This study observed significant reductions of glucose iAUCs following the intake of cold potatoes compared to hot potatoes in 30 overweight or obese women. The study is novel in that a predictive equation to determine baseline characteristics that influenced the glycemic response following a hot or cold potato was developed. The study focused on understanding the glycemic benefits of modifying RS by cooking and refrigeration in a commonly consumed food.
Investigations into the role of RS on glucose homeostasis related to other RS foods mirror the results of the current study. A randomized crossover trial by Nilsson et al.(3) examined a 3-day intervention where high RS bread (barley, ~17 g RS/day) was compared to white bread (2.5 g RS/day) and reported improved glucose in 20 healthy volunteers (85% women). Glucose peaks were reduced following the high-RS bread. Stewart et al.(31) provided acute supplementation of RS type 4 (16.5 g) in a crossover study comparing high fiber vs low fiber scones. After measuring the iAUC for 180 minutes, significant reductions of glucose between 43-45% emerged (31). However, these trials did not examine the influence of food processing on RS levels in whole foods, nor was the influence of the gut microbiome on glycemic response explored. These trials, among many others(1, 2, 6, 7, 31), demonstrate that a higher, acute intake of RS resulted in improved postprandial glucose homeostasis. It is important to note that the present study observed reductions in glucose iAUC following the intake of a cold potato with approximately 13.7 g of RS, which is a lower amount than many of the studies, but was still efficacious. This in part may be due to the combination of RS types, both RS type 2 and RS type 3, present in the cold potato or other fiber components inherent in potatoes. We also studied healthy, overweight females following acute ingestion of a whole food product containing RS, rather than RS as a supplement, which few other studies exclusively investigate.
The usual dietary intake reported by the study participants mostly aligned with the U.S. Dietary Guidelines for Americans 2020-2025 (26). The most concerning dietary pattern from our participants was the lack of fiber intake and excessive added sugar consumption. Studies investigating the postprandial response of RS (either acute or prolonged consumption) usually fail to report the usual dietary intake of the participants in the trial. Our participants’ metabolic phenotype was primed by diets high in available carbohydrate, possibly confounding the interaction between epigenetic factors and the hot potato (higher in available carbohydrates). Phenotypic patterns may have affected the availability of digestive enzymes or activated genes related to carbohydrate metabolism or RS degradation. RS causes genetic alterations in carbohydrate metabolism (32), to where participants consuming a higher-fiber diet may have elicited a different response to the intervention than individuals with consistently lower-fiber intake. The importance of this may be evident in the negative association (though not significant, p=0.14) between insoluble fiber and glucose iAUC determined in our regression model.
The interplay between diet and microorganisms residing in the intestinal tract provides a potential mechanistic concept of how dietary RS can improve PPGR. Although this study did not measure the fermentation of RS and its byproducts nor the microbiome changes resulting from RS intake, we did measure how baseline microbiota can influence the physiological response to RS. Several genera showed relationships with postprandial glucose iAUC following consumption of hot and cold potatoes. The Faecalibacterium genus and Actinobacteria phyla showed moderate, negative correlations following the intake of hot and cold potatoes, respectively. Several studies demonstrate similar findings. Zhang et al. sequenced the microbiome of patients with different levels of glucose intolerance, and Faecalibacteria prausnitzii was most abundant in the normal glucose tolerant group compared to the participants with prediabetes and type 2 diabetes mellitus (T2DM) (33). The importance of Faecalibacterium in determining glucose iAUC became evident in our model as the only significant contributor, other than potato type, associated with PPGR. Other studies have also observed an inverse relationship between Faecalibacterium and glucose sensitivity (34), and a recent review (35) found that four out of five studies found a negative association between Faecalibacterium and T2DM. To our knowledge, this is the first study to demonstrate the inverse association between Faecalibacterium and iAUC glucose following a high-RS whole food.
Various modeling techniques can predict PPGR using baseline characteristics. In a post hoc analysis of 106 healthy Danish adults, Sᴓndertoft et al. used a random forest model focused on clinical features and the microbiome to determine the magnitude of the effect on PPGR (30). The authors noted that a model based solely on microbial components accounted for up to 14% of the variance in PPGR excursions. When clinical features were added to the model, up to 78% of the variance in PPGR excursions was reported. Our model did not have any other clinical laboratory values available, such as serum cholesterol or triglycerides, which may have explained more variance. Because many underlying clinical, physiological, and metabolic features contribute to PPGR, stronger predictors may exist that were not assessed in the present study.
Researchers have replicated the PPGR modeling structure on meal-based interventions. Korem et al.(29) performed a crossover study assessing the PPGR following white bread (low-fiber) versus high-fiber sourdough bread. These authors built a model using stochastic gradient boosting that included solely microbiome features (relative abundance of species, relative abundance of genes, and function) that accurately predicted which type of bread induced a lower glycemic response for individuals (ROC=0.83). Although the analysis of the microbiome in our study did not extend beyond the relative abundance of specific taxa, the microbiome still played a leading role in determining the PPGR following potatoes of different RS concentrations. This further demonstrates the important role the microbiome plays in deciphering the inter-individual PPGR after consuming products with varying fiber content, as we found in our study.
In the present study, a key element in the design was to provide a dietary intervention that could feasibly be achieved in a real-world setting. The amount of potato administered (250 g) was equivalent to a serving size of mashed or baked potato, which can be realistically consumed alone or with a meal. We were unable to determine the actual amount of RS2 and RS3 (RS3 exclusively in the cold potato) for each potato administered to participants, but we did quantify the mean RS in the hot and cold potatoes. Moreover, the volume of potato consumed was equivalent between interventions, yet the proportion of available carbohydrate differed, with a lower amount of available carbohydrate in the cold potato. Another limitation of this study includes that the postprandial time period did not allow for adequate assessment of bacterial fermentation of RS and further stimulation of incretins located lower in the gut. We were also limited to microbial data at the 16S rRNA level, while whole genome sequencing could provide deeper insight into the functional role of key microbiota and specific species associated with lowering postprandial glucose. Despite these limitations, a robust modeling technique incorporated common baseline features and selected variables with the greatest influence on PPGR. Further, because this study recruited volunteers without chronic disease, the identified predictors of PPGR may be applicable to other healthy populations.