Systematic Review of Dysphagia and Pulmonary Compromise in Stroke Patients

Systematic Review Overview and Clinical Importance of Dysphagia

This peer-reviewed systematic review investigates the incidence of dysphagia and its associated pulmonary complications in patients following a stroke. The study is grounded in the observation that the presence of dysphagia is significantly associated with an elevated risk for pulmonary complications and increased mortality rates. There is emerging clinical evidence suggests that the early detection of dysphagia in acute stroke patients not only mitigates these severe medical complications but also reduces the total length of hospital stays and decreases overall healthcare expenditures.

Developing an accurate estimate of the incidence of dysphagia and the subsequent risk for pulmonary consequences in the stroke population is deemed critical for the field. Such data are intended to guide the design of future research efforts aiming to assess the specific benefits and efficacy of various dysphagia interventions. The study defines the scope of dysphagia incidence as the first identification of the condition through clinician testing, excluding any subsequent follow-up identifications for the same patients to maintain a clear incidence metric.

Methodology and Literature Search Strategy

The researchers conducted an exhaustive search of multiple databases covering the period from 1966 through May 2005. The primary databases utilized were Medline, Embase, and Pascal, supplemented by relevant internet addresses and extensive hand-searching of the bibliographies of identified articles. The search terms employed included ‐cerebrovascular disorders‐ and ‐deglutition disorders,‐ with the search limited to human subjects and original articles addressing the frequency of dysphagia or pneumonia.

The inclusion criteria were strictly limited to original research articles that evaluated the swallowing ability of consecutively enrolled adult patients (aged $\ge 18$ years). Patients were included regardless of stroke type or location, and settings included acute, rehabilitation, or chronic facilities. Both retrospective and prospective studies were considered. A critical requirement for inclusion was the use of clearly described methods for reporting the incidence of dysphagia. If a study reported on pneumonia, it was required to provide the specific criteria used to define the condition.

Exclusion criteria were applied to editorials, review articles, case series with fewer than 10 patients, and articles based solely on patient-reported symptoms, which were deemed unreliable. Studies that enrolled patients only after they had already failed a swallow screening or were referred for speech-language pathology assessment were excluded to prevent a selection bias toward higher dysphagia likelihood. Non-English articles were included after translation using standardized software and assistance from native speakers. Authors were contacted to clarify ambiguities when necessary. The methodological quality of each study was appraised using criteria established by Sackett et al., with an emphasis on design strengths, flaws, blinding, and the documentation of diagnostic reliability.

Statistical Analysis and Methodological Quality

Two authors, R.M. and N.F., independently reviewed all accepted articles for methodology and data extraction, resolving any discrepancies through consensus. The researchers utilized the software Review Manager (RevMan) 4.219 to calculate the relative risk ($RR$) and $95\%$ confidence intervals ($CI$) for developing pneumonia associated with dysphagia and aspiration. A random-effects model was selected for the analysis because it produces wider and more conservative confidence intervals. The risk of pneumonia was considered statistically significant if the $95\%\,CI$ did not include the value of one.

Statistical analyses were segmented to assess the pneumonia risk among all patients with dysphagia versus those with more severe dysphagia, as identified by confirmed aspiration. Methodology was scrutinized for blinding and the accounting of all eligible patients. Of the 24 final selected studies, all featured consecutive enrollment; 22 were prospective designs while 2 were retrospective. Only one study reported investigator reliability for diagnostic interpretations, and only six studies declared blinding for outcome assessments.

Evaluation of Literature Retrieved

The initial search yielded a total of 277 citations. Following the initial screening, 91 were eliminated as they were reviews, editorials, or non-peer-reviewed, leaving 186 original abstracts. Further pruning eliminated 82 citations for being case series with <10 patients, having non-adult subjects, or failing to measure dysphagia as an outcome. Of the remaining 104 citations, 68 were eliminated because they involved pre-selected or referred patients. An additional 12 articles were removed due to limited raw data, reliance on patient-reported symptoms, or the use of chart reviews without specified clinician testing. Ultimately, 24 articles were accepted for the final review.

Patient characteristics varied: 12 studies involved mixed acute ischemic strokes; others focused on acute single hemispheric strokes, rehabilitation brain stem strokes, or acute ischemic and hemorrhagic strokes. Seventeen studies limited enrollment to conscious patients, whereas four studies included obtunded patients. To provide a conservative estimate of dysphagia, obtunded patients were excluded from calculations when known, based on the premise that they would naturally be unable to swallow.

Incidence of Dysphagia Following Stroke

Incidence rates were calculated based on the type of assessment used. Nine articles utilized screening tests, primarily the water swallowing technique. Two studies defined failure by choking or coughing; another added wet vocal quality; and others extended monitoring for up to 5 minutes or added volume and time variables. In studies using these screening tests, incidence rates typically ranged between $37\%$ and $45\%$, though outliers existed. For instance, the study by Barer et al. reported the lowest incidence because it only enrolled patients capable of swallowing pills, likely under-reporting severe cases.

Sixteen articles used comprehensive swallowing tests, including clinician testing, videofluoroscopy, clinical-videofluoroscopic combinations, or endoscopy. In acute stroke studies using videofluoroscopy, the incidence of dysphagia was the highest, ranging from $64\%$ to $78\%$. When acute studies used only clinician testing, the reported rates were lower, ranging from $30\%$ to $55\%$. Studies with lower rates often restricted enrollment to hemispheric strokes, excluding more severe impairments. In rehabilitation settings, the incidence ranged widely from $40\%$ to $81\%$, with the highest rate appearing in a prospectively designed cohort study of brain stem stroke patients.

Outcomes Related to Aspiration and Pneumonia

Aspiration is categorized as a severe form of dysphagia; while all patients who aspirate have dysphagia, not all dysphagic patients aspirate. Among studies involving mixed stroke lesions using instrumental testing, the incidence of aspiration ranged from $22\%$ to $52\%$. A study focusing specifically on acute brain stem lesions reported an aspiration incidence of $44\%$.

Nine articles provided data on pneumonia incidence. The risk of pneumonia was found to be significantly higher in patients with dysphagia compared to those without. In acute settings, three studies reported similar pneumonia incidence rates between $16\%$ and $19\%$, while one study reported $33\%$. In rehabilitation cohorts, pneumonia rates ranged from $7\%$ to $29\%$, with the highest incidence occurring in brain stem stroke patients.

Pooled analyses revealed a definitive increase in pulmonary risk based on the severity of the swallowing impairment. The relative risk of pneumonia for patients with dysphagia was $3.17$ ($95\%\,CI, 2.07, 4.87$). For patients with confirmed aspiration, the risk was exponentially higher, with a relative risk of $11.56$ ($95\%\,CI, 3.36, 39.77$) compared to those without aspiration. These findings highlight the critical nature of identifying aspiration as a distinct and severe risk factor within the dysphagic population.