Dysarthria following acute ischemic stroke: Prospective evaluation of characteristics, type and severity
Background
Dysarthria is a common symptom after stroke and a significant source of functional impairment. Estimates of prevalence in the acute phase range from 25\% to 70\%, with about 42\% showing persistent symptoms at 3 months post-stroke. These figures come from multiple studies cited in the paper (e.g., Ali et al. 2015; Lubart et al. 2005; De Cock et al. 2020).
Dysarthria affects psychosocial well-being and daily functioning, highlighting the need for better characterization to aid differential diagnosis and healthcare resource distribution.
Auditory–perceptual assessment is considered the gold standard for evaluating speech characteristics, typically examining subsystems: phonation, resonance, respiration, prosody, and articulation. Common post-stroke features include: artic- ulation inaccuracy, imprecise consonant articulation, harsh/hoarse/ breathy voice qualities, monopitch, loudness changes, and speech pauses.
The Mayo Classification System differentiates dysarthria types (Unilateral Upper Motor Neuron (UUMN), Flaccid, Spastic, Ataxic, Hypokinetic, Hyperkinetic, Mixed) based on lesion location and observed characteristics.
Prior research suggested limited data on post-stroke dysarthria, especially with objective acoustic measures and standardized severity ratings. Some previous work indicated UUMN as a common type after stroke, but evidence was sparse and heterogeneous.
Aims
Primary aim: prospectively evaluate speech characteristics, dysarthria type, and severity in the acute phase following a first-ever ischemic stroke in a comprehensive stroke centre.
Specific objectives:
(1) Assess speech characteristics using auditory–perceptual assessment and objective acoustic measures.
(2) Determine dysarthria type according to the Mayo Classification System.
(3) Evaluate dysarthria severity at function and activity levels using standardized measures and assess early recovery.
Methods
Design: prospective clinical study conducted between March 2018 and October 2019 at Ghent University Hospital Stroke Unit.
Participants: consecutive patients with a first-ever ischemic stroke, admitted within 48 hours of symptom onset, Dutch-speaking, age \ge 18. Exclusions included a history of other diseases affecting swallowing, speech, or language (e.g., dementia, Parkinson’s disease, oral carcinoma) and severe cognitive impairment.
Data collected: demographic information (age, sex, handedness), TOAST stroke etiology, lesion location (MRI/CT), and treatment with acute reperfusion therapy. Ethics approval was obtained and informed consent was collected.
Dysarthria screening (screening for presence of dysarthria):
Performed within 3 days after admission by one of two trained speech–language pathologists (SLPs) using bedside dysarthria screening.
Screening procedures included: non-speech oral motor exercises to evaluate morphology, strength, sensibility, and function of oral musculature; repetition of words with increasing length and complexity; spontaneous speech elicited via open questions to assess respiration, phonation, resonance, prosody, articulation.
If dysarthria was identified, a standardized detailed assessment was performed; all participants also had NIHSS assessment on hospital arrival.
NIHSS sub-item for speech was used to quantify initial severity; assessments occurred at hospital admission, day 3 ± 2, and day 7 ± 2 (or earlier if discharged).
Dysarthria assessment (participants with confirmed dysarthria):
Recordings: audio recordings using an over-ear headset with microphone by one of the two SLPs; standardized assessments performed by E.D.C. or K.B.
Perceptual assessment: subsystems and characteristics evaluated as described by Darley et al. (phonation, articulation, resonance, respiration, prosody).
Acoustic measurements: AMR and SMR, pitch (minimum, maximum, range), maximum loudness, and Maximum Phonation Time (MPT) using Praat software.
Dysarthria type: assigned per Mayo Classification System.
Severity at function and activity levels: six-point RDA scale (0 = no impairment, 5 = very severe impairment).
Speech intelligibility: Dutch Sentence Intelligibility Assessment (DSIA) – participants read randomly generated sentences; transcripts by a blinded SLP; intelligibility reported as percent correct words. If DSIA could not be performed, intelligibility was rated perceptually on a visual analog scale based on spontaneous speech recordings.
Inter-rater process: dysarthria type independently evaluated by both SLPs; disagreements resolved by consensus.
Data analysis:
Descriptive analyses of demographic, clinical, and speech variables conducted in SPSS version 26.
Normality testing influenced whether data were presented as means ± SD or medians with IQR.
Participants without audio recordings excluded from standardized assessment analyses.
Between-group comparisons used Mann–Whitney U-test or independent t-test (continuous variables) and Fisher’s exact test (categorical variables); significance set at p<0.05.
Objective acoustic measures were compared to normative values for healthy Dutch-speaking individuals to identify impairment per parameter (score below the cut-off). Normative references cited include Knuijt et al. 2017 and related sources.
Data availability: data available from corresponding author on reasonable request.
Results
Participants and dysarthria prevalence
Dysarthria present in 67 of 151 first-ever ischemic stroke patients (prevalence 44\%).
Age: mean 69 years (SD 13); females 28/67 (42%).
Median NIHSS on hospital admission: 4 (IQR 2-14).
Acute reperfusion therapy received by 36/67 (54\%) of dysarthric participants:
Thrombolysis only: 20/67 (30%),
Thrombectomy only: 8/67 (12%),
Both thrombolysis and thrombectomy: 8/67 (12%).
Stroke etiology (TOAST) among dysarthric patients: Large artery atherosclerosis 8/67 (12%), Cardio-embolism 18/67 (27%), Small vessel occlusion 6/67 (9%), Other determined source 4/67 (6%), Undetermined 31/67 (46%).
Lesion location: Supratentorial 54/67 (82%); Left hemisphere 27/54 (50%), Right hemisphere 21/54 (39%), Bilateral 6/54 (11%); Infratentorial 7/67 (11%); Mixed 5/67 (7%).
Phase where more detailed data were possible: standardized assessments and audio recordings were feasible in 48/67 (72%) of dysarthric participants.
Reasons for missing audio recordings (n=19): severe aphasia/muteness (n=9), reduced alertness (n=5), in-hospital complications (n=2), rapid transfer to another hospital (n=1), severe verbal apraxia (n=1), participant refused (n=1).
Dysarthric participants without audio recordings tended to be older and had more severe strokes compared with those with audio data (p-values not specified; significant for age and stroke severity).
Auditory–perceptual and objective speech characteristics (n=48 with audio)
Subsystems affected (ladder of impairment among participants):
Articulation: Imprecise consonants observed in 44/48 (92%).
Phonation: Harsh voice observed in 46/48 (96%); hoarse voice in 23/48 (48%); strained-strangled voice in 25/48 (52%).
Respiration: Audible inspiration in 38/48 (79%). Grunt at end of expiration in 2/48 (4%).
Resonance: Hypernasality in 17/48 (35%); hyponasality in 4/48 (8%); nasal emission in 6/48 (13%).
Prosody: Monopitch in 13/48 (27%); rate abnormalities in 33/48 (69%), including short rushes of speech (4/48, 8%), prolongations and pauses; increased/decreased rate in various segments; inappropriate silences and reduced/variable stress; overall more variable prosody contributions.
Amplitude/volume metrics: Maximum loudness and MPT deviated most from normative values.
Diadochokinesis: AMR impaired more often than SMR; AMR/pa: median 5.4 syllables/s (IQR 4.6-6.0) with 26/46 (56.5%) impaired; AMR/ta: 5.4 (IQR 4.4-6.0), 19/43 (44.2%) impaired; AMR/ka: 5.0 (IQR 4.0-5.4), 19/45 (42.2%) impaired; SMR/pataka: 5.6 (IQR 4.7-6.4), 15/46 (32.6%) impaired; MPT median 9 s (IQR 3-13) with 28/44 (64%) impaired.
Overall, perceptual findings of articulation impairment (imprecise consonants) and harsh/strained phonation with audible respiration aligned with known features of UUMN dysarthria.
Dysarthria type (n=48 with audio)
Unilateral upper motor neuron (UUMN) dysarthria was the most prevalent type, present in 25/48 (52\%).
The distribution of dysarthria types is shown in Figure 1 of the paper (not reproduced here).
Dysarthria severity and recovery (n=70 NIHSS speech assessments across time; n=48 with audio for level severity)
Severity by NIHSS sub-item speech at hospital admission: among all dysarthric participants with NIHSS data, 47/70 (67\%) had mild to moderate impairment at admission.
Recovery within 1 week: half of dysarthric participants recovered completely within 1\ week after symptom onset; specifically, 35/70 (50\%) showed complete recovery within 7 days.
Functional impairment (RDA) at function level: 28/48 (approximately 58\%) had no/minimal/mild impairment; function category breakdown indicates the majority were not severely impaired at the functional level.
Activity level impairment (RDA): 34/48 (approximately 71\%) had no/minimal/mild impairment; still, a substantial portion had more than mild impairment at activity level.
Speech intelligibility: median perceptual intelligibility on a visual analogue scale (VAS) was 87\% (IQR 70-96) for the subset with VAS data (n=48); DSIA measured intelligibility at 91\% (IQR 73-97) in the subset with DSIA data (n=39).
Longitudinal interpretation at three time points (Figure 2 and Figure 3 in the paper; described in text):
Figure 2 (dysarthria severity over time by NIHSS speech item) shows that a substantial portion had mild impairment at admission with improvement over time for some patients; overall, many improved within the first 3–7 days.
Figure 3 (RDA function and activity levels) demonstrates that the majority had no/minimal/mild impairment at both function and activity levels in the early phase, corresponding with the observed high rate of early recovery in a subset of patients.
Comparisons and context with prior work
The study notes a relatively high prevalence of dysarthria in the acute ischemic stroke population studied (consisting largely of patients treated in a high-resource, reperfusion-era center).
The observed speech characteristics (articulation deficit, harsh phonation, audible inspiratory effort) and the predominance of UUMN dysarthria are consistent with the idea that left-hemispheric dysarthria may be prominent and that tract involvement affecting articulation, phonation, and respiration underlies many cases.
The authors compare their findings to earlier work, noting similar patterns in articulation and phonation deficits but acknowledging differences in prevalence and severity due to changes in stroke care, population, and assessment methodologies.
Strengths and limitations
Strengths:
Prospective design with standardized assessment of speech characteristics and objective acoustic measurements in a contemporary stroke cohort.
Use of both perceptual (gold standard) and objective acoustic measures to characterize dysarthria and monitor recovery.
First study to prospectively examine dysarthria type in a consecutive group of solely acute ischemic stroke patients.
Limitations:
Standardized assessments and audio recordings could not be performed in 19/67 participants (≈ 28\%), who were older and had more severe strokes; this could bias results toward milder dysarthria.
Inter-rater reliability for dysarthria type was not formally reported beyond consensus resolution of disagreements.
The sample is relatively homogeneous (first acute ischemic stroke) and may not generalize to all stroke populations.
Conclusions and implications
Unilateral upper motor neuron (UUMN) dysarthria was the dominant type in this acute ischemic stroke cohort, and most patients had mild dysarthria.
About half recovered completely within the first week after onset, indicating substantial early recovery potential.
The speech impairment primarily affects articulation, phonation, and respiration subsystems, and objective acoustic measures corroborate perceptual findings.
Clinical implications include improved differential diagnosis (e.g., distinguishing from apraxia of speech or aphasia), more efficient allocation of healthcare resources, and the potential for early monitoring of progress using standardized and objective measures.
The authors advocate longitudinal research with standardized assessments to map recovery trajectories and to understand the long-term impact of post-stroke dysarthria on social participation.
What this paper adds (summary of the authors’ own box in the manuscript)
High prevalence of post-stroke dysarthria after acute ischemic stroke in a contemporary cohort.
Replication of the finding that dysarthria after stroke is predominantly characterized by imprecise articulation of consonants, harsh voice quality, and audible inspiration, with objective measures corroborating these features.
First evidence that UUMN is the most prevalent dysarthria type in an acute ischemic stroke cohort.
Most participants exhibit mild dysarthria, and a substantial proportion recover completely within the first week post-onset.
Objective acoustic parameters (e.g., AMR, MPT, maximum loudness) provide a quantitative basis for monitoring dysarthria and its recovery, supporting the use of such measures in clinical practice.
Early dysarthria assessment can inform target monitoring and resource allocation in the acute post-stroke period.
Introduction (context and rationale)
Post-stroke dysarthria is common, with reported prevalence up to 70\% acutely and 42\% persisting at 3 months (cited references: Ali et al. 2015; De Cock et al. 2020; Lubart et al. 2005).
Dysarthria is a key source of functional impairment and psychosocial impact after stroke (Brady et al. 2011; Dickson et al. 2008).
Auditory–perceptual assessment assesses abnormalities across speech subsystems (phonation, resonance, respiration, prosody, articulation). Common stroke-related perceptual features include articulation inaccuracy, imprecise consonants, harsh/hoarse voice, breathy voice, monopitch, and variable loudness and pauses.
Mayo Classification System distinguishes dysarthria types by lesion location and hallmark perceptual cues (e.g., UUMN vs other types).
Prior literature includes limited objective measurement of acoustic parameters and inconsistent reporting of dysarthria severity; there was a need for prospective data using standardized, objective measures to quantify dysarthria and recovery after acute ischemic stroke, with attention to potential shifts following reperfusion therapies.
Tables and figures (described in text)
Table 1: Demographic variables of dysarthric participants (n=67).
Table 2: Auditory–perceptual assessment of speech characteristics (subsystems and features) in dysarthric participants.
Table 3: Objective analysis of acoustic parameters (pitch range, maximum loudness, MPT, AMR/SMR values) with percentages impaired.
Figure 1: Dysarthria type distribution according to the Mayo Classification System (n=48 with audio; total n=48 in the figure).
Figure 2: Dysarthria severity over time according to the NIHSS speech sub-item at three time points (hospital admission, day 3 ± 2, day 7 ± 2; total n=70).
Figure 3: Dysarthria severity on function and activity levels (RDA) (total n=48).
Key takeaways for exam prep
Post-stroke dysarthria is common in the acute phase; early identification and characterization are crucial for prognosis, differential diagnosis, and resource planning.
In this contemporary, reperfusion-era cohort, UUMN dysarthria was the most frequent type, and the overall severity tended to be mild, with substantial spontaneous recovery within the first week for about half of patients.
Articulation and phonation subsystems are most affected, with respiration also playing a role (audible inspiration).
Objective acoustic measures (AMR/SMR, MPT, maximum loudness, pitch range) align with perceptual findings and provide a quantitative way to monitor changes over time.
Early standardized assessment, including language and speech evaluations, supports better monitoring and could inform rehabilitation planning and discharge decisions.
References (contextual)
The study cites prior work indicating post-stroke dysarthria incidence and outcomes (Ali et al. 2015; Urban et al. 2006; Chand-Mall & Vanaja 2017; Canbaz et al. 2010; Mackenzie 2011), and emphasizes the added value of objective acoustic measurements and standardized outcome measures in line with contemporary Cochrane reviews and diagnostic frameworks.
Limitations and caveats (summary)
About 28\% of dysarthric participants could not be assessed with audio recordings, introducing potential bias toward milder dysarthria since those with more severe strokes were overrepresented in the non-recorded group.
Inter-rater reliability for dysarthria type was not formally quantified beyond consensus after independent ratings.
Generalizability may be limited due to the relatively homogeneous, single-centre patient sample and the exclusion criteria.
Conclusions (brief)
UUMN dysarthria was the dominant type in this acute ischemic stroke cohort, with most participants displaying mild impairment. Approximately half recovered completely within the first week after onset. The speech impairment primarily involved articulation, phonation, and respiration, and objective acoustic measures supported perceptual findings. Longitudinal research is needed to track longer-term recovery and social participation outcomes, and to standardize assessments for routine clinical use.
Summarized Version
Post-stroke dysarthria is a common symptom following stroke, with prevalence estimates ranging from 25\% to 70\% in the acute phase and persisting in about 42\% at three months. It significantly impairs psychosocial well-being and daily functioning, underscoring the need for better characterization.
This prospective clinical study evaluated speech characteristics, dysarthria type, and severity in 151 patients with first-ever ischemic stroke. Dysarthria was present in 44\% (67 patients). Standardized assessments and audio recordings were feasible in 48 of these dysarthric participants (72%).
Key Findings:
Dysarthria Type: Unilateral Upper Motor Neuron (UUMN) dysarthria was the most prevalent type, observed in 52\% (25/48) of participants with audio recordings, making this the first evidence for its dominance in an acute ischemic stroke cohort.
Speech Characteristics: The most affected speech subsystems were:
Articulation: Imprecise consonants in 92\% (44/48).
Phonation: Harsh voice in 96\% (46/48).
Respiration: Audible inspiration in 79\% (38/48).
Objective acoustic measures, such as Maximum Phonation Time (MPT) and maximum loudness, frequently deviated from normative values, corroborating these perceptual findings.
Severity and Recovery:
Among all dysarthric participants with NIHSS data (n=70), 67\% had mild to moderate impairment at hospital admission.
A substantial proportion (50\%, or 35/70) demonstrated complete recovery within one week after symptom onset, highlighting significant early recovery potential.
Strengths and Limitations:
Strengths: The study employed a prospective design with both auditory-perceptual (gold standard) and objective acoustic measurements in a contemporary stroke cohort, providing robust data.
Limitations: Detailed audio assessments could not be performed in 19/67 (\approx 28\%) dysarthric participants, who tended to be older and have more severe strokes. This exclusion could potentially bias the results towards a milder dysarthria profile in the analyzed subgroup.
Conclusion: UUMN dysarthria is the dominant type in acute ischemic stroke, often mild, with a high rate of early spontaneous recovery. The impairment primarily affects articulation, phonation, and respiration, which can be quantitatively monitored using objective acoustic measures. These findings have implications for differential diagnosis, resource allocation, and early rehabilitation planning in post-stroke care.