Page Not Found | University of Alberta
Obesity – Canadian trends
NIH. Clinical guidelines for the identification, evaluation, and treatment of overweight and obesity in adults. National Institutes of Health: 1998.
- Health risks of overweight and obesity
- Morbidity – higher morbidity associated with overweight and obesity for HTN, t2dm, CAD, stroke, gallbladder disease, osteoarthritis, sleep apnea and respiratory problems, and some types of cancer (endometrial, breast, prostate, and colon).
- Obesity associated with complications of pregnancy, menstrual irregularities, hirsutism, stress incontinence, and depression.
- Diabetes – pages 14-15
o Prospective studies in Norway, US, Sweden, and Isreal have shown increased risk of diabetes as weight increases.
o Nurses’ Health Study – risk of developing type 2 diabetes increases with BMI from a BMI as low as 22.
o The RR of diabetes increases by approximately 25% for each additional unit of BMI of 22 kg/m2 (Coditz GA, Willett WC, Rotnitzky A, MansonJE. Weight gain as a risk factor for clinical diabetes melligus in women. Ann Intern Med, 122: 1995.).
o 27% of new cases of diabetes was attributable to weight gain in adulthood of 5kg or more (Ford ES, Williamson DF, Liu S. Weight change and diabetes incidence: Findings from a national cohort of US adults. Am J Epidemiol, 146: 1997.).
o Evidence from both cross-sectional and longitudinal studies showing abdominal obesity is a major risk factor for type 2 diabetes.
Katzmarzyk PT, Ardern CI. Overweight and obesity mortality trends in Canada, 1985-2000. Canadian Journal of Public Health, 95(1): 2004.
- Calculated population-attributable fraction of all-cause deaths due to obesity.
- Used RRs calculated from the Canada Fitness Survey with 13-year follow-up (Katzmarzyk PT, Craig CL, Bouchard C. Underweight, overweight and obesity: Relatinoships with mortality in the 13-year follow-up of the Canada Fitness Survey. J Clin Epidemiol, 54: 2001.). RRs adjusted for age, gender, smoking, and alcohol consumption.
o Normal weight – Reference
o Overweight – RR = 1.16 [0.96, 1.39]
o Obese Class I – RR = 1.25 [0.96, 1.65]
o Obese Classes I and II – 2.96 [1.39, 6.29]
- Used prevalence proportions of overweight, class I obese, and class II obese from 6 national surveys from 1984-2000 (Health Promotion Surveys, National Population Health Surveys, and CCHS 1.1).
- PAR increased from 5.1% in 1985 to 9.3% in 2000.
- Multiplied PAR to total deaths in subjects 20-64 years of age only. Age-limits chosen to focus on premature mortality in non-elderly adults.
- Number of deaths from 2514 to 4321. With linear interpolation for non-survey years, Total accumulated overweight and obesity related deaths was 57181.
- Province-specific relative increase in number of obesity-related deaths provided in map form.
- Obesity and death not directly linked at the individual level.
- PAR and obesity-associated deaths should be considered lower limits, since height is often over represented and weight under-represented on self-report surveys such as those used here.
- Almost 1 in 10 premature deaths among those aged 20 to 64 years can be attributed to overweight and obesity in 2000.
Katzmarzyk PT, Mason C. Prevalence of class I, II and III obesity in Canada. CMAJ, 174(2): 2006.
- National population health surveys from 1985-2003, including the Health Promotion Surveys, National Population Health Surveys, and the CCHS 1.1 and 2.1. Cross-sectional prevalence study.
- Age-adjusted weighted prevalence’s of overweight and obesity calculated – adjusted to what reference population?
- Class III obesity – 0.4% in 1990 to 1.3% in 2003 = 225% relative excess. However, absolute prevalence of class III obesity lower than that of the US.
- Any obesity – 9.6% in 1990 to 15.7% in 2003 = 64% relative excess.
- Overweight (BMI 25-29) – 31.5% in 1990 to 33.9% in 2003 = 8% relative excess.
- Limitations – over-estimation of height, under-estimation of weight in self-reported data. E.g.: CCHS 2004 direct measurement data indicated prevalence of class II and II obesity = 5.1% and 2.7%, compared to 3.0% and 1.3% in the self-reported 2003 survey data.
The Diabetes Prevention Program and diabetes prevention
Diabetes Prevention Program Group. The Diabetes prevention program: Design and methods for a clinical trial in the prevention of type 2 diabetes. Diabetes Care, 22(4): 1999.
- Study in Malmo, Sweden – two groups of middle-aged men with IGT. Those treated with an intensive diet and exercise program for 5 years experienced a rate of diabetes that was only half that of a non-randomized comparison group. Feasibility of the behavioural component (Eriksson K-F, Lindgard F. Diabetologia, 34: 1991).
- RCT in Da-Qing, China. Interventions involving diet, exercise, or both showed a 6-year lower cumulative incidence of diabetes than the non-intervention control group. However, no between intervention differences, and potential imbalances between groups (Pan X, Li G, Hu Y, Wang J, Yang W, An Z, et al. Diabetes Care, 20: 1997).
- RCT with 3 arms. Patient, care provider, and end-point blinded.
- Eligibility – population stratified to be at least half women, 20% elderly, and half composed of ethnic minorities. Main criteria: IGT based on 75g OGTT. No prior diagnosis of diabetes, FPG < 7.0mM, 2-h PG > 7.8mM but < 11.1mM. Additionally, FPG between 95-125mg/dl, with lower limits for American Indians. BMI >/= 24 kg/m2, lower for Asian Americans.
- Eligibility criteria designed to increase chance of progression to type 2 diabetes. Exclusion criteria designed to ensure ability to tolerate study interventions, and freedom from study endpoints.
- Overall, synthesis of previous literature suggests the conversion of IGT to diabetes occurs in 3 per 100 person-years for Caucasians and Mexican-Americans, with much higher rates (e.g.: between 10 and 11 per 100 person years – Asian Indians and Pima Indians) for certain minorities. Overall rate 5.8 per 100 per-years in six cohorts. Higher rate detected for elevated FPG (7.7/100) (Edelstein SL, Knowler WC, Bain RP, Andrew R, Barrett-Connor EL, Dowse GK, et al. Diabetes, 46: 1997).
- Primary outcome: Diabetes, using ADA 1997 definition. FPG measurement. Confirmation with OGTT. Upon outcome, subjects continued to be followed on protocol with intensified monitoring and the goal of achieving optimal glycemic control.
- Secondary outcomes: CVD profile, HbA1c, beta-cell function, insulin sensitivity, renal function, body composition, physical activity and nutrient intake, HRQL, and economics.
- Outcome ascertainment: FPG every 6 months, 75g OGTT every year, and testing as needed given symptoms. Adherence measured by pill counts, medical questionnaires, and activity / nutrition questionnaires, as well as body composition measurements.
- Interventions
o Standard: written information, single session with case manager – recommended that they lose 5-10% of initial weight by diet and exercise, increase activity to 30 min daily 5 days/week, avoid excessive alcohol, stop smoking.
o Standard + metformin.
o Intensive lifestyle: Same goals (lose and maintain weight loss of 7%). Patients undergo 16-session curriculum designed to provide training in behavior modification skills. Support (individual and group) is provided. Protocol is structured but individually flexible. Emphasis on self-esteem, empowerment, and social support. Executed by case managers with contact at least monthly throughout (more intensely at initiation).
o Initially there was a thiazolidinedione (troglitazone) group but this was halted after a fatal case of liver failure.
Adler AI, Turner RC. The Diabetes Prevention Program (Comment). Diabetes Care, 22(4): 1999.
- 50% of newly diagnosed patients with diabetes have evidence of diabetic tissue damage (UKPDS Group. UKPDS Study 6. Complications in newly diagnosed type 2 diabetic patients and their association with different clinical and biochemical risk factors. Diabetes Res, 13: 1990).
- Metformin – could mask diagnosis of diabetes by dropping glucose levels. Suggested that a wash-out occur to distinguish slowing of disease progression vs masking.
Singer F. Rethinking the Diabetes Prevention Program Clinical Trial (Comment). Diabetes Care, 22(10): 1999.
- Subjects who develop diabetes while taking 850 mg metformin twice daily can expect a lowering of their FPG by 30% or more. Thus, using FPG to detect primary outcome in the DPP may mask the development of diabetes in subjects taking metformin.
DPP Research Group. The Diabetes Prevention Program: Evaluation and Management of Diabetes (Comment). Diabetes Care, 22(10): 1999.
- Response to Singer. The difference between masking and slowing disease progression may be semantic, since disease progression is defined, for diabetes, by the diagnosis of diabetes.
- Either way, if one accepts the premise that long term exposure to hyperglycemia is the etiology of diabetes complications, then whether masked or prevented, delaying the diagnosis of diabetes should reduce the incidence of complications.
Tuomilehto J, Lindstrom J, Ericksson JG, et a;. Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. NEJM, 344: 2001.
- Finish Diabetes Prevention Program
- Found results almost identical to those of the US DPP.
- 58% lower incidence of diabetes in the intervention (diet and exercise) group vs control.
- Average weight loss was 4.2 +/- 5.1kg.
Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lcahin JM, Walker EA, Nathan DM. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. NEJM, 346(6): 2002.
- Trial methodology from design paper.
- Blinded treatment phases terminated one year early, in May 2001 – evidence of efficacy on 65% of the planned person-years of observation.
- Recruitment from 1996-1999. 3234 participants
o 1082 placebo
o 1073 metformin
o 1079 intensive lifestyle intervention
- Average follow-up 2.8 years. At study close, 92.5% of participants had attended a scheduled visit within 5 months.
- Adherence
o 7% weight loss in 24 weeks achieved by 50%. Proportion was 38% at study end.
o Physical activity goal 74% at 24 weeks, 58% at study end.
o Study medication, at least 80% adherence – in 77% and 72% of treatment and placebo group.
- Average weight loss higher in lifestyle group (5.6 kg vs 0.1kg placebo and 2.1kg metformin).
- Crude incidence of diabetes: 11.0, 7.8, and 4.8 cases per 100 person years for placebo, metformin, and lifestyle.
o Lifestyle 58% [48, 66] lower vs placebo.
o Metformin 31% [17, 43] lower vs placebo.
o Lifestyle 39% [24, 51] lower vs metformin.
o Group-sequential log-rank test (K-P rates).
- Cumulative incidence of diabetes
o Placebo – 29%.
o Metformin – 22%.
o Lifestyle – 14%.
- FPG increases over time paralleled in all groups. Differences in FPG levels established early, with lifestyle and metformin having similar FPGs, much lower than placebo. Metformin results for HbA1c much closer to being equidistant between placebo and lifestyle for most of the follow-up. No significance testing undertaken.
- Adverse events: Lifestyle – MSK higher; Metformin – GI higher. No deaths, hospitalization, or mortality differences.
- The incidence of diabetes was reduced by lifestyle intervention and metformin compared to placebo. Effects were similar in men and women of all racial and ethnic groups. The findings are applicable to the ethnically and culturally diverse US population. An estimated 10 million US individuals, according to NHANES III, resemble this study population.
Diabetes incidence and prevalence
King H, Aubert RE, Herman WH. Global Burden of Diabetes, 1995-2025. Diabetes Care, 21(9): 1998.
- Data sources
o Wide range of national field studies assessing age-, sex-, and rural/urban specific diabetes prevalence.
o Diabetes diagnosed by venous plasma glucose after 2hr 75g OGTT. Studies using only fasting blood glucose excluded.
o Countries with missing studies assumed to take values from socio-economically and racially similar neighbors.
o Urban prevalence rates known to be higher than rural rates. Missing rural or urban data imputed by assuming urban rates are 2x rural rates, but only in developing countries. Prior studies suggest no such difference in developed countries.
o Demographic estimates from the UN Population Division. Estimates for present and future urbanization patterns were also available from the UNPD.
o Adults 20+ years old only.
o Assume that the factors affecting diabetes prevalence are adequately captured by nation, sex, age, and degree of urbanization. Then within these strata, prevalence rates should be constant over time.
- Prevalence
o 1995 – 4.0%
o 2025 – 5.4% – 35% increase.
o Higher in developed countries. However, proportional increase will be greater in developing countries – 3.3% to 4.9% (48% increase) vs 6.0 to 7.6% (27% increase).
o Highest increase in prevalence will be for China and India.
- Number of patients with diabetes
o 1995 – 135 million
o 2025 – 300 million – 122% increase.
o 51 to 72 million (42% increase) in developed countries.
o 84 to 228 million (170% increase) in developing countries.
o Top 3 countries for numbers: India, China, US.
o Greatest increases seen in India, Middle Eastern Crescent, and Sub-Saharan Africa.
- Age bands affected by diabetes prevalence and numbers
o Numbers – developed countries – oldest age group largest, with greatest increase in numbers. Developing countries – middle age group largest, with greatest increase in numbers. World age structure follows developing countries.
o 1995 – world – more females with diabetes than me (73 vs 62 million) – but equal in developed countries (42 million of both).
- Comments
o Assumed constant prevalence, non-dynamic model. Data sources may be out of date or missing.
o Strength – estimates of total diabetes, including undiagnosed.
o Strength – incorporated projected patterns of urbanization. This is claimed to be a proxy for diabeto-genic lifestyle / environment, covering some of the increases in adjusted diabetes prevalence, to some extent.
Boyle JP, Honeycutt AA, Narayan KM, Hoerger TJ, Geiss LS, Chen H, Thompson TJ. Projection of diabetes burden through 2050: impact of changing demography and disease prevalence in the US. Diabetes Care, 24(11): 2001.
- Data source
o NHIS data, 1980-1998, for prevalence rates of diabetes.
o Race (black vs white), age, and sex-specific prevalence rates estimated.
o Increases in prevalence over time estimated, projected by linear regression (constant increases over time).
o US Census Bureau middle series (Reference).
- Prevalence
o 2000 – 3.99%. these figures are low, and probably represent the overall, not just the adult, population.
o 2050 – 7.21%
- Number of patients with diabetes
o 2000 – 11 million
o 2025 – almost 20 million
o 2050 - >29 million
- Age bands affected by diabetes prevalence and numbers
o The largest increases in race-specific projections occur in the oldest age groups. Men will be affected more than women (+437% vs +271%).
o Whites account for the majority of the projected people with diabetes, but the largest percentage increases occur among blacks (4.6 times as many males with diabetes in 2050 vs 2000).
- Drivers
o Numeric increase of 3.1 million from 2000 to 2010
o 0.8 million due to population growth, 1.2 million due to changing demographics, 1.1 million due to increasing prevalence rates. Demographic changes account for largest share of the increases.
- Comments
o Projection model is not dynamic (no modeling of incidence and removals).
o Prevalence rates increase, but linearly – trend may be conservative.
o No data on other minorities.
o NHIS data – diagnosed and self-reported, not total diabetes.
o Other projections too low – fail to anticipate changing racial composition and increasing prevalence rates, which may capture obesity and lifestyle changes.
o By 2050, there will be an additional 18 million people with diagnosed diabetes in the US, and 37% of this increase will be due to changes in the age, sex, and race structure of the population. 27% will be due to population growth, and 36% will be due to changes in prevalence rates.
Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of diabetes: estimates for the year 2000 and projections for 2030. Diabetes Care, 27(5): 2004
- Data sources
o See King, 1998
o Differed from King in data sets available – more recent ones in some countries.
- Prevalence
o Nice picture of diabetes prevalence by age and sex bands (Fig. 1). Obviously, much higher prevalence in elderly.
o No other prevalence figures provided.
- Numbers
o Developing and developed countries, broken down by three broad age categories >20 years (Figure 2).
o 2000 – 171 million
o 2030 – 366 million
- Age bands affected by diabetes prevalence and numbers
o As for King, 1998.
o Globally, diabetes prevalence is higher in men, but there are more women with diabetes than men due to a greater number of elderly women than men in most populations.
- Comments
o See King, 1998.
o Top 3 countries for numbers still India, China, and US.
Engelgau M, Geiss LS, Saaddine JB, Boyle JP, Benjamine SM, Gregg EW, Tierney EF, Rios-Burrows N, Mokdad AH, Ford ES, Imperatore G, Narayan V. The evolving diabetes burden in the United States. Annals of Internal Medicine, 140(11): 2004.
- Three US national surveys track diabetes prevalence
o NHIS – face-to-face interview, ever diagnosis with diabetes.
o NHANES – face-to-face interview, ever diagnosis with diabetes; additional laboratory examination of glucose levels identifies un-diagnosed diabetes.
o BRFSS – telephone interviews using state-based population samples.
- Diabetes prevanece – 2002 – 6.3% (18.2 million).
o NHIS – 65+ makes up almost 40%, prevalence in elderly ten times that of those youger than 45 years. Prevalence 2-4 times higher in minorities.
o NHIS – 4 to 8-fold increase over the last half-century in prevalence.
o BRFSS – largest increase occurred in persons aged 30-39 and 40-49 (95% and 83% higher, versus 33-42% higher in other age groups > 18 years old) from 1990-2001.
o NHANES – on third of diabetes is undiagnosed.
- Complications
o CDC diabetes surveillance – 37% of patients with diabetes aged 35 years and older reported a diagnosis of CVD in 2000.
o IHD prevalence 14-fold higher in persons 18-44 years old (2.7% vs 0.2%), 3-fold higher in those 45-64 years old (14.3% vs 4.7%), and almost 2-fold higher in those 65 or more years old (20% vs 12%). Diabetes-related RR higher for women than men, although men drive CVD mortality.
o Diabetes = leading cause of blindness (ages 20-64 years) – 12% of all new cases. Prevalence of “considerable” visual impairment in diabetes = 25%, nationally.
o Diabetes nephropathy – 40% of new cases of ESRF, US in 2000.
o Lower-limb disease – will affect an estimated 15% of persons with diabetes in their lifetime.
o NHANES – 8.1% of those with diabetes aged 40 years or older have PAD (Ankle-brachia artery BP ratio < 0.90) versus 4.0% in those without diabetes.
o Neuropathy – NHANES – Diabetes – 3-4 times the prevalence of peripheral neuropathy symptoms (30% vs 10%), insensate feet (26% vs 14%), or either (45% vs 21%).
- Mortality
o Age-adjusted mortality among adults with diabetes is approx. 2-times greater than that of those without – meta anlaysis – 10 prospective studies – RR = 1.9 for men, 2.6 for women (Lee WL, Cheung AM, Cape D, Zinman B. Diabetes Care, 23: 2000).
o ICD coding in 2000 – Diabetes = 6th leading cause of death – underlying cause in 69301 death certificates – under-diagnosed as an underlying cause of death, however.
o Deaths in persons with diabetes – 65% CVD.
o CVD mortality reductions in general population did not apply to US population cohorts from 1971-1984 and 1982-993 – non-diabetes significant men 36% women 27% declines, diabetes NS men 13% women 23% drops in CVD deaths. (See also Thomas RJ, Palumbo PJ, Melton LJ 3rd, Roger VL, Ransom J, O’Brien PC et al.. Arch Intern Med, 163: 2003.)
- Risk factors for complications
o MRFIT – large cohort – 350000, including 5000 patients with diabetes – 12 years. CVD RFs similar for those with and without diabetes – but RRs higher for those with diabetes (Stamler J, Vaccaro O, Neaton JD, Wentworth D. Diabetes Care, 16: 1993.).
o Hyperglycemia – link to CVD – less clear – association in large prospective cohorts. POVS:
▪ Finnish cohort – Fatal or non-fatal AMI – diabetes major RF, within strata of previous heart disease history.
▪ UK study – previous history of MI more important RF than diabetes.
▪ DECODE study – combined 22 prospective cohorts with 11 years follow-up – J-shaped cardiovascular disease mortality rates for the fasting glucose concentration.
- Projections
o Based on 1984-2000 NHIS and census projections, 225% increase in diabetes cases, from 12 to 39 million, from 4.4% to 9.7% (120% increase) prevalence, over 2000-2050 (Honeycutt AA, Boyle JP, Broglio KR, Thopson TJ, Goerger TJ, Geiss LS, et al.. Heatlh Care Manag Sci, 6: 2003.).
o Older affected more, Hispanic affected more. Mostly driven by demographic changes and population growth (model assumptions).
Cowie CC, Rust KF, Byrd-Holt DD, Eberhardt MS, Flegal KM, Engelgau MM, Saydah SH, Williams DE, Geiss LS, Gregg EW. Prevalence of diabetes and impaired fasting glucose in adults in the US population: National Health and Nutrition Examination Survey 1999-2002. Diabetes Care, 29(6): 2006.
- Data source
o NHANES 1999-2002
o Adults aged >20 years only.
o 10291 (83.0%) interviewed over 4 years. FPG with overnight fasting taken from random assignment of 4271 individuals.
o Comparison with NHANES III (1988-1994) data, all standardized to the US 2000 Census population using the direct method.
o T-tests used to test differences in proportions between the two surveys.
- Prevalence – all diabetes
o 1988-1994 – crude prevalence – 7.8%.
o 1999-2002 – crude prevalence – 9.3%.
o Difference significant at p = 0.007.
o Standardized 1988-1994 vs 1999-2002 – 8.2% vs 9.3%, p = 0.06.Change was more prominent (and significant) in men and in non-Hispanic blacks.
- Prevalence – diagnosed diabetes
o 1988-1994 – Diagnosed diabetes – Crude 5.1%
o 1998-2002 – Diagnosed diabetes – Crude = 6.5%, 5.6% in non-Hispanic whites, 10.0% in non-Hispanic blacks, and 6.5% in Mexican Americans. Similar for men and women.
o Age and sex-standardized prevalence – significant racial differences.
o Black – 11.0%
o Mexican Americans – 10.4%
o Whites – 5.2% (Higher prevalence in men than women, 6.0% vs 4.5%, significant for Whites but not for others.)
o 1988-94 vs 1998-2002 – Age and sex standardized prevalence rose from 5.4% to 6.5%. Increase more prominent in older ages, in men, and in non-Hispanic blacks.
- Prevalence – undiagnosed diabetes
o 1988-1994 – Undiagnosed diabetes – crude 2.8%.
o 1998-2002 - Undiagnosed diabetes – crude – 2.8%. Similar standardized prevalence for races. Higher prevalence in men than women for whites.
o Prevalence generally stable from 1988-94 vs 1998-2002 – but in Meximan Americans there was a significant increase inc rude prevalence without a change in standardized prevalence.
- Prevalence - IFG
o IFG – crude – 26.0%. Standardized prevalence in Blacks significantly lower than in Whites and Mexican Americans. Higher in men than women. Stable over time except for Blacks (decreased).
- Diagnosed versus undiagnosed
o 1999-2002 – 30.1% [25.8, 34.9] of total diabetes was undiagnosed. Proportion slightly higher in non-Hispanic whites (34%) than in Mexican Americans (22%), and in men (34%) vs women (26%).
o 1988-1993 – 34.6% [30.6, 38.9] of total diabetes was undiagnosed.
o P = 0.15.
o Standardized prevalence’s – 1988-1994 vs 1999-2002 – p = 0.16 – 34.5% vs 30.1%.
- Number of patients with diabetes
o 19.3 million diagnosed or undiagnosed.
o Comments
o Much more discussion RE age- sex- and gender- related differences in both trends over time and 1998-2002 prevalences. Extensive tables.
o Note non-significant increase in total diabetes, except in men and non-Hispanic blacks. Non-Hispanic blacks and Mexican Americans have almost twice the standardized prevalence of total diabetes of non-Hispanic whites.
Lipscombe LL, Hux JE. Trends in diabetes prevalence, incidence, and mortality in Ontario, Canada 1995-2005: A population-based study. Lancet, 369(9563): 2007.
- Increase in diabetes prevalence attributed to increased incidence, driven by increasing obesity, aging of the population, shifts in lifestyle, and improved survival – however, WHO models do not account for increasing obesity or improved survival, and use urban-rural as a proxy for shifts in lifestyle.
- Data sources
o Ontario diabetes database – diabetes patients only.
o 1991-2005.
o Diabetes washout set to 5 years. Outcome eligibility beginning 1997.
o Death data obtained.
o Rates age- and sex- standardized to the 2001 Canadian census. Diabetes prevalence and incidence calculated with general population denominators.
o Logistic regression on year as an ordinal variable to detect trends, and adjust for age and sex.
- Prevalence - crude
o 1995 – 4.9%.
o 2005 – 8.9% (Increase = 81.6%) (p < 0.0001 for trend by year).
- Prevalence – adjusted
o 1995 – 5%.
o 2005 – 9% (68.6% increase).
o Shows that aging of the population accounts for fairly little of the increase.
o Average increase of 6.2% per year.
- Number of patients with diabetes
o 1995 – 388433
o 2005 – 827419 (113% increase, compared to 17% population growth).
- Incidence – adjusted
o 1997 – 6.6/1000
o 2005 – 8.2/1000
o Relative increase of 30.7%.
- Mortality – adjusted
o 1995 – 3.3/1000
o 2005 – 1.9/1000
o Relative decrease of 42.0%.
- Age bands
o Prevalence higher in older individuals, but increase was 31% more in those aged 20-49 versus those aged 50 or older. Prevalence rate increase (relative excess) was 94% vs 62%.
o Prevalence higher in men than in women. Greatest rise in young women.
o Incidence higher in older than in younger individuals, but increase in incidence higher in the younger than older individuals.
o Incidence higher in men than women, both having similar increases over time.
o Mortality fell in all age-groups, but to a greater extent in the younger than the older age group.
o Mortality slightly higher in men than women, but similar decreases in both sexes.
- Prevalence estimates exceed projections in previous studies.
- Increase prevalence due to increasing incidence (31% rise over 1997-2003, to 8.2/1000 people). During this time, obesity has increased.
- Increased prevalence due to falling mortality. Mortality decline in this study much higher than in other studies – more advanced disease in other studies a possible explanation.
- High Ontario prevalence possibly due to immigration.
Eurich DT, Gamble JM, Simpson SH, Johnson JA. The darkening cloud of diabetes: Do trends in cardiovascular risk management provide a silver lining? Diabetes Care, 31(11): 2008.
- Early population projections appear to have underestimated the true prevalence of diabetes. This may be due to increased incidence and reduced diabetes mortality over time (Lipscombe LL, Hux JE. Lancet, 369: 2007; Thomas JR, Palumbo PJ, Melton LJ III, Roger VL, Ransom J, O’Brien PC, Leibson CL. Arch Intern Med, 163: 2003.).
- Have concurrent changes in pharmacotherapy utilization rates resulted in substantial improvements in the health of people with diabetes?
- Data source
o Saskatchewan Health. Includes prescription drug information for all ages. Registered Indians excluded.
o Outcome = All-cause hospitalization, hospitalization for CVD.
o NDSS diabetes definition, drugs of interest = antihypertensives, including RAS blockade, beta-blockers, CCBs, alpha-blockers, and statins.
o Rates age- and sex- standardized to 2001 Canadian census.
o Joinpoint regression for trend analysis. Can be performed on log risks. Starts with the minimum number of joinpoints and tests whether more joinpoints are statistically significant and must be added to the model, using a Monte Carlo permutation test. Detects apparent changes in trend.
- Prevalence – crude
o 1993 – 5.0%
o 2001 – 7.2%
- Prevalence – age- and sex-standardized
o 1993 – 4.7%
o 2001 – 6.5%
o Annual percent change (from joinpoint analysis) = 4.3% [3.8, 4.8], p < 0.001. Overall, increase = 34%.
- Drug utilization
o Increased. Annual percent change and shape of the trends varied from drug to drug. Annual percent change was from 1.5% then 6.3% (joinpoint at 1996) for anti-hypertensives generally, but was as high as 17.9% for statin therapy (223% relative increase between 1993 and 2001). Largest increase (absolute prevalence) occurred with ACE inhibitor therapy – one third of all diabetic patients received ACE inhibitors.
o Almost half of people with diabetes used an anti-hypertensive agent in 2001.
- Hospitalization rates
o Decreased from 300/1000 to 253/1000 from 1993-2001 (p < 0.05) (16% decrease).
o Decline in the elderly not as large compared with other groups examined, NS.
o Largest reduction in hospitalizations was for cerebrovascular-related events (annual percent change = 5.4%) (overall 36% decrease) (p < 0.05). Cardiac-related hospitalizations also declined 2% annually (overall 19% decrease) (p < 0.05). Similar reductions in both sexes.
- Mortality – “remarkably stable”, irrespective of sex or age. Tests for differences NS.
- Limitations – diagnostic creep? Appropriateness of pharmacotherapy? Patient self-care behaviors and other confounders?
- Despite the darkening cloud of diabetes, improvements in the utilization of evidence-based therapies and in hospitalization rates for cardiovascular events suggest a silver lining.
PHAC. Report for the National Diabetes Surveillance System: Diabetes in Canada, 2009. Public Health Agency of Canada: 2009
- Data source
o NDSS, Canadians aged 1 year or older (case definition includes type 1), data from 2006-07.
- Prevalence
o 2006-07 - 6.2% overall in Canadians aged 1 years or older.
o Highest prevalence (age-standardized) in Newfoundland and Labrador, Nova Scotia, Manitoba, and New Brunswick. Lower in the West, and Quebec.
o BC First Nations – 6.7% age-standardized vs 4.8% other BC residences, prevalence ~40% higher.
- Incidence
o 2006-07 – 6.7 per 1000 population aged 1 and older, higher among males (7.3/1000 vs 6.1/1000). (But higher for females of child-bearing years).
o 2000-01 – 5.3 / 1000
o Increase in age-standardized incidence rates of 9% between 2002-03 and 2006-07.
- Numbers
o 2006-07 – 2 million, slightly higher among males.
o 2012 (Forecasted) – 2.8 million, an annual percent increase of 6% every year.
- Age bands affected by prevalence / numbers
o Quite high prevalence rates in higher age bands.
o Incidence rates peaked in the 70-74 age group.
- Comments
o Data on mortality and costs available.
CDA. An Economic Tsunami: The Cost of Diabetes in Canada. CDA: 2009.
- Data source
o 2005 incidence and mortality rates (diabetes) from the NDSS.
o Statistics Canada’s medium population projection
o Canadian Diabetes Cost Model – economic impact estimated with NDSS utilization data and cost data from Health Canada’s Economic Burden of Illness in Canada report, including indirect costs of lost economic output and premature death attribute to diabetes.
o Incidence rates for males and females up to age 69 assumed to increase (reference case) at 0.75% per annum. This is below the 1.8% national average annual increase observed between 1998 and 2005 – conservative.
o Age-sex band mortality rates fixed at average of the last 3 years of available NDSS data.
- Prevalence
o 2000: 4.2%
o 2010: 7.3%
o 2020: 9.9%
- Numbers of patients with diabetes
o 2000: 1.3 milion
o 2010: 2.5 million
o 2020: 3.7 million
- Age bands affected by changes in prevalence and numbers
o Not reported.
- Drivers
o Population growth – responsible for 9% of the increase over the next decade.
o Age structure – diabetes incidence increases with age (peaks at 65-70 for males and females) – 13%.
o Increased likelihood of developing diabetes – 10%.
o Current demographic structure – Unsure what this means – 68%.
- Comments
o Dynamic model – entry and exit in the disease pool modeled formally.
o Increasing incidence modeled.
o Diagnosed diabetes only.
o No race-specific incidence or mortality. Mortality rates fixed.
Johnson JA, Balko SU. Chapter 2: Epidemiological trends in diabetes. Alberta Diabetes Atlas 2009 (Johnson JA, Ed.). Institute for Health Economics, Edmonton, AB: 2009.
- Prevalence has been increasing. Male age-sex adjusted rate up 63%, female prevalence rate up 54%. Male cases out-number female cases until age 75-79. Provincial prevalence rate in 2007 was 4.6% (all ages).
- Incidence
o Number of cases identified in 2007 double that identified in 1995.
o Incidence appears to be increasing, more so in males.
o Provincial 2007 incidence rate – 4.9/1000.
o More cases diagnosed in males than females, peak incidence appears to be around 65-69 and 70-74 age bands.
Complications of diabetes
Kannel WB, McGee DL. Diabetes and cardiovascular disease: The Framingham study. JAMA, 241(19): 1979
- Framingham study: Cohort of 5209 men and women aged 30 to 62 years in a small town served by a single hospital and a handful of physicians. Clinical cardiovascular endpoints diagnosed from biennial examinations and from hospital admissions, medical examiner’s reports, and other sources. Biennial examinations were also opportunities to reclassify patients on covariates (Markov assumption assumed in analysis, patient risk-time chopped into biennials). Mortality LTFU only 3%. Each biennial exam had 85% participation, with 69% total population participating over all examinations up to and including the tenth biennial examination or until time of death. Cohort initiated in 1949.
- Diabetes – previous history of treatment of oral hypoglycemic agents or a casual blood glucose level higher than 150mg/dL on two successive visits.
- Incidence proportions and relative risks calculated for all two-year intervals within strata. Logistic regression used to adjust for age and other variables. Follow-up = 20 years. Age for this analysis restricted to 45 to 74 years.
- 957 cases of CVD. Male preponderence.
- Adjusted RRs for diabetes lower than crude RRs. Adjusted for age, SBP, cigarettes per day, cholesterol, and LVH-ECG. Diabetes RR shown within strata of other risk factors.
o CVD: Men 2.11 Women 2.03
o CHF: Men 1.82 Women 3.75
o Intermittent claudication: Men 4.16 Women 4.99
o Atherothrombotic brain infarction: Men 2.18 Women 2.17
o Coronary heart disease: Men 1.66 Women 2.06
o Cardiovascular disease death: Men 1.7 Women 3.3.
- Unadjusted rates ranged from 2.48 (CVD) to 4.72 (intermittent claudication) for men, and 3.57 (coronary heart disease) to 8.87 (intermittent claudication) for women. Presumably, all estimates statistically significant.
- PAR calculated to highlight public health importance of diabetes – contributes 7.1% of CVD death in men and 18.3% of CVD death in women; 5.0% of CVD in men and 7.3% of CVD in women. Because PAR depends on absolute rates and prevalence of diabetes, these numbers should not be considered current.
Brancati FL, Whelton PK, Randal BL, Neaton JD, Stamler J, Klag MJ. Risk of end-stage renal disease in diabetes mellitus: a prospective cohort study of men screened for MRFIT. JAMA, 278(23): 1997
- Measure the RR of ESRD related to diabetes mellitus in a large cohort of men from across the US.
- Multiple Risk Factor Intervention Trial (MRFIT), multicenter, RCT. Recruitment between 1973 and 1975. 12866 men randomized. (Other details – need further references).
- Primary outcome of present analysis = all-cause ESRD. Age-adjusted incidence of all-cause ESRD more than 12 times greater in diabetic than nondiabetic men, RR = 12.7 [10.5, 15.4], rates = 199.0 vs 13.7 cases per 100000 person-years.
- Fully adjusted model included race, age, SBP, cholesterol, income, smoking, history of MI. diabetes RR = 9.0 [7.4, 11.0]. Diabetes much stronger predictor (effect modification) in men aged 35 to 39 at baseline (RR 28.1 [16.3, 48.3]). Lowest for men aged 55 to 57 years at baseline (RR 6.7 [4.3, 10.4]). All models fully adjusted.
- Diabetes is a powerful risk factor for the development of ESRD.
- (Combine with prevalence statistics from ADSS).
- (Combine with information on high cost of ESRF – “affects only 0.07% of the population but consumes 2 to 3% of the health care budget” – Cite ADSS Atlas.)
Haffner SM, Lehto S, Ronnemaa T, Pyorala K, Laakso M. Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarction. N Engl J Med, 339(4): 1998.
- Do patients with diabetes and patients with coronary heart disease have a similar risk of cardiovascular events?
- Prospective cohort study. Diabetic patients selected from the central pharmaceutical register of diabetic patients in two regions of Finland (Kuopio University Hospital district and Turku University Central Hospital district). 1059 diabetic patients, Control patients selected randomly using population register from these regions. Selected patients were aged 45 to 64 years. Interviews carried out to determine previous history of cardiovascular risk factors and previous stroke or CVD events. Patients were recruited from 1982 to 1984. A follow-up survey was sent (prospective) in 1990. Follow-up was ended at the end of 1989. Outcome was MI (fatal and nonfatal), stroke (Fatal and non-fatal), and cardiovascular mortality.
- LTFU – non-diabetic subjects without prior myocardial infarction at baseline – 681 available at 7 years. Non-diabetic subjects with prior MI at baseline – 27 available at 7 years. Diabetic subjects – MI 731 at 7 years – no MI 455 at 7 years.
- No TVC analysis reported.
- Diabetic patients were older, had higher BMI, more HTN, higher triglycerides, lower LDL and HDL.
- Previous MI associated with being male, smoking, older age, hypertension, obesity, higher total and LDL cholesterol and tryglyceride levels, and lower HDL levels.
- Table 2 – all P-values for prior MI vs no prior MI and diabetes vs no diabetes significiant, p < 0.001.
- Diabetic patients without prior MI
o MI rate = 3.2/100
o Stroke rate = 1.6/100
o Death = 2.5/100
- Diabetic patients with prior MI:
o MI rate = 7.8/100 person years
o Stroke rate = 3.4/100 person years
o Death = 7.3/100
- Non-diabetic subjects without prior MI
o MI rate = 0.5/100
o Stroke rate = 0.3/100
o Death = 0.3/100
- Non-diabetic subjects with prior MI:
o MI rate = 3.0/100
o Stroke rate = 1.2/100
o Death = 2.6/100
- Diabetes vs non-diabetes rate ratios within strata of prior-MI for each outcome ranged from 2.6 to 8.3. All diabetes vs no diabetes comparisons were significant.
- CPH model comparing diabetic subjects without prior MI with non-diabetic subjects with prior MI showed no differences. Age and sex-adjusted HR = 1.4 [0.7, 2.6]. Fully adjusted HR (LDL, HDL, trigs, smoking, HTN) = 1.2 [0.6, 2.4].
- Patients with type 2 diabetes who have not had a myocardial infarction have a risk of infarction similar to that among non-diabetic patients who have had prior MI. All patients with diabetes should be treated as if they had prior coronary heart disease.
- Effect of diabetes and previous MI alone versus no diabetes or previous MI was similar – effect modification – sub-additivity. In the presence of the other, the effect of diabetes or previous MI falls from 5 to 8-fold to 2 to 3-fold – cumulative IR – CRUDE data from table 2.
Gu K, Cowie CC, Harris MI. Mortality in adults with and without diabetes in a national cohort of the US population, 1971-1993. Diabetes Care, 21(7): 1998
- Estimate total and cause-specific mortality in diabetic and non-diabetic patients according to age, sex, race/ethnicity, and severity of diabetes.
- Age limits: 25-74 years old.
- NHANES I baseline cohort interviewed 1971-1975, with follow-up to 1992-1993 (22 years). Data collected on sex, age, race, ethnicity, education, duration of diabetes, diabetes therapy, previous CVD, physical activity, height, weight, BP, total cholesterol. Analysis by age-standardization (direct method, using 1980 US population), by Kaplan-Meier analysis, and by CPH models (CPH results not reported, focus is on age-standardized mortality effect of diabetes including comorbidities). Underlying and all listed causes of death combined.
- 13830 subjects had vital status determined at study end (96.2% of original cohort), 5.1% with diabetes accounting for 10.6% of the deaths. Rates of mortality higher for men than women for both diabetic and non-diabetic patients, and increased with age. RR (diabetes vs no diabetes) closer to unity as age increased, and higher for women than men in the middle-elderly age groups). RR (diabetes vs no diabetes) similar for ethnic groups.
- All subjects
o Aged 25-44, diabetic rate = 12.4/1000 person-years, RR = 3.6, p $8.14 billion in 2016.
ADA. Economic costs of diabetes in the U.S. in 2007. Diabetes Care, 31(3): 2008
- Estimate the economic burden and cost to society of diabetes. Update the 2003 report with estimates for continued growth in diabetes prevalence, changing practices, technology, and costs, and improvements in data sources and methods.
- Data sources: NHIS, NHANS, MEPS, MCURE, NIS, NAMCS, NHAMCS, NNHS, NHHCS.
- Size of the population with diabetes estimated from NHIS and NHANES (diagnosed diabetes only). Utilization captured from a variety of databases. Diabetes-attributable utilization in each category of utilization and in age-sex-setting-medical conditions bands calculated by multiplying utilization against the etiologic fraction. The etiologic fraction calculated from the Ingenix MCURE database of medical claims for 16.3 million people with private insurance (larger than MEPS, which is underpowered for certain strata) (Poisson regression for RRs). MEPS used to assess whether underlying health status and health behaviors, not accounted for in MCURE, make a difference. Where they do (general medical conditions not identified as chronic conditions associated with diabetes, hypertension, renal failure – the diabetes attributable portion would be overestimated here), MEPS used to calculate a scaling factor based on the ratio of log(RR(full model))/ log(RR(age-sex adjusted model only). MEPS RRs not deployed directly due to larger confidence intervals.
- Extra-mural health services (e.g.: ambulance, home health, podiatric) estimated from MEPS.
- Costs estimated from MEPS and NIS, with some estimates obtained from published sources. Lost productivity days (absenteeism and presenteeism) estimated from NHIS, days of those not in the workforce valued at 75% of average earnings for those in the labor force. Present authors calculate that, after adjusting for the effect of hypertension not attributed to diabetes, the average lost number of workdays per employed person was 1.8 (range 0.9 in age 18-34 to 2.5 in age 45-54). Presenteeism productive loss calculated at 14 days per worker (6.6%). Disability losses also included (e.g.: neuropathy), using NHIS detection of persons receiving social security supplemental insurance (SSI) payments (conservative).
- Mortality estimated from CDC death certification data. Used estimated from Kanya, Grady, Barrett-Connor, Archives of Internal Medicine, 2002: Calcualted that approx. 16% of deaths where CVD is the primary cause of death is due to diabetes. Etiological fractions for ED use deployed as proxies for mortality etiologic fractions in other CODS – 38% of cerebrovascular disease deaths, 57% of renal failure deaths. Lost lives valued using PVFP.
- Total US national costs $174 billion in 2007. $116 billion in medical, $58 billion in reduced productivity. Additional intangible costs not included. Larger than inflation-adjusted 2002 estimate due to growth in diabetes prevalence, medical costs rising faster than general inflation, and methodologic improvements.
- $1 of every $10 health care dollars is attributed to diabetes. Almost half of all HC expenditures attributed to diabetes come from higher rates of hospital admission and longer LOS ($58.3 billion, 14% of the total hospital inpatient care expenditures). Also retail prescriptions (8% of total prescription costs attributable, $12.7 billion).
- 40.7 million of 186 million hospital inpatient days projected for 2007 in the US will occur in patient with diabetes – 24.3 million / 186 million will be attributable to diabetes (13%).
- A large proportion of health resource use is for general medical conditions that are not chronic complications of diabetes – diabetes appears to compound admissions, resulting in longer LOS, more emergency and physician visits.
- Costs related to administration, investment, OTC drugs, disease management, etc. not included.
- Annual medical expenditures over 4x higher for diabetic vs non-diabetic patients ($11744 vs $2935). Thos with diabetes have expenditures 2.3x higher ($11744 vs $5095) than expenditures in the absence of diabetes – diabetes responsible for $6649/year/person excess.
- Lost work days = 15 million, at cost of $2.6 billion. Cost per person per year $493. Presenteeism lost 120 million workdays for $20 billion and $2883 per person per year.
- Despite limitations of the data (mostly towards being too conservative), the present findings demonstrate that the burden of diabetes and its complications on society is significant. Much of this cost may be preventable.
CDA. An Economic Tsunami: The Cost of Diabetes in Canada. Canadian Diabetes Association, Toronto, Ont.: 2009
- Economic burden of diabetes in Canada expected to be $12.2 billion in 2010, an increase of $5.9 billion over the level in 2000. Direct cost currently accounts for 3.5% of public healthcare spending in Canada. Rise in spending is the result of significant increase in the amount of health care utilization, as well as increases in prevalence driven largely by changes in the population structure (model assumptions).
- Direct costs expected to rise from $1.1 billion in 2000, tp $2.1 billion in 2010, to $3.1 billion in 2020. (THAT’S VERY LOW COMPARED TO THE ESTIMATES FROM OHINMAA).
Diabetes – Intervention Trials – UKPDS
UK Prospective Diabetes Study Group. UK Prospective Diabetes Study (UKPDS). VIII. Study design, progress and performance. Diabetologia, 34(12): 1991.
- Aims – UKPDS
o To determine whether treatment aiming for near-normal plasma glucose levels, < 6mM, reduces morbidity and improves life expectancy.
o To determine whether diet alone, sulphonylurea, insulin or metformin therapy has particular effectiveness.
- Aims – Hypertension in Diabetes
o To determine whether treatment aiming for a near-normal blood pressure ( 7.8mM is a risk factor for disease – basis of WHO diagnosis for diabetes. Also epidemiologic evidence linking to CVD.
o UGDP – no clinical benefit from insulin or sulphonylurea therapy. Tolbutamide appeared to increase cardiovascular mortality. FBG levels obtained may have been too high for therapeutic effect.
- Blood pressure – background – Treatment of hypertension in general population – beneficial, but questionable – application to diabetes? No evidence in type 2 diabetes.
- Design – factorial – 3x3 with fewer subjects sampled for the HDS.
- UKPDS and HDS – Recruitment
o Newly-diagnosed diabetes aged 25-65, referred by local GPs. Diabetes confirmed by FPG > 6mM on two occasions. 92% of UKPDS patients meeting the FPG criteria can be expected to meet the >11mM 2h OGTT criteria, based on a simple validation with linear regression.
o HDS entry: Not on antihypertensive therapy and BP >= 160 systolic and/or >= 90 diastolic; or treated and BP >= 150 and/or >= 85.
o Exclusion – History of morbidity, contraindication of study therapies, slightly more restrictive for HDS.
- UKPDS – Randomization – sealed opaque envelopes.
o Initial diet therapy –aim for ideal body weight – 3 months.
o Initial randomization
▪ FPG 15mM, primary diet failure, randomized to active therapy only.
▪ FPG 6mM, main randomization to continued diet, active policy aiming for FPG 6mM, underwent delayed randomization.
o Glucose study I – sulphonylurea – 50% chlorpropamide, 50% glibenclamide.
o Glucose
- UKPDS – Interventions
o Diet policy – ideal body weight, diet okay until FPG > 15mM. Secondary diet failure – FPG >15mM – randomized as in primary diet failure with FPG 15mM – except in sulfonylurea patients after 1990, who were randomized at FPG >6mM, < 15mM to continued sulfonylurea or add-on metformin. If marked hyperglycemia recurred, patients were changed to insulin.
- HDS – Randomization – sealed opaque envelopes.
o Tight control of BP aiming for 30kg/m2 in 35% of patients – newly diagnosed with diabetes.
- Metformin = gibuanide. Decreases blood glucose by lowering, rather than increasing, fasting plasma insulin concentrations – acts by enhancing insulin sensitivity – induces greater peripheral uptake of glucose – decreases hepatic glucose output.
- Biguanides also decrease concentrations of PAI-1 – increased fibrinolytic activity, similar to troglitazone.
- Only long-term biguanides outcome data – UGDP – phenformin – higher mortality from CVD vs placebo – higher total mortality vs insulin and placebo. Single death from lactic acidosis. Phenformin now withdrawn.
- Metformin – 12-20 fold lower risk of lactic acidosis than phenformin.
- UKPDS included two allocations to metformin
o Random allocation of obese active treatment arm patients (primary therapy).
o Random allocation of patients with FPG > 6.0mM at maximal doses of sulfonylureas (Add-on therapy) after 1990.
- At primary randomization, 4209 patients – 2022 non-overweight, 2187 overweight (52%) (>= 120% of ideal body weight).
- Randomization weighted heavily on metformin and diet – this is primary comparison. Randomization weights on sulfonylureas, insulin, and diet were also designed to be proportional to those in non-obese patients, for whom metformin was not a primary randomization treatment option – allows simple combinination of data from obese and non-obese strata with balance for obesity.
- 1704 overweight patients at treatment assignment for primary randomization to metformin.
- 537 patients, both overweight and non-overweight, treated with max doses of sulfonylurea with FPG between 6.1 and 15.0mM randomized 50/50 to early addition of metformin or continued sulfonylurea alone.
- Detected increased mortality in the metformin+sulfonylurea arm of the early metformin trial. Combined data with metformin+sulfonylurea treated patients in the primary randomization. Formal meta-analysis with tests for heterogeneity.
- “Epidemiologic assessment”. 457 met+sul patients – could occur among sulfonylurea patients at >= 15mM or among metformin patients at >= 15mM, in addition to sulfonylurea patients at >= 6mM (early addition of metformin). “All patients received combined therapy as a result of progressive hyperglycemia”. This group compared against each other treatment group for diabetes-related death. CPH models with actual therapy as a TVC and adjustment for age, sex, ethnicity, and baseline FPG.
UK Prospective Diabetes Study Group. Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes (UKPDS 34). Lancet, 352: 1998.
- Results – main randomization to metformin – overweight patients – 1704.
- Median follow-up = 10.7 years. LTFU < 10%.
- Lab data
o FPG, HbA1c – Met – decrease in the first year, subsequent gradual rise.
o Median HbA1c over 10 years – met 7.4%, conventional 8.0%.
o Median HbA1c – first 5 years – met 6.7%, conventional 7.5%.
o Similar control with sulfonylurea or insulin, to metformin.
o Weight gain – similar in metformin and conventional control groups, less than the increase observed for sulfonylureas or insulin.
- Crossing over of diet-treated patents – non-intensive allocation to metformin at FPG >= 15mM – 83% of person-years of follow-up – numbers are inconsistent, elsewhere stated that 44% of person –years were crossed over to non-intensive pharmacological therapies, of which metformin is only one.
- Crossing over of metformin treated patients – switch to monotherapy with insulin – 18% of person years.
- Adverse effects – per-protocol – proportion per year – any hypoglycemic attacks: conventional 0.9%, chlorpropamide 12.1%, glibenclamide 17.5%, insulin 34.0%, metformin 4.2%. Metformin proportion per year of major hypoglycemic attacks = 0%.
- Endpoints - aggregate
o Any diabetes-related end-point – RR = 0.68 [0.53, 0.87], p = 0.0023. Significantly greater risk reduction vs intensive.
o Diabetes-related death – RR = 0.58 [0.37, 0.91], p = 0.017.
o All-cause mortality – RR = 0.64 [0.45, 0.91], p = 0.011. Significantly greater risk reduction vs intensive.
o MI – RR = 0.61 [0.41, 0.89], p = 0.01
o Stroke – RR = 0.59 [0.29, 1.18], p = 0.13.
o PAD – RR = 0.74 [0.28, 2.09]
o Microvascular – RR = 0.71 [0.43, 1.19] – comparable to UKPDS 33 but NS.
o In contrast, intensive therapy was non-significant for all main endpoint – overweight patients only.
- Endpoint – clinical (99% CI)
o No significant differences detected.
o Fatal MI – RR = 0.50 [0.23, 1.09], p = 0.02
o Non-fatal stroke – RR = 0.42 [0.12, 1.45], p = 0.06
o Retinal photocoagulation – RR = 0.69 [0.34, 1.39], p = 0.17.
- Endpoints – surrogate
o Lower rate of progression to retinopathy at 9 years, gone by 12 years.
- NNT = 9.7 for any diabetes-related event – DL calculated from overall proportions.
- NNT = 19.2 for any diabetes-related death – DL calculated from overall proportions.
UK Prospective Diabetes Study Group. Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes (UKPDS 34). Lancet, 352: 1998.
- Results – early addition of metformin to patients receiving sulfonylurea – 537
- Median follow-up = 6.6 years. LTFU < 10%.
- Cross-over – additional metformin – to insulin – 38% of person years.
- Cross-over – sulfonylurea – 25% of years to metformin / insulin.
- Laboratory values
o Median HbA1c over 4 years – additional metformin 7.7%, sulfonylurea alone 8.2%.
o No significant differences in bodyweight or plasma insulin.
- Endpoint – aggregate
o Any diabetes-related endpoint - Sul+met – RR = 0.68 [0.53, 0.87]
o Diabetes-related death – Sul+met – RR = 1.96 [1.02, 3.75], p = 0.039
o All-cause mortality – Sul + met – RR = 1.60 [1.02, 2.52], p = 0.041
o MI – Sul+met – RR = 1.09 [0.67, 1.78], p = 0.73
o Stroke – Sul+met – RR = 1.21 [0.58, 2.55]
o PAD – Sul+met – RR = 2.12 [0.19, 23.3]
o Microvascular disease sul+met – RR = 0.84 [0.43, 1.66]
- Endpoint – clinical
o No significant differences.
o Fatal MI increased, RR = 1.79 [0.64, 4.99], p = 0.14.
o Non-fatal MI decreased, RR = 0.62 [0.26, 1.48], p = 0.15.
o Similarly with stroke.
UK Prospective Diabetes Study Group. Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes (UKPDS 34). Lancet, 352: 1998.
- Results – meta-analysis – intended to account for extremes in the plays of chance (DL comment – inappropriate anyways)
o Heterogeneity tesets confirmed different outcomes between trials for any diabetes-related endpoint, diabetes-related death, and all-cause mortality.
o Endpoint – aggregate
▪ Any diabetes-related endpoint – RR = 0.81 [0.67, 0.98], p = 0.033
▪ Diabetes related death – NS
▪ All-cause mortality – NS
- Results – epidemiologic analysis
o 4417 patients with 45527 person-years of follow-up. 11% of person-years treated with sul+met therapy, accounting for 8% of the 490 diabetes-related deaths.
o Sul+met combination compared with all other treatments – CPH – reduction in risk of diabetes related death – risk reduction of 5% [-33, 32], p = 0.78.
UK Prospective Diabetes Study Group. Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes (UKPDS 34). Lancet, 352: 1998.
- Benefit of metformin over sulphonylurea and insulin therapy not explained by improved glycemic controls. May be due instead to decreases in PAI-1, an inhibitor of fibrinolysis. Metformin also lowers systemic methylglyoxal concentrations. However, these mechanisms may be irrelevant, since addition of metformin to sulfonylurea therapy did not offer any improvements.
- Sulfonylurea treated patients were 5 years older, more hyperglycemic (baseline median FPG 9.1 vs 8.1 after 3 months of diet “wash-out”), less overweight, and were followed up for 5 fewer years at time of randomization to additional metformin compared with overweight patients in the primary metformin randomization.
- Epidemiologic analysis did not corroborate increased mortality risk.
UK Prospective Diabetes Study Group. Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 38. BMJ, 317(7160): 2998.
- 1148 patients enrolled in HDS from 1987-1991. 758 allocated to tight control ( 1/3 with previous CVD.
- Results
o Lab
▪ Overall weighted mean HbA1c difference = 0.88%
▪ Overall weighted mean FPG difference = 1.53 mM
o Endpoints
▪ MACE – HR = 0.91 [0.84, 0.99]
▪ MI – fatal and non-fatal – HR = 0.85 [0.76, 0.94]
▪ Stroke – fatal and non-fatal – HR = 0.96 [0.83, 1.10]
▪ Hospitalized fatal CHF – HR = 1.00 [0.86, 1.16]
▪ All-cause death – HR = 1.04 [0.90, 1.20], I2 = 47.5%.
▪ CVD death – HR = 1.10 [0.84, 1.42]
o Hypoglycemia – HR = 2.48 [1.91, 3.21], I2 = 72.1%.
- Meta-analysis demonstrates a modest reduction in major macrovascular events with greater glucose lowering. The effect is driven primarily by a 15% reduction in the risk of MI.
- The magnitude of these macrovascular risk reductions are consistent with the epidemiological relationship between HbA1c and cardiovascular events reported from observational studies.
- However – high heterogeneity WRT CVD suggests that either patient characteristics, the approach to glucose lowering, or other variables may affect CVD risk.
Diabetes – Intervention Trials – ACCORD, ADVANCE, and VADT
ACCORD Study Group. Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial: Design and methods. Am J Cardiol, 99(12A): 2007.
- Objectives – Can CVD event rates be reduced in patients with type 2 diabetes who are at high risk for CVD events by intensively targeting hyperglycemia, dyslipidemia, and BP?
o Intensive glucose control (HbA1c 7.5%), treated or untreated. High risk for CVD, selected by presence of clinical CVD (e..: prior MI), or risk factors for (e.g.: microalbuminuria, ECG findings) and older age.
o Eligibility for lipid or blood pressure trial.
o Exclusion – factors related to LTFU, imminent death, or intolerance of study treatments.
- Outcomes
o Composite of nonfatal MI, nonfatal stroke, or CVD death (MACE equivalent).
o Secondary – other CVD outcomes, total mortality, microvascular disease.
▪ Macrovascular outcomes – MACE plus revascularization and hospitalization for CHF.
▪ Total mortality
▪ CVD mortality
▪ Total CAD events, fatal and non-fatal, including unstable angina.
▪ Total stroke.
▪ CHF hospitalization or death
▪ Microvascular outcome – progression of diabetic retinopathy (3 stages on ETDRS), photocoagulation, or vitrectomy.
▪ Microvascular – Fatal or non fatal renal failure, or photocoagulation or vitrectomy
▪ HRQL and cost-effectiveness
- Design – double 2x2 (i.e.: 4x2 design). Analysis by multiple variable modeling, intended to disentangle an intervention effect common across cross-cutting strata (marginal, or main, effect).
- The glucose control arms cut across both lipid and blood pressure arms, but the lipid and blood pressure arms do not cut across each other.
- Power – 89% to detect a 15% treatment effect of intensive glycemic control vs standard glycemic control.
- Interventions
o Intensive glycemic control – HbA1c taret < 6.0%.
o Standard glycemic control – HbA1c target 7.0%-7.9%.
o Structured treatment algorithms using metformin, sulfonylureas, meglitinides, thiazolidinediones, alpha-glucosidase inhibitors, insulin, insulin analogues, and lfiestyle intervention, with exenatide added to the formulary in April 2007.
o Lipid lowering therapy – fenofibrate (patient and provider blinded) + simvastatin (open label), vs simvastatin only.
o Intensive SBP target 8%.
o Median 7.5% expected, should provide a contrast of 1.0% to 1.5%
- Intensive glycemic control
o Lifestyle
o Initially - >= 2 classes of agents provided.
o Dosage of >= 1 class increased, or an additional class provided, when HbA1c is >= 6%, or FPG or post-meal PG values exceed normal bounds.
o CDSS implemented through a web-based portal.
- Intervention strategies
o Principles outlined.
o Key is HbA1c, regardless of how it is obtained. Principles underwent updating during the trial.
Action to Control Cardiovascular Risk in Diabetes Study Group. Effects of intensive glucose lowering in type 2 diabetes. New England Journal of Medicine, 358(24): 2008.
- Study terminated early due to evidence of harm (January 8, 2008).
- 10251 patients aged 62.2 (6.8) years old randomized. Mean follow-up 3.5 years. LTFU < 10%.
- Results
o HbA1c
▪ Baseline identical at 8.1%. After 4 months, intensive group HbA1c = 6.7% vs standard group HbA1c = 7.5%.
▪ At 1 year, median HbA1c intensive group = 6.4%, standard group = 7.5%.
▪ Stable over follow-up.
o Drug use – intensive treatment – greater exposure to drugs of all classes, more frequent changes. Additional therapies or dosage increases exceeding 10% occurred 4.4 times / year vs 2/0 times per year, intensive vs standard therapy. Slight imbalance in ACEi use and SBP (both lower in intensive care).
o Adverse effects
▪ Mean weight gain at 3 years – 3.5 kg vs 0.4kg.
▪ Fluid retention – 70% vs 67%.
▪ Hypoglycemia requiring any assistance – 16% vs 5%
▪ Hypoglycemia requiring medical assistance – 11% vs 4%.
▪ Hypoglycemia – medical assistance – per year – 3.1% vs 1.0%.
o Endpoint – primary – MACE – 6.9% vs 7.2% (cumulative), 2.1% vs 2.3% (annual), HR = 0.90 [0.78, 1.04], p = 0.16.
o Endpoints – secondary
▪ All cause death – 1.4% vs 1.1% (annual), HR = 1.22 [1.01, 1.46], p = 0.04. Relative excess = 22%, AR = 1.0%.
▪ CVD death – 0.8% vs 0.6% (annual), HR = 1.35 [1.04, 1.76], p = 0.02.
▪ Nonfatal MI – HR = 0.76 [0.62, 0.92], p = 0.004.
▪ Appears to be more fatal MI in intensive treatment, 0.4% vs 0.3% (cumulative).
- Rates of all-cause mortality began to separate after 1 year and persisted throughout the follow-up period. The HR was consistent within subgroups.
- Number needed to harm (all cause mortality) over 3.5 years = 95.
- Increased mortality may be due to HbA1c levels, magnitude or speed of reduction in HbA1c, changes in drug regimens, rates of hypoglycemia, adverse effects due to undetected drug interactions, or some combination.
- Notably, fewer nonfatal MIs occurred, with the difference emerging 1 or 2 years after the mortality difference.
- Study has identified a previously unrecognized harm of intensive glucose lowering in high-risk patients with type 2 diabetes.
Riddle MC, Ambrosius WT, Brillon DJ, Buse JB, Byington RP, Cohen RM, et al.. Epidemiologic relationships between A1c and all-cause mortality during a median 3.4-year follow-up of glycemic treatment in the ACCORD trial. Diabetes Care, 33(5): 2010.
- Observational study of all-cause mortality and glycemic exposure.
- Four methods of measuring glycemic exposure
o TVC of mean of all 4-month A1c values (cumulative)
o TVC of most recent A1c value
o TVC of the magnitude of the early reduction of glucose levels (cumulative mean every 4 months minus baseline for first 12 months, and then constant) = 12 month decrease in A1c
o Constant for the 4-month decrease in A1c.
- Pre-randomization, randomization, and post-randomization confounders
o Pre randomization – Site characteristics, baseline characteristics
o Post-randomization
▪ TVC for having had a major hypoglycemic episode
▪ TVC for weight change from baseline (categorical)
▪ Assignment to BP or lipid trials.
- CPH models for HR. Curves (spline functions) modeling A1c – mortality relationship used to explore linearity assumption in the CPH model. Poisson regression to estimate rates.
- Strange associations between baseline variables and death in Table 1.
- Of the 4 methods of representing HbA1c, average A1c (cumulative, updated) had the strongest association – larger 1 year decreases in A1c actually appeared to be protective (fully adjusted model only).
o Average HbA1c HR = 1.45 [1.3, 1.63], p < 0.0001 (fully adjusted).
o Last A1c HR = 1.14 [1.05, 1.25], p = 0.0026 (fully adjusted).
o 1-year decrease in A1c HR = 0.85 [0.75, 0.97], p = 0.0127 (fully adjusted).
- Allocation to intensive vs standard was an interaction factor (p = 0.0007).
o Average A1c Intensive HR = 1.66 [1.46, 1.89], Standard HR = 1.14 [0.95, 1.38]
o Setting reference category = risk of mortality at standard treatment with HbA1c of 6% - a spline-smoothed graph of logHR for the intensive vs standard therapy group shows different curves for the two strategies.
▪ Intensive – mortality risk increased steadily with higher average A1c from 6.0 to 9.0% - appears log-linear.
▪ Standard – lowest risk associated with average A1c between 7.9 and 8.0%. Risk of mortality generally lower at all levels than in the intensive group, except at lower A1c, when intensive therapy HR 9.0% - these would appear to be the patients that were more likely to die under intensive therapy.
o Smoothed spline graph of death rate (Poisson regression) as a function of 12-month HbA1c decline
▪ Similar death rates between intensive and standard over most of the range, intensive slightly higher.
▪ Large spike in intensive therapy deaths when the HbA1c decrease was < 0.5% or increased, suggesting that those patients who fail to experience a decrease in HbA1c are those who were more likely to die under intensive therapy.
- Various HbA1c measures were associated with mortality. Average HbA1c had the strongest association.
- Relationships between A1c and mortality differed between treatment strategies – differences other than A1c alone account for higher mortality in the intensive arm.
- Mortality rates diverge the most for the intensive treatment arm when, contrary to the intent of the strategy, A1c levels were > 7%.
- Factors that lead intensively treated A1c to persist above 7% are unknown. Candidates include behavioral issues, mental health, social / financial stresses, and serious medical problems.
- Low A1c is not likely to be the primary mediator of increased risk of death under intensive glycemic control.
- Journal club notes
o Received wisdom now leans towards a hypoglycemia – sympathetic response – destabilized atherosclerotic plaques or myocardium theory.
o However, these results do not support this. Not, though, that this may be spurious due to adjustment for hypoglycemic events.
o DTE –quantities may be im-balanced across strata of post-randomized variables. Also, deliberately forcing patients apart will introduce extraneous variance in the observational anlaysis.
Abraira C, Duckworth W, McCarren M, Emanuele N, Arca D, Reda D, et al.. Design of the cooperative study on glycemic control and complications in diabetes mellitus type 2: Veterans Affairs Diabetes Trial. J Diabetes Complicat, 17 (6): 2003.
- Main clinical benefit of intensive glucose control is reduction in photocoagulation. However, UKPDS did not show effect on visual acuity or renal failure, possibly because of low event numbers. Note, however, that the major cause of vision loss in patients with diabetes is not diabetic retinopathy, but maculopathy, which is common in both diabetic and non-diabetic patients.
- Periodic eye examinations and subsequent intervention alone may prevent 90% of the vision loss in patients with diabetes (Ferris, How effective are treatments for diabetic retinopathy? JAMA, 269: 1993).
- BP control has greater impact on CVD than glycemic control (UKPDS 38).
- Lipid control – benefits in CV endpoints in diabetic patients – exceeds those observed in non-diabetic patients (Pyorala K, Pedersen TR, Kjekshus J, Faergeman O, Olsson AG, Thorgeirsson G. Cholesterol lowering with simvastatin improves prognosis of diabetic patients with coronary heart disease: A subgroup analysis of the Scandinavian Simvastatin Survival Study (4S). Diabetes Care, 20: 1997).
- Glycemic control – no trials showing macrovascular benefits, except UKPDS 34.
- VA-Cooperative Study on Diabetes Mellitus feasibility study
o Abraira C, Colwell J, Nuttall F, Sawin CT, Henderson W, Comstock JP, et al. Arch Intern Med, 157(2): 1997.
o Mean follow-up = 27 months.
o 153 insulin-requiring older obese men with a high prevalence of CV complications
o Mean age 60 +/- 6 years.
o RCT
▪ Standard therapy = 1 insulin injection every morning
▪ Intenve therapy = stepped plan from 1 evening injection of insulin, alone or with glipizide, to multiple daily injections. Target = near normal glycemia levels.
o Separation of 2.1% HbA1c in intensive vs standard treatment arms – no deterioration of HbA1c during trial.
o Results
▪ Mild/moderate hypoglycemic events: 16.5 vs 1.5 per patient per year.
▪ Overall risk of cardiovascular events = 32% vs 20%, p = 0.10.
▪ No difference in total or cardiovascular mortality. CPH model NS for treatment allocation.
o Conclusion – a long-term prospective trial is needed to assess the risk-benefit profile of insulin therapy for type 2 diabetes in the older population.
- What does this mean for older adults, who comprise the vast majority of patients in the VA, and a large proportion of adults in the general population of those with diabetes?
Abraira C, Duckworth W, McCarren M, Emanuele N, Arca D, Reda D, et al.. Design of the cooperative study on glycemic control and complications in diabetes mellitus type 2: Veterans Affairs Diabetes Trial. J Diabetes Complicat, 17 (6): 2003.
- Objectives
o Determine the effect on major CV events of intensive glycemic control compared to standard care.
o Assess differences between treatment groups in other CV endpoints (surrogates), and in total mortality.
o Examine differences in eye disease, nephropathy and neuropathy, HRQL, side effects, costs and benefits.
- Design: Prospective, two-arm RCT.
- Patients
o Veterans >41 years old
o Nonresponsive on insulin or maximum dose of an oral agent, nonresponsive = HbA1c > 7.5%.
o Exclusions – intolerance of study therapies.
o Actual mean HbA1c of selected patients = 9.4%.
- Outcomes
o Primary – CV events = MI, stroke, new or worsening CHF, amputation, invasive intervention for CAD or PAD, CV death.
o Secondary
▪ Eye disease – progression of retinopathy, incidence of photocoagulation, cataract extraction, and reduced visual acuity
▪ Nephropathy – first-void urine ACR and sCR.
▪ Neuropathy – Hx & PE.
▪ Cognitive tests, HRQL, costs.
- Interventions
o Both arms received diet and education, and were treated to standards for blood pressure, lipids, aspirin therapy, and smoking cessation.
o Intensive therapy – HbA1c < 6%.
▪ SBGM.
▪ Metformin (obese) or glimepiride and rosiglitazone (lean), then evening insulin, then multiple daily insulin injections,e tc..
o Standard therapy – HbA1c > 8% but < 9%, prevent symptoms of diabetes.
▪ SBGM
▪ Similar drugs as intensive therapy, without multiple daily insulin injections.
o Since all patients have inadequate control to begin with, both arms will receive combination therapy.
o Study arms are to receive similar distributions of drug classes to address different levels of glycemic control – main difference between groups is expected to be the dose of insulin.
o Normoglycemia chosen as a goal because of epidemiologic evidence indicating no risk across the span of HbA1c levels, with no thresholds.
- Sample size of 1700 required to demonstrate a 21% reduction in CV events with 86% power.
- In all three of the main metabolic risk factors in diabetes, the standard treatment patients will receive better and more effective treatment than under current care. Thus, outcomes such as renal failure or visual acuity are not expected to differ between study arms.
- “The treatment of risk factors for retinopathy and nephropathy, especially hypertension and periodic eye examination, permits the evaluation of additional benefits of intensification of glycemic control beyond levels generally considered to be prudent.”
- It could be said that, patient-wise, and in other ways, VADT begins where UKPDS left off.
Duckworth W, Abraira C, Moritz T, Reda D, Emanuele N, Reaven PD, et al. Glucose control and vascular complications in veterans with type 2 diabetes. NEJM, 360(2): 2009.
- 1791 patients – mean age 60.4 years – mean length of diagnosed diabetes 11.5 years – BMI = 31.3 – HbA1c at baseline = 9.4% - HTN (>140/90) in 72% – previous CVD in 40% – median follow-up = 5.6 years.
- Lipid and BP profiles improved during follow-up, NS for differences at baseline or at follow-up.
- Results
o Weight – 4kg more in intensive arm.
o HbA1c – 6.9% vs 8.4%, between group difference = 1.5%, apparent after 6 months and stable.
o Endpoints – CVD (MACE equivalent)
▪ HR = 0.88 [0.74, 1.05]
o Endpoint – secondary
▪ NS for components of the CVD (MACE) endpoint and secondary outcomes.
▪ All-cause mortality – HR = 1.07 [0.81, 1.42], p = 0.62.
▪ Microvascular – eye – NS on all outcomes. Trend towards benefit WRT diabetic retinopathy (progression of at least two steps) (p = 0.07).
▪ Microvascular – nephropathy and neuropathy
• Worsening of albumin excretion (p = 0.01) and
• Progression to macroalbuminuria (p = 0.04) less frequent in intensive group.
• Trend towards increase autonomic neuropathy in intensive group.
o Adverse events
▪ Hypoglycemia – impaired consciousness – 9% vs 3% per year (p < 0.001).
▪ Hypoglycemia – complete loss of consciousness – 3% vs 1% (p < 0.001).
- The benefit of decreasing the glycated hemoglobin level from 8.4% to 6.9% appeared to be minimal. Appropriate management of hypertension, dyslipidemia, and other cardiovascular risk factors appears to be the most effective approach to preventing cardiovascular morbidity and mortality.
ADVANCE Management Committee. Study rationale and design of ADVANCE: Action in Diabetes and Vascular disease – preterax and diamicron MR controlled evaluation. Diabetologia, 44: 2001.
- RCTs – lowering BP of hypertensive patients leads to substantial reductions in major vascular events (UKPDS 38), with more intense lowering leading to larger reductions (Blood Pressure Lowering Treatment Trialists’ Collaboration. Effects of ACE inhibitors, cacium antagonisis, and other blood-pressure-lowering drugs: results of prospectively designed overviews of randomized trials. Lancet, 355: 2000.)
- BP and glycemia – no threshold observed.
- HOPE study – ACEi therapy confers worthwhile vascular benefits for diabetic patients whether hypertensive or not (Heart Outcomes Prevention Evaluation study Investigators. Lancet, 355: 2000).
- Objectives
o Are there worthwhile benefits of BP lowering for high-risk diabetic patients, regardless of the BP?
o Are these benefits additional to those conferred by background treatment with an ACE inhibitor?
o Does intensive glucose control reduce the risk of major macrovascular disease?
- Design: International factorial 2x2 RCT.
o Double blinded for the BP intervention
o Open label for the glycemia intervention
o Randomization by central computer with stratification.
- Patients
o Broad cross-section of patients at high risk, age > 55 years old.
o High risk
▪ Diabetes duration > 10 years, or
▪ Age 65+, or
▪ History of CVD, CAD, invasive procedures for CAD, severe eye disease, macroalbuminuria, or another major risk factor for vascular disease.
- Interventions
o Perindopril available for all patients by clinical judgement.
o BP lowering – perindopril (ACEi) and indapamide (diuretic).
o Glucose control – gliclazide (sulfonylurea), then non-pharmacologic therapy, other oral agents, and insulin – target HbA1c < 6.5%.
- Outcomes
o Primary
▪ Non-fatal stroke, non-fatal MI, CVD death (MACE)
▪ New or worsening nephropathy or microvascular eye disease.
o Secondary
▪ CVD
▪ CAD
▪ CHF
▪ PAD
▪ Microalbuminuria
▪ Visual deterioration
▪ Neuropathy
▪ Demention
▪ All-cause mortality
- Power = 90% to detect a 16% reduction in the RR of each primary outcome.
ADVANCE Collaborative Group. Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes. NEJM, 358(24): 2008.
- 12877 eligible participants – 11140 randomized – baseline – hbA1 = 7.5% - age = 66 +/- 6 years – previous major CVD in 32% – Mean follow-up = 5.0 years.
- Results
o HbA1c – 6.5% vs 7.3% – time weighted average difference = 0.67%.
o FBG – 1.2mM lower.
o SBP lower in the intensive group by 1.6 mmHg.
o Mean body weight 0.7kg higher in the intensive group (NS).
o Endpoints – primary
▪ MACE or microvascular event – HR = 0.90 [0.82, 0.98], p = 0.01.
▪ MACE – HR = 0.94 [0.84, 1.06]
▪ Major microvascular events – HR = 0.86 [0.77, 0.97], p = 0.01
o NS for interaction with the crossed BP intervention.
o Endpoints – secondary
▪ All-cause mortality – HR = 0.93 [0.83, 1.06]
▪ New or worsening nephropathy – HR = 0;79 [0.66, 0.93], p = 0.006
• New-onset microalbuminuria – HR = 0.91 [0.85, 0.98], p = 0.02
• New-onset macroalbuminuria – HR = 0.70 [0.57, 0.85], p < 0.001.
▪ Hospitalization, any cause – HR = 1.07 [1.01, 1.13], p = 0.03, (45% vs 43%)
▪ NS on other secondary outcomes.
o Hypoglycemia
▪ Severe – 2.7% vs 1.5% (overall) – HR = 1.86 [1.42, 2.40], p < 0.001 – 0.7 vs 0.4 evens per 100 pateients per year.
▪ Minor – 120 vs 90 events per 100 patients per year
▪ Any – overall proportions – 5 years – 53% vs 38%.
- NNT over 5 years for a major macro- or micro-vascular event = 52.
- Main contributor to the 10% reduction in study outcomes was decrease in new or worsening nephropathy.
- Clinical benefit of this outcome? As it turns out, very few patients with diabetes ultimately go to end-stage renal failure … However, indices of renal impairement are strongly associated with the future risk of major vascular events, end-stage renal disease, and death.
- No effect on macrovascular events – but may have turned to have been underpowered. CIs are broad.
- BP difference – may explain at most 3% out of the 10% relative excess observed. May be a nonspecific effect of more frequent health care contact.
- Rate of MACE = 2.2%. Previous studies of patients with type 2 diabetes (UKPDS 33 and HOPES) suggested a rate around about 3% per year. Lower in this study.
- No evidence that the intensive glucose control strategy increased mortality.
- It is clear that the prevention of macrovascular complications of diabetes requires a multifactorial approach addressing all major modifiable risk factors.
Dluhy RG, McMahon GT. Intensive glycemice control in the ACCORD and ADVANCE trials. NEJM, 358(24): 2008.
- Differences in the trials
o ACCORD – no restriction on glucose-lowering treatments.
o ADVANCE – intensive control arm required ro receive gliclazide
o ACCORD – thiazolidinedione treatment (rosiglitazone) in 90% of intensive therapy and 58% of standard therapy arms.
o ADVANCE – thiazolieinediones infrequent.
o ACCORD – follow-up halted early at median 3.5 years.
o ADVANCE – follow-up of 5.0 years.
o ACCORD – primary endpoint – MACE
o ADVANCE – primary endpoint – MACE + major microvascular events – problematic admixture.
o ACCORD – aspirin and statins in 75% and 88%.
o ADVANCE – non-glycemic CVD risk factors not optimally controlled – approx. half received aspirin and half received statins.
o ACCORD – 3.5kg weight gain in intensive therapy group.
o ADVANCE – NS for differences in weight gain – may be due to lower frequency of insulin and thiazolidinedione use.
- Similarities
o Factorial designs
o Less emphasis on diet and lifestyle
o Large number of patients with complete follow-up
o Older patients, approximately one third of whom had previous CVD – more representative of actual patients with diabetes.
- ACCORD – excess deaths
o “Unexpected or presumed CVD” – may have been hypoglycemia related.
o Also rosiglitazone meta-analysis – associated with significant increase in risk of MI, and nearly significant increase in risk of death from CVD (Nissen SE, Wolski K. NEJM, 356: 2007).
- Contribution of glucose lowering to the reduction of macrovascular events in UKPDS, ADVANCE, and ACCORD appears to be minimal.
- Bottom line – The most appropriate target for glycate hemoglobin should remain 7%. Neither the ADVANCE nor the ACCORD trials undermine the importance of meeting the current guidelines for care.
- Decreased absolute event rates highlight the difficulty of showing additional improvements in outcome as care is progressively optimized across the board.
Turnbull FM, Abraira C, Anderson RJ, Byington RP, Chalmers JP, Duckworth WC, Evans GW, Gerstein HC, Holman RR, Moritz TE, Neal BC, Ninomiya T, Patel AA, Paul SK, Travert F, Woodward. Intensive glucose control and macrovascular outcomes in type 2 diabetes. Diabetologia, 52(11): 2009.
- Provide more precise estimates of the effects of glucose-lowering on major cardiovascular events in ACCORD, ADVANCE, UKPDS, VADT, and other large glucose lowering trials (if any).
- Inclusion:
o Trials designed to assess directly the impact of achieving higher vs lower levels of glycemia on CVD outcomes (i.e.: instead of, say, the effect of a particular drug, with treatment to similar glycemia targets).
o Large (>1000 person0years of follow-up in each arm), > 2 years median follow-up, randomized, controlled, double-blinded or PROBE design, ITT, LTFU < 10%.
- No IPD.
- Random effect models for differences in difference (HbA1c) and for CPH HRs, using inverse variance weights. Heterogeneity I2.
- 4 studies: ACCORD, ADVANCE, UKPDS, VADT – 27049 participants randomized, median follow-up from 3.4 years (ACCORD) to 5.6 years (VADT) – average weighted follow-up = 4.4 years.
- Mean age = 62 years, mean duration of diabetes 9 years, > 1/3 with previous CVD.
- Results
o Lab
▪ Overall weighted mean HbA1c difference = 0.88%
▪ Overall weighted mean FPG difference = 1.53 mM
o Endpoints
▪ MACE – HR = 0.91 [0.84, 0.99]
▪ MI – fatal and non-fatal – HR = 0.85 [0.76, 0.94]
▪ Stroke – fatal and non-fatal – HR = 0.96 [0.83, 1.10]
▪ Hospitalized fatal CHF – HR = 1.00 [0.86, 1.16]
▪ All-cause death – HR = 1.04 [0.90, 1.20], I2 = 47.5%.
▪ CVD death – HR = 1.10 [0.84, 1.42]
o Hypoglycemia – HR = 2.48 [1.91, 3.21], I2 = 72.1%.
- Meta-analysis demonstrates a modest reduction in major macrovascular events with greater glucose lowering. The effect is driven primarily by a 15% reduction in the risk of MI.
- The magnitude of these macrovascular risk reductions are consistent with the epidemiological relationship between HbA1c and cardiovascular events reported from observational studies.
- However – high heterogeneity WRT CVD suggests that either patient characteristics, the approach to glucose lowering, or other variables may affect CVD risk.
Diabetes – Intervention Trials – Thiazolidinediones
Dormandy JA, Charbonnel B, Eckland DJ, Erdmann E, Massi-Benedetti, M, Koules I, et al. Secondary prevention of macrovascular events in patients with type 2 diabetes in the PROactive Study (PROspective pioglitAzone Clinical Trial In macroVascular Events): a randomized controlled trial. Lancet, 366(9493): 2005.
- Pioglitazone – PPAR-gamma agonist. General improvement in various CVD RFs. Also reduces levels of inflammatory markers (i.e.: hsCRP).
- Does pioglitazone reduce CVD mortality in patients with type 2 diabetes?
- Methods
o Patients
▪ Aged > 35 years, HbA1c > 6.5% despite standard therapy.
▪ High-risk – recent MI or stroke, or invasive CAD/PAD intervention; ACS; objective evidence of CAD or PAD.
o Simple parallel group RCT, placebo controlled, randomization by central interactive voice response system.
o Intervention – pioglitazone, target HbA1c < 6.5%. Attention to lipid, antiplatelet, and antihypertensive therapy. All investigators required to optimize therapy according to IDF guidelines
o Outcome – primary – MACE = all-cause mortality, non-fatal MI, stroke, ACS, invasive CAD/PAD procedures, amputation.
o Outcome – secondary
▪ All-cause death, MI, stroke
▪ CVD death
▪ Individual components of the primary composite.
o Outcomes – adverse events
o Assuming a 6% annual primary event rate in the placebo group – 91% power for a 20% reduction in hazard.
- 5238 patients randomized – mean age = 61.8 years – median time of diabetes = 8 years – follow-up = 34.5 months = almost 3 years. Discontinued meds in 16% vs 17%. LTFU < 10% (really almost neglibile, in 2 patients).
- Results
o Endpoint – primary – MACE, PAD, and procedures – HR = 0.90 [0.80, 1.02], p = 0.095. Overall placebo proportion approximately 23%, 8% per year.
o Endpoint – secondary – MACE – HR = 0.84 [0.72, 0.98], p = 0.027. Overall placebo proportion approximately 14%.
o Endpoint – individual events – first occurrences – consistent benefits – HRs all NS.
▪ CV mortality – HR = 0.94 [0.74, 1.20]
▪ All-cause mortality – HR = 0.96 [0.78, 1.18]
▪ Non-fatal MI – HR = 0.83 [0.65, 1.06]
▪ Stroke – HR = 0.81 [0.61, 1.07]
▪ Etc.
o Other pre-specified outcomes
▪ MI – fatal/non-fatal – HR = 0.77 [0.60, 1.00], p = 0.046
▪ CV mortality – HR = 0.82 [0.70, 0.97], p = 0.020
o Use of drugs – insulin and metformin more frequent in the placebo group – otherwise similar distribution of medications (affects ability isolate the impact of pioglitazone).
o Body weight - +3.6kg vs -0.4kg (p < 0.0001).
o Laboratory (percent change, all values significant)
▪ HbA1c = -0.8% vs -0.3%
▪ TG = -11.4% vs +1.8%
▪ LDL = +7.2% vs +4.9%
▪ HDL = +19.0% vs +10.1%
▪ LDL/HDL ratio = -9.5% vs -4.2%
▪ Changes in microalbuminuria similar in groups, BP lowering slightly higher (SBP change different by 3mmHg) in pioglitazone treatment.
o Adverse events
▪ Fewer in the pioglitazone patients
▪ Heart failure – any report = 11% vs 8% (p < 0.0001)
▪ Heart failure – death – 1% vs 1%, NS
▪ Heart failure – requiring hospitalization – 5% vs 3% (p = 0.003).
▪ Any hypoglycemia – 28% vs 20%, p < 0.0001.
▪ Some imbalances in neoplasm’s, more bladder tumours, fewer breast cancers.
▪ More frequent edema without heart failure in patients on pioglitazone.
▪ Liver function – no acute toxicity.
o Disease related endpoints decreased, while procedure-related endpoints increased (e.g.: leg revascularization) or were similar (coronary revascularization).
- Over 3 years, the allocation of 1000 patients to pioglitazone would avoid 21 first MIs, strokes, or deaths (secondary MACE outcome) – NNT over 3 years = 48.
- Improvement arose above normal medical care, in a particularly ill group of patients.
- Mechanism for improvement?
o CVD biologic RF profile improved, despite no differences in lipid lowering therapy, etc.. However, HbA1c different despite treatment protocol, and some differences in hypoglycemic agents (appropriate to bring placebo HbA1c down is pioglitazone affects HbA2c). Uncertain what the mediator is, and to what extent these differences are due to pioglitazone alone.
o Reduced insulin use in patients on pioglitazone – an effect of medical care, attempting to reach HbA1c targets, or a physiologic beta-cell sparing effect, or increased insulin sensitivity requiring less insulin?
- Pioglitazone improves CVD outcomes and reduces the need to add insulin to glucose-lowering regimens compared with placebo in high-risk adults with type 2 diabetes.
Fonseca V, Jawa A, Asnani S. Commentary: the PROactive study – the glass is half full. J Clin Endocrinol Metab, 91(1): 2006.
- Generalizablity issues – mainly Caucasian subjects with a relatively low rate of statin use and a high rate of smoking – event rates higher than expected.
- Early end to the study due to achieving the projected number of events – may have reduced study ability to detect an emergent difference in K-M curves – curves were still diverging.
- “Main secondary endpoint” not reported in study protocol paper – use of a secondary endpoint if the primary is NS has been severely criticized (spending of alpha).
- Increase in CHF tempers benefit of pioglitazone. Notably, NYHA class II-IV excluded from study population, when NYHA class II is approved for pioglitazone use. Also, does was quite high in this study – higher than the approved dose – lower dose, different effects?
- Cancers – timeframe – known risk factors in history – improbable that the imbalance of tumors is related to pioglitazone.
- Ultimately, study is reassuring.
Betteridge DJ, DeFronzo RA, Chilton RJ. PROactive: time for a critical appraisal. European Heart Journal, 29: 2008.
- In addition to other pathophysiologic factors (e.g.: hyperglycemia, dyslipidemia, HTN), insulin resistance per se is an independent risk factor of atherosclerotic CVD.
- “Thiazolidinediones are the only glucose-lowering agents that improve insulin signaling/sensitivity in muscle and inhibit the MAP kinase pathway, while improving dyslipidemia, reducing elevated BG levels, inhibiting inflammation, enhancing endothelial function, and ameliorating the hypercoaguable state.”
- Individual thiazolidinediones have beneficial but differing effects on CVD RFs.
o Rosiglitazone – increases LDL and apo-B100, tends to raise TG
o Pioglitazone – LDL neutral, lowers apo B100, reduces plasma TG. More effective in raising HDL, converting LDL lipid subgractions to larger, more buoyant, particles.
- PROactive –underlying hypothesis – pioglitazone would benefit multiple vascular beds beyond standard guideline-driven therapy – hence the very broad primary outcome.
- Pre-defined main secondary endpoint pre-defined, yes, but not reported in the first design/baseline characteristics paper. Study stopped early – event driven.
- Sub-group analysis – main secondary MACE outcome.
o Statin use at baseline - HR = 0.97 [0.75, 1.26]
o No statin use at baseline – HR = 0.77 [0.64, 0.93]
o Statin use itself was associated with a decrease in the main secondary endpoint, HR = 0.84 [0.69, 1.00].
o However, the interaction did not reach statistical significance (p = 0.1547).
o Also NS for differences in patients with renal dysfunction.
- Sub-group analysis – PAD at baseline
o Leg revasc – HR = 1.68, p = 0.0077 in those with PAD at baseline.
- Sub-group analyses – previous CVD or stroke
o Significant risk reductions for recurrent fatal/non-fatal events (28% RRR for MI among those with previous MI; 47% RRR for stroke and reduction in composite of CV mortality, MI and stroke in those with previous stroke) in those treated with pioglitazone.
- Composite endpoints
o Used to increase power, especially where background therapy is gradually optimized – short study, fewer participants – reduced time and costs.
o However, uncertainty in interpretation if benefits are not seen for all components, if components are underpinned by different mechanisms, and if components vary in clinical importance, which may dilue the clinical relevance of the overall finding.
▪ CAD – dyslipidemia is major RF.
▪ Stroke – HTN is major RF.
▪ PAD – smoking and diabetes.
o PROactive endpoint
▪ Multiple vascular beds
▪ Disease related and less objective procedural endpoints.
▪ Inclusion of PAD and amputation in the primary composite appeared to reduce th impact of pioglitazone.
▪ The main secondary endpoint is more traditional and clinically important – this was declared before analysis, but after the protocol publication.
• Just the major hard events
• Facillitates comparisons with other trials
▪ Endpoint changes – uncommon, but not unusual.
o Key issue – how do we interpret the statistically significant main secondary endpoint in the context of an NS but consistent trend in the primary?
▪ Decision between primary and secondary can be somewhat artificial.
▪ Main secondary endpoint would have been a more appropriate choice as the primary outcome.
▪ “Since the main secondary endpoint represented the most clinically relevant components of the primary composite, this might provide an exception where the secondary endpoint might assume primacy over the primary endpoint.”
- Clinical significance
o NNT – primary outcome = 40 per year.
o Contrast simvastatin in the Heart Protection Study – NNT = 75 per year – major vascular events, including revasc and amputations.
o NNT – main secondary outcome (MACE), first occurrences = 143 per year.
o Contrast ramipril in HOPE, MACE – NNT = 91 per year.
o NNT – recurrent MI – patients with previous MI – 149 per year
o Contrast UKPDS metformin NNT in healthier individuals for fatal/non-fatal MI = 150/year.
o NNT – recurrent stroke – previous history of stroke – 63 per year.
o Contrast atorvastatin in SPARCL (Amarenco P, Bogousslavsky J, Callahan A III, Goldstein LB, Hennerici M, et al. NEJM, 355: 2006), NNT = 227 per year.
o The metabolic benefits of pioglitazone are highly clinically relevant – addition advantage of including a thiazolidinedione with other glucose-lowering options.
- Cardiovascular safetly
o Issues with rosiglitazone – Missen and Wolski SR, GSK retrospective analysis, other meta-analyses, RECORD, FDA K-M analysis for all available rosiglitazone data – increase in ischemic events. Nov. 14 black box warning.
o Similar FDA analysis for pioglitazone – significant decrease in CV events, HR = 0.83 [0.72, 0.95] (ohrms/dockets/ac/07/slides/2007-430821-08-fda-graham.ppt)
o Pio is not equal to Rosi.
- Good table of NNTs but results may have been cherry-picked.
Gaede P, Parving H-H, Pedersen O. Untitled correspondence. Lancet, 367: 2006. P.23.
- Cherry-picked outcome – okay, PAD should not have been included – then by the same reasoning, why stroke, since UKPDS showed that this variable did not respond as expected to glucose-lowering therapy? Notable that all the single outcomes were NS … only by combining stroke and MI did we get significance …
- The prediction of a relative risk reduction of 3mmHG systolic may actually account for the entire benefit of pioglitazone. Therefore the role of pioglitazone, in particular, in the therapeutic package for type 2 diabetes remains unsettled.
Homan RR, Retnakaran R, Farmer A, Stevens R. Untitled correspondence. Lancet, 367: 2006. P. 25.
- UKPDS Outcomes Model used to predict difference in 3-year risk of MACE given the differences in HbA1c, SBP, HDL, and body weight observed at PROactive follow-up. Relative risk reduction of 13% obtained – well within the CI of PROactive’s main secondary endpoint.
- Interestingly, the model predicted an 11% RR reduction in CHF, contrasting the 39% RR increase observed in PROactive.
- This analysis supports the explanation that any macrovascular benefits of pioglitazone reflect the modest improvements obtained in established risk factors, with little evidence that changes in novel risk factors affected by pioglitazone (e.g.: insulin resistance, inflammation, endothelial function) have any substantive effect.
- On the other hand, the increase in CHF appears to be specific to pioglitazone.
Home PD, Pocock SJ, Beck-Nielsen H, Curtis PS, Gomis R, Hanefeld M, Jones NP, Komajda M, McMurray JJV. Rosiglitazone evaluated for cardiovascular outcomes in oral agent combination therapy for type 2 diabetes (RECORD): a multicentre, randomized, open-label trial. Lancet, 373(9681): 2009.
- Objective – assess the cardiovascular safety (non-inferiority) of rosiglitazone in combination with metformin or sulfonylurea compared with metformin and sulfonylurea dual therapy.
- Prospective, multicenter, randomized, open-label trial
o Patients
▪ Type 2 diabetes on monotherapy with metformin or sulfonylurea, maximum tolerated dose, less than optimal blood glucose control (HbA1c > 7.0%)
▪ Exclusion: Recent major CVD, treatment for CHF.
o Interventions: Additional rosiglitazone vs additional metformin/sulfonylurea
▪ Additional rosiglitazone – rescue with a third oral agent or transfer to insulin at HbA1c >= 8.5%.
▪ Metformin + sulfonylurea dual therapy – transfer to insulin.
o Outcome measures
▪ Primary = first occurrence of CVD hospitalization or death.
▪ CVD = heart failure, TIA, unplanned revascularizations, etc.
▪ This includes procedures like pacemaker insertion.. A sensitivity analysis with this outcome minus the non-atherosclerotic events was performed (still includes revascularization).
o Design – non-inferiority – non-inferiority margin = 1.20 for the HR.
o Analysis
▪ HR estimated by CPH model stratified for background medications. Cumulative incidence by K-M method. ITT, no censoring for rescue therapy. Changes in HbA1c, etc., estimated using a GEE model with unstructured covariance matrix.
- May 20, 2007 – Nissen and Wolski report published – discontinuations from rosiglitazone therapy increased slightly – 32 people over the placebo group.
- Results
o 4458 patients randomized, 11 excluded subsequently (did not take study meds).
o Mean follow-up = 5.5 years. LTFU in 127 + 394 individuals < 20%.
o Laboratory
▪ HbA1c – significantly lower – study design called for targeting similar HbA1c (< 7.0%) in both groups – did not occur.
▪ Body weight – increased.
▪ HDL – increased.
▪ LDL – reduced less in the rosiglitazone arm.
▪ Results stratified by whether patients were baseline metformin or baseline sulfonylurea, reported in a table.
o Compliance – proportion of follow-up time on dual therapy – 75% and 83%.
o Use of other medications – statins (9% difference) and loop diuretics (5% difference) – greater in the rosiglitazone group than in the control group.
o Endpoint, primary – CVD hospitalizations or death – HR = 0.99 [0.85, 1.16].
o Endpoint, components – Equivalence standard not met. No significant differences detected, though – EXCEPT for CHF.
▪ MACE – HR = 0.93 [0.74, 1.15]
▪ E.g.: All cause death, HR = 0.86 [0.68, 1.08]
▪ E.g.: MI – HR = 1.14 [0.80, 1.63]
▪ E.g.: Stroke – HR = 0.72 [0.49, 1.06]
▪ CHF – fatal and non-fatal – HR = 2.10 [1.35, 3.27], p = 0.0010.
o Results robust within strata of background oral hypoglycemic agent – but non-equivalency, of course, could not be demonstrated due to decreased precision.
o The rosiglitazone arm had more heart-failure related deaths, but fewer stroke-related deaths and other vascular deaths (simple comparison of numbers).
o Sub-group analysis
▪ Previous ischemic heart disease – HR = 1.26 [0.95, 1.68], p = 0.055.
▪ Others NS, mostly around the null value.
o Adverse events – serious adverse events due to …
▪ Heart failure – 3.7% vs 1.9%, p = 0.0003.
▪ Hyperglycemia – 1.2% vs 2.5%, p = 0.0027
▪ Hypoglycemia – 0.7% vs 0.3%, p = 0.076.
▪ Malignancies NS, but pancreatic cancer 24 weeks of drug exposure, similar duration of treatment between groups.
- Exclusion: Studies without MI or CVD deaths in either group.
- ORs and 95% CIs calculated using the Peto method (small cell sizes).
- Sensitivity analysis with active control groups sub-grouped for metformin, sulfonylurea, insulin, and placebo.
- Results
o Relatively young patients (mean 1.0. Aside from the combined comparator drugs, all NS.
▪ Placebo – MI – OR = 1.80 [0.95, 3.39], p = 0.07.
▪ Placebo – CVD death – OR = 1.22 [0.64, 2.34], p = 0.55.
- Increased risk of MI and CVD worrisome given the large rate of CVD in patients with diabetes and the potential widespread exposure to rosiglitazone.
- Statistically consistent OR from smaller, shorter trials suggests that the excess in adverse effects may occur after relatively short exposures.
- Potential mechanisms
o Rise in LDL
o CHF
o Reduction in hemoglobin – increased physiological stress
o Biologic effects largely unknown, PPAR agonists have complex effects – activation or suppression of dozes of genes.
- Reasoning by analogy – muraglitazar – investigational dual PPAR-alpha and gamma agonist – increased adverse CVD events – Phase 2 and 3 testing – not approved, further development halted.
- More than 50 IND applications for novel PPARs have been filed, but no additional drugs have successfully reached the market in more than 6 years. Wide range of different and seemingly unrelated toxicities have emerged during development of other PPAR agents (El-Hage J. Peroxisome proliferator-activated receptor (PPAR) agonists: pre-clinical and clinical cardiac safety considerations. CDER, Rockville, MD: 2006).
o E.g.: cancers, rhabdomyoloysis, nephrotoxicity.
o E.g.: rare, but sometimes fatal liver toxicity in troglitazone.
Nissen SE, Wolski K. Rosiglitazone revisited: an updated meta-analysis of risk for myocardial infarction and cardiovascular mortality. Arch Intern Med: 2010.
- 56 trials, including 35531 patients.
- Fixed effects meta-analysis. Alternate analysis allowed inclusion of studies with zero counts in both active and control groups by pooling on allocation ratios.
- Including RECORD,
o MI - OR = 1.28 [1.02, 1.63], p = 0.04.
o CVD death – OR = 1.03 [0.78, 1.36], p = 0.86.
- Exclusion of RECORD did not affect conclusions, but estimates for ORs increased in magnitude to 1.39 and 1.46, respectively. Alternate pooling found results quite similar to main analysis.
- Eleven years after the introduction of rosiglitazone, the totality of RCTs continue to demonstrate increased risk for MI, although not for CV or all-cause mortality.
Diabetes – Diagnosis
Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care, 26(Suppl 1): 2003.
- Diabetes underestimated by using FPG alone instead of the full diagnostic criteria, which contains 2 other methods of diagnosis.
- Plasma glucose concentrations – continuum. Approximate threshold separating subjects at substantially increased risk for micro-vascular complications.
- Old cut-points (Pre-2003)
o FPG >= 7.8 mM
o 2h OGTT >= 11.1 mM
o Cut-point Venn diagram, the 2h OGTT standard would be much larger, and almost engulfs, the FPG standard. The FPG standard is considered to define a greater degree of hyperglycemia than the 2h OGTT standard – misses more cases, consequently. The diagnosis is a function of the clinical test used.
- New standards – ADA, 2003: FPG cut-point reduced, so that the areas of the Venn diagrams are more equivalent.
- Why the 2h OGTT standard = 11.1mM to begin with?
o Approximates the cut-point separating the two components of the bomodal distribution of 2h OGTT.
o Prevalence of microvascular complications increase dramatically – Pima Indians in the US, Egyptians, and NHANES III.
▪ Pima Indians – retinopathy, sharp rise (flat before) at OGTT >= 10.3 mg/dL, FPG >= 6.4 mM, and HbA1c >= 6.0% (converted from mg/dL by a factor of 18) (McCance DR, Hanson RL, Chales MA, Jacobsson LTH, Pettitt DJ, Bennet PH, Knowler WC. BMJ, 308: 1994).
▪ Egypt – FPG and 2h OGTT strongly and equally associated with retinopathy – prevalence markedly higher at 2h OGT = 11.5mM, FPG = 7.2mM, and HbA1c >= 6.0% (Engelgau MM, Thompson TJ, Herman WH, Boyle JP, Aubert RE, Kenny SJ, Badran A, Sous ES, Ali MA. Diabetes Care, 20: 1997).
▪ NHANES III – 2821 individuals aged 40-74 years received an OGTT, a measurement of HbA1c, and an assessment of retinopathy by fundus photography. Dramatic rise in prevalence at FPG >= 6.7 mM, 2h OGTT >= 10.8 mM, and HbA1c >= 6.2% (Personal communication referenced).
o Value retained because it would have been disruptive to change it – enormous body of clinical and epidemiological data already collected.
- Why lower the FPG from 7.8mM to 7.0mM?
o The FPG and 2h OGGT should diagnose similar conditions, given the similarities of their associations for vascular complications.
o The value of FPG that is equivalent to the 2h OGTT value for distinguishing prevalent retinopathy (ROC analysis), and for distinguishing an equal prevalence of diabetes per the 2h OGTT standard, ranges from 6.7 to 7.9mM.
o The FPG is important because it is more convenient in routine care, and because the repeat test reproducibility of the FPG is higher than that of the 2h OGTT (Mooy JM, Gootenhuis PA, De Vries H, Kostense PJ, Popp-Snijders C, Bouter LM, Heine RJ. Diabetologia, 39: 1996).
- What about HbA1c?
o HbA1c has become the measurement of choice in monitoring the treatment of diabetes, and showed similar threshold characteristics RE: prevalent retinopathy, in the same studies from which the 2h OGTT standard was derived.
o However, the HbA1c test is not yet nationally standardized. HbA1c should continue to be a treatment goal.
- Problems
o Because of new diagnostic criteria, the prevalence of diabetes will fall. This is because the FPG test is now able to substitute for the OGTT, but because a cut-point on the upper range of diabetes prevalence equivalency was chosen, some members of the population with OGTT >= 11.1 will not be captured. The drop in prevalence, using NHANES III data, is 14% (relative decrease). The committee felt that this would be counterbalanced by increased numbers of diagnoses under simplified criteria.
Shaw JE, Zimmet PZ, George K, Alberti MM. Point: impaired fasting glucose: The case for the new American Diabetes Association criteria. Diabetes Care, 29 (5): 2006.
- Why is 5.6 mM the FPG cut-point distinguishing normal from IFT?
o The previous IFG cut-point was >= 6.1mM, but was based on rather arbitrary artifacts of the evidence. The IFG range was narrow, leading to unstable classifications of IFG. Also, IFG was not useful as a screener for OGTT failure, because over 20% of those with undiagnosed diabetes based on an OGTT have FPG < 6.1mM. Similarly, IFG did not identify most patients with IGT, because the median FPG among patients with IGT is 5.4 mM.
o Unlike IGT, IFT is not an independent risk factor for total or CVD mortality. It is the FPG value that came the closest to providing 100% sens and spec for future diabetes. This was also demonstrated in predominantly Asian populations by Shaw et al. and Ko et al. – the particular value reported was 5.6mM (Shaw JE, Zimmet PZ, Hodge AM, de Courten M, Dowse GK, Chitson P, et al. Diabetes Care, 23: 2000) (Ko GT, Chan JC, Yeung VT, Chow CC, Tsang LW, Li JK, et al. Diabetes Care, 21: 1998). However, it has also been shown that there is no natural orobvious cut-point.
o The cut-point was the confirmed by work of the ADA Expert Committee on the Diagnosis and Classification of Diabetes, in a critical review of available IFG data. The optimal cut-points were 5.7mM in a Dutch population, 5.4mM in a Pima-Indian population, 5.2 mM in a Mauritius population, and 5.2mM in a San Antonio population (Genuth S, Alberti KG, Bennett P, Buse J, Defronzo R, Kahn R, et al. Diabetes Care, 26: 2003).
- IFG may be used as a screen for potential IGT or even diabetes according to an OGTT.
- We now know that diabetes can be prevented by treating patients with IGT (but studies did not use patients with IFT).
- However, the lowering of the threshold, of course, created an apparent “pandemic” of IFG in 2003.
Forouhi NG, Balkau B, Borch-Johnsen K, Dekker J, Glumer C, Qiao Q, Spijkerman A, Stolk R, Tabac A, Wareham NJ. The threshold for diagnosing impaired fasting glucose: a position statement by the European Diabetes Epidemiology Group. Diabetologia, 49 (5): 2006.
- Enlarging the range of FPG values considered IFG reduces the category’s association with incident diabetes, taken as the magnitude of the RR.
- Except for Pima Indians (threshold effect at FPG 5.6mM for incident diabetes), most populations do not have a risk threshold of incident diabetes across FPG values.
- What about relation to mortality or CVD events?
o DECOD – 22 Europeann cohorts – nearly 30000 participants – no glycemic threshold for either fasting or 2h OGTT above which total mortality increased sharply. Low and high FPG associated with total and CVD death in a J-shaped relationship, but only a graded association for 2h OGTT.
o Other studies showed a similar gradient.
o One meta-analysis showed a threshold effect with FPG at around 5.6mM, but heterogeneity issues.
o Expanding the definition of IFG, in fact, reduced the predictive value of IFG for CVD and diabetes mortality in several studies.
- But IGT has been related to CVD outcomes …
o IFG not the same as IGT – intima-media thickness studies show significant differences. Also differences in CVD risk factors.
- Could treating IFT reduce diabetes or CVD?
o Evidence applies only to IFG. Single study relevant to IFG did not measure progression to diabetes. Reinforced healthy lifestyle advice in this population did not achieve sustained lifestyle changes, measured by weight and reduction in glycemia at 1 year follow-up, compared with basic healthy lifestyle advice.
o IFG – linked to metabolic syndrome features – but these may be managed on their own.
- Changing the IFG cut-point caused a pandemic of “pre-diabetes”.
o 2 to 5-fold increase in most populations.
o What are the consequences of this for individuals and health care systems?
- The lower bound of IFG is contentious, and the European Diabetes Epidemiology Group argues, instead, for an unchanged lower bound to maximize specificity – identifying and labeling those at higher risk does not necessarily ensure that their health will be improved.
- The decision to draw any cut-off point on a continuous scale should relate the relative harm from being a false positive, compared with a false negative.
- “Non-diabetic hyperglycemia” as a term is preferable to “dysglycemia”, “IFG”, and “pre-diabetes”.
ADA. Diagnosis and classification of diabetes mellitus. Diabetes Care, 31(Suppl 1): 2008.
- Basis of abnormalities in carbohydrate, fat, and protein metabolism in diabetes is deficient action of insulin on target tissues, whether due to inadequate secretion or diminished tissue responses.
- Symptoms of marked hyperglycemia
o Polyuria
o Polydupsia
o Weight loss, sometimes with polyphagia
o Blurred vision
o Impaired growth and susceptibility to certain infections
o Ketoacidosis or nonketotic hyperosmolar syndrome
- Type 1
o 5-10%
o Islet cell autoantibodies, autoantibodies to insulin, to glutamic acid decarboxylase (GAD65), and to the tyrosine phosphatases IA-2 and IA-2beta.
o Strong HLA associations
o C-peptide loss
o Link to other autoimmune disorders – Grave’s disease, Hashimoto’s thyroiditis, Addison’s disease, vitiligo, celiac sprue, autoimmune hepatitis, myasthenia gravis, and pernicious anemia.
- Idiopathic
o Like type 1, but no evidence of autoimmunity
o Strongly inherited, but no HLA association.
- Type 2
o 90-95%
o Most are obese; obesity may cause insulin resistance on its own.
o Older than in type 1, C-peptide preserved. Ketoacidosis rare.
o Physiologic (though possibly not an absolute) deficiency in insulin production given the glucose level.
o Strong genetic predisoposition.
- Genetic defects of the beta-cell (MODY) – autosomal dominant inheritence.
- Genetic defects in insulin action
o Associations with acanthosis nigricans; virilization and large cystic ovaries (women).
o Leprechaunism and Rabson-Mendenhall syndrome – kids – mutations in the insulin receptor gene
- Diseases of the exocrine pancreas
o VINDICATE (D = drugs, E doubles for endocrine or environmental exposure) or VITAMIN D (missing congenital)
o Damage to the pancreas must be extensive for diabetes to occur, except for adrenocarcinomas, which only have to involve a small portion of the pancreas to cause diabetes.
- Endocrinopathies
o GH, cortisol, glucagon, and epinephrine antagonize insulin action.
o Inhibition of insulin secretion by somatostatinomas or aldosteronomas.
- Drugs
o More likely to trigger diabetes in patients with insulin resistance than cause frank diabetes by themselves. Exception = Vacor (a rat poison) and IV pentamidine – permanent beta-cell death.
o May inhibit secretion (see poisons, above)
o May inhibit action
▪ Nicotinic acid
▪ Glucocoritcoids
▪ Etc.
- Infections – congenital rubella, coxsackievirus B, cytomegalovirus, adenovirus, and mumps (rare).
- Uncommon immune-mediated diabetes
o E.g.: stiff man syndrome – high tiers of GAD autoantibodies.
o E.g.: SLE
- Other syndromes (congenital) – association, cause unclear, syndromes accompanied by increased incidence of diabetes.
o Down’s syndrome
o Klinefelter’s syndrome
o Turner’s syndrome
o Wolfram’s syndrome
o Etc.
- Gestational (GDM)
o 4% of all pregnancies in the US
- Diagnostic cut-points
o Normal – FPG < 5.6mM, 2h OGTT < 7.8
o IFG – FPG >= 5.6 mM but < 7.0 mM
o IGT – 2h OGTT >= 7.8 mM but < 11.1 mM
o Diabetes – FPG >= 7.0 mM or 2h OGTT >= 11.1 mM
- Diagnostic criteria – requires confirmation on a subsequent day
o By FPG
o By casual PG >= 11.1 mM with symptoms of hyperglycemia.
o By 2h OGTT with 75g anhydrous glucose dissolved in water.
................
................
In order to avoid copyright disputes, this page is only a partial summary.
To fulfill the demand for quickly locating and searching documents.
It is intelligent file search solution for home and business.
Related searches
- words not found in dictionary
- zsh command not found python3
- file not found python
- npm command not found linux
- command pip not found ubuntu
- pip3 not found windows
- suspected condition not found icd 10
- concern not found icd 10
- pip not found linux
- university of alberta directory
- pycharm module not found error
- react router not found route