Hypertension and Cognitive Function – It’s Not All in the Numbers

The American Heart Association includes having an untreated BP <90th percentile in childhood and <120/<80 mmHg for those over the age of 20 years as one of seven metrics reflective of ideal cardiovascular health.¹ Yet, an estimated 80 million adults living in the United States have hypertension with a disproportionate prevalence among non-Hispanic black women (46.1%) and men (44.9%) compared to whites (30.1% and 32.9%, respectively).² Aside from Mexican-American men, the prevalence of hypertension increased across race-ethic groups in the country between 1988 and 2102.² Hypertension is the single most important treatable risk factor for stroke, with a population attributable risk decreasing from 40% for those age 50-years (relative risk, RR 4.0) to 20% for those age 80-years (RR1.4).³ Even pre-hypertension (BP 121-139/80-89 mmHg), however, is associated with an 66% (RR 1.66, 95% CI 1.51-1.81) higher stroke risk compared to an optimal blood pressure (<120/80 mmHg).⁴ The risk of stroke-related mortality increases with increasing systolic BP>155 mmHg and diastolic BP>75 mmHg for each decade of age between 50-59 years and 80-89 years.⁵

The risk of neurodegenerative disorders including Alzheimer’s Disease (AD) also increases with age with the prevalence rising 15-fold between the ages of 60 and 85-years.⁶ Because stroke can cause a variety of cognitive impairments with up to 30% of survivors having dementia by 6-months post stroke, each condition contributes to the overall burden of dementia epidemiologically and within individual patients.⁷ Distinguishing the roles of vascular and degenerative diseases in specific patients and in population-based studies, therefore, can be challenging. This can be even more difficult in cross sectional epidemiological studies because of reverse causality (i.e., the role of cognitive impairment as a risk factor for stroke compared to the role of stroke as a risk factor for cognitive impairment). For example, a systematic review and meta-analysis found that the pooled relative risk of stroke was nearly 20% higher per standard deviation lower global cognitive performance (RR 1.19, 95% CI 1.12-1.27).⁸

Adding complexity, in addition to overt stroke, subclinical cerebrovascular disease may be associated with cognitive impairment. For example, some studies find that the extent of white matter hyperintensities on brain MRI imaging, which correlate with a variety of vascular risk factors including hypertension, also correlate with cognitive performance.9 Among normal elders, specific cognitive profiles are associated with progression of anterior and posterior white matter hyperintensities.¹⁰ A history of stroke-like symptoms in patients with no prior stroke or TIA is associated with impaired measures of physical functioning.¹¹

Because of the overlap between cerebrovascular and degenerative causes of cognitive impairment and dementia, and because both conditions increase with age, it is not surprising that a variety of vascular risk factors are associated with cognitive aging and AD.^{12, 13} Hypertension is among these risk factors. The writing group for the AHA Scientific Statement, “Impact of Hypertension on Cognitive Function,” carefully reviewed the available literature to evaluate the association between hypertension and cognitive impairment. In this context, it is helpful to consider the evidence based on Bradford Hill’s nine classic criteria for determining causality.¹⁴

The first of Hill’s criteria addresses the strength of the association, i.e., the degree to which hypertension is associated with impaired cognition. As reflected in the Scientific Statement, the answer to this seemingly simple question is not straightforward.¹⁵ Individual studies vary in design, statistical methodology, characteristics of the study populations, and measures of cognition. There are potential interactions with the duration of hypertension, patient age, comorbid conditions, and treatment. Prospective studies may classify a subject as hypertensive based on a single baseline assessment, but not account for changes in blood pressures over time. Although generally consistent, there is surprisingly limited data showing that the presence of hypertension is strongly associated with the degree of cognitive impairment or decline.

The second of Hill’s criteria focuses on the consistency of the relationship among studies. The heterogeneity noted above compromises assessment of the consistency of the hypertension-cognition association. For example, in later life, some studies suggest a beneficial impact of high blood pressure on cognition whereas others find a harmful effect. The most consistent relationship among studies is between blood pressure in midlife and cognitive performance in mid- and late-life.¹⁵

Heterogeneity also affects assessment of the specificity of the relationship between hypertension and cognition. The causes of dementia, however, are clearly multifactorial. Hypertension may be a contributing cause, but it is not the only cause. The relative contributions of different potential risk factors for cognitive impairment and dementia can be difficult to tease apart.

Temporality addresses the problem of reverse causality. Exposure to the risk factor should precede the outcome of interest. As in the example provided above, cognitive impairment alone is associated with a higher risk of subsequent stroke and therefore might be associated with a higher risk of traditional stroke risk factors. Several studies, however, find that pre-existing hypertension is associated with a decline in cognition or a conversion from mild cognitive impairment to AD.^16-18 There is, therefore, evidence that hypertension is often present in patients who later develop cognitive impairments.

There are only limited data evaluating the correlation between the degree of hypertension and the occurrence or degree of cognitive impairment (i.e., Hill’s criteria for a biological gradient). One longitudinal study found that subjects who developed dementia at age 79-85 compared to those who remained cognitively normal had higher mean systolic (178 vs 164 mmHg) and diastolic (101 vs 92 mmHg) blood pressures at age 70 years.¹⁹ Another longitudinal study reported an association between systolic blood pressure and risk of developing vascular dementia for younger (hazard ratio, HR per 20 mmHg higher systolic BP, 1.62; 95% CI 1.13–2.35 at 30–50 years; 1.26, 95% CI 1.18–1.35 at 51–70 years), but not older (HR 0.97, 95% CI 0.92–1.03 at 71–90 years) subjects (p trend=0.006). ²⁰ These data support the presence of a biological gradient between blood pressure and the risk of cognitive impairment.

The AHA writing group thoroughly reviewed data related to the impact of hypertension on the cerebral vasculature in addition to its effects on the pathophysiological processes associated with AD, supporting biologically plausible mechanisms underlying a causal relationship between hypertension and the development of cognitive deficits. Hypertension leads to structural changes in vessel walls in addition to impairments in vascular regulation that can result in vascular insufficiency in subcortical and deep brain regions, including the white matter.15 Hypertension may also increase the accumulation of amyloid- and accumulated amyloid-? may affect cerebrovascular regulatory mechanisms.¹⁵

Hill’s criterion of coherence asks whether a causal relationship between a factor and its putative effect is consistent with other data reflecting the natural history and biology of disease.¹⁴ Coherence for the relationship between blood pressure and cognition is in part supported by data reflecting the of the pathophysiological effects of hypertension, its impact on biomarkers such as MRI brain imaging reflecting vascular injury, and evidence of the temporal relationship between the detection of hypertension and the development of deficits reflected by cognitive tests.

The most exacting of Hill’s criteria is experiment; i.e., do interventions aimed at altering or eliminating a risk factor affect the frequency or magnitude of an outcome? Even if an effect is found, however, this alone is not conclusive. An intervention might have an impact on a marker that is not causally linked to a disease, but parallels its effect on a risk factor. For example, stains decrease both cholesterol levels and measures of inflammation.˜21 The degree to which their effects on vascular risk are related to their effects on each potential mechanism is uncertain. For hypertension. a 2009 systematic review identified four trials that reported the incidence of dementia in 15,936 hypertensive patients (mean age 75.4 years), finding no difference between those randomized to antihypertensive treatment compared to placebo (odds ratio, OR 0.89, 95% CI 0.74-1.07). 22 There was significant heterogeneity among the studies, and they do not address the effect of treatment in middle age on cognition in later life. A subsequent meta-analysis including 14 longitudinal studies including 32,658 treated and 36,905 un-treated subjects found no difference with treatment in the incidence of AD (RR 0.90, 95% CI 0.79-1.03), cognitive decline (RR 0.97, 95% CI 0/92-1.03), or cognitive impairment (RR 0.97, 95% CI 0.92-1.03), but a lower risk of vascular dementia (RR 0.67, 95% CI 0.52-0.87) and any dementia (RR 0.87, 95% CI 0.77-0.96). 23 The AHA writing group concluded that data from clinical trials did not allow conclusive recommendations about treating hypertension at any age to protect cognition.¹⁵

The final Hill criterion is analogy. There is not, however, another putative risk factor analogous to hypertension.

The AHA writing group identified several major gaps in current knowledge and important directions for future research. Unlike the clear causal relationship between increasing usual blood pressure and increasing stroke risk, the relationship between blood pressure and cognition is less clear. Stressing the importance of knowledge of blood pressure numbers is a central public health message and a critical component of the overall strategy for improving cardiovascular health. Although taken together the data are highly suggestive, adding preservation of cognition as another benefit of better blood pressure control awaits further study.