American Journal of Primatology 14:153-166 (1988) Glucose Disappearance Rate in Rhesus Monkeys: Some Technical Considerations K-L.C. JEN' AND B.C. HANSEN' 'Department of Nutrition and Food Science, Wayne State Uniuersity, Detroit; 'Department of Physiology, School of Medicine, University of Maryland Baltimore The intravenous glucose tolerance test (IVGTT) has been widely used as a tool to assist in the diagnosis of diabetes mellitus. Glucose disappearance rate (KG)is calculated as an indicator of relative glucose tolerance; however, a standardized dose, consistent sampling times, and a consistent formula for the calculation of KG have not yet been established for rhesus monkeys. Interpretation of results reported by different laboratories has, therefore, been rendered difficult. In the present study, 48 IVGTTs obtained from 33 male rhesus monkeys ranging widely in glucose tolerance have been analyzed. Various formulas for calculating KG values have been tested in all experiments including a range of different pairs of time points, as well as the tlh. Regression analysis revealed that the log, transformation of the plasma glucose levels obtained after an intravenous glucose load were best fitted with a straight line during the period between five and 20 minutes (R' = 0.97 0.005). The use of time points prior to the five-minute value tended to produce spuriously larger KG values, while sampling points that were later than 30 min occasionally produced an invalid KG because in some monkeys the plasma glucose levels had already returned to basal levels. The advantages of using the five- and 20-minute glucose levels in calculating the KG include (1)the optimal reflection of tissue uptake of glucose; (2) the relatively short sampling time required to obtain an accurate, consistent, and meaningful value for KG; and (3)the relative ease in the calculation of KG. Key words: Macaca mulatta, intravenous glucose tolerance test, KG INTRODUCTION The intravenous glucose tolerance test (IVGTT)has been widely used to identify impaired glucose tolerance and diabetes in humans [Amatuzio et al, 1953; Luft et al, 1981; Streeten et al, 1964; Obenshain et al, 19781 and in other species [Ausman & Gallina, 1978; Howard, 1975; Cunningham et al, 19831. The results of IVGTTs usually have been expressed as the glucose disappearance rate (KG, %/min)-the Received December 31, 1985; revision accepted June 19, 1987. Address reprint requests to Dr. K-L.C. Jen, Department of Nutrition and Food Science, 160 Old Main, Wayne State University, Detroit, MI 48202. 0 1988 Alan R. Liss, Inc. 60 sec na 60 g in 300 ml 0.5 glkg body weight 0.5 g k g body weight 25 g in 83.3 ml volume 25 g in 50 ml volume Human = 0.5 glkg weight Dog = 0.3 g/kg weight 0.5 glkg weight Human Human & dog Human Human Human 4 min 4 min 2.5 min 5-6 rnin 5 min 30,40,50, and 60 min 2 , 4 , 6 , 8 , . . . 16, 19, 22, 27, 32,42,. . . 102, 122,. . . 182 rnin 10,20, 30,40,50, and 60 min 5, 10, 20,30,40, 50,60, 75,90,105, and 120 min 15,30, min, up to 2 hr 5, 10,20, 30,45,60,90, and 120 min 20,30,40,50, and 60 min 4, 12, 20, 28, 36,44, and 72 min 2 min Human Human Human 3,4, 5, 7, 10, 15, 30,45, 60,90, and 120 min 1.0, 2.5, 10,20 g in 50% solution, 25 g for adults 1g/kg for infants Human Sampling time na* Dose Species Infusion interval TABLE I. IVGTT Parameters in Some Selected Experiments* Least-square fit of the values at 30, 40,50, and 60 min Doar et a1  Bergman et a1  Lundback  Amatuzio et a1 119531 Boberg et a1 119761 na B, used incremental glucose, peak value = 4 min B, peakvalue = 20 min A, tl = 10min tz = 42 min Karam et a1  Seltzer et a1  Obenshain et a1 11978) A, t - 10 min t z = 3 'Gin A, tl = 10 min, t z = 50 min na na KG calculation" . 3 t; rp 5, 10, 15,20, 30,40,and 60 min 10, 20, 30,60,90, 120, and 180 min 1-2 min 1min na na 30 sec 0.5 g k g weight or 0.34 g/kg weight 0.3 g k g weight 2.5 g 1.1 g k g weight 0.5 glkg weight 0.5 or 1.0 g k g weight Rhesus monkeys Black Celebes apes information not available in published citation. KG = logeG1-1OgeGz , G1 = glucose level at time tl, Gz = = tz-tl of the peak value. "A *na Rhesus monkeys Rhesus monkeys Rhesus monkeys 1min 3, 15,30, and 45 min na 0.5 glkg weight B, = Howard  Kessler et a1  Kirk et a1  Pitkin et a1  Smith et a1 119731 Kemnitz & Kraemer  Hamilton et a1  Hamilton & Ciaccia  Cornblath et a1  Wilson et a1  Ausman & Gallina  Butterfield et a1 119711 time point when glucose level reached one-half tl = 10 min tz = 30,45, or 60 min B, peak value = 30 min A, B, peak = 5 min B, peak value = 10 min peakvalue = 15 min na B, B, peak value = 5 min na min na A, tl = 5 min, t2 = 20 peakvalue = 13 min or 24 min = glucose level at time tz. B = KG = -69.3 tYz t?4 15, 30, 60,90, 120, and 180 min 10,20,30,45, and 60 min 2, 5, 10, 15,20, 30,40, 50, and 60 rnin 15,30,60,90, and 120 min 1min 0.5 g-1.0 g/kg weight Macaque monkeys Rhesus monkeys 5,10,20,40, and 60 rnin 1min 0.4 g k g weight Rhesus & squirrel monkeys Rhesus monkeys 30 sec 1 , 2 , . . . 10, 13,15,18, 20, 24,25,28, 29,40, 50, and 61 min 2,5,8, 11, 14, 17, 20, up to 90 min at 10-min interval 10,30, and 60 min 0.5 g k g weight 3 min Squirrel monkeys Human . G s cn (D 8 156 J Jen and Hansen slope of the decline of the log,'s of plasma glucose per minute between any two time points. Plasma glucose concentrations after an IV glucose load are described as fitting a straight line when plotted on a semilogarithmic scale [Lundbaek, 19621. Therefore, the rate of fall in glucose level could be expressed as where G1 and G2 are glucose levels at time tl and ta. With this equation, it was expected that results could be compared between studies. However, the published KG results obtained by different laboratories have been difficult to compare due to the fact that different tls and t2s have been used. Table I shows some of the parameters of the IVGTT that have been adopted by different investigators. An alternative KG formula has also been used: 0.693 KG=- t% x 100% per minute with tlh as the time point (min) when glucose concentration drops to one-half either of the peak value or of a certain value [Lundbaek, 19621. Again, different researchers or laboratories have used different time points or different criteria for defining the peak value. In a typical IVGTT, glucose concentration reaches a peak rapidly and then reduces exponentially-reflecting the distribution of the glucose in the glucose space. Concentration then follows a slow decline due to the active uptake of glucose by the cells [Lundbaek, 19621. Thus, it is likely that high KG values could result if very early tl and t 2 are used. On the other hand, a low KG might not represent glucose intolerance but rather might reflect the use of a later tl and t 2 in the calculation. When an excessively long interval between tl and t 2 is used, the assumption that the logarithmic transformation of glucose concentration at tl and t 2 is linear may not be met. There have been few reports that have systematically compared variations in the several parameters of intravenous glucose tolerance testing across various subject groups. Among the testing or analytical methodological parameters that may be considered are (1)the dose of glucose (either fixed at various amounts or fixed on a per kg body weight basis); (2) the time period over which the glucose load is administered; (3) the time points at which plasma (or serum) samples are obtained and (4)the method of calculation of glucose disappearance rate. In the development of a standardized approach to intravenous glucose tolerance testing, it is also important that a method be selected that can be applied to all subjects regardless of current glucose tolerance status, since tolerance status most often is not known until after testing. Further, to be most useful, the method should be equally valid for normal and severely diabetic subjects and should be readily calculated via a consistent formula. Toward these ends, the present study has examined the many methodological approaches used in both human and nonhuman primates and has systematically tested the usefulness of parameters (3) and (4) above, the time points for withdrawal of blood samples, and the formula used for calculation of the glucose disappearance rate. Parameters (1) and (2), glucose load, and time period of infusion were fixed in a manner compatible with the principal that the method should be equally applicable to normal and diabetic subjects and consistently applied. Glucose DisappearanceRate / 157 IVGTT (all subjects) N = 48 300- x ? SE - 250-. - o\ E3 f 200- u 150- 0 100- d 50 , I I 1 1 Fig. 1. The results of 48 intravenous glucose tolerance tests conducted in a group of 33 male rhesus monkeys (X+SE). JYG'ITs were carried out on a group of adult male rhesus monkeys known to vary widely in glucose tolerance from normal to severely diabetic. These tests were conducted in order to develop a standard method of glucose tolerance testing and reporting that can be readily applied to rhesus macaques and that can serve as a diagnostic tool to distinguish rhesus monkeys with very early indications of impaired glucose tolerance (or frank diabetes) from those with normal glucose tolerance. The least-squares fit was examined by using different data sets, each consisting of multiple time points to determine the optimal method; that is, the tl and t 2 for the KG equation that represented the most consistent and reliable estimate of glucose disappearance rate in rhesus monkeys. With the application of consistent calculations carried out between monkeys and within single monkeys over extended periods of time, improvement in the reliability and validity of the estimate of glucose tolerance should result. The change in KG over time may serve as a superior indicator (relative to other measures) of small degrees of deterioration of glucose tolerance in monkeys. MATERIALS AND METHODS Subjects Thirty-three male rhesus monkeys (Macaca mulatta), ranging in age from 10 to 25 years and in weight from 9.6 to 22.0 kg, were selected to represent a wide range of ages, weights, and glucose tolerances. The monkeys were individually caged in stainless-steel primate cages and provided with Purina Monkey Chow-25 (Ralston, MO) and water ad libitum. Of the 33 monkeys, five had clinical symptoms of diabetes mellitus, including glycosuria and polydipsia. The monkeys were maintained in accordance with the NIH Guide for the Care and Use of Laboratory Animals. 158 I Jen and Hansen Procedures Intravenous glucose tolerance tests were conducted following a 16-hour overnight fast on monkeys whose intakes of Purina chow or a standardized diet were within normal limits for the three days preceding the test. The monkeys were lightly anesthetized with ketamine hydrochloride. Our own observations and those of Kemnitz and Kraemer  and of Brady and Koritnik  have shown that ketamine has minimal effects on the results of an intravenous glucose tolerance test. A peripheral intravenous catheter was inserted for the duration of the test. Following the collection of basal blood samples, a 50% glucose solution (dose 0.227 gkg) was infused intravenously over a 30-second period; the cannula was then flushed with enough heparinized saline t o ensure no residual glucose. Blood samples were withdrawn 1, 3, 5, 10, 15, 20, 30, and 60 minutes after the termination of the glucose infusion. Before each sample collection, the dead space, plus 1-2 ml blood, were withdrawn in order to avoid possible dilution of the sample. Heparinized saline was infused between samples in order to maintain the patency of the indwelling catheter. Blood samples were immediately transferred to heparinized and chilled tubes and placed on ice. No more than four samples were collected before the samples were transferred to the centrifuge. Plasma and red blood cells were then separated, and plasma samples were frozen for later assay. Glucose concentration in each plasma sample was determined by the hexokinase and glucose-6-phosphatedehydrogenase method [Slein, 19711. Statistical Methods In order to determine the part in the IVG?T curve that best fitted with a linear regression line of the natural log values of glucose, regression analyses using three to seven data points each were carried out for the following time intervals: 1-15 minutes (5 points), 3-15 minutes (4points), 3-20 minutes (5 points), 3-60 minutes (7 points), 5-20 minutes (4 points), 5-30 minutes (5 points), 10-20 minutes (3 points), 10-30 minutes (4 points), and 10-60 minutes (5 points). The portion of variance accounted for by the linear regression line (R') was then determined. The glucose disappearance rates (KC) calculated for the above-mentionedtime points using both methods of calculation were compared between normal and diabetic groups using the Student's t-test. RESULTS The means and standard errors for each time point are shown in Figure 1. A 3.5-fold rise of plasma glucose levels (from 78.8 f 3.6 to 281.4 f 12.6 mg/dl) was reached one minute after the termination of the glucose infusion. Glucose concentration then dropped rapidly. By 60 minutes after infusion, plasma glucose levels (83.6 5 4.7 mg/dl) were not different from the basal levels. These data are also shown in Figure 2, with the diabetic monkeys plotted separately from the normals. The lower panel represents the absolute plasma glucose levels of the two groups, while the upper panel shows the glucose rise above basal level. Fasting plasma glucose concentrations at basal level were significantly higher in the hyperglycemic and diabetic (133.8 f 12.1 mg/dl) groups compared to the normal group (70.9 f 1.5 mg/dl, P < .001). However, there were no differences between groups in the peak glucose reached (271.3 & 19.2 for the diabetic monkeys and 282.8 i-14.2 mg/dl for the normals). The absolute glucose levels were significantly higher in the diabetic than in the normal monkeys at each sampling point after the first (Ps < .001). When the increase above basal level was considered, the peak response reached in the diabetic animals (137.5 f 24.8 mg/dl) was significantly Glucose Disappearance Rate I 159 uNormal 0-0 Hyperglycernic *p<0.05 **D<o.ol 300 - 250- P 200- I I Ly A 0 2 150- 0 100- 50' I 0 1 10 20 30 40 50 60 MINUTES Fig. 2. The results of intravenous glucose tolerance tests in normal (0;n=42) and diabetic rhesus monkeys (0; n=6). The bottom panel represents the absolute glucose values. The top panel shows the increment in glucose levels above basal. lower than that in the normal animals (211.9 f 14.0 mg/dl, P < .05). Compared to normal monkeys, however, diabetic monkeys had significantly higher incremental glucose levels at 20 minutes (72.7 k 6.8 vs. 50.3 f 3.5 mg/dl, respectively; P < .05); at 30 minutes (51.8 +_ 6.7 vs. 25.8 f 3.3 mg/dl, P < .Ol);and at 60 minutes (20.3 6.4 vs. 2.8 2.1 mg/dl, P < .01). Diabetic monkeys thus evidenced a significantly reduced rate of clearance of glucose after the IV glucose load. A comparison of the individual changes in plasma glucose following an IVGW in a normal and in a diabetic monkey is shown in Figure 3. The least-squares fit for each time period is presented in Table 11. Of the ten regression analyses, the best linear fit was obtained for the period between five and 20 minutes, accounting for 97% of the variance. The regression equation for the periods of 3-15 minutes, 3-20 minutes, 3-30 minutes, 5-30 minutes, 10-20 minutes, 10-30 minutes, and 3-15 minutes gave comparable results, while 1-15 minutes and 10-60 minutes accounted for 80-84% of the variance. The glucose disappearance rates calculated for different time points for the normal and diabetic groups are shown in Table III. While KG calculated for all pairs of time points significantly differentiated the diabetic and normal groups (Ps< .005 to .001), differences were best identified by using the 5-20-minute time points, with or without those points between the 5-20-minute samples. R2 0.934 i 0.013 0.53-1.0 0.809 f 0.004 0.02-0.97 + 0.933 0.012 0.72-1.0 3-20 0.959 f 0.006 0.82-1.00 3-30 0.939 f 0.005 0.83-0.99 3-60 +SE P +SE Range x Diabetic monkeys Range x Normal monkeys 2.39 i 0.238 1.813.05 < -005 4.98 j ,0.295 1.5510.60 1-15 1.74 j,0.119 1.482.19 < ,005 3.50 k 0.207 1.828.66 3-15 1.53 k 0.079 1.351.84 < .005 3.30 f 0.191 1.477.63 3-20 0.970 f 0.005 0.87-1.0 5-20 0.950 f 0.010 0.77-1.0 10-20 0.940 _+ 0.008 0.73-1.0 5-30 0.946 k 0.011 0.65-1.0 10-30 5-20 1.38 0.97 f 0.118 f 0.062 1.110.791.89 1.19 < .001 < .001 1.21 f 0.059 0.991.38 < .001 1.86 3.01 f 0.066 k 0.154 184.108.40.206 5.20 3-60 10-20 1.18 i 0.143 0.841.81 < ,001 1.19 k 0.148 0.891.94 < ,001 2.74 2.86 k 0.129 _+ 0.140 0.941.054.57 4.96 5-30 10-60 t,(lmin) 1.08 f 0.169 0.831.90 < .001 0.79 0.87 f 0.050 k 0.235 0.660.0010.99 1.65 < .001 < .001 2.58 1.57 6.07 & 0.132 0.056 k 0.93 0.870.931.214.50 2.51 25.11 10-30 Calculated From Different Time Points in Normal and Diabetic Groups 2.95 k 0.15 1.305.58 3-30 TABLE 111. Glucose Disappearance Rate %%) *Based on 48 IVGlT experiments using 0.25 g glucosekg injected IV over 30 sec. Range of R2 (X k SE) 3-15 1-15 TABLE 11. Least-Squares Fit (R2) of Glucose Disappearance Rate at Different Time Points* + 0.80 f 0.118 0.321.13 < .001 3.27 f 0.49 1.0717.33 tljL(3min) 0.836 0.021 0.47-0.99 10-60 F 2 9a 3 s. Glucose Disappearance Rate / 161 -Monkey Y-4 0-oMonkey 1-4 3501 300- -? =E - 250- 200- I YI n 0 2 d 150- 100- 50- 0 1, 0 10 20 30 40 50 t 60 MINUTES Fig. 3. A comparison of the results of an intravenous glucose tolerance test in a normal ( 8 )and diabetic (0) rhesus monkey. KG calculated by employing the time points for glucose levels to reach half of the levels at one minute and three minutes postinjection are also shown in Table 111. The ts using the one-minute point had greater mean KG values and larger variance when compared with other calculations for both the normal and diabetic groups. The difference in KG between those groups was statistically significant ( P < .05). The t%based on glucose levels at three minutes reduced the mean KG and variance. The difference between diabetic and normal groups was not, however, as statistically significant as in other calculations (P < .06). The t%based on glucose at five minutes cannot be calculated, since for the majority of the monkeys, half of the glucose value at five minutes was either lower than the initial fasting level or was reached between 30 and 60 minutes, far beyond the linear portion of the glucose curve and thus producing an artificially low KG. The within-subject reproducibility of the KG in replicate tests was highest in those with low KG and most variable in those with high KG values. At the termination of the test, 60 minutes after the glucose injection, glucose levels had returned to or below basal glucose levels in 25 out of the 48 experiments. Therefore, any KG calculation involving time points at or beyond 60 minutes would not produce a good fit. DISCUSSION The IVGTT is a widely used tool in the diagnosis of diabetes mellitus. The NGTT provides a better indicator of glucose disposal rate than the oral glucose tolerance test (OGTT), which is significantly affected by gut motor activity, factors related to gastric emptying rate, and gastrointestinal hormone release [Rehfeld et al, 1970; Thompson et al, 19821. The IVGTT takes less time than the OG!M' to perform (usually up to 60 minutes or 90 minutes instead of two to three hours as in an OGTT). Although the glucose disappearance rate after an IV glucose load has long been recognized as the best way to summarize the IVGTT results [Amatuzio et al, 19531, the time points between which the log-transformed glucose levels have 162 I JenandHansen been considered to be a linear function have been defined variously by different investigators. Therefore, the comparison of results obtained by different laboratories has been discult, as pointed out by Buttefield et a1 . An optimal formula for calculation of the glucose disappearance rate should (1)be easy to perform so that even a large data set can be calculated easily; (2) use convenient sampling time points such as 5, 10, 15, 20 minutes, etc, instead of 4, 8, 12 minutes, or other [Amatuzio et al, 19531; (3)be valid across varying levels of basal glucose, so that the KG can still be calculated with the onset or advancement of diabetes; and (4) be sensitive enough to pick up the earliest changes in glucose tolerance. Inconsistencies in the IVGTT procedures used by various investigators include the rate of infusion of the glucose load, the amount infused, the sampling times, and the method of analysis of data. Table I briefly summarizes some of the IVGTT results reported previously. Such differences in rate of infusion of glucose, glucose load, and sampling times have produced widely varying peak plasma glucose levels and times of peaks. The time needed to inject glucose depends on the volume of glucose solution given, which in turn depends on the dose and the concentration of glucose solution. In humans, the glucose doses employed have included (1)25 g in 50 ml solution-a 50% solution [Lundbaek, 1962; Streeten et al, 19641; (2) 25 g in 83.3 ml solution-a 30% solution [Amatuzio et al, 19531; (3) 60 g in 300 ml solution-a 20% solution [Karam et al, 19631;(4) 1.0,2.5, 10, and 20 g glucose in 50%solution [Lerner & Porte, 19711; (5)0.5 g glucose per kg body weight [Boberg et al, 1976; Seltzer et al, 1967; Doar et al, 19681; and (6) 0.3 gfkg body weight [Bergman et al, 19811. When glucose injected was calculated on a 0.5 gfkg body weight basis, an average 70 kg man would receive 35 g glucose, higher than 25 g but lower than the 60-g glucose dose. Amatuzio et a1  compared the NGTT results of 25 g glucose with those of a 35-g glucose load and found that although the absolute glucose levels were higher with the 35-g glucose load, the glucose disappearance rates of both doses were identical. Lerner and Porte  compared the KG values and the dose of glucose loads. They found that KG increased with the increase in glucose load in a nonlinear fashion. KG did not increase further when glucose load was more than 10 g. With 300 ml of glucose solution injected as reported by Karam et a1 , the effect of the fluid load was of potential concern. In an IVGTT, urinary loss of glucose is a major problem [Butterfield et al, 19711. In the present study a lower glucose dose of 0.25 gfkg was used. This glucose dose was selected in order to minimize urinary excretion of glucose during the course of the test, since 16-hour water deprivation of some diabetic monkeys is not appropriate. Kaneko et a1 , using various doses of glucose, showed that this dose level produced a urinary loss of less than 2% of the glucose load. Howard  commented that glucose levels used to calculate KG should be at least five to ten minutes after glucose load. In the present series of experiments, KG calculated with glucose values between ten and 20 minutes were close to the KG based on five- and 20-minute sampling. Since five- to 20-minute sampling contained more sampling points (four points) than ten- to 20-minute sampling (three points), KG calculations were based on five- through 20-minute sampling times, which involved all of the data points until glucose levels returned to baseline levels. Kessler et a1  calculated KG by using a different method, which involved all of the data points until glucose levels returned to baseline levels. Therefore, the KG could be based upon a 30-, 4 5 , or 60-minute period. In a later publication by this same group [Howard et al, 19861, the KG calculation was changed to use the period from the ten-minute glucose level onward until the “glucose concentration was within 10%of the fasting level.” The concern with this approach is that for diabetic Glucose Disappearance Rate I 163 monkeys, blood glucose sometimes has not returned to baseline values or to within 10%of baseline at 60 minutes (see also Fig. 3). Some KG calculations have not used the difference between glucose levels at two predetermined time points but have instead been calculated on the basis of the time needed for glucose to drop to half of a peak value. In such cases, each ts had to be calculated individually, a tedious procedure when handling a large sample size. Furthermore, this approach, due to differing injection intervals and sampling times, made the peak glucose value hard to define between different laboratories. Table 111 shows that, based on the data obtained from this laboratory, the mean KG based on a ts calculated from the one-minute level would be higher than results obtained by other methods. The variability was also significantly greater. Kemnitz and Kraemer  reported that using this formula, the mean KG changed from 2.78 to 3.25% when 15, 30-, and 45-minute samples were considered instead of 15, 30-, 4 5 , and 60-minute samples. In these two calculations, since the peak value did not change, only the interpolation to find the t s was changed. This implies that for some monkeys, basal glucose levels were reached sometime between 45 and 60 minutes; therefore, the linear regression model did not closely fit the sampling points of 15, 30,45, and 60 minutes. As a result, the mean KG was smaller. Cornblath et a1 119711 studied juvenile monkeys between 18 and 24 months of age, restrained and without anesthesia. The investigators reported what must be considered to be an abnormally low “normal” KG (1.75 +_ .28%/min). Examination of their data show plasma glucose levels at one hour following the IVGm at a level of 100 mg/dl, clearly not a normal fasting or one-hour IVGTT glucose for rhesus monkeys. The results reported by Bergman et a1 [1981, 19851 showed that for some dogs, the glucose concentration reached basal levels by 30 minutes. Nevertheless, they consistently used ten and 42 minutes to calculate the KG. It was apparent that a linear regression model could not fit these cases well. In the present study, calculations based on the half-time method were done by using glucose concentration peak values at only one minute or three minutes. This limitation was necessary because for some monkeys, half of the glucose concentrations at five minutes was below the basal level; therefore, the calculations violated the basic assumption of a linear relationship between glucose level and time postglucose injection. Absolute glucose levels were used in calculating the KG by all of the investigators cited here except Amatuzio et a1 . The latter inserted the glucose incremental values into the formula. With this method, only the early time points could be used to calculate the KG, since after 52 min, glucose levels had already returned to or even below the basal glucose levels. KG values calculated based on absolute glucose values discriminate better between normal and diabetic subjects than do those calculated based on incremental values [Butterfield et al, 19711. In addition, absolute KG values correlated better with fasting glucose levels and glucose levels 60 minutes after loading [Butterfield et al, 19711. The reduction in KG with age has been previously documented [Lundbaek, 19621;however, in the rhesus monkey, this diminution is not linear, with the decline beginning only after age 10 (unpublished observations). Furthermore, when only lean monkeys are considered, the decline is extremely modest, which agrees with the report of Lundbaek . Most of the decline observed was due to the use of obese monkeys with progressively deteriorating glucose tolerance leading ultimately to NIDDM. As shown in Table 111, KG calculated by using a number of these formulas successfully differentiated monkeys with fasting hyperglycemia and diabetic signs from normoglycemic monkeys. However, regression analysis conducted in the pres- 164 1 Jen and Hansen ent study revealed that the log glucose levels following an IV glucose load between five minutes and 20 minutes were the most linear. Ninety-seven percent of the variation in glucose levels could be accounted for by this linear fit. In addition, the glucose concentrations at 30 minutes were already below the baseline levels in four experiments. (One example is shown in Fig. 3.) Kaneko et a1  reported that KG values should not be calculated by using time points with glucose values greater than 200 mg/dl; otherwise falsely high KG values may be obtained. In our experiments, since a lower glucose load of 0.25 g/kg was used, glucose levels at five minutes were already lower than 200 mg/dl. Therefore, we accepted the use of absolute glucose values at five and 20 minutes to calculate the KG. Other advantages resulting from the use of these two time points include (1)avoidance of the very early sampling points (i.e., 1,2, 3, 4 minutes), when the reduction in concentrations may be influenced by the rate of distribution of glucose in the glucose space, and (2) reduction in the long sampling interval recommended by some, i.e., 60, 90, or 120 min. By using the recommended formula, the IVGTT procedure can be significantly shortened. In order to make interlaboratory comparisons feasible, not only is the use of a single KG formula required, but also comparable glucose doses and concentrations of the glucose solution. While we have not studied a range of values for each of these variables, we suggest that for normal as well as for diabetic monkeys 0.25 g glucose/ kg body weight (50% solution) administered IV within one minute be used until further studies systematically examine this question in rhesus monkeys. We further suggest that dietary intake be ascertained to be normal on the immediately preceding days and that a 16-hour overnight fasting period should precede the test. Finally, KG for rhesus monkeys is optimally calculated by using sampling time points of five and 20 minutes following the glucose injection, although times between three and 30 minutes are generally adequate. Attention to these details is critical when IVGTT results from various laboratories are compared. CONCLUSIONS 1. Intravenous glucose tolerance tests using a dose of 0.5 ml per kg of a 50% dextrose solution (0.227 glucose/kg body weight) were performed on both tolerant and intolerant rhesus monkeys. Tests revealed that the glucose disappearance curve was best fitted by a linear regression line between five and 20 minutes following this glucose load. 2. KG calculated with gluocse levels at five and 20 minutes demonstrated that for normal monkeys (normal fasting glucose levels), the KG values ranged between 1.14% and 5.2%, with a mean of 3.01%. 3. For diabetic monkeys (fasting hyperglycemia: glucose > 140 mg/dl), the KG values fell between 0.99% and 1.38%,with a mean of 1.21%. 4. We suggest that consideration be given to the consistent use of this glucose load (0.227 gkg) and to a consistent method of calculating glucose disappearance using the five- and 20-minute time points, since they can be used with both normals and diabetics. This method of calculation has the added advantage of being quicker and easier than some of the other approaches currently in use. ACKNOWLEDGMENTS We thank N. Bodkin, L. Speegle, G. Schielke, D. Harman, and C. Sweeley for their assistance with this work, and A. Thomas for manuscript preparation. This project is supported by NIH grant DK37717. Glucose Disappearance Rate / 165 REFERENCES Amatuzio, D.S.; Stutzman, F.L.; Vanderbilt, M.J.; Nesbitt, S. Interaction of the rapid intravenous glucose tolerance test in normal individuals and in mild diabetes mellitus. JOURNAL OF CLINICAL INVESTIGATION 32:428-435,1953. Ausman, L.M.; Gallina, D.L. Response to glucose loading of the lean squirrel monkey in unrestrained conditions. AMERICAN JOURNAL OF PHYSIOLOGY 234:R20R24,1978. Bergman, R.N.; Finegood, D.T.; Ader, M. Assessment of insulin sensitivity in vivo. ENDOCRINE REVIEW 6:45-86,1985. Bergman, R.N.; Phillips, L.S.; Cobelli, C. Physiological evaluation of factors controlling glucose tolerance in man. JOURNAL OF CLINICAL INVESTIGATION 68:14561467, 1981. Boberg, J.; Hedstrand, H.; Wide, L. The early Serum insulin response to intravenous glucose in patients with decreased glucose tolerance and in subjects with a familial history of diabetes mellitus. SCANDINAVIAN- JOURNAL OF CLINICAL LABORATORY INVESTIGATION 36:145-153. 1976. Brady, A.G.; Koritnik, D.R. The effects of ketamine anesthesia on glucose clearance in African green monkeys. JOURNAL OF MEDICAL PRIMATOLOGY 14:99-107, 1985. Butterfield, W.J.H.; Abrams, M.E.; Whichelow, M.J. The 25-g intravenous glucose tolerance test: A critical appraisal. METABOLISM 3:255-265,1971. Cornblath, M.; Levitsky, L.L.; Kling, A. Response to intravenous glucose in juvenile macaque monkeys. DIABETES 20:156-161, 1971. Cunningham, J.; Calles, J.; Eisikowitz, L.; Zawalich, W.; Felig, P. Increased efficiency of weight gain and altered cellularity of brown adipose tissue in rats with impaired glucose tolerance during diet-induced overfeeding. DIABETES 32:1023-1027,1983. Doar, J.W.H.; Wynn, V.; Cramp, D.G. Blood pyruvate and plasma gIucose levels during oral and intravenous glucose tolerance tests in obese and non-obese women. METABOLISM 17:690-701,1968. Hamilton. C.L.: Ciaccia. P. The course of development o f glucose’ intolerance in the monkey (Macaca mulatta). JOURNAL OF MEDICAL PRIMATOLOGY 7:165-173. 1978. Hamilton, C.L.; Kuo, P.T.; Feng, L.Y. Experimental production of syndrome of obesity, hyperinsulinemia and hyperlipidemia in monkeys. PROCEEDINGS OF THE SOCIETY OF EXPERIMENTAL BIOLOGY AND MEDICINE 140:1005-1008,1972. Howard, C.F., Jr. Spontaneous diabetes in Macaca nigra. DIABETES 21:1077-1090, 1972. Howard, C.F., Jr. Diabetes and lipid metabolism in nonhuman primates. ADVANCES IN LIPID RESEARCH 13:91-134,1975. Howard, C.F., Jr. Diabetes and carbohydrate impairment in non-human primates. Pp 136 in NONHUMAN PRIMATE MODELS FOR HUMAN DISEASES. W.R. Dukelow, ed. Boca Raton, Florida, CRC Press, 1983. Howard, C.F., Jr.; Kessler, M.; Schwartz, S. Carbohydrate impairment and insulin secretory abnormalities among Macaca muZatta from Cay0 Santiago. AMERICAN JOURNAL OF PRIMATOLOGY. 11:147162,1986. Kaneko, J.J.; Matthheev, D.; Rottiers, R.P.; Vanderst, J.; Vermeule, A. Effect of urinary glucose secretion on plasma glucose clearances and plasma-insulin responses to intravenous glucose loads in un-anesthetized dogs. ACTA ENDOCRINOLOGY CA. 871133-138,1978. Karam, J.H.; Grodsky, G.M.; Forsham, P.H. Excessive insulin response to glucose in obese subjects as measured by immunochemical assay. DIABETES 12:197-204, 1963. Kemnitz, J.W.; Kraemer, G.W. Assessment of glucoregulation in rhesus monkeys sedated with ketamine. AMERICAN JOURNAL OF PRIMATOLOGY 3:201-210,1982. Kessler, M.J.; Howard, C.F., Jr.; London, W.T. Gestational diabetes mellitus and impaired glucose tolerance in an aged Macaca mul a t h JOURNAL OF MEDICAL PRIMATOLOGY 14:237-244,1985. Kirk, J.H.; Casey, H.W.; Harwell, J.F., Jr. Diabetes mellitus in two rhesus monkeys. LABORATORY ANIMAL SCIENCE 22: 245-248,1972. Lerner, R.L.; Porte, D., Jr. Relationships between intravenous glucose loads, insulin responses and glucose disappearance rate. JOURNAL OF CLINICAL ENDOCRINOLOGY AND METABOLISM 33:409-417, 1971. Lundbaek, K. Intravenous glucose tolerance as a tool in definition and diagnosis of diabetes mellitus. BRITISH MEDICAL JOURNAL 21507-1513,1962. Luft, R.; Wajngot, A.; Effndie, S. On the pathogenesis of maturity-onset diabetes. DIABETES CARE 458-63,1981. Obenshain, S.S.; Isaacsohn, M.; King, K.; Schwartz, R. Sequential intravenous glucose tolerance: Responses of newborn and adult. BIOLOGY OF THE NEONATE 33:207-216, 1978. Pitkin, R.M.; Van Orden, D.E.; Reynolds, W.A. Plasma insulin response and glucose tolerance in pregnant rhesus monkeys. EN- 166 I Jen and Hansen DOCRINOLOGY 86:435-473,1970. Rehfeld, J.E.; Juhl, F.; Quaade, F. The intestinal insulino-tropic action after jejuno-iliostomy. SCANDINAVIAN JOURNAL OF GASTROENTEROLOGY 5:77-80,1970. Seltzer, H.S.; Allen, E.W.; Herron, A.L., Jr.; Brennan, M.T. Insulin secretion in response to glycemic stimulus: Relation of delayed initial release to carbohydrate intolerance in mild diabetes mellitus. JOURNAL OF CLINICAL INVESTIGATION 46:323-335, 1967. Slein, M.W. D-glucose determinations with h f ~ o k i n a s eand glucose-6-~hosphated e b drogenase. PP 117-123 in METHODS OF ENZYMATIC ANALYSIS. H.U. Bergmeyer, ed. New York, Academic Press, 1971. Smith, G.P.; Gibbs, J.; Strohmayer,A.J.; Root, A.W.; Stokes, P.E. Effect of 2-deoxy-D-glucose on insulin response to glucose in intact and adrenalectomized monkeys. ENDOCRINOLOGY 92:750-754,1973. Streeten, D.H.P; Gerstein, M.M.; Woolfolk, D.; Doisy, R.J. Measurement of glucose disposal rates in normal and diabetic human subjects after repeated intravenous injections of glucose. JOURNAL OF CLINICAL ENDOCRINOLOGY 24:761-774, 1964. Thompson, D.G.; Wingate, D.L.; Thomas, M.; Harrison, D. Gastric emptying as a determinant of the oral glucose tolerance test. GASTROENTEROLOGY 8251-55,1982. Wilson, R.B.; Martin, J.M.; Kelly, H.; Newberne, P.M. Plasma and pancreatic insulin concentrations in adult squirrel and rhesus monkeys. DIABETES 20:151-155,1971.