Background
According to the International Diabetes Federation, 537 million adults (7.1% of the world’s population) in 2021 were living with diabetes, a figure predicted to rise to 783 million by 2045 [
1]. Diabetes caused 6.7 million deaths in 2021, which is equivalent to one death every 5 s [
2]. Prediabetes represents an intermediate state between normal blood glucose levels and clinical diabetes. Individuals with intermediate hyperglycemia are at increased risk of developing type 2 diabetes, but not everyone goes on to develop type 2 diabetes [
3]. There is presently no cure for diabetes, and lifestyle modification with a healthy diet is regarded as the cornerstone of diabetes prevention, potentially conferring a 40–70% relative-risk reduction [
4].
Accumulating evidence has shown that low-grade systemic inflammation plays a causal role in chronic diseases including type 2 diabetes, and dietary patterns are related to inflammation [
5]. In recent years, a few studies have assessed diet quality based on its inflammatory potential and estimated the association between dietary inflammation and chronic disease [
6‐
8]. Among the aforementioned studies, the most common inflammatory biomarker examined was high-sensitivity C-reactive protein (hsCRP) [
8]. However, there is considerable variation in dietary habits across different populations, and characterization of dietary inflammation in a large European population has thus far been limited.
Currently, a few population-based studies have reported a significant association between dietary patterns with higher inflammatory potential and increased risk of type 2 diabetes [
9,
10]. However, the lack of such an association in other studies raises questions [
11,
12]. Additionally, one randomized controlled feeding study with a small sample failed to observe a significant association between an anti-inflammatory diet and prediabetes [
13]. So far, no studies have evaluated the impact of an anti-inflammatory diet on the progression from prediabetes to diabetes.
Genetic and lifestyle-related factors may both contribute to the development of type 2 diabetes [
14]. Moreover, the effect of genetic variants may change in response to alterations in the environment [
15]. Previous studies have demonstrated that genetic variation and interplay between diet and genetic predisposition may account for considerable individual differences in response to dietary prevention of T2D [
14‐
16]. Investigating gene-diet interactions in type 2 diabetes development therefore offers a unique opportunity to identify susceptible populations and determine to what extent they may benefit from personalized nutrition recommendations for type 2 diabetes prevention [
16]. However, given the conflicting and limited research currently available, the question remains whether adherence to a low-inflammatory diet may mitigate genetic predisposition to type 2 diabetes.
In the current study, we sought to (1) calculate an inflammatory diet index (IDI) to assess dietary inflammatory potential; (2) examine the associations between a low-inflammatory diet and risk of type 2 diabetes among normoglycemic and prediabetic participants; and (3) investigate whether a low-inflammatory diet may mitigate diabetes-related genetic risk using data from the large population-based cohort study within the UK biobank.
Discussion
In this large-scale, nationwide prospective study of type 2 diabetes-free adults consisting of normoglycemic and prediabetic individuals, we found that (1) low IDI score (calculated from 16 anti-inflammatory and 18 pro-inflammatory foods) was dose-dependently associated with decreased type 2 diabetes risk; (2) a low-inflammatory diet may delay type 2 diabetes onset by about 2 years among participants with normoglycemia and 1.2 years among prediabetic participants; and (3) a low-inflammatory diet might significantly mitigate the risk of genetic factors on type 2 diabetes development.
A few previous studies have reported inflammation-based dietary patterns in different populations [
6‐
8,
30]. Given that dietary habits vary across populations and the limited investigation of dietary inflammation in European populations, we calculated an IDI using a data-driven method to predict chronic low-grade systemic inflammation in approximately 190,000 European participants. Shivappa et al. developed the dietary inflammatory index (DII) that accounted for 45 pro- and anti-inflammatory food parameters, most of which are specific micronutrients and macronutrients (such as vitamin B12, vitamin C, protein, and n-3 fatty acids) rather than whole foods [
6,
31]. However, the DII is difficult for the general public to easily understand and directly make use of given that we consume foods (consisting of numerous and interacting micro- and macronutrients) rather than isolated nutrients. Our IDI identified some key anti-inflammatory (nuts, vegetables, fruits, wine, coffee) and pro-inflammatory foods (red meat, processed meat, organ meet, sugar-sweetened beverages) which are consistent with previous studies involving food-based indices of dietary inflammation in other populations, such as the Anti-Inflammatory Diet Index created in a Nordic population (women,
n=3503) [
8,
32], the EDIP scores developed and validated in the US population [
7], and a dietary inflammatory potential score constructed in the Chinese population [
30]. Despite variations across studies in dietary habits and the type of inflammatory markers examined, the considerable overlap of our findings with previous reports underscores the involvement of these foods in modulating chronic inflammation.
Chronic low-grade inflammation substantially contributes to the development of type 2 diabetes [
33,
34], and diet is one of the modifiable lifestyle-related factors that might partially modulate inflammation [
6,
35]. Few studies to date have examined the relationship between dietary patterns with higher inflammatory potential and type 2 diabetes risk [
9,
10], and findings have been inconsistent [
11]. A previous longitudinal study from the USA has documented that higher dietary inflammatory potential was strongly related to an increased risk of type 2 diabetes among participants in the Nurses’ Health Study and the Health Professionals Follow-up Study [
9]. Laouali et al. demonstrated that diets with elevated anti-inflammatory potential were associated with a lower risk of type 2 diabetes in a French prospective cohort of women [
10]. A cross-sectional study using the dietary inflammatory index reported a positive association between a pro-inflammatory diet and type 2 diabetes among 1174 adult Mexicans [
36]. By contrast, one Dutch cross-sectional study with a limited sample (
n=1024) found that the adapted dietary inflammatory index was not significantly associated with glycated hemoglobin [
11]. Another cross-sectional investigation among Iranian adults also documented non-significant associations between a pro-inflammatory diet and risk of insulin resistance [
12]. Moreover, only one randomized controlled feeding study conducted in Portland, Oregon, with a small sample (intervention group:
n=20) failed to observe improved blood sugar levels among prediabetic participants following a 6-week anti-inflammatory diet developed by scientists and naturopathic physicians [
13]. Thus, evidence on the relationship between low-inflammatory diets and type 2 diabetes has remained unclear. In the present study, we found that diets lower in inflammation were associated with lower risk of type 2 diabetes among both normoglycemic and prediabetic participants. To the best of our knowledge, this is the first study to demonstrate that a low-inflammatory diet may delay the progression from prediabetes to type 2 diabetes.
It is widely known that type 2 diabetes is a complex genetic and lifestyle-related disorder [
37]. Previously published genetic studies have demonstrated diet may differentially affect type 2 diabetes risk depending on an individual’s genetic risk [
38]. Thus, the consequences of adhering to a healthy diet are considerably complex. A previous prospective study including 357,419 UK Biobank participants reported that higher diet quality (evaluated based on 10 foods predictive of type 2 diabetes risk) was significantly associated with greater reductions in blood HbA1c levels and type 2 diabetes risk among individuals with a higher genetic risk, but not among those at a lower genetic risk [
16]. Another cross-sectional study in 11,657 participants from a community-based population in China identified evidence that fruit intake alleviated the relationship between genetic predisposition and type 2 diabetes risk [
39]. It has also been reported that dietary fiber intake may modify the association between genetic factors and type 2 diabetes incidence [
40]. In the current study, we found that adherence to a low-inflammatory diet might significantly mitigate the risk of genetic factors on type 2 diabetes development.
The biological mechanisms responsible for the decreased type 2 diabetes risk attributable to a low-inflammatory diet, especially among participants at higher genetic risk, are multifactorial and incompletely understood. Consumption of pro-inflammatory foods, especially red meat, processed meat, and sweets which contain disease-promoting components such as saturated fat, advanced glycation end products, heme iron, nitrosamine, sodium nitrite, and nitroso compounds, may have toxic effects on pancreatic
β-cells or impair insulin sensitivity [
6]. Instead, adhering to a low-inflammatory diet may improve long-term hyperglycemia, metabolic disturbances, lipid profile, body composition, blood pressure, insulin sensitivity, and β-cell function [
16,
35], which all play important roles in the development of type 2 diabetes. While genetic variants such as TBC1D4 and TCF7L2 result in up to 50% increased risk of type 2 diabetes by diminishing the incretin effect and impairing glucagon-like peptide 1–induced insulin secretion, a fiber-rich diet may stimulate glucagon-like peptide 1 and mitigate this genetic risk [
38,
40]. More experimental research is needed to provide biological insight into potential gene-diet interactions involved with type 2 diabetes risk.
The major strength of the present study is the large, well-characterized, population-based, prospective cohort with available genetic information. This sample provides a unique opportunity to detect associations between a low-inflammatory diet and type 2 diabetes incidence and whether these associations differ by genetic risk, while also controlling for potential confounders such as socioeconomic characteristics and lifestyle factors. Nonetheless, the limitations in the current study need to be acknowledged. First, as the information of fasting blood glucose were missing in the ascertain of incident diabetes in the UK Biobank, participants with undiagnosed type 2 diabetes might have been misclassified as type 2 diabetes-free, which could have caused an underestimation of the observed associations. Second, inflammatory diets in previous studies were developed based on a valid food frequency questionnaire (FFQ) which included questions about commonly consumed food items over a specified period (such as 12 months). Although dietary assessment in the present study was not based on an FFQ in the UK Biobank (only 29 questions about the consumption frequency of six food groups), the 24-h dietary assessment is more comprehensive and accurate and less likely to cause recall bias. Third, we did not exclude participants (45.6%) who had only one 24-h dietary assessment due to loss of sample size, which may less accurately reflect usual dietary habits [
41]. However, we repeated analyses after excluding participants with only one 24-h dietary assessment, and the results were not meaningfully altered. Fourth, data on dietary patterns were obtained only at baseline. Any variation in dietary habits throughout follow-up were not captured, which could introduce bias. Fifth, data on other inflammatory markers such as TNF-α and IL-6 were not available in the UK Biobank, so we generated the IDI using only hsCRP levels. Sixth, the generalizability of our findings could be limited to the source population because of varied dietary habits in different populations. Finally, the participants were volunteers of entirely white British ancestry, so caution is required in generalizing our findings to individuals of other ethnic backgrounds.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.