HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Kahn, B. F. Articles by Barbour, L. A. Search for Related Content PubMed PubMed Citation Articles by Kahn, B. F. Articles by Barbour, L. A. Related Collections Diabetes Medical complications of pregnancy

Predictors of Glyburide Failure in the Treatment of Gestational Diabetes

From the ¹Department of Obstetrics and Gynecology, Division of Maternal-Fetal Medicine; ²Department of Obstetrics and Gynecology, Section of Basic Reproductive Science; ³Department of Preventive Medicine and Biometrics; ⁴Department of Pediatrics; and ⁵Department of Medicine, Division of Endocrinology, Diabetes, and Metabolism, University of Colorado Health at Denver and Health Sciences Center, Denver Colorado.

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
OBJECTIVE: Our objective was to identify among women with gestational diabetes mellitus (GDM) the patient characteristics that predict treatment failure with glyburide.

METHODS: Historical cohort of 95 GDM women offered glyburide after dietary failure with defined entry criteria.

RESULTS: From November 2000 to May 2005, 118 women had 124 pregnancies and were offered glyburide therapy by the 2 codirectors of our Diabetes Clinic. All but 2 women elected glyburide, and 27 pregnancies were excluded due to criteria defined a priori to the study. A cohort of 95 women with 95 pregnancies were included for analysis. Nineteen percent failed glyburide. Significant predictors of failure were maternal age (34 years compared with 29 years, P = .001), earlier diagnosis of GDM (23 weeks compared with 28 weeks, P = .002), higher gravidity (P = .01) and parity (P = .03), and a higher mean fasting blood glucose (112 compared with 100 mg/dL; P = .045) compared with those successfully treated. After adjustment in the multivariable logistic regression analysis, GDM women diagnosed at a gestational age less than 25 weeks were 8.3 times more likely to fail glyburide compared with those diagnosed after 25 weeks. Maternal and fetal outcomes were favorable with a cesarean delivery rate of 25% and macrosomia rate of 7%.

CONCLUSION: Glyburide was more likely to fail in women diagnosed earlier in pregnancy, of older age and multiparity, and with higher fasting glucoses, suggesting that earlier glucose intolerance and a reduced capacity to respond to an insulin secretagogue may distinguish this group. The time for glyburide as an alternative treatment has come; however, it should be prescribed after careful consideration of these patient characteristics to minimize the likelihood of failure.

LEVEL OF EVIDENCE: II-2

However, the question left unanswered and raised in the editorial comment concerning Jacobson's recent study is whether glyburide is appropriate for all GDM women who fail diet.⁹ The author of the editorial expressed the need for a trial of sufficient sample size that could describe the clinical characteristics associated with glyburide failure.⁹

Previous published trials have been of insufficient sample size to develop a regression model of predictors of glyburide success compared with failure.^4,5 Because decisions are being made by consensus panels whether to endorse glyburide as an alternative treatment for GDM, information to guide the practitioner as to the likelihood of glyburide success or failure would be extremely useful. Such predictors would be valuable to avoid weeks of inadequate glycemic control while doses of glyburide are titrated upward in women destined to fail therapy. We evaluated the predictors of glyburide failure using univariable and multivariable logistic regression analysis in addition to the effectiveness and fetal safety in a cohort of 95 women, of whom about one half were Hispanic. Exclusion and failure criteria were defined before the study and patient management was directly supervised by the 2 clinic codirectors to achieve a standardized approach to care. We set out to define clinical characteristics that predict glyburide failure compared with success in an attempt to describe the most optimal candidates for glyburide treatment.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
After the publication of the Langer et al³ RCT, we offered glyburide to GDM women in whom dietary therapy failed. All women are screened for GDM at 24–28 weeks of gestation and are offered earlier screening if they have a history of GDM, morbid obesity, a strong family history of diabetes, or characteristics of insulin resistance. All GDM patients attend the Diabetes in Pregnancy Clinic at the University of Colorado at Denver and Health Sciences Center, codirected by a perinatologist (J.D.) and an endocrinologist (L.B.). These 2 attending physicians decided at the beginning of the study to offer glyburide to their own clinic patients failing diet if they had no history of pregestational diabetes and did not have a glycosylated hemoglobin (HgbA1c) of more than 7%, suggesting pregestational diabetes. Glyburide therapy was not initiated by the 2 codirectors if the patient's fasting blood glucose was 125 mg/dL or more, due to concerns that glyburide was likely to be ineffective or that these patients might have undiagnosed pregestational diabetes. However, patients referred to the Diabetes in Pregnancy Clinic already on glyburide were included in the analysis irrespective of their fasting blood glucose as long as their fasting blood glucose levels were adequately controlled upon transfer to our clinic. Despite explaining that glyburide was unapproved in pregnancy, all patients but 2 elected glyburide over insulin, both of whom had taken insulin in their prior pregnancies.

Gestational diabetes was diagnosed using the Carpenter and Coustan criteria in which the patients had a least 2 abnormal values on a diagnostic 100-g, 3-hour oral glucose tolerance test (GTT) using a fasting blood glucose 95 mg/dL or more; 1 hour 180 mg/dL or more; 2 hours 155 mg/dL or more; and 3 hours 140 mg/dL or more. Women with a 50-g glucose challenge more than 200 mg/dL and who also had a fasting blood glucose 95 mg/dL or more, were diagnosed as having GDM. Patients were given diet counseling by a diabetes nurse educator and nutritionist and taught to test their own blood glucose levels four times daily, fasting and 1 hour after each meal. Because the majority of the cohort was overweight or obese, a weight gain of 15–25 pounds was recommended.

All GDM patients requiring medical treatment are managed in our clinic and delivered at the University of Colorado Hospital. Patients were seen by obstetric residents and maternal–fetal medicine fellows, but 1 of the 2 clinic codirectors attended and approved management decisions to maintain standardized care for all patients. Women with GDM were seen at least every other week.

Patients were considered to fail diet if after approximately 2 weeks of diet therapy (carbohydrate restriction to 45% total calories), 20% of fasting blood glucose values were 95 mg/dL or more or 1 hour postprandials (1 hour after meals) were 140 mg/dL or more. Patients were instructed to take the glyburide 30 minutes before their breakfast and dinner. The initial dose of glyburide was individualized based on the patient's weight and degree of hyperglycemia using targets of a fasting blood glucose less than 95 mg/dL and 1 hour postprandial less than 140 mg/dL. Failure of glyburide was determined by the 2 clinic codirectors and defined as the inability to maintain at least 80% of the fasting blood glucose and 1-hour postprandial glucoses in the target range using up to a maximal dose of 10 mg twice daily for at least 1 week. Glyburide was titrated upward to a maximum of 20 mg per day to achieve optimal glycemic control if any glycemic improvement was noted on the 10-mg-per-day dose.

Secondary outcome measures included preeclampsia, maternal hypoglycemia, delivery route, birth weight, dystocia, and neonatal complications, including admission to the neonatal intensive care unit (NICU), respiratory distress syndrome, and neonatal hypoglycemia (glucose < 40 mg/dL) that inadequately responded to more frequent oral feeds and required intravenous glucose.

After approval by the Colorado Institutional Review Board, we performed a historical cohort study of all diabetic pregnancies which were delivered between November 2000 and March 2005, during which time this protocol was instituted. To ensure that all GDM patients were captured for analysis, we identified our cohort by International Classification of Diseases, 9th Revision codes through medical records and the billing department as well as by the patient rosters from our Diabetes in Pregnancy Clinic.

Patients were excluded from analysis if they had a multiple gestation, serious preexisting maternal illness, if they delivered at an outside hospital, or if they had fewer than 2 visits to our clinic. All patient visits were dictated, and both the computerized clinic notes as well as the patient charts were reviewed to optimize data retrieval. Demographic data, medication use, pertinent medical and obstetrical history, glucose values, ultrasound results, and maternal and neonatal outcomes were all recorded in a standardized data sheet used by the investigators to facilitate data retrieval.

It is our practice to obtain fetal growth scans at 28–32 weeks gestation. If the fetus demonstrated an abdominal circumference more than 90th percentile or was large for gestational age (LGA) on a 28–32 week growth scan, the codirectors of the clinic factored this into the decision to increase the dose of glyburide in a women whose blood glucoses were suboptimal.¹⁰ However, no patients were switched to insulin for only this indication and must have failed glyburide according to the designated glycemic criteria. Induction of labor was offered after the completion of 39 weeks if the patient met adequate dating criteria, had a favorable cervix, and was without obstetric contraindications. Maternal body mass index (BMI) was calculated using the earliest available weight. Macrosomia was defined as birth weight greater than 4000 g and LGA as greater than the 90th percentile for gestational age. Neonatal hypoglycemia was defined as any glucose measurement (whole blood or heel-stick) of less than 40 mg/dL unresponsive to more frequent feeds and requiring intravenous glucose.

The data were analyzed in SAS 9.1 (SAS Institute, Cary, NC). We initially generated general descriptive data for the cohort. Potential differences in the clinical characteristics of the women who failed glyburide or had success with glyburide were then compared. Associations between categorical variables were tested using the ² or Fisher exact test. Differences in means of continuous variables were tested using the Student t test or its nonparametric equivalent, the Wilcoxon rank sum test (when the data were not normally distributed). Statistics are presented with 95% confidence intervals (CIs) (P < .05). Variables identified in the univariable analysis as being significantly associated or of borderline significant association with the outcome were included in the multivariable logistic regression analysis. Significant continuous variables were categorized along clinically meaningful lines. Multivariable logistic regression analysis was then used to determine the unadjusted and adjusted odds ratio (OR) of these select risk factors for failure of glyburide treatment.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
We reviewed records of 465 pregnancies with all types of diabetes, 124 of which were Class A2 gestational diabetics (GDM pregnancies in which diet alone failed) for which glyburide was offered. Two declined glyburide and of the 122 pregnancies treated with glyburide, 27 exclusions were comprised of 7 women who took glyburide for less than 2 weeks, 2 with serious maternal illness, 2 multiple gestations, 4 who delivered at outside hospitals, 1 woman referred to our practice with an unclear GDM diagnosis, and 5 who delivered at University of Colorado at Denver and Health Sciences Center but had no prenatal records available from referral sources. There were 6 women who had 2 separate pregnancies, both treated with glyburide during the study period, and only the first pregnancy of each was included. The remaining 95 women qualified for inclusion. Seven patients had a fasting blood glucose more than 125 mg/dL, but were treated with glyburide by outside practices before referral into our system. These patients transferred their care to our clinic and were included in our analysis given that they had acceptable control with glyburide at the time of transfer. Data were missing on maternal BMI in 6 women and on the FBS in 12 women. The median starting dose of glyburide was 2.5 mg twice daily, the maximal dose was 10 mg twice daily, and the mean dose at delivery was 5 mg twice daily.

Fifty women (53%) were Hispanic, 32 (34%) Non-Hispanic white, 8 (8%) African American, and the remaining 5 women were of other ethnicities. Other characteristics of the cohort are shown in Table 1. Seventy-three percent were overweight or obese, with a BMI of 26 or more. The mean age of the women was 30 years, and the mean fasting blood glucose on the diagnostic GTT was 102 mg/dL. Statistically significant variables predicting glyburide failure in the analysis included older age, earlier diagnosis of GDM, earlier gestational age at the start of glyburide treatment, higher gravidity and parity, and a higher mean fasting blood glucose on the 3-hour GTT compared with women who were successfully treated (Table 1).

Maternal and fetal outcomes are listed in Table 2. Eighty-one percent (77/95) of pregnancies were treated successfully with glyburide. Overall, 7% of the total cohort developed problems with recurrent hypoglycemia (glucose < 60 mg/dL on several occasions despite dietary manipulation), and 2% were switched to insulin for that reason. Nineteen percent of the cohort failed glyburide, of which 90% were on a dose of at least 10 mg per day.

Of the 18 patients who were switched to insulin, 45% achieved adequate control on insulin (one of whom was controlled on a combination of glyburide and single dose of neutral protamine hagedorn before bedtime for elevated fasting blood glucose levels). Of the 11 patients (55%) with inadequate control despite being switched to insulin, 2 were judged to be noncompliant (by dietary and treatment diaries). One third of them (4/11) were diagnosed before 14 weeks and were thought to have preexisting glucose intolerance, highly suggestive that these women had the most severe diabetes. Hemoglobin A1C values in 3 of 4 of these patients were 5.5 or less, and the other had a value of 6.7 (normal < 6.3).

The incidence of preeclampsia was 12%, and the overall cesarean delivery rate was 25%. There were no perinatal deaths or reports of shoulder dystocia. The macrosomia rate was 7%, and 27% of the infants were LGA. The mean birth weight was 3,309 g at a mean gestational age at delivery of 38 weeks. Neonatal outcomes were favorable, with an incidence of respiratory distress syndrome of 9% and neonatal hypoglycemia requiring intravenous glucose of 12%. There were 3 anomalies reported (ventricular septal defect, hypospadias, and hypoplastic left heart syndrome); however, glyburide was not started until after 22 weeks gestation in these patients.

Select risk factors from Table 1 were included in the multivariable logistic regression analysis. Gestational age at commencement of glyburide was strongly associated with the gestational age of diagnosis of GDM (Pearson correlation coefficient of 0.86). The latter variable was considered more clinically useful so was kept in the model. For similar reasons we chose to include parity over gravidity. The mean fasting blood glucose on the 3-hour oral GTT was not included in the multivariable model due to missing values in 12 patients. Maternal age was dichotomized at the 75th percentile and gestational age of diagnosis at 25 weeks of gestation (Table 3).

After adjustment for older maternal age and multiparity, women who had gestational diabetes diagnosed at less than 25 weeks gestation were 8.3 times more likely to fail glyburide treatment compared with women who had gestational diabetes diagnosed after 25 weeks. Glycosylated hemoglobin (HgA1C) levels were documented in 76% (13/17) of these women, which ranged from 4.7–6.7%.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
The striking finding in our study is that women diagnosed with GDM at less than 25 weeks gestation were greater than 8-fold more likely to fail glyburide, even after adjustment for other covariates (Table 3). These women were unlikely to have had unrecognized pregestational diabetes because we excluded women with an HgA1C more than 7.0 and women with a fasting blood glucose more than 125 mg/dL, with the exception of 7 women with elevated fasting blood glucose values referred to us already on glyburide with acceptable glycemic control. Age 34 years or older also remained highly statistically significant in the multivariable logistic regression model. Other predictors of glyburide failure included higher gravidity and parity and a higher fasting blood glucose on the diagnostic 3-hour GTT. We may have seen an even stronger association with fasting blood glucose and glyburide failure if we had offered glyburide to our own clinic patients with a fasting blood glucose more than 125 mg/dL, but both clinic codirectors were concerned that such patients were likely to fail or could have undiagnosed pregestational diabetes. Seven patients were referred to us with fasting blood glucose more than 125 mg/dL already on glyburide and, because they were in adequate control at transfer of care, they were included in the analysis. The vast majority of our patients were overweight or obese, likely accounting for the lack of a statistically significant association with BMI (P = .22). The characteristics of women who fail glyburide suggest they are more likely to have either preexisting or more severe glucose intolerance and less likely to respond to an insulin secretagogue than women who succeed. We believe the identification of such characteristics associated with failure may be particularly helpful in selecting ideal candidates for glyburide to avoid treatment for an extended period of time with an agent unlikely to be successful.

Eighty-nine percent (16/18) of patients who failed glyburide were on a total dose of at least 10 mg per day and 33% (6/18) were on a maximal dose of 20 mg per day. The patients deemed as glyburide failures on doses less than 20 mg per day had an extremely poor response to the 10 mg per day and further dose escalation was thought highly likely to fail. In Langer et al's³ RCT, 71% of the patients were treated successfully with a 10-mg total dose, and this is consistent with other authors who used a similar dose at the time the patients were considered glyburide failures.^5,11 Furthermore, the maximal daily glyburide dose that achieved acceptable glycemic control in the large cohort of 229 patients treated with glyburide in Jacobson et al's⁸ study was only 5.6 mg ± 4.6 mg/d.

Although a limitation of our historical cohort study is its retrospective retrieval of the data, a strength is that the management of these patients was standardized. The 2 clinic codirectors directly supervised all treatment decisions and determined dose titration and glyburide failure. Furthermore, the protocol was initiated in 2000 before the start of the trial, and the criteria for offering glyburide to our patients was decided a priori by these codirectors. Only 2 women declined glyburide therapy and opted for insulin, despite warnings about its lack of approval in pregnancy. This extremely high patient preference for an oral agent is impressive, similar to other reported studies,^4,5,8,11 and minimizes the possibility of a selection bias confounding the characteristics identified with glyburide failure.

Our failure rate is similar to the majority of studies published after Langer et al's RCT.^3–5,8,11 However, our study is unique to previously published trials exploring predictors of glyburide success due to its larger sample size, design, and patient characteristics. Chmait et al⁴ reported on 69 patients in whom the majority were Hispanic (89%) and noted that dietary failure occurring before 30 weeks was statistically significant in predicting glyburide failure, but there was not a clear attempt to screen out unrecognized pregestational diabetes. Higher fasting and postprandial glucose meter values in the week preceding the initiation of glyburide reached statistical significance, as did older age.

Conway et al⁵ reported their retrospective findings in 75 women in San Antonio, primarily a Hispanic population, treated with glyburide after the Langer et al RCT. Failures also had higher glucose values on their oral GTT but receiver operating characteristic curves revealed no clear cut off points for predicting failure. They also found that women who failed were more likely to have treatment initiated earlier in pregnancy.

Of the 18 patients in our study who failed glyburide and were switched to insulin, approximately one half achieved adequate control. This subgroup might have achieved acceptable glycemic control with earlier institution of insulin, although a prospective study in which insulin is used aggressively in all women who are diagnosed early and require therapy early would be necessary to test this hypothesis.

Our maternal and fetal outcomes compare favorably with other large studies by Langer et al³ and most recently by Jacobson et al.⁸ Our patient population is different from the ones described in these 2 large studies given 83% of Langer's population was Hispanic and the largest ethnic population given glyburide in the Kaiser study by Jacobson et al was Asian (37%). Our overall cesarean delivery rate (25%) was similar to that reported by Langer (23%) and significantly lower than the 39% rate reported by Jacobson. We had a relatively low incidence of macrosomia in our study at only 7%, identical to the Langer series, but significantly lower than that (25%) reported in Jacobson's study. This may be due, in part, to our practice of considering fetal growth in the decision to upward titrate the glyburide dose and to offer induction at approximately 39 weeks for selected patients.¹⁰ Neonatal outcomes were also reassuring in our study (Table 2) and similar to other studies.

Glyburide is definitely preferred by patients in our study as well as in many others, and its relative safety has been confirmed.^{3–5,8,11,12} Clearly, there is a need for an alternative to insulin that is lower in cost associated with high compliance, given that the incidence of GDM is rising dramatically with the obesity epidemic. However, treating GDM patients with high-risk characteristics for glyburide failure could delay the achievement of optimal glycemic control. Given that there is now compelling data from a recent blinded RCT using insulin and diet that optimal GDM treatment clearly improves serious fetal morbidity,² it is particularly important to stratify prospectively which patients are likely to fail glyburide. There are also data that GDM may have long-term adverse implications on the health of the offspring, including adolescent obesity and impaired glucose tolerance,¹³ providing additional arguments for the optimal treatment of GDM mothers.

Approximately 80% of women with GDM may be successfully and safely treated with glyburide and it is an attractive option for many women. However, it is not for everyone, and older, multiparous women with higher fasting blood glucose values demonstrating impaired glucose tolerance earlier in pregnancy are much more likely to fail glyburide. This likely represents a population with more advanced insulin resistance in which an oral secretagogue cannot stimulate a sufficient insulin response to sustain euglycemia. We are in favor of recommending glyburide in the treatment of GDM in a significant subset of patients. However, we would caution its use particularly in women who are diagnosed earlier in pregnancy, to minimize the risk of failure and delay achievement of adequate glycemic control.

	Footnotes

 
The authors thank Dr. Henry Galan for assistance in designing the methodology for data collection, Dr. Albert Schaefer for contributions in developing the glyburide data base, and Brenda Camp for assistance in retrieving data.

Corresponding author: Linda A. Barbour, MD, MSPH, UCDHSC at Fitzsimons, Division of Endocrinology, Metabolism, and Diabetes, MS 8106; PO Box 6511, RC1-South, Room L-18 7103, 12801 East 17th Avenue, Aurora, CO 80045; e-mail: Lynn.Barbour{at}uchsc.edu.

doi:10.1097/01.AOG.0000218704.28313.36

	REFERENCES

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
1. American Diabetes Association. Gestational diabetes mellitus. Diabetes Care 2004;27 suppl:88–90.

2. Crowther CA, Hiller JE, Moss JR, McPhee AJ, Jeffries WS, Robinson JS, et al. Effect of treatment of gestational diabetes mellitus on pregnancy outcomes. N Engl J Med 2005;352:2477–86.[Abstract/Free Full Text]

3. Langer O, Conway DL, Berkus MD, Xenakis EM, Gonzales O. A comparison of glyburide and insulin in women with gestational diabetes mellitus. N Engl J Med 2000;343:1134–8.[Abstract/Free Full Text]

4. Chmait R, Dinise T, Moore T. Prospective observational study to establish predictors of glyburide success in women with gestational diabetes mellitus. J Perinatol 2004;24:617–22.[Medline]

5. Conway DL, Gonzales O, Skiver D. Use of glyburide for the treatment of gestational diabetes: the San Antonio experience. J Matern Fetal Neonatal Med 2004;15:51–5.[Medline]

6. Langer O, Yogev Y, Xenakis EM, Rosenn B. Insulin and glyburide therapy: dosage, severity level of gestational diabetes, and pregnancy outcome. Am J Obstet Gynecol 2005;192:134–9.[Medline]

7. Gabbe S, Gregory R, Power M, Williams S, Schulkin J. Management of diabetes mellitus by obstetrician-gynecologists. Obstet Gynecol 2004;103:1229–34.[Abstract/Free Full Text]

8. Jacobson GF, Ramos GA, Ching JY, Kirby RS, Ferrara A, Field DR. Comparison of glyburide and insulin for the management of gestational diabetes in a large managed care organization. Am J Obstet Gynecol 2005;193:118–24.[Medline]

9. Durnwald C, Landon MB. Glyburide: The new alternative for treating gestational diabetes?. Am J Obstet Gynecol 2005;193:1–2.[Medline]

10. Schaefer-Graf UM, Kjos SL, Fauzan OH, Buhling KJ, Siebert G, Buhrer C, et al. A randomized trial evaluating a predominately fetal growth-based strategy to guide management of gestational diabetes in Caucasian women. Diabetes Care 2004;27:297–302.[Abstract/Free Full Text]

11. Kremer CJ, Duff P. Glyburide for the treatment of gestational diabetes. Am J Obstet Gynecol 2004;190:1438–9.[Medline]

12. 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.[Abstract/Free Full Text]

13. Barbour LA. New concepts in insulin resistance of pregnancy and gestational diabetes: long-term implications for mother and offspring. J Obstet Gynaecol 2003;23:545–9.[Medline]

This article has been cited by other articles:

				D. S Feig, G. G Briggs, and G. Koren Oral Antidiabetic Agents in Pregnancy and Lactation: A Paradigm Shift? Ann. Pharmacother., July 1, 2007; 41(7): 1174 - 1180. [Abstract] [Full Text] [PDF]

				M. D. Lindheimer, A. Khurana, N. Vinayek, C. A. Stratakis, A. Klibanski, and M. F. Greene Case 36-2006: A Pregnant Woman with New Hypertension N. Engl. J. Med., March 1, 2007; 356(9): 966 - 968. [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Kahn, B. F. Articles by Barbour, L. A. Search for Related Content PubMed PubMed Citation Articles by Kahn, B. F. Articles by Barbour, L. A. Related Collections Diabetes Medical complications of pregnancy