Non-Randomized Controlled Trial

C - Click here for explanation of classification scheme.

NEUTRAL: See Quality Criteria Checklist below.

To investigate the relationship between pregnancy outcome and pre-pregnancy overweight or obesity, level of glycemic control and treatment modality (diet or insulin) in GDM women.

All women attending a health clinic from January 1990 through September 1999 who were diagnosed with GDM (two or more abnormal plasma glucose values) before 33 weeks of gestation.

None reported.

• Recruitment: All women attending a health clinic from January 1990 through September 1999 were screened with a one-hour, 50-gram OGTT. Women whose one-hour result was 130mg per dL or more underwent a 100-gram OGTT.
• Design: Non-randomized clinical trial
• Blinding used: Not applicable.

Intervention

• A team of maternal fetal medicine sub-specialists, residents, diabetic nurse educators, dietitians and social workers provided care during pregnancy. The same team using the same diabetic protocol managed all patients.
• Diet treatment: A dietitian instructed participants about a three-meal, four-snack standard daily meal plan. Caloric restriction was prescribed for obese women, based on 25kcal per kg for actual maternal weight in pregnancy and 35kcal per kg for non-obese subjects. Approximately 40% to 45% of calories were derived from CHO.
• Patients who were unable to achieve the established levels of glycemic control on diet therapy were assigned to insulin treatment within two weeks of failed diet therapy, based on physician's assessment of level of glucose control
• Insulin treatment: Starting dose was 0.7 units per kg of actual body weight at admission, administered three times daily and adjusted weekly, as needed.

Statistical Analysis

• Stratified analysis by bivariate (unadjusted), chi-square, logistic regression and ANOVA testing: Four groups, according to glucose control [Good (mean blood glucose less than 100mg per dL); Poor (100mg per dL or more)] and treatment (Diet; Diet Plus Insulin) and three pre-pregnancy BMI categories [Normal Weight (BMI 18.5 to 24.9); Overweight (BMI 25 to 29.9); Obese (BMI 30 or higher)]
• Multiple variable logistic regression: Adjusting for potential covariates and confounders (see Control Variables).

Timing of Measurements

• All patients were assigned memory reflectance glucometers and instructed by a diabetic educator in self-monitoring blood glucose technique (seven times daily: Fasting, pre-meal, two hours post-meal and at bedtime)
• For quality control, blood glucose was measured in the laboratory at weekly clinic visits.

Dependent Variables

Composite outcome

More than one of the following: Metabolic complications, LGA, NICU admission for more than 24 hours and the need for respiratory support (not including oxygen therapy).

For mothers

• Total pregnancy weight gain: Based on pre-pregnancy weight and the last weight measured within a week before delivery
• Labor induction
• Mode of delivery: Elective cesarean delivery is defined as delivery of patients in which vaginal delivery was contraindicated for obstetric or medical indications, e.g., fetal macrosomia or patient request in response to previous cesarean section
• Pre-eclampsia: An SBP over 140mmHg, DBP over 90mmHg and proteinuria (mover 30mg per dL or over 1+ on a dipstick) on more than two occasions at least six hours apart after the 20th week of gestation
• Chronic hypertension: History of hypertension or hypertensive therapy before pregnancy.

For baby

• Gestational age (GA) at delivery: GA was determined by menstrual history in combination with early vaginal examination and early ultrasound examination before 20 weeks of gestation when available
• Birth weight percentile for GA: Large-for-gestational age (LGA) infants had birth weight in at least the 90th percentile, based on growth standards developed for the San Antonio population. Macrosomia was defined as birth weight of at least 4,000g.
• Respiratory and metabolic morbidity: Neonatal respiratory outcomes included presence or absence of hyaline membrane disease and transient tachypnea. Metabolic complications included more than one abnormal factor from polycythemia, hypoglycemia and hyperbilirubinemia.
• Neonatal glucose, hemoglobin and bilirubin determinations: Glucose was measured in the heel blood more than three times during the first hour after birth and thereafter every 30 minutes up to four hours. Hypoglycemia was defined as a blood glucose value of not more than 40mg per dL; polycythemia, as a hematocrit of more than 60%; serum bilirubin was measured when there was clinical evidence of jaundice. Hyperbilirubinemia was defined as a serum bilirubin of more than 12mg per dL.

Independent Variables

• Treatment modality: Diet or insulin
• Level of glycemic control: Well controlled (mean blood glucose less than 100mg per dL) vs. poor control (at least 100mg per dL)
• Dietary compliance was evaluated and reinforced at weekly clinic visits
• Overall mean blood glucose levels were defined as mean blood glucose from diagnosis to delivery. The goal was to maintain levels between 90mg and 100mg per dL.

Control Variables

The potential risk indicators included in the multiple logistic regression:

• Maternal age
• Parity
• Previous macrosomia
• Race and ethnicity
• Obesity
• Disease severity (fasting plasma)
• Weight gain in pregnancy
• Level of glycemic control (mean blood glucose, well vs. poor control)
• Treatment modality (diet or insulin)
• Length of treatment
• Gestational age at delivery.

Initial N

4,001 enrolled.

Attrition (Final N)

4,001.

Age

Significantly different between the BMI groups (P=0.01).

• BMI 18.5 to 24.9: 28.1 years
• BMI 25 to 29.9: 29.5 years
• BMI 30 or greater: 29.7 years of age.

Ethnicity

No difference between the BMI groups

• Hispanic: 84%
• White (non-Hispanic): 12%
• African American: 4%.

Other Relevant Demographics

• Previous macrosomia rate
  • BMI 18.5 to 24.9: 17%
  • BMI 25 to 29.9: 25.4%
  • BMI 30 or greater: 27.7
  • P=0.01 for group differences.
• Family history of GDM
  • BMI 18.5 to 24.9: 50.4%
  • BMI 25 to 29.9: 49.6%
  • BMI 30 or greater: 59.6%
  • P=0.01 for group differences.
• Previous GDM
  • BMI 18.5 to 24.9: 15.5%
  • BMI 25 to 29.9: 20.1%
  • BMI 30 or greater: 24.1%
  • P=0.01 for group differences.
• Weight gain in pregnancy (kd)
  • BMI 18.5 to 24.9: 12.7
  • BMI 25 to 29.9: 9.8
  • BMI 30 or greater: 8.1
  • P=0.01 for group differences.
• Total insulin dose
  • BMI 18.5 to 24.9: 69
  • BMI 25 to 29.9: 88
  • BMI 30 or greater: 121
  • P=0.01 for group differences.
• No significant differences in the mean blood glucose levels between the MCI groups
• No significant differences in the two-hour OGTT levels between the MCI groups.

Anthropometrics

• Mean pregnancy BMI: 21.6
• BMI 18.5 to 24.9: 40% (1,609 of 4,001)
• BMI 25 to 29.9: 29% (1,141 of 4,001)
• BMI 30 or greater: 31% (1,251 of 4,001)
• Among women with BMI greater than 30, 550 (44%) had morbid obesity (defined as a BMI of more than 35); 69% were insulin treated; 31% diet-treated. 66% of these women failed to acheived desired levels of glycemic control.

Location

USA.

I. Results from Bivariate (Unadjusted) Analysis

	Reference Group BMI 18.5-24.9	Diet Well Controlled N=1,317	Diet Poorly Controlled N=427	Insulin Well Controlled N=1,130	Insulin Poorly Controlled N=1,124
		OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)
C-Section Rate   (%)	BMI 25-29.9	0.87 (0.63-1.23)	1.18 (0.67-2.11)	0.85 (0.59-1.22)	0.87 (0.63-1.23)
	BMI>30	1.79 (1.29-2.48)	2.33 (1.35-4.02)	1.03 (0.74-1.46)	1.79 (1.29-2.48)
Composite Outcome (%)	BMI 25-29.9	1.58 (1.15-2.17)	1.28 (0.73-2.22)	1.37 (0.93-2.04)	1.05 (0.70-1.60)
	BMI>30	1.69 (1.20-2.39)	2.49 (1.46-4.24)	1.06 (0.72-1.56)	1.66 (1.15-2.40)
LGA (%)	BMI 25-29.9	2.07 (1.45-2.95)	1.56 (0.85-2.87)	1.31 (0.82-2.11)	1.20 (0.74-1.94)
	BMI>30	1.72 (1.15-2.57)	2.99 (1.15-2.57)	1.23 (0.79-1.94)	2.23 (1.47-3.38)
Fetal Macrosomia (%)	BMI 25-29.9	1.45 (0.88-2.41)	1.23 (0.58-2.59)	1.26 (0.67-2.36)	1.27 (0.61-2.63)
	BMI>30	2.12 (1.29-3.50)	2.87 (1.47-5.61)	0.71 (0.36-1.41)	3.01 (1.64-5.52)
Neonatal Metabolic Complications (%)	BMI 25-29.9	0.92 (0.54-1.55)	0.59 (0.21-1.75)	1.70 (0.99-2.91)	1.35 (0.79-2.31)
	BMI>30	1.59 (1.04-2.63)	1.64 (0.69-3.86)	1.22 (0.77-2.13)	1.35 (0.83-2.18)
Pre-Eclampsia (%)	BMI 25-29.9	0.95 (0.42-2.00)	1.85 (0.76-4.65)	0.93 (0.39-2.18)	2.25 (1.29-3.89)
	BMI>30	1.40 (0.86-2.25)	1.81 (0.71-4.65)	1.38 (0.66-2.88)	2.68 (1.61-4.48)
Chronic HTN (%)	BMI 25-29.9	1.83 (1.32-2.55)	1.89 (1.07-3.34)	2.07 (1.42-3.05)	1.59 (1.10-2.30)
	BMI>30	3.34 (1.66-3.31)	2.48 (1.40-4.37)	3.19 (2.24-4.54)	2.04 (1.46-3.86)

II. Multiple Logistic Regression Summary for Composite Outcome and LGA as Dependent Variables for the Impact of Overweight and Obesity on Preinatal Outcome

	Composite Outcome		LGA
	OR	95% CI	OR	95% CI
Treatment Modality	1.28*	1.10-1.63	1.83*	1.38-2.42
Obesity	1.28*	1.08-1.66	1.83*	1.31-2.55
Overweight	1.08	0.88-1.33	1.17	0.92-1.49
Mean Blood Glucose	1.39*	1.14-2.09	1.71*	0.35-2.16
Fasting OGTT	1.03	0.82-1.30	1.34*	1.02-1.76
Parity	1.72*	1.34-2.18	2.30*	1.69-3.13
Previous Macrosomia	2.31*	1.77-3.00	2.66*	1.99-3.55
Weight Gain	1.02*	1.01-1.03	1.03*	1.02-1.04
Prenatal visits, length of treatment, race/ethnicity, total insulin dose, gestational age for treatment initiation, maternal age, getational age at delivery	NS		NS

 P<0.05.

In summary, our study suggests that obese women with BMI of at least 30, who develop GDM, have a higher risk of adverse perinatal outcome than do normal weight women with GDM. Therefore, achieving glycemic control with insulin therapy may enhance pregnancy outcomes for these patients.

University/Hospital:

University Hospital of Columbia University and University of Texas Health Science Center

• The authors stated that dietary compliance was evaluated and reinforced at the weekly clinic visit, but exactly how the compliance was evaluated was not reported. Only patients with good diabetic control remained in dietary treatment only.
• The 4,001 GDM patients is likely the final sample size. The dropout percentage or withdrawal figure is unclear. How many GDM patients were not included in the analyses is unclear. Although it reads as if they attempted to enroll all women, the participation rate is also unclear. There may have some missing measures for some women, but the infomation is not reported.
• The overall mean blood glucose levels is a poor measure for glycemic control
• The multiple logistic regression is problematic. The coding of variables in the model was not clearly described. For example, how "treatment modality" was associated with a significant increased risk of composite outcome (OR=1.28). However, the refence group is not clear: Diet or insulin? Moreover, both obesity and overweight were included in the model. "Mean blood glucose" was associated with an increased risk of composite outcome (OR=1.39), but the comparison is unstated. The authors did not interpret the odds ratio in the text: They only reported the significance.
• Because patients who were unable to achieve the established levels of glycemic control on diet therapy were assigned to insulin treatment, patients on insulin treatment are obviously very different from patients who remained on the diet treatment only. Thus, no inferences can be made for the effect of treatment modality in ths study.
• All women were instructed in the use of an intensified management protocol using memory relectance glucometers.