The prognostic value of clinical frailty and American Society of Anesthesiology score in patients with chronic limb threatening ischaemia
Abstract
Summary:Background: Frailty is a complex multisystem syndrome associated with increased comorbidity and decreased physiological reserve. There are associations between frailty and adverse outcome in surgical patients. Chronic limb threatening ischemia (CLTI) is increasingly prevalent, with a typically frail patient population. Existing frailty scoring systems focus on functional measures and do not reliably assess comorbidities. The present study aims to describe the prognostic value of multimodal frailty assessment in patients with CLTI. Patients and methods: Patients >50 years old admitted as an emergency with CLTI between May 2020 to June 2021 were included. Frailty was measured using Clinical Frailty Score (CFS), and comorbidities with American Society of Anesthiologists score (ASA). A composite score combining CFS and ASA was derived and the prognostic value compared with each component score. The primary outcome was overall survival. Results: There were 249 eligible patients, 53.4% (n=133) had CFS>4. The mean (95% CI) overall survival for the CFS>4 cohort was 15.9 (13.6–18.3) months vs. 28.5 (26.1–30.9) months for CFS≤4 cohort (p<0.001). Increasing CFS-ASA score was associated with inferior survival on univariate (HR=2.84, 95% CI [1.96–4.11], p<0.001) and multivariate (HR=1.78, 95% CI [1.20–2.64], p<0.01) analyses. ROC-analysis showed comparable prognostic value of CFS and CFS-ASA to predict one-year survival. Conclusions: Frailty is highly prevalent and a poor prognostic indicator in patients with CLTI admitted as an emergency. Our results suggest that incorporating assessment of comorbidities into frailty assessment may offer prognostic value, but comparable to existing clinical frailty assessment. Further work to identify patients with inferior prognosis is required.
Introduction
Peripheral arterial disease (PAD) is a rapidly increasing health burden in the UK [1]. The most severe manifestation of PAD is chronic limb threatening ischemia (CLTI), defined by the 2019 Global Vascular Guidelines (GVG) as ischaemic rest pain, gangrene, or lower limb ulceration of >2 weeks duration [2]. The prevalence of CLTI is increasing and may be up to three times higher in those over 70 years old [2, 3]. Furthermore, it is recognised that patients presenting with CLTI, have poorer outcomes and present a significant healthcare burden [4]. CLTI is associated with mortality, amputation and poor quality of life [2, 4]. Despite the GVG framework for the investigation and treatment of CLTI, there is a growing recognition that treatment options and management decisions are progressively complex in this elderly, comorbid patient cohort [5].
Frailty is a complex multisystem disorder associated with aging and poor functional reserve [6]. The Clinical Frailty Scale (CFS) is a well-validated 9-point clinical scoring tool that provides a quantifiable measure of frailty [7, 8] (Table A1 in Appendix). An association between inferior outcome and frailty in surgical patients has been described [8, 9, 10]. Frailty in vascular surgical patients has been described, however the evidence base is largely retrospective and encompasses multiple vascular pathologies [5, 8].
Current scoring systems to assess frailty are based largely on patients’ functional state. This describes only one component of the frailty syndrome, and neglects to include assessment of comorbidity. The ASA score is already widely used in surgical populations and gives a subjective assessment of patient fitness, including assessment of medical comorbidity [11, 12].
Implementing routine assessment and management of frailty may aid in surgical decision making and treatment strategy, which may have the potential to improve management of frail patients undergoing major surgical intervention [5, 8]. The present study aims to describe the prevalence of frailty in cohort of patients with an unscheduled presentation with CLTI, assess the prognostic value of multimodal frailty assessment in this patient group, and to evaluate a novel scoring system in these patients incorporating both functional state and comorbidity.
Patients and methods
This was a retrospective analysis of a prospectively maintained database carried out at a tertiary referral university teaching hospital in East of Scotland, United Kingdom. Consecutive patients over 50 years old who presented as an unscheduled admission with CLTI to the Department of Vascular Surgery at Ninewells Hospital between 21/05/2020 to 18/06/2021 were entered into a prospectively maintained local database. Patients were diagnosed with CLTI based on the 2019 GVG definition [2]. Severity of disease was clinically assessed by the on-call vascular surgeon using the Rutherford Category [2]. Rutherford Score ≥5 and <5 was considered as a categorical covariate for analyses. Diagnosis was confirmed with ankle-brachial pressure index (ABPI)/arterial duplex ultrasound and computed tomography angiography (CTA) either during or before admission to confirm the presence of significant arterial disease. Institutional Caldicott approval was obtained prior to data collection to govern data handling and therefore individual patient consent was waived.
On admission, all patients were prospectively scored on admission using the 9-point CFS [7] by the admitting doctor and if over 65 years old, triggered routine assessment and scoring by surgical acute frailty team (SAFT). Where present, the latter score was recorded in our data set. As frailty has been identified in patients >50 years old in Scotland, we therefore lowered our inclusion criteria to ensure we captured these patients [13]. To ensure all suitable patients were reviewed, if frailty was identified in patients <65 years old, they were routinely referred to the surgical frailty team as per hospital guidelines. In our institution, SAFT are able to provide input for non-surgical problems commonly occurring in patients with frailty, such as acute delirium, and assisting with long-term discharge planning. Patients were classified as frail or non-frail, defined by a CFS>4 or ≤4 respectively, in keeping with domains of the CFS and previous literature [7, 8]. CFS scoring is a routinely performed in our institution and medical staff receive training in how to perform it at their induction.
In addition, demographic data for each patient was recorded, including age, gender, co-morbidities, ASA score, and medication on admission. In the patients undergoing intervention, this was recorded by a vascular anaesthetist. In those who did not require anaesthetic assessment, ASA was decided by retrospective collaborative discussion between authors following ASA guidelines [14]. Patients were subgrouped by age into >80 and ≤80 for analysis, in keeping with previous literature suggesting >80 is a significant independent predictor of mortality [15]. All patients completed a minimum of 1 year follow-up. Data included date and duration of index admission, death and date of death, inpatient mortality, 30-day survival, length of survival (months), incidence of major adverse cardiovascular events (MACE, including myocardial infarction (MI), stroke or death related to the former), reintervention and readmission. Readmission was defined as patient requiring urgent or unscheduled inpatient admission to hospital for any cause.
The primary outcome was overall survival. Secondary outcomes were rate of MACE, 30-day and one-year survival, readmission, and reintervention.
Statistical analysis
Categorical variables are expressed as frequency and percentages respectively. Continuous data is shown as median with interquartile range. The χ2 test was used to assess the difference between categorical variables and Kruskal-Wallis test for continuous variables. Survival analysis was performed using Kaplan-Meier plot and log rank statistics, with p-value <0.05 denoting significance. Patients were subgrouped based on ASA (≤2/>2, in keeping with previous literature [11, 12]) and CFS (≤4/>4) and outcomes compared between groups. Patients were assigned an integer score (0/1) based on ASA and CFS categories giving a combined CFS-ASA score (0–2). Outcomes were compared based on the combined score.
The effect of clinical covariates on overall survival was examined using a cox proportional hazards model. Covariates with p<0.10 on univariate analysis were incorporated into a multivariate model. To avoid potential multicollinearity between ASA and comorbidity data, specific comorbidities which would typically contribute directly to ASA were not included in the survival model. The effect of CFS subgroup on 30-day and 1-year reported as % survival (% SE), and with a binary logistic regression model, with results reported as Odds Ratio (OR) and 95% confidence intervals (CI).
ROC analysis was used to assess the predictive value of ASA, CFS and CFS-ASA on one-year survival. p<0.05 was considered statistically significant. Patients with missing parameters were selectively excluded from relevant analyses. Statistical analyses were carried out using SPSS Statistics v28.0 (IBM, NY, USA). Data are reported as median (interquartile range (IQR)) unless otherwise specified.
Results
Patient characteristics
The characteristics of the study population are summarised in Table I. There were 249 patients included in this study. All patients were allocated a CFS score of which 133 (53.4%) had a CFS>4. Between the CFS>4 and CFS≤4 cohort, the majority presented with Rutherford Stage 5 disease (80, 75.5% vs. 100, 75.8%), whilst 7 (6.6%) vs. 11 (8.3%) presented with Rutherford Stage 6 disease (p>0.05 for all comparisons). The median age was 72.8±11.3 years old and 73 (29.3%) were over the age of 80. The CFS>4 cohort had a higher proportion of patients >80, a higher incidence of prior stroke and ischaemic heart disease (IHD), a tendency towards higher ASA score, and lower baseline BMI and albumin values (p<0.05 for all comparisons). Other baseline variables were similar between the CFS≤4 and CFS>4 cohorts.
The management strategies are outlined in Table II. In all patients, 194 (77.9%) received intervention during their first admission. 30 (12.0%) were managed conservatively and 25 (10.0%) received palliative “end-of-life” treatment. There was a statistically significant increase in number of frail patients who were directed to “palliative” management compared to the non-frail (17(12.9%) vs. 2(1.9%), p=<0.001). There were 39 (21.2%) major amputations performed during index admission. Of these, 24 (13.0%) were above knee amputation (AKA) and 15 (8.2%) were below knee amputation (BKA). There were 94 (51.1%) endovascular procedures, 35 (19.0%) open revascularisation procedures and 13 (7.1%) combined open/endovascular procedures. A CFS≤4 favoured both endovascular (47, 44.3% vs. 41, 31.1%, p=0.008) and open revascularisation (24, 22.6% vs. 11, 8.3%, p=0.003). The median (IQR) time for admission to intervention was 7 days (±25.9).
Clinical outcomes
The median (IQR) follow-up time was 24.0 (5.0) months and all patients completed a minimum of 12 months follow-up. During this period, 58 patients suffered from MACE; 25 from MI, 9 stroke and 30 acute cardiac deaths (death due to sudden cardiac arrest diagnosed by the attending medical team). There were 112 all-cause deaths during the follow-up period.
The Kaplan-Meier curve for overall survival in subgroups of CFS is shown in Figure 1. The mean (95% CI) overall survival for the CFS>4 cohort was 15.9 (13.6–18.3) months compared with 28.5 (26.1–30.9) months for the CFS≤4 cohort (p<0.001).
Table III shows the Cox Proportional Hazards Model testing the effect of covariates on overall survival. On univariate analysis, CFS>4 (HR=4.26, 95% CI [2.54–7.15], p<0.001), ASA>2 (HR=2.62, 95% CI [1.15–6.00], p<0.05), age>80 (HR=2.99, 95% CI [1.98–4.51], p<0.001), MUST Score ≥2 (HR=1.95, 95% CI [1.08–3.51], p<0.05), albumin<32 g/L (HR=1.69, 95% CI [1.13–2.54], p<0.001) and haemoglobin <120 g/L (HR=1.60, 95% CI [1.13–2.54], p<0.05) were associated with decreased survival.
In the multivariate model, CFS>4 (HR=2.81, 95% CI [1.57–5.03], p<0.001), Age >80 (HR=2.08, 95% CI [1.35–3.21], p<0.001) and albumin <32 g/L (HR=1.94, 95% CI [1.21–3.09], p<0.01) remained significantly associated with decreased survival.
In subgroup analysis of patients with Age>80 (n=73), there was inferior mean (95% CI) survival in the CFS>4 vs. CFS≤4 subgroups (11.5 (8.3–14.7) months vs. 24.0 (16.1–31.8) months, p<0.01). This association was reproduced on a univariate cox proportional hazards model (HR=3.28, 95% CI [1.29–8.30], p<0.05).
Table IV shows the relative prognostic contributions of ASA and CFS categories to survival. The low numbers of patients in the CFS>4 and ASA≤2 subgroup (n=5) prevented meaningful analysis of this subgroup. Patients with CFS>4 and ASA>2 had significantly inferior mean survival compared to patients with CFS≤4 and ASA≤2 (p<0.001).
When ASA and CFS categories were combined into the CFS-ASA score, there were 28 (11.9%) patients with CFS-ASA=0, 81 (34.5%) patients with CFS-ASA=1, and 126 (53.6%) patients with CFS-ASA=2. Survival curves based on the combined score are shown in Figure 2. Mean (95% CI) survival in the CFS-ASA 0 vs. 1 vs. 2 was 30.8 (26.9–34.7) vs. 27.0 (24.1–29.9) vs. 15.9 (13.5–18.4) months (p<0.001). Increasing CFS-ASA score was associated with inferior survival on univariate (HROR=2.84, 95% CI [1.96–4.11], p<0.001) and multivariate (HROR=1.78, 95% CI [1.20–2.64], p<0.01) analyses. Goodness of fit on multivariate analyses was marginally better when the CFS-ASA score was considered vs. CFS alone (Wald statistics 8.11 vs. 7.46).
% 30-day survival (%SE) in the CFS>4 vs. CFS≤4 cohort was 85% (3%) vs. 97% (2%) (p<0.001), and % 1-year survival (%SE) in the CFS>4 vs. CFS≤4 cohort was 52% (4%) vs. 84% (4%) (p<0.001). CFS<4 was associated with increased odds of 30-day (OR=6.80, 95% CI [1.98–23.40], p<0.01) and 1-year (OR=4.73, 95% CI [2.54–8.80], p<0.001) survival.
The mean (SD) length of stay was 16 (±17.5) days. 70 (28.3%) patients were readmitted to hospital within the follow-up period, of which 43 (61.4%) were readmitted under vascular surgery. 64 (25.7%) received further vascular reintervention, 27 (42.2%) of which was major amputation. The secondary outcomes are shown in Table II. At time of follow-up, 10.8% (n=27) of our patients acquired COVID-19 infection during their follow-up period. Of these patients, 10 died, however we do not know whether this was primary cause of death.
Discussion
The present study describes the prognostic value of ASA, CFS, and a novel composite score in patients presenting with CLTI. The CFS-ASA score was superior to ASA alone in prognostication; however its performance was comparable to CFS assessment alone. The composite score and CFS score did not have significant different AUC values for predicting one-year survival; therefore we cannot conclude that the use of the combined score is superior to CFS score alone in clinical practice. This highlights the complex relationship between functional state, comorbidity, and prognosis in frail patients. Previous studies have described novel prognostic factors in vascular surgical patients, including assessment of cachexia and sarcopenia [16]. These parameters are restricted to the research setting at present, however may offer clinical utility in the form of additional prognostic value in frail patients with CLTI.
In our study population, frailty (CFS>4) is present in over half of the patients requiring unscheduled admission with CLTI. Age and multimorbidity are often associated with both frailty and CLTI individually, and the manifestations of CLTI, primarily pain, inhibit functional ability. Other studies have demonstrated high frailty prevalence in patients with CLTI and a systematic review and meta-analysis found frailty to be associated with age, female sex, low BMI and worse outcomes by multiple measures [5, 8]. We were able to support these findings with our results, however we did not observe a relationship between frailty and severity of CLTI.
In current UK clinical practice, the ASA score is a common method of assessing patients’ perioperative risk by a subjective estimate of organ system disease [11, 12]. Our results suggest that patients who fall into both CFS>4 and ASA>2 categories have the lowest survival rates. Despite this, our novel composite score incorporating both CFS and ASA offered only comparable prognostic value to CFS alone. However there is a reasonable disconcordance as most patients were ASA>3, whilst over half were CFS>4. Therefore CFS may provide adequate prognostic value, but a composite score may identify more patients who may benefit for pre-operative optimization. Further work to develop more robust methods of assessment may be required, to provide a more holistic view of a patient’s premorbid state.
The present study demonstrates that frailty is associated with poor clinical outcomes, in particular higher risk of postoperative MACE and all-cause mortality, independent of patient age. Due to their declining physiological reserve, these patients are more likely to suffer from postoperative complications and death as they are less likely to be able to respond to the stress response associated with acute illness and major surgery [5, 9]. As the surgical population is becoming older, this is an important finding that suggests that assessment of frailty is a powerful predictor of poor prognosis and that patients should not be assessed on age alone [17, 18]. Further, there was no statistical significance between frailty and amputation-free survival in our study population. Other studies have reported that frailty is not associated with major amputation, but strongly associated with mortality, with some studies reporting a 33% 2-year mortality rate [4, 5].
We suggest routine frailty screening in vascular surgical patients is warranted and should play a greater role in the consideration of management plans, not currently included in other scoring systems used including WiFI classification [8]. Use of a more patient-centred scoring system that includes frailty may allow more informed treatment choices for patients and may potentially mitigate high morbidity and mortality rates, and direct these patients to conservative/palliative treatment strategies that may have a significant impact on quality of life measures [19, 20]. Further, early involvement of perioperative frailty services will facilitate share decision making on these vulnerable patients [18].
Limitations
We acknowledge there are limitations in this single-centre study with a relatively short follow-up period. Our application of these subjective scoring systems to this specific patient group lacks external validation and is prone to bias. All patients over 65 years old underwent two assessments of frailty. Concordance between clerking CFS and frailty team CFS was not recorded as part of this study and is therefore a limitation, however the clerking team received specific training in implementation of CFS, helping to mitigate the potential risk of bias. Additionally, although the CFS is well validated, there are several other scoring systems regarding frailty [7], which may provide variable insight on its associations with worse outcomes and felt to be outwith the scope of this study.
Further, as an observational study, we cannot comment on how frailty team input has affected outcomes for these patients and we endeavour to examine this in future work. We also acknowledge the implicit bias on how patients are managed by vascular surgeons with knowledge of their frailty score.
The relatively small sample size resulted in low absolute numbers of participants in certain subgroups, limiting the power of the study to draw valid conclusions from subgroup analyses and therefore potentially introducing bias. As an observational study, a formal power and sample size calculation is not typically performed, therefore we chose to use the proxy of number of events (deaths) as a surrogate marker of the likelihood of avoiding a type 2 error.
We excluded any patients who presented on an outpatient basis and therefore will not have captured all patients with CLTI. It is likely that these patients are more likely to have worse outcomes solely on their emergency presentation, as supported by the National Vascular Registry [19, 21]. Our MACE outcomes included in-hospital cardiac arrest and we were unable to record any out-of-hospital cardiac arrest and therefore MACE may be underreported. Further, we used the Rutherford classification system over the Wound Ischaemia Foot Infection (WiFI) score for staging of CLTI [8]. Unfortunately we found the pandemic affected feasibility to obtain toe pressures due to restricted services and acknowledge this as a limitation.
The recruitment period was ongoing throughout the COVID-19 pandemic, therefore there may have been hesitancy from patients to seek medical attention, potentially delaying presentation and leading to more severe disease on admission. We recorded how many inpatients acquired COVID-19 infection, but unable to comment on outpatient rate of infection and its effect on outcome.
Conclusions
Frailty is highly prevalent in patients with CLTI admitted as an emergency and a poor prognostic indicator. Both frailty and ASA>2 are associated with reduced overall survival. We designed a combined CFS-ASA score to identify the patients most at risk, however this offers only comparable prognostic value to frailty assessment (CFS) alone and does not appear to be superior in the present study. The routine use of frailty assessment is rapidly becoming a necessity in vascular surgical patients as a method of prognostication and risk stratification. Novel assessments of frailty and comorbidity are required to optimise this multimodal assessment.
The authors would like to thank the Department of Vascular Surgery, the peri-operative Frailty Service, and Palliative Care Service at Ninewells Hospital (Dundee, UK) for assistance in identifying and helping these patients most at risk.
References
1 High prevalence of peripheral arterial disease and comorbidity in 6880 primary care patients: cross-sectional study. Atherosclerosis. 2004;172(1):95–105.
2 Global vascular guidelines on the management of chronic limb-threatening ischaemia. J Vasc Surg. 2019;69(6S):3S–125S.
3 Comparison of global estimates of prevalence and risk factors for peripheral artery disease in 2000 and 2010: a systematic review and analysis. Lancet. 2013;382(9901):1329–40.
4 The burden of critical limb ischaemia: a review of recent literature. Vasc Health Risk Manag. 2019;15:187–208.
5 Frailty factors and outcomes in vascular surgery patients: a systematic review and meta-analysis. Ann Surg. 2020;272(2):266–76.
6 . The frailty syndrome: definition and natural history. Clin Geriatr Med. 2011;27(1):1–15.
7 A global clinical measure of fitness and frailty in elderly people. CMAJ. 2005;173:489–95.
8 Association of clinical frailty with severity of limb threat and outcomes in chronic limb-threatening ischaemia. Ann Vasc Surg. 2021;000:1–11.
9 . Frailty and post-operative outcomes in older surgical patients: a systematic review. BMC Geriatr. 2016;16(1):157.
10 Association of preoperative patient frailty and operative stress with postoperative mortality. JAMA Surg. 2020;155(1):e194620.
11 . Analysis of the American Society of Anesthesiologists Physical Status Classification System and Caprini Risk Assessment Model in predicting venous thromboembolic outcomes in plastic surgery patients. Aesthet Surg J. 2016;36(4):497–505.
12 . ASA class is a reliable independent predictor of medical complications and mortality following surgery. Int J Surg. 2015;18:184–90.
13 . Improving care and support for people living with frailty in Scotland. Healthcare Improvement Scotland; May 11, 2022. Available from: https://ihub.scot/media/9183/improving-care-and-support-for-people-living-with-frailty-in-scotland.pdf https://ihub.scot/media/9183/improving-care-and-support-for-people-living-with-frailty-in-scotland.pdf
14 . ASA Physical status classification system. American Society of Anesthesiologists; December 13, 2020. Available from: https://www.asahq.org/standards-and-guidelines/asa-physical-status-classification-system
15 Outcomes of octogenarians in open lower extremity revascularisation. EJVES. 2019;58(6):e243–44.
16 . The prevalence of sarcopaenia in a vascular surgical patient cohort and its impact on outcome. Surgeon. 2018;16(6):325–32.
17 Impact of frailty on clinical outcomes in patients with critical limb ischaemia. Circ Cardiovasc Interv. 2018;11(7):e006778.
18 Frailty as a predictor of surgical outcomes in older patients. J Am Coll Surg. 2010;210(6):901–8.
19 . Chronic limb-threatening ischaemia and reframing the meaning of “end”. J Clin Nurs. 2021;30(5–6):687–700.
20 . Non-operative treatment of advanced limb ischaemia: the decision for palliative care. Eur J Vasc Endovasc Surg. 2010;19(3):246–9.
21 . National Vascular Registry: 2021 Report. April 2022.
Appendix
