1344 Diabetes Care Volume 43, June 2020 Jung Ho Gong,1,2 Kenneth Lo,1–3 Qing Liu,1,2 Dietary Manganese, Plasma Jie Li,1–3 Shuiqing Lai,1–3 fl Aladdin H. Shadyab,4 Chrisa Arcan,5 Markers of In ammation, and the Linda Snetselaar,6 and Simin Liu1,2,7 Development of Type 2 Diabetes in Postmenopausal Women: Findings From the Women’s Health Initiative Diabetes Care 2020;43:1344–1351 | https://doi.org/10.2337/dc20-0243 1Department of Epidemiology, Brown University, Providence, RI OBJECTIVE 2Center for Global Cardiometabolic Health, Brown To examine the association between manganese intake and the risk of type 2 University, Providence, RI 3 diabetes in postmenopausal women and determine whether this association is Department of Cardiology, Guangdong Cardio- fl vascular Institute, Hypertension Research Labo- mediated by circulating markers of in ammation. ratory, Guangdong Provincial People’s Hospital, Guangdong Provincial Key Laboratory of Coro- RESEARCH DESIGN AND METHODS nary Heart Disease Prevention, Guangdong Acad- We included 84,285 postmenopausal women without a history of diabetes from the emy of Medical Sciences, South China University national Women’s Health Initiative Observational Study (WHI-OS). Replication of Technology School of Medicine, Guangzhou, China analysis was then conducted among 62,338 women who participated in the WHI- 4Department of Family Medicine and Public Clinical Trial (WHI-CT). Additionally, data from a case-control study of 3,749 women Health, University of California San Diego School nested in the WHI-OS with information on biomarkers of inflammation and of Medicine, La Jolla, CA 5 endothelial dysfunction were examined using mediation analysis to determine the Department of Family, Population, and Preven- tive Medicine, Renaissance School of Medicine, relative contributions of these known biomarkers by which manganese affects Stony Brook University, Stony Brook, NY type 2 diabetes risk. 6Department of Epidemiology, University of Iowa College of Public Health, Iowa City, IA RESULTS 7Hallett Center for Diabetes and Endocrinology, Compared with the lowest quintile of energy-adjusted dietary manganese, WHI-OS Rhode Island Hospital, Warren Alpert Medical School of Brown University, Providence, RI participants in the highest quintile had a 30% lower risk of type 2 diabetes (hazard fi Corresponding author: Simin Liu, simin_liu@brown ratio [HR] 0.70 [95% CI 0.65, 0.76]). A consistent association was also con rmed in .edu CARDIOVASCULAR AND METABOLIC RISK the WHI-CT (HR 0.79 [95% CI 0.73, 0.85]). In the nested case-control study, higher Received3 February2020 and accepted22 March energy-adjusted dietary manganese was associated with lower circulating levels of 2020 inflammatory biomarkers that significantly mediated the association between This article contains Supplementary Data dietary manganese and type 2 diabetes risk. Specifically, 19% and 12% of type 2 online at https://doi.org/10.2337/dc20-1234/ diabetes risk due to manganese were mediated through interleukin 6 and hs-CRP, suppl.12026796. respectively. J.H.G. and K.L. contributed equally to this study. © 2020 by the American Diabetes Association. CONCLUSIONS Readers may use this article as long as the work is Higherintake of manganese was directly associated with a lower type 2diabetes risk properly cited, the use is educational and not for profit, and the work is not altered. More infor- independent of known risk factors. This association may be partially mediated by mation is available at https://www.diabetesjournals inflammatory biomarkers. .org/content/license. care.diabetesjournals.org Gong and Associates 1345 Manganese is an essential element mainly study procedures have been previously included in the analysis (Supplementary obtained from nuts, grains, fruits, green described (15). WHI-OS participants were Fig. 1). Details of the study design have vegetables, and caffeinated drinks (1,2). monitored for an average of 10.8 years been published elsewhere (5,6). The pro- Manganese plays significant roles in mul- until 30 September 2010. Participants spective cohort and the nested case- tiple physiological functions, including were selected at baseline based on the control study were both reviewed and glucose and lipid metabolism and insulin following criteria: 1) no history of di- approved by human subjects review productionandsecretion (3).Manganese abetes, 2) no baseline antidiabetic med- committees at each participating insti- deficiency leads to impaired glucose tol- ication use, 3) complete information of tution, and signed informed consent was erance and increased risk of metabolic food frequency questionnaires at base- obtained from all women enrolled. syndrome through impaired glucose line, and 4) plausible baseline total en- and lipid metabolism (3,4). Manganese ergy intake (between 600 and 5,000 kcal/ Identification of Diabetes at Baseline deficiency also leads to mitochondrial day). Among the participants enrolled in and Follow-up Period oxidative stress by increasing the pro- the WHI-OS, 84,285 met the criteria Diabetes was defined as a physician di- duction of reactive oxygen species (ROS) (Supplementary Fig. 1). Missing values agnosis of treated diabetes. Treated di- (4), a group of physiological products that for covariates were imputed using mul- abetes at baseline and during follow-up contribute to inflammation and endo- tivariate imputation by chained equa- was identified by a self-administered thelial dysfunction (5–8). tions (16). Participants enrolled in the questionnaire at each semiannual con- A mouse study demonstrated that a WHI Clinical Trial (WHI-CT) were included tact, “Since the date given on the front of manganese-rich diet can downregulate for a replication analysis. WHI-CT con- this form, has a doctor prescribed any of ROS generation (9), while in vitro and sisted of three overlapping components: the following pills or treatments?” where animal studies demonstrated reduced Hormone Therapy Trial, Dietary Modifi- the choices were “pills for diabetes” and endothelial dysfunction followed by man- cation Trial, and Calcium and Vitamin D “insulin shots for diabetes.” This identi- ganese supplementation (10). Another Trial. Among the 68,132 participants en- fication method has been used in prior study reported that manganese supple- rolled in the WHI-CT, 62,338 met the publications by the WHI investigators mentation reduced levels of inflamma- selection criteria (Supplementary Fig. 2). (5,17,18). A confirmation study on the WHI tory biomarkers in blood in rats (11). type 2 diabetes questionnaire showed that These studies have provided a physio- Nested Case-Control Study theprevalenceandincidenceoftype2 logical basis for examining manganese in In addition to analyzing data from the diabetes cases were consistent with med- diabetes prevention. To our knowledge, main WHI-OS and CT, we used data ication inventories of oral pills and insulin there is paucity of research on the asso- from a case-control study nested within shots for 77% of WHI-OS and for 79% of ciation between dietary manganese and WHI-OS to explore whether the associ- WHI-CT participants (19). type 2 diabetes. A study of two Chinese pro- ation between dietary manganese and spective cohorts suggested that greater type 2 diabetes was mediated by inflam- Measurement of Dietary Variables dietary manganese intake was associated matory processes or endothelial dysfunc- At baseline, all WHI participants com- with lower risk of type 2 diabetes (12). tion. Among all participants from the pleted a semiquantitative food frequency Association between dietary manga- WHI-OS, 3,781 were included in the questionnaire (FFQ), which was exten- nese and type 2 diabetes in a large cohort nested case-control study designed to sively validated using biomarkers and of postmenopausal women has not yet examine the role of inflammatory bio- dietary records (20). The main section of been investigated. Postmenopausal women markers and endothelial dysfunction in the WHI FFQ includes 122 food groups are associated with a higher risk of type 2 the risk of type 2 diabetes. Control sub- reported in a semiquantitative manner in diabetes compared with premenopausal jects were matched to case subjects by servings/day. Responses to questionnaire women (13), and pathogenesis of men- age (62.5 years), racial/ethnic group items ranged from “Never or less than opausal symptoms has been reported as (white/Caucasian, black/African, Hispanic/ once per month” to a certain number of the result of oxidative stress and ROS Latino, and Asian/Pacific Islander), clin- servings per day depending on the food generation (14). In the current study, we ical center (geographic location), time of group. Nutrient analysis for the WHI was sought to prospectively investigate the blood draw (60.10 h), and the length of conducted at the Fred Hutchinson Cancer association between dietary manganese follow-up. We included these partici- Research Center Nutrition Assessment and type 2 diabetes risk in a large cohort pants to assess the role of insulin re- Shared Resource using the Nutrition Data of postmenopausal women enrolled in sistance, b-cell function, inflammatory System for Research (Nutrition Coordi- the Women’s Health Initiative (WHI) and biomarkers, and endothelial dysfunction nating Center, University of Minnesota, to explore potential explanations for this in the association between dietary man- Minneapolis, MN). This study included association using dietary antioxidant ganese and type 2 diabetes. Among these the following dietary nutrients