論文ID: CJ-24-1030
Background: This study aimed to elucidate the age trends among non-surgical patients requiring intensive care over a 10-year period and the prognostic impact of aging in relation to their underlying etiologies.
Methods and Results: In all, 4,279 non-surgical patients requiring intensive care from 2012 to 2021 were enrolled in the study. Patient backgrounds and prognoses were compared among age 4 groups: Group A, age <60 years (n=910); Group B, age 60–69 years (n=1,062); Group C, age 70–79 years (n=1,355); and Group D, age ≥80 years (n=952). During the study period, the number of patients aged 60–69 years decreased significantly with time, whereas the number aged over 80 years increased significantly. A multivariate Cox regression model identified Group D as an independent predictor of 365-day all-cause mortality (hazard ratio [HR] 2.070; 95% confidence interval [CI] 1.619–2.646) relative to Group A. Multivariate logistic regression analysis indicated that the presence of sepsis was independently associated with 365-day mortality, especially in the cohort aged ≥80 years (HR 1.878; 95% CI 1.270–2.777; P=0.002).
Conclusions: The mean age of patients requiring non-surgical intensive care is increasing annually, and greater age was identified as a significant factor associated with a higher 365-day mortality rate. The presence of sepsis was linked to increased 365-day mortality among older individuals.
Aging is currently a serious social issue worldwide, and aged frail patients more readily require intensive care than do young and/or middle-aged patients. It is expected that by 2030 more than one-quarter of patients in intensive care units (ICUs) will be aged ≥80 years.1 It is also widely acknowledged that the population in Japan is aging at a faster rate than populations in Western countries, with life expectancy increasing every year.2,3 Although age has been suggested to have a prognostic impact for patients in ICUs in various countries,1,4–6 the phenomenon of aging in Japanese ICUs has rarely been discussed. Even with successful ICU treatment, some aged patients experience a decline in their activities of daily living due to muscle wasting during the period of admission; these patients cannot be discharged directly home and must instead undergo additional rehabilitation at another hospital.7 These post-ICU factors can easily lead to adverse long-term outcomes in elderly patients.
Recently, the number of patients with acute heart failure (AHF) and sepsis has increased in the aged cohort.8,9 These 2 conditions are recognized as major concerns in most non-surgical ICUs. In addition, the number of heart failure (HF) patients has rapidly increased to an “HF pandemic” level.10 If current trends persist, HF prevalence and mortality are expected to increase substantially in the next decade.8,10 The prevalence of sepsis, which has been suggested to be an adverse outcome in particular aged categories, is also expected to increase in an aging society.11,12 Therefore, the prognostic impact of differences underlying conditions should be considered. We hypothesized that age of patients admitted to the ICU may have increase over time, and that this would be associated with adverse long-term outcomes due to the inherent severity of the conditions of older admitted patients. Thus, in the present study we evaluated non-surgical ICU patients over a 10-year period and analyzed associations between aging and patient characteristics, management, and short- and long-term prognoses. Furthermore, subgroup analyses were conducted based on etiological differences. The findings of our study may contribute to improving the prognoses of older adults admitted to ICUs in an aging society.
We screened 4,477 patients admitted to the non-surgical ICU from the emergency room and general wards of Nippon Medical School Chiba Hokusoh Hospital in Inzai, Japan, between January 2012 and December 2021. Among these patients, the 198 who were readmitted to the non-surgical ICU during the same hospitalization period were excluded, leaving 4,279 patients included in the study.
The 2 ICU departments (surgical and non-surgical) at the Nippon Medical School Chiba Hokusoh Hospital are closed ICUs. Patients admitted to the surgical ICU (i.e., for trauma, burns, drowning, or cerebrovascular disease) were excluded from the present study. The non-surgical ICU functions as a cardiac and non-cardiac care unit, and all the physicians in the non-surgical ICU are cardiologists. Over 10-year period there have been 8–10 cardiologist working at the ICU: 3–4 double board-certificated intensivists and cardiologists, 3–4 board-certificated cardiologists, and 2–3 senior cardiologists in training. Thus, all the patients enrolled in this study who were all admitted to the non-surgical ICU were treated by a cardiologist. Approximately 70% of the enrolled patients had cardiovascular diseases, with the remaining patients having non-cardiovascular diseases. Patients who were admitted to the non-surgical ICU but had a subsequent cardiovascular surgical procedure (e.g., blow-out or oozing rupture and ventricular septum perforation after acute myocardial infarction, worsening acute aortic disease, and thrombectomy for pulmonary embolism) were enrolled in this study.
Major diseases (e.g., acute coronary syndrome [ACS], AHF, and sepsis) were diagnosed based on newly updated (up to 2024) guidelines.13–15 Patients with the following conditions were admitted to the non-surgical ICU: ACS, AHF, arrhythmia, acute aortic dissection, infectious disease (severe sepsis, septic shock, infective endocarditis, pneumonia, pericarditis, or myocarditis), pulmonary embolism, coronary spasm angina, takotsubo cardiomyopathy, respiratory emergency disease, acute kidney injury, and severe symptoms that required a differential diagnosis. Although there are no definitive criteria for admission to the ICU, the occurrence of these conditions in patients admitted to the non-surgical ICU at Nippon Medical School Chiba Hokusoh Hospital has been consistent for decades.
ProcedurePatients requiring intensive care were categorized into 4 age groups: Group A, age <60 years (n=910); Group B, age 60–69 years (n=1,062); Group C, age 70–79 years (n=1,355); and Group D, age ≥80 years (n=952). The 4 groups were compared in terms of age, sex, reason for ICU admission, medical history (diabetes, hypertension, dyslipidemia, hyperuricemia, hemodialysis before admission, and chronic kidney disease), vital signs and status upon admission (systolic blood pressure, diastolic blood pressure, heart rate, respiratory rate, body temperature, body mass index [BMI], and left ventricular ejection fraction [LVEF]), arterial blood gas (pH, pCO2, pO2, HCO3−, O2 saturation, and lactate), laboratory data (white blood cell count, hemoglobin, blood urea nitrogen, creatinine, sodium, potassium, blood glucose, C-reactive protein, and B-type natriuretic peptide), and mechanical support during the ICU stay (non-invasive positive pressure ventilation, endotracheal intubation, intra-aortic balloon pumping, percutaneous cardiopulmonary support, and continuous hemodiafiltration). Acute Physiology and Chronic Health Evaluation (APACHE II) scores16 were also compared among the 4 groups.
LVEF was calculated using the Teichholz method or Simpson’s method at admission using a Sonos 5500 (Hewlett Packard, Palo Alto, CA, USA) or Vivid I (GE Yokogawa Medical, Tokyo, Japan). The method was chosen on a case-by-case basis. Because LVEF was measured during the acute phase, it was not adequately evaluated in patients with severe orthopnea. LVEF in some patients with left ventricular asynergy could not be adequately measured by the Teichholz method during the acute phase.
Prognostic Evaluation by Year of AdmissionThe primary endpoint of this study was long-term prognosis, including 365-day all-cause mortality. We evaluated the prognostic impact of patient age on 365-day mortality in Groups B, C, and D relative to Group A using a Cox proportional hazards regression model and Kaplan-Meier curves. All-cause mortality was ascertained using information obtained directly from hospital medical records or from other institutions via telephone. Cox regression analysis was used to determine the hazard ratio (HR) for 365-day mortality. Reason for ICU admission, sex, BMI, systolic blood pressure, hemoglobin, estimated glomerular filtration rate (eGFR), C-reactive protein, and LVEF were selected for inclusion in the multivariate logistic regression model. Two etiologies (AHF and sepsis) which had high mortality during 365 days were selected as representative etiologies for ICU admission. Other confounders were selected because they were generally considered as important variables for mortality in various reports. Finally, subgroup analysis using Kaplan-Meier curves was performed to evaluate differences in etiology within the age cohorts (Groups A, B, C, and D) and differences in age within etiologies for ICU admission (i.e., ACS, AHF, acute aortic disease, and sepsis).
Patients with sepsis were divided into 4 groups according to the organ of origin (i.e., respiratory infections, urinary tract infections, gastrointestinal infections and others [e.g. cellulitis, hepatobiliary system, heart and unknown origin]). Patient characteristics and prognoses were also evaluated among these 4 groups.
Statistical AnalysesAll statistical analyses were performed using SPSS 22.0 software (SPSS Japan Institute, Tokyo, Japan). The normality of data distribution was assessed using the Shapiro-Wilk test, In this study, all numerical data are presented as the median with interquartile range (IQR). Comparisons among 4 groups were made using the Jonckheere-Terpstra test, whereas comparisons between 2 groups were made using the Kruskal-Wallis test. The Chi-squared test was used to compare proportions. P<0.05 (two-tailed) were considered statistically significant. Cumulative survival rates for each group were analyzed using Kaplan-Meier curves, with the significance of differences between groups evaluated using log-rank tests. Multivariate Cox regression analysis was used to compare the risk of all-cause mortality among the 4 groups, with patients aged <60 years (Group A) serving as the reference group.
Ethics ReviewThe study protocol was approved by the Research Ethics Committee of Nippon Medical School Chiba Hokusoh Hospital. Because of the retrospective design of the study, the need for written informed consent was waived. Based on the advice of the Ethics Committee, information on this study was presented as a poster displayed at the hospital and published on the hospital’s homepage, where it could be viewed by anyone. All procedures were performed in accordance with the tenets of the Declaration of Helsinki.
The age distribution of patients who required non-surgical intensive care is shown in Figure 1A. The median age of patients was 71 years (IQR 61–79 years), and approximately three-quarters of all patients (n=3,046 patients; 71.2%) were aged 60–84 years. The number of patients admitted each year (i.e., 2012–2013, 2014–2015, 2016–2017, 2018–2019, and 2020–2021) was compared across Groups A, B, C, and D (Figure 1B). The number of patients aged 60–69 years (Group B) decreased significantly with time, whereas the number of patients aged >80 years (Group D) increased significantly.
(A) Age distribution in non-surgical intensive care unit patients. The peak number of patients was observed in the 70–74-year age group, with 511 (11.9%) patients aged 80–84 years, 309 (7.2%) patients aged 85–89 years, and 92 (2.2%) patients aged over 90 years included in the high-age cohort. (B) The number of patients in each age group according to the year of admission (2012–2013, 2014–2015, 2016–2017, 2018–2019, and 2020–2021). The number of patients aged 60–69 years decreased significantly from 2012–2013 to 2020–2021, whereas the number of patients aged ≥80 years increased significantly over time.
The characteristics of patients in each of the 4 age groups are presented in Table 1. The enrolled cohort included 3,055 men (71.4%) and 1,224 women (28.6%). Most male patients were aged 65–74 years (973 patients; 31.8%), and most female patients were aged 75–84 years (401 patients; 32.8%; Figure 2A). In all, 1,840 (43.0%), 976 (22.8%), and 183 (4.3%) patients had ACS, AHF, and acute aortic disease, respectively. In addition, 142 (3.3%), 363 (8.5%), 318 (7.4%), and 338 (7.9%) patients had pulmonary embolism, infectious disease, arrhythmia, and other diseases, respectively. The median age of patients with ACS, AHF, and infectious disease was 68 (IQR 59–76), 76 (IQR 67–82), and 73 (IQR 64–80) years, respectively (Figure 2B). Most patients with ACS were aged 60–79 years (1,069 patients, 58.1%); most of those with AHF were aged 70–89 years (632 patients; 64.8%), and most of those with infectious disease were aged 60–84 years (234 patients; 64.4%).
Associations Between Patient Characteristics and Age
| All patients (n=4,279) |
Age group | P value | ||||
|---|---|---|---|---|---|---|
| Group A (age <60 years; n=910) |
Group B (age 60–69 years; n=1,062) |
Group C (age 70–79 years; n=1,355) |
Group D (age ≥80 years; n=952) |
|||
| Age (years) | 71 [61–79] | 51 [45–56] | 65 [63–67] | 74 [72–77] | 84 [81–87] | <0.001 |
| Male sex | 3,055 (71.4) | 725 (79.7) | 820 (77.2) | 963 (71.1) | 547 (57.5) | <0.001 |
| Diagnosis and classification | ||||||
| Acute coronary syndrome | 1,839 (43.0) | 479 (52.6) | 527 (49.6) | 542 (40.0) | 291 (30.6) | <0.001 |
| AHF | 977 (22.8) | 116 (12.7) | 188 (17.7) | 324 (23.9) | 349 (36.7) | <0.001 |
| Acute aortic disease | 183 (4.3) | 40 (4.4) | 52 (4.9) | 61 (4.5) | 30 (3.2) | 0.248 |
| Pulmonary thromboembolism | 142 (3.3) | 48 (5.3) | 32 (3.0) | 40 (3.0) | 22 (2.3) | 0.002 |
| Arrhythmia | 318 (7.4) | 62 (6.8) | 74 (7.0) | 105 (7.7) | 77 (8.1) | 0.648 |
| Coronary spasms | 50 (1.2) | 21 (2.3) | 17 (1.6) | 9 (0.7) | 3 (0.3) | <0.001 |
| Takotsubo cardiomyopathy | 69 (1.6) | 9 (0.9) | 9 (0.8) | 28 (2.1) | 23 (2.4) | 0.008 |
| Sepsis | 321 (7.5) | 52 (5.7) | 62 (5.8) | 121 (8.9) | 86 (9.0) | <0.001 |
| Other intensive care disease | 380 (8.9) | 83 (9.1) | 101 (9.5) | 123 (9.1) | 73 (7.7) | <0.001 |
| Medical history | ||||||
| Hypertension | 2,967 (69.3) | 496 (54.5) | 730 (68.7) | 981 (72.4) | 760 (79.8) | <0.001 |
| Diabetes | 1,622 (37.9) | 304 (33.4) | 427 (40.2) | 530 (39.1) | 361 (37.9) | 0.011 |
| Dyslipidemia | 2,188 (51.1) | 465 (51.1) | 598 (56.3) | 720 (53.1) | 405 (42.5) | <0.001 |
| Hyperuricemia | 1,224 (28.6) | 246 (27.0) | 311 (29.3) | 392 (28.9) | 275 (28.9) | 0.684 |
| Hemodialysis before admission | 253 (5.9) | 38 (4.2) | 75 (7.1) | 85 (6.3) | 55 (5.8) | <0.001 |
| CKD | 1,219 (28.5) | 149 (16.4) | 246 (23.2) | 425 (31.4) | 399 (41.9) | <0.001 |
| Vital signs and status | ||||||
| SBP (mmHg) | 139 [110–163] | 140 [113–160] | 138 [112–163] | 138 [109–163] | 140 [111–165] | 0.381 |
| DBP (mmHg) | 79 [60–94] | 83 [65–100] | 80 [62–95] | 76 [60–91] | 76 [59–90] | <0.001 |
| Pulse (beats/min) | 87 [70–107] | 88 [72–106] | 86 [69–106] | 87 [70–109] | 88 [68–106] | 0.251 |
| Respiratory rate (beats/min) | 20 [16–27] | 20 [16–25] | 20 [16–26] | 20 [16–27] | 22 [17–29] | <0.001 |
| Body temperature (℃) | 36.4 [35.9–36.8] |
36.5 [36.0–36.9] |
36.4 [35.8–36.8] |
36.4 [35.8–36.8] |
36.3 [35.8–36.8] |
0.003 |
| BMI (kg/m2) | 23.5 [21.0–26.1] |
25.6 [22.9–29.1] |
23.9 [21.6–26.3] |
23.0 [20.7–25.3] |
22.0 [20.0–24.2] |
0.001 |
| LVEF (%) | 51 [39–63] | 56 [40–65] | 50 [38–61] | 51 [36–62] | 50 [38–63] | <0.001 |
| Arterial blood gas | ||||||
| pH | 7.41 [7.34–7.45] |
7.42 [7.38–7.50] |
7.42 [7.35–7.45] |
7.41 [7.33–7.45] |
7.39 [7.30–7.44] |
<0.001 |
| pCO2 (mmHg) | 37 [32–42] | 37 [33–41] | 37 [32–42] | 37 [32–43] | 38 [32–46] | 0.001 |
| pO2 (mmHg) | 113 [82–164] | 119 [86–170] | 113 [83–160] | 111 [82–165] | 107 [78–160] | 0.003 |
| HCO3− (mmol/L) | 22.7 [19.5–25.1] |
23.2 [20.4–25.1] |
22.8 [19.5–25.1] |
22.6 [19.4–25.0] |
22.2 [19.1–25.0] |
0.012 |
| O2 saturation (%) | 98 [96–99] | 98 [97–99] | 98 [96–99] | 98 [96–99] | 98 [95–99] | <0.001 |
| Lactate (mmol/L) | 1.6 [1.1–2.9] | 1.6 [1.1–2.7] | 1.7 [1.1–2.8] | 1.6 [1.0–3.1] | 1.7 [1.0–3.1] | 0.292 |
| Laboratory data | ||||||
| WBC (U/L) | 9,300 [7,100–12,195] |
10,060 [7,000–13,078] |
9,335 [7,273–11,900] |
9,140 [6,900–12,000] |
8,805 [6,790–11,545] |
<0.001 |
| Hemoglobin (g/dL) | 13.1 [11.0–14.7] |
14.5 [12.7–15.8] |
13.7 [12.0–15.0] |
12.8 [10.9–14.2] |
11.5 [9.9–13.0] |
0.001 |
| BUN (mg/dL) | 19.2 [14.7–31.0] |
15.7 [12.1–20.6] |
18.0 [14.3–26.3] |
20.5 [15.8–33.0] |
25.0 [18.1–41.4] |
0.001 |
| Creatinine (mg/dL) | 0.97 [0.74–1.52] |
0.85 [0.69–1.17] |
0.94 [0.73–1.45] |
1.01 [0.75–1.65] |
1.15 [0.81–1.87] |
0.001 |
| Sodium (mmol/L) | 140 [137–142] | 140 [138–142] | 140 [137–142] | 140 [137–142] | 140 [137–142] | 0.971 |
| Potassium (mmol/L) | 4.1 [3.8–4.5] | 4.0 [3.7–4.4] | 4.1 [3.7–4.5] | 4.1 [3.8–4.6] | 4.2 [3.9–4.7] | <0.001 |
| Blood glucose (mg/dL) | 151 [120–218] | 143 [115–210] | 156 [124–227] | 153 [119–217] | 156 [122–218] | 0.001 |
| CRP (mg/dL) | 0.42 [0.10–2.85] |
0.27 [0.09–1.44] |
0.32 [0.09–2.18] |
0.51 [0.11–3.69] |
0.68 [0.15–4.01] |
<0.001 |
| BNP (pg/mL) | 172 [41–618] | 51 [15–243] | 119 [31–501] | 217 [65–673] | 412 [136–966] | 0.001 |
| Scoring | ||||||
| APACHE II (points) | 11 [8–17] | 7 [5–12] | 10 [7–16] | 12 [9–17] | 14 [10–19] | 0.001 |
| Mechanical support during ICU stay | ||||||
| NPPV | 990 (23.1) | 120 (13.2) | 202 (19.0) | 340 (25.1) | 328 (34.5) | <0.001 |
| ETI | 833 (19.5) | 162 (17.8) | 216 (20.3) | 288 (21.3) | 167 (17.5) | 0.067 |
| Pacing | 225 (5.3) | 34 (3.7) | 55 (5.2) | 65 (4.8) | 71 (7.5) | 0.003 |
| IABP | 594 (13.9) | 128 (14.1) | 172 (16.2) | 109 (14.7) | 95 (10.0) | <0.001 |
| PCPS | 218 (5.1) | 58 (6.4) | 63 (5.9) | 72 (5.3) | 25 (2.6) | <0.001 |
| CHDF | 664 (15.5) | 110 (12.1) | 170 (16.0) | 231 (17.0) | 153 (16.1) | 0.012 |
| Short-term prognosis | ||||||
| ICU hospitalization (days) | 3 [2–6] | 3 [2–6] | 3 [2–6] | 3 [2–6] | 4 [2–7] | <0.001 |
| Total hospitalization (days) | 17 [11–30] | 14 [10–26] | 16 [11–27] | 18 [11–31] | 20 [12–36] | <0.001 |
| In-hospital mortality | 561 (13.1) | 80 (8.8) | 124 (11.7) | 197 (14.5) | 160 (16.8) | <0.001 |
Unless indicated otherwise, data are given as the median [interquartile range] or n (%). P values among the 4 groups were determined using the Jonckheere-Terpstra or Chi-squared test. AHF, acute heart failure; APACHE, Acute Physiology and Chronic Health Evaluation; BMI, body mass index; BNP, B-type natriuretic peptide; BUN, blood urea nitrogen; CHDF, continuous hemodiafiltration; CKD, chronic kidney disease; CRP, C-reactive protein; DBP, diastolic blood pressure; ETI, endotracheal intubation; IABP, intra-aortic balloon pumping; ICU, intensive-care unit; LVEF, left ventricular ejection fraction measured on echocardiography; NPPV, noninvasive positive pressure ventilation; PCPS, percutaneous cardiopulmonary support; SBP, systolic blood pressure; WBC, white blood cell.
Age distribution of non-surgical intensive care unit patients according to (A) sex, (B) etiology, (C) body mass index, and (D) estimated glomerular filtration rate (eGFR). (A) Most male patients were aged 65–74 years (973 patients; 31.8%), whereas most female patients were aged 75–84 years (401 patients; 32.8%). (B) Most acute coronary syndrome (ACS) patients were aged 60–79 years (1,069 patients; 58.1%), most AHF patients were aged 70–89 years (632 patients; 64.8%), and most infectious disease patients were aged 60–84 years (234 patients; 64.4%). (C) BMI decreased gradually with increasing age. BMI was highest in group aged 40–44 years (median 26.7 kg/m2; interquartile range [IQR] 23.7–30.1 kg/m2) and lowest in those aged >90 years (median 20.7 kg/m2; IQR 19.4–22.4 kg/m2). (D) eGFR decreased gradually with aging, and was highest in those aged <40 years (median 81.5 mL/min/1.73 m2; IQR 57.5–102.2 mL/min/1.73 m2) and the lowest in aged >90 years (median 41.2 mL/min/1.73 m2; IQR 22.6–51.3 mL/min/1.73 m2). (C,D) Boxes show the IQR, with the median value indicated by the horizontal line; whiskers show the range. AHF, acute heart failure.
Compared with Group A, Group D (age ≥80 years) had significantly fewer men, lower LVEF, lower BMI, a higher number of patients on hemodialysis before admission, and a higher number of patients with chronic kidney disease. ACS was significantly less frequent, whereas AHF and infectious disease were more frequent in Group D than in Group A. Group D also had a significantly higher prevalence of underlying medical conditions, including diabetes and hypertension, than Group A. Laboratory findings indicated that Group D had significantly lower hemoglobin and higher serum blood urea nitrogen, creatinine, potassium, C-reactive protein, and B-type natriuretic peptide concentrations than Group A. BMI and eGFR decreased gradually with aging (Figure 2C,D). In addition, the use of additional mechanical support (intra-aortic balloon pumping and percutaneous cardiopulmonary support) was significantly less frequent in Group D than in Group A (Table 1). APACHE II scores were significantly higher in Group D than in Group A (median 14 [IQR 10–19] vs. 7 [IQR 5–12] points). As a result, the median duration of the ICU stay was higher in Group D than in Group A (median 4 [IQR 2–7] vs. 3 [IQR 2–6] days). These results suggest that the severity of the conditions of non-surgical patients admitted to the ICU was exacerbated by aging, but that physicians refrained from using mechanical support because of the frailty of the aged individuals.
Prognostic Differences by Year of AdmissionThe median follow-up period was 365 days (IQR 161–365 days). There were 561 (13.1%) in-hospital deaths, and 819 (19.1%) patients died within 365 days. In-hospital mortality was significantly higher in Group D (16.8%) than in Group A (8.8%). Kaplan-Meier curves for the different age groups are shown in Figure 3. Survival rates were significantly lower in Group D than in Groups A, B, and C. The multivariate Cox regression model identified Group D as being independently associated with 365-day all-cause mortality (HR 2.070; 95% confidence interval [CI] 1.619–2.646; P<0.001) relative to Group A (Figure 3).
Kaplan-Meier survival curves and Cox regression model for the 4 age groups. The Kaplan-Meier survival curves revealed that all-cause mortality within 365 days was significantly poorer in the group aged ≥80 years than in the groups aged 70–79, 60–69, and <60 years. The Cox regression model indicated that age 60–69, 70–79, and ≥80 years was an independent predictor of 365-day mortality, with hazard ratios of 1.321 (95% confidence interval [CI] 1.024–1.703; P=0.032), 1.586 (95% CI 1.251–2.012; P<0.001), and 2.070 (95% CI 1.619–2.646; P<0.001), respectively). CRP, C-reactive protein.
Multivariate logistic regression analysis identified several factors that were independently associated with 365-day mortality in each cohort (Table 2). Systolic blood pressure (<100 mmHg), hemoglobin (per 1.0-g/dL decrease), C-reactive protein (per 1.0-mg/dL increase), and lower LVEF (<40%) were all associated with increased mortality across all age categories. Interestingly, BMI and eGFR were independently associated with 365-day mortality only in the ≥80-year-old cohort (Group D).
Multivariate Cox Regression Analysis for the Incidence of 365-Day Mortality in Each Category
| 365-day mortality | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Whole cohort | Age <60 years | Age 60–69 years | Age 70–79 years | Age ≥80 years | |||||||||||
| HR | 95% CI | P value |
HR | 95% CI | P value |
HR | 95% CI | P value |
HR | 95% CI | P value |
HR | 95% CI | P value |
|
| EtiologyA | |||||||||||||||
| AHF (yes) | 0.924 | 0.776– 1.102 |
0.380 | 0.589 | 0.311– 1.115 |
0.104 | 0.762 | 0.505– 1.152 |
0.198 | 0.924 | 0.687– 1.242 |
0.598 | 0.979 | 0.728– 1.317 |
0.890 |
| Sepsis (yes) | 1.195 | 0.943– 1.514 |
0.141 | 1.048 | 0.452– 2.429 |
0.914 | 0.995 | 0.595– 1.665 |
0.986 | 0.966 | 0.632– 1.478 |
0.874 | 1.878 | 1.270– 2.777 |
0.002 |
| Male sex | 0.995 | 0.857– 1.155 |
0.946 | 0.697 | 0.445– 1.092 |
0.115 | 1.175 | 0.825– 1.674 |
0.371 | 1.128 | 0.861– 1.477 |
0.381 | 1.033 | 0.805– 1.327 |
0.798 |
| BMI (per 1.0-kg/m2 increase) |
0.977 | 0.961– 0.994 |
0.007 | 1.033 | 0.991– 1.077 |
0.125 | 0.975 | 0.938– 1.013 |
0.189 | 0.994 | 0.965– 1.024 |
0.704 | 0.953 | 0.921– 0.987 |
0.006 |
| SBP <100 mmHg |
3.110 | 2.674– 3.618 |
<0.001 | 5.089 | 3.222– 8.038 |
<0.001 | 3.340 | 2.413– 4.624 |
<0.001 | 3.060 | 2.361– 3.966 |
<0.001 | 2.640 | 1.983– 3.517 |
<0.001 |
| Hemoglobin (per 1.0-mg/dL decrease) |
0.839 | 0.815– 0.864 |
<0.001 | 0.831 | 0.775– 0.892 |
<0.001 | 0.811 | 0.764– 0.862 |
<0.001 | 0.842 | 0.800– 0.885 |
<0.001 | 0.929 | 0.874– 0.987 |
0.018 |
| eGFR (per 1.0-mL/ min/1.73 m2 increase) |
0.994 | 0.992– 0.997 |
<0.001 | 0.995 | 0.990– 1.000 |
0.056 | 0.995 | 0.990– 1.001 |
0.130 | 0.996 | 0.991– 1.001 |
0.103 | 0.993 | 0.987– 0.998 |
0.010 |
| CRP (per 1.0-mg/dL increase) |
1.019 | 1.010– 1.029 |
<0.001 | 0.980 | 0.943– 1.019 |
0.321 | 1.033 | 1.016– 1.049 |
<0.001 | 1.022 | 1.006– 1.038 |
0.007 | 1.022 | 1.004– 1.040 |
0.016 |
| LVEF <40% | 0.468 | 0.403– 0.543 |
<0.001 | 0.307 | 0.195– 0.483 |
<0.001 | 0.470 | 0.338– 0.654 |
<0.001 | 0.429 | 0.333– 0.553 |
<0.001 | 0.573 | 0.440– 0.745 |
<0.001 |
AReason for intensive care unit admission. CI, confidence interval; eGFR, estimated glomerular filtration rate; HR, hazard ratio. Other abbreviations as in Table 1.
Prognostic differences by aging and reason for ICU admission were evaluated further. The 365-day all-cause mortality rate was significantly higher for sepsis than for other diseases across the 60–69-, 70–79-, and ≥80-year age categories (P<0.001; Figure 4). The survival rate during the 365-day follow-up was lowest in patients aged ≥80 years with etiologies including ACS, AHF, and sepsis. The survival rate was exceptionally low in patients with infectious disease who were aged ≥80 years (Figure 5). Notably, multivariate logistic regression analysis revealed that sepsis was independently associated with 365-day mortality only in the cohort aged ≥80 years (Group D; HR 1.878; 95% CI 1.270–2.777; P=0.002; Table 2). Compared with sepsis and others, the patients with sepsis had significantly aged and a higher LVEF. Laboratory findings indicated that serum blood urea nitrogen, creatinine, C-reactive protein, and B-type natriuretic peptide concentrations were significantly higher in patients with sepsis than in those with other etiologies. The use of additional mechanical support (continuous renal replacement therapy and mechanical ventilator support after endotracheal intubation) was significantly more frequent in patients with sepsis than in those with other conditions (Supplementary Table). APACHE II scores were significantly higher in patients with sepsis than those with other conditions (median 16 [IQR 12–21] vs. 11 [IQR 8–16] points). As a result, the median duration of the ICU stay was higher in patients with sepsis than those with other conditions (median 4 [IQR 2–7] vs. 3 [IQR 2–6] days). Among patients with sepsis, respiratory infections were significantly more common in patients in Group D (n=63; 73.3%) and gastrointestinal infections were more common in Group B (n=9; 14.5%; Table 3). Kaplan-Meier curves for survival among patients with sepsis according to organ of origin revealed significantly lower survival rates for patients with gastrointestinal infections than for those with respiratory infections, urinary tract infections, and others (Supplementary Figure). These results suggest that although sepsis was less common in patients aged ≥80 years (92 patients; 25.3%), these patients were more likely to experience adverse outcomes in non-surgical intensive care.
Kaplan-Meier survival curves according to the reason for admission to the intensive care unit (ICU), namely acute coronary syndrome (ACS), acute heart failure (AHF), sepsis, and others, in each age group. (A) In the cohort <60 years of age, 365-day all-cause mortality did not differ significantly according to the etiology for ICU admission. (B–D) In the cohorts aged 60–69 years (B), 70–79 years (C), and ≥80 years (D), 365-day all-cause mortality was significantly higher in patients with sepsis than in those with ACS or AHF (P<0.001 for all). (C) In the cohort aged 70–79 years, 365-day all-cause mortality was significantly higher in patients with sepsis than in those with ACS or AHF (P<0.001).
Kaplan-Meier survival curves for the 4 age groups according to the reason for admission to the intensive care unit (ICU), namely (A) acute coronary syndrome (ACS), (B) acute heart failure (AHF), (C) sepsis, and (D) others. (A) In the ACS cohort, 365-day all-cause mortality was significantly higher in patients aged ≥80 years than in the other age groups (P<0.001). (B) In the AHF cohort, 365-day all-cause mortality was significantly higher in patients aged ≥80 years than in those aged <60 and 60–69 years (P<0.001). (C) In the sepsis cohort, 365-day all-cause mortality was significantly higher in patients aged ≥80 years than in the other age groups (P<0.001). (D) In the “other” cohort, 365-day all-cause mortality was significantly higher in patients aged ≥80 years than in those aged <60 years (P<0.001).
Associations Between Organ of Origin and Age Patients With Sepsis
| All (n=321) |
Age group | P value | ||||
|---|---|---|---|---|---|---|
| Group A (age <60 years; n=52) |
Group B (age 60–69 years; n=62) |
Group C (age 70–79 years; n=121) |
Group D (age ≥80 years; n=86) |
|||
| Respiratory infections | 200 (62.3) | 27 (51.9) | 40 (64.5) | 70 (57.8) | 63 (73.3) | 0.048 |
| Urinary tract infections | 43 (13.4) | 5 (9.6) | 7 (11.3) | 22 (18.2) | 9 (10.5) | 0.272 |
| Gastrointestinal infections | 21 (6.5) | 1 (1.9) | 9 (14.5) | 7 (5.8) | 4 (4.7) | 0.031 |
| Others | 57 (17.8) | 19 (36.5) | 6 (9.7) | 22 (18.2) | 10 (11.6) | <0.001 |
Unless indicated otherwise, data are given as n (%). P values among 4 groups were determined using the Chi-squared test.
Although the distribution of patient age in ICUs has been occasionally evaluated in Western registries, it has rarely been reported in Japanese ICU cohorts. Japan’s life expectancy has steadily increased to the highest globally, and Japanese elderly patients differ from their Western counterparts in terms of their physical characteristics. Discussing the phenomenon of aging in Japanese ICUs is therefore important, and Japanese ICU data may be meaningful for improving future ICUs in an aging society. In the present study, the number of people requiring intensive care increased as the population aged, and aging was determined to be an independent risk factor for recent non-surgical ICU admission.
Over the past decade, the prevalence of sepsis has increased alongside the aging population, establishing serum C-reactive protein levels as another independent factor associated with aging across all age categories. The prognoses of patients with ACS and AHF may be slightly improved by advances in cardiovascular treatment; however, the results of Kaplan-Meier analysis in the present study indicate that the long-term prognosis of patients with sepsis is deteriorating in those aged ≥80 years. Sepsis may be a complication during treatment for cardiovascular disease. Comprehensive patient care during ICU admission may be necessary to help patients and improve prognosis in aging societies.
Aging in Non-Surgical Patients Requiring Intensive CareIn this study, age was found to be an important factor for recently admitted non-surgical patients. Many ICU registries have been established in Western and Asian countries, and the increased age of patients being admitted to ICUs has been noted.1,4–6,17,18 In Australia, the mean age of patients in ICUs is increasing rapidly, and it is forecast that by 2030 26% of all people admitted to Australian ICUs will be aged ≥80 years.1 The proportion of very elderly ICU patients has already exceeded 10% in many European countries.4 Although data from Japan are rare, the Japanese Intensive Care Patient Database (JIPAD) registry gives some indication of the ages encountered in ICUs. The JIPAD reported that mean age upon admission in Japan was 68 years between April 2015 and March 2017,19 which is older than the mean ages reported in other Asian and Western registries for the same admission period.17,18 There are differences in the factors recorded in these registries, such as whether the diagnosis is included (e.g., cardiac arrest, sepsis), the percentage of patients with cardiovascular disease, the type of ICU management (closed or open), and regional characteristics. The mean age of patients in worldwide registries was reported to be under 60 years of age, but in the present study we found that after 2017 the mean age of admitted intensive care patients was 70 years, which is higher than the age of patients in the JIPAD registry. An explanation for this variation is that the prevalence of cardiovascular disease among patients in the present study (70%) was markedly higher than that in other registries. Aging will be a more serious issue in future cardiovascular care units in Japan.
Etiological Differences According to Age in Non-Surgical Patients Requiring Intensive CareAHF and sepsis are critical diseases in aging societies that are associated with a poor prognosis. HF is a significant cardiovascular disease associated with aging and is becoming a serious issue against the background of population aging in Japan.10,20 In our previous study, we found that the age of AHF patients was significantly higher in the 2010s cohort (median 75 years; IQR 67–81 years) than in the 2000s cohort (median 73 years; IQR 64–80 years).21 In the present study, the mean age of AHF patients was much higher than that of patients with any other disease. Studies using data from 3 large-scale AHF registries in Japan (Acute Decompensated Heart Failure Syndromes (ATTEND) / West Tokyo Heart Failure (WET-HF) / Registry Focused on Very Early Presentation and Treatment in Emergency Department of Acute Heart Failure Syndrome (REALITY-AHF)) have demonstrated that the age of AHF patients has been increasing annually.22–24 Thus, the age of AHF patients has increased concurrently with the aging of the population. This may have contributed to the greater number of older patients with cardiovascular disease in the present study.
Approximately 80% of infectious diseases were diagnosed as sepsis in the present study, which is of prognostic importance in intensive care patients. Although the incidence of sepsis in patients aged >80 years is reportedly similar to that in patients in other age groups, mortality is much higher in the former group.25–27 The hospital mortality for patients aged >80 years is almost twice that of patients age ≤50 years.11 Other studies have reported similar findings, with ICU mortality rates of 45.6% in patients aged <60 years, 60.7% in those aged 60–80 years, and 78.9% in those aged >80 years for patients with sepsis in ICU.27 Furthermore, among ICU patients in Australia/New Zealand, the mortality rate among severe sepsis patients was 30.4% in patients aged ≥85 years, compared with 7.3% in those aged ≤44 years.25
In the present study, 365-day survival rates after intensive care steadily worsened with increasing age. Various factors, including systolic blood pressure, hemoglobin, C-reactive protein, and low LVEF, were identified as independent predictors of mortality in the elderly (age ≥80 years) cohort. The presence of sepsis was significantly associated with 365-day mortality only in the ≥80-year age cohort. These findings suggest that the presence of sepsis contributes to adverse outcomes in elderly Japanese patients requiring intensive care. Based on our findings, a prompt and well-structured support strategy may be required for elderly patients with sepsis based on their age. From another perspective, increased non-cardiovascular mortality was recently suggested in patients with cardiovascular disease requiring intensive care. The Tokyo CCU network multicenter registry proposed that the incidence of in-hospital non-cardiac death was significant in patients with acute myocardial infarction, accounting for 15.6% of all in-hospital deaths.28 This means that patients with ACS did not die from direct cardiac causes but rather from non-cardiovascular causes (e.g., sepsis/infectious disease) after a prolonged hospital stay. Even if sepsis is not the main cause of death, it may adversely affect outcomes for patients with various diseases in ICUs.
Some new treatment strategies have been suggested in the cardiovascular field over the past decade. New medications, such as sacubitril/valsartan and sodium-glucose cotransporter 2 inhibitor treatment have been covered by the Japanese National Health Insurance since 2020, with accumulating evidence from their use in daily clinical practice.29 The use of a percutaneous left ventricular assist device (Impella) for total ventricular unloading in cardiogenic shock patients has been established in the past decade.30,31 The technology and equipment for percutaneous coronary intervention in ACS patients are under constant development. Furthermore, direct oral anticoagulants have been introduced for the treatment of pulmonary embolism, and thoracic endovascular aortic repair is considered the go-to option for acute aortic disease treatment. Targeted temperature management protocols have also changed over the past decade. Although it was reported in 2024 that sepsis-related mortality had decreased between 1985 and 2019,32 advances in treatment strategies for sepsis may be less compared with the prognostic advantages of such developments in the case of cardiovascular disease. Although some strategies are available to maintain vital signs in septic shock (e.g., norepinephrine, vasopressin, and hydrocortisone), the fundamental treatment for sepsis has remained unchanged for a long time. We have no choice but to trust antibiotics and steroid therapy. There is a need for a breakthrough treatment that profoundly affects the management of sepsis and transforms the way it is approached.
Study LimitationsThe present study has some limitations. First, it was a retrospective study performed at a single hospital. Unmeasured variables or missing data may have affected the results. In particular, we have no data regarding cognitive function. This may be an important factor when discussing aging in ICU. The definitions of diagnosis, risk, and interpretation of risk factors may have changed between the 2010s and 2020s. Furthermore, the difficulty in standardizing each patient’s care may have influenced the major findings of this study. Second, we have no detailed data after discharge from the ICU. Ideally, these data should be included when adjusting factors for long-term prognosis. Although we have discussed the factors associated with patient prognosis in the Discussion section, the analysis lacks statistical rigor. Third, the treatment strategy was left to the discretion of individual physicians. The lack of invasive treatments for aged patients may have affected prognosis. In fact, continuous renal replacement therapy and mechanical ventilator support after endotracheal intubation were used significantly less often in patients aged ≥80 years. Fourth, ICU beds were sometimes unavailable due to the COVID-19 pandemic in 2020–2021. Patients admitted during this time period are important when considering the aging society,33 so were included in the present study. However, the unusual situation during the pandemic may have affected the present results. Finally, the diagnostic criteria for sepsis changed significantly in 2016. In the present study, sepsis was diagnosed using the Surviving Sepsis Campaign: International Guidelines for Management of Sepsis and Septic Shock 2021.14 Under these updated criteria, some patients with sepsis may not have been diagnosed with sepsis at the time of ICU admission during the period 2012–2015.
Japan has experienced a steady increase in life expectancy, ranking among the highest globally. The mean age of non-surgical patients in Japan who require intensive care has increased over the past decade. Low systolic blood pressure, anemia, inflammation, and low LVEF in non-surgical ICU patients were associated with worse long-term prognoses across all age categories. Notably, BMI, eGFR, and the presence of sepsis were associated with 365-day mortality, especially in the cohort aged ≥80 years. These findings emphasize the importance of the management of sepsis during ICU stays for the aging population in Japan.
The authors thank Uni-edit (https://uni-edit.net/) for editing and proofreading a draft of this manuscript.
This research did not receive any grants from any funding agency in the public, commercial, or not-for-profit sectors.
The authors declare no conflicts of interest in association with the present study.
This study was approved by the Research Ethics Committee of Nippon Medical School Chiba Hokusoh Hospital (Reference no.: 841).
Please find supplementary file(s);
https://doi.org/10.1253/circj.CJ-24-1030
