Jeffrey H. Garland was retained by Walter Rothe for BUI Manslaughter and felony BUI charges which arose from a July 31, 2005 boating accident.
The case was extensively litigated on both blood alcohol testing and causation. Attorney Garland has authored a synopsis of the information developed in the Rothe case relative to blood alcohol testing.
The synopsis refers to “ethanol” which is the correct chemical name of the type of alcohol generally consumed in alcoholic beverages. There are many different types of “alcohols” in the sense the chemists use the term. There are important chemical differences between these alcohols.
Robert Parsons is a criminalist at the Indian River Crime Laboratory in Fort Pierce, Florida. He is permitted by the Florida Department of Law Enforcement (FDLE) to forensically test blood for ethanol content.
Bruce Goldberger, Ph.D., is an expert witness who often testifies, usually for the prosecution, in DUI and BUI cases. Dr. Goldberger is a forensic toxicologist who is employed by the University of Florida. He is a past-president of the American Academy of Forensic Sciences.
Stefan Rose, M.D., is a toxicologist who often testifies, usually for the defense, in DUI and BUI cases. Dr. Rose has testified as a toxicologist in State and federal cases through the Country.
It is hoped that the following synopsis of the Rothe case would be useful to those with similar issues. This synopsis is case-specific to the extent that it relates to the particulars of the Rothe investigation. The principles, however, may be applied to all blood-ethanol investigations.
TABLE OF CONTENTS
Table of Contents…………………………………………………………………………………………………ii
I. MISCELLANEOUS PARSONS ISSUES
The Beausoleil worksheet does not even have a “question mark” next to blank used to indicate the presence of anticoagulant (as did the Brunson worksheet). Still, Parsons noted that the Beausoleil tubes were unclotted. The Beausoleil tubes were also BD #367012 which, as shown above, did not contain an anticoagulant according to the BD website.
Parsons has admitted the possibility of partial blood clotting, but claimed that there would be an increase in viscosity and/or clumpiness. Parsons could only say that he was “pretty good at telling whether or not there is an anticoagulant present based on the condition of the blood”. (7/27/07 hearing 83, 85). One wonders whether “pretty good” is good enough in forensic toxicology where facts must be proved beyond a reasonable doubt, and where the consequences of an error could be many years in prison.
The BD tubes did not belong to the Brunson and Beausoleil NIK #4994 Kit, because they did not contain an anticoagulant. The presence of anticoagulant is specified by NIK for the #4994 Kit and by FDLE Rule 11D-8.012(2) for all blood collection kits.
Brunch v. State, 954 So2d 1242 (Fla. 4th DCA 2007), considered an appeal from the use of a kit which had expired 28 days before. Testifying for the State was Forensic Toxicologist Jesse Bidanset, Ph.D. He testified that the tubes contained both preservative and anti-coagulant. He explained that the expiration date is a warranty of the vacuum’s sufficiency to draw a proper sample. The expiration date does not relate to the preservation of the blood sample or affect its integrity once drawn. Dr. Bidanset said that the tubes withdrew normal sized samples of 9-10 ml, and that there were no other problems noted. The appeal court noted the lack of evidence questioning the reliability of the test results. Id. at 1244.
Brunch observed that FDLE regulations do not prescribe an expiration date for kits or tubes. Therefore, there was no direct violation of FDLE rules. In the absence of evidence that the test results were unreliable, Brunch held there was no due process violation and no per se error for using an expired kit. It was not improper to instruct the jury on the presumption of impairment. Id at 1245.
The nurse at St. Mary’s Hospital testified in deposition that she collected blood samples from Wilfredo Rodriguez in the normal fashion. She remembered nothing out of the ordinary in this exercise. Based upon her extensive experience, the nurse initialed both tubes and placed identifying information on them. She turned the tubes over to the FWC officer. The kit was sealed. Signatures were placed on the outside of the kit. The FWC officer transported the kit to the Indian River Crime Laboratory where it was turned in without any indication of tampering. The person receiving the kit for IRCL accepted the kit without any indication of tampering. Parsons processed the Wilfredo Rodriguez kit and produced a written report on the normal form used by IRCL for forensic ethanol blood test reporting purposes. The formal report, signed by Parsons, specifically said there were two blood collection tubes in the kit.
When Attorney Richard Kibby deposed Parsons, Attorney Kibby had no access to the laboratory records. Parsons asserted to Attorney Kibby that everything about Rothe’s NIK kit was “all very routine” and that there was “no indication of irregularities”. Parsons stated that there was nothing irregular in the Wilfredo Rodriguez kit. Parsons asserted that he tries to be meticulous about noting irregularities. (4/7/06 Parsons depo 32-33). These are fascinating assertions because Rothe’s kit arrived with the tubes from the wrong manufacturer, and Rodriguez’s kit arrived with just one tube.
When the defense, pursuant to multiple subpoenas duces tecum, finally secured the actual laboratory bench notes, it was uncovered that Parsons reported only one tube as being in the Rodriguez kit. Parsons had no explanation what happened to the other blood collection tube. Parsons did not report the second tube as being broken in the kit or broken in the laboratory after the kit was opened. Parsons said at this deposition that Rodriguez’s kit had just one tube, but that he had no knowledge as to why. (12/14/06 Parsons depo 53-54).
Dr. Goldberger said he had received kits with just one tube in the past. However, he was unaware of any manufacturer who would place just one tube in a kit. (12/18/06 Goldberger depo 90-91).
Parsons used a dual column Perkin-Elmer (PE) chromatograph to run his blood tests. The PE chromatograph is capable of retaining data. The data is ordinarily stored on a free-standing computer which is hooked up to the chromatograph. According to Dr. Goldberger, his laboratory retains electronic data and does not systematically purge it. (12/18/06 Goldberger depo 48).
Parsons said that he purges the raw data after each batch is completed in order to make room for the next batch. He said that this purging of data is done routinely even though the data could be saved by using a CD or other storage device. “There is no reason to keep data that you don’t need anymore.” (12/14/06 Parsons depo 35-37). It is deleted after the data has been “verified and printed out” (01/25/07 Parsons depo 9-10).
The defense has hypothesized that Parsons committed “sampling error” in the Rothe test – just as he did the following year when he failed controlled substance proficiency testing. Parsons performed all blood-alcohol testing on his own with no assistance. (4/7/06 Parsons depo 5, 8 – 9). Parsons would mark testing vials with a Sharpie. (12/14/06 Parsons depo 17). This is not the practice utilized in Dr. Goldberger’s laboratory. The UF laboratory always uses two separate technicians in order to minimize sampling error. (12/18/06 Goldberger depo 68).
The chromatograph is very reliable for ethanol testing, but Dr. Goldberger acknowledged that technicians can make mistakes. (12/18/06 Goldberger depo 69). At the time of the deposition, Dr. Goldberger knew that Parsons had failed a controlled substance proficiency test, but assumed that such failure would be unrelated to ethanol testing. Dr. Goldberger was unaware that Parsons had failed a second round of proficiency testing and had been removed from controlled substance testing altogether. (12/18/06 Goldberger depo 73).
Dr. Goldberger acknowledged that sampling error could cause the chromatograms to be incorrect. He recommended retesting as the best way to determine the actual results. (12/18/06 Goldberger depo 74).
Ordinarily, when replicate testing is used, both replicate samples should have results within “shouting distance” of each other. A significant inconsistency would be an indicator of possible sampling error. This “fail safe” would not be useful if the analyst confused both replicate samples with another pair of replicate samples. Dr. Goldberger acknowledged this: “You could have wrong results.” (12/18/06 Goldberger depo 16).
After his first controlled substance proficiency failure, Parsons admitted having multiple samples open during the proficiency testing procedure. He asserted that his “strict practice in case work is to open only one exhibit at a time”. He resolved to “conscientiously treat all proficiency test samples in exactly the same way as I would case samples”. Despite this assertion, Parsons committed another sampling error. (9/6/06 IRCL Corrective Action Report).
In his comments on the Indian River Crime Laboratory (IRCL) Corrective Action Worksheet, Parsons blamed the proficiency test failures on “rushing to meet a deadline”. He attributed the failures to “sampling error”. The initial proficiency failure occurred in June, 2006, but the results were not received until August. Parsons was required to undergo remediation, which included five additional external proficiency tests. During the third round of remedial proficiency testing, Parsons committed a second mistake, which he attributed to “deteriorating mental and physical condition due to a variety of life stressors that had occurred between June and September”. (01/25/07 Parsons depo 56-59).
Parsons agreed that he “mixed up those samples”, although he maintained that he had never done that before in his life. Although the effect on his career and financial security were obvious, Parsons asserts that he would never take such “chances” in casework. He states that, in July, a very close family member died, and that, in August, he was in a very serious car accident. (01/25/07 Parsons depo 59-61).
At the time of the 01/25/07 deposition, Parsons was not performing controlled substance testing. He was not scheduled to return, but claimed that he anticipated returning at some time in the future. (01/25/07 Parsons depo 61).
ICRL received notice that Parsons failed a proficiency test on August 23, 2006. The test had been performed on June 12, 2006. The failure was reported by Collaborative Testing Services, Inc. 06-501. Parsons found alprazalam in item 1 when no controlled substance was present; and correctly identified alprazalam in item 2.
Babu Thomas met with prosecutors at their weekly meeting on September 8, 2006. He advised them that Parsons had failed proficiency testing and was removed from drug cases. The Class I discrepancy was attributed to “human error”. (B. Thomas letter, 9/12/06). Both Thomas and Nippes signed off on an IRCL Corrective Action Report which attributed the failure to sampling/labeling error (9/6/06).
Another Class I discrepancy was reported on October 5, 2006. Nippes gave notice of permanently removing Parsons from controlled substance testing by letter dated October 31, 2006. According to Nippes’ November 2, 2006 letter to Chief Assistant State Attorney, “It was determined by the Senior Criminalist, and agreed by the Laboratory’s Proficiency Review Committee, that the discrepancy was related to a sampling/labeling error (human error) unrelated to analytical procedures”.
Parsons asserted that all of his controlled substance testing in the previous year had been confirmed upon retesting: “There have been no differences in the identifications that I made, all have been confirmed.” This statement was made, even though Babu Thomas was only about half way through the reanalyses at the time of the deposition. (01/25/07 Depo 62).
Parsons was “reclassified” effective November 6, 2006, from Criminalist to Blood-Alcohol Analyst. His pay was cut from $83,999.87 to $60,000.00. 11/6/06 Personnel Action Form. Despite this “misfortune”, Parsons’ September 30, 2006 Progress, Performance and Evaluation Report rated him all A’s and B’s! Despite the reclassification, Parsons continued to analyze blood-alcohol samples (11/30/06 letter from Nippes to Laura Barfield).
On November 30, 2006, Parsons wrote a detailed letter explaining that he made two errors out of nine drug analysis proficiency samples completed. He attributed the mistakes to sampling errors and to “extreme stress, chronic insufficient sleep and exceptional mental and physical fatigue”. The letter goes on to assert that he had never before failed a proficiency test, which is incorrect. He, in fact, failed a blood-alcohol proficiency test in 1994.
Parsons reasserted, in an internet posting, that he never before failed a proficiency test:
In a nut shell, after over 24 years of professional work and the correct completion of hundreds of proficiency tests and test samples, plus well over ten thousand case samples, I made my first error in analysis (that I know of). In fact, I made two errors, among a series of nine drug analysis proficiency test samples completed in a three-month period.
http://tech.groups.yahoo.com/group/forensic-science/message/10958 (1/12/07). When Parsons’ work was actually “graded” by an independent agency, he got 22.2% of analyses wrong! In school, this might have earned Parsons a “C”. In a forensic laboratory, it earned him a suspension, reclassification and $24,000.00 pay cut.
Parsons told Attorney Kibby that no laboratory had ever disagreed with his findings: “they always corroborated my results.” (4/7/06 Parsons depo 26). This is curious, because Parsons failed blood-alcohol proficiency in 1994. Parsons failed controlled substance proficiency testing in the months following the April 7, 2006 deposition.
Parsons was fully aware of an unsatisfactory blood-alcohol proficiency test in the third quarter of 1994. He even commented on it in an e-mail to Tom Wood on June 28, 1999.
On April 17, 2007, Nippes wrote a letter to Chief Assistant State Attorney Thomas Bakkedahl. The retesting concerned cases examined by Parsons between July 1, 2005, and August 14, 2006. Nippes explained that 16 of the cases were not available for reexamination. Nippes identified two cases involving discrepancies. The first was a .9 gram weight difference in a cocaine sample which caused a criminal charge of trafficking to be reduced. The second case failed to identify psilocybin, but said the discrepancy could be attributed to degradation. At a jury trial in Martin County, Parsons testified on March 13, 2008, that all of his retested cases were confirmed: “All of my analytical conclusions both identifying drugs and saying there were no drugs present were all confirmed. So we had insured that I had made no mistakes in case work. The only mistake I ever made was in…that particular series of drug proficiency tests.” In reality, a trafficking in cocaine conviction was set aside on a post-conviction relief motion when the reweighed sample was less than 28 grams.
Parsons failed the third quarter 1994 blood-ethanol proficiency testing. According to September 12, 1994 letter from Tom Wood to Parsons: “Your 1994 Quarter 3 proficiency sample results received by the Implied Consent Program fell outside the two standard deviation boundary from he mean.”
Parsons’ handwritten notes say that his results were .001 below the accepted range. He said he rechecked calibration and it was correct. He speculated that the pipettor may have had a problem or that there was a delay in sending sample to Parsons for testing.
The chromatograph does not specifically identify chemicals. It uses a flame ionization detector. Ionized molecules produce an electric current that is detected and analyzed by a microprocessor. The quantity of a molecule is a function of the intensity of the electric current caused by the ionized molecules. (01/25/07 Parsons depo 3-4).
The electrical potential is not specific to a given molecule. The identification method used by the chromatograph is time retention. Different molecules emerge from the column at different times. The chromatograph has been described as a still with a clock. (01/25/07 Parsons depo 4).
The intensity of the electric potential is indicated by peaks. The molecule is deemed to be identified if the peak appears within 3% “window” on either side of a presumed retention time. (01/25/07 Parsons depo 4-6).
A peak will be identified as ethanol if the peak appears within the window. A single column chromatograph is much less reliable than a dual column. To be identified as ethanol, however, the peaks must appear within the windows for samples run through both columns. Parsons asserts that the chromatograph would exclude other volatile organic molecules (VOMs), because absolute and relative retention times would differ from the times for ethanol. (01/25/07 Parsons depo 6-8; 12/14/06 Parsons depo 32).
Parsons agreed that different molecules could “co-elute” at about the same retention times. He claimed that such co-elution would be shown as a double peak or a “shoulder” to a peak. (01/27/07 Parsons depo 8). Parsons defined a shoulder as asymmetrical hump in a peak. It could be caused by an incompletely resolved second peak or because of incorrectly chosen parameters. (12/14/06 Parsons depo 26-27).
Parsons claimed that the chromatograph’s “Total Chrom” software would identify the molecule only if there were a single peak within the window. Yet Parsons admitted that the Total Chrom software allows an analyst to manually re-examine the raw data. (01/25/07 Parsons depo 8-9). Parsons agreed that an adjustment of the software’s peak separation criteria could impact the instrument’s assumption that there might be one or two peaks within a given window. (01/25/07 Parsons depo 20).
Parsons acknowledged that the Total Chrom software would allow a “reprocess” of raw data in order to distinguish incompletely resolved peaks. According to Parsons:
If you see a co-elution you have incompletely resolved peaks, the instrument does not know which of those peaks to assign that tentative identification to, it could be either one of them, so you have to be able to go back and reprocess to try to mathematically separate those peaks, draw a new baseline between the two so you can look at them one at a time.
(01/25/07 Parsons depo 9). Parsons has acknowledged that, in training, he used techniques to adjust peaks. Such techniques could be used to resolve and separate peaks of two components coming off the column into the detector in a very short period of time. (12/14/06 Parsons depo 27-28).
The Total Chrom Workstation Users Guide (“Users Guide”) describes a variety of factors which may be programmed into the software. Chapter 18 is entitled “Peak Detection”. Of particular interest is the “bunching factor”, which specifies how many sequential data points are grouped into a bunch. The values are then averaged. The goal of this technique is to eliminate “noise”.
The Users Guide states that, ideally, a peak will have about 20 bunched points. The Users Guide warns that setting a bunching factor too high will lead to small, unresolved peaks going undetected. The goal is to keep peaks close to the 20-point optimum. Ch. 18-6 and 18-7, Users Guide. “Total Chrom might not detect smaller peaks at all, especially those that appear as shoulders on the leading edge or trailing edges of larger peaks.” Ch. 18-10, Users Guide.
In the Rothe case, it was impossible for defense experts to re-examine the raw data, because Parsons intentionally deleted the raw data after each batch. Parsons maintained that data retention is not necessary, and that the printed chromatogram is sufficient for subsequent analysis. (01/25/07 Parsons depo 9-10, 37). Despite these assertions, Parsons acknowledged that “exponential skim criteria” was programed into the software as part of the default program. (12/14/06 Parsons depo 31; 01/27/07 Parsons depo 21-22). According to the Perkin-Elmer Total Chrom Manuals, exponential and tangential skim criteria can help the software determine whether to resolve a response into a single or double peak or a shoulder.
Time retention for a specific molecule varies from chromatograph to chromatograph based upon characteristics of the chromatograph, columns, length of columns, type of column “packing”, temperature, pressure and type of gas used. The time retention for any particular molecule on any particular chromatograph is individual to that configuration.
Although Parsons acknowledged the potential problem of co-elution, he has only tested his chromatograph with 5 molecules required by FDLE regulations: n-propanol, ethanol, acetone, methanol and isopropanol. Parsons asserted that his method has been extensively researched and that is no reason for him to repeat such research. (12/14/06 Parsons depo 30, 34; 01/25/07 Parsons depo 12-13).
The software transforms raw data points into a “peak” depicted on a chromatogram via a process called “integration”. (12/14/06 Parsons depo 24). The software creates a curve via a mathematical formula and assumptions about the raw data. The software “uses an algorithm to form the peak or the curve…from the data points”. It was impossible to examine the raw data points in the Rothe case, because the raw data was intentionally deleted. (01/25/07 Parsons depo 22-23; also 25).
The software’s calculation of the presence of a molecule is related to the area under the peak. (12/14/06 Parsons depo 29). The concentration of the molecule is, therefore, directly related to the calculation of the “curve” from a limited number of data points. The calculated concentration would be wrong if the curve has been miscalculated. (12/14/06 Parsons depo 29).
Parsons agreed that the Total Chrom Manual describes a procedure whereby the operator can go back to an “inflection point” and determine where the raw data points might lie. Parsons never went back and reviewed the raw data points. (12/14/06 Parsons depo 26-30, 38).
Parsons has emphasized the accuracy of ethanol testing via headspace chromatography. Perkin Elmer, as manufacturer of the chromatograph and developer of the Total Chrom software, has published an article detailing the limits of detection and quantitation: “Increasing Accuracy of Blood-Alcohol Analysis Using Automated Headspace – Gas Chromatography.” See www.perkinelmer.com. The article concludes that ethanol testing via this method has a limit of detection (LOD) of .005 and a limit of quantitation (LOQ) of .010. The relatively high LOD and LOQ suggest that the presence of other VOMs cannot be excluded.
Parsons rejected the need to preserve data for reprocessing. He asserted in Rothe that there was no irregularity in the chromatogram, that a validated system was used, that certified standards were used, and that there was no indication of a problem. (12/14/06 Parsons depo 39). This is an unusual and perverse attitude in the world of forensic science. Imagine if fingerprints or DNA samples could not be examined by defense experts! In he final analysis, Parsons failed to spot the presence of wrong tubes, failed to notice the absence of anticoagulant, used an expired control and twice failed controlled substance proficiency tests.
Parsons reviewed chemicals listed on materials supplied by DuPont Corp. He asserted that none of these chemicals could be mistaken in a gas chromatograph for ethanol. “My opinion that [they are] irrevelant because [they were] not in the blood stream at the time the blood was drawn at detectable levels.” Parsons justified this conclusion by stating that the chemicals can only be absorbed into the blood stream at low levels and would be “eliminated from the blood stream within minutes of any exposure”. He explained that this is the justification for the 20-minute waiting period in breath testing cases. Although he was unable to cite any research articles justifying opinions on this subject, he claimed to possess numerous research articles. Although requested, he never produced them in the case. (01/25/07 Parsons depo 14-16).
Again referring to the Material Safety Data Sheets, Parsons asserted that the listed chemicals would have to come in physical contact with the subject and be permeable through the skin. He asserted that “very few are”. Parsons was confronted with the Sheets showing the possibility of absorption of certain chemicals at concentrations which could cause serious side effects and, therefore, be dangerous. Parsons reasserted that the Sheets do not establish that chemicals would be present “in the blood stream at a detectable level or that they would affect a blood-alcohol analysis”. (01/25/07 Parsons depo 16-17). Parsons went on to state:
[T]hey cannot possibly be at detectable levels for this method in a living person. Those chemicals, as you point out, are highly toxic. If they were at the levels in the blood stream of blood-alcohol so that they can be detected by this method, the person would be in the morgue, not walking around to mount a defense his criminal charges.
(01/25/07 Parsons depo 18).
Parsons recognized a distinction between chemicals which might be absorbed through the skin as opposed to simply through breathing. He acknowledged that some volatile organic molecules (VOMs) might be permeable through the skin, but asserted that they are “few” in number. Even though he had no specific information on this subject, Parsons considered the absorption of VOMs through the skin “not to be a concern” because of the “extreme unlikelihood of it”. (01/25/07 Parsons depo 16). Parsons further asserted that the concentration of most chemicals on the Material Safety Data Sheets would be “so minuscule that they could not be detected by a gas chromatograph in the blood”. (01/25/07 Parsons depo 18).
Rothe’s chromatogram #4509 shows an ethanol peak at a .99 retention time. He acknowledged an unidentified peak at .87 retention time, which he attributed to “blood breakdown products”. Parsons admitted that he did not know what the .87 peak might be. (12/14/06 Parsons depo 51-53).
Parsons admited that he had conducted no research, himself, into the area of absorption of these chemicals and any effects upon chromatographic testing. (01/25/07 Depo 19).
Dr. Goldberger is familiar with candida albicans as a microbe which can create ethanol. (12/18/06 Goldberger depo 34). Dr. Goldberger “believes” that candida albicans is found on human skin. Dr. Goldberger was asked: “Are you familiar with research showing that blood alcohol concentrations in preserved blood specimens increase even without having been inoculated with bacteria?” Dr. Goldberger said: “No. No, I’m not aware of that work.” (12/18/06 Goldberger depo 97).
Dr. Goldberger was confronted with a book in which he was a contributor: “Medical Legal Aspects of Alcohol”, 4th Edition. Dr. Goldberger acknowledged that ethanol concentrations have increased, but emphasized his belief that those studies used blood which had been “inoculated with bacteria”. (12/18/06 Goldberger depo 95-96; see 7/27/07 Rothe hearing 168). Dr. Goldberger restated his belief that ethanol is generally lost through oxidation or evaporation into the headspace of a blood collection tube.
Parsons has said that EDTA acts both as an anticoagulant and a preservative. (12/14/06 Parsons depo 55). Dr. Goldberger maintained that EDTA is only an anticoagulant and has no preservative functions. (Goldberger depo 42-43).
Blood samples 11 and 12 from the batch run by Parsons did not have an anticoagulant in them – at least according to the tubes. Parsons was asked about a question mark on his handwritten notes. “It meant no anticoagulant was listed on the tube…”. Still, Parsons maintained that the blood was unclotted and that, therefore, there must have been an anticoagulant present. (12/14/06 Parsons depo 56-57). Dr. Goldberger agreed that coagulation would happen in most cases and asserted that an anticoagulant must be present if the blood is not clotted. (Goldberger depo 99). Parsons admitted that both sets of tubes used in 11 and 12 had expired in May, 2004. (7/27/07 Rothe hearing 87).
Numbers 11 and 12 in the batch were supposed to be tubes manufactured by Becton-Dickinson. Dr. Goldberger was confronted with the laboratory bench notes which showed the product number for the Becton-Dickinson tubes. Dr. Goldberger was confronted with Becton-Dickinson printouts which showed that the tubes used should not have contained anticoagulant. Dr. Goldberger acknowledged that Parsons reported that the tubes in tests 11 and 12 were “unclotted”. (12/18/06 Goldberger depo 92-94).
Dr. Goldberger identified a peak for acetaldehyde. He attributed this peak to the breakdown of ethanol after the blood sample was obtained. (12/18/06 Goldberger depo 52-53). This assertion fails to take into account that the acetaldehyde could have been produced by the body before the blood sample was obtained or was a byproduct of ethanol production or destruction by microbes after the sample was obtained. Dr. Goldberger admitted that the acet- aldehyde could have been present from metabolism before the blood sample was taken (12/18/06 Goldberger depo 54) or as a result of oxidation after the blood sample was drawn. (7/27/07 Rothe hearing 192-193).
Parsons asserted that ethanol levels will increase in a blood sample only in the absence of preservative and an intentional inoculation with alcohol-producing organisms. (4/7/06 Parsons depo 41-42). Parsons asserted that there would be no increase in ethanol if there is preservative and without preservative as long as sample is refrigerated . (4/7/06 Parsons depo 15). Despite these assertions, Parsons agreed that there is no way to visually tell if the preservative is present. Neither Parsons nor Dr. Goldberger tested for bacteria or preservative. (4/7/06 Parsons depo 43-44; 7/27/07 Rothe hearing 164, 193-194).
The Rothe samples had been received by the crime lab on August 1, 2005, then refrigerated. They were removed for analysis on August 17, 2005. (4/7/06 Parsons depo 27). Parsons personally turned the samples over to FWC Inv. Douglas Rogerson on October 6, 2005. Thus, the Rothe samples were not refrigerated for the period of time from August 17 until October 6, 2005. (4/7/06 Parsons depo 14-15, 27, 51).
Candida albicans is an efficient producer of ethanol which is not inhibited by sodium fluoride. Candida albicans can be controlled best by refrigeration. “The Effect of Temperature on the Formation of Ethanol by Candida Albicans in Blood”, J. Chang and S.E. Kollman, Journal of Forensic Sciences, Vol. 34, No. 1, pp 105-109, January, 1989.
It is noteworthy that researchers have recommended higher concentration of preservative than used in the Rothe kit, as well as refrigeration. Higher concentrations are readily available. For example, NIK sells Kit #4994 which is “designed to meet the requirements of states mandating by law that sodium fluoride preservative be at least 1% of the 10 ml blood sample”. The NIK Kit #4994 uses 100 mg of sodium fluoride and 20 mg of potassium oxalate. http://www.nikpublicsafety.com/productsdiv/specimen.asp.
Parsons was asked whether the observations of law enforcement officers at the scene of an accident might be useful in assessing whether an individual was under the influence of ethanol. Apparently knowing that the LEOs in Rothe’s case did not make such observations, Parsons did not rely upon those observations or the lack of such observations. He stated that he did not consider those observations “because the observations of other people, whether they’re trained or not, are not necessarily reliable”. (01/25/07 Parsons depo 41-42).
Although at deposition he minimized importance of observations of the subject, Parsons had a different view on a forensic science bulletin board. He said there should be a “full toxicological evaluation”, which would include evaluation of impairment. Info Archive of NNYTech, Mess #2129 (1/30/04). Parsons went on to say ‘the more information the better’.
Parsons was asked about NHTSA studies validating standardized field sobriety tests. He agreed, in a general sense, that the NHTSA-approved tests have a relationship to the presence of alcohol in the person and the level of impairment. He did not recall what level of correlation may exist between performance on the NHTSA-approved tests and BAC. “I would not rely on a field sobriety test to make a determination for any purpose. Its an indication only, its not reliable for drawing conclusions.” (01/25/07 Parsons depo 44-45).
Parsons was asked about the effects of ethanol at different levels. Parsons stated that “impairment…begins with the first drink”. However, he qualified this statement by saying “when it becomes visible is impossible for anyone to predict for a specific individual”. (01/25/07 Parsons depo 48).
While Parsons stated that the effects of ethanol begin with the very first drink, he acknowledged that Florida law defines impairment in terms of the ability to perform ordinary tasks such as walking, talking and making decisions. Parsons said that he would not consider as important the fact that trained, impartial observers did not observe evidence of problems walking, talking or making decisions. Parsons stated that who might have caused the accident is of no consequence in conducting a retrograde extrapolation. (01/25/07 Parsons depo 43-44). These are not factors which Parsons would consider important in determining whether a person is impaired by ethanol. (01/25/07 Parsons depo 45-46).
Initially, Parsons says that the only way to determine whether a person is impaired by ethanol is a chemical test. He denies that a field test is sufficient for that purpose. He says that the individual should be placed into a laboratory and given divided attention tests and mental evaluation tests, none of which can be done in a reliable way in a field setting. (01/25/07 Parsons depo 47).
Parsons said that he would expect a trained observer to be able to see signs of impairment in just about everyone at a .20 level. He qualified this statement with the assertion that long-term alcoholics with even higher levels have “fooled police officers who have field sobriety tested them”. (01/25/07 Parsons depo 48).
Parsons was asked about a statement made in a deposition on April 25, 1994, during which he said that most people will begin to see an alcoholic stupor at levels between 0.20 – 0.30. Parsons said he still believes in the truth of the statement. Stupor was used to refer to loss of conscious control of the body, ability to walk, ability to stay awake, ability to maintain consciousness, and difficult in being aroused from slumber. (01/25/07 Parsons depo 63).
During his hearing testimony, Dr. Bruce Goldberger reported that his laboratory results were .107 and .108, which is approximately 20-30% less than the results reported by Parsons. (7/27/07 Rothe hearing 163, 191). Dr. Goldberger agreed that the blood sample was degrading and decomposing at the time it was tested at the University of Florida (UF). (7/27/07 Rothe hearing 163).
Dr. Goldberger admitted that he was unaware of the storage conditions for the blood during the extended period of time between blood withdrawal and retesting at the UF laboratory. (7/27/07 Rothe hearing 163). Dr. Goldberger agreed that unrefrigerated blood samples will decompose over time. (7/27/07 Rothe hearing 193). Parsons admitted that the blood sample was not refrigerated after testing because “it doesn’t need to be.” It was returned to Parsons’ unrefrigerated evidence locker. (4/7/06 Parsons depo 14, 51).
Dr. Goldberger opined within a reasonable degree of scientific certainty that the UF results were consistent with Parsons’. (7/27/07 Rothe hearing 164). Dr. Goldberger assumed that ethanol is not created in a preserved blood sample. Dr. Goldberger did not run tests to see if preservative was present and, if so, whether it was present in amounts sufficient to impede the growth of microbes which might create ethanol. Dr. Goldberger dismissed studies establishing ethanol increases on the basis that they involve a “purposeful injection of bacteria or microbes into a tube”. (7/27/07 Rothe hearing 192, 193 – 194, 168).
Dr. Goldberger did not explain the core inconsistency in his testimony. If a properly preserved sample will last indefinitely, then why was the Rothe sample degrading and decomposing? There was only an 18-month delay between Parsons’ testing and Dr. Goldberger’s retesting. (7/27/07 Rothe hearing 163). Although Dr. Goldberger maintained that researchers injected microbes into blood samples, he did not dispute that such microbes can produce ethanol. Candida albicans has been found to generate over .10 g/dl of ethanol in just 22 hours. “The Effect of Microbial Contamination of the Blood Sample on the Determination of ethanol Levels in Serum”, P. Blume and D.J. Lakatua, American Journal of Clinical Pathology, Vol. 60, No. 5, pp 700-702, Nov. 1973 (finding also that candida albicans growth is not impeded by sodium fluoride).
Dr. Goldberger concluded that the ethanol concentration only went down in the Rothe case over time. He based his conclusion upon the accuracy of Parsons’ testing. Dr. Goldberger did agree that, hypothetically, the ethanol concentration of a sample could go up or down or stay the same over time. (7/27/07 Rothe hearing 190, 192).
In this case, there was independent evidence that Freddie Rodriguez was drinking heavily, but that Rothe was drinking far less. Both the Rothe and Rodriguez samples were run in the same batch and immediately adjacent to each other. Rodriguez tested zero. Rothe tested .14. If Parsons mixed up the Rothe and Rodriguez samples, then Dr. Goldberger seemed to agree that he would not know whether the ethanol concentration his lab tested had gone up or down or stayed the same in storage.
Parsons said that men and women eliminate ethanol at different rates. The rate for a healthy man is .018 per hour; for a healthy female, it is .02 per hour. (4/25/94 Parsons depo 9, 28; 01/25/07 Parsons depo 32). Parsons further asserted that the elimination rates between normal people do not differ by much. The primary differences involve those who either have impaired liver function or are long-term ethanolics. He asserted that long-term ethanolics can eliminate ethanol at rates two-to-three times greater than normal people. (4/25/94 Parsons depo 13-14). Parsons has recognized studies showing different elimination rates among persons from different ethnic groups. (4/25/94 Parsons depo 28-29).
Parsons considers body weight in calculating a blood-ethanol concentration resulting from consumption of a specific amount of ethanol. However, Parsons does not adjust for a fat percentage. He would use the same estimating procedure for a 150-pound man and a 225-pound man, regardless of height or musculature. (4/25/94 Parsons depo 9-13).
According to Parsons, absorption rates are correlated with only two factors: when the person last ate and how full the stomach might be. There are no other factors which would effect absorption rates. There is no correlation with ethnic factors. (4/25/94 Parsons depo 15).
Parsons asserted that his retrograde extrapolation represents a minimum elimination rate within the range. He asserted that he does not use a higher elimination rate, because he cannot determine what would be reasonable for a given subject. He noted that long-term ethanolics might eliminate at 2-3 times the elimination rate for an average person, but represent a very small portion of the population. (01/25/07 Parsons depo 33).
Parsons speculated that the standard elimination rates do not apply to such long-term ethanolics. Parsons speculated that there might be biochemical markers which could indicate long-term ethanolism, but acknowledged that the existence of such markers would be outside of his field. Parsons finally conceded that he does not assume that Rothe was a long-term ethanolic. (01/25/07 Parsons depo 53-55). Parsons speculated the elimination rate of such ethanolics could range from .01 to .03 per hour (4/7/06 Parsons depo 48).
Parsons is familiar with the term “post-absorptive state”. It is the point at which the concentration of ethanol in the digestive tract is equal to the concentration of ethanol in the blood stream. (01/25/07 Parsons depo 35). The rate of absorption would be effected by the amount of food in the stomach. Most people would absorb most ethanol on an empty stomach within 20-30 minutes of the last drink, but it would be rare for a person to take more than 45 minutes for such absorption. (01/25/07 Parsons depo 36).
Absorption times for persons on full stomachs “are pretty much doubled”. These absorption periods would run approximately 1-1/2 hours, but not to exceed two hours. (01/25/07 Parsons depo 36); (4/7/06 Parsons depo 49-50).
Parsons stated that he could not determine if a person was post-absorptive without knowing when the subject took his last meal and what the composition of the last meal might have been. Without this information, it would be difficult for Parsons to reach a conclusion. (01/25/07 Parsons depo 37-38).
Parsons assumed, in some situations, that the entire drink is still in the stomach at the time of the accident. He did this in order to underestimate the actual BAC at the time. He said he uses a method which never overestimates the actual BAC. He agreed that more information is always better. (01/25/07 Parsons depo 39-41).
Parsons said that effects of ethanol on behavior are “highly individualistic”. The observed effects may depend upon the person’s drinking history. A total non-drinker might start showing behavioral changes with a single drink. According to Parsons, ethanol effects can be observed on judgment and mental processes at just .02, for most people a single beer. (4/25/94 Parsons depo 15-16).
According to Parsons, as BAC rises, there is less self-inhibition and more gregariousness. Above .05, there are observable physical affects. However, there is much variation among individuals. “To say how much of an effect, to what degree or even an average, is very difficult to do.” (4/25/94 Parsons depo 17-18).
Parsons noted that he could not talk specifics about a specific individual without experiments on that person. Still, even one drink has an effect on everybody. There would be an ethanolic stupor at .20-.30 concentrations, unconsciousness at approximately .30, and death at approximately .40. (4/25/94 Parsons depo 18-20).
Parsons seemed to disagree with Florida law on how to evaluate impairment:
Q: Well, if Florida law talks about impairment defined in terms of the ability to perform ordinary tasks such as walking, talking and making decisions, would you agree with that?
Q: And if trained, impartial observers don’t observe evidence of problems walking, talking or making decisions, would that be a factor that might affect a determination whether or not that person would be impaired by ethanol at that given time?
A: No, those observations I’ve explained are not reliable.
(01/25/07 Parsons depo 46).
On the issue of retrograde extrapolation, Parsons acknowledged that it would be “irresponsible to attempt to give a specific blood-alcohol level at a specific time without knowing exactly when each drink was taken and what was in the stomach…it’s a ball park range”. (01/25/07 Parsons depo 49-50).
Parsons reviewed a website posting from January 30, 2004, message number 2129 on the Yahoo Forensic Science Bulletin Board. Parsons acknowledged the document. In the posting, Parsons used the term “gargantuan proportions” to estimate possible error in retrograde extrapolation. Parsons acknowledged the comment, but said that it was an observation of mistakes made by others who were using averages instead of ranges. (01/25/07 Parsons depo 50-51).
Parsons was never asked to do a retrograde extrapolation as of the date of this deposition. He did agree that the blood was drawn at approximately 8:45 P.M. on July 31, 2005. He also agreed that the accident occurred at approximately 6:15 P.M., some 2-1/2 hours before the blood was drawn. (01/25/07 Parsons depo 31-32).
Parsons agreed with the assumption that Rothe was post-absorptive at the time of the accident. If an elimination rate of 0.02 per hour was used, then Parsons agreed that Rothe would have had an ethanol concentration of approximately .19 at the time of the accident Parsons pointed out, however, that he would use a range instead of a specific number. Parsons prefers to use a lower elimination rate in order to avoid “prejudicing” the individual. (01/25/07 Parsons depo 51-53). Parsons denied that the lower elimination rate would prejudice Rothe by reducing the blood-ethanol concentration at the time of the accident. (01/25/07 Parsons depo 53).
Parsons agreed that Freddy Rodriguez’s blood tested “zero”. Based upon this reading, Parsons says that Rodriguez must have finished drinking long enough before the blood sample to have returned to zero at the time of the draw. (4/7/06 Parsons depo 47).
In his deposition, Dr. Goldberger said he always uses a .015 elimination rate per hour. (12/18/06 Goldberger depo 5). Dr. Goldberger acknowledged that there were higher and lower elimination rates, but he chose to use the .015 rate as being the most fair. Using his assumed elimination rate, Dr. Goldberger estimated that Rothe had a BAC of approximately .1775 at the time of the accident. (12/18/06 Goldberger depo 11).
Dr. Goldberger discussed the signs of obvious impairment for somebody with a concentration over a .20. While acknowledging research on standard field sobriety tests, Dr. Goldberger suggested that some extremely intoxicated people do not show overt signs of intoxication, even at .20 or above. He did, however, acknowledge the Marcella Burns’ study prepared for NTSA. (12/18/06 Goldberger depo 6-8).
Law enforcement officers and medical professionals did not observe evidence of impairment by drugs or ethanol. There was testimony about Rothe’s mental status, which most attributed to the death of his best friend and to injuries sustained by other friends who were on the boat. Dr. Goldberger acknowledged that normal people might have a range of emotional responses to an incident resulting in the death of a close friend. Dr. Goldberger agreed that the emotional response could not be predicted based upon the use of ethanol or the lack of use of ethanol. (12/18/06 Goldberger depo 18-19).
Parsons said his office does not do post-mortem analyses. Parsons asserted that post-mortem alcohol samples are much more difficult to interpret. There are a variety of processes that occur at the point of death. Parsons said he would not feel comfortable attempting to do a post-mortem extrapolation, because its not something he has done on a regular basis or at all. (01/25/07 Parsons depo 63-66). The key facts relating to the Charles O’Hara blood sample were related to Parsons. Parsons identified the number of hours without refrigeration as a source of concern, as well as the introduction of transfusive materials into the heart. The putrifactive processes would tend to increase the blood-alcohol concentration, whereas the dilution by transfusion would tend to lower it. (01/25/07 Parsons depo 67-71).
Research has recommended that post-mortem samples be preserved with at least 1% sodium fluoride (for example 100 mg/10ml blood). Other research has recommended 1.5-2% sodium fluoride and refrigeration. Lesser concentrations can allow ethanol producing microorganisms to grow. Importantly, sodium fluoride does not inhibit the growth of candida albicans. See “The Collection and Handling of the Blood Alcohol Specimen”, S. Kaye, American Journal of Clinical Pathology, Vol. 74, pp 743-746, 1980.
In his master’s thesis, Michael Wagner measured ethanol concentrations over a 10-year period for preserved and refrigerated post mortem samples. In the first year, 53% of the samples increased ethanol concentrations from 2.0-20%. In the second year, 27% of the samples increased from 7.1-19.4%. Similar increases in ethanol concentration were reported throughout the 10-year study. “Stability of Drugs of Abuse in Biological Specimens”, www.ccwu.edu/Theses_Wagner.