In the initial search, we retrieved 1373 articles through four database search and the reference lists of included studies and previously published reviews. After removing 563 duplicates, 810 articles were left for further screening. Consistently, reading titles and abstracts of these articles led to the exclusion of 773 articles because they didn’t report the relationship between adherence to DASH diet and risk of DM. Accordingly, the remaining thirty-seven full-text articles were examined for eligibility, and 22 were excluded for the following reasons: 5 did not evaluated DM risk; 1 lacked sufficient data and the corresponding author of this study could not be contacted; 16 did not mention DASH diet score. Finally, fifteen studies(13 cohort and 2 case-control studies) with 557,475 participants and 57,064 cases of DM were included for this systematic review and meta-analysis [6,7,8, 19, 21, 30,31,32,33, 22, 20, 34,35,36,37]. Figure 1. indicated the flow chart of article selection process.

Flow chart of the process of study selection

Characteristics of each eligible study are shown in Table 1. Of these included studies, thirteen were cohort studies [7, 8, 19, 21, 30,31,32,33, 22, 20, 34,35,36], and two were case-control studies [6, 37]. The publication dates of these studies mentioned above varied between 2009 and 2023. Age of study participants ranged from 18 to 84 years. Sample size ranged from 334 to 166500. Eight of the included studies were performed in the United States [7, 8, 19, 21, 30, 22, 32, 33], three in Iran [6, 20, 34], one in Taiwan China [35], one in Singapore [31], one in Brazil [36], and one study in Europe [37]. Fourteen of included studies used food frequency questionnaires(FFQs) to collect dietary data [7, 8, 19, 21, 30,31,32,33, 22, 20, 34,35,36,37], and remaining one study used 24-h dietary recall [6]. In addition, all included articles used methods designed by Fung et al. (7 food groups and sodium) [7, 30,31,32,33, 22, 20, 34, 36], Dixon et al. (7 food groups, saturated fat and alcohol) [6], Günther et al. (8 food groups) [16], Sacks et al. (7 food groups and sodium) [19, 21, 35, 37] to extract DASH diet. Finally, based on the NOS scores, all of the included studies were considered to be of high-quality studies [6,7,8, 19, 21, 30,31,32,33, 22, 20, 34,35,36,37].

Combining 17 effect sizes fSacksrom fifteen articles (557,475 participants and 57,064 cases of DM) were included to evaluate the relationship between adherence to the DASH diet and risk of DM in this study. Figure 2 showed the evidence of a reduced risk of DM in the highest compared with lowest categories of DASH diet (RR:0.82; 95% CI: 0.76, 0.90, P < 0.0001). The high heterogeneity was found among the included studies (P < 0.0001; I2 = 89.1%) and hence the effect was assessed using a random-effects model.

Forest plot of the association between adherence to the DASH diet and risk of DM

Twelve studies [6, 8, 19,20,21, 31, 32, 34,35,36,37] involving 11 cohort studies with DASH diet scores, were included in this dose-response analysis for DM risk. The dose-response meta-analysis showed a linear trend association between adherence to the DASH diet and DM risk (RR:0.99; 95%CI: 0.97–1.02, Pdose-response = 0.546, Pnonlinearity = 0.701) (Fig.3). There was also a linear trend association between DASH diet and DM risk in the analysis of cohort studies (Pnonlinearity = 0.482, Pdose-response = 0.599) (Fig.4).

Dose-response association between adherence to the DASH diet and risk of DM in the analysis of twelve studies