Gestational diabetes mellitus: the burden of disease
The prevalence of gestational diabetes mellitus (GDM) is rising worldwide [1, 2], with recent Australian estimates indicating that between 5.2% and 8.8% of pregnant women develop GDM . Risk factors for GDM include the mother’s own birthweight, maternal obesity, advanced maternal age, ethnicity, family history of diabetes and a previous history of GDM, large babies or unexplained stillbirth [4–6].
In Australia each year, between 6.3% and 7.7%, or between 15,900 and 19,450 pregnant women have a positive oral glucose challenge test (OGCT) on screening for GDM followed by a normal diagnostic oral glucose tolerance test result (OGTT) [7–9]. These women have borderline gestational diabetes with values of glucose tolerance intermediate between normal and those diagnostic of GDM.
There are well-documented risks for the infant of a mother with GDM including fetal macrosomia, birth injuries such as shoulder dystocia , bone fractures and nerve palsies , neonatal hypoglycaemia , and hyperbilirubinaemia . A later complication associated with macrosomia and large-for-gestational age, for female offspring, is premenopausal breast cancer . Long-term adverse health outcomes reported for the infants include obesity [6, 15], impaired intellectual achievement , impairment of glucose tolerance  and increased risk for subsequent diabetes .
For women with GDM there is an increased risk of developing pre-eclampsia and an increased need for induction of labour [18, 19]. Impaired glucose tolerance in pregnancy is highly predictive for the later development of diabetes, with over 50% of women with GDM developing type 2 diabetes within 10 years of the index pregnancy . Although the perinatal risks of GDM are well documented, the impact of borderline gestational diabetes on maternal and infant health outcomes is less clear.
We reported a 10-year audit (1993-2003) on a large cohort of women (16,975) at the Women’s and Children’s Hospital, Adelaide, that examined the influence of differing levels of glucose tolerance on pregnancy complications . Women were offered screening for GDM using an oral glucose challenge test (OGCT). Women who screened positive (plasma glucose concentration ≥7.8 mmol/L) were offered a 75 gram diagnostic oral glucose tolerance test (OGTT). Of the women with a positive OGCT who had an OGTT performed, 1074 (6.3% of all women screened) had normal OGTT results. Women with borderline gestational diabetes (positive OGCT, normal OGTT) had a statistically significant increased risk of pre-eclampsia and caesarean section compared with women with normal glucose tolerance (negative OGCT), while the infants of women with borderline gestational diabetes were at increased risk of hypoglycaemia and hyperbilrubinaemia, compared with the infants of women with normal glucose tolerance .
The results of our audit are consistent with other observational studies in the literature, identifying an increased risk of adverse maternal and infant outcomes with increasing plasma glucose values [21–23]. A retrospective US study of 1813 women reported an association between increasing levels of hyperglycaemia and the risk of pre-eclampsia. Optimisation of glucose control was found to decrease the risk of pre-eclampsia . The Toronto Tri-Hospital Study screened over 4,000 women, using an OGCT, and if positive, an OGTT, between 26 and 28 weeks gestation, and found increasing degrees of carbohydrate intolerance to be associated with increased risks of pre-eclampsia, caesarean section, macrosomia and need for neonatal phototherapy [21, 22]. The international observational study, HAPO, reported an association of increasing hyperglycaemia with greater risks of adverse perinatal outcomes in 25,000 women recruited from over 25 different sites around the world .
We reported on the OGCT results and pregnancy outcomes of 1814 women recruited into the ACTS antioxidant supplementation randomised trial (ACTS Trial) . These women were in their first pregnancy and recruited within Australia between 2001 and 2005. The proportion of women with borderline gestational diabetes (positive OGCT, normal OGTT) was 7.7% of all women screened. Women with borderline gestational diabetes (positive OGCT, normal OGTT) had a statistically significant increased risk of serious adverse health outcomes, pregnancy induced hypertension and caesarean section compared with women with normal glucose tolerance (negative OGCT). Infants born to women with borderline gestational diabetes were at increased risk of macrosomia (birthweight ≥4.5 kg) and need for admission to the neonatal nursery.
We consider that there is now compelling evidence of substantially increased risks of adverse health outcomes for both mother and infant when women have borderline gestational diabetes. The question of whether dietary and lifestyle advice and treatment can reduce these risks now requires urgent consideration.
Treatment of gestational diabetes mellitus improves maternal and infant health
The ACHOIS (Australian Carbohydrate Intolerance Study in Pregnant Women)  randomised trial provided evidence that treatment with dietary advice, blood glucose testing and, if required, insulin given to women with mild GDM (fasting glucose <7.0 mmol/L and/or 2-hour glucose ≥7.8 mmol/L but <11.1 mmol/L) significantly reduced the rate of serious perinatal complications (death, shoulder dystocia, bone fracture and nerve palsy) for the infants from 4% to 1%, relative risk (RR), adjusted for maternal age, race and parity, 0.33; 95% confidence interval (CI) 0.14 to 0.75; p=0.01). The number of women needed to treat to prevent one additional serious outcome in an infant was 34 (95% CI, 20 to 103). More infants in the treatment group were admitted to the nursery but there was a significant reduction in the number of infants with a birthweight over 4 kg (10% vs 21%). Women in the treatment group had a higher rate of induction of labour (39% vs 29%) although rates of caesarean section were similar between treatment groups (31% vs 32%). Three months after pregnancy, treated women had significantly lower rates of depression and better scores for health related quality of life .
The ACHOIS trial confirmed that mild GDM is a pathological entity that, untreated, is associated with relatively rare but nonetheless significant adverse perinatal outcomes, which can be avoided or reduced with treatment consisting of individualised dietary and lifestyle advice, with insulin treatment as necessary . Cost consequences were in the range acceptable to healthcare funders . The ACHOIS trial provided the first substantial evidence supporting detection and treatment of mild gestational diabetes  and has led to proposals for routine screening to become the standard for detection of gestational diabetes in Australia and elsewhere [27–30].
While the ACHOIS trial clarified that treatment of women with mild GDM is beneficial, uncertainty remains whether treatment of women with more borderline gestational diabetes offers similar benefits. Current clinical practice for the management of pregnant women, who screen positive on OGCT but whose subsequent OGTT is normal, is to leave such women ‘untreated,’ with reassurance that their results have not reached the required cut-off for the diagnosis of mild or more severe GDM.
Critical appraisal of the literature: a Cochrane review
We conducted a systematic review, using the best evidence currently available to assess whether treatment of pregnant women with borderline levels of glucose intolerance (defined as positive OGCT, normal OGTT) improves maternal and infant outcomes . Published randomised controlled trials and cluster-randomised trials comparing alternative management strategies for women with borderline GDM were eligible for inclusion. Pregnant women with hyperglycaemia who did not meet diagnostic criteria for GDM, based on OGTT test results as defined variously by individual trialists according to local health authorities and professional organizations, were considered eligible. Interventions included dietary advice (standard or individualized), exercise and lifestyle advice (standard or individualized) and drug treatment including insulin and oral drugs. Outcomes included both maternal and fetal/neonatal health outcomes, as well as outcomes extending into childhood and adulthood, and health service costs. We used the search strategy of the Cochrane Controlled Trials Register (CENTRAL), MEDLINE, EMBASE, hand-searches of 30 journals and proceedings of major conferences as well as weekly current awareness alerts for a further 44 journals (Date of last search 30 September 2011).
Four studies involving 543 women and their babies were included [32–35]. Two of the four studies were from the United States [33, 35], one was from Canada  and one from Italy . One study  was found to have a low to moderate risk of bias, with the remaining three studies at moderate to high risk of bias. The babies of women receiving management for borderline gestational diabetes were less likely to be macrosomic (birthweight >4000g) (three trials, 438 infants, RR 0.38, 95% CI 0.19 to 0.74) or large-for-gestational age (three trials, 438 infants, RR 0.37, 95% CI 0.20 to 0.66) when compared with those in the routine care group. No significant differences in rates of caesarean section (three trials, 509 women, RR 0.93, 95% CI 0.68 to 1.27) or operative vaginal birth (one trial, 83 women, RR 1.37, 95% CI 0.20 to 9.27) were found.
This review found that for pregnant women with hyperglycaemia who did not meet the diagnostic criteria to be classified as having GDM or type 2 diabetes, interventions such as dietary counselling, blood glucose monitoring and insulin therapy resulted in a reduction in the rates of macrosomia and large for gestational age babies.
Although the results of this systematic review suggest benefit in treating women with borderline gestational diabetes, the review concluded that larger trials are needed of sufficient power and in other populations to assess the effects of management of such women on maternal and infant health outcomes .
Aims and objectives of this trial
It is now clear from the ACHOIS trial  that treatment of pregnant women with mild GDM, formerly defined as impaired glucose tolerance, is beneficial for women and their infants. It is still uncertain whether the benefits of similar treatment for women with borderline gestational diabetes outweigh any harms from such treatment.
The aims of this multicentre randomised clinical trial are to assess whether dietary and lifestyle advice and treatment given to pregnant women who have borderline gestational diabetes on screening for GDM (defined as a positive OGCT followed by a normal OGTT), reduces neonatal complications and maternal risks.
The primary hypothesis is that dietary and lifestyle advice and treatment given to women who have borderline gestational diabetes on screening for GDM will reduce the incidence of large for gestational age infants, defined as birthweight above the 90th centile for gestation and fetal sex on standardised birthweight charts.
The secondary hypotheses are that dietary and lifestyle advice and treatment given to women who have borderline gestational diabetes will reduce the risk of death or serious health outcome for the infant; reduce the risk of serious health outcome for the woman; reduce the risk of other causes of infant morbidity; and reduce the risks of other adverse health outcomes for the woman.