Introduction

Gestational diabetes mellitus (GDM) is commonly defined as “any degree of glucose intolerance with onset or first recognition during pregnancy” [1]. It has been suggested that women who develop GDM may have preexisting β-cell dysfunction as a result of chronic insulin resistance [2–4]. This reduces their ability to further increase insulin secretion to compensate for the decreasing insulin sensitivity as their pregnancy progresses [2, 5, 6], and glucose tolerance is impaired as a result. However, once the pregnancy-related impairment of insulin sensitivity has disappeared, glucose tolerance is likely to return to normal levels, although pregnant women who have had GDM are at a high risk of developing type 2 diabetes in future years. Worldwide, the prevalence of GDM has continued to increase. This is likely to increase further after the adoption of the new International Association of Diabetes and Pregnancy Study Groups (IADPSG) guidelines [7–9], which have lower glucose diagnostic levels [10]. The causes and adverse outcomes of GDM are summarized in Fig. 1.

Fig. 1
figure 1

The causes and adverse outcomes of gestational diabetes mellitus. Shaded boxes indicate factors/outcomes where a low glycemic index diet may be beneficial. GDM, gestational diabetes mellitus; IR, insulin resistance; T2DM, type 2 diabetes mellitus; CVD, cardiovascular diseases; FHx, family history; BGL, blood glucose level

Full size image

Pedersen hypothesized that maternal blood glucose was the main fuel for fetal growth [11], and subsequently, it has been shown that women with even a mildly elevated blood glucose level have a higher risk of giving birth to a large for gestational age (LGA; defined as birth weight ≥90th percentile of the relevant population) infant [10]. LGA infants have higher risks of childhood obesity [12, 13], cardiovascular diseases [14], and diabetes [12] later in life; therefore, one of the main objectives of GDM treatment is to maintain maternal blood glucose levels as close to normal as possible. Since carbohydrate (CHO) is the only macronutrient that significantly increases postprandial blood glucose response, moderation of CHO intake is usually recommended as the main and first-line strategy to achieve postprandial euglycemia [15].

Carbohydrates and Gestational Diabetes Mellitus

The main purpose of dietary management of GDM is to achieve postprandial euglycemia. A euglycemic GDM diet is usually one with a restricted CHO content [16, 17] that is evenly distributed throughout the day. This is because a diet with a higher proportion (>55 %) of energy from CHOs had been shown to cause higher levels of postprandial glycemia [18].

Despite the significant effect of CHO on postprandial glycemia, complete avoidance of CHO-containing foods is heavily discouraged, since many of them are nutritious, such as milk, yoghurt, and fruits. In addition, overrestriction of CHO in GDM may actually increase the risk of fetal macrosomia [19]. Therefore, instead of complete avoidance or heavy restriction of CHOs, moderation of CHO intake is usually recommended as the first-line strategy to achieve postprandial euglycemia [15, 20]. The American Diabetes Association recommends a diet with ≤40 % of energy derived from CHOs for pregnant women with GDM [17].

The Glycemic Index and Glycemic Load

Glycemic index (GI) and the related dietary term glycemic load (GL) are systematic and physiologically based measures of the glycemic burden of dietary CHOs. Specifically, the GI is an in vivo measure of the blood glucose response to a standard amount of CHOs from a food, relative to a reference food. It is defined methodologically as the average incremental area under the curve in response to the test food (numerator) versus the reference food (denominator), usually pure glucose or white bread, in 10 or more individuals. Examples of common low- and high-GI foods are given in Table 1. So far, the limited evidence available does not suggest that the GI of a food is different in pregnancy [21].

Table 1 Examples of common low and high glycemic index (GI) foods
Full size table

The Potential Benefits of Low Glycemic Index in Pregnancy

The first evidence supporting the beneficial effect of a low-GI diet in pregnancy came from Clapp [22], who investigated the effect of a low-GI diet on the pregnancy outcome of healthy gravidas. After following a low-GI weight maintenance diet from before pregnancy until 8 weeks gestation, the 12 subjects were randomized either to continue the low-GI diet (“aboriginal” CHO diet) or to a high-GI (“cafeteria” CHO) diet that is isoenergetic. Mothers on the high-GI diet were found to have gained more weight than were those who followed a low-GI diet (mean ± SEM: low GI, 11.8 ± 2.3 kg, vs. high GI, 19.7 ± 1.2 kg; p < .01). Infants born to mothers in the high-GI group also had higher birth weight (mean ± SEM: low GI, 3.27 ± 0.12 kg, vs. high GI, 4.25 ± 0.11 kg; p < .01) and higher fat mass (mean ± SEM: low GI, 301 ± 50 g, vs. high GI, 402 ± 80 g; p < .01).

Epidemiological evidence from the Camden study [23] also found that a low-GI diet during pregnancy was associated with lower birth weight. They found that infants of mothers who had a dietary GI < 50—that is, a low-GI diet—had a 116-g lower birth weight than did infants of mothers who had a dietary GI > 60 (p < .05). The authors, however, did not find significant association between dietary GI and risk of LGA.

To test the hypothesis that low-GI diets may improve pregnancy outcomes—in particular, birth weight—Moses et al. [24] in 2006 conducted a randomized controlled trial in 70 healthy women with a singleton pregnancy recruited in weeks 12–16 of gestation, investigating the effect of a low-GI diet on outcomes of pregnancy. There was a lower prevalence of LGA infants in the low-GI group (3 % vs. 33 % in the conventional diet group; p = .01), and the prevalence of SGA was similar (9 % vs. 7 %; nonsignificant). This has led to the hypothesis that a low-GI diet may be beneficial in GDM, where blood glucose control is strongly associated with pregnancy outcomes [25–27].

Evidence from well-designed, large-scale randomized trials suggests that treating even mild GDM results in marked improvement in pregnancy outcomes [25, 28]. By reducing maternal postprandial hyperglycemia in diabetic pregnancies through pharmacological and dietary strategies, risks of macrosomia (birth weight > 4 kg) and later obesity in childhood could be reduced [13, 24, 29].

Apart from the quantity of CHO consumed, postprandial glycemia is also affected by the rate of CHO digestion and absorption. Low-GI foods, which are digested and absorbed more slowly than high-GI foods, have been found to reduce postprandial glycemia in healthy individuals [30]. Therefore, by following a low-GI meal pattern, postprandial hyperglycemia in pregnancy may be better controlled with minimal need of CHO restriction [29].

In addition to its potential beneficial effect on postprandial blood glucose responses, a low-GI diet may also lower the risk of developing GDM, via better weight control [31–35] and improved insulin sensitivity [33, 36] (Fig. 1). Evidence from the Nurses’ Health Study II [37] found that women who had a dietary GI higher than 57 units had a 30 % increased risk of developing GDM, as compared with subjects whose dietary GI was below 51. When total CHO intake is taken into account as well—that is, dietary GL—the increase in risk for the group with the highest GL (greater than 138 units), as compared with the lowest quintile (GL < 104 units), jumped from 30 % to 61 %.

Recent Evidence from Animal Studies

Smith et al. [38•] in 2009 conducted a novel study that provided significant evidence in a sheep model, which is considered comparable to human models in the studies of pregnancy and fetal development [38•], that a high-GI diet may be detrimental in pregnancy. They randomly assigned 104 ewes to receive oral administration of either 100 mL propylene glycol (n = 51), which produces similar postprandial effect as a high-GI meal, or 100 mL water (n = 53) twice per day. The authors found that lambs born to the ewes in the propylene glycol group had significantly higher birth weight (p = .032), ponderal index (calculated as birth weight in kilograms, divided by the cube of height in meters; p = .043), and plasma glucose levels (p < .001). These lambs also grew quicker, as evidenced by the achievement of the same carcass weight at an earlier age than the control lambs (p = .039). The authors therefore concluded that transient high glycemic intake was associated with higher birth weight and faster postnatal growth.

Recent Evidence from Intervention Studies

The number of interventional studies supporting the benefits of a low-GI diet during pregnancy complicated by GDM is limited. There are three published human studies in the last 3 years directly investigating the effect of a low-GI diet in pregnancy complicated with GDM.

Following the success in reducing the prevalence of LGA infants in healthy pregnancy using a low-GI diet [24], Moses et al. [39••] conducted a study to investigate the effect of a low-GI diet in GDM on postprandial glycemic control. In their study, 63 women with GDM were randomized to follow either a low-GI (n = 31) or a conventional high-fiber diet with a higher GI (n = 32) from 28 to 32 weeks of pregnancy until delivery. Dietary assessment at 36–37 weeks by a 3-day food record revealed an 8-unit difference in dietary GI between the groups (low-GI group, 48 ± 1, vs. higher GI, 56 ± 1; p = .018). As compared with the low-GI group, a significantly higher proportion of women in the higher GI group met the treatment criteria to commence insulin (low-GI group, 29 %, vs. high-GI group, 59 %; p = .023). For 19 women in the higher GI group who met the criteria to start insulin and were switched to a low-GI diet, 9 of them avoided the need of insulin to maintain optimal glycemic control. This suggests that a low-GI diet is effective in controlling postprandial blood glucose spikes. Infants of mothers in the low-GI group had similar birth weight and birth weight percentiles as their counterparts in the higher GI group.

A more recent randomized controlled trial by Louie et al. [40••] provided evidence supporting the implication. In that study, 99 women with GDM 26–42 years of age were randomized to follow either a low-GI or conventional high-fiber diet from a mean of 28 weeks of gestation until the end of their pregnancy. The target difference between the diets was 10 unit (50 vs. 60), although at the end, only a 6-unit difference was achieved (47 vs. 53; p < .001), possibly due to the high awareness of GI, as well as high education level, among the study population. Women in both groups had similar pregnancy outcomes, including birth weight, birth weight percentile, and prevalence of macrosomia, as well as the need for insulin treatment. The authors concluded that a low-GI diet could be offered as a safe alternative dietary management strategy to women with GDM. Following a low-GI diet from early pregnancy, especially for women at high risk, may be beneficial, because the harm to the fetus caused by maternal hyperglycemia may be done before GDM is diagnosed at the end of second trimester, where universal GDM testing occurs in most countries [41].

A secondary analysis [42] of the baseline food intake data of the study subjects revealed that the majority of them were not meeting the nutritional requirement for fiber, folate, iodine, and iron, as well as having excessive amounts of sodium and folate. The analysis also found that lower dietary GI, but not GL or grain intake, was significantly associated with a higher intake of various micronutrients. That study, albeit having a small sample size, suggested that women with GDM may be at risk of nutrient inadequacy and that following a low-GI diet may reduce the risk.

The latest randomized trial of the effect of low-GI diet in GDM was conducted by Grant et al. [43••]. In their study, 47 pregnant women diagnosed with GDM or impaired glucose tolerance in pregnancy were randomized to follow a low-GI (n = 23) or a control (n = 24) diet from 28 weeks of pregnancy until delivery. They were also instructed to self-monitor their fasting and 2-h postprandial blood glucose level, using a home glucose meter, and these were reviewed by a clinician against clinical targets. Dietary assessment by 3-day food records revealed that the low-GI group achieved a 9-unit lower dietary GI than the control group at the end of the intervention (49 ± 0.8 vs. 58 ± 0.5; p = .001), and 7 g more fiber (30 ± 1.6 vs. 23 ± 1.0; p = .001). They found that subjects in the low-GI group had fewer postprandial blood glucose values above the target (25.9 % vs. 30.3 %; p = .003). However, unlike the findings of Moses et al. [39••], there was no significant difference between the proportions of women requiring insulin, which could be due to the different clinical targets used to decide on insulin commencement (1 h postprandial ≥ 8.0 mmol/L in Moses et al. vs. 2 h postprandial > 6.6 mmol/L in Grant et al.). A high-GI meal has been shown to cause a spike in postprandial blood glucose level close to 60 min after meals, with the 2-h postprandial level returning close to the fasting level in diabetic subjects [44]. Therefore, when GI of the foods/meals is concerned, 1-h postprandial blood glucose level may be more relevant. Lastly, unlike in the other studies [39••, 40••], infants of mothers in the low-GI group were ~200 g lower on average (3,124 ± 124 g vs. 3,320 ± 220 g), although this did not reach statistical significance with the number of subjects available.

Implications for Clinical Care

Current evidence, although still limited, suggests that pregnant women with GDM may benefit from following a low-GI meal pattern. A low-GI diet was found to be safe in GDM and produces similar pregnancy outcomes as a conventional high-fiber diet. The studies by Moses et al. [39••] and Grant et al. [43••] provided reasonable evidence that postprandial glycemia in GDM may be better controlled with a low-GI diet. Clinicians may therefore look at replacing high-GI CHO from their patients’ diet with low-GI CHO before initiating insulin treatment. Because many nutritious foods, such as dairy foods and fruits, are naturally low GI, it may also be easier to plan a nutritionally adequate GDM diet by choosing more low-GI CHO foods over high-GI, highly processed foods.

Future Directions and Conclusions

On the basis of the current available evidence, we conclude that pregnant women may benefit, with no apparent disadvantage, from following a low-GI meal pattern. By reducing maternal glucose levels within the normal range, a healthy low-GI diet could reduce the risk of adverse pregnancy outcomes. Further studies that intervene at an earlier stage of pregnancy are required.