Longitudinal associations of dietary fiber and its source with 48-week weight loss maintenance, cardiometabolic risk factors and glycemic status under metformin or acarbose treatment: a secondary analysis of the March randomized trial

The global surge in type 2 diabetes mellitus (T2DM) presents a significant public health challenge. T2DM, characterized by insulin resistance, hyperglycemia, and hyperinsulinemia, is strongly linked to various comorbidities, notably cardiovascular disease. Effective management requires a combination of pharmaceutical interventions and lifestyle modifications, particularly dietary changes. While glucose-lowering drugs like metformin and acarbose are cornerstones of T2DM therapy, the synergistic effects of these medications with dietary factors, especially dietary fiber, remain insufficiently understood. Most previous studies have examined these elements separately. The Metformin and Acarbose in Chinese as the initial Hypoglycemic treatment (MARCH) trial provided a unique opportunity to investigate the interaction between dietary factors and antidiabetic therapies. Previous analyses of MARCH data have shown the impact of macronutrient intake, particularly carbohydrates, on glycemic control and weight management. However, the quality of carbohydrates, specifically the role of dietary fiber, needs further exploration. Dietary fiber, particularly from various sources, offers immense potential in T2DM prevention and management. Its impact varies based on source (whole grains, legumes, fruits, vegetables) and its proportion within total carbohydrate intake. Moreover, the effects of dietary fiber can be influenced by the type of medication used (metformin and acarbose have distinct mechanisms of action). Given the lack of research focusing on the combined effects of dietary fiber and specific antidiabetic drugs, this study aimed to examine the longitudinal associations between dietary fiber intake from various sources, the carbohydrate-to-fiber ratio, and clinical outcomes (weight loss, glycemic control, and cardiovascular risk factors) in newly diagnosed T2DM patients receiving metformin or acarbose treatment over 48 weeks within the MARCH trial.

Existing epidemiological studies and clinical trials often evaluate pharmaceutical interventions or dietary intakes in isolation, neglecting their potential synergistic or interactive effects on T2DM management. Studies have shown the benefits of increased dietary fiber consumption in alleviating T2DM, but the complexity of fiber sources and their varying effects based on structure and physicochemical properties pose a challenge in drawing definitive conclusions. Metformin, a first-line treatment, reduces glucose absorption, inhibits hepatic gluconeogenesis, and enhances peripheral glucose uptake by increasing insulin sensitivity. Acarbose, an α-amylase inhibitor, prevents postprandial glucose spikes by slowing carbohydrate digestion. The bioavailability of these oral drugs can be influenced by the presence of dietary fibers in the gastrointestinal tract, making it crucial to understand their interaction. While the MARCH trial compared acarbose and metformin, prior research lacked studies exploring the combined effects of dietary fiber and these drug regimens. This secondary analysis aimed to fill this knowledge gap by investigating the complex interplay between specific dietary fiber sources, total fiber intake, carbohydrate-to-fiber ratio, and therapeutic outcomes.

This secondary analysis utilized data from the MARCH trial, a 48-week, multicenter randomized controlled trial (RCT) comparing acarbose and metformin as initial therapy in newly diagnosed Chinese T2DM patients. Initially, 1099 patients were screened, with 784 initiating treatment (393 metformin, 391 acarbose). After excluding participants with insufficient dietary data or implausible energy intake, the final analysis included 551 participants (286 acarbose, 265 metformin). Dietary data were collected at baseline, 24 weeks, and 48 weeks using a 24-hour dietary recall method, assessing intake of total dietary fiber, vegetable fiber, fruit fiber, legume fiber, whole grain fiber, and the carbohydrate-to-fiber ratio. Clinical and biochemical measurements included glucose metabolism indicators (HbA1c, FBG, 2hPPG), insulin resistance markers (HOMA-IR, HOMA-β, EISI, AUC insulin, WBISI), hormone secretion (AUC glucagon, GLP-1), and cardiometabolic risk factors (BMI, TC, LDL-C, HDL-C, triglycerides, SBP, DBP). Statistical analysis using SPSS and R involved Friedman’s test to assess changes in dietary measures over time, and mixed-effects linear and restricted cubic spline (RCS) regression analyses to examine longitudinal associations between dietary measures and clinical outcomes, adjusting for relevant covariates. A two-tailed P<0.05 indicated statistical significance.

Analysis of dietary changes showed significant reductions in legume fiber in the acarbose group and vegetable fiber in the metformin group over 48 weeks. The carbohydrate-to-fiber ratio significantly increased over time in both groups. In the acarbose group, higher total and whole grain fiber intake was positively associated with improved β-cell function (HOMA-β) and postprandial glycemic control (2hPPG). Higher whole grain fiber was also linked to better insulin sensitivity (WBISI). However, higher fruit fiber correlated with increased DBP, TC, and LDL-C, suggesting potential cardiovascular risk. Higher vegetable fiber intake reduced DBP but increased HbA1c. Legume fiber showed a positive correlation with reduced 2hPPG. RCS analysis revealed non-linear, S-shaped associations between the carbohydrate-to-fiber ratio and 2hPPG and HDL-C in the acarbose group. In the metformin group, legume fiber intake was positively associated with weight loss (BMI), FBG, and 2hPPG reductions. However, higher whole grain fiber showed a positive correlation with DBP. RCS analysis showed non-linear relationships between carbohydrate-to-fiber ratio and AUC for serum insulin and BMI in the metformin group.

The findings suggest that different dietary fiber sources exert varying effects on metabolic outcomes, depending on the type of antidiabetic medication. In the acarbose group, whole grains and total fiber promoted improved β-cell function and glycemic control. Legume fiber demonstrated consistent benefits in both treatment groups for glycemic control, with additional weight loss benefits under metformin. The negative impact of a high-carbohydrate, low-fiber diet under acarbose treatment and the positive effects under metformin highlight the complex interactions between diet, medication, and individual responses. These results support the concept that optimizing dietary fiber intake alongside medication may improve T2DM management. The differing effects of various fiber types could be due to their solubility, impact on gut transit time, and effects on satiety. The gut microbiome's role in mediating the effects of dietary fiber and drugs needs further investigation, as fiber can modulate gut microbiota composition, impacting SCFA production and potentially influencing drug bioavailability or efficacy. The inconsistent findings regarding fruit and vegetable fiber emphasize the need for further research to identify specific components and their individual effects on lipid metabolism. The non-linear relationship between carbohydrate-to-fiber ratio and metabolic outcomes reinforces the importance of carbohydrate quality in addition to quantity.

This study demonstrates the added benefits of combining metformin or acarbose with specific dietary fiber sources in newly diagnosed T2DM patients. Whole grain and legume fibers showed consistent positive effects on glycemic control, with legume fiber also contributing to weight loss under metformin treatment. These results highlight the potential for synergistic effects between dietary fiber and antidiabetic medications. Future research should investigate the role of the gut microbiome in mediating these interactions, clarify the effects of specific fiber types on lipid profiles, and develop more nuanced dietary recommendations for optimized T2DM management.

This study's limitations include the use of 24-hour dietary recall, which may be subject to recall bias and may not fully represent habitual dietary patterns. The sample size, while substantial, might limit the generalizability of findings to other populations. The exclusion of participants with insufficient dietary data raises the potential for selection bias. Future studies should incorporate more comprehensive dietary assessment methods, larger and more diverse participant populations, and incorporate gut microbiome analysis to further elucidate the complex interactions between diet, medication, and the gut microbiome.