Cognitive issues associated with dementia are well established, but around 90% of the patients with dementia may also experience non-cognitive symptoms such as increased aggression, anxiety, apathy, agitation, depression, delusions, hallucinations and sleep disturbances (NICE, 2018).

Herewith, patients with (more advanced) dementia could experience neuropsychiatric symptoms (e.g. agitation, aggression, distress, psychosis). These symptoms could be caused by the progression of dementia, by a psychiatric cause or by somatic causes. These symptoms significantly impact patients, families, and carers. If these symptoms could be treated successfully, it would have a big impact on the quality of life for everyone concerned.

However, the effectiveness and specific type of antipsychotics is unclear. Therefore, this module aims to evaluate the effectiveness and identify the best antipsychotic treatments for managing neuropsychiatric symptoms in dementia patients.

Subquestion 1. Placebo comparisons

1.1 Typical antipsychotics

Population: Patients with dementia and neuropsychiatric symptoms

Intervention: typical antipsychotics; Comparator: Placebo

_{CMAI: Cohen-Mansfield Agitation Inventory;  NPI - NH: Neuropsychiatric Inventory Home Version; SMD: Standardized mean difference; MD: Mean difference.}

_{^1. Risk of Bias: very serious. All studies were rated at high risk of bias in at least one of the following domains: selection bias (comparability of study groups), attrition bias (incomplete outcome data), and other bias (use of a run-in period); Imprecision: serious. Wide confidence intervals.}

_{^2. Risk of Bias: serious. Due to high dropout rates, baseline differences, and missing information on multiple domains; Imprecision: serious. Wide confidence intervals.}

_{^3. Risk of Bias: serious. Due to high dropout rates, baseline differences, and missing information on multiple domains.}

_{^4. Risk of Bias: very serious. All studies were rated at high risk of bias in at least one of the following domains: selection bias (comparability of study groups), attrition bias (incomplete outcome data), and other bias (use of a run-in period); Imprecision: serious. Confidence interval indicates both an important effect and an effect with no clinical relevance.}

_{^5. Risk of Bias: serious. Due to missing information on multiple domains and because the study was rated at high risk of bias at the following domain: other sources of bias (use of a run-in period, commercial funding). Imprecision: very serious. Wide confidence intervals (crossing one border of clinical relevance) and the low number of patients.}

_{^6. Risk of Bias: serious. Due to missing information on multiple domains and because the studies were rated at high risk of bias at the following domain: other sources of bias (use of a run-in period, commercial funding); Imprecision: serious. Wide confidence intervals and the low number of patients.}

1.2 Atypical antipsychotics

Population: Patients with dementia and neuropsychiatric symptoms

Intervention: Atypical Antipsychotics; Comparator: Placebo

_{CMAI: Cohen-Mansfield Agitation Inventory;  NPI - NH: Neuropsychiatric Inventory Home Version; SMD: Standardized mean difference; MD: Mean difference.}

_{^1. Risk of Bias: serious. All studies were rated at high risk of bias in at least one of the following domains: selection bias (comparability of study groups), attrition bias (incomplete outcome data), reporting bias (selective outcome reporting), and other bias; Imprecision: serious. Wide confidence intervals.}

_{^2. Risk of bias: serious. All studies were rated at high risk of bias in at least one of the following domains: selection bias (comparability of study groups), attrition bias (incomplete outcome data), reporting bias (selective outcome reporting), and other bias; Imprecision: serious. Wide confidence intervals.}

_{^3. Risk of Bias: serious. All studies were rated at high risk of bias in at least one of the following domains: selection bias (comparability of study groups), attrition bias (incomplete outcome data), reporting bias (selective outcome reporting), and other bias; Imprecision: serious. Wide confidence intervals (crossing border of clinical relevance).}

_{^4. Risk of Bias: serious. All studies were rated at high risk of bias in at least one of the following domains: selection bias (comparability of study groups), attrition bias (incomplete outcome data), reporting bias (selective outcome reporting), and other bias;  Imprecision: serious. Wide confidence intervals (crossing border of clinical relevance).}

_{^5. Risk of Bias: serious. Due to missing information about randomization procedure and blinding, and high risk of attrition bias (incomplete outcome data). Inconsistency: serious. Conflicting results. Imprecision: serious. Wide confidence intervals and the low number of patients.}

_{^6. Risk of Bias: serious. All studies were rated at high risk of bias in at least one of the following domains: selection bias (comparability of study groups), attrition bias (incomplete outcome data), reporting bias (selective outcome reporting), and other bias;  Imprecision: serious. Wide confidence intervals (crossing border of clinical relevance).}

Subquestion 2. Head-to-head comparisons

2.1 Typical antipsychotics

Population: Patients with dementia and neuropsychiatric symptoms

Intervention: Typical Antipsychotics; Comparator: Head-to-head

Not reported.

2.2 Atypical antipsychotics

Population: Patients with dementia and neuropsychiatric symptoms

Intervention: Atypical Antipsychotics; Comparator: Head-to-head

Quetiapine vs risperidone

_{SAS: Simpson Angus scale; CMAI: Cohen-Mansfield Agitation Inventory;  NPI - NH: Neuropsychiatric Inventory Home Version; MMSE: Mini-Mental State Examination; MD: mean difference; SD: standardized difference}

_{^1. Risk of Bias: serious. Imprecision: serious. Wide confidence intervals.}

_{^2. Risk of Bias: serious. Imprecision: serious. Low number of patients.}

_{^3. Risk of Bias: serious. Imprecision: serious. Low number of patients.}

_{^4.Risk of Bias: serious. Imprecision: serious. Low number of patients.}

_{^5.Risk of Bias: serious. Imprecision: serious. Low number of patients.}

_{^6.Risk of Bias: serious. Imprecision: serious. Low number of patients.}

Olanzapine vs Quetiapine

_{BPRS: Brief Psychiatric Rating Scale; NPI- NH: Neuropsychiatric Inventory Home Version; MD: mean difference.}

_{^1.Risk of Bias: serious. Imprecision: serious. Wide confidence intervals.}

Olanzapine vs risperidone

_{BPRS: Brief Psychiatric Rating Scale; NPI: Neuropsychiatric Inventory Home Version; MD: mean difference; 1.Risk of Bias: serious. Imprecision: serious. Wide confidence intervals.}

Quetiapine vs Haloperidol

_{BPRS: Brief Psychiatric Rating Scale; CGI-S: Clinical Global Impressions–Severity of Illness; NPI - NH: Neuropsychiatric Inventory Home Version; MD: mean difference; SD: standardized difference}

_{^1. Risk of Bias: serious. Imprecision: very serious. Low number of patients and wide confidence intervals;}

_{^2. Risk of Bias: serious. Imprecision: serious. Low number of patients.}

PICO 1. Description of studies

One systematic review was included in the analysis of the literature. Important study characteristics and results are summarized in Table 1. The assessment of the risk of bias is summarized in the risk of bias tables (under the tab ‘Evidence tabellen’).

Mühlbauer (2021) performed a Cochrane systematic review that assessed the efficacy and safety of antipsychotics for the treatment of agitation and psychosis in people with Alzheimer's disease and vascular dementia. On 7 January 2021, a search was performed in various international databases (ALOIS, the Cochrane Dementia and Cognitive Improvement Group's register, MEDLINE (Ovid Sp), Embase (Ovid SP), PsycINFO (Ovid SP), CINAHL (EBSCOhost), Web of Science Core Collection (ISI Web of Science), LILACS (BIREME), ClinicalTrials.gov and the World Health Organization's meta-register, and the International Clinical Trials Registry Portal). Studies that were included were randomized, placebo-controlled trials comparing the effects of antipsychotics and placebo for the treatment of agitation or psychosis in people with dementia due to Alzheimer's disease or vascular dementia. Authors excluded studies comparing different antipsychotics head-to-head and antipsychotic withdrawal trials.

The search yielded 8233 separate hits of which 24 trials met the eligibility criteria including 6090 patients. Six trials tested a typical antipsychotic (haloperidol, thiothixene), four for agitation and two for psychosis. Twenty trials tested an atypical antipsychotic (e.g. risperidone, olanzapine, aripiprazole, quetiapine), eight for agitation and twelve for psychosis. Pooled data were presented and included in the literature analyses. Reported outcomes measures were psychosis, adverse events (somnolence and extrapyramidal symptoms), agitation, carer burden or carer quality of life, and cognitive function.

PICO 2. Description of studies

One systematic review and network meta-analysis was included in the analysis of the literature. Important study characteristics and results are summarized in Table 2. The assessment of the risk of bias is summarized in the risk of bias tables (under the tab ‘Evidence tabellen’).

Lü (2024) conducted a systematic review and network meta-analysis on the efficacy, acceptability, and tolerability of second-generation antipsychotics, also known as atypical antipsychotics, on behavioural and psychological symptoms of dementia. A search was performed in four databases (PubMed, Embase, Web of Science, Cochrane Trial Register) from inception to December 2023. Patients were included with Alzheimer’s dementia (AD), vascular dementia, or mixed dementia. Patients with Parkinson’s dementia, Lewy body dementia, other mental comorbidities unrelated to dementia (e.g. depression, delirium, schizophrenia), or uncontrolled physical illness (e.g. any physical disease in the acute phase, cardiovascular- or cerebrovascular accidents) or poorly controlled chronic disease (eg, poorly controlled hypertension and diabetes, residual symptoms of cerebral vascular events convalescence) were excluded. There were no restrictions based on patient’s age or the severity of dementia.

The search yielded 874 separate hits of which 19 trials met the eligibility criteria for the NMA. Briefly, these studies were published between 1999 and 2023, and their sample sizes varied from 40 to 652 participants with a total of 6374 individuals with intervention lengths ranging from 6 weeks to 36 weeks. 

The NMA only included three closed loops Quetiapine vs Olanzapine vs placebo, risperidone vs Quetiapine vs placebo and risperidone vs olanzapine vs placebo). Thus, three out of 19 studies included head-to-head comparison:

• Rainer, 2007: Quetiapine versus risperidone
• Schneider, 2006: Olanzapine versus Quetiapine versus risperidone
• Tariot, 2006: Quetiapine versus Haloperidol

The other studies included in Lü (2024) answered subquestion 1.1: placebo comparison, for which Muhlbauer (2021) was already included. In addition, Lü (2024) had an inadequate search strategy for inclusion as an NMA. Therefore, the three studies that performed head-to-head comparison were included rather than the full NMA.

Table 1. Characteristics of included studies

_{*For further details, see risk of bias table in the Appendix}

_{** Lü (2024) reported low risk of bias for the study Schneider (2006). However, due to baseline differences in the NPI scores, the RoB for the outcome psychosis was concerned with moderate bias}

_{*** Lü (2024) reported low risk of bias for the study Tariot (2006). However, due to missing information on the blinding and randomization, the RoB for this study was concerned with moderate bias}

Results

1. Subquestion - placebo

1.1 Typical antipsychotics

Psychosis

Two studies included in Mühlbauer (2021) reported on psychosis, presented in units of NPI-NH (higher is worse) (Devanand, 1998; Tariot, 2006). The analysis resulted in a SMD of -0.29  (95% CI: -0.55 to -0.03), favouring the typical antipsychotic group. This difference was considered a small clinically relevant effect. Results are shown in a forest plot (Figure 1).

Figure 1. Psychosis: Forest plot of the SMD on the effect of typical antipsychotics versus placebo

_{¹ Devanand (1998): Haloperidol (low dose of 0.50-0.75 mg/day, or standard dose of 2-3 mg/day)}

_{² Tariot (2006): Haloperidol (0.5-12 mg/day)}

Somnolence

Three studies included in Mühlbauer (2021) reported on somnolence, which was assessed with different instruments (Allain, 2000; Tariot, 2006; Teri, 2000). In the antipsychotics group, 53 out of 229 patients experienced somnolence compared to 17 out of 237 patients in the placebo group. This resulted in a risk ratio of 2.75 (95% CI: 0.85 to 8.92) favouring the placebo group. This difference was clinically relevant. Results are shown in a forest plot (Figure 2).

Figure 2. Somnolence: Forest plot of the risk ratio on the effect of typical antipsychotics versus placebo

_{¹ Allain (2000): Haloperidol (2 mg/day; dose could be progressively increased to a maximum of 6 mg/day)}

_{² Tariot (2006): Haloperidol (0.5-12 mg/day)}

_{³ Teri (2000): Haloperidol (0.5-2 mg/day)}

Extrapyramidal symptoms

Three studies included in the SR Mühlbauer (2021) reported on the outcome extrapyramidal symptoms, which was assessed with different instruments (Allain, 2000; Tariot, 2006; Teri, 2000). In the antipsychotics group, 77 out of 229 patients experienced extrapyramidal symptoms compared to 35 out of 238 patients in the placebo group. This resulted in a risk ratio of 2.26 (95% CI: 1.58 to 3.23) favouring the placebo group. This difference was clinically relevant. Results are shown in a forest plot (see Figure 3).

Figure 3. Extrapyramidal symptoms: Forest plot of the risk ratio on the effect of typical antipsychotics versus placebo

_{¹ Allain (2000): Haloperidol (2 mg/day; dose could be progressively increased to a maximum of 6 mg/day)}

_{² Tariot (2006): Haloperidol (0.5-12 mg/day)}

_{³ Teri (2000): Haloperidol (0.5-2 mg/day)}

Agitation

Four studies included in Mühlbauer (2021) reported on the outcome agitation, presented in units on CMAI (higher is worse) (Allain, 2000; Devanand, 1998; Finkel, 1995; Teri, 2000). The analysis resulted in a SMD of -0.40 (95% CI: -0.77 to -0.03), favouring the typical antipsychotic group. This difference was considered a small clinically relevant effect. Results are shown in a forest plot (Figure 4).

Figure 4. Agitation: Forest plot of the SMD on the effect of typical antipsychotics versus placebo

_{¹ Allain (2000): Haloperidol (2 mg/day; dose could be progressively increased to a maximum of 6 mg/day)}

_{² Devanand (1998): Haloperidol (low dose of 0.50-0.75 mg/day, or standard dose of 2-3 mg/day)}

_{³ Finkel (1995): Thiothixene (mean dose 4.6 mg/day with a range of 0.25 to 18 mg/day)}

_{⁴ Teri (2000): Haloperidol (0.5-2 mg/day)}

Carer burden or carer quality of life

One study from the SR of Mühlbauer (2021) reported on caregiver burden (Teri, 2000). The mean difference of the caregiver burden with patients treated with typical antipsychotics (Haloperidol; 0.5-2 mg/day) (n=34, -1.88 ± 8.89) versus caregiver burden with patients treated with placebo (n=36, -2.58 ± 9.67), was 0.70 (95% CI: -3.65 to 5.05), favouring the placebo group. This difference was not clinically relevant.

Health-related quality of life

Mühlbauer (2021) did not report this outcome.

Cognitive function (important)

Two studies included in Mühlbauer (2021) reported on the outcome cognitive functioning, presented in units of MMSE (Tariot, 2006; Teri, 2000). The mean difference between patients treated with typical antipsychotics (n=97) and patients treated with placebo (n=108) was -0.25 (95% CI: -1.27 to 0.77). This difference was not clinically relevant. Results are shown in a forest plot (Figure 5).

Figure 5. Cognitive function: Forest plot of the mean difference on the effect of typical antipsychotics versus placebo

_{¹ Tariot (2006): Haloperidol (0.5-12 mg/day)}

_{² Teri (2000): Haloperidol (0.5-2 mg/day)}

1.2 Atypical antipsychotics

Psychosis

Twelve studies included in Mühlbauer (2021) reported on psychosis, presented in units of NPI-NH (higher is worse) (Ballard, 2018; Deberdt, 2005; De Deyn, 2004; De Deyn, 2005; Mintzer, 2006; Mintzer, 2007; NCT00187742, 2006; Paleacu, 2008; RIS-INT-83, 2003; Schneider, 2003; Streim, 2008; Tariot, 2006). The analysis resulted in a SMD of -0.11  (95% CI: -0.18 to -0.03), favouring the atypical antipsychotic group. This difference was not clinically relevant. Results are shown in a forest plot (Figure 6).

Figure 6. Psychosis: Forest plot of the SMD on the effect of atypical antipsychotics versus placebo

_{¹ Ballard (2018): Pimavanserin (2x17 mg tablets daily)}

_{² Deberdt (2005) F1D MC HGGU: risperidone (0.5-2 mg/day)}

_{³ Deberdt (2005) F1D MC HGGU: Olanzapine (2.5-10 mg/day)}

_{⁴ De Deyn (2004) F1D MC HGIV: Olanzapine (1.0, 2.5, 5.0, or 7.5 mg/day)}

_{⁵ De Deyn (2005): Aripiprazole (2-15 mg/day)}

_{⁶ Mintzer (2006) RIS USA 232: risperidone (1.0-1.5 mg/day)}

_{⁷ Mintzer (2007): Aripiprazole (2, 5, or 10 mg/day)}

_{⁸ NCT00287742 (2006): risperidone (0.5-2 mg/day)}

_{⁹ Paleacu (2008): Quetiapine (50-300 mg/day)}

_{¹⁰ RIS-INT-83 (2003): risperidone (1-1.5 mg/day)}

_{¹¹ Schneider (2003) RIS USA 63: risperidone (1, 2, or 4 mg/day)}

_{¹² Streim (2008): Aripiprazole (2-15 mg/day)}

_{¹³ Tariot (2006): Quetiapine (25-600 mg/day)}

Somnolence

Thirteen studies included in Mühlbauer (2021) reported on somnolence, which was assessed with different instrument (Allain, 2000; Ballard, 2018; Brodaty, 2003; Deberdt, 2005; De Deyn, 2005; Grossberg, 2020b; Mintzer, 2006; Mintzer, 2007; Paleacu, 2008; Schneider, 2006; Streim, 2008; Tariot, 2006; Zhong, 2007). In the antipsychotics group, 352 out of 2353 patients experienced somnolence compared to 108 out of 1525 patients in the placebo group. This resulted in a risk ratio of 2.39 (95% CI: 1.67 to 3.40) favouring the placebo group. This difference was clinically relevant. Results are shown in a forest plot (see Figure 7).

Figure 7. Somnolence: Forest plot of the risk ratio on the effect of atypical antipsychotics versus placebo

_{¹ Allain (2000): Tiapride (100 mg/day; dose could be progressively increased to a maximum of 300 mg/day)}

_{² Ballard (2018): Pimavanserin (2x17 mg tablets daily)}

_{³ Brodaty (2003) RIS-AUS-05: risperidone (0.5-2 mg/day)}

_{⁴ Deberdt (2005) F1D MC HGGU: risperidone (0.5-2 mg/day)}

_{⁵ Deberdt (2005) F1D MC HGGU: Olanzapine (2.5-10 mg/day)}

_{⁶ De Deyn (2005): Aripiprazole (2-15 mg/day)}

_{⁷ Grossberg (2020b): Brexpiprazole (0.5-2 mg/day)}

_{⁸ Mintzer (2006) RIS USA 232: risperidone (1.0-1.5 mg/day)}

_{⁹ Mintzer (2007): Aripiprazole (2, 5, or 10 mg/day)}

_{¹⁰ Paleacu (2008): Quetiapine (50-300 mg/day)}

_{¹¹ Schneider (2006): risperidone (tablets of 0.5 and 1.0 mg. Mean last dose 1.0 mg/day)}

_{¹² Schneider (2006): Quetiapine (tables of 25 and 50 mg. Mean last dose 56.6 mg/day)}

_{¹³ Schneider (2006): Olanzapine (tablets of 2.5 and 5.0 mg. Mean last dose 5.5 mg/day)}

_{¹⁴ Streim (2008): Aripiprazole (2-15 mg/day)}

_{¹⁵ Tariot (2006): Quetiapine (25-600 mg/day)}

_{¹⁶ Zhong (2007): Quetiapine (100-200 mg/day)}

Extrapyramidal symptoms

Fifteen studies included in the SR Mühlbauer (2021) reported on the outcome extrapyramidal symptoms, which was assessed with different instrument (Allain, 2000; Brodaty, 2003; Deberdt, 2005; De Deyn, 2005; Grossberg, 2020a; Grossberg, 2020b; Mintzer, 2006; Mintzer, 2007; NCT00187742, 2006; Paleacu, 2008; RIS-INT-83, 2003; Schneider, 2006; Streim, 2008; Tariot, 2006; Zhong, 2007). In the antipsychotics-group, 361 out of 2582 patients experienced extrapyramidal symptoms compared to 131 out of 1598 patients in the placebo group. This resulted in a risk ratio of 1.39 (95% CI: 1.14 to 1.68) favouring the placebo group. This difference was clinically relevant. Results are shown in a forest plot (see Figure 8).

Figure 8. Extrapyramidal symptoms: Forest plot of the risk ratio on the effect of atypical antipsychotics versus placebo

_{¹ Allain (2000): Tiapride (100 mg/day; dose could be progressively increased to a maximum of 300 mg/day)}

_{² Brodaty (2003) RIS-AUS-05: risperidone (0.5-2 mg/day)}

_{³ Deberdt (2005) F1D MC HGGU: Olanzapine (2.5-10 mg/day)}

_{⁴ Deberdt (2005) F1D MC HGGU: risperidone (0.5-2 mg/day)}

_{⁵ De Deyn (2005): Aripiprazole (2-15 mg/day)}

_{⁶ Grossberg (2020a): Brexpiprazole (0.5-2 mg/day)}

_{⁷ Grossberg (2020b): Brexpiprazole (0.5-2 mg/day)}

_{⁸ Mintzer (2006) RIS USA 232: risperidone (1.0-1.5 mg/day)}

_{⁹ Mintzer (2007): Aripiprazole (2, 5, or 10 mg/day)}

_{¹⁰ NCT00287742 (2006): risperidone (0.5-2 mg/day)}

_{¹¹ Paleacu (2008): Quetiapine (50-300 mg/day)}

_{¹² RIS-INT-83 (2003): risperidone (1-1.5 mg/day)}

_{¹³ Schneider (2006): risperidone (tablets of 0.5 and 1.0 mg. Mean last dose 1.0 mg/day)}

_{¹⁴ Schneider (2006): Quetiapine (tables of 25 and 50 mg. Mean last dose 56.6 mg/day)}

_{¹⁵ Schneider (2006): Olanzapine (tablets of 2.5 and 5.0 mg. Mean last dose 5.5 mg/day)}

_{¹⁶ Streim (2008): Aripiprazole (2-15 mg/day)}

_{¹⁷ Tariot (2006): Quetiapine (25-600 mg/day)}

_{¹⁸ Zhong (2007): Quetiapine (100-200 mg/day)}

Agitation

Seven studies included in Mühlbauer (2021) reported on the outcome agitation, presented in units on CMAI (higher is worse) (Allain, 2000; Ballard, 2005; Grossberg 2020a; Grossberg, 2020b; Schneider, 2006; Zhong, 2007). The analysis resulted in a SMD of -0.21 (95% CI: -0.30 to -0.12), favouring the atypical antipsychotic group. This difference was considered a small clinically relevant effect. Results are shown in a forest plot (Figure 9).

Figure 9. Agitation: Forest plot of the SMD on the effect of atypical antipsychotics versus placebo

_{¹ Allain (2000): Tiapride (100 mg/day; dose could be progressively increased to a maximum of 300 mg/day)}

_{² Ballard (2005): Quetiapine (50-100 mg/day)}

_{³ Grossberg (2020a): Brexpiprazole (0.5-2 mg/day)}

_{⁴ Grossberg (2020b): Brexpiprazole (0.5-2 mg/day)}

_{⁵ Schneider (2006): Olanzapine (tablets of 2.5 and 5.0 mg. Mean last dose 5.5 mg/day)}

_{⁶ Schneider (2006): Quetiapine (tables of 25 and 50 mg. Mean last dose 56.6 mg/day)}

_{⁷ Schneider (2006): risperidone (tablets of 0.5 and 1.0 mg. Mean last dose 1.0 mg/day)}

_{⁸ Zhong (2007): Quetiapine (100-200 mg/day)}

_{⁹ Brodaty (2003) RIS-AUS-05: risperidone (0.5-2 mg/day)}

Carer burden or carer quality of life

Mühlbauer (2021) did not report this outcome.

Health-related quality of life

One study from the SR of Mühlbauer (2021) reported on caregiver burden (Schneider, 2006). This study studied three atypical antipsychotics: Olanzapine, Quetiapine and risperidone. Overall, the mean difference in health-related quality of life in patients treated with typical antipsychotics versus placebo was -0.95 (95%CI: -6.04 to 4.14). This difference was considered a small clinically relevant effect. Results are shown in a forest plot (Figure 10). The mean values in the figure have been converted to negative values to facilitate correct interpretation within the context of the figure.

Figure 10. Health-related quality of life: Forest plot of the mean difference on the effect of atypical antipsychotics versus placebo

_{¹ Schneider (2006): risperidone (tablets of 0.5 and 1.0 mg. Mean last dose 1.0 mg/day)}

_{² Schneider (2006): Olanzapine (tablets of 2.5 and 5.0 mg. Mean last dose 5.5 mg/day)}

_{³ Schneider (2006): Quetiapine (tables of 25 and 50 mg. Mean last dose 56.6 mg/day)}

Cognitive function

Eleven studies included in Mühlbauer (2021) reported on the outcome cognitive functioning, presented in units of MMSE (Ballard, 2005; Ballard, 2018; Deberdt, 2005; De Deyn, 2005; Grossberg 2020a; Grossberg, 2020b; Mintzer, 2007; Schneider, 2006; Streim, 2008; Tariot, 2006; Zhong, 2007). The SMD between patients treated with atypical antipsychotics (n=1759) and patients treated with placebo (n=939) was 0.11 (95% CI: 0.01 to 0.22). This difference was not clinically relevant. Results are shown in a forest plot (Figure 11). The mean values in the figure have been converted to facilitate correct interpretation within the context of the figure.

Figure 11. Cognitive function: Forest plot of the SMD on the effect of atypical antipsychotics versus placebo

_{¹ Ballard (2005): Quetiapine (50-100 mg/day)}

_{² Ballard (2018): Pimavanserin (2x17 mg tablets daily)}

_{³ Deberdt (2005) F1D MC HGGU: Olanzapine (2.5-10 mg/day)}

_{⁴ Deberdt (2005) F1D MC HGGU: risperidone (0.5-2 mg/day)}

_{⁵ De Deyn (2005): Aripiprazole (2-15 mg/day)}

_{⁶ Grossberg (2020a): Brexpiprazole (0.5-2 mg/day)}

_{⁷ Grossberg (2020b): Brexpiprazole (0.5-2 mg/day)}

_{⁸ Mintzer (2007): Aripiprazole (2, 5, or 10 mg/day)}

_{⁹ Schneider (2006): Quetiapine (tables of 25 and 50 mg. Mean last dose 56.6 mg/day)}

_{¹⁰ Schneider (2006): Olanzapine (tablets of 2.5 and 5.0 mg. Mean last dose 5.5 mg/day)}

_{¹¹ Schneider (2006): risperidone (tablets of 0.5 and 1.0 mg. Mean last dose 1.0 mg/day)}

_{¹² Streim (2008): Aripiprazole (2-15 mg/day)}

_{¹³ Tariot (2006): Quetiapine (25-600 mg/day)}

_{¹⁴ Zhong (2007): Quetiapine (100-200 mg/day)}

2. Subquestion - Head-to-head comparison

2.1 Typical antipsychotics

Lü (2024) does not report any head-to-head comparisons on typical antipsychotics. The comparison of the typical antipsychotic Haloperidol to the atypical antipsychotic Quetiapine is reported in 1.2.2.

2.2 Atypical antipsychotics

Psychosis

Quetiapine vs risperidone

Two studies reported on psychosis (Rainer, 2007; Schneider, 2006). The two studies used different scales, and Rainer (2007) did not report the mean (SD), thus data could not be pooled.

Rainer (2007) reported psychosis in units of NPI-NH, which decreased from 25.6 to 17.5 with Quetiapine (n=34; 50-400 mg/day) and from 25.7 to 16.6 with risperidone (n=31; 0.5-4 mg/day). Mean (SD) was not reported, thus the mean difference could not be calculated and clinical relevance could not be assessed.

Schneider (2006) reported psychosis with the BPRS and NPI. NPI score was reduced with 16.6 ± 18.3 in the Quetiapine group (n=31; mean last dose 56.5 mg/day) and 16.4 ± 15.0 in the risperidone group (n=32; mean last dose 1.0 mg/day). This resulted in a mean difference of 0.20 (95%CI: -8.84 to 9.24), in favour of the Quetiapine group, which was not clinically relevant. BPRS score was reduced with 7.4 ± 9.8 in the Quetiapine group (n=30) and 8.5 ± 12.2 in the risperidone group (n=32). This resulted in a mean difference of 1.10 (95%CI: -4.39 to 6.59), in favour of the risperidone group, which was not clinically relevant.

Olanzapine vs Quetiapine

One study reported on psychosis (Schneider, 2006) measured with the BPRS and NPI. The NPI score was reduced with 14.0 ± 18.7 point in the Olanzapine group (n=40; mean last dose 5.5 mg/day) and 16.6 ± 18.3 in the Quetiapine group (n=31; mean last dose 56.5 mg/day). This resulted in a mean difference of 2.20 (95%CI: -6.46 to 10.86), in favour of the Quetiapine group, which was not clinically relevant. BPRS score was reduced with 7.4 ± 9.8 in the Olanzapine group (n=40) and 7.4 ± 9.8 in the Quetiapine group (n=30). This resulted in a mean difference of 0.00 (95%CI: -4.64 to 4.64), which was not clinically relevant.

Olanzapine vs risperidone

One study reported on psychosis (Schneider, 2006) measured with the BPRS and NPI. NPI score was reduced with 14.0 ± 18.7 point in the Olanzapine group (n=40; mean last dose 5.5 mg/day) and 16.4 ± 15.0 in the risperidone group (n=32; mean last dose 1.0 mg/day). This resulted in a mean difference of 2.40 (95%CI: -5.58 to 10.18), in favour of the risperidone group, which was not clinically relevant. BPRS score was reduced with 7.4 ± 9.8 in the Olanzapine group (n=40) and 8.5 ± 12.2 in the risperidone group (n=32). This resulted in a mean difference of 1.10 (95%CI: -4.10 to 6.30), in favour of the risperidone group, which was not clinically relevant.

Quetiapine vs Haloperidol

One study reported on psychosis (Tariot, 2006), measured with the BPRS and CGI-S after 10 weeks. The BPRS score was reduced by 9.06 ± 11.07 in the Quetiapine group (n=85; 25-600 mg/day) and 7.13 ± 10.85 in the Haloperidol group (n=86; 0.5-12 mg/day). This resulted in a mean difference of 1.93 (95%CI: -1.36 to 5.22), in favour of the Haloperidol group, which was not clinically relevant. The CGI-S score was reduced by 0.60 ± 1.05 in the Quetiapine group (n=85) and 0.52 ± 1.00 in the Haloperidol group (n=86). This resulted in a mean difference of 0.08 (95%CI: -0.23 to 0.39), in favour of the Quetiapine group, which was not clinically relevant.

Somnolence

Quetiapine vs risperidone

One study reported on somnolence (Rainer, 2007). Two of the 38 patients with Quetiapine (50-400 mg/day) (5.3%) experienced somnolence compared to zero of the 34 patients with risperidone (0.5-4 mg/day) (0%). This resulted in a risk ratio of -0.05 (95%CI: -0.14 to 0.03), in favour of the risperidone group, which was not clinically relevant.

Quetiapine vs Haloperidol

One study reported on somnolence (Tariot, 2006). 23 of the 91 patients with Quetiapine (25-600 mg/day) (25.3%) experienced somnolence compared to 34 of the 94 patients with Haloperidol (0.5-12 mg/day) (36.2%). This resulted in a risk ratio of -0.11 (95%CI: -0.24 to 0.02), in favour of the Quetiapine group, which was not clinically relevant.

Extrapyramidal symptoms

Quetiapine vs risperidone

One study reported on extrapyramidal symptoms (Rainer, 2007). The incidence of extrapyramidal symptoms, as measured by the mean change from baseline to endpoint in SAS scores, increased with a score of (5.00 – 5.06) 0.06 in the Quetiapine group (n=38; 50-400 mg/day) and a score of (5.68 – 6.03) 0.35 in the risperidone group (n=34; 0.5-4 mg/day). Mean (SD) was not reported, thus the mean difference could not be calculated and clinical relevance could not be assessed.

Agitation

Quetiapine vs risperidone

One study reported on agitation (Rainer, 2007). Both Quetiapine and risperidone showed some efficacy against agitation, as shown by a reduction in CMAI scores from baseline to Week 8, with no significant differences between treatments (Quetiapine, 59.21 to 55.67, n=33 (50-400 mg/day); risperidone, 54.48 to 48.97, n=31 (0.5-4 mg/day)). Mean (SD) was not reported, thus the mean difference could not be calculated and clinical relevance could not be assessed.

Quetiapine vs Haloperidol

One study reported on agitation (Tariot, 2006), measured with the BPRS-agitation and NPI-NH agitation. After 10 weeks the BPRS agitation scale was reduced by 2.14  ± 3.58 in the Quetiapine group (n=85; 25-600 mg/day) and 2.37  ± 3.43 in the Haloperidol group (n=86; 0.5-12 mg/day). This resulted in a mean difference of 0.23 (95%CI: -0.82 to 1.28), in favour of the Haloperidol group, which was not clinically relevant. After 10 weeks the NPI-NH agitation scale was reduced by 2.63 ± 4.99 in the Quetiapine group (n=86) and 2.43  ± 4.45 in the Haloperidol group (n=86). This resulted in a mean difference of 0.20 (95%CI: -1.36 to 1.76), in favour of the Quetiapine group, which was not clinically relevant.

Carer burden or carer quality of life

Quetiapine vs risperidone

One study reported on burden to the carer, presented in units of NPI-NH (Rainer, 2007). The NPI score was reduced from 30.2 to 27.7 with Quetiapine (n=34; 50-400 mg/day) and from 30.3 to 26.7 with risperidone (n=31; 0.5-4 mg/day). Mean (SD) was not reported, thus de mean difference could not be calculated and clinical relevance could not be assessed.

Health-related quality of life

None of the studies reported on health-related quality of life

Cognitive function

Quetiapine vs risperidone

One study reported on cognitive function (Rainer, 2007) measured with MMSE. The MMSE score remained stable from baseline to Week 8 with Quetiapine (n=30; 50-400 mg/day) (18.50 to 18.47) and risperidone (n=26; 0.5-4 mg/day) (18.00 to 18.65). Mean (SD) was not reported, thus the mean difference could not be calculated and clinical relevance could not be assessed.

Quetiapine vs Haloperidol

One study reported on cognitive function (Tariot 2006) measured with MMSE. The MMSE score was reduced by 1.58 ± 2.98 in the Quetiapine group (n=69; 25-600 mg/day) and 1.06 ± 4.26 in the Haloperidol group (n=63; 0.5-12 mg/day). This resulted in a mean difference of 0.52 (95%CI: -0.75 to 1.79), in favour of the Quetiapine group, which was not clinically relevant.

A systematic review of the literature was performed to answer the following question(s):

PICO 1 What is the effectiveness of antipsychotics in comparison to placebo in patients with dementia and neuropsychiatric symptoms?

PICO 2 What is the effectiveness of various antipsychotics in patients with dementia and neuropsychiatric symptoms?

Relevant outcome measures

The guideline panel considered psychosis and adverse events as critical outcome measures for decision making; and agitation, carer burden or carer quality of life, health-related quality of life, and cognitive function as important outcome measures for decision making.

A priori, the guideline panel did not define the outcome measures listed above but used the definitions used in the studies.

For all outcome measures, the guideline panel defined a 25% difference for dichotomous outcomes (0.8 ≥ RR ≥ 1.25) and 0.5 SD for continuous outcomes as a minimal clinically (patient) important difference.

If studies used different rating instruments (e.g. psychosis could be measured with NPI, BEHAVE-AD) to measure the same outcome measure, then the standardized mean difference (SMD) is calculated. According to Cohen, SMD values of 0.2-0.5 represent a small clinically relevant effect, values of 0.5-0.8 a medium clinically relevant effect, and values >0.8 represent a large clinically relevant effect (Cohen, 1988).

Search and select (Methods)

The databases Medline (via OVID) and Embase (via Embase.com) were searched with relevant search terms from 2010 until 30 August 2024. The detailed search strategy is listed under the tab ‘Literature search strategy’. The systematic literature search resulted in 4086 hits. Studies were selected based on the following criteria:

• Systematic reviews in which searches were performed in at least two databases, with a detailed search strategy, risk of bias assessment and results of individual studies available, randomized controlled trials or (observational) comparative studies.
• Full-text English or Dutch language publication and
• Studies according to the PICO.

Initially, 31 studies were selected based on title and abstract screening. After reading the full text, 29 studies were excluded (see the exclusion table under the tab ‘Evidence tabellen’). For example, reasons for exclusion were a less recent search (Ma, 2014), search performed in a Chinese database in which the Chinese full text articles were not accessible (Zhang, 2024), insufficient methodological quality (Gerlach, 2020). Eventually, two systematic reviews were selected for inclusion:

• Subquestion 1 Placebo comparison: Mühlbauer (2021).
• Subquestion 2 Head-to-head comparison: Lü (2024).