Rivastigmine for Alzheimer's disease

Summary of findings for the main comparison. Rivastigmine compared to placebo for Alzheimer's disease

Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm² (9.8 mg/day) patch) compared to placebo for Alzheimer's disease
Patient or population: patients with Alzheimer's disease, mild to moderate Settings: multicentre, mostly in Europe or United States Intervention: rivastigmine (capsules 6 to 12 mg/day in 2 divided doses or 10 cm² patch) for 24 to 26 weeks Comparison: placebo for 24 to 26 weeks
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo	Rivastigmine (capsules 6 to 12 mg/day b.i.d. or 10 cm² patch)
Cognitive function (change from baseline at 26 weeks using ADAS‐Cog)		The mean score in the rivastigmine group was 1.79 lower (2.21 to 1.37 lower)		3232 (6 studies)	⊕⊕⊕⊝ moderate^1,2,	ADAS‐Cog score has a maximum of 70 points, the lower score of the rivastigmine group indicates greater improvement
Cognitive function (change from baseline at 26 weeks using MMSE)		The mean score in the rivastigmine group was 0.74 higher (0.52 to 0.97 higher)		3205 (6 studies)	⊕⊕⊕⊝ moderate^1,2	MMSE has a maximum score of 30 points, a lower score indicates greater impairment. treatment effect was in favour of rivastigmine
Activities of daily living (change from baseline at 26 weeks measured using various scales)		The mean score in the rivastigmine group was 0.2 standard deviations higher (0.13 to 0.27 higher)		3230 (6 studies)	⊕⊕⊕⊝ moderate¹	SMD 0.2 (0.13 to 0.27) A SMD of 0.2 is considered a small effect size. Treatment effect in favour of rivastigmine
Physician rated global impression tests (no change or worse compared with baseline, measured using Global Impression of Change at 26 weeks)	810 per 1000	744 per 1000 (712 to 773)	OR 0.68 (0.58 to 0.8)	3338 (7 studies)	⊕⊕⊕⊝ moderate¹	Treatment effect was in favour of rivastigmine
Behavioural symptoms (change from baseline at 26 weeks measured using various scales)		The mean score in the rivastigmine group was 0.04 standard deviations lower (0.14 lower to 0.06 higher)		1529 (3 studies)	⊕⊕⊕⊝ moderate^1,3	SMD ‐0.04 (‐0.14 to 0.06) A SMD of 0.2 is considered a small effect size. The size of this SMD and its small confidence interval suggests that there is no difference between the two groups
Acceptability of treatment (as measured by withdrawals from trials before end of treatment at 26 weeks)	149 per 1000	260 per 1000 (230 to 293)	OR 2.01 (1.71, 2.37)	3569 (7 studies)	⊕⊕⊕⊝ moderate¹	Withdrawals significantly more frequent in rivastigmine group compared with placebo group
Incidence of adverse events (at least one adverse event by 26 weeks)	761 per 1000	870 per 1000 (850 to 888)	OR 2.14 (1.80 to 2.53)	3587 (7 studies)	⊕⊕⊕⊝ moderate¹	Adverse events significantly more frequent in rivastigmine group compared with placebo group
Quality of life of patients or carers (measured using NPI‐D carer distress scale (change from baseline at 24 weeks)		The mean score in the rivastigmine group was 0.1 higher (0.91 lower to 1.11 higher)		529 (1 study)	⊕⊕⊕⊝ moderate¹	The size of this MD and its small confidence interval suggests that there is no difference between the two groups
The assumed risk* used the median control group risk across studies. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; OR: Odds ratio.
¹ Confidence in estimate of effect lowered due to relatively high dropout rates across studies, which are higher in the treatment group. The ITT analysis in these studies used LOCF (last observed carried forward) imputations. In addition, results are available up to only 26 weeks, longer term data would be more applicable. ² There was high heterogeneity the ADAS‐Cog outcome due to B352, which had high dropout rates and showed a difference of 3.8 points, compared to 1.2 to 1.6 points for the other studies. However, evidence not further downgraded; removal of this study from the analysis will only result in a small change of estimate by about 0.35 points. ³ Three studies (IDEAL; Lopez‐Pousa 2005; Nakamura 2011) reported a scale measuring behavioural disturbance. ⁴ The protocol for most studies had some measures related to quality of life or impact on carers, but only one study reported this (IDEAL).

Background

Description of the condition

Alzheimer's disease (AD), alone or in combination with other brain conditions, is the commonest cause of dementia affecting older people. It is associated with the loss of cholinergic neurons in parts of the brain subserving aspects of memory. As the disease progresses, people lose the ability to remember, communicate, think clearly and perform their usual daily activities. Their behaviour or personality may also change. In severe AD, people lose the ability to care for themselves and require full time care.

Currently there is no cure available for AD, but a few pharmacological interventions are available to alleviate symptoms.

Description of the intervention

Acetylcholinesterase inhibitors, such as rivastigmine, delay the breakdown of acetylcholine released into synaptic clefts and may enhance cholinergic neurotransmission.

Tacrine, the first of the acetylcholinesterase inhibitors to undergo extensive trials for this purpose, was associated with significant disadvantages, including low oral bioavailability and metabolism involving hepatic microsomal enzymes with a consequent risk of interactions with other drugs. Tacrine was also associated with adverse effects including hepatotoxicity. Several other acetylcholinesterase inhibitors, including rivastigmine, galantamine, and donepezil, have now been introduced. They are believed to have superior properties in terms of specificity of action and low incidence of adverse effects.

Rivastigmine is a 'pseudo‐irreversible' inhibitor of acetyl and butyrylcholinesterases with a phenylcarbamate structure, the metabolism of which is almost totally independent of the hepatic cytochrome P450 system. After binding to cholinesterase, the carbamate portion of rivastigmine is slowly hydrolysed, cleaved, conjugated to a sulphate and excreted. Rivastigmine has an oral bioavailability of 0.355 and low (40%) binding to plasma proteins. Its elimination half‐life is around two hours. Its disposition is essentially unaltered in patients with renal or hepatic impairment (Jann 2000) and the risk of interactions with other drugs is low (Grossberg 2000). This is of particular relevance for elderly patients with AD, some of whom may also need medications for other conditions. The drug is selective both to the central nervous system (CNS) and within it. In studies in human volunteers the inhibition of central acetylcholinesterase was substantially greater than the inhibition of peripheral acetylcholinesterase or butyrylcholinesterase (Kennedy 1999). Evidence from animal studies suggests that rivastigmine is a more potent inhibitor of acetylcholinesterase in the cortex and hippocampus, the brain regions most affected by AD (Polinsky 1998). Rivastigmine also preferentially inhibits the G1 enzymatic form of acetylcholinesterase, which predominates in the brains of patients with AD (Polinsky 1998). Rivastigmine is long‐acting and readily penetrates the CNS after parenteral or oral administration. The duration of cholinesterase inhibition by rivastigmine is approximately 10 hours.

Rivastigmine can be administered orally as capsules or liquid or from a transdermal patch, which has been developed more recently. Based on pharmacokinetic principles, the transdermal patch form was postulated to have advantages over the oral form. Adherence was expected to be improved by once daily dosing. Tolerance was also expected to be improved as the patch delivers a more steady concentration of rivastigmine to the body and has a lower equivalent dose to the oral form (9.5 mg as a transdermal patch is equivalent to 12 mg daily in the oral form).

Why it is important to do this review

Large multicentre trials have been completed in the USA, Canada, Europe, Australia and South Africa. Rivastigmine has received approval for use in 60 countries including all the member states of the European Union and in the USA, where it received approval from the Food and Drugs Administration (FDA) in April 2000. It is important to assess the safety and efficacy of this intervention in a systematic review.

Objectives

To determine the clinical efficacy and safety of rivastigmine for patients with dementia of Alzheimer's type
To compare the efficacy and safety of the oral and transdermal formulations of rivastigmine

Methods

Criteria for considering studies for this review

Types of studies

We included double‐blind, randomised controlled trials in which rivastigmine was administered for 12 weeks or longer and compared with placebo; or rivastigmine patches were compared with rivastigmine capsules. Trials in which the allocation to treatment was not randomised, or in which treatment allocation was not concealed, were excluded. This was because prior knowledge of treatment allocation may lead to biased allocation of patients (Schulz 1995).

Types of participants

The patients in trials to be included were diagnosed with probable AD according to internationally accepted criteria such as the Diagnostic and Statistical Manual of Mental Disorders DSM‐IV (DSM IV) and National Institute of Neurological and Communicative Disorders and Stroke‐Alzheimer's Disease and Related Disorders Association (NINCDS‐ADRDA) criteria (McKhann 1984).

Types of interventions

Objective 1

Intervention: rivastigmine given at any dose, using any method of administration

Comparison: placebo

Objective 2

Intervention: rivastigmine patches at the manufacturer's recommended dose

Comparison: rivastigmine capsules at the manufacturer's recommended dose

Types of outcome measures

In the original protocol and during the review, we looked for all the following outcomes:

cognitive function (as measured by psychometric tests);
functional performance;
global impression;
behavioural disturbance;
acceptability of treatment as measured by withdrawal from trials;
incidence of adverse effects;
effect on carers;
death;
institutionalisation rates;
quality of life;
dependency.

Search methods for identification of studies

Electronic searches

We searched ALOIS (www.medicine.ox.ac.uk/alois), the Cochrane Dementia and Cognitive Improvement Group Specialized Register, on 2 March 2015. The search terms used were: Rivastigmine OR exelon OR ENA OR "SDZ ENA 713".

ALOIS is maintained by the Trials Search Co‐ordinator of the Cochrane Dementia and Cognitive Improvement Group and contains studies in the areas of dementia prevention, dementia treatment and cognitive enhancement in healthy people. The studies are identified from:

monthly searches of a number of the major healthcare databases, MEDLINE, EMBASE, CINAHL, PsycINFO and LILACS;
monthly searches of a number of the trial registers, ISRCTN, UMIN (Japan's Trial Register), the World Health Organization (WHO) Clinical Trials Registry Platform portal (which covers ClinicalTrials.gov, ISRCTN, the Chinese Clinical Trials Register, the German Clinical Trials Register, the Iranian Registry of Clinical Trials, and the Netherlands National Trials Register, plus others);
quarterly searches of the Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library;
six‐monthly searches of a number of grey literature sources, ISI Web of Knowledge Conference Proceedings, Index to Theses, Australasian Digital Theses.

To view a list of all sources searched for ALOIS see About ALOIS on the ALOIS website.

Details of the search strategies used for the retrieval of reports of trials from the healthcare databases, CENTRAL and conference proceedings can be viewed in the ‘methods used in reviews’ section within the editorial information about the Dementia and Cognitive Improvement Group.

Additional searches were performed in many of the sources listed above to cover the timeframe from the last searches performed for ALOIS to ensure that the search for the review was as up‐to‐date and as comprehensive as possible. The search strategies used can be seen in Appendix 1.

The latest search for this review (March 2015) retrieved a total of 17 results for consideration.

Searching other resources

In addition, the search engines Copernic and Google were used to find evidence of unreported or unpublished trials using the word rivastigmine and its synonyms. Novartis websites, the Food and Drug Administration (FDA), European Medicines Agency (EMEA) and National Institute for Health and Care Excellence (NICE) websites were searched for data and evidence of trials.

1. Reference searching
The references of all identified studies were inspected for more studies.

2. Pharmaceutical companies
Novartis, the developer of rivastigmine, was contacted for information about any unpublished and published trials.

Data collection and analysis

Selection of studies

Irrelevant citations were discarded by a review of the title of the publication and its abstract. In the presence of any suggestion that the article could possibly be relevant, it was retrieved in full for further assessment. In the later versions of the review, one review author (JSB) selected the trials for inclusion in the review from the culled citation list.

There were multiple publications for most of the industry sponsored trials, often reporting different aspects (outcomes) of the studies or different lengths of follow up.

Data extraction and management

Data were extracted from the published reports in journals and unpublished company reports using data collection forms. One review author (JSB) extracted information from the reports of each study.

In addition to extracting pre‐specified information about study characteristics and aspects of methodology relevant to risk of bias, the following summary statistics, required for each trial and each outcome, were extracted.

For continuous data, mean change from baseline, the standard deviation, and the number of patients for each treatment group at each assessment. Where changes from baseline were not reported, the mean, standard deviation and number of patients for each treatment group at each time point were extracted, if available.
For binary data, the numbers in each treatment group and the numbers experiencing the outcome of interest were sought.
For ordinal variables which can be approximated to continuous variables, the main outcomes of interest were the assessment score at the time point being considered and the change from baseline (i.e. pre‐randomisation or at randomisation) at this time point. For some binary and ordinal outcomes the endpoint category relative to the baseline category was the outcome of interest. For other categorical outcomes, such as the Clinical Global Impression of Change (CIBIC‐Plus), the endpoint itself was of clinical relevance as all patients had begun, by definition, at the same baseline score.

The baseline assessment score was the latest available score, no longer than two months prior to the randomisation. Studies may have included a titration period prior to the randomisation phase of the study. Data from any open follow‐on phase, after the randomised phase, were not used to assess safety or efficacy.

For each outcome measure, data were sought on every patient assessed. To allow an intention‐to‐treat analysis (ITT), the data were sought irrespective of compliance and whether or not the patient was subsequently deemed ineligible or otherwise excluded from treatment or follow up. If ITT data were not available, an analysis of patients who completed treatment was conducted.

Assessment of risk of bias in included studies

The risk of bias assessment was conducted using the standard recommended approach for assessing the risk of bias in studies included in Cochrane reviews. The Cochrane Collaboration risk of bias tool is available in RevMan 5.2 and assesses the following domains:

sequence generation;
allocation concealment;
blinding of participants and personnel;
blinding of outcomes assessment;
incomplete outcome data;
selective outcome reporting;
'other bias'.

We made a judgement about the risk of bias in each domain, assigning it to one of three categories: 'high', 'low' or 'unclear' risk of bias. These assessments were based on the criteria for making judgements that are listed in section 8.5 of the Cochrane Handbook for Systematic Reviews of Interventions. The criteria focus on whether the risk is of importance (that is whether the presence of the risk could have an important impact on the results or the outcomes of the trial) rather than whether a risk of bias is present or not (Higgins 2011). The levels of risk may be different for different outcomes and this was considered during the assessment.

If insufficient detail was reported to make a judgement, this was usually considered as an ‘unclear' risk of bias. An ‘unclear’ judgement was also used in situations where it was clear what happened in the study but its likely impact on the study results was not known.

Measures of treatment effect

For dichotomous outcomes (where the outcome of interest was either present or absent), the estimate of treatment effect of the intervention was expressed as the Peto odds ratio (OR) together with the 95% confidence interval (CI).

For continuous data the measure of treatment effect was the mean difference (MD) or the standardised mean difference (SMD).

Unit of analysis issues

The review only included parallel‐group, double‐blinded randomised controlled trials (RCTs), with individual patients randomised. No unit of analysis issues were expected or encountered.

Dealing with missing data

Where data were missing from the published report of a trial, the authors or the study sponsors were contacted to obtain the data and to clarify any uncertainty.

We made no attempts at data imputation, except for the estimation of standard deviations for continuous data using the methods detailed in section 7.7.3 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Where possible we reported ITT analyses. We conducted sensitivity analyses to compare methods of dealing with missing data.

Assessment of heterogeneity

Potential differences between the included studies in the types of participants, interventions or control used were assessed before pooling data. No subgroup analyses were planned.

We assessed heterogeneity between studies using the Chi² test (with a significance level set at P < 0.10) and the I² statistic, which calculates the percentage of variability due to heterogeneity rather than to chance, with I² values over 50% suggesting substantial heterogeneity (Higgins 2011).

Assessment of reporting biases

Outcomes reported in a trial were compared with the protocol, whenever possible, to examine whether all of the study's pre‐specified outcomes that were of interest to the review had been reported.

Data synthesis

For ordinal variables, such as psychometric test scores, functional and quality of life scales, where there are a large number of possible scores, the measure was treated as continuous and the mean difference or the SMD was calculated.

For ordinal variables with only a small number of possible values, such as the Clinical Global Impression of Change, the data were reduced to a binary variable. The two classes were improvement compared with no change or worse. For all binary variables the Peto method of the typical OR was used.

The duration of the trials varied between 12 weeks and 1 year. Separate meta‐analyses were conducted for endpoints of 12 weeks, 24 to 26 weeks and 52 weeks. Some trials contributed data to more than one meta‐analysis if multiple assessments had been done.

A weighted estimate of the typical treatment effect across trials was calculated. Overall estimates of the treatment difference are presented. In all cases the overall estimate from a fixed‐effect model was presented.

Subgroup analysis and investigation of heterogeneity

Heterogeneity were examined both visually and using the I² statistic. Where there was evidence of heterogeneity of the treatment effect between trials then sensitivity analyses were conducted, where only homogeneous results were pooled.

There were no pre‐identified subgroups for subgroup analysis.

Sensitivity analysis

This review sought to analyse data using ITT data whenever possible. Some studies reported both an ITT analysis that included all patients randomised and a per protocol analysis. The ITT analysis results reported in studies often involved data imputation techniques such as the last observation carried forward (LOCF) for patients who did not complete the study. The impact of different ways of dealing with missing data were investigated using a sensitivity analysis of as observed, ITT and per protocol analyses. These results were tabulated and any important discrepancies discussed.

Summary of findings table

We summarised the data on the efficacy and safety of the currently recommended dose of rivastigmine (6 to 12 mg/day orally or 9.5 mg/day transdermally) in the summary of findings table using GRADE methods (Guyatt 2008) to assess the overall quality of the evidence.

Results

Description of studies

Results of the search

The updated searches performed in 2011, 2013, 2014 and 2015 retrieved a total of 112 references. The full texts of 42 references were read and, of these, 10 were of studies that could be included or additional reports of studies already included, and 32 were of studies that were excluded.

Included studies

The characteristics of the 13 included trials are summarised in Characteristics of included studies.

Design, participants, samples sizes and setting

Important details of study design (number of participants, duration of follow up, mean Mini‐Mental State Examination (MMSE) of participants at baseline and description of interventions) are summarised in Table 1 and the objectives of the trials in Table 2.

Table 1. Description of the included studies at baseline

Study	Duration (weeks)	Participants	Mean age (SD)	% males	Mean MMSE (SD)	country	Number of centres	Treatment groups
Oral (different doses versus placebo )
B103 (Phase II)	13	402	69.4	44	‐	Europe	54	4 mg/day b.i.d 6 mg/day b.i.d placebo
B104 (Phase II)	18	114	71.2 (7.5)	39	19.5 (3.7)	Belgium, France, UK, Norway, Canada	11	6 to 12 mg/day b.i.d. 6 to 12 mg/day t.i.d placebo
B303/B305* (Phase III)	26	725	72.0 (8.1)	41	20.0 (4.5)	France, Germany, Austria, Switzerland, Canada, USA	44	1 to 4 mg b.i.d 6 to 12 mg/day b.i.d., placebo
B304* (Phase III)	26	677	71.4 (8.2)	41	18.5 (4.5)	UK, Ireland, Australia, Canada, RSA, Italy	37	2 to 12 mg/day b.i.d. 2 to 12 mg/day t.i.d. placebo
B351* (Phase III)	26	702	74.1 (8.3)	44	20.0 (4.4)	USA	14	3 mg/day t.i.d 6 mg/day t.i.d 9 mg/day b.i.d placebo
B352* (Phase III)	26	699	74.5 (7.4)	39	19.7 (4.5)	USA	22	1 to 4 mg per day b.i.d 6 to 12 mg/day b.i.d. placebo
Ballard 2005	26	93	83.8 (7.7)	20	‐	UK	‐	6 to 12 mg/day b.i.d placebo
Karaman 2005*	52	44	73.8	45	12.2	Turkey	1	6 to 12 mg/day b.i.d. placebo
Lopez‐Pousa 2005*	26	218	77.6	23	8.8	Spain	21	6 to 12 mg/day b.i.d. placebo
Mowla 2007	12	122	69.2	46	16.1 (4.0)	Iran	‐	6 to 12 mg/day b.i.d. placebo
Tai 2000	26	80	‐	‐	‐	Taiwan	‐	3 to 6 mg/day b.i.d. placebo
Oral and patches
IDEAL* (Phase III)	24	1195	73.3 (7.8)	33	16.5 (3.0)	North, Central and South America, Asia, Europe	100	patch 9.5 mg/day patch 17.4 mg/day capsules 6 to 12 mg/day b.i.d. placebo
Patches
Nakamura 2011	24	859	74.6 (7.2)	31.7	16.6 (3.0)	Japan	multicentre	patch 4.6 mg/day patch 9.5 mg/day placebo

* These studies met the inclusion criteria of the main analysis comparing rivastigmine at the therapeutic doses versus placebo.

b.i.d = bis in die in Latin, this means that a medication is taken two times a day, dividing the total daily dose into two doses.

t.i.d = ter in die in Latin, this means that a medication is taken three times a day, dividing the total daily dose into three doses.

MMSE = Mini‐Mental Health State Examination. The score range from 0 ( severe impairment) to 30 (normal).

Table 2. Objectives of included studies

Study	Objective
B103	To assess the short term (3 months) symptomatic efficacy and tolerability of rivastigmine 4 and 6 mg/day compared with placebo in patients with AD
B104	Primary: to determine the maximum tolerated dose (MTD) of rivastigmine in patients with mild to moderate dementia of the Alzheimer type (DAT)
B104	Secondary: to determine ‐ a) whether tolerability is different when the drug is administered twice daily (b.i.d.) or three times daily (t.i.d.) ‐ b) if nausea and vomiting, associated with cholinesterase inhibition, can be controlled with antiemetics thereby increasing the MTD, and ‐ c) to assess the efficacy of rivastigmine at its MTD in comparison with that of placebo in the treatment of DAT
B303/B305	Primary 1: to evaluate the efficacy of two non‐overlapping dose ranges of rivastigmine (1 to 4mg daily and 6 to 12 mg daily) versus placebo over a 26 week treatment period as assessed by two primary measures of outcome; change from baseline in ADAS‐Cog score and the CIBIC‐Plus score at week 26
	Primary 2: to evaluate the safety of the study medication as assessed by incidence of adverse events, clinical laboratory evaluations , vital signs, ECG recordings, and the results of physical examination made at baseline and throughout the study
	Secondary: to assess dose‐efficacy and dose‐safety relationships for rivastigmine
B304	Primary: to evaluate the efficacy and safety of individual highest well‐tolerated doses (range 6 to 12 mg daily) of rivastigmine given b.i.d. or t.i.d. for 26 weeks compared with placebo in the therapy of patients with probable Alzheimer's disease
B304	Secondary: to compare the twice daily and three times daily dosing regimens with respect to efficacy and safety to evaluate changes in activities of daily living (ADL)
B351	Primary: to evaluate the efficacy and safety of three fixed doses of rivastigmine (3, 6 and 9 mg/day) and placebo for 26 weeks of treatment
	Secondary: to assess the dose‐efficacy and dose‐safety relationships for rivastigmine
	Tertiary: to explore the pharmacokinetics of rivastigmine at doses of 3, 6 and 9 mg daily
B352	Primary: to evaluate the efficacy and safety of two non‐overlapping dose ranges of rivastigmine (1 to 4 mg daily and 6 to 12 mg daily) and placebo for 26 weeks of treatment
	Secondary: to assess the dose‐efficacy and dose‐safety relationships of rivastigmine. To investigate the relationship between plasma concentrations of rivastigmine and efficacy and safety
	Tertiary: to explore the pharmacokinetics of rivastigmine at doses of 1 to 4 and 6 to 12 mg daily
IDEAL	To compare the efficacy,safety and tolerability of a novel rivastigmine transdermal patch with conventional rivastigmine capsules and placebo in patients with AD
Karaman 2005	To evaluate the efficacy of rivastigmine for a period of 12 months in patients with advanced moderate AD
Lopez‐Pousa 2005	To evaluate the safety and efficacy of rivastigmine in patients with more advanced AD
Mowla 2007	To assess the effect of serotonin augmentation on cognition and ADL of patients with AD
Ballard 2005	To determine whether rivastigmine was better than placebo for agitation and cognition
Tai 2000	To evaluate the safety and efficacy of Exelon compared with placebo in patients with probable Alzheimer's disease who had dementia ranging from mild to moderate degree
Nakamura 2011	To evaluate the efficacy, safety, and tolerability of the 5 cm² (9 mg loading dose, 4.6 mg/24 h delivery rate) and 10 cm² (18 mg loading dose, 9.5 mg/day delivery rate) rivastigmine patch in Japanese patients with AD

Only randomised, double‐blinded placebo controlled trials or studies comparing different formulations were included in this review. Thirteen studies met the inclusion criteria of the review.

Six trials, phase II and III, were all supported by Novartis Pharmaceuticals Corporation and were completed by 1996. They are identified by their Novartis or ADENA code (ADENA was the name given by Novartis to the Exelon Phase III clinical trials programme). The two phase II trials were designed to assess the tolerability, efficacy and safety of rivastigmine over three to four months. The four phase III trials were designed to assess the efficacy and safety of rivastigmine in patients with mild to moderately severe AD over six months. The trials had many features in common. They were all multicentre, randomised, double‐blind, parallel‐group trials. All trials compared rivastigmine with placebo, with at least two treatment groups of different rivastigmine regimens.

Of the seven later trials, three were also sponsored by Novartis (IDEAL; Lopez‐Pousa 2005; Nakamura 2011). The key information about these seven trials is summarised as follows.

There is limited information available about Tai 2000, which has been published only as an abstract. This trial appeared to be an independent trial carried out in Taiwan. Eighty participants with mild to moderate AD were treated with rivastigmine or placebo for 26 weeks. No data were available to include in the meta‐analyses.
Ballard 2005 was a small 26 week trial (n = 93) with three treatment arms, rivastigmine, quetiapine and placebo, of equal size. The objective was to compare the efficacy of rivastigmine and quetiapine for agitation in people with possible or probable AD who were living in institutions. We did not include any data from this trial in the meta‐analyses because of concerns about a high risk of attrition bias and exclusion of the most severely impaired patients from the analyses.

Karaman 2005 and Lopez‐Pousa 2005 aimed to investigate the efficacy of rivastigmine for patients with more advanced disease than those previously tested.

Karaman 2005 was a small 12 month trial (n = 44, mean baseline MMSE = 12.2). We did not include data from this trial in our meta‐analyses due to concern about a high risk of bias.
Lopez‐Pousa 2005 was a 6 month trial (n = 218, mean baseline MMSE = 8.8). In addition to the outcomes of cognitive function, activities of daily living and global clinical change, Lopez‐Pousa 2005 was the earliest included trial to assess behavioural symptoms.
Mowla 2007 was a 12 week trial in mild to moderate AD with three treatment groups, rivastigmine, rivastigmine plus fluoxetine and placebo. The rivastigmine plus fluoxetine group was not included in this review. There were 82 participants in total in the rivastigmine and placebo groups. We were not able to include any data from this trial in the meta‐analyses due to incomplete reporting of results.

IDEAL and Nakamura 2011 were the only trials to include transdermal rivastigmine.

IDEAL was a 6 month study (n = 1195) in mild to moderate AD, with 4 treatments arms, rivastigmine capsules, 2 doses of transdermal rivastigmine and placebo.
Nakamura 2011 was a 24 week dose finding trial in mild to moderate AD (n = 859) with 3 treatment arms, 2 doses of transdermal rivastigmine and placebo..

All studies used current diagnostic criteria for dementia (DSM‐IV) and probable AD (NINCDS‐ADRDA) (McKhann 1984) except Tai 2000, which did not give its diagnostic criteria. The severity of disease was mostly assessed by the MMSE rating scale, and patients that were included had MMSE scores of 10 to 26 inclusive apart from 2 studies (Karaman 2005; Lopez‐Pousa 2005), which randomised patients with MMSE scores of 3 to 12. The list of exclusions was not extensive. Patients with severe and unstable illnesses (cardiovascular or pulmonary disease, unstable diabetes mellitus, peptic ulceration within the preceding five years, evidence of alcohol or substance abuse) were excluded, as were individuals taking medications such as anticholinergic drugs, acetylcholine precursor health food supplements, memory enhancers, insulin and psychotropic drugs. The procedures followed were in accordance with the ethical standards of the relevant institutional committees on human experimentation and with the Declaration of Helsinki (Helsinki declaration).

Interventions

Information about treatment groups and actual doses achieved are tabulated in Table 1 and Table 3 respectively.

Table 3. Mean daily dose (mg/day) of rivastigmine achieved in the studies at different time points

Time (weeks)	treatment group	B103	B104	B303/B305	B304	B351	B352	IDEAL	Karaman 2005	Lopez‐Pousa 2005	Nakamura 2011
10 to 12	low b.i.d.	4	‐	3.8	‐	2.9	3.6
	medium b.i.d.	6	‐	‐	‐	5.7	‐
	high b.i.d.	‐	9.6	10.4	9.5	8.8	10.1		6.1
	high t.i.d.	‐	10.2	‐	9.7	‐	‐
26	low b.i.d.	‐	‐	3.7	‐	2.8	3.5
	medium b.i.d.	‐	‐	‐	‐	5.7	‐
	high b.i.d.	‐	‐	10.4	9.3	8.5	9.7	9.7	8.3	9.8
	high t.i.d.	‐	‐	‐	9.6	‐	‐
	low patch										4.6
	medium patch							9.5			9.5
	high patch							16.5
48	medium patch
	high patch
52	high b.i.d.								10.7

Exact doses not available for B103, Ballard 2005, Tai 2000, Mowla 2007.

Twelve studies investigated the oral form of rivastigmine, and one of these studies also included an arm randomised to a rivastigmine patch (IDEAL).

Earlier industry sponsored trials investigated a range of doses, from 2 mg/day to 12 mg/day in two or three divided doses. In later trials (Ballard 2005; Karaman 2005; Lopez‐Pousa 2005; Mowla 2007; IDEAL) only the dose range of 6 to 12 mg/day was used to compare against placebo. Tai 2000 investigated doses of 3 to 6 mg/day in two divided doses. All studies with high oral doses achieved a mean daily dose of between 9.3 to 10.7 mg/day, except for Karaman 2005 (8.3 mg/day) and B351 (8.5 mg/day). The mean daily doses achieved for medium doses were between 5.7 and 6 mg/day. Further information on the doses achieved was not available for four trials (B103; Ballard 2005; Mowla 2007; Tai 2000).

Two studies evaluated the safety and efficacy of patches. IDEAL investigated 6 to 12 mg/day capsules in 2 doses and the other 2 arms tested rivastigmine patches, a 10 cm² patch which delivered 9.5 mg/day and a 20 cm² patch which delivered 17.4 mg/day. Patients were titrated to their target dose in four week steps. Patients in the patch groups started with a 5 cm² patch until the target dose was achieved; in the capsule group they began with 3 mg/day, increased by steps of 3 mg/day. All patients had a rivastigmine or placebo patch once a day and a rivastigmine or placebo capsule twice a day. Nakamura 2011 investigateda 10 cm² patch which delivered 9.5 mg/day, a 5 cm² patch which delivered 4.6 mg/day and a placebo arm. Patients were titrated to their target patch dose over four week intervals, followed by an eight week maintenance period.

Outcomes

The trials examined cognitive, functional and global effects, behavioural symptoms, as well as the safety and tolerability of rivastigmine.

Apart from the outcome measures related to safety or adverse effects, all the outcomes for the effectiveness of rivastigmine were measured by questionnaires or psychometric tests. Different types of instruments were utilised to measure each outcome. The details of the outcomes measured and reported in each trial are summarised in Table 4.

Table 4. Measured outcomes

Outcomes assessed	Cognitive function			Activities of daily lIving		Behavioural symptoms	Physician rated global impression of change		Other domains
Study	ADAS‐Cog	MMSE	Others	PDS	Others	Behavioural symptoms	CIBIC‐Plus	Others	Other domains
B103		X	OE, TMT, NOSGER, DSST, VRT					CGIC
B104	X		Wechsler psychometric tests, NOSGER				X
B303/B305	X	X	ADAS‐CogA	X	CAS		X	GDS
B304	X	X	ADAS‐CogA	X	CAS		X	GDS
B351	X	X	ADAS‐CogA	X	CAS		X	GDS
B352	X	X	ADAS‐CogA	X	CAS		X	GDS
Ballard 2005			SIB			CMAI
Karaman 2005	X	X		X	ACDS‐ADL, DAD		X	GDS
IDEAL	X	X	CLOCK DRAWING, TMT		ACDS‐ADL	NPI‐12		ADCS‐CGIC
Lopez‐Pousa 2005		X	SIB, BLESSED DEMENTIA SCALE		ACDS‐ADL	NPI‐10, NPI‐4		GDS ADCS‐CGIC
Mowla 2007			WMS‐III,		ADL			CGI	Hamilton score
Tai 2000		X	NPT				X	GDS
Nakamura 2011	X	X	MENFIS		DAD	BEHAVE‐AD	X

x indicated that the study measured this outcome.

The full names of these scales and their properties are described in Types of outcome measures.

1. Cognitive function

Alzheimer's Disease Assessment Scale (ADAS‐Cog) (Rosen 1984). ADAS‐Cog comprises 11 individual tests: spoken language ability (0 to 5), comprehension of spoken language (0 to 5), recall of test instructions (0 to 5), word finding difficulty (0 to 5), following commands (0 to 5), naming object (0 to 5), construction drawing (0 to 5), ideational praxis (0 to 5), orientation (0 to 8), word recall (0 to 10) and word recognition (0 to 12). The total score ranges from 0 to 70, the higher the score indicating greater impairment.
The ADAS‐CogA total score is the ADAS‐Cog plus the attention item from the ADAS‐Noncog.
The Mini‐Mental State Examination (MMSE) (Folstein 1975) evaluates cognition in five areas: orientation, immediate recall, attention and calculation, delayed recall and language. The test takes only 15 minutes to administer and the scores range from 0 (severe impairment) to 30 (normal).
The Severe Impairment Battery (SIB) (Panisset 1994; Saxton 1990) is a 40‐item questionnaire designed to assess the severity of cognitive dysfunction in advanced AD and is divided into 9 domains: memory, language, orientation, attention, praxis, vasospastically, construction, orientation to name and social interaction. The score ranges from 0 (greatest impairment) to 100 (no impairment).
The Revised Wechsler Memory Scale (WMS‐R) (Wechsler 1987) comprises a series of brief subtests, some taken from the WMS and each measuring a different facet of memory, which are summarised into five composite scores and finally two major scores using weights prescribed by Wechsler. Some of the tests were used in B103.
The Fuld Object‐Memory Test (OME) (Fuld 1981) evaluates short term memory and learning by measuring the recall of 10 previously viewed objects.
The Benton Visual Retention Test (VRT) (Benton 1974) evaluates visual memory by assessing the accuracy of reproduction of each of 10 designs shown briefly to the individual.
The Trail Making Test (TMT) (Reitan 1958) assesses the time taken to connect a series of 25 numbered dots.
The Alzheimer's Disease Cooperative Study‐Clinical Global Impression of Change (ADCS‐CGIC) (Schneider 1997) provides a single global rating of change from baseline, rated by an independent observer who has no access to the other efficacy or safety data.
The Ten‐Point Clock Drawing Test (Watson 1993) assesses visuospatial and executive functions.
The Mental Function Impairment (MENFIS) (Homma 1991) evaluates core symptoms of dementia including cognitive, motivational and emotional aspects based on an interview with the patient and carer. The score ranges from 0 to 78 (greater functional deficit).
Digital substitution test (DSST).

2. Activities of daily living

The Progressive Deterioration Scale (PDS) (DeJong 1989) is an instrument with 29 items assessing the activities of daily living as rated by a carer. Each item is scored on a visual analogue scale of 0 to 100, and the total score is the mean item score. The score of 0 to 100 decreases with severity of dementia.
The Alzheimer's Disease Cooperative Study activities of daily living inventory for severe Alzheimer's disease (ADCS‐ADL) (Galasko 1997). This is a 19‐item scale for basic and complex abilities validated in patients with moderate to severe dementia. The total score ranges from 0 to 54 (no impairment). Items include basic activities of daily living (eating, bathing) and complex activities (operating taps, switching lights).
The Caregiver Activity Survey (CAS) is completed by the caregiver and includes six items for which the caregiver estimates the amount of time spent in the previous 24 hours helping the patient with activities of daily living.
The Nurses' Observation Scale for Geriatric Patients (NOSGER) (Brunner 1990) is designed to assess various cognitive functions and behaviour as related to activities of daily living and as assessed by a caregiver who sees the patient frequently. The NOSGER contains 6 x 5 = 30 items which were selected to assess the following dimensions: (a) memory, (b) self‐care, (c) instrumental activities of daily life, (d) mood, (e) disturbing behaviour, (f) social behaviour.The Disability Assessment for Dementia (DAD) is a 46‐item structured interview for the carer, scored 0 to 100 (least impairment), to evaluate activities of daily living (Gelinas 1999).

3. Behavioural symptoms

The Neuropsychiatric Instrument (NPI) (Cummings 1994) is a 12‐item, carer rated instrument to evaluate behavioural and neuropsychiatric symptoms, including delusions, hallucinations, agitation and aggression, depression or dysphoria, anxiety, elation or euphoria, apathy, disinhibition, irritability, aberrant motor behaviour, night‐time behaviour and appetite or eating disorder. The frequency is rated from 1 (occasional, less than once a week) to 4 (very frequent) and severity from 1 (mild) to 3 (severe). The product of frequency and severity ranges from 1 to 12, with a total score ranging from 12 to 120 for the 10 domains summed. A lower score indicates improvement.
The Cohen‐Mansfield Agitation Inventory (CMAI) (Cohen‐Mansfield 1995) scale, range from 29 to 203, is widely used in nursing homes to assess agitation. The scale examines 29 types of agitated behaviour, including pacing, verbal or physical aggression, performing repetitious mannerisms, screaming, and general restlessness. The frequency of these behaviours is measured on a 7‐point scale, ranging from 1 (never occurs) to 7 (occurs several times an hour, and includes cluster scores for physical and verbal aggression, and total aggression.
The Behavioural Pathology in AD (BEHAVE‐AD) assesses potentially remediable behavioural problems (agitation, aggression, affect, psychosis) in patients with AD. It consists of 22 symptoms grouped into 7 categories, each scored by a carer on a 4‐point scale (Reisberg 1989).

4. Physician rated global impression tests

A Clinician's Interview‐Based Impression of Change scale (CIBIC‐Plus) (Reisberg 1994) includes information supplied by the caregiver and patient. It provides a global rating of patient function in four areas: general, cognitive, behaviour and activities of daily living. All patients are scored as 4 at baseline; subsequent assessment on a scale of 1 to 7 is relative to baseline, with 1 showing marked improvement and 7 marked worsening.
The Global Deterioration Scale (GDS) (Reisberg 1982) is reported as a score from 1 to 7, 1 indicating normality to 7 indicating very severe dementia, and is a global assessment carried out by a clinician who has access to all information about a patient.
The Clinical Global Impression of Change (CGIC) (Guy 1976) is a global rating of all domains of a patient's current condition in comparison with baseline. It is a 7‐point scale ranging from 1 (very much improved) to 7 (very much worse), with 4 indicating no change. The assessment is conducted by the same clinician at both time points with input from relatives or carers.

5. Acceptability of treatment, as measured by withdrawal from trial

In anticipation of the typical gastrointestinal adverse events associated with cholinesterase inhibitors, which can be dose‐dependent, the various arms of the older trials compared both different doses and twice or thrice daily dosage schedules. Three fixed doses were tested in B351, but the other trials aimed for a maximum tolerated dose within a prescribed range. The period of titration was longer for larger doses and varied between 3 and 12 weeks. The later trials tested a transdermal patch formulation which provided continuous delivery of the drug with the objective of improving tolerability. The mean daily doses of rivastigmine at different time points are presented in Table 3. Safety and tolerability were evaluated by recording adverse events and serious adverse events. In addition, routine physical examinations with blood and urine analyses were performed and vital signs and electrocardiograms were checked at all clinic visits. Seven trials reported the withdrawal rate at 26 weeks (B303/B305; B304; B351; B352; IDEAL; Lopez‐Pousa 2005; Nakamura 2011).

6. Incidence of adverse events

The studies reported the types of adverse events reported by patients, and the number of patients experiencing these events, usually focusing on the most commonly experienced adverse events. A wide range of adverse events which were consistent with the anticholinergic properties of rivastigmine were reported, including gastrointestinal adverse events such as nausea, vomiting, abdominal pain or discomfort, and diarrhoea. Other adverse events reported included falls, insomnia, agitation, weight loss, headache, dizziness, and cutaneous adverse events where patches were used.

The same seven studies which reported on withdrawal from the trial before completion of the study also reported the number of patients who experienced at least one adverse event. Most of these studies had defined a safety population which is the basis for the adverse events analyses.

7. Quality of life of patients and carers

Only one study reported changes in the NPI‐D carer distress scale. This study reported the change from baseline at 24 weeks (IDEAL).

Excluded studies

Please see Characteristics of excluded studies.

Risk of bias in included studies

(Figure 1 and Figure 2)

Figure 1

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Figure 2

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

All the trials sponsored by Novartis were considered to be at low risk of bias for randomisation and allocation concealment, other than Lopez‐Pousa 2005 where it was difficult to be sure whether allocation was concealed effectively.

Of the independent trials, Ballard 2005 had a low risk of allocation bias, with clearly described procedures. However, the risk of bias in this domain was unclear for Tai 2000 (an abstract), Karaman 2005 and Mowla 2007 because there were no descriptions of methods. Karaman 2005 was of particular concern as only "participants who tolerated the drug well and perceived benefit were invited to continue rivastigmine treatment" after eight weeks.

Blinding

All trials were double‐blinded and placebo controlled, with precautions taken to maintain the blinding such as ensuring the placebo was identical in appearance to the active treatment. However, in B104 the placebo group received the treatment twice daily whereas one of the treatment arms received the intervention three times daily. There were no descriptions of additional steps taken to mask this. The difference in the number of times the capsules were taken could have unmasked the three times per day group. The effectiveness of double‐blinding in Mowla 2007 was also unclear because all patients in this study had received the placebo during the six week pre‐randomisation run‐in period.

Of the two studies testing patches, IDEAL was considered to be at low risk of bias for blinding as a double dummy was used. Nakamura 2011 stated that "patients, investigator staff, persons performing the assessments and data analysts are all blinded", but it was unclear how this was achieved since the study had used different patch sizes (2.5, 5, 7.5 and 10 cm²) to achieve the target dose.

Incomplete outcome data

Attrition bias was a major concern. There were substantial losses from Ballard 2005 where 19% (6/31) of those randomised to rivastigmine did not start treatment compared with 6% of those randomised to placebo. Only 18/31 in the rivastigmine group completed the trial compared with 27/31 in the placebo group. Those with a low baseline score on the Severe Impairment Battery (SIB) were not included in the analyses. These concerns led us to exclude data from Ballard 2005 from the meta‐analyses.

Karaman 2005, although the longest duration included trial (52 weeks), lost very few patients: only 3 of 24 in the rivastigmine group and none from the placebo group. This was a much lower rate of loss than for any other trial.

For the other 11 studies missing assessments caused major problems in the analysis and interpretation of the results. Approximately 17% of patients from the 1 to 4 mg daily and placebo groups and 35% of patients from the 6 to 12 mg daily groups left the trial before completing treatment. If patients dropped out at random from each group, that is the dropout was not associated with the treatment, the comparisons between groups are not biased but estimates of differences are reduced in precision. However, the dropout rates were not random and were related to treatment. Various methods were used in the trials for dealing with missing data.

The older trials (B303/B305; B304; B351; B352) reported in detail the methods using for dealing with missing data. Approximately a third of the patients who dropped out contributed endpoint data (retrieved drop out (RDO)). The ITT analyses included the completers (observed cases (OC)) data and the RDO data, and for the remainder of the patients the last available assessment (last observation carried forward (LOCF)). This remainder comprised approximately 6% of the patients in the placebo and 1 to 4 mg daily groups, and 24% for the 6 to 12 mg daily group at 26 weeks. An overestimate of the outcome effect would be expected.

In order to compare the different methods of dealing with missing assessments, for two outcomes (ADAS‐Cog and CIBIC‐Plus) we conducted meta‐analyses on three different groups of patients: OC only, RDO + OC, and ITT (OC + RDO + LOCF). The results are presented in Table 5. These analyses showed that compared with OC or RDO + OC, the ITT analyses did not produce results favouring rivastigmine, indeed the opposite was true but the differences between results were small. Therefore, the ITT analyses were considered satisfactory and were reported for all other outcomes. Further analysis of the data from the ITT, the OC and RDO + OC analyses to investigate the size and direction of the bias due to differential dropouts from the arms of the rivastigmine trials (Birks 2008) led to the conclusion that the absolute size of the bias was small and the direction could not be ascertained.

See: Summary of findings for the main comparison Rivastigmine compared to placebo for Alzheimer's disease

Table 5. Comparison of different methods of dealing with missing values

Time point	population	rivastigmine n	placebo n	result	probability level	95% confidence limits
1 to 4 mg daily versus placebo, ADAS‐Cog measured as change from baseline
12 weeks	ITT	650	643	favours rivastigmine WMD ‐0.31	0.30	‐0.87, 0.25
	OC	589	598	favours rivastigmine WMD ‐0.46	0.14	‐1.08, 0.15
	RDO + OC	616	615	favours rivastigmine WMD ‐0.37	0.20	‐0.96, 0.23
18 weeks	ITT	650	643	favours rivastigmine WMD ‐1.07	0.0004	‐1.66, ‐0.48
	OC	558	552	favours rivastigmine WMD ‐1.19	0.0005	‐1.86, ‐0.52
	RDO + OC	573	572	favours rivastigmine WMD ‐1.33	0.00008	‐1.99, ‐0.67
26 weeks	ITT	650	644	favours rivastigmine WMD ‐0.84	0.01	‐1.48, ‐0.19
	OC	519	526	favours rivastigmine WMD ‐0.96	0.01	‐1.72, ‐0.21
	RDO + OC	559	564	favours rivastigmine WMD ‐1.07	0.004	‐1.80, ‐0.34
6 to 12 mg daily versus placebo, ADAS‐Cog measured as change from baseline
12 weeks	ITT	1054	863	favours rivastigmine WMD ‐1.49	<0.00001	‐1.96, ‐1.01
	OC	843	803	favours rivastigmine WMD ‐1.80	<0.00001	‐2.33, ‐1.27
	RDO + OC	967	828	favours rivastigmine WMD ‐1.38	<0.00001	‐1.89, ‐0.88
6 to 12 mg daily versus placebo, ADAS‐Cog measured as change from baseline
18 weeks	ITT	1054	863	favours rivastigmine WMD ‐1.79	<0.00001	‐2.30,‐ 1.29
	OC	732	742	favours rivastigmine WMD ‐2.36	<0.00001	‐2.96, ‐1.76
	RDO + OC	837	772	favours rivastigmine WMD ‐2.12	<0.00001	‐2.69, ‐1.55
26 weeks	ITT	1054	863	favours rivastigmine WMD ‐2.09	<0.00001	‐2.65, ‐1.54
	OC	670	709	favours rivastigmine WMD ‐2.62	<0.00001	‐3.29, ‐1.94
	RDO + OC	788	759	favours rivastigmine WMD ‐2.39	<0.00001	‐3.03, ‐1.74
1 to 4 mg daily versus placebo, CIBIC‐Plus measured as no change or worse
12 weeks	ITT	608	612	favours rivastigmine Peto OR 0.93	0.60	0.72, 1.21
	OC	583	596	favours rivastigmine Peto OR 0.95	0.70	0.72, 1.23
	RDO + OC	609	612	favours rivastigmine Peto OR 0.94	0.60	0.72, 1.22
18 weeks	ITT	614	620	favours rivastigmine Peto OR 0.98	0.90	0.75, 1.26
	OC	556	554	favours placebo Peto OR 1.04	0.80	0.80, 1.37
	RDO + OC	570	576	favours placebo Peto OR 1.02	0.90	0.78, 1.34
26 weeks	ITT	614	623	favours rivastigmine Peto OR 0.71	0.01	0.55, 0.93
	OC	513	523	favours rivastigmine Peto OR 0.67	0.006	0.50, 0.89
	RDO + OC	544	549	favours rivastigmine Peto OR 0.68	0.008	0.52, 0.91
1 to 4 mg daily versus placebo, CIBIC‐Plus measured as no change or worse
12 weeks	ITT	950	825	favours rivastigmine Peto OR 0.74	0.008	0.60, 0.92
	OC	831	799	favours rivastigmine Peto OR 0.72	0.005	0.58, 0.91
	RDO + OC	952	825	favours rivastigmine Peto OR 0.75	0.01	0.60, 0.93
18 weeks	ITT	970	835	favours rivastigmine Peto OR 0.81	0.06	0.65, 1.01
	OC	720	741	favours rivastigmine Peto OR 0.72	0.005	0.57, 0.91
	RDO + OC	820	772	favours rivastigmine Peto OR 0.77	0.02	0.62, 0.97
26 weeks	ITT	973	839	favours rivastigmine Peto OR 0.68	0.0007	0.55, 0.85
	OC	660	693	favours rivastigmine Peto OR 0.63	0.0004	0.49, 0.81
	RDO + OC	784	758	favours rivastigmine Peto OR 0.65	0.0003	0.51, 0.82

The results for two outcomes, ADAS‐Cog and CBIC at 12, 18 and 26 weeks, have been pooled for 3 studies, B303/B305, B351. B352. These studies reported results for 3 populations, intention‐to‐treat (ITT), completers (OC), and completers + retrieved dropout (RDO + OC). The table reports the results of the meta‐analyses for 2 comparisons (1 to 4 mg daily versus placebo and 6 to 12 mg/day versus placebo) for the 3 populations at the 3 time points.

Selective reporting

For most of the studies the risk of reporting bias across all outcomes was difficult to judge. A few of the studies had listed the Caregiver Activities Survey (CAS) as an outcome in their protocols but these were not reported in the study results. In addition B304 and B351, two large randomised trials, were not published. Our data were obtained from information provided by Novartis Ltd.

For three of the studies (B104; B304; Nakamura 2011) sufficient information was available from the study protocols and we considered these as low risk of bias. However, there was insufficient information to assess the risk of reporting bias in the other studies.

Other potential sources of bias

Out of these 13 studies included in the review, only four (Ballard 2005; Karaman 2005; Mowla 2007; Tai 2000) were conducted without direct sponsorship or funding from the manufacturer, Novartis Pharma, but none provided data that could be included in the review.

Karaman 2005 reported standard deviations for the outcome measures that were an order of magnitude smaller than those seen in any other trial. We have asked the authors for clarification of these unusual findings but have not received a reply.

Effects of interventions

There are 13 included trials but 4 (Ballard 2005; Karaman 2005; Mowla 2007; Tai 2000) did not contribute to the analyses. Data from Ballard 2005 was excluded because of the high attrition rate from the rivastigmine group and concern over the elimination from the analyses of patients with a low baseline score. Data from Mowla 2007 could not be included due to incomplete reporting. No data could be used from Tai 2000 as the trial report provided insufficient information. The data from Karaman 2005 were of concern because of the potential for biased results and were omitted from the analyses. Although the longest duration trial, 52 weeks, only 3 of 24 in the rivastigmine group and none from the placebo group were lost. This was a much lower rate of loss than for any other trial. The numbers randomised were not reported but it was stated that patients were excluded at eight weeks if they did not appear to benefit.

In order to meet the objectives of the review we conducted analyses comparing various doses and formulations of rivastigmine with placebo or comparing different formulations of rivastigmine.

The rating scales and cognitive tests used differ in the direction representing improvement. A decrease in score indicates clinical improvement with the ADAS‐Cog, the CIBIC‐Plus and the GDS, while an increase shows improvement for the PDS and MMSE.

Comparison of rivastigmine (6 to 12 mg/day twice daily capsules or 10 cm² (9.5 mg/day) patch) with placebo

Cognitive function

The meta‐analysis, using weighted mean differences (WMDs), revealed a benefit on cognitive function as measured by the ADAS‐Cog test scores for rivastigmine compared with placebo at 26 weeks (ITT analysis, WMD ‐1.79; 95% CI ‐2.21 to ‐1.37, P < 0.00001, 6 studies).

The MMSE showed similar results in favour of rivastigmine at 26 weeks compared with placebo (ITT analysis, WMD 0.74; 95% CI 0.52 to 0.97, P < 0.00001, 6 studies).

Activities of daily living

The meta‐analysis, using standardised mean differences (SMDs), showed an improvement associated with rivastigmine compared with placebo at 26 weeks (ITT analysis, WMD 0.20; 95% CI 0.13 to 0.27, P < 0.00001, 6 studies).

Global assessment

The seven‐point CIBIC‐Plus scale, or the ADCS‐CGIC scale, measuring global clinical state was dichotomized by counting those showing no change or decline against those showing improvement. There were benefits associated with rivastigmine compared with placebo at 26 weeks (ITT analysis, 1339/1848 rivastigmine, 1197/1490 placebo) (OR 0.68; 95%CI 0.58 to 0.80, P < 0.00001, 7 studies).

Behavioural symptoms

Three studies (IDEAL; Lopez‐Pousa 2005; Nakamura 2011) assessed behavioural symptoms using the Neuropsychiatric Instrument (NPI‐10 and NPI‐12). There was no difference between rivastigmine and placebo at 26 weeks.

Withdrawals before the end of treatment

The meta‐analysis of withdrawals before the end of treatment showed a significant difference in favour of placebo compared with rivastigmine 26 weeks (571/2038 rivastigmine, 240/1531 placebo) (OR 2.06; 95%CI 1.74 to 2.45, P < 0.00001, 7 studies).

Adverse events

The meta‐analysis of numbers of patients with at least one adverse event showed that at 26 weeks there was a significant difference between the rivastigmine and placebo groups in favour of placebo (1637/2025 rivastigmine, 1123/1562 placebo) (OR 2.16; 95%CI 1.82 to 2.57, P < 0.00001, 7 studies).

Quality of life of carers

One study reported the changes in NPI‐D carer distress scale from baseline and this was reported at 24 weeks (IDEAL). No significant difference was detected (MD 0.10; 95% CI ‐0.91 to 1.11, 1 study).

Comparison of rivastigmine (1 to 4 mg/day and 6 to 12 mg/day twice daily capsules) with placebo

Cognitive function

The meta‐analysis, using WMDs, revealed a benefit on cognitive function as measured by ADAS‐Cog test scores for the lower dose rivastigmine compared with placebo at 26 weeks, but not at 12 weeks; and for the higher dose at 12 and 26 weeks:

rivastigmine 1 to 4 mg/day at 12 weeks (ITT analysis, WMD ‐0.31; 95% CI ‐0.87 to 0.25, P = 0.01, 3 studies);
rivastigmine 6 to 12 mg/day at 12 weeks (ITT analysis, WMD ‐1.49; 95% CI ‐1.96 to ‐1.01, P < 0.00001, 4 studies);
rivastigmine 1 to 4 mg/day at 26 weeks (ITT analysis, WMD ‐0.84; 95% CI ‐1.48 to ‐0.19, P = 0.01, 3 studies);
rivastigmine 6 to 12 mg/day at 26 weeks (ITT analysis, WMD ‐1.99; 95% CI ‐2.49 to ‐1.50, P < 0.00001, 5 studies).

The MMSE showed similar results in favour of lower dose rivastigmine at 26 weeks and higher dose rivastigmine at 26 weeks, compared with placebo:

rivastigmine 1 to 4 mg/day at 26 weeks (ITT analysis, WMD 0.43; 95% CI 0.08 to 0.78, P = 0.02, 3 studies);
rivastigmine 6 to 12 mg/day at 26 weeks (ITT analysis, WMD 0.82; 95% CI 0.56 to 1.08, P < 0.00001, 5 studies).

One study (Lopez‐Pousa 2005) used the Severe Impairment Battery (SIB), which showed benefit associated with higher dose rivastigmine compared with placebo at 26 weeks (MD 4.53; 95% CI 0.47 to 8.59, P = 0.03).

Activities of daily living

The PDS (carer assessment of the activities of daily living) showed an improvement associated with higher dose, but not lower dose, rivastigmine compared with placebo at 12 and 26 weeks:

rivastigmine 1 to 4 mg/day at 12 weeks (WMD ‐0.77; 95% CI ‐1.84 to 0.30, 3 studies);
rivastigmine 1 to 4 mg/day at 26 weeks (WMD ‐0.38; 95% CI ‐1.61 to 0.84) (3 studies);
rivastigmine 6 to 12 mg/day at 12 weeks (WMD 1.08; 95% CI 0.19 to 1.98, P = 0.02, 4 studies);
rivastigmine 6 to 12 mg/day at 26 weeks (WMD 2.15; 95% CI 1.13 to 3.16, P < 0.0001, 4 studies).

One study (IDEAL) assessing activities of daily living (ADL) using the ADCS‐ADL scale showed benefit for 6 to 12 mg/day at 24 weeks (MD 1.80; 95% CI 0.20 to 3.40, P = 0.03).

Global assessment

The seven‐point CIBIC‐Plus scale, or the ADCS‐CGIC scale, measuring global clinical state was dichotomized by counting those showing no change or decline against those showing improvement (as set out in the study protocols by Novartis) and analysed using the Peto OR. There were benefits associated with lower dose rivastigmine compared with placebo at 26 weeks, but not at 12 weeks; and benefits with the higher dose at both 12 and 26 weeks compared with placebo:

rivastigmine 14 mg/day at 12 weeks (ITT analysis, 456/608 rivastigmine, 466/612 placebo) (OR 0.93; 95% CI 0.72 to 1.21, 3 studies);
rivastigmine 6 to 12 mg/day at 12 weeks (ITT analysis, 688/950 rivastigmine, 645/825 placebo) (OR 0.74; 95% CI 0.60 to 0.92, P = 0.008, 4 studies);
rivastigmine 1 to 4 mg/day at 26 weeks (ITT analysis, 457/614 rivastigmine, 500/623 placebo) (OR 0.71; 95% CI 0.55 to 0.93, P = 0.01, 3 studies);
rivastigmine 6 to 12 mg/day at 26 weeks (ITT analysis, 957/1330 rivastigmine, 971/1223 placebo) (OR 0.66; 95% CI 0.55 to 0.79, P < 0.00001, 6 studies).

The GDS (global assessment) carried out at 26 weeks by a clinician who had access to all information about a patient was dichotomized by counting those showing moderately severe, severe or very severe dementia against those showing moderate or mild dementia. Using the Peto OR to compare with placebo, there were benefits associated with 6 to 12 mg daily rivastigmine (ITT analysis, 579/1056 on rivastigmine showed the worse condition compared to 511/868 on placebo) (OR 0.78; 95% CI 0.64 to 0.94, P = 0.01, 4 studies) but not with 1 to 4 mg daily rivastigmine.

Behavioural symptoms

Two studies (IDEAL; Lopez‐Pousa 2005) assessed behavioural symptoms using the NPI (NPI‐10 and NPI‐12). There was no difference between rivastigmine and placebo:

rivastigmine 6 to 12 mg/day at 26 weeks (ITT analysis, WMD ‐0.06; 95% CI ‐0.20 to 0.09, 2 studies).

Withdrawals before the end of treatment

The meta‐analyses of withdrawals before the end of treatment showed no significant differences between withdrawals from the 1 to 4 mg daily rivastigmine group and from the placebo group at 12 and 26 weeks. There were significant differences for the higher dose group in favour of placebo at 12 and 26 weeks:

rivastigmine 1 to 4 mg/day at 12 weeks (17/136 rivastigmine, 8/133 placebo) (OR 2.15; 95% Cl 0.95 to 4.89, 1 study);
rivastigmine 1 to 4 mg/day at 26 weeks (113/644 rivastigmine, 113/646 placebo) (OR 1.01; 95% CI 0.75 to 1.34, 3 studies);
rivastigmine 6 to 12 mg/day at 12 weeks (20/133 rivastigmine, 8/133 placebo) (OR 2.60; 95% CI 1.19 to 5.68, P = 0.02, 1 study);
rivastigmine 6 to 12 mg/day at 26 weeks (448/1458 rivastigmine, 1194/1243 placebo) (OR 2.19; 95% CI 1.83 to 2.63, P < 0.00001, 6 studies).

Adverse events

Most adverse events occurred within the titration period. The meta‐analyses of numbers of patients with at least one adverse event showed that by the end of the titration period and at 26 weeks there were no significant differences between the lower dose rivastigmine and placebo groups. There were, however, significant differences between the higher dose rivastigmine and placebo groups in favour of placebo by the end of the titration period and at 26 weeks:

rivastigmine 1 to 4 mg/day at the end of the titration period (440/644 rivastigmine, 437/646 placebo) (OR 1.04; 95% Cl 0.82 to 1.31, 3 studies);
rivastigmine 1 to 4 mg/day at 26 weeks (509/644 rivastigmine, 518/646 placebo) (OR 0.93; 95% CI 0.71 to 1.23, 3 studies);
rivastigmine 6 to 12 mg/day at the end of the titration period (920/1072 rivastigmine, 584/878 placebo) (OR 2.96; 95% CI 2.39 to 3.68, P < 0.00001, 4 studies);
rivastigmine 6 to 12 mg/day at 26 weeks (1242/1450 rivastigmine, 901/1276 placebo) (OR 2.49; 95% CI 2.05 to 3.02, P < 0.00001, 6 studies).

A similar pattern was seen for the number of patients with at least one severe adverse event. The rivastigmine 1 to 4 mg daily group did not differ significantly from the placebo group, but there were significant differences between the rivastigmine 6 to 12 mg daily and placebo groups in favour of the latter for the titration period:

rivastigmine 1 to 4 mg/day at the end of the titration period (48/644 rivastigmine, 51/646 placebo) (OR 0.94; 95% CI 0.62 to 1.42, 3 studies);
rivastigmine 6 to 12 mg/day at the end of the titration period (130/1052 rivastigmine versus 61/868 placebo) (OR 1.88; 95% CI 1.39 to 2.55, P < 0.0001, 4 studies).

There were many types of adverse events reported and only the significant results are reported here. There were significant differences in favour of placebo for the rivastigmine 6 to 12 mg daily group by the end of the titration period and by 26 weeks for the number of patients suffering nausea, vomiting, diarrhoea, anorexia, headache, syncope, abdominal pain and dizziness. There were significant differences in favour of placebo for the rivastigmine 1 to 4 mg daily group compared to placebo by the end of the titration period and by 26 weeks for the number of patients suffering nausea, vomiting, diarrhoea and anorexia.

Withdrawals before the end of treatment due to adverse events

The meta‐analyses of withdrawals at 26 weeks due to adverse events showed no significant differences in withdrawals from the lower dose rivastigmine and placebo groups. There were, however, significant differences between the rivastigmine 6 to 12 mg daily and placebo groups in favour of placebo (291/1453 versus 94/1276) (OR 2.73, 95% CI 2.19 to 3.41, P < 0.00001, 6 studies).

Comparison of rivastigmine (20 cm² (17.4 mg/day) patch) with placebo

Cognitive function

The meta‐analysis, using MDs, showed that rivastigmine had a benefit compared with placebo for cognitive function as measured by the ADAS‐Cog at 24 weeks:

rivastigmine (ITT analysis, MD ‐2.60; 95% CI ‐3.72 to ‐1.48, P < 0.00001, 1 study).

The MMSE showed similar results in favour of rivastigmine at 26 weeks, compared with placebo:

rivastigmine (ITT analysis, MD 0.90; 95% CI 0.32 to 1.48, P = 0.002, 1 study).

The TMT‐A showed similar results in favour of rivastigmine at 26 weeks, compared with placebo:

rivastigmine (ITT analysis, MD ‐14.20; 95% CI ‐24.11 to ‐4.29, P = 0.005, 1 study).

There was no significant difference between rivastigmine and placebo for the clock drawing test.

Activities of daily living

The ADCS‐ADL showed benefit in favour of rivastigmine compared with placebo at 24 weeks:

rivastigmine (ITT analysis, MD 2.30; 95% CI 0.52 to 4.08, P = 0.01, 1 study).

Behavioural symptoms

One study assessed behavioural symptoms using the NPI (NPI‐12). There was no difference between rivastigmine and placebo (ITT analysis, MD ‐0.60; 95% CI ‐2.88 to 1.68, 1 study).

Withdrawals before the end of treatment

There was a significant difference between rivastigmine and placebo in favour of placebo for total withdrawals before the end of treatment (62/303 rivastigmine compared with 36/302 placebo) (OR 1.90; 95% CI 1.22 to 2.97, P = 0.005).

Adverse events

There was a significant difference between rivastigmine and placebo in favour of placebo for the total number of patients that had at least one adverse event by 24 weeks (200/303 rivastigmine compared with 139/302 placebo) (OR 2.28; 95% CI 1.64 to 3.16, P < 0.00001).

There was a significant difference between rivastigmine and placebo in favour of placebo for the total number of patients that had at least one adverse event of dizziness (21/303 compared with 7/302) (ITT analysis, OR 3.14; 95% CI 1.31 to 7.50, P = 0.01), nausea (64/303 compared with 15/302) (OR 5.12; 95% CI 2.85 to 9.22, P < 0.00001), vomiting (57/303 compared with 10/302) (ITT analysis, OR 6.77; 95% CI 3.38 to 13.53, P < 0.00001), weight decrease (23/303 compared with 4/302) (ITT analysis, OR 6.12; 95% CI 2.09 to 17.92, P = 0.0009), and decreased appetite (15/303 compared with 3/302) (ITT analysis, OR 5.19; 95% CI 1.49 to 18.12, P = 0.01, 1 study).

Withdrawals before the end of treatment due to adverse events

The meta‐analyses of withdrawals at 26 weeks due to adverse events showed no significant differences in withdrawals from the rivastigmine and placebo groups (26/303 rivastigmine compared with 15/302 placebo) (OR 1.80; 95% CI 0.93 to 3.46, 1 study).

Quality of life of carers

One study assessed the NPI‐D carer distress scale at 24 weeks (IDEAL). No significant difference between rivastigmine and placebo was detected (ITT analysis, MD 0.00; 95% CI ‐1.07 to 1.07).

Comparison of rivastigmine (10 cm² (9.5 mg/day) patch) with placebo

Cognitive function

The meta‐analysis, using WMDs and MDs, showed a benefit of the 10 cm² rivastigmine patch on cognitive function as measured by the ADAS‐Cog, MMSE, TMT‐A and MENFIS at 24 weeks:

ADAS‐cog (ITT analysis, WMD ‐1.34; 95% CI ‐2.02 to ‐0.66, P = 0.0001, 2 studies);
MMSE (ITT analysis, WMD 0.64; 95% CI 0.26 to 1.02, P = 0.0009, 2 studies);
TMT‐A (ITT analysis, MD ‐20.0; 95% CI ‐29.8 to ‐10.2, P < 0.0001, 1 study);
MENFIS (ITT analysis, MD ‐1.30; 95% CI ‐2.32 to ‐0.28, P = 0.01, 1 study).

Activities of daily living

The ADCS‐ADL showed benefit in favour of rivastigmine at 24 weeks (ITT analysis, MD 2.20; 95% CI 0.62 to 3.78, P = 0.006, 1 study).

The DAD showed benefit in favour of rivastigmine at 24 weeks (ITT analysis, MD 2.3; 95% CI 0.34 to 4.26, P = 0.02, 1 study).

Global assessment

The seven‐point CIBIC‐Plus scale measuring global clinical state was dichotomized by counting those showing no change or decline against those showing improvement and analysed using the Peto OR. There was no difference between rivastigmine and placebo at 24 weeks (382/518 rivastigmine, 426/545 placebo) (ITT analysis, OR 0.77; 95% CI 0.58 to 1.02, P = 0.07, 2 studies).

Withdrawals before the end of treatment

There was a significant difference between rivastigmine and placebo in favour of placebo for total withdrawals before the end of treatment (123/580 rivastigmine compared with 82/590 placebo) (OR 1.67; 95% CI 1.23 to 2.26, P = 0.001, 2 studies).

Adverse events

There were significant differences between rivastigmine and placebo in favour of placebo for the total number of patients that had at least one adverse event by 24 weeks (395/578 rivastigmine compared with 361/588 placebo) (OR 1.39; 95% CI 1.08 to 1.80, P = 0.01, 2 studies) and withdrawals due to adverse events (62/580 rivastigmine compared with 36/590 placebo) (OR 1.84; 95% CI 1.20 to 2.82, P = 0.005, 2 studies).

There were significant differences between rivastigmine and placebo in favour of placebo for the total number of patients that had at least one adverse event at the application site: erythema (113/287 compared with 55/286) (OR 2.73; 95% CI 1.87 to 3.98, P < 0.00001, 1 study), application site pruritis (100/287 compared with 61/286) (OR 1.97; 95% CI 1.36 to 2.86, P = 0.0004, 1 study), application site oedema (31/287 compared with 7/286) (OR 4.83; 95% CI 2.09 to 11.15, P = 0.0002, 1 study), application site exfoliation (11/282 compared with 4/286) (OR 3.68; 95% CI 1.20 to 1.33, P = 0.02), contact dermatitis (68/287 compared with 40/286) (OR 1.91; 95% CI 1.24 to 2.94, P = 0.003, 1 study), nausea (41/578 compared with 24/588) (OR 1.80; 95% CI 1.07 to 3.02, P = 0.03, 2 studies) and vomiting (41/578 compared with 21/588) (OR 2.06; 95% CI 1.20 to 3.53, P = 0.009, 2 studies).

Withdrawals before the end of treatment due to adverse events

There was a significant difference between rivastigmine and placebo in favour of placebo for withdrawals due to adverse events (62/580 rivastigmine compared with 36/590 placebo) (OR 1.84; 95% CI 1.20 to 2.82, P = 0.005, 2 studies).

Comparison of rivastigmine (5 cm² (4.6 mg/day) patch) with placebo

This comparison was made in one study (Nakamura 2011).

Cognitive function

There was no difference between rivastigmine and placebo at 24 weeks for cognitive function measured using the ADAS‐Cog scale (ITT analysis, MD 0.80; 95% CI ‐1.62 to 0.02), MMSE ITT analysis, MD 0.00; 95% CI ‐0.52 to 0.52) and MENFIS (ITT analysis, MD ‐0.70; 95% CI ‐0.70, 95% CI ‐1.72 to 0.32).

Activities of daily living

There was no difference between rivastigmine and placebo at 24 weeks for activities of daily living measured using the DAD scale (ITT analysis, MD 1.20; 95% CI ‐0.73 to 3.13).

Global assessment

There was no difference between rivastigmine and placebo at 24 weeks for global assessment measured using the CIBIC‐plus J scale (212/269 rivastigmine, 226/267 placebo) (ITT analysis, OR 0.67; 95% CI 0.43 to 1.05).

Behavioural symptoms

There was no difference between rivastigmine and placebo at 24 weeks for behavioural symptoms measured using the BEHAVE‐AD scale (ITT analysis, MD 0.00; 95% CI ‐0.67 to 0.67).

Withdrawals before the end of treatment

There was a significant difference between rivastigmine and placebo in favour of placebo for total withdrawals before the end of treatment (64/284 rivastigmine compared with 46/288 placebo) (OR 1.53; 95% CI 1.01 to 2.33, P = 0.05).

Adverse events

There was a significant difference between rivastigmine and placebo in favour of placebo for the total number of patients that had at least one adverse event at 24 weeks (243/282 rivastigmine compared with 222/286 placebo) (OR 1.80; 95% CI 1.16 to 2.78, P = 0.009), but no difference for deaths.

There were significant differences between rivastigmine and placebo in favour of placebo for the total number of patients that had at least one adverse event at the application site: erythema (106/282 compared with 55/286) (OR 2.53; 95% CI 1.73 to 3.70, P < 0.00001, 1 study), application site pruritis (92/282 compared with 61/286) (OR 1.79; 95% CI 1.23 to 2.60, P = 0.003, 1 study), application site oedema (35/282 compared with 7/286) (OR 5.65; 95% CI 2.46 to 12.94, P < 0.0001, 1 study), application site exfoliation (14/282 compared with 4/286) (OR 3.68; 95% CI 1.20 to 11.35, P = 0.02), contact dermatitis (69/282 compared with 40/286) (OR 1.99; 95% CI 1.30 to 3.06, P = 0.002, 1 study); but no difference between rivastigmine and placebo for adverse events of nasopharyngitis, nausea, vomiting and diarrhoea.

Withdrawals before the end of treatment due to adverse events

There was a significant difference between rivastigmine and placebo in favour of placebo for withdrawals due to adverse events (38/284 rivastigmine compared with 21/288 placebo) (OR 1.96; 95% CI 1.12 to 3.44, P = 0.02).

Comparison of rivastigmine (10 cm² (9.5 mg/day) patch) with rivastigmine (6 to 12 mg/day twice daily) capsules

Cognitive function

One study (IDEAL) showed no difference between the rivastigmine patch and rivastigmine capsules on cognitive function as measured by the ADAS‐Cog, MMSE, TMT‐A and MENFIS at 24 weeks:

ADAS‐cog (ITT analysis, MD 0.0; 95% CI ‐1.10 to 1.10, P = 1.0, 1 study);
MMSE (ITT analysis, MD 0.30; 95% CI ‐0.27 to 0.87, P = 0.30, 1 study);
TMT‐A (ITT analysis, MD ‐2.6; 95% CI ‐13.5 to 8.3, P = 0.64, 1 study);
clock drawing (ITT analysis, MD 0.1; 95% CI ‐0.5 to 0.7, P = 0.73, 1 study).

Activities of daily living

The ADCS‐ADL showed no difference between the rivastigmine patch and rivastigmine capsules at 24 weeks (ITT analysis, MD 0.40; 95% CI ‐1.23 to 2.03, P = 0.63, 1 study).

Global assessment

The seven‐point CIBIC‐Plus scale measuring global clinical state was dichotomized by counting those showing no change or decline against those showing improvement and analysed using the Peto OR. There was no difference between the rivastigmine patch and rivastigmine capsules at 24 weeks (171/248 rivastigmine patch, 161/267253 rivastigmine capsules) (ITT analysis, OR 1.27; 95% CI 0.88 to 1.84, P = 0.21, 1 study).

Behavioural symptoms

One study assessed behavioural symptoms using the NPI (NPI‐12). There was no difference between the rivastigmine patch and rivastigmine capsules (ITT analysis, MD 0.50; 95% CI ‐1.55 to 2.55, P = 0.63, 1 study).

Withdrawals before the end of treatment

There was no significant difference between rivastigmine and placebo for withdrawals before the end of treatment (64/293 compared with 63/297) (OR 1.09; 95% CI 0.70 to 154, P = 0.85, 1 study).

Adverse events

There was a significant difference between the rivastigmine patch and rivastigmine capsules in favour of the patch for the total number of patients that had at least one adverse event by 24 weeks (147/291 rivastigmine compared with 186/294 placebo) (OR 0.59; 95% CI 0.43 to 0.82, P = 0.002, 1 study).

There were significant differences between the rivastigmine patch and rivastigmine capsules in favour of the patch for the total number of patients that had at least one adverse event of decreased appetite (2/291 compared with 12/294) (OR 0.16; 95% CI 0.04 to 0.73, P = 0.02, 1 study), dizziness (7/291 compared with 22/294) (OR 0.30; 95% CI 0.13 to 0.72, P = 0.007, 1 study), asthenia (5/291 compared with 17/294) (OR 0.28; 95% CI 0.10 to 0.78, P = 0.01, 1 study), nausea (21/291 compared with 68/294) (OR 0.26; 95% CI 0.15 to 0.43, P < 0.001, 1 study) and vomiting (18/291 compared with 50/294) (OR 0.32; 95% CI 0.18 to 0.57, P < 0.001, 1 study).

Withdrawals before the end of treatment due to adverse events

There was no significant difference between rivastigmine and placebo for withdrawals due to adverse events (28/293 rivastigmine compared with 24/297 placebo) (OR 1.20; 95% CI 0.68 to 2.13, P = 0.53, 1 study).

Discussion

Summary of main results

The results of the review showed the following main findings.

The currently recommended doses of rivastigmine (6 to 12 mg/day in two divided doses for capsules and 9.5 mg/day for transdermal patches) have some benefits compared to placebo at 26 weeks for cognitive function, activities of daily living and the physician rated global impression scales. No difference was found for behavioural symptoms or the impact on carers. Patients on rivastigmine are about twice as likely (OR of about 2) to experience adverse events or to withdraw from the trial before the end of the study.
Limited evidence from one trial suggests that the transdermal formulation (9.5 mg/day) is as effective as the oral formulation (6 to 12 mg/day) and is associated with a lower incidence of adverse events but does not affect the rate of withdrawals due to adverse events.

Outcomes

The two cognitive tests used, the MMSE and ADAS‐Cog, assess similar domains and a high correlation between the results would be expected. The results from 5 studies show that 6 to 12 mg daily of oral rivastigmine improved the cognitive function of patients with mild to moderate probable Alzheimer's disease treated over a period of 26 weeks, by 0.8 points on the MMSE (range 0 to 30) and by 2.0 points on the ADAS‐Cog (range 0 to 70), when compared with placebo. The results from 2 studies show that the 9.5 mg/day of rivastigmine in a transdermal patch improved cognitive function by 0.6 points on the MMSE and 1.4 points on the ADAS‐cog when compared with placebo. Pooling the data showed a treatment effect of 0.7 points on the MMSE and 1.8 points on the ADAS‐Cog. There was a smaller effect on cognitive function in the 1 to 4 mg daily oral treatment group.

Four studies assessed the effect of 6 to 12 mg daily oral rivastigmine on activities of daily living as reported by a carer using the PDS rating scale (range 0 to 100). Rivastigmine showed a benefit of 2.2 points compared with placebo, but the difference between placebo and 1 to 4 mg daily rivastigmine was not significant. The 10 cm² (9.5 mg/day) patch showed a benefit of 2.2 points on the ADCS‐ADL scale (range 0 to 54) when compared with placebo.

The US Food and Drug Administration (FDA) requires an independent clinician to assess global clinical state after interviewing the patient and the carer at baseline and the endpoint. When the results of global impression measures were dichotomized to compare the number of patients who improved with the numbers who showed no change or whose condition had deteriorated, the 6 to 12 mg daily group was significantly better than the placebo group at 12 and 26 weeks, and there was a similar significant difference favouring the 1 to 4 mg daily group over placebo at 26 weeks. The 10 cm² (9.5 mg/day) patch was also significantly better than placebo at 24 weeks.The clinician and carer, whilst following the guidelines for the application of the CIBIC‐Plus, are essentially making an assessment of whether the patient has improved or not based on criteria relevant to them. This is perhaps closest to what is commonly meant by the term 'meaningful improvement'.

Minimal clinically important differences (MCID), patient derived scores that represent changes in a score that have meaning for patients, have been suggested for the ADAS‐cog (3 points in severe AD (Howard 2011)) and MMSE (1.4 points in mild AD (Schrag 2012)). Comparing our findings with these we might conclude that the treatment effects for cognitive function are unlikely to be clinically relevant.

Adverse effects

When taking capsules, a fairly lengthy titration period of up to 12 weeks is needed to develop tolerance and to minimize adverse effects such as nausea, vomiting, diarrhoea, abdominal pain, dizziness, headache and anorexia. The target was to treat patients with a maximum tolerated dose administered in two divided doses, the upper limit being 12 mg per day. There were significantly more total dropouts and dropouts due to adverse events from the 6 to 12 mg daily dose groups than from placebo groups and therefore adverse effects remain a clinical issue. There was no hepatotoxicity associated with rivastigmine and no statistically or clinically significant changes in vital signs.

The continuous dose patch was introduced to improve tolerability. One study (IDEAL) tested two sizes of rivastigmine patch, one delivering a higher dose than previously tested in a 20 cm² patch (17.4 mg/day) and one 10 cm² patch (9.5 mg/day), a dose similar to the usual oral dose. Another study (Nakamura 2011) tested 5 cm² (4.6 mg/day) and 10 cm² (9.5 mg/day) patches. The smallest patch showed no treatment effect when compared with placebo for cognition, global function and activities of daily living. The efficacy of the 9.5 mg/day patch was comparable to that of the capsules with a similar daily dose, but was associated with significantly fewer adverse events of nausea, vomiting, dizziness and asthenia. There was no difference in the number of withdrawals due to adverse events. Therefore, the 9.5 mg/day patch appears to have advantages compared with both the higher dose patches and the 6 to 12 mg/day capsules in terms of the overall incidence of adverse events, but it may not reduce the incidence of the more serious events that lead to cessation of treatment.

Overall completeness and applicability of evidence

We were able to include evidence from both published and unpublished trials in this systematic review. There were 3319 participants. Most participants were in industry sponsored trials. Data from two independent trials (n = 162) were not available and we excluded data from two other independent trials (n = 75) from our analyses because of concerns about risk of bias. The participants in the included trials had mainly mild to moderate dementia due to Alzheimer's disease. They were not highly selected with respect to their general health so that all but the seriously ill were included. Only two trials included patients with severe dementia, and we excluded the data from one of these leaving data on only 218 patients with severe dementia included in the analyses.

The main limitations in the completeness and applicability of the evidence were the lack of long term data beyond 26 weeks and the limited range of outcomes measured. Beyond 26 weeks, some trials continued as an open label, extension phase. There were very few data on outcomes important to patients and carers, such as quality of life.

Quality of the evidence

The quality of the evidence at 26 weeks is moderate for most outcomes. Our main concern for the evidence is that only seven studies have contributed data to the meta‐analysis, and all of these studies were either industry sponsored or industry funded. In addition, withdrawals from these studies were of concern.

The results from B352 nearly always showed greater benefits for rivastigmine on each outcome than demonstrated in B303/B305, B304 and B351. B352 was responsible for the heterogeneity between trials that was reported for some of the measures of cognitive function. There are no obvious differences between B352 and the other phase III trials. It was conducted only in the US, but so was B351. The doses reached by patients in B352 were higher than those of B351, by 1.2 mg per day on average. Results from B352 have been more extensively reported than the other three phase III trials but there is no reason to suppose that this trial is of any more importance in the overall assessment of rivastigmine. We have not downgraded the evidence based on this heterogeneity concern since the impact on the overall pooled results is small and does not change the interpretation of any of the results.

Outcomes such as behavioural symptoms and quality of life are important to patients but these were only reported by three studies and one study, respectively. Hence, the quality of evidence for these was lower.

Potential biases in the review process

The initial protocol of the review, which was published in 1998, had aimed to include all double‐blinded randomised controlled trials (RCTs) of rivastigmine with a minimum study period of two weeks, regardless of the doses or formulations used. However, this resulted in a large number of possible comparisons. In addition, studies often used multiple instruments to report the same outcome, for example cognitive function was measured using the MMSE, ADAS‐Cog and other tests. For some of these outcomes we decided to use only the most commonly used tests in the main analysis.

In this update we decided to concentrate on the currently recommended doses (6 and 12 mg/day for oral doses, and 9.5 mg/day for transdermal patches), and a minimum treatment duration of six months for the main analysis. We considered the decision to focus on longer term data was clinically sensible since a titration period was required to reach the target doses.

Mowla 2007, Karaman 2005, Ballard 2005 and Tai 2000, all non‐industry funded studies, did not provide data that could be included in the review.

Agreements and disagreements with other studies or reviews

Most patients from the four phase III trials continued in an open label phase for a further 26 weeks during which the maximum tolerated dose was administered. Results from these extension phases have been described as showing a possible beneficial effect of rivastigmine on disease progression (Product monograph, Novartis 1998). Reported results showed that patients who had received placebo or rivastigmine 1 to 4 mg daily in the randomised phase showed initial improvement on the ADAS‐Cog before declining at the same rate as the 6 to 12 mg daily group, although remaining more impaired by approximately 1.5 points. These results must be interpreted with caution. The randomised, double‐blind conditions no longer prevailed. There had been differential dropout from the groups and there was no placebo group for the comparison. An imputed rate of decline for placebo patients was obtained by extrapolating from the randomised phase and not from actual observations.

There is much interest in the identification of patient characteristics that might predict a response to a cholinesterase inhibitor. Burns 2004 reported that cholinesterase inhibitors may be effective in patients with more severe disease. Data were pooled from three studies (B303/B305; B351; B352) for those with a baseline MMSE of 10, 11 or 12 (n = 117), in the group treated with rivastigmine (6 to 12 mg/day) or placebo, and the analysis showed that rivastigmine benefited those with more severe disease. This result has not added anything substantial to what was known already. The analysis of the total dataset from these trials demonstrated that rivastigmine was of benefit to the population randomised.

Erkinjuntti 2002, funded by Novartis Pharmaceuticals, investigated the response to rivastigmine of those without hypertension compared to those with using the data from B303/B305. They reported that particular benefits may be observed in those with vascular risk factors. These results are based on retrospective analysis of the study data and there has been no study confirming this finding using prospective data.

Farlow 2003a retrieved dropouts from the studies B303/B305, B351 and B352. These patients stopped treatment before the end of the trials but were invited back for assessment at the endpoint. Farlow concluded that those who had been in the rivastigmine groups had deteriorated less than those from the placebo groups, and therefore rivastigmine had provided a beneficial delay in disease progression. The two groups cannot be compared. The participants belong to a highly selected group, those who stopped treatment and agreed to return. The placebo group was much smaller than the rivastigmine group (38 compared with 88). Those who left the trial from the placebo group may have done so because their illness was more severe. This may have applied to some of those in the rivastigmine group who left but, in addition, there were those who left because they suffered from adverse effects. It is not possible to compare these two groups in a meaningful way.

Grossberg 2000 was a Novartis funded extension study examining the data from the four phase III studies and the related open label extension studies. Those who had been taking rivastigmine continuously for two years were compared with historical controls and the study concluded that rivastigmine has a beneficial effect on cognitive performance for up to two years in patients with Alzheimer's disease. These results must be treated with caution as the two groups are not comparable.

Several reviews of rivastigmine have been published. Schneider 1998 and Spencer 1998 both limited analysis and interpretation to the three trials B352, B303/B305 and B351. Spencer reported that "individual and pooled results indicate that rivastigmine usually produces cognitive, global and functional changes that indicate significantly less deterioration than was observed with placebo". Schneider reported that "the pooled analyses confirm the efficacy of rivastigmine in the treatment of both the cognitive and functional deficits of mild to moderately severe AD". Clegg 2002 was the report from NICE (National Institute of Clinical Excellence, UK) of the systematic review on which the decision was made that the cholinesterase inhibitors would be available on the National Health Service to treat those with Alzheimer's disease. Williams 2003 was a review of all aspects of rivastigmine, and summarised the clinical trials but without meta‐analyses.

Hauber 2000 calculated the potential savings costs using rivastigmine compared with no treatment for Alzheimer's disease. Hauber used a disease stage model. Results from two phase III trials of rivastigmine, together with extrapolation beyond the six month duration of the trial, identified the stage of disease using the MMSE assessments. Costs of healthcare resource use was estimated as a function of MMSE, using data from Canadian sources. Rivastigmine was judged to be cost effective due to the delay in disease progression. The analysis was repeated in a UK and US setting. These results are not based on randomised evidence and rest on many assumptions. It would be unwise to base decisions on whether rivastigmine should be prescribed to patients on the basis of cost‐effectiveness studies such as these. Fillit 2004 presents an excellent summary of the cost‐effectiveness studies and of the assumptions on which they are based. Fillit concludes that the results from these studies are not reliable and that outcomes related to costs and healthcare resource use must be assessed in randomised clinical trials.

Figure 1

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Figure 2

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm2 (9.5 mg/day) patch) versus placebo, Outcome 1 ADAS‐Cog (change from baseline at 24‐26 weeks) ITT.

Analysis 1.1

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm² (9.5 mg/day) patch) versus placebo, Outcome 1 ADAS‐Cog (change from baseline at 24‐26 weeks) ITT.

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm2 (9.5 mg/day) patch) versus placebo, Outcome 2 MMSE (change from baseline at 24‐26 weeks) ITT.

Analysis 1.2

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm² (9.5 mg/day) patch) versus placebo, Outcome 2 MMSE (change from baseline at 24‐26 weeks) ITT.

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm2 (9.5 mg/day) patch) versus placebo, Outcome 3 Activities of daily living (change from baseline at 24‐26 weeks ) ITT.

Analysis 1.3

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm² (9.5 mg/day) patch) versus placebo, Outcome 3 Activities of daily living (change from baseline at 24‐26 weeks ) ITT.

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm2 (9.5 mg/day) patch) versus placebo, Outcome 4 Clinical Global Impression (no change or worse at 24‐26 weeks ) ITT.

Analysis 1.4

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm² (9.5 mg/day) patch) versus placebo, Outcome 4 Clinical Global Impression (no change or worse at 24‐26 weeks ) ITT.

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm2 (9.5 mg/day) patch) versus placebo, Outcome 5 Behavioural symptoms (change from baseline at 24‐26 weeks) ITT.

Analysis 1.5

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm² (9.5 mg/day) patch) versus placebo, Outcome 5 Behavioural symptoms (change from baseline at 24‐26 weeks) ITT.

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm2 (9.5 mg/day) patch) versus placebo, Outcome 6 Withdrawals before end of treatment at 24‐26 weeks.

Analysis 1.6

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm² (9.5 mg/day) patch) versus placebo, Outcome 6 Withdrawals before end of treatment at 24‐26 weeks.

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm2 (9.5 mg/day) patch) versus placebo, Outcome 7 at least one adverse event by 24‐26 weeks.

Analysis 1.7

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm² (9.5 mg/day) patch) versus placebo, Outcome 7 at least one adverse event by 24‐26 weeks.

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm2 (9.5 mg/day) patch) versus placebo, Outcome 8 NPI‐D carer distress scale (change from baseline at 24‐26 weeks) ITT.

Analysis 1.8

Comparison 1 Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm² (9.5 mg/day) patch) versus placebo, Outcome 8 NPI‐D carer distress scale (change from baseline at 24‐26 weeks) ITT.

Analysis 2.1

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 1 ADAS‐Cog (change from baseline at 12 weeks) ITT.

Analysis 2.2

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 2 ADAS‐Cog (change from baseline at 26 weeks) ITT.

Analysis 2.3

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 3 MMSE (change from baseline at 26 weeks) ITT.

Analysis 2.4

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 4 SIB (change from baseline at 26 weeks).

Analysis 2.5

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 5 ADCS‐ADL (change from baseline at 26 weeks) ITT.

Analysis 2.6

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 6 PDS (change from baseline at 12 weeks ) ITT.

Analysis 2.7

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 7 PDS (change from baseline at 26 weeks ) ITT.

Analysis 2.8

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 8 Clinical Global Impression (no change or worse at 12 weeks) ITT.

Analysis 2.9

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 9 Clinical Global Impression (no change or worse at 26 weeks ) ITT.

Analysis 2.10

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 10 GDS( moderately severe, severe, or very severe dementia at 26 weeks) ITT.

Analysis 2.11

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 11 CGIC (little or no improvement, or worse at 12 weeks) ITT.

Analysis 2.12

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 12 Behavioural disturbance NPI‐10 or NPI‐12 (change from baseline at 26 weeks) ITT.

Analysis 2.13

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 13 withdrawals before end of treatment at 12 weeks.

Analysis 2.14

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 14 withdrawals before end of treatment at 26 weeks.

Analysis 2.15

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 15 at least one adverse event by the end of titration period.

Analysis 2.16

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 16 at least one adverse event by 26 weeks.

Analysis 2.17

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 17 dropouts due to adverse events by 12 weeks.

Analysis 2.18

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 18 dropouts due to adverse events by 26 weeks.

Analysis 2.19

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 19 at least one adverse event of decreased appetite by 26 weeks.

Analysis 2.20

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 20 at least one adverse event of weight decrease by 26 weeks.

Analysis 2.21

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 21 at least one adverse event of nausea by the end of titration period.

Analysis 2.22

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 22 at least one adverse event of nausea by 26 weeks.

Analysis 2.23

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 23 at least one adverse event of vomiting by the end of titration period.

Analysis 2.24

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 24 at least one adverse event of vomiting by 26 weeks.

Analysis 2.25

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 25 at least one adverse event of diarrhoea by the end of titration period.

Analysis 2.26

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 26 at least one adverse event of diarrhoea by 26 weeks.

Analysis 2.27

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 27 at least one adverse event of anorexia by the end of titration period.

Analysis 2.28

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 28 at least one adverse event of anorexia by 26 weeks.

Analysis 2.29

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 29 at least one adverse event of headache by the end of titration period.

Analysis 2.30

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 30 at least one adverse event of headache by 26 weeks.

Analysis 2.31

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 31 at least one adverse event of insomnia by the end of titration period.

Analysis 2.32

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 32 at least one adverse event of insomnia by 26 weeks.

Analysis 2.33

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 33 at least one adverse event of syncope by the end of titration period.

Analysis 2.34

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 34 at least one adverse event of syncope by 26 weeks.

Analysis 2.35

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 35 at least one adverse event of abdominal pain by the end of titration period.

Analysis 2.36

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 36 at least one adverse event of abdominal pain by 26 weeks.

Analysis 2.37

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 37 at least one adverse event of dizziness by the end of titration period.

Analysis 2.38

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 38 at least one adverse event of dizziness by 26 weeks.

$Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 39 at least one adverse event of bone fracture by the end of titration period.$

Analysis 2.39

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 39 at least one adverse event of bone fracture by the end of titration period.

$Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 40 at least one adverse event of bone fracture by 26 weeks.$

Analysis 2.40

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 40 at least one adverse event of bone fracture by 26 weeks.

Analysis 2.41

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 41 at least one adverse event of asthenia by 26 weeks.

Analysis 2.42

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 42 at least one severe adverse event by the end of titration period.

Analysis 2.43

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 43 at least one serious adverse event by 26 weeks.

Analysis 2.44

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 44 deaths before end of treatment at 12 weeks.

Analysis 2.45

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 45 deaths before end of treatment at 26 weeks.

Analysis 2.46

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 46 CIBIC‐Plus (no change or worse at 12 weeks) OC.

Analysis 2.47

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 47 CIBIC‐Plus (no change or worse at 26 weeks) OC.

Analysis 2.48

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 48 CIBIC‐Plus (no change or worse at 12 weeks) OC+RDO.

Analysis 2.49

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 49 CIBIC‐Plus (no change or worse at 26 weeks)OC+RDO.

Analysis 2.50

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 50 CIBIC‐Plus (no change or worse at 12 weeks) ALL+OC.

Analysis 2.51

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 51 CIBIC‐Plus (no change or worse at 26 weeks) ALL+OC.

Analysis 2.52

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 52 ADAS‐Cog (change from baseline at 12 weeks) OC.

Analysis 2.53

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 53 ADAS‐Cog (change from baseline at 26 weeks) OC.

Analysis 2.54

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 54 ADAS‐Cog (change from baseline at 12 weeks) OC+RDO.

Analysis 2.55

Comparison 2 Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo, Outcome 55 ADAS‐Cog (change from baseline at 26 weeks) OC+RDO.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 1 ADAS‐Cog (change from baseline at 24 weeks) ITT.

Analysis 3.1

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 1 ADAS‐Cog (change from baseline at 24 weeks) ITT.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 2 TMT‐A (change from baseline at 24 weeks) ITT.

Analysis 3.2

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 2 TMT‐A (change from baseline at 24 weeks) ITT.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 3 clock drawing (change from baseline at 24 weeks) ITT.

Analysis 3.3

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 3 clock drawing (change from baseline at 24 weeks) ITT.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 4 MMSE (change from baseline at 24 weeks) ITT.

Analysis 3.4

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 4 MMSE (change from baseline at 24 weeks) ITT.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 5 ADCS‐ADL (change from baseline at 24 weeks) ITT.

Analysis 3.5

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 5 ADCS‐ADL (change from baseline at 24 weeks) ITT.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 6 NPI‐12 (change from baseline at 24 weeks) ITT.

Analysis 3.6

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 6 NPI‐12 (change from baseline at 24 weeks) ITT.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 7 withdrawals before end of treatment at 24 weeks.

Analysis 3.7

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 7 withdrawals before end of treatment at 24 weeks.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 8 at least one adverse event by 24 weeks.

Analysis 3.8

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 8 at least one adverse event by 24 weeks.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 9 withdrawals due to an adverse event before end of treatment at 24 weeks.

Analysis 3.9

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 9 withdrawals due to an adverse event before end of treatment at 24 weeks.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 10 at least one adverse event of dizziness by 24 weeks.

Analysis 3.10

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 10 at least one adverse event of dizziness by 24 weeks.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 11 at least one adverse event of nausea by 24 weeks.

Analysis 3.11

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 11 at least one adverse event of nausea by 24 weeks.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 12 at least one adverse event of vomiting by 24 weeks.

Analysis 3.12

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 12 at least one adverse event of vomiting by 24 weeks.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 13 at least one adverse event of weight decrease by 24 weeks.

Analysis 3.13

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 13 at least one adverse event of weight decrease by 24 weeks.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 14 at least one adverse event of decreased appetite by 24 weeks.

Analysis 3.14

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 14 at least one adverse event of decreased appetite by 24 weeks.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 15 at least one adverse event of headache by 24 weeks.

Analysis 3.15

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 15 at least one adverse event of headache by 24 weeks.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 16 at least one adverse event of asthenia by 24 weeks.

Analysis 3.16

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 16 at least one adverse event of asthenia by 24 weeks.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 17 deaths before end of treatment at 24 weeks.

Analysis 3.17

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 17 deaths before end of treatment at 24 weeks.

Comparison 3 Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo, Outcome 18 NPI‐D carer distress scale (change from baseline at 24 weeks) ITT.

Analysis 3.18

Comparison 3 Rivastigmine 20 cm² patch (17.4 mg/day) versus placebo, Outcome 18 NPI‐D carer distress scale (change from baseline at 24 weeks) ITT.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 1 ADAS‐Cog (change from baseline at 24 weeks) ITT.

Analysis 4.1

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 1 ADAS‐Cog (change from baseline at 24 weeks) ITT.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 2 MMSE (change from baseline at 24 weeks) ITT.

Analysis 4.2

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 2 MMSE (change from baseline at 24 weeks) ITT.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 3 clock drawing (change from baseline at 24 weeks) ITT.

Analysis 4.3

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 3 clock drawing (change from baseline at 24 weeks) ITT.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 4 TMT‐A (change from baseline at 24 weeks) ITT.

Analysis 4.4

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 4 TMT‐A (change from baseline at 24 weeks) ITT.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 5 Mental Function Impairment MENFIS (change from baseline at 24 weeks) ITT.

Analysis 4.5

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 5 Mental Function Impairment MENFIS (change from baseline at 24 weeks) ITT.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 6 ADCS‐ADL (change from baseline at 24 weeks) ITT.

Analysis 4.6

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 6 ADCS‐ADL (change from baseline at 24 weeks) ITT.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 7 Disability Assessment for Dementia (DAD) (change from baseline at 24 weeks) ITT.

Analysis 4.7

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 7 Disability Assessment for Dementia (DAD) (change from baseline at 24 weeks) ITT.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 8 BEHAVE‐AD (change from baseline at 24 weeks) ITT.

Analysis 4.8

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 8 BEHAVE‐AD (change from baseline at 24 weeks) ITT.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 9 NPI‐12 (change from baseline at 24 weeks) ITT.

Analysis 4.9

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 9 NPI‐12 (change from baseline at 24 weeks) ITT.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 10 Clinical Global Impression (no change or worse at 24 weeks).

Analysis 4.10

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 10 Clinical Global Impression (no change or worse at 24 weeks).

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 11 withdrawals before end of treatment at 24 weeks.

Analysis 4.11

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 11 withdrawals before end of treatment at 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 12 at least one adverse event by 24 weeks.

Analysis 4.12

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 12 at least one adverse event by 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 13 withdrawals due to an adverse event before end of treatment at 24 weeks.

Analysis 4.13

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 13 withdrawals due to an adverse event before end of treatment at 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 14 at least one adverse event of application site erythema by 24 weeks.

Analysis 4.14

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 14 at least one adverse event of application site erythema by 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 15 at least one adverse event of application site pruritis by 24 weeks.

Analysis 4.15

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 15 at least one adverse event of application site pruritis by 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 16 at least one adverse event of application site edema by 24 weeks.

Analysis 4.16

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 16 at least one adverse event of application site edema by 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 17 at least one adverse event application site exfoliation by 24 weeks.

Analysis 4.17

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 17 at least one adverse event application site exfoliation by 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 18 at least one adverse event of dermatitis contact by 24 weeks.

Analysis 4.18

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 18 at least one adverse event of dermatitis contact by 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 19 at least one adverse event of nasopharyngitis by 24 weeks.

Analysis 4.19

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 19 at least one adverse event of nasopharyngitis by 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 20 at least one adverse event of nausea by 24 weeks.

Analysis 4.20

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 20 at least one adverse event of nausea by 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 21 at least one adverse event of vomiting by 24 weeks.

Analysis 4.21

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 21 at least one adverse event of vomiting by 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 22 at least one adverse event of diarrhoea by 24 weeks.

Analysis 4.22

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 22 at least one adverse event of diarrhoea by 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 23 at least one adverse event of weight decrease by 24 weeks.

Analysis 4.23

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 23 at least one adverse event of weight decrease by 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 24 at least one adverse event of dizziness by 24 weeks.

Analysis 4.24

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 24 at least one adverse event of dizziness by 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 25 at least one adverse event of decreased appetite by 24 weeks.

Analysis 4.25

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 25 at least one adverse event of decreased appetite by 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 26 at least one adverse event of headache by 24 weeks.

Analysis 4.26

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 26 at least one adverse event of headache by 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 27 at least one adverse event of asthenia by 24 weeks.

Analysis 4.27

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 27 at least one adverse event of asthenia by 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 28 deaths before end of treatment at 24 weeks.

Analysis 4.28

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 28 deaths before end of treatment at 24 weeks.

Comparison 4 Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo, Outcome 29 NPI‐D carer distress scale (change from baseline at 24 weeks) ITT.

Analysis 4.29

Comparison 4 Rivastigmine 10 cm² patch (9.5 mg/day) versus placebo, Outcome 29 NPI‐D carer distress scale (change from baseline at 24 weeks) ITT.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 1 ADAS‐J Cog (change from baseline at 24 weeks) ITT.

Analysis 5.1

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 1 ADAS‐J Cog (change from baseline at 24 weeks) ITT.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 2 MMSE (change from baseline at 24 weeks) ITT.

Analysis 5.2

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 2 MMSE (change from baseline at 24 weeks) ITT.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 3 Mental Function Impairment MENFIS (change from baseline at 24 weeks) ITT.

Analysis 5.3

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 3 Mental Function Impairment MENFIS (change from baseline at 24 weeks) ITT.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 4 Disability Assessment for Dementia (DAD) (change from baseline at 24 weeks) ITT.

Analysis 5.4

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 4 Disability Assessment for Dementia (DAD) (change from baseline at 24 weeks) ITT.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 5 CIBIC‐Plus J (no change or worse at 24 weeks) ITT.

Analysis 5.5

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 5 CIBIC‐Plus J (no change or worse at 24 weeks) ITT.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 6 BEHAVE‐AD (change from baseline at 24 weeks) ITT.

Analysis 5.6

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 6 BEHAVE‐AD (change from baseline at 24 weeks) ITT.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 7 withdrawals before end of treatment at 24 weeks.

Analysis 5.7

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 7 withdrawals before end of treatment at 24 weeks.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 8 at least one adverse event by 24 weeks.

Analysis 5.8

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 8 at least one adverse event by 24 weeks.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 9 withdrawals due to an adverse event before end of treatment at 24 weeks.

Analysis 5.9

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 9 withdrawals due to an adverse event before end of treatment at 24 weeks.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 10 at least one adverse event of application site erythema by 24 weeks.

Analysis 5.10

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 10 at least one adverse event of application site erythema by 24 weeks.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 11 at least one adverse event of application site pruritis by 24 weeks.

Analysis 5.11

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 11 at least one adverse event of application site pruritis by 24 weeks.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 12 at least one adverse event of application site edema by 24 weeks.

Analysis 5.12

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 12 at least one adverse event of application site edema by 24 weeks.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 13 at least one adverse event application site exfoliation by 24 weeks.

Analysis 5.13

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 13 at least one adverse event application site exfoliation by 24 weeks.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 14 at least one adverse event of dermatitis contact by 24 weeks.

Analysis 5.14

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 14 at least one adverse event of dermatitis contact by 24 weeks.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 15 at least one adverse event of nasopharyngitis by 24 weeks.

Analysis 5.15

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 15 at least one adverse event of nasopharyngitis by 24 weeks.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 16 at least one adverse event of nausea by 24 weeks.

Analysis 5.16

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 16 at least one adverse event of nausea by 24 weeks.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 17 at least one adverse event of vomiting by 24 weeks.

Analysis 5.17

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 17 at least one adverse event of vomiting by 24 weeks.

Comparison 5 Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo, Outcome 18 deaths before end of treatment at 24 weeks.

Analysis 5.18

Comparison 5 Rivastigmine 5 cm² patch (4.6 mg/day) versus placebo, Outcome 18 deaths before end of treatment at 24 weeks.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 1 ADAS‐Cog (change from baseline at 24 weeks) ITT.

Analysis 6.1

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 1 ADAS‐Cog (change from baseline at 24 weeks) ITT.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 2 MMSE (change from baseline at 24 weeks) ITT.

Analysis 6.2

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 2 MMSE (change from baseline at 24 weeks) ITT.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 3 clock drawing (change from baseline at 24 weeks) ITT.

Analysis 6.3

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 3 clock drawing (change from baseline at 24 weeks) ITT.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 4 TMT‐A (change from baseline at 24 weeks) ITT.

Analysis 6.4

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 4 TMT‐A (change from baseline at 24 weeks) ITT.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 5 ADCS‐ADL (change from baseline at 24 weeks) ITT.

Analysis 6.5

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 5 ADCS‐ADL (change from baseline at 24 weeks) ITT.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 6 Clinical Global Impression (no change or worse at 24 weeks).

Analysis 6.6

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 6 Clinical Global Impression (no change or worse at 24 weeks).

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 7 NPI‐12 (change from baseline at 24 weeks) ITT.

Analysis 6.7

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 7 NPI‐12 (change from baseline at 24 weeks) ITT.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 8 withdrawals before end of treatment at 24 weeks.

Analysis 6.8

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 8 withdrawals before end of treatment at 24 weeks.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 9 at least one adverse event by 24 weeks.

Analysis 6.9

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 9 at least one adverse event by 24 weeks.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 10 withdrawals due to an adverse event before end of treatment at 24 weeks.

Analysis 6.10

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 10 withdrawals due to an adverse event before end of treatment at 24 weeks.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 11 at least one adverse event of nausea by 24 weeks.

Analysis 6.11

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 11 at least one adverse event of nausea by 24 weeks.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 12 at least one adverse event of vomiting by 24 weeks.

Analysis 6.12

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 12 at least one adverse event of vomiting by 24 weeks.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 13 at least one adverse event of diarrhoea by 24 weeks.

Analysis 6.13

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 13 at least one adverse event of diarrhoea by 24 weeks.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 14 at least one adverse event of weight decrease by 24 weeks.

Analysis 6.14

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 14 at least one adverse event of weight decrease by 24 weeks.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 15 at least one adverse event of dizziness by 24 weeks.

Analysis 6.15

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 15 at least one adverse event of dizziness by 24 weeks.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 16 at least one adverse event of decreased appetite by 24 weeks.

Analysis 6.16

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 16 at least one adverse event of decreased appetite by 24 weeks.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 17 at least one adverse event of headache by 24 weeks.

Analysis 6.17

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 17 at least one adverse event of headache by 24 weeks.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 18 at least one adverse event of asthenia by 24 weeks.

Analysis 6.18

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 18 at least one adverse event of asthenia by 24 weeks.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 19 deaths before end of treatment at 24 weeks + 30 days.

Analysis 6.19

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 19 deaths before end of treatment at 24 weeks + 30 days.

Comparison 6 Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 20 NPI‐D carer distress scale (change from baseline at 24 weeks) ITT.

Analysis 6.20

Comparison 6 Rivastigmine 10 cm² patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses), Outcome 20 NPI‐D carer distress scale (change from baseline at 24 weeks) ITT.

Summary of findings for the main comparison. Rivastigmine compared to placebo for Alzheimer's disease

Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm² (9.8 mg/day) patch) compared to placebo for Alzheimer's disease
Patient or population: patients with Alzheimer's disease, mild to moderate Settings: multicentre, mostly in Europe or United States Intervention: rivastigmine (capsules 6 to 12 mg/day in 2 divided doses or 10 cm² patch) for 24 to 26 weeks Comparison: placebo for 24 to 26 weeks
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo	Rivastigmine (capsules 6 to 12 mg/day b.i.d. or 10 cm² patch)
Cognitive function (change from baseline at 26 weeks using ADAS‐Cog)		The mean score in the rivastigmine group was 1.79 lower (2.21 to 1.37 lower)		3232 (6 studies)	⊕⊕⊕⊝ moderate^1,2,	ADAS‐Cog score has a maximum of 70 points, the lower score of the rivastigmine group indicates greater improvement
Cognitive function (change from baseline at 26 weeks using MMSE)		The mean score in the rivastigmine group was 0.74 higher (0.52 to 0.97 higher)		3205 (6 studies)	⊕⊕⊕⊝ moderate^1,2	MMSE has a maximum score of 30 points, a lower score indicates greater impairment. treatment effect was in favour of rivastigmine
Activities of daily living (change from baseline at 26 weeks measured using various scales)		The mean score in the rivastigmine group was 0.2 standard deviations higher (0.13 to 0.27 higher)		3230 (6 studies)	⊕⊕⊕⊝ moderate¹	SMD 0.2 (0.13 to 0.27) A SMD of 0.2 is considered a small effect size. Treatment effect in favour of rivastigmine
Physician rated global impression tests (no change or worse compared with baseline, measured using Global Impression of Change at 26 weeks)	810 per 1000	744 per 1000 (712 to 773)	OR 0.68 (0.58 to 0.8)	3338 (7 studies)	⊕⊕⊕⊝ moderate¹	Treatment effect was in favour of rivastigmine
Behavioural symptoms (change from baseline at 26 weeks measured using various scales)		The mean score in the rivastigmine group was 0.04 standard deviations lower (0.14 lower to 0.06 higher)		1529 (3 studies)	⊕⊕⊕⊝ moderate^1,3	SMD ‐0.04 (‐0.14 to 0.06) A SMD of 0.2 is considered a small effect size. The size of this SMD and its small confidence interval suggests that there is no difference between the two groups
Acceptability of treatment (as measured by withdrawals from trials before end of treatment at 26 weeks)	149 per 1000	260 per 1000 (230 to 293)	OR 2.01 (1.71, 2.37)	3569 (7 studies)	⊕⊕⊕⊝ moderate¹	Withdrawals significantly more frequent in rivastigmine group compared with placebo group
Incidence of adverse events (at least one adverse event by 26 weeks)	761 per 1000	870 per 1000 (850 to 888)	OR 2.14 (1.80 to 2.53)	3587 (7 studies)	⊕⊕⊕⊝ moderate¹	Adverse events significantly more frequent in rivastigmine group compared with placebo group
Quality of life of patients or carers (measured using NPI‐D carer distress scale (change from baseline at 24 weeks)		The mean score in the rivastigmine group was 0.1 higher (0.91 lower to 1.11 higher)		529 (1 study)	⊕⊕⊕⊝ moderate¹	The size of this MD and its small confidence interval suggests that there is no difference between the two groups
The assumed risk* used the median control group risk across studies. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; OR: Odds ratio.
¹ Confidence in estimate of effect lowered due to relatively high dropout rates across studies, which are higher in the treatment group. The ITT analysis in these studies used LOCF (last observed carried forward) imputations. In addition, results are available up to only 26 weeks, longer term data would be more applicable. ² There was high heterogeneity the ADAS‐Cog outcome due to B352, which had high dropout rates and showed a difference of 3.8 points, compared to 1.2 to 1.6 points for the other studies. However, evidence not further downgraded; removal of this study from the analysis will only result in a small change of estimate by about 0.35 points. ³ Three studies (IDEAL; Lopez‐Pousa 2005; Nakamura 2011) reported a scale measuring behavioural disturbance. ⁴ The protocol for most studies had some measures related to quality of life or impact on carers, but only one study reported this (IDEAL).

Summary of findings for the main comparison. Rivastigmine compared to placebo for Alzheimer's disease

Table 1. Description of the included studies at baseline

Study	Duration (weeks)	Participants	Mean age (SD)	% males	Mean MMSE (SD)	country	Number of centres	Treatment groups
Oral (different doses versus placebo )
B103 (Phase II)	13	402	69.4	44	‐	Europe	54	4 mg/day b.i.d 6 mg/day b.i.d placebo
B104 (Phase II)	18	114	71.2 (7.5)	39	19.5 (3.7)	Belgium, France, UK, Norway, Canada	11	6 to 12 mg/day b.i.d. 6 to 12 mg/day t.i.d placebo
B303/B305* (Phase III)	26	725	72.0 (8.1)	41	20.0 (4.5)	France, Germany, Austria, Switzerland, Canada, USA	44	1 to 4 mg b.i.d 6 to 12 mg/day b.i.d., placebo
B304* (Phase III)	26	677	71.4 (8.2)	41	18.5 (4.5)	UK, Ireland, Australia, Canada, RSA, Italy	37	2 to 12 mg/day b.i.d. 2 to 12 mg/day t.i.d. placebo
B351* (Phase III)	26	702	74.1 (8.3)	44	20.0 (4.4)	USA	14	3 mg/day t.i.d 6 mg/day t.i.d 9 mg/day b.i.d placebo
B352* (Phase III)	26	699	74.5 (7.4)	39	19.7 (4.5)	USA	22	1 to 4 mg per day b.i.d 6 to 12 mg/day b.i.d. placebo
Ballard 2005	26	93	83.8 (7.7)	20	‐	UK	‐	6 to 12 mg/day b.i.d placebo
Karaman 2005*	52	44	73.8	45	12.2	Turkey	1	6 to 12 mg/day b.i.d. placebo
Lopez‐Pousa 2005*	26	218	77.6	23	8.8	Spain	21	6 to 12 mg/day b.i.d. placebo
Mowla 2007	12	122	69.2	46	16.1 (4.0)	Iran	‐	6 to 12 mg/day b.i.d. placebo
Tai 2000	26	80	‐	‐	‐	Taiwan	‐	3 to 6 mg/day b.i.d. placebo
Oral and patches
IDEAL* (Phase III)	24	1195	73.3 (7.8)	33	16.5 (3.0)	North, Central and South America, Asia, Europe	100	patch 9.5 mg/day patch 17.4 mg/day capsules 6 to 12 mg/day b.i.d. placebo
Patches
Nakamura 2011	24	859	74.6 (7.2)	31.7	16.6 (3.0)	Japan	multicentre	patch 4.6 mg/day patch 9.5 mg/day placebo
* These studies met the inclusion criteria of the main analysis comparing rivastigmine at the therapeutic doses versus placebo. b.i.d = bis in die in Latin, this means that a medication is taken two times a day, dividing the total daily dose into two doses. t.i.d = ter in die in Latin, this means that a medication is taken three times a day, dividing the total daily dose into three doses. MMSE = Mini‐Mental Health State Examination. The score range from 0 ( severe impairment) to 30 (normal).

Table 1. Description of the included studies at baseline

Table 2. Objectives of included studies

Study	Objective
B103	To assess the short term (3 months) symptomatic efficacy and tolerability of rivastigmine 4 and 6 mg/day compared with placebo in patients with AD
B104	Primary: to determine the maximum tolerated dose (MTD) of rivastigmine in patients with mild to moderate dementia of the Alzheimer type (DAT)
B104	Secondary: to determine ‐ a) whether tolerability is different when the drug is administered twice daily (b.i.d.) or three times daily (t.i.d.) ‐ b) if nausea and vomiting, associated with cholinesterase inhibition, can be controlled with antiemetics thereby increasing the MTD, and ‐ c) to assess the efficacy of rivastigmine at its MTD in comparison with that of placebo in the treatment of DAT
B303/B305	Primary 1: to evaluate the efficacy of two non‐overlapping dose ranges of rivastigmine (1 to 4mg daily and 6 to 12 mg daily) versus placebo over a 26 week treatment period as assessed by two primary measures of outcome; change from baseline in ADAS‐Cog score and the CIBIC‐Plus score at week 26
	Primary 2: to evaluate the safety of the study medication as assessed by incidence of adverse events, clinical laboratory evaluations , vital signs, ECG recordings, and the results of physical examination made at baseline and throughout the study
	Secondary: to assess dose‐efficacy and dose‐safety relationships for rivastigmine
B304	Primary: to evaluate the efficacy and safety of individual highest well‐tolerated doses (range 6 to 12 mg daily) of rivastigmine given b.i.d. or t.i.d. for 26 weeks compared with placebo in the therapy of patients with probable Alzheimer's disease
B304	Secondary: to compare the twice daily and three times daily dosing regimens with respect to efficacy and safety to evaluate changes in activities of daily living (ADL)
B351	Primary: to evaluate the efficacy and safety of three fixed doses of rivastigmine (3, 6 and 9 mg/day) and placebo for 26 weeks of treatment
	Secondary: to assess the dose‐efficacy and dose‐safety relationships for rivastigmine
	Tertiary: to explore the pharmacokinetics of rivastigmine at doses of 3, 6 and 9 mg daily
B352	Primary: to evaluate the efficacy and safety of two non‐overlapping dose ranges of rivastigmine (1 to 4 mg daily and 6 to 12 mg daily) and placebo for 26 weeks of treatment
	Secondary: to assess the dose‐efficacy and dose‐safety relationships of rivastigmine. To investigate the relationship between plasma concentrations of rivastigmine and efficacy and safety
	Tertiary: to explore the pharmacokinetics of rivastigmine at doses of 1 to 4 and 6 to 12 mg daily
IDEAL	To compare the efficacy,safety and tolerability of a novel rivastigmine transdermal patch with conventional rivastigmine capsules and placebo in patients with AD
Karaman 2005	To evaluate the efficacy of rivastigmine for a period of 12 months in patients with advanced moderate AD
Lopez‐Pousa 2005	To evaluate the safety and efficacy of rivastigmine in patients with more advanced AD
Mowla 2007	To assess the effect of serotonin augmentation on cognition and ADL of patients with AD
Ballard 2005	To determine whether rivastigmine was better than placebo for agitation and cognition
Tai 2000	To evaluate the safety and efficacy of Exelon compared with placebo in patients with probable Alzheimer's disease who had dementia ranging from mild to moderate degree
Nakamura 2011	To evaluate the efficacy, safety, and tolerability of the 5 cm² (9 mg loading dose, 4.6 mg/24 h delivery rate) and 10 cm² (18 mg loading dose, 9.5 mg/day delivery rate) rivastigmine patch in Japanese patients with AD

Table 2. Objectives of included studies

Table 3. Mean daily dose (mg/day) of rivastigmine achieved in the studies at different time points

Time (weeks)	treatment group	B103	B104	B303/B305	B304	B351	B352	IDEAL	Karaman 2005	Lopez‐Pousa 2005	Nakamura 2011
10 to 12	low b.i.d.	4	‐	3.8	‐	2.9	3.6
	medium b.i.d.	6	‐	‐	‐	5.7	‐
	high b.i.d.	‐	9.6	10.4	9.5	8.8	10.1		6.1
	high t.i.d.	‐	10.2	‐	9.7	‐	‐
26	low b.i.d.	‐	‐	3.7	‐	2.8	3.5
	medium b.i.d.	‐	‐	‐	‐	5.7	‐
	high b.i.d.	‐	‐	10.4	9.3	8.5	9.7	9.7	8.3	9.8
	high t.i.d.	‐	‐	‐	9.6	‐	‐
	low patch										4.6
	medium patch							9.5			9.5
	high patch							16.5
48	medium patch
	high patch
52	high b.i.d.								10.7
Exact doses not available for B103, Ballard 2005, Tai 2000, Mowla 2007.

Table 3. Mean daily dose (mg/day) of rivastigmine achieved in the studies at different time points

Table 4. Measured outcomes

Outcomes assessed	Cognitive function			Activities of daily lIving		Behavioural symptoms	Physician rated global impression of change		Other domains
Study	ADAS‐Cog	MMSE	Others	PDS	Others	Behavioural symptoms	CIBIC‐Plus	Others	Other domains
B103		X	OE, TMT, NOSGER, DSST, VRT					CGIC
B104	X		Wechsler psychometric tests, NOSGER				X
B303/B305	X	X	ADAS‐CogA	X	CAS		X	GDS
B304	X	X	ADAS‐CogA	X	CAS		X	GDS
B351	X	X	ADAS‐CogA	X	CAS		X	GDS
B352	X	X	ADAS‐CogA	X	CAS		X	GDS
Ballard 2005			SIB			CMAI
Karaman 2005	X	X		X	ACDS‐ADL, DAD		X	GDS
IDEAL	X	X	CLOCK DRAWING, TMT		ACDS‐ADL	NPI‐12		ADCS‐CGIC
Lopez‐Pousa 2005		X	SIB, BLESSED DEMENTIA SCALE		ACDS‐ADL	NPI‐10, NPI‐4		GDS ADCS‐CGIC
Mowla 2007			WMS‐III,		ADL			CGI	Hamilton score
Tai 2000		X	NPT				X	GDS
Nakamura 2011	X	X	MENFIS		DAD	BEHAVE‐AD	X
x indicated that the study measured this outcome. The full names of these scales and their properties are described in Types of outcome measures.

Table 4. Measured outcomes

Table 5. Comparison of different methods of dealing with missing values

Time point	population	rivastigmine n	placebo n	result	probability level	95% confidence limits
1 to 4 mg daily versus placebo, ADAS‐Cog measured as change from baseline
12 weeks	ITT	650	643	favours rivastigmine WMD ‐0.31	0.30	‐0.87, 0.25
	OC	589	598	favours rivastigmine WMD ‐0.46	0.14	‐1.08, 0.15
	RDO + OC	616	615	favours rivastigmine WMD ‐0.37	0.20	‐0.96, 0.23
18 weeks	ITT	650	643	favours rivastigmine WMD ‐1.07	0.0004	‐1.66, ‐0.48
	OC	558	552	favours rivastigmine WMD ‐1.19	0.0005	‐1.86, ‐0.52
	RDO + OC	573	572	favours rivastigmine WMD ‐1.33	0.00008	‐1.99, ‐0.67
26 weeks	ITT	650	644	favours rivastigmine WMD ‐0.84	0.01	‐1.48, ‐0.19
	OC	519	526	favours rivastigmine WMD ‐0.96	0.01	‐1.72, ‐0.21
	RDO + OC	559	564	favours rivastigmine WMD ‐1.07	0.004	‐1.80, ‐0.34
6 to 12 mg daily versus placebo, ADAS‐Cog measured as change from baseline
12 weeks	ITT	1054	863	favours rivastigmine WMD ‐1.49	<0.00001	‐1.96, ‐1.01
	OC	843	803	favours rivastigmine WMD ‐1.80	<0.00001	‐2.33, ‐1.27
	RDO + OC	967	828	favours rivastigmine WMD ‐1.38	<0.00001	‐1.89, ‐0.88
6 to 12 mg daily versus placebo, ADAS‐Cog measured as change from baseline
18 weeks	ITT	1054	863	favours rivastigmine WMD ‐1.79	<0.00001	‐2.30,‐ 1.29
	OC	732	742	favours rivastigmine WMD ‐2.36	<0.00001	‐2.96, ‐1.76
	RDO + OC	837	772	favours rivastigmine WMD ‐2.12	<0.00001	‐2.69, ‐1.55
26 weeks	ITT	1054	863	favours rivastigmine WMD ‐2.09	<0.00001	‐2.65, ‐1.54
	OC	670	709	favours rivastigmine WMD ‐2.62	<0.00001	‐3.29, ‐1.94
	RDO + OC	788	759	favours rivastigmine WMD ‐2.39	<0.00001	‐3.03, ‐1.74
1 to 4 mg daily versus placebo, CIBIC‐Plus measured as no change or worse
12 weeks	ITT	608	612	favours rivastigmine Peto OR 0.93	0.60	0.72, 1.21
	OC	583	596	favours rivastigmine Peto OR 0.95	0.70	0.72, 1.23
	RDO + OC	609	612	favours rivastigmine Peto OR 0.94	0.60	0.72, 1.22
18 weeks	ITT	614	620	favours rivastigmine Peto OR 0.98	0.90	0.75, 1.26
	OC	556	554	favours placebo Peto OR 1.04	0.80	0.80, 1.37
	RDO + OC	570	576	favours placebo Peto OR 1.02	0.90	0.78, 1.34
26 weeks	ITT	614	623	favours rivastigmine Peto OR 0.71	0.01	0.55, 0.93
	OC	513	523	favours rivastigmine Peto OR 0.67	0.006	0.50, 0.89
	RDO + OC	544	549	favours rivastigmine Peto OR 0.68	0.008	0.52, 0.91
1 to 4 mg daily versus placebo, CIBIC‐Plus measured as no change or worse
12 weeks	ITT	950	825	favours rivastigmine Peto OR 0.74	0.008	0.60, 0.92
	OC	831	799	favours rivastigmine Peto OR 0.72	0.005	0.58, 0.91
	RDO + OC	952	825	favours rivastigmine Peto OR 0.75	0.01	0.60, 0.93
18 weeks	ITT	970	835	favours rivastigmine Peto OR 0.81	0.06	0.65, 1.01
	OC	720	741	favours rivastigmine Peto OR 0.72	0.005	0.57, 0.91
	RDO + OC	820	772	favours rivastigmine Peto OR 0.77	0.02	0.62, 0.97
26 weeks	ITT	973	839	favours rivastigmine Peto OR 0.68	0.0007	0.55, 0.85
	OC	660	693	favours rivastigmine Peto OR 0.63	0.0004	0.49, 0.81
	RDO + OC	784	758	favours rivastigmine Peto OR 0.65	0.0003	0.51, 0.82
The results for two outcomes, ADAS‐Cog and CBIC at 12, 18 and 26 weeks, have been pooled for 3 studies, B303/B305, B351. B352. These studies reported results for 3 populations, intention‐to‐treat (ITT), completers (OC), and completers + retrieved dropout (RDO + OC). The table reports the results of the meta‐analyses for 2 comparisons (1 to 4 mg daily versus placebo and 6 to 12 mg/day versus placebo) for the 3 populations at the 3 time points.

Table 5. Comparison of different methods of dealing with missing values

Comparison 1. Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm2 (9.5 mg/day) patch) versus placebo

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 ADAS‐Cog (change from baseline at 24‐26 weeks) ITT Show forest plot	6	3232	Mean Difference (IV, Fixed, 95% CI)	‐1.79 [‐2.21, ‐1.37]

2 MMSE (change from baseline at 24‐26 weeks) ITT Show forest plot	6	3205	Mean Difference (IV, Fixed, 95% CI)	0.74 [0.52, 0.97]

3 Activities of daily living (change from baseline at 24‐26 weeks ) ITT Show forest plot	6	3230	Std. Mean Difference (IV, Fixed, 95% CI)	0.20 [0.13, 0.27]

4 Clinical Global Impression (no change or worse at 24‐26 weeks ) ITT Show forest plot	7	3338	Odds Ratio (M‐H, Fixed, 95% CI)	0.68 [0.58, 0.80]

5 Behavioural symptoms (change from baseline at 24‐26 weeks) ITT Show forest plot	3	1529	Std. Mean Difference (IV, Fixed, 95% CI)	‐0.04 [‐0.14, 0.06]

6 Withdrawals before end of treatment at 24‐26 weeks Show forest plot	7	3569	Odds Ratio (M‐H, Fixed, 95% CI)	2.06 [1.74, 2.45]

7 at least one adverse event by 24‐26 weeks Show forest plot	7	3587	Odds Ratio (M‐H, Fixed, 95% CI)	2.16 [1.82, 2.57]

8 NPI‐D carer distress scale (change from baseline at 24‐26 weeks) ITT Show forest plot	1	529	Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.91, 1.11]

Comparison 1. Rivastigmine (capsules 6 to 12 mg/day in two divided doses or 10 cm2 (9.5 mg/day) patch) versus placebo

Comparison 2. Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 ADAS‐Cog (change from baseline at 12 weeks) ITT Show forest plot	4		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

1.1 rivastigmine (1‐4 mg/d) vs placebo	3	1293	Mean Difference (IV, Fixed, 95% CI)	‐0.31 [‐0.87, 0.25]
1.2 rivastigmine (6‐12 mg/d) vs placebo	4	1917	Mean Difference (IV, Fixed, 95% CI)	‐1.49 [‐1.96, ‐1.01]
2 ADAS‐Cog (change from baseline at 26 weeks) ITT Show forest plot	5		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

2.1 rivastigmine (1‐4 mg/d) vs placebo	3	1293	Mean Difference (IV, Fixed, 95% CI)	‐0.84 [‐1.48, ‐0.19]
2.2 rivastigmine (6‐12 mg/d) vs placebo	5	2451	Mean Difference (IV, Fixed, 95% CI)	‐1.99 [‐2.49, ‐1.50]
3 MMSE (change from baseline at 26 weeks) ITT Show forest plot	5		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

3.1 rivastigmine (1‐4 mg/d) vs placebo	3	1297	Mean Difference (IV, Fixed, 95% CI)	0.43 [0.08, 0.78]
3.2 rivastigmine (6‐12 mg/d) vs placebo	5	2458	Mean Difference (IV, Fixed, 95% CI)	0.82 [0.56, 1.08]
4 SIB (change from baseline at 26 weeks) Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

4.1 Rivastigmine 6‐12 mg/day	1	210	Mean Difference (IV, Fixed, 95% CI)	4.53 [0.47, 8.59]
5 ADCS‐ADL (change from baseline at 26 weeks) ITT Show forest plot	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

5.1 rivastigmine (6‐12 mg/d) vs placebo	1	535	Mean Difference (IV, Fixed, 95% CI)	1.80 [0.20, 3.40]
6 PDS (change from baseline at 12 weeks ) ITT Show forest plot	4		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

6.1 rivastigmine (1‐4 mg/d) vs placebo	3	1288	Mean Difference (IV, Fixed, 95% CI)	‐0.77 [‐1.84, 0.30]
6.2 rivastigmine (6‐12 mg/d) vs placebo	4	1912	Mean Difference (IV, Fixed, 95% CI)	1.08 [0.19, 1.98]
7 PDS (change from baseline at 26 weeks ) ITT Show forest plot	4		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

7.1 rivastigmine (1‐4 mg/d) vs placebo	3	1288	Mean Difference (IV, Fixed, 95% CI)	‐0.38 [‐1.61, 0.84]
7.2 rivastigmine (6‐12 mg/d) vs placebo	4	1912	Mean Difference (IV, Fixed, 95% CI)	2.15 [1.13, 3.16]
8 Clinical Global Impression (no change or worse at 12 weeks) ITT Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

8.1 rivastigmine (1‐4 mg/d) vs placebo	3	1220	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.93 [0.72, 1.21]
8.2 rivastigmine (6‐12 mg/d) vs placebo	4	1775	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.74 [0.60, 0.92]
9 Clinical Global Impression (no change or worse at 26 weeks ) ITT Show forest plot	6		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

9.1 rivastigmine (1‐4 mg/d) vs placebo	3	1237	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.71 [0.55, 0.93]
9.2 rivastigmine (6‐12 mg/d) vs placebo	6	2553	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.66 [0.55, 0.79]
10 GDS( moderately severe, severe, or very severe dementia at 26 weeks) ITT Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

10.1 rivastigmine (1‐4 mg/d) vs placebo	3	1296	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.90 [0.71, 1.14]
10.2 rivastigmine (6‐12 mg/d) vs placebo	4	1923	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.78 [0.64, 0.94]
11 CGIC (little or no improvement, or worse at 12 weeks) ITT Show forest plot	1		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

11.1 rivastigmine (1‐4 mg/d) vs placebo	1	269	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.03 [0.60, 1.77]
11.2 rivastigmine (6‐12 mg/d) vs placebo	1	266	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.72 [0.43, 1.22]
12 Behavioural disturbance NPI‐10 or NPI‐12 (change from baseline at 26 weeks) ITT Show forest plot	2		Std. Mean Difference (IV, Fixed, 95% CI)	Subtotals only

12.1 Rivastigmine (6‐12 mg/day) vs placebo	2	744	Std. Mean Difference (IV, Fixed, 95% CI)	‐0.06 [‐0.20, 0.09]
13 withdrawals before end of treatment at 12 weeks Show forest plot	1		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

13.1 rivastigmine (1‐4 mg/d) vs placebo	1	269	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.15 [0.95, 4.89]
13.2 rivastigmine (6‐12 mg/d) vs placebo	1	266	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.60 [1.19, 5.67]
14 withdrawals before end of treatment at 26 weeks Show forest plot	6		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

14.1 rivastigmine (1‐4 mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.01 [0.75, 1.34]
14.2 rivastigmine (6‐12 mg/d) vs placebo	6	2701	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.19 [1.83, 2.63]
15 at least one adverse event by the end of titration period Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

15.1 rivastigmine (1‐4 mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.04 [0.82, 1.31]
15.2 rivastigmine (6‐12 mg/d) vs placebo	4	1920	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.96 [2.39, 3.68]
16 at least one adverse event by 26 weeks Show forest plot	6		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

16.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.93 [0.71, 1.23]
16.2 rivastigmine (6‐12mg/d) vs placebo	6	2726	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.49 [2.05, 3.02]
17 dropouts due to adverse events by 12 weeks Show forest plot	1		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

17.1 rivastigmine (4mg/d) vs placebo	1	269	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.70 [1.06, 6.84]
17.2 rivastigmine (6mg/d) vs placebo	1	266	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.11 [1.28, 7.56]
18 dropouts due to adverse events by 26 weeks Show forest plot	6		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

18.1 rivastigmine (1‐4 mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.03 [0.69, 1.52]
18.2 rivastigmine (6‐12 mg/d) vs placebo	6	2729	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.73 [2.19, 3.41]
19 at least one adverse event of decreased appetite by 26 weeks Show forest plot	1		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

19.1 rivastigmine (6‐12 mg/d) vs placebo	1	596	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.51 [1.26, 9.79]
20 at least one adverse event of weight decrease by 26 weeks Show forest plot	1		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

20.1 rivastigmine (6‐12mg/d) vs placebo	1	596	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.55 [1.46, 8.66]
21 at least one adverse event of nausea by the end of titration period Show forest plot	5		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

21.1 rivastigmine (1‐4mg/d) vs placebo	4	1559	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.85 [1.36, 2.52]
21.2 rivastigmine (6‐12 mg/d) vs placebo	5	2186	Peto Odds Ratio (Peto, Fixed, 95% CI)	5.57 [4.59, 6.75]
22 at least one adverse event of nausea by 26 weeks Show forest plot	6		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

22.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.74 [1.28, 2.36]
22.2 rivastigmine (6‐12mg/d bid) vs placebo	6	2726	Peto Odds Ratio (Peto, Fixed, 95% CI)	5.36 [4.50, 6.40]
23 at least one adverse event of vomiting by the end of titration period Show forest plot	5		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

23.1 rivastigmine (1‐4mg/d) vs placebo	4	1559	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.97 [1.22, 3.16]
23.2 rivastigmine (6‐12 mg/d) vs placebo	5	2187	Peto Odds Ratio (Peto, Fixed, 95% CI)	5.72 [4.48, 7.29]
24 at least one adverse event of vomiting by 26 weeks Show forest plot	6		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

24.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.65 [1.08, 2.52]
24.2 rivastigmine (6‐12mg/d) vs placebo	6	2726	Peto Odds Ratio (Peto, Fixed, 95% CI)	5.15 [4.20, 6.32]
25 at least one adverse event of diarrhoea by the end of titration period Show forest plot	5		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

25.1 rivastigmine (1‐4mg/d) vs placebo	4	1559	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.99 [0.68, 1.42]
25.2 rivastigmine (6‐12 mg/d) vs placebo	5	2186	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.97 [1.51, 2.57]
26 at least one adverse event of diarrhoea by 26 weeks Show forest plot	5		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

26.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.93 [0.67, 1.31]
26.2 rivastigmine (6‐12mg/d) vs placebo	5	2516	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.76 [1.39, 2.24]
27 at least one adverse event of anorexia by the end of titration period Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

27.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.21 [1.24, 3.95]
27.2 rivastigmine (6‐12 mg/d) vs placebo	4	1920	Peto Odds Ratio (Peto, Fixed, 95% CI)	4.94 [3.56, 6.85]
28 at least one adverse event of anorexia by 26 weeks Show forest plot	5		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

28.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.13 [1.29, 3.52]
28.2 rivastigmine (6‐12mg/d) vs placebo	5	2130	Peto Odds Ratio (Peto, Fixed, 95% CI)	4.46 [3.34, 5.95]
29 at least one adverse event of headache by the end of titration period Show forest plot	5		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

29.1 rivastigmine (1‐4mg/d) vs placebo	4	1559	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.97 [0.69, 1.37]
29.2 rivastigmine (6‐12 mg/d) vs placebo	5	2186	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.64 [1.26, 2.14]
30 at least one adverse event of headache by 26 weeks Show forest plot	5		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

30.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.17 [0.84, 1.64]
30.2 rivastigmine (6‐12mg/d) vs placebo	5	2516	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.72 [1.34, 2.21]
31 at least one adverse event of insomnia by the end of titration period Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

31.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.04 [0.64, 1.67]
31.2 rivastigmine (6‐12 mg/d) vs placebo	4	1920	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.40 [0.94, 2.09]
32 at least one adverse event of insomnia by 26 weeks Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

32.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.05 [0.70, 1.58]
32.2 rivastigmine (6‐12mg/d) vs placebo	4	1920	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.33 [0.95, 1.87]
33 at least one adverse event of syncope by the end of titration period Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

33.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.50 [0.43, 5.20]
33.2 rivastigmine (6‐12 mg/d) vs placebo	4	1920	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.16 [0.99, 4.68]
34 at least one adverse event of syncope by 26 weeks Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

34.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.00 [0.37, 2.69]
34.2 rivastigmine (6‐12mg/d bid) vs placebo	4	1920	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.72 [0.96, 3.11]
35 at least one adverse event of abdominal pain by the end of titration period Show forest plot	5		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

35.1 rivastigmine (1‐4mg/d) vs placebo	4	1559	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.16 [0.72, 1.88]
35.2 rivastigmine (6‐12mg/d) vs placebo	5	2186	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.50 [1.80, 3.48]
36 at least one adverse event of abdominal pain by 26 weeks Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

36.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.20 [0.77, 1.87]
36.2 rivastigmine (6‐12mg/d) vs placebo	4	1920	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.24 [1.65, 3.05]
37 at least one adverse event of dizziness by the end of titration period Show forest plot	5		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

37.1 rivastigmine (1‐4mg/d) vs placebo	4	1559	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.99 [0.70, 1.39]
37.2 rivastigmine (6‐12 mg/d) vs placebo	5	2186	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.38 [1.86, 3.04]
38 at least one adverse event of dizziness by 26 weeks Show forest plot	5		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

38.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.25 [0.91, 1.72]
38.2 rivastigmine (6‐12mg/d) vs placebo	5	2516	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.24 [1.78, 2.82]
39 at least one adverse event of bone fracture by the end of titration period Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

39.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.83 [0.25, 2.72]
39.2 rivastigmine (6‐12 mg/d) vs placebo	4	1920	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.95 [0.37, 2.46]
40 at least one adverse event of bone fracture by 26 weeks Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

40.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.60 [0.27, 1.34]
40.2 rivastigmine (6‐12mg/d) vs placebo	4	1920	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.70 [0.34, 1.42]
41 at least one adverse event of asthenia by 26 weeks Show forest plot	1		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

41.1 rivastigmine (6‐12mg/d) vs placebo	1	596	Peto Odds Ratio (Peto, Fixed, 95% CI)	4.37 [1.79, 10.65]
42 at least one severe adverse event by the end of titration period Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

42.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.94 [0.62, 1.42]
42.2 rivastigmine (6‐12 mg/d) vs placebo	4	1920	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.88 [1.39, 2.55]
43 at least one serious adverse event by 26 weeks Show forest plot	6		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

43.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.98 [0.70, 1.36]
43.2 rivastigmine (6‐12mg/d) vs placebo	6	2726	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.17 [0.93, 1.47]
44 deaths before end of treatment at 12 weeks Show forest plot	1		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

44.1 rivastigmine (1‐4mg/d) vs placebo	1	269	Peto Odds Ratio (Peto, Fixed, 95% CI)	7.34 [0.76, 71.14]
44.2 rivastigmine (6‐12 mg/d) vs placebo	1	266	Peto Odds Ratio (Peto, Fixed, 95% CI)	7.45 [0.46, 119.66]
45 deaths before end of treatment at 26 weeks Show forest plot	6		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

45.1 rivastigmine (1‐4mg/d) vs placebo	3	1290	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.98 [0.20, 19.15]
45.2 rivastigmine (6‐12mg/d) vs placebo	6	2737	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.16 [0.40, 3.37]
46 CIBIC‐Plus (no change or worse at 12 weeks) OC Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

46.1 rivastigmine (1‐4mg/d) vs placebo	3	1179	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.95 [0.72, 1.23]
46.2 rivastigmine (6‐12mg/d) vs placebo	4	1630	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.72 [0.58, 0.91]
47 CIBIC‐Plus (no change or worse at 26 weeks) OC Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

47.1 rivastigmine (1‐4mg/d) vs placebo	3	1036	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.67 [0.50, 0.89]
47.2 rivastigmine (6‐12mg/d) vs placebo	4	1353	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.63 [0.49, 0.81]
48 CIBIC‐Plus (no change or worse at 12 weeks) OC+RDO Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

48.1 rivastigmine (1‐4mg/d) vs placebo	3	1221	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.94 [0.72, 1.22]
48.2 rivastigmine (6‐12mg/d) vs placebo	4	1777	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.75 [0.60, 0.93]
49 CIBIC‐Plus (no change or worse at 26 weeks)OC+RDO Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

49.1 rivastigmine (1‐4mg/d) vs placebo	3	1093	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.68 [0.52, 0.91]
49.2 rivastigmine (6‐12mg/d) vs placebo	4	1542	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.65 [0.51, 0.82]
50 CIBIC‐Plus (no change or worse at 12 weeks) ALL+OC Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

50.1 rivastigmine (1‐4mg/d) vs placebo	3	1293	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.00 [0.77, 1.30]
50.2 rivastigmine (6‐12mg/d) vs placebo	4	1917	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.90 [0.72, 1.13]
51 CIBIC‐Plus (no change or worse at 26 weeks) ALL+OC Show forest plot	4		Peto Odds Ratio (Peto, Fixed, 95% CI)	Subtotals only

51.1 rivastigmine (1‐4mg/d) vs placebo	3	1297	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.73 [0.55, 0.96]
51.2 rivastigmine (6‐12mg/d) vs placebo	4	1921	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.88 [0.69, 1.12]
52 ADAS‐Cog (change from baseline at 12 weeks) OC Show forest plot	4		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

52.1 rivastigmine (1‐4mg/d) vs placebo	3	1187	Mean Difference (IV, Fixed, 95% CI)	‐0.46 [‐1.08, 0.15]
52.2 rivastigmine (6‐12mg/d) vs placebo	4	1646	Mean Difference (IV, Fixed, 95% CI)	‐1.80 [‐2.33, ‐1.27]
53 ADAS‐Cog (change from baseline at 26 weeks) OC Show forest plot	4		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

53.1 rivastigmine (1‐4mg/d) vs placebo	3	1045	Mean Difference (IV, Fixed, 95% CI)	‐0.97 [‐1.72, ‐0.21]
53.2 rivastigmine (6‐12mg/d) vs placebo	4	1379	Mean Difference (IV, Fixed, 95% CI)	‐2.62 [‐3.29, ‐1.94]
54 ADAS‐Cog (change from baseline at 12 weeks) OC+RDO Show forest plot	4		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

54.1 rivastigmine (1‐4mg/d) vs placebo	3	1231	Mean Difference (IV, Fixed, 95% CI)	‐0.37 [‐0.96, 0.23]
54.2 rivastigmine (6‐12mg/d) vs placebo	4	1795	Mean Difference (IV, Fixed, 95% CI)	‐1.38 [‐1.89, ‐0.88]
55 ADAS‐Cog (change from baseline at 26 weeks) OC+RDO Show forest plot	4		Mean Difference (IV, Fixed, 95% CI)	Subtotals only

55.1 rivastigmine (1‐4mg/d) vs placebo	3	1123	Mean Difference (IV, Fixed, 95% CI)	‐1.07 [‐1.80, ‐0.34]
55.2 rivastigmine (6‐12mg/d) vs placebo	4	1547	Mean Difference (IV, Fixed, 95% CI)	‐2.39 [‐3.03, ‐1.74]

Comparison 2. Rivastigmine oral capsules (1 to 4 mg/day or 6 to 12 mg/day in two divided doses) versus placebo

Comparison 3. Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 ADAS‐Cog (change from baseline at 24 weeks) ITT Show forest plot	1	543	Mean Difference (IV, Fixed, 95% CI)	‐2.6 [‐3.72, ‐1.48]

2 TMT‐A (change from baseline at 24 weeks) ITT Show forest plot	1	496	Mean Difference (IV, Fixed, 95% CI)	‐14.2 [‐24.11, ‐4.29]

3 clock drawing (change from baseline at 24 weeks) ITT Show forest plot	1	520	Mean Difference (IV, Fixed, 95% CI)	0.2 [‐0.34, 0.74]

4 MMSE (change from baseline at 24 weeks) ITT Show forest plot	1	543	Mean Difference (IV, Fixed, 95% CI)	0.90 [0.32, 1.48]

5 ADCS‐ADL (change from baseline at 24 weeks) ITT Show forest plot	1	544	Mean Difference (IV, Fixed, 95% CI)	2.3 [0.52, 4.08]

6 NPI‐12 (change from baseline at 24 weeks) ITT Show forest plot	1	544	Mean Difference (IV, Fixed, 95% CI)	‐0.60 [‐2.88, 1.68]

7 withdrawals before end of treatment at 24 weeks Show forest plot	1	605	Odds Ratio (M‐H, Fixed, 95% CI)	1.90 [1.22, 2.97]

8 at least one adverse event by 24 weeks Show forest plot	1	605	Odds Ratio (M‐H, Fixed, 95% CI)	2.28 [1.64, 3.16]

9 withdrawals due to an adverse event before end of treatment at 24 weeks Show forest plot	1	605	Odds Ratio (M‐H, Fixed, 95% CI)	1.80 [0.93, 3.46]

10 at least one adverse event of dizziness by 24 weeks Show forest plot	1	605	Odds Ratio (M‐H, Fixed, 95% CI)	3.14 [1.31, 7.50]

11 at least one adverse event of nausea by 24 weeks Show forest plot	1	605	Odds Ratio (M‐H, Fixed, 95% CI)	5.12 [2.85, 9.22]

12 at least one adverse event of vomiting by 24 weeks Show forest plot	1	605	Odds Ratio (M‐H, Fixed, 95% CI)	6.77 [3.38, 13.53]

13 at least one adverse event of weight decrease by 24 weeks Show forest plot	1	605	Odds Ratio (M‐H, Fixed, 95% CI)	6.12 [2.09, 17.92]

14 at least one adverse event of decreased appetite by 24 weeks Show forest plot	1	605	Odds Ratio (M‐H, Fixed, 95% CI)	5.19 [1.49, 18.12]

15 at least one adverse event of headache by 24 weeks Show forest plot	1	605	Odds Ratio (M‐H, Fixed, 95% CI)	2.66 [0.94, 7.56]

16 at least one adverse event of asthenia by 24 weeks Show forest plot	1	605	Odds Ratio (M‐H, Fixed, 95% CI)	3.05 [0.82, 11.38]

17 deaths before end of treatment at 24 weeks Show forest plot	1	605	Odds Ratio (M‐H, Fixed, 95% CI)	1.67 [0.40, 7.06]

18 NPI‐D carer distress scale (change from baseline at 24 weeks) ITT Show forest plot	1	544	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐1.07, 1.07]

Comparison 3. Rivastigmine 20 cm2 patch (17.4 mg/day) versus placebo

Comparison 4. Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 ADAS‐Cog (change from baseline at 24 weeks) ITT Show forest plot	2	1062	Mean Difference (IV, Fixed, 95% CI)	‐1.35 [‐2.03, ‐0.66]

2 MMSE (change from baseline at 24 weeks) ITT Show forest plot	2	1028	Mean Difference (IV, Fixed, 95% CI)	0.64 [0.26, 1.02]

3 clock drawing (change from baseline at 24 weeks) ITT Show forest plot	1	514	Mean Difference (IV, Fixed, 95% CI)	0.4 [‐0.17, 0.97]

4 TMT‐A (change from baseline at 24 weeks) ITT Show forest plot	1	499	Mean Difference (IV, Fixed, 95% CI)	‐20.0 [‐29.80, ‐10.20]

5 Mental Function Impairment MENFIS (change from baseline at 24 weeks) ITT Show forest plot	1	537	Mean Difference (IV, Fixed, 95% CI)	‐1.30 [‐2.32, ‐0.28]

6 ADCS‐ADL (change from baseline at 24 weeks) ITT Show forest plot	1	528	Mean Difference (IV, Fixed, 95% CI)	2.20 [0.62, 3.78]

7 Disability Assessment for Dementia (DAD) (change from baseline at 24 weeks) ITT Show forest plot	1	536	Mean Difference (IV, Fixed, 95% CI)	2.30 [0.34, 4.26]

8 BEHAVE‐AD (change from baseline at 24 weeks) ITT Show forest plot	1	537	Mean Difference (IV, Fixed, 95% CI)	‐0.20 [‐0.92, 0.52]

9 NPI‐12 (change from baseline at 24 weeks) ITT Show forest plot	1	529	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐2.16, 2.16]

10 Clinical Global Impression (no change or worse at 24 weeks) Show forest plot	2	1063	Odds Ratio (M‐H, Fixed, 95% CI)	0.77 [0.58, 1.02]

11 withdrawals before end of treatment at 24 weeks Show forest plot	2	1170	Odds Ratio (M‐H, Fixed, 95% CI)	1.67 [1.23, 2.26]

12 at least one adverse event by 24 weeks Show forest plot	2	1460	Odds Ratio (M‐H, Fixed, 95% CI)	1.63 [1.29, 2.06]

13 withdrawals due to an adverse event before end of treatment at 24 weeks Show forest plot	2	1170	Odds Ratio (M‐H, Fixed, 95% CI)	1.84 [1.20, 2.82]

14 at least one adverse event of application site erythema by 24 weeks Show forest plot	1	573	Odds Ratio (M‐H, Fixed, 95% CI)	2.73 [1.87, 3.98]

15 at least one adverse event of application site pruritis by 24 weeks Show forest plot	1	573	Odds Ratio (M‐H, Fixed, 95% CI)	1.97 [1.36, 2.86]

16 at least one adverse event of application site edema by 24 weeks Show forest plot	1	573	Odds Ratio (M‐H, Fixed, 95% CI)	4.83 [2.09, 11.15]

17 at least one adverse event application site exfoliation by 24 weeks Show forest plot	1	573	Odds Ratio (M‐H, Fixed, 95% CI)	2.81 [0.88, 8.93]

18 at least one adverse event of dermatitis contact by 24 weeks Show forest plot	1	573	Odds Ratio (M‐H, Fixed, 95% CI)	1.91 [1.24, 2.94]

19 at least one adverse event of nasopharyngitis by 24 weeks Show forest plot	1	573	Odds Ratio (M‐H, Fixed, 95% CI)	1.03 [0.62, 1.73]

20 at least one adverse event of nausea by 24 weeks Show forest plot	2	1166	Odds Ratio (M‐H, Fixed, 95% CI)	1.80 [1.07, 3.02]

21 at least one adverse event of vomiting by 24 weeks Show forest plot	2	1166	Odds Ratio (M‐H, Fixed, 95% CI)	2.06 [1.20, 3.53]

22 at least one adverse event of diarrhoea by 24 weeks Show forest plot	1	593	Odds Ratio (M‐H, Fixed, 95% CI)	1.93 [0.87, 4.24]

23 at least one adverse event of weight decrease by 24 weeks Show forest plot	1	593	Odds Ratio (M‐H, Fixed, 95% CI)	2.11 [0.63, 7.07]

24 at least one adverse event of dizziness by 24 weeks Show forest plot	1	593	Odds Ratio (M‐H, Fixed, 95% CI)	1.04 [0.36, 3.00]

25 at least one adverse event of decreased appetite by 24 weeks Show forest plot	1	593	Odds Ratio (M‐H, Fixed, 95% CI)	0.69 [0.11, 4.16]

26 at least one adverse event of headache by 24 weeks Show forest plot	1	593	Odds Ratio (M‐H, Fixed, 95% CI)	2.11 [0.71, 6.26]

27 at least one adverse event of asthenia by 24 weeks Show forest plot	1	593	Odds Ratio (M‐H, Fixed, 95% CI)	1.74 [0.41, 7.36]

28 deaths before end of treatment at 24 weeks Show forest plot	2	1170	Odds Ratio (M‐H, Fixed, 95% CI)	1.02 [0.28, 3.81]

29 NPI‐D carer distress scale (change from baseline at 24 weeks) ITT Show forest plot	1	529	Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.91, 1.11]

Comparison 4. Rivastigmine 10 cm2 patch (9.5 mg/day) versus placebo

Comparison 5. Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 ADAS‐J Cog (change from baseline at 24 weeks) ITT Show forest plot	1	531	Mean Difference (IV, Fixed, 95% CI)	‐0.8 [‐1.62, 0.02]

2 MMSE (change from baseline at 24 weeks) ITT Show forest plot	1	487	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐0.52, 0.52]

3 Mental Function Impairment MENFIS (change from baseline at 24 weeks) ITT Show forest plot	1	536	Mean Difference (IV, Fixed, 95% CI)	‐0.70 [‐1.72, 0.32]

4 Disability Assessment for Dementia (DAD) (change from baseline at 24 weeks) ITT Show forest plot	1	536	Mean Difference (IV, Fixed, 95% CI)	1.20 [‐0.73, 3.13]

5 CIBIC‐Plus J (no change or worse at 24 weeks) ITT Show forest plot	1	536	Odds Ratio (M‐H, Fixed, 95% CI)	0.67 [0.43, 1.05]

6 BEHAVE‐AD (change from baseline at 24 weeks) ITT Show forest plot	1	536	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐0.67, 0.67]

7 withdrawals before end of treatment at 24 weeks Show forest plot	1	572	Odds Ratio (M‐H, Fixed, 95% CI)	1.53 [1.01, 2.33]

8 at least one adverse event by 24 weeks Show forest plot	1	568	Odds Ratio (M‐H, Fixed, 95% CI)	1.80 [1.16, 2.78]

9 withdrawals due to an adverse event before end of treatment at 24 weeks Show forest plot	1	572	Odds Ratio (M‐H, Fixed, 95% CI)	1.96 [1.12, 3.44]

10 at least one adverse event of application site erythema by 24 weeks Show forest plot	1	568	Odds Ratio (M‐H, Fixed, 95% CI)	2.53 [1.73, 3.70]

11 at least one adverse event of application site pruritis by 24 weeks Show forest plot	1	568	Odds Ratio (M‐H, Fixed, 95% CI)	1.79 [1.23, 2.60]

12 at least one adverse event of application site edema by 24 weeks Show forest plot	1	568	Odds Ratio (M‐H, Fixed, 95% CI)	5.65 [2.46, 12.94]

13 at least one adverse event application site exfoliation by 24 weeks Show forest plot	1	568	Odds Ratio (M‐H, Fixed, 95% CI)	3.68 [1.20, 11.33]

14 at least one adverse event of dermatitis contact by 24 weeks Show forest plot	1	568	Odds Ratio (M‐H, Fixed, 95% CI)	1.99 [1.30, 3.06]

15 at least one adverse event of nasopharyngitis by 24 weeks Show forest plot	1	568	Odds Ratio (M‐H, Fixed, 95% CI)	0.67 [0.38, 1.19]

16 at least one adverse event of nausea by 24 weeks Show forest plot	1	568	Odds Ratio (M‐H, Fixed, 95% CI)	0.33 [0.09, 1.24]

17 at least one adverse event of vomiting by 24 weeks Show forest plot	1	568	Odds Ratio (M‐H, Fixed, 95% CI)	1.01 [0.43, 2.38]

18 deaths before end of treatment at 24 weeks Show forest plot	1	572	Odds Ratio (M‐H, Fixed, 95% CI)	1.01 [0.06, 16.29]

Comparison 5. Rivastigmine 5 cm2 patch (4.6 mg/day) versus placebo

Comparison 6. Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 ADAS‐Cog (change from baseline at 24 weeks) ITT Show forest plot	1	501	Mean Difference (IV, Fixed, 95% CI)	0.0 [‐1.10, 1.10]

2 MMSE (change from baseline at 24 weeks) ITT Show forest plot	1	506	Mean Difference (IV, Fixed, 95% CI)	0.30 [‐0.27, 0.87]

3 clock drawing (change from baseline at 24 weeks) ITT Show forest plot	1	491	Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.46, 0.66]

4 TMT‐A (change from baseline at 24 weeks) ITT Show forest plot	1	481	Mean Difference (IV, Fixed, 95% CI)	‐2.60 [‐13.48, 8.28]

5 ADCS‐ADL (change from baseline at 24 weeks) ITT Show forest plot	1	501	Mean Difference (IV, Fixed, 95% CI)	0.40 [‐1.23, 2.03]

6 Clinical Global Impression (no change or worse at 24 weeks) Show forest plot	1	501	Odds Ratio (M‐H, Fixed, 95% CI)	1.27 [0.88, 1.84]

7 NPI‐12 (change from baseline at 24 weeks) ITT Show forest plot	1	501	Mean Difference (IV, Fixed, 95% CI)	0.50 [‐1.55, 2.55]

8 withdrawals before end of treatment at 24 weeks Show forest plot	1	590	Odds Ratio (M‐H, Fixed, 95% CI)	1.04 [0.70, 1.54]

9 at least one adverse event by 24 weeks Show forest plot	1	585	Odds Ratio (M‐H, Fixed, 95% CI)	0.59 [0.43, 0.82]

10 withdrawals due to an adverse event before end of treatment at 24 weeks Show forest plot	1	590	Odds Ratio (M‐H, Fixed, 95% CI)	1.20 [0.68, 2.13]

11 at least one adverse event of nausea by 24 weeks Show forest plot	1	585	Odds Ratio (M‐H, Fixed, 95% CI)	0.26 [0.15, 0.43]

12 at least one adverse event of vomiting by 24 weeks Show forest plot	1	585	Odds Ratio (M‐H, Fixed, 95% CI)	0.32 [0.18, 0.57]

13 at least one adverse event of diarrhoea by 24 weeks Show forest plot	1	585	Odds Ratio (M‐H, Fixed, 95% CI)	1.15 [0.57, 2.29]

14 at least one adverse event of weight decrease by 24 weeks Show forest plot	1	585	Odds Ratio (M‐H, Fixed, 95% CI)	0.49 [0.21, 1.17]

15 at least one adverse event of dizziness by 24 weeks Show forest plot	1	585	Odds Ratio (M‐H, Fixed, 95% CI)	0.30 [0.13, 0.72]

16 at least one adverse event of decreased appetite by 24 weeks Show forest plot	1	585	Odds Ratio (M‐H, Fixed, 95% CI)	0.16 [0.04, 0.73]

17 at least one adverse event of headache by 24 weeks Show forest plot	1	585	Odds Ratio (M‐H, Fixed, 95% CI)	0.55 [0.25, 1.20]

18 at least one adverse event of asthenia by 24 weeks Show forest plot	1	585	Odds Ratio (M‐H, Fixed, 95% CI)	0.28 [0.10, 0.78]

19 deaths before end of treatment at 24 weeks + 30 days Show forest plot	1	590	Odds Ratio (M‐H, Fixed, 95% CI)	2.56 [0.49, 13.31]

20 NPI‐D carer distress scale (change from baseline at 24 weeks) ITT Show forest plot	1	501	Mean Difference (IV, Fixed, 95% CI)	0.10 [‐0.96, 1.16]

Comparison 6. Rivastigmine 10 cm2 patch (9.5 mg/day) versus rivastigmine capsules (6 to 12 mg/day in two divided doses)