Normal serum aminotransferase concentration and risk of mortality from liver diseases: prospective cohort study

Editor – Kim and co-workers (1) have investigated and suggested
reference limits for aminotransferase concentrations and mortality from
liver disease. However, for health derived reference limits to be used
there must be a high degree of analytical and population group
transferability (2). It would therefore be essential to report on the
analytical methods employed when commenting on specific results as per the
STARD (Standards for Reporting of Diagnostic Accuracy) initiative which
was co-published by this journal (3). Due to the lack of assay
standardization (2, 4) and therefore non-comparability, comparing absolute
values (e.g. 40 IU/L) may be misleading and confusing. For example, three
different ALT assays harmonised to the International Federation of
Clinical Chemistry method and routinely employed by accredited
laboratories have significantly different reference intervals: 21–72 IU/L
for the Vitros 250® (Ortho Diagnostics, www.orthoclinical.com), <_45 iu="iu" l="l" for="for" the="the" au640="au640" olympus="olympus" diagnostics="diagnostics" www.olympusamerica.com="www.olympusamerica.com" dsg_section="dsg_section" dsg_chemistrysys.asp="dsg_chemistrysys.asp" and="and" _3065="_3065" dimension="dimension" rxl="rxl" dade="dade" behring.="behring." furthermore="furthermore" reference="reference" intervals="intervals" differ="differ" two="two" different="different" enzymes="enzymes" ast="ast" alt="alt" _2="_2" therefore="therefore" use="use" of="of" a="a" common="common" single="single" upper="upper" limit="limit" _1="_1" is="is" discouraged.="discouraged." p="p"/> The biological variation for ALT is approximately 25%. Taking the ALT
concentration at the suggested (1) upper reference limit of 40 IU/L, the
95% confidence interval of a single estimate would be 20–60 IU/L. This
would render the utilisation of a population cut-off for ALT as
impractical for routine clinical practice when investigating individuals
with single serum estimations. It would also be important to report on,
and control for, other factors such as body mass index (lower range in the
Far East), viral hepatitis (higher incidence in the Far East), age,
gender, recent exercise, muscle injury and haemolysis as such issues may
significantly influence aminotransferase values (4).

It needs to be emphasized that if the upper normal limit is taken at
40 IU/L, 97.5% of values would be equal to or below this, but this would
only be equal to two SD above the mean if the distribution was Gaussian.
The distribution of aminotransferases show positive skewness and would
require transformation before one can state that the upper normal limit is
two SD above the mean level (5).

References:

1. Kim HC, Nam CM, Jee SH, Han KH, Oh DY, Suh I. Normal serum
aminotransferase concentration and risk of mortality from liver disease:
prospective cohort study. BMJ 2004; 328:983-86

2. Burtis CA, Ashwood ER. Tietz textbook of clinical chemistry. 3rd
edition. WB Saunders: Philadelphia; 1999.

3. Patrick M. Bossuyt, Johannes B. Reitsma, David E. Bruns, Constantine A.
Gatsonis, Paul P. Glasziou, Les M. Irwig, et al. Towards complete and
accurate reporting of studies of diagnostic accuracy: the STARD
initiative. BMJ 2003; 326:41-44.

4. National Academy of Clinical Biochemistry. D. Robert Dufour, John A.
Lott, Frederick S. Nolte, David R. Gretch, Raymond S. Koff, and Leonard B.
Seeff
Diagnosis and Monitoring of Hepatic Injury. I. Performance Characteristics
of Laboratory Tests. Clin. Chem. 2000; 46:2027-2049.

5. Roderick P. Commentary: Liver function tests: defining what’s normal.
BMJ 2004; 328:987.

Competing interests:
None declared