Urinary ethyl glucuronide testing detects alcohol consumption in alcoholic liver disease patients awaiting liver transplantation



This study compared measurement of urinary ethyl glucuronide (EtG), a conjugated minor ethanol metabolite with a longer detection window than ethanol itself, with breath alcohol testing and self-report as ways to disclose recent drinking by 18 liver transplant candidates with an alcoholic liver disease diagnosis that underwent an addiction group therapy program. At each therapy session, patients were questioned about any alcohol consumption in the intervening time, and they also performed a mandatory breath alcohol test, while observed urine samples for measurement of EtG were delivered on a voluntary basis. None of the patients ever admitted to intake of alcohol, and only 1 of 127 breath alcohol tests turned out positive. However, 9 patients showed positive EtG results in 24 (49%) of 49 urine samples. The individual frequency of urine samples being positive for EtG ranged from 22% to 100% with a mean value of 57%. Because 6 patients refused to provide urine on a total of 18 occasions, alcohol use might have been even more common. These results underscore the uncertainty of self-report data and the low sensitivity of breath alcohol testing as ways to disclose recent drinking, and underline the necessity of introducing sensitive and specific objective measures of recent alcohol consumption, such as EtG, in the transplantation setting. Liver Transpl 13:757–761, 2007. © 2007 AASLD.

End-stage alcoholic liver disease (ALD) has long been considered a controversial indication for liver transplantation but is today an accepted and common treatment option.1–3 About 20-30% of all liver transplantations are presently conducted because of ALD in the final stage of liver insufficiency, and the overall clinical outcomes compare favorably to those documented for other causes of liver disease.4 Although chronic alcohol abuse is the underlying cause for transplantation in ALD patients, a considerable number (∼20-25%) return to harmful drinking after the operation.5, 6 Alcohol recidivism after transplantation may lead to a reduced long-term survival rate.7–9 Activities to promote abstinence from alcohol in ALD transplant patients are therefore highly recommended.9

Recently relapse rates of 15-25% have been reported among ALD patients on the liver transplant waiting list.10 Besides an increased risk for return to heavy drinking after transplantation, ALD patients should refrain from alcohol before transplantation to improve their health. Abstinence can reduce severe complications of the liver disease,11 sometimes even to the extent where transplantation can be delayed or is no longer required.

Most studies on alcohol consumption patterns have relied on patient self-reports as records of drinking, although it is recognized that this information is often inaccurate.12, 13 Measurement of ethanol in breath or body fluids is a specific alcohol test, but it has low sensitivity because of rapid ethanol elimination (∼0.1 g/kg/hour) and will only detect very recent intake.14 Alcohol biomarkers are also used for objective monitoring of alcohol use,15 but they have not been widely used in transplant studies. Markers of chronic abuse or alcohol-related organ damage, such as carbohydrate-deficient transferrin and γ-glutamyltransferase, are not sensitive enough to spot occasional or sustained moderate intake.16, 17

A new laboratory method that has gained popularity as a sensitive and specific way to detect recent drinking is ethyl glucuronide (EtG).15, 18 After alcohol intake, >95% of the ingested ethanol dose is metabolized in a 2-stage oxidation process by the action of hepatic alcohol and aldehyde dehydrogenase enzymes, while the remainder is mainly excreted unchanged in urine, sweat, and breath. A very small amount (<0.1%) of the ethanol undergoes conjugation reactions to produce EtG as well as ethyl sulfate (EtS), which are excreted in the urine.18–20 Compared with ethanol, the detection time for EtG in the urine is considerably longer: >6 hours after intake of a low ethanol dose (7 g),21 >20 hours after ∼25-40 g,19 and typically >50 hours after drinking at least 100 g ethanol.22 Accordingly, even in cases when ethanol is no longer detectable, a positive finding of EtG and EtS in urine provides a strong indication that the person was recently drinking alcohol.23

Alcohol dependence is a chronic medical illness that necessitates an early and continuous monitoring and treatment. This study compared urinary EtG measurement with breath alcohol testing and self-report as ways to disclose recent drinking in ALD patients undergoing addiction group therapy before liver transplantation.


ALD, alcoholic liver disease; EtG, ethyl glucuronide; EtS, ethyl sulfate.


The study enrolled liver transplant candidates with an ALD diagnosis admitted to the University Hospital of Essen. Participation in the study was voluntary, and all patients signed written informed consent. The study was approved by the ethics committee at the University Hospital of Essen.

All transplant patients are initially evaluated for their suitability as recipients on the basis of a consensus of clinical, biochemical, and psychosomatic examinations. The classification into eligible, at-risk, or high-risk patients depends on the length of the abstinence period before admission and on existing risk factors for recidivism, as detailed in Figure 1. Patients with ongoing heavy alcohol use are transferred to inpatient rehabilitation, and those classified as at risk or high risk are recommended to take part in a manualized addiction rehabilitation group therapy.24, 25 For the high-risk patients, participation is a precondition for being listed. The therapy program consists of 12 sessions (90 minutes every 2 weeks) with psychoeducation and problem-solving training, and it aims to reach abstinence from alcohol.

Figure 1.

Schematic illustration of the method used for classifying alcoholic liver disease (ALD) candidates for liver transplantation into high-risk, at-risk, and eligible patients.

Between February and August 2005, 18 patients (9 men and 9 women; mean age 51 years; range 31-68) took part in the group therapy and were included in the study. Of the participants, 10 were classified as high-risk patients and 6 as at-risk patients, and 2 were eligible for transplantation but took part in the therapy because they were interested. Ten patients met ICD-10 criteria for an alcohol dependence syndrome and 2 for harmful alcohol use, and 13 (72%) reported being abstinent for less than 6 months.

At the beginning of each session, the patients were asked whether they had consumed any alcohol in the intervening time, and they also underwent a mandatory breath alcohol test. The breathalyzer device (Dräger Alcotest 7410) has a detection limit of ∼0.01 g/L and was calibrated every 6 months. Results of the breath ethanol measurement were communicated directly to the patients. Because the therapy aims to achieve abstinence from alcohol, a positive breath test leads to further counseling, such as regular contact with the therapist or inpatient addiction rehabilitation, and their place on the waiting list may also be frozen.

Before starting the group session, urine samples for measurement of EtG were collected on a voluntary basis under observation by 2 study nurses. Voluntary sampling and not communicating the EtG results to the group therapists, transplant physicians, or to the patients themselves were stipulated by the ethics committee, to protect patients from any disadvantage until the EtG test was approved for routine use. The treatment was not influenced or delayed on the basis of the EtG results, and they have not become part of the medical records. The urine specimens were stored at −70°C until shipped in dry ice for measurement of EtG by a routine liquid chromatography–mass spectrometry method.21 All positive results were verified by liquid chromatography–mass spectrometry or mass spectrometry by the presence of the product ions of EtG (m/z 113, 85, and 75 from m/z 221).


The disincentives for drinking by ALD patients in the liver transplant setting are very high. Patients are not put on the waiting list until they are abstinent from alcohol. In case of slip or relapse drinking while on the waiting list, their status may be frozen for some time, or they can even be removed from consideration for transplantation. It may consequently be especially hard to get honest self-report information about alcohol use in this population.26 None of the participants in the present study ever admitted to intake of alcohol.

Besides being asked about any intake of alcohol in the intervening time, mandatory breath alcohol tests were performed in connection with each session. The patients never refused to undergo a breath test, and only 1 out of a total of 127 breath tests turned out positive (Table 1). Accordingly, the self-report data and results of breath alcohol tests indicated that alcohol use was not common in this population.

Table 1. Attendance Rates and Results of Alcohol Tests in Candidates With Alcoholic Liver Disease for Liver Transplantation Undergoing an Addiction Group Therapy Program
PatientsnGroup session attendance, mean (median; range)Breath ethanol tests (n)Positive breath ethanol test, n (%)Urine samples (n)Positive urinary EtG, n (%)Urinary EtG level, mean (median; range) (mg/L)Patients with positive EtG, n (%)
  • Abbreviation: EtG, ethyl glucuronide.

  • *

    Ethanol concentration, 0.8 g/L (0.08%).

Total187.1 (7.5; 1-11)1271 (0.8%)*9624 (25%)12.1 (1.0; 0.1-130)9 (50%)
Eligible24.5 (4.5; 3-6)9053 (60%)2.3 (0.5; 0.4-6.0)1 (50%)
At risk66.0 (5.5; 1-11)360336 (18%)0.5 (0.4; 0.1-1.3)3 (50%)
At high risk108.2 (9.0; 1-9)821 (1.2%)5815 (26%)18.7 (4.0; 0.1-130)5 (50%)

Nonetheless, the results of the urinary EtG measurements demonstrated that half of the liver transplant candidates had been drinking alcohol at least once during the period of the group therapy, but this was not admitted when questioned, nor was it identified by the breath alcohol testing. Altogether, 96 voluntary urine specimens (range 1-10 samples per patient; median 6 samples) from 16 of the 18 patients were available for measurement of EtG. For 8 patients, all samples delivered (n = 47) were negative for EtG, whereas 24 (49%) of 49 samples obtained from the other 9 patients showed positive results (Fig. 2).

Figure 2.

Distribution of positive ethyl glucuronide (EtG) values in voluntary observed urine samples (n = 16) collected from ALD candidates for liver transplantation while undergoing an addiction group therapy program. The EtG results for one high-risk patient for whom all urine samples (n = 8) were positive are provided separately (inset).


Because 6 patients refused to provide urine on a total of 18 occasions, alcohol use might have been even more common, assuming that this occurred at times when they had been drinking. If EtG testing had been mandatory, more reliable information about alcohol consumption would have been obtained. Another limiting factor was the low attendance in the 12 group therapy sessions (individual range 1-11 sessions; mean attendance rate 59%) (Table 1), which was mainly the result of frequent inpatient medical treatment.

The frequency of positive EtG results for the 9 recidivistic patients ranged from a low of 22% in 1 subject (2 of 9 samples) up to 100% in 3 subjects (1 of 1, 2 of 2, and 8 of 8 samples), with an overall mean value of 57% (median 50%). Except for 1 high-risk patient who showed 8 of 8 positive urine samples in the range 2.1-130 mg/L EtG (mean 34.3 mg/L; median 24.2 mg/L) (Fig. 2, inset), the remaining 16 results were between 0.1-6.0 mg/L, with a mean of 1.02 mg/L (median 0.36 mg/L) (Fig. 2). Patients with an alcohol dependence diagnosis (n = 16) showed a higher frequency of EtG positive samples (84%) than those with harmful alcohol use (n = 2; 33%). That the EtG results observed in clinical urine samples are usually in the low concentration range agrees with previous observation21 and indicated that the ALD patients planned their alcohol intake to be able to present a negative breath test when attending the scheduled therapy sessions.13 However, because EtG remains detectable for many hours longer than ethanol, they still showed positive, albeit mostly low, urinary EtG results (Fig. 2). It should be pointed out that a low EtG level could indicate recent intake of a small dose of alcohol as well as a large intake ∼1-2 days ago.18, 21, 23 In addition, because EtG is sensitive to urine dilution,19 samples may routinely be checked for their creatinine content to reduce the risk for falsely low or false-negative results due to ingestion of large volumes of fluid before sampling.

An important finding of this study was that urinary EtG testing disclosed alcohol use in half of the ALD candidates for liver transplantation undergoing an addiction group therapy program, although this was not detected by breath alcohol testing or admitted by the patients when questioned. Regardless of the risk classification group, the frequency of positive EtG results was the same. These results underscore the uncertainty of self-report data and low sensitivity of breath alcohol testing as methods to monitor abstinence. Furthermore, the present results support previous reports that alcohol dependence is a common chronic medical illness, and addiction counseling is also recommended in the transplant setting. The number of patients with alcohol dependence was even higher than previously reported,6, 26 and all except the 2 volunteers were classified as at risk or high risk for recidivism. Accordingly, application of a more sensitive alcohol test to identify recent drinking, such as urinary EtG and EtS, is needed in the transplantation setting for early identification of alcohol dependence and early detection of relapse, for motivational feedback of verified abstinence to the patients and as objective outcome measures of recidivism rates. Whether liver disease affects the validity of the EtG test has not been examined specifically, but the results of the present study verified that EtG is a valuable test for recent alcohol intake in patients who also have severe ALD.