The relationship between carbohydrate intake and glucose tolerance in pregnant women

Transcription

1 Acta Obstet Gynecol Scand 2003: 82: Copyright # Acta Obstet Gynecol Scand 2003 Printed in Denmark. All rights reserved Acta Obstetricia et Gynecologica Scandinavica ISSN ORIGINAL ARTICLE The relationship between carbohydrate intake and glucose tolerance in pregnant women MOTOI TAKIZAWA 1,TAKASHI KANEKO 2,KEIKO KOHNO 1,YUKIHITO FUKADA 1 AND KAZUHIKO HOSHI 1 From the Departments of 1 Obstetrics and Gynecology and 2 Environmental Health, Yamanashi Medical University, Yamanashi, Japan Acta Obstet Gynecol Scand 2003; 82: # Acta Obstet Gynecol Scand Objective. We verified whether a misdiagnosis of gestational diabetes mellitus can result in pregnant women when glucose tolerance has deteriorated after a low-carbohydrate meal, and tried to elucidate the mechanism behind the different outcome of the test. Study design. Twenty-seven pregnant women were given directions for their evening meal the day before each of two 75-g oral glucose tolerance tests (OGTT). The evening meal was either a low-carbohydrate meal (carbohydrate, 6.7%; Low), or a high-carbohydrate meal (carbohydrate, 85.7%; High). Results. The OGTT showed that the glucose tolerance was significantly impaired after Low than after High, with a significant increase of fasting plasma non-esterified fatty acids (NEFA) level. Moreover, the insulinogenic index (I-I) after High significantly decreased than that after Low. Conclusions. The present data suggests that there is a risk of misdiagnosis of impaired glucose tolerance with only one intake of this extremely low-carbohydrate meal on the evening before testing. The decrease of insulin secretion and the activation of glucose-fatty acid cycle may be considered as the mechanism. Key words: 75-g oral glucose tolerance test; carbohydrate intake; gestational diabetes mellitus; glucose-fatty acid (Randle) cycle; insulinogenic index Submitted 29 July, 2002 Accepted 27 January, 2003 It is well known that glucose tolerance deteriorates during pregnancy. This impairment of glucose tolerance with pregnancy is called gestational diabetes mellitus (GDM), an intriguing pathological state that shows specificity for glucose metabolism in pregnant women. The incidence rate of GDM in Japan is reported to be about % (1). In Abbreviations: GDM: gestational diabetes mellitus; DM: diabetes mellitus; OGTT: oral glucose tolerance test; GCT: glucose challenge test; NEFA: non-esterified fatty acids; BMI: body mass index; g-auc: area under the plasma glucose concentration-time curve; Low: low-carbohydrate meal; High: high-carbohydrate meal;. HOMA: homeostasis model assessment; I-I: insulinogenic index. general, the impairment of glucose tolerance is less severe than that in pregnant women with comorbid diabetes mellitus (DM), but because of the risk of fetal and neonatal abnormalities [e.g. fetal and neonatal death, fetal macrosomia (1), and respiratory distress syndrome], similar to those for pregnant women with DM, early measures are essential for the diagnosis of GDM in the clinical setting. Additionally, given the high probability of future progression to DM for women with GDM, great importance is also attached to early diagnosis from the perspective of predicting the development of DM. The diagnostic technique that is widely used in the field of obstetrics is the oral glucose tolerance

2 Carbohydrate intake and glucose tolerance 1081 test (OGTT). The American Diabetes Association recommends the following procedure: screen pregnant women with a 50-g glucose challenge test (GCT) at weeks of gestation, and select those whose plasma glucose level at 60 min after loading exceeds 140 mg/dl (¼7.7 mmol/l) (2). Make a definitive diagnosis for patients yielding a positive result with a 100-g OGTT. The Japan Society of Obstetrics and Gynecology also recommends the 50-g GCT as a screening method, and the 75-g OGTT to confirm the diagnosis in patients producing a positive result at screening (3,4). Because patients are typically directed to fast from 21:00 hours the evening before the day of the OGTT, they tend to make do with a light evening meal on the day before the test (for example, a garden salad). However, it is important to ensure a sufficient carbohydrate intake from the evening meal the day before the test. If not, the test results may be assessed as indicating diabetes or impaired glucose tolerance (5). It has also been reported that the plasma concentration of non-esterified fatty acids (NEFA) specifically increases after low-carbohydrate meals (6). In the present work, we sought to verify whether a misdiagnosis of GDM can result in pregnant women when glucose tolerance has deteriorated after a low-carbohydrate meal, and tried to elucidate the mechanism behind the different outcome of the test. Materials and methods Subjects We recruited the volunteers from the pregnant women at weeks of gestation who visited the Department of Obstetrics and Gynecology of Yamanashi Medical University as outpatients from 1999 to The subjects for the tests were sampled from among pregnant women in whom DM was not identified as a complication before pregnancy. The objectives of the research were fully explained to 27 Japanese pregnant women [age, years; body mass index (BMI), kg/m 2 : given in means SD] who gave their consent to participate in the study. They were given directions for their evening meal the day before each of two 75-g OGTT. The tests were performed 5 14 days apart. The present study was approved by the university ethics committee. Diet description and 75-g oral glucose tolerance test After their usual meals at breakfast and lunchtime on the day before testing, subjects were directed to adhere rigorously to a prescribed menu for the evening meal, ensuring they ate between 18:30 and 20:00 hours. Calorific intake was forbidden from 22:00 hours onwards. On the day of testing, a 75-g OGTT was performed from 09:00 hours. First, fasting blood samples were withdrawn (before glucose loading), then subjects were given a drink containing 75 g of carbohydrate to be taken over a period of 5 min. Venous blood samples were then successively withdrawn at 30, 60 and 120 min for measurements of plasma glucose levels. The concentrations of insulin and c-peptide in serum were measured before loading, and at 30 and 120 min. The plasma concentrations of triglycerides (TG) and NEFA were measured in preloading blood samples. Glucose tolerance was evaluated according to WHO categories: normal, impaired glucose tolerance, or diabetic (7). The area under the plasma glucose concentration-time curve (g-auc) was determined by the trapezoidal rule. The insulinogenic index (I-I) (increment in serum insulin concentration for 30 min/increment in plasma glucose concentration for 30 min after loading) was calculated to evaluate the insulin secretion response (8). The homeostasis model assessment (HOMA) index was used as an indicator of insulin sensitivity (9). The evening meal was either a lowcarbohydrate meal of 505 kcal, of which the carbohydrate content was 6.7% (Low), or a high-carbohydrate meal of 518 kcal, of which the carbohydrate content was 85.7% (High). On the day of the test, the investigator confirmed that the test meal had been eaten. All subjects ate the meals as directed. Throughout the testing period, vigorous exercise was restricted. No subjects complained of poor physical condition during the testing period. Subjects were randomized to either the low-carbohydrate meal on the day before the first test then the high-carbohydrate meal on the day before the second test (Low/ High), or vice versa (High/Low). Briefly, in accordance with a randomized cross-over design, 13 subjects were assigned to the Low/High testing sequence and 14 to the High/Low sequence. There were no differences between the Low/ High group and High/Low group with respect to the characteristics of age, parity, BMI, gestational weeks of first test, or interval between testing (Table I). Statistical analysis Data were presented as means SD for 27 subjects, and the data were statistically analyzed by two-way ANOVA with a computer software

3 1082 T. Kaneko et al. Table I. Characteristics of subjects Number of patients Age (years) Parity (times) BMI (kg/m 2 ) Gestational weeks of 1st test Interval days between testing Overall Low/High High/Low Plus/minus values are means SD. (StatView 5.0, Abacus Co., Berkeley, California, USA), and between-group comparisons were performed using Student s paired t-test. Differences were regarded as significant when the p-level was below Results The results obtained from 75-g OGTT after highcarbohydrate (High) and low-carbohydrate evening meals (Low) the day before testing are presented below. Before glucose loading, the mean plasma glucose concentration was mmol/l after the High meal and mmol/l after the Low meal. At 60 min, the concentration was mmol/l after the High meal and mmol/l after the Low meal. At 120 min, the concentration was mmol/l after the High meal and mmol/l after the Low meal. The paired t-test was used to analyze plasma glucose levels before and at 60 and 120 min after glucose loading. Before loading, the plasma glucose level after the Low meal was significantly lower than the corresponding level after the High meal. At 60 min after loading, the plasma glucose level after the Low meal was significantly higher than that after the High meal. At 120 min after loading, the plasma glucose level after the Low meal was significantly higher than that after the High meal (Fig. 1). Data on the mean serum insulin concentration before loading and mean serum insulin concentration at 30 min are shown in Table II. Statistical analysis of the serum insulin levels measured after the High meal and after the Low meal disclosed no significant differences between the values before loading or at 30 min (before loading, p ¼ 0.050; 30 min, p ¼ 0.079). Plasma glucose * High CH Low CH ** Before 60 (time, min) ** 120 * p < 0.05 ** p < 0.01 Fig. 1. Plasma glucose responses after different evening meals (a high-carbohydrate meal, High CH and a low-carbohydrate meal, Low CH) on the day before the 75-g oral glucose tolerance test (OGTT). Table II. Plasma glucose and serum insulin levels before and at 30 min after glucose loading, insulinogenic index, fasting plasma non-esterified fatty acids levels, and homeostasis model assessment index, after high-carbohydrate (High CH) or low-carbohydrate meals (Low CH) Glucose before 30 min Insulin before 30 min (pmol/l) I-I (pmol/mmol) NEFA (meq/l) HOMA High CH Low CH * NS * NS NS I-I: insulinogenic index; NEFA: non-esterified fatty acid; HOMA: homeostasis model assessment. NS: not significant. *p < 0.05; p < Plus/minus values are means SD.

4 Carbohydrate intake and glucose tolerance 1083 Investigation of the HOMA index disclosed no significant differences between the two groups (p ¼ 0.730). We also investigated the mean serum insulin concentration and change in insulin concentration relative to the change in plasma glucose level at 30 min, i.e. I-I. The mean I-I was 146 after the High meal, significantly different from the figure of 58.8 after the Low meal. In brief, these data suggested that the insulin secretion response to the increased plasma glucose level was lower after the Low meal than after the High meal. To evaluate the causal factors behind the meal effects, we also measured the plasma concentrations of TG and NEFA before glucose loading. The TG concentrations were not significantly different between the two groups. On the other hand, the NEFA concentration after the High meal was significantly lower than that after the Low meal (Tables II and III). In 5 of the 27 subjects tested, the OGTT findings were in concordance with the diagnostic criteria for GDM (Japan Society of Obstetrics and Gynecology) after the Low meal, but lay outside the diagnostic limits after the High meal (Fig. 2). To analyze any potential difference in findings due to increased duration of pregnancy between the first and second OGTT, we also stratified the results for the individual meal sequences (Table III). The fact that plasma glucose levels were low before loading and increased after loading in the Low group, and that I-I and NEFA concentrations in the Low group were lower than those in the High group, indicated that reversing the sequence of low- and high-carbohydrate meals had no effect on outcome. Discussion In research in which predetermined meals were fed to healthy young males for 7 days, Himsworth concluded that the only dietary factor that determined glucose tolerance and insulin sensitivity in healthy individuals was carbohydrate intake, and that carbohydrate intake ratio, lipid intake and lipid intake ratio had no effect (10 12). Furthermore, since Wilkerson et al. reported the absence of a significant relationship between carbohydrate intake and glucose tolerance in healthy adults (13), the importance of carbohydrate intake the day before glucose tolerance testing has not been greatly emphasized. To shed further light on this question, Kaneko et al. re-evaluated the association between carbohydrate intake the day before OGTT and glucose tolerance in healthy adult volunteers (5,6). They found that all subjects in the high-carbohydrate diet group exhibited normal glucose tolerance, but those in the low-carbohydrate diet group had significantly worse glucose tolerance. g-auc in the low-carbohydrate diet group was significantly higher and the I-I was significantly lower than that in the high-carbohydrate diet group. Furthermore, the fasting plasma concentration of NEFA in the low-carbohydrate diet group was significantly higher. These findings led to the suggestion that the deterioration in glucose tolerance due to lowcarbohydrate intake may have been caused by decreased insulin secretion (I-I) and activation of the Randle (glucose-fatty acid) cycle. The fact that there was also a significant correlation between fasting plasma NEFA concentration and g-auc suggested that the increase in fasting plasma NEFA concentration might serve as an indicator for predicting deterioration in glucose Table III. OGTT findings; plasma glucose levels before and at 60 and 120 min after glucose loading, insulinogenic index (I-I), and fasting plasma non-esterified fatty acids levels (NEFA) Low/High sequence Low/High (13 subjects) Before 60 min 120 min I-I (pmol/mmol) NEFA (meq/l) High Low p-value High/Low sequence High/Low (14 subjects) Before 60 min 120 min I-I (pmol/mmol) NEFA (meq/l) High Low p-value Plus/minus values are means SD.

5 1084 T. Kaneko et al. Fig. 2. Five subjects whose 75-g oral glucose tolerance test (OGTT) findings were outside the diagnostic criteria of gestational diabetes mellitus (GDM) after a high-carbohydrate meal (High CH), despite conforming to the diagnostic criteria of GDM after a low-carbohydrate meal (Low CH). tolerance. This finding led to the hypothesis that persons who eat a low-carbohydrate meal before undergoing OGTT might be misdiagnosed with impaired glucose tolerance (5,6). The most significant change caused by intake of a low-carbohydrate meal is the increase in fasting plasma NEFA concentration. Randle et al. discovered the existence of a link between carbohydrate and lipid metabolism, naming it the Randle cycle (14,15), which has become a new focus of worldwide attention. The findings of our research can also be explained by these hypotheses. Many testing centers encourage the intake of meals containing at least 150 g/day of carbohydrates for 3 days before OGTT, because they believe that such preparatory meals will reduce false GDM positivity. Some researchers claim that restrictions on carbohydrate intake are not required to address this problem. In a prospective study, Salmon et al. found no significant differences with respect to mean plasma glucose concentration between groups of patients that ate preparatory meals and those that did not, with no differences seen in the number of subjects with OGTT abnormalities (16). They concluded that dietary prescriptions would unnecessarily delay the diagnosis of GDM, given that the OGTT findings were not significantly altered in healthy pregnant women. Major et al. investigated the effects of restrictions on carbohydrate intake (one group with a carbohydrate content of less than 42% and the other with a content greater than 45%) on perinatal prognosis for GDM patients receiving dietary therapy (17). They found that restricting carbohydrate intake in GDM patients led to improved control of plasma glucose. They also reported that the number of indications for insulin therapy diminished, and that the incidence rate of large for gestational age, and cesarean sections due to cephalopelvic disproportion or fetal macrosomia were reduced. Peterson et al. conducted a cross-over study of dietary content (carbohydrate ratio) and metabolic response in obese women whose body weight had been reduced by diet therapy, and found that low-carbohydrate, high-fat diet therapy produces an excellent lipid profile in obese women (18). At the same time, these researchers also investigated the effects of a past history of GDM. Their results suggest that maximum OGTT values, fasting insulin levels and insulin resistance are higher in obese women with a history of GDM, compared with findings for obese women without such a history. Entrekin et al. investigated the effects of high-carbohydrate preparatory meals on 3-h OGTT in pregnant women (19). In an investigation of pregnant women with abnormal findings on plasma glucose screening tests, they analyzed the subjects in three groups, according to the carbohydrate content of meals eaten before testing. Given that the clinical outcomes were the same for fasting plasma glucose, and the 1, 2, and 3-h levels in the three groups, it was concluded that high-carbohydrate meals had no greater effects than normal meals in pregnant women undergoing OGTT. In all the above reports in which it was concluded that there is no need to restrict carbohydrate intake, the carbohydrate content of the meals was about 40 50%. The low-carbohydrate meal used in the present study was 10% carbohydrate. The present data suggests that there is a risk of misdiagnosis of impaired glucose tolerance with only one intake of this extremely lowcarbohydrate meal on the evening before testing.

6 Carbohydrate intake and glucose tolerance 1085 The present finding in women with normal OGTT is applied to women in the lower range of glucose intolerance, which was demonstrated in an additional experiment (data not shown, personal communication). References 1. Sato K, Fujimoto S. [Evidence-Based Medicine.] (in Japanese). In: Sato K, Fujimoto S (eds). Tokyo: Medical View Corporation, 1999: Stephen R, Carr MD. Screenig for gestational diabetes mellitus. Diabetes Care 1988; 21: B Japan Society of Obstetrics and Gynecology, Subcommittee to diagnose or control the gestational diabetes mellitus in Japan. [Report by a committee of nutrition and metabolism: a suggestion to control the gestational diabetes mellitus and the pregnancy with diabetes mellitus.] (in Japanese with an English Abstract) Acta Obstet Gynaecol Jpn 1985; 37: Japan Society of Obstetrics and Gynecology, A committee of the perinatal medicine in Japan. [A report about the gestational diabetes mellitus (GDM).] (in Japanese with an English Abstract) Acta Obstet Gynaecol Jpn 1995; 47: Kaneko T, Wang P-Y, Tawata M, Sato A. Low carbohydrate intake before oral glucose-tolerance tests. [Res Letter] Lancet 1998; 352: Wang P-Y, Kaneko T, Wang Y, Tawata M, Sato A. A low carbohydrate intake before an oral glucose tolerance test impairs the glucose tolerance of normal adults. Tohoku J Exp Med 1999; 189: Alberti KGMM. The diagnosis and classification of diabetes mellitus. Diabetes Voice 1999; 44: Seltzer HS, Allen EW, Herton AL, Brennan MT. Insulin secretion in response to glycemic stimulus: relation of delayed initial release to carbohydrate intolerance in mild diabetes mellitus. J Clin Invest 1967; 46: Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and b-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985; 28: Himsworth HP. Dietetic factor determining glucose tolerance and sensitivity to insulin of healthy men. Clin Sci 1935; 2: Himsworth HP, Marshall EM. The diet of diabetics prior to the onset of the disease. Clin Sci 1935; 2: Himsworth HP. Diet and the incidence of diabetes. Clin Sci 1935; 2: Wilkerson HLC, Hyman H, Kaufman M, McCuistion AC, Francis JO S. Diagnostic evaluation of oral glucose tolerance tests in nondiabetic subjects after various levels of carbohydrate intake. N Engl J Med 1960; 262: Randle PJ, Garland PB, Hales CN, Newsholm EA. The glucose fatty acid cycle. Its role in insulin sensitivity and the metabolic disturbances of diabetes mellitus. Lancet 1963; I: Sugden MC, Sharples SC, Randle PJ. Carcass glycogen as a potential source of glucose during short-term starvation. Biochem J 1976; 160: Salmon MC. Oral glucose tolerance test and preparatory diet. Am J Obstet Gynaecol 2000; 182: Major CA, Henry MJ, DeVeciana M, Morgan MA. The effect of carbohydrate restriction in patients with dietcontrolled gestational diabetes. Obstet Gynaecol 1998; 91: Peterson CM, Jovanovic-Peterson L. Randomized crossover study of 40% vs. 55% carbohydrate weight loss strategies in women with previous gestational diabetes mellitus and non-diabetic women of % ideal body weight. J Am Coll Nutr 1995; 14: Entrekin K, Work B, Owen J. Does a high carbohydrate preparatory diet affect the 3-hour oral glucose tolerance test in pregnancy? J Maternal-Fetal Med 1998; 7: Address for correspondence: Takashi Kaneko Department of Environmental Health Yamanashi Medical University Shimokato 1110 Tamaho Yamanashi Japan tkaneko@y-eisei