- Email: [email protected]
The postprandial glucose profile in the diabetic pregnancy
American Journal of Obstetrics and Gynecology (2004) 191, 576e81
www.elsevier.com/locate/ajog
The postprandial glucose profile in the diabetic pregnancy Avi Ben-Haroush, MD,a,* Yariv Yogev, MD,a,b Rony Chen, MD,a Barak Rosenn, MD,b Moshe Hod, MD,a Oded Langer, MDb Department of Obstetrics and Gynecology, Rabin Medical Center, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel a; Department of Obstetrics and Gynecology, St. Luke’s-Roosevelt Hospital Center, New York, University Hospital of Columbia University, New York, NYb Received for publication November 5, 2003; revised January 13, 2004; accepted January 29, 2004
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– KEY WORDS Gestational diabetes mellitus Postprandial peak Glucose
Objective: A controversy exists regarding the time to monitor blood glucose in the diabetic pregnancy (60 or 120 minutes after meals). Using a novel approach that provides continuous measurement of blood glucose, we sought to determine postprandial glucose profile in the diabetic pregnancy. Study design: Subjects were connected to a continuous glucose monitoring system for 72 consecutive hours. A continuous glucose monitoring system measures the interstitial glucose levels in subcutaneous tissue every 5 minutes. Women were instructed to record the time of each meal during the study period. For each meal, the first 240 minutes were analyzed. Results: Sixty-five women participated in the study: 26 women were treated by diet alone; 19 women received insulin therapy, and 20 women had type 1 diabetes mellitus. The time interval from meal to peak postprandial glucose levels was similar in all the evaluated types of diabetic pregnancies and in good and poor control insulin-treated patients with gestational diabetes mellitus (approximately 90 minutes). Failure to return to preprandial glucose values within a 3-hour observation period was identified in approximately 50% of the patients. A similar postprandial glucose peak time was obtained for breakfast, lunch, and dinner in all study groups. Postprandial hypoglycemia events were noted in approximately 10% of the meals and occurred about 160 minutes after mealtime. Conclusion: The time interval for postprandial glucose peak in diabetic pregnancies is approximately 90 minutes after meals throughout the day and is not affected by the level of glycemic control. This information should be considered in the treatment of diabetes mellitus in pregnancy. Ó 2004 Elsevier Inc. All rights reserved.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
* Reprint requests: Avi Ben-Haroush, MD, Perinatal Division, Department of Obstetrics and Gynecology, Rabin Medical Center, Beilinson Campus, Petah Tiqva 49100, Israel. E-mail: [email protected] 0002-9378/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ajog.2004.01.055
Postprandial glucose values appear to be the most effective for the determination of the likelihood of macrosomia and other adverse pregnancy outcomes.1,2 Controversy surrounds the appropriate timing for evaluation: at 60 minutes or at 120 minutes after a meal.
577
Ben-Haroush et al Some investigators recommend the use of a 1-hour postprandial value for optimal care3,4; other investigators have shown that monitoring either 1 hour or 2 hours postprandially led to similar outcomes.5 As a result, women can choose the most convenient time for their postprandial monitoring.6 Additionally, a differential measurement (1 hour after breakfast and 2 hours after dinner) has been suggested for imposing a stricter criterion for controlling blood glucose levels.7 No data exist regarding glycemic profile, which is characterized by continuous blood glucose measurements. Therefore, using an original approach,8-10 we sought to investigate the postprandial glucose profile in pregnancies that were compromised by diabetes mellitus.
Material and methods Subjects Gravid women with type 1 diabetes mellitus and gestational diabetes mellitus (GDM) were recruited from our diabetic population for this prospective study (GDM, 24 women; type 1 diabetes mellitus, 11 women [Rabin Medical Center] and GDM, 21 women; type 1 diabetes mellitus, 9 women [St. Luke’s-Roosevelt Hospital Center]). Type 1 diabetes mellitus was diagnosed before the onset of the current pregnancy. GDM was diagnosed according to the criteria of Carpenter and Coustan11 and was treated by diet and insulin as indicated. The diet control involved caloric restriction and was prescribed on the basis of 20 to 25 kcal/kg (obese) to 35 kcal/kg (nonobese) of actual pregnancy weight. Subjects were counseled by nutritionists about a diet regimen of 3 meals and 4 snacks daily. Diet adherence was reinforced at subsequent clinic visits. Insulin was added to the therapeutic regimen of patients with GDM after diet therapy alone failed to maintain the fasting plasma glucoseO105 mg/dL or 2-hour values O120 mg/dL or both. The starting insulin dose was 0.7 unit per kilogram of actual body weight at admission, given subcutaneously 3 times daily and increased weekly as necessary. The insulin treatment did not include any of the insulin analogues, and none of the subjects used an insulin pump. The local ethics committees at both centers approved the study protocol.
The continuous glucose monitoring system The continuous glucose monitoring system (MiniMed, Sylmar, Calif ) was used in all cases. The system measures glucose levels in subcutaneous interstitial tissue. It is composed of a disposable subcutaneous glucosesensing device and an electrode that was impregnated with glucose oxidase that was connected by a cable to a light-weight monitor that is worn over the clothing or on a belt. The system takes a glucose measurement
every 10 seconds (on the basis of the electrochemical detection of glucose by its reaction with glucose oxidase) and stores an average value every 5 minutes, for a total of 288 measurements each day. A communication device enables the data that are stored in the monitor to be downloaded and reviewed on a personal computer. The time delay between glucose values of venous plasma and subcutaneous concentrations is given with maximal 5 minutes.12 The software for the download of the sensor data takes this delay into consideration, which avoids the need for further corrections. In a previous study, we showed that the correlation coefficient (r) between the glucose measurements by the sensor and meter was 0.93 G 0.04 and between the plasma glucose by reflectance meter monitoring and sensor recording was 0.91 G 0.02.8 The patients were instructed to code the time of food intake at the beginning of the meal, insulin injections, exercise periods, and symptomatic hypoglycemic events into the monitor. The patients’ daily routine and level of physical activity were not standardized. Subjects were connected for 72 consecutive hours; during this period, they also performed fingerstick capillary glucose monitoring using a reflectance meter in the morning after overnight fasting and 2 hours after meals (6-8 times per day) and self-coded the data into the continuous glucose monitor for synchronized data analysis and for quality control. Patients were unaware of the results of the sensor measurements during the monitoring period and did not use the fingerstick capillary glucose measurements for treatment adjustment during the study period. At the end of the study for quality control, a comparison between the reflectance meter and the continuous glucose monitoring values was performed.
Data analysis For each subject, 3 meals were analyzed; for each meal, the first 240 minutes were analyzed (Figure). Mean blood glucose, the time and value of postprandial glucose peak and nadir, the time interval required to return to preprandial glucose value, and the number of postprandial hypoglycemic events (glucose !50 mg/dL for O15 minutes, detected after the glucose peak time) were calculated. The delta (postprandial minus preprandial glucose values) at 60, 120, and 180 minutes after a meal and the percentage of meals in which the patient did not reach the preprandial glucose value within a 3-hour period were also calculated. Further analysis was performed to assess possible differences in the postprandial glucose profile between breakfast, lunch, and dinner and between insulin-treated patients with GDM in good versus poor (mean blood glucose,O105 mg/ dL) glycemic control.
578
Ben-Haroush et al
Statistical analysis Statistical analysis was performed with the SPSS for Windows (version 10.0; SPSS Inc, Chicago, Ill) statistical package and NCSS/PASS 2000 (NCSS Statistical Software, Kaysville, Utah) software. Results are expressed in means G SD or rate. The Student t test and univariate analysis of variance tests were used for normally distributed continuous data. For abnormally distributed continuous variables, the Mann-Whitney, Kruskal-Wallis, or Friedman’s nonparametric (for repeated measures) test was used. The chi-squared test was used for categoric data. The coefficient of variation with Levene’s test for equality of variances was calculated for postprandial blood glucose values for each group. For the main dependent variable (time interval from meal to peak glucose), 22 meals for each group will result with 80% power to detect a difference of 30 minutes, with a significant level of .05. A probability value of !.05 was considered significant.
Results Sixty-five women participated in the study: 26 women with GDM that was treated solely by diet; 19 women with GDM that was treated with insulin, and 20 women with type-1 diabetes mellitus. The insulin-treated patients with GDM were characterized by older maternal age, higher gravidity and parity, and higher body mass index, compared with type-1 diabetes mellitus or diet-treated patients with GDM (Table I). Patients with type-1 diabetes mellitus had higher mean blood glucose and a larger coefficient of variation of blood glucose values (41.4% in postprandial measurements; Table II). The time interval that was required for a return to preprandial glucose value was significantly different between the groups; in approximately 50% of meals, the patient did not reach the preprandial glucose value within a 3-hour observation period (Table II). The time interval from meal to peak postprandial glucose levels varied greatly from patient to patient (coefficient of variation of 43%, 37.8%, and 47%, in patients with type 1 diabetes mellitus, diet-treated GDM, and insulin-treated GDM, respectively), but overall the time interval was similar in all groups (mean time of 93 G 40, 82 G 31, and 85 G 40 minutes, in patients with type 1 diabetes mellitus, diet-treated GDM, and insulintreated GDM, respectively; Table II). Further analysis of glucose characteristics revealed that, at 120 minutes after meals, most of the patients (77%, 88%, and 72% of patients with type 1 diabetes mellitus, diet-treated GDM, and insulin-treated GDM, respectively; P = .015) still had a positive delta value (postprandial minus preprandial glucose values) in blood glucose values (ie, glucose values above preprandial values). At the same time, negative delta values in blood glucose values (glucose values below preprandial
Figure Schematic postprandial glycemic profile. Patients were connected for 72 hours; the first 240-minute period after every meal was used for evaluation.
values) were found in 17%, 12%, and 28% of patients with type 1 diabetes mellitus, diet-treated GDM, and insulin-treated GDM, respectively (P = .015). Women with type-1 diabetes mellitus had the largest positive and negative delta values (60 G 57 and 33 G 16 mg/ dL, respectively). When data were stratified by mealtime, similar postprandial glucose peak times were obtained for breakfast, lunch, and dinner in all study groups (Table III). Postprandial hypoglycemia events were noted in 11.6% of the meals of patients with type-1 diabetes mellitus and in 8.1% of the meals of patients with GDM (P = .43). These asymptomatic events occurred 161 G 46 and 170 G 30 minutes after mealtime, respectively (P = .32). Analysis of the patients with insulin-treated GDM by level of glycemic control showed that the time interval from meal to peak postprandial glucose levels was similar in patients with poor or good control (mean blood glucose: 164 G 53 and 103 G 26 mg/dL; P ! .001; postprandial glucose peak time: 89 G 41 and 84 G 39 min; P = .58, respectively). When patients with type 1 diabetes mellitus were stratified further by gestational age (second vs third trimester), the postprandial peak time was 80 G 44 and 88 G 34 minutes, respectively (P = .213). Similarly, when data were stratified according to body mass index (%27 vsO27 kg/m2), the postprandial peak time was 80 G 44 and 88 G 34 minutes, respectively (P = .275).
Comment The following findings were the key findings in our study: (1) The use of continuous glucose monitoring enabled us to obtain glucose determinates at 5-minute intervals to provide a more accurate postprandial glucose profile. (2) The time interval from meal to peak postprandial glucose levels varies greatly from patient to patient.
579
Ben-Haroush et al Table I
Patient’s selected characteristics
Total number of glucose measurements in 72 hours Number of analyzed meals Number of 4-hour postprandial glucose measurements Maternal age (years) Gestational age (weeks) Gravidity Nulliparity BMI (Kg/m2)***
Type-1 diabetes (n=20)
Diet-treated GDM (n=26)
Insulin-treated GDM (n=19)
14762
19890
14303
60 2828
78 3788
57 2798
29.0 G 4.1 (19-36) 25.3 G 5.6 (19-36) 2.1 G 1.1 11 (55%) 26.2 G 1.4
32.2 G 5.0 (23-41) 31.6 G 3.2 (26-36) 3.0 G 1.5 7 (27%) 28.5 G 2.8
35.6 G 4.8 (27-43) 32.0 G 3.4 (27-38) 3.9 G 2.6 4 (21%) 28.6 G 2.5
P value*
0.001 !0.001 0.005 0.051** 0.001
* Kruskal Wallis test. ** Chi-square test. *** Body mass index- weight (Kg)/ height (m)2.
Table II
Postprandial glucose profile in study groups Type-1 diabetes (n=20) Diet-treated GDM (n=26) Insulin-treated GDM (n=19) P value*
Mean blood glucose of 72-hour measurements (mg/dl) (Coefficient of variation**) Mean blood glucose of 4-hour postprandial measurements (mg/dl) (Coefficient of variation**) Preprandial glucose (mg/dl) Postprandial glucose peak time (min) Postprandial glucose peak time coefficient of variation** Postprandial glucose nadir time (min) Postprandial glucose peak (mg/dl) Postprandial glucose nadir (mg/dl) Number of meals in which the patient did not reach the preprandial glucose value within a 3-hour period Time interval to preprandial glucose value (min)
116 G 45 (38.8%)
94 G 22 (23.4%)
110 G 32 (29.1%)
!0.001 (!0.001***)
133 G 55 (41.4%)
102 G 28 (27.5%)
122 G 48 (39.3%)
!0.001 (!0.001***)
103 G 14 93 G 40 (40-185) 43%
84 G 18 82 G 31 (30-195) 37.8%
101 G 46 85 G 40 (30-205) 47 %
0.005 0.423 0.041***
199 G 44 (75-240)
182 G 43 (95-240)
195 G 48 (90-240)
0.018
182 G 58 (86-400)
131 G 29 (90-224)
148 G 45 (90-257)
!0.001
95 G 41 (40-195)
83 G 27 (40-173)
98 G 37 (48-234)
0.048
31/58 (53.4%)
29/75 (38.7)
24/46 (52.2%)
0.16
128 G 33
134 G 32
111 G 30
0.022
* Kruskal Wallis test. ** (standard deviation/mean)*100%. *** Levene’s test for equality of variances.
Nevertheless, in women with GDM and type 1 diabetes mellitus, the time interval is approximately 90 minutes. (3) Failure to return to preprandial glucose values within a 3-hour observation period was identified overall in ap-
proximately 50% of the patients (in approximately 80%, 75%, and 65% of patients at 1, 2, and 3 hours after mealtime, respectively). 4. For good and poorly controlled patients with insulin-treated GDM, the time inter-
580
Ben-Haroush et al
Table III
Postprandial glucose peak time for every meal Postprandial peak time
Group
Breakfast
Lunch
Postprandial peak value Diner
P value* Breakfast
Type-1 diabetes (n=20) 88.8 G 42.5 103.3 G 44.4 88.6 G 33.1 0.88 Diet-treatedGDM (n=26) 77.6 G 31 87.8 G 34 87.6 G 28 0.10 Insulin-treatedGDM (n=25) 75.6 G 30.3 83.7 G 41.4 102.8 G 46.6 0.10
Lunch
Diner
P value*
180.1 G 68 196.2 G 68 168.6 G 55 0.48 130.5 G 22 127.4 G 30 140.0 G 33 0.30 133.7 G 39.5 147.1 G 48.4 159.2 G 49.8 0.004
* Friedman’s non-parametric test for repeated measures
val from meal to peak postprandial glucose levels was similar (approximately 90 minutes). (5) Similar postprandial glucose peak times were obtained for breakfast, lunch, and dinner in all study groups. (6) Postprandial hypoglycemic events (glucose !50 mg/dL for O15 minutes, detected after the glucose peak time) were noted in approximately 10% of meals and occurred about 160 minutes after mealtime. Controversy exists regarding the time and the threshold value for postprandial glucose determination. The Fourth International Workshop on GDM recommended lowering capillary blood glucose concentrations to 140 mg/dL at 1 hour and 120 mg/dL at 2 hours,13 whereas the American Diabetes Association14 recommended the option of measuring 1-hour postmeal values with a cut-off of 120 mg/dL. However, few studies to date have addressed this issue.4,5,7 In the current study, we used continuous glucose monitoring, with approximately 9500 glucose measurements in 65 patients. This is in contrast to the previous studies that used 1 single determination, which reflects a ‘‘snapshot’’ of glucose evaluation, rather than a ‘‘video’’ of continuous glucose profile. Our study revealed that the time interval from meal to peak postprandial glucose levels is similar in all types of diabetic pregnancies, and that this interval is not affected by the level of glycemic control and/or by mealtime (breakfast, lunch, or dinner). The question remains whether the peak postprandial glucose or the time to return to preprandial value will be a better reflection of quality of control. However, because all the thresholds that are recommended currently (140 mg/dL at 1 hour and 120 mg/dL at 2 hours) are not related to preprandial value, it will be reasonable to speculate that the glucose value at the 90-minutes interval should be the reflection of achievement of level of control. This idea is supported by the finding in our study that, in patients with well-controlled diabetes mellitus, the peak glucose value was 103 G 26 mg/dL, and in patients whose diabetes mellitus is poorly controlled, the peak value is 164 G 53 mg/dL. We recognize that future studies should look for the association between postprandial glucose values 90 minutes after meals and pregnancy outcome. This is especially relevant because several authors have demonstrated the importance of postprandial glucose determinations and pregnancy out-
come.2,15 Furthermore, it may be argued that adverse outcome in diabetic pregnancies can be minimized by effective therapy that aims at mimicking the glucose profiles of healthy pregnant women. However, there is little published data about what constitutes a glucose profile in normal pregnancy; Parretti et al16 studied 51 nonobese, nondiabetic pregnant women and used multiple fingerstick glucose measurements to evaluate their glucose profile. They found that the mean peak postprandial glucose response is at 1 hour and that the mean postprandial glucose levels never exceeded 105 mg/dL. Of note is the fact that approximately 50% of the patients could not reach the preprandial glucose values even 3 hours after a meal, which again supports the concept that peak glucose values may be more representative of the glycemic profile than the time to returning to preprandial value. We also found that approximately 10% of the patients had hypoglycemia, approximately 160 minutes after a meal. This may be attributable to the tendency toward accelerated starvation in pregnancy and, more specifically, in pregnancies that are complicated by diabetes mellitus.17 We feel that this finding should be considered when the meal plan for the pregnant patient with diabetes mellitus is designed; the addition of several between meal snacks may prevent this occurrence. An additional approach to the improvement of postprandial peak glucose values and the reduction of the number of hypoglycemic episodes may be considered. Although not addressed in this study, by using insulin analogues, which characteristically have shorter half life, it has been reported that their use was associated with a lower incidence of hypoglycemic events.18 Human insulin (used in the current study) has a peak time of 3 to 4 hours. The new insulin analogues have much shorter peak times, approximately 90 minutes,19 which is parallel to the mean glucose peak time that was found in our study. Thus, it is possible that insulin analogues may mimic the physiologic characteristics of the glucose profile more efficiently. In conclusion, the continuous glucose monitoring system enabled us to characterize postprandial glucose profile in diabetic pregnancies. Our study determined that the mean time to peak postprandial glucose in 3 categories of women with diabetes mellitus is approximately
Ben-Haroush et al 90 minutes. This information should be considered in the treatment of diabetes mellitus in pregnancy.
References 1. American College of Obstetricians and Gynecologists. Gestational diabetes. Washington (DC): The College; 2001. Practice Bulletin No.: 30. 2. Jovanovic-Peterson L, Peterson CM, Reed GF, Metzger BE, Mills JL, Knopp RH, et al. Maternal postprandial glucose levels and infant birth weight: the Diabetes in Early Pregnancy Study [The National Institute of Child Health and Human DevelopmenteDiabetes in Early Pregnancy Study]. Am J Obstet Gynecol 1991;164: 103-11. 3. Combs CA, Gunderson E, Kitzmiller JL, Gavin LA, Main EK. Relationship of fetal macrosomia to maternal postprandial glucose control during pregnancy. Diabetes Care 1992;15:1251-7. 4. Huddleston JF, Cramer MK, Vroon DH. A rationale for omitting two-hour postprandial glucose determinations in gestational diabetes. Am J Obstet Gynecol 1993;169:257-62. 5. Moses RG, Lucas EM, Knights S. Gestational diabetes mellitus: At what time should the postprandial glucose level be monitored? Aust N Z J Obstet Gynaecol 1999;39:457-60. 6. Leguizamon G, Krupitzki H, Glujovsky D, Olivera Ravasi M, Reece EA. Blood glucose monitoring in gestational diabetes mellitus: 1- versus 2-h blood glucose determinations. J Matern Fetal Neonatal Med 2002;12:384-8. 7. Sivan E, Weisz B, Homko CJ, Reece EA, Schiff E. One or two hours postprandial glucose measurements: Are they the same? Am J Obstet Gynecol 2001;185:604-7. 8. Yogev Y, Chen R, Ben-Haroush A, Phillip M, Jovanovic L, Hod M. Continuous glucose monitoring for the evaluation of gravid women with type 1 diabetes women. Obstet Gynecol 2003;101:633-8.
581 9. Chen R, Yogev Y, Ben-Haroush A, Peled Y, Bar J, Phillip M, et al. The use of continuous glucose monitoring in the evaluation of gravid women with gestational diabetes. J Matern Fetal Neonatal Med 2003;14:256-60. 10. Yogev Y, Ben-Haroush A, Chen R, Kaplan B, Phillip M, Hod M. Continuous glucose monitoring for treatment adjustment in diabetic pregnancies: a pilot study. Diabet Med 2003;20:558-62. 11. Carpenter MW, Coustan DR. Criteria for screening tests for gestational diabetes. Am J Obstet Gynecol 1982;144:768-73. 12. Thennadil SN, Rennert JL, Wenzel BJ, Hazen KH, Ruchti TL, Block MB. Comparison of glucose concentration in interstitial fluid, and capillary and venous blood during rapid changes in blood glucose levels. Diabetes Technol Ther 2001;3:357-64. 13. Metzger BE, Coustan DR. Summary and recommendations of the fourth international workshop-conference on gestational diabetes mellitus. Diabetes Care 1998;21(suppl):B161-7. 14. American Diabetes Association. Gestational diabetes mellitus. Diabetes Care 2003;26(suppl):S103-5. 15. de Veciana M, Major CA, Morgan MA, Asrat T, Toohey JS, Lien JM, et al. Postprandial versus preprandial blood glucose monitoring in women with gestational diabetes mellitus requiring. N Engl J Med 1995;9(333):1237-41. 16. Parretti E, Mecacci F, Cioni R, Carignani L, Mignosa M, Torre PL, et al. Third-trimester maternal glucose levels from diurnal profiles in nondiabetic pregnancies. Diabetes Care 2001;24: 1319-23. 17. Carpenter MW. Metabolic changes in gestational diabetes. Clin Perinatol 1993;20:583-91. 18. Simmons D. The utility and efficacy of the new insulins in the management of diabetes and pregnancy. Curr Diab Rep 2002;2: 331-6. 19. Pettitt DJ, Ospina P, Kolaczynski JW, Jovanovic L. Comparison of an insulin analog, insulin aspart, and regular human insulin with no insulin in gestational diabetes mellitus. Diabetes Care 2003;26:183-6.
www.elsevier.com/locate/ajog
The postprandial glucose profile in the diabetic pregnancy Avi Ben-Haroush, MD,a,* Yariv Yogev, MD,a,b Rony Chen, MD,a Barak Rosenn, MD,b Moshe Hod, MD,a Oded Langer, MDb Department of Obstetrics and Gynecology, Rabin Medical Center, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel a; Department of Obstetrics and Gynecology, St. Luke’s-Roosevelt Hospital Center, New York, University Hospital of Columbia University, New York, NYb Received for publication November 5, 2003; revised January 13, 2004; accepted January 29, 2004
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– KEY WORDS Gestational diabetes mellitus Postprandial peak Glucose
Objective: A controversy exists regarding the time to monitor blood glucose in the diabetic pregnancy (60 or 120 minutes after meals). Using a novel approach that provides continuous measurement of blood glucose, we sought to determine postprandial glucose profile in the diabetic pregnancy. Study design: Subjects were connected to a continuous glucose monitoring system for 72 consecutive hours. A continuous glucose monitoring system measures the interstitial glucose levels in subcutaneous tissue every 5 minutes. Women were instructed to record the time of each meal during the study period. For each meal, the first 240 minutes were analyzed. Results: Sixty-five women participated in the study: 26 women were treated by diet alone; 19 women received insulin therapy, and 20 women had type 1 diabetes mellitus. The time interval from meal to peak postprandial glucose levels was similar in all the evaluated types of diabetic pregnancies and in good and poor control insulin-treated patients with gestational diabetes mellitus (approximately 90 minutes). Failure to return to preprandial glucose values within a 3-hour observation period was identified in approximately 50% of the patients. A similar postprandial glucose peak time was obtained for breakfast, lunch, and dinner in all study groups. Postprandial hypoglycemia events were noted in approximately 10% of the meals and occurred about 160 minutes after mealtime. Conclusion: The time interval for postprandial glucose peak in diabetic pregnancies is approximately 90 minutes after meals throughout the day and is not affected by the level of glycemic control. This information should be considered in the treatment of diabetes mellitus in pregnancy. Ó 2004 Elsevier Inc. All rights reserved.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
* Reprint requests: Avi Ben-Haroush, MD, Perinatal Division, Department of Obstetrics and Gynecology, Rabin Medical Center, Beilinson Campus, Petah Tiqva 49100, Israel. E-mail: [email protected] 0002-9378/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ajog.2004.01.055
Postprandial glucose values appear to be the most effective for the determination of the likelihood of macrosomia and other adverse pregnancy outcomes.1,2 Controversy surrounds the appropriate timing for evaluation: at 60 minutes or at 120 minutes after a meal.
577
Ben-Haroush et al Some investigators recommend the use of a 1-hour postprandial value for optimal care3,4; other investigators have shown that monitoring either 1 hour or 2 hours postprandially led to similar outcomes.5 As a result, women can choose the most convenient time for their postprandial monitoring.6 Additionally, a differential measurement (1 hour after breakfast and 2 hours after dinner) has been suggested for imposing a stricter criterion for controlling blood glucose levels.7 No data exist regarding glycemic profile, which is characterized by continuous blood glucose measurements. Therefore, using an original approach,8-10 we sought to investigate the postprandial glucose profile in pregnancies that were compromised by diabetes mellitus.
Material and methods Subjects Gravid women with type 1 diabetes mellitus and gestational diabetes mellitus (GDM) were recruited from our diabetic population for this prospective study (GDM, 24 women; type 1 diabetes mellitus, 11 women [Rabin Medical Center] and GDM, 21 women; type 1 diabetes mellitus, 9 women [St. Luke’s-Roosevelt Hospital Center]). Type 1 diabetes mellitus was diagnosed before the onset of the current pregnancy. GDM was diagnosed according to the criteria of Carpenter and Coustan11 and was treated by diet and insulin as indicated. The diet control involved caloric restriction and was prescribed on the basis of 20 to 25 kcal/kg (obese) to 35 kcal/kg (nonobese) of actual pregnancy weight. Subjects were counseled by nutritionists about a diet regimen of 3 meals and 4 snacks daily. Diet adherence was reinforced at subsequent clinic visits. Insulin was added to the therapeutic regimen of patients with GDM after diet therapy alone failed to maintain the fasting plasma glucoseO105 mg/dL or 2-hour values O120 mg/dL or both. The starting insulin dose was 0.7 unit per kilogram of actual body weight at admission, given subcutaneously 3 times daily and increased weekly as necessary. The insulin treatment did not include any of the insulin analogues, and none of the subjects used an insulin pump. The local ethics committees at both centers approved the study protocol.
The continuous glucose monitoring system The continuous glucose monitoring system (MiniMed, Sylmar, Calif ) was used in all cases. The system measures glucose levels in subcutaneous interstitial tissue. It is composed of a disposable subcutaneous glucosesensing device and an electrode that was impregnated with glucose oxidase that was connected by a cable to a light-weight monitor that is worn over the clothing or on a belt. The system takes a glucose measurement
every 10 seconds (on the basis of the electrochemical detection of glucose by its reaction with glucose oxidase) and stores an average value every 5 minutes, for a total of 288 measurements each day. A communication device enables the data that are stored in the monitor to be downloaded and reviewed on a personal computer. The time delay between glucose values of venous plasma and subcutaneous concentrations is given with maximal 5 minutes.12 The software for the download of the sensor data takes this delay into consideration, which avoids the need for further corrections. In a previous study, we showed that the correlation coefficient (r) between the glucose measurements by the sensor and meter was 0.93 G 0.04 and between the plasma glucose by reflectance meter monitoring and sensor recording was 0.91 G 0.02.8 The patients were instructed to code the time of food intake at the beginning of the meal, insulin injections, exercise periods, and symptomatic hypoglycemic events into the monitor. The patients’ daily routine and level of physical activity were not standardized. Subjects were connected for 72 consecutive hours; during this period, they also performed fingerstick capillary glucose monitoring using a reflectance meter in the morning after overnight fasting and 2 hours after meals (6-8 times per day) and self-coded the data into the continuous glucose monitor for synchronized data analysis and for quality control. Patients were unaware of the results of the sensor measurements during the monitoring period and did not use the fingerstick capillary glucose measurements for treatment adjustment during the study period. At the end of the study for quality control, a comparison between the reflectance meter and the continuous glucose monitoring values was performed.
Data analysis For each subject, 3 meals were analyzed; for each meal, the first 240 minutes were analyzed (Figure). Mean blood glucose, the time and value of postprandial glucose peak and nadir, the time interval required to return to preprandial glucose value, and the number of postprandial hypoglycemic events (glucose !50 mg/dL for O15 minutes, detected after the glucose peak time) were calculated. The delta (postprandial minus preprandial glucose values) at 60, 120, and 180 minutes after a meal and the percentage of meals in which the patient did not reach the preprandial glucose value within a 3-hour period were also calculated. Further analysis was performed to assess possible differences in the postprandial glucose profile between breakfast, lunch, and dinner and between insulin-treated patients with GDM in good versus poor (mean blood glucose,O105 mg/ dL) glycemic control.
578
Ben-Haroush et al
Statistical analysis Statistical analysis was performed with the SPSS for Windows (version 10.0; SPSS Inc, Chicago, Ill) statistical package and NCSS/PASS 2000 (NCSS Statistical Software, Kaysville, Utah) software. Results are expressed in means G SD or rate. The Student t test and univariate analysis of variance tests were used for normally distributed continuous data. For abnormally distributed continuous variables, the Mann-Whitney, Kruskal-Wallis, or Friedman’s nonparametric (for repeated measures) test was used. The chi-squared test was used for categoric data. The coefficient of variation with Levene’s test for equality of variances was calculated for postprandial blood glucose values for each group. For the main dependent variable (time interval from meal to peak glucose), 22 meals for each group will result with 80% power to detect a difference of 30 minutes, with a significant level of .05. A probability value of !.05 was considered significant.
Results Sixty-five women participated in the study: 26 women with GDM that was treated solely by diet; 19 women with GDM that was treated with insulin, and 20 women with type-1 diabetes mellitus. The insulin-treated patients with GDM were characterized by older maternal age, higher gravidity and parity, and higher body mass index, compared with type-1 diabetes mellitus or diet-treated patients with GDM (Table I). Patients with type-1 diabetes mellitus had higher mean blood glucose and a larger coefficient of variation of blood glucose values (41.4% in postprandial measurements; Table II). The time interval that was required for a return to preprandial glucose value was significantly different between the groups; in approximately 50% of meals, the patient did not reach the preprandial glucose value within a 3-hour observation period (Table II). The time interval from meal to peak postprandial glucose levels varied greatly from patient to patient (coefficient of variation of 43%, 37.8%, and 47%, in patients with type 1 diabetes mellitus, diet-treated GDM, and insulin-treated GDM, respectively), but overall the time interval was similar in all groups (mean time of 93 G 40, 82 G 31, and 85 G 40 minutes, in patients with type 1 diabetes mellitus, diet-treated GDM, and insulintreated GDM, respectively; Table II). Further analysis of glucose characteristics revealed that, at 120 minutes after meals, most of the patients (77%, 88%, and 72% of patients with type 1 diabetes mellitus, diet-treated GDM, and insulin-treated GDM, respectively; P = .015) still had a positive delta value (postprandial minus preprandial glucose values) in blood glucose values (ie, glucose values above preprandial values). At the same time, negative delta values in blood glucose values (glucose values below preprandial
Figure Schematic postprandial glycemic profile. Patients were connected for 72 hours; the first 240-minute period after every meal was used for evaluation.
values) were found in 17%, 12%, and 28% of patients with type 1 diabetes mellitus, diet-treated GDM, and insulin-treated GDM, respectively (P = .015). Women with type-1 diabetes mellitus had the largest positive and negative delta values (60 G 57 and 33 G 16 mg/ dL, respectively). When data were stratified by mealtime, similar postprandial glucose peak times were obtained for breakfast, lunch, and dinner in all study groups (Table III). Postprandial hypoglycemia events were noted in 11.6% of the meals of patients with type-1 diabetes mellitus and in 8.1% of the meals of patients with GDM (P = .43). These asymptomatic events occurred 161 G 46 and 170 G 30 minutes after mealtime, respectively (P = .32). Analysis of the patients with insulin-treated GDM by level of glycemic control showed that the time interval from meal to peak postprandial glucose levels was similar in patients with poor or good control (mean blood glucose: 164 G 53 and 103 G 26 mg/dL; P ! .001; postprandial glucose peak time: 89 G 41 and 84 G 39 min; P = .58, respectively). When patients with type 1 diabetes mellitus were stratified further by gestational age (second vs third trimester), the postprandial peak time was 80 G 44 and 88 G 34 minutes, respectively (P = .213). Similarly, when data were stratified according to body mass index (%27 vsO27 kg/m2), the postprandial peak time was 80 G 44 and 88 G 34 minutes, respectively (P = .275).
Comment The following findings were the key findings in our study: (1) The use of continuous glucose monitoring enabled us to obtain glucose determinates at 5-minute intervals to provide a more accurate postprandial glucose profile. (2) The time interval from meal to peak postprandial glucose levels varies greatly from patient to patient.
579
Ben-Haroush et al Table I
Patient’s selected characteristics
Total number of glucose measurements in 72 hours Number of analyzed meals Number of 4-hour postprandial glucose measurements Maternal age (years) Gestational age (weeks) Gravidity Nulliparity BMI (Kg/m2)***
Type-1 diabetes (n=20)
Diet-treated GDM (n=26)
Insulin-treated GDM (n=19)
14762
19890
14303
60 2828
78 3788
57 2798
29.0 G 4.1 (19-36) 25.3 G 5.6 (19-36) 2.1 G 1.1 11 (55%) 26.2 G 1.4
32.2 G 5.0 (23-41) 31.6 G 3.2 (26-36) 3.0 G 1.5 7 (27%) 28.5 G 2.8
35.6 G 4.8 (27-43) 32.0 G 3.4 (27-38) 3.9 G 2.6 4 (21%) 28.6 G 2.5
P value*
0.001 !0.001 0.005 0.051** 0.001
* Kruskal Wallis test. ** Chi-square test. *** Body mass index- weight (Kg)/ height (m)2.
Table II
Postprandial glucose profile in study groups Type-1 diabetes (n=20) Diet-treated GDM (n=26) Insulin-treated GDM (n=19) P value*
Mean blood glucose of 72-hour measurements (mg/dl) (Coefficient of variation**) Mean blood glucose of 4-hour postprandial measurements (mg/dl) (Coefficient of variation**) Preprandial glucose (mg/dl) Postprandial glucose peak time (min) Postprandial glucose peak time coefficient of variation** Postprandial glucose nadir time (min) Postprandial glucose peak (mg/dl) Postprandial glucose nadir (mg/dl) Number of meals in which the patient did not reach the preprandial glucose value within a 3-hour period Time interval to preprandial glucose value (min)
116 G 45 (38.8%)
94 G 22 (23.4%)
110 G 32 (29.1%)
!0.001 (!0.001***)
133 G 55 (41.4%)
102 G 28 (27.5%)
122 G 48 (39.3%)
!0.001 (!0.001***)
103 G 14 93 G 40 (40-185) 43%
84 G 18 82 G 31 (30-195) 37.8%
101 G 46 85 G 40 (30-205) 47 %
0.005 0.423 0.041***
199 G 44 (75-240)
182 G 43 (95-240)
195 G 48 (90-240)
0.018
182 G 58 (86-400)
131 G 29 (90-224)
148 G 45 (90-257)
!0.001
95 G 41 (40-195)
83 G 27 (40-173)
98 G 37 (48-234)
0.048
31/58 (53.4%)
29/75 (38.7)
24/46 (52.2%)
0.16
128 G 33
134 G 32
111 G 30
0.022
* Kruskal Wallis test. ** (standard deviation/mean)*100%. *** Levene’s test for equality of variances.
Nevertheless, in women with GDM and type 1 diabetes mellitus, the time interval is approximately 90 minutes. (3) Failure to return to preprandial glucose values within a 3-hour observation period was identified overall in ap-
proximately 50% of the patients (in approximately 80%, 75%, and 65% of patients at 1, 2, and 3 hours after mealtime, respectively). 4. For good and poorly controlled patients with insulin-treated GDM, the time inter-
580
Ben-Haroush et al
Table III
Postprandial glucose peak time for every meal Postprandial peak time
Group
Breakfast
Lunch
Postprandial peak value Diner
P value* Breakfast
Type-1 diabetes (n=20) 88.8 G 42.5 103.3 G 44.4 88.6 G 33.1 0.88 Diet-treatedGDM (n=26) 77.6 G 31 87.8 G 34 87.6 G 28 0.10 Insulin-treatedGDM (n=25) 75.6 G 30.3 83.7 G 41.4 102.8 G 46.6 0.10
Lunch
Diner
P value*
180.1 G 68 196.2 G 68 168.6 G 55 0.48 130.5 G 22 127.4 G 30 140.0 G 33 0.30 133.7 G 39.5 147.1 G 48.4 159.2 G 49.8 0.004
* Friedman’s non-parametric test for repeated measures
val from meal to peak postprandial glucose levels was similar (approximately 90 minutes). (5) Similar postprandial glucose peak times were obtained for breakfast, lunch, and dinner in all study groups. (6) Postprandial hypoglycemic events (glucose !50 mg/dL for O15 minutes, detected after the glucose peak time) were noted in approximately 10% of meals and occurred about 160 minutes after mealtime. Controversy exists regarding the time and the threshold value for postprandial glucose determination. The Fourth International Workshop on GDM recommended lowering capillary blood glucose concentrations to 140 mg/dL at 1 hour and 120 mg/dL at 2 hours,13 whereas the American Diabetes Association14 recommended the option of measuring 1-hour postmeal values with a cut-off of 120 mg/dL. However, few studies to date have addressed this issue.4,5,7 In the current study, we used continuous glucose monitoring, with approximately 9500 glucose measurements in 65 patients. This is in contrast to the previous studies that used 1 single determination, which reflects a ‘‘snapshot’’ of glucose evaluation, rather than a ‘‘video’’ of continuous glucose profile. Our study revealed that the time interval from meal to peak postprandial glucose levels is similar in all types of diabetic pregnancies, and that this interval is not affected by the level of glycemic control and/or by mealtime (breakfast, lunch, or dinner). The question remains whether the peak postprandial glucose or the time to return to preprandial value will be a better reflection of quality of control. However, because all the thresholds that are recommended currently (140 mg/dL at 1 hour and 120 mg/dL at 2 hours) are not related to preprandial value, it will be reasonable to speculate that the glucose value at the 90-minutes interval should be the reflection of achievement of level of control. This idea is supported by the finding in our study that, in patients with well-controlled diabetes mellitus, the peak glucose value was 103 G 26 mg/dL, and in patients whose diabetes mellitus is poorly controlled, the peak value is 164 G 53 mg/dL. We recognize that future studies should look for the association between postprandial glucose values 90 minutes after meals and pregnancy outcome. This is especially relevant because several authors have demonstrated the importance of postprandial glucose determinations and pregnancy out-
come.2,15 Furthermore, it may be argued that adverse outcome in diabetic pregnancies can be minimized by effective therapy that aims at mimicking the glucose profiles of healthy pregnant women. However, there is little published data about what constitutes a glucose profile in normal pregnancy; Parretti et al16 studied 51 nonobese, nondiabetic pregnant women and used multiple fingerstick glucose measurements to evaluate their glucose profile. They found that the mean peak postprandial glucose response is at 1 hour and that the mean postprandial glucose levels never exceeded 105 mg/dL. Of note is the fact that approximately 50% of the patients could not reach the preprandial glucose values even 3 hours after a meal, which again supports the concept that peak glucose values may be more representative of the glycemic profile than the time to returning to preprandial value. We also found that approximately 10% of the patients had hypoglycemia, approximately 160 minutes after a meal. This may be attributable to the tendency toward accelerated starvation in pregnancy and, more specifically, in pregnancies that are complicated by diabetes mellitus.17 We feel that this finding should be considered when the meal plan for the pregnant patient with diabetes mellitus is designed; the addition of several between meal snacks may prevent this occurrence. An additional approach to the improvement of postprandial peak glucose values and the reduction of the number of hypoglycemic episodes may be considered. Although not addressed in this study, by using insulin analogues, which characteristically have shorter half life, it has been reported that their use was associated with a lower incidence of hypoglycemic events.18 Human insulin (used in the current study) has a peak time of 3 to 4 hours. The new insulin analogues have much shorter peak times, approximately 90 minutes,19 which is parallel to the mean glucose peak time that was found in our study. Thus, it is possible that insulin analogues may mimic the physiologic characteristics of the glucose profile more efficiently. In conclusion, the continuous glucose monitoring system enabled us to characterize postprandial glucose profile in diabetic pregnancies. Our study determined that the mean time to peak postprandial glucose in 3 categories of women with diabetes mellitus is approximately
Ben-Haroush et al 90 minutes. This information should be considered in the treatment of diabetes mellitus in pregnancy.
References 1. American College of Obstetricians and Gynecologists. Gestational diabetes. Washington (DC): The College; 2001. Practice Bulletin No.: 30. 2. Jovanovic-Peterson L, Peterson CM, Reed GF, Metzger BE, Mills JL, Knopp RH, et al. Maternal postprandial glucose levels and infant birth weight: the Diabetes in Early Pregnancy Study [The National Institute of Child Health and Human DevelopmenteDiabetes in Early Pregnancy Study]. Am J Obstet Gynecol 1991;164: 103-11. 3. Combs CA, Gunderson E, Kitzmiller JL, Gavin LA, Main EK. Relationship of fetal macrosomia to maternal postprandial glucose control during pregnancy. Diabetes Care 1992;15:1251-7. 4. Huddleston JF, Cramer MK, Vroon DH. A rationale for omitting two-hour postprandial glucose determinations in gestational diabetes. Am J Obstet Gynecol 1993;169:257-62. 5. Moses RG, Lucas EM, Knights S. Gestational diabetes mellitus: At what time should the postprandial glucose level be monitored? Aust N Z J Obstet Gynaecol 1999;39:457-60. 6. Leguizamon G, Krupitzki H, Glujovsky D, Olivera Ravasi M, Reece EA. Blood glucose monitoring in gestational diabetes mellitus: 1- versus 2-h blood glucose determinations. J Matern Fetal Neonatal Med 2002;12:384-8. 7. Sivan E, Weisz B, Homko CJ, Reece EA, Schiff E. One or two hours postprandial glucose measurements: Are they the same? Am J Obstet Gynecol 2001;185:604-7. 8. Yogev Y, Chen R, Ben-Haroush A, Phillip M, Jovanovic L, Hod M. Continuous glucose monitoring for the evaluation of gravid women with type 1 diabetes women. Obstet Gynecol 2003;101:633-8.
581 9. Chen R, Yogev Y, Ben-Haroush A, Peled Y, Bar J, Phillip M, et al. The use of continuous glucose monitoring in the evaluation of gravid women with gestational diabetes. J Matern Fetal Neonatal Med 2003;14:256-60. 10. Yogev Y, Ben-Haroush A, Chen R, Kaplan B, Phillip M, Hod M. Continuous glucose monitoring for treatment adjustment in diabetic pregnancies: a pilot study. Diabet Med 2003;20:558-62. 11. Carpenter MW, Coustan DR. Criteria for screening tests for gestational diabetes. Am J Obstet Gynecol 1982;144:768-73. 12. Thennadil SN, Rennert JL, Wenzel BJ, Hazen KH, Ruchti TL, Block MB. Comparison of glucose concentration in interstitial fluid, and capillary and venous blood during rapid changes in blood glucose levels. Diabetes Technol Ther 2001;3:357-64. 13. Metzger BE, Coustan DR. Summary and recommendations of the fourth international workshop-conference on gestational diabetes mellitus. Diabetes Care 1998;21(suppl):B161-7. 14. American Diabetes Association. Gestational diabetes mellitus. Diabetes Care 2003;26(suppl):S103-5. 15. de Veciana M, Major CA, Morgan MA, Asrat T, Toohey JS, Lien JM, et al. Postprandial versus preprandial blood glucose monitoring in women with gestational diabetes mellitus requiring. N Engl J Med 1995;9(333):1237-41. 16. Parretti E, Mecacci F, Cioni R, Carignani L, Mignosa M, Torre PL, et al. Third-trimester maternal glucose levels from diurnal profiles in nondiabetic pregnancies. Diabetes Care 2001;24: 1319-23. 17. Carpenter MW. Metabolic changes in gestational diabetes. Clin Perinatol 1993;20:583-91. 18. Simmons D. The utility and efficacy of the new insulins in the management of diabetes and pregnancy. Curr Diab Rep 2002;2: 331-6. 19. Pettitt DJ, Ospina P, Kolaczynski JW, Jovanovic L. Comparison of an insulin analog, insulin aspart, and regular human insulin with no insulin in gestational diabetes mellitus. Diabetes Care 2003;26:183-6.