To Name :
To Email :
From Name :
From Email :
Comments :

Evidence-Based Reviews

‘I’m sober, Doctor, really’: Best biomarkers for underreported alcohol use

When and how to use highly specific combinations to assess withdrawal risk.

Vol. 7, No. 9 / September 2008

Hospitalized patients who are not truthful about their alcohol consumption may be at risk for an unplanned withdrawal. Self-reports of alcohol use—such as CAGE and the Alcohol Use Disorders Identification Test (AUDIT)—are valid, inexpensive, and noninvasive, but patients easily can feign results.1 Biochemical measures are more objective, and combinations of markers are an effective tool to detect recent heavy drinking in the 10% to 25% of patients who underreport alcohol use.2

Biochemical measures can detect acute alcohol intoxication and recent prolonged drinking. Because marker levels return to normal after long-term abstinence, ongoing monitoring can help detect a relapse before a patient admits to it.3

This article presents 3 cases in which biochemical markers helped prevent alcohol withdrawal in patients who denied alcohol abuse. We discuss why we ordered biochemical tests and which combinations provided highly sensitive results.

CASE 1: Depression and substance abuse

Ms. C, age 39, presents with bleeding gums due to excessive warfarin, which she takes prophylactically for a history of deep vein thrombosis. She is seen by the psychiatric consultation service for depression—which she says she has experienced since “the day I was born”—and substance abuse that includes a history binge drinking. Ms. C says she has stopped drinking and remained abstinent for the past year because she is fearful of further damaging her kidneys. She also denies psychosis. She does not have a history or symptoms of hepatobiliary or hematologic disease.

Challenge. Despite Ms. C’s self-reported 1 year of sobriety, her history of binge drinking and depression calls for evaluating her alcohol withdrawal risk. Laboratory markers of alcohol abuse are the only means to assess her recent drinking behavior.

Discussion. Lab results include serum albumin of 3.4 g/dL, total bilirubin of 0.3 mg/dL, total protein of 6.3 g/dL, aspartate aminotransferase (AST) of 13 U/L, alanine aminotransferase (ALT) of 19 U/L, alkaline phosphatase of 136 U/L, and blood ammonia level of 37 μg/dL. Gamma-glutamyl transferase (GGT) is elevated at 104 U/L (normal range for women: 0 to 45 U/L). Mean corpuscular volume (MCV) is elevated at 101 fL (normal range 80 to 100 fL).

The combination of elevated MCV and GGT has a 95% sensitivity for alcohol abuse.4 GGT levels become elevated after 24 hours to 2 weeks of heavy alcohol consumption and return to normal within 2 to 6 weeks of abstinence, which allows them to detect binge drinking. MCV takes 6 to 8 weeks of heavy drinking—we which we define as consuming ≥40 grams of alcohol/day5—to become elevated and returns to normal within 3 months of abstinence.

These data provide evidence that Ms. C recently consumed substantial amounts of alcohol. As a result, we start her on alcohol withdrawal precautions (AWP).

Markers of alcohol abuse

Biochemical markers commonly used to detect alcohol abuse (Table 1) include:

  • blood alcohol level (BAL)
  • MCV
  • liver function tests (LFTs) such as ALT, AST, and GGT
  • carbohydrate deficient transferrin (CDT).

Table 1

By the numbers: Biomarkers of excessive alcohol consumption









Blood test normal range

<60 mg/L

Women: 0 to 45 U/L Men: 0 to 53 U/L

10 to 34 U/L

8 to 37 U/L

80 to 100 fL

Blood test abnormal range

>1.3% of total transferrin concentration

Women: >45 U/L Men: >53 U/L

Levels rarely exceed 500 U/L

Levels rarely exceed 300 U/L

>100 fL

Time to elevation

2 to 3 weeks

24 hours to 2 weeks

3 to 7 days

3 to 7 days

After 6 weeks

Time to descent to normal levels

2 to 4 weeks of abstinence

2 to 6 weeks of abstinence

Half-life 12 to 24 hours

Half-life 37 to 57 hours

3 months

Dose-response of alcohol

60 g/d

80 to 200 g/d

≥40 g/d

≥40 g/d

≥40 g/d


55% to 90%a-e

37% to 85%b, f, g

AST:ALT ratio >2:1 has a 70% sensitivity and 92% to 100% specificity for alcoholic-induced liver diseaseh-j

20% to 70%b,k

Relapse sensitivity

55% to 76%a,l,m




92% to 97%a,b

18% to 93%a,b,e

64% to 66%b,k,n

Positive predictive value

46% to 75%c,g



Negative predictive value

72% to 98%a,c,g

69% to 92%a,e,g


AST: aspartate aminotransferase; ALT: alanine aminotransferase; CDT: carbohydrate deficient transferrin; GGT: gamma-glutamyl transferase; MCV: mean corpuscular volume

Source: Reference Citations: click here

BAL can document acute alcohol intoxication, but its use is limited because alcohol has a 4-hour half-life and an elimination rate of 7 grams/hour—equivalent to 1 drink/hour.6 (A “drink” typically is defined as a 12-ounce bottle of beer or wine cooler, a 5-ounce glass of wine, or 1.5 ounces of 80-proof distilled spirits.) Therefore, BAL will identify as false negatives alcohol-dependent patients who abstain from alcohol within 24 hours of testing.

MCV is an index of the average volume of erythrocytes. Macrocytosis occurs when the volume exceeds 100 fL. Elevated MCV is the most typical morphologic abnormality associated with excessive alcohol consumption7,8 and macrocytosis—sometimes without associated anemia—is often evident in persons with alcoholism. MCV elevates after 6 weeks of alcohol misuse and may remain elevated for up to 3 months after a person has stopped drinking.9

Because patients with disorders unrelated to alcohol use can have elevated MCV, alone it is not a useful screening marker for alcohol abuse.10 Additionally, because macrocytosis can persist under strictly controlled alcohol abstinence, MCV is not a reliable clinical indicator of relapse.11

LFTs measure enzymes and proteins. ALT, AST, and GGT are the most relevant for detecting heavy drinking. An AST:ALT ratio >2:1 supports a suspicion of alcohol abuse.12 More than 90% of patients with an AST:ALT ratio of 2:1 have alcoholic liver disease. This increases to more than 96% if the ratio is 3:1.13

GGT is an enzyme concentrated in the liver, bile ducts, and kidneys; normal range is 0 to 45 U/L (for females) or 53 U/L (for males).14 GGT levels >30 U/L correlate with alcohol consumption of >4 drinks per day.15 GGT has a half-life of 14 to 26 days and remains elevated for 4 to 6 weeks after drinking cessation, which make it useful for monitoring abstinence in treatment programs.16 Sensitivity ranges from 37% to 85% and specificity is as high as 93% in nonmedical populations.17 Although nonalcoholic liver disease can elevate GGT in persons who do not abuse alcohol, 50% to 72% of GGT elevations can be explained by excessive alcohol consumption.18

CDT is a newer biomarker used to monitor alcohol consumption. The most accurate way to express CDT level is as a percentage of total transferrin concentration. This method accounts for individual variations in transferrin levels, thus minimizing false positives.18 In persons who consume >4 or 5 drinks per day for 2 weeks or more, CDT is >1.3% of total transferrin.19 Unfortunately, because it is expensive and requires sophisticated test methodology, CDT testing is not available at most hospitals.20

Combinations improve detection

Each biochemical measure has strengths and weaknesses as a marker for determining patients’ alcohol consumption (Table 2). CDT and GGT show the highest sensitivity for heavy drinking, and CDT has a higher specificity than GGT (Table 3).21,22 Relapse to alcohol use after abstinence may be best identified by a simultaneous 30% increase in CDT and GGT.5

Because GGT has a longer half-life than CDT, its diagnostic efficiency in detecting alcohol relapse may not develop until 4 weeks after alcohol detoxification, whereas CDT may become clinically useful for detecting relapse as early as 1 week after detoxification.23

Table 2

Biomarkers of alcohol use: Strengths and weaknesses





High specificity for alcohol use; few factors cause false positives
High sensitivity in distinguishing alcoholics from social drinkers
Marker of relapse and abstinence from drinking
Confirmatory test for patients suspected of alcohol abuse

Low sensitivity; more valuable to confirm than exclude heavy drinking
Cost (average $30/assay) and low availability of testing
Likely less sensitive for women and younger patients compared with men
Poor screening tool for alcohol use in general population


Elevations precede alcohol-induced liver damage
High specificity in patients with suspected alcohol abuse
Effective marker for patients suspected of binge drinking
Inexpensive (<$10)

Can be falsely elevated by liver and biliary disease, smoking, obesity, and medications that induce microsomal enzymes
Low sensitivity makes it a poor screening tool in general population
Poor marker of relapse

AST:ALT >2:1

Highly sensitive and specific for alcohol-induced liver damage

Enzyme elevations can be detected only after periods of heavy drinking
Elevations secondary to liver damage at the hepatocellular level (after fatty changes)


Accuracy similar in male and female patients
Elevations in suspected cases of alcohol use indicate chronicity of drinking
Routine laboratory test

Poor biomarker for relapse
False positives caused by liver disease, hemolysis, bleeding disorders, anemia, folate deficiency, and medications that reduce folate
Low sensitivity and specificity for alcohol use make it a poor screening tool for alcohol abuse

AST: aspartate aminotransferase; ALT: alanine aminotransferase; CDT: carbohydrate deficient transferrin; GGT: gamma-glutamyl transferase; MCV: mean corpuscular volume

Table 3

Interpreting diagnostic test performance





Percent of persons with disease who test positive

High value is desirable for ruling out disease (low false-negative rate)


Percent of persons without disease who test negative

High value is desirable for ruling in disease (low false-positive rate)

Positive predictive value

Percent of positive test results that are true positives

Probability that a person with a positive test result has the disease

Negative predictive value

Percent of negative test results that are true negatives

Probability that a person with a negative test result is disease-free

Source: References 21,22

There is evidence that combining tests can improve alcohol use detection.24 For example, Dolman et al25 found that the ability of the AUDIT questionnaire to correctly predict which patients would experience alcohol withdrawal increases when it is used in combination with biochemical markers. Specifically, the positive predictive value of an AUDIT score ≥8 increased from 17% to 47% when found in combination with ≥2 abnormal biochemical marker levels; the study looked at GGT, ALT, AST, and MCV. Sensitivity was 94% and specificity was 98%.

Similarly, combinations of biochemical markers—especially CDT and GGT—have improved detection of alcohol use and subsequent risk of withdrawal.26 Table 4 provides a summary of studies that evaluated using combinations of biochemical markers.4,5,27-31

Table 4

Combining biomarker tests: An effective approach





Morgan et al4



Hietala et al5


Mundle et al29


Bell et al30


Sillanaukee et al31


GGT + AST:ALT >2:1

Gluud et al27


Morgan et al4


MCV + AST:ALT >2:1

Kawachi et al28


Morgan et al4


GGT + MCV + AST:ALT >2:1

Morgan et al4



Sillanaukee et al31


* Sensitivity for detecting excessive alcohol consumption

AST: aspartate aminotransferase; ALT: alanine aminotransferase; CDT: carbohydrate deficient transferrin; GGT: gamma-glutamyl transferase; MCV: mean corpuscular volume

Consider patients’ comorbidities

Patients at risk for underreporting alcohol use include those with unemployment histories, previous alcohol treatment, and higher scores on the Alcohol Dependence Scale (18.5, SD=8.1).2 Interpret biochemical testing results in the context of a patient’s overall clinical picture.

The following 2 case patients denied or underreported recent alcohol use but we determined they were at high risk for an alcohol disorder because of their medical and/or psychiatric histories. Analysis of biochemical markers helped assess the risk of alcohol withdrawal.

CASE 2: Altered mental status

Family members bring Mr. N, age 44, to the hospital because of his odd behavior. He presents with paranoid delusions and an inappropriate elated mood. His medical history includes acquired immune deficiency syndrome (AIDS). After cerebrospinal fluid analysis, computed tomography of the head, electroencephalogram, and metabolic workup are within normal limits, the patient is diagnosed with human immunodeficiency virus (HIV) mania and is admitted.

On admission, Mr. N denies alcohol use. A blood alcohol/urine toxicity screen is negative. One day after admission, Mr. M develops elevated blood pressure and tachycardia and reports headache and nausea.

Did you miss this content?
Botulinum toxin for depression? An idea that’s raising some eyebrows