The contribution of behavioral features to caregiver burden in FTLD spectrum disorders

Abstract

Introduction

Caregivers of patients with frontotemporal lobar degeneration (FTLD) spectrum disorders experience tremendous burden, which has been associated with the neuropsychiatric and behavioral features of the disorders.

Methods

In a sample of 558 participants with FTLD spectrum disorders, we performed multiple‐variable regressions to identify the behavioral features that were most strongly associated with caregiver burden, as measured by the Zarit Burden Interview, at each stage of disease.

Results

Apathy and disinhibition, as rated by both clinicians and caregivers, as well as clinician‐rated psychosis, showed the strongest associations with caregiver burden, a pattern that was consistent when participants were separated cross‐sectionally by disease stage. In addition, behavioral features appeared to contribute most to caregiver burden in patients with early dementia.

Discussion

Caregivers should be provided with early education on the management of the behavioral features of FTLD spectrum disorders. Interventions targeting apathy, disinhibition, and psychosis may be most useful to reduce caregiver burden.

1. BACKGROUND

Patients with neurodegenerative illness are uniquely dependent on external support. Progressive declines in cognition, behavior, language, and motor functioning associated with neurodegeneration impede the ability of many patients to function independently. ¹ Most patients are cared for at home, usually by spouses or other family members. ² Care demands are high, and caregivers often experience tremendous burden. ³ , ⁴

Caregivers’ experiences of burden vary depending on patient factors, including age, with the burden higher for patients with early‐onset versus late‐onset dementias, ³ , ⁵ , ⁶ and disease severity, with the burden typically increasing as the disease progresses. ³ , ⁷ , ⁸ Caregiver burden also varies based on caregiver characteristics including sex, relationship to patient, relationship quality, and social connectedness. Women report higher levels of burden, ² , ³ , ⁹ as do immediate family members. ² , ⁹ Burden is lower in caregivers with a strong relationship to the patient ³ , ¹⁰ or with social support. ² , ⁴ Some studies observe a linear increase in caregiver burden over time, ⁷ whereas others find that burden plateaus or diminishes. ¹¹

In particular, the presence of neuropsychiatric and behavioral symptoms (hereafter referred to as “behavioral features”) predicts high levels of caregiver burden in neurodegenerative disorders. ² , ³ , ⁴ , ⁵ , ⁷ , ⁸ , ⁹ , ¹² Although behavioral features are common across neurodegenerative diseases, ¹³ they are a diagnostic criterion of behavioral variant frontotemporal dementia (bvFTD). ¹⁴ BvFTD is a clinical diagnosis associated with frontotemporal lobar degeneration (FTLD) neuropathology, along with a cluster of diseases characterized by the degeneration of the frontal and temporal lobes. Other FTLD spectrum disorders include semantic and non‐fluent variants of primary progressive aphasia (svPPA and nfvPPA), corticobasal syndrome (CBS), progressive supranuclear palsy (PSP), and FTD with amyotrophic lateral sclerosis (FTD‐ALS). ¹⁵ These disorders differ in their clinical presentations, with varying degrees of behavioral, language, and motor changes. ¹⁵ Most cases are sporadic, but 10% to 30% follow an autosomal dominant inheritance pattern. ¹⁵ FTLD disorders have high economic costs—often occurring while the patient and caregiver are still working—and are difficult to diagnose, both of which contribute to caregiver burden. ³

Although behavioral features are most characteristic of bvFTD, ¹⁴ , ¹⁶ they are common in other FTLD spectrum disorders. ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ svPPA can show rates of behavioral features comparable to those of bvFTD, ²² and unique emotional and behavioral features in svPPA, such as restrictive eating, somatic complaints, and emotional distress, have been described. ²³ Psychotic features are common in FTD with amyotrophic lateral sclerosis (FTD‐ALS). ²⁴

Given the high rates of behavioral features and other factors, caregiver burden tends to be higher in FTLD than in other neurodegenerative diseases. ³ , ⁶ , ¹² , ¹⁴ , ¹⁶ , ²⁵ Studies have directly associated this caregiver burden with behavioral features in bvFTD, ³ , ⁵ svPPA, ⁷ and CBS. ¹² , ²⁶ In patients with ALS and FTD‐ALS, behavioral features are a stronger predictor of caregiver burden than motor impairment. ⁷ , ²⁷ , ²⁸

A limitation of the studies cited above is that they have either examined individual or a subset of FTLD diagnoses. Furthermore, although burden has been shown to increase with disease progression, it is not established whether associations between behavioral features, specifically, and caregiver burden vary depending on the patient's stage of illness. Thus in the current study, we explored the contribution of individual behavioral features to caregiver burden across FTLD spectrum disorders and examined outcomes cross‐sectionally in questionable, mild, and moderate dementia.

2. METHODS

2.1. Participants

Participants were enrolled in the Advancing Research and Treatment for Frontotemporal Lobar Degeneration (ARTFL; U54 NS092089) or Longitudinal Evaluation of Familial Frontotemporal Dementia Subjects (LEFFTDS; U01 AG045390) studies (since combined into the ARTFL LEFFTDS Longitudinal Frontotemporal Lobar Degeneration consortium [ALLFTD]; U19 AG063911). Patients with sporadic or familial FTLD spectrum disorders were evaluated at 1 of 18 centers in North America. An extensive research evaluation was completed with each participant; full study procedures are published elsewhere. ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷

Each participant was accompanied by a study partner, who was interviewed about the participant's symptoms and their own experiences of burden. Study partner demographics are in Table 1. The majority of study partners (85%) were spouses or partners, and the remainder were other family members or friends. Eighty‐eight percent resided with the patients, and those who did not saw patients at least weekly. These study partners will herein be referred to as caregivers.

TABLE 1.

Patient and caregiver demographics

		CDR plus NACC FTLD
	Whole Sample	0.5	1	2
Patients	(n = 558)	(n = 102)	(n = 214)	(n = 242)
Mean age, years (SD)	64.16 (9.04)	62.95 (10.64)	64.48 (7.87)	64.38 (9.28)
Sex (M/F)	307/251	46/56	133/81	128/114
Education, years (SD)	15.75 (2.57)	15.59 (2.53)	15.99 (2.58)	15.60 (2.57)
Race (white/non‐white)	528/28	99/3	207/7	222/18
Diagnosis (count)
bvFTD	231	13	86	132
svPPA	64	11	31	22
nfvPPA	50	16	21	13
PSP	89	16	34	39
CBS	78	20	34	24
FTD‐ALS	24	4	8	12
MCI‐behavior	10	10	0	0
MCI‐cognitive	12	12	0	0
MoCA total, mean (SD)	19.45 (6.41)	23.94 (3.79)	20.86 (5.44)	16.32 (6.52)
Genetic status (carrier/noncarrier/unknown)*	104/339/115	31/53/18	28/142/44	45/144/53
Caregivers
Sex (M/F)	204/351	43/59	70/143	91/149
Education, years (SD)	15.98 (2.57)	15.77 (2.69)	16.03 (2.52)	16.02 (2.57)
Relationship
Spouse/partner	473	86	181	206
Child	38	7	17	14
Sibling	21	3	7	11
Other relative	15	–	7	8
Friend	9	6	2	1

^*Distribution of genetic carriers: MAPT n = 29, C9orf72 n = 54, GRN n = 21.

Informed consent was obtained from all participants or their appointed surrogates. Procedures were conducted according to the Declaration of Helsinki. Ethical approval was obtained from local institutional review boards at each center.

The final sample for our study included 558 participants: 12 with mild cognitive impairment (MCI‐cognitive), 10 with mild behavioral impairment (MCI‐behavior), 231 with bvFTD, 64 with svPPA, 50 with nfvPPA, 78 with CBS, 89 with PSP, and 24 with FTD‐ALS. Note that “MCI‐cognitive” refers to what is typically termed “MCI” in the Alzheimer's literature, and represents early cognitive changes that are possibly attributable to a neurodegenerative disorder but without significant functional impairment. ³⁸ , ³⁹ “MCI‐behavior” is a novel term that describes early behavioral changes that may be associated with a diagnosis of bvFTD (eg, disinhibition, apathy) but that do not meet full criteria for a bvFTD diagnosis. ²⁹ Demographic details are presented in Table 1.

2.2. Inclusion criteria

All patients with a confirmed FTLD‐associated pathogenic mutation and the presence of cognitive, language, or behavioral changes, as defined by a Global Clinical Dementia Rating + National Alzheimer's Coordinating Center FTLD (CDR + NACC FTLD) score of ≥0.5, ie, “symptomatic”, were included in our sample. Asymptomatic genetic carriers (Global CDR + NACC FTLD score of 0) were not included in the sample. Patients without a known genetic mutation who were assigned a Global CDR + NACC FTLD score of ≥0.5 and a primary clinical phenotype of bvFTD, svPPA, nfPPA, CBS, PSP, or FTD‐ALS were also included if confidence in their primary phenotype was at least 50%.

HIGHLIGHTS

• Behavioral features in frontotemporal lobar degeneration (FTLD) disorders contribute greatly to caregiver burden.

• In particular, apathy, disinhibition, and psychosis show the strongest associations.

• Association estimates are highest in caregivers of patients with early dementia.

• Our findings motivate early education and management of behavioral changes in FTLD.

RESEARCH IN CONTEXT

1. Systematic review: Authors reviewed the literature using traditional sources (eg, PubMed) to examine publications on caregiver burden and its ties to neuropsychiatric and behavioral features of frontotemporal lobar degeneration (FTLD) spectrum disorders and other dementias. Recent publications indicate a high prevalence of burden in caregivers of patients with FTLD spectrum disorders and strong associations between the presence of neuropsychiatric and behavioral features and caregiver burden.

2. Interpretation: Consistent with the described literature, we found that burden in caregivers of patients with FTLD spectrum disorders was associated with the presence of neuropsychiatric and behavioral features. Strongest associations were found for apathy, disinhibition, and psychosis.

3. Future directions: Our findings motivate inquiry into the efficacy of caregiver education on and/or interventions for apathy, disinhibition, and psychosis in reducing caregiver burden in FTLD spectrum disorders.

Participants were excluded from analyses if their caregivers did not complete the Zarit Caregiver Burden Interview, if they saw their caregivers less often than weekly, or if their caregivers were paid caretakers. Participants were also excluded if they had severe dementia (Global CDR + NACC FTLD score of 3), as individuals in this group were more likely to have severe impairment in activities of daily living, which may overshadow behavioral features, ⁸ or to live in institutional care (23%). 

2.3. Measures

2.3.1. Diagnostic evaluation

Diagnoses were made by site study clinicians after participants completed all research procedures, which included medical and family history, a standard neurologic interview and exam, neuropsychological testing, and a caregiver interview. Primary phenotypes per published criteria ¹⁴ , ²⁹ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ were recorded following clinical and neurological examination by a trained study physician. Physician confidence in the primary phenotype was self‐rated from 0 to 100%. ²⁹

Participants’ global dementia severity was evaluated using the standard NACC CDR ⁴⁴ and supplementary FTLD module. ⁴⁵ Ratings from 0 to 3 (0 = none, 0.5 = questionable, 1 = mild, 2 = moderate, 3 = severe) were assigned for the six CDR domains of Memory, Orientation, Judgment & Problem Solving, Community Affairs, Home & Hobbies, Personal Care, and two supplemental domains of Behavior and Language. An algorithm designed for disease staging in FTLD was used to calculate a global score (0‐3) based on all domains (hereto referred to as “CDR + NACC FTLD”). ²⁹ , ³⁴ Note that a global score of 0.5, or “questionable dementia,” indicates the presence of subtle cognitive, language, or behavioral changes, which are either inconsistent or very mild, but which the clinician believes may be attributable to a neurodegenerative process.

Motor symptoms, including tremor, gait, rigidity, bradykinesia, and speech, were evaluated with the Unified Parkinson's Disease Rating Scale (UPDRS) ⁴⁶ Part III.

2.3.2. Behavioral features

Behavioral features were evaluated by study physicians, based on participant evaluation and caregiver report. Physicians rated the presence or absence of apathy/withdrawal, depression, psychosis, disinhibition, irritability, and agitation (NACC B9 form).

Caregivers additionally rated behavioral features using the Neuropsychiatric Inventory Questionnaire (NPI‐Q), a 12‐item questionnaire. ⁴⁷ Severity was rated from 0 to 3 (0 = absent, 1 = mild, 2 = moderate, 3 = severe). Items included delusions, hallucinations, agitation/aggression, depression/dysphoria, anxiety, elation/euphoria, apathy/indifference, disinhibition, irritability/lability, motor disturbance, nighttime behaviors, and appetite/eating.

2.3.3. Caregiver burden

Caregivers completed the Zarit Caregiver Burden Interview. ⁴⁸ They were asked about 22 feelings of burden, for example, “Do you feel stressed between caring for your relative and trying to meet other responsibilities?”, “Do you feel your health has suffered because of your involvement with your relative?”, and “Do you feel embarrassed over your relative's behavior?” Study partners responded on a Likert‐type scale (0 = Never, 1 = Rarely, 2 = Sometimes, 3 = Quite Frequently, 4 = Nearly Always). Possible total scores ranged from 0 to 88, with higher numbers indicating greater burden.

2.4. Data preparation and statistical analyses

2.4.1. Behavioral predictors regression analysis

Linear modeling was conducted to determine the strength of association between individual behavioral features and caregiver burden. Total Zarit Burden Interview scores were used as the outcome measure. Predictor variables consisted of individual behavioral feature ratings on the clinician‐rated NACC B9 and caregiver‐rated NPI‐Q scales. We performed separate analyses with each scale in order to provide a more robust set of information, as the two scales differ by rater, item‐level wording, and rating scale, thus using both in our analyses allowed an examination of the relationship between symptoms and caregiver burden that was less affected by the limitations of a specific measurement.

Covariates were selected a priori, were included in all models, and consisted of participant age, caregiver sex, relationship of caregiver to participant, UPDRS score, and participant genetic status (carrier/non‐carrier/unknown, that is, inadequate information to classify as a carrier or non‐carrier of a hereditary FTLD mutation). Participant age, caregiver sex, and relationship of caregiver have previously been associated with caregiver burden. UPDRS scores were included to enable us to examine behavioral associations after removing burden associated with motor changes, as this was not the focus of this study. Genetic status was included to examine differential caregiver burden between familial and sporadic forms of the disease, as caregiving demands and stressors may differ between these groups.

Each behavioral predictor was entered into a separate model with the covariates specified above to determine the strength of its association with caregiver burden. These single predictor models were performed to examine associations between each variable and caregiver burden, to provide a starting point for understanding these relationships. In addition, to determine which features were uniquely associated with burden from within the sets of clinician‐ and caregiver‐rated features, two multiple predictor regression models were defined in advance: one including all symptoms from the clinician‐rated NACC B9 form as predictors and one including all features from the caregiver‐rated NPI‐Q scale as predictors. These multiple predictor models were performed to explore how much independent information was associated with each predictor out of a set, for example, was additional variance explained when predictor y was added to predictor z. All covariates were added to both types of models. To avoid multicollinearity, and because the symptoms often co‐occur and can be difficult for caregivers to distinguish, delusions and hallucinations were combined into a single “psychosis” item on both scales. None of the estimated correlations between predictors within either model exceeded 0.5.

Regression analyses were performed on the whole sample and again separated by CDR + NACC FTLD global score (0.5 = questionable dementia, 1 = mild dementia, and 2 = moderate dementia). For whole‐group analyses that include disease stage as an additional predictor, see Supplemental Material. Phenotype was not included as a variable in these analyses because there is significant overlap in behavioral features among FTLD phenotypes, ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ participants may meet criteria for multiple phenotypes as their disease progresses, and symptom profiles change over time. ⁴⁹ , ⁵⁰ Participants with missing values for any predictor variables or covariates were excluded from analyses (≈10% of our sample). The greatest impact of missing data was due to missing UPDRS scores, leading to the exclusion of 38 participants. There was no discernable pattern in the remaining missing data. In all comparisons, statistical significance was defined at the $\alpha =0.05$ level (ie, $P\le .05$).

3. RESULTS

3.1. Percentages

The percentages of individual behavioral features within our sample are reported in Table 2. Each feature was observed in between 7% and 62% of the sample.

TABLE 2.

Percentage of sample with behavioral features

	Clinician Rated				Caregiver Rated*
	Whole Sample n (%)	CDR + NACC FTLD = 0.5 n (%)	CDR + NACC FTLD = 1 n (%)	CDR + NACC FTLD = 2 n (%)	Whole Sample n (%)	CDR = 0.5 n (%)	CDR = 1 n (%)	CDR = 2 n (%)
Apathy	340 (61.6)	33 (33.0)	120 (56.9)	187 (77.6)	336 (60.4)	44 (43.6)	129 (60.3)	162 (67.6)
Disinhibition	278 (50.5)	14 (14.3)	104 (49.3)	160 (66.4)	242 (43.4)	19 (18.6)	93 (43.5)	130 (53.7)
Depression	115 (21.1)	16 (16.2)	48 (23.0)	51 (21.4)	235 (40.3)	45 (44.1)	92 (43.0)	88 (45.0)
Anxiety	117 (21.4)	10 (10.2)	52 (24.6)	55 (23.1)	244 (43.9)	39 (38.2)	98 (45.8)	107 (44.6)
Irritability	238 (43.0)	30 (30.0)	91 (43.1)	117 (48.3)	273 (48.9)	40 (39.2)	109 (50.9)	124 (51.2)
Agitation	127 (23.0)	55 (55.0)	36 (17.1)	86 (35.5)	235 (40.3)	25 (24.5)	91 (42.5)	109 (45.0)
Psychosis	41 (7.3)	2 (2.0)	19 (8.9)	20 (8.3)	75 (13.4)	5 (4.9)	30 (14.0)	40 (16.5)
Nighttime behaviors	–	–	–	–	224 (40.2)	30 (29.4)	97 (45.3)	97 (40.2)
Elation/euphoria	–	–	–	–	90 (16.1)	9 (8.8)	34 (15.9)	47 (19.4
Motor disturbance	–	–	–	–	216 (38.8)	13 (12.7)	82 (38.5)	131 (50.2)
Appetite/eating	–	–	–	–	293 (52.6)	37 (36.3)	109 (50.9)	147 (61.0)

Percent of sample endorsing presence of symptom.

^*This includes endorsement of mild, moderate, or severe symptoms.

CDR + NACC FTLD = Clinical Dementia Rating plus National Alzheimer's Coordinating Center Supplementary FTLD module.

3.2. Single‐predictor plus covariate models

Statistics are presented in Table 3. Within whole‐sample analyses, all behavioral features (both clinician and caregiver rated) entered independently as predictors had statistically significant associations with caregiver burden (Table 3). Statistically significant associations in cross‐sectional analyses within each disease stage are also shown in Table 3.

TABLE 3.

Single predictor associations with caregiver burden

				CDR + NACC FTLD
	All (n = 503)			0.5 (n = 90)			1 (n = 197)			2 (n = 216)
Clinician Rated	B	95% CIs	p	B	95% CIs	p	B	95% CIs	p	B	95% CIs	p
Apathy/withdrawal	10.27	(7.54, 12.99)	<0.001	11.13	(5.53, 16.74)	<0.001	6.62	(2.65, 10.60)	0.001	4.89	(0.12, 9.65)	0.044
Disinhibition	12.75	(9.98, 15.52)	<0.001	10.10	(1.94, 18.26)	0.015	9.81	(5.49, 14.14)	<0.001	6.30	(1.82, 10.79)	0.006
Depression	4.79	(1.26, 8.32)	0.008	3.45	(‐4.52, 11.43)	0.396	5.37	(0.39, 10.34)	0.035	4.07	(‐0.95, 9.10)	0.112
Anxiety	3.82	(0.32, 7.33)	0.033	6.08	(‐3.30, 15.46)	0.204	‐0.20	(‐5.12, 4.72)	0.936	3.59	(‐1.22, 8.41)	0.144
Irritability	6.26	(3.42, 9.10)	<0.001	2.73	(‐3.73, 9.19)	0.407	5.61	(1.34, 9.89)	0.010	3.77	(‐0.33, 7.86)	0.072
Agitation	6.22	(2.90, 9.53)	<0.001	1.22	(‐11.31, 13.74)	0.849	2.66	(‐2.77, 8.09)	0.337	1.70	(‐2.48, 5.89)	0.425
Psychosis	10.87	(5.50, 16.24)	<0.001	27.32	(7.37, 47.28)	0.007	9.30	(1.83, 16.76)	0.015	5.34	(‐1.83, 12.52)	0.144

				CDR + NACC FTLD
	All (n = 513)			0.5 (n = 94)			1 (n = 201)			2 (n = 218)
Caregiver Rated	B	95% CIs	p	B	95% CIs	p	B	95% CIs	p	B	95% CIs	p
Apathy/Indifference	5.53	(4.35, 6.71)	< 0.001	7.37	(4.59, 10.15)	<0.001	5.20	(3.30, 7.10)	<0.001	2.76	(1.09, 4.44)	0.001
Disinhibition	5.59	(4.36, 6.82)	< 0.001	7.86	(4.40, 11.33)	<0.001	5.59	(3.71, 7.46)	<0.001	2.28	(0.49, 4.07)	0.013
Depression	2.59	(0.99, 4.19)	0.001	1.29	(‐2.19, 4.77)	0.466	3.35	(1.05, 5.64)	0.004	3.00	(0.73, 5.28)	0.010
Anxiety	3.82	(2.33, 5.31)	<0.001	5.16	(0.82, 9.50)	0.020	3.91	(1.76, 6.06)	<0.001	2.13	(0.16, 4.11)	0.034
Irritability/Lability	3.98	(2.65, 5.31)	<0.001	4.11	(0.92, 7.30)	0.012	5.35	(3.26, 7.45)	<0.001	1.65	(‐0.14, 3.43)	0.070
Agitation/Aggression	3.82	(2.33, 5.30)	<0.001	2.89	(‐0.57, 6.35)	0.102	4.55	(2.42, 6.68)	<0.001	1.71	(‐0.38, 3.79)	0.108
Psychosis	4.96	(2.86, 7.06)	<0.001	10.61	(3.77, 17.46)	0.002	7.32	(4.07, 10.56)	<0.001	0.88	(‐1.74, 3.49)	0.512
Nighttime behaviors	4.54	(3.21, 5.87)	<0.001	3.19	(‐0.21, 6.58)	0.066	4.17	(2.26, 6.08)	<0.001	4.02	(2.19, 5.86)	<0.001
Euphoria/Elation	4.72	(2.76, 6.68)	<0.001	11.66	(4.77, 18.54)	0.001	6.16	(3.25, 9.08)	<0.001	0.87	(‐1.65, 3.39)	0.497
Motor disturbance	4.50	(3.16, 5.84)	<0.001	10.66	(5.71, 15.61)	<0.001	2.26	(0.08, 4.44)	0.042	2.00	(0.19, 4.70)	0.030
Appetite/Eating	4.43	(3.19, 5.67)	<0.001	3.91	(0.82, 6.70)	0.013	2.98	(1.06, 4.91)	0.002	3.10	(1.31, 4.90)	0.001

Associations between individual behavioral features and caregiver burden (measured by the Zarit Burden Interview) were tested one feature at a time, adjusting for the following covariates: “UPDRS total score,” “Patient Age,” “Patient Genetic Status,” “Caregiver Sex,” and “Caregiver Relationship.”

Clinician‐rated predictors are binary (yes/no) and caregiver‐rated predictors are on a four‐point scale (absent, mild, moderate, severe).

CI = Confidence Interval.

CDR + NACC FTLD = Clinical Dementia Rating plus National Alzheimer's Coordinating Center Supplementary FTLD module.

CDR 0.5 = Questionable dementia, 1 = mild dementia, 2 = moderate dementia.

B values represent regression parameter estimates.

To further characterize the contribution of each behavioral feature to burden, we calculated the estimate for each feature, that is, the estimated difference in mean caregiver burden per unit change in the predictor (eg, a change from “no apathy” to “apathy” on the clinician rating form or a change from “moderate apathy” to “severe apathy” on the caregiver rating form). For context, a 4‐point change in total burden would represent a change from “never” to “always” on a single item on the Zarit Burden Interview and a 20‐point change would represent a change from “mild burden” to “moderate burden” on the total Zarit Burden score. Even for those predictors that did not reach statistical significance, all estimates were in the expected direction, making the overall pattern of results fit well with expectations. Overall, estimates were lower for the caregiver‐rated scale, likely driven by the fact that a one‐unit change represents a more graded transition (eg, change from “mild” to “moderate”) as opposed to a change from “absent” to “present” as on the clinician‐rated scale.

3.3. Multiple‐predictor models

Statistics are presented in Table 4. Models were pre‐specified and forced entry, that is, no model selection procedure was utilized. In the whole sample—within the set of clinician‐rated behavioral features—apathy, disinhibition, and psychosis were statistically significant predictors of caregiver burden and had the largest statistical estimates. From the set of caregiver‐rated features, apathy, disinhibition, anxiety, and nighttime behaviors were statistically significant predictors of caregiver burden and had the largest statistical estimates. All estimates were in the expected direction.

TABLE 4.

Multiple‐predictor regression models

				CDR + NACC FTLD
	All (n = 503)			0.5 (n = 90)			1 (n = 197)			2 (n = 216)
Clinician‐Rated	B	95% CIs	P	B	95% CIs	P	B	95% CIs	P	B	95% CIs	P
Apathy	6.20	(3.41, 9.00)	<.001	8.66	(2.90, 14.43)	.003	3.90	(−0.14, 7.94)	.058	3.31	(−1.58, 8.19)	.185
Disinhibition	9.84	(6.93, 12.75)	<.001	5.52	(−2.06, 13.09)	.153	7.42	(2.90, 11.94)	.001	5.60	(0.91, 10.30)	.019
Depression	1.94	(−1.42, 5.30)	.257	0.69	(−7.20, 8.57)	.864	4.71	(−0.50, 9.91)	.077	2.59	(−2.52, 7.70)	.320
Anxiety	0.36	(−2.94, 3.65)	.833	0.95	(−8.09, 9.99)	.837	−2.96	(−8.02, 2.10)	.252	1.70	(−3.18, 6.58)	.494
Irritability	1.99	(−0.97, 4.95)	.187	0.57	(−5.70, 6.83)	.859	3.21	(−1.39, 7.80)	.171	1.59	(−3.07, 6.25)	.504
Agitation	1.28	(−2.09, 4.65)	.456	0.81	(−10.39, 12.00)	.888	−1.10	(−6.64, 4.44)	.697	−0.55	(−5.14, 4.05)	0.816
Psychosis	6.76	(1.80, 11.72)	.008	23.48	(4.70, 42.25)	.014	4.86	(−2.30, 12.02)	.183	4.86	(−2.23, 11.94)	0.179
Covariates:
UPDRS Score (motor)	0.09	(−0.01, 0.19)	.082	0.09	(−0.14, 0.31)	.445	0.01	(−0.15, 0.16)	.951	−0.02	(−0.18, 0.14)	.845
Patient Age	−0.07	(−0.23, 0.08)	.356	0.19	(−0.11, 0.49)	.222	−0.17	(−0.43, 0.09)	.203	−0.17	(−0.40, 0.07)	.159
Patient Genetic Status (Carrier vs Non‐carrier)	−3.77	(−7.27, −0.28)	.035	−1.56	(−8.36, 5.24)	.653	−4.61	(−10.58, 1.35)	.130	−1.26	(−6.73, 4.21)	.651
Patient Genetic Status (Carrier vs Unk a )	−5.57	(−9.70, −1.45)	.008	1.80	(−7.59, 11.19)	.707	−8.32	(−15.21, −1.45)	.018	−2.66	(−8.65, 3.34)	.385
Caregiver Sex (M vs W)	−3.30	(−5.94, −0.66)	.018	1.50	(−3.73, 6.73)	.574	−6.08	(−10.37, −1.80)	.005	−3.81	(−7.94, 0.32)	.071
Caregiver Relationship b
Model Statistics	R = 0.49; R2 = 0.24; F(16, 486) = 9.65, P < .001			R = 0.53; R2 = 0.29; F(15, 74) = 1.97, P = .029			R = 0.49; R2 = 0.24; F(16, 180) = 3.60, P < .001			R = 0.35; R2 = 0.12; F(16, 199) = 1.77, P = .037

				CDR + NACC FTLD
	All (n = 513)			0.5 (n = 94)			1 (n = 201)			2 (n = 218)
Caregiver Rated	B	95% CIs	P	B	95% CIs	P	B	95% CIs	P	B	95% CIs	P
Apathy	2.70	(1.37, 4.03)	<.001	4.50	(0.84, 8.16)	.016	2.47	(0.43, 4.51)	.018	1.42	(−0.43, 3.27)	.132
Disinhibition	2.60	(1.21, 3.99)	<.001	1.70	(−2.69, 6.08)	.448	2.40	(0.33, 4.47)	.023	1.21	(−0.75, 3.17)	.226
Depression	0.17	(−1.49, 1.82)	.844	−2.04	(−5.69, 1.62)	.275	0.74	(−1.71, 3.17)	.555	1.48	(−1.03, 3.98)	.248
Anxiety	1.67	(0.18, 3.16)	.028	3.55	(−0.49, 7.59)	.085	0.69	(−1.55, 2.93)	.548	1.05	(−1.12, 3.22)	.342
Irritability/Lability	1.18	(−0.25, 2.60)	.105	2.71	(−0.75, 6.17)	.125	2.64	(0.36, 4.93)	.024	0.28	(−1.67, 2.24)	.777
Agitation/Aggression	0.17	(−1.45, 1.78)	.839	−0.36	(−4.24, 3.52)	.855	0.47	(−1.85, 2.79)	.690	0.21	(−2.11, 2.53)	.858
Psychosis	1.67	(−0.30, 3.64)	.097	2.26	(−4.63, 9.14)	.521	3.43	(0.28, 6.58)	.033	−0.31	(−2.85, 2.24)	.813
Nighttime behaviors	1.74	(0.38, 3.11)	.012	−0.95	(−4.16, 2.26)	.561	1.10	(−0.92, 3.12)	.286	3.08	(1.08, 5.09)	.003
Euphoria/Elation	1.08	(−0.84, 2.99)	.271	0.61	(−7.23, 8.45)	.879	2.88	(−0.10, 5.85)	.058	−0.09	(−2.65, 2.47)	.945
Motor disturbance	0.78	(−0.65, 2.21)	.285	5.27	(−0.51, 11.04)	.074	−0.33	(−2.41, 1.75)	.757	−0.63	(−2.66, 1.40)	.544
Appetite/Eating	0.59	(−0.78, 1.95)	.401	1.05	(−1.74, 3.84)	.459	−0.71	(−2.76, 1.35)	.499	1.15	(−0.88, 3.18)	.267
Covariates:
UPDRS Score (motor)	0.03	0.07, −0.13)	.594	0.29	(0.08, 0.50)	.006	−0.03	(−0.17, 0.11)	.695	−0.10	(−0.25, 0.05)	.205
Patient Age	0.04	(−0.12, 0.19)	.649	0.01	(−0.27, 0.30)	.924	0.02	(−0.24, 0.27)	.906	−0.10	(−0.33, 0.13)	.380
Patient Genetic Status (Carrier vs Non‐carrier)	−2.65	(−6.04, 0.73)	.124	−5.31	(−12.04, 1.43)	.122	−1.76	(−7.26, 3.74)	.530	0.77	(−4.33, 5.88)	.767
Patient Genetic Status (Carrier vs Unk a )	−3.43	(−7.41, 0.54)	.09	−4.70	(−13.14, 3.75)	.276	−4.65	(−11.15, 1.86)	.162	−0.28	(−6.09, 5.54)	.925
Caregiver Sex (M vs W)	−3.30	(−5.90, −0.70)	.013	−1.11	(−6.34, 4.12)	.678	−3.10	(−7.01, 0.82)	.073	−3.87	(−7.84, 0.10)	.056
Caregiver Relationship b
Model Statistics	R = 0.53; R2 = 0.28; F(20, 492) = 9.70, P < .001			R = 0.63; R2 = 0.40; F(19, 74) = 2.583, P = .002			R = 0.60; R2 = 0.36; F(20, 180) = 4.99, P < .001			R = 0.43; R2 = 0.19; F(20, 197) = 2.26, P = .002

Associations between individual behavioral features and caregiver burden (measured by the Zarit Burden Interview) were tested using multivariable regressions, adjusting for covariates. Two models were determined a priori, one with the full set of behavioral features from the clinician‐rated scale and one with the full set of behavioral features from the caregiver‐rated scale.

Clinician‐rated predictors are binary (yes/no) and caregiver‐rated predictors are on a four‐point scale (absent, mild, moderate, severe).

CI = Confidence Interval.

CDR + NACC FTLD = Clinical Dementia Rating plus National Alzheimer's Coordinating Center Supplementary FTLD module.

CDR 0.5 = Questionable dementia, 1 = mild dementia, 2 = moderate dementia.

B values represent regression parameter estimates.

^a Unknown genetic status indicates that there was inadequate information to classify as a carrier or non‐carrier of a hereditary FTLD mutation, ie, no or limited clinical genetic testing was performed.

^b See Table 5 for estimated means.

When analyses were separated by disease stage, clinician‐rated apathy and psychosis and caregiver‐rated apathy were statistically significant predictors of burden in patients with questionable dementia. The largest statistical estimates were observed in clinician‐rated psychosis, apathy, and disinhibition and in caregiver‐rated motor disturbance, apathy, and anxiety. In patients with mild dementia, clinician‐rated disinhibition along with caregiver‐rated apathy, disinhibition, irritability, and psychosis were statistically significant predictors of burden. The largest observed estimates were in clinician‐rated disinhibition, depression, and psychosis and in caregiver‐rated psychosis, euphoria, and irritability. In patients with moderate dementia, clinician‐rated disinhibition and caregiver‐rated nighttime behaviors were statistically significant predictors of burden and had the largest observed estimates, along with clinician‐rated apathy and psychosis. The majority of estimates were in the expected direction, with some exceptions. Estimates not in the expected direction were small in magnitude and not clinically or statistically significant.

In addition, associations were found between several covariates and caregiver burden. Within the clinician‐rating model, the estimate of participant genetic status had the largest absolute value of all covariate estimates and was a statistically significant predictor in the whole sample, with caregivers of participants with known pathogenic mutations reporting less burden than those with no mutation or unknown status. Caregiver sex was also a statistically significant predictor of burden, with male caregivers experiencing less burden than female caregivers.

Within the caregiver‐rating model, estimates of both patient genetic status and caregiver sex again had the largest absolute value of the covariates, with findings in the same direction as in the clinician‐rating model. Caregiver sex was statistically significantly associated with caregiver burden in the whole group. Associations did not reach statistical significance when analyses were separated by disease stage, but estimates continued to reflect the above pattern, albeit with the confidence intervals crossing zero, indicating the possibility of no effect. Caregiver relationship to participant was statistically significantly associated with caregiver burden in both models, but with no clear pattern. (See Table 5 for mean caregiver burden for each caregiver and Global CDR + NACC FTLD score categories.)

TABLE 5.

Mean Zarit Burden Interview score by informant relationship

		CDR + NACC FTLD
	Whole Sample: Mean (CIs)	CDR = 0.5: Mean (CIs)	CDR = 1: Mean (CIs)	CDR = 2: Mean (CIs)
Spouse	33.80 (32.35, 35.25)	22.59 (19.50, 25.68)	33.01 (30.77, 35.25)	39.17 (37.14, 41.20)
Child	32.63 (27.35, 37.91)	17.86 (3.65, 32.07)	33.06 (26.69, 39.43)	39.50 (30.05, 48.95)
Sibling	32.62 (24.38, 40.86)	9.33 (‐16.52, 35.19)	33.43 (18.84, 48.02)	38.45 (27.15, 49.76)
Other Relative	29.20 (20.50, 37.90)	–	22.29 (7.90, 36.67)	35.25 (23.53, 46.97)
Friend	20.56 (8.08, 33.03)	16.17 (2.41, 29.92)	24.50 (‐233.27, 272.27)	–

Mean score on Zarit burden interview shown (range = 0‐88), with 95% confidence intervals in parentheses. Confidence intervals = binomial calculation, lower and upper bounds shown.

CI = Confidence Interval

CDR + NACC FTLD = Clinical Dementia Rating plus National Alzheimer's Coordinating Center Supplementary FTLD module.

4. DISCUSSION

Our results show that behavioral features are associated with caregiver burden in FTLD spectrum disorders. When examined individually, each behavioral feature reported by the clinician or caregiver was statistically significantly associated with caregiver burden (Table 3), with estimates in the expected directions. Multiple‐predictor models allowed us to identify which features from each scale were the strongest predictors of caregiver burden (Table 4). Apathy and disinhibition as rated by both clinicians and caregivers, and clinician‐rated psychosis, showed the strongest associations with burden, both within the whole group and cross‐sectionally at each stage of disease. It should be noted that our results are not adjusted for multiple comparisons, and thus we rely on scientific judgment rather than formal adjustment methods to indicate where caution is warranted due to multiple comparisons.

The features found to be the most burdensome in multiple‐predictor models were not necessarily the most frequent features in our sample (Table 2), indicating that some common behavioral features such as depression and agitation may be more manageable to caregivers, whereas features such as psychosis may be uniquely burdensome despite being rare. There were also discrepancies in the frequency of some clinician‐rated versus caregiver‐rated symptoms, notably anxiety, depression, and agitation. One possible explanation is that caregivers and clinicians may define some symptoms differently. For example, we have observed anecdotally in our lab that caregivers often interpret repetitive behaviors such as pacing, which are common in FTD, as prima facie indications of anxiety. Similarly, caregivers often interpret apathy and emotional withdrawal in the absence of anhedonia as depression. Another possible explanation is that differences in wording between the two scales caused differences in reporting. For example, in the NPI‐Q, the item on agitation asks, “Is the patient resistive to help from others at times, or hard to handle?”, whereas on the B9 form it asks “Does the subject have trouble sitting still? Does s/he shout, hit, and/or kick?” Thus the items may be tapping into slightly different aspects, or different levels of severity, of the same larger construct. Correlations of overlapping clinician‐ and caregiver‐rated symptoms are included in the Supplemental Material and show that while all correlations reach statistical significance, the Pearson's coefficients fall in the “moderate” range, from 0.296 to 0.539.

The behavioral features associated with burden in the whole group were not always associated statistically significantly at each disease stage, potentially due to reduced sample size and power within the subgroup analyses, but estimates for these same features tended to be of similar magnitude and direction at each disease stage, particularly for apathy, disinhibition, and psychosis. Inspection of the statistical estimates indicates that these features likely contribute to burden throughout disease, which aligns with existing evidence that behavioral symptoms continue to be prominent in mild and moderate FTD. ²¹ Estimates tended to be highest in participants with questionable dementia in both single‐predictor and multiple‐predictor models, although confidence intervals overlapped between groups. This inspection of statistical estimates raises the possibility that behavioral features are more burdensome when patients are at the earliest stages of dementia—when features could be first developing and when caregivers may not have a good understanding of the disease nor have a support network in place. By the time patients reach moderate dementia, caregivers may be more adept at managing symptoms, or more emotionally adjusted to them. Furthermore, as their cognitive capacity declines, patients may be less able to act out particular behaviors, or non‐behavioral symptoms that interfere with activities of daily life may predominate. ⁵¹

The association of apathy and disinhibition to caregiver burden is consistent with the past literature. ³ , ⁸ , ¹² Both are common across FTLD spectrum disorders. ¹² , ¹⁴ , ⁵² Apathy is most prevalent, ¹² , ⁵³ characterized by a lack of motivation and interest in activities, and often occurs alongside emotional withdrawal. ⁵³ Apathy can be uniquely distressing to family members who feel that patients are no longer themselves ⁸ , ⁵³ and may be particularly challenging in adults who retain job and family responsibilities. Disinhibition refers to disrupted social awareness and accompanying socially inappropriate behaviors, as well as impulsivity or recklessness. ⁵⁴ These behaviors can be difficult to manage, embarrassing to caregivers, ⁸ and have severe consequences such as being arrested or financial loss. Apathy and disinhibition may represent broader categories of lesser and greater activity with illness as opposed to more distinct symptoms such as euphoria or hallucinations. This may account, in part, for the commonality of apathy and disinhibition and their strong association with caregiver burden.

Psychosis, which included delusions and hallucinations, was strongly associated with caregiver burden, most notably in the clinician‐rated model. This is consistent with previous findings. ⁸ Unlike apathy and disinhibition, psychosis is relatively infrequent, appearing in 7% to 14% of our sample. Psychotic features may contribute uniquely to burden because of their strangeness, or because delusions or hallucinations could be distressing to patients.

Although not the primary focus of our investigation, we also found two covariates to be associated with burden. In our sample, male caregivers reported less burden than female caregivers. This is consistent with past reports. ² , ³ , ⁹ It is possible that women are saddled with more care demands than men, or that their coping strategies differ. ⁵⁵ An additional hypothesis is that men are less comfortable reporting the emotional components of burden during a research visit, as has been suggested by others. ⁵⁶ We also found that caregivers of patients with a pathogenic mutation reported less burden on average than those of patients with no mutation or with unknown genetic status. This was a surprise, as caregivers of genetic carriers may have to care for multiple family members with disease, may have children at risk of disease, or may themselves be mutation carriers, all of which could contribute to burden. However, patients with a family history may also be diagnosed earlier, have better access to education about their disease (ie, greater knowledge and preparedness from personal experience), or have more time to come to terms with the prospect of their illness, which could potentially explain reduced caregiver burden in this group. Notably, we did not investigate whether the association between genetic status and caregiver burden differed by specific mutation type, and thus were not able to evaluate the presence of between‐gene differences in this association.

Finally, it is worth mentioning where clinically meaningful estimates were not found. Despite the large number of patients with motor phenotypes in our sample, the estimates across all models for the UPDRS score were too small to be clinically meaningful, representing less than half a point change on the burden scale. This is consistent with previous studies in patients with ALS and FTD‐ALS, where behavioral features contributed more to caregiver burden than motor symptoms did. ⁷ , ²⁷ , ²⁸ Our results extend similar findings to the broad category of FTLD syndromes.

There were several limitations to our study. First, the demographic characteristics of the participants in our sample were relatively homogenous, with high education and nearly 95% of subjects identifying as white. Burden may be differentially experienced in other demographic groups. Second, although we had access to basic demographic information about our caregiver sample, we did not have access to more complex variables such as socioeconomic status, social support, hours per week spent providing care, and coping style, all of which have been associated with caregiver burden ⁸ and may therefore modulate the relationships we report. Third, we had to exclude a relatively high number (n = 38) of participants for whom our measure of motor symptoms was missing. On closer examination, some of these participants were unable to complete the examination due to severe motor symptoms. Thus it is possible that the estimated relationship between motor symptoms and caregiver burden was biased by this missing data. Fourth, due to limited sample sizes, we were not able to run separate analyses within each phenotypic group. Behavioral features may be differently associated with caregiver burden depending on phenotype. Similarly, because of the reduction of sample size in our subgroup analyses, we were limited in our ability to make statistical inferences. Finally, we did not have sufficient data to perform longitudinal analyses of change in caregiver burden over time.

In summary, our analyses showed that behavioral features are strongly associated with caregiver burden across FTLD spectrum disorders. In multiple‐predictor models, these associations appear to be best explained by apathy, disinhibition, and psychosis. Although this pattern is generally consistent at each disease stage, the relationships between behavioral features and caregiver burden are strongest in patients with questionable dementia. These findings have several direct clinical interpretations. Clinicians should probe for the presence of behavioral features in patients with all FTLD spectrum disorders, not just those with bvFTD, and should provide caregiver education accordingly. Education and interventions for behavioral features may be most effective for patients in the earliest stages of dementia. For caregivers of patients with multiple behavioral features who are experiencing significant burden, interventions for apathy and disinhibition may be most effective in reducing burden. At present there are no effective pharmacological treatments for apathy, although the use of oxytocin is under investigation, ⁵⁷ and disinhibition has limited treatments, which may have significant side effects. ⁵⁸ In the absence of disease‐modifying trials, symptomatic clinical trials targeting these symptoms, especially in early dementia, may have the most promise for reducing caregiver burden. In addition, non‐pharmacologic interventions such as behavioral management and caregiver support are recommended. ⁵⁹

B.S. Appleby: Has received research funding from CDC, NIH, Ionis, and Alector. Consulting for Acadia and Ionis. Received royalties from Wolters Kluwer. Received payment for expert testimony for Kaufman & Canoles. Received support for meetings/travel from the CJD Foundation. Holds unpaid leadership roles for the Alzheimer's Association and CJD Foundation.

K.L. Devick: Has received grant funds as a Co‐I or core personnel from the National Institutes of Health (NIH). Received compensation for serving as a teaching assistant for the Causal Mediation Analysis course as a part of the Erasmus Summer Programme.

Bradford C. Dickerson: Receives support from funding by the NIH. Received royalties for book publishing from Oxford University Press and Cambridge University Press. Received consulting fees for participation in Senior Advisory Boards for Acadia, Alector, Arkuda, Biogen, Denali, Lilly, Merck, Novartis, Takeda, and Wave Lifesciences. Received consulting fees from continuing medical education companies for lectures from Projects in Knowledge, InTouch Media. Chair‐elect for Medical Advisory Council, Association for FTD (unpaid); Chair, Medical and Scientific Advisory Council, MA/NH Chapter, Alzheimer's Association (unpaid).

J.A. Fields: Receives research funding from the NIH (NIH/NIA ‐ U54AG 44170‐8; institution NIH/NINDS ‐ UH3NS 95495‐4; institution NIH/NINDS ‐ U01NS 100620‐4; institution NIH/NIA ‐ RF1AG 57547‐4; institution NIH/NIA ‐ U19AG 63911‐2; institution NIH/NIA ‐ R01AG 68128‐1; institution NIH/NIA ‐ R44AG 65088‐2; institution NIH/NIA ‐ U01AG 45390‐5; institution NIH/NIA ‐ R43AG 65088‐1; institution PCORI ‐ R43AG 65088‐1; institution). Received ExercisAbilities as direct payment for lecture. Paid member of the NIA SWAN‐Aging DSMB.

N. Ghoshal: Has participated or is currently participating in clinical trials of anti‐dementia drugs sponsored by the following companies: Bristol Myers Squibb, Lilly/Avid Radiopharmaceuticals, Janssen, Novartis, Pfizer, Wyeth, SNIFF (The Study of Nasal Insulin to Fight Forgetfulness) study, and A4 (The Anti‐Amyloid Treatment in Asymptomatic Alzheimer's Disease) trial. She receives research support from the Tau Consortium and Association for Frontotemporal Dementia and is funded by the NIH. Received support for meetings/travel to Residency Examination Advisory Panel (appointed member); American Academy of Neurology; and the Rainwater Foundation Biannual Tau Consortium Meetings. Unpaid board member on the Northwestern University Feinberg School of Medicine Alumni Board.

M. Grossman: Support from NIH (AG0066597, AG054519, AG066418, AG052943), Department of Army, Newhouse Foundation, Wyncote Foundation, Peisach Family Foundation, and Robinson Family Foundation; consultant to Bracco and Passage Bio; participation in trials sponsored by Avid Radiopharmaceuticals, Life Molecular imaging, Biogen, Passage Bio, Prevail, and Alector.

J. Kornak: Has provided expert witness testimony for (1) Teva Pharmaceuticals in Forest Laboratories Inc. et al. v. Teva Pharmaceuticals USA, Inc., Case Nos. 1:14‐cv‐00121 and 1:14‐cv‐00686 (D. Del. filed Jan. 31, 2014 and May 30, 2014) regarding the drug memantine; (2) for Apotex/HEC/Ezra in Novartis AG et al. v. Apotex Inc., No. 1:15‐cv‐975 (D. Del. filed Oct. 26, 2015, regarding the drug Fingolimod; (3) on behalf of Puma Biotechnology in Hsingching Hsu et al. vs. Puma Biotechnology, Inc., et al. 2018 regarding the drug Neratinib; and (4) on behalf of Hikma Pharmaceuticals in Amarin Pharma, Inc vs. Hikma Pharmaceuticals in 2019. Received consulting fees from legal companies. Received honorarium from UC Davis. He receives research support from the NIH.

M.I. Lapid: Receives research funding from the NIH. Unpaid Assembly Representative, American Psychiatric Association (APA). Unpaid Member, Board of Directors, International Psychogeriatric Association (IPA).

I. Litvan: Supported by NIH grants: 2R01AG038791‐06A, U01NS090259, U01NS100610, U01NS80818, R25NS098999, P20GM109025; U19 AG063911‐1, and 1R21NS114764‐01A1; the Michael J Fox Foundation, Parkinson Foundation, Lewy Body Association, Roche, Abbvie, Biogen, Centogene, EIP‐Pharma, and Biohaven Pharmaceuticals. She was a member of the Scientific Advisory Board of Lundbeck and is a Scientific advisor for Amydis. She receives her salary from the University of California San Diego and as Chief Editor of Frontiers in Neurology.

I.R. Mackenzie: Scientific advisory board member for Prevail Therapeutics. Patent license paid by Prevail Therapeutics. Receives institutional funding from the Alzheimer's Association. Received honorarium for plenary talk at Canadian Association of Neuropathologists annual meeting, 2019. Travel costs reimbursed to attend Canadian Association of Neuropathologists annual meeting as plenary speaker.

M.F. Mendez: Receives support from NIH funding. Paid UpToDate Section Editor. Received payment from Medical Speaker's Bureau.

C.U. Onyike: Research support from the NIH, National Ataxia Foundation, Bluefield Foundation, Biogen Inc., Alector, the Jane Tanger Black Fund for Young‐Onset Dementias, and the Robert and Nancy Hall Brain Research Fund. Is a consultant for Acadia Pharmaceuticals and Alector Inc. Provided paid expert testimony for GOODELL, DEVRIES, LEECH & DANN, LLP. Holds unpaid leadership roles on the ISFTD and AFTD.

B. Pascual: Receives funding from the NIH (R01 AG054131‐Institution U19 AG063911‐Institution R01 AG068260‐Institution).

M.C. Tartaglia: Receives support from the NIH. Received payment from the AAN. Holds an unpaid leadership position from the Alzheimer's Society of Toronto.

B.F. Boeve: Has served as an investigator for clinical trials sponsored by GE Healthcare and Axovant. He receives royalties from the publication of a book entitled “Behavioral Neurology of Dementia” (Cambridge Medicine, 2009, 2017). He serves on the Scientific Advisory Board of the Tau Consortium. He receives research support from NIH, the Mayo Clinic Dorothy and Harry T. Mangurian Jr. Lewy Body Dementia Program, and the Little Family Foundation. Receives support from Biogen, Alector, and EIP Pharma for clinical trials to institution.

A.L. Boxer: Receives research support from NIH, the Tau Research Consortium, the Association for Frontotemporal Degeneration, the Rainwater Charitable Foundation, the Alzheimer's Association, and the Bluefield Project to Cure Frontotemporal Dementia. He has served as a consultant for AGTC, Alector, Arkuda, Arvinas, Bioage, Ionis, Lundbeck, Passage BIO, Samumed, Ono, Sangamo, Stealth, Transposon, UCB, Wave, Oligomerix, Roche, AGTC, AZTherapies, and GSK; and has received research support from Avid, Eisai, Biogen, and Roche. Holds stock in Alector, Arvinas, Arkuda, and AZTherapies.

H. Rosen: Has received research support from Biogen Pharmaceuticals, has consulting agreements with Wave Neuroscience and Ionis Pharmaceuticals, has received consulting fees from Biogen, and receives research support from the NIH and CA DPH. Unpaid member of a DSMB for Oregon Health Sciences University.

S. Cosentino: Receives funding from the NIH (R01 AG054525; R01 NS086736; U19 AG063893; R25AG059557; R01 MH120794; 5P50AG008702; 2R01AG007370; U24AG056270; R01AG058918; R01AG064614; and U19AG063911). Receives consulting fees from SAGE Pharmaceuticals, the Association for Frontotemporal Degeneration, and the NIH‐funded Pre‐symptomatic Familial ALS study. Received honoraria for speaking at the University of Bern as part of a graduate training seminar. Travel to scientific meetings is supported by NIH grant funding.

E.D. Huey: Receives research funding from the NIH.

M.S. Barker: Received consulting fee for statistical analyses from the Association for Frontotemporal Degeneration.

J. Goldman: Receives research support from NIH, HDSA, and the New York State Department of Health (RFA #1510130358).

H.E. Silverman, J.M. Ake, M. Manoochehri, D. Brushaber, L.K. Forsberg, Neil R. Graff‐Radford, H.W. Heuer: Nothing to disclose.

Supporting information

Supporting information

Supporting information

ALLFTD CONSORTIUM MEMBERS AND COLLABORATORS

Name	Affiliation
Andrea Bozoki, M.D.	Division of Memory and Cognitive Disorders, Department of Neurology, University of North Carolina, Chapel Hill, NC, USA
Kimiko Domoto‐Reilly, M.D.	Department of Neurology, University of Washington, Seattle, WA, USA
Daniel Kaufer, M.D.	Department of Neurology, University of North Carolina, Chapel Hill, NC, USA
David S. Knopman, M.D.	Department of Neurology, Mayo Clinic, Rochester, MN, USA
Douglas R. Galasko, M.D.	Department of Neuroscience, University of California, San Diego, San Diego, CA, USA
Ging‐Yuek Hsiung, M.D.	Division of Neurology, Department of Medicine, University of British Columbia, Vancouver, British Columbia, Canada
Walter Kremers, Ph.D.	Department of Neurology, Mayo Clinic, Rochester, MN, USA
Gabriel Leger, M.D.	Department of Neurology, University of California, San Diego, San Diego, CA, USA
Joseph C. Masdeu, M.D., Ph.D.	Nantz National Alzheimer Center, Houston Methodist Neurological Institute, Houston, TX, USA and Weill Cornell Medicine, NY, USA
Aaron Ritter, M.D.	Cleveland Clinic Lou Ruvo Center for Brain Health, Las Vegas, NV, USA
Erik D. Roberson, M.D., Ph.D.	Department of Neurology, Alzheimer's Disease Center, University of Alabama at Birmingham, Birmingham, AL, USA
Sandra Weintraub, Ph.D.	Northwestern Feinberg School of Medicine, Mesulam Center for Cognitive Neurology and Alzheimer's Disease

Silverman HE, Ake JM, Manoochehri M, et al., The contribution of behavioral features to caregiver burden in FTLD spectrum disorders. Alzheimer's Dement. 2022;18:1635–1649. 10.1002/alz.12494