Eligible patients were those aged ≥60 years who were hospitalized for AIS from August 2021 to December 2022 and for whom written informed consent could be obtained from the patient or his/her family. AIS was defined as a case wherein a lesion was identified on a brain MRI diffusion-weighted image taken following the onset of new neurological symptoms or wherein AIS was the most appropriate cause of the neurological symptoms (e.g., older adult–onset medial longitudinal fasciculus syndrome) spanning >24 h. AIS severity was assessed using the National Institutes of Health Stroke Scale (NIHSS) on admission. The NIHSS classifies neurological symptoms caused by stroke into 15 items; each category is scored from 0 to 2 or 3 points, and the total score is calculated (15). The score ranges from 0 to 40; higher scores indicate more severe stroke symptoms (15).

Patients’ medical histories were reviewed for hypertension, diabetes, dyslipidemia, stroke (ischemic and/or hemorrhagic), smoking, and alcohol use. Patients who had smoked for >1 year, smoking most days of the week, were considered to have a smoking history. Patients who drank alcohol most days of the week were considered to have a drinking history.

Blood data were referenced to that at the time of hospitalization. After obtaining informed consent, height, weight, muscle mass, grip strength on the nonparalyzed side, and frailty were evaluated. In addition, carotid ultrasonography was performed during hospitalization to assess internal carotid artery stenosis, with measurements of maximum intima–media thickness (IMT) and peak systolic velocity (PSV). A PSV exceeding 200 cm/s is considered indicative of significant stenosis, defined as ≥70% narrowing according to the criteria of the North American Symptomatic Carotid Endarterectomy Trial (16). Furthermore, the complications of pneumonia and urinary tract infection during hospitalization were evaluated. The outcome was evaluated using the length of hospital stay and the modified Rankin Scale (mRS) at discharge. The mRS is frequently employed for evaluating the motor function of patients with stroke and is rated on a scale of 0–6, with 0 representing no symptoms and 6 indicating mortality (17).

For the evaluation of WMH, brain MRI T2 fluid-attenuated inversion recovery (FLAIR) images were obtained using the following equipment: Ingenia 3.0 T [Philips, Netherlands; repetition time (TR) = 10,000 ms; echo time (TE) = 125 ms; inversion time (TI) = 2,700 ms; field of view = 290 × 192 mm; slice thickness = 5 mm; interslice gap = 1 mm] or Ingenia 1.5 T (Philips, Netherlands; TR = 11,000 ms; TE = 120 ms; TI = 2,800 ms; field of view = 240 × 153 mm; slice thickness = 5 mm; interslice gap = 1 mm). Based on the images, WMHs were categorized into grades 0–3 according to the Fazekas classification (18). Deep white matter hyperintensity (DWMH) was graded on a scale of 0–3 [0, absent; 1, punctate foci (maximum size < 3 mm); 2, beginning confluence of foci (size ≥ 3 mm); and 3, large confluent areas]. Similarly, periventricular hyperintensity (PVH) was graded on a scale of 0–3 (0, absent; 1, a “cap” or pencil-thin lining; 2, a smooth “halo”; and 3, irregular PVH extending into the deep white matter). Two neurologists performed the classification; in case of disagreement in the evaluation, the patient was classified into the appropriate grade by consensus.

Frailty was diagnosed on the basis of the following items: shrinking, physical function, physical activity, and exhaustion (8). However, walking speed, which is frequently used as an evaluation item for physical function and physical activity, is affected by AIS. In this study, the “frailty screening index,” a method for evaluating frailty based on interviews, was used for diagnosing frailty (19). The frailty screening index asks patients five questions. If the patient answers “Yes” to the following three questions, 1 point is added: “Have you lost 2 kg or more in the past 6 months?” “Do you think you walk slower than before?” “In the past 2 weeks, have you felt tired without a reason?” Furthermore, 1 point is added when a patient answers “No” to the following two questions: “Do you go for a walk for your health at least once a week?” “Can you recall what happened 5 min ago?” A total of ≥3 points is defined as frailty (19). In this study, the question was prefaced with “Please tell me about your condition before you were hospitalized for this stroke.”

Sarcopenia was assessed by muscle mass and nonparalytic side grip strength according to the Asian Working Group for Sarcopenia 2019 criteria (9). Muscle mass was assessed using bioelectrical impedance analysis (Inbody S10, Seoul, South Korea), and low muscle mass was considered when the muscle mass was <7.0 and <5.4 kg/m² for males and females, respectively. Grip strength was assessed as low when it was <26 and <18 kg for males and females, respectively. Patients with both low muscle mass and low grip strength were diagnosed with sarcopenia (9).

In this study, the Geriatric Nutritional Risk Index (GNRI) was employed for evaluating nutritional risk (20). The GNRI is a simple tool for assessing the nutrition-related risks of morbidity and mortality in hospitalized older adults, calculated using serum albumin levels and the ratio of the actual weight to the ideal weight. The GNRI is calculated using the following formula: GNRI = 14.89 × albumin (g/dL) + 41.7 × (weight/ideal body weight). According to the original study, four nutritional risk levels were defined as follows: major risk (GNRI < 82), moderate risk (82 ≤ GNRI < 92), low risk (92 ≤ GNRI ≤ 98), and normal (GNRI > 98) (20). In this study, the ideal body weight was calculated as the weight at a body mass index (BMI) of 22 kg/m².

To investigate whether the variables demonstrated a monotonic increasing or decreasing trend with respect to the PVH and DWMH grades, the Jonckheere–Terpstra test was applied for continuous variables, and the Cochran–Armitage test was used for nominal scales. The presence or absence of frailty and sarcopenia was analyzed using Fisher’s exact test. For nominal scales. The Shapiro–Wilk test was applied to both ordinal and continuous variables. Variables with normal distributions were analyzed using Student’s t-test, whereas variables without normal distributions were analyzed using the Mann–Whitney U test. When analyzing GNRI nutritional risk as a continuous variable, it was substituted as follows: normal, 1; low risk, 2; moderate risk, 3; and major risk, 4. Spearman’s rank correlation coefficient was calculated for the correlation between PVH and DWMH grades. Furthermore, binomial logistic regression analysis using the forced entry method was performed to explore the associations between frailty, sarcopenia, and PVH. The dependent variable was set as either frailty, sarcopenia, or PVH of grade 2 or higher, while the other two variables (e.g., when frailty was the dependent variable, sarcopenia and PVH grade ≥2 were used as independent variables), along with age, sex (male: 1; female: 2), and GNRI, were included as independent variables. For the independent variables included in the model, variance inflation factors were calculated using linear regression. All values were less than 10, indicating no evidence of multicollinearity. Statistical analysis was performed using SPSS 22.0 (IBM Corp., Armonk, NY, United States), with p < 0.05 considered statistically significant. To assess statistical power, post hoc power analyses were conducted using G*Power 3.1 for each outcome variable in the logistic regression models, focusing on categorical independent variables that demonstrated statistically significant associations. The analyses were based on the observed odds ratios, an alpha error of 0.05, and the corresponding sample sizes.

During the observation period, 172 patients with AIS were hospitalized. Of these patients, 27 were aged <60 years, 33 did not consent to the study, 1 did not have an MRI, 1 had missing weight data, and 1 had AIS recurrence within the study period and were therefore excluded from the analysis. Finally, 109 patients (median age, 79 years; 59.6% male) were included in the analysis. As 13 patients who had difficulty with the interview were excluded, the assessment of frailty was performed on 96 patients. Sarcopenia was assessed in 100 patients, excluding nine patients whose grip strength or muscle mass was not measured. Carotid ultrasonography examinations were performed in 101 patients. A PSV exceeding 200 cm/s, indicative of significant stenosis, was observed in only one patient. In this study, maximum IMT was analyzed as an indicator of carotid artery stenosis.

Of the eligible patients, 9, 21, 35, and 44 exhibited DWMH grades 0, 1, 2, and 3, respectively, and 7, 59, 13, and 30 had PVH grades 0, 1, 2, and 3, respectively. The 4 × 4 table (Table 1) shows a comparison of the grades of the two WMHs. The correlation coefficient between the DWMH and PVH grades was 0.662 (p < 0.001).

The DWMH grade was higher in older adult patients (p = 0.041) and lower in males (p = 0.008). Physical measurements revealed that the higher the grade, the lower the height (p = 0.002), grip strength (p = 0.004), and muscle mass (p = 0.013); however, no significant relationship with BMI or sarcopenia rate was observed (Table 2). Moreover, no significant differences in the total frailty screening index and the proportion of frail individuals were noted. Regarding stroke severity, a trend for higher NIHSS scores with higher grades was observed (p = 0.006). No significant differences in medical history (hypertension, diabetes, dyslipidemia, and stroke); smoking history; drinking history; albumin; creatinine; C-reactive protein; hemoglobin A1c; maximum IMT; GNRI; or GNRI nutritional risk were noted. Regarding complications during hospitalization, higher DWMH grades indicated a higher pneumonia incidence (p = 0.028); however, no significant association with the incidence of urinary tract infection was observed. Additionally, higher grades were associated with longer length of stay (p = 0.005) and higher mRS at discharge (p = 0.001).

Similar to DWMH (Table 3), higher PVH grades were associated with older patients (p < 0.001), higher proportion of female patients (p = 0.035), higher NIHSS scores (p < 0.001), lower height (p = 0.003), lower grip strength (p = 0.001), and lower muscle mass (p < 0.001). Conversely, unlike DWMH, significant differences in body weight (p < 0.001), BMI (p = 0.017), total frailty screening index (p < 0.001), proportion of frailty (p = 0.003), and proportion of sarcopenia (p = 0.003) were observed. No significant differences in the rates of hypertension, diabetes, dyslipidemia, stroke or smoking history were noted. The higher PVH-grade group demonstrated lower alcohol consumption rates (p = 0.049). Regarding blood data and nutritional risk indicators, higher PVH grades were associated with lower albumin levels (p = 0.027) and GNRI (p = 0.004) as well as higher GNRI nutritional risk grade (p = 0.005). No significant differences were observed in maximum IMT among the PVH grade groups. Regarding the course of the patients following hospitalization, no significant difference in pneumonia and urinary tract infection complication rates was noted; however, the length of stay and mRS at discharge were both high (all p < 0.001).

Frailty was assessed in 96 patients, and 26 (27.1%) were diagnosed as frail on the basis of a frailty screening index of ≥3. Sarcopenia was observed in 47 of 100 patients (47%). A univariate analysis was conducted by stratifying patients according to the presence or absence of frailty and sarcopenia. The detailed results of this analysis are presented in Supplementary Tables 1, 2. Among patients with AIS, those with frailty exhibited a significantly higher sarcopenia prevalence (p = 0.001). The following were the common characteristics among patients with AIS having frailty and sarcopenia: older age (frail, p = 0.022; sarcopenic, p < 0.001), lower grip strength (p = 0.030 and p < 0.001), higher PVH grade (both p = 0.003), higher NIHSS score at admission (p = 0.008 and p = 0.001), longer hospital stay (p = 0.024 and p = 0.010), and higher mRS score at discharge (p = 0.010 and p < 0.001). Frailty was more common in males (p = 0.011) and was associated with a higher pneumonia incidence during hospitalization (p = 0.044). Patients with sarcopenia showed lower body weight, BMI, and muscle mass (all p < 0.001); lower albumin levels (p = 0.002); and lower GNRI but higher GNRI nutritional risk (both p < 0.001).

In the ordinal logistic regression with the PVH grade as the dependent variable, the p-values for all of the threshold values were ≥0.05 (data not shown). Therefore, the analysis was conducted after converting the data to binary data, that is, whether the PVH grade was ≥2 or not (Tables 4–6). Frailty was significantly associated with age [p = 0.020; odds ratio (OR), 1.098; 95% confidence interval (CI), 1.015–1.188], male sex (sex was coded as male: 1 and female: 2; p = 0.003; OR, 0.117; 95% CI, 0.029–0.475), a PVH grade of ≥2 (p = 0.024; OR, 3.753; 95% CI, 1.186–11.879), and sarcopenia (p = 0.034; OR, 3.742; 95% CI, 1.106–12.655). In the analysis using sarcopenia as the dependent variable, GNRI nutritional risk (p = 0.002; OR, 7.937; 95% CI, 2.172–28.996) and frailty (p = 0.027; OR, 4.026; 95% CI, 1.173–13.823) were identified as significant factors. When a PVH grade of ≥2 was set as the dependent variable, frailty was the only significant predictor (p = 0.023; OR, 3.728; 95% CI, 1.195–11.628). The post hoc power analysis revealed that all relevant models demonstrated sufficient statistical power, with calculated power exceeding 0.9 in each case. Specifically, this included models with frailty as the dependent variable and sex, PVH grade, or sarcopenia as independent variables; sarcopenia as the dependent variable and frailty as the independent variable; and PVH grade as the dependent variable and frailty as the independent variable. Figure 1 illustrates the associations among frailty, sarcopenia, grade ≥2 PVH, and nutritional risk based on the results of multivariate analysis.