Long-Term Efficacy and Toxicity of Cholinesterase Inhibitors in the Treatment of Alzheimer Disease

Abstract

Though the symptoms of Alzheimer disease go on for years, the phase 3 trials of the cholinesterase inhibitors (ChEIs), the current mainstay of symptomatic pharmacotherapy for this condition, were typically of only 3- to 6-months’ duration. We have limited data on long-term (that is, a year or more) therapy with these agents. In this review, we explore the available information on the biological and clinical effects of long-term ChEI therapy, what happens when these agents are discontinued, and examine what others have recommended. An individualized approach to deciding on whether to carry on with a ChEI should be taken. If continued, treatment goals should be clarified and patients monitored over time, for both drug-related benefits and adverse effects.

Cholinesterase inhibitors, such as donepezil, galantamine, and rivastigmine, are the most commonly used medications for the symptomatic treatment of AD in Canada. In 2012, ChEIs were the seventh most costly agent covered by publically funded drug benefit programs for older Canadians, with $129.4 million spent on them.¹ RCTs indicate that their use may lead to modest cognitive and global benefits.²

AD is a progressive form of dementia that goes on for years after the onset of symptoms.³ For this condition, as with many other chronic diseases, it is not uncommon for pharmacotherapy to be taken for a longer period of time than the duration of the trials that led to marketing approval. Most of the phase 3 RCTs of ChEIs were 6 months or less in duration,² which is shorter than the average time span that these drugs are taken by patients with AD in Canada.^4,5

The 2008 Clinical Practice Guidelines on the pharmacologic treatment of dementia of the American College of Physicians and the American Academy of Family Physicians concluded there was insufficient evidence on which to determine the optimal duration of therapy with ChEIs and that research on this topic was needed.⁶ Likewise, the second list of Five Things Physicians and Patients Should Question developed by the American Geriatrics Society notes that the risks and benefits of long-term therapy with ChEIs are not well-established.⁷ Notwithstanding this uncertainty, we have to advise our patients and their families on how long treatment should persist and to respond when they ask about the relative benefits and safety of long-term use of ChEIs. While at times dubious about their clinical utility, family members of people with a dementia often fear that stopping them may be accompanied by deterioration in the clinical status of their relative.⁸

To help address this issue, a qualitative review of the available literature on the relative benefits and risks of long-term (for this review, defined as 1 or more years) ChEI therapy was performed. The main themes relevant to the topic were identified followed by purposive (nonexhaustive) searches of the relevant literature. The papers referred to in this review were either previously known to the author or identified by a series of PubMed word searches. The focus was on the biomedical effects of long-term use. A pharmacoeconomic analysis, based on an examination of available data, was not done, though it should be noted that their current availability as generics has lessened (though not eliminated) concerns about the cost-effectiveness of these agents.

Biological Effects of Long-Term Cholinesterase Inhibitor Therapy

In 1976, Davies and Maloney⁹ suggested that AD arose from a “cholinergic system failure.”^{p 1403} This was based on finding a marked reduction in the activity of enzymes involved in the metabolism of ACh within the brains of patients dying with AD. This observation, coupled with data from experimental studies showing an important role for ACh in learning and memory, led to the cholinergic hypothesis for age-associated memory loss and AD.¹⁰ This proposes that degeneration of cholinergic neurons in the basal forebrain leads to the loss of cholinergic neurotransmission in the cortex, hippocampus, and other areas of the brain that, in turn, contributes to the impaired cognition seen. Just as Parkinson disease was considered a dopaminergic deficiency condition, AD was conceptualized as a cholinergic one.¹¹ This had an important influence on early AD drug development and led to the identification, study, and eventual approval of ChEIs for the treatment of AD.^11,12 The beneficial effects of ChEIs are still felt to arise primarily from their ability to improve cholinergic neurotransmission by slowing the degradation of ACh by the enzyme AChE.

While the ChEIs are approved as symptomatic forms of therapy for AD, some feel that they may also have disease-modifying (that is, exert an impact on the underlying pathophysiology of the disorder) effects.^13,14 Numerous mechanisms have been proposed for how this may occur, including the antiinflammatory properties of cholinergic upregulation, effects of these agents on AChE isoform and nicotinic receptor expression, reduced amyloid precursor protein processing, attenuation of beta-amyloid toxicity, and maintenance of ACh-mediated compensatory pathways.^15–17 Unfortunately, convincing proof that ChEIs have a neuroprotective effect in patients with AD is lacking, and it appears that these agents have limited, if any, disease-modifying activity.¹⁸ In support of this is the lack of any significant changes in CSF AD biomarkers (that is, amyloid beta 1–42, total tau, and phosphorylated tau) among patients with this form of dementia treated for 6 to 12 months with ChEIs.^19–21 An autopsy study suggested that, if anything, the use of these agents may accelerate disease progression. There was significantly greater accumulation of phosphorylated tau in the cerebral cortex of patients treated with ChEIs, compared with untreated ones.²² Though this finding requires verification, the authors of the report speculated this may help explain diminishing responsiveness to these agents with long-term therapy.

Highlights.

• Long-term therapy with ChEIs is associated with a small but statistically significant increased risk of bradycardia and syncope and their consequences (for example, pacemaker implantation and hip fracture).

• Discontinuing these agents can lead to the loss of beneficial cognitive and functional effects that may not have been evident.

• Decisions about whether to continue therapy should be individualized. There is general agreement that consideration of stopping should take place when the patient’s dementia has progressed to a severe stage.

Treatment with rapidly reversible ChEIs, such as donepezil and galantamine, is associated with significant upregulation of AChE activity in the CSF of patients with AD, while use of rivastigmine, a pseudo-irreversible ChEI, is associated with a significant decrease in both AChE and butyrylcholinesterase activity.^21,23 These differing effects have not been correlated with either favourable or adverse clinical outcomes.

Clinical Effects of Long-Term Cholinesterase Inhibitor Therapy

Data from the RCTs of ChEIs provide little information about their long-term effects. As previously noted, most of these studies were of short (3 or 6 months) duration.² As these results were considered sufficient for regulatory approval of the ChEIs, there was little incentive for pharmaceutical firms to do longer studies. While there are a handful of 1-year-or-more, placebo-controlled trials of ChEIs for patients with AD,^24–28 their interpretation is complicated by restricted inclusion criteria, not surprising relatively high (that is, 25% or more) discontinuation rates, and (or) questionable statistical analyses (for example, using last observation carried forward to deal with missing data). Even if funding could be found, ethical and feasibility considerations now mean it is unlikely that 1-year-or-longer RCTs of ChEIs could be conducted in our country.

There are numerous open-label extension and observational studies that have followed treated patients for often more than a year. These studies have been used to look for evidence of long-term efficacy and safety. The former would be assessed by comparing the cognitive and (or) functional outcomes of patients who have continuously taken a ChEI with historical control subjects, predictions from mathematical models,²⁹ or contemporaneous patients who had never been treated, used the drugs briefly, or stopped taking them. These so-called real-world studies almost always report benefits and good tolerability with prolonged therapy.^30,31 Because these reports are uncontrolled, with drop-out and survivor bias, their results have to be interpreted with caution as there is substantial risk that erroneous conclusions about both efficacy and safety will be drawn.^32–36 As an example of the reduction that can be seen in the number of subjects from enrolment to completion of an open-label study, 286 participants were randomized at the start of the double-blind component of the Donepezil Nordic study, 157 (54.9% of all those randomized) entered the open-label phase, and 114 (39.9%) completed the full 2 years of it.³⁷ Open-label extension and observational studies do have their defenders, who feel they provide useful and otherwise unavailable long-term information.^30,31,38 As a minimum, they show that some patients can safely take these drugs for a long period of time.

In opposition to the generally favourable results reported with observational studies, the Alzheimer’s Disease Neuoimaging Initiative and the Australian Imaging, Biomarkers and Lifestyle studies both showed greater cognitive decline among participants with mild cognitive impairment or AD taking a ChEI, compared with not receiving treatment.^39,40 As with the favourable studies, the known and unknown potential sources of biases undermine our ability to draw any firm conclusions from these results. For example, clinicians may have selectively prescribed ChEIs to patients with more aggressive disease.³⁹

While adverse effects from these agents generally arise during initiation and upward drug titration, several studies using administrative data have raised safety concerns with long-term use. For example, the prescription of ChEIs is associated with an increased risk of receiving an anticholinergic drug to manage urinary incontinence.⁴¹ This is felt to be an example of a prescribing cascade, where the first agent causes an adverse drug effect (that is, urinary incontinence) that is treated by the prescription of a second medication to deal with the drug-induced problem. The concurrent prescription of procholinergic and anticholinergic agents, which could blunt each other’s therapeutic effects leading to a pharmacological stalemate, is a particular concern with this specific combination of drugs.⁴² Of greater worry, though, is the finding that the prescription of ChEIs is associated with a small absolute but statistically significant increased risk of bradycardia leading to hospitalization, syncope, pacemaker insertion, and hip fracture (presumably from presyncopal and [or] syncopal falls).^43–45 Adverse cardiovascular effects are not unexpected, as ChEIs can induce sinus bradycardia, sinoatrial block, and aggravate preexisting sinus node disease and atrioventricular blocks. A meta-analysis of RCT results found an increased risk of syncope with their use.⁴⁶ There have also been several case reports of QT interval prolongation with torsades de pointes ventricular tachycardia with ChEIs.⁴⁷ Patients on ChEIs should be routinely asked about presyncope and (or) syncope and have their pulse checked for bradycardia.⁴⁸ An electrocardiogram should be done if the patient has been prescribed another drug that could prolong their QT interval (for example, antiarrhythmic, antihistamine, amtimicrobial, tricyclic antidepressant, and neuroleptic) or if some other factor that may increase the time required for ventricular repolarization (for example, organic heat disease and metabolic abnormalities) is present.^49,50

Cohort studies have allayed some of the worries about ChEIs. These agents could theoretically lead to or aggravate bronchoconstriction. While a small study of hospitalized patients on a geriatric unit receiving ChEIs suggested an increased risk for pulmonary disorders,⁵¹ a much larger population-based one found no evidence of an increased likelihood of serious pulmonary complications in older people treated with a ChEI who had concomitant chronic obstructive airway disease.⁵² These drugs also may increase gastric acid production, but a recent report found that upper gastrointestinal bleeding was not associated with the use of ChEIs.⁵³ Though reassuring at a population level, this does not mean that individual patients will not encounter these problems if treated with these agents.

Involuntary weight loss has attracted attention as a potential adverse effect of long-term use of these agents. While gastrointestinal side effects, including anorexia and weight loss, were relatively common in the RCTs of ChEIs, interpreting a change in weight during long-term therapy with a ChEI is difficult because weight loss commonly occurs in untreated AD.⁵⁴ Though highlighted as a potential concern in a case series,⁵⁵ 2 observational studies did not find a greater risk for weight loss in patients with AD receiving a ChEI.^56,57 This is another area where individual patients may be at risk. Weight should be monitored during long-term therapy.

The above adverse effects were predictable from the known pharmacology of ChEIs. Postmarketing surveillance is needed to help identify unpredictable and rare ones. Though causality is not confirmed, one that has been suggested as arising from ChEI use is the Pisa syndrome (that is, abnormal flexion of the body and head to one side accompanied by slight axial rotation of the trunk).⁵⁸ This is felt to arise from a dopaminergic–cholinergic imbalance.

An important factor influencing the tolerability of ChEIs is the use of other drugs. Drug–drug interactions were present in about one-third of spontaneous reports of adverse effects with ChEIs in France.⁵⁹ The most frequently encountered interactions were with medications that, on their own, can lead to bradycardia (for example, beta blockers) and anticholinergics. In Canada, nearly two-thirds of people 65 and older are prescribed 5 or more medications during a year, with one-quarter receiving 10 or more.¹ Safe prescribing of ChEIs requires being knowledgeable of all the medications the patient may be taking and being circumspect about prescribing additional drugs.

Effects of Stopping Long-Term Cholinesterase Inhibitor Therapy

There are a few case reports of patients with AD experiencing withdrawal symptoms after stopping a ChEI.^60,61 A small open-label study of patients with either Lewy-body or Parkinson disease dementia found that abrupt withdrawal of donepezil produced acute cognitive and behavioural decline.⁶² A retrospective cohort study of nursing home residents who discontinued a ChEI found they were significantly more likely than those who carried on with therapy to show aggression and (or) repetitive behaviour. They also spent significantly less time on leisure activities.⁶³

Two double-blind RCTs provide information on the effects of discontinuing a ChEI after prolonged use in patients with AD. One was a 24-month, double-blind, randomized, placebo-controlled withdrawal trial that followed a 12-month, open-label study of galantamine.⁶⁴ Patients randomized to placebo were more likely to discontinue the double-blind phase prematurely (overall and specifically because of a perceived lack of efficacy), but were no more likely to show a 4+ decline on the Alzheimer Disease Assessment Scale—cognitive subscale (commonly referred to as ADAS-cog). As well, there was no significant difference in the Clinical Interview Based Impression of Changes—Plus Caregiver Input (commonly referred to as CIBIC-plus) scores of the 2 groups. In the Donepezil and Memantine in Moderate to Severe Alzheimer’s Disease (commonly referred to as the DOMINO-AD Trial) study,²⁸ participants with moderate-to-severe AD treated with donepezil for at least 3 months (87% had been on donepezil for 12 or more months) were randomly allocated to continuing donepezil, stopping it, stopping it and starting memantine, or continuing donepezil and starting memantine. After a year, participants who carried on with donepezil scored, on average, 1.9 points higher on a standardized MMSE, compared with those who stopped it and did 3.0 points better on the Bristol Activities of Daily Living Scale (commonly referred to as BADLS). There is an ongoing discontinuation trial of ChEIs that should shed further light on the issue.⁶⁵

If therapy is discontinued, the dose should be tapered before stopping, and the patient should be monitored during the next 1-to-3 months, with consideration given to reinstating therapy if there is a significant deterioration in the their condition.⁶⁶ An older open-label study suggested that ChEI treatment effects that are lost after prolonged (6-week) discontinuation are not fully recovered when drug treatment is restarted.⁶⁷ The validity and clinical importance of this observation is debatable and should not, in itself, dissuade the practitioner from consideration of stopping therapy. Reassurance that permanent harm is not being done to patients by temporarily stopping ChEIs to see if they are benefiting from therapy comes from a Québec population-based observational study that found the occurrence of treatment gaps of 6 or more weeks during the first year of therapy with a ChEI among those who resumed treatment was not associated with an increased risk of institutionalization or death.⁶⁸

Recommendations of Others

A few years ago, Canadian dementia experts were surveyed about their perceptions on discontinuing ChEIs. There was near unanimous (96% of respondents) agreement that these agents should not be discontinued after a year because of the lack of long-term data.⁶⁹ Most felt that they could not specify a length of time at which point these drugs should be discontinued. In other words, duration of therapy was not viewed as a stand-alone criterion for stopping.

As AD is a progressive illness, over time patients treated with symptomatic drugs, such as ChEIs, will unfortunately continue to decline. In deciding whether to discontinue because of a perceived lack of efficacy, a challenge is trying to estimate the expected rate of decline on cognitive and (or) functional measures if the patient was untreated. Focusing on the MMSE as an example of a cognitive measure that may be used to try and make this determination, a meta-analysis reported that the overall annual rate of change in the MMSE among untreated patients with AD was 3.3 (95% CI 2.9 to 3.7).⁷⁰ Declines on even brief cognitive measures, such as the MMSE, though, are not linear. In general, faster deterioration is seen at a moderate degree of cognitive impairment, compared with milder or more severe deficits, and the rate of change is influenced by various other factors.²⁹ There is even evidence of a secular trend. Subjects assigned to the placebo arm in more recent RCTs of AD showed slower rates of cognitive decline, compared with older studies.⁷¹ Finally, it has to be noted that the overall annual rate of change in the MMSE is close to the standard deviation of the measurement error for the test.⁷² These factors make it difficult to come up with an accurate estimate of expected MMSE decline, to compare it with what was seen, and to decide if there are meaningful differences at the level of an individual patient. While uncertainty on how best to determine greater-than-expected decline was evident in the survey, most respondents agreed that progression to a severe stage (that is, Global Deterioration Scale stage 7, development of swallowing difficulties) should trigger consideration of stopping.⁶⁹

Recently published consensus recommendations and clinical practice guidelines^66,73–75 have advocated for an individualized approach,⁷⁵ with a search for evidence that continued therapy is having a worthwhile effect.⁷⁴ There was support for stopping when the patient has progressed to a severely impaired stage (for example, to profound disease or Global Deterioration Scale stage 7).^66,73

Conclusions

Decisions about whether to continue a ChEI should be individualized and not arbitrarily based on factors such as a MMSE score, length of treatment, or being institutionalized. It goes without saying that deliberations about when to start, continue, and stop these agents must not detract from other aspects of the care that should be provided to a patient with dementia and their family.⁷⁶ At this stage in the life cycle of ChEIs, it is unlikely that additional long-term controlled trials will be conducted.

Though this can be seen in untreated patients,⁷⁷ we often ascribe beneficial effects to ChEI therapy if the patient’s dementia appears to be progressing slowly (however defined) while on long-term therapy. In this situation, treatment is generally continued as long as the patient (or their proxy decision maker, if they lack capacity to consent to health care) is willing to continue with the agent, remains reasonably adherent, does not have comorbidities that would make continued therapy unacceptably risky or futile, is not having intolerable adverse drug effects, and has not progressed to a severely impaired stage.⁶⁶ If one persists with the ChEI, treatment goals should be agreed on with the patient and their family. Patients would then be seen periodically to monitor for both perceived benefits and adverse effects.⁷

Abbreviations

ACh

acetylcholine

AChE

acetylcholinesterase

AD

Alzheimer disease

ChEI

cholinesterase inhibitor

CSF

cerebrospinal fluid

MMSE

Mini-Mental State Examination

RCT

randomized controlled trial