Which Medication to Reduce First: Antipsychotic or Depakote in Elderly Dementia Patients
In elderly patients with dementia taking both an antipsychotic and valproic acid (Depakote), the antipsychotic should be reduced first. This recommendation is based on the substantially higher mortality risk associated with antipsychotics compared to valproic acid, and the stronger evidence supporting antipsychotic deprescribing in this population.
Evidence Supporting Antipsychotic Reduction First
Mortality Risk Comparison
Antipsychotics carry a 1.6–1.7-fold increased mortality risk compared to placebo in elderly dementia patients, with absolute mortality increases ranging from 2.0% to 3.8% over 180 days depending on the specific agent 1, 2.
Haloperidol shows the highest mortality risk (3.8% absolute increase, NNH=26), followed by risperidone (3.7%, NNH=27), olanzapine (2.5%, NNH=40), and quetiapine (2.0%, NNH=50) when compared to matched nonusers 2.
When compared directly to antidepressants, antipsychotics show even higher mortality increases: haloperidol 12.3% (NNH=8), risperidone 5.4%, olanzapine 4.7%, and quetiapine 3.2% (NNH=31) 2.
Valproic acid shows lower mortality risk than antipsychotics, with a relative risk of 0.91 (95% CI=0.78–1.06) compared to risperidone as reference 3.
Guideline-Based Deprescribing Recommendations
The Mayo Clinic Proceedings explicitly recommends attempting antipsychotic taper within 3–6 months to determine the lowest effective maintenance dose in agitated dementia patients, as approximately 47% of patients continue receiving antipsychotics after discharge without clear indication 4, 1.
Antipsychotics should only be continued if the patient has severe, dangerous agitation with psychotic features that threatens substantial harm to self or others 1.
Daily in-person examination is required to evaluate ongoing need for antipsychotics and assess for side effects including extrapyramidal symptoms, falls, metabolic changes, and QT prolongation 1.
Dose-Response Mortality Risk
Atypical antipsychotics show a dose-response increase in mortality risk, with 3.5% greater mortality (95% CI=0.5%–6.5%) in high-dose subgroups relative to low-dose groups 2.
When compared directly with quetiapine, dose-adjusted mortality risk is increased with both risperidone (1.7%; 95% CI=0.6%–2.8%) and olanzapine (1.5%; 95% CI=0.02%–3.0%) 2.
Evidence Regarding Valproic Acid in Dementia
Efficacy and Safety Profile
Valproic acid has shown mixed results in controlled trials for behavioral and psychological symptoms of dementia, with five controlled studies failing to confirm efficacy despite numerous open-label studies showing benefit 5.
Valproic acid is generally well tolerated in geriatric populations, with no notable major side effects reported except possible excessive sedation; hematologic and hepatic effects are not more frequent than in the general population 5.
The FDA label warns of somnolence in elderly dementia patients treated with valproate, with significantly higher proportions showing somnolence and dehydration compared to placebo, and a trend toward reduced nutritional intake and weight loss 6.
Specific Geriatric Considerations
In elderly patients, valproate dosage should be increased more slowly with regular monitoring for fluid and nutritional intake, dehydration, somnolence, and other adverse reactions 6.
Dose reductions or discontinuation should be considered in patients with decreased food or fluid intake and in patients with excessive somnolence 6.
Clinical Algorithm for Medication Reduction
Step 1: Assess Current Clinical Status
Determine if the patient still meets criteria for antipsychotic use: severe agitation, distress, or threatening substantial harm to self or others 1.
Evaluate for reversible medical causes of behavioral symptoms including pain, urinary tract infections, constipation, dehydration, and metabolic disturbances 1.
Review current behavioral symptoms using quantitative measures such as the Cohen-Mansfield Agitation Inventory or NPI-Q to establish baseline 1.
Step 2: Initiate Antipsychotic Taper
Begin gradual antipsychotic withdrawal over a period greater than 1 month to minimize potential discontinuation effects including dyskinesias, parkinsonian symptoms, dystonias, and neuroleptic malignant syndrome 4.
Monitor closely during taper for worsening behaviors, and be prepared to re-escalate dosing if persisting withdrawal symptoms result in distress to the patient 4.
Implement intensive non-pharmacological interventions during the taper including caregiver education employing redirection and reorientation techniques, environmental interventions, simplifying tasks, participation in activities, optimizing sensorial input, ensuring social engagement, and maintaining the sleep-wake cycle 4.
Step 3: Consider Valproic Acid Adjustment Only After Antipsychotic Taper
If behavioral symptoms remain controlled after antipsychotic discontinuation, consider whether valproic acid is still providing benefit or can also be tapered 4.
If valproic acid taper is indicated, do so gradually over 2–4 weeks with monitoring for worsening behaviors 4.
Monitor specifically for somnolence, dehydration, and reduced nutritional intake during valproic acid taper in elderly patients 6.
Common Pitfalls to Avoid
Do not discontinue both medications simultaneously, as this makes it impossible to determine which medication was providing benefit and increases the risk of behavioral decompensation 4.
Do not abruptly discontinue the antipsychotic, as this is associated with withdrawal dyskinesias, parkinsonian symptoms, and neuroleptic malignant syndrome 4.
Do not continue antipsychotics indefinitely without attempting taper, as the American Geriatrics Society recommends reviewing need at every visit and tapering if no longer indicated 1.
Do not assume that because a patient has been stable on both medications, both are necessary; many patients can be successfully tapered without worsening of behavioral symptoms 4.
Special Circumstances Requiring Caution
In patients with a history of severe behavioral crises, the antipsychotic taper should be even more gradual with very close monitoring 4.
In patients with baseline low albumin, lower valproate clearance, and higher BUN, there is increased risk of somnolence and dehydration with valproic acid, warranting closer monitoring if this medication is continued 6.
In patients taking carbapenem antibiotics, valproate levels may drop to subtherapeutic levels, potentially unmasking the need for the antipsychotic if behavioral symptoms worsen 6.