Glucose that comes from the mother provides fuel for fetal growth [1]. High maternal blood glucose levels (BGL), even within the current recommended range, have been associated with higher infant body fat [2]. The peak postprandial BGL (PBGL) occurs later in pregnant women than in the non-pregnant state [3], i.e., at 60 vs. 30 min, and is an important contributor to the risk of fetal overgrowth [4,5]. Treatment of gestational diabetes mellitus (GDM), where maternal glucose homeostasis is impaired, is therefore generally based on a combination of pre-meal BGL and PBGL 1 or 2 h after meals as monitored by self-blood glucose monitoring [6,7].

The glycemic index (GI) is a measurement of the glycemic quality of the carbohydrates in foods, where a low GI indicates that the carbohydrates are digested and absorbed slowly [8]. The limited evidence available suggests that GI of meals is of relevance in the setting of GDM [9]. A low GI diet was shown to reduce the requirement for insulin in the glycemic management of women with GDM [10], and was equally effective in improving pregnancy outcomes in GDM when compared to a conventional high fibre diet [11].

While the GI concept may be valid in diabetic pregnancy, scepticism of the efficacy of a low GI diet in GDM remains. This could partly be attributed to the lack of scientific evidence to demonstrate that low GI meals actually reduce, but not delay, the peak PBGL in GDM. A recent analysis of glycemic responses to over 1000 foods indicated that the timing of the peak was the same for high vs. low GI foods in non-diabetic individuals [12]. Whether this is true in pregnancy, especially those complicated by GDM, remains unclear because changes in gastric motility [13] and insulin sensitivity [14] during pregnancy may alter digestion and absorption.

In this study we investigated whether two bread-based breakfasts with different GI produced different postprandial peaks and peak time points in a group of women with GDM. Our hypothesis was that a low GI bread-based breakfast would produce a lower but not delayed PBGL peak in GDM when compared to an energy and macronutrient matched high GI bread-based breakfast.

The study sample included five South Asians (Indian, Nepalese and Bangladeshi), four Chinese, and one Caucasian. The subjects attended the first and second session at a mean ± SEM of 33.5 ± 0.5 and 35.1 ± 0.6 weeks of gestation respectively. There was no difference in BMI (first visit: 24.9 ± 0.4 kg/m² vs. second visit: 26.1 ± 0.9 kg/m²; p = 0.212) and mean 7-day fasting BGL (first visit: 4.8 ± 0.1 mmol/L vs. second visit: 4.8 ± 0.1 mmol/L; p = 0.461) on the test days. No subject started insulin therapy during the study period.

Overall the consumption of a high GI bread-based breakfast resulted in higher postprandial glycemia (Figure 1). The mean ± SEM iAUC_glucose after a low GI breakfast was significantly lower than that after the consumption of a high GI breakfast (212.7 ± 22.9 vs. 340.8 ± 23.4; p = 0.001). The mean ± SEM peak BGL was 6.7 ± 0.3 mmol/L for the low GI breakfast and 8.6 ± 0.3 mmol/L for the high GI breakfast (p < 0.001). However, there was large inter-subject variability in the timing of the peak between the two test meals (Figure 2): the peak occurred between 45 and 75 min for the low GI breakfast (mean ± SEM: 60.0 ± 4.0 min), and between 30 and 60 min for the high GI breakfast (mean ± SEM: 43.1 ± 3.4 min; p = 0.015). In the eight subjects who provided data for both breakfasts, six had a delayed peak PBGL time after consuming the low GI breakfast when compared to that of the high GI breakfast. There was no significant difference in subjective satiety throughout the 2-h test period (Figure 3).

This study is the first attempt to examine the timing of the peak PBGL in women with GDM after consumption of breakfasts of different GI values. We expected to see differences in the peak concentration but similar timing of the peak. However we found that the low GI breakfast produced a peak at closer to 60 min after the start of the meal compared with ~45 min after the high GI breakfast.

It has been previously shown that a low GI diet produces comparable outcomes as a conventional high fibre diet in pregnancy complicated with GDM [11]. The limited evidence on the efficacy of a low GI diet for the management of GDM has suggested that a low GI diet may improve postprandial glycemia [10,16], therefore reducing excessive transfer of maternal blood glucose to the fetus. In GDM, breakfast was shown to have the greatest variability in postprandial glycemic response [17]. The results from our study suggested that a low GI breakfast may be of benefit in the management of post-breakfast glycemia.

Since peak PBGL has been shown to be more strongly associated with outcomes of GDM pregnancy [18] than 2-h PBGL, it is important to accurately time the postprandial SBGM testing to capture the peak BGL. In healthy, non-pregnant subjects low and high GI meals reach peak concentrations at the same time, i.e., at 30 min [12]. However during pregnancy, especially those complicated with GDM, changes in gastric motility [13] and insulin sensitivity [14] can be expected to alter the rate of carbohydrate digestion and absorption, and hence shape of the postprandial glucose curve. Indeed, Wolever et al. [19] showed that the peak PBGL after a standard test meal occurred later and later as glucose tolerance worsened in subjects with diabetes.

A previous study had found that the peak PBGL in GDM occurs at about 60 min post meal [7]. Therefore SBGM that tests the 2 h PBGL may miss the glucose peak, and the clinical decision to commence insulin therapy was usually based on a cut-off of 1-h PBGL [6]. We found that while a low GI breakfast delayed the peak PBGL in GDM, the peak occurs closer to 60 min after the start of the meal than a high GI breakfast, which produced a peak at ~45 min post meal. Therefore our results suggested the findings of the study by Moses et al. [10], which demonstrated that a low GI diet reduces the need for insulin in GDM for BGL management, was indeed due to the fact that low GI meals produce a lower peak PBGL at 60 min post meal. Our results also suggest that the 1-h postprandial SBGM reading is likely to underestimate the actual postprandial glycemic response to a high GI meal. Therefore our finding raises the question whether the GI of the patient’s diet should be considered in the interpretation of SBGM results.

Our study has limitations, including a small sample size and a high proportion of Asian subjects, which compromises the generalizability of the findings. In addition, the non-continuous nature of blood sample collection did not allow accurate determination of the actual peak time. However, more frequent fingerprick blood sampling would be impractical. Although continuous glucose monitoring might be helpful in this context, it measures changes in the interstitial fluid rather than capillary blood, and by nature is subject to a delay [20,21,22].

The low GI breakfast produces lower postprandial glycemia than a macronutrient matched high GI breakfast. The timing of the peak BGL varies within and between breakfasts of different GI. The peak PBGL after a high GI breakfast occurs at ~45 min, 15 min earlier than that of a low GI breakfast. The peak PBGL of a low GI breakfast occurs closer to the time recommended for PBGL monitoring in GDM. Since many women consume high GI meals throughout pregnancy, there are implications for clinical practice.