Implications of testing for LFTs and Alcohol Markers

1
Implications of testing for LFTs and Alcohol
Markers
Dr. Peter Miller AMU/ASU 21 May, 2008
2
History of LFTs and life insurance

• AIDS in early 80s created need for blood HIV
  testing
• Why not add other tests if youre already taking
  blood,?
• Blood chemistry profile (BCP) included glucose,
  lipids, renal function, serum proteins and liver
  function tests, plus urinalysis
• The Milliman and Roberts report on HIV testing
  and the BCP justified the benefits of testing
• Now came the challenge as to how to underwrite
  the abnormal results

3
Early underwriting elevated LFTs

• Confusion due to ignorance on significance of
  elevations
• We had never heard of Bayes' theorem, named after
  the Reverend Thomas Bayes (17021761),
  sensitivity, specificity, prevalence, pre-test
  probability and post-test probability
• Elevated GGT became the marker of alcohol abuse,
  so that we were rating or declining many cases
• In fact, what was the extra mortality of elevated
  LFTs?
• Some simple solutions were needed

4
Easy solutions to the problems of elevated LFTs

• Recognition of the causes of mild/moderate
  elevation
• Hepatic causes fatty liver, alcohol abuse,
  chronic HBV HCV, haemochromatosis, autoimmune
  hepatitis, Wilsons disease, a-1-antitrypsin
  deficiency
• Nonhepatic causes celiac disease, inherited
  disorders of muscle metabolism, acquired muscle
  diseases, strenuous exercise
• The major extra mortality is associated with
  fatty liver, chronic viral hepatitis and alcohol
  abuse
• Identifying alcohol abuse is very difficult, even
  in clinical practice
• The simple way would be to perform a reflex test
  of the serum sample that insurance labs were soon
  storing, e.g. HBV
• Also led to introduction of haemoglobin
  associated acetaldehyde (HAA) and carbohydrate
  deficient transferrin (CDT) as reflex tests
• Add an LFT calculator so the underwriter doesnt
  have to think and the problem was solved!!

5
What is the mortality of elevated LFTs?

• Mom, I cant believe it was so quick
• those liver function tests are killers!

6
The AALE studyAlcohol abuse and liver enzyme
study

• AAIM, SOA, ALU study of 82,262 insured lives with
  policies issued 1989-95 and exposed to 1997
  mortality experience compared to the 1990-95
  Select Basic Tables
• MIB codes for alcohol criticism (use significant
  to health and longevity), adverse driving record,
  abnormalities of ALT /or AST and abnormalities
  of GGTP
• Males only (too few female deaths for meaningful
  analysis)
• Mortality did not vary significantly with age,
  though numbers were not meaningful enough for age
  stratification for alcohol abuse
• Mortality rates were not significant higher in
  substandard vs. standard issue, 147 vs. 175

Titcomb et al. Alcohol Abuse and Liver Enzymes
(AALE) Results of an Intercompany Study of
Mortality. J Insur Med 200133277289
7
AALE - summary reportwith our without other
non-study impairment or test codes
Titcomb et al. Alcohol Abuse and Liver Enzymes
(AALE) Results of an Intercompany Study of
Mortality. J Insur Med 200133277289
8
Prevalence of fatty liver or hepatic steatosis

• Most common of all liver disorders cause of
  chronic liver disease
• MRI detection
• US adults 31
• Potential liver donors 33
• Ultrasound detection
• Koreans 23.4
• Japan 14 (19 years ago) 29 (2002 study)
• China 15 (2005)
• Italy (non-obese, nondrinkers) 16.4
• Liver donors with normal aminotransferase 20
• Prevalence increases with age
• Children 2.6
• At ages 40-59, 26
• Diabetes and obesity up to 75
• Diabetes with morbid obesity almost 100
• How can you rate for a condition that 20 of the
  population has?

9
Liver Enzymes in NAFLD

• Diagnosis NAFLD requires the exclusion of alcohol
  as cause 20 g/d for women and 30 g/d for men
• Normal values in up to 79 at one time or other
  may fluctuate
• Typically mild elevation of ALT AST, 2 to 4
  fold
• ALT usually greater than AST ratio
• AST gt ALT may indicates cirrhosis
• May be elevations of GGT, alkaline phosphatase
  ferritin
• The diagnosis of NASH requires a liver biopsy,
  but liver enzymes do not correlate with liver
  histology
• Full range of NAFLD may be seen with normal LFTs

10
Mortality from NAFLD

• 420 patients at the Mayo Clinic from 1980-2000
• Imaging alone (348), imaging and biopsy (65),
  biopsy alone (5) and cryptogenic cirrhosis with
  metabolic syndrome (2)
• Mean follow up 7.6 years (0.1-23.5)
• SMR all patients versus general population, 1.34
• SMR of those with at least 10 years follow up
  1.55
• Liver disease 3rd leading cause of death v. 13th
  in general Minnesota population
• A 26 prevalence in the population makes it a
  challenge to rate

The natural history of nonalcoholic fatty liver
disease a population-based cohort study. Adams
LA, Lymp JF, St Sauver J, Sanderson SO, Lindor
KD, Feldstein A, Angulo P. Gastroenterology.
2005 Jul129(1)113-21
11
Chronic hepatitis B and C infection

• Low prevalence in the UK
• HBV lt0.3
• HCV 1 (Europe)
• HBV common in immigrants from areas of high
  prevalence
• Also IV drug abusers, frequent sexual partners
  and male to male sex.
• HCV mostly from contaminated blood or medical
  equipment
• HBV HCV easily identified as cause of LFT
  elevation

12
General Household Survey, 2002Alcohol
consumption in the last week
One unit of alcohol is obtained from half a pint
of normal strength beer, lager or cider, a single
measure of spirits, one glass of wine, or one
small glass of port, sherry or other fortified
wine.
13
Average weekly alcohol consumption, by gross
weekly household income
Alcohol consumption General Household Survey, 2002
14
Risk assessment of alcohol use

• Inconsistent taxonomy terminology of alcohol
  usage
• DSM IV alcohol abuse and dependence are based on
  consequences and not usage
• Wide range of alcohol consumption and mortality
  patterns by gender, age, ethnicity, income level,
  occupation, smoking status and socio-economic
  status
• Quantification of alcohol usage in research
  studies and surveys relies almost entirely on
  self-reporting
• The beneficial effect on survival with some
  levels of alcohol usage

15
DSM-IV alcohol use disorders
16
Alcohol consumption v. diagnosis of alcohol abuse
or dependence

• The DSM-IV categorical approach was used to
  determine alcohol diagnoses for 435 highly
  educated young adult men
• Alcohol dependence group
• 4.5 mean drinks/day, max 15.5 drinks
• 16.5 mean drinking days/month, max 23.5 days
• Alcohol abuse group
• 3.4 mean drinks/day, max 12.5 drinks
• 15.6 mean drinking days/month, max 22.5 days
• No alcohol problem over 15 years
• 2.4 mean drinks/day, max 7.4 drinks
• 12.8 mean drinking days/month, max 14.4 days

Schuckit, M. A., T. L. Smith, et al. (2000). "The
5-year clinical course of high-functioning men
with DSM-IV alcohol abuse or dependence." Am J
Psychiatry 157(12) 2028-35.
17
Relative risk of death according to alcohol use
18
Relative risk of death from all causes and
various diagnoses
19
Identifying excess alcohol use

• High level of suspicion
• Social and personal history
• Questionnaires, screening instruments
• Medical history
• Physical exam
• Laboratory assessment

20
Social characteristics of alcohol misuse heavy,
binging

• Accident prone
• Driving criticism
• Cigarette smoker
• Frequent job changes or poor employment record
• Frequent mood shifts or violent behavior
• Jobs chosen to provide easy access to alcohol
• Marital instability

21
Alcohol screening instruments

• National Institute on Alcohol Abuse and
  Alcoholism (NIAAA) quality and frequency
  questions
• On average, how many days per week do you drink
  alcohol?
• On a typical day when you drink, how many drinks
  do you have?
• What is the maximum number of drinks you have had
  on any given occasion during the last month?
• CAGE, AUDIT, TWEAK, MAST, BMAST, RAPS4
• Single question screening (2005) When was the
  last time you had more than X drinks in 1 day?
  (X5 for men, 4 for women)

22
Medical history with excessive alcohol use

• Trauma recurrent injuries, accidents
• Cardiovascular hypertension, atrial
  fibrillation
• Gastrointestinal gastritis, bleeding
• Psychiatric depression, anxiety
• Neurologic single seizure, peripheral
  neuropathy
• Pulmonary frequent, recurrent infections, TB
• Falls in elderly - impaired judgment,
  hypotension, myopathy, neuropathy

23
Laboratory findings with excess alcohol use

• Liver function test elevations AST, ALT, GGT
• Positive alcohol markers CDT, HAA
• Complete blood count high MCV, low hemoglobin,
  platelets, and white blood count
• Lipid profile high HDL-C, high triglycerides
• High uric acid, and high serum ferritin
• Low BUN, low blood glucose, low albumin, low
  magnesium, and low phosphorus
• Blood alcohol concentration (BAC)
• Biologic markers less sensitive specific than
  questionnaires

24
LFTs and excess alcohol use

• LFT elevations in 14.5 insurance applicants and
  53 alcohol rehabilitation patients (Stout, 1998)
• AST and ALT both may be elevated with alcohol
  use. Sensitivity AST 60-80, ALT 50, specificity
  low
• AST/ALT ratio gt 2 suggests excess alcohol use
• GGT may be elevated with 4-5 drinks daily for gt4
  weeks, decreases after 4-5 weeks abstinence.
  Sensitivity 40-60, specificity 80. May not
  identify low levels of consumption or binge
  drinking. Poor screening test for alcohol.
• Combined elevation of AST and GGT is more
  suggestive of excess alcohol use

25
Alcohol markers

• Carbohydrate-deficient transferrin (CDT) -
  elevated by gt 4 drinks/day over 1-2 weeks,
  normalizes with cessation in 1-3 weeks
• CDT sensitivity 60-90, lower in women,
  specificity 90-100, for regular chronic moderate
  to heavy drinking, but not to binge drinking
• Hemoglobin-associated acetaldehyde (HAA) -
  combination of Hgb with acetaldehyde, elevated
  with gt 5 drinks daily, or in the presence of
  blood alcohol
• Mean corpuscular volume (MCV) gt 100 - index of
  red blood cell size, sensitivity low (30-60),
  higher in alcohol dependent women, specificity
  gt90, detects earlier drinking after a long
  period of abstinence (Mundle et al, 2000)

26
Abnormal laboratory values and CDT in insurance
applicants

• Positive CDT in 0.5 in 396 random applicants
• Elevated liver enzymes 11,477/123,513 applicants
• ALT 8.6, AST 2.9, GGT 8.3
• Reflexed to CDT
• Reflex CDT pos in 293/11,477 (2.6 ) with
  elevated LFTs
• GGT elevated in 70
• GGT and/or AST elevation in 97 of positive CDTs
• ALT alone elevated in only 2.7 of CDT positive
  applicants
• CDT positive in 2 with HDL gt 75 (n100), and in
  5 of positive tobacco applicants (n101)

Stout, J Insur Med, 1998
27
Reflex CDT in insurance applicants

• Reflex CDT positive in 1.96 of 33,082 applicants
  with positive BAC, elevated LFTs or HDL
• 31/132 applicants with positive blood alcohol
  (23.5)
• 493/29,091w/LFT elevations, mostly GGT(1.7)
• 127/3,979 with HDL gt 70 mg/dL (3.2)
• CDT correlates with smoking, BAC, GGT, AST HDL
• Recommend reflex testing with
• positive BAC, elevated GGT and/or AST with HDL gt
  50 mg/dL, BUN lt 7.6 mg/dL, HDL gt 70 mg/dL, and
  for cause

Stout, CRL News Views, 2000
28
Reflex CDT and HAA in insurance applicants

• 50,000 HAA and CDT, reflexed, increased positive
  rates in smokers, sensitivity 50-60, specificity
  95-98, postulate CDT and HAA act differently,
  occur independently
• Reflex alcohol markers in 9,935 samples, 81.4
  male, CDT tested in 85, (n8,460), HAA in 28,
  (n2,810), and both tested in 13.4, (n1,335)
• Positive CDT in 474/8,460 (5.6), positive HAA in
  129/2,810 (4.6), both positive in 22/1,335
  (1.6)
• Recommended testing GGT gt 65 U/L, AST/ALT gt 1,
  AST gt 33 U/L, HDL gt 80 mg/dL, best results with
  both CDT and HAA
• High AST/ALT ration more likely to give positive
  CDT

Daniel, OTR 1997
Lowden Wagner, OTR 2002
29
Maximum screen positive rates of CDT for various
laboratory tests

• Test
• GGT
• ALT
• AST
• AST GGT
• AST/ALT gt2
• HDL
• Cotinine
• Blood alcohol
• Triglycerides (low)
• BUN (low)

• Maximum positive CDT
• 4
• 2
• 7
• 10
• 8
• 8
• 6
• 23
• 4
• 20

63 (17-90) screen tests (Bio-Rad) confirmed
positive with capillary electrophoresis
Stout RL, CRL News Views 1998 2000
30
Problems with CDT

• Sensitivity 60-90 for heavy regular consumption
  of gt5 drinks per day but low for binge drinkers
• Outdated assays had high false positives and
  there is still no standardised assay
• False positives due to chronic active hepatitis,
  liver cancer, iron deficiency anaemia,
  haemochromatosis, congenital disorders of
  glycosylation, specimen aging and genetic
  variants of transferin (up to 3 of general
  population)
• In a low prevalence population, majority of
  results will be false positives for detecting
  alcohol abuse
• Positive CDT ? alcohol abuse or dependence
• Conclusion poor screening test but may have use
  as a confirmatory test where there is a high
  index of suspicion

31
Excessive detection of alcohol consumption (EDAC)
and alcohol markers

• 134 (7) of 1680 samples positive for EDAC, a
  profile of 15 tests including liver enzymes,
  lipids, proteins and blood sugar
• 22/134 (16) positive for CDT and/or HAA. Suggest
  EDAC may identify excess alcohol use in
  applicants with normal liver enzymes (9/22) but
  positive CDT and/or HAA (Bean et al, J Insur Med,
  2001)
• Similar study, EDAC-CDT vs. traditional
  screening but this did not include GGT (On the
  Risk, 2002)
• 5774 samples, 810 positive for traditional screen
  (with GGT), 22/810 (3) CDT 324 positive EDAC,
  44/324 (14) CDT (On the Risk, 2005)
• Is a proprietary product, so difficult to
  peer-review or reproduce

32
Alcohol markers in screening insurance applicants

• A good screening test should be reliable and
  accurate, provide good sensitivity, specificity
  and predictive value
• It should be clinically available and practical
  for widespread application
• Clinical CDT studies conducted in
  alcohol-dependent groups, sensitivity is not good
  for heavy or risky drinking
• Data limited on HAA and EDAC
• HAA alcohol 8 papers, 4 same author (last in
  Feb 2003), EDAC alcohol 7 in 8 years (all same
  authors) in PubMed
• No studies of relationship between positive
  markers and morbidity or mortality endpoints
• Positive CDT ? alcohol abuse or dependence
• Conclusion CDT is a poor screening test but may
  have use as a confirmatory test where there is a
  high index of suspicion

33
Alcohol risk assessment

• Risk varies and depends on many factors including
  pattern, quantity and frequency of alcohol use,
  individual susceptibility and characteristics,
  and external factors
• Quantity alone does not identify the mortality
  risk as there are trade offs in the various
  causes of death
• Effective underwriting, that utilizes multiple
  sources and factors, provides the best assessment
  of risk