LOS ANGELES – A modified frailty index strongly predicts morbidity and mortality after esophagectomy, based on a study of more than 2,000 patients in a national database.
Both the rate of serious complications and the rate of death rose steadily and significantly with the index, which is derived from 11 common clinical variables.
Compared with their counterparts having the lowest index, patients having the highest index were 31 times more likely to die in the first postoperative month even after confounders were taken into account, according to data reported at the annual meeting of the Society of Thoracic Surgeons (STS).
"The modified frailty index appears to be an important predictor of morbidity and mortality. It can be used to guide surgeons in making more accurate predictions of complications. It also provides an objective guide for patients in decision making," commented lead investigator Dr. Arielle Hodari, a thoracic surgery resident at the Henry Ford Hospital in Detroit.
In the study, the investigators calculated a preoperative modified frailty index for 2,095 patients who underwent esophagectomy between 2005 and 2010. The index contains 11 variables chosen by mapping those used in the Canadian Study of Health and Aging Frailty Index (CMAJ 2005;173:489-95) onto those available in the National Surgical Quality Improvement Program (NSQIP) database.
The 11 variables in the index were (1) nonindependent functional status, (2) diabetes, (3) chronic obstructive pulmonary disease or pneumonia, (4) congestive heart failure, (5) myocardial infarction, (6) percutaneous coronary intervention, cardiac surgery, or angina, (7) hypertension, (8) peripheral vascular disease, (9) impaired sensorium, (10) transient ischemic attack or stroke without residual deficit, and (11) stroke with deficit.
The study patients had indexes ranging from a low of 0 (meaning 0 of 11 variables) to a high of 0.45 (5 of 11 variables). The large majority clustered at the lower end of the spectrum; only about 1% had an index of 0.45.
The main results showed that as the index increased from 0 to 0.45, there was also an increase in the rate of Clavien 4 complications from 18% to 62% (P less than .001), which are life-threatening complications leading to ICU admission, as well as an increase in the rate of 30-day mortality from 2% to 23% (P less than .001).
The findings were similar for respiratory complications and for cardiovascular complications individually, according to Dr. Hodari.
In a multivariate analysis, patients with the highest index had sharply elevated odds of death relative to their counterparts with the lowest index (odds ratio, 31.8; P = .015).
"The NSQIP data does not differentiate between pathology (squamous cell carcinoma vs. adenocarcinoma) or surgical modality," Dr. Hodari commented, discussing the study’s limitations. "We also need validation in a prospective study."
During the discussion following the presentation of the data, Dr. Hodari was asked about the number of patients who had received neoadjuvant/induction therapy before coming to surgery.
"What I find in my own personal practice is that it’s hard sometimes to say no [to surgery] altogether. So if somebody is T3 or N1, I’ll let them get some induction therapy and then get them back to reevaluate them," he said. "I’d be particularly interested in this type of analysis being applied then, because a lot of times [the patients] haven’t really been improved by that process, and maybe it just shows that I’ve procrastinated in making a difficult decision and probably made it even harder by putting it off until after their induction therapy. Do you know anything about the status of induction therapy in these patients, whether the frailness that was measured was in part due to the induction therapy and how that might affect your results?"
"Unfortunately, that is not available in NSQIP, so we don’t know which patients had neoadjuvant treatment," Dr. Hodari replied.
Regarding questions about the use of 30-day mortality as an endpoint, he responded, "I think it’s worthwhile – maybe it’s not in the database – to look at these figures at 90 days, because, of course, the high-risk patients die within 30 days. But there is a big increase in mortality, almost doubling, at 90 days."
An attendee noted that the original frailty index includes unintentional weight loss, but the modified one does not. "Why was that done for this particular patient population, because with esophageal cancer and dysphagia and unintentional weight loss being a big component of the comorbidity these patients have, wouldn’t you think that would be important to include that?"
That choice was made on the basis of data constraints, Dr. Hodari replied. "We chose variables by going through NSQIP and finding something that corresponds with the Canadian system," she explained.