# Blood Sample Analysis - Is the Result Based on Whole Blood or Serum?

Posted by attorney Patrick Mahaney

Statutory limits in Alabama for ethanol in the blood are set at 0.02 g %, 0.04 g %, 0.05 g % and 0.08 g % in whole blood, depending on whether the state's "zero tolerance" law for under 21 years of age, the .04% limit set for commercial vehicle operators, the threshold level of .05% set in the Chemical Test for Intoxication Act, or the .08% per se level established in the DUI statute is the issue before the court. Therefore, one must consider what the whole blood ethanol concentrations would have been even if another specimen was the basis for the analytical result. The most common alternate specimen for ethanol analyses is serum, which is typically analyzed in hospital and clinical laboratories. Because there is a predictable and measurable difference in water content between serum and whole blood, a result in serum may be “converted" into an equivalent whole blood result. Conversion requires a change of units from mg/dL to g/100 mL and a reduction of 10% to 15% in concentration due to the lower water content of whole blood. There are several ways to calculate this conversion. One is based upon the HCT of the subject, the theory being the HCT is a reciprocal reflection of the water content of whole blood (Shajani, et al., Can. Soc. Forens. Sci. J. 22(4):317-320 [1989]). Another is based upon a comprehensive statistical examination of blood ethanol determined with headspace GC versus concurrent serum ethanol concentrations determined with clinical instrumentation (Barnhill, et al., J. Anal. Toxicol. 31(1):23-30 [2007]). The simplest, however, is to reduce the serum ethanol content by the relative difference determined for parallel analyses of serum and plasma (15%; Charlebois, et al., J. Anal. Toxicol. 20(1):171-178 [1996]). In this calculation, the serum ethanol result in the units, mg/dL, is divided by 1150 to provide a whole blood ethanol equivalent in g/100 mL (g%). It is also prudent to truncate the calculated result to two decimal places. This applies a bias toward a lower calculated result. For example, a serum ethanol result of 139 mg/dL would convert to 0.139 g/100 mL in serum, which is the equivalent of approximately 0.12 g/100 mL in whole blood (139 mg/dL ÷ 1150 = 0.121 g/100 mL → 0.12 g/100 mL). Clinical reports may include interpretational notes such as “legal limit" for a serum ethanol concentration of 80 mg/dL because this result converts to 0.08 g/100 mL, a per se limit. However, this notation is incorrect because only the units were converted and not the ethanol concentration difference between serum and whole blood. A clinical ethanol result of 80 mg/dL is actually the equivalent of a whole blood ethanol result of approximately 0.07 g/100 mL. For a whole blood result to be 0.08 g/100 mL, a serum ethanol result could be as high as 102 mg/dL (102 mg/dL ÷ 1150 = 0.089 g/100 mL → 0.08 g/100 mL). Reliance upon the serum result without a correct and complete conversion could lead to a misinterpretation of the result. A critical question that must be asked concerning the sample that was analyzed: Was the sample whole blood or serum or plasma? This question is of particular importance if the blood alcohol determination was performed in a hospital lab. Whole blood is composed of cellular material, plasma and fibrinogen (clotting agent). Hospitals test serum or plasma (whole blood is filtered to remove the cellular material and fibrinogen). Medical blood draws are primarily concerned as to whether alcohol is present and not concerned about the specific amount or concentration of the alcohol. A forensic or evidentiary blood draw is concerned with precise concentration of alcohol in the sample. Testing plasma or serum is less cumbersome than testing whole blood, but there are problems associated with plasma (e.g. when you centrifuge the blood sample, you take the solid, cellular material out, but you leave the same amount of alcohol in a smaller volume of liquid). This process artificially increases the concentration of alcohol in the liquid-which can lead to erroneously high test result. Hospitals test serum or plasma but report it as “blood alcohol." However, whole blood ethanol is not the same as serum ethanol or plasma ethanol. Serum is plasma with the fibrinogen (clotting material) removed. Serum is collected when the blood sample is allowed to actually clot. When the blood clots, a clear liquid (serum) forms over the blood. While serum and plasma alcohol concentrations are not significantly different, both of these tests produce results that are very different from the results produced by the analysis of whole blood. When plasma or serum analysis is used, the blood alcohol content will appear on average 14 % to 16 % higher than whole blood. In a ground-breaking study of the difference between whole blood ethanol and serum ethanol and plasma ethanol reported in the Journal of Analytical Toxicology, Vol. 11, November/December 1987, “Comparison of Plasma, Serum, and Whole Blood Ethanol Concentrations" by forensic scientists Charles L. Winek and Mark Carfagna, fifty subjects who had consumed alcohol were tested by taking a whole blood sample and simultaneously taking a second sample for blood serum. The average relative difference in reported ethanol concentration between whole blood and serum or plasma was 1.14 ± 0.02 across all fifty subjects. In other words, the plasma-serum series of tests showed a higher ethanol result by 14% as compared to the whole blood tests, with both samples measured by direct injection gas chromatography. As Winek and Carfagna noted: “A person with an ethanol concentration of 92 mg/dL in whole blood could have a reported concentration above 100 mg/dL if either serum or plasma is analyzed." In other words, according to Winek and Carfagna, a reported hospital lab result of 0.10 % BAC [mg/dL] using blood serum would be actually 0.092 % BAC if whole blood was analyzed. In view of the significant penalties for a conviction of alcohol related traffic collision, it is imperative that competent counsel understand which method — whole blood or blood serum/blood plasma- was used to determine the subject’s reported blood alcohol level. A second scientific treatise, Clinical Chemistry, Volume 39, No. 11, 1993, confirmed the earlier Winek and Carfagna study by running tests of blood samples taken from 211 persons. Professor Petrie M. Rainey of the Department of Laboratory Medicine, Yale University School of Medicine, using the gas chromatography method of analysis at the Yale University School of Medicine laboratory facility, determined the median ratio to be 1.15. However, the range of deviation among the 200+ individuals examined was 1.14 to 1.17. The report noted to a 95% degree of confidence, the range was 1.14 to 1.17, and the median ratio of all subjects was 1.15. As Professor Rainey noted: “Clinical laboratories have traditionally measured ethanol concentrations in serum or plasma. All state laws that define driving while intoxicated are written in terms of whole-blood concentrations. Because treatment of injuries takes precedence over collection of evidence, alcohol concentrations obtained in the emergency department are often the only measurements available on injured motorists. These measurements may be used as legal evidence in both civil and criminal proceedings. However, differences between serum and whole-blood alcohol concentrations have created difficulty in interpreting serum concentrations under legal statutes." Professor Rainey’s study determined, on average, the correct method to convert serum or plasma into the grams per cent by weight calculation was to divide by 1.15. “The median whole-blood alcohol concentration can be calculated by dividing the serum alcohol by 1.15 for a result in mg/dL or by 1150 for a result in weight percent." However, this computation is not precise for all persons, and there is a degree of variance in the concentration of ethanol between whole blood and blood plasma. Clinical Chemistry, Vol. 39, No. 11 at pages 2288-2292.