1 Introduction

Primary progressive apraxia of speech (PPAOS) is a form of frontotemporal lobar degeneration defined by the insidious onset of disturbances in motor speech planning or programming in the absence of other initial difficulties [1-4]. Neuropathologically, PPAOS is typically associated with four-repeat tauopathies, either corticobasal degeneration (CBD) or progressive supranuclear palsy (PSP) pathology [5]. This is likely the reason that many individuals with PPAOS later develop features of atypical parkinsonian conditions, including PSP and corticobasal syndrome [6-10]. Dysphagia, or difficulties with chewing and swallowing, is well-established as a common symptom that can be associated with mortality in individuals with atypical parkinsonian conditions [11-15]. Fatal complications include choking and aspiration pneumonia [16]. Using pathology-defined groups, Müller and colleagues found that 83% of individuals who had PSP pathology and 31% of individuals with CBD had subjective swallowing difficulties, with onset latencies of 42 and 64 months, respectively [11]. Latency to dysphagia was longer than the latency to dysarthria in nearly every participant and was highly correlated with survival time. Similarly, dell'Aquila and colleagues found that early dysphagia symptoms shortened survival time for individuals who presented clinically with PSP [17]. Josephs et al. reported that six of their 13 participants with PPAOS had developed dysphagia (median disease duration of 6.9 years for the full group) and that most of them had also developed dysarthria in the context of a rapidly evolving clinical picture that resembled PSP (note that there is an overlap between the cohort reported by Josephs et al. and the participants included here) [18]. Nevertheless, the relationship between dysphagia and other symptoms (e.g., parkinsonism, dysarthria) and its impact on mortality in PPAOS are still unknown.

2 Methods

2.1 Participants

Retrospective data are presented from 56 participants who were recruited by the Neurodegenerative Research Group (NRG) at Mayo Clinic Rochester (see Table 1). Written informed consent was obtained from all participants at each visit in accordance with the Mayo Clinic IRB. Median age at enrollment was 72.6 years (IQR: 14.4 years). All participants had a PPAOS diagnosis at enrollment, meaning that they had unequivocal apraxia of speech (AOS) in the absence of broader language, cognitive, or neurologic symptoms that would meet criteria for another diagnosis. Consensus diagnoses were reached independently for each visit following a comprehensive evaluation by a board-certified neurologist, certified speech-language pathologist, and psychometrist. At each visit, the presence or absence of dysphagia symptoms was also determined by asking the participant or study partner if they had noticed any difficulties with chewing or swallowing. Clinical data for the first visit at which they endorsed dysphagia symptoms is presented in Table S1. Thirty-six (64%) of the participants had died prior to the data freeze (see Figure 1). Cause of death was available for 20 (55%) of these individuals, established by reviewing medical charts and family member reports.

TABLE 1. Demographic and assessment results for the enrollment visit as a function of dysphagia symptoms.

	No Dysphagia (n = 44)	Dysphagia (n = 12)	p
Age at AOS onset (Years)	68 (60, 73)	72 (62, 76)	0.412
Females	23 (52%)	7 (58%)	0.963
Education (Years)	16 (15, 18)	16 (13, 17)	0.696
Disease duration at enrollment	2.8 (1.8, 4.5)	4.0 (3.0, 4.9)	0.137
AOS severity (/4)	1.25 (1.00, 2.00)	2.00 (1.50, 3.00)	0.013
AOS subtype
Phonetic	14 (32%)	3 (25%)	0.404
Prosodic	23 (52%)	5 (42%)
Mixed	7 (16%)	4 (33%)
Dysarthria	7 (16%)	6 (50%)	0.036
Dysarthria type
UUMN	0 (0%)	1 (17%)	0.721
Spastic	4 (57%)	3 (50%)
Hypokinetic	1 (14%)	1 (17%)
Hyperkinetic	1 (14%)	0 (%)
Mixed	1 (14%)	1 (17%)
WAB aphasia quotient (/100)	97.6 (96.6, 99.2)	98.1 (96.5, 98.9)	0.992
MoCA (/30)	27 (25, 28)	28 (26, 29)	0.179
FAB (/18)	17.00 (16.00, 17.00)	17.00 (16.00, 17.25)	0.562
MDS-UPDRS III (/132)	6 (4, 12)	13 (12, 17)	0.007
WAB praxis (/60)	58.00 (57.75, 60.00)	58.50 (57.00, 59.00)	0.331

• Note: Data reflect median (first quartile, third quartile) or n (%). Disease durations were calculated from initial symptom (i.e., AOS) onset, as reported by the participant and care partner, to the time of enrollment. Bolded text indicates significant differences at the p < 0.05 level.
• Abbreviations: ASRS-3 = Apraxia of Speech Rating Scale – Version 3, FAB = Frontal Assessment Battery, MDS-UPDRS III = Movement Disorders Society Sponsored revision of the Unified Parkinson's Disease Rating Scale Part III, MoCA = Montreal Cognitive Assessment, WAB = Western Aphasia Battery – Revised.

3 Results

3.2 Multistate Model

Of the 36 participants who died, 31 had documented evidence of dysphagia prior to dying (see Figure 1). The remaining five were lost to follow-up several years before death; therefore, it is unclear if they developed dysphagia later in their disease course. For this reason, the transition intensity between the alive without dysphagia state and death approached zero. The mean sojourn time in the no dysphagia state was 5.98 years (before transitioning to the dysphagia or death states; 95% CI: 4.45–8.05 years) and the mean sojourn time in the dysphagia state was 5.43 years (before transitioning to the death state; 95% CI: 3.89–7.60 years). Model estimates for the cumulative probability of transitioning from and to each state can be found in Figure 2.

4 Discussion

Although the initial difficulties experienced by individuals with PPAOS are limited to motor speech planning and programming, they experience a broad range of symptoms as the disease progresses. Before dying, 31/36 (86%) of this cohort endorsed dysphagia; the remaining five may have developed dysphagia but were lost to follow-up. The cross-sectional comparison between participants who enrolled in the research study with dysphagia symptoms versus without dysphagia symptoms demonstrated that swallow function is associated with motor speech function. Based on previous research, we expected that almost all participants who endorsed dysphagia symptoms would have dysarthria, whereas only 60% of them did [11, 13]. Across participants, dysphagia complaints included both the oral (e.g., slow, inefficient chewing) and pharyngeal (e.g., coughing on liquids or secretions) phases in addition to gorging (e.g., eating or drinking too much and/or too quickly). The latter may have contributed to the endorsement of dysphagia symptoms in participants who did not have neuromuscular features that would have also caused dysarthria, but the nature of dysphagia warrants further investigation. AOS severity and motor subtest scores on the MDS-UPDRS also differed between groups, which could both be indicators of a slight difference in disease severity between groups at enrollment.

The multistate model has the benefit of drawing on the full longitudinal dataset to inform prognostication over time. Model estimates suggest that the average time of being dysphagia-free after initial symptom onset is 5.98 years, and the average time of living with dysphagia is 5.43 years. This model does not establish causality. However, the cause of death data suggest that complications from dysphagia contributed to mortality for 35% of the participants for whom those data were available. Most, if not all, of them were making diet modifications, but only six had documentation of receiving enteral nutrition. Survival past feeding tube placement was never longer than 8 months. These results have direct implications for clinical decision-making, education, and prognostication.

For the first time, these results underline the importance of monitoring for dysphagia in individuals with PPAOS, but a few limitations warrant consideration. For one, the timelines are limited to data from annual visits and retrospective chart review, which limits precision, but these data are well-suited for a multistate model. As with any longitudinal study, information about early symptom development (e.g., prior to diagnosis and study enrollment) is especially sparse and the presented timelines rely on estimates of symptom onset. Only indexing dysphagia symptoms by self-report could also be considered a limitation, but this has the benefit of being easily translated to the clinic and implemented by any member of the care team. Future research would benefit from the addition of instrumental measures of swallowing because the extent to which subjective reports accurately reflect objective findings is unclear [26-29]. Instrumental studies would also allow for trials of dysphagia management strategies that might mitigate the risk of fatal complications.

Author Contributions

G.M.: study design, data collection, statistical analysis, writing of the first draft, editing of the manuscript; N.T.T.P.: statistical analysis, editing of the manuscript; S.M.B.: data collection, editing of the manuscript; H.M.C.: data collection, editing of the manuscript; J.R.D.: data collection, editing of the manuscript; J.L.W.: editing of the manuscript; H.B.: statistical analysis, editing of the manuscript; K.A.J.: editing of the manuscript; R.L.U.: study design, data collection, statistical analysis, editing of the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.