† These authors contributed equally.
Academic Editor: Manuel Martínez-Sellés
The aim of this observational study was to assess long-term prognosis of
a contemporary octogenarian population admitted to an Intensive Cardiac Care Unit
with acute myocardial infarction (MI), and the prognostic value of two simple
biomarkers obtained at admission: glucose blood level (ABG) and estimated
glomerular filtration rate (eGFR). A total of 293 consecutive patients were
included (202 with ST elevation MI and 91 with non-ST elevation MI) with median
age 83.9 years, 172 (58.7%) male. The optimal cut-off points for all-cause death
defined by ROC curves were ABG
Due to a longer life expectancy, the number of people above age 80 years is
growing rapidly in Europe and Northern America. Consequently, elderly patients
account for an increasing proportion of patients admitted to Intensive Cardiac
Care Units (ICCU) with acute coronary syndromes (ACS). Currently about one-third
of ICCU admissions are of patients aged
The oldest patients, and particularly octogenarians, have been excluded from
most randomized clinical trials conducted before 2010. The available data in
patients with ACS aged
Elevated blood glucose and renal function are well established prognostic risk factors in ACS [3, 4, 5, 6]. Their contribution to long term prognosis in octogenarians is less known and with controversial results [7, 8]. Therefore, the aim of our study was to assess long-term prognosis of a contemporary octogenarian population admitted to an ICCU with acute myocardial infarction, and the predictive value of two simple parameters obtained at admission: glucose blood level and estimated glomerular filtration rate (eGFR).
The University Hospital Joan XXIII from Tarragona, Spain, is the primary medical centre for a population of 250,140 inhabitants and serves as a secondary cardiac centre performing coronary angiography and percutaneous coronary intervention (PCI) for a population about 802,547, with primary PCI service 24 hour/7 days a week.
Data from all patients consecutively admitted to our ICCU are prospectively collected in the RENACI (REgistro NAcional de Cardiopatía Isquémica) database, approved by the Ethics Committee (CEIm65/2008). In this study we analyse all patients who were 80 or more years old, with a final diagnosis of ST segment elevation myocardial infarction (STEMI) or non-ST segment elevation myocardial infarction (NSTEMI), admitted between January 2015 and December 2019. For patients with several hospital stays, only the first stay was considered in analysis. Patients presenting as cardiac arrest were excluded.
Data on age, gender, risk factors for coronary artery disease, past medical
history including previous coronary heart disease, cardiac failure, obstructive
pulmonary disease, history of atrial fibrillation, peripheral or cerebrovascular
disease, hepatic disease or neoplasm and chronic kidney disease (CKD) were
recorded. CKD was defined as an eGFR
The primary end-point was all-cause mortality with a maximum follow-up of 3.5 years. Secondary end-points were readmission for MI or heart failure and a composite of major adverse cardiac events (MACE: first admission for non-fatal MI or heart failure or all-cause death). Vital status and events at follow-up were obtained from electronic medical records.
Categorical variables are expressed as numbers and percentages whereas
continuous variables are expressed as medians and interquartile range (IQR).
Comparisons of categorical data were performed with chi-squared tests or Fisher’s
exact test when expected frequencies were
To determine if patients’ groups were associated with primary and secondary
endpoints, univariate and multivariate Cox regressions were performed with the
backward stepwise procedure. In the multivariate analysis, clinically relevant
and significant variables in the univariate analysis were included. Clinical
model 1 was adjusted by age, male sex, Charlson index, dependence status and
medical history of myocardial infarction, heart failure, cerebrovascular disease,
chronic kidney disease and chronic pulmonary disease and model 2 was adjusted by
age, male sex, atrial fibrillation/flutter at admission, Killip class
Finally, to estimate the ability of admission glucose concentration and eGFR to
improve long-term risk prediction of all-cause death beyond clinical model 1 and
2 we performed ROC curve analyses and the Hosmer-Lemeshow test. The clinical
models were compared before and after adding glycemia concentration
Between 1 January 2015 and 31 December 2019 there were 3582 admissions to the ICCU, of which 603 (17%) were 80 or more years old. The leading cause (61.5%) for admission in this population was acute myocardial infarction (STEMI and NSTEMI). After excluding recurrent episodes, patients presenting with cardiac arrest and patients lost to follow-up (11 patients), the final study cohort included 293 patients (Fig. 1).
Study flowchart. ABG, admission blood glucose; eGFR,
estimated glomerular filtration rate; ICCU, Intensive Cardiac Care Unit; MI,
myocardial infarction; NSTEMI, non-ST elevation myocardial infarction; STEMI, ST
elevation myocardial infarction. Group 1: Low ABG (
Median (IQR) patient age was 83.9 (81.8–86.8) years and 172 (58.7%) were male.
Twenty-four patients (8.2%) were 90 or more years old. Of all patients, 202
(68.9%) were admitted with STEMI and 91 (31.1%) with NSTEMI. Median (IQR)
admission glycemia concentration was 153 (125–222) mg/dL and the best cut-off
point for the prediction of all-cause death was 186 mg/dL (area under the curve
[AUC] 0.585 (95% confidence interval [CI] 0.526–0.642); sensitivity 44%;
specificity 77%). Median (IQR) eGFR was 60.9 (39.0–78.1) mL/min/1.73
m
Baseline characteristics in the overall population and comparison between the three groups are presented in Table 1. Patients in group 3 were more frequently female, non-smoker, diabetic, with medical history of CKD and worse Charlson index.
Variable | Overall | Group 1 | Group 2 | Group 3 | p Value | |
(N = 293) | (N = 145) | (N = 96) | (N = 52) | |||
Demographics | ||||||
Age, years | 83.9 (81.8–86.8) | 84.0 (82.0–87.1) | 83.2 (81.6–86.5) | 84.0 (81.8–86.5) | 0.424 | |
Male sex | 172 (58.7) | 91 (62.8) | 61 (63.5) | 20 (38.5) | 0.005 | |
Cardiovascular risk factors | ||||||
Current or past smoker | 90 (30.7) | 52 (35.9) | 29 (30.2) | 9 (17.3) | 0.045 | |
Hypertension | 235 (80.2) | 111 (76.6) | 80 (83.3) | 44 (84.6) | 0.294 | |
Diabetes mellitus | 110 (37.5) | 32 (22.1) | 46 (47.9) | 32 (61.5) | ||
Hypercholesterolemia | 164 (56.0) | 78 (53.8) | 61 (63.5) | 25 (48.1) | 0.148 | |
Medical history | ||||||
Myocardial infarction | 66 (22.5) | 33 (22.8) | 23 (24.0) | 10 (19.2) | 0.802 | |
Heart failure | 21 (7.2) | 8 (5.5) | 8 (8.3) | 5 (9.6) | 0.445 | |
Cerebrovascular disease | 37 (12.6) | 15 (10.3) | 13 (13.5) | 9 (17.3) | 0.409 | |
Peripheral arterial disease | 33 (11.3) | 10 (6.9) | 14 (14.6) | 9 (17.3) | 0.057 | |
Chronic kidney disease | 58 (19.8) | 7 (4.8) | 28 (29.2) | 23 (44.2) | ||
Chronic pulmonary disease | 66 (22.5) | 28 (19.3) | 25 (26.0) | 13 (25.0) | 0.423 | |
Atrial fibrillation/flutter | 35 (12.0) | 13 (9.0) | 13 (13.5) | 9 (17.3) | 0.237 | |
Charlson index | 1 (0–3) | 1 (0–2) | 2 (1–4) | 2 (1–4) | ||
Partial or total dependence | 51 (17.4) | 22 (15.2) | 16 (16.7) | 13 (25.0) | 0.269 | |
Dementia | 18 (6.1) | 9 (6.2) | 4 (4.2) | 5 (9.6) | 0.440 | |
Data represent the number (percentage) or median (interquartile range).
Group 1: ABG |
Hyperglycaemia
Overall | Group 1 | Group 2 | Group 3 | p Value | ||
(N = 293) | (N = 145) | (N = 96) | (N = 52) | |||
Physical examination at admission | ||||||
Atrial fibrillation/flutter | 50 (17.1) | 22 (15.2) | 18 (18.8) | 10 (19.2) | 0.694 | |
Killip class |
129 (44.0) | 48 (33.1) | 44 (45.8) | 37 (71.2) | ||
Laboratory findings at admission | ||||||
Glycemia (mg/dL) | 153 (125–222) | 134 (113–157) | 166 (135–244) | 246 (223–337) | ||
eGFR (mL/min per 1.73 m |
60.9 (39.0–78.1) | 75.5 (62.3–84.2) | 49.3 (30.9–67.5) | 34.7 (27.2–40.9) | ||
Haemoglobin (g/dL) | 12.3 (10.8–13.4) | 12.5 (11.4–13.7) | 12.5 (11.0–13.5) | 11.1 (9.8–12.8) | 0.002 | |
GRACE score | ||||||
GRACE score | 163 (144–183) | 159 (139–177) | 168 (148–185) | 179 (153–214) | ||
Coronary disease | ||||||
Significant three vessels or left main stenosis | 78 (26.6) | 29 (20.0) | 34 (35.4) | 15 (28.9) | 0.027 | |
PCI | 248 (84.6) | 127 (87.6) | 77 (80.2) | 44 (84.6) | 0.298 | |
Fibrinolysis | 3 (1.0) | 2 (1.4) | 1 (1.0) | 0 (0.0) | 1.000 | |
LVEF | ||||||
LVEF |
89 (31.0) | 34 (23.6) | 34 (36.6) | 21 (42.0) | 0.020 | |
Intensive management during hospitalization | ||||||
Non-invasive mechanical ventilation | 15 (5.1) | 3 (2.1) | 7 (7.3) | 5 (9.6) | 0.037 | |
Invasive mechanical ventilation | 18 (6.1) | 2 (1.4) | 9 (9.4) | 7 (13.5) | 0.001 | |
Vasoactive drugs | 43 (14.7) | 8 (5.5) | 19 (19.8) | 16 (30.8) | ||
Intra-Aortic balloon pump | 7 (2.4) | 2 (1.4) | 5 (5.2) | 0 (0.0) | 0.112 | |
Discharge diagnostic | ||||||
STEMI | 202 (68.9) | 101 (69.7) | 61 (63.5) | 40 (76.9) | 0.236 | |
NSTEMI | 91 (31.1) | 44 (30.3) | 35 (36.5) | 12 (23.1) | 0.236 | |
Death during hospitalization | ||||||
All-cause death | 32 (10.9) | 6 (4.1) | 11 (11.5) | 15 (28.9) | ||
Data represent the number (percentage) or median (interquartile range).
Group 1: ABG |
In the subgroup of 202 STEMI patients, a primary PCI was performed in 179 (88.61%), a delayed PCI in 9 (4.45%) and only one underwent rescue PCI — a total of 189 (93.6%) invasively treated STEMI patients. In the subgroup of 91 NSTEMI patients a PCI was performed in 59 (64.84%) and only one patient was referred for coronary by-pass surgery. Finally, 32 deaths (10.9%) were seen during hospitalization with the highest mortality in groups 2 and 3 (11.5% and 28.9%, respectively) and a relatively low mortality in group 1 (4.1%).
During 3.5 years of follow-up (median 2.2 [IQR 0.8–3.5] years), 129 (44%)
patients died, with the highest mortality in group 2 and group 3 (Table 3,
Fig. 2). Univariate and multivariate Cox regression analysis were performed. The
variables that remained statistically significant in the multivariate analysis
were age, previous MI or cerebrovascular disease, GRACE score, LVEF
All-cause death and MACE cumulative survival (A and B) and cumulative incidence of readmission for myocardial infarction and heart failure (C and D). Group 1 (blue line): AGL
Variable | Overall | Group 1 | Group 2 | Group 3 | p Value | |
(N = 293) | (N = 145) | (N = 96) | (N = 52) | |||
MACE | ||||||
1 year | 112 (38.2) | 36 (24.8) | 44 (45.8) | 32 (61.5) | ||
2 years | 142 (48.5) | 50 (34.5) | 56 (58.3) | 36 (69.2) | ||
3.5 years | 162 (55.3) | 58 (40.0) | 63 (65.6) | 41 (78.8) | ||
All-cause death | ||||||
1 year | 78 (26.6) | 23 (15.9) | 30 (31.3) | 25 (48.1) | ||
2 years | 106 (36.2) | 33 (22.8) | 42 (43.8) | 31 (59.6) | ||
3.5 years | 129 (44.0) | 40 (27.6) | 51 (53.1) | 38 (73.1) | ||
Myocardial infarction | ||||||
1 year | 31 (10.6) | 13 (9.0) | 12 (12.5) | 6 (11.5) | 0.255 | |
2 years | 37 (12.6) | 15 (10.3) | 16 (16.7) | 6 (11.5) | 0.143 | |
3.5 years | 43 (14.7) | 18 (12.4) | 19 (19.8) | 6 (11.5) | 0.082 | |
Heart failure | ||||||
1 year | 21 (7.2) | 5 (3.4) | 9 (9.4) | 7 (13.5) | 0.003 | |
2 years | 31 (10.6) | 11 (7.6) | 12 (12.5) | 8 (15.4) | 0.021 | |
3.5 years | 36 (12.3) | 14 (9.7) | 14 (14.6) | 8 (15.4) | 0.037 | |
Data represent the number (percentage) or median (interquartile range).
Group 1: ABG |
MODEL 1 | Univariate cox regression | Multivariate cox regression | |||
Variables | HR (95% CI) | p Value | HR (95% CI) | p Value | |
Age | 1.07 (1.02–1.12) | 0.008 | 1.08 (1.03–1.13) | 0.002 | |
Male sex | 0.92 (0.65–1.31) | 0.652 | - | - | |
Previous myocardial infarction | 1.50 (1.02–2.19) | 0.039 | 1.55 (1.05–2.29) | 0.026 | |
Previous heart failure | 1.86 (1.08–3.18) | 0.025 | - | - | |
Previous cerebrovascular disease | 1.77 (1.13–2.78) | 0.013 | 1.70 (1.08–2.68) | 0.023 | |
Chronic kidney disease | 1.61 (1.09–2.39) | 0.017 | - | - | |
Chronic pulmonary disease | 1.56 (1.07–2.27) | 0.022 | - | - | |
Charlson index | 1.13 (1.04–1.22) | 0.004 | - | - | |
Partial or total dependence | 1.53 (1.13–2.09) | 0.006 | - | - | |
Group 2 | 1.47 (1.03–2.09) | 0.034 | 2.46 (1.62–3.73) | ||
Group 3 | 2.88 (1.97–4.21) | 4.19 (2.67–6.57) | |||
MODEL 2 | Univariate cox regression | Multivariate cox regression | |||
Variables | HR (95% CI) | p Value | HR (95% CI) | p Value | |
Age | 1.07 (1.02–1.12) | 0.008 | 1.08 (1.03–1.14) | 0.003 | |
Male sex | 0.92 (0.65–1.31) | 0.652 | 1.60 (1.08–2.37) | 0.019 | |
Atrial fibrillation/flutter | 1.65 (1.09–2.51) | 0.019 | - | - | |
Killip class |
2.38 (1.68–3.38) | - | - | ||
Haemoglobin | 0.81 (0.74–0.89) | 0.86 (0.78–0.95) | 0.005 | ||
GRACE score | 1.02 (1.01–1.02) | 1.01 (1.00–1.02) | 0.004 | ||
PCI | 0.38 (0.25–0.56) | 0.37 (0.25–0.56) | |||
LVEF |
2.67 (1.87–3.80) | 1.73 (1.17–2.55) | 0.006 | ||
Non–invasive mechanical ventilation | 3.33 (1.87–5.93) | 2.12 (1.11–4.02) | 0.022 | ||
Invasive mechanical ventilation | 4.59 (2.62–8.05) | 3.33 (1.71–6.47) | |||
Vasoactive drugs | 3.70 (2.47–5.54) | - | - | ||
Group 2 | 1.47 (1.03–2.09) | 0.034 | 1.80 (1.17–2.79) | 0.008 | |
Group 3 | 2.87 (1.97–4.21) | 2.56 (1.55–4.22) | |||
CI, confidence interval; GRACE, Global Registry of Acute Coronary Events
score; Group 2: admission blood glucose |
During follow-up, in 162 (55.3%) patients the composite of MACE events was observed. Distribution among groups is presented in Table 3. Patients with high ABG and/or low eGFR (group 2 and 3) presented the highest number of events (Table 3, Fig. 2). In multivariate Cox regression analysis, group 2 and especially group 3 were independently related with a higher risk of MACE (Supplementary Table 1).
A total of 43 (14.7%) patients developed a new MI during follow-up and 36 (12.3%) required hospitalization for heart failure (Table 3). Multivariate competing risk analysis revealed a statistically significant association only for heart failure admission in group 3. Cumulative incidence of readmission for myocardial infarction and heart failure is shown in Fig. 2.
ROC curves were evaluated with AUC to determine whether admission glycemia
concentration
ROC curves for predicting all-cause death before
(blue line) and after (red line) the addition of ABG
Reclassification and discrimination were also improved in both clinical models
after the addition of glycemia concentration
The main findings of our study are:
(1) In this contemporary cohort of octogenarian patients with ACS admitted to an ICCU a high proportion of patients was treated invasively, especially the subgroup with STEMI.
(2) Despite that, nearly half of the patients died within 3.5 years (median follow up 2.2 years) after admission.
(3) An ABG
(4) These variables, when added to a medical history based (Model 1) or an index event based predictive model (Model 2), improved long-term risk prediction of all-cause death.
At present, a universal definition for “elderly” is lacking. Even if
biological age may not correspond with chronological age, most now consider
“elderly” as those aged
Randomized trials have failed to demonstrate a mortality benefit with an invasive approach in elderly population, probably due to the fact that this invasive strategy is still limited to the lower risk population and less comorbid patients and that only half of the patients undergoing angiography will undergo revascularization, the procedure with a potential impact on mortality [17, 18, 19]. However, consistent mortality reductions have been observed in real-world registries of elderly patients with Non-ST Elevation Acute Coronary Syndrome (NSTE-ACS) over the last 20 years, concomitant with an increase in the use of an early invasive approach [20, 21]. Based upon this evidence, the ESC 2020 guidelines for management of ACS without persistent ST-segment elevation recommend to apply for older people the same interventional strategies used for the younger ones, although with a more personalized approach [22].
We have included consecutive octogenarian patients, with a median age of 84
years old, admitted to an ICCU with STEMI and NSTE-ACS. Although NSTEMI is the
most frequent clinical presentation of ACS in the elderly, in our study two
thirds were STEMI patients, reflecting the fact that our hospital is the
reference site for primary PCI for a large population. Index event related
variables also exhibited high-risk features. Nearly 50% of patients presented
Killip class
Scarce information is available with respect to long-term survival of elderly
patients with ACS. Kvakkestad et al. [24] analysed in-hospital mortality and
long-term survival in a cohort of STEMI patients according to age. In the
subgroup of patients
On admission to hospital, it is recommended to evaluate glycaemic status in all patients with ACS, regardless of a history of diabetes, and to assess kidney function by eGFR for prognostic reasons and to identify patients at risk of contrast-induced nephropathy [16, 22]. In the present study, we have applied the best cut-off points for all-cause death prediction for admission blood glucose and eGFR and segregated the cohort into 3 groups, according to the presence of both, either or no values beyond the cut-offs, allowing discrimination of the subgroups (group 2 and group 3) of patients with a higher long-term mortality and MACE risk. We have focused on these two biomarkers because, unlike some scores, they can be easily obtained and are linked to diabetes and CKD, two established conditions related to the prognosis and severity of presentation of ACS.
Chronic kidney disease is a common condition in elderly ACS patients. The proportion of patients with CKD is increasing constantly as elderly people are living longer, but also because of an increasing prevalence of hypertension and diabetes mellitus. CKD is associated with adverse outcomes through the entire spectrum of ACS [4, 8, 26, 27]. An advanced age and CKD, frequently associated with comorbidities and geriatric syndromes, are often used as reasons to withhold patients from undergoing angiography and PCI [4, 8, 27], despite evidence of improvement in one year [4] and long-term survival with PCI in NSTEMI patients with CKD [27].
Assessment of renal function in elderly people is debated, the issue being
whether the lower eGFR values observed at older ages can be attributed to
physiologic aging or is associated with kidney disease and comorbidities. There
is not a general agreement about which is the best method to estimate GFR rate at
advanced ages. We have decided to use the recommended CKD-EPI formula [9] that
has also demonstrated good performance with respect to measured GFR in acute
STEMI patients [28]. In our cohort, a lower threshold (eGFR
Several studies suggest that the association of a lower eGFR with mortality is
weaker with increasing age [7, 26]. In the LONGEVO-SCA registry, including NSTEMI
patients aged
Hyperglycaemia during ACS may reflect stress-induced hyperglycaemia (SIH),
worsening glycaemic control among diabetic patients or identify previously
undiagnosed diabetes. SIH is considered an acute response of the body to many
critical illnesses, including ACS, and many observational studies have documented
that hyperglycaemia occurs frequently in this context. The association between
blood glucose levels on admission and long-term mortality in patients with ACS
has been previously recognised regardless of diabetic status [5, 32]. In patients
Several limitations to this study should be considered. This is a single centre study that includes patients 80 or more years old with ACS, predominantly STEMI, admitted to an ICCU. The population studied probably excluded patients with poor general health and cognitive status, who would have been treated non-invasively outside the ICCU, especially in the subgroup of NSTEMI patients. It is an observational study and we cannot exclude a selection bias and/or uncontrolled confounders. Data regarding medical treatment on admission and at follow-up are lacking. Glucose and creatinine were measured on admission but we are unaware if the time between symptom onset and sample extraction could improve or worsen the observed results. We have not used validated scales for geriatric syndromes or frailty measurements, although STEMI was the more frequent type of ACS seen in our population (more than two thirds of included patients) and a comprehensive geriatric assessment is very difficult in the STEMI scenario. Because our study was limited to octogenarian patients the results may not apply to younger patients. However, we believe that despite these limitations our study provides interesting data regarding risk stratification and prognosis in this subgroup of patients.
Our study evaluates the predictive value for all-cause death of the combination
of ABG and eGFR in octogenarian patients with STEMI and NSTEMI. Long-term
mortality was high despite PCI being performed in a significant proportion of
patients. A high glycaemia concentration on admission (
The scope for further reductions in long-term mortality is likely to be much greater for older than for younger patients with acute myocardial infarction. The variables identified in this study provide more opportunities for risk stratification. Treating non-recognised diabetic patients, prescribing guideline adherent therapies in CKD patients and providing access to cardiac rehabilitation teams, together with the use of other specific instruments for geriatric population such as nurse intervention [35], simplifying treatments, and enhancing care transitions, could help improve outcomes in this population.
ESG, OP and GB—conceived and designed the study; JRL, CS, MFG, KV and ARN—collected data; OP and GB—analysed the data; ESG, OP and GB—wrote the paper; AB revised the manuscript. All authors read and approved the final manuscript.
The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee (CEIm65/2008).
We would like to thank all the professionals from the ICCU who take care daily of elderly people with kindness and dedication. Thanks to all the peer reviewers for their opinions and suggestions.
This research received no external funding.
The authors declare no conflict of interest.