Association between Neuropsychiatric Improvement and Neurocognitive Change in Alzheimer’s Disease: Analysis of the CATIE-AD Study

Abstract

Background/Objective:

To assess associations between improvements in neuropsychiatric symptoms (NPS) and neurocognitive change in patients with Alzheimer’s disease (AD) during treatment using the Clinical Antipsychotic Trials of Intervention Effectiveness–Alzheimer Disease (CATIE-AD) dataset.

Methods:

AD outpatients with NPS who needed pharmacological treatment (n = 421) were followed up with antipsychotics, citalopram, or placebo for up to 36 weeks (mean±SD = 252±52 days). The study aim was to investigate associations between improvement in each NPS evaluating scale (by Clinical Global Impression of Change [CGI-C], Neuropsychiatric Inventory [NPI], or Brief Psychiatric Scale [BPRS]) at endpoint (week 36 or early termination [ET], n = 340) and neurocognitive change (change score in the Mini-Mental State Examination [MMSE] between endpoint and baseline during the treatment). Multiple logistic regression analyses were performed on the associations between each NPS improvement and neurocognitive change as well as socio-clinico-demographic variables of interest.

Results:

At endpoint, NPS improvement rates were 76.1%, 70.8%, and 58.1% in CGI-C, NPI, and BPRS, respectively, while MMSE score change was –2.3±3.8. NPS improvement was significantly related to more severe psychotic symptoms at baseline and preserved levels of neurocognition (smaller MMSE score change) among several variables.

Conclusions:

Our findings suggested that neurocognitive preservation may be associated with attaining optimal benefits from any treatment against NPSs in a longitudinal treatment course of patients with AD.

Keywords

Alzheimer’s disease antipsychotic CATIE-AD neurocognitive impairment neuropsychiatric symptom

INTRODUCTION

Alzheimer’s disease (AD) is a neurodegenerative disease characterized by neurocognitive impairment as core symptoms, which progressively impairs activities of daily living (ADL) including self-care, housework, and social activity [1]. As conventional marginal symptoms, neuropsychiatric symptoms (NPS), including psychosis, aggressiveness, depression, and apathy appear in AD patients, which in turn increases mental and physical distress of their caregivers and accelerates their institutionalization in the long-term course [2]. Relationships between neurocognition and NPS are complicated in patients with AD [3 –5].

The Clinical Antipsychotic Trials of Intervention Effectiveness–Alzheimer’s Disease (CATIE-AD) was a 36-week double-blind placebo-controlled study for to determine the longitudinal effectiveness of atypical Antipsychotics (AAP) on psychotic or aggressive symptoms in 421 patients with AD [6]. The trial demonstrated that the proportion of treatment improvement and discontinuation were not different between patients’ initial intake of AAP and placebo [6]. By re-analyzing the trial dataset, we previously reported that lower neurocognitive level at baseline was among risk factors for psychosis in the cross-sectional study [7], which is consistent with other studies concluding that more rapid cognitive declines were associated with development of psychosis/aggression in course of AD patients [8, 9]. We also found that treatment response at week 8 was predicted by lower neurocognitive status at baseline [10]. However, to our knowledge, no study has yet investigated associations between changes in NPS and neurocognitive function in patients with AD during longitudinal treatment.

Thus, the present study sought to determine associations between neurocognitive change and NPS improvement in AD patients during 36-week treatment, using CATIE-AD dataset, which may resolve the complex relationship.

METHODS

Study design and participants

Previously, study design of the CATIE-AD was detailed in a protocol article [11], and its dataset was used in the present study as follows. Briefly, four hundred and twenty-one AD outpatients with psychosis or agitation, who were ambulatory and living at home or assigned facilities, were enrolled in the CATIE-AD study (Table 1). Regarding psychiatric disorders, patients with drug abuse, delirium, endogenic psychiatric disorders (e.g., schizophrenia or mood disorder), and other dementias (e.g., dementia with Lewy body and vascular dementia) were excluded. Also, patients were excluded if they were taking benefit from any psychotropic before the trial. Severity of psychotic symptoms or aggressive behaviors in eligible patients with AD was very serious, such that they caused trouble in everyday life and patients needed antipsychotic treatment as a result. Patients were randomly assigned to three AAPs (olanzapine, quetiapine, risperidone) or placebo in phase 1 (Table 1). When a new medication would be considered more beneficial, patients proceeded to a next phase. From phase 2, citalopram was added as a randomly assigned treatment medication. Phase 3 was an open-label randomized study where patients were assigned to one of the medications which they did not previously receive. After patients showed insufficient response to an assigned medication, they could directly enter the open-choice phase from any of the three phases within 36 weeks. If patients terminated the trial for some reason before week 36 week, that was defined as ‘Early Termination’ (ET). In this analysis, we used the dataset which comprised the data of patients whose symptoms were assessed from baseline to week 36 or ET (i.e., endpoint). If patients discontinued the trial due to any reasons and did not match the definition of ET, they were excluded from this analysis. The informants/caregivers had lived together or visited the patients for at least 8 hours per week over 3 days per week. Concerning a reward of study entry, participants and their caregivers were performed all evaluation with no cost to maintain fairness of the study.

Table 1

Demographic data of subjects at baseline and week-36/early termination (ET)

	Baseline (n = 421)	Week-36/ET (n = 340)
Continuous variables	Mean±SD (range)
Age, y (n = 421)	77.9±7.5 (range: 51–103)
Education, y (n = 421)	12.3±3.4 (range: 0–21)
MMSE score (n = 416⟶297)	15.0±5.8 (range: 4–29)	13.2±6.6 (range: 0–30)
MMSE score change		–2.3±3.8 (range: –14–9)
ADCS score (n = 413⟶316)	39.0±17.2 (range: 2–76)	31.0±18.3 (range: 0–71)
GMHR score (n = 420⟶323)	3.3±0.7 (range: 2–4)	3.2±0.8 (range: 1–4)
CGI-C score at w-36/ET (n = 330)		2.8±1.4 (range: 1–7)
NPI score (n = 414⟶321)	36.9±18.3 (range: 3–104)	19.3±16.0 (range: 0–96)
BPRS score (n = 419⟶322)	27.8±12.3 (range: 3–72)	19.8±11.4 (range: 0–67)
Burden interview (n = 409⟶304)	34.4±16.0 (range: 0–76)	30.5±15.2 (range: 0–82)
Duration of study; days (n = 340)		252±52 (range: 4–407)
Categorical Variables	n (%)
Gender, Female/Male	235 (55.8) /186 (44.2)
Race, Caucasian/non-Caucasian	331(79) / 88(21)
Marital status, married/not-married	249 (59.1) / 172 (40.9)
Residence, own home/not-own home	307 (72.9) / 114 (27.1)
Care service, receiving/not-receiving	161 (38.5) / 257 (61.5)
Diabetes Mellitus, Yes/No	59 (14) / 362 (86)
Hypertension, Yes/No	219 (52) / 202 (48)
Cardiac disorders, Yes/No	118 (28) / 303 (72)
Cerebrovascular accidents, Yes/No	43 (10.2) / 378 (89.8)
Anticholinesterases, taking/not-taking	205 (48.7) / 216 (51.3)
Psychotropics, taking/not-taking	136 (32.3) / 285 (67.7)
Kind of the AAP or Placebo in phase 1
Olanzapine, taking/not-taking	100 (23.8) / 321 (76.2)
Quetiapine, taking/not-taking	94 (22.3) / 327(77.7)
Risperidone, taking/not-taking	85 (20.2) / 336 (79.8)
Placebo, taking/not-taking	142 (33.7) / 279 (66.3)
CGI-C score, improver/non-improver (Improver was 0–3 point)		251 (76.1) / 79 (23.9)
NPI score improver, improver/non-improver (Improver was > = 8 points reduction)		225 (70.8) / 93 (29.2)
BPRS score, improver/non-improver
(Improver was > = 25% reduction of total score)		186 (58.1) / 134 (41.9)

MMSE, Mini-Mental State Examination; MMSE score change, (MMSE score at week-36/ET)-(MMSE score at BL); ADCS, Alzheimer’s Disease Cooperative Study; GMHR, General Medical Health Rating; CGI-C at w-36/ET, Clinical Global Impressions Rating at week-36/ET; NPI, Neuropsychiatric Inventory; BPRS, Brief Psychiatric Rating Scale; Anticholinesterases, donepezil (n = 151), galantamine (n = 26), rivastigmine (n = 26), combination therapy (donepezil and memantine [n = 2]); Psychotropics, include antipsychotics, antidepressant, anxiolytics, hypnotics/sedatives, or antiepileptics; AAP, atypical antipsychotic.

Outcome measures

The present study used three representative scales to evaluate the NPS of patients with AD based on previous studies [12 –14]. The Clinical Global Impression of Change (CGI-C, scores range from 1 to 7, where a higher score reflects a worse impression of change by clinicians) was utilized to evaluate patients’ clinical alteration, as determined by the clinician’s impression [15]. The severity of general NPSs was evaluated by the Neuropsychiatric Inventory (NPI, scores range from 0 to 120, that a higher score reflects a severer neuropsychiatric condition) [12]. The severity of psychotic and behavioral symptoms was measured with the Brief Psychiatric Scale (BPRS, scores range from 0 to 108, where a higher score reflects a severer psychotic level) [14]. While the CGI-C score was used only at endpoint, NPI and BPRS scores were measured throughout the study.

Neurocognition, ADL, and general medical health state measures

Subjects’ global neurocognitive function was measured with the Mini-Mental State Examination (MMSE, scores range from 0 to 30, where a higher score reflects a higher neurocognitive level) [16]. ADL were assessed with the Alzheimer Disease Cooperative Study (ADCS, scores range from 0 to 78, where a higher score reflects a higher ADL level) [17]. General medical health state was evaluated according to the General Medical Health Rating (GMHR, scores range from 1 to 4, where a higher score reflects a healthier state) [18]. Caregivers’ burden or distress was evaluated with the Burden Interview (scores range from 0 to 88, where a higher score reflects a severer caregiver burden) [19]. The present study analyzed these scores measured at baseline and endpoint. According to previous longitudinal studies with large samples, scores in MMSE gradually reduce in spite of antidementia drug intake in patients with AD [20 –22]. Thus, we defined a ‘neurocognitive change’ as follows: (MMSE score at endpoint) – (MMSE score at baseline).

Definition of NP symptom improvement

In the present study, we investigated treatment improvement in the aforementioned three outcome measures. The treatment improvement in the present study means a conventional ‘minimum improvement’ in a longitudinal treatment course [10 , 23]. Based on confirmed consensus gained from previous studies, three definitions of treatment improvement included 1) from 1 (very much improved) to 3 (minimally improved) in the CGI-C score at endpoint that is so called ‘minimum improvement’ [10]; 2) reduction of 8 points that is 0.4×(standard deviation [SD = 19.7] of NPI score change between baseline and endpoint in the present study) or more in the NPI total score from baseline, or 0 point at endpoint when their baseline NPI total score was less than 8 points that is so called ‘minimum clinically important differences’ in a previous trial [23]; and 3) reduction rate of 25% or higher in the BPRS total score from baseline that is so called ‘minimum improvement’ [14].

Other socio-clinico-demographic characteristics as independent variables of interest

Based on previous studies, we examined other socio-clinico-demographic variables which potentially influenced treatment outcomes [7, 10]. These variables are as follows: continuous variables included age (years), duration of education (years), baseline’s scores on the MMSE, MMSE score change, BPRS, ADCS, GMHR, and Burden Interview, and duration of study (number of days from baseline to the date of study completion). Categorical variables included sex, race (Caucasian or non-Caucasian), present marital status (married or not), residence (own home or not), care service (receiving or not), medical history (diabetes mellitus, hypertension, any cardiac disorders [any coronary artery disease, aortic or mitral valve disease, any arrhythmia, cardiac failure congestive, and inflammatory heart disease], and cerebrovascular accidents), intake of medications within 14 days prior to baseline (any anticholinesterase and any psychotropics), and intake of trial medications (olanzapine, quetiapine, risperidone, placebo, or citalopram) in phase 1 and endpoint. Trial medicine (olanzapine: 2.5 mg or 5.0 mg, quetiapine: 25 mg or 50 mg, risperidone: 0.5 mg or 1.0 mg, or placebo) was randomly assigned to patients in phase 1.

Statistical analysis

As primary investigations, to compare the score change of MMSE between NPS (based on CGI-C, NPI, or BPRS) improvers and non-improvers, t-tests were used. Next, we investigated potential associations among socio-clinico-demographic factors and NPS improvement in the rating scales which we found was significantly related to neurocognitive change by the aforementioned primary analyses. To compare NPS improvers and non-improvers, t-tests were used when the factors were continuous variables (i.e., age year, education years, baseline scores on the MMSE, BPRS, ADCS, GMHR, and Burden Interview, and study duration), and χ² tests were used when the factors were categorical variables (sex, race, present marital status, residence, care service, medical history, intake of medication: any anticholinesterase and psychotropics before baseline, and intake of trial medications). Subsequently, multiple logistic regression models were performed to verify associations among each NPS improvement in the rating scales which we found was related to neurocognitive change by the aforementioned primary analyses and the variables found to be significant by the preliminary analyses. Moreover, age, education years, and scores in the MMSE, BPRS, ADCS, and GMHR at baseline were also added as independent variables based on previous reports [7, 10]. The Hosmer-Lemeshow tests were performed for an evaluation of each multiple logistic regression model adjustment or fitting. Discriminative accuracy of the final model was assessed using the area under the receiver operator characteristic curve (AUC). A p value <0.05 was considered statistically significant for preliminary analyses. On the other hand, a Bonferroni-corrected p value (<0.05/3 = 0.016) was considered statistically significant for the primary χ² tests and multivariable regression analyses because of multiple comparisons. IBM SPSS Statistics for Windows, Version 22.0 (Armonk, NY: IBM Corp.) was used for all the statistical analyses.

RESULTS

Patient characteristics

Table 1 depicts that socio-clinico-demographic characteristics at baseline and endpoint, and the patients’ rates of each NPS treatment improver during the about 36 weeks. While 421 patients with AD (235 females [55.8%]; age = 77.9±7.5 years) were included at baseline, 340 patients reached endpoint (among them, 77 patients completed the full 36 weeks), which comprises the dataset for this study. The duration of study was 252±52 days. NPS treatment improvement rates were 76.1%, 70.8%, and 58.1% in CGI-C, NPI, and BPRS scores, respectively. MMSE score change was –2.3±3.8. Table 2 shows that each assigned trial’s medicine in the 340 patients until endpoint.

Table 2

The assigned trial medicine of the 340 patients until endpoint (week 36/ET)

	Phase 1		Phase 2		Phase 3		At endpoint (week 36W/ET)
	number	(%)	number	(%)	number	(%)	number	(%)
Olanzapine	81	23.8	43	20.2	27	25.7	87	25.6
Quetiapine	72	21.2	50	23.5	25	23.8	75	22.1
Risperidone	74	21.8	43	20.2	26	24.8	79	23.2
Placebo	113	33.2	–	–	–	–	36	10.6
Citalopram	–	–	77	36.2	27	25.7	63	18.5
Total	340	100	213	100	105	100	340	100

ET, early termination.

NPI subscores at baseline and endpoint

Each NPI subscore at baseline was as follows: delusions: 5.5±4.0; hallucination: 2.9±3.7; agitation: 5.5±3.6; depression: 2.7±3.2; anxiety: 3.8±3.8; euphoria: 0.4±1.4; apathy: 4.3±3.8; disinhibition: 2.2±3.3; irritability: 5.1±3.9; aberrant motor behavior: 4.5±4.4; sleep/nighttime behavior disorders: 3.4±4.2; appetite/eating disorders: 2.3±3.6. On the other hands, each NPI subscore at endpoint was as follows delusions: 2.0±3.0; hallucination: 1.3±2.5; agitation: 2.5±3.2; depression: 1.2±2.3; anxiety: 2.1±2.9; euphoria: 0.3±1.3; apathy: 3.6±4.0; disinhibition: 1.1±2.3; irritability: 2.6±3.3; aberrant motor behavior: 2.6±3.5; sleep/nighttime behavior disorders: 1.3±2.7; appetite/eating disorders: 1.7±3.1.

Psychotropics and anti-dementia drugs within 14 days prior to baseline

Within 14 days prior to baseline, 136 patients took the following psychotropics: antipsychotics (risperidone [n = 21]; haloperidol [n = 18]; quetiapine [n = 14]; olanzapine [n = 7]; tiotixene [n = 1]), antidepressants (trazodone [n = 29]; sertraline [n = 10]; paroxetine [n = 10]; citalopram [n = 6]; fluoxetine [n = 4]; venlafaxine [n = 3]; amitriptyline [n = 1]; nefazodone [n = 1]), anxiolytics (lorazepam [n = 39]; alprazolam [n = 4]; clonazepam [n = 3]; buspiron [n = 3]; oxazepam [n = 2]; hydroxyzine [n = 2]), hypnotics and sedatives (zolpidem tartrate [n = 5]; temazepam [n = 1]; triazolam [n = 1]) and antiepileptics (valproate semisodium [n = 3]; gabapentin [n = 2]; carbamazepine [n = 1]; phenobarbital [n = 1]). Anti-dementia drugs were administered in 205 patients within 14 days prior to baseline (donepezil [n = 151]; galantamine [n = 26]; rivastigmine [n = 26]; combination therapy: donepezil and memantine [n = 2]).

The comparison of MMSE score change between improver and non-improver of NPS

MMSE change was significantly different between improvers and non-improvers in NPS (Table 3). Especially, between improver and non-improver of CGI-C and BPRS scores, MMSE changes were significantly different with using Bonferroni correction (Table 3).

Table 3

Comparisons of MMSE score change^a between each treatment improvement

	CGI-C score Improver (Minimum Improvement: 0–3)		NPI score Improver (> = 8 points reduction of total score)		BPRS score Improver (> = 25% reduction of total score)
	Improver (n = 229)	Non-Improver (n = 66)	Improver (n = 207)	Non-Improver (n = 84)	Improver (n = 171)	Non-Improver (n = 122)
MMSE score change	–1.9±3.8	–3.4±3.7	–2.0±3.8	–3.1±3.7	–1.7±3.7	–3.1±3.9
	t- score = 2.7, p = 0.008^**¶		t-score = 2.3, p = 0.024^*		t- score = 3.1, p = 0.002^**¶

CGI-C, Clinical Global Impressions Rating of Change; NPI, Neuropsychiatric Inventory; BPRS, Brief Psychiatric Rating Scale; MMSE, Mini-Mental State Examination. Each association was analyzed using t-test. ^*p < 0.05, ^**p < 0.01, ^¶p<0.016 = 0.05/3 (Bonferroni correction in CGI-C or BPRS improver). Values in bold font are significant results at each Bonferroni correction. ^aMMSE score change, (MMSE score at week-36/ET)-(MMSE score at BL).

Preliminary analyses of socio-clinico-demographic variables related to NPS improvement

CGI-C improvement was related to receiving dementia-related care service (Table 4). BPRS improvement was associated with higher BPRS score at baseline and living in not-own home (Table 4).

Table 4

Table4

	CGI-C score Improver (Minimum Improvement: 0–3)				BPRS score Improver (> = 25% reduction of total score)
Variables	Improver	Non-Improver	t-score	p	Improver	Non-Improver	t-score	p
Continuous variables	Mean±SD (range)				Mean±SD (range)
Age, y	78.0±7.6	77.2±6.9	0.86	0.393	78.0±7.0	77.4±7.9	0.71	0.479
Education, y	12.2±3.2	12.4±3.7	–0.35	0.730	12.2±3.5	12.3±3.3	–0.46	0.649
MMSE score of BL	15.2±5.8	14.8±5.6	0.51	0.611	15.2±5.6	14.9±5.9	0.54	0.592
BPRS score of BL	27.1±11.7	27.7±13.6	–0.40	0.690	29.5 ± 11.9	24.0 ± 11.8	4.14	<0.001^***
ADCS score of BL	40.0±16.7	39.1±17.4	0.25	0.805	40.2±16.1	39.1±17.8	0.58	0.565
GMHR score of BL	3.4±0.7	3.4±0.6	0.12	0.902	3.4±0.7	3.4±0.7	0.08	0.939
Burden interview score of BL	33.9±16.0	32.3±15.4	0.76	0.447	33.5±16.5	33.5±15.3	–0.03	0.976
Duration of study; days	260.0±31.1	249.8±62.6	1.38	0.173	259.2±34.5	259.7±43.2	–0.11	0.91
Variables	Improver	Non-Improver	χ² score	p	Improver	Non-Improver	χ² score	p
Categorical variables	n				n
Gender, Female/Male	135/116	46/33	0.48	0.489	106/80	71/63	0.51	0.477
Race, Caucasian/non-Caucasian	191/59	67/12	2.51	0.113	141/45	110/23	2.20	0.138
Marital status, married/not-married	147/104	54/25	2.42	0.120	108/78	86/48	1.22	0.269
Residence, own home/not-own home	178/73	60/19	0.76	0.384	124/62	108/26	7.58	0.006 ^**
Care service, receiving/not-receiving	100/149	20/59	5.70	0.017 ^*	72/113	48/86	0.32	0.573
Diabetes Mellitus, Yes/No	39/212	7/72	2.23	0.135	29/157	17/117	0.53	0.465
Hypertension, Yes/No	121/130	44/35	1.35	0.246	89/97	72/62	1.08	0.299
Cardiac disorders, Yes/No	63/188	28/51	3.22	0.073	51/135	36/98	0.01	0.913
Cerebrovascular accidents, Yes/No	26/225	5/74	1.15	0.284	17/169	14/120	0.15	0.696
^aAnticholinesterases, taking/not-taking at baseline	129/122	44/35	0.45	0.504	94/92	77/57	1.50	0.220
^bPsychotropics, taking/not-taking at baseline	83/168	25/54	0.06	0.814	56/130	45/89	0.44	0.509
Atypical Antipsychotic or Placebo in phase 1
Olanzapine, taking/not-taking	60/191	19/60	0.00	0.979	43/143	35/99	0.38	0.537
Quetiapine, taking/not-taking	55/196	12/67	1.68	0.195	39/147	24/110	0.46	0.497
Risperidone, taking/not-taking	54/197	18/61	0.06	0.811	42/144	28/106	0.13	0.719
Placebo, taking/not-taking	82/169	30/49	0.75	0.385	62/124	47/87	0.11	0.746
Atypical Antipsychotic, Citalopram, or Placebo in endpoint (week 36/ET)
Olanzapine, taking/not-taking	59/192	24/53	1.83	0.176	44/142	37/95	0.78	0.378
Quetiapine, taking/not-taking	55/196	15/62	0.21	0.649	39/147	30/102	0.14	0.708
Risperidone, taking/not-taking	63/188	15/62	1.03	0.311	50/136	27/105	1.74	0.187
Placebo, taking/not-taking	26/225	8/69	0.00	0.994	19/167	14/118	0.01	0.910
Citalopram, taking/not-taking	48/203	15/62	0.01	0.945	34/152	24/108	0.00	0.982

CGI-C, Clinical Global Impressions Rating of Change; BL, baseline; ET, early termination; MMSE, Mini-Mental State Examination; ADCS, Alzheimer’s Disease Cooperative Study; GMHR, General Medical Health Rating; BPRS, Brief Psychiatric Rating Scale. Continuous variables were analyzed using t-test, and categorical variables were analyzed using χ² test. ^aAnticholinesterases: donepezil (n = 151); galantamine (n = 26); rivastigmine (n = 26); combination therapy (donepezil and memantine [n = 2]). ^bPsychotropics include antipsychotics, antidepressant, anxiolytics, hypnotics/sedatives, or antiepileptics. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001.

Conceptual model on the association between neurocognitive change and NPS improvements

Both models’ adjustments were not rejected (p > 0.05 for each outcome by Hosmer-Lemeshow test). Smaller score reduction of MMSE (OR = 1.11, 95% CI = 1.03–1.20, p = 0.010) (AUC = 0.61, 95% CI = 0.54–0.69, p = 0.006) were significantly associated with CGI-C improvement after Bonferroni correction (Table 5). Next, Higher BPRS score at baseline (OR = 1.05, 95% CI = 1.02–1.07, p = 0.001) (AUC = 0.63, 95% CI = 0.57–0.70, p < 0.001) and smaller score reduction of MMSE (OR = 1.10, 95% CI = 1.03–1.18, p = 0.007) (AUC = 0.61, 95% CI = 0.54–0.67, p = 0.002) were significantly related to BPRS improvement after Bonferroni correction (Table 5).

Table 5

Significant variable effect on treatment improver in Multiple Logistic Regression Models

	CGI-C score Improver (Minimum Improvement: 0–3)			BPRS score Improver (> = 25% reduction of total score)
Variables	OR	p	95%CI	OR	p	95%CI
Continuous variables
Age, y	1.03	0.257	0.98–1.07	1.02	0.331	0.98–1.06
Education, y	1.00	0.944	0.91–1.10	1.04	0.384	0.96–1.12
MMSE BL score	1.01	0.765	0.94–1.08	1.01	0.649	0.96–1.08
BPRS BL score	0.99	0.531	0.96–1.02	1.05	0.001 ^**¶	1.02–1.07
ADCS BL score	1.00	0.886	0.97–1.02	1.02	0.157	0.99–1.04
GMHR BL score	1.21	0.448	0.74–1.96	1.12	0.604	0.73–1.70
MMSE score change^a	1.11	0.010^*¶	1.03–1.20	1.10	0.007 ^**¶	1.03–1.18
Categorical variables
Residence (own home)				0.46	0.020^*	0.24–0.89
Care services (receiving)	1.90	0.061	0.97–3.72

BPRS, Brief Psychiatric Rating Scale; BL, Baseline; CGI-C, Clinical Global Impressions of Change; GMHR, General Medical Health Rating; MMSE, Mini-Mental State Examination; OR, odds ratio; CI, Confidence Interval ^*p < 0.05, ^**p < 0.01, ^¶p < 0.016. Values in bold font are significant results at each Bonferroni correction. Age, Education years, and scores in the MMSE, BPRS, ADCS, and GMHR at baseline were added to selected variables in each conceptual regression model. ^aMMSE score change, (MMSE score at week-36/ET)-(MMSE score at BL).

DISCUSSION

The present study was the first to longitudinally investigate the associations between each NPS improvement and neurocognitive change during treatment in patients with AD who needed interventional treatment for their psychotic symptoms or agitation. In the long-term course of approximately 36 weeks, we found that: 1) rates of each NPS improvement were over 50%, and a reduction score in the MMSE reflecting global neurocognitive was about two point per nine months; 2) NPS improvements were associated with the smaller neurocognitive reductions, but not assigned trial medicine (olanzapine, quetiapine, risperidone, placebo, or citalopram); and 3) NPS improvement as per the CGI-C was associated with the smaller neurocognitive reductions while NPS improvement as per the BPRS was linked to greater severity of initial psychotic symptoms and the smaller neurocognitive reductions.

NPS improvement rates in the present study are keeping in line with treatment response according to previous reports [10 , 23]. For example, a previous randomized control trial using citalopram for agitation in patients with AD noted that minimum improvement in CGI-C score at week 9 was about 50% with placebo and 70% with citalopram [24]. While the CATIE-AD study did not show the superiority of AAP over placebo in terms of effectiveness on NPS, NPS assessed with the CGI-C improved in patients regardless of medications or placebo [6]. Pharmacological treatment studies have reported that NPI total score decreased by more than half in patients with dementia in 8 or 12 months. The same longitudinal open-label trial also showed that the antipsychotics responder rate based on BPRS score was 50% per 5 months [25]. These findings are compatible with the result of present study: average NPI score decreased from 36.9 to 19.3 per about 36 weeks (Table 1) [26, 27]. Furthermore, in the present study, those NPS improvements at about week 36 were not influenced by kinds of assigned medicine (AAPs, placebo, or citalopram) (Table 4), which supported a previous study using the same database [6]. Taken together, these findings including short-term trials suggest that NPS may improve during short-term duration in a time-dependent manner, regardless of which pharmacological interventions are conducted. It was reported that citalopram was effective for agitation in patients with AD in a previous study [24]. Citalopram was also used as an assigned medicine from phase 2 of the present study. However, we found no relationship between intake of citalopram and improvement of NPS measured with the CGI-C and BPRS (Table 4). This may at least in part be attributable to the fact that the present study enrolled subjects with psychotic symptoms or aggressive behaviors that needed interventional treatments rather than depression, anxiety, and apathy.

To our knowledge, this was the first study demonstrating the significant associations between neurocognitive change and NPS improvements with any interventions, which was found with the primary analyses and multiple regression analyses. Previous longitudinal studies have reported that more rapid declines of MMSE score was among risk factors of psychosis/aggression onset in patients with AD [8 , 28]. Another study found that the more severe NPS at baseline predicted the faster declines in neurocognition and ADL in patients with AD [29, 30]. These findings support our results of the associations between NPS improvement and neurocognitive change in patients with AD. We speculate that only those who can preserve neurocognition or do not have neurocognitive deterioration may show NPS improvement regardless of treatment intervention modalities [31]. However, the causality between NPS improvement and neurocognitive preservation during any treatment interventions remains still unclear [32, 33]. Of another note, a recent investigation reported that non-pharmacological intervention contributed to NPS improvement without worsening neurocognition in neurocognitively declined people for 6 months, therefore future study is warranted to assess the association between NPS improvement and neurocognitive preservation in a course of non-pharmacological interventions [31].

CGI-C improvement, which was based on physicians’ impression, was associated with smaller neurocognitive reduction. Given that global cognitive impairment predicts later ADL reductions [34], these two factors might be synonym for the physicians’ impression. Our previous study demonstrated that CGI-C response at week 8 was predicted by lower MMSE score at baseline while the present study found that the MMSE score at baseline was not associated with CGI-C improvement at week 36 (Table 4) [10]. Therefore, we speculate that the neurocognitive change (score difference) rather than baseline score may influence treatment outcomes for NPS in the long run. The global rapid neurocognitive alteration including executive dysfunction is considered as a risk factor causing distorted comprehension or dysregulated feelings in this population [35 –37]. Therefore, currently available treatments for NPS may modify the misinterpretation or self-controlling emotion leading to NPS improvement.

BPRS improvement was associated with more severe initial psychotic problems and smaller neurocognitive reduction in patients with AD. To support our finding, previous reports have noted that initial severity of psychotic symptoms was frequently reported as a predictor for AAP effects on psychotic symptoms [10, 38]. This paradoxical result may be corroborated by the so-called ‘law of initial value hypothesis’ that the higher initial value causes the greater organism response in a course of treatment in patients with severe psychopathology [31, 38]. As the CGI-C improvement, in the long run, BPRS improvement was associated with smaller neurocognitive change but not its initial level at baseline in AD, which might support previous studies concluding that more rapid cognitive declines were associated with development of psychosis/aggression in course of AD patients [8, 9]. Taken together, in a course of improving psychotic symptoms in patients with AD, initial psychotic severity and neurocognitive preservation may influence the prognosis, regardless of influences on neurocognitive or ADL status.

The present study has several limitations. The sample size was comparatively small, and the Bonferroni correction was used to avoid the type I errors in the multiple regression analyses. Therefore, we also did not perform a regression analysis on treatment improvement using NPI score. Second, to evaluate neurocognitive status in patients with AD, we used only MMSE scale which is an easily administered test that can be completed at bedside within 10 minutes to evaluate the global cognitive status [16]. Another representative scale, Alzheimer’s Disease Assessment Scale (ADAS), was also performed to evaluate the global cognitive status of patients in the previous study using CATIE-AD database [39]. In the present study, however, 85 patients (25%) did not complete the ADAS among 340 patients at week-36, which did not allow us to analyze the ADAS data statistically. Thirdly, the database does not have the information on the duration of illness in the subjects enrolled in the present study, thus we could not investigate the association between duration of illness and pharmacological effect. Fourthly, the mean difference scores of NPI and BPRS between 36 week and baseline was –17.3±19.7 and –7.3±11.5, respectively, thus the score changes were relatively larger compared with MMSE change scores: –2.3±3.8. Given the score ranges of the NPI and BPRS are greater than that of the MMSE, the score changes in the NPI and BPRS could be easy to be influenced by their initial raw scores [31, 38]. Thus, the two scales (NPI and BPRS) were dichotomized according to their change scores. Finally, the CATIE-AD trial enrolled subjects with psychotic symptoms or aggressive behaviors that needed interventional treatments, which was a clinic-based sample selected under limited condition. Thus, the results of the present study may not be generalizable to all patients with AD.

In conclusion, in the 36-week trial for pharmacological treatment in AD patients with psychosis or aggressiveness symptoms, the present study found: 1) NPS improver rates were more than a half in patients with AD, 2) NPS improvement was associated with preserved neurocognition, and 3) NPS improvement was influenced by initial severity of psychosis, but not neurocognitive status at baseline. As a viewpoint of treatment for NPS in the longitudinal course of patients with AD, our results suggest the importance focusing neurocognitive preservation among several predictors. Moreover, moving forward, to further investigate the relationship between trajectories of neurocognition interacted with other socio-clinico-demographic factors and NPS may help elucidate the pathophysiology of NPS and determine treatment strategies or care plans for NPS in patients with AD.

DISCLOSURE STATEMENT

Authors’ disclosures available online (https://www.j-alz.com/manuscript-disclosures/18-0304r4).

References

McKhann

, Drachman

, Folstein

, Katzman

, Price

, Stadlan

(1984) Clinical diagnosis of Alzheimer’s disease: Report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s Disease. Neurology 34, 939–944.

de Vugt

, Stevens

, Aalten

, Lousberg

, Jaspers

, Verhey

(2005) A prospective study of the effects of behavioral symptoms on the institutionalization of patients with dementia. Int Psychogeriatr 17, 577–589.

Lövheim

, Sandman

, Karlsson

, Gustafson

(2008) Behavioral and psychological symptoms of dementia in relation to level of cognitive impairment. Int Psychogeriatr 20, 777–789.

Mega

, Cummings

, Fiorello

, Gornbein

(1996) The spectrum of behavioral changes in Alzheimer’s disease. Neurology 46, 130–135.

Piccininni

, Di Carlo

, Baldereschi

, Zaccara

, Inzitari

(2005) Behavioral and psychological symptoms in Alzheimer’s disease: Frequency and relationship with duration and severity of the disease. Dement Geriatr Cogn Disord 19, 276–281.

Schneider

, Tariot

, Dagerman

, Davis

, Hsiao

, Ismail

, Lebowitz

, Lyketsos

, Ryan

, Stroup

, Sultzer

, Weintraub

, Lieberman

, CATIE-AD Study Group (2006) Effectiveness of atypical Antipsychotic drugs in patients with Alzheimer’s disease. N Engl J Med 355, 1525–1538.

Nagata

, Nakajima

, Shinagawa

, Plitman

, Graff-Guerrero

, Mimura

, Nakayama

(2016) Psychosocial or clinico-demographic factors related to neuropsychiatric symptoms in patients with Alzheimer’s disease needing interventional treatment: Analysis of the CATIE-AD study. Int J Geriatr Psychiatry 232, 1264–1271.

Gauthier

, Vellas

, Farlow

, Burn

(2006) Aggressive course of disease in dementia. Alzheimers Dement 2, 210–207.

Ropacki

, Jeste

(2005) Epidemiology of and risk factors for psychosis of Alzheimer’s disease: A review of 55 studies published from 1990 to 2003. Am J Psychiatry 162, 2022–2030.

10.

Nagata

, Nakajima

, Shinagawa

, Plitman

, Nakayama

, Graff-Guerrero

, Mimura

(2017) Baseline predictors of antipsychotic treatment continuation and response at week 8 in patients with Alzheimer’s disease with psychosis or aggressive symptoms: An analysis of the CATIE-AD study. J Alzheimers Dis 60, 263–272.

11.

Schneider

, Tariot

, Lyketsos

, Dagerman

, Davis

, Hsiao

, Jeste

, Katz

, Olin

, Pollock

, Rabins

, Rosenheck

, Small

, Lebowitz

, Lieberman

(2001) National Institute of Mental Health Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) Alzheimer disease trial methodology. Am J Geriatr Psychiatry 9, 346–360.

12.

Cummings

(1997) The Neuropsychiatric Inventory: Assessing psychopathology in dementia patients. Neurology 48(5 Suppl 6), S10–S16.

13.

Ismail

, Emeremni

, Houck

, Mazumdar

, Rosen

, Rajji

, Pollock

, Mulsant

(2013) A comparison of the E-BEHAVE-AD, NBRS, and NPI in quantifying clinical improvement in the treatment of agitation and psychosis associated with dementia. Am J Geriatr Psychiatry 21, 78–87.

14.

Leucht

, Kane

, Etschel

, Kissling

, Hamann

, Engel

(2006) Linking the PANSS, BPRS, and CGI: Clinical implications. Neuropsychopharmacology 31, 2318–2325.

15.

Schneider

, Olin

, Doody

, Clark

, Morris

, Reisberg

, Schmitt

, Grundman

, Thomas

, Ferris

(1997) Validity and reliability of the Alzheimer’s disease cooperative study-clinical global impression of change. The Alzheimer’s disease cooperative study. Alzheimer Dis Assoc Disord 11(Suppl 2), S22–S32.

16.

Folstein

, Folstein

, McHugh

(1975) “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

17.

Galasko

, Bennett

, Sano

, Ernesto

, Thomas

, Grundman

, Ferris

(1997) An inventory to assess Activities of Daily Livingfor clinical trials in Alzheimer’s disease. Alzheimer Dis Assoc Disord 11(Suppl 2), S33–S39.

18.

Lyketsos

, Galik

, Steele

, Steinberg

, Rosenblatt

, Warren

, Sheppard

, Baker

, Brandt

(1999) The “General Medical Health Rating” (GMHR): A bedside global rating of medical comorbidity in patients with dementia. J Am Geriatr Soc 47, 487–491.

19.

Zarit

, Reever

, Bach-Peterson

(1980) Relatives of the impaired elderly: Correlates of feeling burden. Gerontologist 20, 649–655.

20.

Winblad

, Wimo

, Engedal

, Soininen

, Verhey

, Waldemar

, Wetterholm

, Haglund

, Zhang

, Schindler

(2006) 3-year study of donepezil therapy in Alzheimer’s disease: Effects of early and continuous therapy. Dement Geriatr Cogn Disord 21, 353–363.

21.

Barocco

, Spallazzi

, Concari

, Gardini

, Pelosi

, Caffarra

(2017) The progression of Alzheimer’s disease: Are fast decliners really fast? A four-year follow-up. J Alzheimers Dis 57, 775–786.

22.

Clark

, Sheppard

, Fillenbaum

, Galasko

, Morris

, Koss

, Mohs

, Heyman

(1999) Variability in annual Mini-Mental State Examination score in patients with probable Alzheimer disease: A clinical perspective of data from the Consortium to Establish a Registry for Alzheimer’s disease. Arch Neurol 56, 857–862.

23.

Howard

, McShane

, Lindesay

, Ritchie

, Baldwin

, Barber

, Burns

, Dening

, Findlay

, Holmes

, Hughes

, Jacoby

, Jones

, McKeith

, Macharouthu

, O’Brien

, Passmore

, Sheehan

, Juszczak

, Katona

, Hills

, Knapp

, Ballard

, Brown

, Banerjee

, Onions

, Griffin

, Adams

, Gray

, Johnson

, Bentham

, Phillips

(2012) Donepezil and memantine for moderate-to-severe Alzheimer’s disease. N Engl J Med 366, 893–903.

24.

Porsteinsson

, Drye

, Pollock

, Devanand

, Frangakis

, Ismail

, Marano

, Meinert

, Mintzer

, Munro

, Pelton

, Rabins

, Rosenberg

, Schneider

, Shade

, Weintraub

, Yesavage

, Lyketsos

, CitAD Research Group (2014) Effect of citalopram on agitation in Alzheimer disease: The CitAD randomized clinical trial. JAMA 311, 682–691.

25.

Devanand

, Pelton

, Cunqueiro

, Sackeim

, Marder

(2011) A 6-month, randomized, double-blind, placebo-controlled pilot discontinuation trial following response to haloperidol treatment of psychosis and agitation in Alzheimer’s disease. Int J Geriatr Psychiatry 26, 937–943.

26.

Devanand

, Mintzer

, Schultz

, Andrews

, Sultzer

, de la Pena

, Gupta

, Colon

, Schimming

, Pelton

, Levin

(2012) Relapse risk after discontinuation of risperidone in Alzheimer’s disease. N Engl J Med 367, 1497–1507.

27.

Rouch

, Pongan

, Trombert

, Fabre

, Auguste

, Sellier

, Freulon

, Jacqueline

, Federico

, Mouchoux

, Martin-Gaujard

, Krolak-Salmon

, Laurent

, Dorey

(2017) One-year evolution of behavioral and psychological symptoms of dementia in patients initially hospitalized in cognitive behavioral units: The EVITAL prospective cohort. J Alzheimers Dis 57, 147–155.

28.

Emanuel

, Lopez

, Houck

, Becker

, Weamer

, Demichele-Sweet

, Kuller

, Sweet

(2011) Trajectory of cognitive decline as a predictor of psychosis in early Alzheimer disease in the cardiovascular health study. Am J Geriatr Psychiatry 19, 160–168.

29.

Peters

, Schwartz

, Han

, Rabins

, Steinberg

, Tschanz

, Lyketsos

(2015) Neuropsychiatric symptoms as predictors of progression to severe Alzheimer’s dementia and death: The Cache County Dementia Progression Study. Am J Psychiatry 172, 460–465.

30.

Poulin

, Bergeron

, Dickerson

, Alzheimer’s Disease Neuroimaging Initiative (2017) Risk factors, neuroanatomical correlates, and outcome of neuropsychiatric symptoms in Alzheimer’s disease. J Alzheimers Dis 60, 483–493.

31.

Hsu

, Tsai

, Hwang

, Chen

(2017) Predictors of non-pharmacological intervention effect on cognitive function and behavioral and psychological symptoms of older people with dementia. Geriatr Gerontol Int 17(Suppl1), 28–35.

32.

Herrmann

, Gauthier

, Lysy

(2007) Clinical practice guidelines for severe Alzheimer’s disease. Alzheimers Dement 3, 385–397.

33.

Howard

, Juszczak

, Ballard

, Bentham

, Brown

, Bullock

, Burns

, Holmes

, Jacoby

, Johnson

, Knapp

, Lindesay

, O’Brien

, Wilcock

, Katona

, Jones

, DeCesare

, Rodger

, CALM-AD Trial Group (2007) Donepezil for the treatment of agitation in Alzheimer’s disease. N Engl J Med 357, 1382–1392.

34.

Liu-Seifert

, Siemers

, Price

, Han

, Selzler

, Henley

, Sundell

, Aisen

, Cummings

, Raskin

, Mohs

, Alzheimer’s Disease Neuroimaging Initiative (2015) Cognitive impairment precedes and predicts functional impairment in mild Alzheimer’s disease. J Alzheimers Dis 47, 205–214.

35.

Blackwood

, Howard

, Bentall

, Murray

(2001) Cognitive neuropsychiatric models of persecutory delusions. Am J Psychiatry 158, 527–539.

36.

Geda

, Schneider

, Gitlin

, Miller

, Smith

, Bell

, Evans

, Lee

, Porsteinsson

, Lanctôt

, Rosenberg

, Sultzer

, Francis

, Brodaty

, Padala

, Onyike

, Ortiz

, Ancoli-Israel

, Bliwise

, Martin

, Vitiello

, Yaffe

, Zee

, Herrmann

, Sweet

, Ballard

, Khin

, Alfaro

, Murray

, Schultz

, Lyketsos

, Neuropsychiatric Syndromes Professional Interest Area of ISTAART (2013) Neuropsychiatric symptoms in Alzheimer’s disease: Past progress and anticipation of the future. Alzheimers Dement 9, 602–608.

37.

Martyr

, Clare

(2012) Executive function and activities of daily living in Alzheimer’s disease: A correlational meta-analysis. Dement Geriatr Cogn Disord 33, 189–203.

38.

Carbon

, Correll

(2014) Clinical predictors of therapeutic response to antipsychotics in schizophrenia. Dialogues Clin Neurosci 16, 505–524.

39.

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.