Table 2.
Characteristics of studies included in the current systematic review owing to reporting the association between dietary inflammatory index (DII) and central obesity
| First author, reference | Year | Country | Study design | Sex | Age range (years) | Sample size/population | Number of cases/controls | Dietary assessment/index | Results | Adjusted variables | Quality of the study |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Shivappa(50) | 2018 | Italy | Cross-sectional | Both | ≥35 | 20 823/general population | 4164/4164 | FFQ/DII | Individuals in highest DII quintile had lower prevalence of HTN, lower SBP and lower FBS compared with lowest. No difference in DBP was observed | Age, sex | 9 |
| Ren(21) | 2018 | China | Cross-sectional | Both | 18–75 | 1712/general population | 579/566 | 24 h recall record/DII | Individuals in highest tertile were 1·40 (95 % CI 1·03, 1·89) times more likely to have HTN compared with lowest tertile. No association with DII and high blood glucose was observed. In gender-stratified results, the DII–HTN association was observed only among women (1·17; 95 % CI 0·80, 1·70) | Age, gender, city, education, family monthly expenditure on food, smoking, BMI | 8 |
| Phillips(25) | 2018 | Ireland | Cross-sectional | Both | 50–69 | 2047/adult general population | 664/664 | Self-completed FFQ/DII | No significant association between DII and HTN was observed | – | 7 |
| Phillips(22) | 2018 | Ireland | Cross-sectional | Both | 50–69 | 1992/general population | 996/996 | FFQ/E-DII | Individuals in higher than median E-DII had higher FPG and SBP values compared with lowest. No significant difference in other glycaemic markers and DBP by DII was reported | – | 7 |
| Park(26) | 2018 | USA | Cross-sectional | Both | 20–90 | 1815/metabolically healthy overweight and obese adults | 634/570 | 24 h dietary recalls/DII | Individuals in highest tertile had higher HOMA-IR compared with lowest. Other parameters were not significantly different | – | 7 |
| Park(26) | 2018 | USA | Cross-sectional | Both | 20–90 | 1918/metabolically unhealthy overweight and obese adults | 610/674 | 24 h dietary recalls/DII | Individuals in highest tertile had lower DBP compared with lowest. Other parameters were not significantly different | – | 7 |
| Farhangi(11) | 2018 | Iran | Cross-sectional | Both | 35–80 | 454/patients, candidates for CABG | 113/113 | FFQ/DII | No significant difference between the prevalence of HTN and serum values of HbA1c in different quartiles | Age, gender, BMI, educational attainment, diabetes, MI | 9 |
| Denova-Gutiérrez(60) | 2018 | USA | Cross-sectional | Both | 20–69 | 1174/general population | 234/235 | FFQ/DII | Individuals in the top quintile of DII had significantly higher FBS and HbA1c compared with the lowest | – | 7 |
| Abdurahman(60) | 2018 | Iran | Cross-sectional | Both | 19–59 | 300/individuals with obesity | 75/75 | FFQ/DII | Non-significant elevation of FPG, SBP and DBP among highest v. lowest quartile of DII. No significant association between DII and HTN or hyperglycaemia in logistic model | Age, sex, PA, BMI, history of chronic diseases | 8 |
| Vissers(44) | 2017 | Australia | Cross-sectional | Women | 52 | 7169/general population | 1664/5505 | FFQ/DII | A pro-inflammatory diet was significantly associated with a higher risk of incident HTN in comparison to the anti-inflammatory diet, with a 24 (95 % CI 6, 46) % higher risk. The prevalence of HTN was also higher among women with more pro-inflammatory diet | Energy intake, age, diabetes, smoking, education, menopausal PA, BMI | 7 |
| Vahid(19) | 2017 | Iran | Case–control | Both | Mean = 47 | 414/pre-diabetics and healthy matched controls | 138/138 | FFQ/DII | DII was associated with higher FPG and HbA1c concentrations | Age, BMI, education, smoking, alcohol, diabetes, LDL-C, TAG | 7 |
| Shivappa(45) | 2017 | USA | Cross-sectional | Both | ≥19 | 12 438/general population | 4119/4183 | 24 h dietary recall/DII | Prevalence of HTN in highest tertile of DII was significantly higher than in the lowest (35·0 v. 32·3 %) | – | 8 |
| Nikniaz(54) | 2018 | Iran | Cross-sectional | Both | 18–64 | 606/general population | 151/151 | FFQ/DII | No significant difference in FBS, SBP, DBP in different DII quartiles; in multivariate logistic model, OR of high FBS was 2·56 times higher in 4th quartile compared with 1st | Smoking, PA, sex, age, BMI | 8 |
| Mirmajidi(28) | 2019 | Iran | Cross-sectional | Both | 18–60 | 171/abdominal obese | 85/86 | FFQ/DII | FBS was significantly higher among individuals with higher than median DII. In regression model, DII was positively associated with FBS | Age, sex, PA, energy intake | 8 |
| Mazidi(43) | 2018 | USA | Cross-sectional | Both | ≥18 | 21 874/general population | 5504/5473 | FFQ/DII | Higher FBS, SBP, DBP, HOMA-IR, insulin, HbA1c and 2 h glucose in highest DII quartile compared with lowest. Being in top quartile of DII made individuals 1·21 times more likely to have HTN | Age, race, sex, income/poverty ratio, education, marriage, BMI | 9 |
| Mazidi(63) | 2018 | USA | Cross-sectional | Both | ≥18 | 21 649 | 5128/5153 | FFQ/E-DII | Prevalence of HTN in 4th DII quartile significantly higher than in lowest (34·1 v. 28·1 %). FBS, SBP and DBP in highest DII quartile were significantly higher | – | 8 |
| Kim(55) | 2018 | Korea | Cross-sectional | Men | 19–65 | 3682 | 921/920 | FFQ/DII | In multivariate logistic regression, being in 4th quartile of DII made men 1·30 times more likely to develop hyperglycaemia | Age, BMI, education, alcohol, smoking, PA, energy intake | 8 |
| Kim(55) | 2018 | Korea | Cross-sectional | Women | 19–65 | 5609 | 1402/1403 | FFQ/DII | No significant association between DII, HTN and hyperglycaemia | Age, BMI, education, alcohol, smoking, PA, energy intake | 8 |
| Sokol(20) | 2016 | Poland | Cross-sectional | Men | 45–65 | 1290/general population | 458/213 | FFQ/DII | No significant association between higher DII and HTN or hyperglycaemia was observed | Age, BMI | 7 |
| Sokol(20) | 2016 | Poland | Cross-sectional | Women | 45–65 | 2572/general population | 507/751 | FFQ/DII | More pro-inflammatory diet was associated with decreased prevalence of HTN and hyperglycaemia was observed | Age, BMI | 7 |
| Moslehi(51) | 2016 | Iran | Cross-sectional | Both | 19–75 | 12 523/general population | 744/743 | FFQ/DII | Prevalence of HTN among top quartile of DII was significantly lower than in the lowest. No significant association between markers of glycaemic status and DII was observed | Sex, age, smoking, PAL, family history of diabetes, HTN, glucose- and lipid-lowering medication use, BMI | 8 |
| Ramallal(47 ) | 2015 | Spain | Cohort | Both | Mean = 38 (sd 12) | 18 794/general population | 4698/4699 | FFQ/DII | Baseline prevalence of HTN in different DII quartiles was not different. Individuals in highest DII quartile had higher OR of HTN compared with lowest | Baseline and family history of diabetes, HTN, CVD, hypercholesterolaemia, special diets, smoking, energy, PA, BMI, education, alcohol, snacking, sitting time, time watching TV | 8 |
| Neufcourt(49); baseline analysis | 2015 | France | Cohort | Both | 35–60 (women), 45–60 (men) | 3726/general population | 932/930 | 24 h dietary records/DII | Significantly higher serum glucose and SBP values in highest v. lowest DII quartile. No significant difference in DBP values | – | 8 |
| Neufcourt(49); after follow-up analysis | 2015 | France | Cohort | Both | 35–60 (women), 45–60 (men) | 3726/general population | 932/930 | 24 h dietary records/DII | Significantly higher SBP and DBP values after 13-year follow-up in highest v. lowest DII quartile. No significant difference in FBS | Age gender, supplementation group, energy, number of 24 h records, education, smoking, PA, baseline values | 8 |
| Alkerwi(59) | 2015 | Luxembourg | Cross-sectional | Both | 18–69 | 1352/general population | 338/338 | FFQ/DII | No significant association between DII, SBP, DBP and glycaemic biomarkers was reported | Age, sex, education level, smoking status, PA, energy intake | 9 |
| Wirth(61) | 2014 | USA | Cross-sectional | Both | Mean = 42·4 (sd 8·5) | 447/police officers | 112/111 | FFQ/DII | No significant difference in insulin and FBG between different DII quartiles was observed; however, odds of hyperglycaemia among individuals in 4th quartile was 2·03 times more than in 1st quartile | Age, sex, alcoholic drinks per week | 5 |
| Alkerwi(52) | 2014 | Luxembourg | Cross-sectional | Both | 18–69 | 1352/general population | 450/450 | FFQ/DII | Prevalence of HTN in highest DII tertile was significantly lower than in lowest. SBP in highest tertile was significantly lower than in lowest. No significant in difference in DBP, glucose, insulin and HOMA-IR was observed. Also, no significant association between DII and HTN or hyperglycaemia was observed | Age, sex, education, income, smoking, PA | 8 |
| Woudenbergh(62) | 2013 | Netherlands | Cross-sectional | Both | Mean = 64 (sd 9) | 1024/general population | 341/341 | FFQ/DII-ADII | ADII was adversely associated with HOMA-IR, fasting glucose and post-load glucose but not with HbA1c | Age, sex, cohort, PA, smoking, family history of diabetes, use of lipid-lowering medication, HTN, energy intake | 8 |
| Sánchez-Villegas(53) | 2015 | Spain | Cross-sectional | Both | Mean ≈ 38 | 15 093/university graduates | 3018/3019 | FFQ/DII | HTN prevalence in the highest quintile was significantly lower than in the lowest | – | 6 |
| Naja(27) | 2017 | Lebanon | Cross-sectional | Both | >18 | 331/general population | 66/67 | FFQ/DII | No significant difference in mean DII of individuals with HTN or hyperglycaemia compared with healthy individuals was reported. No significant association was observed between DII and hyperglycaemia or HTN in logistic regression | Age, sex, marital status, education, crowding index, PA, smoking | 9 |
| Hayden(57) | 2017 | USA | Cross-sectional | Women | 65–79 | 7085/older women | 1467/2694 | FFQ/DII | No significant difference in the prevalence of hypertension between lowest v. highest quartile was observed | – | 6 |
| Camargo-Ramos(58) | 2017 | Colombia | Cross-sectional | Both | Mean = 39·7 (sd 6.9) | 90/overweight and sedentary adults | 77/13 | 24 h dietary record/DII | Elevated HbA1c in individuals with pro-inflammatory v. anti-inflammatory diet. Glucose, SBP and DBP were non-significantly higher | Age, sex | 7 |
| Bodén(46) | 2017 | Sweden | Case–control | Men | Mean = 50 | 5284/general population | 1321/1321 | FFQ/DII | Significantly higher SBP values in highest v. lowest DII quartile | – | 7 |
| Bodén(46) | 2017 | Sweden | Case–control | Women | Mean = 50 | 1600/general population | 400/400 | FFQ/DII | No significant difference in SBP between different DII quartiles | – | 7 |
| Wirth(48) | 2016 | USA | Cross-sectional | Men | >20 | 7566/general population | 2097/1586 | 24 h recall/DII | No significant association between DII and HTN was observed | Family member smoking, age, BMI | 8 |
| Wirth(48) | 2016 | USA | Cross-sectional | Women | >20 | 8047/general population | 1818/2326 | 24 h recall/DII | Significant association between DII and HTN was observed | Family member smoking, age, BMI | 8 |
CABG, coronary artery bypass grafting; E-DII, energy-adjusted dietary inflammatory index; ADII, adapted dietary inflammatory index; HTN, hypertension; SBP, systolic blood pressure; FBS, fasting bold sugar; DBP, diastolic blood pressure; FPG, fasting plasma glucose; HOMA-IR, homeostatic model assessment of insulin resistance; HbA1c, glycated Hb; MI, myocardial infarction; PA, physical activity; LDL-C, LDL-cholesterol; PAL, physical activity level; TV, television.