Frailty Before Cardiac Surgery Common, Linked to Adverse Outcomes

(UPDATED) The prevalence of frailty prior to cardiac surgery is high, varying widely depending on the tool used to capture it, and it’s significantly associated with a higher risk of patient-defined adverse cardiovascular and noncardiovascular events (PACE), according to the results of a new study.

Of three frailty tools tested, the Hospital Frailty Risk Score (HFRS) appeared to perform the best, with the highest area under the receiver operating characteristic (ROC) curve for estimating the risk of PACE during follow-up.

Given the new data, researchers say it’s time to start thinking about frailty before cardiac patients head into the operating room.

“We recommend regularly incorporating a frailty assessment in the preoperative visit,” lead investigator Louise Sun, MD, SM (University of Ottawa Heart Institute, Canada), told TCTMD. “Frailty is independently associated with poor short- and long-term outcomes, especially after cardiac surgery, which is a much higher-risk population than those undergoing noncardiac surgery. Frailty as a comorbid condition certainly elevates the risk of adverse outcomes and procedure failure.”

Although frailty is a well-known determinant of patient health and outcomes after medical procedures, less is known about how frailty can be incorporated into resource planning and risk stratification, said Sun. For example, neither the STS nor EuroSCORE risk calculators, which calculate the predicted morbidity and mortality associated with various cardiac surgeries, include frailty as part of their equations. A preoperative frailty assessment, however, has been shown to improve risk stratification when added to traditional scores.

Surgeon Mark Ferguson, MD (University of Chicago, IL), who wasn’t involved in the analysis, said they routinely screen for frailty prior to cardiothoracic surgery in patients 50 years and older and that frail men tend to outnumber frail women (unlike in the present analysis). At their center, the incidence of frailty is close to 70%, which is in line with prevalence detected using the HFRS.

In thoracic surgery, as opposed to cardiac surgery, frail patients theoretically have the ability to undergo prehabilitation because urgency is less of a factor. Moreover, patients undergoing thoracic surgery aren’t limited by severe heart disease symptoms. “It’s been shown that prehab is quite valuable in terms of reducing the risk of complications of surgery, particularly after thoracic operations such as a lung resection, and to some extent, an esophagectomy,” Ferguson told TCTMD. “It’s an important investment in time and energy on the part of the patient.”

If frailty is diagnosed in a patient with the severe heart disease, the advantages of prehabilitation are less well defined than in those undergoing a more elective operation, he said.   

Three Different Frailty Scores

To TCTMD, Sun said there are a number of different frailty indices, some that require patient visits and others that are based on administrative data. Given the abundance of tools available, the researchers wanted to assess the performance of three different frailty instruments that rely on routinely collected patient data. 

Published September 9, 2022, online in JAMA Network Open, the analysis included 88,456 patients (mean age 66.3 years; 25.9% female) who underwent index CABG; aortic, mitral, or tricuspid valve surgery; or thoracic aorta surgery between 2008 and 2017 in Ontario, Canada. As part of the study, frailty was assessed with the HFRS, the Johns Hopkins’ Adjusted Clinical Groups (ACG) frailty indicator, and the Preoperative Frailty Index (PFI). All three are multidimensional tools that capture elements of frailty and denote a susceptibility to adverse outcomes in situations of stress, such as cardiac surgery, according to the researchers. The ACG indicator is a dichotomous frailty score (yes/no), whereas the HFRS and PFI are continuous frailty measures.

Based on the ACG, HFRS, and PFI, there was some evidence of frailty seen in 16.9%, 71.3%, and 86.8% of patients, respectively. These patients tended to be older and female; were more likely to live in rural areas; had lower incomes, LVEF, and body mass index; had a higher comorbidity burden; and were more likely to require emergent surgery. 

Over a median follow-up duration of 6.2 years (maximum 11.5 years), 22.7% of patients who met the AGC criteria for frailty developed PACE—the primary composite endpoint that included stroke requiring hospitalization for at least 14 days or inpatient rehabilitation, new-onset dialysis, heart failure, admission to long-term care, or ventilator dependence—versus 15.2% of those who were not frail. Similarly, the HFRS and PFI scores were significantly higher in those with PACE during follow-up than in those without frailty.

The risk of PACE was 10% and 75% higher with each 0.1-point increment in the HFRS and PFI scores, while patients had a 66% higher risk of PACE if defined as frail with the ACG. Mortality also was significantly more common in frail patients, regardless of the score used to assess it, than in nonfrail patients.

The HFRS had the highest area under the ROC for estimating PACE in the first 2 years and mortality in the 4 years after surgery. Beyond 4 years, however, the PFI had the highest area under the ROC. The ACG frailty score didn’t perform very well overall, said Sun, noting it was only slightly better than a coin toss for estimating PACE over any time point.

“We looked at whether there is a time dependency to the ability of each of these instruments to estimate long-term patient-centered outcomes,” said Sun. “Some of these instruments are probably better at estimating perioperative and short-term outcomes whereas some of the instruments are better at estimating long-term outcomes.”

Given these findings, the different frailty scores could be “tailored to specific outcomes and follow-up durations to better inform patient-centered decision-making, preoperative optimization, and health resource planning,” they state.

“You want to make sure you’re using the right instrument with the highest predictive ability for [however] far out into the future that you want to predict,” said Sun. One of the strengths of the study, she added, is the use of the validated PACE endpoint, which consists of clinical events important to patients.

Ferguson pointed out that the frailty instruments selected—ACG, HFRS, and PFI—are those ideally suited for retrospective analyses because they rely on administrative information contained in electronic medical records or databases. While the HFRS was shown to be the best tool in the first few years, with PFI deemed better beyond 4 years, neither of the frailty scores performed particularly well as assessed by the area under the ROC, said Ferguson. In terms of individuals components of PACE, no frailty tool estimated the risk of HF well (area under the ROC ranging from 0.52-0.57), but the HFRS was relatively good for predicting dialysis and admission to long-term care (0.79 and 0.72, respectively), he said.

Ferguson commended the investigators for the large analysis and extensive follow-up, but pointed out that those strengths can also bring challenges in that it can be easier to show statistical significance with selected measurements and clinical outcomes. “That can translate to good predictability for large populations, but poor predictability for individual patients,” he said.

Clinical Utility of Frailty Scores

In an editorial accompanying the study, Jeffrey Jackson, MD, MPH, and Cory Ganshert, DO (Clement J. Zablocki VA Medical Center, Milwaukee, WI), say the study makes a “strong case for incorporating frailty into preoperative risk stratification for cardiovascular surgery.” While there is a lot of promise with including frailty as part of that assessment, there are still some problems to be worked out. For example, most of the adverse events were not predicted by any of the three risk scores, which is the challenge when trying to predict rare events.

“While adverse outcomes were more common in those with higher scores, most adverse outcomes occurred in those without elevated risk,” write Jackson and Ganshert. “It is difficult to find the needle in the haystack.”

A second problem is that both the HFRS and PFI labeled most patients as frail, “which suggests that these two measures are capturing more than just frailty,” they add. “What’s unclear is what the operative team should do with this information. Whether interventions, either with the patient prior to surgery or by the operative team to improve outcomes, is unknown and worthy of study.” 

At their center, physicians historically used the FRAIL scale and Physical Frailty Phenotype, the latter also developed at Johns Hopkins, to assess patient frailty, said Ferguson.