From the *Department of Pathology and Forensic Medicine, University Hospital Centre Split; and †Department Forensic Medicine, University of Split School of Medicine, Split, Croatia.
Manuscript received May 15, 2013; accepted October 27, 2013.
The authors report no conflicts of interest.
Reprints: Davorka Sutlovic, BSc, PhD, Department of Pathology and Forensic Medicine, University Hospital Centre Split, Spinciceva 1, 21 000 Split, Croatia. E-mail: email@example.com.
The typical postmortem toxicological analysis begins with the preliminary identification of drugs or chemicals present in postmortem specimens.1 Measurements of blood and urine alcohol concentration (BAC and UAC) in postmortem samples are often performed while determining the cause of death or in forensic practice (eg, after traffic accidents). The qualitative and quantitative determination of alcohol in postmortem specimens is a fast and accurate analytical method.2 According to the regulation, samples need to be stored at least 6 months from the time of the first analysis end.3
The main problem in repeated BAC determining from postmortem samples is its stability. In fact, there are a number of reasons for the change in the actual BAC: the loss of ethanol concentration due to oxidation and/or evaporation or increase in ethanol concentrations as consequence of degradation under the influence of microorganisms. The amount of generated ethanol depends on the species of microorganisms present, the available substrates, the temperature and time of storage, and the presence of preservatives added to the specimens.4,5
When presenting results in court, it is necessary to prove that the result in question is reliable, even if the sample was analyzed for several months after sampling. Knowledge about the stability of the BAC, within legally required storage time of 6 months, especially if the sample was stored longer than 6 months, is an important factor in interpreting repeated analysis results. Differences in stability for several substances including BAC have been demonstrated in several past studies.6–8
The aim of our study was to evaluate the stability of alcohol concentration in postmortem blood samples stored in a refrigerator at −20°C during 6 months.
MATERIALS AND METHODS
During autopsy, blood samples were collected into sterile and chemically clean containers. Sodium fluoride was added to the blood samples as a preservative. Samples were kept in the refrigerator at 4°C until analyzed. Blood alcohol concentrations were measured twice: first within 1 to 4 days after being taken at the Laboratory of Forensic Toxicology and second after some period of storage at −20°C (Table 1).
Blood alcohol (ethanol) concentration was measured by Shimadzu 2010 gas chromatography with headspace and flame ionization detector. Headspace conditions are as follows: thermostatting 12 minutes at 75°C with rotation speed 500 of revolutions/min and needle transfer temperature 95°C. Ultrapure-grade helium was used as the carrier gas at a flow rate of 11.70 mL/min. The chromatographic column was RTX-BAC2 (fused silica, 30 m, and 0.53-mm internal diameter, with a film 0.20 µm thick). Injection temperature was 200°C, column condition was 3 minutes at 60°C, and flame ionization detector temperature was 200°C. Each blood sample was measured 3 times.
The linear calibration of ethanol concentration was obtained using 5 standard solutions in the range 0.5 to 2.50 g/kg. The precision, in terms of repeatability, was assessed at 1 concentration level, 0.5 g/kg, measured 10 times on the same day.
Continuous data were expressed as average, median, and SD. The significance of the differences between values was tested. The differences between BAC1 and BAC2 values were plotted. Statistical analyses were performed using GraphPad Prism 4 for Windows (GraphPad Software, San Diego, Calif), SPSS for Windows version 11.03 statistical software (SPSS Inc, Chicago, Ill), and MS Office Excel 2010 packages, Microsoft Excel computer software (Microsoft, Redmond, Wash).
For all statistical tests, a significance level of 95% (P ≤ 0.05) was used. Bland-Altman plot was used to calculate bias and imprecision between the measurements.9
Relative SDs of response factor (RSDRF) of 5 replicate samples for standard solutions were 8.5450%. The correlation coefficient (R) obtained by linear fitting to the calibration points was 0.9996. Limit of ethanol quantification was 0.01 g/kg. The precision was assessed at a concentration of 0.5 g/kg. Relative SD was 3.3489%.
Blood Alcohol Concentration
The stability of alcohol in the postmortem blood samples was tested by repeated measurement of the BAC in samples stored at −20°C in a period ranging from 191 to 468 days. Storing at −20°C mainly causes decrease in alcohol concentration. The average percent drop was −9.56%. In 55 (69.62%) of the 79 analyzed samples, concentrations reduced between 0.3% (from 1.29 to 1.28 g/kg) and 60.5% (from 3.60 to 1.42 g/kg). In 12 of the 79 analyzed samples (15.19%), concentrations increased between 0.45% (0.42–0.43 g/kg) and 56.98% (0.93–1.46 g/kg). In 12 samples, there were no concentration changes. From those 12 samples, 11 had an initial concentration of 0 g/kg, and 1 sample had 1.56 g/kg. The results are shown in Table 1.
Because of the comparability, a difference between 2 measurements was expressed as a percentage. Minimum, maximum, average, and median value of those differences were as follows: 0%, 100%, 12.10%, and 7.32%, respectively. SD of differences between 2 measurements was 16.3563.
In 15 of 79 samples, the presence of drugs and/or drugs of abuse was confirmed. From those 15 samples, 3 samples had alcohol concentration of 0 g/kg, and concentration did not change. In the remaining 12 samples, alcohol concentration changed. The highest change occurred in sample 54, which contained both methadone and diazepam. Its alcohol concentration increased from 0.93 to 1.46 g/kg.
Results obtained in this study show good agreement between the experimental measurements (Fig. 1). About 90% of the results (slope average) lie within 95% limits, and 10% are outside. The 2 measurements had very similar average results, and the bias was only −0.14. In 95% of the results, the difference lies between −0.75 and 0.47.
Higher correlation was observed in samples that have been stored from 310 to 468 days (Fig. 2). Pearson correlation factor, for samples, which were stored from 191 to 300 days, was 0.06 versus the correlation factor of 0.36 for samples stored from 310 to 468 days. Despite these positive correlations, the results of some tested blood samples showed a high BAC variation. Changes in BAC higher than 10% were observed in 39% of the samples. The said difference is not acceptable when dealing with forensic samples. In court proceedings, the discrepancy can cause problems. But, according to regulations, the samples need to be stored at least 6 months from the time of the analysis end.3 The storage time for samples in this study supersedes that storage minimum.
Samples 33 and 76 showed significant losses in ethanol concentration. A possible explanation for ethanol reduction in those samples might be improper sample storage. In fact, polypropylene tubes were not filled to the top. Excess headspace above could lead to ethanol oxidation in the blood samples. Sodium fluoride was added to all samples, as a preservative, but it did not prevent loss of ethanol by oxidation. Olsen and Hearn7 investigated in their study changes in blood ethanol concentration of samples stored in 50- and 10-mL tubes. They found that the stability of ethanol in the gray-top Vacutainer samples (10 mL) could be attributed to less headspace in the smaller volume container. One mean of protection from ethanol loss is storing the blood sample in a tube of suitable volume.
Shan et al10 performed a study testing the influence of long-term storage of blood samples (range, 13–39 months), with variations of temperature. The results showed that the blood, which was stored for a long time, had a lower alcohol concentration compared with the alcohol results before storing. Samples with higher alcohol concentrations to begin with showed greater deviation (Fig. 1B). On the other hand, there had been blood samples in which increases in the BAC were observed. The explanation might be that, in the presence of a suitable substrate (as glucose in diabetic patients), the production of endogenous ethanol by microorganisms through the fermentation of carbohydrates is possible. Positive results obtained in this manner can lead to false accusations of ethanol use.11
According to the Guidelines for Obtaining Specimens for Postmortem Toxicological Analysis, blood and urine samples should be collected into a separate tube containing 2% sodium fluoride.12 To inhibit microorganisms from producing ethanol out of glucose and ethanol oxidation, it is necessary to add more than 2% sodium fluoride in blood and urine samples of diabetic patients.
Previous articles report testing the stability of the abused drugs’ concentration in the presence of alcohol, but there were no tests of blood alcohol stability in the presence of abused drugs.8,13 Atanasov et al8 have studied diazepam stability in the presence of alcohol in different storage conditions. The results showed that storage temperature has the most important role on the stability of benzodiazepines. Also, regardless of the concentration, ethanol influences the benzodiazepine stability. Data from our study showed no significant changes in alcohol concentration in presence of drug abuse (the only exception was sample 29).
There were good agreements in the BAC in 2 performed measurements, but the observed deviation in few cases was up to 10% (confirmed in 39% samples) and was not acceptable when dealing with forensic samples. It is necessary to store the blood sample in a tube of suitable volume with minimal or no headspace to defer alcohol oxidation. Also, if a sample comes from a diabetic patient, the analyst is advised to add more than 2% of sodium fluoride to inhibit microorganisms from producing ethanol out of glucose and ethanol oxidation in general.14
The authors thank Mrs Tanja Viskovic and Mrs Tajana Pocrnja Babic for their technical assistance and Dr Dijana Gugic for the language advice.
1. Karch SB . Postmortem Toxicology of abused drugs. 2nd ed. Boca Raton, FL: CRC Press; 2008; .
2. Karch SB . Drug Abuse Handbook. Boca Raton, FL: CRC Press; 1998;.
Jones AW, Pounder DJ. Measuring blood-alcohol concentration for clinical and forensic purposes; pp 327–356
3. Croatian regulation on the blood and urine sampling procedure for determining the alcohol, drug and psychoactive substances concentration. Narodne novine. 1999; 107
4. Clark MA, Jones JW . Studies on putrefactive ethanol production: I. Lack of spontaneous ethanol production in intact human bodies. J Forensic Sci. 1982; 27: 366–371.
5. Petkovic S, Savic S, Zgonjanin D, et al. Ethanol concentrations in antemortem blood samples under controlled conditions. Alcohol Alcohol. 2008; 43: 658–660.
6. Karinen R, Oiestad EL, Andresen W, et al. Comparison of ethanol and other drugs of abuse concentrations in whole blood stored in Venoject® glass and plastic and Venosafe™ plastic evacuated tubes. J Anal Toxicol. 2010; 34:(7): 420–428.
7. Olsen T, Hearn WL . Stability of ethanol in postmortem blood and vitreous humor in long-term refrigerated storage. J Anal Toxicol. 2003; 27: 517–519.
8. Atanasov VN, Stoykova S, Runiov A, et al. Stability of diazepam in blood samples at different storage conditions and in the presence of alcohol. Forensic Sci Int. 2012; 215:(1–3): 159–163.
9. Bland JM, Altman DG . Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986; i: 307–310.
10. Shan X, Tiscione NB, Alford I, et al. A study of blood alcohol stability in forensic antermortem blood samples. Forensic Sci Int. 2011; 211:(1–3): 47–50.
11. Sulkowski HA, Wu AH, McCarter YS . In-vitro production of ethanol in urine by fermentation. J Forensic Sci. 1995; 40: 990–993.
13. Fjeld B, Burns ML, Karinen R, et al. Long-term stability of GHB in post-mortem samples and samples from living persons, stored at −20°C, using fluoride preservatives. Forensic Sci Int. 2012; 222:(1–3): 47–51.
14. Sutlovic D, Nestic M, Kovacic Z, et al. Microbial ethanol production in postmortem urine sample. Med Sci Law. 2013; 53:(4): 243246