Hypoglycemia During the 100-g Oral Glucose Tolerance Test: Incidence and Perinatal Significance

During the 100-g Oral Glucose Tolerance Test: Incidence and Perinatal Significance Amir Weissman, MD, Ido Solt, MD, Moshe Zloczower, MD, and Peter Jakobi, MD OBJECTIVE: To estimate and report the incidence and perinatal significance of hypoglycemia during the 100-g oral glucose tolerance test in pregnant women. METHODS: Over a 3-year period, we analyzed the incidence and perinatal outcome of pregnant women who experienced hypoglycemia, defined as a plasma glucose level of 50 mg/dl or less while undergoing the 100-g oral glucose tolerance test. The study group included women who delivered singletons at term. Women who underwent the 100-g oral glucose tolerance test during the same period and had no hypoglycemia served as the control group. RESULTS: A total of 805 women were included in the study, which comprised 51 women (6.3%) who experienced hypoglycemia during the test and 754 women in the control group. Gestational diabetes mellitus was diagnosed in 5/51 (9.8%) women in the study group, compared with 216/754 (28.6%) women in the control group (P <.03), and the neonates born to these women had significantly lower birth weights. CONCLUSION: The incidence of reactive hypoglycemia during the 100-g oral glucose tolerance test in our population is 6.3%. Women who experience hypoglycemia during the test have a significantly lower incidence of gestational diabetes and neonatal birth weights. (Obstet Gynecol 2005; 105:1424 8. 2005 by The American College of Obstetricians and Gynecologists.) LEVEL OF EVIDENCE: III The 3-hour, 100-g oral glucose tolerance test (GTT) is the accepted method for diagnosing gestational diabetes mellitus (GDM) during pregnancy. 1,2 It is a common knowledge among health caregivers that a significant number of women experience symptoms of hypoglycemia during the test, including tachycardia, faintness, nausea, and perspiration. In some, very low blood glucose levels may be detected concomitantly. The finding of low blood glucose levels or symptomatic hypoglycemia during the test may cause anxiety and apprehension for both the woman and the medical personnel. However, we did not find in the literature or among the From the Department of Obstetrics and Gynecology, Rambam Medical Center, Haifa, Israel. experts who deal with diabetes answers to 2 simple questions: 1) What is the incidence of hypoglycemia during the oral GTT? and 2) Do women who react with hypoglycemia have different perinatal outcome than women without hypoglycemia? We hypothesized that women who react with hypoglycemia make up a unique group whose physiological characteristics may be different from women without hypoglycemia, and therefore their perinatal outcome may be different. There is no clear cutoff blood glucose level for experiencing hypoglycemic symptoms; some patients may exhibit a hypoglycemic reaction at a normal glucose range whereas others may be unaware of hypoglycemia even at very low blood glucose concentrations. 3,4 In various reports, 50 mg/dl was considered to be a suitable glucose concentration for the diagnosis of hypoglycemia. 3,5 7 The goal of the present study was to estimate the incidence of hypoglycemia (plasma glucose levels 50 mg/dl) during the 100-g oral GTT and to report its possible effect on perinatal outcome. MATERIALS AND METHODS We analyzed the perinatal data of all women who underwent the 3-hour oral GTT and delivered in our institution between December 1998 and December 2001. All these women had the 50-g glucose screening test at 24 28 weeks of gestation, and those with an abnormal test result ( 140 mg/dl) 2 underwent the 3-hour 100-g oral GTT. Inclusion criteria included singleton pregnancies and term deliveries ( 37 weeks). Exclusion criteria included a past history of gestational diabetes. The test was performed at 8 9 AM after an overnight fast of not less than 8 hours and after 3 days of unrestricted diet with at least 150 g of carbohydrates daily. Following withdrawal of a fasting plasma sample, a solution of 100 g glucose was mixed in 200 ml of water and swallowed in 10 minutes. Patients who could not swallow the solution or vomited afterward were excluded from the study. Venous blood samples were drawn at 60, 120, and 180 minutes. The plasma glucose samples were analyzed using the glucose oxidase method. The diagno- VOL. 105, NO. 6, JUNE 2005 1424 2005 by The American College of Obstetricians and Gynecologists. 0029-7844/05/$30.00 Published by Lippincott Williams & Wilkins. doi:10.1097/01.aog.0000159577.28448.f9

sis of GDM was made if 2 or more of the following venous plasma measurements were met or exceeded: fasting 95 mg/dl, 1 hour 180 mg/dl, 2 hours 155 mg/dl, 3 hours 140 mg/dl. 2 The study included 814 consecutive women who met the inclusion criteria. Nine women (1.1%) were excluded because their records were not complete and not all relevant data could be retrieved. Therefore, 805 women were included in the final analyses. The study group comprised women who were found to have plasma glucose levels of 50 mg/dl or less during the test (with or without clinical symptoms). All the other women served as controls. We chose a plasma glucose cutoff level of 50 mg/dl for the definition of hypoglycemia for the following reasons: 1) Many nonpregnant patients manifest at least one of the common symptoms of hypoglycemia (tachycardia, sweating, tremor, dizziness, headache, and faintness) at a plasma glucose level of 50 mg/dl or below. 3,4 2) Although there are no accepted cutoff values for the definition of hypoglycemia and different cutoff levels were reported, 50 mg/dl was considered by several authors to be a suitable glucose concentration for the diagnosis of hypoglycemia. 3,5 7 3) The 50-mg/dL plasma glucose level was found to be below the 10th percentile of our study population during the oral GTT. The gestational age, birth weight, mode of delivery, and the indications for cesarean delivery for each patient were retrieved from the labor room records. Large for gestational age (LGA) was defined as birth weight greater than the 90th percentile, and small for gestational age (SGA) was defined as birth weight less than the 10th percentile. Macrosomia was defined as an estimated fetal weight of 4,000 g or more. 8,9 Indications for cesarean delivery or failure to progress during labor and/or macrosomia were analyzed separately as being among the common complications of gestational diabetes. Ultrasound examination is performed on all parturients at admission to our delivery ward. In accordance with the guidelines of the Israel Society of Obstetricians and Gynecologists, we propose elective cesarean delivery to women with a diagnosis of GDM and an estimated fetal weight of 4,000 g or more and to women without GDM but with an estimated fetal weight of 4,500 g or more. The Student t test was used for continuous variables and 2 test with Yates correction for discrete variables. The Fisher exact test was used where appropriate. Multiple logistic regression analyses were performed to estimate the relationship between hypoglycemia and the perinatal variables that were studied. Significance was set at P.05. The Statgraphics 5.1 (Manugistics, Rockville, MD) statistical software was used in this study. Table 1. Plasma Glucose Levels in Study and Control Groups During 50-g and 100-g Oral Glucose Tolerance Tests (n 51) No (n 754) 50-g Glucose screening 159 19.1 162 19.2 test (mg/dl) 100-g Glucose tolerance test (mg/dl) Fasting 80.2 9.1 87.4 31.9 1 hour 162.5 35.7 166.4 37.3 2 hours 113.4 27.7 137.4 35.9* 3 hours 44.0 4.7 96.1 27.8 Data are presented as mean standard deviation. * P.05. P.001. RESULTS A total of 805 women were included in the final analyses. Table 1 presents the blood glucose levels in the study and control groups during the 50-g and 100-g oral GTT. The plasma glucose values during the 50-g glucose screening test were similar in both groups (159 19.1 versus 162 19.2 mg/dl). As can be also seen, the blood glucose levels are very similar at fasting and at 1 hour of the 100-g oral GTT. However, a significant difference is present at 2 hours, followed by an even higher difference at 3 hours. Plasma glucose values of 50 mg/dl or less were found in 51 (6.3%) patients (mean blood glucose level 44 mg/dl, 95% confidence interval 42.7 45.3). In all patients, the nadir in plasma glucose levels occurred 3 hours after ingestion of the glucose solution, and none had a fasting 1- or 2-hour plasma glucose level of 50 mg/dl or less. The perinatal data are summarized in Table 2. Despite the similar plasma glucose values during the 50-g glucose screening test, women with hypoglycemia during the 3-hour, 100-g oral GTT had a significantly lower rate of GDM than did patients who did not have hypoglycemia (9.8% versus 28.6%, P.03). Although the gestational age at delivery was similar in both groups, the newborn birth weights were significantly lower in the study group than in the control group. There were no differences in the rates of SGA infants between the 2 groups, but there were fewer LGA infants in the study group. There were twice as many cesarean deliveries performed in the control group because of macrosomic infants (2% versus 5.4%), but this difference did not reach statistical significance because of the small size of the study group. We further analyzed our cohort according to the results of the oral GTT. Table 3 presents the data for the women in whom GDM was excluded by the oral GTT. Women who reacted with hypoglycemia during the test had a VOL. 105, NO. 6, JUNE 2005 Weissman et al on Oral Glucose Tolerance Test 1425

Table 2. Perinatal Data of the Study and Control Groups (n 51) No (n 754) Maternal age (y) 29.9 4.6 30.9 5.2 Parity 1.96 0.9 2.35 1.4 Diagnosis of GDM 5 (9.8)* 216 (28.6)* Gestational age at 39.4 1.3 39.5 1.2 delivery (wk) Birth weight (g) 3,281 557 3,434 474 LGA infants 3 (5.9) 77 (10.2) ( 90th percentile) SGA infants 6 (11.8) 78 (10.3) ( 10th percentile) Macrosomia 3 (5.9) 89 (11.8) ( 4,000 g) Cesarean deliveries 10 (20) 203 (26.9) (total) Cesarean deliveries for macrosomia 1 (2) 41 (5.4) GDM, gestation diabetes mellitus; LGA, large for gestational age; SGA, small for gestational age. Data are presented as mean standard deviation or n (%). * P.03. P.02. lower rate of LGA infants and fewer cesarean deliveries, although this difference did not reach statistical significance. The newborns were lighter by 100 g on average in women who reacted with hypoglycemia compared with the controls. The study group with hypoglycemia during the test and with the diagnosis of GDM was very small (only 5 women), so statistical analyses were not performed and their data are not given. Table 3. Perinatal Data of Women for Whom Gestational Diabetes Mellitus Was Excluded (n 46) No (n 538) Maternal age (y) 30.4 5.2 29.7 4.7 Parity 1.98 0.98 2.3 1.3 Gestational age at 39.4 1.3 39.6 1.2 delivery (wk) Birth weight (g) 3,309 569 3,422 458 LGA infants 3 (6.5) 59 (11) ( 90th percentile) SGA infants 5 (10.8) 47 (8.7) ( 10th percentile) Macrosomia 3 (6.5) 61 (11.3) ( 4,000 g) Cesarean deliveries 8 (17.4) 136 (25.3) (total) Cesarean deliveries for macrosomia 1 (2.2) 28 (5.2) LGA, large for gestational age; SGA, small for gestational age. Data are presented as mean standard deviation or n (%). There were no significant differences between the study and control groups in any of the variables. To evaluate the perinatal significance of hypoglycemia, multiple logistic regression analyses were performed to describe the relationship between hypoglycemia and the different perinatal variables. The analysis showed that the infant s weight and the diagnosis of GDM were significantly associated with hypoglycemia (P.03 and P.005, respectively). DISCUSSION The occurrence of hypoglycemia and occasional hypoglycemic reaction during the 3-hour oral GTT is a well-known phenomenon. However, its incidence and possible effect on perinatal outcome have not been directly addressed. In the curves presented in the original report by O Sullivan and Mahan 10 (although numerical values were not given in a tabular form), glucose values below 50 mg/dl in the third hour can be deduced in approximately 25% of the patients. Hypoglycemic values can also be assumed from the study of Abell and Beisher, 11 who reported that, during a 50-g oral GTT, the fifth percentile for the 3-hour capillary glucose value was 55 mg/dl. In that study, glucose values below the fifth percentile at 3 hours were also associated with SGA infants. In our study population, we found an incidence of 6.3% of hypoglycemia during the 100-g oral GTT. All hypoglycemic events occurred 3 hours after the glucose ingestion, and there were no cases of fasting hypoglycemia (after fasting of at least 8 hours). Hypoglycemic episodes are more common during pregnancy because of the physiological changes that take place: basal insulin levels are increased, while glucagon release is suppressed by estrogen, progesterone, human placental lactogen, and probably other mediators. 12 The result is that pregnant women may experience postprandial hypoglycemia more often. Some pregnant women may be more prone to this condition and may react with an exaggerated response. There is no clear cutoff blood glucose level for experiencing hypoglycemic symptoms. Although some patients may exhibit a hypoglycemic reaction even at a normal glucose range, others may be unaware of hypoglycemia even with very low blood glucose concentrations. 3 The association between glucose intolerance during pregnancy and increased birth weight with its related obstetric complications is well documented. 1,13 Even women with only one abnormal value in the oral GTT or women with a positive 50-g screening test and a negative oral GTT may have a higher incidence of macrosomic infants. 13,14 By analogy, frequently occurring low blood glucose levels, as encountered in women with reactive postprandial hypoglycemia, may inversely affect fetal overgrowth. This may occur in pregnant 1426 Weissman et al on Oral Glucose Tolerance Test OBSTETRICS & GYNECOLOGY

women whose blood glucose levels are repeatedly lowered after meals, even without symptomatic hypoglycemic episodes because of pancreatic oversecretion of insulin. We hypothesize that such a phenomenon may occur more frequently in women who experience hypoglycemia during the oral GTT, probably an indication of pancreatic sensitivity or hyperreactivity to large glucose load. However, it is important to note that the results of the 50-g screening test glucose values were similar in both the study and control groups (Table 1) and offered no prognostic assistance in determining whose GTT test would be abnormal and predicting the eventual perinatal outcome. Even during the 100-g oral GTT, both groups began with similar fasting and 1-hour blood glucose levels (Table 1). However, from that time point, the 2 groups deviate significantly. After a lag time during which the absorption of glucose takes place and insulin secretion begins, the women in the study group start to undergo a significant decrease in blood glucose levels, probably as a reaction to an overshoot of insulin secretion. Our results show that women who have hypoglycemia during the 100-g oral GTT have a significantly lower rate of GDM than expected, compared with the expected GDM rate in the control group, despite the similar 50-g glucose screening values. The newborns birth weights in women with hypoglycemia were significantly lower than in the control group although the gestational ages at deliveries were similar. This difference was not due to a higher rate of SGA, as suggested by Abel and Beischer, 11 but to a lower rate of LGA infants. The cesarean delivery rate for indication of macrosomia was lower in the study group, but because of the very small number of women who reacted with hypoglycemia and had macrosomic infants, statistical analyses were flawed by a error. These results are of interest because they were found in women whose 50-g challenge test was abnormal and were therefore at higher risk for GDM and macrosomia. All of these findings indicate that carbohydrate intolerance in pregnant women seems to behave as a sliding scale, which may significantly affect fetal growth in a progressive manner and not merely as a binary disease process, ie, you either have it and are at risk or you do not have it and are fine. The limitation of our study is that it is retrospective. Thus, some variables that could potentially affect fetal growth (obesity, weight gain during pregnancy), could not be retrieved. However, the relatively large numbers of our cohort may partly compensate for that drawback. Even with the large cohort of women we included, the study group ended up relatively small, 4 times smaller then we had expected and calculated from the data in O Sullivan s study. 10 Increasing the cohort of women 4-fold would result in a 12-year survey and, if it were performed at one center, would amount to a time period during which major changes in obstetric practice could occur, thereby rendering such a survey impractical. Concerns about short-term maternal hypoglycemic events in gestational diabetic women under treatment were recently raised. 7 Previous studies in both animals and humans suggested that relative maternal hypoglycemia is associated with fetal growth restriction. 15 18 Our study shows that 6.3% of women react with hypoglycemia after a 100-g oral GTT. These women may do so repeatedly after heavy meals. However, their perinatal outcome was excellent, and even though all were at high risk for macrosomia, there was actually a lower incidence of macrosomia in that group without an increase in the rate of SGA infants. Experiencing a hypoglycemic episode while undergoing the 3-hour oral GTT may cause anxiety and be upsetting for both medical personal and the pregnant women. Based on our study, however, the patient can be reassured that such a phenomenon is not unusual, is transitory, and carries a favorable prognosis in terms of obstetric outcome. This group of women is expected to have a lower incidence of GDM, lower birth weights, and a lower rate of cesarean deliveries for macrosomia during labor. A large multicentric study could provide us answers about the implications of various glucose blood levels for perinatal outcome and may even lead us to revise our approaches to glucose screening during pregnancy. REFERENCES 1. Coustan DR, Carpenter MW. The diagnosis of gestational diabetes. Diabetes Care 1998;21(suppl 2):B5 8. 2. Metzger BE, Coustan DR. Summary and recommendations of the Fourth International Workshop Conference on Gestational Diabetes Mellitus. Diabetes Care 1998; 21(suppl 2):B161 7. 3. Unawareness of hypoglycemia editorial. N Engl J Med 1995;333:1771 2. 4. Palardy J, Havrankova J, Lepage R, Matte R, Belanger R, D Amour P, et al. Blood glucose measurements during symptomatic episodes in patients with suspected postprandial hypoglycemia. N Engl J Med 1989;321:1421 5. 5. Field JB. : definition, clinical presentations, classification, and laboratory tests. Endocrinol Metab Clin North Am 1989;18:27 43. 6. Brun JF, Fedou C, Mercier J. Postprandial reactive hypoglycemia. Diabetes Metab 2000;26:337 51. 7. Yogev Y, Ben-Haroush A, Chen R, Rosenn B, Hod M, Langer O. Undiagnosed asymptomatic hypoglycemia: diet, insulin, and glyburide for gestational diabetic pregnancy. Obstet Gynecol 2004;104:88 93. VOL. 105, NO. 6, JUNE 2005 Weissman et al on Oral Glucose Tolerance Test 1427

8. Schwartz R, Teramo KA. What is the significance of macrosomia? Diabetes Care 1999;22:1201 5. 9. Sanchez-Ramos L, Bernstein S, Kaunitz AM. Expectant management versus labor induction for suspected fetal macrosomia: a systematic review. Obstet Gynecol 2002; 100:997 1002. 10. O Sullivan JB, Mahan CM. Criteria for the oral glucose tolerance test in pregnancy. Diabetes 1964;13:278 85. 11. Abell DA, Beischer NA. Evaluation of the three-hour oral glucose tolerance test in detection of significant hyperglycemia and hypoglycemia in pregnancy. Diabetes 1975;24: 874 80. 12. Phelps RL, Metzger BE, Freinkel N. Carbohydrate metabolism in pregnancy: diurnal profiles of blood glucose, insulin, free fatty acids, triglycerides, cholesterol, and individual amino acids in late normal pregnancy. Am J Obstet Gynecol 1981;140:730 6. 13. Brody SC, Harris R, Lohr K. Screening for gestational diabetes: a summary of the evidence for the U. S. Preventive Services Task Force. Obstet Gynecol 2003;101: 380 92. 14. Berkus MD, Langer O. Glucose tolerance test: degree of glucose abnormality correlates with neonatal outcome. Obstet Gynecol 1993;81:344 8. 15. Bevier WC, Fischer R, Jovanovic L. Treatment of women with an abnormal glucose challenge test (but a normal oral glucose tolerance test) decreases the prevalence of macrosomia. Am J Perinatol 1999;16:269 75. 16. Owens JA, Falconer J, Robinson JS. Glucose metabolism in pregnant sheep when placental growth is restricted. Am J Physiol 1989;257(suppl):R350 7. 17. Langer O, Damus K, Maiman M, Divon M, Levy J, Bauman W. A link between relative hypoglycemia-hypoinsulinemia during oral glucose tolerance tests and intrauterine growth retardation. Am J Obstet Gynecol 1986; 155:711 6. 18. Khouzami VA, Ginsburg DS, Daikoku NH, Johnson JW. The glucose tolerance test as a means of identifying intrauterine growth retardation. Am J Obstet Gynecol 1981; 139:423 6. Address reprint requests to: Dr. A. Weissman, Department of Obstetrics and Gynecology, Rambam Medical Center, POB 9602, Haifa 31096 Israel; e-mail: wamir@netvision.net.il. Received November 12, 2004. Received in revised form January 25, 2005. Accepted February 3, 2005. 1428 Weissman et al on Oral Glucose Tolerance Test OBSTETRICS & GYNECOLOGY