CLIN. CHEM.37/5, 720-723
(1991)
Body Mass Index and Liver Enzyme Activity in Serum
Antonmo Salvaggio,’ Maurizio Periti, Lino Miano,2 Monica Tavanelli, and Daniela Marzorati2
The association between body mass index (BM1) and
serum liver enzyme activity [y-glutamyltransf erase (GGT),
alanine aminotransferase (ALT), and aspartate aminotransferase (AST)] was studied in 3167 subjects, 2373
men and 794 women. The subjects were managers and
employees, ages 18-64 years, who were examined during a program of preventive medicine. Analysis of covanance was used to compare the serum liver enzyme
activities (expressed as natural logarithms) of the subjects, who were subdivided according to BMI, while also
considering age, alcohol and cigarette consumption, and
physical activity. In men, the percentage increase in the
geometric mean of liver enzyme activity of the obese
subjects (BM1 >30 kg/rn2) compared with that of the
normal subjects (BM1 25 kg/m2) was 47.7% (P <0.001)
for GGT, 55.3% (P <0.001) for ALT, and 19.7% (P
<0.001) for AST; in women, the increase was 63.2% (P
<0.01) for GGT, 58.4% (P <0.001) for ALT, and 7.3% (P
>0.05) for AST. Thus, our observations demonstrate a
BMI and serum liver enzyme activity.
relation between
Keyphrases:
vanation, source of
ences
y-g!utamyltransferase
alanine
aspartate aminotransferase
Additional
sex-related differaminotransferase
Data from blood donors suggest that increased serum
activity of various liver enzymes may be related to overweight (1-3). Confirmation of such a relation has recently
been provided
by studies on populations of employees (4,
5); these studies, however, included only male subjects.
We analyzed data from employees
and managers of
both sexes examined
during a program of preventive
medicine. The relation between body mass index (BMI)
and serum
activity
of liver enzymes was analyzed,
considering
at the same time age, alcohol consumption,
cigarette
smoking, and physical activity.3
Materials
and Methods
The series consisted of 3167 subjects, managers
and
employees of both sexes, ages 18-64 years, examined
between June 1987 and September
1988 during a program of preventive medicine arranged between various
companies and Centro Diagnostico
Italiano.
A questionnaire was completed
by a doctor during a physical
examination.
Some laboratory tests were performed, including
determination of y-glutamyltransferase
(GGT) (6), alanine
aminotransferase
(ALT),
and aspartate
aminotransferase
by optimized methods based on the recommendations of the German Society for Clinical Chemistry (7), at 37 #{176}C.
Our conventional reference ranges
(U/L) were as follows: GGT, 43 for women and 59 for
men; ALT, 60; AST, 58. A logarithmic transformation (natural logarithm) of the serum enzyme activity
was adopted to take into account the positive skewness
of the enzyme activity distribution.
The BMI value, defined as weight(kg)/[height(m)12,
was used as a weight index; during weighing, subjects
wore only a standard bathrobe and no shoes or clothes.
The subjects were divided into three groups according to
their BMI: 25, 25-30, and >30 kglm2.
Alcohol consumption, expressed as fractions of a liter
for wine and beer and as number of glasses for spirits
(30-mL glasses for strong liquors, 50-mL glasses for
others), was transformed into grams of ethanol per day
(8). The participants were thus classified as nondrinkers
or as drinkers of 30, 31-60, or >60 g of ethanol per
day. The smokers were classified as smoking 20 or
>20 cigarettes per day. Also, the subjects were divided
by physical activity: none or irregular, regular but
noncompetitive, or intense (more than 4 h of intense
training weekly).
Medical and paramedical
personnel examining the
subjects did not know the aim of the study. Data on
serum enzyme activities in the subjects, divided accord(AST),
Table 1. CharacteristIcs of the Subjects Examined,
Expressed as Mean ±1 SD
Men
of subjects
Age, years
BMI, kg/rn2
GGT, LJ/L
Italy.
2Centro Diagnostico Italiano, Milan, Italy.
3Nonstandard
abbreviations: GGT, y.glutamyltransferase (EC
2.3.2.2); AST, aspartate aminotransferase
(EC 2.6.1.1); ALT, alanine amunotransferase
(EC 2.6.1.2); and BMI, body mass index.
Received October 16, 1990; accepted March 1, 1991.
720
CLINICAL CHEMISTRY, Vol. 37, No. 5,
1991
41.7
25.2
±
36.0
±
±
794
9.3
3.0
28.9
(5.48)8
InGGT
ALT, U/L
3.38 ± 0.59
(1.06)
35.5 ± 28.4
(5.93)
InALT
AST, U/L
Clunica Medics Generale, University of Milan, and Medicina
Riabilitativa,
Ospedale “L. Sacco,” Via Celio, 2, 20148 Milano,
Women
2373
No.
mAST
3.41
±
0.50
(1.18)
22.4 ± 11.9
(12.47)
3.04 ± 0.32
(2.49)
36.1
22.5
±
±
9.6
3.4
17.3
±
13.4
(3.10)
2.69 ± 0.50
(0.57)
21.0 ± 12.1
(5.14)
2.95
±
0.40
(0.83)
18.1 ± 10.1
(6.05)
2.84 ± 0.26
(1.76)
8ParenUs enclose the skewness coefficients. If data come from normal
distributions,the skewnesscoefficients have an expected value of zero, but if
high values of enzyme activities extend far above their means, these coefficents willbe positive.
Table 2. Geometric Mean Values for GGT, ALT, and AST (U/L) In Subjects Grouped by Age, BMI, Cigarette and
Alcohol Consumption, and Physical Activity
Men
No.
Age (years)
s20
21-30
31-40
41-50
51-60
GGT
ALT
AST
No.
GOT
ALT
AST
5
16.6
321
22.4
28.5
32.1
30.9
18.0
20.5
16
257
266
193
54
8
12.2
12.9
15.6
18.4
15.6
16.9
15.2
15.6
18.5
19.5
20.5
18.4
<0.0001
25.0
16.9
16.9
21.3
<0.0001
23.3
28.2
30.6
31.8
28.8
27.4
0.0004
19.1
<0.0001
17.1
<0.0001
25.3
33.1
27.1
33.1
39.6
<0.0001
42.1
<0.0001
14.2
17.3
23.8
18.4
21.1
30.0
16.9
17.6
18.4
<0.0001
<0.0001
0.1238
28.8
30.9
29.7
29.7
14.7
14.7
18.9
19.1
17.3
16.9
730
907
372
38
>60
31.2
P8
BMI (kg/rn2)
1207
25
25-30
1034
>30
132
P
Smoking(cigarettesper day)
0
1517
1-20
614
>20
242
28.8
33.4
0.0007
P
Alcohol(g/day)
0
1-30
30-60
542
1186
P
Physicalactivity
None or irregular
Regular
Intense
1072
1271
30
20.7
21.1
21.1
20.9
0.7205
19.9
21.5
23.8
21.1
20.5
20.1
0.0370
42.9
<0.0001
30.3
29.7
31.2
34.8
20.5
20.5
21.5
23.1
0.0008
<0.0001
30.9
27.9
31.5
29.4
23.3
27.9
20.9
20.7
21.3
P
<0.0001
0.0012
‘The P values reportedrelate to the explicatorycontributionoffered by the various
logarithm, ANOVA; Ftest).
ing to BMI, were examined
covariance
by analysis
of variance
and
(9).
Finally, subjects were divided according to whether
they had normal or above-normal
serum enzyme activity (reference values were the 95th percentile
for sub-
Table
647
121
26
<0.0001
0.1970
24.0
28.2
34.5
476
169
>60
Women
521
235
38
16.6
18.5
16.8
0.3338
0.8863
0.5629
437
14.0
18.7
16.9
321
15.5
18.9
23.1
19.3
18.7
31.8
0.0621
17.3
17.1
23.6
32
4
0.0002
479
15.0
19.1
16.9
14.4
18.9
17.6
10.9
12.9
13.9
0.6242
0.4955
0.6073
0.0784
classifications(analysisof varianceof data transformedaccordingto natural
314
1
jects with normal
weight and for nondrinkers).
The
proportions of subjects with above-normal
enzyme concentrations were related to BMI and alcohol consumption by a linear logistic model, fitted by the maximum
likelihood method (10).
3. RelatIon between BMI and Serum Activities of GGT, ALT, and AST (Expressed as the Natural Logarithm):
Geometric Meansa and Mean Differences ±SE
Men
GOT,
ALT,
Women
AST,
GOT,
mAST
25
25.5
25-30
31.5
38.0
>30
P8
P,,.,’
0.0701
U/L
(3.24)”
27.2
(3.30)
20.1
(3.00)
14.2
± 0.020 33.1
0.19 ± 0.02#{176}
21.7 0.08 ± 0.O1C 16.4
0.39 ± 0.05c 42.2 0.44 ± 0.050 23.9 0.18 ± 0.03c 23.1
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
<0.0001
0.21
ALT,
InGOT
U/L
mALT
AST,
U/L
(2.65)
18.4
(2.91)
17.1
0.15 ± 0.05#{176}
20.7 0.12 ± 004d
17.5
0.49 ± O.lO’ 29.1 0.46 ± 0.08#{176}18.3
<0.0001
<0.0001
(2.84)
0.02 ± 0.03
0.07 ± 0.05
<0.0001
<0.0001
‘Geometric means adjusted, by analysis of covailance, for age (a linear and a quadratic term), cigarette smoldng (binary variables: ‘20
cigarettes/day), alcohol consumption (binary variables: 30, 31-60, and >60 9/day), and physical activity (binary variables: regularand intensive).
b Parentheses enclose the absolute mean value, notadjusted,of the subjectswithBMI ‘25 kg/m2(the reference group for differences).
0.3470
0.0596
and >20
#{176}P
<0.001.
P <0.01.
#{149}
The P values listed relateto the explicatory contribution offeredby the classificationbased on BMI (I and to the explicatory contribution offeredbyBMI If
introduced as a quantitative variable, linearly relatedtothe logarithmof the enzyme activity (Ikj.
CLINICAL CHEMISTRY, Vol. 37, No. 5, 991
721
Resufts
Table
studied.
1 shows
some
characteristics
of the subjects
skewness coefficients (9)
data are reported.
For enzyme activities,
for original and log-transformed
Tables 2 and 3 report the activity of GGT, ALT, and
AST in the subjects, classified by various variables. Table
3 shows the data on enzyme activities in the various
categories of BME, adjusted for alcohol and cigarette consumption, physical activity, and age and then compared
by analysis of covariance. We performed multiple regressions in which, in addition to continuous variables such as
age (a linear and a quadratic
term), binary variables related to the various classification criteria were introduced.
Results of linear logistic fitting are shown in Figure 1.
In men, the proportion of subjects with above-normal
enzyme activity, adjusted for age, was positively related
to BMI for GOT (P <0.001), ALT (P <0.001), and AST (P
<0.001), but only to GOT (P <0.01) and ALT (P <0.001)
in women (AST, P >0.05). The number of women with
high alcohol consumption was small; thus, in Figure 1,
only data for nondrinking women and women who
drank <30 mL of ethanol per day were considered.
Discussion
Our observations demonstrate an important relation
between body weight (expressed as body mass index)
elevated
Men,
1,0
OCT
Women,
elevated
GGT
-,
AlohoI(g/d.y)
#{176}
0.8
0
0.6
0
0.4
0.4
0.2
0.0
15
20
25
30
35
40
15
20
25
SM!
Men, elevated
1.0
-
0.6
-
ALT
Women,
0
0
-
0.2
-
40
elevated
ALT
1.0
0
0.4
35
BMI
-
0.8
30
‘
0.0
0.8
0.6
0.4
0.2
0.0
15
20
25
SM!
Men,
elevated
30
35
40
SM
AST
Women,
elevated
AST
1.0
0
0.8
0
0.6
0
0.8
0.6
0
0.4
0.4
0.2
0.2
0.0
#{176}‘
0.0
15
20
25
30
35
40
_________
,‘
15
20
25
GM!
4-,
30
35
40
GM!
Fig. 1. Proportions of men and womenwith above-normal concentrationsof serum enzyme activity in groups that vary in BMI and
alcohol consumption
Symbols indicate the obseived proportions in three intervals of BMI (25,
25-30, and >30). Lines identify curves of the expected values that were
calculated and adjusted for age according to linear logistic models. Only
observedproportions for groups of at least25 subjects were reported;thus, in
women, onlydata on subjects with alcohol consumption ranging from 0 to 30
g/day were reported (aggregating dataof subjectswith BMI >30)
722
CLINICAL CHEMISTRY, Vol. 37, No. 5, 1991
and serum activity rates of liver enzymes. This relation
is still evident after age, physical activity, and cigarette
and alcohol consumption
are adjusted for.
Relations between serum enzyme activities and other
variables were similar to those observed in other studies.
Enzyme activities increase with age (11, 12). GOT and
ALT concentrations in men decrease with physical activity (5). Again only in men, GOT concentration
is higher in
smokers (5). Alcohol consumption
has been associated
with above-normal
activities of liver enzymes (1,2,13-19).
In our data, this relation was evident for GOT, ALT, and
AST in men, whereas in women, it reached statistical
significance only for GOT, probably because of the small
number of women with substantial alcohol consumption.
BMI was related to serum activities of GOT, ALT, and
AST in men, but only to GOT and ALT in women. The
mean enzymatic activity rose with an increase in BMI,
being greater in subjects with higher BMI (>30 kg/m2);
again, a significant linear relation between the log of
the activity and BMI was found for all the enzymes
considered except for AST in the women.
Conceivably, obesity alone might lead to increased
liver enzymes. Nevertheless, it is possible that underreporting of true alcohol consumption by the subjects
produced a spurious association caused by interactions
between alcohol abuse, obesity, and increased activity of
serum liver enzymes (and, possibly, fatty liver). Even so,
the relation between BMI and above-normal serum
enzyme activity was evident in moderate alcohol consumers as well as in heavy drinkers (Figure 1).
We cannot yet draw firm conclusions about the possible
causes of the relationships found in cross-sectional studies
such as this one. However, some pathological
findings may
help to explain our observationsor at least explain why
there should be a causal relation between BME and serum
activity of the liver enzymes. Fat vacuolation was observedin liver parenchymal cells from seven of nine blood
donors who underwent biopsy because of persistently
above-normal ALT concentrations in serum (20). Fatty
changes were observed in 85% of 61 obese subjects (21), in
accordance with other reports (22). Finally, significantly
increased AST concentrations occurred in subjects with
moderate or severe fatty changes (21) and in those with
marked hepatic steatosis (23).
On the basis of such observations,
Wejstal et al. (24)
recently suggested that ALTvalues should be corrected for
body weight, especially when ALT measurements are
used as a surrogate
test in screening for non-A, non-B
hepatitis (3,24). According to another proposal (25), AST
values also should be correctedfor body weight because in
various animal species,ranging from mice to cattle, the
expected enzymatic activity can be expressed as a function
of a power (0.85) of the weight (25).
In conclusion, although alcohol is certainly one of the
most serious causes of high serum activities of the
hepatic enzymes in subjects
examined during preventive medicine programs, marked
overweight is potentially another important factor in the increase of liver
enzymatic activities.
References
1. GilIon J, Hussey AJ, Howe SP, Beckett GJ, Prescott RJ. Posttransfusional non-A, non-B hepatitis: significance of raised ALT
and anti HBc in blood donors. Vox Sang 1988;54:148-53.
2. Friedman LS, Dienstag JL, Watkins E, et al. Evaluation of
blood donors with elevated serum alanine aminotran8ferase
levels.
Ann Intern Med 1987;107:137-44.
3. Briere RO. Effect of sex, race, and obesity on unit rejection rate.
Transfusion
1988;28:392-3, 1989;29:133.
4. Nomura F, Ohnishi K, Satomura
Y, et al. Liver function in
moderate obesity-study
in 534 moderately obese subjects among
4613 male company employees. mt J Obes 1986;1O:349-54.
5. Robinson D, Whitehead TP. Effect of body mass and other
factors on serum liver enzyme levels in men attending for well
population screening. Ann Clin Biochem 1989;26:393-400.
6. Persijn JP, van der Silk W. A new method for the determination
of gamina-glutamyltransferase
in serum. J Clin Chem Clin Bio-
chem 1976;14:421.
7. Recommendations
of the German Society for Clinical Chemistry: standardization
of methods for the estimation of enzyme
activity
in biological fluids. Z Klin Chem Kim Biochem
1970;8:659.
8. Mallor CS. Nomogram
ferent beverages. Br Med
9. Snedecor GW, Cochran
IA: Iowa State University
10. Cox
DR. Analysis
for calculating mass of alcohol in difJ 1970;iii:703.
WG. Statistical methods, 7th ed. Ames,
Press, 1980.
of binary
data.
London:
Chapman
and Hall,
1970.
11. Kahn RA, Johnson G, Aach RD, et al. The distribution of
serum alanine aminotransferase levels in a blood donor population. Am J Epidemiol 1982;115:929-40.
12. Siest G, Schiele F, Galteau MM, et al. Aspartate aminotransferase and alanine aminotransferase
activities in plasma: statistical distributions,
individual variations, and reference values. Clin
Chem 1975;21:1077-87.
13. Poleeky HF, House K, Hanson M. Factors affecting ALT levels
in blood donors [Abstract]. Transfusion
1981;21:606.
14. Sherman KE, Dodd RY. Alanine and aminotransferase levels
among volunteer blood donors: geographic variation and risk
factors. J Infect Dis 1982;145:383-6.
15. Nordby G, Foss OP. The alanine aminotransferase
activity as
screening test for blood donors.Vox Sang 1974;26:265-71.
16. Brandt KH, Meulendijk PN, Poulie NJ, et al. Data on the
determination
of SGOT and SGPT activity in donor blood for the
possible prevention of post-transfusion
hepatitis. Acta Med Scand
1965;177:321-5.
17. Rollason JG, Pincherle
G, Robinson D. Serum gamma glu-.
tamyl transpeptidase
in relation to alcohol consumption. Clin
Chim Acta 1972;39:75-80.
18. Rosalki SB. Gamma-glutamyl transpeptidase.
Adv Clin Chem
1975;17:53-107.
19. Kristensson H, Trell E, Eriks8on S, Hanik L, Hood B. Serum
.glutamyltransferase
in alcoholism. Lancet 1977;i:609.
20. Sampliner RE, Beluk D, Harrow EJ, Rivers S. The persistence
and significance of elevated alanine aminotransferase
levels in
blood donors. Transfusion
1985;25:102-4.
21. Anderson T, Chrisoffersen P, GluudC.The liver in consecutive
patients with morbid obesity: a clinical, morphological, and biochemical study. mt j Obes 1984:8:107-15.
22. Anderson T, Gluud C. Liver morphology in morbid obesity: a
literature study. hit J Obes 1984;8:97-106.
23. Manes JL, Taylor IIB, Starkloff GB. Relationship between
hepatic morphology and clinical and biochemical findings in morbidly obese patients. J Clin Pathol 1973;26:776-83.
24. Wejstal R, Hansson G, Lindholm A, Norkrans G. Persistent
alanine aminotransferase
elevation in healthy Swedish blood
donors mainly caused by obesity. Vox Sang 1988;55:152-6.
25. PappasNJ Jr, Quereshi AR. Liver aspartate aminotransferase
activity as a power function of body weight. Biochem Med Metab
Biol 1988;39:121-5.
CLIN. CHEM. 37/5, 723-728 (1991)
Factors Affecting Determinations of Manganese in Serum by Atomic Absorption
Spectrometry
Jean
N#{232}ve’
and Nathalie Leclercq
A method was developed for the routine determination of
manganese in serum from healthy human subjects by
graphite furnace atomic absorption spectrometry. We
controlled the experimental conditions rigorously, from
sampling to analysis, to minimize contamination and to
conserve diluted samples. We optimized the procedure
with two serum Reference Matenals, one of which had a
manganese concentration very close to what is thought to
be the physiological concentration in humans. The best
analytical performance was obtained by directly injecting
into the furnace serum diluted with an equal volume of a
solution containing Triton X-100 and sodium EDTA and
calibrating by the standard additions procedure or by a
Department
of Pharmaceutical
Organic Chemistry, Free University of Brussels, Institute of Pharmacy, Brussels, Belgium.
‘Address for correspondence:
Umversit#{233}
Libre de Bruxelles,
Institut de Pharmacie, Campus Plaine 205-5, B-1050 Bruxelles,
Belgium.
Received November 21, 1990; accepted March 5, 1991.
calibration graph constructed in a similar matrix. The
Zeeman backgroundcorrection produced better accuracy
and precision than did the classical deuterium correction.
Within-run CV for a manganese concentration of 12.7
nmol/L in serum was 7.9%, and between-run
precision
was 16.1%. The mean (SD) serum manganese concentration in 31 healthy adults was 10.8 (SD 3.0) nmol/L. Sex
and age of subjects did not affect concentrations.
AddItIonalKeyphrase.: reference values
trace elements
Determination of manganese in biological materials
because manganese
is involved in bone and
tissue formation and in carbohydrate and lipid metabolism. Manganese also plays a role in animal reproduction (1). More recently, its presence was proven to be
essential in human subjects (1). Although manganese
deficiency is rare in adults because the element is
widely available in the usual food supply, deficiency
is important
CLINICAL CHEMISTRY, Vol. 37, No. 5, 1991
723