(2014) Antihypertensive treatment in people with dementia. Journal of the American Medical Directors Association, 15 (9).

The range and magnitude of potential benefits and harms of antihypertensive treatment in people with dementia has not been previously established. Method A scoping review to identify potential domains of benefits and harms of antihypertensive therapy in people with dementia was undertaken. Systematic reviews of these domains were undertaken to examine the magnitude of the benefits or harms. Results Potential outcome domains identified in the 155 papers in the scoping review were cardio-vascular events, falls, fractures and syncope, depression, orthostatic hypotension, behavioural disturbances, polypharmacy risks, kidney problems, sleep problems, interactions with cholinesterase inhibitors and pain. The systematic reviews across these domains identified relatively few studies done in people with dementia, and no convincing evidence of safety, benefit or harm across any of them. There is no justification for materially different guidance for the treatment of hypertension in people with dementia, but sufficient evidence to warrant particular caution and further research into treatment in this group of patients.


Introduction
The WHO 2012 report "Dementia: a public health priority" estimated that there are 35.6 million people with dementia living at present worldwide 1 . About half of people over 50 years of age have hypertension 2 . While the protective effects of antihypertensive treatment against cardiovascular events have been established in numerous clinical trials for different age groups [3][4] including the very old 5 , people with dementia have been consistently excluded from these studies. The favourable benefit to risk ratio seen in studies of hypertension treatment in fit older people might not apply if people with dementia are at a higher risk from the side effects of anti-hypertensive treatment than people with normal cognition (for example through falls or other adverse drug effects), and if they are less likely to be recognised or reported than in those with normal cognition. Therefore management of hypertension in people with dementia may need to be different from in those with normal cognition 6 .
However, current guidelines for the treatment of hypertension do not provide specific advice for people with dementia.
There have been no reviews to comprehensively examine studies investigating the range of outcomes (including harmful effects) of antihypertensive treatment in people with dementia. Such work is important to clarify whether there is a case for altered guidance for the management of hypertension in people with dementia.

Methods
A two stage process was undertaken. In the first stage, a scoping review 7 was undertaken to identify the range of outcomes of antihypertensive treatment in people with dementia. In the second stage, systematic reviews were undertaken for each of the main areas of harmful or adverse outcomes identified in the first stage.

Stage 1: scoping review Protocol
The scoping review was based on a pre-defined protocol to search and identify relevant research articles.

Eligibility
Inclusion: All articles reporting original research regarding the treatment of hypertension in people with dementia were included. No publication date limit was applied Exclusion: All studies investigating hypertension as a risk factor for dementia; effects on carers; specific non-dementia diseases (e.g. cancer) and rarer forms of dementia (Korsakoff's syndrome, dementia due to HIV, normal pressure hydrocephalus, CADASIL); animal studies and non-English language articles. As effects of antihypertensive drugs on the progression of dementia have been reviewed elsewhere [8][9] , they were excluded from the literature analysis.

Information sources
Pubmed, Embase, Web of Science and the Cochrane library databases were searched.

Search
The search for articles took place from July 2012 and was updated in February 2013. Searches were limited to English language human research articles. Search terms included combinations of 'hypertension', 'antihypertensive drug', 'antihypertensive treatment' or 'blood pressure' in combination with 'dementia' or 'Alzheimer's disease'

Selection
Duplicates were removed, titles and abstracts were examined by one person (Vv/dW) and inclusion and exclusion criteria applied. Full text versions of the remaining papers were obtained and assessed again based on inclusion and exclusion criteria. The remaining relevant articles were used to identify the range of outcomes examined in scientific articles.

Data collection
Main outcome variables were recorded and categorized into topics.

Analysis
The number of papers examining each outcome was tabulated.

Stage 2 systematic reviews Protocol
The systematic reviews were based on a pre-defined protocol to search and identify relevant research papers.

Eligibility
Inclusion: All articles reporting original research regarding antihypertensive treatment in people with dementia and the topic identified in stage 1 were included. No publication date limit was applied.
Exclusion: As in stage 1. Studies which did not include antihypertensive treatment in any form in the analysis (as composite variable or individual medication classes) were also excluded.

Information sources
Pubmed, Embase, Web of Science and the Cochrane library databases were searched.

Search
Articles were searched until June 2013 publication dates. Searches were limited to English language and, where possible, human research articles. Search terms for stage 2 included 'antihypertensive' or antihypertensive drug classes ('diuretics', 'beta blockers', 'calcium channel blockers', 'angiotensin receptor blockers' and 'ACE inhibitors') in combination with 'dementia' or 'Alzheimer's disease' and the topics identified in stage 1 (see flow diagram 1).

Selection
Duplicates were removed, titles and abstracts were examined by one person (VvdW) and inclusion and exclusion criteria applied. Full text versions of the remaining papers were obtained and assessed based on inclusion and exclusion criteria. The remaining relevant articles were included and the number of papers examining each topic was recorded.

Data collection
Methods and results were extracted to record purpose of the study, design, sample size, main measurements and results of each study.

Analysis
Method sections of papers were examined to determine level of evidence. Results were compared and analysed regarding the effect of antihypertensive treatment on the topic in question. Figure 1 shows that 155 papers entered this review. Figure 1 Scoping review  Table 1 shows a summary of the findings of the searches undertaken in the systematic reviews of the topics identified in the scoping review.

Cardio-and cerebrovascular events
In total, four studies examined the effects of antihypertensive medication on cardio-and cerebrovascular events in people with dementia (see table 2), with antihypertensives either as main independent variable (three studies) or as control variable (one study).
No randomized controlled placebo trials have been conducted in this area. In people with dementia, two randomized controlled trials investigated the effects of calcium channel blockers (CCBs) in an intention-to-treat, a efficacy and a safety analysis [10][11][12] . For both trials, main outcome was cognition but the safety analysis included cardio-and cerebrovascular events, which were reported as adverse events. The studies compared a control group against a treatment group using CCBs. Both groups were allowed to continue with unspecified antihypertensive treatment as prescribed before the beginning of the study except alpha methyldopa. The results showed that the control group had significantly more adverse (cerebrovascular) events than the treatment group [11][12] . For cardiovascular events the trend was similar but reached significance in only one of the studies 12 . In a subgroup analysis of participants with sub-cortical vascular dementia 13 , no significant differences in these adverse events were found, although again the same trend was observed (fewer events in the treatment group). One study was a non-randomized six weeks open label trial including people with dementia into the safety of Nivladipine (CCB) 14 .
There was no significant difference in adverse events, but this was a very short period of follow up.
In addition, a prognostic cohort study including a large group of people diagnosed with dementia investigated the effects of angiotensin receptor blockers and ACE inhibitors on the risk of hospitalization and mortality while controlling for cerebro-and cardiovascular events 10 . The results showed that while there was no significant association between angiotensin receptor blockers and risk for hospitalisation or mortality, ACE inhibitors significantly increased the risk for mortality but not for hospitalisation. This association remained significant after controlling for coronary heart disease, stroke and congestive heart failure.

Falls, fractures and syncope
The risk of falls and fractures is higher in people with dementia compared to those without [15][16][17] . In addition, some antihypertensive treatments have been shown to increase the risk for falls and fractures in older people [18][19] . However, this review found no evidence from randomized controlled trials for the risk of falls due to antihypertensive medication in people with dementia. This review identified four prospective cohort studies, two of those investigating the association between falls risk and antihypertensive medication in people with dementia [20][21][22] , and a further two examining the association of falls risk with cardiovascular medication [23][24] (Table 2). Cardiovascular medication included antihypertensive drugs as well as anticoagulants and antiplatelet agents.
The findings showed conflicting results in people with dementia. Examining risk factors for falls in people with and without dementia, Eriksson et al. [20][21] found that taking diuretics was a significant risk factor for falls for people with dementia in a univariate analysis, but this was non-significant when controlled for significant covariates, but taking more than four drugs was a significant risk factor for falling in people with dementia after controlling for significant covariates. Pellfolk et al. 22 found no significant difference in the use of antihypertensive medication between fallers and non-fallers in people with dementia and therefore did not enter antihypertensive medication in their logistic regression analysis of independent risk factors of falls. However, Allan et al. 23 reported that cardiovascular medication was a significant predictor of falls in an analysis of all types of dementia combined (Alzheimer's disease, vascular dementia, dementia with Lewy bodies and Parkinson's with dementia), as well as when stratified per dementia type. When significant and potentially modifiable predictors were entered into a multivariate analysis, which was stratified by dementia type, cardiovascular medication remained significantly associated with falls. Asada et al. 24 , on the other hand, found that cardiovascular medication was not a significant contributor to falls-related injuries in people with dementia.

Behavioural problems
Only the effects of beta (Propranolol/Pindolol) and alpha blockers (Prazosin) on behavioural symptoms in people with dementia have been investigated in randomised controlled trials ( Table 2).
All three trials had very small sample sizes (between 6 and 31 participants) but showed a significant beneficial effect of centrally active antihypertensive medication on behavioural problems. Herrmann et al. 25 showed that verbal aggression was significantly reduced in people using beta-blockers, without a significant change in total aggression. Peskind, Tsuang and Bonner 26 found no significant association between use of beta blockers and any specific behavioural subscale but a significant effect of beta blocker use on the total behavioural score. A small randomised controlled trial to assess the effect of alpha blockers on behaviour showed that non-significantly greater improvements in the treatment group than in the control group 27 . A case series study to assess the effect of beta blockers on behaviour in people with dementia 28 indicated significant improvements for all behavioural measures (except for wandering, which improved on only one scale). Behavioural improvements after using beta blockers were also found in another case study including people with dementia 29 . Meta-analyses showed that beta blockers have relatively little effect on cerebro-and cardiovascular events and mortality compared to other antihypertensive drugs [30][31] , and are more likely to be discontinued 32 .
Other antihypertensive treatments have not been investigated for their effect on behavioural problems.

Depression
Antihypertensive treatment, in particular the use of highly lipid-soluble beta blockers, has been associated with depression 33 . One randomised controlled trial and one cohort study investigated the relationship between antihypertensive treatment and depression in people with dementia. The randomised controlled trial investigated the effect of Nimodipine (CCB) on depression in patients with diagnosed dementia as well as those with mild to moderate cognitive decline on the Global Deterioration scale 34 . The treatment and control group were allowed to use other medication including antihypertensive treatment. Scores on the Hospital Anxiety and Depression scale improved significantly more in the treatment group than in the control group. In a large cohort study examining the prevalence of depression in people with dementia, Newman 35 found that beta blocker use was not a significant covariate in the relationship between dementia and diagnosis of major depression. No other antihypertensive drugs were included in the analysis.

Sleep problems, pain, kidney problems and incontinence
Although these topics emerged during the first stage of this review, there were only a very few studies investigating the effect of antihypertensive medication in people with dementia ( Table 2).
Sleep problems are common in people with dementia [36][37] , and sleep disturbances, in particular obstructive sleep apnoea, have been shown to improve with antihypertensive treatment 38 . While there are no studies examining the effects of antihypertensive treatment on sleep quality, one prospective cohort study investigated the relationship between sleep disturbances and nocturnal reduction of blood pressure in people with dementia whilst controlling for antihypertensive treatment 39 . The results showed no significant difference in use of antihypertensive medication between dippers (hypertensive patients who experience a nocturnal reduction in blood pressure) and non-dippers (those who do not).
Other sleep disturbances have not been investigated in people with dementia in relation to use of antihypertensive medication.
Some antihypertensive drugs such as beta blockers, angiotensin receptor blockers and ACE inhibitors seem to have beneficial effects on pain in older people without dementia [40][41] . For people with dementia, however, the association between pain and antihypertensive treatment has not yet been investigated.
The relationship between kidney problems and antihypertensive medication has not been assessed in people with dementia. One cohort study including geriatric emergency patients investigated the association. The results showed that antihypertensive drugs may contribute to an increased risk of a clinically relevant impairment of renal function 42 . However, although 56% of the participants were diagnosed with dementia and no separate analysis for this group was conducted.
The effect of antihypertensive drugs on incontinence has been examined by Ruby et al. 43 in a large cohort study of older people. They reported that use of alpha blockers was significantly associated with urinary incontinence. Ridgeway et al. 44 investigated the relationship between incontinence and use of antihypertensive medication in a long-term care population with 79% being diagnosed with dementia. The results indicated that while dementia was significantly associated with urinary and faecal incontinence, antihypertensive medication was not individually related to incontinence.
However, a sub-analysis including only people with dementia was not reported.

Use of Cholinesterase inhibitors
Cholinesterase inhibitors are commonly used in the mild to moderate stages of Alzheimer's disease.
Several studies investigated the interaction effects between cholinesterase inhibitors and antihypertensive medication (Table 2) although the evidence is limited to cohort studies and population database studies, no randomised controlled trials have been conducted to assess these effects.
The results of a prospective cohort study 45 showed significant better cognition in those taking antihypertensive drugs compared to those not taking them. Other studies, however, demonstrated that the joint use of antihypertensive medication and cholinesterase inhibitors was associated with a significantly increased risk of serious adverse drug reactions. Grossberg et al. 46 found no significant interaction between rivastigmine and antihypertensive medication as a group in their cohort study but when analysed separately, the use of diuretics and rivastigmine was related to a significantly higher rate of serious adverse events while centrally acting antihypertensives, ACE inhibitors and calcium channel blockers were not. Furthermore, the results of two database studies showed that use of antihypertensive treatment was significantly associated with serious adverse drug reactions in patients treated with cholinesterase inhibitors 47 and beta blockers were among the most frequently encountered drugs involved in drug-drug interactions with cholinesterase inhibitors 48 .
These studies indicate that the combination of cholinesterase inhibitors and antihypertensive medication might result in benefits (improvement of cognition) as well as harms (serious adverse drug reactions), which might depend on drug class.

Polypharmacy and inappropriate medications
All antihypertensive medications contribute to polypharmacy and its harmful effects 49 , but only two cohort studies investigated antihypertensive drug use in the context of polypharmacy in people with dementia (

Orthostatic hypotension
Orthostatic hypotension is associated with poorer cognition [52][53] and occurs significantly more often in people with dementia than without 54 . However, the role of antihypertensive medication in this context is unclear. Again, no randomised controlled trials have been conducted to examine the relationship between antihypertensive treatment and hypotension in people with dementia. The non-randomised open label trial, which looked at safety of Nivladipine for people with dementia 14 , found no significant difference between treatment and control group in the incidence of orthostatic hypotension.
In a cross-sectional cohort study, Anderson et al. 54 found that differences in incidence of orthostatic hypotension between people with Alzheimer's disease, Lewy Body dementia and without dementia were independent of the use of antihypertensive treatment. Mehrabian et al. 55 confirmed these results in a cohort study including people with memory complaints including dementia when looking at use or no use of antihypertensive treatment (Table 2). While worse cognitive function and number of different antihypertensives were significantly associated with orthostatic hypotension, it was not reported if the number of different antihypertensive drugs was significantly related to orthostatic hypotension for the dementia group separately. Patients with OH used a sig. higher number of AHT than those without OH. OH patients had sig. poorer cognitive function. This was independent of AHT. OH was significantly more often present in VaD, AD, MCI than in cognitively non-impaired individuals in an overall test (no individual differences or effect sizes reported).

Discussion
The This review could be subject to a publication bias towards positive outcomes and benefits, and against the reporting of adverse consequences of treatment, which means that the possible benefits the review identified of antihypertensive treatment should be interpreted with caution.
This review adds to the little that is known specifically about the treatment of hypertension in people with dementia. Previous reviews have proposed but not proven that such treatment might delay the onset of dementia [8][9] . Another review concluded that there is no convincing evidence of the effectiveness of antihypertensive therapy in the reduction of cardiovascular and cerebrovascular events in people with dementia 56 , which our review has confirmed. This review, however, illustrates the range of potential harms and benefits of antihypertensive therapy in people with dementia beyond those mentioned in most trials of antihypertensive agentsalthough it found little firm evidence about them.
Given the lack of firm evidence of increased adverse outcomes arising from antihypertensive therapy in people with dementia, there is currently no reason to produce materially different guidelines for the treatment of hypertension in people with dementia, and justifies the current presumption that the favourable evidence drawn from the treatment of non-demented people should be extrapolated to those with dementia. As part of this guidance, the side effects of antihypertensive drugs should be sought on a routine basis: those looking after hypertension in people with dementia may need to gather a collateral history to do so adequately. Our review does not contradict guidance about prescribing in general in frailty states, which would apply to many people in the later stages of dementia: this includes withdrawing preventative medication in those believed to be in their last months of life when the potential benefits diminish in magnitude, and confirming whether the consent to treatment made by patients before the onset of dementia is still valid when they have lost mental capacity 57 . Our review also should not discourage physicians caring for acutely ill people with dementia from withdrawing antihypertensive drugs temporarily while they hare cardiovascular instability or are at high risk of problems such as acute kidney injury.
Nevertheless, the lack of evidence of harm or benefit in this group means that there is at least a concern about the safety of antihypertensive treatment and a gap in the evidence base. Given the growing numbers of people with dementia and hence with dementia and hypertension, and growing concerns over both polypharmacy and frailty, we propose that clinicians should audit carefully their treatment of hypertension in this group of patients, trials of safe withdrawal of antihypertensive therapy should be considered for this population (as have been done for the general population) [58][59] taking the wide range of outcomes identified in this review into account, and more work is required to clarify the factors that should determine when antihypertensive therapy should be withdrawn in the later stages of the dementia process.