The impact of cardiovascular disease on people with diabetes: the potential for prevention

Abstract

The welcome reductions in recent years of the specific complications of diabetes mellitus—ie, blindness from diabetic retinopathy, end-stage renal failure, and lower-limb amputation—emphasise and may even contribute to the high residual burden of cardiovascular disease in these patients. Most of this burden is coronary heart disease, with an age-adjusted prevalence in people with diabetes in the USA between 30% and 51%.1Wingard DL Barrett-Connor E Heart disease and diabetes.in: National Diabetes Data Group Diabetes in America. 2nd edn. NIH, Bethesda1995: 429-448Google Scholar By contrast, the age-adjusted prevalences for peripheral vascular disease and stroke are 9% and lo%, respectively. The increased cardiovascular risk from diabetes is likely to be multifactorial in origin, due partly to the increased frequency of known risk factors for cardiovascular disease, partly to the larger impact of any given risk factor in people with diabetes, and partly to the additional risk factors specific for diabetes. The approach to coronary heart disease prediction, protection, and prevention for people with diabetes is correspondingly multifactorial. In most societies, diabetes increases the risk of coronary heart disease between two and four times over the prevailing population rates, and large variations between countries in the prevalence and incidence of coronary heart disease are consistently reflected in their diabetic subpopulations. For example, in the lo-year follow-up from the WHO Multinational Study of Vascular Disease in Diabetes,4Fuller JH Stevens LK Wang SL International variations in cardiovascular mortality associated with diabetes mellitus: the WHO Multinational Study of Vascular Disease in Diabetes.Ann Med. 1996; 28: 319-322Google Scholar the age-adjusted coronary-heart-disease mortality rates (International Classification of Disease codes 410–414) for Japanese people with non-insulin-dependent diabetes (NIDDM) in Tokyo were 0·0 and 0·3 per 1000 person-years for men and women, respectively. The corresponding and significantly different rates for people with NIDDM in London, UK, were 9·9 and 4·5, and in Switzerland, 7·1 and 3·3. Thus, although the ‘metabolic’ component of coronary heart disease in diabetes—call it the doubling effect—seems to be fairly constant, the ‘geographic’ component is probably largely determined by the environment and is highly variable. This latter effect can be seen when ethnic groups migrate from regions with a low prevalence of coronary heart disease to areas with a higher prevalence—for example, Japanese people with NIDDM living in Hawaii have a rate of coronary heart disease similar to the indigenous Hawaiian population. Similar findings from other populations suggest that these environmental factors are, to some extent, potentially preventable, reducible, or reversible. Longitudinal observations on non-diabetic and diabetic western populations, such as the Multiple Risk Factor Intervention Trial,5Stamler J Vaccaro O Neaton JD Wentworth D Diabetes, other risk factors, and 12-yr cardiovascular mortality for men screened in the Multiple Risk Factor Intervention Trial.Diabetes Care. 1993; 16: 434-444Google Scholar provide crude estimates of the potential impact on coronary-heart-disease mortality from correction of the metabolic effect, the environmental effect, or both. The implications of effective interventions are, however, complex. In many western societies, greater cardiovascular health gains are likely to be achieved by reversal of the environmental (general) risk factors. These reductions would have less effect on the indigenous Japanese person with diabetes who is already at a substantially lower risk of coronary heart disease than his or her non-diabetic European counterpart. The relation between specific and non-specific determinants of mortality in diabetes is dominated by the incremental risk of cardiovascular disease associated with diabetic renal disease. Improved glycaemic control alone could substantially delay the progression of diabetic renal failure but thereby extend the time during which there is an increased risk of coronary heart disease. Conversely, a reduction of coronary-heart-disease mortality would prolong the time the patient is vulnerable to the risk of diabetic renal disease. In one cohort of patients with insulin-dependent diabetes mellitus (IDDM), a 37-fold excess in cardiovascular mortality was found in patients with proteinuria, whereas non-proteinuric patients had only a four-fold excess mortality compared with the general population.6Borch-Johnsen K Kreiner S Proteinuria: value as predictor of cardiovascular mortality in insulin dependent diabetes mellitus.BMJ. 1987; 294: 1651-1654Google Scholar Proteinuria at baseline, in patients with NIDDM in the WHO Multinational Study of Vascular Disease in Diabetes,4Fuller JH Stevens LK Wang SL International variations in cardiovascular mortality associated with diabetes mellitus: the WHO Multinational Study of Vascular Disease in Diabetes.Ann Med. 1996; 28: 319-322Google Scholar increased the risk of cardiovascular death three to four times—an effect independent of other major risk factors for mortality.7Stephenson JM Kenny S Stevens LK Fuller JH Lee E Proteinuria and mortality in diabetes: the WHO Multinational Study of Vascular Disease in Diabetes.Diabet Med. 1995; 12: 149-155Google Scholar In the NIDDM population, microalbuminuria (urine albumin excretion rates between 30 and 300 mg per day) has emerged as a powerful risk factor for both total and cardiovascular mortality.8Mattock MB Morrish NJ Viberti GC Keen H Fitzgerald AP Jackson G Prospective study of microalbuminuria as predictor of mortality in NIDDM.Diabetes. 1992; 41: 736-741Google Scholar Microalbuminuria is an independent risk factor but interacts with arterial blood pressure, dyslipidaemia, smoking habits, and other risk factors, and heralds a general increase in vascular vulnerability, and perhaps permeability, in diabetes. A multivariate analysis of coronary-heart-disease mortality in NIDDM found microalbuminuria to be the leading predictor with an odds ratio (OR) of 10·02 (95%CI 4·28–23·47), followed by smoking (OR 6·52), diastolic blood pressure (OR 3·20), and serum cholesterol (OR 2·32).9Mattock MB Keen H Barnes DJ et al.Microalbuminuria: a risk factor for coronary heart disease in non-insulin dependent diabetic men.in: Cardiovascular Disease Prevention III. Hampton Medical Conferences, Teddington, UK1997: 30Google Scholar Antihypertensive therapy can decrease urinary albumin excretion but, as yet, there is no strong evidence of a concomitant decrease in the associated cardiovascular risk.10Marre M Fabbri P Berrut G Bouhanick B The concept of incipient diabetic nephropathy and effect of early antihypertensive intervention.in: Mogensen CE The kidney and hypertension in diabetes mellitus. 3rd edn. Kluwer Academic Publishers, London1996: 351-360Google Scholar We explored the preventive potential of interventions on both general and diabetes-specific risk factors in a multivariate simulation model based on data from the Framingham study11Anderson KM Odell PM Wilson PWF Kannel WB Cardiovascular disease risk profiles.Am Heart J. 1990; 121: 293-298Google Scholar and a chronic disease model of NIDDM developed by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDI<).12Eastman RC Javitt JC Herman WH et al.Model of complications of NIDDM. I. Model construction and assumptions.Diabetes Care. 1997; 20: 725-734Google Scholar The lo-year risk of cardiovascular disease, which includes myocardial infarction, angina pectoris, congestive heart failure, stroke, transient ischaemic attacks, and peripheral vascular disease, is calculated for people with and without diabetes and with single or multiple risk factors. Variables included in the model were age, sex, systolic arterial pressure, total serum cholesterol and high-density-lipoprotein (HDL) cholesterol, and smoking status (yes or no). Two values of systolic blood pressure (120 and 140 mm Hg) and total cholesterol (6 and 7 mmol/L) are used—the same as those used in Pyörälä's model of cardiovascular-disease prevention.13Pyörälä K CHD prevention in clinical practice.Lancet. 1996; 348: s26-s28Summary Full Text Full Text PDF Scopus (26) Google Scholar An average HDL cholesterol (1·25 mmol/L for women and 1·04 mmol/L for men) based on data from people with diabetes in the NHANES-II study,14Cowie CC Harris MI Physical and metabolic characteristics of persons with diabetes.in: National Diabetes Data Group Diabetes in America. 2nd edn. NIH, Bethesda1995: 117-164Google Scholar is also entered. The Framingham model is incorporated into a computer simulation model of NIDDM developed by the NIDDK that is similar to the model of the effects of glycaemic control on IDDM derived from the Diabetes Control and Complications Trial.15DCCT Study Group. Lifetime benefits and costs of intensive therapy as practiced in the Diabetes Control and Complications Trial: an economic evaluation. JAMA (in press).Google Scholar Hypothetical patients are created by sampling demographic profiles of people diagnosed with NIDDM each year in the USA. The average age at diagnosis is 51 years, with equal numbers of men and women, and 70% are non-Hispanic white.16Kenny SJ Aubert RE Geiss LS Prevalence and incidence of noninsulin dependent diabetes.in: National Diabetes Data Group Diabetes in America. 2nd edn. NIH, Bethesda1995: 47-68Google Scholar The cardiovascular-disease risk profile is simulated by values randomly drawn from probability distributions (specific for age, sex, and ethnic group) for systolic blood pressure, total and HDL-cholesterol, and smoking, based on data from NHANES II, HHANES, the Strong Heart Study of Native Americans, and Japanese American studies.14Cowie CC Harris MI Physical and metabolic characteristics of persons with diabetes.in: National Diabetes Data Group Diabetes in America. 2nd edn. NIH, Bethesda1995: 117-164Google Scholar People with diabetes are also at increased risk of non-cardiovascular-disease mortality owing to infectious and metabolic complications.17Moss SE Klein R Klein BEF Cause-specific mortality in a population-based study of diabetes.Am J Public Health. 1991; 81: 1158-1162Google Scholar Between the ages of 25 and 65, there is a 3·5-fold increase in the age-adjusted risk of death from influenza and pneumonia.18Geiss LA Thompson TJ Are persons with diabetes more likely to die from influenza and pneumonia.Diabetes. 1995; 44 (abstr): 456Google Scholar Infections and metabolic complications account for about 50% of the deaths in diabetic renal failure.19US Renal Data System 2nd edn. USRDS 1994 Annual Data Report. National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda1994Google Scholar Non-cardiovascular-disease mortality is important because of its direct effect on mortality and its indirect effect on cardiovascular-disease outcomes through changes in the life expectancy of people with NIDDM. In our model, non-cardiovascular-disease mortality is first calculated for people without diabetes by subtraction of the average cardiovascular disease mortality in the Framingham model from the total-mortality risk in the US population.20Kochanek KD Hudson BL Advance report of final mortality statistics, 1992.Monthly Vital Statistics Report. 1994; 43: 17Google Scholar A relative risk of 2·75 is applied to non-cardiovascular-disease mortality for people with diabetes: this was the mean standard-mortality ratio weighted for cause of non-cardiovascular-disease death in men and women diagnosed with diabetes at an older age in the Wisconsin Epidemiologic Study of Diabetic Retinopathy.17Moss SE Klein R Klein BEF Cause-specific mortality in a population-based study of diabetes.Am J Public Health. 1991; 81: 1158-1162Google Scholar The result of this calculation is a lo-year decrease in life expectancy for people with diabetes— similar to that observed in populations with diabetes.21Geiss LS Herman WH Smith PJ Mortality in non-insulin dependent diabetes.in: National Diabetes Data Group Diabetes in America. 2nd edn. NIH, Bethesda1995: 233-258Google Scholar Our analysis cannot account for the survival benefit of smoking cessation on non-cardiovascular-disease mortality, which we did not model. Table 1 shows the estimated effect of diabetes, alone and in combination with other risk factors, on the absolute 10-year risk of a cardiovascular-disease event in men and women aged 50 years. For any given value of a risk factor, the incremental risk with diabetes is much greater for women than for men. Diabetes alone confers a 10-year risk of a cardiovascular-disease event of 16·4% and 14·6% for men and women, respectively; this increases to 38·5% and 35·6% for smokers with hypercholesterolaemia and the higher systolic blood pressure (140 mm Hg). Overall, the risks are 1·5 to 2·4 times greater than in people without diabetes. For some outcomes (eg, fatal coronary heart disease), the relative risk is much greater for those with diabetes—especially women—although the absolute risk of this event at age 50 is low.Table 1Absolute 10-year risk of cardiovascular events for an individual 50 years-old calculated from the Framingham multivariate modelMaleFemaleMaI.FemaleMaleFemaleMaleFemaleMaleFemaleVariablesSmoking (0=No, 1=Yes)0000001111Systolic blood pressure (mm Hg)120120120120140140120120140140Total cholesterol (mmol/L)6677666677HDL-cholesterol (mmol/L)1·041·251·041·251·041·251·041·251·041·25Without diabetesStroke or TIA0·9%0·6%0·9%0·6%1·5%1·1%1·6%1·1%2·8%2·1%Myocardial infarction3·6%1·2%4·7%14%5·4%2·1%9·5%4·5%15·0%8·2%Coronary heart disease (CHD)8·8%4·9%10·8%6·2%11·3%6·5%14·1%8·5%20·5%13·2%Cardiovascular disease (CVD)9·9%6·2%11·5%7·3%14·3%9·4%18·7%12·7%27·8%20·1%Fatal CHD1·2%0·1%1·6%0·2%2·0%0·3%2·6%0·5%5·2%1·2%Fatal CVD1·1%0·3%1·4%0·4%2·0%0·6%2·2%0·7%4·4%1·6%With diabetesStroke or TIA1·4%1·5%1·4%1·5%2·5%2·7%2·5%2·7%4·5%4·9%Myocardial infarction6·5%4·5%8·2%5·8%9·2%6·6%14·5%11·1%20·9%17·0%Coronary heart disease (CHD)12·0%10·2%14·4%12·3%15·1%12·9%18·4%16·0%25·6%22·7%Cardiovascular disease (CVD)16·4%14·6%18·6%16·6%22·4%20·1%27·9%25·3%38·5%35·6%Fatal CHD1·5%1·1%2·1%1·6%2·6%2·0%3·3%2·6%6·3%5·2%Fatal CVD1·6%1·3%2·0%1·6%2·7%2·2%3·1%2·5%5·9%4·9%TIA=transient ischaemic attack: myocardial Infarction includes silent and unrecognised events; CHD=myocardial infarction, angina pectoris, coronary insufficiency, and death from coronary heart disease; fatal CHD includes sudden and non-sudden death: CVD includes all of the above plus congestive heart failure and peripheral vascular disease: fatal CVD includes all deaths from CVD. Open table in a new tab TIA=transient ischaemic attack: myocardial Infarction includes silent and unrecognised events; CHD=myocardial infarction, angina pectoris, coronary insufficiency, and death from coronary heart disease; fatal CHD includes sudden and non-sudden death: CVD includes all of the above plus congestive heart failure and peripheral vascular disease: fatal CVD includes all deaths from CVD. The putative effects of interventions on life expectancy and various cardiovascular disease outcomes after diagnosis of diabetes at age 51 are shown in table 2. In this simulation model, the median survival is 18·3 and 20·0 years with standard and intensive glycaemic control, compared with 29·3 years for those entered in the model at the same age without diabetes. The correction of a single risk factor—stopping smoking, a decrease in either systolic pressure (to 120 mm Hg) or total cholesterol (to 6 mmol/L or less), or an increase in HDL cholesterol to the non-diabetic reference range—each reduces the predicted cardiovascular disease risk by a small amount. However, the treatment of all risk factors simultaneously with standard diabetes care eliminates about half the excess cardiovascular-disease risk—about the same effect as the elimination of the diabetes factor from the Framingham cardiovascular-disease model. Thus, diabetes per se seems to account for about half the excess cardiovascular-disease morbidity associated with diabetes.Table 2Model predictions of survival and life-time cumulative incidence of cardiovascular events after diagnosis of diabetesYears of life remainingStrokeMyocardfal infarctionCHDCVDFatal CHDFatal CVDStandard care18.35.0%14.3%27.1%47.2%6.3%16.2%Stop smoking18.55.0%11.9%25.3%43.5%5.2%14.1%Treat SBP to <120 mm Hg18.42.1%12.0%24.6%41.4%4.4%14.1%Treat total serum cholesterol to <5.2 mmol/L18.35.4%13.4%26.7%4%25.1%6.0%15.0%Increase HDL cholesterol to non.diabetic range18.55.4%13.4%25.1%44.9%2.6%15.0%Do all of the above18.52.2%11.2%19.7%35.9%2.6%12.6%Do all of the above and eliminate excess non.CVD mortality risk24.33.7%11.2%28.1%50.6%4.1%20.3%Standard care with diabetes coefficients eliminated from CVD model18.52.6%7.5%20.3%34.0%3.2%13.3%Intensive care20.26.1%15.5%26.0%45.0%7.1%9.7%All risk factors treated as above20.22.7%8.8K18.1%32.7%2.9%5.4%All risk factors treated as above and excess non.CVD mortality rate eliminated28.15.1%12.8%27.1%47.6%5.3%9.8%Intensive care with diabetes coefficients eliminated from CVD model20.33.0%7.8%18.8%29.7%3.8%6.0%Person without diabetes29.32.3%5.7%18.1%27.3%2.5%3.6%SBP=systolic blood pressure; CVD=cardiovascular disease; CHD=coronary heart disease. Open table in a new tab SBP=systolic blood pressure; CVD=cardiovascular disease; CHD=coronary heart disease. Although fatal coronary heart disease and fatal cardiovascular disease are favourably affected by multiple risk-factor correction or the elimination of diabetes-specific risk, the predicted effect on life expectancy is only 0·2 years (table 2) irrespective of whether patients are modelled under standard or intensive glycaemic control. The survival of people with diabetes approaches that of non-diabetic people only when cardiovascular-disease risk is reduced by both multiple risk-factor intervention and the elimination of excess mortality from non-cardiovascular disease (table 2). Under these conditions, the life-expectancy after diagnosis is 24·3 years from standard care and increases to 28·1 years if the risk from renal failure is also reduced by intensive care. Survival does not quite reach the non-diabetic expectation because even under intensive care 5% of people with diabetes are predicted to develop renal failure. The elimination of the excess non-cardiovascular-disease risk from the model has a paradoxical, but not unexpected, effect on morbidity and mortality. The predicted cumulative incidence of cardiovascular disease is greater and the proportion of deaths due to fatal cardiovascular disease increases (table 2 and figure)—despite the assumption of multiple risk-factor treatment. This outcome is the result of a longer exposure to cardiovascular-disease morbidity and mortality risk and is a realistic prospect if vaccination programmes reduce infectious-disease mortality, improved glycaemic control reduces non-cardiovascular-disease renal mortality, and gains from patient and professional education programmes reduce metabolic deaths in people with diabetes. Our simulation analysis confirms and, to an extent, quantifies the clinical conviction that the greatest survival benefit for people with diabetes will be obtained from aggressive correction of smoking, dyslipidaemia, and high arterial pressure combined with intensive glycaemic control and protection against non-cardiovascular-disease threats to life. The analysis also warns that, unless the diabetes-specific risk of cardiovascular disease, either direct or that associated with renal failure, is reduced, the morbidity of cardiovascular disease may actually increase. If non-cardiovascular-disease mortality alone is reduced, cardiovascular disease is predicted to increase from the interaction of time, glycaemia, and cardiovascular-disease risk factors.