Abstract
Background
Aspiration pneumonia (ASP) during hospitalization is associated with poor patient outcomes. We examined the factors associated with ASP onset in older patients in an acute-care hospital in Japan.
Methods
This retrospective cohort study analyzed data from the administrative claims database of patients aged > 65 years hospitalized and discharged from a general acute-care hospital between July 2016 and March 2021. We conducted multivariable logistic regression analyses to identify factors associated with ASP incidence, with independent variables including sex, age (65–74, 75–84, ≥ 85 years), place of residence before admission, income, main disease causing admission, dysphagia at admission, Barthel Index at admission, Charlson Comorbidity Index (0, ≥ 1), Hospital Frailty Risk Score (0–4, ≥ 5), presence of rehabilitation services, surgical treatment, planned hospitalization, enteral nutrition, and total parenteral nutrition.
Results
The ASP incidence during hospitalization among the 11,504 included patients (mean age 81.1 years) was 1.6%. ASP incidence was associated with female sex (adjusted odds ratio [aOR]: 0.451, 95% confidence interval [CI]: 0.326–0.625), age ≥ 85 years (aOR: 2.901, 95% CI: 1.471–5.718), dysphagia at admission (aOR: 3.810, 95% CI: 1.410–10.293), low activities of daily living level (aOR: 0.971, 95% CI; 0.963–0.979), comorbidities (aOR: 1.839, 95% CI: 1.298–2.605), frailty (aOR: 2.918, 95% CI: 2.089–4.077), unplanned hospitalization (aOR: 3.068, 95% CI: 1.593–5.906), and receiving rehabilitation services during hospitalization (aOR: 0.435, 95% CI: 0.256–0.739).
Conclusion
This study identified the factors associated with ASP onset in older patients in an acute-care hospital in Japan. These findings may help identify older inpatients who are more likely to develop ASP, supporting the planning of targeted care strategies during hospitalization.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12877-025-06556-7.
Keywords: Aspiration pneumonia, Hospital-acquired pneumonia, Dysphasia, Rehabilitation, Diagnosis procedure combination, Risk factors, Frailty, Activities of daily living
Introduction
Pneumonia is common and a major cause of death among older adults [1–4]. Aspiration pneumonia (ASP) is the sixth-leading cause of death in Japan, accounting for approximately 56,000 deaths annually [1]. ASP comprises a large proportion of pneumonia cases, and the ratio of ASP to total pneumonia cases increases with age [2]. According to Japan’s Ministry of Health, Labor and Welfare, approximately 70% of patients with pneumonia in hospitals are aged 75 years or older, and most cases of pneumonia are classified as ASP [4]. A study in the United States reported 76% of the deaths related to ASP occurred in the group of adults aged 75 years or older [3]. As the aging population grows rapidly worldwide [5], ASP is becoming a critical issue.
ASP during hospitalization leads to a prolonged length of stay, increased readmission, and decreased patient quality of life [6]. Moreover, ASP leads to a decline in physical function, and an increased need for assistance in daily activities is inevitable [7, 8]. Interventions to prevent ASP during hospitalization include early rehabilitation and multidisciplinary interventions [9, 10]. Therefore, the examination of factors associated with ASP incidence during hospitalization is important for identifying patients who may benefit from preventive intervention.
Patients who had acquired ASP during hospitalization may have already developed ASP upon hospital admission [2]. However, most studies on the factors associated with ASP incidence during hospitalization did not adequately consider whether patients had respiratory disease upon admission. Furthermore, few studies have comprehensively examined factors including patient conditions such as frailty, comorbidities, and activities of daily living (ADL) and medical care services associated with ASP incidence during hospitalization. Although some recent studies have investigated factors for ASP, they have mainly focused on specific diseases such as Parkinson’s disease and cerebrovascular disease or on specific conditions such as cognitive impairment; moreover, these studies did not consider a wider range of patient characteristics [11, 12].
Therefore, this study aimed to identify the factors associated with the incidence of ASP in older inpatients using medical claims data, including comprehensive factors, excluding patients diagnosed with respiratory disease at the time of admission.
Methods
Study design and patients
This retrospective cohort study analyzed a large-scale anonymized medical claims database of a large general acute-care hospital in Tokyo, Japan. This hospital is a secondary emergency medical center that provides urgent care and uses the Diagnosis Procedure Combination (DPC) case-mix patient classification system, similar to many other acute-care hospitals in Japan. These data comprised patient-level demographic characteristics and medical records, including diagnoses during clinical encounters with insurance claims. Major diagnoses, causes of admission, most and second-most medical-resource-intensive diagnoses, comorbidities at admission, and complications after admission were recorded using the International Classification of Disease, Tenth Revision (ICD-10) codes from the DPC data.
This study included patients hospitalized and discharged between July 2016 and March 2021. We excluded patients < 65 years of age, those who died in the hospital, those admitted for respiratory disease according to the diagnosis code of the cause of admission, those without diagnosis at the time of admission, and those discharged from the hospital within two days.
The study protocol was approved by the ethics committee of the Tokyo Metropolitan Institute for Geriatrics and Gerontology (Approval number: R18–20). The requirement for individual informed consent was waived by the committee, as the study used anonymized secondary data and employed an opt-out approach. All analyses were performed in accordance with the Ethical Guidelines for Medical and Health Research involving Human Subjects established by the Japanese Government. The study was also conducted in accordance with the ethical principles outlined in the Declaration of Helsinki.
Dependent variable
ASP incidence during hospitalization was defined according to the ICD-10 codes of major diagnoses, most and second-most medical-resource-intensive diagnoses, and complications after admission (ICD-10: J69) from the DPC data, as described previously [13].
Independent variables
Information on patient sex, age (65–74 years, 75–84 years, ≥ 85 years), place of residence before hospitalization (home, hospital, or long-term care facility), income level (patient’s copayment; 10% or 20%, 30%, unknown), primary diagnosis, presence of dysphagia at the time of admission, Barthel Index (BI) at the time of admission, Charlson Comorbidity Index (CCI) (0, ≥ 1), Hospital Frailty Risk Score (HFRS) (0–4, ≥ 5), rehabilitation during hospitalization, whether surgical treatment was performed, planned or unplanned hospitalization status, the use of enteral nutrition, and the use of central venous nutrition. The use of enteral nutrition and total parenteral nutrition was defined as present only if it was initiated within the first two days of hospitalization. We used an annual income that was determined using the copayment rate. In Japan, insurance copayment rates depend on annual household income. For example, patients aged ≥ 75, or 70–74 years, pay 10% or 20%, respectively, if their household income is < 3,700,000 yen. Patients aged ≥ 70 years with annual income of ≥ 3,700,000 yen pay 30%. For those aged ≤ 69 years, the copayment rate is always 30%, and the annual income is not considered. Therefore, we treated their copayment status as “unknown” in this study. We identified the main diseases that caused admission using the corresponding ICD-10 codes as covariates in the analysis. As previously described [14], we calculated the CCI and HFRS using DPC data. The CCI was calculated using comorbidities present at the time of admission, which served to quantitatively evaluate the burden of pre-existing conditions, primarily as a predictor of mortality and disease severity [15]. To compute the HFRS, we utilized ICD-10 codes that were associated with the primary diagnosis at admission, comorbidities at admission, main diagnosis during hospitalization, the most resource-intensive diagnosis and the second resource-intensive diagnosis during hospitalization. We followed the algorithm developed by Gilbert et al., which assigns predefined weights to specific conditions [16]. The total score helps to identify patients with frailty using ICD-10 codes in the claims database. Furthermore, we included the BI, which is used to assess the degree of assistance required by patients in performing 10 basic activities of daily living [17]. Dysphagia was defined according to the diagnosis code (ICD-10: R13) according to comorbidities present at the time of hospitalization and the primary diagnosis at the time of admission. Surgical treatment, use of enteral nutrition, and total parenteral nutrition were identified as covariates of the health process from DPC data. The BI was divided into two groups: BI–Low, with scores ≤ 25 and requiring assistance in almost all ADL levels [18] and BI_Middle-High, with scores ≥ 30 and ranging from requiring some assistance to independent.
Statistical analysis
Patient characteristics were summarized separately for the ASP and Non-ASP groups. The proportions of participants within each category are expressed as percentage. Continuous variables are presented as mean with standard deviation or median with interquartile range, as appropriate. We then examined the factors associated with ASP occurrence during hospitalization using multivariable logistic regression analysis adjusted for all independent variables (sex, age, place of residence before admission, income, main disease causing admission, dysphagia at admission, BI at admission, CCI, HFRS, rehabilitation services, surgical treatment, planned hospitalization, enteral nutrition, and total parenteral nutrition). To address the potential impact of missing income data (unknown income) on estimating ASP incidence, we performed multiple imputation by chained equations, generating 20 imputed datasets [19]. The imputation model included the outcome (ASP) and all covariates. Model discrimination was assessed with the c-statistic, calculated as the mean value across the 20 imputed datasets. Next, we conducted additional analyses using logistic regression models to examine the associations of rehabilitation service and ASP incidence in subgroups based on the BI, as the effect of rehabilitation intervention on health outcomes among older adults has been considered to be affected by ADL status [20]. Patients were initially categorized into two groups: low BI (0–25) and middle and high BI (30–100), based on a previously proposed cutoff of 27.7 points for predicting discharge eligibility in older adults [18]. To assess the impact of threshold variation, a sensitivity analysis was conducted using an alternative cutoff of 30 points. In this analysis, patients were divided into a low BI group (0–30) and a middle and high BI group (35–100). Given the relatively low number of patients with ASP in the middle and high BI category, we combined the middle and high BI groups to ensure statistical stability and to enhance interpretability by following the same approach used in the main BI group analysis. We constructed a multivariate logistic regression model for each BI group. The models were adjusted for sex, age group, annual income, place of residence before admission, main diagnosis at admission (musculoskeletal, digestive system, congestive heart failure, cerebrovascular disease, and other), dysphagia at admission, CCI, HFRS, surgical treatment, unplanned admission, and rehabilitation services.
In addition, we performed three sensitivity analyses. First, we excluded patients with respiratory diseases included in the ICD-10 codes for comorbidities at the time of admission, and factors associated with ASP that newly developed after admission were examined. Second, to assess the impact of missing income data and ensure appropriate estimation, we excluded 5,813 patients with unknown income (mostly patients aged < 70 years) as a complete case analysis. In this subsample, 504 patients whose primary diagnosis was cerebrovascular disease were further excluded, as no ASP cases occurred in this group, resulting in complete separation and unreliable logistic estimation. Third, to evaluate potential collinearity between two variables, we tested two models: one without HFRS, and one without the primary diagnosis at admission. Adjusted odds ratios (aORs) and their 95% confidence intervals (CIs) were calculated for each variable, and a p < 0.05 was considered statistically significant. All analyses were conducted using IBM SPSS Statistics for Windows, version 28.0.
Results
Figure 1 shows the flow chart of patient selection. We first identified 24,923 candidate participants admitted to the hospital and discharged during the study period. We excluded 2,931 patients aged < 65 years and 1,184 patients who died during hospitalization. We further excluded 7,659 patients discharged within two days, 1,645 patients diagnosed with respiratory diseases, and no index diagnosis at admission. The final sample comprised 11,504 patients.
Fig. 1.
Flow chart of patient selection
The overall mean age was 81.0 years (standard deviation: 7.9 years), with women accounting for more than half (56.2%) of the study population (Table 1). Table 1 also shows characteristics between patients with and without ASP during hospitalization. A greater proportion of patients in the ASP group were admitted from nursing homes (23.9% vs. 6.2%) and had lower rates of discharge to home (41.1% vs. 72.2%) than those in the Non-ASP groups. Patients with ASP tended to be older, with a higher proportion of those aged ≥ 85 years (68.3% vs. 35.1%). Furthermore, this group more frequently showed a higher comorbidity burden (CCI ≥ 1: 70.6% vs. 38.5%) and higher frailty scores (HFRS ≥ 5: 56.7% vs. 12.6%). Dysphagia at admission was more frequently observed in the ASP group (3.3% vs. 0.5%), and these patients had a lower BI (median [IQR]: 0 [0–20] vs. 55 [15–85]), more unplanned admissions (92.8% vs. 57.2%), less surgical treatment (6.7% vs. 22.8%) and rehabilitation (11.1% vs. 26.0%), and longer hospital length of stay (median [IQR]: 28 [18–41] vs. 15 [8–28] days).
Table 1.
Characteristics of patients with ASP and Non-ASP
| Variable | Total | ASP | Non-ASP | |
|---|---|---|---|---|
| Total, N (%) | 11,504 (100) | 180 (1.6) | 11,324 (98.4) | |
| Demographics | ||||
| Sex, N (%) | Men | 5037 (43.8) | 92 (51.1) | 4945 (43.7) |
| Women | 6467 (56.2) | 88 (48.9) | 6379 (56.3) | |
| Age, mean (SD) | 81.0 (7.9) | 86.9 (6.8) | 80.9 (7.9) | |
| Age groups, N (%) | 65–74 years | 2645 (23.0) | 10 (5.6) | 2635 (23.3) |
| 75–84 years | 4762 (41.4) | 47 (26.1) | 4715 (41.6) | |
| ≥ 85 years | 4097 (35.6) | 123 (68.3) | 3974 (35.1) | |
| Before admission, N (%) | Home | 10,398 (90.4) | 128 (71.1) | 10,270 (90.7) |
| Hospital | 356 (3.1) | 9 (5.0) | 347 (3.1) | |
| Nursing home | 750 (6.5) | 43 (23.9) | 707 (6.2) | |
| Discharge destination, N (%) | Home | 8247 (71.7) | 74 (41.1) | 8173 (72.2) |
| Hospital | 708 (6.2) | 36 (20.0) | 672 (5.9) | |
| Nursing home | 2368 (20.6) | 65 (36.1) | 2303 (20.3) | |
| Geriatric Intermediate care home | 181 (1.6) | 5 (2.8) | 176 (1.6) | |
| Annual income, N (%) | < 3,700,000 yen | 5601 (48.7) | 103 (57.2) | 5498 (48.6) |
| ≥ 3,700,000 yen | 716 (6.2) | 10 (5.6) | 706 (6.2) | |
| Unknown | 5187 (45.1) | 67 (37.2) | 5120 (45.2) | |
| Diagnosis at admission, N (%) | Musculoskeletal | 717 (6.2) | 5 (2.8) | 712 (6.3) |
| Digestive system | 1430 (12.4) | 13 (7.2) | 1417 (12.5) | |
| Cancer | 1236 (10.7) | 1 (0.6) | 1235 (10.9) | |
| Neurologic | 476 (4.1) | 7 (3.9) | 469 (4.1) | |
| Congestive heart failure | 644 (5.6) | 8 (4.4) | 636 (5.6) | |
| Diabetes | 398 (3.5) | 3 (1.7) | 395 (3.5) | |
| Cerebrovascular disease | 970 (8.4) | 1 (0.6) | 969 (8.6) | |
| Fracture | 576 (5.0) | 5 (2.8) | 571 (5.0) | |
| Other | 5057 (44.0) | 137 (76.1) | 4920 (43.4) | |
| Dysphagia at admission, N (%) | No | 11,441 (99.5) | 174 (96.7) | 11,267 (99.5) |
| Yes | 63 (0.5) | 6 (3.3) | 57 (0.5) | |
| Barthel Index, median [IQR] | 55 [10; 85] | 0 [0; 20] | 55 [15; 85] | |
| CCI, N (%) | 0 | 7013 (61.0) | 53 (29.4) | 6960 (61.5) |
| ≥ 1 | 4491 (39.0) | 127 (70.6) | 4364 (38.5) | |
| HFRS, N (%) | 0–4 | 9973 (86.7) | 78 (43.3) | 9895 (87.4) |
| ≥ 5 | 1531 (13.3) | 102 (56.7) | 1429 (12.6) | |
| Unplanned admission N (%) | No | 4856 (42.2) | 13 (7.2) | 4843 (42.8) |
| Yes | 6648 (57.8) | 167 (92.8) | 6481 (57.2) | |
| Surgical treatment, N (%) | No | 8909 (77.4) | 168 (93.3) | 8741 (77.2) |
| Yes | 2595 (22.6) | 12 (6.7) | 2583 (22.8) | |
| Rehabilitation, N (%) | No | 8543 (74.3) | 160 (88.9) | 8383 (74.0) |
| Yes | 2961 (25.7) | 20 (11.1) | 2941 (26.0) | |
| Enteral nutrition, N (%) | No | 11,418 (99.3) | 175 (97.2) | 11,243 (99.3) |
| Yes | 86 (0.7) | 5 (2.8) | 81 (0.7) | |
| Total parenteral nutrition, N (%) | No | 11,314 (98.3) | 178 (98.9) | 11,136 (98.3) |
| Yes | 190 (1.7) | 2 (1.1) | 188 (1.7) | |
| Hospital length of stay, median [IQR] | 15 [8; 28] | 28 [18; 41] | 15 [8; 28] | |
Abbreviations: ASP Aspiration pneumonia, SD Standard deviation, IQR Interquartile range, CCI Charlson Comorbidity Index, HFRS Hospital Frailty Risk Score
The factors positively associated with ASP incidence during hospitalization included age ≥ 85 years (aOR: 2.901, 95% CI: 1.471–5.718, reference: 65–74 years), presence of dysphagia at admission (aOR: 3.810, 95% CI: 1.410–10.242), CCI ≥ 1 (aOR: 1.839, 95% CI: 1.298–2.605), HFRS ≥ 5 (aOR: 2.918, 95% CI: 2.089–4.077), and unplanned admission (aOR: 3.068, 95% CI: 1.593–5.906) (Table 2). Additionally, Bl level at admission (aOR: 0.971, 95% CI: 0.963–0.979), surgical treatment (aOR: 0.469, 95% CI: 0.253–0.868), and rehabilitation during hospitalization (aOR: 0.435, 95% CI: 0.256–0.739) were negatively associated with ASP onset. These analyses were based on multiply imputed datasets (n = 5187; unknown income as missing data); pooled estimates were reported. The mean c-statistics across 20 imputations was 0.906 (range: 0.905–0.907). A sensitivity analysis that excluded patients with respiratory disease as a comorbidity did not significantly change the results of the multivariate logistic regression analysis. The main associated factors remained statistically significant, and these findings supported the primary analysis; the details are presented in Additional file (1) Moreover, in the sensitivity analysis after excluding patients with unknown income and those with cerebrovascular disease (n = 5813), the main associated factors, rehabilitation, BI, dysphagia at admission, CCI, HFRS, and surgical treatment, remained significantly associated with ASP. In contrast, the associations with age groups were no longer statistically significant, and a new significant negative association was observed for digestive system diseases. These results are presented in Additional file (2) Furthermore, the sensitivity analyses that excluded either HFRS or primary diagnosis at admission showed similar results, and the estimates did not change substantially. These variables did not significantly affect the model estimates (Additional file 3).
Table 2.
Adjusted odds ratio of patient characteristics associated with ASP incidence
| Characteristics | aORs (95% CIs)* | p* | |
|---|---|---|---|
| Sex | Men | 1.000 | |
| Women | 0.451 (0.326, 0.625) | < 0.001 | |
| Age, years | 65–74 | 1.000 | |
| 75–84 | 1.493 (0.737, 3.022) | 0.266 | |
| ≥ 85 | 2.901 (1.471, 5.718) | 0.002 | |
| Annual income, yen | < 3,700,000 yen | 1.000- | |
| ≥ 3,700,000 yen | 0.646 (0.301, 1.386) | 0.258 | |
| Before Admission | Home | 1.000 | |
| Hospital | 1.984 (0.914, 4.304) | 0.083 | |
| Nursing home | 1.237 (0.835, 1.832) | 0.289 | |
| Diagnosis at admission | Musculoskeletal | 1.000 | |
| Digestive system | 0.393 (0.132, 1.174) | 0.094 | |
| Cancer | 0.137 (0.016, 1.214) | 0.074 | |
| Neurologic | 0.575 (0.169, 1.960) | 0.377 | |
| Congestive heart failure | 0.510 (0.160, 1.629) | 0.256 | |
| Diabetes | 2.678 (0.575, 12.475) | 0.210 | |
| Cerebrovascular disease>† | 0.033 (0.004, 0.293) | 0.002 | |
| Fracture | 0.393 (0.107, 1.449) | 0.161 | |
| Other | 0.815 (0.317, 2.093) | 0.670 | |
| Barthel Index | 0.971 (0.963, 0.979) | < 0.001 | |
| Dysphagia at admission | No | 1.000 | |
| Yes | 3.810 (1.410, 10.293) | 0.008 | |
| CCI | 0 | 1.000 | |
| ≥ 1 | 1.839 (1.298, 2.605) | < 0.001 | |
| HFRS | 0–4 | 1.000 | |
| ≥ 5 | 2.918 (2.089, 4.077) | < 0.001 | |
| Surgical treatment | No | 1.000 | |
| Yes | 0.469 (0.253, 0.868) | 0.016 | |
| Unplanned admission | No | 1.000 | |
| Yes | 3.068 (1.593, 5.906) | < 0.001 | |
| Rehabilitation | No | 1.000 | |
| Yes | 0.435 (0.256, 0.739) | 0.002 | |
| Enteral Nutrition | No | 1.000 | |
| Yes | 1.719 (0.581, 5.087) | 0.328 | |
| Total parenteral Nutrition | No | 1.000 | |
| Yes | 0.314 (0.074, 1.328) | 0.115 | |
Estimates were pooled after multiple imputation by chained equations (m = 20). Model discrimination was evaluated using the c-statistic; the mean value across the 20 imputed datasets was 0.906 (range: 0.905–0.907)
Abbreviations: ASP Aspiration pneumonia, aOR adjusted odds ratio, CCI Charlson Comorbidity Index, CI Confidence interval, HFRS Hospital Frailty Risk Score
*Multivariable logistic regression analysis adjusted for all covariates (sex, age group, annual income, place of residence before admission, primary diagnosis at admission, BI at admission, dysphagia at admission, CCI, HFRS, surgical treatment, unplanned hospitalization, rehabilitation services, enteral nutrition, and total parenteral nutrition)
†Because ASP was observed in only one patient with cerebrovascular disease, the very low aOR should be interpreted with caution given the potential for quasi-complete separation
In the low BI group, rehabilitation during hospitalization (aOR: 0.280, 95% CI: 0.132–0.597) was significantly associated with a lower incidence of ASP compared with patients who did not receive rehabilitation (Table 3). However, rehabilitation services were not associated with ASP incidence during hospitalization among patients with middle and high BI. For the sensitivity analysis, we repeated the multivariable logistic regression using an alternative BI cutoff of 30 points (low: 0–30; middle and high: 35–100). As shown in Additional file 4, rehabilitation was significantly associated with a lower incidence of ASP in the low BI group (aOR: 0.325, 95% CI: 0.166–0.638). In contrast, the association was not statistically significant in the middle-to-high BI group (aOR: 0.768, 95% CI: 0.206–2.867), and the confidence interval was wide. These results suggest that the findings were generally consistent with our main analysis and supported the robustness of the relationship between rehabilitation and ASP, particularly among patients with lower functional status.
Table 3.
Adjusted odds ratios for the association between rehabilitation and ASP incidence by BI category
| BI category | n (%) | OR (95% CI) | aOR (95% CI)* | |
|---|---|---|---|---|
| Low (n = 3831) | ||||
| Rehabilitation | No | 2288 (59.7) | 1.000 | 1.000 |
| Yes | 1543 (40.3) | 0.164 (0.096, 0.282) | 0.280 (0.132, 0.597) | |
| Middle and high (n = 7673) | ||||
| Rehabilitation | No | 6224 (81.5) | 1.000 | 1.000 |
| Yes | 1413 (18.5) | 0.710 (0.276, 1.830) | 0.848 (0.301, 2.386) | |
In this model, data from 8,246 patients (low, n = 2523; middle and high, n = 5723) were included for analysis. The low BI and middle and high BI categories were defined by the BI of 0–25 and 30–100, respectively. Estimates were pooled after multiple imputation by chained equations (m = 20). Model discrimination was evaluated using the c-statistic; the mean value across the 20 imputed datasets was 0.801 (range: 0.799–0.807) for low BI and 0.888 (range: 0.886–0.890) for middle and high BI
Abbreviations: ASP Aspiration pneumonia, BI Barthel Index, OR Odds ratio, aOR adjusted odds ratio, CCI Charlson Comorbidity Index, CI Confidence interval, HFRS Hospital Frailty Risk Score
*Multivariable logistic regression analysis adjusted for sex, age group, annual income, place of residence before admission, primary diagnosis at admission, (musculoskeletal, digestive system, congestive heart failure, cerebrovascular disease, other), BI at admission, dysphagia at admission, CCI, HFRS, surgical treatment, unplanned admission, and rehabilitation services
Discussion
This study analyzed DPC data to identify factors associated with ASP incidence during hospitalization. These factors included male sex, advanced age, presence of dysphagia on admission, unplanned admission, low ADL levels, comorbidities, frailty, and lack of rehabilitation services during hospitalization. In addition, the use of rehabilitation services was associated with a lower incidence of ASP among older patients with ADL impairment. These findings may help hospitals identify patients who would need more attention to a lower incidence of ASP during hospitalization.
In this study, dysphagia was identified as a factor associated with ASP during hospitalization, consistent with previous findings [13, 21]. In addition, the prevalence of older inpatients with dysphagia at admission (0.6%) in this study was similar to those (0.5%) in a prior study using DPC data [13]. However, this prevalence may be underestimated compared with that estimated using clinical assessment tools [22]. As dysphagia tends to be unrecognized in acute-care settings [23], improving the assessment of dysphagia on admission may be important for identifying patients who could benefit from interventions aimed at reducing the incidence of ASP during hospitalization.
The results of the current study demonstrated that lower ADL, higher CCI, and increased frailty risk were associated with ASP incidence, consistent with previous reports [24]. Patients with a high CCI or frailty may have a weakened immune system and increased risks of infection [25, 26]. In addition, physical functional status tends to be accompanied by a decline in swallowing function, which may lead to ASP during hospitalization [21]. Therefore, the comprehensive assessment of function and comorbidity status is important for identifying older patients who are more likely to develop ASP.
In the present study, patients with lower ADL status who received rehabilitation services were associated with a lower incidence of ASP during hospitalization, while no association was observed between rehabilitation service and ASP incidence among patients with higher ADL status. Previous studies reported the effectiveness of rehabilitation interventions in preventing ASP among older patients with lower ADLs levels [27, 28]. Extended bed rest is also a risk factor for ASP [29]. Rehabilitation services help improve swallowing function by decreasing bed rest time among older adults with functional decline [30]. Moreover, a multidisciplinary team approach to decreasing bed rest time plays a crucial role in ASP prevention among older adults with ADL decline [9, 31]. Therefore, interventions to decrease bed rest time, including rehabilitation services and a multidisciplinary team approach, could be considered as part of comprehensive care strategies for older adults with ASP and low ADL status during hospitalization.
The present study found an association between lower ASP incidence and undergoing surgical treatment or having cerebrovascular disease. Surgical treatment [32] and cerebrovascular disease [33] are common risk factors for ASP. However, the incidence rate of ASP would be lower in emergency endoscopy and gastrointestinal surgery as a study in Australia reported ASP in only 19 (0.3%) of 6,431 inpatients admitted to the general surgery team [34]. The most common surgical procedure in this study was gastrointestinal surgery. Therefore, as seen in previous studies, this type of treatment may be associated with a lower incidence of ASP (Additional file 5). Patients requiring less urgent or invasive procedures are often in good general condition; therefore, ASP incidence may be relatively low. In the present study, the patients who underwent surgery were generally younger and did not have ADL impairments compared with those who did not undergo surgery (Additional file 5), which might partly explain the observed lower incidence of ASP. Furthermore, the Stroke Care Unit in the study hospital provides specialized management for admitted patients diagnosed with cerebrovascular diseases, which includes 24-hour care from the early stages of the disease by a dedicated team of physicians, nurses, and rehabilitation therapists with specialized knowledge and expertise. Early intervention by specialists and interprofessional collaboration have been reported to be effective in improving function and preventing ASP [35]. In the present study as well, such an approach may have helped prevent a decline in physical function during hospitalization and thereby contributed to a lower likelihood of ASP.
This study is the first to comprehensively examine the risk factors associated with ASP incidence during hospitalization using DPC data among older patients without respiratory disease. However, this study has several limitations. First, because this study had a retrospective observational design and used administrative data, causal relationships cannot be established, and the associations observed should be interpreted with caution. Second, the findings may not be directly generalizable to other hospitals in Japan because it used data from a single acute-care hospital in Tokyo. Third, patients who died during hospitalization were excluded from the analysis. We examined the incidence rate of ASP in the death data (n = 1182) and found that the ASP incidence during hospitalization was 2.8%. Although the cause of death could not be determined by the DPC, 43.1% of hospitalized patients had cancer as their underlying condition at the time of admission. In this study, the exclusion of 1182 patients may have resulted in an underestimation of the incidence rate of ASP and may have introduced selection bias by excluding the most severe cases. Therefore, the identified risk factors may not be generalizable to severe or fatal ASP. Fourth, details on rehabilitation services were unknown because the DPC data did not include this information. Future studies should examine differences in the effects of rehabilitation services provided by physical, occupational, and speech therapists. Fifth, the exact timing of ASP onset could not be identified in the database. Therefore, we could not clearly confirm whether some variables were collected before or after the onset of ASP. This makes it difficult to clarify whether the observed factors preceded or followed the development of ASP. Finally, unmeasured confounding cannot be completely ruled out in this study. Factors, such as eating function and oral care practices, that were not captured in the DPC data have been reported to be associated with ASP [36, 37]. Future studies would benefit from including these variables.
Conclusion
The factors associated with ASP incidence included female sex, age ≥ 85 years, presence of dysphagia on admission, unplanned admission, low ADL level, comorbidities, frailty, and lack of rehabilitation services during hospitalization. These findings may help identify older hospitalized patients who are more likely to develop ASP and thus support the planning of targeted care strategies during hospitalization.
Supplementary Information
Supplementary Material 1. Adjusted odds ratios of patient characteristics associated with ASP incidence: Excluding patients with respiratory comorbidities
Supplementary Material 2. Adjusted odds ratios of patient characteristics associated with ASP incidence: Excluding patients with unknown income and those with cerebrovascular diseases
Supplementary Material 3. Adjusted odds ratios of patient characteristics associated with the ASP incidence: Excluding HFRS (Model 1) or primary diagnosis at admission (Model 2)
Supplementary Material 4. Adjusted odds ratios for the association between rehabilitation and ASP incidence by BI category (low BI: 0–30; middle and high BI: 35–100)
Supplementary Material 5. Patient characteristics and differences between patients with surgical and non-surgical treatments
Acknowledgements
Not applicable.
Abbreviations
- ASP
Aspiration pneumonia
- ADL
Activities of Daily Living
- DPC
Diagnosis Procedure Combination
- BI
Barthel Index
- ICD-10
International Statistical Classification of Diseases and Related Health Problems, Tenth Revision
- CCI
Charlson Comorbidity Index
- HFRS
Hospital Frailty Risk Score
- aORs
Adjusted odds ratios
- CIs
Confidence intervals
- SD
Standard deviation
- IQR
Interquartile range
Authors’ contributions
S.Y.: Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Project administration, Validation, Visualization, Writing – original draft, Writing – review & editing; S.M.: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Resources, Software, Writing – review & editing; T.H.: Conceptualization, Data curation, Methodology, Project administration, Resources, Supervision, Writing – review & editing.
Funding
This study was supported by KAKENHI Grants-in-Aid for Scientific Research (WAKATE) (Grant Numbers 24K20156) and Scientific Research (B) (Grant Number 20H03924) from the Japan Society for the Promotion of Science.
Data availability
The dataset analyzed in the current study is not publicly available due to ethical restrictions and institutional policies. In accordance with the Ethical Guidelines for Medical and Health Research involving Human Subjects established by the Japanese Government, data sharing is not permitted outside the institution, in order to protect the confidentiality and privacy of study participants.
Declarations
Ethics approval and consent to participate
The study protocol was approved by the ethics committee of the Tokyo Metropolitan Institute for Geriatrics and Gerontology (Approval number: R18–20). The requirement for individual informed consent was waived by the committee, as the study used anonymized secondary data and employed an opt-out approach. All analyses were performed in accordance with the Ethical Guidelines for Medical and Health Research involving Human Subjects established by the Japanese Government. The study was also conducted in accordance with the ethical principles outlined in the Declaration of Helsinki.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Ministry of Health, Labour and Welfare, Japan. Current Population Survey. 2022. Available from: https://www.mhlw.go.jp/toukei/saikin/hw/jinkou/geppo/nengai22/dl/gaikyouR4.pdf.
- 2.Teramoto S, Fukuchi Y, Sasaki H, Sato K, Sekizawa K, Matsuse T. High incidence of aspiration pneumonia in community- and hospital-acquired pneumonia in hospitalized patients: a multicenter, prospective study in Japan. J Am Geriatr Soc. 2008;56(3):577–9. [DOI] [PubMed] [Google Scholar]
- 3.Gupte T, Knack A, Cramer JD. Mortality from aspiration pneumonia: incidence, trends, and risk factors. Dysphagia. 2022;37(6):1493–500. [DOI] [PubMed] [Google Scholar]
- 4.Ministry of Health, Labour and Welfare, Japan. Second meeting of the working group on home medical care and medical-care collaboration: Ministry of Health, Labour and Welfare. 2016. Available from: https://www.mhlw.go.jp/file/05-Shingikai-10801000-Iseikyoku-Soumuka/0000135467.pdf.
- 5.United Nations DoEaSA, Population Division. World population prospects 2022: summary of results. New York, NY: United Nations; 2022. [Google Scholar]
- 6.Wakabayashi T, Hamaguchi S, Morimoto K, Adult Pneumonia Study Group J. Clinically defined aspiration pneumonia is an independent risk factor associated with long-term hospital stay: a prospective cohort study. BMC Pulm Med. 2023;23(1):351. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Komiya K, Rubin BK, Kadota JI, Mukae H, Akaba T, Moro H, et al. Prognostic implications of aspiration pneumonia in patients with community acquired pneumonia: a systematic review with meta-analysis. Sci Rep. 2016;6:38097. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Taylor JK, Fleming GB, Singanayagam A, Hill AT, Chalmers JD. Risk factors for aspiration in community-acquired pneumonia: analysis of a hospitalized UK cohort. Am J Med. 2013;126(11):995–1001. [DOI] [PubMed] [Google Scholar]
- 9.Simpson AJ, Allen J-L, Chatwin M, Crawford H, Elverson J, Ewan V, et al. BTS clinical statement on aspiration pneumonia. Thorax. 2023;78(Suppl 1):s3–21. [DOI] [PubMed] [Google Scholar]
- 10.Nakamura T, Kurosaki S. Effects of early dysphagia rehabilitation by speech-language-hearing therapists on patients with severe aspiration pneumonia. Prog Rehabil Med. 2020;5:20200020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Zhao T, Zhang Y, Wang K, Yu H, Lin L, Qin X, et al. Identifying risk factors for aspiration in patients hospitalized with community-acquired pneumonia. Int J Clin Pract. 2023;2023:2198259. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Chua WY, Wang JDJ, Chan CKM, Chan L-L, Tan E-K. Risk of aspiration pneumonia and hospital mortality in Parkinson disease: a systematic review and meta-analysis. Eur J Neurol. 2024;31(12):e16449. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Nonaka S, Fujii S, Hara M, Morita S, Sueoka E, Node K, et al. Incidence of aspiration pneumonia during hospitalization in Japanese hospitalized cases did not increase whereas concern factors were exacerbated in a time-dependent manner: analysis of diagnosis procedure combination (DPC) data. J Clin Biochem Nutr. 2018;63(1):66–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Mitsutake S, Yano S, Ishizaki T, Furuta K, Hatakeyama A, Sugiyama M, et al. Association of functional and cognitive impairment severity with discharge to long-term care facilities in older patients admitted to a general acute care hospital from home. Arch Gerontol Geriatr. 2023;115:105111. [DOI] [PubMed] [Google Scholar]
- 15.Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi J-C, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005;43(11):1130–9. [DOI] [PubMed] [Google Scholar]
- 16.Gilbert T, Neuburger J, Kraindler J, Keeble E, Smith P, Ariti C, et al. Development and validation of a hospital frailty risk score focusing on older people in acute care settings using electronic hospital records: an observational study. Lancet. 2018;391(10132):1775–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Mahoney FI, Barthel DW. Functional evaluation: The Barthel Index. Md State Med J. 1965;14:61–5. [PubMed] [Google Scholar]
- 18.Strini V, Piazzetta N, Gallo A, Schiavolin R. Barthel index: creation and validation of two cut-offs using the BRASS index. Acta Biomed. 2020;91(2–s):19–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Graham JW, Olchowski AE, Gilreath TD. How many imputations are really needed? Some practical clarifications of multiple imputation theory. Prev Sci. 2007;8(3):206–13. [DOI] [PubMed] [Google Scholar]
- 20.Greysen SR. Activating hospitalized older patients to confront the epidemic of low mobility. JAMA Intern Med. 2016;176(7):928–9. [DOI] [PubMed] [Google Scholar]
- 21.Manabe T, Teramoto S, Tamiya N, Okochi J, Hizawa N. Risk factors for aspiration pneumonia in older adults. PLoS ONE. 2015;10(10):e0140060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Clavé P, Shaker R. Dysphagia: current reality and scope of the problem. Nat Rev Gastroenterol Hepatol. 2015;12(5):259–70. [DOI] [PubMed] [Google Scholar]
- 23.Ekberg O, Hamdy S, Woisard V, Wuttge-Hannig A, Ortega P. Social and psychological burden of dysphagia: its impact on diagnosis and treatment. Dysphagia. 2002;17(2):139–46. [DOI] [PubMed] [Google Scholar]
- 24.Lanspa MJ, Jones BE, Brown SM, Dean NC. Mortality, morbidity, and disease severity of patients with aspiration pneumonia. J Hosp Med. 2013;8(2):83–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Aubert CE, Schnipper JL, Fankhauser N, Marques-Vidal P, Stirnemann J, Auerbach AD, et al. Patterns of multimorbidity associated with 30-day readmission: a multinational study. BMC Public Health. 2019;19(1):738. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Iwai-Saito K, Shobugawa Y, Aida J, Kondo K. Frailty is associated with susceptibility and severity of pneumonia in older adults (A JAGES multilevel cross-sectional study). Sci Rep. 2021;11(1):7966. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Ebihara T. Comprehensive approaches to aspiration pneumonia and dysphagia in the elderly on the disease Time-Axis. J Clin Med. 2022;11(18):5323. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Pitton Rissardo J, Fornari Caprara AL. A comprehensive literature review of dysphagia screening protocols for stroke. Preprints: Preprints; 2024. [Google Scholar]
- 29.Yatera K, Yamasaki K. Management of the diagnosis and treatment of pneumonia in an aging society. Intern Med. 2025;64(4):503–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Ishii M, Nakagawa K, Yoshimi K, Okumura T, Hasegawa S, Yamaguchi K, et al. Time spent away from bed to maintain swallowing function in older adults. Gerontology. 2023;69(1):37–46. [DOI] [PubMed] [Google Scholar]
- 31.Barczi SR, Sullivan PA, Robbins J. How should dysphagia care of older adults differ? Establishing optimal practice patterns. Semin Speech Lang. 2000;21(4):347–61. [DOI] [PubMed] [Google Scholar]
- 32.Studer P, Räber G, Ott D, Candinas D, Schnüriger B. Risk factors for fatal outcome in surgical patients with postoperative aspiration pneumonia. Int J Surg. 2016;27:21–5. [DOI] [PubMed] [Google Scholar]
- 33.Lidetu T, Muluneh EK, Wassie GT. Incidence and predictors of aspiration pneumonia among stroke patients in Western Amhara Region, North-West ethiopia: a retrospective follow up study. Int J Gen Med. 2023;16:1303–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Lindfield KJ, Little A. Identification of pre-operative, intra-operative and post-operative risk factors for aspiration pneumonia in patients undergoing abdominal surgery. Global Journal of Perioperative Medicine. 2019;3(1):001–6. [Google Scholar]
- 35.Armstrong JR, Mosher BD. Aspiration pneumonia after stroke: intervention and prevention. Neurohospitalist. 2011;1(2):85–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Terpenning M. Geriatric oral health and pneumonia risk. Clin Infect Dis. 2005;40(12):1807–10. [DOI] [PubMed] [Google Scholar]
- 37.Langmore SE, Terpenning MS, Schork A, Chen Y, Murray JT, Lopatin D, et al. Predictors Aspiration Pneumonia: How Important Is Dysphagia? Dysphagia. 1998;13(2):69–81. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Material 1. Adjusted odds ratios of patient characteristics associated with ASP incidence: Excluding patients with respiratory comorbidities
Supplementary Material 2. Adjusted odds ratios of patient characteristics associated with ASP incidence: Excluding patients with unknown income and those with cerebrovascular diseases
Supplementary Material 3. Adjusted odds ratios of patient characteristics associated with the ASP incidence: Excluding HFRS (Model 1) or primary diagnosis at admission (Model 2)
Supplementary Material 4. Adjusted odds ratios for the association between rehabilitation and ASP incidence by BI category (low BI: 0–30; middle and high BI: 35–100)
Supplementary Material 5. Patient characteristics and differences between patients with surgical and non-surgical treatments
Data Availability Statement
The dataset analyzed in the current study is not publicly available due to ethical restrictions and institutional policies. In accordance with the Ethical Guidelines for Medical and Health Research involving Human Subjects established by the Japanese Government, data sharing is not permitted outside the institution, in order to protect the confidentiality and privacy of study participants.