Clinical Efficacy of Comprehensive Geriatric Assessment in Patients Undergoing Hemodialysis: A Retrospective Observational Study

1. Introduction

The increasing number of older adults worldwide has increased the impact of age-related conditions and complete healthcare needs. Comprehensive geriatric assessment (CGA) provides a multidisciplinary approach for diagnosing frailty in older adults [1,2]. This process identifies medical, psychosocial, and functional limitations in order to develop a coordinated plan that maximizes overall health and well-being. Frailty is a biological syndrome that increases vulnerability to poor homeostasis and is associated with a higher risk of adverse health outcomes such as hospitalization and mortality [3,4]. Frailty is highly prevalent among older adults [5,6]. The hemodialysis population has been aging, and declines in muscle strength, oral frailty, and instrumental activities of daily living (IADL) have been recognized as factors that impair patients’ quality of life [7,8]. In addition to these physical impairments, patients undergoing maintenance hemodialysis (MHD) often experience mental health problems, such as anxiety and depression [9]. The effectiveness of CGA, a multidimensional process, in patients admitted to hospitals is well known [10,11]; however, the evaluation of CGA is time-consuming [12] and is not routinely performed in clinical practice. There are no available data on whether CGA is beneficial in patients undergoing MHD.

Shinseikai Dai-ichi Hospital located in the eastern part of Nagoya City, Japan, has 144 hospital beds and 100 dialysis beds and belongs to the HOSPY group. The HOSPY group comprises of one acute care hospital equipped with an emergency room and an intensive care unit (six beds), one subacute hospital (Shinseikai Dai-ichi Hospital), and six dialysis clinics. Approximately 1,300 patients undergo hemodialysis in our group. Given that the average age of patients on MHD is 73.3 years, frailty is commonly observed and IADL is frequently reduced [13].

Therefore, in this study, we performed CGA in patients admitted to our hospital and evaluated its clinical efficacy in those undergoing MHD, Patients were first categorized into the MHD and non-dialysis groups, and CGA-related variables were compared between them. The association between the CGA score and clinical outcomes was then examined within the MHD group.

2. Methods

2.1. Study Design and Participants

Inpatients admitted to Shinseikai Dai-ichi Hospital aged ≥65 years from October 2024 to September 2025 were enrolled in this study. Upon admission, the CGA, including the Tokyo Metropolitan Institute of Gerontology Index of Competence (TMIG-IC), Hasegawa’s Dementia Scale-Revised (HDS-R), vitality index, and Geriatric Depression Scale-15 (GDS-15), was performed [13,14]. The TMIG-IC consists of 13 items assessing: IADL (five items), intellectual activity (four items), and social roles (four items). An IADL score of ≤4 was defined as impaired IADL. Higher scores indicated greater independence. The GDS-15 consists of 15 questions evaluating mood, interest, vitality, and anxiety regarding the future. Scores were interpreted as follows: ≥5 points: depressive tendency and ≥10 points: depressive state

2.2. Statistical Analysis

This was a retrospective observational study. Data were analyzed using descriptive statistics. Comparisons between the two groups were performed using Fisher’s exact test or the Mann-Whitney U test, as appropriate. Spearman’s rank correlation coefficient was used to evaluate correlations among CGA-related variables. Variables with p < 0.10 in univariate analyses were entered into multivariate logistic regression models. Since the total TMIG-IC score and its subscale scores were highly correlated, they were analyzed in separate multivariable models to avoid multicollinearity. In addition, HDS-R was excluded from the multivariable analyses because of a substantial proportion of missing data. Multicollinearity was assessed using variance inflation factors (VIFs), with a VIF < 5 considered acceptable. All p -values were two-sided, and p –values < 0.05 were considered statistically significant. Statistical analyses were performed via EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R (R Foundation for Statistical Computing, Vienna, Austria) [15]. This study was approved by the Ethics Committee of Shinseikai Dai-ichi Hospital (S2025-#003; approved on December 4, 2025).

3. Results

In total, 433 patients (328 in the dialysis group and 105 in the non-dialysis droup) were included in this study. The median (range) age was 80 (65 - 97) years in the MHD group and 82 (65 - 100) years in the non-dialysis group (Table 1). Clinical outcomes were examined and compared between the MHD and non-dialysis groups; the length of hospital stay (days) did not differ between the two groups [MHD: 38 (1 - 546) versus non-dialysis: 38 (3 - 343), p = 0.867], while the number of admissions to nursing homes after discharge was significantly greater in the non-dialysis group than in the MHD group (p < 0.001). Although inpatients without hemodialysis were significantly older than those with MHD (p = 0.028), the number of discharge due to death was higher in the MHD group than in the non-dialysis group. To investigate the factors related to mortality, 45 patients undergoing dialysis and 28 patients without dialysis were excluded from the subsequent analysis because of unknown outcomes (e.g., nursing home admission, change of physician, transfer to another hospital, or relocation). As a result, 283 participants underwent MHD and 77 participants did not. Regarding the sex distribution, the former group included 171 men and 112 women, whereas the latter group included 32 men and 45 women, indicating a male predominance in the MHD group. The rate of mortality within one-year of discharge was higher in the MHD group (30.0%; 85/ 283) than in the non-dialysis group (19.5%; 15/ 77), but the difference was not significant. Total TMIG-IC scores were significantly lower in the MHD group than in the non-dialysis group [4 (0 - 13) versus 7 (0 - 13); p = 0.005]. Declines in IADL, intellectual activity, and social roles were observed in the MHD group ([MHD versus non-dialysis: 1 (0 - 5) versus 2 (0 - 5); p = 0.011, 1 (0 - 4) versus 2 (0 - 4); p = 0.01, 2 (0 - 4) versus 2.5 (0 - 4); p = 0.01]. The HDS-R and vitality index scores were also significantly lower in patients undergoing MHD than in the non-dialysis group. In the GDS-15 assessment, patients on MHD demonstrated higher scores than those without hemodialysis [7 (0 - 15) versus 5 (0 - 15); p = 0.001]. The prevalence of depression and depressive tendencies (GDS-15 scores > 5) was significantly higher in the MHD group than in the non-dialysis group (67.8% versus 52.2%; p = 0.022), although that of depression was not statistically different between the two groups of inpatients (p = 0.063). These findings suggest that MHD inpatients frequently experience psychological distress.

Because CGA-related variables were correlated with mortality in patients undergoing MHD, we conducted a univariate analysis to determine which factors affected the clinical outcomes of this population. Table 2 shows that sex, TMIG-IC, vitality index, HDS-R, and GDS-15 were significantly different between the group of inpatients with MHD alive at one- year after discharge (survivors) and those who died within one year (non-survivors), whereas age, BMI, and length of hospital stay were not different. As for the inpatients without hemodialysis, the score of TMIG-IC in the survivors significantly differ from that in the non-survivors. However, the age, sex distribution, length of hospital stay, HDS-R score, vitality index, and GDS-15 scores did not.

We next performed multivariate analysis in the MHD group, which revealed that female sex (OR 0.37, 95% CI 0.20 - 0.71, p = 0.003), TMIG-IC (OR 0.87, 95% CI 0.78 - 0.96, p = 0.005), and vitality index (OR 0.84, 95% CI 0.72 - 0.99, p = 0.035) were significantly associated with one- year mortality rate after discharge. Among the individual TMIG-IC items, only the social role was significantly associated with mortality rate (OR 0.70, 95% CI 0.50 - 0.97, p = 0.034), whereas IADL (OR 0.83, 95% CI 0.67 - 1.04, p = 0.106) and intellectual activity (OR 1.05, 95% CI 0.81 - 1.36, p = 0.735) did not.

Table 3. Multivariate analysis for the inpatients with MHD, showing the association with the mortality.

	Odds Ratio	95% CI	P value
Age	1.02	0.98 – 1.06	0.337
Female sex	0.37	0.20 – 0.71	0.003
TMIG-IC total score	0.87	0.78 – 0.96	0.005
Vitality index	0.84	0.72 – 0.99	0.035
GDS-15 score	1.01	0.94 – 1.08	0.891

MHD: Maintenance hemodialysis. TMIG-IC: Tokyo Metropolitan Institute of Gerontology Index of Competence. GDS-15: Geriatric Depression Scale-15.

Table 3. Multivariate analysis for the inpatients with MHD, showing the association with the mortality.

	Odds Ratio	95% CI	P value
Age	1.02	0.98 – 1.06	0.337
Female sex	0.37	0.20 – 0.71	0.003
TMIG-IC total score	0.87	0.78 – 0.96	0.005
Vitality index	0.84	0.72 – 0.99	0.035
GDS-15 score	1.01	0.94 – 1.08	0.891

MHD: Maintenance hemodialysis. TMIG-IC: Tokyo Metropolitan Institute of Gerontology Index of Competence. GDS-15: Geriatric Depression Scale-15.

4. Discussion

The application of CGA to evaluate functional capacity and psychological status revealed that inpatients undergoing MHD had significantly lower total TMIG-IC scores, IADL, and social role scores than non-dialysis inpatients. Our observation suggests that biological age as determined by CGA offers a more accurate reflection of mortality risk than chronological age alone [16,17]. In addition, the GDS-15 scores were significantly higher in the MHD group, even though non-dialysis patients were older in this study. These impairments may be associated with restrictions in daily life caused by dialysis, reduced opportunities to go out, and insufficient social communication. In particular, deterioration in social roles has been reported to reduce quality of life [18] and social participation. In the current study, we found that social roles were significantly associated with mortality within one- year of discharge in inpatients with MHD. Therefore, community-based interventions aimed at redesigning mobility and practical support systems, improving living environments, and promoting social participation may be warranted for patients with MHD.

Regarding psychological status, evaluation using the GDS-15 demonstrated that approximately 70% of patients undergoing MHD had depression or depressive tendencies. This finding suggests that the dialysis treatment itself may impose a psychological burden. Anxiety and distress related to the continuation of dialysis therapy are commonly observed among the inpatients with MHD, and systems or measures enabling timely psychological support may be important.

This study had several limitations. First, the study was conducted at a single institution, which may have limited the generalizability of the findings. Thus, further multicenter studies are warranted to clarify whether CGA could be useful in predicting mortality in patients undergoing hemodialysis. Second, the data were obtained from observational studies, which may have been prone to selection and confounding biases. Therefore, our results should be interpreted with caution. Some authors have suggested that serial assessment is a useful method to monitor the disease course including clinical outcomes. Third, we had limited demographic and clinical data that might have been associated with health outcomes, such as marital status, educational level, socioeconomic status, remaining renal function, or other clinical parameters of inflammatory markers (CRP, Hb, and Ht), cardiovascular conditions, and nutritional status. In particular, elevated CRP levels and functional decline were recently reported to offer greater predictive power than chronological age in assessing survival probability [19]; however, we did not analyze the association between mortality and inflammatory markers in the present study. Finally, for patients undergoing MHD, social support is more available to those who are married, live with family members, or have social support networks. Therefore, it is reasonable to assume that enhancing social support plays a significant role in improving clinical outcomes. The role of social support was not evaluated in the current analysis, although higher levels of social support have been found to be associated with higher QOL [20] and the enhancement of QOL should contribute to extending the lifespan in patients undergoing MHD.

5. Conclusion

The present study for the inpatients identified the following characteristics among inpatients undergoing MHD.

(1) Functional decline, as determined by the TMIG-IC, was more prominent in MHD inpatients than in those without dialysis.

(2) Psychological distress was also observed in this population.

(3) There was a significant association between mortality and sex, vitality index, and TMIG-IC scores, especially the social role score.

In conclusion, CGA may be useful for assessing vulnerability and predicting mortality among MHD inpatients.