Fast Quantification of Ethanol in Whole Blood Specimens by the ...

65 downloads 27 Views 536KB Size Report
The linear range increased and reaction end point decreased with increasing NAD .... of a blank blood specimen and a blank aqueous specimen were the other ...
Journal of Analytical Toxicology,Vol. 29, January/February2005

[ TechnicalNote

FastQuantification of Ethanolin Whole Blood Specimensby the EnzymaticAlcohol Dehydrogenase Method. Optimization by ExperimentalDesign Lena Kristoffersen1,*, Bjorn Skuterud 1, Bente R. Larssen1, Svetlana Skurtveit 2, and Anne Smith-Kielland 1 1Norwegian Instituteof Public Health Division of Forensic Toxicologyand Drug Abuse, P.O. Box 4404 Nydalen, N-0403 Oslo, Norway and 2Norwegian Instituteof Public Health Division of Epidemiology, P.O. Box 4404 Nydalen, N-0403 Oslo, Norway

Abstract ] A sensitive,fast, simple, and high-throughput enzymatic method

for the quantification of ethanol in whole blood (blood) on Hitachi 917 is presented. Alcohol dehydrogenase(ADH) oxidizes ethanol to acetaldehyde usingthe coenzyme nicotinamide adenine dinucleotide (NAD), which is concurrently reduced to form NADH. Method development was performed with the aid of factorial design, varying pH, and concentrationsof NAD § and ADH. The linear range increasedand reaction end point decreased with increasing NAD § concentration and pH. The method was linear in the concentration range 0.0024-0.4220 g/dL. The limits of detection and quantification were 0.0007 g/dL and 0.0024 g/dL, respectively. Relative standard deviations for the repeatability and within-laboratory reproducibility were in the ranges 0.7-5.7% and 1.6-8.9%, respectively. The correlation coefficient when compared with headspace gas chromatography-flame ionization detection methods was 0.9903. Analysis of authentic positive blood specimens gave resultsthat were slightly lower than those of the reference method.

Introduction National Institute of Forensic Toxicology,now named Norwegian Institute of Public Health Division of Forensic Toxicology and Drug Abuse (NIPH), receives blood specimens related to suspected driving under the influence of drugs and alcohol cases. In 2003, about 42% of the received specimens were ethanol negative (n = 7400). GC is the reference method for determination of ethanol, other alcohols and volatile substances in forensic laboratories (1-3), but it is time-consuming and expensive. Rapid enzymatic ethanol assays are used in clinical (4-7), forensic (8) and research (9) laboratories and were first described by Bonnichsen and Theorell (10) and Bficher and Redetzki (11) in 1951. ADH oxidizes ethanol to * Author to whom correspondence should be addressed. E-mail: [email protected],

66

acetaldehyde using the coenzyme NAD§ which is concurrently reduced to form NADH: ADH Ethanol + NAD+ ~ Acetaldehyde + NADH

Eq. 1

NADH is measured spectrophotometrically at 340 nm. To force the reaction towards completion, semicarbazide is added, reacting with acetaldehyde to form semicarbazone. The commercial enzymatic assays used in the clinical laboratory generally have a high limit of quantitation (LOQ) and/or are not developed for analysis of whole blood, which is the standard matrix in forensic analysis (4-7). The objectiveof this study was to adapt the system of Eq. 1 for use in forensic analysis on the Hitachi 917 instrument (Roche, Mannheim, Germany). This instrument is already used for drug screening in our laboratory. Variables like NAD§ ADH, and semicarbazide concentrations, temperature, and pH influence the enzymatic reaction (10,12). Using factorial design (13), variables can be changed systematically to extract the maximum amount of information in the smallest number of experimental runs. In this paper, we present a sensitive, fast, simple, and highthroughput assay (200 specimens/h) for quantification of ethanol in blood in the concentration range 0.0024-0.4220 g/dL. Full factorial design has been used in the method development. The presented method allows quantification of ethanol in forensic as well as in clinical specimens. The method is not specific for ethanol and a positive result should always be confirmed with GC when used for forensic purposes.

Materials and Methods Reagents and solutions

ADH from yeast was obtained from Roche (Roche Diagnostics GmbH, Mannheim, Germany). NAD§ from yeast (97-99%)

Reproduction(photocopying)of editorialcontentof thisjournalis prohibitedwithoutpublisher'spermission.

Journal of Analytical Toxicology, Vol. 29, January/February 2005

was purchased from Sigma Chemical (St. Louis, MO), and pro analysi (p.a.) semicarbazid from Ftuka (Steinheim, Switzerland). Tetra sodium diposphate decahydrate, glycine, sodium hydroxide (NaOH), and perchloric acid (PCA) (all p.a.), and ethanol 99.8% were obtained from Merck (Darmstadt, Ge many). Deionized water (< 5~tS/cm)was used. Blood with the addition of sodium fluoride (7 g NaF/L) was obtained from The Blood Centre at Ullevaal University Hospital (Oslo, Norway). It was screened for ethanol and other alcohols to ensure alcohol free specimens prior to spiking. Forensic antemortem blood specimens were received in 5-mL Vacutainer | vials containing 20 mg Na-fluorid and Na-heparin 75 USP units (BD Vacutainer Systems, Plymouth, England). A stock buffer with pH 8.8 containing 0.092 M tetra sodium diphosphate, 0.075M semicarbazide, and 0.13M glycine was used to prepare reagents. In addition a 0.031M NAD§ stock solution was used to prepare the Hitachi reagent 1 (R1). All solutions were stored at 4~ for up to 2 weeks. R1 was made by mixing buffer/NAD+ (0.935:0.0645), and Hitachi reagent 2 (R2) was made by mixing buffer/ADH (0.62:0.38). R1 and R2 were prepared immediately before analysis.

Instrumentation and ethanol analysis The instrument utilizes a dual wavelength principle; the primary and secondary wavelengths in this method were 340 nm (reaction) and 505 nm (non-reaction), respectively. The reaction template of the Hitachi 917 is presented in Figure 1, absorbance being read about every 18 seconds, for 10 min. Temperature was 37~ An extra alkaline washing step with 4% NaOH of the specimen probe was used. Aliquots of 100-1JL specimen (blood or aqueous standards/controls) were stored in barcoded, capped tubes (12-mm • 75-mm Neutrex Disposable culture tubes, Scherf, MeiningenDreissigacker, Germany). Before analysis the aliquots were mixed with 900 I~L0.38M PCA (Hamilton dispenser HA4, Microlab | plus 1000, Hamilton, Bonaduz, Switzerland) on a Whirlimixer and then centrifuged at 3000 rpm (1670 g) for 10 rain (Megafuge 2.0, Heraeus Sepatech GmbH, Osterode/Harz, Germany). The instrument collected a 12-/.tLaliquot of the supernatant, which was automatically mixed with 270 IlL of R1. About 1.5 rain after addition of R1, 40 I~L R2 was added (reaction solution) and the reaction was followed for 8.5 min (Figure 1). The absorbance difference between the primary and secondary wavelength was registered. A calibration curve was established daily by analyzing a blank aqueous specimen and an aqueous calibration standard at the ethanol concentration of 0.1559 g/dL, with two replicates at each calibration point. The procedures gave a good separation between the initial and final absorbance (e.g., A 1.9000 absorbance units for a blood specimen spiked to 0.4115 g/dL). Aqueous control specimens with ethanol concentrations of 0.0046 and 0.0252 g/dL were analyzed automatically after the calibration and then for every 20th specimen. Also, two blood controls and two aqueous controls with ethanol concentrations of 0.0000 (bl), 0.0074 (bl), 0.1008 (aq) and 0.3982 (aq) g/dL were analyzed at the start and end of the analysis. External quality control specimens were analyzed monthly.

Experimentaldesign Temperature and semicarbazide concentration were kept constant. Three variables (NAD§ ADH, and pH) and four responses were evaluated in a full factorial design 23 (Table I). The correlation coefficients (r2) for the calibration curves were calculated by least-squares method with aqueous ethanol standards in the concentration range 0.0000-0.3982 g/dL (8 concentration levels). The reaction end point was defined as the reading point on the reaction curve that had an absorbance (primary-secondary) that was 2% smaller than the absorbance in the last reading point (no. 34), using an aqueous ethanol standard of 0.3294 g/dL. The measured ethanol concentration of a blank blood specimen and a blank aqueous specimen were the other responses. A false-positive result was defined as a result above the LOQ for blood. The factorial design experiments were carried out in randomized order. Three center points experiments were used (Exp. No. 9-11). The confidence level was 95%. The response models were obtained with multiple linear regression and validated by cross-validation. Experimental design analysis was performed by Unscrambler 7.8 software from Camo Ltd. (Oslo, Norway).

Method validation The limit of detection (LOD) and L0Q were determined by

analyzing 5 different ethanol-negative blood specimens on 10 successive days and were set to the mean of the negative specimens plus 3 and 10 standard deviations (sd), respectively. Linearity was evaluated by the correlation coefficient obtained from the center points in the factorial design. Repeatability was estimated analyzing spiked blood (n = 10) specimens (0.0074 and 0.2179 g/dL) and aqueous (n = 10) specimens (0.0046, 0.1008, and 0.3982 g/dL) in one run. Within-laboratory reproducibility was estimated by analyzing the same specimens (n = 5) on 10 successive days, one replicate each day. Accuracy was defined as the percent difference between measured and theoretical concentration. Specificity was determined by spiking blank blood to a concentration of 0.23-0.24 g/dL of acetone, methanol, 1-butanol, 2-propanol, or 1-propanol. The apparent ethanol concentration was measured. Specimens were stored uncapped in the instrument. The evaporation of ethanol during a run of 200 successive spec1

340 nm

&9 = ,I

Q (,1

Add&mix Reagent 2

A&&A&AA&AAA&

AAA&A&&&&

34

9 9

"= AdO,mix9149 ~9149 o460eooooeeoeoooo9