Impaired

Glucose D. Jallut,

Tolerance

and Diabetes in Obesity: of Glucose Metabolism

A. Golay, R. Munger,

P. Frascarolo,

A 6-Year

Follow-up

Study

Y. Schutz, E. Jbquier, and J.P. Felber

To investigate the time course of glucose metabolism in obesity 33 patients (21 to 69 years old; body mass index [BMI], 25.7 to 53.3 kg/m’) with different degrees of glucose intolerance or diabetes who had been studied initially and 6 years later were submitted to the same 100-g oral glucose tolerance test (OGTTI with indirect calorimetry. From a group of 13 obese subjects with normal glucose tolerance (NGT), four developed impaired glucose tolerance UGT); from a group of nine patients with IGT, three developed non-insulin-dependent diabetes mellitus (NIDDM); five of six obese NIDDM subjects with high insulin response developed NIDDM with low insulin response. Five patients had diabetes with hypoinsulinemia initially. As previously seen in a cross-sectional study, the 3-hour glucose storage measured by continuous indirect calorimetry remained unaltered in patients with IGT, whereas it decreased in NIDDM patients. A further decrease in glucose storage was observed with the lowering of the insulin response in the previously hyperinsulinemic diabetics. These results confirm cross-sectional studies that suggest successive phases in the evolution of obesity to diabetes: A, NGT; B, IGT (the hyperglycemia normalizing the glucose storage over 3 hours); C, diabetes with increased insulin response, where hyperglycemia does not correct the resistance to glucose storage anymore: and D. diabetes with low insulin response, with a low glucose storage and an elevated fasting and postload glycemia. 8 1990 by W.B. Saunders Company.

0

BESITY IS CONSIDERED as a risk factor for noninsulin-dependent diabetes mellitus (NIDDM).’ Epidemiological studies show a high prevalence of NIDDM among obese subjects. 2.3It is known that diabetes generally appears many years after the beginning of obesity,4 but longitudinal studies are scarce. Sasaki et al’ provided evidence for the role of obesity in the progression from normal glucose tolerance (NGT) to impaired glucose tolerance (IGT) and overt diabetes in a 7-year follow-up study. In a S- to 12-year follow-up study, Kadovaki et al6 reported that multivariate analysis indicates that obesity and IGT are independent risk factors for the development of diabetes. As reported previously,’ glucose and lipid metabolisms were studied by means of continuous indirect calorimetry in the course of a 100-g oral glucose tolerance test (OGTT) in a group of 82 obese subjects with different degrees of IGT. The duration of obesity and the ages were greater in the subgroup presenting NIDDM (32 + 2 and 51 * 1 year, respectively) than in the subgroup with IGT (17 k 2 and 32 + 2 years) and that with NGT (8 k 1 and 28 f 2 years), suggesting an evolution from obesity with NGT to obesity with IGT and later with NIDDM. When the subjects were stratified into nine S-year classes of duration of obesity (from 0 to 45 years), the prevalence of both IGT and overt diabetes increased with obesity duration.’ The occurrence of diabetes was associated with a drastic reduction in glucose storage, estimated as nonoxidative glucose disposal. This reduction in

From the Division of Endocrinology and Clinical Biochemistry, Department of Medicine, University Hospital, Lausanne. Switzerland; and the Institute of Physiology, University of Lausanne. Lausanne. Switzerland. Supported in part by the Raymond Berger Foundation for Diabetes Research in Lausanne and the Swiss National Science Foundation, Grant No. 3.707-0.87. Address reprint requests to Professor J.P. Felber. Institute of Physiology. Rue du Bugnon 7. CH-1005 Lausanne/Switzerland. o I990 by W.B. Saunders Company. 0026-0495/90/3910-0012$3.00/0

1068

glucose storage occurred simultaneously with the appearance of fasting hyperglycemia. The decrease in the insulin response to glucose and the increase in post-OGTT plasma glucose levels were associated with the duration of obesity. A longitudinal study is needed to confirm these crosssectional data; therefore, a 6-year follow-up was performed in 33 subjects who had been previously investigated using the same protocol of a 100-g OGTT associated with continuous indirect calorimetry. The aim was to assess the evolution of the different states of abnormal glucose metabolism in obesity in relation to the defects in glucose utilization. MATERIALS AND METHODS

From the 82 obese subjects who had taken part in the previous study,‘-’ 33 could be reinvestigated after an interval of 6.3 k 0.2 years. The study was not intended initially to be a longitudinal one, so that the drop-out rate was important. The subjects body mass index (BMI) was initially above 25 kg/m*. The protocol was submitted to and accepted by the ethical committee of the Department of Medicine of Lausanne University Hospital. The subjects were divided into three groups, NGT, IGT, and NIDDM, according to the criteria of the National Diabetes Data Group,’ although an oral glucose load of 100 g instead of 75 g was used in the study. The physical and biochemical characteristics of the three groups are given in Table 1. The NIDDM group was later subdivided into two subgroups, according to their immunoreactive insulin (IRI) response to the glucose load. “,” The high responders presented AIR1 > 70 rU/mL, and the low responders, AIR1 4 30 pU/mL. No values were found in between these groups. All these cases originate from the indigenous population in the Lausanne area. Experimental

Protocol

As in the first study, started in 1979, a 100-g OGTT was performed during continuous indirect calorimetry. The subjects were studied after an overnight fast. The medication of the diabetic patients (Table 1) was removed for at least 24 hours before the test. After a half-hour period of bed rest, during which basal measurements were made, the subjects were administered a 100-g oral glucose load in 400 mL lemon-flavored water, and the measurements were performed for a further 3-hour period. Continuous indirect calorimetry was performed throughout the study period. Blood was withdrawn from an antecubital vein every 30 minutes starting 30

Metabolism, Vol39, No 10 (October), 1990: pp 1068-1075

1069

FOLLOW-UP OF GLUCOSE METABOLISM IN OBESITY

Table 1. Characteristics of Subjects With Initial NGT. IGT, or NIDDM 1st

SubiectNo.

Sex/Age (vd

2nd Test

Test

Weight (kg)

BMI (kg/m*)

Interval (V)

A Weight (kg1

A BMI (kg/m2)

Medication

NGT 1

F/33

78

31.6

6

-11.5

-4.2

2

M/38

86

27.5

6

-0.4

-4.0

3

F/48

79

32.5

6

2.5

4

F/35

75

26.6

6

3.5

1.4

5

F/24

85.5

30.3

6

14.0

5.0

6

F/2 1

76.5

29.9

6

7

F/36

69

25.7

6

2.5

1.2

8

M/34

78

26.4

5

6.0

2.1

-8.5

1.1

-4.9

9

F/22

92.5

31.3

5

10

Ml43

88.5

27.9

8

6.7

2.1

11

M/32

98

31.3

8

1.5

1.2

12

M/69

93

33.3

5

5.0

1.9

13

M/37

124.5

40.7

5

20.5

6.3

Mean

36.3

86.4

30.4

6.0

2.7

0.5

3.5

3.9

0.3

2.4

1.0

14

F/38

86.5

28.9

6

-3.0

15

F/32

94.0

35.4

6

- 19.0

-6.8

16

F/44

140.0

53.3

6

-27.0

-9.2

17

F/34

92.5

31.3

6

8.5

2.4

18

M/49

101.0

31.9

6

1.5

1.2

+SEM

1.1

-6.5

-2.2

IGT 0.7

19

F/5 1

78.5

30.1

6

1.5

1.5

20

F/44

93.5

35.2

6

0.5

0.2

21

M/58

140.0

45.2

10

-8.5

-3.2

22

F/46

121.5

46.9

5

-15.0

-5.7

Mean

44.1

105.3

37.6

6.3

-6.7

-2.1

2.9

7.6

2.9

0.5

+_SEM

3.9

1.4

NIDDM 23 24

M/49 F/4 1

141

47.1

6

-18.0

-6.0

None

110

44.6

8

-13.5

-4.1

None

25

M/52

102

34.5

7

-8.5

-2.5

Insulin

26

M/67

119

40.7

7

-17.0

-5.0

Sulfonylureas

27

M/52

110

31.8

7

28

Ml49

111.5

43.6

4

29

M/48

82

27.7

8

14.0

5.7

Insulin

30

M/60

75.5

27.7

6

0.0

1.2

Sulfonylureas

31

F/49

104.5

41.3

6

- 12.0

- 1.7

Biguanides

32

F/62

83

32.4

6

- 15.5

-5.0

Sulfonylureas

33

Ml50

85.5

30.3

7

-1.5

-0.9

Insulin

Mean

52.6

102.2

36.5

6.6

-6.2

- 1.6

tSEM

2.2

5.9

2.1

0.3

minutes before the glucose load, for measurements of glucose, IRI, and free fatty acids (FFA). Urine excreted during the test was collected at the end to measure urinary glucose and nitrogen excretion. Gas exchange measurements, allowing calculation of glucose and lipid oxidation rates, were performed during a 30minute control period immediately before glucose ingestion and for 3 hours after the glucose load using a computerized open-circuit indirect calorimetry as previously described.” A transparent plastic ventilated hood was placed over the subject’s head and made airtight around the neck. To avoid air loss, a slight negative pressure was maintained in the hood. Ventilation was measured with a digital pneumotachograph (Hewlett Packard, 47303A, Palo Alto, CA). A constant fraction of the air flowing out of the hood was automatically collected for analysis; the oxygen concentration was continuously

4.0 -0.5

3.1

1.2 -0.3

insulin + sulfonylureas Insulin

1.1

measured by means of a differential analyzer (Magnos 4G, Hartmann and Braun, Frankfurt, FRG) and carbon dioxide concentration by means of an infrared analyzer (Uras 3G, Hartmann and Braun). The method allows the estimation of the total quantity of glucose and lipid oxidized under basal conditions and during the 3-hour period following glucose ingestion.‘” The amount of glucose oxidized in response to the oral load (ie, suprabasal oxidation) was calculated by subtracting the basal rate of glucose oxidation, calculated for 3 hours, from the total amount of glucose oxidized within 3 hours following glucose ingestion. The total amount of glucose disposed of by nonoxidative pathways (corresponding mainly to glycogen synthesis and glycolysis) during the 3-hour period following the oral glucose load was obtained by subtracting from the 100 g ingested glucose: (1) the total quantity of

1070

JALLUT ET AL

glucose oxidized, (2) the urinary glucose loss, and (3) the excess glucose remaining in the glucose space. The glucose space was taken to be 25% of body weight,” a mean value between those reported by different investigators. “~‘5It has been shown that calculation of the glucose space as a percent of body weight does not vary between nondiabetic and diabetic humans.16 It should be noted that the above calculation for nonoxidative glucose uptake assumes that the entire (100 g) glucose load is absorbed. ” However, recent studies indicate that only 80% to 85% of the ingested glucose load is absorbed within a 3-hour period.‘8,‘9Thus, the absolute amount of glucose taken up in a nonoxidative way is expected to be slightly overestimated, but the error can be assumed to be the same for all subjects; the comparisons of the calculated values between subjects remain valid. Analytical Procedures Plasma and urinary glucose were measured by the glucose oxidase method, and plasma IRI concentration was measured according to the method described by Herbert et aL2’ Plasma FFA were extracted with the method of Dole and Meinertz” and determined according to Heindel et aL2* Urinary nitrogen was measured using the Kjeldahl method? Data Analysis For each measured metabolic parameter, the basal (fasting) value was the mean of at least three determinations. After glucose ingestion, mean values for plasma glucose, insulin, and FFA levels were calculated by averaging the measurements obtained at 30minute intervals between 0 and 180 minutes. Statistical Methods All data are given as mean f SEM. A one-way ANOVA, with a multiple comparison test (Tuckey test) was used to differentiate the three groups. Paired t tests were used for comparing the individual values between the two tests. Unpaired t tests were used to compare the mean values between the different groups. Table 2. Biochemical

end Metabolic

Charactertistics

RESULTS

The three groups of obese subjects are shown in Table 1. The mean asight of the group with NGT was lower than that of the groups with IGT and NIDDM: 86.4 + 3.9 versus 105.3 -r- 7.6 (PC .05) and 102.2 5 5.9 kg (PC .05),

respectively. From the group of 13 subjects who initially presented a NGT, four became intolerant to glucose. From the group of tlinc s;:b;_, ‘s who initially presented IGT, three became diabetic, and one (case no. 15) whose BMI decreased from 35.4 to 28.6 kg/m2 with a weight loss of 19 kg, returned to NGT. The mean duration of obesity at the time of the first test was lower in the NGT group (8.3 * 1.3 years) than in the IGT (20.1 + 1.6 years, P < .Ol) and NIDDM (28.9 f 2.2, P < .Ol) groups. Similarly, the age was lower in the NGT group (36.3 + 3.5 years) than in the IGT (44.1 + 2.9, NS) and NIDDM (52.6 2 2.2, P < .Ol) groups. The evolution of glucose metabolism in the three groups is shown in Table 2. In the NGT group, fasting plasma glucose and the mean incremental value of plasma glucose during the OGTT (AG) were increased in the second test. No change was found in basal IRI, the increment over basal IRI values (AIRI), and suprabasal glucose oxidation. Glucose storage slightly decreased from 67.6 2 2.1 to 60.1 ? 3.0 g/3 hours (P c .005). In the IGT group, fasting plasma glucose was increased from 93.0 + 3.3 to 106.6 + 4.1 mg/dL (P < .Ol) in the second test. No significant change was observed in the other parameters. In the NIDDM group, AIR1 showed an important decrease (89.0 + 26.8 to 24.3 * 7.8 pU/mL, P < .05). Glucose storage decreased markedly from 39.3 + 3.2 to 21.0 i- 5.0 g/3 hours (P c .OOS). Comparing the increment in glycemia above basal levels

of the NGT IN = 131. IGT (N = 9). and NIDDM (N = II) 1st Test

NGT Fasting plasma glucose (mg/dL)

2nd Test

Groups P

93.9

+ 1.8

99.4

2 2.1