The first studies linking hypertension and greater risk of cognitive decline in older adults date back from the early nineties. One of the first carefully designed and controlled study was the Framingham study that showed a negative correlation between untreated BP in middle age adults and cognitive functioning 12–14 years later [16]. Cognitive functions were measured using several standardized neuropsychological tests from the Wechsler adult intelligence and memory scales. Results showed that averaged blood pressure and chronicity of hypertension were associated with verbal and visual memory, as well as working memory performances. The association between BP and cognitive state has also been observed within a shorter follow-up period. Indeed, in the Epidemiology of vascular aging (EVA) study [27], 1373 individuals aged between 59 and 71 years were followed for only 4 years. Authors observed that those showing high BP were 4.3 times more likely to show cognitive decline, as determined by a decrease of four points on the MMSE. The risk of cognitive decline decreased to 1.9 times in individuals who were taking antihypertensive medication.

Other studies further explored the temporal relationship between BP and cognitive decline, while taking into account the severity of hypertension, as to whether it was controlled or not. Alves de Moraes and collaborators followed 8058 participants from the Atherosclerosis Risk in Communities (ARIC) study [28]. BP and cognitive functions were measured at two different time points, 6 years apart. Cognitive testing targeted memory, processing speed, and verbal fluency. Participants were classified in one of the following BP categories: (a) normal BP, defined as having no prevalent hypertension and normal BP at both visits; (b) incident hypertension, defined as having normal BP at the first visit, but hypertension at the second visit; (c) controlled hypertension, defined as having prevalent hypertension but normal BP at both visits; (d) partially controlled hypertension, defined as having prevalent hypertension, and abnormal BP at one of the two visits; and (e) uncontrolled hypertension, defined as having prevalent hypertension with high BP at both visits. Results showed that, at baseline, normotensive participants had higher scores for processing speed than those in the three other groups. Verbal fluency performance was also higher in normotensive individuals when compared to subjects with partially controlled hypertension. Among the hypertensive groups, processing speed and verbal fluency were lower in partially controlled and uncontrolled hypertension categories. Longitudinal analyses on a sample of 10,963 adults from the same cohort revealed that after controlling for demographic factors, hypertension at baseline was associated with decline on a measure of processing speed over 6 years (Digit Symbol Substitution test of the Wechsler adult intelligence scale), but not in memory or word fluency performance [29]. Analysis of change between time 1 and time 2 yielded only one significant difference between normotensives and uncontrolled hypertensive patients, observed in processing speed score. Further, this difference was limited for older participants. These results suggest that the temporal change in cognition is not perfectly related to changes in BP and may be age-dependent. These results also suggest that some domains of cognition are more sensitive to differences in hypertension status than others.

Other studies reporting on longer follow-up periods also support the notion that high BP is associated with cognitive decline in older adults. For instance, an ambitious study examined the longitudinal association between BP and cognition in a large Sweden population-based cohort of 999 men followed over 20 years [30]. Ambulatory 24-h BP monitoring was recorded at the beginning of the study, when participants were 50 years of age, and then 20 years later. Cognitive function was measured only at 70 years with a composite score that included the MMSE and the Trail-making test, a test that taps psychomotor speed and executive control (i.e., switching). Longitudinal analyses revealed that men with DBP ≤ 70 mmHg showed the highest cognitive performance, whereas men with DBP ≥ 105 mmHg had the lowest cognitive performance. SBP was not related to cognition in this sample. Cross-sectional analyses at age 70 also revealed a significant inverse relationship between DBP and cognitive score, with stronger associations in untreated men. In fact, an increase of 1 SD of 24-h DBP raised the risk of cognitive impairment (i.e., being in the lowest quintile of cognitive performance) by 1.45. The composite score precludes conclusion as to whether decline was general or impacted more specific cognitive functions.

In a secondary analysis of the data from the ACTIVE intervention trial, impacts of BP and diabetes mellitus on cognitive performance was investigated in 2802 participants aged 65–94 [31]. Cognitive functioning was assessed yearly over 3 years using a collection of neuropsychological tests that targeted global cognitive function, memory, reasoning, and speed of processing. Using BP at baseline, participants were classified into four BP groups: normal BP (< 120/80 mmHg), prehypertension (120–139/80–89 mmHg), Stage 1 hypertension (140–159/90–99 mmHg), or Stage 2 hypertension (> 160/100 mmHg). Cross-sectional results showed that high BP was specifically associated with reasoning abilities but not memory performance. In longitudinal analyses, although a test–retest improvement of cognitive performance was observed from baseline to time 2, individuals with Stage 1 and Stage 2 hypertension had a faster decrease in reasoning performance than normotensive participants from time 2 to time 3. Interestingly, the acceleration of cognitive decline associated with high BP was specific to frontally-mediated cognitive functions.

Other studies suggest however that verbal episodic memory could also be negatively affected by hypertension. In fact, in Swan, Carmelli, and Larue’s report from the Western Collaborative Group Study [32], SBP was tracked over 30 years to examine if its change was associated with neuropsychological performance after adjustment for age, education, depression, stroke, and use of antihypertensive medication. They found that subjects whose SBP remained high (≥ 140 mmHg) from midlife to follow-up, had lower scores on a composite measure of verbal memory than those whose SBP had a normal trajectory, defined as being low (< 120 mmHg) or medium (120–139 mmHg) throughout life, or showing an increase from low to medium or medium to high over the follow-up period (average increase of 20.1 %). Subjects whose SBP decreased over the life span (average change of −6.7 %) performed less well on a composite measure of processing speed than subjects from the normal group. Participants from the decliners’ group also showed higher prevalence of depression and coronary heart disease, which both have been related to cognition [33, 34].

Some studies have demonstrated a more complex relationship between BP and cognition throughout life span. For instance, a longitudinal population-based study on 2068 men and women aged 65–102 years showed that BP had a U-shaped association with performance on a mental status questionnaire, meaning that subjects with low SBP (< 130 mmHg) and high SBP (≥ 160 mmHg) made more errors than participants with optimal SBP (130–159 mmHg) [35] (see also [36] and [37]).

It thus seems that several longitudinal studies support the notion that hypertension, most commonly indexed by abnormal BP, is associated with lower cognitive performance compared to normotensive state and with higher risk of further cognitive decline over relatively short period of time, sometimes as early as 4 years follow-up. A definitive conclusion as to whether this longitudinal observation holds for global cognitive status or is more specific to some cognitive domains can only be supported by a limited number of studies, but it seems that executive control and controlled attention are more sensitive to hypertension status than other functions, although some studies also identified episodic memory as being negatively affected by hypertension. Moreover, it is worth mentioning that not all studies support a relationship between BP and cognitive changes. For instance, in the Chicago Health and Aging Project, BP at baseline was not associated with the 6-year change in cognitive function in the large community-based sample of 4284 older adults [38]. Further studies could certainly help clarify the link between BP and change in BP over years and the profile of cognitive decline.

Results from cross-sectional studies provide valuable data in order to better understand how hypertension and BP status impacts cognition globally or differentially according to the cognitive domains that are being investigated. In a large population-based cohort of 19,836 individuals aged 45 and older (mean age: 64.6 ± 9.5), a linear relationship between higher DBP and impaired cognition (measured by the 6-item screener, derived from the MMSE) was observed [39]. Each 10 mmHg increase in DBP was associated with a 7 % augmentation in odds ratio for cognitive impairment. In this study, SBP was also related to cognitive performance, but adjustment for confounding variables, such as demographic characteristics, vascular risk factors, depressive symptoms, and antihypertensive medication suppressed the significance of the association. Using a more comprehensive battery of cognitive tests (six subtests of the Wechsler Adult Intelligence Scale [WAIS]), Robbins and collaborators [40] assessed the relationship between blood pressure and cognitive performance in a sample of 1563 participants aged 18–79 years (mean age: 49.1 ± 15.0). The results showed that both SBP and DBP were associated with performance in all tests, but regression coefficients indicated that performance at the Digit Symbol Substitution test, a measure of psychomotor speed, was more strongly and consistently related to SBP and DBP than any other cognitive outcome. Performance at the Similarities subtest, a test of abstract reasoning capacities, was also strongly predicted by BP (SBP and DBP) but to a lesser extent. It could thus be the case that to some extent BP would have a more specific impact on speed of processing and/or tests that require controlled and effortful attention, but definitive conclusion requires further observation with a broader variety of cognitive tests. Interestingly in this study, the negative effect of high BP level on cognition was independent of age but not race, as significant interactions of race with SBP and DBP were observed in most cognitive outcomes. In fact, although African–American sometimes showed greater impact of high BP on cognition, the relation between BP and cognitive performance was observed in both racial groups. Moreover, the specificity of the cognitive domains altered, with speed of processing and abstract reasoning being more impaired than other cognitive tests, holds for all participants independently of age and racial groups.

Vicario et al.’s study [41] further explored the impact of hypertension on executive functions by using more than one test of attention and executive control. They observed that 46 % of hypertensive patients were unable to complete Part B of the Trail-making test within the time limit, whereas only 13 % of normotensives failed to do so. This suggests a negative impact of hypertension on cognitive flexibility or attentional switching. The authors also observed significant differences between groups in the Stroop task, which targets inhibition, as well as delayed recall, a measure of long-term episodic memory. However, they did not reproduce the effect of hypertension on MMSE scores that was observed in other studies. Saxby and colleagues employed a comprehensive neuropsychological battery to evaluate the effect of hypertension on attention, memory, and executive functioning in a sample of 223 individuals aged 70–89 [42], and used factor analysis techniques to reduce the number of cognitive dimensions studied. When compared to the normotensive group, hypertensive participants showed lower performance on composite measures of speed of cognition, executive functions, episodic memory, and working memory, but not continuity of attention. Hypertension appeared to affect many cognitive functions, sparing only measures of vigilance and task accuracy that were included in the continuity of attention score.

Overall, cross-sectional studies suggest that older adults with hypertension tend to show lower cognitive performances than those without hypertension. However, not all cross-sectional studies support this claim. In DiCarlo et al.’s study, that involved a large sample of 3425 participants aged 65–84 years, hypertension was not related to cognitive impairment (measured with the MMSE) [43]. Another study with 936 adults ranging from 24 to 81 years of age, showed that BP was not a significant predictor of cognitive performance, measured with five cognitive tests, even in participants that were not taking antihypertensive medication [44]. Analysis with participants stratified for age also did not yield significant results. Nevertheless, given the number of studies reporting a significant link between hypertension or BP level on cognition, overall findings support the notion that hypertension and elevated BP can negatively impact cognition. Several cognitive domains have been pointed out as showing vulnerability to high BP, especially processing speed and executive functions (e.g., abstract reasoning, cognitive flexibility), but memory may also be affected.

Given the well-established age-dependent effect on cognition and the high prevalence of hypertension in older adults, the potential interaction effect of age on the association between hypertension and/or BP and cognition has been investigated, but studies have led to mixed results. In 357 older men, Brady et al. [45] showed an age-related decline in most neuropsychological measures, but did not find the expected main effect of BP category (normotensive, controlled hypertensive, uncontrolled hypertensive, and untreated hypertensive) on cognitive functioning. Interestingly, their results revealed an interaction between BP and age on performance in tests of verbal fluency and immediate recall. Only older uncontrolled hypertensive men performed less well than normotensive men. The authors suggested that deficits in episodic memory recall and verbal fluency may be due to deficits in the strategies used for information retrieval, which depend on the effectiveness of executive functions. In the same line, Bucur and Madden [46] showed a selective BP-related decline in executive functions only in their oldest group (60–79 years) of participants. In this study, speed of processing was not influenced by BP nor by the interactive effects of age and BP. Thus, it has been hypothesized that elevated BP may exacerbate age-related cognitive decline in frontally mediated functions, namely executive control. However, some studies that tested the interaction of age and BP on various cognitive functions did not yield significant results [39, 47], or showed the inverse pattern. In Suhr, Stewart, and France [48], there was an interaction between SBP and age on performance in a verbal learning/attention measure, where only participants younger than 40 years showed the significant association between SBP and cognition. However, the sample did not include adults older than 59 years.

While some studies shed light on the possible age-accelerating effect of BP on cognition, another body of evidence showed what resembles to a protective effect of high BP in the oldest portion of the older population. For instance, a prospective investigation from the Rotterdam study and the Leiden 85-plus study showed that elevated baseline BP was associated with a greater risk of cognitive decline 11 years later, but only in 65–74-year-old participants, whereas younger (< 65) did not show any significant relationship between BP and future cognition [49]. However, in the oldest group (≥ 75), higher BP was associated with better cognitive functioning a decade later. In Obisesan et al.’s study [50], hypertension was associated with impaired performance in the short-portable MMSE, but not in the oldest group. Indeed, hypertensive subjects aged 80 years and older showed better cognitive performance than their normotensive counterparts. Interestingly, while hypertension has consistently been shown to increase the risk of dementia in middle-aged or young-old adults, this relationship is not observed in the oldest old for whom the inverse association is sometimes observed [51–55]. Further, in Sabayan et al. [56], participants aged 85 and older with high BP had less cognitive and physical disability (measured with the MMSE and assessment of activities of daily living [ADL]) than those with low BP. Longitudinal analyses from this cohort also showed that high BP was associated with less cognitive and physical decline, especially among participants who had preexisting physical disability. Together, these results support the idea that mild hypertension may have a protective effect on cognitive functioning in the oldest portion of older adults. In their exhaustive review of cross-sectional and longitudinal studies on the effect of low and high BP on cognitive functioning and dementia across different age groups, Qiu et al. [18] conclude that elevated BP in midlife is an important risk factor for developing late-life cognitive impairment and dementia. Among older adults, very high SBP (> 180 mmHg) is also associated with an increased risk of dementia, but low DBP (< 70 mmHg) augments the risk as well. The authors go further and suggest that high BP in middle age coupled with low BP in old age may be a marker of AD (see also [17]). Several explanations may account for the potential protective effect of high BP on cognition in the oldest-old. The most largely admitted hypothesis state that higher BP is necessary to overcome age-related arterial stiffness or vascular damage and maintain adequate perfusion in the brain [57], hence the protective effect of mild hypertension on cognition.