The definition of normal pregnancy glucose levels, both fasting and in the postprandial state, has been investigated by a variety of methods, including detailed hospital inpatient studies, ambulatory studies using capillary glucose meters and continuous glucose monitoring (CGMS) over many years. A comprehensive analysis of this literature was published by Hernandez et al. in 2011 [19] and concluded that mean glucose levels in normal weight pregnant women were lower than previously described (Mean ± SD Fasting 71 ± 8 mg/dL or 3.9 ± 0.4 mmol/L; 1 h post meal 109 ± 10 mg/dL or 6.1 ± 0.6 mmol/L; 2 h post meal 99 ± 10 mg/dL or 5.5 ± 0.6 mmol/L). The commonly used statistical upper limits of these observations (Mean + 2 SD) are also substantially lower than current, largely empirical, glucose treatment targets advocated in GDM [20–25] (Table 28.1) The determination of optimal glucose targets remains contentious, although the current recommendations of major professional bodies are largely congruent. The values mentioned in Table 28.1 are the maximal glucose values in the fasting and post-prandial states considered acceptable by the relevant organizations, or reported as treatment targets by the randomized controlled trials. The abbreviations and data sources for Table 28.1 are as follows: ADA—American Diabetes Association [20, 22]; ADIPS—Australasian Diabetes in Pregnancy Society [21]: ACOG—American College of Obstetricians and Gynecologists [22]; NICE—National Institute of Clinical Excellence (UK) [22]: DIPSI—Diabetes in Pregnancy Society India [23]: CDA—Canadian Diabetes Association [24, 25]; ACHOIS—Australian Carbohydrate Intolerance Study [26]; MFMN—Maternal Foetal Medicine Networks study [27]

Although the clinical importance of treating GDM was hotly debated for many years, two recent randomized controlled trials have now been conducted, with largely congruent results. The ACHOIS study [26], conducted principally in Australia, randomized 490 women to an intervention including dietary advice, home blood glucose monitoring and insulin therapy as required and compared them to 510 women assigned to standard care. The primary composite outcome of serious perinatal complications (death, shoulder dystocia, bone fracture or nerve palsy) was reduced from 4 to 1 % (p = 0.01) with intervention. Only 20 % of women required adjunctive insulin treatment. The rates of LGA and macrosomia (defined as birthweight ≥ 4,000 g) were also reduced, with no increase in small for gestational age (SGA) babies. Quality of life was improved with treatment and hypertensive complications of pregnancy were less frequent. Treated women had higher rates of induction of labour, but similar rates of caesarean delivery. Admission of neonates to the neonatal nursery was more frequent with intervention, possibly related to hospital policies.

The USA based Maternal-Foetal Medicine Units Network trial of the treatment of mild GDM [27] randomly assigned 485 women to intervention vs. 473 assigned to standard care. This study excluded women with fasting glucose ≥ 95 mg/dL, but required them to have 2 / 3 other OGTT values in excess of ADA thresholds for the diagnosis of gestational diabetes [22]. The composite outcome used in this trial included both perinatal mortality and neonatal outcomes associated with maternal hyperglycaemia (hypoglycaemia, hyperbilirubinemia, hyperinsulinemia and birth trauma. Fully 92 % of treated women were managed by lifestyle interventions, with only37/485 or 7.6 % needing adjunctive insulin therapy. As with the ACHOIS study, the rates of LGA and macrosomia were reduced with treatment as was neonatal fat mass. Hypertensive disorders of pregnancy, caesarean section, maternal weight gain from enrolment and shoulder dystocia were all reduced by active intervention. The composite neonatal outcome did not differ between the treatment groups.

Considering these two trials together, we conclude that active treatment of women with GDM reduces serious pregnancy complications including hypertensive disorders of pregnancy and excessive foetal growth and its consequences. Thus, after many years of debate, active treatment of GDM appears justified.

Lifestyle modification, including medical nutrition therapy and encouragement of physical activity, forms the primary mode of therapy for GDM. As noted above, such therapy proved sufficient in 80–90 % of women enrolled in the two major randomized trials. The American Dietetic Association evidence based nutrition practice guidelines for GDM [28] provide a sound overall framework for nutritional interventions. They advocate review by a registered dietitian within 1 week of GDM diagnosis and a minimum of two follow up visits. Whilst dietary recommendations are individualized, the general guidance favours mild energy restriction to approximately 70 % of recommended daily intake for overweight/obese women and reduction of total carbohydrate intake to less than 45 % of total caloric intake. Other recent studies have specifically examined the role of glycaemic index in nutritional therapy for GDM [29] and have suggested that this may enhance the effects of standard treatment and assist some women in meeting glycaemic targets and obviating the need for pharmacotherapy, although overall pregnancy outcomes have been reported as similar to those achieved with a high-fibre diet [30].

Insulin therapy remains the cornerstone of treatment for women who fail to meet glycaemic goals after lifestyle modification. The glucose targets currently recommended by various international organizations and those used in the two major randomized trials are listed in Table 28.1. The precise insulin regimen used for an individual woman depends on the pattern of elevated glucose readings seen on home glucose monitoring. Women with predominant fasting hyperglycaemia may respond well to a single evening injection of intermediate or long-acting insulin. Those with predominant elevation of post-prandial glucose are generally treated with soluble insulin or a rapid-acting analogue insulin prior to meals. Women with a mixed pattern of elevated glucose readings may also be treated with premixed insulins, generally twice daily at breakfast and dinner time. A recent detailed review [31] has concluded that there is minimal evidence available to choose between specific insulin regimens.

Oral hypoglycaemic agents, principally glibenclamide (glyburide in the USA) and metformin, have also been trialled in gestational diabetes. The first major randomized controlled trial of glyburide, conducted by Langer and colleagues [32] reported equivalent glucose control and no substantial differences in outcomes in women treated with glyburide as compared to insulin after failure of lifestyle management. Langer’s study was not powered to examine major pregnancy outcomes. Subsequently, a systematic review of this and subsequent randomized trials [33] reported no differences in terms of glycaemic control or pregnancy outcomes when comparing glyburide and insulin therapy. However a large recent (non-randomized) cohort report of 10,682 women from the “Sweet Success” programme in California reported increased rates of LGA and neonatal intensive care unit admission in those women treated with glyburide [34].

Metformin has also been evaluated in one large [35] and two smaller randomized trials [36, 37]. These studies included women who failed to meet glycaemic targets after lifestyle interventions for GDM. They demonstrated very similar glucose control and pregnancy outcomes with metformin therapy when compared to standard insulin treatment. The rate of supplemental insulin therapy was high overall—46 % in the MiG study and 32 % in the Finnish report but surprisingly no patients in the study from New Mexico USA required supplemental insulin therapy. Patient preferences were clearly in favour of metformin therapy.

Uncontrolled reports suggested that metformin reduces pregnancy complications in women with polycystic ovarian syndrome treated during pregnancy [38–40], but a recent randomized trial failed to confirm any benefit [41].

Two randomized trials [42, 43] comparing metformin with glyburide/glibenclamide in the treatment of GDM have shown somewhat discordant results. In a report from Brazil [42, 44] Silva et al. reported comparable efficacy between these two medications, with supplemental insulin required in around 25 % of each group. In contrast, Moore et al. [43] reported that 35 % of metformin treated women required supplemental insulin as compared to 16 % of those treated with glyburide.

The use of oral hypoglycaemic agents in the treatment of GDM varies widely on a country by country basis, with glyburide favoured in the USA [45], whilst metformin is recommended more widely in the United Kingdom and Australasia [46–48]. Some practitioners remain concerned by the possibility of adverse foetal effects of oral drugs. Glyburide has been reported only at very low levels in cord blood, due to limited transplacental passage and active counter transport [49, 50]. Metformin crosses the placenta readily [51–53] but has shown no adverse foetal or early childhood effects to date [54, 55].

As noted above, many of the adverse effects of GDM relate to excessive foetal growth, driven by nutrient excess and consequent foetal hyperinsulinism as outlined in the Pederson hypothesis [9] and its subsequent modification by Freinkel and Metzger [5]. These have also been presented in Fig. 28.1. Direct assessment of amniotic fluid insulin concentrations, pioneered by Weiss et al. [56] demonstrated that foetal hyperinsulinemia was associated with increased risks of pregnancy complications. In turn, Schafer Graf et al. [57] noted that foetal abdominal circumference (AC) on ultrasound could be used to non-invasively predict foetal hyperinsulinism and identify those babies most at risk of diabetic fetopathy.

Subsequently, an alternative approach to treatment of GDM, based on ultrasound assessment of foetal growth, in particular foetal AC, has been proposed and evaluated in four randomized studies [58–62]. According to this approach, summarized by Kjos et al. [60], glucose lowering treatment may be intensified in those pregnancies where the foetus shows evidence of accelerated growth, generally determined by foetal AC > 75th centile for gestational age on ultrasonographic assessment. These protocols suggest that “low risk” pregnancies, as defined by normal (<75th centile) foetal AC, may require less stringent glycaemic control and conversely that detection of an increased AC should lead to intensification of glucose lowering therapy.

A summary of the traditional (glucocentric) and alternative (USS guided) pathways for GDM therapy is presented in Fig. 28.3.

Despite a lack of high-level evidence of definite benefit, foetal ultrasound is frequently used to assess foetal well bring and to estimate foetal weight to assist in determining the timing and mode of delivery. Most units have extrapolated their practices in this area from those developed for pre-gestational or pre-existing diabetes.