Is a subjective assessment of functional capacity predictive of perioperative complications?

Welcome to Depth of Anesthesia. I'm David Hao and I'm an anesthesia resident at the Massachusetts General Hospital. This episode will be featured on anesthesiologynews.com. Anesthesiology News recently launched a new podcast series called The Etherist that takes a deep dive into the story behind the ongoing drug shortage crisis You can listen on iTunes or wherever you get your podcast. Just search for anesthesiology news.

My guest today is Dr. Alyssa Walsh. Dr. Walsh is one of my co-residents here in the Department of Anesthesia, Critical Care and Pain Medicine. She presented a highly thought-provoking talk at our journal club session a few weeks ago, and I thought that this would be a great topic for depth of anesthesia. She's joined by Dr. Lori Shapiro, the Medical Director of the Pre-anesthetic Testing Area here at the Massachusetts General Hospital. Dr. Walsh and Dr. Shapiro, welcome to the show.

Thank you. Thank you. So our case today is that of a 75-year-old man with a history of hypertension. hyperlipidemia, insulin dependent diabetes, chronic kidney disease, and a newly discovered pancreatic mass, who's presenting for a Whipple procedure. He has never seen a cardiologist before On the day of surgery, the anesthesiologist asks the patient if he can climb two flights of stairs without getting short of breath. The patient shrugs because he also has a bad knee, so he has to take the stairs slowly The anesthesiologist marks that the patient is able to achieve four metabolic equivalents of activity and thus can proceed to surgery safely.

Dr. Walsh, what do you think are some of the claims in this case? So David, it seems like the claims are as follows. Number one, that a subjective assessment is an accurate way to measure functional capacity Number two, the self-reported ability to climb two flights of stairs is the best subjective method to assess functional capacity. And number three, that exercise tolerance greater than or equal to four metabolic equivalents predicts the risk of postoperative complications in any major non-cardiac surgery To our listeners, what's the level of your agreement with the primary claim that a subjective assessment is an accurate metric of functional capacity? And what is the level of evidence for what you believe?

Before we get started, we're going to review the results of a recent Twitter poll that we conducted. We asked, what is your primary method of assessing pre-operative functional capacity? 53% said two flights of stairs. 44 % said formats and only 3% said stress testing. Any thoughts? So, this is about what I expected. At our hospital, our current practice is usually just that.

We have a pair of subjective assessments. Each patient who's presenting for surgery will receive a multiple choice question regarding their exercise tolerance, ranging from the spectrum of short of breath at rest, or unable to exercise, to able to run two miles or more, as well as their ability to lie flat This is combined with a subjective assessment by the anesthesiologist on the day of surgery. The value that we're trying to ascertain with those questions is the maximum number of metabolic equivalents that a patient can achieve before symptomatic limitation. Excluding Emergency Procedures and Acute Coronary Syndrome, the 2014 AHAAC Guidelines recommend that any patient with cardiac risk factors presenting for major non-cardiac surgery should undergo an assessment of their exercise tolerance.

If the patient has a high functional capacity, then they can defer further cardiac workup prior to the operation. The threshold they give is a functional capacity of four metabolic equivalents. So I definitely hear metabolic equivalents thrown around a lot, right? So METS. So a metabolic equivalent, or MET, corresponds to the basal metabolic rate or the amount of oxygen consumed by a 40-year-old 70-kilo man sitting at rest.

This is about 3.5 milliliters of oxygen per kilo per minute, thus 4 metz is about 14 milliliters of oxygen per kilo per minute. There actually exists a compendium of physical activities, which is updated every decade, and will give you the approximate METS awarded for any given activity. It is excruciatingly detailed and not particularly scientific. The AHA ACC guidelines specifically use climbing two flights of stairs walking on level ground at four miles per hour, and performing heavy work around the house as all being equivalent to four mets. It's interesting that we hear about four mets all the time.

Where does that even come from? The funny thing is no one seems to know. I asked the author of the Meth study that we'll describe later in this podcast, as well as several cardiologists. It started appearing in the preoperative assessment literature in the 1980s and 1990s. The first one we found was Carliner et al. in 1985, who did a prospective study of two hundred patients scheduled for elective major noncardiac surgery who underwent exercise testing preoperatively.

They actually defined poor functional capacity as less than five metz They found that only six of two hundred patients, or three percent of their entire patient population, had death or myocardial infarction, but five out of six of those patients had less than five MATS And so they concluded that a functional capacity of less than 5 mets had a low sensitivity but high specificity for adverse perioperative events So knowing that this is the case, are we even good at assessing where that threshold is if we don't do formal exercise testing? That's a great question. So in 1999, Riley et al. published an article in JAMA Internal Medicine looking at the relationship between self-reported exercise tolerance and serious perioperative complications in six hundred patients undergoing major non cardiac surgery.

They assessed functional capacity by asking the patients to estimate the number of blocks that they could walk, or the stairs that they could climb, before becoming symptomatic. They chose these metrics because they were simple and familiar to the average clinician. They then looked at the sensitivity, specificity, and likelihood ratios for predicting perioperative complications using various combinations of number of blocks walked and stairs climbed. They ended up defining poor exercise tolerance as inability to walk four blocks or to climb two flights of stairs, because it provided the highest sensitivity, seventy one percent, with a lower specificity of forty seven percent, and a likelihood ratio of one point three, for predicting serious post op complications. They ended up finding that 20.4 % of patients who reported poor exercise tolerance by these metrics had perioperative complications, compared to 10.4 % of those who did not have a poor functional capacity.

Note that this is an incredibly high rate of complications in either functional capacity group. That alone may be a marker of how far we've come in two decades in perioperative medicine. and we question whether this is still applicable to contemporary institutions. A number of studies began to use either formats or climbing two flights of stairs as the ideal metric of functional capacity based on this study. This also became the standard of clinical practice. Unfortunately, the subsequent data, frankly, has been fairly poor It turns out that the subjective assessment of METS neither correlates with quantitative measurements of exercise capacity at the VO2 max, nor consistently predicts postoperative complications.

In the METS study in 2018, only 16% of patients with a VO2 max corresponding to under-4 METS was actually recognized as having a poor functional capacity. To me, it's not enough to just learn the evidence. Learning the limitations of the evidence is just as important, if not more important, and the evidence for all this is pretty limited I really love that because I think that reflects what we're trying to do here with the podcast is not just identify evidence, but to evaluate it critically and talk about some of its limitations. So it sounds like with functional capacity there might be a need for a different approach. What are some other measures that we have available?

So beyond the subjective assessment, the methods we have available to assess functional capacity are objective physiologic measurements and more structured self-reported assessments. There may also be objective serum biomarkers for predicting postoperative risk. Which is essentially what you're trying to get at when you're asking about the functional capacity. And you mentioned VO2 Max earlier. What is that?

VO2 max is peak oxygen consumption, or the maximum rate of O2 consumption during incremental exercise. It's equivalent to the product of cardiac output times the difference in arterial and venous oxygen content at peak exercise. Normal VO2 max is about 27 to 31 milliliters of O2 per kilo per minute in women and 35 to 40 milliliters of O2 per kilo per minute in men. It can also be translated into Mz units, where again, 4 metz is about equivalent to 14 ml of oxygen per kilo per minute. So if I wanted to measure the VO2max for one of my patients, how would I even go about getting that done?

VO2 max is calculated using cardiopulmonary exercise testing, or CPAT. which essentially involves placing your patient on an ergometer and strapping on a Baynesque mask to their face to measure real-time minute ventilation oxygen uptake, CO two production, saturation, heart rate, blood pressure, and EKG. You attain baseline spirometry and hemodynamics, then send the patient through a standardized exercise regimen where the resistance is increased for the same speed, every minute. The test will end when the patient says they cannot go any further or has any safety concerns such as arrhythmia, malignant hypertension, or hypotension. So that sounds like quite a process. Is there any easier way to get that kind of information?

Currently, CPED is the only proven technique. But interestingly, several contemporary smartwatches also have estimated measurements of the VO2 max. The algorithm is of course proprietary. but is based on anthropometric data such as age, height, etc. According to the website for Garmin, the accuracy of their method when applied for running is reported as 95% accurate. Based on a database of 2,690 freely performed runs from 79 runners whose VO2 Max was tested four times during their six to nine month preparation period for a marathon Obviously, there are some caveats to this. They didn't have this data published in a rigorous journal, but also they're heavily enriching for a patient population that does not exactly approximate the typical patient presenting for major non-cardiac surgery.

Still, it's an interesting thought for the future. It definitely is. What about alternative self-reported assessments Because when I ran the Twitter poll, Kenneth Cummings actually posted a reply asking, how about DASI I was so excited to see that comment. So the Duke Activity Status Index, or the Dossi questionnaire, was actually mentioned in the 2014 AHA ACC guidelines is an alternative to the traditional subjective assessment. It was first published by Flatkey et al. in 1989 in the American Journal of Cardiology.

The authors took 50 patients undergoing exercise testing and asked them about their ability to perform a variety of common tests They then developed a 12-item scale, the DASI, that correlated well with peak oxygen uptake with a correlation coefficient of 0.8. They then tested their new scale in an independent group of fifty patients, who took the test and then underwent exercise testing as well. This time the dossier correlated significantly with peak oxygen uptake with a moderate coefficient of 0.58. The questionnaire includes many of the questions we often ask patients in regards to stairs climbed or blocks walked, but also incorporates activities of daily living, chores, yard work, sports, and even sexual activity Each action is given a weighted point score and the total ranges from zero to fifty-eight point two.

So I think that this questionnaire is really exciting because our traditional subjective assessment is often asked in a very charitable, socially acceptable sort of way, asking, oh, I hate to embarrass you, but do you think you could do a flight of stairs or two? Or do you think you could walk a couple blocks? The DASI questionnaire gives people multiple opportunities to confront their own psychological defenses about how their health and fitness may not be what it used to be, and thus they may be at an elevated risk for surgery. Of note, the DOSI does only go to 9.8 mets total, even with a full score of 58.2 I think it's a really great point because I often find myself having the same kind of issues when I talk to patients.

We often ask leading type questions and so I always try to ask questions like what's the most active thing you'd have done in the past week to try to get around that. So what about biomarkers for predicting increased perioperative risk? The hot topic has been N-terminal pro D type naturic peptide, or easier to say, NT pro BMP. Its release is triggered by the RAAS system activation, leading to increased ventricular preload and wall stretch, whereas BMP leads to naturesis and vasodilation NT ProBMP is actually an inactive substance. I think many of us are familiar with its use in identifying patients in acute congestive heart failure.

However, Reading It All published a meta-analysis in 2009 in anesthesiology that looked at 15 prospective observational studies for the BMP or anti -proBMP Have been measured preoperatively in patients undergoing major non-cardiac surgery. They found that short-term MACE, or major adverse cardiac events, occurred in 283 of 862 or 32.8% of patients with elevated BMP compared to 119 of 3049 or 3.9% of patients who did not. This corresponds to an odds ratio of 19.77. When the authors broke it down by BMP versus anti -bro BMP The result was similar with an odds ratio of 25.45 for BMP and 15.65 for NT ProBMP. Only two studies reported longer-term MACE data were in MACE occurred in 76 out of 202, or 37.6% of patients with an elevated anti-proBMP. compared to eighteen of three hundred fifty one in those who did not.

This corresponds to an odds ratio of seventeen point seven. There certainly are limitations. There were other differences between studies in terms of demographics, for example. that may have had a confounding effect. Additionally, no study actually defined an optimal threshold value for BMP or NT ProBMP to use in practice Still, it's remarkably consistent. The authors state the precise reason for the strong association between BMP and postoperative outcome is not addressed by this analysis, but it's likely that BMP elevation identifies patients with impaired cardiac function or a significant ischemic burden who may not be able to withstand the hemodynamic and pro-inflammatory stresses of general anesthesia and major surgery.

That's really interesting. And so now we've discussed a lot of material, I think, about different ways to assess functional capacity or to predict post-operative risk. Can we just summarize some of the main points before we move ahead? Absolutely. So beyond our traditional subjective assessment, you can perform quantitative measurements such as the VO2 Max, or a more structured, detailed subjective assessment such as the DASI.

You can also do serum biomarker testing such as antipropium P as a potential marker for increased perioperative risk So I think the big question now is have these been compared head-to-head? So we mentioned the MET study briefly before. The MET study, or measurement of exercise tolerance before surgery, a great acronym. was published by Oiju Sidera et al. in the Lancet in twenty eighteen. Essentially it addressed exactly this question, comparing different methods of assessing functional capacity before major non cardiac surgery. This was a multi-center international trial of a total of 25 hospitals in Canada, the UK, Australia, and New Zealand between 2013 and 2016.

It was designed as a prospective cohort study. Patients were aged forty years or greater scheduled for elective major non cardiac surgery, under general or regional anesthesia, or both, staying in the hospital at least one night and had to carry at least one risk factor for cardiac complications or coronary artery disease. Any notable exclusion criteria for this study? Overall, they were reasonable Presence of an ICD was one of the exclusion criteria, and it makes me wonder as to whether they excluded a population that by definition was very sick and could have changed the statistics. Yeah, I think that's a fair point.

So how did they actually run this study? So in terms of how they ran this study They would start off with a visit with the patient where they would do a subjective assessment in the traditional fashion and also acquire consent for the study. Sometime prior to surgery, the patients would undergo CPET testing, as well as have a measured NT ProBMP. On post op day zero, the patients would undergo both a DOSI questionnaire as well as the traditional subjective assessment. Subsequently, between post-op day zero and post-op day three, or discharge, whichever came first, the patients would have troponins, creatinine and an electrocardiogram measured in house every day.

Subsequently, at thirty days postoperatively, the patients were assessed for complications, and at one year the patients were also assessed for any complications. Of note, patients providers and the outcome adjudicors were blinded to the CPED and NTPropMP data The providers, outcome educators, were blinded to Dossi scores, but obviously the patients were not. So what were some of the primary and secondary outcomes of this study? The primary outcome was death or myocardial infarction within 30 days after the surgery. Secondary outcome was death within one year after surgery Other outcomes measured were death or myocardial injury within thirty days after surgery, and moderate or severe, including fatal complications, during the index admission.

So stuff we definitely care about. Absolutely. A word quickly on their sample size and outcomes in this study. They began the study with a goal of 1180 patients to achieve 90% statistical power. assuming that the event rate of their primary outcomes was about 8%, among other factors. Unfortunately, halfway through the event rate was actually only about 5%. and they found that only about twenty percent of patients actually made it to CPAD or surgery.

Thus, they updated their goal number to account for those patients who had failed to qualify for the primary outcome, based on the actual event rate. But took a hit per s took a hit for statistical power to eighty percent. So what's the takeaway from that? The takeaway from that is that they may have a higher rate of false results. Got it.

So what are some of the key takeaways of this study? This pragmatic study achieved a relatively realistic major non cardiac surgery population. Fifteen percent had only regional anesthesia, twenty five percent ended up in the ICU or similarly monitored unit. A significant number of patients required early termination of CPAT or had adverse events, 10%. Mean VO2 max was 19.2 ccs per kilo per minute, 6.5 standard deviation.

As a reminder, a VO2 max of 14 correlates to about a maps of 4 And a normal VO2 max for women is about 27 to 31, and for men it's 35 to 40. So in general, these patients did have a reduced functional capacity based on VO2 max. So when they looked at the subjective assessment versus the DOSI score, generally speaking, both were lower in patients with confirmed poor functional capacity However, as we mentioned before, only 16% of patients with a VO2 max under 14 were subjectively identified as having a poor functional capacity. This corresponds to a sensitivity of nineteen point two percent and a specificity of ninety-four point seven percent. The authors also saw that there was a moderate but significant correlation of the DOSI score with the VO2 max with a correlation coefficient of 0.43 and a weak but significant correlation of the serum NT Probium P with the VO2 max, with a negative correlation coefficient of negative 0.21 Awesome.

So now we're going to head into the part of the study where we actually look at the different measures of pre-operative functional capacity and how they related to some of the outcomes. Excellent. So what the authors saw was that the DOSI score alone significantly improved prediction of 30-day death or myocardial infarction, their primary outcome. And that NT probium P was the only factor that significantly improved the prediction of one-year death, their secondary outcome In terms of their other outcomes, both the DOSI score and the NT ProbiMP significantly improved prediction of death or myocardial injury at 30 days. And VO2 max was the only two was the only factor that significantly improved prediction of moderate or severe postoperative complications.

Alright, so what's the punchline for this study? The punchline, pre-operative subjective assessment, neither accurately identified patients with poor fitness, nor predicted postoperative morbidity or mortality The DASI Questionnaire and NT Pro BMP predicted perioperative cardiac risk. VO2 Max predicted postoperative complications, but not perioperative cardiac risk So taking into context all the things we just talked about with regards to this study, how can we start to operationalize that if we are to apply some of the evidence that we just discussed here? So a theoretical way that we could use some of these findings is that a patient with a high DOSI score and a low preop NT ProBMP might be somebody that we think about as a candidate for an earlier discharge. and maybe placed in a lower acuity unit, whereas somebody with a reverse, a low DOSI score, and a high pre -op NT ProBMP Maybe somebody for whom we think they may need to be monitored in the hospital overnight, or even require ICU monitoring, and we have an increased index of suspicion for perioperative cardiac events.

Got it. What are some of the limitations of this study? Primary and secondary outcomes were both rarer than anticipated, which limited statistical power. There was a low consent rate among otherwise eligible patients, about 27%. CPEP measurements and troponin assays were determined by individual sites.

And there's definitely a need for a more rapid or easily applied risk stratification tool. It's worth noting as well that this was an international study that did not include U. S. hospitals. The culture in the US is that it's a fee-for-service model, and as a result, we do often offer surgery to patients with a higher comorbidity burden compared to other countries This study may have missed some of the patients that we might see in the US, for example, because they may not have even been offered surgery as they'd already been assessed as being higher risk. So what are some of the future directions we're heading in with respect to pre-operative testing?

I think there's a lot of study work that um this study invites. For example defining optimal risk thresholds for DOSI scores for NT Pro BMP levels. I think there's room for exploring novel markers of perioperative cardiac risk, for example, exaggerated exercise mediated heart rate responses. And use of data from non-invasive monitoring in the perioperative period, including personal fitness devices, is certainly something that can be and will be explored in the future So ultimately at our own institution, we're hoping to move forward with codifying the DOSI questionnaire as our default subjective questionnaire. And we're considering the utility of NT ProBMP as a routine preoperative marker to predict postoperative complications It's certainly not financially or logistically feasible for everyone to undergo CPET testing, but there may be some roads ahead with non-invasive estimates of VO2 max and anaerobic threshold.

So I would say that based on the evidence that we discussed over the course of this show, there are concerns with the current ways that we do things, and I think there are ways that we can get better As anesthesiologists and perioperative physicians, in terms of assessing perioperative risk, I think it's a really exciting time to be part of this. The data that supports perioperative risk assessment is spotty and dated And I think that moving ahead with the DASI score and replicating and expanding the work done by Weiji Sandara in Canada is the way to go in perioperative risk assessment. So heading back to our original case, uh so to our listeners again, what's the level of your agreement with the primary claim that a subjective assessment is an accurate metric of functional capacity?

And what is the level of evidence for what you believe? Dr. Walsh and Dr. Shapiro, thank you for joining me this week on the show. Thank you so much, David. Really excited to be part of this. Thank you very much, Dr. David Hao.

As always, please share this podcast with anybody in the field of anesthesia. We hope to continue to encourage thoughtful and critical discussion about our clinical practices. And before we head out, I know Dr. Shapiro, you have a pretty funny story about pre-op testing for us, right? We talk about how Asking a patient about their functional capacity is subjective data that's difficult to use a hundred percent as reliable risk assessment. We had a patient who was having their preoperative interview and he was asked, can you walk up a flight of stairs?

And he said, yeah, I can do that And then he was asked the second question, which was, can you walk up two flights of stairs? And he said, I can do that. And the interviewer Asked him when was the last time you did that? And the patient said, Oh, years ago. But I know I could if I wanted to.