Search

While diabetes prevalence in U.S. adults has increased in the last three decades, less is known about trends in ocular complications of diabetes, particularly in adults aged Disclosure E. A. Lundeen: None. M. Kim: None. D. B. Rein: None. J. Wittenborn: None. C. Cai: None. J. B. Saaddine: None.

Objective: Rates for non-traumatic lower-extremity amputations (LEA) have been declining, but concerns exist over disparities in rates between races/ethnicities. Our objective is to track changes in major LEA (MLEA) rates before and after the Affordable Care Act (ACA) among Medicare beneficiaries residing in the Diabetes Belt compared to those in the surrounding areas. Methods: We used Medicare claims files for a sample of ~1 million Fee-for-Service beneficiaries in 2006 - 2015 to compute rates per 1000 patients with diabetes. The Diabetes Belt was identified by the CDC as 644 counties across Appalachian and southeastern counties in the US that had diabetes prevalence ≥ 11% in 2008. Results: MLEA rates for the Diabetes Belt and surrounding counties as well as rates for Non-Hispanic (NH) Black and NH White are in Figure. Patients in the Belt experienced one full MLEA more per 1000 patients than those in the surrounding area. Although amputation rates declined rapidly in both areas, NH blacks consistently had > 3 times higher rates than NH whites in the Belt. Trends did not show noticeable changes in MLEAs before and after the ACA. Discussion: Our data show persistent disparities in MLEA rates between the Diabetes Belt and surrounding counties. Racial disparities were much larger in the Belt. Targeted policies to prevent MLEAs among NH black patients are needed to reduce the disparities in this region. Disclosure J. M. Lobo: None. S. Kim: None. H. Kang: None. T. L. Mcmurry: None. M. Sohn: None. Funding National Institutes of Health (R01DK113295)

Overall prevalence of diabetes has increased since the year 2000 in the US, disproportionately affecting low-income populations. Recent changes in income-related inequalities (IRI) in diabetes prevalence are unknown. We estimated IRI in diagnosed diabetes in 2001−2018 among US adults aged ≥ 18 years using data from the National Health Interview Survey. We assessed IRI using the Concentration Index (CI). The CI ranges from -1 to 1, with negative values indicating diabetes was concentrated among low-income groups. Trends were assessed using Joinpoint regression. During 2001−2018, all CIs were negative (Figure). The degree of inequality (absolute value of CIs) decreased from 0.16 in 2001 to 0.12 in 2011 (annual percentage change [APC] -2.5, p=0.013), and then increased to 0.18 in 2018 (APC 4.7, p=0.004). The degree of IRI in diabetes differed by sex and age group. Diabetes was more concentrated in low-income groups among females than males. For females, the degree of inequality decreased during 2001−2012 (APC -2.5, p=0.024) and increased after 2012 (APC 4.7, p=0.097); for males, the trend of the inequality was not statistically significant. Among age groups, adults aged 45-64 years had the highest level of IRI in diabetes. Also, the degree of inequality increased in each age group during 2001−2018 (all p Disclosure Y. Chen: None. X. Zhou: None. K. M. Bullard: None. P. Zhang: None. G. Imperatore: None. D. B. Rolka: None.

Diabetes-related end-stage kidney disease (ESKD-D) disproportionately affects U.S. racial/ethnic minority populations compared with whites. With the recent plateau in diabetes prevalence, we assessed ESKD-D trends by race/ethnicity to determine if disparity gaps have changed. From the U.S. Renal Data System, we obtained the number of adults (whites, blacks, Hispanics, Asians, and Native Americans [NAs]) aged ≥18 years with newly treated ESKD-D (with diabetes listed as primary cause) between 2000 and 2016. ESKD-D incidence rates by race/ethnicity were calculated using Census population estimates and age-adjusted to the 2000 U.S. standard population. Average annual percentage change (AAPC) in rates was estimated using joinpoint regression. From 2000 to 2016, the age-adjusted ESKD-D rates decreased by 53% for NAs (66.7 to 31.2 per 100,000, AAPC= -4.5%, p Disclosure N. Burrows: None. Y. Zhang: None. I.A. Hora: None. M.E. Pavkov: None. G. Imperatore: None.

Background: To our knowledge, no estimates of lifetime risk of diabetes have been produced for India, a country with a high propensity to developing diabetes and a low overall body weight compared to high-income countries. Data: We estimated the lifetime risk of diabetes in metropolitan India, and its variation by sex, age and BMI. We used government mortality statistics, diabetes prevalence from the Indian Council for Medical Research INdia DIABetes study (2008-15) and incidence from the Centre for Cardiometabolic Risk Reduction in South Asia (2010-2018). We limited estimates Indians aged 20 or more years. Results: Overall lifetime risk of diabetes in women in metropolitan cities aged 20 was 74.7% (95%CI: 65.4-83.3%), and 69.0% (95%CI: 58.9-78.7%) in men. Remaining risk is highest among the obese (92.9% (95%CI: 88.0-96.0%) in women and 92.4% (95%CI: 86.6-96.1%) in men aged 20) and is lowest among underweight/normal weight people (56.4% (95%CI: 46.9-66.3%) in women and 54.2% (95%CI: 44.5-64.6%) in men). Lifetime risk declines with age (to 67.3% (95%CI: 57.4-77.1%) in women aged 40 to 42.5% (95%CI: 33.8-52.9%) at age 60). Conclusion: In metropolitan India, 1 in 2 underweight/normal weight, and 9 in 10 obese, people aged 20 are projected to develop diabetes during their life. This portends a serious epidemic and indicates what may develop elsewhere in India in the absence of urgent action. Table. Lifetime risk of diabetes by sex, age and BMI Disclosure S. Luhar: None. S. Kinra: None. V. Mohan: None. D. Kondal: None. R. Anjana: None. S.A. Patel: None. M.K. Ali: Research Support; Self; Merck & Co., Inc. D. Prabhakaran: None. N. Tandon: None. K. Narayan: None. Funding Economic and Social Research Council (ES/J500021/1)

Authors developed an interactive map visualizing county level availability of Diabetes Self-Management Education and Support (DSMES) site or program locations. Location data were overlaid with diabetes prevalence data and socio-economic status data. Over 4200 locations were geocoded including 3471 sites (American Diabetes Association recognized) and 802 program locations (American Association of Diabetes Educators accredited). Counties were characterized as having or not having DSMES class locations. Counties were additionally characterized based on tertiles of diabetes prevalence and socio-economic disadvantage (using the Area Deprivation Index (ADI) score). This analysis showed that 48 percent of U.S. counties (1515/3142 counties and county equivalents) had a DSMES program with at least one site or program location. Approximately 39% of counties in the highest diabetes prevalence tertile (greatest diabetes burden) had DSMES programs compared to 53% in the middle tertile of diabetes prevalence and 54% in the lowest tertile of diabetes prevalence. Approximately 26% of counties in the tertile with highest ADI scores (representing greatest socio-economic disadvantage) had DSMES programs compared to 48% in the middle tertile and 69% percent in the tertile with the lowest ADI scores (least socio-economic disadvantage). This interactive map is a tool that can be used at a national level and by state and local health departments and others to prioritize establishment and expansion of programs in underserved counties with higher rates of diabetes prevalence and greater socio-economic disadvantage. Disclosure B. Jayapaul-Philip: None. S. Dai: None. E. Woghiren: None. G.E. Rutledge: None.

Populations in economic transition have a high diabetes prevalence even at lower body mass indices (BMI). We compared diabetes incidence in South Asians ≥45 years in urban India (Delhi and Chennai) and Pakistan (Karachi) with blacks and whites in U.S. Prospective analyses were done using data from the Center for Cardiometabolic Risk Reduction in South Asia Study (CARRS, South Asians, n=3,136), and the Atherosclerosis Risk in Communities Study (ARIC, blacks, n=3,059; whites, n=9,924). We defined diabetes as fasting glucose ≥126 mg/dl, HbA1c ≥ 6.5%, or medication use. South Asians were less obese than blacks and whites (BMI, kg/m2: 24.9 vs. 28.2 vs. 26.0; Waist, cms 87.5 vs.97.5 vs. 95.2). Age-standardized diabetes incidence (cases/1000 person-years, 95% CI) in South Asian men was similar to blacks and 1.6 times than whites (26.0, 22.2-29.8 vs. 26.2, 22.7-29.7 vs. 16.1, 14.8-17.4); and in South Asian women slightly higher than blacks and 3 times than whites (31.9, 27.5-36.2 vs. 28.6, 25.7-31.6 vs. 11.3, 10.2-12.3). In those with BMI Adults in urban India/Pakistan have higher diabetes incidence than U.S. whites. Even nonobese South Asian adults have markedly higher risk of diabetes compared to whites. Disclosure K. Narayan: None. D. Kondal: None. N.R. Daya: None. S.A. Patel: None. M. Deepa: None. R. Anjana: None. L.R. Staimez: None. U. Gujral: None. S. Kobes: None. R. Shivashankar: None. M.K. Ali: None. M. Kadir: None. D. Prabhakaran: None. R.L. Hanson: None. V. Mohan: None. E. Selvin: None. N. Tandon: None. Funding National Heart, Lung, and Blood Institute; Department of Health and Human Services (HHSN268200900026C); United Health Group, Minneapolis, MN; National Institute of Diabetes and Digestive and Kidney Diseases (P30DK111024)

Diabetes mellitus (DM) is the 7th leading cause of death in the U.S. in 2015. According to Center of Disease Control (CDC), prevalence of DM has increased in recent years affecting 9.4% of U.S. population and 12.5% of New York City (NYC) population in 2015. Increased DM prevalence is a worldwide trend associated with increasing age of the population. Aim: To determine prevalence and severity of diabetes/prediabetics in NYC Chinese population in comparison with those of non-Chinese patients. Methods: We obtained and analyze HMO A1C data of 276,854 Chinese (age 18-98 with median of 51 years, 55% female, 45% male, representing the largest population studies) and non-Chinese sampling population in NYC from January 2015 to October 2018 using BioReference Lab. Patients (57,612 Chinese, 49,548 non-Chinese) classified as prediabetics with A1C between 5.7-6.4% upon initial A1C test had follow-up test after management according to CDC Prediabetic management protocol. Results: Of 197,487 Chinese who had A1C test indicated prevalence of 18.5% diabetes, 45.6% of prediabetic and 35.6% nondiabetics. The male group had a higher incidence of diabetes than females (21% vs. 16.5%). Among 1,341,054 non-Chinese clients, there were 51.6% nondiabetics, 31.4% prediabetics and 18.0% diabetes. The severity of diabetes with respect to A1C data in 36,596 Chinese were 6.5-7.4, 57%; 7.5-9.0, 27%; >9.0, 16% whereas in 241,180 non-Chinese they were 44%, 29% and 28%, respectively. Among Chinese prediabetics, 30% reversed to nondiabetics and 5.7% progressed to diabetic categories, whereas among non-Chinese prediabetics, 25.8% reverse to nondiabetics and 11.1% progressed to diabetics during the study time. Conclusion: Prevalence of diabetes in NYC population is higher than that found in 2015. More severe diabetes with higher A1C levels was found in non-Chinese than Chinese patients. The NYC Chinese had higher incidence of prediabetics but had better management outcome than non-Chinese community. Disclosure G. Liu: None. J.D. George: None. J.W. Sharp: None. A. Hu: None. T. Chang: None.

Diabetes prevalence in older people is increasing and older people have an increased risk of hypoglycemia. Results are shown here of subgroup analyses in adults aged ≥65 years (pre-planned) and ≥70 years (post hoc) of the BRIGHT study in insulin-naïve adults with T2DM. Between treatment groups, baseline characteristics were similar; expected differences were observed between age subgroups: diabetes duration was longer, renal function was worse and BMI slightly lower in the ≥65 and ≥70 years groups vs. the overall population (Table). HbA1c reductions from baseline to week 24 were similar between treatment groups in the overall and ≥65 years populations. Heterogeneity of treatment effect was seen across the Gla-300 provided similar HbA1c reduction vs. IDeg-100 in older people with T2DM, and greater reductions vs. IDeg-100 in those aged ≥70 years with no increased hypoglycemia risk in this frail population. Disclosure B. Charbonnel: Consultant; Self; AstraZeneca, Merck Sharp & Dohme Corp., Novo Nordisk A/S, Sanofi, Servier. Speaker’s Bureau; Self; AstraZeneca, Eli Lilly and Company, Merck Sharp & Dohme Corp., Novo Nordisk A/S, Sanofi, Takeda Pharmaceutical Company Limited. V.R. Aroda: Consultant; Self; ADOCIA, AstraZeneca, Becton, Dickinson and Company, Novo Nordisk Inc., Sanofi, Zafgen, Inc. Employee; Spouse/Partner; Merck & Co., Inc. Research Support; Self; AstraZeneca, Calibra Medical, Eisai Inc., Janssen Research & Development, Novo Nordisk Inc., Sanofi, Theracos, Inc. J. Westerbacka: Employee; Self; Sanofi. Stock/Shareholder; Self; Sanofi. F. Lauand: Employee; Self; Sanofi. E. Boelle: Employee; Self; Sanofi Research & Development. A.Y. Cheng: Advisory Panel; Self; Abbott, AstraZeneca, Boehringer Ingelheim International GmbH, Eli Lilly and Company, HLS Therapeutics, Janssen Pharmaceuticals, Inc., Merck & Co., Inc., Novo Nordisk A/S, Sanofi. Research Support; Self; Boehringer Ingelheim International GmbH, Sanofi. Speaker’s Bureau; Self; Abbott, AstraZeneca, Boehringer Ingelheim International GmbH, Eli Lilly and Company, Janssen Pharmaceuticals, Inc., mdBriefCase Group Inc., Merck & Co., Inc., Novo Nordisk A/S, Sanofi. J. Rosenstock: Research Support; Self; AstraZeneca, Bristol-Myers Squibb Company, Genentech, Inc., GlaxoSmithKline plc., Lexicon Pharmaceuticals, Inc., Melior Pharmaceuticals, Inc., Bukwang Pharm. Co., Ltd., Merck & Co., Inc., Oramed Pharmaceuticals, PegBio Co., Ltd., Pfizer Inc. Other Relationship; Self; Boehringer Ingelheim International GmbH, Eli Lilly and Company, Intarcia Therapeutics, Inc., Janssen Pharmaceuticals, Inc., Novo Nordisk Inc., Sanofi. G.B. Bolli: Advisory Panel; Self; Sanofi. Research Support; Self; A. Menarini Diagnostics, Medtronic. Funding Sanofi (NCT02738151)

Objective: Diabetes screening and management guidelines have changed over the last 15 years. We compared diabetes prevalence, management, metabolic control and complications in 2 surveys, conducted 15 years apart. Methods: Repeat cross-sectional studies 2001-2003 (XRDS1) and 2016-2018 (XRDS2) across 4 rural Australian towns with previously low access to primary care, involving a household survey followed by biomedical assessments (including OGTTs) among randomly selected individuals. Results: Overall, household survey/clinic responses were >60% involving 4464/1042 and 2315/748 residents in XRDS1 and 2 respectively. The prevalence of known diabetes increased overall (5.4% to 10.4% p Conclusions: Over 15 years, diabetes prevalence increased, CVD risk management improved but complications were unchanged or increased. Further work is required to understand the extent to which changing survival, migration, access to care and ageing have contributed to these patterns. Disclosure D. Simmons: Speaker's Bureau; Self; Sanofi-Aventis. Other Relationship; Self; Medtronic. K. Glenister: None. D.J. Magliano: None. L. Bourke: None. Funding National Health and Medicine Research Council of Australia

Introduction: Increase in diabetes prevalence in the U.S. is likely due to demographic and dietary changes, but epidemiological studies do not show large enough changes in either to explain the magnitude of diabetes increase. We used a validated diabetes model to decompose the impact of diet on diabetes prevalence from 2001-2014, while considering demographics data and new diabetes diagnosis criteria. Methods: We used a published model of diabetes that considers macronutrient intake, demographics and individual physiology to simulate a virtual population matched to U.S. demographics in 2001 and advanced it annually by updating demographic distribution (matched to U.S. Census) and diet (estimated by training the model to CDC-estimated diabetes prevalence from ‘01-’06). Prevalence was forecast for ‘07-’14 assuming the estimated diet trend continued linearly. Results: Predicted prevalence was within 0.24% (absolute error) of CDC estimates between ‘07-‘14. Diabetes rates were largely explained by increase in carbohydrate (5%) and decrease in fat intake (11%) between ‘01-’14. The apparent plateau in diagnosed diabetes after 2010 (Figure) is inconsistent with model prediction unless both high HbA1c and glucose are used to diagnose diabetes, suggesting possible misuse of the ‘10 criteria in clinical practice. Disclosure S. Mohanty: None. C.M. Chelini: Employee; Self; Pwc. P. D'Alessandro: None. G. Dwivedi: None.

Background: Our objective was to assess trends on incidence and prevalence rates of pharmacologically-treated diabetes in France between 2010 and 2016. Methods: Diabetes cases were identified in the National Health Data System (SNDS), covering the whole of France, i.e., 66 million people, through a validated algorithm based on antidiabetic drugs reimbursements. Incident cases were identified as new diabetes cases in people free of diabetes during the 2 previous years. For a given year, mean number of residents in France was used to calculate prevalence and incidence rates. Finally, Poisson models were applied to assess possible effects of age, sex and geographical region on time trends. Results: Prevalence slightly increased from 4.3% in 2010 to 4.9% in 2016, with a significant increasing annual time trend for the period of 1.9% (95% CI: 1.7%-2.1%). After adjustment, the increasing time trend was more important among people under 45 and people over 75 years. In 2012, incidence was 4.2 while in 2016 it was 3.9 cases per 1000 person-years. A non-significant decreasing annual time trend was observed for the period 2012-2016 (-1.3%, 95% CI -2.8%, 0.2%). This decreasing time trend was stronger in adults over 45 years and among women, after adjustment. Conclusion: Between 2010 and 2016, the prevalence of diabetes increased while incidence of diabetes remained stable. Further efforts on diabetes prevention are required. Disclosure S. Fuentes: None. S. Fosse-Edorh: None. P. Bernillon: None. C. Bonaldi: Employee; Spouse/Partner; Novartis France. E. Cosson: Board Member; Self; Abbott, Boehringer Ingelheim Pharmaceuticals, Inc., LifeScan, Inc., Lilly Diabetes, Medtronic, Merck Sharp & Dohme Corp., Novartis France, Novo Nordisk A/S, Roche Diagnostics France, Sanofi. Research Support; Self; Air Liquide, Lilly Diabetes, Novo Nordisk A/S, Roche Diagnostics France, Roche Foundation, Sanofi. Funding Santé Public France

Republic of Macedonia (RoM) is estimated to have third highest diabetes prevalence in Europe. National eHealth System (NeHS) was introduced since 2015, enabling physicians to optionally record metabolic parameters. Our aim was to evaluate national metabolic control in insulin-treated pts, based on data derived from NeHS. NeHS was searched for all insulin-treated pts, as of 01-May-2017, with data in their Electronic Healthcare Records (EHR) for any of following parameters: HbA1c, BMI, Total Cholesterol (TC), LDL Cholesterol (LDL), Triglycerides (TG), Systolic Blood Pressure (SBP) and Diastolic Blood Pressure (DBP). Mean values were calculated and number of pts identified to pre-specified thresholds: HbA1c - 7% and 8%; BMI - 25 kg/m2 and 30 kg/m2; TC - 5 mmol/l; LDL - 2.6 mmol/l; TG - 1.7 mmol/l; SBP - 140 mmHg and DBP - 90 mmHg. Total number of insulin-treated pts in RoM was 37,011 out of 84,568 diagnosed pts (43.8%). Of those, 6.204 (16.8%), mean age 63.8 ± 11.7 years, 54.7% women, were identified as having data in EHR for any metabolic parameter. Mean HbA1c was 7.8 ± 1.8% (n=6,204); 2,346 pts (37.8%) ≤ 7%, 1,607 (25.9%) >7% and ≤ 8%; and 2,251 (36.3%) >8%. Mean BMI was 30.4 ± 5.2 kg/m2 (n=1,920); 273 pts (14.2%) ≤ 25 kg/m2, 739 (38.5%) >25 and ≤ 30 kg/m2, and 908 (47.3%) >30 kg/m2. Mean TC was 5.3 ± 1.4 mmol/l (n=1,838); 988 pts (53.8%) >5 mmol/l. Mean LDL was 3.2 ± 1.1 mmol/l (n=433); 290 pts (67.0%) >2.6 mmol/l. Mean TG was 2.3 ± 1.6 mmol/l (n=1,588); 938 pts (59.1%) >1.7 mmol/l. Mean SBP was 134.5 ± 17.3 mmHg (n=1,055); 232 pts (22.0%) >140 mmHg. Mean DBP was 81.1 ± 8.5 mmHg (n=1,055); 65 pts (6.2%) >90 mmHg. These are the first metabolic control results in insulin-treated pts from RoM, derived from NeHS. There is a need for improvement of glycemic control (36.3% of pts with HbA1c >8%), weight and lipid management. Since 16.8% of pts have any metabolic data in EHR, change from optional to mandatory recording is necessary to improve individual and national metabolic control. Disclosure I. Smokovski: Advisory Panel; Self; Sanofi. T. Milenkovic: None.

Aims and Objectives: Diabetes is a serious and increasing global epidemic, and accurate estimates are essential for more efficient allocation of resources. The International Diabetes Federation (IDF) North America and Caribbean Region (NAC) consist of the U.S.A, Mexico and Canada, as well as 25 Caribbean countries and territories. The new edition of the International Diabetes Federation (IDF) Atlas (IDF, 2017) provides estimates of the numbers of people (18-99 years) living with diabetes in the IDF North America and Caribbean (NAC) Region. Methodology: Estimates for diabetes in adults were taken from 24 data sources in the IDF NAC region, representing 14 out of 28 countries. Selected characteristics of these data sources are as follows: Barbados, Mexico, Suriname, Trinidad and Tobago and the U.S. had studies conducted within the last five years; Belize, Haiti, Mexico and the U.S. Virgin Islands had studies that performed oral glucose tolerance tests. Prevalence rates for other countries may be underestimated due to limited data sources on oral glucose tolerance test. The details of the methodology were described in IDF Diabetes Atlas 8th Ed. Results and Conclusion: Approximately 50.1 (41.8-56.1) million people or 13.1% (10.9-14.6%) of adults aged 18-99, are living with diabetes in the IDF NAC region in 2017. The age-adjusted prevalence is 10.8%, which is highest among all seven IDF regions. If the trend continues, the number of people with diabetes is projected to reach 73.4 (58.8-83.4) million in 2045. About 37.6% of those people living with diabetes in this region are estimated to be undiagnosed in 2017. The number of deaths attributed to diabetes from age 20 to 99 years is 378,720 in 2017. The total healthcare expenditure related to diabetes in NAC region is USD 439,858 million in 2017 and will reach USD 508,315 million by 2045. Diabetes exerts a heavy burden in this region. Therefore effective diabetes prevention and management programs should be implemented in order to control diabetes prevalence. Disclosure Y. Huang: None. S. Karuranga: None. E.J. Boyko: None. B. Malanda: None.

The combination of high diabetes prevalence, declining mortality trends and high chronic kidney disease (CKD) prevalence has raised concerns about the future burden of CKD in the American Indian population. We modeled the future prevalence of diabetes and CKD in a Southwestern American Indian population well-characterized for type 2 diabetes using data from research examinations conducted on 4,476 participants between 1965 and 2007. Diabetes was assessed by 2-hour glucose tolerance testing, CKD was defined by estimated GFR (eGFR) In conclusion, the model provides assessment of the future burden of diabetes and diabetes-related CKD, projecting a marked increase in the proportion of people with diabetes who will have CKD over the next two decades. Disclosure M.E. Pavkov: None. A. Hoyer: None. E.W. Gregg: None. R. Brinks: None. P.H. Bennett: None.

As diabetes prevalence rises in Africa, tests to detect abnormal glucose tolerance (abnl-GT) must be optimized. Recent reports suggest A1C performs well as a diagnostic test for abnl-GT in the OB. In contrast, glycated albumin (GA) contributes to the detection of abnl-GT in the nonobese. To determine if BMI affects sensitivity of A1C and GA in Africans, we evaluated the sensitivity of A1C and GA separately and combined to detect abnl-GT in 88 OB blacks (age 42±10y, BMI 33.4±2.9, range 30.0-42.4) and 232 nonobese blacks (age 38±10y, BMI 25.5±2.6, range 18.2-29.8) who were born in Africa and live in the U.S.A. Abnl-GT was determined by glucose criteria for the OGTT. Thresholds for A1C and GA were defined by the cut-off at their upper quartile (A1C≥5.7%; GA ≥13.97%). Prevalence of abnl-GT in OB and nonobese Africans were: 43% v 35%, P=0.20, resp. In the OB, sensitivities of A1C, GA and the combined tests were: 67%, 29% and 79% resp (Figure A). In the OB, sensitivity of A1C + GA was similar to A1C alone (P=0.13). For the nonobese, sensitivities of A1C, GA and the combined tests were: 35%, 38% and 62% resp (Figure B). Sensitivity increased when the tests were combined because abnl-GT was detected in 22 nonobese Africans not identified by A1C. Patterns were similar by sex and region of Africa. Data from Africans living in the U.S.A suggest detection of abnl-GT in the nonobese is improved by combining A1C with GA. Disclosure J. Kabakambira: None. S.M. Briker: None. R.L. Baker: None. C. DuBose: None. L. Mabundo: None. S.T. Chung: None. D.B. Sacks: Other Relationship; Self; Sebia, Trinity Biotech. A.E. Sumner: None.

The nearly 3,000 local health departments (LHDs) nationwide are the frontline of public health, and are positioned to implement evidence-based interventions (EBIs) for diabetes control. This study used a national online survey to determine the prevalence and correlates of four CDC Community Guide recommended EBIs in LHDs. Among 240 LHDs, each EBI was delivered directly or with key partners by ≥60% (Figure 1); 96 LHDs (40%) offered all four EBIs, and 15 (7%) offered none. Diabetes prevalence in the state was positively associated with offering the Diabetes Prevention Program (OR=1.28 (95% CI: 1.02-1.62)), diabetes self-management education (OR=1.32 (95% CI: 1.04-1.67)), and identifying patients and determining treatment (OR=1.27 (95% CI: 1.05-1.54)) (Table). Implementation of these EBIs by more LHDs can help control diabetes.Table. Selected Associations between Local Health Department (LHD) Characteristics and Delivering (directly and/or with key partners) Diabetes Evidence-Based Interventions (EBIs) (odds ratio [OR] (95% confidence interval [95% CI]))DSME1DPP2Identify3CHWs4LHD and respondent characteristicsJurisdiction Population categories ( Disclosure R.G. Tabak: Other Relationship; Self; Envolve PeopleCare/Centene Corporation. R.G. Parks: None. P.M. Allen: None. R.R. Jacob: None. S. Mazzucca: None. K. Stamatakis: None. M. Chin: Research Support; Self; Merck Foundation. R. Brownson: None.

Background: Little is known about the magnitude of excess mortality due to diabetes in low and middle income countries. Brazilian data to that end have been limited by reliance on death certificates and their subjective assignment of diabetes as a cause of death. We provide estimates of excess mortality due to diabetes by combining mortality risks from the ELSA-Brasil cohort with nationally representative surveys and databases. Methods: We calculated the standardized mortality rate ratio for those with (vs. without) self-reported diabetes in ELSA-Brasil by Poisson regression adjusting for age, sex, race and education. We obtained similarly adjusted, self-reported diabetes prevalence from the 2013 National Health Survey. Mortality and population projections were from national statistics. We combined these data to model the excess deaths due to diabetes. Results: In 2013, among adults with diabetes aged 35 to 74 (table), 73,092 deaths (13% of total) could have been avoided if the mortality rate in people with diabetes were the same as in those without diabetes. The percent ranged from 5.9% in persons aged 35-49 to more than 16% in those 60 or older. In the national mortality system, only 31,117 deaths had diabetes as a main cause. Age GroupStandardized Mortality Rate Ratio Population NTotal Deaths NDeaths among People with Diabetes NExcess Deaths Associated with Self-Reported Diabetes With diabetesWithout diabetes Assuming diabetes mortality rateAssuming non-diabetes mortality rateExcess deaths among people with diabetes N% of total deaths35-492.40 (1.13 - 5.08)746,790 (3.7 %)19,538,124 (96.3%)120,20011,7224,6777,0455.950-591.94 (1.24 - 3.04)992,878 (10.2%)8,781,526 (79.8%)140,57730,86012,91417,94612.860-691.93 (1.35 - 2.74)941,815 (16.3%)4,853,378 (83.7%)182,45055,78725,15930,62816.870-741.58 (0.94 - 2.66)338,769 (20.3%)1,333,776 (79.7%)106,63735,53518,06217,47316.4Total1.91 (1.51 - 2.41)3,020,252 (8.1%)34,506,804 (92.9%)549,864133,90460,81273,09213.3 Conclusion: This calculation of excess deaths related to known diabetes in Brazil suggests a major diabetes mortality burden, which is considerably underestimated by calculations based only on death certificates. Disclosure P. Bracco: None. E.W. Gregg: None. D.B. Rolka: None. M.I. Schmidt: Research Support; Self; Eli Lilly and Company. Research Support; Spouse/Partner; Eli Lilly and Company. S. Barreto: None. P.A. Lotufo: Consultant; Self; AbbVie Inc.. E.C. Melo: None. B.B. Duncan: Research Support; Self; Eli Lilly and Company. Research Support; Spouse/Partner; Eli Lilly and Company.

The Dominican Republic (DR) is a Caribbean country of 10 million with estimated diabetes prevalence at ∼11%. While the public health system has a diabetes program, there are few reliable surveillance data to portray diabetes outcomes in the DR. The HbA1c is not a widely available test for metabolic control in the DR. In 2010, a non-profit group, Chronic Care International (CCI), began a diabetes and hypertension program with 2 clinics in poor, rural DR. CCI developed a model of quality improvement for care and self-management with its Dominican medical team, including doctors, nurses and community health workers. We present process and outcome data collected over 6 years utilizing electronic medical records. In addition to quality improvement analyses, we fit a random-effects linear model to the data to assess trends over time. By May 2017, subjects (N=1,031) were adults: mean age 60.1 years, 45% men, 74% with T2DM. Of those with diabetes (n=758), 76% had an HbA1c Disclosure E.A. Walker: None. H.J. Dethlefs: None. R.A. Dowd: None. C. Schechter: None. C. Filipi: None.

OBJECTIVE Glycated hemoglobin (HbA1c), used to monitor and diagnose diabetes, is influenced by average glycemia over a 2- to 3-month period. Genetic factors affecting expression, turnover, and abnormal glycation of hemoglobin could also be associated with increased levels of HbA1c. We aimed to identify such genetic factors and investigate the extent to which they influence diabetes classification based on HbA1c levels. RESEARCH DESIGN AND METHODS We studied associations with HbA1c in up to 46,368 nondiabetic adults of European descent from 23 genome-wide association studies (GWAS) and 8 cohorts with de novo genotyped single nucleotide polymorphisms (SNPs). We combined studies using inverse-variance meta-analysis and tested mediation by glycemia using conditional analyses. We estimated the global effect of HbA1c loci using a multilocus risk score, and used net reclassification to estimate genetic effects on diabetes screening. RESULTS Ten loci reached genome-wide significant association with HbA1c, including six new loci near FN3K (lead SNP/P value, rs1046896/P = 1.6 × 10−26), HFE (rs1800562/P = 2.6 × 10−20), TMPRSS6 (rs855791/P = 2.7 × 10−14), ANK1 (rs4737009/P = 6.1 × 10−12), SPTA1 (rs2779116/P = 2.8 × 10−9) and ATP11A/TUBGCP3 (rs7998202/P = 5.2 × 10−9), and four known HbA1c loci: HK1 (rs16926246/P = 3.1 × 10−54), MTNR1B (rs1387153/P = 4.0 × 10−11), GCK (rs1799884/P = 1.5 × 10−20) and G6PC2/ABCB11 (rs552976/P = 8.2 × 10−18). We show that associations with HbA1c are partly a function of hyperglycemia associated with 3 of the 10 loci (GCK, G6PC2 and MTNR1B). The seven nonglycemic loci accounted for a 0.19 (% HbA1c) difference between the extreme 10% tails of the risk score, and would reclassify ∼2% of a general white population screened for diabetes with HbA1c. CONCLUSIONS GWAS identified 10 genetic loci reproducibly associated with HbA1c. Six are novel and seven map to loci where rarer variants cause hereditary anemias and iron storage disorders. Common variants at these loci likely influence HbA1c levels via erythrocyte biology, and confer a small but detectable reclassification of diabetes diagnosis by HbA1c.

During study of diabetes prevalence in six rural Tanzanian communities, a repeat oral glucose tolerance test (OGTT) was carried out in 514 subjects ≥ 15 yr of age within 1 wk of an initial 75-g OGTT. In 498 subjects, blood glucose was measured 2 h after the glucose load on both occasions, and in 175 subjects, fasting blood glucose measurement was also repeated. Of the 498 subjects, 245 had normal glucose tolerance in the first test and were selected at random for further testing; 223 subjects had impaired glucose tolerance (IGT), and 30 had diabetic values. Diabetes and IGT were diagnosed on the basis of the 2-h blood glucose values. In the second test, 241 (98.4%) of the 245 subjects with normal tolerance continued in this category and 4 (1.6%) showed IGT. Of the 223 with IGT in the first test, 171 (76.2%) reverted to normal on the second test, 7 (3.1%) had diabetic values, and 45 (20.2%) persisted with IGT. Of the 30 subjects diagnosed as diabetic in the first test, 8 (26.7%) remained with diabetic values, 11 (36.7%) had IGT, and 11 (36.7%) were normal. Based on the second test, the population-prevalence rates of diabetes and IGT would have been 0.5 and 3.3% vs. 1 and 7.6% based on the first test. There was a significant downward trend in the mean 2-h blood glucose values in all three diagnostic groups. Regression toward the mean could not account for the downward shift in blood glucose values observed on retesting. It is postulated that a phenomenon occurred similar to the acclimatizing reflex observed in measurement of blood pressure. Therefore, single 2-h postload values may be inadequate for the accurate assessment of IGT and diabetes-prevalence rates within communities, particularly those unused to blood sampling. Because of the instability of the category of IGT, it may be advisable to discontinue use of this term or to restrict the term to those who show impaired tolerance on more than one occasion.

Rural-urban comparisons of diabetes prevalence were made in the Polynesian population of Western Samoa. The prevalence of diabetes in the urban population was almost three times that in the rural (10.1% versus 3.6%). While the urban male and female subjects were significantly more obese than their rural counterparts, the difference in prevalence rate could not be wholly explained on this basis. Diabetes prevalence was still approximately double in urban subjects when we compared the rural and urban populations after removing the differences in obesity and age. The results suggest that, apart from age and obesity, other factors, e.g., differences in diet, physical activity, or stress (or a combination of these), may participate in the rural to urban difference in diabetes prevalence.

Venous plasma glucose concentrations before and two hours after a 75 g carbohydrate load were determined on 597 Nauruans, a Micronesian population with one of the highest diabetes prevalence rates yet recorded in the literature. A marked increase in both the mean fasting and postload plasma glucose concentrations with age for both sexes was noted for the total population. In this population the frequency distributions of fasting and two-hour plasma glucose show bimodality, separating the population into normal and hyperglycemic subpopulations. Analysis of mean fasting and two-hour glucose concentrations of subjects in the normal and hyperglycemic components shows that the increase in mean glucose with age in the Nauruan population is mainly the result of an increasing proportion of subjects falling into the hyperglycemic (diabetic) component with increasing age. Only small changes in mean glucose concentrations with age were noted in subjects in the normal component. Percentile distributions of both fasting and postload glucose concentrations indicate that there is a dramatic age-related increase in glucose concentrations in the higher percentiles (70th and 90th) in both males and females. This finding is consistent with the hypothesis of two subpopulations in which the glucose concentrations increase significantly with age in one (higher percentiles) and remain relatively constant in the other (lower percentiles). The data indicate that, while there is a rise in glucose concentrations with increasing age, it does not necessarily represent diabetes. The major contribution to this age-related rise is due to the increasing number of diabetics with age, and the diabetics can be separated from the rest of the population on the basis of percentile analysis and bimodality of glucose distributions.

While frequency distributions of glucose concentrations in Caucasian populations are unimodal, bimodality has been described in the Pima Indians, a population with an extremely high prevalence of diabetes. Venous plasma glucose concentrations at fasting and after a 75-gm. oral glucose load were determined in 596 Nauruans, a Micronesian population with a diabetes prevalence of the same order as the Pima Indians. In both sexes and in subjects 10 to 19 years, the frequency distributions of the logarithms of the fasting and two-hour glucose values were clearly unimodal. In most sex and age groups of 20 years and older, the frequency distributions of fasting and two-hour glucose values were bimodal and consistent with a model of two overlapping Gaussian distributions. This population is characterized by marked obesity. However, there was no significant difference in the degree of obesity between subjects in the first and second curves of the bimodal distribution. This makes it unlikely that the bimodality is a consequence of the marked obesity seen in both the Pima and Nauru populations. The data show that among Nauruans, as with the Pimas, the frequency distribution of glucose concentrations can be used to separate the population into normal and hyperglycemic groups.