Improved survival in both men and women with diabetes between 1980 and 2004 – a cohort study in Sweden

Cardiovascular Diabetology Original investigation Improved survival in both men and women with diabetes between 1980 and 2004 - a cohort study in Sweden Mats Eliasson 2 3 Mats Talbck 1 Mns Rosn 0 2 0 Swedish Council on Technology Assessment in Health Care , Stockholm , Sweden 1 Centre for Epidemiology, National Board of Health and Welfare , Stockholm , Sweden 2 Department of Public Health and Clinical Medicine, Umea University , Umea , Sweden 3 Department of Internal Medicine, Sunderby Hospital , Lulea , Sweden Background: In Sweden, diabetes prevalence is increasing in spite of unchanged incidence, indicating improved survival. In recent US studies mortality in diabetic subjects has decreased over three decades, but only in men. Our aim was to study mortality over time in diabetic subjects. Methods: The annual Swedish Living Conditions Survey from 1980 to 2004 has been record-linked to the Cause of Death Register in order to study trends in mortality risk for those reporting diabetes as a chronic illness. Survival and the relative mortality risk within 5 years of follow-up have been calculated for a random sample of men and women aged 40-84 years with (n = 3,589) and without diabetes (n = 85,685) for the period 1980 to 2004. Poisson regression models were used. Results: The age-adjusted mortality risk relative to non-diabetics within 5 years of follow-up for men was doubled during all periods. The relative risk for women was initially about 2.5, with a substantial drop in mortality in 1995-1999 to 1.45 although it increased to 1.90 in the last period. Using models that took into consideration the presence of heart disease, hypertension, daily smoking, and socio-economic status at the initial interview did not change the relative mortality risk. The age-adjusted 10-year observed survival rate for men with diabetes increased from 41.4% 1980-1984 to 51.5% in 1995-1999. The observed survival for women increased from 43.7% to 61.0%. Conclusion: Survival rates have improved in subjects with diabetes since the early 1980s, more so in women than in men, thereby decreasing the gap to non-diabetic women. - Background Patients with diabetes have a markedly increased mortality which arises mainly from cardiovascular disease and end-stage renal disease [1-3]. The classical risk factors for cardiovascular disease, smoking, hypertension, and hypercholesterolemia, contribute strongly to the mortality [4]. Treatment methods for diabetes focus on improved glucose control and cardiovascular prevention. These methods have undergone several changes during the last decades, and international guidelines have been established. There is a consensus among researchers that diabetes prevalence is increasing. The debate is rather whether this is due to an increasing age-specific prevalence due to an increase in incidence or to a longer longevity among diabetics [5] or whether a higher total prevalence can solely be ascribed to an aging population. Recent Swedish population studies do not support a higher incidence during the last 30 years [6,7] which corroborates pharmaco-epidemiological data from the Danish island of Fyn [8]. In Sweden, a higher prevalence over time seems restricted to the elderly [9]. Trends in all causes of death in Sweden where diabetes is either reported as an underlying or contributory cause of death show a clear sex difference, i.e. a small increase in diabetes mortality for men, but a steady decrease in mortality for women between 1970 and 2004 [9]. Conflicting results have been published concerning diabetes survival. Studies from the US and England have shown a clear improvement in diabetes survival up to 1999 [10,11]. A recent report from the National Health and Nutrition Examination Survey (NHANES) in the US showed that among diabetic men all-cause mortality decreased by more than forty percent between 1971 and 2000, accompanying the decrease in the non-diabetic population [12]. However, among women with diabetes, no decline in mortality was found, and that lead to a doubling of the mortality rate difference between diabetic and non-diabetic women. A recent report from the population based HUNT study in Norway described similar improvements in mortality in coronary heart disease among subjects with or without diabetes between the mid 1980's and the mid 1990's with a persistent doubled mortality risk in diabetic subjects [13]. The aim of this study was to report trends in diabetes survival in Sweden from 1980 to 2004 and to analyse whether these trends were different for the sexes and whether they were influenced by changes in risk factor patterns or by socio-economic status. Methods Statistics Sweden has conducted annual surveys since 1975 of the living conditions of a random sample of about 7,000 people aged 1684 years of age [14]. These surveys collect individual data on health, lifestyle, social and economic conditions, etc. It is possible to extract information from these surveys for those reporting diabetes as a chronic illness or if they regularly take medicines for diabetes. We linked records from the Living Conditions Surveys from 1980 to 2004 to corresponding records in the national Cause of Death Register from 1980 to 2004. This allowed us to compare survival among those reporting diabetes with survival among the rest of the survey sample (the general population) adjusting for heart disease, hypertension, smoking, and socio-economic status at the time of the interview (baseline). We restricted the analysis to the age group 4084 years at the initial interview in order to obtain a more homogenous group of diabetes patients, mainly patients with Type 2 diabetes. The results presented here are based on 89,274 interviews: Table 1 presents baseline characteristics concerning the subjects. The average age of those reporting diabetes in the surveys was 65 years for men and 69 years for women and did not change over time. We used Poisson regression models with mortality as the dependent variable to analyse the relative mortality risk within 5 years of follow-up among diabetics and non-diabetics, for 5-year time periods from 19801984 to 2000 Table 1: Base-line characteristics for subjects reporting diabetes as a chronic illness in the Swedish Living Conditions Survey 1980 2004. CHD or hypertension1 among diabetics CHD or hypertension1 among non-diabetics Daily smokers among diabetics Daily smokers among non-diabetics Number of observations (No.) and percentages (%). 1. ICD-9: 401 429 Hypertension and coronary heart disease. 19801984 No % 19851989 2004. The data for men and for women were analysed separately, and we included adjustments for age, reported heart disease and hypertension (ICD-9: 401429), daily smoking, and socio-economic status level at the time of the interviews. Results Multivariate analyses showed that the diabetics had a three- to four-fold higher mortality risk than non-diabetics (Table 2). Adjusting the results for age of the subjects (...truncated)