A Liquid-Stable Reagent for Lactic Acid Levels Application to the Hitachi 911 and Beckman CX7

Transcription

1 Clinical Chemistry / LACTIC ACID DETERMINATION A Liquid-Stable Reagent for Lactic Acid Levels Application to the Hitachi 911 and Beckman CX7 Joseph D. Artiss, PhD, 1,2 Raymond E. Karcher, PhD, 3 Kevin T. Cavanagh, PhD, 4 Sandra L. Collins, MT(ASCP), 3 Valerie J. Peterson, MBA, 3 Sudhir Varma, MD, 3 and Bennie Zak, PhD 1 Key Words: Stable; Lactic acid; Serum; Plasma; Lactate oxidase; Hitachi 911; Beckman CX7 Abstract We evaluated the use of a new lactate oxidase based reagent for the determination of serum and plasma lactic acid levels with the Hitachi 911 (Roche Diagnostics, Indianapolis, IN) and the Beckman CX7 (Beckman Instruments, Brea, CA). Evaluation studies demonstrated on-board stability of at least 3 months and a calibration stability of more than 5 months. Within- and between-day imprecision of this reagent was less than 2% for both applications. The reagent is free of the deleterious effects of triglyceride up to levels of 1,4 mg/dl (15.8 mmol/l), bilirubin to concentrations of 24.6 mg/dl (42 µmol/l), and hemoglobin, from lysed erythrocytes, to levels of more than.3 g/dl (3. g/l). When used on the Hitachi 911 for the determination of plasma lactate concentrations, the reagent correlates with the Dade aca III (Dade International, Deerfield, IL). When applied to the Beckman CX7 for the determination of serum lactate levels, the method correlates with the Beckman method. Traditionally, venous lactic acid concentrations have been used as an indicator of hypoxia or lactic acidosis caused by decreased perfusion or increased production of lactate. 1 The observation that lactate levels are elevated in patients who have slipped into shock has led to an increased demand for this test. In addition, Schmiechen et al 2 reported that elevated lactate levels obtained from patients in the emergency department, in conjunction with electrocardiographic findings, have a strong predictive value for acute myocardial infarction (sensitivity, 96%; specificity, 55%). The negative predictive value was 98%. Furthermore, they determined that hyperlactatemia correlates with mortality and the need for intensive care management. Lactate concentrations also have been demonstrated to have predictive value for the development of shock following a myocardial infarction. 3 As well, lactate concentrations have been demonstrated to have prognostic value in patients with septic shock, 4 inferior vena cava injuries, 5 and in children following cardiac surgery. 6,7 These more recent reports may lead to further increases in the demand for lactic acid testing. Considering how critically ill the patients are, we should expect that these requests will almost certainly be stat. Until recently, most commercially available reagents for the determination of lactic acid concentrations involved the nicotinamide adenine dinucleotide (NAD)-coupled oxidation of lactate to pyruvate, as follows (NADH is the reduced form of NAD): Lactate L-Lactate + NAD + Pyruvate + NADH + H + Dehydrogenase Am J Clin Pathol 2;114:

2 Artiss et al / LACTIC ACID DETERMINATION Reagents based on this reaction sequence have adequate sensitivity for serum lactate measurements but have relatively poor open-bottle or on-board stability. Furthermore, because an ammonium-based buffer is often used for the lactate dehydrogenase enzyme preparation, the potential for contamination on instruments that reuse cuvettes and also are used to measure ammonia necessitates that analysis for lactate be batched. 8 (While this manuscript was in preparation, Roche Diagnostics, Indianapolis, IN, began to offer a lactic acid reagent analogous to the PSI reagent. 9 Consequently, there is no need to batch these determinations, if the new reagent is used.) Both of these circumstances have led to the wastage of reagent and labor. Furthermore, it is difficult to fulfill stat requests if the testing must be batched. A lactate oxidase based reagent for the determination of plasma and cerebral spinal fluid lactic acid levels has been reported. 1 The principle on which the current assay is based is similar to that reported by Bozimowski et al. 1 The earlier approach has been modified to incorporate N-ethyl- N-(2-hydroxy-3-sulfopropyl)-3-methylaniline (TOOS) as a cosubstrate for peroxidase with 4-aminoantipyrene (AAP). This substitution effectively reduces the sensitivity of the indicator reaction, thus extending the dynamic range of the assay. The reaction sequence is as follows: L-lactate + O 2 Lactate Oxidase As this reagent does not contain ammonia, it is more applicable to random access type equipment. Our early experiences with this reagent also suggested that it exhibits much greater stability than lactate dehydrogenase based reagents. We established operating parameters and evaluated the performance of this reagent on 2 of the more commonly available clinical chemistry random access analyzers. Materials and Methods Pyruvate + H 2 O 2 H 2 O 2 + TOOS + AAP Peroxidase Chromogen (555 nm) Venous serum or plasma was collected in red- or graystoppered tubes, respectively, according to standard laboratory procedures, and the gray-stoppered tubes were transported on wet ice to the laboratory. All specimens were centrifuged at 3,g for 5 minutes, and the plasma was separated from the cells as quickly as possible. For the purposes of this study, all samples were taken from leftover specimens that had been stored at 4 C for 1 week and were destined for disposal. The Dade aca III (Dade International, Deerfield, IL), Hitachi 911 (Roche Diagnostics), and Beckman CX7 (Beckman Instruments, Brea, CA) reagents were used as described by the respective manufacturers. The Pointe Scientific lactate reagent (Pointe Scientific Inc, Lincoln Park, MI) (PSI) is supplied as a 2-part liquid-stable reagent identified as R1 and R2. As reported by the manufacturer, R1 contains the following per liter: 1 mmol of tris(hydroxymethyl)aminomethane (TRIS) buffer, 1.7 mmol of AAP, more than 1 ku of peroxidase (horseradish), 9 mg of sodium azide, surfactant, and stabilizer. As reported by the manufacturer, R2 contains the following per liter: 1 mmol of TRIS buffer, more than 1 ku of lactate oxidase (microbial), 1.5 mmol of TOOS, 9 mg of sodium azide, surfactant, and stabilizer. The reagents were used as supplied by PSI without modification. Although these reagents are available in packaging that is compatible with the Hitachi 911, they were purchased in bulk for the Beckman CX7. We filled the Beckman CX7 user-defined reagent cartridge with 24 ml of PSI Lactate Oxidase reagent (R1) in compartment A, 12 ml of AAP reagent (R2) in compartment B, and 4 ml of AAP reagent (R2) in compartment C. This approach provided the maximum number of tests given the volume restrictions of the various compartments. The parameters that we developed for the application of these reagents to the Hitachi 911 and Beckman CX7 are available from PSI. Control materials, Liquicheck levels I and II, were purchased from Bio-Rad Diagnostics Group (Hercules, CA). Linearity of the reagent on the Hitachi 911 and Beckman CX7 was evaluated by making serial dilutions of a 18-mg/dL (2-mmol/L) concentration of lactate calibrator. Specimens saved for patient comparison studies were stored frozen at 2 C and assayed concurrently as a batch by both methods. Specimens used for interference studies were not originally lactate requests but were left over from routine testing and were selected for their elevated levels of potentially interfering substances. For this reason, the lactate levels of these samples are all within the reference range. Results and Discussion As initial accelerated stability studies indicated that this reagent was very stable, we wanted to evaluate its onboard stability. After routine calibration, the performance of the first Hitachi 911 was monitored daily with 2 levels of commercial control material, BioRad Liquicheck levels I and II. Figure 1 shows the results of this study. The Hitachi 911 had sufficient reagent storage capacity to allow us to operate for approximately 4 months, at which time a new lot of reagent was placed on the instrument without 14 Am J Clin Pathol 2;114:

3 Clinical Chemistry / ORIGINAL ARTICLE Lactate Concentration (mmol/l) Mean = 4.6 mmol/l SD =.11 mmol/l CV = 2.3% N = Mean = 1.12 mmol/l.5 SD =.5 mmol/l CV = 4.2%. N = 135 5/1/98 6/1/98 7/1/98 8/1/98 Date August 6, 1998 New lot of reagent Instrument was not calibrated 9/1/98 1/1/98 Figure 1 Levy-Jennings type plot of lactate results on 2 levels of control material during a 5-month period on the Hitachi 911 (Roche Diagnostics, Indianapolis, IN). The data include a change of lot number on August 6, 1998, that did not require recalibration of the instrument. To convert millimoles per liter to milligrams per deciliter, divide by.111. recalibration. These data show that without recalibration the instrument-reagent system was sufficiently stable to yield SDs of.99 and.45 mg/dl (.11 and.5 mmol/l) at lactate levels of 41.4 and 9.9 mg/dl (4.6 and 1.1 mmol/l, respectively; reference range, mg/dl [ mmol/l]). It is important to note that this study was conducted during the period of 5 months with 2 different lot numbers of reagent. We are not, however, recommending that reagent be placed onto any instrument without calibrating the system. The total time required for each test is dictated to some extent by the nuances of the automated equipment being used. In the package insert, PSI recommends a total incubation time Table 1 Imprecision * Level I of 5.5 minutes. This is consistent with our findings that the total measurement time is 7.5 minutes on the Beckman CX7 and 5 minutes on the Hitachi 911. This amount of time fits well with other routine tests that are likely to be requested at the same time. Because of its longer stability and larger on-board reagent container volume, the PSI method extends the operating time between reagent refills and, therefore, calibrations on the Beckman CX7 compared with the Beckman application. Filling the Beckman user-defined container according to the protocol described herein provides enough reagent for a total of 16 determinations, whereas the Beckman application provides only 35 tests per reagent-container refill. Beckman s reagent, in contrast with the proposed reagent, is stable for only 14 days, and the method requires calibration every 7 days. The PSI method seems to require calibration only at the time reagents are replenished, which is dictated by the Beckman CX7 operating protocols, thus reducing reagent consumption and related costs. Additional imprecision data were collected for the Hitachi 911 and the Beckman CX7 instruments in accordance with the National Committee for Clinical Laboratory Standards protocol. 11 These data are given in Table 1. Two Hitachi 911s in different laboratories demonstrated linearity to at least 18 mg/dl (2 mmol/l) of lactate. The Beckman CX7 was linear to at least 135 mg/dl (15 mmol/l) of lactate. Concern has been raised about the use of samples spiked with total parenteral nutrition for the evaluation of the effects of elevated lipids on instrument and reagent systems. 12 In our experience, we have found that the effects of bilirubin on peroxidase-coupled reactions are very complex and that the practice of spiking samples with purified bilirubin is prone to producing misleading results. Likewise, by concentrating on the possible deleterious effects of hemoglobin, we might overlook the fact that there are a host of materials, including peroxide-consuming catalase, that are Level II Within Day Between Day Within Day Between Day Hitachi 911 (n = 4) Mean (mmol/l) SD (mmol/l) CV (%) Beckman CX7 (n = 4) Mean (mmol/l) SD (mmol/l) CV CV, coefficient of variation. * The n is the number of tests. Hitachi 911, Roche Diagnostics, Indianapolis, IN, and Beckman CX7, Beckman Instruments, Brea, CA. n = 39. Am J Clin Pathol 2;114:

4 Artiss et al / LACTIC ACID DETERMINATION released into the sample when erythrocytes are lysed. For these reasons, we have chosen to perform our interference studies by collecting normal pools of leftover serum and making serial dilutions with samples that contain high levels of a potential interfering substance. If the dilutions fall on a straight line, it may be assumed that there is no detectable interference by the substance in question. Conversely, if at some point the line deviates from linearity, we can predict the concentration at which interference will occur. It is, however, necessary to assume that there is no proportional error to be found at low concentrations of the potential interfering substance, ie, that at very low concentrations of the potentially interfering substances, there is absolutely no interference. This assumption is probably reasonable with these 3 materials. We found no interference with triglyceride to levels of 1,4 mg/dl (15.8 mmol/l), bilirubin to concentrations of 24.6 mg/dl (42 µmol/l), and hemoglobin, from lysed erythrocytes, to levels of more than.3 g/dl (3. g/l). The data obtained from these studies also may be used in a manner similar to a recovery study. That is, the theoretical lactate concentration of each dilution may be calculated based on the measured concentrations of the low and the high pool samples. The measured lactate concentration then may be compared with the calculated value. If there is interference in the high samples, the recovery data will look poor (<95%, >15%). All of the dilutions that we studied demonstrated recoveries that ranged from 99.9% to 1.1%. Dopamine has been reported to cause a negative interference with peroxidase-coupled reactions. 13,14 Although we have not studied this hormone or drug as a potential interfering substance, we would predict, based on reaction conditions and similar sample/reagent ratios, that similar effects would be experienced with this assay. However, we do not believe that dopamine interference should be a concern, as the lowest levels seen to cause interference in these 2 studies were 5, 14,15 to 4, 13,15 times the physiologic level. The correlation data with plasma samples, ranging in lactate concentration from 2.7 to 93.7 mg/dl ( mmol/l), on the Dade aca III are shown in Figure 2. The correlation data for serum samples determined on the Beckman CX7 and Hitachi 911 are shown in Figure 3. Since each reagent was calibrated with its supplied calibrator, we looked to these materials as sources of the observed proportional bias. In running the various calibrators as samples, we found that as much as 4% of the bias may be accounted for by small differences in the calibrators. Conclusions In our experience, the PSI reagent proved very stable relative to the more common lactate dehydrogenase based PSI 911 Plasma Lactate Concentration (mmol/l) Dade aca Plasma Lactate Concentration (mmol/l) Figure 2 Plasma lactate correlation between the Dade aca (Dade International, Deerfield, IL) and the Hitachi 911 (Roche Diagnostics, Indianapolis, IN) with the Pointe Scientific (PSI, Lincoln Park, MI) lactate reagent. Linear regression analysis performed on 57 measurements yielded the following: Hitachi/PSI =.965aca +.1 and r 2 =.996. To convert millimoles per liter to milligrams per deciliter, divide by.111. PSI Lactate Concentration (mmol/l) CX7 Lactate Concentration (mmol/l) Figure 3 Serum lactate concentration correlation data for 19 samples with the Hitachi 911 (Roche Diagnostics, Indianapolis, IN) with the Pointe Scientific (PSI, Lincoln Park, MI) lactate reagent vs Beckman CX7 (Beckman Instruments, Brea, CA) (circle); Beckman CX7 with PSI lactate reagent vs Beckman CX7 (triangle); and Hitachi 911 with PSI lactate reagent vs Beckman CX7 with PSI lactate reagent (square). Linear regression analysis performed on 19 measurements yielded the following: Hitachi/PSI reagent =.83Beckman CX7 +.26, r 2 =.992; CX7/PSI reagent =.89Beckman CX7 +.1, r 2 =.996; and Hitachi/PSI reagent =.93CX 7/PSI reagent +.2, r 2 =.994. To convert millimoles per liter to milligrams per deciliter, divide by Am J Clin Pathol 2;114:

5 Clinical Chemistry / ORIGINAL ARTICLE reagent. As there is no ammonium-based buffer used in this reagent, we encountered no carryover problems with the Hitachi 911. Both of these features, as well as the packaging of the reagent, made it convenient to use and apply to the Hitachi 911 and the Beckman CX7. Imprecision, both within- and between-day, was acceptable. The PSI reagentinstrument systems seemed to be free of interference from triglyceride to levels of at least 1,4 mg/dl (15.8 mmol/l), bilirubin to levels of at least 24.6 mg/dl (42 µmol/l), and hemolysis to hemoglobin levels of at least.3 g/dl (3. g/l). When applied to the Hitachi 911 and Beckman CX7, the reagent correlates well with the instrument manufacturer s reagent with serum samples, and the Hitachi 911 application correlates well with the Dade aca III when comparing plasma samples. The ease with which we applied the PSI reagent to these 2 instruments suggests that it should be readily applicable to many of the commonly available clinical analyzers. From the 1 Department of Pathology and 2 Center for Molecular Genetics, Wayne State University School of Medicine, Detroit, MI; 3 Clinical Pathology Department, William Beaumont Hospital, Royal Oak, MI; and 4 Clinical Laboratory, Ingham Regional Medical Center, Lansing, MI. Supported in part by Pointe Scientific, Lincoln Park, MI. The reagent described herein was developed at Wayne State University and was transferred to Pointe Scientific, under the terms of a licensing agreement. Dr Artiss provides consulting services for Pointe Scientific. Address reprint requests to Dr Artiss: Wayne State University School of Medicine, 54 E Canfield, Detroit, MI References 1. Sacks DB. Carbohydrates. In: Burtis CA, Ashwood ER, eds. Tietz Textbook of Clinical Chemistry. 2nd ed. Philadelphia, PA: Saunders; 1986: Schmiechen NJ, Han C, Milzman DP. ED use of rapid lactate to evaluate patients with acute chest pain. Ann Emerg Med. 1997;3: Mavric Z, Zaputovic L, Zagar D, et al. Usefulness of blood lactate as a predictor of shock development in acute myocardial infarction. Am J Cardiol. 1991;67: Bernardian G, Pradier C, Tiger F, et al. Blood pressure and arterial lactate level are early indicators of short-term survival in human septic shock. Intensive Care Med. 1996;22: Rosengart MR, Smith DR, Melton SM, et al. Prognostic factors in patients with inferior vena cava injuries. Am Surg. 1999;65: Duke T, Butt W, South M, et al. Early markers of major adverse events in children after cardiac operations. J Thorac Cardiovasc Surg. 1997;114: Siegel LB, Dalton HJ, Hertzog JH, et al. Initial postoperative serum lactate levels predict survival in children after open heart surgery. Intensive Care Med. 1996;22: Hitachi 911 Analyzer Application Code 117 [product application]. Indianapolis, IN: Boehringer Mannheim; Dorn AR, Hunt CJ, Mountain LD. Evaluation of a liquid lactate reagent for use on the BM/Hitachi series analyzers [abstract]. Clin Chem. 1998;44:A Bozimowski D, Artiss JD, Zak B. Sensitive determination of cerebrospinal fluid pyruvate, lactate and glucose concentrations. Clin Chim Acta. 1985;153: Kennedy JW, Carey RN, Coolen RB, et al. Evaluation of Precision Performance of Clinical Chemistry Devices. 2nd ed. Wayne, PA: National Committee for Clinical Laboratory Standards; NCCLS Document EP5-T2, Vol 12(4). 12. Cobbaert C, Tricarico A. Different effect of Intralipid and triacylglycerol rich lipoproteins on the Kodak Ektachem cholesterol determination. Eur J Clin Chem Clin Biochem. 1993;31: Karon BS, Daly TM, Scott MG. Mechanisms of dopamine and dobutamine interference in biological tests that use peroxide and peroxidase to generate chromophore. Clin Chem. 1998;44: Koprowicz KT, Ooi DS, Donnelly JG. Influence of dopamine on peroxidase-based assays [letter]. Clin Chem. 1996;42: Matsen JM, Kjeldsberg CR, Ash KO, et al. User s Guide. Salt Lake City, UT: ARUP Laboratories; 1996:216. Am J Clin Pathol 2;114: