Functional Disabilities and Psychiatric Symptoms in Primary Progressive Aphasia

Abstract

Objective:

The aims of this study are to describe the chronology of functional disabilities in primary progressive aphasia (PPA), and to examine associations between psychiatric comorbidities and functional disabilities.

Methods:

We conducted a retrospective data analysis using subjects enrolled at Alzheimer's Disease Research Centers between 2005 and 2019. Data were obtained from the National Alzheimer’s Coordinating Center database. We included subjects whose primary diagnosis was PPA. Functional status was coded as a binary variable for the following functions: ambulation, transaction skills, verbal communication, meal preparation, and self-care. Behavioral data derived from the Neuropsychiatric Inventory Questionnaire. Descriptive statistics and cox proportional hazard analyses were used to characterize the emergence of disabilities and their association with psychiatric comorbidities.

Results:

Data included 91 subjects with a clinical dementia rating scale of zero at baseline. At the initial visit, no individuals had impairments in self-care, while 7% had impairments in transactions, 3% in ambulation, and 2% in meal preparation. Ninety-three percent had language impairments at the onset of the study, and all by visit 4. By visit 5, 41% of patients had impairments in ambulation and in self-care, 49% were impaired in meal preparation and 70% had impairment in transactions. The presence of anxiety, depression, sleep disturbance and psychosis were all significantly associated with an increased risk for multiple functional disabilities.

Conclusion:

These findings provide clinicians with guidance for forecasting disabilities and targeting interventions in PPA.

INTRODUCTION

Frontotemporal dementia (FTD) encompasses a group of heterogeneous neurodegenerative syndromes and includes the primary progressive aphasia (PPA) phenotypes primary nonfluent aphasia (PNFA) and semantic dementia (SD). PNFA presents with labored and agrammatic speech alongside impaired sentence comprehension, with sparing of word and object knowledge, in association with left posterior fronto-insular atrophy or hypometabolism.¹ Patients who have SD exhibit fluent and vacuous speech, featuring dysnomia and agnosia for words and objects, in association with focal atrophy or hypometabolism of the anterior temporal lobe.¹ PNFA and SD develop in early midlife, about 60 years of age, and have estimated prevalence of 0.5–3.9/100,000 and 2.5–7.3/100,000, respectively.^2-6 There is variability with respect to illness duration, with estimated mean life expectancy of 8–9 years.⁷

While impaired language and communication are the predominant disabilities in PPA, in time additional disabilities accumulate, eroding quality of life and increasing the burden of care.^8-12 Eventually all forms of PPA converge on a dementia syndrome in which the deficiencies in cognition extend beyond language.¹³ In other words, as illness progresses, deficiencies in executive function, memory, orientation, praxis, and other cognitive functions develop and worsen. The result is the progressive erosion of skill necessary for transactions, meal preparation, self-care, ambulation and other functions. A few studies have documented the correlation between executive dysfunction, behavioral symptoms, and disability in PPA,^11,12 whereas the chronology of functional decline has not been described.

This study is intended to begin addressing the dearth of data on the devolution of functional capacity in PPA. The aim is to describe the chronology of disability and examine the temporal association between medical and psychiatric comorbidities and specific disabilities. It is hoped that studies of this kind will develop data for forecasting functional decline in individuals who have PPA and other FTD syndromes, in service of targeted care and better-informed guidance for patients and families. Such data will also facilitate the identification of opportunities to optimize quality of life, promote or support autonomy and ease the burden of care for family and other caregivers.

METHODS

Study Sample

Data were drawn from the Uniform Data Set maintained by the National Alzheimer’s Coordinating Center (NACC), a database funded by the National Institute on Aging.¹⁴ We included 1,944 subjects with a primary clinical diagnosis of PPA at the latest visit, 91 of whom had an initial Clinical Dementia Staging (CDR)^15,16 score of zero. Most were not designated a PPA subtype in the dataset; 18% were nonfluent/agrammatic PPA, 12% semantic dementia, and 70% were not designated a PPA subtype. Given this limitation, we did not segregate PPA subtypes in the analysis.

Demographics/Clinical Data

Illness severity was defined on the basis of the CDR Dementia Staging Instrument (hereafter referred to as CDR). We also examined the change in illness severity as defined by the CDR plus NACC FTLD Behavior & Language Domains (i.e., CDR plus NACC FTLD)—a modified version developed for measuring severity in FTD syndromes. Functional status was defined by variables for ambulation (gait and falls), transaction skills (shopping, taxes, and bills), verbal communication, meal preparation, and self-care. Coding for the variables was based on data from the following Uniform Data Set forms: Clinician Judgement of Symptoms (for gait disturbance, falls and verbal communication); Functional Activities Questionnaire (FAQ, for transaction skills and meal preparation); and the CDR plus NACC FTLD, for verbal communication and self-care.

Data from the Neuropsychiatric Inventory Questionnaire (NPI-Q) captured the following neuropsychiatric states: depression, anxiety, psychosis, apathy, disinhibition, irritability, and sleep disturbance/nighttime behaviors. Medical comorbidity was defined using variables for diabetes, hypertension, hypercholesterolemia, and parkinsonism. Functional status and comorbidities were assessed at each visit. Visits within the dataset were spaced at approximately one year time intervals. Functional, neuropsychiatric, and medical states were coded as binary variables, as the objective was to describe disability transitions using survival analyses.

Pathology data were available for a subset of the subjects. Frontotemporal lobar degeneration (FTLD) pathology was defined by presence of tau, TDP-43, FUS, ubiquitin, or p62 positive inclusion bodies. AD pathology was based on Thal phase for amyloid plaques, Braak stage for neurofibrillary degeneration, density of neocortical neuritic plaques, AD neuropathologic change score and density of diffuse plaques.

Statistical Methods

We report descriptive statistics, for sample characterization and the distribution of disability over time. Cox regression models were used to determine the extent to which medical and psychiatric comorbidities were associated with the probability of specific disabilities over time. Statistical significance was based on p values <0.05. We used STATA SE 15 (StataCorp LP, College Station, TX) for all analyses.

RESULTS

Demographics

Sample characteristics are shown in Table 1. The subjects were 97.9% white, and the majority were women (56.5%). Mean age was 68.8 years, and there was no difference between the mean age of the men and that of the women. About 46% of subjects were known to have died within the study period, with more women dying than men. FTD or Alzheimer disease (AD) pathology data were available for 75% of study subjects who died; 39% of these had only FTD pathology, 39% had both FTD and AD pathology, and 21% had only AD pathology. Average years of education was 15.5 and was higher for men than for women with men averaging 16.4 years and women averaging 14.9 (p <0.001). Men had higher rates of apathy, and of hypertension, hypercholesterolemia and parkinsonism.

TABLE 1.

Baseline Demographics

Frontotemporal Dementia With initial CDR = 0(n = 91)	Men	Women	All	p Valuea
Sex (% male)	43.5	56.5	—	—
Age (years)	68.0	69.4	68.8	0.11
Race (% white)	100	96.2	97.9	0.01
Education (years)	16.4	14.9	15.5	<0.001
Died (% known to have died)	36.8	52.4	45.6	0.003
Reason for study entry				0.44
Participate in research study (%)	45.4	43.4	44.3
Seeking clinical evaluation (%)	54.6	55.7	55.2
Participate in research and seeking clinical evaluation (%)	0.0	0.94	0.53
Medical comorbidities
Diabetes	1.8	0.5	1.1	0.06
Hypertension	15.3	9.0	11.7	0.03
Hypercholesterolemia	15.3	8.0	11.2	0.02
Parkinsonism	3.1	9.9	6.9	0.03
Psychiatric comorbidities
Psychosis	4.3	0.94	2.4	0.08
Depression	31.9	40.1	36.5	0.15
Anxiety	30.7	27.8	29.1	0.64
Apathy	40.5	24.5	31.5	0.004
Disinhibition	20.9	17.0	18.7	0.30
Irritability	33.1	35.9	34.7	0.24
Sleep disturbance	25.8	17.5	21.1	0.08

Notes: Bolded values indicated statistical significance, i.e., p <0.05.

^ap value derived from t test for continuous variables and chi-square test of categorical variables.

Functional Disabilities

Table 2 shows the distribution of functional disabilities at successive visits throughout the course of illness. A large majority (93%) had impairments of language and verbal communication at the first visit, whereas impairments in other domains were uncommon—and none had impairments in self-care. At the subsequent visits, impairments in other domains were increasingly common, and communication impairment was universal by the fourth visit. A majority of subjects with a CDR score of zero at visit 1 had a score of one or higher by visit 5; 19% had a score of 3 (Table 3). Seventy-one percent of subjects had a CDR plus NACC FTLD score of 0.5 or higher at visit 1. By visit 5, 24% of subjects had a score of 3 (Table 4).

TABLE 2.

Disability Progression Overtime in PPA

Frontotemporal Dementia (Initial CDR = 0) Disability	Visit Number
	1	2	3	4	5
Ambulation	3.3	13.9	21.5	31.4	40.5
Transactional	6.6	29.1	50.8	66.7	70.3
Language	93.4	96.2	98.5	100.0	100.0
Self-care	0.0	7.6	16.9	31.4	40.5
Meal Preparation	2.2	15.2	24.6	41.2	48.7
Total patients	91	79	65	51	37

Data in percentages using the total number of patients not lost to follow-up as the denominator.

TABLE 3.

CDR Progression Over Time

Frontotemporal Dementia (Initial CDR = 0) CDR Score	Visit Number
	1	2	3	4	5
0	100	36.7	23.1	7.8	2.7
0.5	0	51.9	56.9	52.9	46.0
1	0	6.3	12.3	17.7	18.9
2	0	2.5	1.5	5.9	13.5
3	0	2.5	6.2	15.7	18.9
Total patients	91	79	65	51	37

Data in percentages showing the proportion of patients not lost to follow-up that fall into each CDR score category at each subsequent visit.

TABLE 4.

CDR FTLD Progression Over Time

Frontotemporal Dementia (Initial CDR = 0) CDR FTLD Score	Visit Number
	1	2	3	4	5
0	1.1	1.3	1.5	0	0
0.5	59.3	46.8	30.8	23.5	8.1
1	11.0	31.7	40.0	39.2	27.0
2	1.1	6.3	16.9	19.6	40.5
3	0	5.1	9.2	17.7	24.3
Missing	27.5	8.9	1.5	0	0
Total patients	91	79	65	51	37

Data in percentages showing the proportion of patients not lost to follow-up that fall into each CDR score category at each subsequent visit.

Comorbidities

Figure 1 and Table 4 show associations between neuropsychiatric states and functional disabilities at subsequent visits. Psychosis was associated with impairments in self-care (hazard ratio [HR] 5.0, confidence interval [CI] [2.4–10.6], p <0.001), and meal preparation (HR 3.8, CI [2.0–7.3], p <0.001) at the subsequent visit. Depression and anxiety were associated with ambulation impairment at the subsequent visit (HR 1.6, CI [1.0–2.4], p = 0.05 and HR 1.8, CI [1.2–2.8], p = 0.009, respectively). Anxiety was also associated with transactional and meal preparation impairments (HR 1.8, CI [1.3–2.6], p = 0.001 and HR 1.7, CI [1.2–2.6], p = 0.006, respectively). Apathy was associated with ambulation and meal preparation impairments (HR 2.5, CI [1.4–4.2], p = 0.001 and HR 3.9, CI [2.5–6.3], p < 0.001, respectively). Sleep disturbance was associated with impairments in transactional skills (HR 1.6, CI [1.1–2.3], p = 0.009), verbal communication (HR 1.5, CI [1.0–2.2], p = 0.04), self-care (HR 2.0, CI [1.3–3.1], p = 0.002) and meal preparation (HR 1.7, CI [1.1–2.6], p = 0.02).

FIGURE 1.

Association of psychiatric comorbidities and functional disability at visit prior to disability onset.^{† †}Cox proportional hazard ratios adjusted for age, sex, and education. *Statistically significant at p <0.05.

Associations of medical conditions with functional disability at subsequent visits are shown in Table 5. The development of parkinsonism was associated with impairments in ambulation and meal preparation at the subsequent visit (HR 3.9, CI [2.3–6.6], p <0.001 and HR 3.8, CI [1.2–12.2], p = 0.03, respectively). Hypertension was associated with self-care impairment (HR 2.3, CI [1.2–4.5], p = 0.02) at the subsequent visit, and hypercholesterolemia with transactional impairments (HR 2.2, CI [1.3–3.7, p = 0.004; Table 6).

TABLE 5.

Association of Psychiatric Comorbidities and Functional Disability at Visit Prior to Disability Onset

	Ambulation		Transactions		Language		Self-Care		Meal Prep
	HR[95% CI]	p Value	HR[95% CI]	p Value	HR[95% CI]	p Value	HR[95% CI]	p Value	HR[95% CI]	p Value
Psychosis	1.3 [0.5–3.8]	0.58	1.9 [0.6–6.3]	0.29	1.0	—	5.0a [2.4–10.6]	<0.001	3.8a [2.0–7.3]	<0.001
Depression	1.6a [1.0–2.4]	0.05	0.98 [0.7–1.4]	0.9	0.7a [0.6–0.9]	0.006	1.0 [0.6–1.6]	0.94	1.2 [0.8–1.9]	0.33
Anxiety	1.8a [1.2–2.8]	0.009	1.8a [1.3–2.6]	0.001	0.9 [0.7–1.1]	0.33	1.6a [1.0–2.6]	0.03	1.7a [1.2–2.6]	0.006
Apathy	2.5a [1.4–4.2]	0.001	1.3 [0.9–1.7]	0.14	1.2 [0.9–1.7]	0.17	1.4 [0.9–2.3]	0.11	3.9a [2.5–6.3]	<0.001
Disinhibition	0.53a [0.29–1.0]	0.04	0.9 [0.6–1.3]	0.63	1.1 [0.7–1.6]	0.79	0.75 [0.5–1.2]	0.22	0.7 [0.5–1.2]	0.21
Irritability	1.0 [0.64–1.6]	0.98	1.0 [0.8–1.4]	0.81	0.9 [0.7–1.1]	0.4	0.7 [0.4–1.1]	0.2	1.2 [0.8–1.9]	0.36
Sleep disturbance	1.2 [0.75–1.9]	0.43	1.6a [1.1–2.3]	0.009	1.5a [1.0–2.2]	0.04	2.0a [1.3–3.1]	0.002	1.7a [1.1–2.6]	0.02

Notes: Model adjusted for age, sex, education.

Bolded values indicated statistical significance, i.e., p <0.05

^aStatistically significant at p <0.05.

TABLE 6.

Association of Medical Comorbidities and Functional Disability at Visit Prior to Disability Onset

	Ambulation		Transactions		Language		Self-Care		Meal Prep
	HR[95% CI]	p Value	HR[95% CI]	p Value	HR[95% CI]	p Value	HR[95% CI]	p Value	HR[95% CI]	p Value
Diabetes	1.0	—	1.0	—	1.0	—	1.0	—	1.0	—
Hypertension	0.84 [0.36–2.0]	0.68	1.3 [0.81–2.1]	0.27	1.0	—	2.3a [1.2–4.5]	0.02	0.88 [0.43–1.82]	0.73
Hypercholesterolemia	0.77 [0.36–1.6]	0.49	2.2a [1.3–3.7]	0.004	1.0	—	1.2 [0.69–2.2]	0.49	2.7 [0.64–11.0]	0.18
Parkinsonism	3.9a [2.3–6.6]	<0.001	0.97 [0.55–1.7]	0.93	1.5 [0.79–2.9]	0.21	1.9 [0.95–3.9]	0.07	3.8a [1.17–12.2]	0.03

Notes: Model adjusted for age, sex, education.

Bolded values indicated statistical significance, i.e., p <0.05.

^aStatistically significant at p <0.05.

DISCUSSION

We examined the chronology of functional disabilities in PPA and found, as expected, that impairments in verbal communication are ubiquitous at the beginning of the illness. Impairments in other domains are increasingly common from the second visit on. We observed a hierarchy in the frequency of functional disabilities at successive visits that, in general, follow the order: verbal communication > transaction > meal preparation > self-care > ambulation. The early development of verbal communication impairments fits the expected clinical course of PPA, as language deficits define the condition and its presentation.¹⁷ The subsequent development of transactional impairments, however, represents a less obvious but important finding. Executive dysfunction has been shown to be a prominent feature in PPA and impairments in transaction skills are a common sign of early executive dysfunction. Executive dysfunction is correlated with frontotemporal atrophy in PPA,^18,19 and is not uncommon early in the clinical course.²⁰ In other words, impairments in transaction skills may serve as a behavioral marker of executive dysfunction. The early development of transactional impairments carries with it the potential for frequent and potentially serious financial missteps. Awareness of this impairment and the associated risks, on the part of clinicians, patients, family, and carers affords the opportunity for safe-guards and prompt intervention.

The eventual development of impairments in meal preparation, ambulation, and self-care is consistent with the gradual devolution of PPA from a focal syndrome to a dementia state characterized by “across-the-board” cognitive and functional impairments.^13,21,22 From a care perspective, explication of the chronology of disability highlights intervention opportunities. Rehabilitative services such as physiotherapy and occupational therapy, alongside environment manipulations, can preserve functioning and life participation, particularly when applied early.²³ Early physiotherapy and environment modifications are especially valuable for those cases complicated by parkinsonism, which is not uncommon in PPA—where it may present as an akinetic-rigidity that has poor response to levodopa treatment.²⁴ Parkinsonism has a strong association with imbalance, gait disorder and falls, and early and regular utilization of physiotherapy improves motor function and reduces risks for falls and injury.

Earlier studies have described neuropsychiatric states in PPA,^25-29 but have not linked these to the development and progression of functional disability. In this study, depression and anxiety were the most common neuropsychiatric states; 32% of patients had depressive symptoms at baseline, and 24% had anxiety symptoms. These frequencies are consistent with estimates of depression and anxiety in PPA patients reported in earlier studies, where depression rates varied from 38%–56% and anxiety rates from 15% to 50%.^27,30 In our study, depression was associated with ambulation impairments at the subsequent visit, while anxiety was associated with the development of impairments in ambulation, transactions, self-care, and meal preparation. The basis for these associations is not clear, though they may indicate that depression and anxiety are signals of progressing disease. Regardless of the explanation, these findings suggest that treating anxiety and depression in PPA may be valuable in improving quality of life and mitigating the impact of neurodegeneration on functional status. There has not been much study of depression or anxiety treatment in PPA.³¹ A review summarized evidence that SSRI antidepressants are safe and effective for treating depression in FTD,³² which suggests these agents can be considered for depression and anxiety in PPA. Some reports have documented success with trazodone treatment, and stimulants, in ameliorating the behavioral symptoms of some FTD cases.³² Behavioral therapies, which have a lower risk profile, may also ameliorate depression and anxiety in PPA—but this has not been systematically evaluated. Given the high prevalence of depression and anxiety in PPA, and their associations with functional disability, it is reasonable to propose early psychotropic and psychotherapeutic treatments in order to improve quality of life in the patients.

Sleep disturbance was common in the study population, with 21% reporting sleep-related symptoms. Trazodone is a reasonable treatment for sleep disturbances given the efficacy data from FTD and post-stroke studies.³² There is also evidence from the stroke literature that sleep disturbances impair recovery after stroke, and their treatment facilitates functional recovery.³³ In other words, it is possible that improving sleep quality in PPA facilitates function. Cognitive Behavioral Therapy, well established for insomnia treatment, has been shown to be effective in post-stroke and dementia related sleep disorders and is therefore a reasonable consideration for the PPA population.^34,35

While less common than depression and anxiety, psychosis is not a rare occurrence in PPA as prevalence estimates are around 5%.^27,36 In our study, 2% exhibited symptoms of psychosis, substantially less frequent than depression and anxiety but still significant. Psychosis was strongly associated with subsequent impairments in self-care and meal preparation, which has significant implications on quality of life and carer burden. The data supporting the use of neuroleptics for psychosis in PPA, and in FTD generally, are mixed.³⁷ Antipsychotics may help ameliorate agitated or dangerous behavior, but is not recommended as first-line treatment due to an especial risk for extrapyramidal side-effects in patients who have dementia-related parkinsonism, and the general risk for all-cause mortality in elders.^38,39 SSRIs can ameliorate agitation in FTD patients and are a safer alternative.^38,40 The more recently developed pimavanserin can be used safely in patients who have parkinsonism.⁴¹ Antiepileptics and acetylcholinesterase inhibitors and memantine have also been studied in FTD populations, but efficacy has not been demonstrated.³² The development of medications for psychosis, agitation, aggression, and related states in PPA and FTD remains an open area for research.

While our findings suggest that treatment of neuropsychiatric states in PPA may mitigate functional decline, the proposal would be hypothetical as we have not analyzed data on treatment outcomes. We would recommend clinical trials, of psychotropic medicines and behavioral therapies to evaluate the effect of treating neuropsychiatric symptoms on functional outcomes in PPA.

We observed heterogeneity in the tempo of disease progression, in that time from progression from a low CDR rating (0 or 0.5) to higher ratings (2 or 3) varied considerably. Nearly half of the subjects had CDR ratings lower than 1 at visit 5 despite developing significant functional disabilities by visit 5. The CDR is known to underestimate illness severity in PPA and other FTD syndromes, primarily because it does not account for impairments in language, social cognition, behavior, and motor function.⁴² The CDR plus NACC FTLD,^43,44 designed to address this problem, adds to the standard CDR assessments of language and behavior domains. Our comparison of the CDR plus NACC FTLD scores to the CDR scores demonstrates, as expected, the underestimation of severity by the latter instrument. Whereas all subjects had a CDR global score of zero at the first visit, most subjects (71%) had CDR plus NACC FTLD ratings higher than one. By visit 5, CDR plus NACC FTLD scores were high (2 or 3) for 65% of the patients, while 32% of patients had standard CDR scores of 2 or 3 at visit 5. In other words, CDR plus NACC FTLD scores are more sensitive indicators of early illness and better estimators of illness severity in PPA than CDR scores.

We have used a large multisite database of subjects with PPA, which facilitated exploration of associations. Limitations include the use of a retrospective cohort and reliance on clinical diagnosis of PPA rather than on pathologically confirmed cases. However, in our study population the majority of cases for whom autopsy data were available had FTLD pathology. Significant numbers of these had mixed FTD/AD pathology and most of the rest had AD pathology. It is to be noted that clinicopathologic studies have demonstrated that two of the three PPA variants, PNFA and SD, are typically associated with FTLD pathologies, whereas logopenic PPA is typically associated with AD pathology.^45,46 Our findings highlight the clinical and pathological heterogeneity of PPA and the challenging nature of the clinical diagnosis—owing to variable clinical presentations, the PPA variants sharing of characteristics, and ambiguous features such as apraxia of speech.^45,46 In considering the findings of this study, it is important to note the emphasis is on the clinical diagnosis as the focus is a clinical care question; i.e., describing probabilities of functional disabilities associated with the phenotype. Finally, we note that while of substantial size, the study population was predominately white and findings may not generalize to other racial/ethnic groups. We have not adjusted for multiple comparisons, as the number is not particularly large and this is a preliminary study.

CONCLUSION

This study provides preliminary insights into the chronology of functional disabilities in PPA, and the predictive associations between psychiatric symptoms and functional disabilities in this population. While preliminary, the findings point to opportunities to promote quality of life with psychiatric interventions. These findings will need to be buttressed and expanded upon in studies that provide the data for conditional probabilities for functional impairments on which to base personalized forecasts. Work of this type has the potential to open new opportunities for early intervention in PPA care that are focused on nonpharmacologic and pharmacologic interventions to forestall functional disabilities. While PPA is an inexorably progressive neurodegenerative condition, we can seek opportunities to preserve quality of life, mitigate patient and carer hardship, and decrease the societal costs of the illness. This study is our preliminary step in a vision to provide patients, families and their physicians (or other health providers) with data-based predictions for care management, with the objective of facilitating timely interventions that optimize care and quality of life.

Highlights.

• What is the primary question addressed by this study?

  We describe the chronology of functional disabilities in primary progressive aphasia and examine the temporal association of psychiatric comorbidities on disability progression.

• What is the main finding of this study?

  Impairments in communication occur early in the course of primary progressive aphasia, followed closely by impairments in transactions while impairments in self-care, meal preparation, and ambulation occur later. Several psychiatric comorbidities including depression, anxiety, psychosis, and sleep disturbances are temporally associated with the development of new functional disabilities in primary progressive aphasia.

• What is the meaning of the finding?

  This study provides new insights into the chronology of functional disabilities in primary progressive aphasia and highlights the association between psychiatric symptoms and subsequent functional disabilities in this population.