These results support the following conclusions. First, hospitalization due to TAAD occurred at a significantly lower rate among individuals with diabetes than among nondiabetic inpatients. Second, diabetes was inversely associated with TAAD independent of age, income, region, hospital type, or other clinical characteristics. Third, the inverse relationship between diabetes and TAAD was present in men and women, in nonwhites and whites, and in cases with and without aortic dissections. Fourth, the inverse relationship between diabetes and TAAD was strongest among discharges with chronic complications of diabetes, suggesting a possible link to hyperglycemia in terms of duration, severity, or susceptibility to vascular injury. Finally, this relationship was quite specific to TAAD.
The strengths of our study included the selection of case and control subjects from a uniform nationally representative database, making information bias unlikely, a rigorous case definition and the selection of a random sample of all control subjects. The principal limitation of this study was our use of hospitalized rather than population-based controls, which may introduce significant bias into our estimates due to enrichment of diabetes among inpatients. We found that this is unlikely, because the overall prevalence of diabetes among unique 2006 NIS discharges (21.6%) was not significantly different than the reported age-adjusted prevalence of diabetes among US adults in contemporary National Health and Nutrition Examination survey data (20.3%).25,26
The relationship between diabetes and TAAD also remained constant in an independent cohort of locally recruited inpatients with more prevalent diabetes (35.9%). Our hypothesis is based on cross-sectional associations and requires validation in cohort studies in order to establish a causal relationship between diabetes and TAAD. However, moderate effects of diabetes on clinical endpoints are not likely to be detected in previously published TAAD cohorts due to relatively short-term followup periods and small numbers of patients. This is the largest study of TAAD to date and the only study that is adequately powered to address this question.
The NIS does not contain detailed clinical information on the severity of illness or indications for procedures. This raises concerns about the potential for differential misclassification. We used the same ICD
-9-CM codes for TAAD as a prior study of nationwide data, and we verified that cases were enriched for procedure codes that are exclusive to TAAD.8
Cardiovascular diagnoses such as CHF, hyperlipidemia, ischemic coronary disease, and diabetes were verified to be highly specific when clinical and claims data were compared,27
while demographic variables are unlikely to be misclassified in a differential manner with respect to TAAD. Our results may underestimate the impact of diabetes on TAAD because cases may have been more likely to be diagnosed with diabetes than control subjects. This type of misclassification could occur because diabetes may be scrutinized in TAAD patients as a well-known risk factor for vascular disease and would tend to bias our results toward the null hypothesis. Nonetheless, a significant association remained between diabetes and TAAD. Furthermore, we found a positive association between diabetes and other types of vascular disease in the same individuals.
Because no unique personal data were included, we could not identify individuals with multiple hospitalizations. Therefore our results are mainly confined to discharges rather than individuals, and we report discharge rates per population rather than incidence or prevalence per population. When we limited our analysis to a subset of unique discharges, the associations did not substantially change. Although our results with regard to diabetes and TAAD are unprecedented, our findings are consistent with prior studies showing that the rate of abdominal aortic aneurysm progression is reduced in diabetic patients.
Diabetes was recently shown to decrease the progression of aortic disease by direct metabolic effects as well as modulation of inflammation in the aortic wall. In mice with experimentally induced hyperglycemia, the expansion rate of abdominal aortic aneurysms was significantly attenuated.11
In clinical studies, diabetes was found to be associated with a reduced prevalence and expansion rate of abdominal aortic aneurysms, and this was correlated with decreased secretion of metalloproteinases by aortic inflammatory cells from diabetic patients.12,28
Hyperglycemia is also associated with reduced adventitial neovascularization and decreased infiltration of inflammatory cells into the medial layer of the aorta.29
These processes could also inhibit the progression of TAAD by reduction of vascular smooth muscle cell death and extracellular matrix degradation. Alternatively, it is possible that TAAD may protect against the development of diabetes. TAAD is associated with increased circulating concentrations of insulin-like growth factor 1, an endocrine peptide with potent antidiabetic effects.30,31
Further experimental evidence will be required to determine the plausibility of these various pathophysiologic explanations for our findings.
We found that diabetes is associated with a 40% to 80% reduction in the rate of hospitalization due to TAAD and is as important as hypertension as a predictor in our multivariate model. Using NIS data, we estimate that TAAD was diagnosed in at least 15 per 100 000 US adults in 2006. This compares favorably with annual estimates of 10.5 to 14.8 cases of TAAD per 100 000 persons in Rochester, Minnesota between 1980 and 1994 and 10.7 to 16.3 cases per 100 000 persons in Sweden between 1987 and 2002.32,33
The prevalence of TAAD is increased 10-fold in hospitalized patients compared with the general population, indicating that TAAD is a major cause of morbidity and mortality. The principal implication of our findings is that diabetes is negatively associated with hospitalization due to TAAD. Patients with diabetic complications related to the severity or persistence of diabetes have the lowest rates of TAAD. Future research should seek to determine whether there is a direct vascular protective effect of hyperglycemia on the aortic wall and to disentangle the molecular pathways that mediate this effect. Another implication is that future studies of TAAD treatment should consider diabetes in predictive models of aneurysm expansion or dissection.