Plant-based dietary patterns and fasting insulin: a cross-sectional study from NHANES 2017–2018

BACKGROUND: Previous studies have created plant-based diet indices to assess the health effects of specific dietary patterns. OBJECTIVE: To examine the association between the plant-based content of diet and fasting insulin in adults from the NHANES 2017–2018 database. METHODS: Demographic, dietary, lab and clinical data and fasting insulin were obtained from the NHANES 2017–2018 database. From two 24-h dietary recalls, we created a plant-based diet index (PDI) and a healthy plant-based diet index (hPDI). A high PDI score indicated more plants were consumed versus animal foods. A high hPDI score indicated healthier, plant materials (whole grains, whole fruits, vegetables, legumes, vegetable oils, seeds and nuts) were consumed. The relationships between the natural log of fasting insulin, PDI, and hPDI were analyzed using multiple linear regression adjusting for body mass index (BMI) and alanine aminotransferase (ALT). RESULTS: Analyses were based on 1,714 participants, 897 women and 817 men with a median age of 52 years. In this sample, 610 (35.6%) were white, 407 (23.8%) were black, 231 (13.5%) were Mexican, 207 (12.1%) were Asian, 157 (9.2%) were other Hispanic, and 102 (6%) were other or mixed race. Median fasting insulin was 9.74 μU/mL (IQR: 6.2, 15.56). For every 1 unit increase in PDI, the natural log of fasting insulin decreased 0.0068 ± 0.003 μU/mL (CI: -0.00097, -0.013) (p = 0.02). After adjusting for BMI and ALT, the PDI did not significantly predict fasting insulin as the association was not robust due to multicollinearity. The hPDI was inversely and significantly associated with the natural log of fasting insulin (-0.0027 ± 0.00134, CI: -0.000087, -0.0053) (p = 0.043) in a multivariable model including BMI and ALT. CONCLUSION: A healthy plant-based diet is associated with a decrease in fasting insulin levels. Healthfulness of the diet is an important factor when considering the benefit of a plant-based diet.