Hide Sidebar

Mortality risk of commonly prescribed antipsychotic medications.

Mortality risk of commonly prescribed antipsychotic medications.

Critique date: Tue, 09/07/2010 - 12:30pm
Institution: 
Mount Sinai School of Medicine
   
Article Citation
Citation: 
Rossom, R. C., Rector, T. S., Lederle, F. A., & Dysken, M. W. (2010). Are all commonly prescribed antipsychotics associated with greater mortality in elderly male veterans with dementia? Journal of the American Geriatrics Society, doi:10.1111/j.1532-5415.2010.02873.x
   
Clinical Bottom Lines

1. Commonly prescribed doses of quetiapine (25 mg to 50 mg daily) were not associated with increased mortality within 30 days of initiation.

2. Haldol, olanzapine, and risperidone at commonly prescribed doses were associated with higher 30-day mortality compared to their respective control groups. 

   
Methods
Type of Study: 
Harm
Study Design: 
Cohort study
Follow-up Period: 
360 days
Setting: 
Veterans national healthcare data in Austin, Texas
Patient Population: 

Number of participants: Haloperidol Control (C)= 8,867 Exposed (E) = 2,217; Olanzapine C = 13,536, E= 3,384; Quetiapine C= 17,109 E= 4.277; Risperidone C= 32,996 E= 8,249

Mean Age: All were 78 +/- 6 except for the control group for Haloperidol = 77 +/- 6

Sex: Males = 98%, Females = 2-3%

 

Inclusions:

Veterans in the VHA Austin, Texas database that had an inpatient or outpatient encounter that had listed a diagnosis of dementia from October 1999 to September 2005

ICD-9 Dementia Diagnoses Included: dementia, senile dementia, dementia not otherwise specified, vascular dementia, unspecified senile psychotic conditions, alcohol-induced dementia, dementia in conditions classified elsewhere, Alzheimer’s disease, frontotemporal dementia, Lewy body dementia

Significant Exclusions: 

Veterans <65 years

Veterans with these diagnoses: schizophrenia, bipolar affective disorder, delusional disorder, other nonorganic psychoses

Intervention/Exposure: 

Intervention/Exposure:

Exposure= Haloperidol, Olanzapine, Risperidone or Quetiapine prescription.

No specific doses were required to have been included in the exposed group in the study. However, study noted the median (25th – 75th percentile) dose prescribed for each medication; namely haloperidol 1 mg (1-2); olandazpine 5 mg (2.5-5), quetiapine 50 mg (25-50), risperidone 1 mg (0.5-1).

 

Control = Veterans who had a diagnosis of dementia and did not receive an anti-psychotic prescription during the period of study.

For each exposed subject, all potential control subjects with their first diagnosis of dementia in the same quarter of the same year and still at risk when the antipsychotic exposure began were identified, four were randomly selected as control subjects. The date the index antipscychotic prescription was dispensed to the exposed subject was assigned as index date for the four controls. Controls were selected without replacement, many unexposed subjects were selected as controls for more than one anti-psychotic.

Outcome Measures: 

 Primary Outcome: Mortality rate in five follow-up periods: day 0-30, day 31-90, day 91-180, day 181-270, day 271-360

Secondary Analysis:

Excluding inpatients and subjects discharged within 30 days

Adjusting for Baseline Covariates

Prior Conditions: Ischemic heart disease, heart failure, Peripheral Vascular Disease, Cerebrovascular Accident, Diabetes Mellitus, COPD, Cancer, Renal Disease

Prior Inpatient Stay which was classified as follows: none, on index date, within 90 days, within 91 to 365 days

Prior Prescriptions: cholinesterase inhibitor, memantine, benzodiadepine, Parkinson’s drug, anticonvulsant, antidepressant, opiate, NSAID, other analgesic, diuretic, ACE-I or ARB, Beta-blocker, Calcium channel blocker, digitalis, nitrate, antiplatelet drug, warfarin, alpha blocker, lipid reducer, diabetes drug, respiratory drug

Participant Follow-up: 
Patients were censored when a subject’s last prescription for the initial antipsychotic was consumed, when a different antipsychotic was initiated, or on September 30, 2005, whichever occurred first. Subjects were followed for 360 days.
   
Conclusion
Results: 

 

1. The percentage dying was greater for the exposed cohorts only in the initial 30-day follow up period. The haloperidol cohort had the highest mortality absolute risk increase at 3.7% followed by Olanzapine 1%, and Risperidone 0.8%. (Table I)

2. The quetiapine cohort did not demonstrate a mortality absolute risk increase in any of the follow-up periods, including the initial 30-day period. (Table I)

3. At any prescribed dose, the haloperidol, olanzapine, and risperidone cohorts (unadjusted) demonstrate statistically significant increased mortality risk; after adjusting for baseline covariates, only the haloperidol and olanzapine cohorts demonstrate this increased mortality risk.  (Table II)

4. Haloperidol, Olanzapine, and Risperidone exhibit a dose-response gradient. Higher prescribed doses have statistically significant increased mortality risk compared to lower prescribed doses. (Table II)

 

I. Unadjusted Mortality Percentages Within Specified Follow-Up Intervals*

 

Haloperidol

Olanzapine

Quetiapine

Risperidone

 

Control

Exposed

ARI

NNH

Control

Exposed

ARI

NNH

Control

Exposed

ARI

NNH

Control

Exposed

ARI

NNH

n

8,867

2,217

 

 

13,536

3,384

 

 

17,109

4,277

 

 

32,996

8,249

 

 

Follow-Up Interval, Days

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0–30

1.7%

5.4%

3.7%

27

1.7%

2.7%

1%

100

1.7%

1.7%

0%

N/A

2%

2.8%

0.8%

125

ARI- absolute risk increase, NNH - number needed to harm

*Unadjusted mortality percentages within the first 30 days for all doses prescribed  

 

II. Mortality Within the First 30 Days After an Antipsychotic Medication Was Dispensed

Analysis

Haloperidol

Olanzapine

Quetiapine

Risperidone

Any Prescribed Dose, unadjusted

HR (95% (CI)

3.2 (2.5–4.1)

1.7 (1.3–2.2)

1.0 (0.7–1.2)

1.4 (1.2–1.7)

P-value

<0.001

<0.001

0.8

<0.001

Any prescribed dose, adjusted

HR (95% (CI)

2.2 (1.7–2.9)

1.3 (1.0–1.7)

0.8 (0.6–1.1)

1.2 (1.0–1.4)

P-value

<0.001

0.03

0.14

0.08

Any prescribed dose excluding recent inpatients, adjusted*

HR (95% (CI)

2.3 (1.6–3.3)

1.4 (1.0–1.9)

0.7 (0.5–1.0)

1.2 (1.0–1.4)

P-value

<0.001

0.03

0.03

0.13

Lower prescribed doses, adjusted

HR (95% (CI)

1.5 (1.0–2.2)

1.0 (0.7–1.6)

0.7 (0.5–1.0)

1.1 (0.9–1.3)

P-value

0.07

0.79

0.03

0.47

Higher prescribed doses, adjusted

HR (95% (CI)

3.2 (2.2–4.5)

1.5 (1.1–2.0)

1.2 (0.7–1.8)

1.6 (1.1–2.2)

P-value

<0.001

0.01

0.5

0.01

notes:

Adjusted  - indicates adjustment for differences in baseline covariates

* excludes subjects who were in a healthcare facility within 30 days of the first prescription of the antipsychotic

Lower prescribed doses were haloperidol <= 1mg, olanzapine <=2.5 mg, quetiapine <=50 mg, risperidone <=1 mg.

Higher prescribed doses were above lower prescribed doses as listed above.

Higher doses of quetiapine had increased HR; however, P value is  0.5 and CI is wide ( 0.7 to 1.8 )
Concerns Regarding Methodology, Applicability to Older Adults, etc.: 

 

1. The groups were not similar at the outset. While the exposed group likely has higher risk of mortality due to covariates, it is unclear whether comparison between the different antipsychotics would be affected by this selection bias.

2. Patients receiving different anti-psychotics were compared to a control group selected using ICD-9 dementia diagnoses (described above in inclusions) diagnosed in the same quarter of the same year as the exposed subject (described in detail above in control) which may not be adequate matching. It is unclear what effect this would have in the comparison across antipsychotics. Furthermore, the authors also state that nearly all comparisons of baseline characteristics were statistically significant even when seemingly inconsequential and they attribute this to the large cohorts.

3. The study is limited by using prescribed doses, the authors were unable to ascertain whether these prescribed doses were the same doses taken by subjects.

4. The authors mention a dose-dependent gradient for quetiapine reflected in an increased HR of 1.2 (doses >50 mg) compared to the lower dose with HR 0.7 (Table 4). However, the HR p value for higher doses of quetiapine is 0.5 and the CI is wide (0.7-1.8).  Hence, a dose-response is not clearly demonstrated.

5. Prior studies have demonstrated harm of antipsychotics in patients with dementia. A randomized control trial comparing the different antipsychotics would help answer which antipsychotic is less harmful. But given the potential harm of antipsyhotics, an RCT may not be practical; and with low event rate and risk differences the RCT would need very large sample sizes. A retrospective cohort design is an alternative but bias in study design may lead to erroneous conclusion.More studies are needed to identify which antipyschotics are safer for use in patients with dementia. However, currently given lack of other evidence to guide treatment decisions, clinicians may consider when all things being equal to favor quetiapine

Funding Source and Role: 
Clinical and Health Services Research and Development Grants from Veteran Health Administration; Minnesota Medical Foundation at the University of Minnesota; the Minnesota Veterans Research Institute
Created By: 
Maria Kristina P. Gestuvo, Geriatrics Fellow, Brookdale Department of Geriatrics, Mount Sinai Medical Center
This is a review of the validity of a single study; the ‘bottom lines’ do not reflect comparison with the rest of the literature on this subject.