Atypical Antipsychotics Augmentation in Patients with Depressive Disorder and Risk of Subsequent Dementia: A Nationwide Population-Based Cohort Study

Abstract

Background:

While atypical antipsychotic medications are widely used for treating depressive disorders, their long-term effects on the risk of subsequent dementia have not been studied adequately.

Objective:

To investigate whether the risk of dementia differs according to the use of atypical antipsychotic drugs, and compare the effects of antipsychotic agents on dementia risk in individuals with late-life depressive disorders.

Methods:

A nationwide population-based retrospective cohort study was conducted using data from the National Health Insurance Service—Senior Cohort of South Korea. Atypical antipsychotic dosages were standardized using a defined daily dose, and the cumulative dosage was calculated. Participants were observed from January 2008 to December 2015. Cox proportional hazard regression analysis was used to estimate the hazard ratios.

Results:

The cohort included 43,788 elderly adults with depressive disorders: 9,901 participants (22.6%) were diagnosed with dementia. Findings showed that atypical antipsychotics were prescribed to 1,967 participants (4.5%). Compared with non-users, users of atypical antipsychotics experienced a significantly higher risk for dementia with an adjusted hazard ratio (aHR) of 1.541 (95% confidence interval [CI], 1.415–1.678). A cumulative dose-response relationship was observed (test for trend, p < 0.0001). Among atypical antipsychotics, risperidone displayed the highest risk for dementia (aHR 1.767, [95% CI, 1.555–2.009]).

Conclusion:

In this study of elderly individuals with depressive disorders, atypical antipsychotic use was associated with a significantly higher risk of subsequent dementia. Healthcare professionals should be aware of this potential long-term risk. A limitation that should be mentioned is that we could not exclude patients with bipolar depression.

Keywords

Antipsychotic agents dementia depression mood disorders

INTRODUCTION

With increases in the elderly population, more people are suffering from dementia, causing it to become one of the most significant health issues of the 21st century. By 2050, more than 100 million people are expected to be living with dementia [1]. Along with the increasing prevalence rate, societal costs have also increased, with one study estimating the societal cost of dementia to be greater than that of cancer or heart disease [2]. Due to the lack of any effective disease-modifying treatment, dementia prevention is essential. Therefore, many studies have focused on the identification of potentially modifiable risk factors [3].

There is growing evidence identifying an association between depression and dementia. A meta-analysis of community-based prospective cohort studies concluded that late-life depression was associated with 1.85 times increase in risk for all-cause dementia compared to those with no depressive episodes [4]. Furthermore, studies have found that severity and frequency of depressive episodes were associated with an increased risk of developing dementia [5, 6].

While several clinical studies have shown that depression is associated with subsequent dementia, whether this is due to depression being an actual risk factor or a prodrome of dementia remains controversial [7]. To address this issue, clinical studies have attempted to use long-term follow-up periods in longitudinal studies or to distinguish between late-life and earlier depressive episodes; however, this has shown conflicting results [8 –10]. Numerous mechanisms have been proposed, including alterations in the hypothalamic-pituitary-adrenal (HPA) axis, amyloid pathology, or depression and dementia as comorbid conditions of vascular pathologies [11]. It is likely that these mechanisms are not mutually exclusive, and may interact with each other.

Despite such controversy, many studies have suggested the possibility of decreasing the risk or delaying the progression of dementia by appropriately treating depression. In animal models, antidepressants reduced the amyloid plaque burden [12], and lithium blocked both stress-induced depression-like behaviors and hippocampal apoptotic changes [13]. Long-term treatment with antidepressants or lithium for patients with mood disorders has also been effective in reducing subsequent dementia risk [14, 15]. Moreover, a cohort study found that long-term treatment with selective serotonin reuptake inhibitors (SSRIs) for elderly individuals with a lifetime history of depression reduced the risk of conversion from mild cognitive impairment to Alzheimer’s dementia [16].

In addition to antidepressants and mood stabilizers, atypical antipsychotic medications are considered as an alternative treatment option for mood disorder patients. Both their approved and off-label use for mood disorders have become common in clinical practices. Studies have revealed the efficacy of atypical antipsychotic drugs in bipolar depression, treatment-resistant depression, and depressive episodes with mixed features [17 –19]. Therefore, it is reasonable to speculate that by achieving a higher remission rate and reducing residual symptoms of depression, atypical antipsychotic drugs may have a beneficial effect on lowering the risk for subsequent dementia.

Despite increased evidence of their efficacy in treating depressive disorders, atypical antipsychotic drugs are prescribed cautiously, due to their potential adverse effects. The same study that reported the efficacy of these medications in depressive disorder treatment also showed increased adverse metabolic outcomes, such as weight gain and abnormal metabolic laboratory results [19]. Furthermore, drugs with anticholinergic effects, which are also common adverse effects of atypical antipsychotic drugs, are known to increase the risk of dementia [20]. These effects complicate the association between the treatment of depressive disorders with antipsychotic medications and the risk of subsequent dementia.

The authors of this study have not discovered any previous research that examined whether the use of atypical antipsychotic drugs in late-life depressive disorders increases the risk of dementia. Only one study has revealed an inverse correlation between the use of clozapine and the severity of dementia in a small population of bipolar disorder patients [21]. Studies conducted among patients with other psychiatric disorders, such as schizophrenia and posttraumatic stress disorder (PTSD), have produced conflicting results [22, 23].

This study hypothesized that the risk of dementia will differ according to atypical antipsychotic augmentation therapy in individuals with late-life depressive disorders. Additionally, the augmentation treatment effects of atypical antipsychotic drugs related to the risk of dementia in late-life depression were compared.

MATERIALS AND METHODS

Data source

Data were obtained from the National Health Insurance Service–Senior Cohort (NHIS-Senior) of South Korea [24]. The NHIS-Senior is a nationwide retrospective administrative data cohort established by the NHIS, which is the universal health insurance provider for all Korean citizens. By random sampling, approximately 550,000 older adults aged≥60 years were selected from the total eligible population of 5.5 million older adults in the 2002 NHIS claims database. The cohort was observed until 2015 for all participants.

Study population

Among individuals in the NHIS-Senior, participants who had been diagnosed with depression between January 1, 2002 and December 31, 2007 were included as the study population. A history of depression was defined as obtaining both a diagnosis of a unipolar depressive disorder (the International Classification of Diseases 10th revision [ICD-10] codes F32.0–F32.9, F33.0–F33.9, and F34.1) and a documented prescription history of antidepressant medications. Among those with depression (n = 63,607), participants who died (n = 8,948) or had been diagnosed with any type of dementia before 2008 (n = 10,871) were excluded. We could not exclude patients with the diagnostic code for bipolar disorder in the study periods due to the lack of information on the diagnostic codes of other psychiatric disorders. This study was approved by the institutional review board of the Seoul National University Bundang Hospital (X-1907/550-910).

Independent variable

The independent variable for the study was the use of second-generation antipsychotic drugs for the treatment of depression. Prescription records of participants were screened for all second-generation antipsychotic drugs available in South Korea (Supplementary Table 1). Participants were defined as users if they were prescribed any atypical antipsychotic drugs at least once between 2002 and 2007. Atypical antipsychotic drugs were further categorized into olanzapine, risperidone, quetiapine, and other atypical antipsychotic drugs, considering each sample size of participants who used each type of medication.

To examine dose-response relationships, dosages for each antipsychotic agent were standardized using a defined daily dose (DDD). DDDs for antipsychotic drugs are provided in the World Health Organization’s data on the anatomical therapeutic chemical (ATC) classification system. Using DDD values, the total standardized daily dose (TSDD) for each participant was calculated by dividing the total prescribed dose of the agent by the DDD. Then, the continuous TSDDs were categorized into four levels based on the distribution of data and interpretability, as < 30 TSDD, 30–89 TSDD, 90–179 TSDD, and≥180 TSDD, which were used for primary analyses.

Outcome variable

The outcome variable was defined as the incidence of dementia. Participants were regarded as having dementia when they obtained both a recorded diagnosis for ICD-10 codes of dementia (F00–F03, G30, and G31) and a documented history of being prescribed cognitive enhancer medication for dementia (e.g., donepezil, rivastigmine, galantamine, and memantine).

Covariates

Covariates were selected based on prior research and upon their availability in the cohort data. Demographic factors included age, gender, and income status. As income status was not provided, insurance premiums were referenced as an indicator of socioeconomic status. Healthcare visit frequency was also considered to account for the ascertainment bias because participants would be more likely to be diagnosed with dementia when they receive frequent medical evaluations. The healthcare visit frequency variable was constructed with four levels according to the quartiles of all visits by all participants. Coexisting morbidities included in the Charlson comorbidity index and history of medications (antidepressants, anticonvulsants, lithium, benzodiazepine, HMG-CoA reductase inhibitors, antidiabetic medications, antihypertensive medications, and antiplatelet medications) were also considered.

Statistical analysis

A chi-squared test was performed to compare baseline demographic characteristics between antipsychotic drug users and the no-use group. Survival time of participants was calculated from January 1, 2008, until the date of incidence of dementia, death, or follow-up end date (December 31, 2015). Cox proportional hazard regression analysis was performed to estimate the hazard ratios (HRs) of incident dementia associated with the use of all antipsychotic drugs. Subgroup analyses by type of antipsychotic drugs were also conducted to estimate type-specific HRs. Furthermore, to investigate dose-response relationships, HRs for categorical TSDD variables for antipsychotic drugs, compared to the reference group (no-use group) were estimated. Then, p-values for linear trends were calculated, by treating the median values of each categorical group as continuous variables. The effect of antipsychotic drugs was first evaluated in the sex- and age-adjusted models and subsequently evaluated in the fully adjusted model, which was controlled for demographic characteristics, healthcare visit frequency, co-morbidities, and other medication history. In the fully adjusted models for associations by type of antipsychotics, adjustments were made for all other types of antipsychotic drugs.

In the sensitivity analysis, a lag period of one to three years from the beginning of the follow-up period was applied. This sensitivity analysis was conducted to exclude participants who, prior to 2008, were not diagnosed with dementia, but were on the course of developing dementia. Further, the analysis was repeated with a diagnosis of Alzheimer’s dementia and other causes of dementia as an outcome to determine whether the effects of antipsychotic drugs differ by the type of dementia. Additionally, subgroup analysis was performed for participants with or without psychotic features during mood episodes, and for participants with or without histories of mood stabilizer use, to investigate possible interactions by severity and bipolar disorder diagnosis, respectively. All statistical analyses were two-sided with a significance threshold set at 0.05. SAS Enterprise Guide version 7.2 (SAS Institute Inc) and R Studio version 1.0.136 (RStudio Inc: packages survival version 2.43-3 and Survminer version 0.4.3) were used to perform the analyses.

RESULTS

Figure 1 shows the flow of the study population. Among the 558,147 participants in the cohort, 43,788 met the inclusion and exclusion criteria of the study. At the follow-up for data as of December 31, 2015, a total of 9,901 participants were diagnosed with dementia.

Fig. 1

Flow diagram of the study population.

Table 1 shows the baseline descriptive characteristics of the study population. Approximately 4.5% of the participants were prescribed one of the second-generation antipsychotic drugs at least once. Compared with the no-use group, participants in all antipsychotic drug use groups were more likely to be male, received a higher income, more frequently sought healthcare services, and were more likely to be medicated with SSRIs or mood stabilizers.

Table 1

Descriptive Characteristics of the Study Population. No. (%) of participants

	Total		Antipsychotics no-use group		All antipsychotics use group		p^a
	(n = 43,788)		(n = 41,821)		(n = 1,967)
	N	%	N	%	N	%
Sex							< 0.0001
Female	30,488	69.6	29,289	70.0	1199	61.0
Age group							0.847
65 to 69 y	14,433	33.0	13,784	33.0	649	33.0
70 to 74 y	14,288	32.6	13,660	32.7	628	31.9
75 to 79 y	8,800	20.1	8,403	20.1	397	20.2
Older than 80 y	6,267	14.3	5,974	14.3	293	14.9
Income, deciles							0.0002
Medical Aid	3,683	8.4	3,555	8.5	128	6.5
Low (1st–3rd decile)	7,614	17.4	7,314	17.5	300	15.3
Middle (4th–7th decile)	10,933	25.0	10,436	25.0	497	25.3
High (8th–10th decile)	21,558	49.2	20,516	49.1	1042	53.0
Healthcare visit frequency^b							< 0.0001
First quartile	10,866	24.8	10,415	24.9	451	22.9
Second quartile	10,751	24.6	10,316	24.7	435	22.1
Third quartile	11,212	25.6	10,731	25.7	481	24.5
Fourth quartile	10,959	25.0	10,359	24.8	600	30.5
Other medication history
Antidepressants
SSRIs	16,919	38.6	15,607	37.3	1312	66.7	< 0.0001
Non-SSRIs	39,061	89.2	37,303	89.2	1758	89.4	0.804
Anticonvulsants	9,139	20.9	8,524	20.4	615	31.3	< 0.0001
Lithium	343	0.8	204	0.5	139	7.1	< 0.0001
Benzodiazepine	41,517	94.8	39,617	94.7	1900	96.6	0.0003

TSDD, total standardized daily dose; SSRIs, Selective serotonin reuptake inhibitors; HMG-CoA, β-Hydroxy β-methylglutaryl-CoA ^aChi-squared test was done between antipsychotics no-use group and all antipsychotics use group. ^bThe fourth quartile group had the highest frequency of medical visits.

Any antipsychotic drug use and dementia risk

Table 2 shows the adjusted HRs for dementia associated with antipsychotic drug use. Participants in the “any antipsychotic drug use” group showed a statistically significant higher risk for dementia (adjusted HR 1.541, 95% CI [1.415–1.678]), compared with the no-use group. In the analysis using the categorical TSDD variable, participants with higher TSDD exhibited a higher risk for dementia; particularly participants with the highest dose (TSDD ≥180) exhibited the highest risk for dementia (adjusted HR 1.984; 95% CI [1.521–2.591]). The linear trend of the dose-response relationship was significant (p < 0.0001).

Table 2

Adjusted hazard ratio (aHR) of dementia associated with antipsychotics use^a .

		Antipsychotics use group
	Antipsychotics no-use group	Antipsychotics use group		< 30 TSDD	30–89 TSDD	90–179 TSDD	≥180 TSDD	p for trend
	(n = 41,821)	(n = 1,967)		(n = 1,375)	(n = 291)	(n = 124)	(n = 177)
			CI
Any atypical antipsychotics use dementia events
	9,318	583		394	88	44	57
Person-years	264,922.3	10,706.5		7,485.9	1,592.8	957	970.8
Incidence^a	3,517.3	5,445.3		5,263.2	5,524.9	4,597.7	5,871.4
Age-and-sex adjusted hazard ratios	(reference)	1.642	1.510–1.786	1.547	1.723	2.02	2.077	< 0.0001
Adjusted hazard ratios^b	(reference)	1.541	1.416–1.678	1.458	1.614	1.847	1.985	< 0.0001
Olanzapine
dementia events	9,804	97		52	20	8	17
Person-years	27,3793.5	1,835.3		1,017.8	319.3	199.5	298.6
Incidence^a	3,580.8	5,285.2		5,109.1	6,263.7	4,010.0	5,693.2
Age-and-sex adjusted hazard ratios	(reference)	1.606	1.315–1.962	1.488	1.997	1.21	1.933	0.001
Adjusted hazard ratios^c	(reference)	1.255	1.022–1.543	1.22	1.465	0.962	1.347	0.187
Risperidone
dementia events	9,637	264		183	42	19	20
Person-years	27,1486.2	4,142.6		2,895.1	627.8	264.2	355.5
Incidence^a	3,549.7	6,372.8		6,321.0	6,690.0	7,191.5	5,625.9
Age-and-sex adjusted hazard ratios	(reference)	1.986	1.757–2.245	1.928	2.1	2.359	2.005	< 0.0001
Adjusted hazard ratios^c	(reference)	1.767	1.555–2.009	1.719	1.874	1.973	1.878	< 0.0001
Quetiapine
dementia events	9,712	189		150	22	8	9
Person-years	272,298.7	3,330.1		2,649.9	364.7	151.6	163.9
Incidence^a	3,566.7	5,675.5		5,660.6	6,032.4	5,277.0	5,491.2
Age-and-sex adjusted hazard ratios	(reference)	1.677	1.452–1.937	1.651	1.953	1.257	2.153	0.002
Adjusted hazard ratios^c	(reference)	1.352	1.165–1.569	1.357	1.635	0.845	1.377	0.3
Other atypical antipsychotics
dementia events	9,739	162		115	29	9	9
Person-years	272,178.5	3,450.3		2630.1	507	199.8	113.4
Incidence^a	3,578.2	4,695.2		4372.5	5,719.9	4,504.5	7,936.5
Age-and-sex adjusted hazard ratios	(reference)	1.398	1.197–1.633	1.28	1.684	1.565	2.947	< 0.0001
Adjusted hazard ratios^c	(reference)	1.208	1.032–1.413	1.138	1.335	1.255	2.243	0.004

CI, confidence interval; TSDD, total standardized daily dose ^aIncidence was defined as events/100,000 person-years ^bAdjusted for sex, age group, income, depression diagnosis, co-morbidities, healthcare visit frequency, and other medication history ^cAdjusted for sex, age group, income, depression diagnosis, co-morbidities, healthcare visit frequency, other medication history and other atypical antipsychotics use.

The results of the sensitivity analysis supported the robustness of the results. The statistical significance of the higher risk for dementia was not altered, even after the lag period was applied from one to three years (Table 3, adjusted HRs 1.431, 1.390, 1.380, respectively). In the subgroup analysis (Supplementary Table 2), mood stabilizer usage or depressive episodes with psychotic features did not affect the outcome. Changes in outcome variables, such as the presence or absence of Alzheimer’s dementia, did not yield differences in HRs.

Table 3

Sensitivity analysis: Adjusted hazard ratio (aHR) of the risk of dementia according to various lag periods

		Antipsychotics use group
	Antipsychotics no-use group	Antipsychotics use group		< 30 TSDD	30–89 TSDD	90–179 TSDD	≥180 TSDD	p for trend
		aHR	CI	aHR	aHR	aHR	aHR
Any atypical antipsychotics use
Lag1	(reference)	1.431	1.305–1.568	1.345	1.607	1.559	1.844	< 0.0001
Lag2	(reference)	1.390	1.258–1.536	1.298	1.551	1.531	1.873	< 0.0001
Lag3	(reference)	1.380	1.236–1.540	1.316	1.476	1.334	1.865	< 0.0001
Olanzapine
Lag1	(reference)	1.207	0.966–1.508	1.136	1.395	0.945	1.434	0.143
Lag2	(reference)	1.181	0.927–1.505	1.071	1.253	1.099	1.555	0.08
Lag3	(reference)	1.121	0.853–1.473	1.031	1.294	0.962	1.358	0.306
Risperidone
Lag1	(reference)	1.644	1.428–1.891	1.612	1.714	1.716	1.774	0.0007
Lag2	(reference)	1.581	1.355–1.846	1.585	1.528	1.306	1.903	0.003
Lag3	(reference)	1.502	1.261–1.788	1.556	1.195	1.224	1.839	0.022
Quetiapine
Lag1	(reference)	1.257	1.068–1.480	1.247	1.599	0.823	1.342	0.382
Lag2	(reference)	1.224	1.025–1.462	1.210	1.572	0.951	1.129	0.609
Lag3	(reference)	1.199	0.985–1.460	1.211	1.420	0.806	1.087	0.857
Other atypical antipsychotics
Lag1	(reference)	1.243	1.056–1.463	1.181	1.502	1.115	1.721	0.043
Lag2	(reference)	1.260	1.060–1.497	1.134	1.744	1.285	2.012	0.006
Lag3	(reference)	1.321	1.098–1.588	1.167	1.974	1.116	2.465	0.002

TSDD, total standardized daily dose; aHR, adjusted hazard ratio; CI, confidence interval; Lag1, Lag period of 1 year; Lag2, lag period of 2 years; Lag3, lag period of 3 years.

Specific antipsychotic drug use and dementia risk

Risks for dementia associated with specific anti-psychotic drug use are shown in Table 2. All the second-generation antipsychotic drugs that were examined showed significantly higher risks of dementia compared with the no-use group. The risperidone use group exhibited the highest risk of dementia (adjusted HR 1.767, 95% CI [1.555–2.009]) among second-generation antipsychotic drugs. In the analysis using the categorized TSDD variables, risperidone and other atypical antipsychotics drug use group indicated a significant dose-response relationship (test for trend, p < 0.0001 and p = 0.004 respectively, Table 2). Results for the olanzapine and quetiapine use groups were different. Although not statistically significant, the higher risks diminished when olanzapine and quetiapine were used in moderate amounts (90–179 TSDD).

When the lag period was applied from one to three years, the higher risk for dementia became statistically insignificant for some antipsychotic drugs, but the trend for higher risk persisted (Table 3). Specifically, the results for the risperidone and other atypical antipsychotic drug use groups remained statistically significant, whereas those for the olanzapine and quetiapine use groups became non-significant for all lag periods and a lag period of three years, respectively.

In the subgroup analysis, although there were no evident interactions, the results differed according to antipsychotic agents (Supplementary Table 2). While risks for dementia of the risperidone use group showed statistically higher risks in every subgroup, those of other groups did not. In groups other than the risperidone use group, there was a shared tendency for higher risks to become insignificant in mood stabilizer use and depression with psychotic feature subgroups. However, the results of the subgroups and sensitivity analyses should be cautiously interpreted, owing to the small sample size.

DISCUSSION

In this nationwide population-based retrospective study of elderly individuals with depressive disorders, atypical antipsychotic drug use was associated with a higher risk of dementia incidence. A dose-response relationship was also discovered: a higher cumulative dose of atypical antipsychotic drugs exhibited a higher risk of dementia. When the effects of using specific antipsychotic medications were analyzed, results varied according to antipsychotic agents. While risperidone use was associated with the highest risk of dementia and evident dose-response relationship, quetiapine and olanzapine usage was associated with a relatively low risk of dementia and an insignificant dose-response relationship. These results support the hypothesis of this study that atypical antipsychotic augmentation therapy in elderly patients with mood disorders will affect the risk of subsequent dementia, and that the risk will differ according to specific antipsychotic agents.

To the authors’ best knowledge, this study was the first to examine the effect of atypical antipsychotic drugs on the risk of dementia in patients with depressive disorders. Several studies have examined the association between antidepressants and subsequent dementia in patients with depression. Clinical studies have suggested long-term treatment with antidepressants can have a potential protective effect on the risk of dementia [16 , 26]. This was also supported by evidence from animal studies which showed promising effects on amyloid pathology [12, 27]. In contrast, research exploring the effect of antipsychotic drugs on cognitive impairment in non-schizophrenic patients is scarce. A meta-analysis of randomized placebo-controlled trials of atypical antipsychotic drug use for dementia warned that there is a deleterious effect on cognition [28]. In a recent 36-week longitudinal study of 421 individuals diagnosed with Alzheimer’s disease, atypical antipsychotic medication use was associated with accelerated cognitive decline represented by a difference of 2.46 points for the Mini-Mental State Examination (MMSE), compared with the placebo group [29]. Although the study populations were different, the results of the current study are in line with this previous study in that the use of antipsychotic medications was associated with cognitive decline. Furthermore, to enhance the reliability of the current results, adjustments were made for various covariates, including baseline characteristics, comorbidities, and medications. Furthermore, the increased risk of dementia in relation to use of any of antipsychotic drugs, and risperidone specifically, was found to be robust in a sensitivity analysis using various lag periods.

One plausible explanation for the current results is that severity of depression might be a confounding factor. Both major depressive disorder and bipolar disorder are related to a higher risk of incident dementia, with increased risk correlated with more severe and frequent episodes [5 , 30]. As atypical antipsychotic drugs are used to augment the treatment of mood disorders, supplementing a medication regimen with atypical antipsychotic drugs might imply that the patient has more severe or intractable symptoms. In this case, the higher risk of dementia in the atypical antipsychotic drug treatment group may be due to the higher severity of depressive episodes, rather than the use of antipsychotic drugs. This study conducted a subgroup analysis for patients with or without psychotic features, as episodes with psychotic features tend to reveal a higher level of severity [31]. The subgroup analysis yielded significantly higher risks in both groups, with an insignificant interaction between groups. Although the presence of psychotic features did not capture every aspect of the severity, severity alone seems to be insufficient in explaining the increased risk of dementia.

There is a possibility that atypical antipsychotic drug treatment might have affected the underlying mechanisms of dementia. Several studies presented possible mechanisms by which antipsychotic drugs might affect the brain in a manner that results in cognitive impairment. A study has shown that patients with bipolar disorder who were prescribed antipsychotic drugs had a higher CSF/serum albumin ratio than those with no history of antipsychotic drug use [32]. This result implies that antipsychotic drugs are associated with blood-cerebrospinal fluid barrier dysfunction. In another study of individuals with intractable depression, increased amyloid-β (Aβ) and total tau burden pathology were found in the antipsychotic drug treatment group [33]. Additionally, there has been speculation that using drugs with anticholinergic activity can lead to increased dementia risk via increased Aβ burden [20]. Moreover, a recent cross-sectional study for patients with bipolar disorder confirmed that the use of atypical antipsychotic drugs was associated with a lower brain volume [34]. The authors would like to mention that since this current study did not focus on exact mechanisms, further studies are warranted to elucidate underlying mechanisms.

It is noteworthy that this study showed a different pattern of the dementia risk for risperidone usage, compared with the use of other atypical antipsychotic drugs, such as olanzapine or quetiapine. Not only did risperidone users present with the highest risk of incident dementia but they were also the only group that revealed a statistically significant dose-response relationship. A recent literature review proposed that while second-generation antipsychotic medications might have a neuroprotective effect to some extent, the neuroprotective effect of risperidone diminishes at higher doses [35]. Moreover, when risperidone was used to treat neuropsychiatric symptoms of dementia, Aβ levels of cerebrospinal fluid were found to decrease, indicating increased amyloid pathology [36]. Considering olanzapine and quetiapine are known to have a higher tendency to cause metabolic side effects compared to risperidone [37], these results indicate that metabolic side effect profiles according to drug class contributed little to dementia in this study. Compared to users of olanzapine or quetiapine, risperidone users are more susceptible to hyperprolactinemia [38]. Hyperprolactinemia, through increased platelet aggregation and orthostatic hypotension, through direct cardiovascular effects, have been proposed as potential mechanisms of ischemic strokes [39] or cerebrovascular adverse events associated with antipsychotic medication usage [40]. This increased cerebrovascular risk may be partially responsible for incident dementia [41]. The authors would also like to call attention to the possibility of a selection bias due to the observational nature of the study; risperidone may have been prescribed to individuals more prone to a higher risk of dementia.

One of the strengths of this study is that it was the first to investigate the association of the use of antipsychotic medications in elderly depressive disorder patients with subsequent dementia risk. Additionally, the nationwide population-based nature of the study, combined with a large number of participants, adds reliability to the results. Another strength is that the results are not affected by the recall bias, since the study was based on administrative data. Furthermore, various lag periods were applied to assure that reverse causality does not explain the results. Still, while the sensitivity analyses in the present study did not yield a different result, the lag periods might not have been long enough to exclude cases of pre-existing dementia, especially considering the insidious course of dementia.

However, there are several limitations to the study. First, no adjustment could be made for possible confounding factors, including years of education, APOE allele status, and lifestyle factors, since that information was absent from the data. However, other covariates were included, such as health insurance premiums or various comorbid diseases. Second, although the study included a large number of participants, individuals included in each of the TSDD groups or subgroups under subgroup analysis were small (e.g., the number of patients diagnosed with depression with psychotic features was 768, and less than 100 participants were medicated with aripiprazole). Therefore, statistical significances of those results were somewhat uncertain with broad confidence intervals. Third, exposure was defined based on the prescription data, and exposure measurement error due to non-compliance cannot be excluded. Fourth, data lacked information on cognitive scores in neuropsychological tests, thus necessitating the operational criteria for the outcome to be dependent upon diagnoses and prescribed medications. Although since NHIS mandates doctors to enter scores for clinical scales, such as Clinical Dementia Rating or MMSE, when prescribing cognitive-enhancing medications for dementia, it limits the possibility of misdiagnosis. Fifth, due to the nature of the administrative data based retrospective study, certain information on characteristics of the depression were insufficient. A previous study reported that late-onset depression was more linked to dementia than early-onset depression [10]. However, we could not include participant age at depression onset in our analysis. Other characteristics that could not be adjusted include the total number of depressive episodes or longest episode duration, which are also known to influence dementia risk [6]. Another factor concerns the reasons why atypical antipsychotic medications were prescribed. Although, in subgroup analyses, we found that psychotic features did not affect the association between use of antipsychotic drugs and subsequent dementia, adjustments could not be made for other possible reasons, such as treatment-resistant features or agitation. Sixth, information on whether participants had bipolar depression or comorbid psychiatric disorders was unavailable because of the privacy policy of the NHIS. Therefore, we could not confirm that we excluded the patients with bipolar disorder. Adjustments were made using a history of mood stabilizer use, and a subgroup analysis was conducted between mood stabilizer users and non-users, yielding consistent results that support the results of this study. However, there is frequent off-label use of mood stabilizers for the treatment of depressive disorder. Thus, prescription history could not be a sufficient criterion for excluding bipolar disorder. Lastly, this analysis was conducted in elderly individuals with depressive disorders; therefore, there are constraints on generalizability to younger age groups or patients with other psychiatric disorders.

Along with antidepressants and mood stabilizers, atypical antipsychotic medications constitute a major pharmacological treatment option for mood disorders. As their efficacy has been shown for depressive episodes with mixed features [42, 43] and both bipolar mania and depression [17], more atypical antipsychotic drugs are prescribed for mood disorders, including depression without psychotic features [44]. However, their side effects in patients with mood disorders have also been reported, such as weight gain, metabolic side effects [19], and even brain volume loss [34]. The findings of this study raise concerns regarding the long-term use of atypical antipsychotic drugs in older depressive disorder patients. However, the results of this study should be viewed in light of several limitations, especially the study’s retrospective observational design which could complicate the interpretation of the results. Compared to a randomized clinical trial, the design of the current study does not provide explicit information on whether any specific antipsychotic medications should be used or not. Therefore, conclusions cannot be drawn based on the results of our study that the use of antipsychotic medications for mood disorder patients should be avoided. However, it should also be noted that conducting a randomized clinical trial to confirm long-term medication effects would be very challenging. Future studies, including more precisely designed cohort studies and randomized clinical trials, should be conducted to gain evidence on the long-term benefits and risks of using atypical antipsychotic medications in the treatment of mood disorders.

In conclusion, this population-based retrospective cohort study of elderly individuals presents findings that indicate for the first time that the use of atypical antipsychotic medications for depressive disorders may be a possible modifiable risk factor of dementia. Healthcare professionals must be aware of these results, as atypical antipsychotic medications are commonly prescribed for depressive disorders, and subsequent dementia can greatly affect the well-being of patients and their caregivers. These results should be viewed along with several limitations, and further studies are warranted to confirm the results and reveal potential underlying mechanisms.

Footnotes

ACKNOWLEDGMENTS

This work was supported by the National Research Foundation of Korea (NRF) of Korea Grant funded by the Ministry of Science and ICT (Information and Communication Technologies), South Korea [grant number NRF-2018R1C1B6001708; W Myung], and by the Basic Science Research Program through the NRF of Korea, funded by the Ministry of Education, South Korea [grant number NRF-2018R1D1A1B07049034; H Lee]. This research was also supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number: HI19C0416; W Myung). This work was also supported by the National Research Foundation of Korea Grant funded by the Korean Government (NRF-2012S1A6A3A01033504; W Myung). This work was also supported by ALVOGEN KOREA. This study used NHIS data (NHIS-2019-2-275) from the Korean NHIS. The funding source had no involvement in the study design; the collection, analysis, and interpretation of data; in the writing of the paper; and in the decision to submit the article for publication.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Prince

, Bryce

, Albanese

, Wimo

, Ribeiro

, Ferri

(2013) The global prevalence of dementia: A systematic review and metaanalysis. Alzheimers Dement 9, 63–75 e62.

Wimo

, Jonsson

, Gustavsson

, McDaid

, Ersek

, Georges

, Gulacsi

, Karpati

, Kenigsberg

, Valtonen

(2011) The economic impact of dementia in Europe in 2008-cost estimates from the Eurocode project. Int J Geriatr Psychiatry 26, 825–832.

Baumgart

, Snyder

, Carrillo

, Fazio

, Kim

, Johns

(2015) Summary of the evidence on modifiable risk factors for cognitive decline and dementia: A population-based perspective. Alzheimers Dement 11, 718–726.

Diniz

, Butters

, Albert

, Dew

, Reynolds

(2018) Late-life depression and risk of vascular dementia and Alzheimer’s disease: Systematic review and meta-analysis of community-based cohort studies. Br J Psychiatry 202, 329–335.

Wilson

, Barnes

, De Leon

, Aggarwal

, Schneider

, Bach

, Pilat

, Beckett

, Arnold

, Evans

(2002) Depressive symptoms, cognitive decline, and risk of AD in older persons. Neurology 59, 364–370.

Kessing

, Andersen

(2004) Does the risk of developing dementia increase with the number of episodes in patients with depressive disorder and in patients with bipolar disorder? J Neurol Neurosurg Psychiatry 75, 1662–1666.

Livingston

, Huntley

, Sommerlad

, Ames

, Ballard

, Banerjee

, Brayne

, Burns

, Cohen-Mansfield

, Cooper

, Costafreda

, Dias

, Fox

, Gitlin

, Howard

, Kales

, Kivimäki

, Larson

, Ogunniyi

, Orgeta

, Ritchie

, Rockwood

, Sampson

, Samus

, Schneider

, Selbæk

, Teri

, Mukadam

(2020) Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet 396, 413–446.

Holmquist

, Nordstrom

(2020) The association of depression with subsequent dementia diagnosis: A Swedish nationwide cohort study from 1964 to 2016. PLoS Med 17, e1003016.

Green

, Cupples

, Kurz

, Auerbach

, Go

, Sadovnick

, Duara

, Kukull

, Chui

, Edeki

, Griffith

, Friedland

, Bachman

, Farrer

(2003) Depression as a risk factor for Alzheimer disease: The MIRAGE Study. Arch Neurol 60, 753–759.

10.

Singh-Manoux

, Dugravot

, Fournier

, Abell

, Ebmeier

, Kivimaki

, Sabia

(2017) Trajectories of depressive symptoms before diagnosis of dementia: A 28-year follow-up study. JAMA Psychiatry 74, 712–718.

11.

Butters

, Young

, Lopez

, Aizenstein

, Mulsant

, Reynolds

III CF

, DeKosky

, Becker

(2008) Pathways linking late-life depression to persistent cognitive impairment and dementia. Dialogues Clin Neurosci 10, 345.

12.

Cirrito

, Disabato

, Restivo

, Verges

, Goebel

, Sathyan

, Hayreh

, D’Angelo

, Benzinger

, Yoon

, Kim

, Morris

, Mintun

, Sheline

(2011) Serotonin signaling is associated with lower amyloid-beta levels and plaques in transgenic mice and humans. Proc Natl Acad Sci U S A 108, 14968–14973.

13.

Silva

, Mesquita

, Bessa

, Sousa

, Sotiropoulos

, Leao

, Almeida

, Sousa

(2008) Lithium blocks stress-induced changes in depressive-like behavior and hippocampal cell fate: The role of glycogen-synthase-kinase-3beta. Neuroscience 152, 656–669.

14.

Kessing

, Sondergard

, Forman

, Andersen

(2009) Antidepressants and dementia. J Affect Disord 117, 24–29.

15.

Nunes

, Forlenza

, Gattaz

(2007) Lithium and risk for Alzheimer’s disease in elderly patients with bipolar disorder. Br J Psychiatry 190, 359–360.

16.

Bartels

, Wagner

, Wolfsgruber

, Ehrenreich

, Schneider

, Alzheimer’s Disease Neuroimaging Initiative (2018) Impact of SSRI therapy on risk of conversion from mild cognitive impairment to Alzheimer’s dementia in individuals with previous depression. Am J Psychiatry 175, 232–241.

17.

Derry

, Moore

(2007) Atypical antipsychotics in bipolar disorder: Systematic review of randomised trials. BMC Psychiatry 7, 40.

18.

Zhou

, Keitner

, Qin

, Ravindran

, Bauer

, Del Giovane

, Zhao

, Liu

, Fang

, Zhang

, Xie

(2015) Atypical antipsychotic augmentation for treatment-resistant depression: A systematic review and network meta-analysis. Int J Neuropsychopharmacol 18, pyv060.

19.

Spielmans

, Berman

, Linardatos

, Rosenlicht

, Perry

, Tsai

(2013) Adjunctive atypical antipsychotic treatment for major depressive disorder: A meta-analysis of depression, quality of life, and safety outcomes. PLoS Med 10, e1001403.

20.

Gray

, Anderson

, Dublin

, Hanlon

, Hubbard

, Walker

, Yu

, Crane

, Larson

(2015) Cumulative use of strong anticholinergics and incident dementia. JAMA Intern Med 175, 401–407.

21.

Angst

, Gamma

, Gerber-Werder

, Zarate

Jr. , Manji

(2007) Does long-term medication with lithium, clozapine or antidepressants prevent or attenuate dementia in bipolar and depressed patients? Int J Psychiatry Clin Pract 11, 2–8.

22.

Lin

, Chung

, Chen

, Chi

(2018) Increased risk of dementia in patients with Schizophrenia: A population-based cohort study in Taiwan. Eur Psychiatry 53, 7–16.

23.

Mawanda

, Wallace

, McCoy

, Abrams

(2017) PTSD, psychotropic medication use, and the risk of dementia among US Veterans: A retrospective cohort study. J Am Geriatr Soc 65, 1043–1050.

24.

Kim

, Kim

, Yoon

, Won

, Ha

, Cho

, Park

, Bae

, Lee

, Park

, Lee

, Jeong

, Kang

(2019) Cohort Profile: National health insurance service-senior (NHIS-senior) cohort in Korea. BMJ Open 9, e024344.

25.

Bartels

, Belz

, Vogelgsang

, Hessmann

, Bohlken

, Wiltfang

, Kostev

(2020) To be continued? Long-term treatment effects of antidepressant drug classes and individual antidepressants on the risk of developing dementia: A German case-control study. J Clin Psychiatry 81, 19m13205.

26.

Brodrick

, Mathys

(2016) Antidepressant exposure and risk of dementia in older adults with major depressive disorder. J Am Geriatr Soc 64, 2517–2521.

27.

Sheline

, West

, Yarasheski

, Swarm

, Jasielec

, Fisher

, Ficker

, Yan

, Xiong

, Frederiksen

, Grzelak

, Chott

, Bateman

, Morris

, Mintun

, Lee

J-M

, Cirrito

(2014) An antidepressant decreases CSF Aβ production in healthy individuals and in transgenic AD mice. Sci Transl Med 6, 236re234.

28.

Schneider

, Dagerman

, Insel

(2006) Efficacy and adverse effects of atypical antipsychotics for dementia: Meta-analysis of randomized, placebo-controlled trials. Am J Geriatr Psychiatry 14, 191–210.

29.

Vigen

, Mack

, Keefe

, Sano

, Sultzer

, Stroup

, Dagerman

, Hsiao

, Lebowitz

, Lyketsos

, Tariot

, Zheng

, Schneider

(2011) Cognitive effects of atypical antipsychotic medications in patients with Alzheimer’s disease: Outcomes from CATIE-AD. Am J Psychiatry 168, 831–839.

30.

Chen

, Hu

, Wei

, Qin

, McCracken

, Copeland

(2008) Severity of depression and risk for subsequent dementia: Cohort studies in China and the UK. Br J Psychiatry 193, 373–377.

31.

Gaudiano

, Young

, Chelminski

, Zimmerman

(2008) Depressive symptom profiles and severity patterns in outpatients with psychotic vs nonpsychotic major depression. Compr Psychiatry 49, 421–429.

32.

Zetterberg

, Jakobsson

, Redsater

, Andreasson

, Palsson

, Ekman

, Sellgren

, Johansson

, Blennow

, Landen

(2014) Blood-cerebrospinal fluid barrier dysfunction in patients with bipolar disorder in relation to antipsychotic treatment. Psychiatry Res 217, 143–146.

33.

Clarke

, Hartmann

, Jones

, Ballard

, Francis

(2011) Antipsychotic medication is associated with selective alterations in ventricular cerebrospinal fluid Abeta 40 and tau in patients with intractable unipolar depression. Int J Geriatr Psychiatry 26, 1283–1291.

34.

Birner

, Bengesser

, Seiler

, Dalkner

, Queissner

, Platzer

, Fellendorf

, Hamm

, Maget

, Pilz

, Lenger

, Reininghaus

, Pirpamer

, Ropele

, Hinteregger

, Magyar

, Deutschmann

, Enzinger

, Kapfhammer

, Reininghaus

(2020) Total gray matter volume is reduced in individuals with bipolar disorder currently treated with atypical antipsychotics. J Affect Disord 260, 722–727.

35.

Chen

, Nasrallah

(2019) Neuroprotective effects of the second generation antipsychotics. Schizophr Res 208, 1–7.

36.

Bloniecki

, Aarsland

, Blennow

, Cummings

, Falahati

, Winblad

, Freund-Levi

(2017) Effects of risperidone and galantamine treatment on Alzheimer’s disease biomarker levels in cerebrospinal fluid. J Alzheimers Dis 57, 387–393.

37.

Gianfrancesco

, Grogg

, Mahmoud

, Wang

R-H

, Meletiche

(2003) Differential effects of antipsychotic agents on the risk of development of type 2 diabetes mellitus in patients with mood disorders. Clin Ther 25, 1150–1171.

38.

Leucht

, Cipriani

, Spineli

, Mavridis

, Örey

, Richter

, Samara

, Barbui

, Engel

, Geddes

, Kissling

, Stapf

, Lässig

, Salanti

, Davis

(2013) Comparative efficacy and tolerability of 15 antipsychotic drugs in schizophrenia: A multiple-treatments meta-analysis. Lancet 382, 951–962.

39.

Shin

, Choi

, Jung

, Lee

, Kwon

, Park

(2013) Risk of ischemic stroke with the use of risperidone, quetiapine and olanzapine in elderly patients: A population-based, case-crossover study. J Psychopharmacol 27, 638–644.

40.

Mittal

, Kurup

, Williamson

, Muralee

, Tampi

(2011) Risk of cerebrovascular adverse events and death in elderly patients with dementia when treated with antipsychotic medications: A literature review of evidence. Am J Alzheimers Dis Other Demen 26, 10–28.

41.

Desmond

, Moroney

, Sano

, Stern

(2002) Incidence of dementia after ischemic stroke: Results of a longitudinal study. Stroke 33, 2254–2260.

42.

Suppes

, Silva

, Cucchiaro

, Mao

, Targum

, Streicher

, Pikalov

, Loebel

(2016) Lurasidone for the treatment of major depressive disorder with mixed features: A randomized, double-blind, placebo-controlled study. Am J Psychiatry 173, 400–407.

43.

Patkar

, Pae

, Vohringer

, Mauer

, Narasimhan

, Dalley

, Loebel

, Masand

, Ghaemi

(2015) A 13-week, randomized double-blind, placebo-controlled, cross-over trial of ziprasidone in bipolar spectrum disorder. J Clin Psychopharmacol 35, 319–323.

44.

Gerhard

, Akincigil

, Correll

, Foglio

, Crystal

, Olfson

(2014) National trends in second-generation antipsychotic augmentation for nonpsychotic depression. J Clin Psychiatry 75, 490–497.