barts Quantitation () Precision results for Hb A 2,HbA 1c, and Hb F quantitation are contained in Tables 2, 3, and 4, respectively. Linearity results for Hb A 2 /E are illustrated in Figure 3. Linearity results for Hb F are shown in Figures 4 and 5. Linear regression results and line of best-fit estimates compared with current methodology for Hb A 2,HbA 1c, Hb F, and Hb Barts quantitation are shown in Figures 6, 7,, and 9, respectively. Values for Hb A 2 reference interval validation ranged from 2.5% to 3.% (0.0 to 0.03). One individual with an Hb A 2 level of 3.% (0.03) exceeded the present laboratory reference interval of 1.% to 3.5% (0.01 to 0.035) and the manufacturer s recommended reference interval of 2.0% to 3.2% (0.02 to 0.032). Values for Hb F ranged from not detected () to 1.0% (0.010). No results exceeded the present laboratory reference interval of less than 1.2% (0.012) and the manufacturer s recommended reference interval of less than 1.4% (0.014). Arch Pathol Lab Med Vol 126, December 2002 Dual Hb A 2 /A 1c Kit Lafferty et al 1495

3 Figure 3. Hb A 2 /E quantitation linearity, dual kit -Thal method. Hb indicates hemoglobin. Figure 1. Hb A 1c linearity, dual kit Hb A 1c method. Hb indicates hemoglobin. Figure 2. Line of best-fit and linear regression results for Hb A 1c quantitation, current Hb A 1c kit versus the dual kit Hb A 1c method (n 6, line of best-fit estimate 0.97). Hb indicates hemoglobin. Detection of Hb Variants The results obtained in patients with the Hb S, Hb C, and Hb E variants encountered during the study are shown in Tables 5 and 6. Hemolysates with the following rare Hb variants were analyzed and showed abnormal peaks with the dual kit -Thal method: Hb Lepore, Hb D Punjab,HbI,HbO Padova,HbG Galveston,HbG,HbJ,HbH, Hb Constant Spring, and 3 unidentified -globin chain Hb variants. Two unidentified -globin chain Hb variants were run and were not detected by the dual kit -Thal method. COMMENT Hb A 1c Quantitation (Hb A 1c Method) TheHbA 1c results obtained with the dual kit using the Hb A 1c method showed excellent within-run and between- Table 2. n Precision Analysis of Hb A 2 Quantitation by the * Hb A 2 Normal HB A 2 Increased Hb A 2 Normal HB A 2 Increased HB A 2 Increased Table 3. n Precision Analysis of Hb A 1c Quantitation by the * Hb A 1c Normal HB A 1c Increased Hb A 1c Normal HB A 1c Normal HB A 1c Increased Arch Pathol Lab Med Vol 126, December 2002 Dual Hb A 2 /A 1c Kit Lafferty et al

4 Table 4. n Precision Analysis of Hb F Quantitation by the * Hb F Normal HB F Increased Hb F Normal HB F Slightly Increased HB F Markedly Increased Figure 4. Hb F quantitation linearity, normal to moderately increased range, dual kit -Thal method. Hb indicates hemoglobin. Figure 5. Hb F quantitation linearity, normal to markedly increased range, dual kit -Thal method. Hb indicates hemoglobin. run precision for both normal and increased Hb A 1c values. Coefficients of variation ranged from a low of 0.6% to a high of 3.5% (Table 1). Linearity from a normal value of 5.3% (0.053) to an increased value of 12.3% (0.123) was excellent (Figure 1). Hb A 1c values obtained on the dual kit correlated well with those obtained with the current Hb A 1c kit by linear regression (Figure 2) with a line of best-fit estimate of Reference interval validation correlated well with the reference interval used with the current Hb A 1c kit. The presence of Hb S, Hb C, or Hb E was readily detected by the dual kit using the Hb A 1c method. The presence of these common Hb variants in the heterozygous form had no effect on the Hb A 1c values obtained on this assay. In these cases, the Hb A 1c percentage is determined as a proportion of the Hb A detected. The variant and glycated variant peaks (eg, Hb S and Hb S 1c ) are excluded from the analysis. The current and dual methods are unable to measure Hb A 1c in patients with no Hb A (eg, patients with homozygous or compound heterozygous hemoglobinopathies) or in patients with increased levels of Hb F (ie, 30% [0.30]) or rare Hb variants that co-elute with Hb A 1c. 1,7 In our experience, these scenarios are easily recognizable and occur rarely (3 cases in analyses). Hb A 2 Quantitation () TheHbA 2 results obtained with the dual kit using the -Thal method showed good precision within the run and between runs. Coefficients of variation ranged from a low of % to a high of 6.0% (Table 2). Linearity from a normal value of 2.7% (0.027) to a high Hb A 2 /E value of.1% (0.1), obtained for a patient with Hb E trait, was acceptable (Figure 3). Hb A 2 quantitation showed acceptable correlation with the current -Thal kit by linear regression (Figure 6), although the line of best-fit estimate (0.9) was not as good as that obtained for Hb F. Reference interval validation correlated well with the reference interval used with the current -Thal kit. One healthy individual subsequently identified as having Hb C trait presented with a slightly increased Hb A 2 concentration of 3.% (0.03). In our experience, patients with Hb C trait can have falsely elevated Hb A 2 levels with the current method. This phenomenon is most likely due to contamination of the Hb A 2 peak with modified Hb C (eg, Hb C 1c ). This finding suggests that the same phenomenon occurs with the dual kit. Generally speaking, this is not a problem as the presence of significant amounts of Hb A (ie, 50% [0.50]) excludes the possibility of Hb C/ -thalassemia. In our experience, patients with Hb C/ -thalassemia present with Hb A 2 levels in excess of 5.2% (0.052). Hb A 1c Quantitation () One key difference between the current -Thal kit and the dual kit is that Hb A 1c results can be obtained when using the -Thal method in the latter. Hb A 1c values obtained this way showed good precision within the run and between runs. Coefficients of variation ranged from a low of 0.5% to a high of % (Table 3). Hb A 1c values obtained with the dual kit using the -Thal method showed ac- Arch Pathol Lab Med Vol 126, December 2002 Dual Hb A 2 /A 1c Kit Lafferty et al 1497

5 Figure 6. Line of best-fit estimate and linear regression results for Hb A 2 quantitation, current -Thal kit versus the dual kit -Thal method (n 77, line of best-fit estimate 0.9). Hb indicates hemoglobin. Figure 7. Line of best-fit estimate and linear regression results for Hb A 1c quantitation, dual kit Hb A 1c method versus the dual kit -Thal method (n 36, line of best-fit estimate 0.3). Hb indicates hemoglobin. Figure. Line of best-fit estimate and linear regression results for Hb F quantitation, current -Thal kit versus the dual kit -Thal method (n 32, line of best-fit estimate 0.99). Hb indicates hemoglobin. Figure 9. Line of best-fit estimate and linear regression results for Hb Barts quantitation, current in-house cation exchange HPLC versus the dual kit -Thal method (n 16, line of best-fit estimate 1.19). Hb indicates hemoglobin; HPLC, high-performance liquid chromatography. Table 5. Hb S, Hb A 2, and Hb F Results for Patients with Hb S; Versus the * Genotype Hb S disease (posttransfusion) Hb S/ -thalassemia Hb S disease Hb S disease Hb S disease * Hb indicates hemoglobin;, not detected. %HB S %HbA %HBF Genotype Hb C trait Hb C trait Hb C trait Table Hb E and Hb C Results; Versus the * %HbC Genotype Hb E trait Hb E trait Hb E trait Hb E/ -thalassemia (posttransfusion) Hb E trait (cord blood) %HbA 2 /E 149 Arch Pathol Lab Med Vol 126, December 2002 Dual Hb A 2 /A 1c Kit Lafferty et al

6 ceptable correlation with values obtained with the dual kit using the Hb A 1c method (Figure 7). Hb F Quantitation () The Hb F results obtained with the dual kit using the -Thal method showed acceptable precision within the run and between runs. Coefficients of variation ranged from a low of 0.0% to a high of 17.6% (Table 4). The high coefficients of variation of 7.9% and 17.6% were obtained on series that ranged from 1.0% to 1.2% (0.010 to 0.012) and 0.3% to 0.6% (0.003 to 0.006), respectively, and do not indicate unacceptable precision. Linearity from a normal Hb F value to a moderately increased Hb F value of 15.3% (0.153) and a markedly increased Hb F value of 7.9% (9) was excellent (Figures 4 and 5). Hb F values obtained with the dual kit using the -Thal method showed excellent correlation by linear regression (Figure ) with a line of best-fit estimate of Reference interval validation correlated well with the reference interval used with the current -Thal kit. Hb Barts Quantitation () A second key difference between the current -Thal kit and the dual kit is the ability to quantify Hb Barts on cord or neonatal blood samples with the latter. We analyzed 16 cord blood samples that showed evidence of Hb Barts with the dual kit using the -Thal program. Results obtained on the dual kit correlated fairly well with those obtained using an in-house cation exchange HPLC method (Figure 9). Results from the dual kit did tend to be higher as the percentage of Hb Barts increased. Red cells used for the in-house method are washed before analysis, whereas cells used in the dual kit are not washed. The slightly increased values obtained with the dual kit may be the result of plasma contaminants or modified Hb F. In spite of this, the values correlate reasonably well. Even small amounts of Hb Barts (eg, 1.0% to 2.0% [0.010 to 0.020]) in newborn samples can be indicative of a deletional form of -thalassemia; therefore, the ability to detect and quantify Hb Barts improves the utility of this method for newborn screening programs. 9 Cord blood specimens were difficult to run on the dual kit due to the high levels of Hb F, which tended to overwhelm the integration software. This resulted in an inability to identify the Hb F peak and, in random cases, software errors that necessitated aborting the run. Detection of Hb Variants () Common Hb variants (eg, Hb E, Hb S, and Hb C) were readily detected by the dual kit using the -Thal program. Quantities of the variants detected compared well between the current kit and the dual kit (Tables 5 and 6); however, the retention times differed ( -Thal kit: Hb E, 3.65 minutes; Hb S, 4.54 minutes, and Hb C, 5.13 minutes; dual kit: Hb E, 2.90 minutes; Hb S, 3.50 minutes; and Hb C, 4.50 minutes). 6,7 With the exception of the Hb E problems discussed subsequently, the retention time differences posed no difficulty as the variant peaks were clearly identified by the CDM software. A third key difference between the present kit and the dual kit is the ability to avoid falsely elevated Hb A 2 values in the presence of Hb S with the latter approach. All patients with Hb S had elevated Hb A 2 values with the current method, whereas, with the exception of one unexplained case of, all patients without concomitant -thalassemia had normal Hb A 2 values with the dual kit. The one case of Hb S/ -thalassemia had an elevated Hb A 2 level with the dual kit. A problem was noted with the identification and integration of Hb E. With the current kit, Hb E is identified by the presence of a markedly increased Hb A 2 peak. With the dual kit, the increased Hb A 2 peak was present, but the program was unable to identifyitashba 2. Analysis of fresh and frozen samples with rarer variants showed that the dual kit is able to detect other important Hb variants (eg, Hb Lepore, Hb D Punjab, Hb H, and Hb Constant Spring). In addition, the dual kit was able to detect a variety of rare -globin and -globin Hb variants. A minor concern was the inability to detect the 2 rare -globin chain Hb variants encountered. Although clinically insignificant, the detection of these variants is sometimes helpful in identifying patients with concomitant -thalassemia trait in whom the Hb A 2 level is falsely low because half of the Hb A 2 is contained in the undetected -globin chain variant. In summary, the dual kit provides precise Hb A 1c,Hb A 2, and Hb F quantitation. The results show good linearity and correlate well with the current Hb A 1c and -Thal kits. There is no appreciable difference in specimen throughput between the current kits and the dual kit using either the Hb A 1c or -Thal methods. The kit detects all clinically important Hb variants and a wide variety of rare variants. We noted a number of minor problems with the dual kit. The difficulty in integrating the Hb A 2 peak in patients with Hb E and the difficulty in running cord blood specimens should be fixable with minor software adjustments. The inability to detect -globin chain variants has the potential, in rare circumstances, to mask -thalassemia trait. The dual kit has several distinct advantages over the present methodology. First, it eliminates the need for a timeconsuming and problem-inducing switch over between thehba 1c and -Thal kits. A switch over with the dual kit requires a matter of 5 to 10 minutes as opposed to 1 to 2 hours with the current kits. Second, the ability to detect and quantify Hb Barts improves the utility of this technology for newborn screening programs. Third, the ability to quantify Hb A 1c using the -Thal method may be of value in diabetic patients being evaluated for hemoglobinopathies. Finally, the ability to accurately quantify Hb A 2 in the presence of Hb S eliminates the need to have a separate Hb A 2 reference interval for patients with this variant. 10 The authors acknowledge the following individuals for their support in conducting this study: Linda Halchuk, Judith Ireland, Mila Vacovsky, and the Provincial Hemoglobinopathy Laboratory; Josephine Baldwin, Research Coordinator, and Dr M. J. McQueen, Hamilton Regional Laboratory Medicine Program; and Greg Coburn, Tom St Amant, and Arthur Wilkins, Bio-Rad Laboratories, Mississauga, Ontario. References 1. Variant II Hemoglobin A 1c Program Instruction Manual. Hercules, Calif: Bio- Rad Laboratories; March Dacie JV, Lewis SM. Investigation of abnormal haemoglobins and thalassemia. In: Practical Haematology. th ed. Edinburgh, Scotland: Churchill Livingstone; 1995: Lafferty J. Good Practice Guidelines for the Laboratory Investigation of Hemoglobinopathies. QMP-LS Document. Toronto, Ontario: Quality Management Program Laboratory Services, Hematology-General Binder. October 2001;3: Riou J, Godart C, Hurtrel D, et al. Cation-exchange HPLC evaluated for presumptive identification of hemoglobin variants. Clin Chem. 1997;43(1): Campbell M, Henthorn J, Davies S. Evaluation of cation-exchange HPLC Arch Pathol Lab Med Vol 126, December 2002 Dual Hb A 2 /A 1c Kit Lafferty et al 1499

7 compared with isoelectric focusing for neonatal hemoglobinopathy screening. Clin Chem. 1999;45(7): Variant II -thalassemia Short Program Instruction Manual. Hercules, Calif: Bio-Rad Laboratories; July Variant II HbA 2 /HbA 1c Dual Program Instruction Manual. Munich, Germany: Bio-Rad Laboratories; February National Committee for Clinical Laboratory Standards. How to Define and Determine Reference Intervals in the Clinical Laboratory; Approved Guideline. 2nd ed. Wayne, Pa: NCCLS; 2000;15(4):23. NCCLS document C2-A. 9. Higgs DR, Vickers MA, Wilkie AOM, et al. A review of the molecular genetics of the human alpha-globin gene cluster. Blood. 199;73: Shokrani M, Terrell F, Turner E, del Pilar Aguinaga M. Chromatographic measurements of hemoglobin A 2 in blood samples that contain sickle hemoglobin. Ann Clin Lab Sci. 2000;30(2): Arch Pathol Lab Med Vol 126, December 2002 Dual Hb A 2 /A 1c Kit Lafferty et al