Does fasting interval affect the glucose challenge test?

Does fasting interval affect the glucose challenge test? Michael D. Berkus, MD,. Michael P. Stern, MD,b Braxton D. Mitchell, PhD,b Edward R. Newton, MD,. and Oded Langer, MD" San Antonio, Texas The relationship between fasting interval and glucose screening was assessed in a prospective study of 153 non-diabetic pregnant patients undergoing a standard 50g glucose challenge test. An interval of less than 3 hours after the last meal was found to be associated with a significantly greater insulin response as opposed to a fasting interval of more than 3 hours (121.3 pmol/I vs. 83.5 pmol/I, p < .001) and a greater insulin/glucose index (0.92 vs. 0.66, P < .001), with no difference in plasma glucose. Overall, obese patients had a higher glucose response than non-obese, but only higher insulin levels (107 pmol/I vs. 69 pmol/I, p < .001) and insulin/glucose index (0.79 vs. 0.59, P < .02) when the fasting interval was greater than 3 hours. Our data suggest that the fasting interval can influence insulin response during a glucose challenge test and the aberration in insulin secretion may effect screening results, especially in the high-risk gravida with glucose abnormality. (AM J OBSTET GVNECOL 1990;163:1812-7.)

Key words: Gestational diabetes, glucose screening, fasting interval, insulin, obesity The American College of Obstetrics and Gynecologists Technical Bulletin recommended that a 50 gm oral glucose screening should be performed on all pregnant women 30 years of age or older, and on any women with risk factors such as a family history of diabetes or obesity. This screening should be performed between 24 and 28 weeks of gestation "without regard to time of last meal or time of day" and if the plasma glucose level measured 1 hour after the 50 gm load exceeds 140 mg/dl, a 3-hour oral glucose tolerance test should be administered.' Although extensive research has been performed regarding methods of glucose screening to determine the best threshold for identifying patients at risk for gestational diabetes,2-' a paucity of information exists as to the influence of a fasting or fed state on screening values," and no study has examined the influence of fasting interval on insulin response to a glucose challenge test. Previously, in the report of Coustan et al.," it was found that the glucose response was significantly higher if patients with gestational diabetes had fasted than if they had been given a standard 600 kcal meal 1 hour before testing. They had hypothesized that the screening test value was lower in the fed state because previous ingestion of a mixed meal dampened the usual state of insulin resistance observed in fasting pregnant women." From the Department of Obstetrics and Gynecologya and the Department of Medicine,' The University of Texas Health Science Center at San Antonio. Supported in part by National Institutes of Health grant HL-24799. Presented at the Thirty-seventh Annual Meeting of the Society for Gynecologic Investigation, St Louis, Missouri, March 21-24,1990. Reprint requests: Michael D. Berkus, MD, Department of Obstetrics and Gynecology, The University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Dr., San Antonio, TX 78284-7836. 6/6/24719

1812

They did not measure insulin levels in that study to confirm their hypothesis. The screening of pregnant patients after a standard meal may not be applicable to tests administered after the ingestion of unknown foodstuffs. Additionally, a I-hour time interval may not be representative of other time intervals. Of significance is that examination of nonpregnant women with type II diabetes has shown an effect of protein loading on plasma glucose response to a 50 gm screening test after 4 hours and no significant effect at 1 hour. Also, the insulin response after a mixed loading was much greater than after a 50 gm glucose challenge alone. 6 Finally, Cheney et al. 7 have shown that obese women with gestational diabetes manifest higher glucose and insulin values after a 400 kcal breakfast meal, as compared with nonobese patients with gestational diabetes. Thus it seems that pregnant patients with gestational diabetes respond to a glucose challenge test differently whether they are in the fasting or fed state and that this response is further modified by maternal size. However, it is not clear in subjects with an elevated screening result but a normal oral glucose tolerance test result to what extend insulin resistance plays a role in insulin response. Furthermore, there is no information regarding the relationship of fasting interval and obesity with glucose and insulin secretion when only the glucose challenge test result is abnormal. Since the recommendation is to screen patients with a glucose challenge test without regard to timing or content of the last caloric intake, we chose to study mothers under this guideline' to discern the impact on screening results. Thus the purpose of this study was to examine the effect of fasting interval on glucose and insulin response to a glucose challenge test, to ascertain the influence of obesity and fasting interval on screen-

Fasting interval and glucose challenge test 1813

Volume 163 Number 6, Part 1

Table I. Analysis of variance of hourly fasting intervals Fasting interval (hr)

n

0-1 1-2 2-3 >3

20 24 25 84

Glucose (mg/dl)

118 123 120 120 (F =

Insulin (pmoIlL)

28 24 27 23 0.168)

129 114 III 80

82* 65* 91* 51 (F = 4.28)

± ± ± ±

± ± ± ±

Insulin/glucose index

1.04 0.89 0.85 0.65

0.6t 0.3t 0.5t 0.4 (F = 5.47) ± ± ± ±

*0 to I hour, I to 2 hours, or 2 to 3 hours versus> 3 hours, p < 0.04. to to I hour, I to 2 hours, or 2 to 3 hours versus> 3 hours, p < 0.05.

ing results, and to determine whether alterations in glucose and insulin secretion seen in women with gestational diabetes are present with minor degrees of glucose abnormality. Material and methods

Nonhypertensive pregnant women were recruited from the obstetrics clinic of The University of Texas Health Science Center at San Antonio (UTHSCSA) and from other San Antonio-area clinics. The subjects were between 24 and 28 weeks of gestation or late registrants after 28 weeks and were not receiving insulin at the time of the study. The subjects included presumed normal pregnant women about to undergo a routine 50 gm glucose screening for gestational diabetes. Patients were enrolled without regard to their fasting state on an availability basis. The study was approved by the Institutional Review Board of the UTHSCSA and all subjects gave informed consent. Body mass index was calculated as weight (in kilograms) divided by height (in meters) squared. For the purpose of analysis patients were considered to be obese if their body mass index was ?27. Patients' glucose screening results were considered to be normal if the values were <140 mg/dl, to be abnormal if the values were ?140 mg/dl, and to have gestational diabetes when two values on the oral glucose tolerance test were abnormal. Those mothers with gestational diabetes were excluded from the analysis. The patients were asked the number of hours since they had last eaten or drunk liquids other than water to determine the fasting time-to-screening interval. This value was then used as the fasting interval for the study. All testing was completed during morning clinics. Blood samples were drawn 1 hour after a 50 gm load of a glucose equivalent (Koladex or Orangedex, Custom Laboratories, Baltimore) was administered. Plasma glucose level was determined with the use of the glucose oxidase method measured with a Paramax chemistry analyzer (American Dade Chemical Systems, Irvine, Calif.). Serum insulin concentrations were measured with a commercial solid-phase radioimmunoassay procedure (Diagnostic Products, Los Angeles). The coefficient of variation between du plicate aliquots measured

at the same time was 8.1 % for the I-hour concentrations of insulin." The insulin/ glucose index was calculated to adjust for differing glucose levels. Data analysis included bivariate and multivariate statistical techniques with parametric and non parametric tests. Categoric data were analyzed with X2 and Fisher's exact tests. Continuous data were analyzed by Student's t test. One-way analysis of variance was used to compare different fasting intervals. Results

A total of 163 patients participated in this prospective study. There were 119 women with a normal glucose screen « 140 mg/ dl) and 44 with a glucose value ?140 mg/dl. Of the 45 subjects with an abnormal screen, 10 had two abnormal values on the oral glucose tolerance test and were excluded from the analysis; this resulted in a total of 153 patients for evaluation. Examination of fasting intervals of all subjects by means of analysis of variance for hourly groupings yielded significant differences for insulin and insulin/ glucose index levels but not for plasma glucose values (F = 4.28, 5.47, and 0.16, respectively) (Table I). Subsequent post hoc t tests between hourly mean insulin response values yielded significantly greater values between 0 and 1 hour, 1 and 2 hours, and 2 and 3 hours and intervals >3 hours but no differences between intervals of 0 to 1 hour, 1 to 2 hours, or 2 to 3 hours. Therefore for the purpose of analysis the patients were stratified into two groups by a fasting interval of 0 to 3 hours, representing the "fed" patients, and ?3 hours, representing "fasting" patients. As shown in Table II, there were no significant demographic differences between groups, i.e., the "fed" group that ate within 3 hours of the test and the "fasting" group that had eaten 3 to 20 hours before being screened. All the patients were from the routine obstetrics clinic and were of similar age, parity, racial composition, gestational age, incidence of obesity, incidence of abnormal screening results, and plasma glucose level. Notably, both insulin level and the insulin/ glucose index were significantly elevated in the "fed" group «3 hours). The differences in insulin response were due

1814 Berkus et al.

December 1990 Am J Obstet Gynecol

Table II. Patient characteristics and effect of 3-hour fasting interval "Fed" (mean = 1.6 ± 0.8 hr) (n = 69)

"Fasting" (mean = 10.9 ± 5.0 hr) (n = 84)

p Value

25.0 ± 6 6.5 38.6 32.7 30.2 ± 5 26 (37.7%) 18(26.1%) 120 ± 26 117 ± 79 0.92 ± 0.5

23.6 ± 5 3.9 51 39.9 29.6 ± 6 29 (34.9%) 16 (19.0%) 120 ± 23 84 ± 63 0.70 ± 0.5

NS NS NS NS NS NS NS NS

Age (yr) Primiparous (%) Multiparous (%) Hispanic (%) Gestational age (wk) Obesity (No.) Abnormal screen (No.) Glucose (mg/dl) Insulin (pmoIlL) Insulin/glucose index

0.005 0.009

Table III. Effect of 3-hour fasting interval on insulin response n

Insulin (pmoIlL) Normal (glucose < 140 mg/dl) Abnormal (glucose ~ 140 mg/dl) Insulin/glucose index Normal (glucose < 140 mg/dl) Abnormal (glucose ~ 140 mg/dl)

"Fed" 3 hr)

«

"Fasting" (> 3 hr)

p Value NS

119 34

88 ± 43 198 ± 101

82 ± 66 94 ± 44

0.0006

119 34

0.8 ± 0.4 1.26 ± 0.6

0.7 ± 0.4 0.6 ± 0.3

0.0004

entirely to patients with abnormal glucose screens (~140 mg/dl) (Table III). As in insulin response the values for the corrected index (insulin/ glucose) are significantly greater only for women with an abnormal screening. Since obesity has been shown to have an influence on insulin resistance, the additional effect of obesity on glucose and insulin response was analyzed. The glucose response was higher for obese patients when the fasting interval was >3 hours (Fig. I). Dividing glucose response for normal and abnormal glucose groups yielded significantly elevated values only for the normal "fasting" obese group. The insulin response in obese patients was greater than in nonobese women but only for the "fasting" group (Fig. 2). This was true whether the glucose screening was normal «140 mg/dl) or abnormal. Also, for nonobese or lean patients in the abnormal group, the insulin response was significantly more for "fed" lean than "fasting" lean mothers, despite the small number of patients in these categories. When the insulin/ glucose index or ratio was examined, the significance pattern was similar; however, the difference in the abnormal "fasting" group (glucose ~ 140 mg / dl) was only marginally significant for obese versus lean (0.73 ± 0.3 vs 0.53 ± 0.2, P = 0.09) (Fig. 3).

Comment A screening test should reliably detect individuals at risk for a disease in a safe, inexpensive, and convenient manner. The 50 gm oral glucose challenge test has been shown to meet these criteria! Giving the test without

NS

regard to time of last meal adds to the convenience of the test and allows for more rapid assessment when necessary in advanced gestation. However, no previous studies have attempted to determine the effect, during glucose screening, of a variable time interval from the last meal on glucose and insulin results. This is the first report of the insulin response to a glucose challenge under the recommendation of The American College of Obstetricians and Gynecologists. I The key findings in this study support the concept that the fasting interval can influence insulin response during a glucose challenge test and that maternal size may further modulate this effect. A fasting interval of up to 3 hours was found to be the critical interval in this study. Stratifying the data according to these results «3 hours and ~3 hours) yielded significant differences in insulin but not in glucose response. This difference was present only in the abnormal screening group, i.e., plasma glucose level ;:::140 mg/dl. Obese subjects did have higher plasma glucose values than non obese women when "fasting" and also had a greater insulin response than lean patients when the fasting interval was ~3 hours. When the glucose screen was abnormal, both obese and nonobese gravid women had decreased insulin response ~3 hours or more after eating. When the screen was normal (glucose <140 mg/dl), only lean subjects had a decreased response when "fasting." Thus fasting interval did affect insulin response and resistance as reflected by the insulin/glucose index, and obesity further modified the results. The final question addressed by the study was to determine whether the

Fasting interval and glucose challenge test

Volume 163 :\ umber 6, Part I

1815

-FASTING-Obese va. FASTING-Lean, P'.04

Fig. 1. Effect of fasting interval and obesity on glucose response, FED-Obese, Fasting interval <3 hours and body mass index 2:27, FED-Lean, Fasting interval <3 hours and body mass index <27. FASTING-Obese, Fasting interval 2:3 hours and body mass index 2:27. FASTING-Lean, Fasting interval 2:3 hours and body mass index <27.

-FASTING-Obese vs. FASTING-Lean, P'.02

-FED-Lean VI. FASTING-Lean. P'.004

Fig. 2. Effect of fasting interval and obesity on insulin response. FED-Obese, Fasting interval <3 hours and body mass index 2:27. FED-Lean, Fasting interval <3 hours and body mass index <27. FASTING-Obese, Fasting interval 2:3 hours and body mass index 2:27. FASTING-Lean, Fasting interval 2:3 hours and body mass index <27.

glucose and insulin response found in women with gestational diabetes was present in patients with only minor glucose abnormality, i.e., only aberrant results of the glucose challenge test (the abnormal group in this study). As there were insufficient cases of diabetes in

our original sample to have the statistical power necessary for an adequate comparison, data were gathered from the current literature for this evaluation. Coustan et aI.' found a significantly lower plasma glucose screening value when the 50 gm challenge was

1816

Berkus et al.

December 1990 Am J Obstet Gynecol

1.2

0.8 0.6

FE,D'''{)bese

0.4 0.2

o -FASTING-Obese vs. FASTING-Lean, pc.05

-FED-Lean vs. FASTING-Lean, pc.003

Fig. 3. Effect of fasting interval and obesity on insulin/glucose index. FED-Obese, Fasting interval <3 hours and body mass index 2::27. FED-Lean, Fasting interval <3 hours and body mass index <27. FASTING-Obese, Fasting interval 2::3 hours and body mass index 2::27. FASTING-Lean, Fasting interval 2::3 hours and body mass index <27.

administered 1 hour after a standard 600 kcal meal was eaten, but this was true only in subjects known to have gestational diabetes. They hypothesized that this lower value occurred in the "fed" state as compared to the "fasting" state because eatng had "dampened" the normal insulin resistance. We found the insulin resistance, as reflected by the insulin/glucose ration, to be increased in the "fed" mothers in our abnormal group, but there was no difference in the glucose response of "fed" versus "fasting" gravid women. Thus patients with only an abnormal glucose challenge test behaved similarly to the "presumed normal" patients in the study of Coustan et aI., i.e., with no difference in glucose response, but they had increased insulin values probably because of the protein present in their last meal. A dampened insulin resistance was not seen; rather the "fed" subjects with gestational diabetes most likely received stimulation to produce more insulin. This lowered the glucose level significantly, but not to the level of the normal patients, as reflected by higher glucose values (154.8 ± 24.1 vs 115.8 ± 23.4 gm/ dl) in that study.5 In a well-done study by Nuttall et al. 6 untreated, nonpregnant individuals with diabetes were given mixed protein and glucose meals. With increasing protein supplement, they found greater insulin response without any increase in glucose level. Clearly, screening soon after a mixed meal, including protein, as may have occurred in the "fed" group of this study, could explain the increased insulin response obtained. This effect of

protein would necessarily be prolonged in pregnancy, as a result of the increased gastric emptying time, and this physiologic phenomenon might explain the 3-hour stratification found in this report. Cheney et al. 7 studied obese and nonobese women with gestational diabetes after a standard 400 kcal mixed meal. They found significantly increased glucose and insulin levels for obese patients with gestational diabetes but decreased insulin levels in lean women with gestational diabetes when compared with normal subjects. They concluded that the obese patients with gestational diabetes had increased insulin resistance, whereas the lean patients with gestational diabetes had insulin deficiency. Catalano et al. 9 showed that this resistance was present even after delivery when obese subjects with gestational diabetes were tested with the hyperinsulinemic-euglycemic clamp. In our study the "fed" patients are similar to the post-mixed-meal patients of Cheney et al. In the abnormal group, glucose, insulin, and insulin/glucose index (i.e., insulin resistance) were increased as compared with values in normal gravid women. This was true for both obese and lean patients in this group. Thus the lean patients with minor glucose abnormalities in our study did not show any deficiency in insulin response. A number of investigators 10·11 have shown that insulin sensitivity declines when an individual becomes obese. Ryan et al. 15 demonstrated, by means of euglycemic glucose clamp techniques, that nondiabetic pregnant patients have increased insulin resistance over non-

Volume 163 Number 6, Part I

pregnant ones, and gestational-onset diabetic women have even greater resistance. The findings in this study of increased insulin and insulin / glucose index in obese patients in the "fasting" group are consistent with these findings. The relatively greater increase in insulin response of nonobese patients, when "fed," seemed to overshadow the effect of obesity, even in the group with glucose values 2:140 mg/dl, where both obese and lean patients have increased insulin output. Thus the "fed" state increased the insulin response to a greater extent than obesity altered insulin resistance in these mothers with minimal glucose abnormalities. In summary, when compared with patients who undergo a standard oral 50 gm glucose challenge test after a fasting interval of >3 hours, pregnant women who have eaten more recently ("fed" women) have a greater insulin response if the glucose screening level is 2: 140 mg/dl, whereas obese subjects have evidence of increased insulin resistance only when "fasting," regardless of glucose level. REFERENCES I. American College of Obstetricians and Gynecologists. Management of diabetes in pregnancy. Washington: ACOG, 1986; ACOG Tech Bull no 92. 2. O'Sullivan JB, Mahan CM, Dandrow RY. Screening criteria for high-risk gestational diabetic patients. AM J OBSTET GYNECOL 1973; 116:895-900. 3. National Diabetes Data Group. Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 1979;28: 1039-57. 4. Summary and recommendations of workshop-conference on gestational diabetes. Diabetes Care 1980;3:499-501.

Fasting interval and glucose challenge test

1817

5. Coustan DR, Wid ness JA, Carpenter MW, et al. Should the fifty-gram, one-hour plasma glucose screening test for gestational diabetes be administered in the fasting or fed state? AM J OBSTET GYNECOL 1986; 154: 1031-5. 6. Nuttall FQ, Mooradian AD, Gannon MC, Billington C, Krezowdki P. Effect of protein ingestion on the glucose and insulin response to a standardized oral glucose load. Diabetes Care 1984;7:465-70. 7. Cheney C, Shragg P, Hollingsworth D. Demonstration of heterogeneity in gestational diabetes by a 400-kcal breakfast meal tolerance test. Obstet Gynecol 1985;65: 17-23. 8. Haffner SM, Stern MP, Hazuda HP, Pugh JA, Patterson JK. Hyperinsulinemia in a population at high risk for noninsulin-dependent diabetes mellitus. N Engl J Med 1986; 315:220-4. 9. Catalano PM, Bernstein 1M, Wolfe RR, et al. Subclinical abnormalities of glucose metabolism in subjects with previous gestational diabetes. AM J OBSTET GYNECOL 1986; 1155: 1255-62. 10. DeFronzo RA, Sherwin RS, Hendler R, Fe!ig P. Insulin binding to monocytes and insulin action in human obesity, starvation, and refeeding. J Clin Invest 1978;62:204-13. II. Bonadonna R, Groop L, Kraemer N, DeFronzo RA. Obesity and insulin resistance in man: a dose response study. Metabolism [In press]. 12. Golay A, Felber JP, Jequier E, DeFronzo RA, Ferrannini E. Metabolic basis of obesity and non-insulin dependent diabetes mellitus. Diabetes Metab Rev 1988;4:727-47. 13. Kolterman OG, Inse! J, Saekow M, Olefsky JM. Mechanism of insulin resistance in human obesity-evidence for receptor and post-receptor defects. J Clin Invest 1980; 65: 1273-84. 14. Bogardas C, Lillioja S, Mott D, Reaven GR, Kashiwagi A, Foley J. Relationship between obesity and maximal insulin-stimulated glucose uptake in vivo and in vitro in Pima Indians. J Clin Invest 1984;73:800-5. 15. Ryan E, Brower E, O'Sullivan M, Skyler J. Insulin resistance in pregnancy. Studies with the euglycemic clamp technique. Diabetes 1985;34:380.