After consumption of alcoholic beverages, the bulk of the ethanol dose (95%–98%) is eliminated in a 2-stage oxidation process mainly in the liver, first to acetaldehyde by alcohol dehydrogenase and then further to acetic acid by aldehyde dehydrogenase. The remainder is excreted unchanged in urine, sweat, and expired air (1). In addition, a very small fraction (<0.1%) of the ingested ethanol undergoes phase II conjugation reactions to produce ethyl glucuronide (EtG) and ethyl sulfate (EtS) (2)(3), catalyzed by uridine diphosphate-glucuronosyltransferase or sulfotransferase, respectively. EtG and EtS are eventually excreted in the urine. As both of these nonoxidative direct ethanol metabolites show much longer elimination times than ethanol itself (3), the interest in EtG and EtS has focused largely on their use as sensitive and specific biomarkers of recent alcohol intake with clinical and forensic applications (4)(5). A positive finding of EtG and/or EtS provides a strong indication that the person was recently drinking alcohol, even when the ethanol concentration has returned to 0 or is no longer measurable.
Glucuronide and sulfate conjugates of endogenous and exogenous origin are cleaved by β-glucuronidase and sulfatase, enzymes that are widely distributed among animals and plants. β-Glucuronidase is also present with high activity in most strains of Escherichia coli (6). Because this characteristic is rather unique for E. coli compared with other bacterial species, β-glucuronidase assays with chromogenic and fluorogenic substrates have been developed for the rapid and specific identification of E. coli in clinical microbiological diagnostics and for testing contamination of food and water (7)(8). Sulfatase activity has been detected in many different bacteria (9), but not in E. coli (10)(11), or only in very low amounts (12).
E. coli is the most common bacterium isolated in clinical laboratories and is also the predominant pathogen (∼80%) in urinary tract infections (UTIs) (13). This study, therefore, evaluated whether the presence of E. coli or other common pathogens in urine specimens, resulting from UTIs or possible contamination during sampling and handling, could give false-negative EtG and EtS results in the detection of recent alcohol consumption because of hydrolysis by bacterial β-glucuronidase and sulfatase.
Fresh clinical urine specimens (n = 46; stored refrigerated) containing confirmed bacterial growth at a density of 103 to >105 colony-forming units (CFU)/mL and with >80% of samples containing >105 CFU/mL, as identified by culture on standard solid media, were obtained from the microbiology laboratory at the Karolinska University Hospital. Specimens were collected consecutively from the routine pool of infected urine samples and were also selected to include different pathogens. The samples were supplemented with 1 mg/L each of EtG (Medichem Diagnostics) and EtS (TCI); they were then split into 3 tubes (without preservatives), which were placed at −20, 4, and 22 °C. Urine without the addition of EtG and EtS served as controls. At the start and after 1, 2, and 5 days of storage at 4 and 22 °C, samples were placed at −20 °C until taken for analysis of EtG and EtS by a sensitive and specific direct electrospray liquid chromatographic–mass spectrometric (LC-MS) method (3). LC-MS analysis was performed in the negative ion mode, with selected ion monitoring of the pseudomolecular ions at m/z 125 for EtS (Mr 126.1) and m/z 130 for EtS-D5 (pentadeuterated internal standard, prepared by reaction of ethanol-D6 with chlorosulfonic acid; Sigma-Aldrich) and at m/z 221 and m/z 226 for EtG (Mr 222.1) and EtG-D5 (internal standard; Medichem Diagnostics), respectively. The EtG and EtS concentrations in unknown samples were calculated from peak-area ratios to the internal standard by reference to a calibration curve (single determinations). The detection and quantification limits were ∼0.05 mg/L and ∼0.1 mg/L, respectively, for both compounds (3).
The different pathogens identified among the 46 clinical urine specimens included in this study are listed in Table 1⇓ . E. coli was the predominant bacterial species, and it is also the primary causative agent of UTI (∼80% of cases), although other uropathogens, such as Staphylococcus (10%–15%), Klebsiella, Enterobacter, Enterococcus, Proteus, and Pseudomonas species infrequently cause disease (13). For both EtG and EtS, 9 urine specimens were found positive before supplementation (range, 0.3–39.3 mg/L), and for only EtG (0.2 mg/L), 1 was found positive.
In the majority (68%) of the urine specimens containing E. coli, a marked decrease in the EtG concentration was observed over time after storage at 22 °C, but not at 4 °C or −20 °C (see examples in Fig. 1⇓ ). These results agree with those of previous studies, showing that most, but not all, E. coli strains possess β-glucuronidase activity (14)(15)(16). In 3 specimens, complete hydrolysis of the added EtG (1 mg/L) was noted after 24 h of storage at 22 °C. In the 2 specimens that contained the highest EtG concentrations, 37.3 and 39.3 mg/L before supplementation, the EtG concentrations were decreased to 11.2 and 0.7 mg/L, respectively, after storage for 5 days at 22 °C.
In 1 of 3 urine specimens containing Klebsiella pneumoniae and the single specimen containing Enterobacter cloacae, a gradual disappearance of EtG was also observed after storage at 22 °C, but the rate of EtG hydrolysis was slower than for most specimens with E. coli tested under the same conditions (data not shown). These observations are consistent with previous results, which indicate that a few other pathogens causing UTI, including some strains of Klebsiella and Enterobacter, also possess low β-glucuronidase activity (15)(16).
To determine whether chemical preservatives could prevent bacterial hydrolysis of EtG, we added 1 mg/L each of EtG and EtS to 8 UTI urine specimens confirmed positive for E. coli and incubated them in tubes containing sodium fluoride (10 mg NaF/mL of urine) or without preservatives. In 6 of these specimens, marked or complete hydrolysis of EtG was observed in the tubes without preservatives after storage for 5 days at 22 °C, whereas EtG was found to be stable on storage in the tubes containing NaF.
No decrease in the EtS concentration on storage was observed in any of the UTI specimens examined in this study (Table 1⇓ ). Likewise, 2 commercial preparations of sulfatase (type H1 from Helix pomatia and type VI from Aerobacter aerogenes; Sigma-Aldrich) were found not to hydrolyze EtS in urine samples.
EtG (4)(17)(18) and, more recently, EtS (3)(19)(20) have been introduced as sensitive and specific biomarkers for detection of recent alcohol consumption, with major advantages compared with conventional ethanol testing, including a much longer detection time (i.e., improved sensitivity) with retained high specificity. Our results demonstrated, however, that EtG may not be stable on storage if the urine specimens taken for analysis are infected with pathogens possessing β-glucuronidase activity (typically E. coli). In contrast, EtS was indicated to be completely stable to bacterial hydrolysis. The disappearance rate of EtG was temperature dependent, and refrigeration or freezing of samples, or use of sampling tubes containing fluoride preservatives, was effective in preventing, or markedly reducing, hydrolysis.
In some urine specimens, the tested EtG concentration (1 mg/L) was no longer detectable after storage at room temperature for 1 day. This concentration is in the range commonly observed in clinical practice (21), where ∼40% of all EtG-positive urine samples collected for testing of recent alcohol consumption showed values <10 mg/L. An even more apparent disappearance of EtG was noted in the 2 urine specimens containing high concentrations before supplementation (37.3 and 39.3 mg/L), in which much or most of the EtG had been hydrolyzed after storage for 5 days at room temperature. Accordingly, if biological specimens are stored under appropriate conditions (e.g., room temperature) or transported or mailed to the laboratory without refrigeration, freezing, or use of preservatives, the possibility of false-negative and falsely low EtG results always has to be considered. Even if precautions are taken to prevent bacterial degradation of EtG in the collected specimens, hydrolysis could possibly have taken place before sampling, e.g., in urine retained in the bladder during sleep of patients with UTIs or other pathologic conditions (22) or between time of death and autopsy in postmortem cases.
In summary, the results of the present study demonstrated that EtG, but not EtS, is sensitive to bacterial hydrolysis, particularly when specimens are infected by E. coli, which is the most common source of UTIs. Given that UTIs caused by E. coli are among the most frequent bacterial infections encountered in clinical practice (13), this represents an obvious risk factor for false-negative and falsely low EtG results. Because EtG and EtS show similar windows of detection after alcohol consumption (3), it may therefore be advantageous and recommended to measure EtS instead of EtG for detection or confirmation of recent drinking, or at least to combine EtG with EtS analysis, which is possible by LC-MS (3)(23). In any case, ensuring that specimens are refrigerated or frozen and/or that sampling tubes contain fluoride preservatives to prevent bacterial growth is recommended practice.
This work was supported by grants from the Karolinska Institutet (to Anders Helander).
- © 2005 The American Association for Clinical Chemistry