Trajectory patterns of blood pressure change up to six years and the risk of dementia: a nationwide cohort study

Introduction

Dementia is a common public health challenge worldwide, affecting nearly 47 million people in 2015, and this figure is expected to rise to 131 million by 2050 [1]. In view of the lack of effective treatment on cognitive decline for patients with dementia and the poor prognosis of progressive exacerbation, exploring possible preventive measures to reduce the prevalence of dementia or delay its onset may be a reasonable clinical strategy with important practical significance [2]. Given this, vascular factors have aroused an increasing interest from researchers, while controlling these factors provides a feasible method to prevent cerebrovascular impairment [2].

Blood pressure (BP) as one of the modifiable vascular factors was observed to be associated with dementia outcomes in various forms, such as mean measure [3], prehypertension [4], orthostatic hypotension [5], and pulse pressure (PP) [6]. Recently, accumulated evidence indicates that BP variability, which contributes to cerebrovascular damage as an independent factor of mean BP [7], is positively associated with the risk of dementia [2, 8–11]. The possible mechanism may be that a higher BP variability is correlated with increased hyperintensity lesions of white matter on brain imaging [12], increased intima-media thickness of carotid artery [13], and accelerated progression of early carotid atherosclerosis [2, 14]. However, the patterns of BP change over a certain period, which cannot be described by BP variability or any single measure, have not been comprehensively elucidated. BP trajectory is helpful for capturing the dynamic change of BP within a certain period and visualizing the direction and size of BP variability. Recent longitudinal studies suggest that subgroups with distinct trajectories of BP change may be associated with a variety of clinical outcomes such as cardiovascular events [15–18], stroke [18, 19], and all-cause mortality [17, 18, 20]. Given the close association between the cardiovascular and cerebrovascular burden and dementia [21, 22], we hypothesized that BP trajectory patterns would have predictive significance for the risk of subsequent dementia outcomes.

At present, limited results have been reported regarding the effect of BP change trajectory on incident dementia. Most of the existing studies were focused on the effects of longitudinal measurement, baseline value, or BP decline on the risk of subsequent dementia outcomes, and BP trajectory was consistently described across a number of studies based on whether the aged adults were diagnosed of dementia in a later time [23–29]. By following up with 707 middle-aged women for 37 years, the Prospective Population Study of Women in Gothenburg reported that subjects who developed dementia had a steeper decline in systolic BP (SBP) during the five years before diagnosis than those who did not [24]. Another 32-year cohort study (n=1,890) observed that male subjects who developed dementia had a greater increase in SBP followed by a greater decline after the age of 78 than those who did not [25]. A longitudinal study from the Kungsholmen Project revealed that individuals who developed dementia showed a greater decline in BP, especially in SBP, compared with those who remained free of dementia [26, 27]. In addition, the Adult Changes in Thought Study (n=2,356) reported that the mean SBP of subjects aged 64-75 who developed dementia was consistently higher and declined more sharply during the two years before diagnoses compared with those who did not [29]. However, some of the changes in BP trajectory may be caused by suffering from or progressing to dementia, which weakens their predictive significance for the development of dementia, and these changes are not formed naturally in the general population. From aforementioned studies, it is infeasible to identify meaningful subgroups of BP trajectory as well as their respective characteristics.

Moreover, the effects of BP change trajectory on the incidence of dementia in certain specific populations, such as the oldest-old (≥ 80 years old), women or men, and patients with hypertension, have not been well established. Previous studies have reported an age-varying association between mean BP and risk of dementia [29, 30]. It was found that the mean SBP was higher in subjects who developed dementia at an age lower than 75, but not in those who developed dementia after the age of 75 [29]. A lower BP was associated with a higher risk of dementia in individuals aged over 75 [30]. An earlier study suggested that there were gender differences in BP trajectory throughout the life, which might result in an inconsistent association between BP trajectory and dementia for men and women [24]. Apart from age and gender, the diagnosis of hypertension is also related to antihypertensive treatment and stable high BP levels. Existing studies have shown that antihypertensive drugs are associated with a decreased risk of dementia [31, 32]. Since high BP levels are deemed to affect the development of dementia [3], this factor needs to be well controlled. According to the general practice, only two repeated BP values are measured within one day, which may lead to errors due to various reasons. Therefore, the diagnosis of hypertension may be a better indicator for a stable BP level than the two BP measurements.

In the present study, we aimed to investigate the association between the trajectory of BP change and the risk of dementia and to explore the differences in age, gender, and hypertension subgroups using a nationally representative sample of adults aged ≥60 years. We hypothesized that BP trajectory patterns would have predictive significance for subsequent dementia events and there were differences in age, gender, and hypertension subgroups.

Results

Discussion

In this nationwide cohort study in China, stabilized SBP was found to be associated with a higher risk of dementia compared with normal SBP and elevated SBP in the total sample, but not in the subgroups of the oldest-old, women, and subjects without hypertension at baseline. Similarly, stabilized PP was associated with a higher risk of dementia compared with normal PP and elevated PP in the total sample, but not in the subgroups of the oldest-old, women, and subjects with hypertension at baseline. However, there was no strong evidence for the direct association between DBP trajectory and dementia.

In this study, we identified four unique BP/PP trajectories: normal, stabilized, elevated, and persistently high BP/PP. The proportion of normal BP in the present study was higher than that in prior studies [17, 18], probably because our subjects were enrolled from longevity areas and had fewer harmful BP changes than the general population. Consistent with our study, the Cardiovascular Health Study reported similar normal, stabilized, and elevated SBP trajectories among 4,067 subjects with a median age of 77 years [17]. The differences in age, gender, body mass index, and risk of cardiovascular and cerebrovascular disease among SBP trajectory groups were also observed as consistent as in our study [17]. Our findings suggest that the occurrence of stabilized SBP should not be considered as a complication of heart disease, diabetes, and cerebrovascular diseases, but higher body mass index resulted in an elevated and persistently high PP. The mechanisms by which these abnormal trajectories are generated need to be further studied.

Consistent with our findings, previous studies also demonstrated that SBP tended to decline a few years before the onset of dementia [24–26, 33, 34]. Besides, the SBP remained consistently at a level above the normal value before it began to decline [24–26, 33]. It was repeatedly reported that individuals with a higher SBP in midlife involved a significantly increased risk of dementia in their later life [3, 4]. Several decades-long follow-up studies demonstrated that the SBP of those subjects who developed dementia tended to increase more significantly from midlife to late-life and thereafter declined more in the years before the dementia onset [24, 25, 35]. An earlier study suggested that the trajectory of change, rather than the current BP, might be most useful in identifying a subsequent diagnosis of dementia [36], but the causal association between SBP trajectory and neurodegeneration remained unclear since dementia had a decades-long prodrome. The pathology of Alzheimer’s disease may have been observed as early as 20 years ahead of diagnosis [37]. It is undeniable that the progression to dementia may cause a decline in SBP, which needs to be confirmed by pathological studies in the prodromal stage of dementia. However, stabilized SBP still has predictive significance for the occurrence of dementia. Inconsistent with previous studies, the present research did not find any detrimental effect of persistently high or elevated SBP on dementia [29, 33]. In contrast, we found that elevated SBP was associated with a reduced risk of dementia in subjects without hypertension, heart disease, or cognitive impairment, compared with those having a normal SBP. This finding conflicted with the accumulated evidence that BP variation was positively associated with the risk of dementia regardless of the direction [8–11]. In a practical sense, SBP should not be elevated by intervention for the purpose to reduce the risk of dementia. Repeated studies are required to confirm this finding.

One mechanism may exist that mid-life hypertension and late-life hypotension independently affect dementia. Long-term high BP starting from midlife can lead to a series of cerebrovascular diseases, such as white matter damage, asymptomatic cerebral infarction, and clinical stroke. The ischemic brain injuries caused by these diseases may act alone or in combination with neurodegenerative changes in late life to promote the clinical manifestations of dementia syndrome [38]. Higher blood pressure would result in more severe cerebral atherosclerosis, which is also related to cerebral neurodegenerative diseases and the clinical manifestations of dementia [39, 40]. There may be two pathways linking low BP with dementia. First, neuroimaging studies reported that low BP was associated with more severe white matter lesions and atrophy of the hippocampus [41, 42], which significantly increased the risk of dementia [43, 44]. Second, intermittent or persistent hypotension may further damage cerebral blood perfusion and lead to more extensive cerebral ischemia, which accelerates the clinical manifestation of dementia syndrome [26]. Another mechanism may be attributed to the combined effects of mid-life hypertension and late-life hypotension on dementia. When the systemic BP is low, the impairment of brain autoregulation caused by chronic hypertension will lead to a decreased ability in maintaining stable blood flow, especially prone to the decrease of cerebral blood flow [33, 45]. The latter has been associated with pathogenic brain changes [33, 46].

This study found that, in the group of hypertension at entry, elevated DBP was associated with a higher risk of dementia than normal DBP. Inconsistent with our finding, previous studies reported an association between declined DBP and a higher risk of dementia. A recent meta-analysis supported that midlife high DBP was associated with an increased risk of dementia, while late-life low DBP was associated with an increased risk of dementia [47], which suggests that there may be a certain declining pattern of DBP affecting the risk of dementia. The Kungsholmen analyses and the Honolulu-Asia Aging Study reported that individuals who developed dementia showed a decline in DBP during the 3-year and 6-year period before diagnosis, respectively [25, 26]. Probably because only a few participants are classified into the elevated DBP trajectory group, unusually large statistical effects were detected in this study, suggesting a certain possibility of errors. Consistent with our results, the Gerontological Regional Database study observed that incident dementia cases exhibited a greater decline in PP over 5 years among the very old [48]. The mechanisms underlying the association between the decline in PP and an increased dementia risk remain unclear. A decline in PP may indicate a decrease in blood ejection and stroke volume, which can be associated with dementia through impaired cerebral blood flow [49]. Besides, the decline in PP may also be associated with cerebral vascular lesions and dysregulation of BP caused by severe cerebral atherosclerosis, which may lead to dementia [49].

In subgroup analyses, the association between stabilized SBP/PP and dementia was observed only in subjects aged ≥ 80 years but not in those aged 60-79 years. According to the existing evidence, there have been mixed results regarding the relationship between a declined SBP and dementia in the same age group. A decades-long cohort study showed that the pattern of midlife hypertension and late-life hypotension was associated with incident dementia only in the younger group [33]. The Adult Changes in Thought Study found that SBP declined more significantly in those subjects who developed dementia at a younger age, but not in those who developed dementia at an older age [29]. Consistent with our findings, the Gerontological Regional Database study reported that individuals who developed dementia exhibited a greater SBP decline in the very old subjects [48]. This may be partially explained by the age differences in the incidence of dementia. It has been well established that the incidence rate of dementia is much higher in the very old than in the young old. These findings suggest that age is likely to be an important factor for consideration when estimating the risk of BP decline on dementia.

Few studies have thus far specifically examined the gender differences in the association between SBP trajectory and dementia. Out of our expectation, statistically significant results were only observed in older women but not in older men. A 37-year cohort study confirmed that older women who developed dementia had a steeper decline in SBP during the five years before diagnosis [24]. In contrast, a 30-year cohort study showed that older men with a decreased SBP had poor cognitive performance [50]. In another 32-year prospective study, older men who developed dementia showed a greater SBP decline during the six years before diagnosis [25]. This is possibly attributed to the difference in the SBP trajectory composition between female and male groups in our study: there were more subjects with stabilized SBP in the female group than in the male group. With more subjects being classified as stabilized SBP, it was more likely to observe statistically significant results in the female group.

In our study, stabilized SBP was also found to increase the risk of dementia in subjects without hypertension, indicating that the trajectory of SBP change could affect dementia in a way independent of late-life hypertension. Besides, the effect in the hypertension group was stronger than that in the non-hypertension group, suggesting that late-life hypertension can also affect the relationship between the trajectory of SBP change and dementia. Consistent with our study, an association between a subsequent steep decline in SBP and an increased risk of dementia was reported in individuals without hypertension and/or not receiving antihypertensive therapy at midlife [34]. In subjects who developed dementia, a steeper decline in SBP during the five years before diagnosis was observed in individuals receiving antihypertensive treatment than in individuals never receiving antihypertensive treatment [24]. However, the Honolulu-Asia Aging Study reported no association between SBP trajectory and dementia in subjects who were currently on antihypertensive treatment [25], probably because this study only examined male subjects. In addition, stabilized PP was found to increase the risk of dementia in comparison with normal and persistently high PP only in the hypertension group, suggesting that late-life hypertension could affect the association between PP trajectory and dementia. In the non-hypertensive group, normal PP was found to increase the risk of dementia in comparison with elevated PP, which requires more comprehensive research to confirm.

One of the strengths of our study lies in the use of a large and nationally representative sample, which equips our results with a higher statistical power. The larger sample also provides the possibility of subgroup analyses. Second, our sensitivity analyses showed similar results when subjects involving a high risk of dementia were excluded, providing further evidence on the robustness of our findings. Third, to the best of our knowledge, this is the first study that explored the relationship between the trajectory of PP change and subsequent dementia. Fourth, this study determined the BP trajectory patterns up to 6-year follow-up among older adults free of dementia at entry. Compared with the previous researches in which BP trajectories were grouped based on whether dementia occurred or not, our study provided stronger evidence for the association between the decline in SBP from hypertension and the subsequent dementia.

Nevertheless, potential limitations should be acknowledged as well. First, since the median length of follow-up was only 6 years in this study, we could not confirm whether stabilized SBP/PP acted as a consequence of neurodegeneration or a risk factor for later dementia. Secord, because of data unavailability, this study was unable to detect any associations between BP trajectory and different types and degrees of dementia. Previous studies have found that certain specific BP characteristics had a unique effect on the vascular- and Alzheimer-specific pathology [51]. Individuals with the most advanced dementia were found to have the lowest BP [25]. Third, this study only measured the BP trajectory at three-time points. If more time points were covered, the measurements of BP trajectory would be more accurate. Fourth, dementia was assessed based on self- or proxy-reported hospital diagnosis, but it had not been confirmed by a doctor’s examination. Fifth, as a survival sample, the survival bias might affect some of our findings. Both hypertension and dementia can influence the risk of mortality, and thus people with baseline hypertension might have been excluded as participants due to death. Therefore, this study might underestimate the association between BP trajectory and dementia. Sixth, most of the measurements were reported by participants or their proxies, so the possibility of underreporting or misreporting might exist due to recall bias. Seventh, measurement bias might occur due to possible incorrect and missed diagnoses of dementia, as well as the lack of BP monitoring data. Finally, the effect of specific antihypertensive drugs on the risk of dementia could not be adjusted due to data insufficiency.

In the present nationwide cohort study, stabilized SBP was found to be associated with a higher risk of dementia only in the groups of the oldest-old, women, and subjects without hypertension at baseline, in comparison with normal SBP and elevated SBP. It was also found that stabilized PP was associated with a higher risk of dementia only in the groups of the oldest-old, women, and subjects with hypertension at baseline, in comparison with normal PP and elevated PP. These findings suggest that stabilized SBP and PP in late life have predictive significance for the occurrence of dementia. Age, gender, and late-life hypertension should be comprehensively considered when estimating the risk of BP decline on dementia. Further follow-ups from middle to late life are required to reveal the associations between the whole process of BP trajectory and dementia.

Materials and Methods

Supplementary Materials

Supplementary Figures

Supplementary Tables 1, 2, and 3

Supplementary Table 4

Supplementary Tables 5 and 6

Supplementary Table 7

Supplementary Tables 8 to 12

Supplementary Table 13

Supplementary Table 14