Anda di halaman 1dari 6

Resting Metabolic Rate in Severely Obese

Diabetic and Nondiabetic Subjects


Kuo-Chin Huang,* Nic Kormas,* Katharine Steinbeck,* Georgina Loughnan,* and Ian D. Caterson*
Abstract
HUANG, KUO-CHIN, NIC KORMAS, KATHARINE
STEINBECK, GEORGINA LOUGHNAN, AND IAN D.
CATERSON. Resting metabolic rate in severely obese
diabetic and nondiabetic subjects. Obes Res. 2004;12:
840845.
Objectives: To compare the resting metabolic rate (RMR)
between diabetic and nondiabetic obese subjects and to
develop a predictive equation of RMR for these subjects.
Research Methods and Procedures: Obese adults (1088;
mean age 44.9 12.7 years) with BMI 35 kg/m
2
(mean BMI 46.4 8.4 kg/m
2
) were recruited. One
hundred forty-two subjects (61 men, 81 women) were di-
agnosed with type 2 diabetes (DM), giving the prevalence of
DM in this clinic population as 13.7%. RMR was measured
by indirect calorimetry, and several multivariate linear re-
gression models were performed using age, gender, weight,
height, BMI, fat mass, fat mass percentage, and fat-free
mass as independent variables.
Results: The severely obese patients with DM had consis-
tently higher RMR after adjustment for all other variables.
The best predictive equation for the severely obese was
RMR 71.767 2.337 age 257.293 gender
(women 0 and men 1) 9.996 weight (in kilo-
grams) 4.132 height (in centimeters) 145.959 DM
(nondiabetic 0 and diabetic 1). The age, weight, and
height-adjusted least square means of RMR between dia-
betic and nondiabetic groups were significantly different in
both genders.
Discussion: Severely obese patients with type 2 diabetes
had higher RMR than those without diabetes. The RMR of
severely obese subjects was best predicted by an equation
using age, gender, weight, height, and DM as variables.
Key words: resting metabolic rate, prediction, diabetes,
severely obese
Introduction
The prevalence of obesity is increasing in many coun-
tries, and obesity is a major global health problem (1,2). In
Australia, 19.3% of men and 22.2% of women are obese,
and the prevalence of obesity is 2.5 times higher than in
1980 (3). Obesity increases the risk of many medical dis-
orders and mortality. Among these disorders, type 2 diabe-
tes has a close relationship with severity of obesity (4,5).
The rationale for treating type 2 diabetic patients with
pharmacotherapy and diet control is to improve glycemia
and, thereby, reduce the risk of diabetic complications (6,7).
However, for obese subjects, weight reduction remains the
optimal method for the prevention and management of type
2 diabetes. In two prospectively randomized and controlled
studies, intensive lifestyle modification with mild weight
loss has been shown to reduce the incidence of diabetes by
58% in obese subjects with impaired glucose tolerance
compared with a similar control group (8,9). Orlistat, an oral
antiobesity agent, has been found to improve oral glucose
tolerance and diminish the rate of progression to the devel-
opment of impaired glucose tolerance and type 2 diabetes in
obese subjects (10). Furthermore, the use of antiobesity
agents in obese subjects with type 2 diabetes also demon-
strates that weight reduction is associated with improved
control of blood glucose and amelioration of cardiovascular
risk factors (11,12).
Resting metabolic rate (RMR)
1
is the main component of
daily energy expenditure, accounting for 60% to 70% of
total energy expenditure in most individuals, and a minor
Received for review July 7, 2003.
Accepted in final form March 16, 2004.
The costs of publication of this article were defrayed, in part, by the payment of page
charges. This article must, therefore, be hereby marked advertisement in accordance with
18 U.S.C. Section 1734 solely to indicate this fact.
*Metabolism and Obesity Services, Department of Endocrinology, Royal Prince Alfred
Hospital, Missenden Road, New South Wales, Australia; Departments of Family Medicine,
National Taiwan University Hospital, Taipei, Taiwan; and Human Nutrition Unit, School
of Molecular and Microbial Biosciences and Department of Medicine, University of
Sydney, New South Wales, Australia.
Address correspondence to Ian D. Caterson, Human Nutrition Unit, School of Molecular and
Microbial Biosciences, University of Sydney, NSW 2006, Australia.
E-mail: I.Caterson@mmb.usyd.edu.au
Copyright 2004 NAASO
1
Nonstandard abbreviations: RMR, resting metabolic rate; DM, type 2 diabetes; FM, fat
mass; FM%, fat mass percentage; FFM, fat-free mass.
840 OBESITY RESEARCH Vol. 12 No. 5 May 2004
change in RMR could lead to a significant energy imbalance
and huge change of body weight over a long period (1316).
Measuring RMR is time consuming and costly; therefore, a
number of recognized prediction equations to calculate
RMR have been developed in clinical practice. These can
provide the basis for prescribing an individualized energy
intake to attain a desired level of energy deficit. However,
RMR seems to be inaccurate in obese subjects, with over-
estimation of RMR by these prediction equations (1719).
In addition, obese subjects with diabetes are generally ex-
cluded or not specified in these equations. In the present
study, we aimed to compare the difference in measured
RMR between diabetic and nondiabetic subjects in a large
population of severely obese patients. The second aim was
to generate a predictive equation of RMR in these subjects
using variables inclusive of diabetic status.
Research Methods and Procedures
Subjects and Characterization
The study was a retrospective analysis of adult obese
patients with BMI 35 kg/m
2
who attended the Metabo-
lism and Obesity Service in Royal Prince Alfred Hospital
from January 1999 to May 2003. Approval for analyzing the
demographic and biological data in these subjects was ob-
tained from the Central Sydney Area Health Service Ethics
Review Committee. The subjects were all referred from
primary care physicians for weight management. At the
initial visit, a series of assessments including anthropomet-
ric measurements, pathology tests, and RMR were per-
formed for every patient. Trained staff measured height,
waist circumference (measured to the nearest 0.1 cm), and
weight (measured to the nearest 0.1 kg). Waist circumfer-
ence was taken midway between the inferior margin of the
last rib and the crest of the ilium in a horizontal plane. BMI
was calculated as weight (kilograms) divided by height
squared (meters squared). A venous blood sample was taken
after a 12-hour fast for measuring plasma glucose by an
automated spectrophotometer (Roche/Hitachi 917 Auto-
mated Chemistry analyzer; Roche Diagnostics, Indianapo-
lis, IN), and type 2 diabetes was defined as fasting plasma
glucose 7.0 mM. Body composition was measured using
the bioelectrical impedance analysis method with a four-
terminal bioimpedance analyzer (Bodystat 1500; Bodystat
Ltd., Tampa, FL). RMR was measured for 40 minutes under
standardized conditions, using ventilated hood indirect cal-
orimetry (Deltatrac, Datex Division Instrumentarium Corp.,
Helsinki, Finland), and calibrated using a precision gas
mixture before each measurement. The mean value of the
data used in the calculation of RMR was obtained from a
measurement taken within the mid-20-minute period. RMR
was derived without taking urinary nitrogen excretion into
account due to its minimal effect on RMR (20). Eighteen
participants had RMR performed on 2 consecutive days, and
the test-retest coefficient of variation was 1.0 0.8%.
Statistical Analysis
Data were presented as mean and SD. Statistical analyses
including two-sample Students t test, correlation analyses,
and multivariate linear regression analyses were performed
by the SPSS/PC statistical program (version 10.0 for Win-
dows; SPSS, Inc., Chicago, IL). Partial correlation of the
clinical characteristics and RMR with adjustment for age
and gender was performed in the diabetic and nondiabetic
obese. Several multivariate linear regression models were
performed using plasma RMR as the dependent variable and
using age, gender, weight, height, BMI, fat mass (FM), fat
mass percentage (FM%), and fat-free mass (FFM) as inde-
pendent variables. The least square (LS) means of RMR
between the diabetic and nondiabetic obese with the adjust-
ment for age, weight, and height among these subjects in
each gender were tested by ANOVA.
Results
The baseline characteristics of the subjects are shown in
Table 1. The mean age and mean plasma glucose of the
diabetic obese subjects were significantly greater than in the
nondiabetic obese (p 0.001) for both genders. There were
no statistically significant differences in weight, height, FM,
FM%, lean mass, and predictive RMR between diabetic and
nondiabetic groups. RMR was higher in the diabetic obese
than the nondiabetic obese in women (p 0.005) and men
(p 0.013). The predicted RMR using the Harris-Benedict
equation (21) was found to be overestimated in the nondi-
abetic and diabetic men and underestimated in the diabetic
women. The percentage difference between measured RMR
and predicted RMR was higher in the nondiabetic obese
than in the diabetic obese (p 0.001). After adjustment for
age and gender, measured RMR was positively correlated
with weight, height, BMI, WC, FM, FM%, FFM, plasma
glucose levels, and predicted RMR in the diabetic and
nondiabetic groups (Table 2). The percentage difference
from measured RMR was negatively correlated with RMR
(p 0.001).
Using multivariate linear regression analysis, the diabetic
severely obese had higher RMR than the nondiabetic se-
verely obese after adjustment for other variables in different
models (Table 3). The coefficient of the FM variable was
found to be negative when the FM variable was incorpo-
rated in model 1 in Table 3 (data not shown). Among these
models, the best predictive equation of RMR in these sub-
jects (R
2
0.750) was RMR 71.767 2.337 age
257.293 gender (women 0 and male 1) 9.996
weight (in kilograms) 4.132 height (in centimeters)
145.959 DM (nondiabetic 0 and diabetic 1). To
examine whether the equation was accurate or not in an
Resting Metabolic Rate in Severe Obese, Huang et al.
OBESITY RESEARCH Vol. 12 No. 5 May 2004 841
independent population, a predictive equation of RMR us-
ing age, gender, weight, height, and DM as predictors was
derived from 710 subjects (two-thirds of our patients), and
then the predicted RMR from this equation and the mea-
sured RMR in the other 328 subjects (one-third of our
patients) were compared by paired Students t test. We
found that there was no statistically significant difference
between the predicted and measured RMR.
The age, weight, and height-adjusted LS means of the
RMR by gender are shown in Figure 1. The LS means of
RMR between diabetic (2546.8 40.9 kcal/d in men and
2007.6 25.6 kcal/d in women) and nondiabetic groups
(2423.9 20.3 kcal/d in men and 1849.3 8.6 kcal/d in
women) were significantly different in both genders (p
0.01 in men and p 0.001 in women, respectively).
Discussion
In the present study, we demonstrated that severely obese
subjects with type 2 diabetes had a higher measured RMR
than obese people without DM after adjustment for other
variables. This finding that RMR was greater in obese type
2 diabetic subjects compared with nondiabetic obese sub-
jects is similar to previous studies (2224). The etiology of
a greater RMR in diabetics has been suggested to be the
result of abnormal protein metabolism (24) and high insulin
resistance (25). However, the exact mechanism still remains
unclear. We also propose, for the first time to our knowl-
edge, that the best predictive equation for RMR in the
severely obese subjects is RMR 71.767 2.337 age
Table 1. Comparison of general characteristics categorized by gender between the diabetic and nondiabetic obese
patients
Variables
Men (n 279) Women (n 759)
Diabetic
(n 61)
Nondiabetic
(n 218) p values
Diabetic
(n 81)
Nondiabetic
(n 678) p values
Age (years) 51.9 11.7 43.9 12.9 0.001 51.6 11.9 43.7 12.4 0.001
BW* (kg) 148.6 26.9 146.4 32.3 NS* 123.1 25.7 121.2 24.1 NS
BH* (cm) 175.7 6.0 176.1 8.6 NS 161.0 8.0 162.3 7.7 NS
BMI* (kg/m
2
) 48.0 7.9 47.1 9.2 NS 47.4 8.8 46.0 8.2 NS
WC* (cm) 142.0 13.6 136.9 21.8 0.035 121.4 26.4 116.6 18.4 NS
FM%* (%) 44.1 4.9 42.5 7.4 NS 53.2 6.7 52.0 6.8 NS
FM* (kg) 66.0 18.1 62.5 21.9 NS 67.0 20.4 63.8 18.8 NS
FFM* (kg) 81.5 9.6 81.0 12.0 NS 57.2 10.1 57.3 9.5 NS
Glucose (mmol/l) 9.53 3.13 5.34 0.72 0.001 10.02 3.50 5.17 0.65 0.001
RMRm* (Kcal/day) 2538.0 439.7 2426.2 424.9 NS 2006.8 436.5 1849.4 326.7 0.005
RMRp* (Kcal/day) 2593.8 408.0 2664.7 524.0 NS 1907.4 292.3 1916.3 259.8 NS
RMRd* (%) 3.3 14.7 10.2 12.4 0.002 3.1 11.6 4.9 11.4 0.001
Data are means SD; statistics were computed by Students t tests.
* BW, body weight; BH, body height; WC, waist circumference; RMRm, measured RMR; RMRp, predictive RMR by Harris-Benedict
equation; RMRd, {(RMRp RMRm)/RMRm} 100%; NS, not significant.
Table 2. Correlation coefficients of RMRm* and dif-
ferent variables* among diabetic and non-diabetic
obese patients after adjustment for age and gender
Variables Diabetic (n 142) Nondiabetic (n 896)
BW 0.694 0.759
BH 0.330 0.364
BMI 0.590 0.644
WC 0.427 0.617
FM% 0.462 0.361
FM 0.666 0.663
FFM 0.475 0.592
Glucose 0.052 0.150
RMRp 0.623 0.734
RMRd 0.701 0.545
* Abbreviations are as defined in Table 1.
p 0.05.
p 0.01.
p 0.001. Statistics were tested by Pearsons correlations.
Resting Metabolic Rate in Severe Obese, Huang et al.
842 OBESITY RESEARCH Vol. 12 No. 5 May 2004
257.293 gender (women 0 and male 1) 9.996
weight (in kilograms) 4.132 height (in centimeters)
145.959 DM (nondiabetic 0 and diabetic 1).
A measurement of RMR in the obese is important but is
not generally available in clinical practice, and there is a
reliance on predictive equations. A predictive equation of
RMR for the severely obese is important to provide the
basis for an individualized treatment plan for weight loss.
Data from previous studies of several predictive equations
in different populations show substantial differences. RMR
seems to be inaccurate and mostly overestimated by such
prediction equations in obese subjects (1619). In addition,
RMR is generally estimated by predictive formulae based
on weight, height, age, and gender. For instance, the Harris-
Benedict equation is the most common method for calcu-
lating RMR (21). In our results, the predicted RMR by
Harris-Benedict equation overestimated RMR in the both
nondiabetic obese and diabetic men and underestimated it in
diabetic women (Table 1). In multiple linear regression
models with four independent variables of age, gender,
Table 3. Multivariate linear regression models showing regression coefficients (SE) with measured RMR* as
dependent variable, and listed variables* as independent variables
Independent
variables
Model 1
(R
2
0.750)
Model 2
(R
2
0.737)
Model 3
(R
2
0.723)
Model 4
(R
2
0.657)
Model 5
(R
2
0.647)
Model 6
(R
2
0.588)
Constant 71.767 60.655 521.995 1384.640 886.220 788.810
(183.803) (187.416) (68.231) (47.999) (63.293) (82.774)
Age 2.337 1.440 1.515 6.184 6078 2.870
(0.657) (0.658) (0.696) (0.744) (0.742) (0.869)
Gender 257.293 273.821 220.863 541.048 550.221 37.156
(23.415) (23.415) (29.301) (21.169) (20.903) (34.325)
BW (kg) 9.996 10.158
(0.323) (0.330)
BH (cm) 4.132 3.933
(1.139) (1.161)
DM 145.959 171.074 149.951 190.961
(23.295) (28.184) (27.627) (30.862)
FFM (kg) 14.118 20.545
(0.916) (1.046)
FM (kg) 9.367 11.537
(0.443) (0.462)
BMI (kg/m
2
) 26.807
(1.071)
* Abbreviations are as defined in Table 1. Gender: female, 0; male, 1.
DM, diabetic 1, nondiabetic 0.
p 0.05.
p 0.01.
p 0.001.
Figure 1: The measured RMR values (LS mean SE) with the
adjustment for the age, body weight, and height among diabetic
(2546.8 40.9 kcal/d in men and 2007.6 25.6 kcal/d in women)
and nondiabetic (2423.9 20.3 kcal/d in men and 1849.3 8.6
kcal/d in women) groups categorized by gender in an ANOVA
model. p 0.05 (*), p 0.01 (**), p 0.001 (#).
Resting Metabolic Rate in Severe Obese, Huang et al.
OBESITY RESEARCH Vol. 12 No. 5 May 2004 843
weight, and height, RMR could be estimated from model 2
in Table 3. In model 1, addition of the DM variable im-
proved the prediction accuracy, and the equation with all
five variables was the best model for estimating RMR in
severely obese subjects.
FFM, reflecting the amount of metabolically active
tissue, has been found to be closely correlated with RMR
(26). In the present study, body weight had a stronger
correlation with RMR than FFM and was a better deter-
minant in the predictive model for RMR than was FFM,
a finding similar to that of Karhunen et al. (27) and
Mifflin et al. (28) in less obese subjects. The positive
correlation between FM and RMR in Tables 2 and 3
could be explained by close relationship between weight
and FM. In fact, the coefficient of the FM variable was
found to be negative when the FM variable was put into
model 1 in Table 3 (data not shown).
It is possible that there are other factors that may con-
tribute to predicting RMR in severely obese subjects. Insu-
lin and sulfonylurea therapies in diabetic patients, for ex-
ample, have been noted to be associated with reduction of
RMR (23,29,30). Improvement in the prediction of RMR
among obese subjects with type 2 diabetes has been dem-
onstrated by factoring glycemia, not hemoglobin A
1c
, into
the equation (31). Furthermore, RMR has been found to be
lower in African Americans than whites (32). Whether the
addition of these variables can improve the accuracy of
predicting RMR in the severely obese deserves further
study.
In conclusion, this study demonstrated that severely
obese subjects with type 2 diabetes had a higher RMR
than the obese without DM after adjustment for other
variables. The RMR of severely obese subjects could be
predicted by age, gender, weight, height, and DM vari-
ables. It is possible that the addition of other variables in
this predictive equation might further improve its accu-
racy in the severely obese and allow a better individual
prescription of diet and exercise in a weight management
program.
Acknowledgment
There was no outside funding/support for this study. We
thank Julie Hetherington in the Metabolism and Endocrine
Unit for technical assistance.
References
1. World Health Organization. Obesity: Preventing and Man-
aging the Global Epidemic: Report of a WHO Consultation on
Obesity. Geneva, Switzerland: World Health Organization;
1998.
2. Flegal KM, Carroll MD, Ogden CL, Johnson CL. Preva-
lence and trends in obesity among US adults, 1999-2000.
JAMA. 2002;288:17237.
3. Cameron AJ, Welborn TA, Zimmet PZ, et al. Overweight
and obesity in Australia: the 1999-2000 Australian Diabetes,
Obesity and Lifestyle Study (AusDiab). Med J Aust. 2003;
178:42732.
4. Kopelman PG. Obesity as a medical problem. Nature. 2000;
404:63543.
5. Huang KC, Lin WY, Lee LT, et al. Four anthropometric
indices and cardiovascular risk factors in Taiwan. Int J Obes
Relat Metab Disord. 2002;26:10608.
6. Turner RC, Cull CA, Frighi V, Holman RR. Glycemic
control with diet, sulfonylurea, metformin, or insulin in pa-
tients with type 2 diabetes mellitus. JAMA. 1999;281:2005
12.
7. Stratton IM, Adler AI, Neil HA, et al. Association of gly-
caemia with macrovascular and microvascular complications
of type 2 diabetes (UKPDS 35): prospective observational
study. BMJ. 2000;321:40512.
8. Tuomilehto J, Lindstrom J, Eriksson JG, et al. Prevention
of type 2 diabetes mellitus by changes in lifestyle among
subjects with impaired glucose tolerance. N Engl J Med.
2001;344:134350.
9. Diabetes Prevention Program Research Group. Reduction
in the incidence of type 2 diabetes with lifestyle intervention
or metformin. N Engl J Med. 2002;346:393403.
10. Heymsfield SB, Segal KR, Hauptman J, et al. Effects of
weight loss with orlistat on glucose tolerance and progression
to type 2 diabetes in obese adults. Arch Intern Med. 2000;160:
13216.
11. Hollander PA, Elbein SC, Hirsch IB, et al. Role of orlistat
in the treatment of obese patients with type 2 diabetes: a
1-year randomized double-blind study. Diabetes Care. 1998;
21:128894.
12. Gokcel A, Karakose H, Ertorer EM, Tanaci N, Tutuncu
NB, Guvener N. Effects of sibutramine in obese female
subjects with type 2 diabetes and poor blood glucose control.
Diabetes Care. 2001;24:195760.
13. Ravussin E, Lillioja S, Anderson TE, Christin L, Bogardus
C. Determinants of 24-hour energy expenditure in man: meth-
ods and results using a respiratory chamber. J Clin Invest.
1986;78:156878.
14. Ravussin E, Lillioja S, Knowler WC, et al. Reduced rate of
energy expenditure as a risk factor for body-weight gain.
N Engl J Med. 1988;318:46772.
15. Schulz LO, Schoeller DA. A compilation of total daily en-
ergy expenditures and body weights in healthy adults. Am J
Clin Nutr. 1994;60:67681.
16. Foster GD, McGuckin BG. Estimating resting energy expen-
diture in obesity. Obes Res. 2001;9:36772s.
17. Heshka S, Feld K, Yang MU, Allison DB, Heymsfield SB.
Resting energy expenditure in the obese: a cross-validation
and comparison of prediction equations. J Am Diet Assoc.
1993;93:10316.
18. Frankenfield DC, Muth ER, Rowe WA. The Harris-Bene-
dict studies of human basal metabolism: history and limita-
tions. J Am Diet Assoc. 1998;98:43945.
19. Horgan GW, Stubbs J. Predicting basal metabolic rate in the
obese is difficult. Eur J Clin Nutr. 2003;57:33540.
20. Ferrannini E. The theoretical bases of indirect calorimetry: a
review. Metabolism. 1988;37:287301.
Resting Metabolic Rate in Severe Obese, Huang et al.
844 OBESITY RESEARCH Vol. 12 No. 5 May 2004
21. Harris J, Benedict F. A Biometric Study of Basal Metabolism in
Man. Washington, DC: Carnegie Institute of Washington; 1919.
22. Nair KS, Webster J, Garrow JS. Effect of impaired glucose
tolerance and type 2 diabetes on resting metabolic rate and
thermic response to a glucose meal in obese women. Metab-
olism. 1986;35:6404.
23. Bogardus C, Taskinen MR, Zawadzki J, Lillioja S, Mott D,
Howard BV. Increased resting metabolic rates in obese sub-
jects with non-insulin-dependent diabetes mellitus and the
effect of sulfonylurea therapy. Diabetes. 1986;35:15.
24. Gougeon R, Pencharz PB, Marliss EB. Effect of NIDDM on
the kinetics of whole-body protein metabolism. Diabetes.
1994;43:31828.
25. Perseghin G, Mazzaferro V, Benedini S, et al. Resting
energy expenditure in diabetic and nondiabetic patients with
liver cirrhosis: relation with insulin sensitivity and effect of
liver transplantation and immunosuppressive therapy. Am J
Clin Nutr. 2002;76:5418.
26. Nelson KM, Weinsier RL, Long CL, Schutz Y. Prediction
of resting energy expenditure from fat-free mass and fat mass.
Am J Clin Nutr. 1992;56:84856.
27. Karhunen L, Franssila-Kallunki A, Rissanen A, Kervinen
K, Kesaniemi YA, Uusitupa M. Determinants of resting
energy expenditure in obese non-diabetic Caucasian women.
Int J Obes Relat Metab Disord. 1997;21:197202.
28. Mifflin MD, St Jeor ST, Hill LA, Scott BJ, Daugherty SA,
Koh YO. A new predictive equation for resting energy ex-
penditure in healthy individuals. Am J Clin Nutr. 1990;51:
2417.
29. Nair KS, Halliday D, Garrow JS. Increased energy expen-
diture in poorly controlled Type 1 (insulin-dependent) diabetic
patients. Diabetologia. 1984;27:136.
30. Welle S, Nair KS, Lockwood D. Effect of a sulfonylurea and
insulin on energy expenditure in type II diabetes mellitus.
J Clin Endocrinol Metab. 1988;66:5937.
31. Gougeon R, Lamarche M, Yale JF, Venuta T. The predic-
tion of resting energy expenditure in type 2 diabetes mellitus
is improved by factoring for glycemia. Int J Obes Relat Metab
Disord. 2002;26:154752.
32. Gannon B, DiPietro L, Poehlman ET. Do African Ameri-
cans have lower energy expenditure than Caucasians? Int J
Obes Relat Metab Disord. 2000;21:413.
Resting Metabolic Rate in Severe Obese, Huang et al.
OBESITY RESEARCH Vol. 12 No. 5 May 2004 845

Anda mungkin juga menyukai