Prognostic factors for 30-days mortality in eighty years aged and older critically ill patients: a single center retrospective cohort study

Background/aim Nowadays, with the rise in average life expectancy, the rate of hospitalization of the older population in intensive care unit (ICU) is gradually increasing. Unfortunately, there are no ideal combination of prognostic factors predicting the mortality in older patients admitted to the ICU. In the present study, we aim to determine the prognostic factors and their impacts on short-time mortality in older critically ill patients. Materials and methods This retrospective cohort study was performed between January 2019 and February 2020. We included 133 patients aged ≥80 years and hospitalized ≥24 h in the ICU. Results A total of 133 critically ill patients enrolled in the present study. And, the median age of the patients was 85 (80–106) years. 30-days and overall ICU mortality rates were found 30.1% and 34.6%, respectively. The patients were grouped as survivors (n = 94) and nonsurvivors (n = 39). Hospital length of stay before the ICU admission was found significantly longer in nonsurvivors (p = 0.001). Sequential organ failure assessment (SOFA) score and acute physiology and chronic health evaluation-II (APACHE-II) score were significantly higher in nonsurvivors (p < 0.001, p < 0.001). Also, blood lactate level and glucose level were respectively significantly higher in nonsurvivors (p < 0.001, p = 0.006). We found that modified nutrition risk in critically ill (mNUTRIC) score and prehospital clinical frailty scale (CFS) were independent prognostic factors for the older critically ill patients (HR = 9.19, 95% CI=1.47–57.32, p = 0.018, HR = 20.16, 95% CI = 2.63–54.07, p =0.004). Conclusion mNUTRIC score and prehospital CFS score were the most important prognostic factors in the admission of older patients to intensive care units.


Study design
This single-center retrospective cohort study was performed in a 12-beds adult tertiary ICU of Malatya Training and Research Hospital between January 2019 and February 2020.A total of 717 patients were admitted to the ICU during the study period.A total of 222 patients aged 80 years and older in the ICU were assessed for eligibility.Of these, 89 were excluded from the study due to various reasons.And a total of 133 patients were enrolled in the study (Figure 1).

Data collection and definitions
The following data were recorded and analyzed: all patients' demographic and clinical data, laboratory findings, types of respiratory support, the reason for admission, admission source, treatment options, hospital length of stay (LOS) before ICU admission, ICU length of stay (LOS), outcomes, scores on the acute physiology and chronic health evaluation-II (APACHE-II) and sequential organ failure assessment (SOFA), modified nutrition risk in critically ill (mNUTRIC) and clinical frailty scale (CFS).Patients' data and scores reported in this study were collected within the first 24 h following the ICU admission.
The nutritional status of the patients was assessed according to the mNUTRIC score within the first 24 h of the ICU admission.mNUTRIC score includes the following variables: age, number of comorbidities, days from hospital to ICU admission, APACHE-II score, and SOFA score.And the patients with mNUTRIC scores ≥5 were defined as high nutritional risk [9].
The prehospital frailty status of the patients was assessed according to the CFS [10].And the patients who had CFS ≥5 were defined as frail.CFS was calculated based on patients' medical records and the interview of the patients and/or their relatives.
APACHE-II and SOFA scores were used for the assessment of the severity of illness.These scores were calculated based on the patients' worst clinical and laboratory findings observed during the first 24 h following the admission of ICU [11,12].And also, patients' Glasgow coma scale (GCS) was calculated at the time to the admission of ICU.

Measurement of outcome
All of the patients were followed during their ICU stay or until death.All-cause of short time mortality was defined as death within 30 days after the ICU admission.Patients' mortality data were collected from the hospital medical record system.

Statistical analysis
SPSS for Windows 22.0 software (SPSS Inc., Chicago, IL, USA) was used for the statistical analysis of the data obtained from the hospital database.All results were analyzed with a confidence interval level of 95% and a significance level of p < 0.05.The homogeneity and distribution of the variables were assessed using the Skewness-Kurtosis.Frequencies and percentages were used for the categorical data, continuous variables were presented as mean +/-standard deviation or median (min-max) according to the distribution normality of data.We compared the variables between survivors and nonsurvivors.Categorical variables were compared using the chi-squared test.The independent samples t-test was used for the analysis of the two independent groups parametric data while Mann-Whitney U test was used for the analysis of nonparametric data.We used the receiver operating characteristic (ROC) curve to estimate the appropriate cut-off values of SOFA, APACHE-II, CFS, and mNUTRIC score for determining the 30-days mortality.The Kaplan-Meier method was used to determine the overall survival rates of the patients at day 30 and the longrank test was used to compare the differences in survival between the groups.After the univariate survival analysis, Cox regression analysis was used for the assessment of the multivariate survival analysis.

Characteristics of the patients
A total of 133 critically ill patients were included in the study.Seventy-five (56.4%) of patients were female and the median age of the patients was 85 (80-106) years.Hypertension (68.4%), coronary artery disease (41.3%), dementia (35.3%), and chronic obstructive pulmonary disease (24.4%) were the most common comorbidities.The emergency department (55.6%) was the most frequent patient's source of admission to the ICU.Sepsis (32.3%) was the most frequent cause of the admission to the ICU followed by acute cardiogenic pulmonary edema (19.5%) and postoperative respiratory failure (12.8%).Hospital LOS before the ICU admission was found 1.95 ± 4.40 days and ICU LOS was found 11.40 ± 12.76 days.
The patients were grouped as survivors (n = 94) and nonsurvivors (n = 39).We found that there were no statistical differences between the survivors' and the nonsurvivors' group with respectively age, sex and diagnosis at ICU admission.However, hospital LOS before the ICU admission was found significantly longer in nonsurvivors (p = 0.001).And also, ICU LOS was found significantly longer in survivors' group (p = 0.015).We found that SOFA, APACHE-II, CFS and mNUTRIC score were significantly higher in nonsurvivors (p < 0.001).In addition, blood lactate level and glucose level were respectively significantly higher in nonsurvivors (p < 0.001, p = 0.006).Comparison of the baseline clinical characteristics and laboratory findings of the two groups are presented in Tables 1 and 2.
We found that a total of 43 (32.4%)patients underwent invasive mechanical ventilation and 33 (24.8%) patients underwent noninvasive mechanical ventilation.Nonsurvivors had more invasive mechanical ventilation requirements (p < 0.001).Also, the reintubation rate was found higher in nonsurvivors' group (5.3% vs. 10.2%), but it was not statistically significant (p = 0.3).The use of vasoactive agents was found significantly higher in nonsurvivors (p = 0.004).Treatment options of the patients were presented in Table 3.

Short time survival analysis of the patients
We performed ROC curve analysis for finding the optimal cut-off value for determining the 30-days mortality in the older critically ill patients (Figure 2).The cut-off values of SOFA, APACHE-II, CFS, and mNUTRIC scores were presented in Table 4.
In the present study, we found that 30-days and overall ICU mortality rates were found at 30.1% and 34.6%, respectively.Prognostic factors affecting the 30days mortality in the older patients are presented in Table 5.The effects of the clinical characteristics and laboratory data on the 30-days survival showed that patients with hyperglycemia and hyperlactatemia during the ICU admission had significantly shorter survival times respectively (p = 0.001, p = 0.023).Also, we found significantly shorter survival time in patients with CFS ≥5 (Figure 3A), mNUTRIC score ≥5 (Figure 3B), SOFA score >5 (Figure 3C), and APACHE-II scores ≥23 (Figure 3D) (p < 0.001).

Discussion
In the present study, we evaluated the clinical characteristics and the outcomes of the older critically ill patients.The main finding of our study showed that mNUTRIC score and prehospital CFS score were the most important independent prognostic factors in the admission of the older patients to intensive care units.
Nowadays, the rate of hospitalization of the older population in ICU is gradually increasing with the increase in the average life expectancy [13].Nielson et al. showed in their study in 2014 that 12.6% of the patients who were admitted to the ICU consisted of patients aged 80 and over, and that patients aged 80 and over who were admitted to intensive care increased by 18% over the years [14].In our study, 21.1% of the patients who were admitted to the ICU during 13-months aged 80 years and over.The high percentage of ≥80-year-old patients in the present study can be explained by the characteristic of the population in the city.
Problems are experienced in the follow-up of the older patients in ICU due to the consideration that life expectancy  will be short as well as due to the underlying comorbidities [13,15].It has been demonstrated in numerous clinical studies that mortality increases in parallel with the age of patients who are admitted to the ICU, and age is an independent risk factor for mortality [3,16].However, recent studies have shown that chronological age is not an independent risk factor for mortality.Besides, it has been underscored that rather than chronological age, biological age is more important for survival [6, 15,17].
Many studies evaluating the survival in critically ill older patients are single-centered retrospective cohort studies and include different age and disease groups.Therefore, 30-days and ICU mortality rates vary in studies.In previous studies, the mortality rate soars to 50% in critically ill older patients in ICU [1,4,6,[15][16][17].In the present study, we found that 30-days and overall ICU mortality rates were found 30.1% and 34.6%, respectively.
In the present study, the general ward was the primary admission source for the older patients admitted to the ICU.Also, we found that hospital LOS before the ICU admission was 1.95 ± 4.40 days and ICU LOS was 11.40 ± 12.76 days.Clinical studies that evaluate the association between hospital LOS before the ICU admission and survival of the patients transferred to ICU showed that patients with longer hospital LOS before the ICU admission had worse outcomes and survival.Thus, to improve outcomes in critically ill older patients who have been admitted to ICU, it may be beneficial to establish rapid response teams to rapidly recognize the clinical deterioration of patients who are followed up in general wards, and to make early interventions [18,19].Sepsis, defined as life-threatening organ dysfunction resulting from the dysregulated response of the host to infection, is one of the most important causes of admission to ICU and mortality.Advanced age is a significant risk factor for the development of sepsis, and a dramatic increase is seen in the incidence of sepsis, particularly in those aged 80 and over [6,15].In the present study, we found that sepsis (32.3%) was the most frequent cause of admission to the ICU followed by acute cardiogenic pulmonary edema (19.5%) and postoperative respiratory failure (12.8%).
There are no prognostic factors and combinations that can be used during the admission of the older critically ill patients to intensive care units, that are accepted by all clinicians, and whose validity and reliability have been proven in the literature.The fact that the clinical characteristics, which are used in conventional disease severity scorings, such as the APACHE-II score, may be inadequate to accurately predict the survival of critically ill patients, is increasingly gaining acceptance [6,8,20].We found out that SOFA and APACHE-II scores were not an independent and significant prognostic factor for 30-days survival, albeit we determined that patients with higher SOFA and APACHE-II scores in our study had higher mortality and shorter survival rates.Also, we found that patients with higher serum levels of glucose and lactate had a shorter survival time.However, these two variables were not independent and significant prognostic factors for the 30-days survival.
Traditional disease severity scoring systems may not be able to detect significant data about a patient's preillness state, such as cognitive impairment, decreased functional capacity, and frailty.However, these data may be important markers of mortality and morbidity, especially in the older patients.Assessing the clinical frailty of patients before admission to the ICU, especially in critically ill patients, could facilitate better clinical decisions [4,6,8].
The concept of clinical frailty is defined as a multivariate syndrome characterized by loss of physical, physiological, and cognitive reserves rather than an acute disease state that increases with age but is not specific to advanced age.Frail elders experience difficulty in adapting to various stressful situations such as acute illness and trauma.Frailty can be assessed using the CFS developed by Rockwood et al.CFS is a 9-point scale and a patient with a score ≥ of 5 is defined as frail [4,8,10,21,22].The prevalence of prehospital frailty varies in previous studies and it increases with age.More importantly, it has been suggested that an increased CFS score (CFS score of ≥5) is a significant factor in short-term and long-term mortality in the older patients who have been admitted to ICU [3,4,8,21,22].In the present study, a total of 82 (61.6%) critically ill patients had a CFS score ≥5.Also, we found that the CFS score was significantly higher in nonsurvivors (p < 0.001) and independent prognostic factors of 30-days mortality in the older patients admitted to the ICU included a high CFS score (CFS score ≥5).
The majority of the patients who apply to the ICUs and particularly those who receive mechanical ventilation support are at nutritional risk.In particular, this risk increases even more in fragile and critically ill older patients.Hence, the nutritional risk should be assessed in all patients who have been admitted to the ICU without wasting time [1,6,21].Based on the mNUTRIC score developed by Heyland et al., patients are divided into low (0-4) and high (5-9) risk groups to evaluate the nutritional risk in critical patients, and early nutritional support is recommended for patients in the high-risk group [9,23].Moreover, various studies have revealed that a high     mNUTRIC score (mNUTRIC score ≥5) is also associated with increased mortality and unfavorable outcomes.For this reason, critically ill patients with high nutritional risk should be identified without delay and early nutritional support should be initiated [24,25].In the present study, a total of 89 (66.9%) critically ill patients had mNUTRIC score ≥5.Also, we found that patients with mNUTRIC score ≥5 had a shorter survival time.
In conclusion, prehospital CFS and mNUTRIC score were the independent and significant prognostic factors for the older critically ill patients.In addition to traditional scoring systems that assess organ failure during admission of the older patients to the ICU, assessment of prehospital frailty and nutritional risk status could be more effective in predicting short and long-term mortality.Therefore, we recommend that prehospital frailty and the nutritional risk assessment of the older patient should be routinely performed for the more rational use of intensive care unit beds and sufficient prognostic evaluation in these patient groups.

Figure 1 .
Figure 1.Flowchart of the study.

Figure 2 .
Figure 2. Receiver operating characteristic curves of different scores predicting short time mortality.

Figure 3A .
Figure 3A.Kaplan-Meier curves of 30-days survival analysis shows the impact of clinical frailty scale in the older critically ill patients.

Figure 3B .
Figure 3B.Kaplan-Meier curves of 30-days survival analysis shows the impact of modified nutrition risk in critically ill score in the older critically ill patients.

Figure 3C .
Figure 3C.Kaplan-Meier curves of 30-days survival analysis shows the impact of sequential organ failure assessment score in the older critically ill patients.

Figure 3D .
Figure 3D.Kaplan-Meier curves of 30-days survival analysis shows the impact of acute physiology and chronic health evaluation-II score in the older critically ill patients.

Table 1 .
Baseline characteristics of the patients.

Table 2 .
Baseline laboratory findings of the patients.

Table 3 .
Treatment options of the patients in the ICU.

Table 4 .
The optimal cut-off value of clinical frailty scale, mNUTRIC score, SOFA score, APACHE-II scores for the prediction of mortality.
APACHE-II: acute physiology and chronic health evaluation II; SOFA: sequential organ failure assessment, CFS: clinical frailty scale, mNUTRIC: modified the nutrition risk in critically ill, AUC: area under curve, CI: confidence interval.

Table 5 .
Comparison of mean survival time according to demographic characteristics and ICU parameters.
CI: confidence interval, SE: standard error; SOFA: sequential organ failure assessment, APACHE-II: acute physiology assessment and chronic health evaluation II, mNUTRIC: modified the nutrition risk in critically ill, CFS: clinical frailty scale, COT: conventional oxygen therapy; NIMV: noninvasive mechanical ventilation; IMV: invasive mechanical ventilation.