Key Messages

• In older adults with mild to moderate clinical Alzheimer’s-type dementia (CATD), evidence for donepezil versus placebo showed:

  • Small statistically significant improvements for cognition (low strength of evidence [SOE]), global staging (moderate SOE), and clinical impression of change (moderate SOE).
  • Increased risk for serious adverse events (low SOE) and withdrawals due to adverse events (low SOE) at standard doses (targeting 10 mg/day).
  • Insufficient evidence for function or quality of life.
• In older adults with moderate to severe CATD, evidence for donepezil versus placebo showed:

  • Small statistically significant improvements for cognition (moderate SOE), function (low SOE), and clinical impression of change (low SOE).
  • No difference in serious adverse events (low SOE), but increased withdrawals due to adverse events (low SOE) at standard doses.
  • Insufficient evidence for quality of life and global staging.
• Efficacy compared with placebo was increased for both 10 mg/day and 5 mg/day donepezil; though 10 mg/day donepezil increased risk of serious adverse events and withdrawals due to adverse events compared with placebo, 5 mg/day donepezil and placebo did not differ for these harms.

Eligible Studies

Based on a high-quality systematic review published in 2018,¹⁹⁵ we identified 24 eligible publications reporting 17 unique trials of ≥24 weeks duration that compared donepezil with placebo, reported results for cognition, function, quality of life, global staging, clinical impression of change, or harms, and were considered low or medium risk of bias.^196–219 We found one additional eligible study with low or medium risk of bias (ROB) not included in this prior review,²²⁰ bringing to 18 the total included and analyzed in the current review.

Four trials that directly compared donepezil 5 mg/day with 10 mg/day^{204, 211} and slow-release donepezil 23 mg/day with standard-release donepezil 10 mg/day^221–223 are reported in a section of this chapter titled “Donepezil Dosage Comparisons.”

Appendix Tables E.1-E.10 provide evidence tables, summary ROB assessments, and strength of evidence for key comparisons and outcomes.

Baseline Study Characteristics

Characteristics of the 4,742 participants enrolled in the 18 analyzed donepezil versus placebo trials are shown in Table 6.1. Participants were primarily community dwelling, but two trials enrolled participants in residential care.^{213, 217} Most study participants were categorized with mild or moderate CATD (e.g., Mini-Mental State Exam [MMSE] score 10 to 26) and approximately one-quarter of participants were in studies restricted to individuals with severe CATD.^{196, 204, 205, 219} All trials lasted 24 to 26 weeks except for two of 52 to 54 weeks.^{209, 217} Most often, donepezil dosing was titrated up to 10 mg/day, but three trials compared fixed doses of 5 mg/day and 10 mg/day to placebo.^{197, 204, 211}

Table 6.1

Baseline characteristics of donepezil versus placebo trials.

Outcomes

Table 6.2 summarizes the primary efficacy and harms results between donepezil and placebo. For efficacy, in patients with mild to moderate CATD treated for 24 weeks, donepezil improved several outcomes compared with placebo. For treatment up to 10 mg/day, these included small mean improvements in cognition (weighted mean difference in MMSE change, 1.0 [95% confidence interval (CI), 0.7 to 1.4]; standardized mean difference (SMD), 0.30 [95% CI, 0.16 to 0.44]), increased likelihood that cognition improved (absolute risk difference [ARD], 19%; number needed to treat [NNT], 5.3) or did not worsen (ARD, 23%; NNT, 4.3), and increased likelihood clinical impression of change improved (ARD, 12%; NNT, 8.3) or did not worsen (ARD, 14%; NNT, 7.2). However, donepezil did not increase the likelihood of a moderate or marked improvement in clinical impression of change and evidence was insufficient to draw conclusions about function or quality of life. In a 1-year trial, donepezil was associated with less worsening in MMSE (-0.5 vs. -2.0, p<0.001),²¹⁷ while a single trial reporting on change in patient residence to a different level of independence found that through 54 weeks two participants (0.9%) each in the donepezil and placebo groups entered a nursing facility.²⁰⁹ For patients with moderate to severe CATD treated for 24 to 26 weeks, compared with placebo, donepezil resulted in small mean improvements in cognition and function, and increased likelihood that clinical impression of change improved (ARD, 10%; NNT, 10) or did not worsen (ARD, 11.2%; NNT, 8.9). However, no studies of patients with moderate to severe CATD reported data on brief multidomain cognitive batteries, quality of life or staging. For treatment with 5 mg/day, efficacy compared with placebo appeared similar in magnitude to that for 10 mg/day compared with placebo, both in patients with mild to moderate CATD and moderate to severe CATD.

For harms, donepezil up to 10 mg/day increased risk of withdrawals due to adverse events in patients with mild to moderate CATD and those with moderate to severe CATD (both low SOE), and increased serious adverse events only in patients with mild to moderate CATD (low SOE) (Table 6.2). Donepezil and placebo did not statistically differ for confusion (5.8% vs. 6.3%; RR, 0.91 [95% CI, 0.61 to 1.38], n=4 trials),^{197, 199, 213, 217} falls (7% in each arm, n=3 trials),^{204, 212, 219} restlessness (2% in each arm, p=0.44; n=2 trials),^{203, 204} tremor (8% vs. 2%, p=0.07, n=1 trial),²¹³ abnormal gait (12% vs. 8%, p=0.33, respectively; n=1 trial)PN Tariot, JL Cummings ²¹³ or mortality (1.4% vs. 2.3%; Peto odds ratio [OR], 0.68 [95% CI, 0.43 to 1.08], n=12 trials).^{196, 197, 199, 204, 205, 207, 209, 211, 213, 217, 220, 224} Donepezil 5 mg/day did not increase risk of either serious adverse events or withdrawals due to adverse events compared with placebo. No studies reported data on somnolence or stroke.

Appendix Tables E.4-E.10 provide detailed evidence tables and assessments of strength of evidence for key comparisons and outcomes.

Table 6.2

Summary of findings for primary outcomes: donepezil versus placebo.

Variation in Outcomes by Participant Characteristics

No studies reported whether the effect of donepezil versus placebo on cognition, function, quality of life, global staging, clinical impression of change, or harms varied as a function of participant characteristics. Two trials performed an exploratory analysis based on disease severity at baseline as defined by MMSE score.^{207, 220} One of these trials reported that donepezil improved ADAS-Cog scores compared with placebo in participants with baseline MMSE ≤18, but not in those with baseline MMSE >18.²⁰⁷ The second trial reported that donepezil and placebo did not differ for change from baseline for Alzheimer’s Disease Assessment Scale-Cognitive Subscale (ADAS-Cog) or Clinician’s Interview-Based Impression of Change with caregiver input (CIBIC-Plus) in either participants with baseline MMSE 20-26 or baseline MMSE 10-19.²²⁰ However, neither trial reported a test for interaction.

Key Messages

• In older adults with CATD, the evidence for donepezil 23 mg/day compared with 10 mg/day showed:

  • No difference in cognition, function, or clinical impression of change (all low SOE).
  • Insufficient evidence to draw conclusions about quality of life.
  • Increased risk of withdrawals due to adverse events (moderate SOE), but no difference for serious adverse events (low SOE).
• In older adults with mild to moderate CATD, the evidence for donepezil 10 mg/day compared with 5 mg/day showed:

  • Mixed results for cognition (low SOE).
  • No difference in clinical impression of change (low SOE).
  • Insufficient evidence for function, quality of life, or staging.
  • Increased risk of serious adverse events (low SOE) and withdrawals due to adverse events (moderate SOE).
• In older adults with moderate to severe CATD, the evidence for donepezil 10 mg/day compared with 5 mg/day showed:

  • No difference in cognition or function (both low SOE).
  • Improvement for clinical impression of change (low SOE).
  • Insufficient evidence to draw conclusions about quality of life, staging, serious adverse events, or withdrawals due to adverse events.

Eligible Studies

Based on a high-quality systematic review published in 2018,¹⁹⁵ we identified five eligible publications of five unique trials of at least 24 weeks duration that compared donepezil doses and reported results for cognition, function, quality of life, staging, clinical impression of change or harms. These included two that compared high-dose 23 mg/day donepezil to standard-dose 10 mg/day donepezil,^{221, 223} and three that compared 10 mg/day donepezil to 5 mg/day donepezil.^{197, 204, 211} We found no additional eligible studies published after the May 2017 search date of this previous systematic review. Appendix Tables E.4-E.10 provide evidence tables, summary ROB assessments, and strength of evidence for key comparisons and outcomes.

Baseline Study Characteristics

Characteristics of the 2,886 participants enrolled in the five analyzed trials are shown in Table 6.3. Most participants were community-dwelling and lived with a caregiver. All trials were 24 weeks in duration. Two trials compared sustained-release high-dose 23 mg/day donepezil with standard-dose 10 mg/day donepezil.^{221, 223} Three trials compared 10 mg/day donepezil with 5 mg/day donepezil.^{197, 204, 211}

Table 6.3

Baseline characteristics of donepezil versus donepezil trials.

Outcomes

Table 6.4 summarizes the primary efficacy and harms results from trials that directly compared different donepezil doses. For efficacy, donepezil 23 mg/day and 10 mg/day did not differ for cognition, function, or clinical impression of change (all low SOE). By comparison, donepezil 10 mg/day increased the likelihood of improvement for brief multidomain cognitive tests and clinical impression of change compared with 5 mg/day when evaluated using categorical responder analyses but not as continuous outcome measures (all low SOE). Donepezil 10 mg/day and 5 mg/day did not differ for brief cognitive tests commonly used as stand-alone tests or for function tests (all low SOE). No studies that directly compared different donepezil doses reported data on quality of life, disease staging, or change in residence to a different level of independence.

For harms, when directly compared, donepezil 23 mg/day and 10 mg/day did not differ for serious adverse events (low SOE), but the higher dose increased risk of withdrawals due to adverse events (moderate SOE). Donepezil 23 mg/day and 10 mg/day doses did not statistically differ for risk of confusion (1.3% vs. 1.1%, p=0.82; 2 trials),^{221, 223} somnolence (2.1 % vs. 0.8%, p=0.08; 2 trials),^{221, 223} falls considered serious treatment-emergent adverse events (0.6% vs. 0.4%),^{221, 223} or mortality (0.7% vs. 0.9%, p=0.46; 2 trials).^{221, 223} No studies comparing these doses reported data on extrapyramidal symptoms or stroke. When directly compared with donepezil 5 mg/day, the higher dose of 10 mg/day appeared to increase risk of both serious adverse events and withdrawals due to adverse events. However, donepezil 10 mg/day and 5 mg/day doses did not statistically differ for risk of confusion (7% vs. 6%, 1 trial),¹⁹⁷ falls (6.3% vs. 6.9%; 1 trial)²⁰⁴ or mortality (1.0% vs. 0.6%; Peto odds ratio 1.69 [95% CI, 0.42 to 6.80]; 3 trials).^{197, 204, 211} No studies comparing these doses reported data on somnolence, extrapyramidal symptoms or stroke.

Appendix Tables E.4-E.10 provide detailed evidence tables and assessments of strength of evidence for key comparisons and outcomes.

Table 6.4

Summary of findings for primary outcomes: donepezil dose comparisons.

Variation in Outcomes by Participant Characteristics

In a post hoc analysis in one trial in individuals with moderate to severe CATD, 23 mg/day donepezil improved Severe Impairment Battery (SIB) scores more than 10 mg/day donepezil among participants with baseline MMSE ≤16 (p<0.001), but not among those with MMSE 17 to 20 (p<0.94).²²¹ However, no test of interaction was reported.

Key Messages

• In older adults with mild to moderate CATD, evidence for galantamine compared with placebo showed:

  • Improvement in cognition (low SOE).
  • Increased risk for withdrawals due to serious adverse events (low SOE).
  • Insufficient evidence for individual cognitive domains, function, quality of life, clinical impression of change, or serious adverse events.
• In older adults with moderate to severe CATD, evidence for galantamine compared with placebo was insufficient for all efficacy and harms outcomes.
• In older adults with mild to moderate CATD, evidence directly comparing different galantamine doses was insufficient for all efficacy and harms outcomes.

Eligible Studies

We identified nine eligible publications of seven unique trials reporting results for cognition, function, quality of life, global staging, clinical impression of change, or harms.^225–233 Four unique trials were assessed as having high ROB, and were therefore not used in our analysis.^228–233 We analyzed the findings of the three remaining medium ROB trials, including two that compared galantamine with placebo,^{225, 226} and one that directly compared different dosages of galantamine.²²⁷ Appendix Tables E.11-E.15 provide evidence tables, summary ROB assessments, and strength of evidence for key comparisons and outcomes.

Baseline Study Characteristics

Characteristics of the 1,060 participants enrolled in the two analyzed trials that compared galantamine with placebo are shown in Table 6.5. One of these trials enrolled community-dwelling participants and compared galantamine 24 mg/day and 32 mg/day with placebo,²²⁶ while the other enrolled individuals residing in residential homes, nursing homes, or geriatric residences and compared galantamine 24 mg/day with placebo.²²⁵ Both trials lasted 26 weeks.

Table 6.5

Baseline characteristics of galantamine versus placebo trials.

One 28-week open-label trial (n=34) enrolled participants with mild to moderate CATD and directly compared galantamine 16 mg/day with galantamine 24 mg/day.²²⁷

Outcomes

Table 6.6 summarizes the primary efficacy and harms results between galantamine and placebo. For efficacy, galantamine improved cognition compared with placebo (low SOE), but evidence was insufficient for function or global change measures. No studies reported data on quality of life, staging or change in patient residence to a different level of independence.

For harms, galantamine increased risk of withdrawals due to adverse events in individuals with mild to moderate CATD (low SOE), but evidence was insufficient in those with severe CATD. Evidence was insufficient to draw conclusions about differences between galantamine and placebo for serious adverse events. In one trial reporting, galantamine and placebo did not statistically differ for falls (12% vs. 11%),²²⁵ but mortality was lower in the galantamine group (4.0% vs. 11.0%, p=0.01).²²⁵ No studies reported data on confusion, somnolence, extrapyramidal symptoms or stroke.

The single trial that directly compared galantamine 16 mg/day and 24 mg/day reported that treatment groups did not differ for cognition or function but provided no numerical data.²²⁷

Appendix Tables E.12-E.15 provide detailed evidence tables and strength of evidence for key comparisons and outcomes.

Table 6.6

Summary of findings for primary outcomes: galantamine versus placebo.

Variation in Outcomes by Participant Characteristics

One trial of 653 participants with mild to moderate CATD (baseline MMSE scores 11 to 24) found that the relative benefits of galantamine compared with placebo on the 11-item ADAS-Cog scale appeared larger in patients with baseline MMSE <18 than in those with baseline MMSE ≥18.²²⁶ A second trial conducted in 407 participants with severe CATD (baseline MMSE 5 to 12) reported that the effect of galantamine versus placebo on mean total SIB score at 6 months did not vary as a function of baseline MMSE score (p=0.92).²²⁵ However, authors did not report a test for interaction. No studies assessed whether participant characteristics modified the effects of galantamine versus placebo on other efficacy or harms outcomes.

Key Messages

• In older adults with mild to moderate CATD, the evidence for rivastigmine 12 mg/day oral or 9.5 mg/day transdermal patch compared with placebo showed:

  • Small improvement in cognition (low SOE for brief cognitive tests commonly used as individual stand-alone tests, moderate SOE for brief multidomain batteries), function (low SOE), staging (low SOE), and clinical impression of change (moderate SOE), and insufficient evidence for attention.
  • Increased withdrawals due to adverse events (moderate SOE), but no difference in serious adverse events (low SOE).
• In older adults with mild to moderate CATD, the evidence for rivastigmine 4 mg/day oral or 4.6 mg/day transdermal patch compared with placebo showed:

  • No difference in cognition (moderate SOE for brief cognitive tests commonly used as individual stand-alone tests, low SOE for brief multidomain batteries), function (low SOE), and insufficient evidence in global staging, but a less than small improvement in clinical impression of change (low SOE).
  • No difference in serious adverse events or withdrawals due to adverse events (low SOE).
• In older adults with moderate to severe CATD, the evidence for rivastigmine 12 mg/day oral or 9.5 mg/day transdermal patch compared with placebo showed small improvement in clinical impression of change (low SOE) and insufficient evidence for cognition.

Eligible Studies

Based in part on a high-quality 2015 systematic review (search date March 2015),²³⁴ we identified 26 eligible publications of eight unique randomized controlled trials of at least 24 weeks duration, that compared rivastigmine with placebo, and reported results for cognition, function, quality of life, global staging, clinical impression of change, or harms.

Of the eight eligible trials, three were rated high ROB (two for high attrition) and not analyzed.^235–237 The five remaining trials were included in analyses.^238–242 Appendix Tables E.16-E.23 provide evidence tables, summary ROB assessments, and strength of evidence for key comparisons and outcomes.

Baseline Study Characteristics

Characteristics of the 3,674 participants enrolled in the five analyzed trials are shown in Table 6.7. Participants were primarily community dwelling at baseline. Four trials enrolled participants with baseline MMSE scores of either 10 to 26, or 10 to 20.^{239–241, 243} One trial (n=218) enrolled only participants with a MMSE score of 5 to 12 and a Global Deterioration Scale (GDS) stage of 5 to 6.²⁴² Treatment duration for all five trials was 24 or 26 weeks.

Table 6.7

Characteristics of the rivastigmine versus placebo trials.

Rivastigmine was administered orally or as a transdermal patch. Three trials evaluated oral rivastigmine only,^{239, 242, 243} one evaluated the patch only,²⁴¹ and one evaluated both delivery methods.²⁴⁰ Targeted oral doses were 4 mg/day and 12 mg/day in divided doses, and targeted patch doses were 4.6 mg/day, 9.5 mg/day, and 17.4 mg/day. The 9.5 mg/day to 12 mg/day doses were categorized as standard and the 4 mg/day to 4.6 mg/day doses were categorized as low. The 17.4 mg/day dose exceeds the maximum FDA recommended for the rivastigmine patch (13.3 mg/day), so results are included only in Appendix Tables E.18-E.23 and are not discussed further in the report text.

Outcomes

Table 6.8 summarizes the primary efficacy and harms outcomes from the five studies with low or medium risk of bias that compared rivastigmine with placebo. For efficacy, in patients with mild to moderate CATD treated for 24 to 26 weeks, rivastigmine 12 mg/day oral and 9.5 mg/day patch each improved several outcomes compared with placebo. These included brief cognitive tests commonly used as individual stand-alone tests (low SOE), brief multidomain cognitive batteries (moderate SOE), function (low SOE), staging (low SOE), and clinical impression of change (moderate SOE). Evidence was insufficient for attention. However, rivastigmine 4 mg/day oral and 4.6 mg/day patch did not differ from placebo for brief cognitive tests commonly used as individual stand-alone tests (moderate SOE), brief multidomain cognitive batteries (low SOE), evidence was inconsistent for clinical impression of change (low SOE), and evidence was insufficient for staging. No studies reported data comparing rivastigmine with placebo for quality of life or change in patient residence to different level of independence. In patients with moderate to severe CATD treated for 26 weeks, rivastigmine 12 mg/day oral showed small improvement in clinical impression of change (low SOE) and insufficient evidence for cognition.

For harms, no rivastigmine dose differed with placebo for risk of serious adverse events (low SOE). However, rivastigmine 12 mg/day oral or 9.5 mg/day patch compared with placebo appeared to increase risk of withdrawals due to adverse events (moderate SOE). Rivastigmine 4 mg/day oral or 4.6 mg/day patch did not differ with placebo for risk of withdrawals due to adverse events (low SOE). Rivastigmine and placebo did not appear to differ for risk of mortality, whether for 12 mg/day oral or 9.5 mg/day patch (0.6% vs. 0.4%; peto OR, 1.18 [95% CI, 0.40 to 3.51]),^240–243 or for 4 mg/day oral or 4.6 mg/day patch (2 deaths in 1,054 participants in all treatment groups combined^{241, 243}). In another trial that compared rivastigmine 12 mg/day oral administered either twice or three times daily with placebo, no deaths occurred.²³⁹ No studies reported information about confusion, somnolence, falls, extrapyramidal symptoms or stroke.

Appendix Tables E.18-E.23 provide detailed evidence tables and strength of evidence for key comparisons and outcomes.

Table 6.8

Summary of findings for primary outcomes: rivastigmine versus placebo.

Variation in Outcomes by Participant Characteristics

No studies reported whether participant characteristics modified the effect of rivastigmine on cognition, function, quality of life, global staging, clinical impression of change, or harms.

Key Messages

• In older adults with mild to moderate CATD, evidence for standard-dose rivastigmine (12 mg/day oral or 9.5 mg/day patch) compared with low-dose rivastigmine (4 mg/day oral or 4.6 mg/day patch) showed:

  • No difference in clinical impression of change (low SOE) or serious adverse events (low SOE).
  • Insufficient evidence about cognition, function, global staging, quality of life, or withdrawals due to adverse events.
• In older adults with mild to moderate CATD, evidence for rivastigmine 12 mg/day oral compared with 9.5 mg/day or 13.3 mg/day patch showed:

  • No difference in clinical impression of change (low SOE) or cognition tests for attention (low SOE).
  • Insufficient evidence about brief cognitive tests commonly used as individual stand-alone tests, brief multidomain cognitive batteries, cognition tests for executive function, function, staging, quality of life, serious adverse events, and withdrawals due to adverse events.
• In older adults with mild to moderate CATD, evidence for 12 mg/day oral rivastigmine in three divided doses compared with two divided doses was insufficient to draw conclusions for all efficacy and harms outcomes.

Eligible Studies

Based in part on a high-quality 2015 systematic review (search date March 2015),²³⁴ we identified 33 eligible publications of 11 unique randomized controlled trials of at least 24 weeks duration, that directly compared rivastigmine with another rivastigmine dose or delivery route, and reported results for cognition, function, quality of life, global staging, clinical impression of change, or harms. Of the 11 eligible trials, six were rated high risk of bias (most commonly for high attrition) and not analyzed.^{235, 237, 244–247} The five remaining trials were included in analyses.^{239–241, 243, 248} Appendix Tables E.18-E.23 provide evidence tables, summary risk of bias assessments, and strength of evidence for key comparisons and outcomes

Baseline Study Characteristics

Characteristics of the 2,973 participants in the rivastigmine arms in the five analyzed trials are shown in Table 6.9. Participants were primarily community dwelling at baseline. Treatment duration for four trials was 24 or 26 weeks. One trial had an open-label phase following which “decliners” were enrolled in a 48-week double-blind phase; we analyzed the blinded data only at 24 weeks due to high attrition with longer followup.²⁴⁸ All trials enrolled participants with mild to moderate CATD severity at baseline (MMSE entry criteria ranged from 10-20 to 10-26).^{239–241, 243, 248} However, the open-label decliners trial likely included a minority of individuals with MMSE <10 at baseline of its double-blind phase better categorized with moderate to severe CATD.²⁴⁸

Rivastigmine was administered orally or as a transdermal patch. Two trials evaluated oral rivastigmine only,^{239, 243} two evaluated patch only,^{241, 248} and one evaluated both.²⁴⁰ Targeted oral doses were 4 mg/day and 12 mg/day in divided doses, and targeted patch doses were 4.6 mg/day, 9.5 mg/day, 13.3 mg/day (maximum recommended by the Food and Drug Administration) and 17.4 mg/day. As stated in the Rivastigmine versus Placebo section above, results for the 17.4 mg/day dose are included only in Appendix Tables E.18-E.23.

Table 6.9

Characteristics of the rivastigmine dose comparison trials.

Outcomes

Table 6.10 summarizes the primary efficacy and harms from the five trials that compared different rivastigmine doses. For comparative effectiveness, evidence was insufficient to draw conclusions between any rivastigmine dose comparisons for brief cognitive tests commonly used as individual stand-alone tests, brief multidomain cognitive batteries, function, staging, or clinical impression of change. No studies reported data on quality of life or change in patient residence to a different level of independence.

For harms, evidence was mostly insufficient about any differences between rivastigmine doses for risk of serious adverse events or withdrawals due to adverse events. In one trial, participants randomized to 13.3 mg/day rivastigmine patch and 9.5 mg/day patch did not appear to differ for risk of confusion (1.8% vs. 2.1%)²⁴⁸ or falls (4.3% vs. 3.5%).²⁴⁸ In four trials that reported on mortality, two deaths occurred in the 12 mg/day oral group compared with five in the 9.5 mg/day patch group,²⁴⁰ two deaths occurred in the 12 mg/day oral group compared with none in the 4 mg/day oral group,²⁴³ no deaths occurred in the 9.5 mg/day patch group compared with one death in the 4.6 mg/day patch group,²⁴¹ and no deaths occurred in either of two 12 mg/day oral dosing groups.²³⁹ No studies reported data on somnolence, extrapyramidal symptoms or stroke.

Appendix Tables E.18-E.23 provide detailed evidence tables and strength of evidence for key comparisons and outcomes.

Table 6.10

Summary of findings for primary outcomes: rivastigmine dose comparisons.

Variation in Outcomes by Participant Characteristics

No studies reported whether participant characteristics modified the effect of rivastigmine dose on cognition, function, quality of life, global staging, clinical impression of change, or harms.

Key Messages

• In older adults with mild to moderate CATD who were not being treated with a cholinesterase inhibitor, the evidence for memantine compared with placebo showed:

  • No difference for function (low SOE).
  • Small improvement for clinical impression of change (low SOE).
  • Insufficient evidence about cognition, serious adverse events, or withdrawals due to adverse events.
  • No evidence about quality of life or staging of disease.
• In older adults with mild to moderate CATD who were being treated with a cholinesterase inhibitor, evidence for memantine compared with placebo showed:

  • No difference in clinical impression of change (low SOE).
  • Insufficient evidence about cognition, function, serious adverse events, or withdrawals due to adverse events.
  • No evidence about quality of life or staging of disease.
• In older adults with moderate to severe CATD who were being treated with a cholinesterase inhibitor, the evidence for memantine compared with placebo showed:

  • Inconsistent effects on cognition (low SOE for improvement on brief battery, insufficient SOE for brief tests commonly used as individual stand-alone tests).
  • No difference in function (low SOE).
  • Small improvement in clinical impression of change (low SOE).
  • Insufficient evidence about serious adverse events and withdrawals due to adverse events.
  • No evidence about quality of life or staging of disease.

Eligible Studies

We identified 31 eligible publications of 16 unique randomized controlled trials that compared memantine with placebo in adults with CATD and reported results for cognition, function, quality of life, global staging, clinical impression of change, or harms.^249–279 Ten trials had high ROB and were excluded from analyses.^{249, 251–256, 258, 261, 264} The most common reason for high ROB was high attrition, including for all four eligible trials of 52 weeks or longer that did not report shorter-term outcomes.^{255, 256, 261, 264} The remaining six trials had medium ROB and were analyzed.^{250, 257, 259, 260, 262, 263} Appendix Tables E.24-E.29 provide evidence tables, summary ROB assessments, and strength of evidence for key comparisons and outcomes.

Baseline Study Characteristics

Characteristics of the 2,227 participants in the six analyzed trials are shown in Table 6.11, organized by category of treatment comparison: (1) memantine versus placebo in patients untreated with a cholinesterase inhibitor;²⁵⁰ (2) add-on memantine versus placebo in patients continuing stable pre-study (≥3-6 months) cholinesterase inhibitor treatment;^{257, 260, 262, 263} and (3) randomization of individuals receiving pre-study cholinesterase inhibitor treatment (≥3 months) to continued cholinesterase inhibitor monotherapy, discontinuation of cholinesterase inhibitor, replacement of cholinesterase inhibitor with memantine monotherapy, or continuation of cholinesterase inhibitor plus add-on memantine.²⁵⁹

All trials included participants with probable CATD, with one ²⁵⁷ also including participants with possible AD.²⁵⁹ Two trials enrolled participants with mild to moderate CATD (n=836), with a mean baseline MMSE of 17 (range 10-22).^{250, 262} The other four trials enrolled participants with moderate to severe CATD (n=1,391),^{257, 259, 260, 263} with a mean baseline MMSE of 10 (range 3-17). Nearly all study participants were community-dwelling,²⁶⁰ and all were required to have a competent caregiver. Mean participant age was 77 years, 64 percent were female, and 93 percent were white.

Five trials evaluated memantine doses of 20 mg once daily, or 10 mg taken twice daily^{250, 259, 260, 262, 263} and one trial evaluated extended-release memantine 28 mg once daily.²⁵⁷

Table 6.11

Baseline characteristics of memantine versus placebo trials by type of drug comparison, with and without continued cholinesterase inhibitor.

Memantine Versus Placebo (Without Cholinesterase Inhibitor)

Outcomes

Table 6.12 summarizes the primary efficacy and harms results between memantine and placebo in patients not receiving a cholinesterase inhibitor. Results were based on a single eligible trial that randomized 403 participants with mild to moderate CATD to memantine 20 mg daily versus placebo for 24 weeks.²⁵⁰

For efficacy, evidence was insufficient to draw conclusions about between group differences for change in brief multidomain cognitive batteries and low strength evidence showed no between group differences for change in function (low SOE). Although likelihood of clinical impression of change not worsening (i.e., either any improvement or no change) was greater with memantine than placebo (low SOE), no studies reported on likelihood of improvement or of moderate or marked improvement. Further, no studies reported data on brief cognitive tests commonly used as individual stand-alone tests, quality of life, staging, or change in patient residence to a different level of independence.

For harms, evidence was insufficient to draw conclusions about between treatment group differences for serious adverse events or withdrawals due to adverse events. Although memantine increased risk of somnolence versus placebo (7% vs. 1%; RR, 7.03 [95% CI, 1.62 to 30.56]), treatment groups did not statistically differ for risk of confusion (5% vs. 3.5%; RR, 1.44 [95% CI, 0.56 to 3.70]), falls (7.4% vs. 7.5%; RR, 1.00 [95% CI, 0.50 to 2.00]) or mortality (0.5% vs. 0.5%; RR, 1.00 [95% CI, 0.06 to 15.96]). No studies reported data on extrapyramidal symptoms or stroke.

Table 6.12

Summary of findings for primary outcomes: memantine versus placebo (none receiving cholinesterase inhibitor).

Memantine Versus Placebo (With Cholinesterase Inhibitor)

Outcomes

Table 6.13 summarizes the primary efficacy and harms from the five trials that compared memantine with placebo in adults with CATD who were receiving a stable cholinesterase inhibitor. For efficacy, in adults with mild to moderate CATD, add-on memantine and add-on placebo did not differ for function or clinical impression of change (both low SOE), and evidence was insufficient to draw conclusions about differences in cognition. No studies reported data on quality of life, global staging, or change in patient residence to a greater level of dependence. In adults with moderate to severe CATD, add-on memantine resulted in small improvements in clinical impression of change and brief multidomain cognitive batteries, but not function (all low SOE). Evidence in this population was insufficient to draw conclusions about between-treatment differences in brief cognitive tests commonly used as individual stand-alone tests or domain-level cognitive tests typically part of a larger battery, and no studies reported data on quality of life, staging, or change in patient residence to a different level of independence.

For harms, evidence was insufficient to draw conclusions about differences between add-on memantine and add-on placebo for serious adverse events (9.9% vs. 9.3%; ARD, 0.6% [95% CI, -2.9% to 4.1%]; RR, 1.06 [95% CI, 0.73 to 1.53]; 2 trials)^{, 257, 262} or withdrawals due to adverse events (8.1% vs. 8.4%; ARD, ‑0.2% [95% CI, ‑3% to 2.5%]; RR, 0.92 [95% CI, 0.49 to 1.71]; 3 trials).^{257, 262, 263} In addition, treatment groups did not statistically differ for confusion (4.5% vs. 2.7%; ARD, 1.8% [95% CI, ‑0.1% to 3.7%]; RR, 1.62 [95% CI, 0.70 to 3.76]; 3 trials),^{257, 262, 263} somnolence (2.9% vs. 1.2%; ARD, 1.7% [95% CI, ‑0.4% to 3.9%]; RR, 2.46 [95% CI, 0.78 to 7.75]; 1 trial),²⁵⁷ falls (7.4% vs. 8.5%; ARD, ‑1.1% [95% CI, ‑3.6% to 1.4%]; RR, 0.87 [95% CI, 0.56 to 1.36]; 4 trials),^{257, 259, 263} stroke (1.2% vs. 1.7%; ARD, ‑0.5% [95% CI, ‑1.9% to 1.1%]; RR, 0.73 [95% CI, 0.25 to 2.12]; 2 trials)^{257, 259} or mortality (3.4% vs. 4.3%; ARD, ‑0.9% [95% CI, -2.9% to 1.1%]; RR, 0.76 [95% CI, 0.44 to 1.34]; 3 trials).^{257, 259, 262} No studies reported data on extrapyramidal symptoms.

Table 6.13

Summary of findings for primary outcomes: memantine versus placebo (all continuing stable pre-trial cholinesterase inhibitor), by baseline CATD severity.