Hypertension and diabetes versus the risk of aortic disease: a new look on prevention?

This editorial refers to ‘Dose-dependent relationship of blood pressure and glycaemic status with risk of aortic dissection and aneurysm’, by Y. Suzuki et al., https://doi.org/10.1093/eurjpc/zwac205.

Aortic dissection (AD) and aortic aneurysm (AA) are both rare life-threatening diseases requiring extensive and expensive therapy, while prevention and risk factors are not well understood. The incidence of AD has been reported as 3–6 cases per 100 000 population and even higher incidence rates were reported in AA, with a 1–3% prevalence of abdominal AA in men over 60 years of age.^1,2 Historically, risk factors for AD and AA were advanced age, male sex, untreated hypertension, smoking, COPD, as well as environmental factors.^3,4 More recently, patients with diabetes were associated with lower risk of AD and AA,⁵ but such relationship has always been challenged and controversial without independent external validation. In this issue of the European Journal of Preventive Cardiology, Suzuki and colleagues analyse a large-scale health check-up and claims database to shed light on the association among hypertension, hyperglycaemia, and the risk of AD and AA.⁶

The first observation in the article by Suzuki et al., confirming current knowledge, was that hypertension is a significant risk factor for AD and AA in a dose–response relationship. Compared with a reference group [normal/elevated blood pressure (BP)], hazard ratios of the American College of Cardiology/American Heart Association (ACC/AHA)-defined Stage 1 (systolic BP of 130–139 mmHg or diastolic BP of 80–89 mmHg) and Stage 2 (systolic BP ≥140 mmHg or diastolic BP ≥90 mmHg) hypertension were 1.89 [95% confidence interval (CI), 1.60–2.22] and 5.87 (95% CI, 5.03–6.84) for AD and 1.37 (95% CI, 1.23–1.52) and 2.17 (95% CI, 1.95–2.42) for AA, respectively, in an unadjusted model. Significant associations were stable in two adjusted models, one including glycaemic status, age, and sex, while the other added body mass index (BMI), systolic BP, dyslipidaemia, cigarette smoking, and alcohol consumption to the mix. In restricted cubic spline curves adjusted for covariables, the risk of AD and AA increased monotonically, which were consistent with a previous large-cohort study and meta-analyses.^7,8

The second findings suggested that hyperglycaemia impacts negatively on the incidence of AD and AA after adjustment. Interestingly, in unadjusted data, hyperglycaemia was associated with increased risk of AD and AA. The hyperglycaemic population had roughly twice the incidence of both AD and AA, with fasting plasma glucose (FPG) level (as a continuous and categorical variable) positively related to increased risk of AD and AA. Miraculously, however, after covariable adjustment with age and sex, this effect decreased to an insignificant difference (Regression Model 2) and the magnitude was decreased with further adjustment including BMI, systolic BP, dyslipidaemia, cigarette smoking, and alcohol consumption in addition to Model 2 variables. In restricted cubic spline curves adjusted for the same covariates, the risk of AD and AA decreased as FPG level increases.

For this two-sided association of glycaemic status with risk of AD and AA discovered in the current study, a deep-diving re-examination of risk conditions had been requested before drawing conclusions.⁹ Of note, hypertension and diabetes are both primarily lifestyle-related conditions sharing similar metabolic risk factors and often develop together.¹⁰ As the authors acknowledged, most patients in the diabetic cohort had hypertension as well (66.5% of the diabetic population in the study cohort) with a striking male predominance (around 80% of the diabetic population in the study cohort were male) and at advanced age (median age of 47–52 years while 43 years in the entire cohort). Without adjustment, diabetes patients in fact had a higher incidence of aortic diseases and only after adjustment with these covariables, the association with diabetes was favourable statistically. Further detailed investigation on diabetic patients without hypertension is needed to clarify the nature of this ‘blessed diabetic population’ who enjoyed a lower risk of AD and AA. In the light of rare occurrence of aortic diseases, it would be highly valuable to identify both the ‘blessed’ subset with diabetes and the true high-risk population in an attempt to eventually prevent disease progression.

While the study provides epic insights on the epidemiology of these aortic diseases, it has some limitations. Large administrative cohort database can be subject to information bias, which may weaken the effect size of observed relationship.¹¹ The current study population may consist of relatively healthier population as ‘Kempo’ health insurance is for employees, which may also attenuate risk compared with a true population incidence of aortic diseases. On the other hand, insurance claims data on both outpatient and hospital care were used as study outcomes, which may inflate the risk by adding disease codes for insurance claim purpose (especially when hospitals perform tests for aortic diseases). Nonetheless, the study added noticeable observations, such as associations of hypertension and hyperglycaemia with AD and AA with detailed analyses including a series of sensitivity analyses.

In summary, the impact of BP on the risk of AD and AA is confirmed and consolidated in the current large cohort study, which certainly increases awareness of elevated BP for the incidence of aortic diseases. Moreover, risk stratification guided by epidemiological data seems to be an effective tool to focus on high-risk subsets of population for low-incidence conditions such as aortic diseases. Further investigations into hyperglycaemic status and aortic disease are warranted, and in future may open an avenue to an innovative preventative strategy against life-threatening aortic diseases.

References

Author notes