We compared the post mortem diagnostic value of gamma-glutamyltransfer
ase (GGT), carbohydrate-deficient transferrin (CDT), alcoholic liver d
isease (ALD), blood alcohol concentration (BAG), the presence of multi
ple bruises and poor hygiene of the feet as markers of chronic alcohol
ism (heavy continuous drinking) in 32 alcoholics with 32 age-sex match
ed controls drawn from a forensic autopsy population. Alcoholics and c
ontrols were selected on the basis of positive and negative medical hi
story but controls were excluded if BAC exceeded 70 mg%. Femoral venou
s blood, urine and vitreous humour alcohol concentrations were determi
ned by headspace gas chromatography (GC). BAC was positive in 19 alcoh
olics (mean 234 mg%, range 2-570 mg%) and six controls (mean 32 mg%, r
ange 2-52 mg%). Serum GGT was measured by a kinetic photometric method
, and CDT by both isoelectric focusing/laser densitometry and by a com
mercial radioimmunoassay kit (CDTeCt(TM)). Features of alcoholic liver
disease were graded histologically using two weighted scoring systems
. Eleven alcoholics tested positive for GGT, CDTq and ALD, nine were p
ositive for two tests, five for one test and three were negative for a
ll three tests. No controls were positive for all three tests but six
were positive for two tests and nine for only one test; 17 were negati
ve for all three tests. Using the normal clinical cut-off values GGT,
CDTq and CDTect(TM) gave poor specificity which was improved at modera
te cost to sensitivity by raising cut off values for each test. Compar
ison of receiver operating characteristic curves, likelihood ratios an
d post-test odds showed CDT to be the best individual test, followed b
y ALD and GGT. Quantitation of CDT by IEF/laser densitometry performed
slightly better than MAEC/RIA by CDTect(TM). CDT shows considerable p
romise as a post mortem marker of chronic alcoholism.