Diabetes mellitus (DM) has become a global epidemic that affects millions of people worldwide. Dietary intervention is obviously effective in the treatment of DM and its associated complications. Multiple factors may contribute to the development of DM, and among them, chronic inflammation is widely considered an underlying pathophysiological mechanism involved in the development of DM (1). Studies suggest that multiple factors that affect inflammation may be associated with DM, such as sex, age, smoking, alcohol, physical activity, the application of some medications, and diet (2). The premise that diet can increase or reduce inflammation has been extensively investigated (3–5), but the degree to which modulating diet can have desirable effects in DM is unclear. It is therefore important to clarify relationships between diet, inflammation, and DM.

The dietary inflammatory index (DII) is a useful tool created by Cavicchia et al. (6) in 2009 and Shivappa et al. (7) in 2014 that has been developed for use in different populations. It can be used to estimate the overall inflammatory potential of an individual’s diet by utilizing a 24-h dietary record interview or food frequency questionnaires (FFQs) (7). The DII is based on 1943 eligible articles that reported the effects of 45 dietary parameters on serum levels of six biomarkers; the proinflammatory biomarkers interleukin (IL)-1β, IL-6, tumor necrosis factor alpha, and C-reactive protein, and the anti-inflammatory biomarkers IL-4 and IL-10 (7). Positive DII scores indicate a proinflammatory diet that significantly increases serum proinflammatory biomarker levels, whereas negative DII scores indicate the contrary. Therefore, DII scores can constitute a quantitative measure for the assessment of the relationships between diet and health outcomes. They can be used to identify the associations between diet and parameters ranging from serum concentrations of inflammatory factors to various types of chronic disease indicators (8) such as cardiovascular diseases (9), metabolic syndrome (10), obesity (11), cancer (12), and DM (13).

The relationship between the DII score and DM is controversial. Some studies suggest a positive relationship between the DII score and DM (13–15), whereas other studies indicate no positive association (16, 17) or a negative association (18). To date, only one narrative review by Phillips et al. has assessed the relationships between DII and DM (19), but no systematic review and meta-analysis has been reported. A systematic review and meta-analyses using published study data may be of great value with respect to clarifying the associations between DII and DM and may provide a theoretical basis for dietary management in DM patients.

The current systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (20).

In the present meta-analysis, individuals with a more proinflammatory diet were at a 32% higher risk of DM. Similarly, in a review by Catherine et al., it was concluded that the DII can be used as an effective tool to evaluate the overall inflammatory nature of diet, and that it helps to identify relationships between diet, inflammation, and DM. However, that review only reported qualified synthesis, whereas the quantitative rate of the relationship remains unclear (19). On the whole, previous evidence about potential associations between DII scores and DM has been limited largely because of a lack of studies, small sample sizes, and differences in participants’ characteristics including age, sex, and race. The present systematic review and meta-analysis incorporated results from five published studies and, thus, identified stronger associations.

Diet is widely believed to be a main contributor to chronic inflammation (5). A Western diet is generally rich in proinflammatory foods such as sugar, red meat, dairy products, refined carbohydrates, and fried foods, and this may increase the serum levels of inflammatory markers (22) and the risk of DM in both males (23) and females (24). In contrast, a Mediterranean diet is generally rich in vegetables, olive oil, and nuts, and low in red meat, with poultry and fish replacing beef and lamb. Several epidemiological studies have assessed the protective effects of a Mediterranean diet with respect to several diseases associated with chronic low-grade inflammation such as cancer, diabetes, obesity, metabolic syndrome, and atherosclerosis (25). Therefore, we investigated and summarized the current evidence on associations between several normal dietary patterns and inflammation, as derived from epidemiological studies (26–31) ( Table 4 ).

The correlation between diet and DM risk can be partly explained by chronic inflammation that may lead to IR and β cell dysfunction (32). The medium function of inflammation may involve multiple mechanisms, as described below. Activation of c-Jun N-terminal kinase by the excessive production of inflammatory factors induces phosphorylation of insulin receptor substrate-1 (IRS-1) at Ser307 leading to its inactivation, thus suppressing insulin receptor signaling (33, 34). Inflammatory factors also reduce insulin-dependent glucose transport and peripheral glucose utilization respectively by inducing the suppressor of cytokine signaling-3 and inhibiting lipoprotein lipase activity (35). Oxidative stress damage induced by inflammation may cause β cell apoptosis and deficiency of insulin secretion. In addition, sustained activation of mTORC1/S6K1 under conditions of nutrient overload (glucose, amino acids—especially branched-chain amino acids such as leucine) via AMP-activated protein kinase generates a feedback loop that suppresses IRS-1 mRNA and protein levels and IRS-1 activity. mTORC1/S6K1 plays a dual role in β cells, regulating β cell size and inhibiting insulin/insulin-like growth factor 1 signal transduction, which may lead to increased apoptosis (36, 37).

The original DII was created as a useful tool for categorizing the diets of individuals on a continuum from maximally anti-inflammatory to maximally proinflammatory, and it can be used in epidemiologic studies to estimate the potential burden of DM related to diet. All dietary indices prior to the DII such as the Healthy Eating Index, Alternative Healthy Eating Index, Dietary Approaches to Stop Hypertension, and Mediterranean Dietary Index were not specifically designed to evaluate the overall inflammatory potential of diet (38–42). In contrast, the DII was developed to reflect all evidence from a wide range of populations as well as laboratory animal and cell culture experiments.

The results of subgroup analyses indicated that individuals with a higher BMI are more susceptible to the effects of inflammatory stimuli. The strong associations between obesity, inflammation, and DM can be partly explained by the recruitment of macrophages into adipose tissue. Excessive inflammatory factors produced by macrophages and adipocytes may lead to the activation of c-Jun N-terminal kinase and subsequently inhibit phosphorylation of IRS-1 serine residues that induce IR  (34). Studies that used FFQs identified stronger associations than those that used dietary records. This may be because DII scores can be affected by dietary patterns and dietary assessment tools. Although completing an FFQ requires more time than completing a dietary record, it is more accurate, enabling investigators to estimate relatively long-range dietary intake, focus on specific foods or nutrients consumed by a given population, and understand relationships between diet and health outcomes (43, 44).

The current meta-analysis had several limitations. Only five studies were eligible for the analysis of DM risk. Because the DII was first created in 2009, to date, few studies have used it, particularly studies investigating associations between DII scores and DM. Different DII score categorical cutoff points (such as quintile and quartile) rendered comparisons less straightforward; thus, we could only obtain a quantitative rate of the association between the DII and DM but not a specific range. As a part of individual investigations and the basis of DII scores, the availability of FFQ data is offset by well-documented limitations such as recall bias, limitations associated with assessing specific cultural food items, and lack of validation in different research settings. Lastly, substantial heterogeneity was observed in the current meta-analysis. One possible explanation for this is the differences in energy intake adjustment used in the studies included. Another may relate to the different demographic characteristics and designs of the studies. Despite the above-described limitations, the DII calculation method used was the same in each study, thus increasing comparability. The present meta-analysis also provides greater statistical power and a more robust pooled estimate of associations by pooling results from all of the included studies.

Diets resulting in high DII scores are associated with an increased risk of DM, particularly in obese individuals. However, the results of the current meta-analysis were based on a small number of studies, and further high-quality prospective studies with larger sample sizes conducted in different populations and countries are required to elucidate this association.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding authors.

YX and C-LG designed this study. Q-QT and X-YD extracted and confirmed the data. Q-QT performed the statistical analysis. Q-QT was responsible for manuscript preparation. YX and C-LG commented on the manuscript. All authors contributed to the article and approved the submitted version.

This work was supported by the Natural Science Foundation of China (NO.81900764, NO.81970676), Sichuan Science and Technology Program (NO.2019YFS0537, NO.2020YFS0456), and Luzhou-Southwest Medical University cooperation project (NO.2020LZXNYDJ32).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.