Swallowing Trouble Strikes Nearly 4 in 10 Stroke Patients, Landmark Swedish Study Finds

Every year, millions of people worldwide survive a stroke, only to face a complication that rarely makes headlines but can shape the entire course of their recovery: the inability to swallow safely. Now, one of the largest and most rigorous investigations of this condition to date has delivered a strikingly precise answer to a question that has divided clinicians for decades. In a prospective multicentre cohort study published in Frontiers in Neurology, researchers led by Jenny Selg and Patricia Hägglund of Umeå University found that 38.8 percent of patients experienced dysphagia—swallowing impairment—in the acute phase after a first-ever stroke. The figure, drawn from more than 1,000 consecutively recruited patients assessed with a validated screening tool within 96 hours of stroke onset, lands squarely in the middle of a bewildering range of previous estimates that have spanned anywhere from 19 to 81 percent.

The scale of the study is what sets it apart. Between June 2021 and May 2025, the research team screened 1,093 patients admitted to stroke units at four Swedish hospitals: Umeå University Hospital, a tertiary care centre serving northern Sweden, along with Danderyd Hospital, Stockholm South General Hospital, and Kungälv Hospital. The cohort included 468 women, or 42.8 percent of participants, and had a median age of 75 years. By recruiting consecutive admissions rather than selected cases, the investigators minimised the selection bias that has plagued earlier studies, which often excluded the sickest patients—the very people most likely to have swallowing problems.

At the heart of the methodology lies the Gugging Swallowing Screen, known as GUSS, administered in its validated Swedish version by six trained research speech-language pathologists. Unlike simple water-swallow tests, GUSS is a step-wise instrument that evaluates vigilance, saliva management, voluntary cough, and the ability to swallow semi-solid, liquid, and solid boluses in a predefined hierarchical order. Total scores range from 0 to 20, with 20 indicating no dysphagia and progressively lower scores signalling mild, moderate, or severe impairment. The tool has demonstrated excellent sensitivity and specificity against instrumental gold standards such as videofluoroscopic swallowing studies and flexible endoscopic evaluation of swallowing—assessments that would have been impossible to deploy across a cohort of this size.

Even with such careful design, acute stroke medicine poses a practical dilemma: many patients in the first days after a stroke are too drowsy, aphasic, or cognitively impaired to cooperate with formal screening. To avoid excluding them, the researchers applied an ethically approved proxy method—a bedside water-swallowing assessment performed by stroke unit nurses using the Standardized Swallowing Assessment, which correlates strongly with GUSS. In total, 57 percent of patients were assessed directly with GUSS-S, while 43 percent entered the study through the proxy route. The mean time from admission to assessment was 37 hours, well within the 96-hour window that defines the acute phase.

Beyond establishing the headline incidence figure, the study identified which patients are most at risk. In multivariable regression models that adjusted for multiple factors simultaneously, five independent predictors of dysphagia emerged. Dysarthria—slurred or weakened speech resulting from damage to the motor speech network—carried an adjusted odds ratio of 1.85, meaning affected patients were nearly twice as likely to have swallowing impairment. This makes physiological sense: the muscles and neural circuits governing speech and swallowing overlap substantially, so impaired articulation can serve as a visible clinical flag for hidden swallowing dysfunction. Greater stroke severity, measured per point on the National Institutes of Health Stroke Scale, raised the odds by 16 percent with each increment, reflecting the wider disruption of cortical and subcortical swallowing networks as brain injury extends.

The remaining predictors were equally telling. Women faced 39 percent higher odds of dysphagia than men, a finding consistent with some prior research suggesting sex differences in stroke outcomes and swallowing physiology, though the authors note the literature remains mixed. Strokes located infratentorially—below the tentorium, in the brainstem and cerebellum—doubled the odds compared with supratentorial lesions, a result that aligns with decades of neuroanatomical evidence showing that the brainstem houses critical swallowing centres. Older age, modelled with a flexible cubic spline to capture its non-linear relationship with risk, was also independently associated with dysphagia, consistent with age-related decline in the strength and coordination of swallowing musculature. Notably, reperfusion therapies such as thrombolysis and thrombectomy, delivered to a substantial share of patients at the comprehensive stroke centre, were not independently linked to dysphagia once other factors were accounted for.

The consequences of dysphagia proved as consequential as its causes. Patients with swallowing impairment remained in hospital on average 1.64 times longer than those who could swallow normally—a difference that ripples through both individual recovery and healthcare economics. They were more than seven times as likely to require tube feeding, whether via nasogastric, nasoduodenal, nasojejunal, or surgically inserted gastrostomy tubes. Their odds of being discharged anywhere other than home were dramatically elevated: 4.84-fold for nursing homes, 4.69-fold for short-term care facilities, and roughly two to three times higher for transfers to other hospitals or wards. Most sobering of all, in-hospital death was associated with a 16.67-fold increase in odds among patients with dysphagia, even after adjustment for age, stroke severity, sex, and treatment centre.

Pneumonia, the most feared complication of swallowing impairment, told a similar story. When saliva, food, or liquid slips past the vocal folds into the trachea—a process called aspiration—bacteria can seed the lungs and ignite infection. In this cohort, pneumonia developed in 5.1 percent of all patients, but the burden fell unevenly: roughly one in ten patients with dysphagia developed pneumonia, compared with just 2 percent of those with normal swallowing, and dysphagia remained independently associated with the infection after adjustment. The overall pneumonia rate is lower than in many older studies, which the authors attribute partly to modern stroke care, including early mobilisation and routine dysphagia screening—practices that previous research has shown to reduce pneumonia rates when performed by trained staff.

Placed in historical context, the findings carry a subtle but important message. More than two decades ago, a landmark review by Martino and colleagues reported dysphagia rates of 37 to 45 percent when screening-based assessments were used, and 64 to 78 percent with instrumental methods. Despite revolutionary advances since then—acute reperfusion therapy, early mobilisation, coordinated multidisciplinary stroke unit care—the incidence observed in 2020s Sweden sits within that same screening-based range. This suggests that while stroke survival and acute management have improved enormously, the early burden of swallowing impairment has remained largely unchanged, likely because the underlying neuroanatomy of swallowing is highly individual and lesion-dependent, defying both generalisation and precise prediction of individual outcomes.

The authors are candid about their study’s limitations. Assessors could not be blinded to clinical information, since verifying eligibility required reviewing medical records that contained swallowing-related data, creating a potential for observer bias. Screening tools, however well validated, are less sensitive than instrumental videofluoroscopic or endoscopic examinations, so the true incidence may be somewhat higher. Some variables, including stroke severity scores, reperfusion data, and body-mass index, were missing for portions of the cohort, and tube feeding was identified solely through procedure codes, which may have led to underestimation. Yet sensitivity analyses using penalised regression produced consistent results, and the sheer scale, prospective design, and multicentre scope of the work make it one of the most robust estimates available. For clinicians, the takeaway is unambiguous: dysphagia affects more than a third of acute stroke patients, it announces itself through recognisable warning signs such as dysarthria and severe neurological deficits, and it demands early identification and targeted management to shorten hospital stays, prevent pneumonia, and keep patients on the path back home.

Subject of Research: Incidence and predictors of post-stroke dysphagia in acute stroke patients

Article Title: Incidence of dysphagia in acute stroke patients—a prospective multicentre cohort study

Article References: Selg, J., Mirkoska, V., Holmlund, T., Jäghagen, L., Wester, P., & Hägglund, P. (2026). Incidence of dysphagia in acute stroke patients—a prospective multicentre cohort study. Frontiers in Neurology, 17, Article 1886254. https://doi.org/10.3389/fneur.2026.1886254

Image Credits: AI Generated

DOI: 10.3389/fneur.2026.1886254

Keywords: stroke, dysphagia, swallowing disorders, aspiration, pneumonia, GUSS screening, speech-language pathology, stroke severity, NIHSS, dysarthria, brainstem stroke, cohort study