The goal of management in type 1 diabetes is to maintain blood glucose and hemoglobin (Hb)A1c levels as near to normal as possible. This goal is difficult to achieve in all patients and poses an even greater challenge in pregnant women.
Improved glycemic control during diabetic pregnancy has been found to decrease perinatal morbidity and mortality, 1,2 as well as the congenital malformation rate. 3 On the one hand, however, too-tight glycemic control in this patient group may be accompanied by a high incidence of hypoglycemia, ranging from 36% to 71%. 4,5 On the other hand, as maternal glycemia increases, the risk of macrosomia, the most common and significant perinatal complication clearly associated with diabetes in pregnancy, increases as well. 6 Patients with diabetic pregnancy are treated with multiple insulin injections and are advised to perform self–blood glucose monitoring by fingerstick measurements 4–8 times per day. The optimal frequency of blood glucose testing in gravid patients with type 1 diabetes has not been established.
The self–blood glucose monitoring method has an important limitation; it provides only a single value during the day and does not allow for continuous, longitudinal monitoring. As such, it may be missing both hypoglycemic and hyperglycemic events. To counter this problem, a new technology of continuous glucose monitoring was recently developed (Mastortotoro J, Levy R, Georges LP, White N, Mestman J. Clinical results from a continuous glucose sensor multi-center study [abstract]. Diabetes 1998;47:A61). The system measures interstitial glucose levels in subcutaneous tissue, within a range of 40–400 mg/dL. Glucose values obtained with continuous glucose monitoring have been shown to correlate with laboratory measurements of plasma glucose levels 7 and with home glucose meter values (Mastortotoro J, et al. Diabetes 1998;47:A61).
The purpose of the present study was to compare the daily glycemic profile reflected by continuous and intermittent blood glucose monitoring in pregnant women with type 1 diabetes and to determine whether treatment strategy protocols based on the two monitoring methods differ.
MATERIALS AND METHODS
The initial study sample consisted of 41 consecutive gravid women with type 1 diabetes who were recruited for this prospective study during a routine clinical visit to the Diabetes in Pregnancy Center of the Perinatal Division Unit, Rabin Medical Center between November 2001 and March 2002. Of these, 34 women (82.9%) gave consent to participate after receiving a comprehensive explanation of the study. The local ethics committee approved the study protocol.
In all cases, type 1 diabetes mellitus was diagnosed before the onset of the current pregnancy. Gestational age ranged from 16 to 32 weeks. All patients were being treated with insulin and were under the care of a registered dietitian for individualized counseling and instructions.
At entry to the study, the patient's chart was reviewed for demographic data, gravidity, parity, and body mass index (BMI). Prior to sensor placement, levels of HbA1c, fructosamine, and plasma glucose were measured.
The MiniMed continuous glucose monitoring system (MiniMed, Sylmar, CA) was used in all cases for 3 days. The system measures glucose levels in subcutaneous interstitial tissue. It is composed of a disposable subcutaneous glucose-sensing device and an electrode impregnated with glucose oxidase connected by a cable to a lightweight monitor, which is worn over the clothing or on a belt. The system takes a glucose measurement every 10 seconds, based on 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 stored in the monitor to be downloaded and reviewed on a personal computer. The patients are unaware of the results of the sensor measurements during the monitoring period.
The same trained nurse placed all continuous glucose monitoring sensors. Prior to placement, the site on the flank was prepared with an alcohol sponge, and the sensor was calibrated according to the manufacturer's instructions.
The patients were shown how to code the time of food intake, insulin injections, exercise periods, and symptomatic hypoglycemic events into the monitor.
Patients were instructed to wear the continuous glucose monitoring device for 72 consecutive hours. During this period, they also performed fingerstick capillary glucose measurements in the morning after overnight fasting and 2 hours after meals (6–8 times per day) using a glucometer (Ames Glucometer Elite, Bayer Corp., Elkhart, IN) and self-coded the data into the monitor. At the end of the study period, before the nurse disconnected the women from the sensor, plasma and glucometer glucose values were measured. Quality control measures of glucose levels from the meter, sensor, and plasma glucose were also performed at the time of connection to the continuous glucose monitoring system and again at study completion. The data collected by self–blood glucose monitoring and continuous glucose monitoring for each patient were evaluated separately by a single experienced physician. The average time (in minutes) per day of sensor glucose readings of less than 50 mg/dL and greater than 140 mg/dL was calculated from the daily graphs. A hypoglycemic event was defined as a greater than 30-minute asymptomatic reading below 50 mg/dL or symptomatic hypoglycemia detected by meter or monitoring records.
The decision regarding the insulin regimen was made twice, first on the basis of the self–blood glucose monitoring data and then on the basis of the continuous glucose monitoring data. To prevent bias, the two types of data were presented to the physician on different occasions. The physician was blinded to the identity of the individual patients.
Continuous parameters are given as means ± standard deviations. Pearson correlation coefficient (r) and the significance for it (P) were calculated between the variables. Reliability coefficient was used to quantitate the consistency of the two measuring methods (self–blood glucose monitoring and continuous glucose monitoring) applied to each subject. Paired t test was used to determine the statistical significance of differences in mean continuous parameters. A P value equal to or less than .05 was considered statistically significant.
Mean patient age was 26 ± 4.7 years (range 21–36 years), and mean gestational age was 25 ± 6.2 weeks (range 16–32 weeks). Mean gravidity and parity were 2.4 ± 1.1 and 1.2 ± 0.9, respectively. Mean BMI was 26.2 ± 4.7 kg/m2, mean HbA1c level 6.1 ± 1.2% (normal range 4.5–5.7%), and mean fructosamine level 276 ± 29 mg/dL (normal range 205–285 mg/dL).
All patients completed the 3-day study. There were no adverse events associated with the use of continuous glucose monitoring. None of the patients experienced irritation or infection at the insertion site. Although patients were blinded to the continuous glucose monitoring readings during the study period, they reported high satisfaction using the device concerning potential future benefits of continual monitoring.
An average of 780 ± 54 glucose measurements was recorded for each patient with continuous glucose monitoring. The mean total time of hyperglycemia (glucose level greater than 140 mg/dL) undetected by the finger-stick method was 192 ± 28 minutes per day. Nocturnal hypoglycemic events (glucose level less than 50 mg/dL) were recorded in 26 patients; in all cases, there was an interval of 1–4 hours before clinical manifestations appeared or the event was revealed by random blood glucose examination. Mean glucose level by self–blood glucose monitoring and continuous glucose monitoring was 101 ± 13 mg/dL and 121 ± 13 mg/dL, respectively (P = .02). The individual glucose levels varied widely during each 24-hour period, but the overall 3-day profile of each patient remained consistent in many occasions during this time period (Figure 1).
Analysis of the whole study group for the total 102 days of continuous monitoring showed that all 34 patients had undetected hyperglycemia by self–blood glucose monitoring, with a range of 74–303 minutes per day (mean 192 ± 28 minutes per day, median 166 minutes).
Nocturnal hypoglycemic events (at bedtime, during the night, and in the early morning) were recorded in 26 of the 34 patients in 58 nights; symptoms occurred in 28 episodes in 17 patients. In all affected patients, there was an interval of 1–4 hours before clinical manifestations appeared or the event was revealed by random blood glucose examination.
There was no statistically significant correlation between HbA1c and fructosamine levels and the occurrence of hypoglycemic events.
Figure 2 demonstrates a 24-hour continuous recording in one of the patients.
In 24 of the 34 patients (70%), the physician recommended that the insulin regimen formulated on the basis of the self–blood glucose monitoring data be changed on subsequent evaluation of the continuous glucose monitoring data. The most common change made was in decreasing the long or intermediate-acting insulin dosage at night (mean reduction by 25% in the nighttime dose of insulin).
The correlation coefficient (r) between the glucose measurements by the sensor and meter was .93 ± .04, and between the plasma glucose, meter monitoring, and sensor recording, .91 ± .02. The reliability coefficient was .88.
Despite years of meticulous study, there is still a paucity of information regarding the optimal level of glycemia in diabetic pregnancy that clinicians should target to safely reduce maternal and perinatal morbidity. Strict metabolic control in patients with type 1 diabetes has been associated with an increased risk of maternal hypoglycemia. In our study, continuous monitoring of blood glucose in women with diabetic pregnancies confirmed the high occurrence rate of nocturnal hypoglycemia suspected in earlier studies. Rosenn et al 5 reported significant hypoglycemia, defined as hypoglycemia requiring assistance from another person, in 71% of gravid patients with type 1 diabetes, with a peak incidence in the first trimester. The impact of maternal hypoglycemia on human fetal development and neonatal outcome has not been extensively studied. Although concern about the hazards of hypoglycemia is related primarily to the pregnant mother, the potential effects on the developing fetus need to be considered as well.
Studies in rat and mice embryos point to a possible fetal risk of malformations in the presence of short- and long-term maternal hypoglycemia. 8
These findings were not apparent with self–blood glucose monitoring. Indeed, Data derived from continuous glucose monitoring led the evaluating physician to change the insulin regimen, usually by decreasing the nighttime dose of intermediate-acting insulin. In more than half the cases, the hypoglycemic events were subclinical, diagnosed only by continuous glucose monitoring, and in one fifth of the patients, more than one hypoglycemic event occurred during the night.
These findings may indicate that continuous glucose monitoring is a better monitoring method than self–blood glucose monitoring in detecting hypoglycemic events, which are usually asymptomatic and occur at night. Whereas stringent glycemic control may lead to hypoglycemia, too-loose control poses a risk of macrosomia, the most common and significant perinatal complication associated with diabetic pregnancy, which can lead to an increased risk of birth injuries and asphyxia. The risk of macrosomia rises as maternal glycemia increases. In addition, intensified management of gestational diabetes reduces the rate of perinatal complications, normalizes birth weight, 9 and has a positive influence on the congenital malformation rate. 3
Our study showed that glucose levels were above the upper normal threshold for many hours during the day. These events were related to unscheduled meals and between-meal snacks and were not detected by conventional self–blood glucose monitoring protocols. Despite the recent introduction of intensified treatment protocols, high rates of macrosomia and perinatal morbidity have persisted. On the basis of our study, we speculate that this may be due to the imperfect evaluation of the daily glucose profile by self–blood glucose monitoring, which underestimates hyperglycemic events. Polansky et al 10 reported that a large dinner might be associated with a sustained postprandial state for 4.7 hours. In addition, the most rigorous monitoring protocols only require postprandial glucose measurements three times per day. Many patients indulge in large between-meal snacks, and these may be the cause of the hidden hyperglycemia.
High levels of HbA1c and fructosamine and elevated BMI (greater than 30 kg/m2) were all unrelated to the total diurnal time of hyperglycemia. The lack of a strong correlation between HbA1c and glucose levels by continuous glucose monitoring may indicate that plasma blood glucose levels vary significantly day by day, and although continuous monitoring is more informative than sporadic, nonlongitudinal glucose monitoring, it cannot adequately describe daily glucose profiles over an 8–10-week period. It is possible that HbA1c is a better predictor of preprandial than postprandial glucose levels, as more hours are spent in interprandial and nocturnal periods than in the postprandial periods.
In our study, we used continuous glucose monitoring as a guide for therapy adjustment. Recently, Kaufman et al 11 showed that continuous glucose monitoring could serve as a clinical tool for clinical decision making and glycemic control in children with type 1 diabetes. In another recent work, Hershkovitz et al 12 demonstrated the clinical implications of continuous glucose monitoring use to assess and manage asymptomatic hypoglycemic events in children with glycogen storage disease.
Parretti et al 13 assessed the daily glycemic profile of nondiabetic, nonobese pregnant women using multiple daily fingerstick measurements. A comparison by continuous glucose monitoring in our department in a nondiabetic, nonobese gravid population yielded a remarkably similar glycemic profile (unpublished data).
The present prospective study is the first to test the application of continuous glucose monitoring on gravid women with type 1 diabetes. It showed that the diurnal glucose profile undergoes extreme changes, including both hypoglycemic and hyperglycemic events, which are unrecognized by the conventional self–blood glucose monitoring technique. Continuous monitoring profiles allow the physician to identify glucose patterns and to better target insulin treatment. The treatment changes in our series would not have been made on the basis of meter data alone.
At this point, we do not recommend that continuous glucose monitoring replace self–blood glucose monitoring, but we suggest that intermittent application of continuous glucose monitoring can be used to fine-tune glycemic control, assess patient compliance, if necessary, and prevent nocturnal hypoglycemic events.
Because the incidence of hypoglycemia is higher and achieving strict glycemic control is more difficult in the first trimester of pregnancy, the frequency of continuous glucose monitoring should be increased during this period. Later applications vary individually.
Importantly, this study does not indicate a clinical difference in outcome from self–blood glucose monitoring, and the clinical utility of continuous glucose monitoring in improving perinatal outcome remains unknown.
A large, prospective study on maternal and neonatal outcome is needed to evaluate the clinical implications of this new monitoring technique.
1. Karlsson K, Kjellmer I. The outcome of diabetic pregnancies in relation to the mother's blood sugar level. Am J Obstet Gynecol 1972;112:213–20.
2. Landon MB, Gabbe SG, Piana R, Menutti MT, Main EK. Neonatal morbidity in pregnancy complicated by diabetes mellitus: Predictive value of maternal glycemic profiles. Am J Obstet Gynecol 1987;156:1089–95.
3. Damm P, Molsted-Pedersen L. Significant decrease in congenital malformations in newborn infants of an unselected population of diabetic women. Am J Obstet Gynecol 1989;161:1163–7.
4. Cryer PE. Iatrogenic hypoglycemia as a cause of hypoglycemia-associated autonomic failure in IDDM: A vicious cycle. Diabetes 1992;41:255–60.
5. Rosenn BM, Miodovnik M, Holcberg G, Khoury JC, Siddiqi TA. Hypoglycemia: The price of intensive insulin therapy for pregnant women with insulin-dependent diabetes mellitus. Obstet Gynecol 1995;85:417–22.
6. Jovanovic L, Reed GF, Metzger BE, Mills JL, Knopp RH, Aarons JH. Maternal postprandial glucose levels and infant birth weight: The Diabetes in Early Pregnancy Study. The National Institute of Child Health and Human Development—Diabetes in Early Pregnancy Study. Am J Obstet Gynecol 1991;164:103–11.
7. Rebrin H, Steil GM, Van Antwep WP, Mastortotoro JJ. Subcutaneous glucose predicts plasma glucose independent of insulin: Implications for continuous monitoring. Am J Physiol 1997;277:E561–71.
8. Akazawa S, Akazawa M, Hashimoto M, Yamaguchi Y, Kuriya N, Toyama K, et al. Effects of hypoglycaemia on early embryogenesis in rat embryo organ culture. Diabetologia 1987;30:791–6.
9. Jovanovic L, Bevier W, Peterson CM. The Santa Barbara County Health Care Services Program: Birth weight change concomitant with screening for and treatment of glucose-intolerance of pregnancy: A potential cost-effective intervention. Am J Perinatol 1997;14:221–8.
10. Polansky KS, Given BD, Hirsch LJ, Tillil H, Shapiro ET, Beebe C, et al. Abnormal patterns of insulin secretion in non-insulin-dependent diabetes mellitus. N Engl J Med 1988;318:1231–9.
11. Kaufman FR, Gibson LC, Halvorson M, Carpenter S, Fisher LK, Pitukcheewanont P. A pilot study of the continuous glucose monitoring system: Clinical decisions and glycemic control after its use in pediatric type 1 diabetic subjects. Diabetes Care 2001;24:2030–4.
12. Hershkovitz E, Rachmel A, Ben-Zaken H, Phillip M. Continuous glucose monitoring in children with glycogen storage disease type-1. J Inherit Metab Dis 2001;24:863–9.
13. Parretti E, Mecacci F, Papini M, Cioni R, Carignani L, Mignosa M, et al. Third-trimester maternal glucose levels from diurnal profiles in nondiabetic pregnancies: Correlation with sonographic parameters of fetal growth. Diabetes Care 2001;24:1319–23.