Characteristics, management and outcomes of critically ill patients who are 80 years and older: a retrospective comparative cohort study

Background Older age is associated with chronic illnesses and disability, which contribute to increased admission to the intensive care unit (ICU). Our primary objective was to compare the characteristics, ICU management and outcomes of critically ill patients ≥ 80 year-old with those of younger patients. Methods This was a retrospective cohort study of patients admitted to a tertiary-care ICU from 1999 to 2011. The characteristics, ICU management and outcomes of patients ≥ 80 year-old were compared with those 50–64.9 and 65–79.9 year-old. Multivariate analysis was performed to determine the adjusted risk of Do-Not-Resuscitate orders and hospital mortality in patients ≥ 80 year-old compared with the younger groups. Results During the study period, patients aged ≥ 80 years (N = 748) represented 7.9% of all ICU admissions and 12.8% of patients aged ≥ 50 years. Chronic cardiac (32.2%) and respiratory (21.8%) diseases were more prevalent in them than the younger groups (p < 0.0001). The most common reasons for their ICU admission were cardiovascular (30.9%) and respiratory (40.4%) conditions. Sepsis was commonly present in them on admission (32.9%). Mechanical ventilation and renal replacement therapy were commonly provided (76.9% and 16.0%, respectively). During ICU stay, Do-Not-Resuscitate orders were more frequently written for patients aged ≥ 80 years (35.0%) compared with 21.9% for 50–64.9 year-old group, p < 0.0001, and 25.4% for the 60–79.9 year-old group, p < 0.0001. On multivariate analysis, patients aged ≥ 80 years were more likely to receive these order compared with the 50–64.9 year-old patients (adjusted OR, 1.83; 95% CI, 1.45-2.31) and the 65–80 year-old patients (adjusted OR, 1.64; 95% CI, 1.32-2.04). The hospital mortality increased gradually with age and was highest (54.6%) in patients ≥ 80 year-old (p < 0.0001). Patients ≥ 80 year-old had higher risk of hospital mortality compared with patients aged 50–64.9 years (adjusted OR, 2.16; 95% CI, 1.73-2.69) and with those aged 65–79.9 years (adjusted OR, 1.51; 95% CI, 1.23-1.86). Conclusions Patients ≥ 80 year-old represented a significant proportion of ICU admissions. Although they received life sustaining measures similar to younger groups, they had higher adjusted mortality risk compared with the younger groups.


Background
Old age is associated with chronic medical diseases and functional impairment, which may lead to increased incidence and severity of acute critical illnesses and to admission to the intensive care unit (ICU). For severe sepsis, the incidence generally increases with age during adulthood [1]. Incidence of hospitalization secondary to community-acquired pneumonia doubles in patients aged > 60 years [2]. A study from the United Kingdom found that the ratio of ICU admissions to local population in people ≥ 60 years rose linearly by 2.62 admissions per 10,000 population per year over a six consecutive year-period with the increase being highest in patients ≥ 80 years [3].
Rationing health care based upon age has been reported. In a systematic review, Sinuf et al. studied rationing of ICU resources and found that age and severity of illness were most strongly associated with a refusal to admit to the ICU [4]. A recent prospective cohort study of patients > 85 year-old presenting to the emergency departments of 15 Parisian hospitals found significant variability in ICU admission even after adjustment for patient characteristics [4]. This was likely related to the belief that older age was associated with poor outcomes after ICU admission [5]. This issue has been investigated in multiple studies mostly from Western countries [6][7][8]. However, it is also believed that age explains only a small part of the outcomes of critical illness and that prior functional status, co-morbidities and the level of therapeutic support are important factors [9].
Knowing the outcomes and prognosis determinants of patients aged ≥ 80 years who are admitted to the ICU is important for clarification of perceptions of intensive care providers and possibly for proper allocation of resources. The objectives of this study were to describe the characteristics, management and outcomes of critically ill patients ≥ 80 year-old and to determine if age ≥ 80 years was an independent predictor of ICU management and of hospital mortality in a tertiary-care center in Saudi Arabia.

Patients and setting
This was a retrospective analysis of a cohort of adult patients admitted between January 1, 1999 and December 31, 2011 to the ICU of King Abdulaziz Medical City, a 900-bed tertiary-care teaching hospital in Riyadh, Saudi Arabia that had been accredited by the Joint Commission International. The ICU was a 21-bed medical-surgical closed unit and was staffed by board-certified critical care physicians on a 24 hours per days, 7 days a week basis [10] with residents from different specialties rotating periodically. In this study, we compared patients ≥ 80 year-old with those 50-64.9 and 65-79. 9 year-old because the younger patients (< 50 year-old) were thought to have completely different comorbid conditions, reasons for ICU admission and end-of-life care. For patients who had multiple ICU admissions in the same hospitalization, we included only the first ICU admission. The Institutional Review Board of King Saud bin Abdulaziz University for Health Sciences, Riyadh, Saudi Arabia approved this study, granted waiver of consent and allowed the review of the patients' medical records.

Collected data
Our Intensive Care Department had a comprehensive database in which trained coordinators prospectively collected demographic and clinical data and followed patients for predefined outcomes. The following data were extracted from the database: age, gender, body mass index, functional status before hospitalization based on the modified Rankin Scale [11], Acute Physiology and Chronic Health Evaluation (APACHE) II score [12], the presence of chronic health illnesses as defined by APACHE II system, the main reason for ICU admission (as per APACHE II definitions), presence of sepsis on admission, diagnosis of the following on ICU admission: myocardial ischemia such as acute coronary syndrome and acute myocardial infarction, community-or hospital-acquired pneumonia and new stroke, admission Glasgow Coma Scale score and admission platelet count, creatinine, lactate, bilirubin and International Normalized Ratio (INR).
We also noted the following ICU management aspects: use of vasopressors within the first 24 hours after ICU admission, requirement for mechanical ventilation, renal replacement therapy and tracheostomy during ICU stay, performance of cardiopulmonary resuscitation for cardiac arrest while in the ICU and the practice of Do-Not-Resuscitate orders. During the study period, advanced directives were not practiced in Saudi Arabia. However, our hospital had a policy in which a Do-Not-Resuscitate order can be written if three qualified physicians agreed that a patient would not benefit from cardiopulmonary resuscitation in case of cardiac arrest because of he/she was terminally ill or had a severe illness with high predicted mortality, especially if significant functional disability and/or dementia preexisted. This medical decision would be then explained to the patient or surrogate decision maker who generally had to agree before its implementation. His or her signature was not required. Do-Not-Resuscitate orders generally precluded intubation but not the use of noninvasive ventilation.
In this study, hospital mortality was the primary outcome. Other outcomes included ICU and post-ICU discharge mortality, length of stay in the ICU and hospital and duration of mechanical ventilation. We also calculated the predicted mortality based on the Mortality Probability Model (MPM) II at 0 and 24 hours [13] and on APACHE II score for the three age groups.

Statistical analysis
Statistical analysis was done using Statistical Analysis System (SAS, version 9.0; SAS Institute, Cary, NC). Continuous variables were reported as means with standard deviations and if clinically relevant as medians with the first and third quartiles. Categorical variables were presented as frequencies with percentages. The standardized mortality ratio for APACHE II was calculated for each group by dividing the actual by predicted hospital mortality and reported with its 95% confidence intervals (CI) [14]. The Chi-squared test was used to assess the differences between categorical variables and the Student's t-test to assess the differences between continuous variables. Multivariate analysis was performed to evaluate older patients' adjusted risk of hospital mortality, having a Do-Not-Resuscitate order during ICU stay and need for renal replacement therapy. The association between hospital mortality and age was also evaluated in patients admitted to the ICU because of cardiovascular, respiratory, neurological, other medical, trauma-related and postoperative reasons and in patients with the diagnosis of myocardial ischemia, community-or hospital-acquired pneumonia and new stroke on ICU admission. Multivariate analysis was also performed to determine the predictors of hospital mortality in patients ≥ 80 year-old. The independent variables entered in these analyses were gender, nonage-related APACHE II score, the main reason for admission, chronic illnesses, functional status before hospitalization (moderately severe or severe disability versus more active status), mechanical ventilation, length of ICU stay, creatinine and INR. The results were presented as adjusted odds ratio (aOR) with 95% CIs. A P-value < 0.05 was considered statistically significant in all analyses.

Characteristics of patients
During the 13 year study period, 9493 patients were admitted to the ICU. The patients aged ≥ 80 years (N = 748) represented 7.9% of them and 12.8% of the 5832 patients who were 50 years and older. Their mean age was 85.1 ± 4.9 years (Q1-Q3: 81-88 years). Figure 1A describes the age distribution of the cohort according to gender. Men aged ≥ 80 years accounted for 8.1% of admitted men and women aged ≥ 80 years 7.6% of admitted women. Figure 1B describes the percentages of patients aged ≥ 80 years admitted from 1999 to 2011 and shows random variation between years. Additionally, Table 1 describes the characteristic of the three age groups. The oldest (≥80 year-old) patients were predominantly males (64.2%) with APACHE II score = 27.2 ± 8.2. Chronic cardiac (32.2%) and respiratory (21.8%) diseases were more prevalent in them than in the younger groups (p < 0.0001). More patients aged ≥ 80 years had moderate or severe physical disability before hospitalization as assessed by the modified Rankin Scale. Disorders of the respiratory (40.4%) and cardiovascular (30.9%) systems were their most common reasons for admission to the ICU. Sepsis was present on admission in 32.9%, the highest among the groups (p < 0.0001). Myocardial ischemia, community-acquired pneumonia and new stroke on admission were also more prevalent.
Intensive care unit management Table 2 describes certain aspects of ICU management for the three age groups. Use of vasopressors was similar in patients aged ≥ 80 years compared with the two younger groups. Patients aged ≥ 80 years had the highest rate of mechanical ventilation (76.9%) and the lowest rate of renal replacement therapy (16.0%). Multivariate analysis showed that the OR for providing renal replacement therapy was not different in patients aged > 80 years compared to 50-65 year-old patients (aOR, 0.824; 95% CI, 0.613-1.107) and 60-79. 9 year old patients (aOR, 0.87; 95% CI, 0.65-1.15). Tracheostomy was performed more frequently in patients aged ≥ 80 years even though there was no difference in the duration of mechanical ventilation among the three groups. Figure 2 describes the percentages of patients in the different age groups who had Do-Not-Resuscitate orders during their ICU stay. Do-Not-Resuscitate orders were more frequently written for patients aged ≥ 80 years (35.0%) compared with 21.9% for 50-64. 9 year-old group, p < 0.0001, and 25.4% for the 60-79. 9 year-old group, p < 0.0001 (Table 2). In addition, Do-Not-Resuscitate orders were written earlier (median day after ICU admission, 3; Q1-Q3: 1-8 days) than for patients < 80 year-old (median, 5; Q1-Q3: 1-10 days for the 50-64. 9 year-old and median, 5; Q1-Q3, 1-11 days for the 65-79.9 yearold). After controlling for other factors and compared with patients 50-64.9 years, the practice of Do-Not-Resuscitate orders was similar in patients who were 65-80 year-old (aOR, 1.16; 95% CI, 0.98-1.37) but higher for patients aged ≥ 80 years (aOR, 1.83; 95% CI, 1.45-2.31). Additionally, patients aged ≥ 80 years were more likely to have Do-Not-Resuscitate order compared with patients who were 65-80 year-old (aOR, 1.64; 95% CI, 1.32-2.04). Table 3 describes the outcomes of the patient groups. The lengths of stay in the ICU and hospital were similar in the three groups. Although there was no difference in ICU mortality between the three age groups, hospital mortality increased gradually with age and was highest (54.6%) in patients ≥ 80 years old (p < 0.0001). Moreover, the hospital mortality rates for patients who had Do-Not-Resuscitate order while in the ICU were 92.6% for the younger group, 89.7% for patients aged 65-79.9 years and 85.1% for the very elderly. The mortality rates of patients who had cardiopulmonary resuscitation for cardiac arrest while in the ICU were 82.7% for the younger group‚ 79.5% for patients aged 65-79.9 years and 80.6% for the very elderly. Figure 3 describes the hospital mortality of the three age groups stratified by the reason of admission and generally shows that the mortality rates increased with age. This was not observed in patients admitted to the ICU because of hospital acquired pneumonia and neurologic disease. After adjusting for imbalances in the baseline characteristics, patients ≥ 80 year-old had higher risk of hospital mortality compared to patients aged 50-64.9 years (aOR, 2.16; 95% CI, 1.73-2.69) and those aged 65-79.9 (aOR, 1.51; 95% CI, 1.23-1.86). This increased risk was not observed in patients with ICU admission because of neurologic disorders, new stroke or hospital-acquired pneumonia (Figure 3).

Discussion
The main findings of this study were the following: patients aged ≥ 80 years represented a significant portion of patients admitted to the ICU; had significant comorbid conditions; were admitted mainly because of acute dysfunction of the cardiac and respiratory systems; and compared with the younger groups, received similar life sustaining treatments but were more likely to have Do-Not-Resuscitate orders and to die in the hospital than the younger groups.
The age structure of the world population has shifted and will continue to do so with the proportion of old people increasing in both developed and developing countries [15]. In parallel with this shift, more old patients are admitted to ICU. In Australia and New Zealand, the proportion of patients aged > 80 years was 13% of the  the Saudi population is much younger than those of developed countries. In 2011, people aged ≥ 80 years represented 0.6% of the population compared with 3.8% in Australia, 4.2% in Denmark and 3.8% in the United States [18]. Physicians frequently consider old age when deciding on the provision of life-sustaining measures. An observational simulation study found that 86, 78 and 62% of participating physicians (predominantly males without religious beliefs; median ICU experience = 9 years) felt that noninvasive mechanical ventilation, invasive mechanical ventilation or renal replacement therapy was warranted, respectively for patients aged ≥ 80 years [19]. On multivariate analysis, age < 85 years, self-sufficiency and bed availability were associated with ICU admission [19]. In a systematic review of 10 observational studies of seriously ill patients considered for ICU admission during periods of reduced bed availability, Sinuff et al. found that age and severity of illness were most strongly associated with a refusal to admit to the ICU [4]. A recent prospective study of patients > 85 years presenting to the emergency departments of 15 Parisian hospitals found significant variability in ICU admission even after adjustment for patients' characteristics [5]. The geographic variation in ICU use for patients ≥ 85 years old was also seen in another study where it was less common in England (1.3%) than the United States (11.0%) [20]. In the current study, we have observed that patients aged ≥ 80 years were frequently provided with life support measures, such as mechanical ventilation and renal replacement therapy, like younger patients. However, age ≥ 80 years was found to be an independent risk factor for the practice of Do-Not-Resuscitate orders after controlling for co-morbid conditions.
Low physiological reserve and comorbidities often place very old people in a situation of greater complexity, which may impact outcome. However, it is thought that they have lived to that age because they are resilient to acute illnesses. Studies generally show higher critical illnessassociated mortality in the old and very old patients. A ICU, intensive care unit. *P-value between patients aged 50-64.9 and those 65-79.9 years. ** P-value between patients aged 50-64.9 and those ≥ 80 years. ***P-value between patients aged 65-79.9 and those ≥ 80 years.     [7]. A secondary analysis of data from a randomized trial comparing the effects of dopamine and norepinephrine in patients with shock found that the mortality rates were higher in the old (75-84 years) and very old (≥ 85 years) patients at 28 days, at hospital discharge, and after 6 and 12 months [21]. Most very old patients were dead at 6 (92%) and 12 months (97%) with mortality rates increasing with age in all types of shock [21].  [24]. In the current study, the ICU mortality was similar in the three age groups, but the hospital mortality was significantly higher in patients aged ≥ 80 years having a higher adjusted mortality risk compared to younger age groups (50-64.9 and 65-79.9 years). This was observed in different admission types, except for admissions due to neurologic disease and hospital-acquired pneumonia. This could be because patients aged ≥ 80 years had high prevalence of chronic illnesses and functional disability in our study.

% who had DNR order during ICU stay
The findings of this study should be interpreted in the light of its strengths and limitations. Strengths include the large sample size. Limitations include being a monocenter study and lack of data on post-ICU care processes, which may have affected hospital outcome, and on longterm outcomes such as cognitive function and disability. Critical illness in old people has been associated with decline in cognitive function. Analysis of data from a prospective cohort study of 2929 individuals ≥ 65 year-old without dementia showed that adjusted hazard ratio for incident dementia was 1.4 following a noncritical illness hospitalization (95% CI, 1.1-1.7; p = 0.001) and 2.3 following a critical illness hospitalization (95% CI, 0.9-5.7; p = 0.09) [25]. This may be one of the reasons for increased mortality after ICU discharge.

Conclusions
In conclusion, we have found that patients aged ≥ 80 years accounted for a significant proportion of patients admitted to a tertiary-care ICU in Saudi Arabia. However, their proportion was lower than those of developed countries likely due to the younger Saudi population. The treating intensivists supported them with life sustaining interventions, such as mechanical ventilation and renal replacement therapy, similar to younger groups, but the Do-Not-Resuscitate practice was more common in them. More than half of them died in the hospital with age ≥ 80 years being an independent risk factor for hospital mortality.

Key messages
Patients aged ≥ 80 years represented a significant portion of ICU admissions (7.9% of all admissions and 12.8% of patients aged ≥ 50 years). They received life support measures such as mechanical ventilation and renal replacement therapy similar to younger patients. However, they had higher frequency of Do-Not-Resuscitate orders. Patients aged > 80 years had higher hospital mortality than the younger patients mostly after ICU discharge. Age ≥ 80 years was associated with almost 2 times increase in the adjusted hospital mortality risk compared with age 50-64.9 years. important intellectual content, approved the final version to be published. MRS: Data collection, interpretation of data, critical revision of the manuscript for important intellectual content, approved the final version to be published. YMA: Conception and design, analysis and interpretation of data, drafting of the manuscript, critical revision of the manuscript for important intellectual content, approved the final version to be published. All authors read and approved the final manuscript.