Seeing the entire elephant: The challenges of frailty assessment for peripheral artery disease

Commentary on: Al-Damlugi et al. Frailty and outcomes following revascularization of lower-extremity peripheral artery disease: Insights from the Vascular Quality Initiative (VQI). Vasc Med. 2022; 27: 252-258.

According to an ancient Indian parable, a group of blind men were asked to imagine the appearance of an elephant solely through touch. ¹ Each blind man felt a different part of the elephant’s body and came to a different conclusion about the elephant’s overall physique, ultimately resulting in mistrust of the opinions of the other men. This story illustrates the folly of claiming absolute truth based on one’s subjective experience; the story also provides a useful analogy for framing the discussion about a complex clinical disorder such as frailty.

Frailty can be defined as ‘a state of increased vulnerability and reduced ability to maintain homeostasis after a stressful event resulting from impairment in multiple physiologic systems’. ² Over 20 different scales for frailty exist, ³ in part reflecting the complexity of defining this clinical syndrome. Frailty as a construct has been consistently associated with increased risk of short and long-term mortality, prolonged hospitalization, and poor quality of life.^{4 –7} This relationship has been present across different frailty measures, and disease states, as well as being independent of clinical risk factors.^{4 –6} As a closer approximation to one’s biologic age, ² frailty has been shown to have a stronger association with risk of adverse outcomes than chronologic age in several studies.^8,9 And yet, despite the wealth of data illustrating frailty’s powerful role in promoting disease and prognostication, formal frailty assessment remains underutilized in clinical practice, ¹⁰ and one’s frailty status is often not captured in clinical registries or used for risk adjustment. ⁵ To compound matters, some studies have suggested that efforts to reverse frailty, such as prerehabilitation prior to invasive procedures, may not prevent these adverse outcomes from occurring.^11,12 If so, then how can we as clinicians best identify and advocate for our frail patients? Are we only encountering part of the elephant in our efforts to see the whole?

As a physician’s subjective assessment of a patient’s frailty status does not predict postprocedural risk, ¹³ several formalized, well-established frailty tools have been developed to capture the relevant prognostic information, which can be broadly encompassed under two frameworks: the Rockwood definition, which considers frailty as an accumulation of deficits over time, and the Fried definition, which considers frailty as a biologic syndrome marked by impairment in five domains – shrinking (i.e., weight loss), exhaustion, weakness, slowness, and low physical activity. ¹⁴ Nevertheless, which of these frailty scales best predicts clinical risk may differ based on the given disease state and the procedure under consideration. ⁴

In this issue of Vascular Medicine, Al-Damluji and colleagues highlight the importance of selecting an appropriate frailty scale for peripheral artery disease (PAD) in an analysis of the Vascular Quality Initiative (VQI) national database. ¹⁵ The VQI is a multicenter registry that collects demographic, clinical, procedural, and outcome variables for patients undergoing vascular procedures, with approximately 600 participating sites, distributed across 18 regional vascular groups throughout the United States. From this cohort, the authors identified those undergoing peripheral revascularization (peripheral vascular intervention (PVI) or surgical bypass) for lower-extremity PAD between 2010 and 2019. Patients with acute limb ischemia or nonhealing amputation were excluded.

Frailty was assessed in this population using the five-item modified frailty index (mFI-5) and the VQI-derived risk analysis index (RAI), both of which are based on a Rockwood framework. The mFI-5 was derived from an original 16 variables cross-walked from the Canadian Study of Health and Aging Frailty Index, including hypertension, diabetes, heart failure, chronic obstructive pulmonary disease, and functional status deficit. The VQI-RAI represents an adaptation of the RAI, a scale validated among veterans, including information on age, sex, cancer, unintentional weight loss, poor appetite, renal and heart failure, dyspnea, dependence in activities of daily living, and cognitive deterioration. The VQI-adapted RAI uses body mass index to replace the nutritional domain (unintentional weight loss and poor appetite) and replaces dyspnea with presence of chronic obstructive pulmonary disease. Moreover, as the VQI was lacking in information on cognitive impairment and cancer, these two variables were not included in the VQI-RAI.

Using these two scales, the authors analyzed the outcomes of 135,916 individuals undergoing PVI and 41,561 patients undergoing lower-extremity bypass. They identified frailty in 69% of individuals by the mFI-5 and 16% by the RAI, with overall poor agreement between scales (kappa statistics 0.13–0.17). In adjusted analysis, both scales were associated with a higher mortality risk regardless of revascularization type. However, only those deemed frail by the VQI-RAI had higher odds (1.5-fold risk) of unplanned amputations in the bypass cohort. Although the addition of a frailty index to traditional PAD risk factors improved prediction of mortality and unplanned amputations, the degree of risk reclassification was marginal.

This report is an important step forward in measuring frailty in the PAD population, but the mixed results are not surprising. As the mFI-5 includes several conditions that are commonplace in PAD (e.g., hypertension, diabetes, chronic obstructive pulmonary disease, and heart failure), it is foreseeable that more than two-thirds of the cohort was deemed frail by this definition, thus resulting in marginal incremental benefit in risk prediction. Moreover, in the VQI adaptation of the RAI, important risk variables such as weight loss, cancer, and cognitive status (which encompasses cognitive frailty) were not included. Finally, as outcomes were assessed in-hospital, they may be impacted by length of stay, and may not reflect 30-day periprocedural outcomes.

Despite these limitations, it remains remarkable that these two frailty scales nevertheless were highly predictive of mortality. Thus, rather than highlighting the deficiencies of frailty as a concept for risk prediction in PAD, these results additionally argue for further research to develop specific frailty assessments, derived and tested among patients with PAD, that could potentially augment risk prediction and improve on the results observed with such generic instruments. Moreover, they broadly support a role for frailty assessment in the PAD population using scales that are anchored to well-established frailty conceptualizations to understand the individual factors of relevance in driving risk in this population. For example, urinary incontinence may be associated with frailty and be highly predictive of risk in one population, but it may predict poorly in another. We have previously shown that frailty, measured under the Fried definition, is an important and independent predictor of revascularization strategy and longitudinal outcomes in Medicare patients hospitalized for chronic limb-threatening ischemia. ⁴ As frailty is an important source of traditionally unmeasured risk, a PAD-specific frailty instrument could improve prognostication, identify individuals for nutritional support and intensive rehabilitation, and even potentially improve decision-making on the optimal revascularization strategy for a given individual. ⁴ Moreover, identifying specific conditions that drive risk in frail patients with PAD could potentially help guide interventions to improve preprocedural risk. ⁴

Although frailty can be measured in more than 20 ways, the first step towards using it to improve outcomes is to measure it in any way. The current study by Al-Damluji et al. provides the roadmap for doing so in PAD. Though not all parts of the elephant are equivalent, finding the ones that matter the most in PAD may be the solution to improving short- and long-term outcomes in this high-risk population.