This study determined whether maternal glucose tolerance had a progressive effect on the length of gestation in singleton pregnancies and whether there was an increasing tendency towards spontaneous preterm birth with increasing glucose intolerance.

• In this study, we first obtained from the registry the serial numbers and names of women who had the antenatal OGTT over a 24-month period (1996-1997) and who delivered at our hospital.
• The study was confined to women who did not require insulin in order to avoid any possible effect or side effect of insulin treatments, such as hypoglycemia, on the initiation of labor.

• All women with significant medical complications and those who were taking medications such as corticosteroids were managed in the High-Risk Clinic of another hospital under the same department and were not included in this study.

Recruitment method was not detailed

Design: Historical Cohort

Blinding used:  not applicable 

Intervention 

• Women with GDM were referred to a dietitian and put on diet control (30 kcal/kg of ideal weight) and then assessed with pre- and 2-hour postprandial blood glucose profiles.
• Insulin therapy was started for inadequate control (fasting plasma glucose >5.9 mmol/L or postprandial glucose >7mmol/L) if dietary readjustment failed to normalize the blood glucose profile.
• Obstetrical intervention was based on clear-cut-clinical indications only, and labor induction or elective cesarean section is not performed for suspected large for gestational age (LGA) fetuses in pregnancies complicated by GDM unless other complications, such as malpresentations, were also present.

Statistical Analysis  

• One-way analysis of variance to determine differences in continuous variables  among the six groups
• Duncan’s multiple range test (set at 5% level) was used to identify the groups that were different from the others
• X² test with Spearman’s correlation coefficient (p) for categorical variables
• Multiple regression, using a commercially available statistical package (statistical package for the Social Sciences 10.0; SPSS Inc, Chicago, IL ).  

Timing of Measurements

The OGTT was performed in the morning after an overnight fast (8-12 hours), and all were instructed to continue with their usual diet before the OGTT.

Dependent Variables

Diagnostic criteria

• All women referred for antenatal care undergo a routine ultrasound scanning at 18-20 weeks’ gestation for both dating and exclusion of fetal anomalies, irrespective of whether there were was prior ultrasound confirmation of their dates.
• At the time of the first antenatal visit, women identified to have risk factors for the development of GDM , such as advanced maternal age(>35 years), relevant past obstetric and family history, obesity, and recurrent or significant glycosuria , undergo the 75-g OGTT.
• The results of the OGTT was interpreted by the original World Health Organization (WHO) criteria - women who have impaired glucose tolerance (IGTT) (OGTT 2-hour value > 8.0 mmol/L) or diabetes mellitus (2-hour value > 11.0 mmol/L) were both categorized as having GDM.
• All low-risk women have random blood glucose testing at the beginning of the third trimester(28-30 weeks) to screen for glucose intolerance. Those with abnormal spot glucose(>5.8 mmol/L if less than 2 hours postprandial and > 5.0 mmol/L if more than 2 hours postprandial) also undergo the 75-g OGTT.
• Each woman was classified into one of 6 groups according to the 2-hour glucose value of the OGTT (in mmol/L) as follows: 

• For the women who had a repeat OGTT in the third trimester, only the result of the last test was used for classification.
• The women considered non-diabetic were analyzed as three groups to determine whether there were any differences among them
• The women diagnosed with IGT were classified as two groups for analysis, with 9.0 mmol/L as the cutoff value, because 9.0 mmol/L has also been proposed as the definition of GDM.
• The women diagnosed with diabetes mellitus were analyzed as one group

Neonatal Outcome:

• Caesarean delivery
• Incidence of total preterm delivery (< 37 weeks; < 32 weeks)
• Macrosomia
• Large for Gestational Age (LGA, birthweight >90^th percentile for the local population)
• Small for gestational age (SGA, birth weight< 400 g )   

Independent Variables

Maternal anthropometric parameters

• age (y) 
• weight (kg)
• height (cm)
• BMI (kg/m²)

Infant

• gestation (wk)
• birth weight (g)
• BMI (kg/m²)

Control Variables

Initial N: 2,168 included women delivering in the hospital and had complete records, including 1169 women with GDM treated with diet restriction only.

Attrition (final N):  as above

Age: See Table I

Ethnicity: ethnic Chinese

Other relevant demographics:

Anthropometrics

Location: Department of Obstetrics and Gynecology, The University of Hong Kong and Department of Obstetrics and Gynecology, Queen Mary Hospital, Hong Kong People's Republic of China.

Of the 2,168 women in the final study cohort, 1169 (53.9%) had GDM, and the distribution of these subjects according to the OGTT 2-hour glucose values is shown in Table 1. For the mothers, there were significant differences in the age, weight, height, and body mass index (BMI).

Table 1. Maternal and Infant Characteristics in Relation to the OGTT 2-Hour Glucose Values for the Entire Cohort 

Results are mean + standard deviation

*  P<.0001 by one-way analysis of variance

¶   P<.05 vs. group 1 by Duncan test.

|| P<.05 by vs. groups 1 and 2 by Duncan test.                      

♦  P<.05 vs. groups 1-3 by Duncan test.

£  P<.05 vs. groups 1-4 by Duncan test.

§  P<.05 vs. group 2 by Duncan test.

#  P<.005 by one-way analysis of variance

When pregnancy outcome was analyzed, the incidence of cesarean delivery was significantly different among the six groups and was positively associated with the OGTT 2-hour glucose values (Table 2).

Table 2. Pregnancy Outcome in Relation to the OGTT 2-Hour Glucose Values for the Entire Cohort 

Results are expressed as percentage

* x²test P=.002

¶ P=.002 by Spearman correlation

 || P =.017 by Spearman correlation

♦ P = .009 by Spearman correlation

The incidence of spontaneous labor was examined, there were significant differences (P<0.0001) among the six groups. These resulting percentage figures were inversely correlated with the OGTT 2-hour glucose values (P<0.0001)(Table 3).

Table 3. Maternal and Infant Characteristics in Relation to the OGTT 2-Hour Glucose Values for the Cohort With Spontaneous Labor 

Results are mean + standard deviation

* P<.0001 by one-way analysis of variance

¶ P<.05 vs. group 1 by Duncan test.                    

|| P<.05 vs. groups 1-3 by Duncan test.

 ♦ P<.05 vs. groups 1-5 by Duncan test.

 £ P<.02 by one-way analysis of variance

 § P<.05 vs. group 2 by Duncan test.

#  P < .05 vs. group 1 and 2 by Duncan test.

** P<.01 by one-way analysis of variance

When the incidences of cesarean and spontaneous delivery before 37 and 32 weeks were analyzed, significant differences were found among the 6 groups (Table 4).

Table 4. Pregnancy Outcome in Relation to OGTT 2-Hour Glucose values for the Cohort With Spontaneous Labor 

 Results are expressed as percentage

*P= .003 by x² test.

¶ P= .015 by x² test.

|| P=.009 by Spearman correlation

♦ P = .018 by x² test.

 £ P= .005 Spearman correlation

Other Findings

To determine whether maternal parity had any effect on the incidence of spontaneous birth, the analysis of the spontaneous labor cohort was stratified for parity (nulliparous or multiparous) and confined to delivery before 37 weeks because of the few deliveries before 32 weeks. The incidence of preterm delivery was still significantly different for both the nulliparous and multiparous subjects. 

Gestational glucose intolerance affected gestational length and incidence of preterm births, which should be considered a confounding factor in the analysis of the neonatal outcome of GDM pregnancies.

Limitation

There was no information on the incidence of previous preterm birth in the database and thus there was no examination of the contribution of this factor to the outcome of the index pregnancy.

In a historical cohort study, data on exposure and occurrence of disease are collected after the events have taken place—the cohorts of exposed and non-exposed subjects are assembled from existing records, or health care registries. Studies of this kind provide a better opportunity than one time cross sectional studies to examine whether certain behaviors do in fact lead to (or cause) the disease.

The limitations and critique of the study, as stated by the authors appear to be very appropriate.