Mortality From Chronic Liver Diseases in Diabetes
Mortality From Chronic Liver Diseases in Diabetes
As detailed in Figure 1, 176,771 diabetic subjects were initially identified in December 2007 from the regional electronic archive of exemptions from medical charge of the Veneto Region. Out of them, 209 were excluded because of incomplete data and 1,302 because they died in the last weeks of the year. Out of the cohort of 173,260 patients with complete data and alive on 31 December 2007, final analyses were restricted to the 167,621 subjects aged 30–89 years at the beginning of the follow-up, corresponding to 5.0% of the total regional population of the same age band. Of the cohort subjects, 54.6% were men, 38.2% were aged 30–64 years, 32.6% were aged 65–74 years, and 29.2% were aged 75–89 years, respectively. There were 473,374 person-years of follow-up in the period from 2008 to 2010, and 17,134 (10.2%) deaths occurred.
(Enlarge Image)
Figure 1.
Schematic design of the enrollment and follow-up of the diabetic subjects.
The crude mortality rate was higher in men than in women and, as expected, increased progressively with aging, reaching a crude mortality rate of 76.6 per 1,000 subjects among those aged 75–89 years (Table 1).
The main causes of death were cardiovascular diseases and neoplasms accounting for 34.6% and 31.0% of all deaths, respectively. CLD accounted for 2.3% of total deaths, and they were responsible for about half (47.2%) of the mortality from digestive diseases. Notably, as shown in Table 2, the diabetic cohort had a significantly greater mortality risk than the general population, with an overall SMR of 1.49 (95% confidence interval=1.46–1.52) in men and 1.53 (1.49–1.56) in women, respectively. The excess in mortality of diabetic subjects with respect to the general population was much larger among younger age classes. All the main causes of death showed a significantly higher risk in diabetes (Figure 2, Table 2). The risk of dying from CLD was markedly increased among diabetic subjects with respect to the general population, with a SMR of 2.47 (2.19–2.78) for men and of 2.70 (2.24–3.23) for women.
(Enlarge Image)
Figure 2.
Age-standardized mortality rates (standard European population) for circulatory diseases, neoplasms, and chronic liver diseases among diabetics and the general population of the Veneto region aged 30–89 years.
When all diseases reported in death certificates were analyzed, CLD was mentioned as either a causing or a contributing factor in 6.9% of the total deaths that occurred in diabetic subjects. The 2.5-fold relative risk estimated by analyses restricted to the underlying cause of death was confirmed. Interestingly, as reported in Table 3, NVNA-related CLD, which represented about two-thirds of cases of all CLD-related deaths among diabetics, had a SMR of 2.86 (2.65–3.08), whereas SMRs for virus-related and alcohol-related CLD were 2.17 (1.90–2.47) and 2.25 (1.98–2.54), respectively. As shown in Table 4, when the analysis of mortality from CLD was stratified by sex and age, the SMRs for NVNA-related CLD remained the highest in all age classes and in both sexes. It is noteworthy that the SMR for NVNA-related CLD was higher in younger subjects (30–64 class of age) and decreased across the three classes of age. Also worth mentioning is that the SMR for NVNA-related CLD death was remarkably higher in women than in men.
Finally, when analyses were carried out in the sub-cohort with at least 7 years of diabetes duration, the estimated mortality risk for CLD only slightly decreased in the underlying cause of death analysis (SMR=2.33, 1.99–2.71), and remained unchanged in the multiple causes analysis both for all CLD (SMR=2.55, 2.34–2.77) and for NVNA-related CLD (SMR=2.89, 2.59–3.22, data not shown).
Results
As detailed in Figure 1, 176,771 diabetic subjects were initially identified in December 2007 from the regional electronic archive of exemptions from medical charge of the Veneto Region. Out of them, 209 were excluded because of incomplete data and 1,302 because they died in the last weeks of the year. Out of the cohort of 173,260 patients with complete data and alive on 31 December 2007, final analyses were restricted to the 167,621 subjects aged 30–89 years at the beginning of the follow-up, corresponding to 5.0% of the total regional population of the same age band. Of the cohort subjects, 54.6% were men, 38.2% were aged 30–64 years, 32.6% were aged 65–74 years, and 29.2% were aged 75–89 years, respectively. There were 473,374 person-years of follow-up in the period from 2008 to 2010, and 17,134 (10.2%) deaths occurred.
(Enlarge Image)
Figure 1.
Schematic design of the enrollment and follow-up of the diabetic subjects.
The crude mortality rate was higher in men than in women and, as expected, increased progressively with aging, reaching a crude mortality rate of 76.6 per 1,000 subjects among those aged 75–89 years (Table 1).
The main causes of death were cardiovascular diseases and neoplasms accounting for 34.6% and 31.0% of all deaths, respectively. CLD accounted for 2.3% of total deaths, and they were responsible for about half (47.2%) of the mortality from digestive diseases. Notably, as shown in Table 2, the diabetic cohort had a significantly greater mortality risk than the general population, with an overall SMR of 1.49 (95% confidence interval=1.46–1.52) in men and 1.53 (1.49–1.56) in women, respectively. The excess in mortality of diabetic subjects with respect to the general population was much larger among younger age classes. All the main causes of death showed a significantly higher risk in diabetes (Figure 2, Table 2). The risk of dying from CLD was markedly increased among diabetic subjects with respect to the general population, with a SMR of 2.47 (2.19–2.78) for men and of 2.70 (2.24–3.23) for women.
(Enlarge Image)
Figure 2.
Age-standardized mortality rates (standard European population) for circulatory diseases, neoplasms, and chronic liver diseases among diabetics and the general population of the Veneto region aged 30–89 years.
When all diseases reported in death certificates were analyzed, CLD was mentioned as either a causing or a contributing factor in 6.9% of the total deaths that occurred in diabetic subjects. The 2.5-fold relative risk estimated by analyses restricted to the underlying cause of death was confirmed. Interestingly, as reported in Table 3, NVNA-related CLD, which represented about two-thirds of cases of all CLD-related deaths among diabetics, had a SMR of 2.86 (2.65–3.08), whereas SMRs for virus-related and alcohol-related CLD were 2.17 (1.90–2.47) and 2.25 (1.98–2.54), respectively. As shown in Table 4, when the analysis of mortality from CLD was stratified by sex and age, the SMRs for NVNA-related CLD remained the highest in all age classes and in both sexes. It is noteworthy that the SMR for NVNA-related CLD was higher in younger subjects (30–64 class of age) and decreased across the three classes of age. Also worth mentioning is that the SMR for NVNA-related CLD death was remarkably higher in women than in men.
Finally, when analyses were carried out in the sub-cohort with at least 7 years of diabetes duration, the estimated mortality risk for CLD only slightly decreased in the underlying cause of death analysis (SMR=2.33, 1.99–2.71), and remained unchanged in the multiple causes analysis both for all CLD (SMR=2.55, 2.34–2.77) and for NVNA-related CLD (SMR=2.89, 2.59–3.22, data not shown).
