The possible influence of micro-organisms and putrefaction in theproduction of GHB in post-mortem biological fluid

Simon Elliotta,*, Pauline Loweb, Amanda Symondsb

aRegional Laboratory for Toxicology, City Hospital, Dudley Road, Birmingham B18 7QH, UKbDepartment of Microbiology, City Hospital, Dudley Road, Birmingham B18 7QH, UK

Accepted 30 October 2003

Abstract

In recent years, the post-mortem production of the drug of abuse gamma-hydroxybutyric acid (GHB) in biological fluids

(e.g. blood and urine) has caused various interpretative problems for toxicologists. Previously, other researchers have shown

certain microbial species (Pseudomonas spp. and Clostridium aminobutyricum) possess the necessary enzymes to convert

GABA to GHB.

A preliminary investigation involving putrefied post-mortem blood indicated there was no observed relationship between

‘‘endogenous’’ GHB concentrations and concentrations of common putrefactive markers (tryptamine and phenyl-2-ethylamine).

Microbiological analysis identified the presence of various micro-organisms: Clostridia spp., Escherichia coli, Proteus vulgaris,

Enterococcus faecalis and Aeromonoas spp.

Equine plasma, human blood and urine samples were inoculated with these and an additional micro-organism (Pseudomonas

aeruginosa) and incubated at 22 8C for 1 month. Following comparison with control samples and pre-inoculation concentra-

tions, the data indicated an apparent production of GHB in unpreserved P. aeruginosa inoculated blood (2.3 mg/l). All other

fluoride-preserved and unpreserved samples (including controls) had GHB concentrations <1 mg/l. Although this concentration

is lower than is typically associated with ‘‘endogenous’’ post-mortem GHB concentrations, this paper proposes a potential

microbial production of GHB with time.

# 2003 Elsevier Ireland Ltd. All rights reserved.

Keywords: GHB; Micro-organisms; Endogenous; Toxicology

1. Introduction

The therapeutic agent and drug of abuse gamma-hydro-

xybutyric acid (GHB) has been found to be an endogenous

compound in animals and humans [1]. Endogenous human

concentrations in life have been found to be typically less

than 10 mg/l in urine and less than 4 mg/l in blood/plasma

[2,3]. Following exogenous ingestion, these concentrations

are greatly increased and can be 10–1000-fold higher in

such specimens. However, due to the rapid metabolism and

excretion of GHB, blood/plasma and urine concentrations

can be comparable to endogenous levels within 8–12 h

following ingestion [4].

GHB has also been found to be present in post-mortem

biological fluid in fatalities where GHB ingestion has not

been suspected [5]. Such findings have indicated concentra-

tions in blood up to 200 mg/l; however, the data may be

technique specific and influenced by the storage conditions.

Incubation experiments at various temperatures in the pre-

sence and absence of sodium fluoride indicated GHB con-

centrations were greater in unpreserved samples with time

and at increased temperatures [6]. In addition, although it

was initially believed that GHB was not present in post-

mortem urine, this has now been found not to be the case and

GHB appears to be present in various biological fluids

(including vitreous humour), albeit at concentrations less

than blood [7]. Increasing data have been obtained and

presented using a commonly used method of GHB analysis

involving the use of gas chromatography with mass spectro-

metry (GC–MS) following trimethylsilane derivatisation

Forensic Science International 139 (2004) 183–190

* Corresponding author. Tel.: þ44-121-507-5204;

fax: þ44-121-507-6021.

E-mail address: simontox@yahoo.co.uk (S. Elliott).

0379-0738/$ – see front matter # 2003 Elsevier Ireland Ltd. All rights reserved.

doi:10.1016/j.forsciint.2003.10.018

with deuterated GHB (GHB-D6) as an internal standard.

Studies indicate that GHB is present in post-mortem blood at

concentrations less than 30 mg/l and in post-mortem urine at

concentrations less than 20 mg/l [7]. In a previous study by

this author involving 40 GHB-unrelated fatalities, GHB was

found in preserved/unpreserved post-mortem blood at a

mean concentration of 12 mg/l (range 2–29 mg/l) and in

preserved/unpreserved urine at a mean concentration of

5 mg/l (range 0–18 mg/l). Although there was no overall

difference in the concentrations measured in preserved and

unpreserved samples, in cases of an extended post-mortem

interval (>21 days, death to analysis), preserved blood

concentrations were lower than the unpreserved samples

but this difference was not observed in urine.There has been

much speculation regarding the source of endogenous GHB.

In life, the biochemical pathway of GHB biosynthesis and

metabolism is thought to involve the conversion of gamma

aminobutyric acid (GABA) to succinic semialdehyde (via

GABA transaminase) and subsequent conversion to succi-

nate (by succinic semialdehyde dehydrogenase) or to GHB

by succinic semialdehyde reductase [8]. Conversely, GHB

can be converted to succinic semialdehyde (SSA) by GHB

dehydrogenase (in the cytosol) or by GHB-ketoacid trans-

hydrogenase (in the mitochondria) [9]. A further proposal

involves the conversion of possible endogenous 1,4-butane-

diol to GHB via alcohol dehydrogenase (ADH) and alde-

hyde dehydrogenase, however this is thought to be unlikely

[10]. After death, a rapid rise in GHB is well documented in

animals [11], and in humans concentrations are invariably

higher in post-mortem fluid than ante-mortem fluid. Various

authors have indicated the possibility that a post-mortem

reduction in Krebs cycle activity could result in an accu-

mulation of substrates such as succinate and consequently

succinic semialdehyde being shifted to produce GHB [12].

Rat studies by Eli and Cattabeni using microwave irradiation

or decapitation, indicated the involvement of enzymes in

post-mortem GHB production [11]. Higher GHB concentra-

tions were attained following decapitation, purportedly due

to the concomitant cessation of enzymic activity at the time

of death. Drugs that interfere with GABA metabolism also

reduced GHB production, supporting a potential role of

GABA conversion to GHB after death [11]. Furthermore,

there may also be an enzymatic conversion of GABA (result-

ing from post-mortem decomposition) to GHB. A potential

source of GABA has been suggested to occur via the con-

version of putrescine (1,4-butanediamine) to GABA in var-

ious multi-step processes [13]. Putrescine is present in all

cells and has been noted to increase during putrefaction.

Nonetheless, at present, the exact reason for this rise is

unclear and could involve any of the processes stated above.

For compounds such as ethanol, microbial fermentation

has been found to play a significant role in post-mortem

production [14]. Until now, the potential role of microbial

fermentation in post-mortem GHB production has not been

specifically studied. However, in the late 1950s, microbiol-

ogists discovered certain micro-organisms had the necessary

enzymes to utilise GABA or GHB as their main carbon

source [15,16]. One species of Pseudomonas was found to

yield succinate from GABA (via SSA) [16]. It was also

found that another species of Pseudomonas (that can grow in

a GHB medium) and Clostridium aminobutyricum (that

grows in a GABA medium) had systems for the conversion

of GABA to SSA (by GABA transaminase) and subsequent

conversion to GHB catalysed by GHB dehydrogenase under

appropriate conditions (Fig. 1). Unfortunately, the identities

of both Pseudomonas spp. were not cited in the papers.

Based on this information it is therefore possible that after

death, if present, such micro-organisms could utilise GABA

and produce GHB.

The purpose of this study was to firstly determine the

concentrations of GHB in putrefied post-mortem blood and

relate such concentrations to the presence of any micro-

organisms and known products of putrefaction (tryptamine

and phenyl-2-ethylamine). Secondly, the possible influence

of such micro-organisms on the production of GHB was

evaluated in human blood in vitro.

Fig. 1. Pathway showing the possible production of GHB from GABA by microbes (C. aminobutyricum and unspecified Pseudomonad)

possessing the necessary enzymes.

184 S. Elliott et al. / Forensic Science International 139 (2004) 183–190

2. Materials and methods

2.1. Chromatographic equipment

Gas chromatography with mass spectrometry (GC–MS)

analysis was performed using a Hewlett Packard HP 6890

GC system fitted with a HP 5793 series mass selective

detector. A J & W Scientifics DB5 MS capillary column,

30 m � 0:25 mm, 0.25 mm film thickness was used (J & W

Scientifics, Bracknell, UK).

HPLC–DAD analysis was performed using a P580 low

pressure pump, STH 585 column oven, ASI-100 auto-

sampler and a UVD340S diode array detector all from

Dionex (Camberley, UK). A Waters Spherisorb S5OD/CN

4:6 mm � 150 mm cartridge column (Elstree, UK), pro-

tected by a 4 mm � 10 mm guard column of Spherisorb

S5ODS2 was used for the analysis. Data acquisition was

handled by a Dionex Chromeleon software package with the

diode-array detector recording spectral data between 200

and 595 nm.

2.2. Materials

Sulphuric acid and sodium carbonate were supplied by

BDH Chemicals (Poole, UK). Ethyl acetate was obtained

from Merck Eurolab (Poole, UK). 1.0 M triethylammonium

phosphate (TEAP) buffer (pH 3.0) was supplied by Fluka

(Dorset, UK). The HPLC-grade acetonitrile was supplied by

Rathburns Chemicals Ltd., (Walkerburn, UK). HPLC-grade

1-chlorobutane was obtained from Fisher Scientific Inter-

national, (Loughborough, UK). Bis(trimethylsilyl)-trifluor-

oacetamide (BSTFA) containing 1% trimethylchlorosilane

(TMCS) derivatising agent and nalbuphine were supplied

by Sigma–Aldrich Chemicals (Poole, UK). Pure 0.1 mg/ml

methanolic GHB-D6 sodium salt was supplied by LGC

Promochem (Welwyn Garden City, UK). Pure GHB sodium

salt, tryptamine and phenyl-2-ethylamine were obtained

from Sigma–Aldrich (Poole, UK) and were used to prepare

reference and calibration standards for the formal identifica-

tion and quantitation of these compounds in the specimens

analysed. GHB calibration standards of 1, 2.5, 5, 10, 25

and 50 mg/l were prepared in blank equine plasma (pre-

screened). An internal quality control standard of 7.5 mg/l

was also produced in equine plasma. Tryptamine and

phenyl-2-ethylamine calibration standards of 0.125, 0.25,

0.5 and 1 mg/l were prepared in blank equine plasma (pre-

screened). Post-mortem samples were diluted with blank

plasma prior to analysis.

2.3. Extraction methods for biological specimens

For GHB analysis: in a 2 ml Eppendorf capped tube, 50 ml

of 5 mg/l GHB-D6 internal standard (in cold 0.05 M H2SO4)

was added to 100 ml aliquots of blood, urine or microbial

broth. A total of 500 ml acetonitrile extraction solvent was

added and vortex mixed for 30 s followed by centrifugation

at 13,000 rpm for 5 min. A total of 500 ml of upper (solvent)

layer was transferred to a 12 ml glass vial and evaporated to

dryness at 45 8C under air. A total of 75 ml of BSTFA 1%

TMCS derivatising agent was added followed by brief vortex

mixing and incubation at 90 8C for 5 min in a heating

block. After cooling, this was transferred to a GC–MS vial

insert ready for injection. The injection volume was 1 ml.

This method does not involve the conversion of GHB to

GBL.

For tryptamine and phenyl-2-ethylamine HPLC analysis:

0.5 ml of 0.2 M Na2CO3 buffer (including 2 mg/l nalbuphine

internal standard) was added to 0.5 ml of post-mortem blood

in a 12 ml polypropylene tube and vortexed briefly. A total of

5 ml of 1-chlorobutane was added followed by mechanical

shaking for 3 min and centrifugation at 4500 rpm for 3 min.

The upper (solvent) layer was transferred to a second

polypropylene tube and 100 ml of 0.05 M H2SO4 added.

Following mechanical shaking for 3 min and centrifugation

at 4500 rpm for 3 min, the upper (solvent) layer was aspi-

rated and the remaining lower (aqueous) acid layer trans-

ferred to a HPLC vial for injection. The injection volume

was 30 ml.

2.4. Chromatography conditions

GC–MS analysis was based on a DB5-MS capillary

column (30 m, 0.32 mm i.d.) with a temperature gradient

starting at 60 8C (for 2 min) and ramping to 180 8C (20 8C/

min for 6 min) then ramping to 250 8C (50 8C/min for 1 min)

and finally 70 8C post run. Total run time 9 min with mass

spectrometer data acquired in the full scan mode between 6.2

and 7.0 min. Derivatised GHB (GHB-diTMS) and GHB-D6

(GHB-D6-diTMS) were detected using the 233 and 239 ion

fragments, respectively. Identification of GHB was con-

firmed using mass spectral matching of a library standard

entry. GHB-diTMS eluted at 6.70 min and GHB-D6-diTMS

eluted at 6.68 min.

For HPLC-DAD analysis, the tryptamine and phenyl-2-

ethylamine quantitation procedure was based on 10% acet-

onitrile (in 25 mM TEAP buffer) isocratic elution at a flow

rate of 2 ml/min. The column temperature was maintained

at 25 8C.

2.5. Analysis of putrefied post-mortem blood

Six fatalities unrelated to GHB ingestion were routinely

investigated where there had been evidence of decomposi-

tion of the body/specimens. Findings from one fatality where

there was no obvious evidence of putrefaction could be

used as a ‘‘control’’ specimen. All specimens were stored

at �20 8C prior to analysis. The concentrations in the post-

mortem blood of two common markers of putrefaction were

measured; tryptamine and phenyl-2-ethylamine, in addition

to GHB. Furthermore, microbiological analysis was per-

formed in order to determine the identity/presence of any

micro-organisms.

S. Elliott et al. / Forensic Science International 139 (2004) 183–190 185

2.6. Microbiological analysis

The identification of micro-organisms detected in the

blood and urine specimens was undertaken using standard

microbiological procedures. It has been previously found

that various indole producing micro-organisms have been

identified in putrefactive samples, thus the ‘‘indole test’’ was

included in the initial investigation [17].

Based on the initial identification of certain microbial

species in post-mortem samples and the historical data

outlining the possible ability of Pseudomonas spp. and

Clostridium aminobutyricum to produce GHB from GABA,

eight microbes were chosen for the study. All organisms

were grown in a suitable media/broth and incubated under

the appropriate conditions to facilitate growth (e.g. Clos-

tridia sp. were maintained under anaerobic conditions).

However, Clostridium aminobutyricum could not be grown

successfully to enable inoculation and was therefore not

included in the study.

1. Aeromonas hydrophila.

2. Clostridium perfringens (anaerobic).

3. Clostridium sordellii (anaerobic).

4. Enterococcus faecalis (ATCC 92912 strain).

5. Escherichia coli (NCTC 10418 strain).

6. Proteus vulgaris.

7. Pseudomonas aeruginosa (NCTC 10662 strain).

In order to assess the possible production of GHB in the

presence of micro-organisms, 9.5 ml unpreserved and fluor-

ide-preserved (0.2 %) equine plasma, fresh human blood and

urine (donated by the author) were inoculated separately

with 0.5 ml (high titre, 1 � 109 ml�1) of the seven chosen

organisms to provide a final titre of 1 � 107 per ml of matrix.

Each specimen was then incubated at room temperature

(22 8C).

Initial aliquots of each inoculation broth, equine plasma,

human blood and human urine ‘‘controls’’ were analysed for

the presence of GHB (time 0). Subsequent aliquots of each

inoculated specimen (plasma, blood and urine—preserved

and unpreserved) were analysed after 30 days along with the

uninoculated ‘‘control’’ samples. In addition, equine plasma,

human blood and urine ‘‘inoculated’’ with equivalent sterile

Table 1

Results of putrefied post-mortem biological fluid analysis for the presence micro-organisms, putrefactive compounds and GHB

Number Comments Sample Tryptamine

(mg/l)

Phenyl-2-ethylamine

(mg/l)

GHB

(mg/l)

Micro-organisms

detected

Indole test

(þ/�)

1 Heroin OD Blood (fluoride) 0.79 1.69 27 Clostridium sordellii þive

2 Drowning Blood (plain) 0.07 1.16 3 Aeromonas spp. þive

3 Methadone OD Blood (plain) ND 13.92 13 Escherichia coli þive

4 Drowning Blood (plain) 3.83 1.00 16 Clostridia spp.,

Clostridium acetobutyricum

�ive

Blood (fluoride) 2.75 0.66 9

5 (Control) Drowning Blood (plain) ND ND 12 None �ive

Blood (fluoride) ND ND 4

Urine (plain) ND ND 1

6 Alcohol OD Blood (plain) ND 28.38 NA Escherichia coli þive

7 Drowning Blood (plain) 0.09 0.96 NA Proteus vulgaris þive

ND: not detected (limit of detection ¼ 0.05 mg/l); NA: not analysed (insufficient sample).

Table 2

GHB concentrations measured in human blood, human urine and

equine plasma following microbial inoculation and incubation at

22 8C for 1 month

Organism Matrix þ/�fluoride (0.2%)

GHB concentration (mg/l)

þ1 month (þ/� fluoride,

respectively)

P. aeruginosa Blood þ, � 1.3, 2.3

Plasma þ, � ND, ND

Urine þ, � 1.3, 1.7

P. vulgaris Blood þ, � ND, ND

Plasma þ, � ND, ND

Urine þ, � 1.7, 1.8

E. coli Blood þ, � ND, ND

Plasma þ, � ND, ND

Urine þ, � 1.8, 2.4

E. faecalis Blood þ, � ND, <1

Plasma þ, � ND, ND

Urine þ, � 1.1, 1.5

A. hydrophila Blood þ, � <1, ND

Plasma þ, � ND, ND

Urine þ, � <1, <1

C. sordellii Blood þ, � <1, <1

Plasma þ, � ND, <1

Urine þ, � 1.2, 1.1

C. perfringens Blood þ, � ND, ND

Plasma þ, � ND, ND

Urine þ, � <1, <1

ND: not detected (limit of detection: 0.5 mg/l).

186 S. Elliott et al. / Forensic Science International 139 (2004) 183–190

(microbe-free) broth was analysed after 30 days. Further-

more, all aliquots were re-cultured to determine if the

relevant microbes were still viable and to identify any

microbial contaminants.

3. Results

3.1. Analysis of putrefied post-mortem blood

In six fatalities where there was pathological evidence of

advanced post-mortem putrefaction, phenyl-2-ethylamine,

tryptamine and GHB were identified in the post-mortem

blood. No phenyl-2-ethylamine or tryptamine was detected

in ‘‘control’’ post-mortem specimens where there had been

no evidence of putrefaction or decomposition, however

GHB was detected. In all cases, GHB ingestion was not

suspected. The concentrations of phenyl-2-ethylamine, tryp-

tamine and GHB detected in these specimens are shown in

Table 1. A maximal GHB concentration of 27 mg/l was

measured in a preserved blood specimen that exhibited

extensive putrefaction but with corresponding relatively

low tryptamine and phenyl-2-ethylamine concentrations

(0.79 and 1.69 mg/l, respectively). Following microbial

analysis of all cases, Clostridia spp. (including C. sordellii

and C. Acetobutylicum), E. coli, P. vulgaris, E. faecalis and

Aeromonoas spp. were isolated. No such micro-organisms

were identified in the ‘‘control’’ blood sample.

3.2. GHB production in the presence of micro-organisms

The results of GHB analysis in the presence/absence of

seven microbial species in plasma, blood and urine are

shown in Table 2. Overall, there was little GHB detected

in any of the samples (limit of quantitation 1 mg/l). At time

0, following analysis of pre-inoculated equine plasma,

human blood and human urine, GHB was only detected

in the urine at a concentration consistent with normal

endogenous levels (2 mg/l). After 1 month incubation, no

GHB was detected in any of the equine plasma inoculations

and the urine concentrations were not significantly affected.

In the blood aliquots, a higher GHB concentration was

detected (2.3 mg/l) in unpreserved blood inoculated with

P. aeruginosa compared to the ‘‘control’’ blood samples

(<1 mg/l) (Table 3 and Fig. 2a and b). However, in the

corresponding fluoride-preserved blood inoculated with P.

aeruginosa, a GHB concentration of 1.3 mg/l was detected.

Nonetheless, in the other blood samples inoculated with the

other micro-organisms studied, GHB concentrations were

less than the limit of quantitation (<1 mg/l). Microbiological

assessment after 1 month following inoculation indicated

that the microbes in the blood and plasma samples were still

Table 3

GHB concentrations measured in ‘‘control’’ samples at time 0 (prior to inoculation) and 1 month following inoculation and incubation at 22 8C(where appropriate)

Organism Matrix þ/�fluoride (0.2%)

GHB concentration

(mg/l), time 0

GHB concentration

(mg/l) þ 1 month

None Blood ND <1

None Plasma ND ND

None Urine 2.4 2.9

C. perfringens Anaerobe Broth 1.2 1.4

C. sordellii Anaerobe Broth ND ND

E. coli Aerobe Broth <1 ND

P. aeruginosa Aerobe Broth ND ND

A. hydrophila Aerobe Broth ND ND

P. vulgaris Aerobe Broth ND ND

E. faecalis Aerobe Broth ND ND

None (Anaerobe Broth) Blood � <1

Blood þ ND

Plasma � ND

Plasma þ ND

Urine � 1.3

Urine þ 1.5

None (Aerobe Broth) Blood � ND

Blood þ ND

Plasma � ND

Plasma þ ND

Urine � 1.7

Urine þ 1.7

ND: not detected (limit of detection: 0.5 mg/l).

S. Elliott et al. / Forensic Science International 139 (2004) 183–190 187

Fig. 2. (a) Selected ion chromatogram of unpreserved human blood inoculated with Pseudomonas aeruginosa following 1 month incubation at

22 8C. Sample and GHB-D6 internal standard extracted using acetonitrile. GHB concentration measured to be 2.3 mg/l with calibration

standards. (b) Ion fragmentation mass-spectrum comparison of the peak at 6.70 min and a commercially available (NIST) library entry for

trimethylsilane derivatised GHB (4-butanoic acid). This confirmed the presence of GHB in the P. aeruginosa inoculated blood sample.

188 S. Elliott et al. / Forensic Science International 139 (2004) 183–190

viable. However, those in the urine samples were not viable

and had stopped growing. As the microbes remained viable

in the blood even in the presence of sodium fluoride, this

suggests that the concentration of fluoride present (0.2%)

may not have been sufficient to prevent microbial growth

(at this titre) and hence potentially inhibit GHB production.

A fluoride concentration of 1–2% may be more appropriate.

4. Discussion

Preliminary study of post-mortem blood specimens that

exhibited morphological features of advanced putrefaction

(e.g. highly clotted, high viscosity, distinctive smell, dar-

kened colour) detected the presence of GHB and known

putrefactive compounds tryptamine and/or phenyl-2-ethyla-

mine (Table 1). In addition, microbiological analysis

identified Clostridia spp. (including C. sordellii and

C. Acetobutylicum), E. coli, E. faecalis, P. vulgaris and

Aeromonas spp. Although there was no obvious correlation

between the tryptamine, phenyl-2-ethylamine and GHB

concentrations themselves, the presence of tryptamine

appeared to be associated with Clostridia spp. However,

there was no such potential microbial relationship observed

with either phenyl-2-ethylamine or GHB. In particular, the

GHB concentrations were comparable to endogenous con-

centrations previously found in post-mortem blood and do

not necessarily imply a directly proportional relationship

between post-mortem putrefaction and GHB production (i.e.

the greater the degree of putrefaction, the greater the GHB

concentration). Specifically, the concentrations found in the

non-putrefied ‘‘control’’ samples (12 mg/l unpreserved,

4 mg/l preserved) were similar to those found in extensively

putrefied samples with higher tryptamine and phenyl-2-

ethylamine concentrations. The limited data (Cases 4 and

5) do however, further support the possible influence of

sodium fluoride on reducing endogenous GHB concentra-

tions in post-mortem blood. This is likely to be due to the

disruption of the microbial and/or non-microbial enzymatic

systems involved in putrefaction and GHB production.

The data from the preliminary microbiological study

allowed the initial identification of microbes potentially

involved in post-mortem GHB production. Subsequent

inoculation of equine plasma, human blood and human urine

with the microbial species initially identified (Clostridia

spp., E. coli, P. vulgaris, E. faecalis and Aeromonas spp.)

and another common microbe (P. aeruginosa) indicated an

increased GHB production in blood in vitro in the presence

of P. aeruginosa after 1 month incubation (2.3 and 1.3 mg/l

in unpreserved and fluoride-preserved blood, respectively)

(Fig. 2a and b and Tables 2 and 3). There was no additional

GHB production in the equine plasma or urine after 1 month.

As previously mentioned, unspecified Pseudomonas organ-

isms have been found to possess the necessary enzymes for

conversion of GABA to GHB [16]. The conversion of

the intermediate SSA to GHB is an equilibrium reaction

catalysed by GHB dehydrogenase, which was found to have

optimal activity at pH 7 in the presence of diphosphopyr-

idine nucleotide. It is therefore conceivable that such a

reaction has occurred with P. aeruginosa utilising the GABA

in the blood to produce GHB. However, even at the high titre

used (1 � 109 ml�1), the resultant GHB concentration is not

as high as that observed in typical instances of endogenous

GHB post-mortem (e.g. up to �30 mg/l, measured using the

same technique described in this paper). This could be due to

various reasons, for example; study of an inappropriate

Pseudomonad (i.e. other Pseudomonads may be more effi-

cient at GHB production), too high titre and/or the absence

of particular synergistic co-factors or additional microbes/

yeast. In fact, although P. aeruginosa is a common environ-

mental contaminant, it is not always present in the body and

is usually present as a result of pathogenic infection (parti-

cularly in those with suppressed immune systems). It is

unlikely, therefore, that this organism is the sole (if at all)

microbial source of GHB post-mortem, as GHB appears to

be virtually ubiquitous in such specimens. It does not rule

out, however, the possibility of ‘‘infection’’ after death and

resultant GHB production. In addition, although C. Amino-

butyricum could not be grown successfully to allow inclu-

sion in this study, the possibility of its involvement in GHB

synthesis remains. Nevertheless, these preliminary findings

and the data previously published by Nirenberg and Jakoby,

indicate a possible role of micro-organisms in the production

of GHB after death. In particular, as previously speculated, a

microbial production of GHB would explain the apparent

increase in GHB concentrations with time and apparent

inhibition in the presence of sodium fluoride. It is proposed,

therefore, that at the time of death or shortly afterwards, a

biochemical disturbance (such as the disruption of the Krebs

cycle) could produce an initial increase in GHB production.

Subsequently, microbial utilisation of any GABA present by

certain microbial species (such as Pseudomonas spp.) could

produce more GHB with time. As the microbes tested did not

appear to be viable long-term in urine, and as urine has fewer

enzymes and less GABA than blood, this would also explain

the lower GHB concentrations found in post-mortem urine

even following an extended post-mortem interval. This

supports the preferred use of urine (and potentially vitreous

humour) in interpreting GHB concentrations in post-mortem

biological fluid.

5. Conclusions

Analysis of extensively putrefied samples (as indicated by

morphological observations and the presence of known

putrefactive compounds) did not indicate a proportional

relationship between GHB concentration and the extent of

putrefaction. In vitro inoculation of plasma, blood and urine

with common micro-organisms and those species found in

the putrefied samples, indicated that Pseudomonas spp. (e.g.

P. aeruginosa) may have a role in the production of GHB in

S. Elliott et al. / Forensic Science International 139 (2004) 183–190 189

blood with time. Nevertheless, further confirmatory study is

necessary before definitive conclusions can be made.

Acknowledgements

The authors would like to thank Dr. Robin Braithwaite

and Mr. Shashi Parmar for their assistance during the studies

and preparation of this manuscript.

References

[1] S.P. Bessman, W.N. Fishbein, Gamma-hydroxybutyrate, a

normal brain metabolite, Nature 200 (4912) (1963) 1207–1208.

[2] D.T. Yeatman, K. Reid, A study of urinary endogenous

gamma-hydroxybutyrate (GHB) levels, J. Anal. Toxicol. 27

(2003) 40–42.

[3] S.P. Elliott, Gamma-hydroxybutyric acid (GHB) concentra-

tions in humans and factors affecting endogenous production,

Forensic Sci. Int. 133 (1) (2003) 9–16.

[4] P.V. Kavanagh, P. Kenny, J. Feely, The urinary excretion of

g-hydroxybutyric acid in man, J. Pharm. Pharmacol. 53

(2001) 399–402.

[5] E.L. Fieler, D.E. Coleman, R.C. Baselt, GHB concentrations

in pre and post-mortem blood and urine (letter), Clin. Chem.

44 (1998) 692.

[6] R.G. Stephens, D.E. Coleman, R.C. Baselt, In vitro stability

of endogenous GHB in post-mortem blood, J. Forensic Sci. 44

(1) (1999) 231.

[7] S.P. Elliott, Further evidence for the presence of GHB in post-

mortem biological fluid: implications for the interpretation of

findings, J. Anal. Toxicol., in press.

[8] R.H. Roth, Formation and regional distribution of g-

hydroxybutyric acid in mammalian brain, Biochem. Pharmac.

19 (1970) 3013–3019.

[9] E.E. Kaufman, T. Nelson, C. Goochee, L. Sokoloff, The

purification and characterisation of an NADPþ-linked alcohol

oxido-reductase which catalyses the interconversion of

GHB and succinic semialdehyde, J. Neurochem. 32 (1979)

699–712.

[10] K. Sakurada, M. Kobayashi, H. Iwase, M. Yoshino, H.

Mukoyama, T. Takatori, K. Yoshida, Production of gamma-

hydroxybutyric acid in post-mortem liquid increases with

time after death, Toxicol. Lett. 129 (3) (2002) 207–217.

[11] M. Eli, F. Cattabeni, Endogenous g-hydroxybutyrate in rat

brains areas: postmortem changes and effects of drugs

interfering with g-aminobutyric acid metabolism, J. Neuro-

chem. 41 (2) (1983) 524–530.

[12] J.D. Doherty, S.E. Hattox, O.C. Snead, R.H. Roth, Identifica-

tion of endogenous g-hydroxybutyrate in human and bovine

brain and its regional distribution in human, J. Pharamcol.

Exp. Ther. 207 (1978) 130–139.

[13] L.J. Marinetti, g-Hydroxybutyric acid and its analogs,

g-butyrolactone and 1,4-butanediol, in: S.J. Salamone (Ed.),

Forensic Science: Benzodiazepines and GHB: Detection and

Pharmacology, Humana Press, Totowa, NJ, 2001.

[14] H.A. Sulkowski, A.H.B. Wu, Y.S. McCarter, In vitro

production of ethanol in urine by fermentation, J. Forensic

Sci. 40 (6) (1995) 990–993.

[15] J.K. Hardman, T.C. Stadtman, Metabolism of g-amino acids:

fermentation of g-aminobutyric acid by Clostridium amino-

butyricum, J. Bacteriol. 79 (1960) 544–548.

[16] M.W. Nirenberg, W.B. Jakoby, Enzymatic utilization of g-

hydroxybutyric acid, J. Biol. Chem. 235 (4) (1960) 954–960.

[17] J. Ramsey, P. Hayward, Technical note: Indole, TIAFT Bull.

(1978) 26–27.

190 S. Elliott et al. / Forensic Science International 139 (2004) 183–190