Prognostic value of frailty in elderly patients with acute coronary syndrome: a systematic review and meta-analysis

Background Frailty is common and associated with poorer outcomes in the elderly, but its prognostic value in acute coronary syndromes (ACS) requires clarification. We thus undertook a systematic review and meta-analysis to evaluate the relationship between frailty and poor prognosis in patients with ACS. Methods We systematically searched PubMed, Embase to find literatures which studied the prognostic value of frailty in elderly patients with ACS. Our main endpoints were the all-cause mortality, cardiovascular disease (CVD), major bleeding and readmissions. We pooled studies using random-effect generic inverse variance method, and conducted three pre-specified subgroup analyses. Results Of 1216 identified studies, 15 studies were included in our analysis. Compared with the normal group, frailty (HR = 2.65; 95%CI: 1.81–3.89, I2 = 60.2%) and pre-frailty (HR = 1.41; 95%CI: 1.19–1.66, I2 = 0%) were characterized by a higher risk of mortality after adjustment. Frailty also was associated with increased risk of any-type CVD, major bleeding and hospital readmissions in elderly patients with ACS. The pooled effect sizes in frail patients were 1.54 (95%CI: 1.32–1.79), 1.51 (95%CI: 1.14–1.99) and 1.51 (95%CI: 1.09–2.10). Conclusions Frailty provides quantifiable and significant prognostic value for mortality and adverse events in elderly ACS patients, helping doctors to appraise the comprehensive prognosis risk and to applicate appropriate management strategies. Electronic supplementary material The online version of this article (10.1186/s12877-019-1242-8) contains supplementary material, which is available to authorized users.


Background
The accomplishments of cardiovascular prevention have decreased the incidence of acute coronary syndromes (ACS) and have delayed the onset age of ACS. The progresses in the treatment of ACS (dual antithrombotic therapy and invasive strategy, e.g. the drug eluting stent) significantly lowers the mortality of ACS and lead to a swift growth in the portion of elderly patients. With the fast population ageing, the mean onset age of ACS patients rises stably in the last decades [1]. Advanced age is one of the forceful prognosticators of mortality and morbidity of ACS. Age as a crucial prognostic marker is presented in the majority of ACS risk scores, including the Thrombolysis in Myocardial Infarction (TIMI) risk score and the Global Registry of Acute Coronary Events (GRACE) score [2].
Frailty, one of the most important health problems in geriatrics, commonly exists in ACS patients, in part due to mutual risk factors. A frail phenotype representing decreased physiological reserve and increased vulnerability reflects better biological age, therefore, it may cause the heterogeneity in clinical consequences within the elderly patients [3]. Frailty has become a substantial factor in assessment of several special medical situations and been embedded into clinical decision making, such as evaluation of surgical risk and cancer treatment. However, this has not yet become incorporated as part of routine management of ACS.
The well-defined pathways for the management of ACS, largely based on randomized controlled trial (RCT) evidence, may not be generalizable to elderly frail patients. Indeed, to date, there is no international guidelines as to how frail ACS patients should be treated. Previous studies have reported frailty was connected with higher mortality or adverse events in ACS patients with some inconsistent information. We therefore undertook a systematic review and meta-analysis to examine the significance of frailty on ACS prognosis, including mortality, cardiovascular events, major bleeding and readmission.

Data sources
The methods of this systematic review and meta-analysis were performed in accordance with the Meta-analysis Of Observational Studies in Epidemiology (MOOSE) Statement [4] (Additional file 1: Text S1). A comprehensive literature search was performed up to July 1, 2018. The language was restricted to English. The primary sources were the electronic databases of PubMed and Embase, using various combinations of Medical Subject Headings (MeSH) and non-MeSH terms: "Frailty" combined with "Myocardial Infarction", "Acute Coronary Syndrome" and "Heart Attacks". Additional file 1: Text S2 presents the full search strategy. All results were exported to EndNote for the removal of duplicates. We also screened reference lists of published reviews to identify additional relevant studies.

Study selection and data extraction
The titles/abstracts and full texts were screened by two investigators (Qingyu Dou and Wen Wang) independently. Studies met following criteria were included: (1) performed a well-defined cohort design; (2) used frailty as a major exposure in elderly ACS patients; (3) elderly population with the age of 65 or greater (4) displayed hazard ratio (HR) or relative risk (RR) for outcomes with a 95% confidence interval (CI) or provided sufficient information to calculate these data. Exclusion criteria included: (1) case reports, reviews and conference abstracts; (2) insufficient parameters concerning main outcomes. (3) Using only one item (e.g. low gait speed) as a marker of frailty.
The related information and parameters from all included studies were extracted by two investigators including the first author's name, the year of publication, the place of study, the design of study, population, sample size, follow-up time, assessment methods of frailty, prevalence for frailty/prefrailty, and HR or RR of main outcomes with 95% CI. The main outcomes included all-cause mortality, cardiovascular events (re-infarction and stroke/TIA), composite outcome of death and cardiovascular events, major bleeding and readmission during follow-ups. When a cohort was represented by two or more studies, all available articles were included to make the most complete analyses. Discrepancies were addressed by discussion with a third author (Hui Wang).

Quality assessment
Two authors independently assessed the quality of studies, and disagreements were re-evaluated by a third author. The Newcastlee-Ottawa Quality Assessment Scale [5] was used to evaluate the quality of the literature. The scale assessed the selection of cohorts, comparability of cohorts and the quality of outcomes by 9 parameters. Scores range from 0 to 9. The quality of studies were graded as follows: good (≥ 8 stars); fair (5-7 stars); and poor (< 5 stars).

Statistical analysis
We conducted analyses of adjusted and unadjusted estimates separately. We used relative risk (RR) or hazard risk (HR) as the effect measure for the association of frailty and adverse outcomes. HR or RR was pooled using random-effect generic inverse variance method. Statistical heterogeneity among studies were evaluated with chisquared and I-squared statistics [6]. We explored sources of heterogeneity using three pre-specified subgroup hypotheses: type of patients (frailty vs pre-frailty); type of ACS (ST-segment elevation myocardial infarction vs non-ST-segment elevation myocardial infarction); follow-up time (during hospital/within 1 month vs ≤ 1 year vs > 1 year). Subgroup analyses were performed if there were at least two studies in each subgroup category. We detected publication bias using visually examining symmetry of funnel plots and Egger's tests [7]. We performed a sensitivity analysis by omitting retrospective studies. All statistical tests will be performed with the STATA14.0 software. All statistical tests are two sided, P < 0.05 is considered statistically significant.

Search strategy and research characteristics
Overall, 1216 publications for possible inclusion were revealed by the initial systematic search of the databases. After the removal of duplicates, the remaining titles/abstracts were examined and irrelevant research were excluded, mainly because they were conference abstracts, reviews and case series. Then 24 left articles were carefully reviewed in full texts. Nine of them were excluded for different reasons. Fifteen studies [8][9][10][11][12][13][14][15][16][17][18][19][20][21][22] including 8554 patients were finally identified. No additional studies were found after manual inspection of the references. Figure 1 showed the flow chart of literature search strategy and detailed reasons of exclusion.

Discussion
In the present systematic review and meta-analysis, case series, conference abstracts or reviews, and studies used different exposure or outcome assessment were excluded. All the 15 included studies were of relatively high methodological quality. In the sensitivity analysis, the results were stable after the retrospective study by Sujino was removed. Our results demonstrated in elderly ACS patients, frailty significantly increased the all-cause mortality risk by 2.65-fold, any-type CVD risk by 1.54-fold, major bleeding risk by 1.51-fold and hospital readmissions risk by 1.51-fold.  A previous systematic review reported that the overall prevalence of frailty, in community-dwelling adults aged 65 and older, is on average 10.7%. Moreover, prevalence of frailty increases with age, reaching 15.7% in individuals aged 80 to 84 and 26.1% in those aged 85 or more [23]. The proportion of frailty and pre-frailty in our meta-analysis were significantly higher than community population. Both ACS and frail patients had higher rates of common existed CVD risk factors, like hypertension, type 2 diabetes and lack of exercise. However, regardless of these possible mixed CVD risk factors, frailty itself brings about increased risk of CVD [24]. The pathophysiologic mechanism of frailty, including elevated inflammatory state (interleukin-6 and C-reactive protein [25]), higher markers of thrombosis (D-dimer [25]) and endocrine unbalances (lower insulin-like growth factors − 1 [26]) could act a part in the onset and outcome of ACS. Furthermore, frailty has accelerated biological aging modifications(e.g. higher oxidative stress levels [27], impaired autophagy [28] and shorter telomere length [29] that further give impetus to the development and poor prognosis of ACS. Frailty could impose obvious influence on the management of elderly ACS patients, especially on revascularization. Since the risk of operational complication rises in frail patients, a less invasive strategy may be preferred. Observational study shows primary percutaneous coronary intervention (PCI) was performed less frequently in patients with frailty compared with their non-frail counterparts [30]. For STEMI patients, timely reperfusion in all patients with ischemia symptoms and persistent ST-segment elevation within 12 h, is the cornerstone of treatment. Studies have demonstrated performing PCI reduced in-hospital mortality, even in patients ≥80 years with frailty [19,22]. Since our meta-analysis found frailty increased the risk of major bleeding, primary PCI is safer than fibrinolytic therapy especially in high risk elderly STEMI patients with frailty. However, frailty greatly increases all-cause mortality by 6.51-fold in STEMI patients in our study, which may reduce the ability to benefit from interventions. In these extremely frail STEMI patients with a high mortality despite intervention, the death risk should be fully informed and requires a shared decision by doctors and families. The decision not to accept interventional therapy is understandable and reasonable.
The role of an invasive strategy in frail NSTEMI patients is still worth exploring. An early invasive strategy was found to have more benefit in the elderly than younger patients, unless there were extensive and complicated comorbidities [31,32]. Nowadays, there are studies aimed to judge whether the prognostic impact of PCI in NSTEMI differs across frailty status. The LONGEVO-SCA registry included unselected NSTEMI patients aged ≥80 years. The incidence of cardiac events was more common in patients managed conservatively after adjusting for confounding factors.
However, this association was not significant in patients with established frailty criteria [33]. In recent Nuñez's study [34], a prospective observational study of 270 elderly patients hospitalized for NSTEMI, at a median follow-up of 4.4 years, in patients with Fried ≥3, PCI was associated with a significant reduction of risk of all-cause and cardiovascular-rehospitalizations without reducing all-cause mortality. In 2015 ESC guidelines for the management of NSTEMI, revascularization in elderly patients should be took into consideration after cautious weighing up benefits and risks, including comorbidities, frail state, predicted life expectancy, quality of life, and patient preferences [35].
With regard to medical treatment, as bleeding risk increases with age, comorbidities, polypharmacy and declined renal function, elderly patients are at particular risk of bleeding. In addition to routine evaluation of bleeding risk (HAS-BLED bleeding score), based on our study, we also recommend that the frailty assessment be considered to tailor the individualized antithrombotic treatment for elderly patients. It is essential to reduce bleeding risk according to the degree of frailty; these include employing proper dosage of antithrombotic drugs, avoidance of a glycoprotein (GP) IIb/IIIa-inhibitor; adding a proton pump inhibitor; avoiding the use of non-steroid anti-inflammatory drugs and using radial access whenever possible [36].
Our study found frailty increases both the risk of major bleeding and the risk of cardiovascular events. It is impossible to reduce the incidence of cardiovascular events by increasing the intensity of antiplatelet therapy. Frail patients are also less likely to take angiotensin-converting enzyme inhibitors/angiotensin receptor blockers or βblockers, since they are more likely to have adverse drug reactions from medical therapy. The frailty as a therapeutic goal intervened by non-pharmacological means is the future hotspot of research. The earlier stage of frailty is reversible and could be remedied. The non-pharmacological interventions (e.g. cardiac rehabilitation, physical exercise, and removing unnecessary medications may postpone and reduce the risk of CVD [37,38]. A recent study showed patients with diabetes mellitus who have undergone PCI, cardiac rehabilitation participation was associated with significantly reduced all-cause mortality and composite end point of mortality, myocardial infarction, or revascularization [39]. Meanwhile, nutritional supplement with 25-30 g of high-quality protein per meal have slowed or prevented sarcopenia, a manifestation of prefrailty [40]. In particular, physical activity interventions might play a pivotal role in the prevention of both CVD and frailty. More studies are required to confirm its function and establish standard exercise prescriptions.

Limitation
There are following limitations in our meta-analysis. First, even though we performed subgroup analysis according to type of ACS and follow-up time, there were not sufficient data to detect the statistical difference between subgroups. Second, the included criterion of our review was based on the age of 65 years or more. However, two studies just recruited participants older than 80 years and another study only included STEMI patients older than 85 years, which may result in the heterogeneity of our studies. Last, none of included studies in our meta-analysis reported results concerning quality of life. Future studies should use more patient-centered consequences such as activity of daily living (ADL) as primary outcome.

Conclusion
Our study suggests that in elderly ACS patients, frailty assessment should be integrated into the current existing management to better appraise the comprehensive prognosis risk. The identification of frailty help doctors to applicate appropriate management strategies including invasive therapy and antithrombotic medication, and help patients make properly informed choices. Further, the value of frailty as a therapeutic target should be given full attention. Currently, there is little evidence that frailty management could improve outcomes of elderly ACS patients. It is necessary to conduct more studies related the effect of the frailty intervention on elderly ACS prognosis in the future.