Metabolic equivalent of task (MET) in the preoperative assessment in aortic surgery: a retrospective analysis


 Background: Reliable prediction of the preoperative risk is of crucial importance for patients undergoing aortic repair. In this retrospective cohort study we evaluated the metabolic equivalent of task (MET) in the preoperative risk assessment with clinical outcome in a cohort of consecutive patients.Methods: Retrospective analysis of prospectively collected data in a single center unit of 296 patients undergoing open or endovascular aortic repair. Out of these, 276 patients had a preoperative statement of their functional capacity in metabolic units and were evaluated concerning their postoperative outcome. Results: Mean survival of the whole cohort was 74 months. There was no significant difference between patients with a functional capacity of more or less than 4 MET respectively (74.5 vs 65.4 months; p=0.64). The infrarenal cohort was subdivided based on the operative technique (open or endovascular) and the preoperative MET status. Survival analysis of all four subgroups showed no significant differences. The median follow-up of the patients’ cohort was 10.8 months. In 46 patients with >4MET (20.9%) perioperative complications occurred compared to the group with <4MET with 18 patients (32.1%) (p=0.075). The risk to miss a potential need for cardiac optimization in patients >4MET was 7%.Conclusion: Functional preoperative evaluation by MET in patients undergoing aortic surgery is used as standardized tool for patients’ assessment and is meant to act as a useful surrogate marker of perioperative performance. However, it could not discern patients in need for cardiopulmonary testing.Trial registration: clinicaltrials.gov, registration number NCT03617601 (retrospectively registered)


Introduction
For several reasons a reliable preoperative risk evaluation is of particular importance in vascular patients.
Foremost the majority of indications is made for patients in sixth to eighth decade of their lives and a large portion of it remains of preventive intention such as elective aneurysm repair or carotid surgery for asymptomatic occlusive disease. Furthermore the presence of atherosclerotic disease is supposed to bear an own increased perioperative risk 1, 2 . Almost 40% of all perioperative complications are made up by cardiac incidences 3 . In high-risk patients undergoing non-cardiac operations the 30-day mortality for cardiovascular death or myocardial infarction has been estimated to be over 5% 4 .
A prophylactic coronary revascularization in patients with stable coronary artery disease by percutaneous coronary intervention (PCI) or bypass operation did not show any bene t concerning the reduction of mortality or occurrence of perioperative myocardial infarction when compared to medical treatment 5, 6 .
The preoperative assessment of the metabolic equivalent of task (MET) is an easy clinical evaluation of functional capacity or exercise tolerance of an individual. A MET is de ned as the resting metabolic rate, that is the amount of the consumed oxygen at rest (approximately 3.5 ml = 0 2 /kg/min i.e. 1.2 kcal/min for a 70 kg person) 7 . According to the MET concept a patient would be considered as " t for surgery" when the stairs of two ights can be climbed and the housework can be fully managed by oneself.
Preoperative assessment of MET also is used for the evaluation of the perioperative risk for cardiac complications.
The current guidelines of the European Society of Cardiology and European Society of Angiology (ESC/ESA) restrain from preoperative cardiac testing in non-cardiac surgery as it has failed to improve perioperative outcome 8 . Alike, the current guidelines of the American College of Cardiology and American Heart Association (ACC/AHA) from 2014 recommend no further cardiac testing in patients undergoing non-cardiac surgery with moderate risk for a major cardiac event (MACE) with a functional capacity that exceeds 4 MET with class IIb evidence 9 . Although these guidelines are widely supported and implemented, they are still not well validated for speci c vascular interventions i.e. aortic operations.
In this retrospective cohort study we investigated the MET concept in the preoperative risk assessment with clinical outcome parameters in a cohort of consecutive patients who received open or endovascular abdominal, thoraco-abdominal or thoracic aortic repair. We hypothesize that patients with a preoperative status of 4 MET and more had less perioperative complications, suffered less from postoperative myocardial infarction and stroke and had a lower in-hospital mortality compared to patients with a status under 4 MET. This might support the recommendation of non-indicated cardiac testing in patients who are t for surgery.

Patients and Data collection
All relevant patients' data were retrospectively extracted from the Swiss Vasc Registry, a prospective, mandatory nationwide computer-based vascular registry in Switzerland. After the extraction of the centerassociated clinical data all les were completed with the current patients' follow-up and rechecked for obvious entry mistakes in the hospital software Medfolio. From May 2009 till March 2016, 296 patients underwent open or endovascular aortic repair and were divided into four main groups depending on infrarenal, juxta-and para-renal, thoraco-abdominal and thoracic pathology (Table 1). Beside demographic parameters, type of operation and urgency, co-morbidities and cardiovascular risk factors were documented. Perioperative data included postoperative complications and reoperations, myocardial infarction and stroke, in-hospital mortality, and the need of later cardiovascular interventions.
The study was approved by the Cantonal Ethics Committee of Zurich under the protocol number 2017 − 00801 in June 2017 and adheres to the principles of reported research on human beings set forth in the Helsinki Declaration. The study has been registered under clinicaltrials.gov with the registration number NCT03617601. Operative technique in open infra-, juxta-, and thoraco-abdominal repair: All infrarenal aneurysms received preferably baloon occlusion of the iliac arteries in order to avoid clamping damage. An aneurysm was labeled juxtarenal, if at least one renal artery had to be clamped suprarenal in order to perform the proximal anastomosis. If there was suprarenal clamping of both renal arteries, renal cold perfusion was established whenever possible. In case of thoraco-abdominal aortic repair patients were positioned in a right lateral decubitus position to facilitate aortic exposure via thoraco-phrenico-lumbotomy. Airway management was achieved with a double lumen endotracheal tube to allow unilateral right lung ventilation. Operation was performed with partial left heart bypass with suprainguinal cannulation of the external iliac arty and vein without hypothermia 11 . All patients received spinal cord monitoring with sensory evoked potentials (SEPs) and motor evoked potentials (MEPs) and prophylactic spinal cord drainage and pressure monitoring prior to surgery 12 .
Operative strategy in infrarenal and thoracic endovascular repair: All patients receiving endovascular repair were operated with a femoral cut-down. Indication for endovascular repair was within the instructions for use (IFU) of the chosen endoprosthesis. Endovascular repair of aortic pathologies in landing zone < 3 according to Mitchell and Ishimaru were operated in functional cardiac arrest with rapid pacing 13 . If treatment length exceeded 20 cm and required deployment in landing zone 2, or in the presence of prior aortic surgery, respectively carotid-subclavian bypass was carried out prior to endovascular repair in order to minimize the risk of spinal cord malperfusion 14 .
Follow-up of patients: All patients with open infrarenal aortic repair had a duplex scan from the referring angiologist 3 months postoperative and clinical evaluation by the operating surgeon in order to evaluate clinical status and rule out incisional hernia. From then on, ultrasound and a physical examination was done every year. Patients with thoraco-abdominal open repair received an Angio-CT scan 3 months after surgery and then were followed with ultrasound. All endovascular repairs underwent postoperative CT scan after 3 months and one year. If sac diameter was in regression and endoleaks were absent it was switched to an ultrasound or contrast enhanced duplex scan.

Statistical Methods
Descriptive statistics were generated by counting and calculation of percentages of the nominal and ordinal variables. Numeric variables were described with mean and standard deviation. Survival analysis was done according to the method described by Kaplan and Meier. For the analysis of differences between the groups the log rank test was used. Twenty patients out of 296 were emergencies without any information about preoperative MET status. These patients were excluded from the comparison between patients with > 4MET and < 4MET according demographic data, survival and follow up. Perioperative complications, postoperative cardiac interventions and postoperative severe events (stroke, myocardial infarction) were calculated separately for the emergency group without MET status. In-hospital mortality was calculated for all patient regardless of their MET status focusing on the localization of the procedure (infrarenal, juxtarenal, thoraco-abdominal, thoracic). Statsdirect software (Version 2.7.3, Statsdirect Ltd, Cheshire, UK) was used for all statistical analyses.
Sensitivity/speci city analysis: Based on additional cardiac assessment we were able to calculate the sensitivity and speci ty of the diagnostic validity of MET. Sensitivity being the proportion of patients with a pathologic result of the assessment among patients with < 4MET, speci city the proportion with a normal result among patients with > 4MET.
Sample size calculation: Based on results of the study -power of the results of the retrospective risk analysis -we performed a sample size calculation. This gives us an indication of the number of patients needed to detect a statistically signi cant difference, assuming that our hypothesis is true. Power: 100% -probability of not detecting the "true" difference. For this calculation we set theβ -error at 20%, power therefore being 80%.

Results
Demographic parameters of patients with > 4MET and < 4MET Patients with < 4MET had a signi cant higher incidence of diabetes mellitus (p = 0.0002), peripheral artery occlusive disease (p < 0.0001), history of smoking (p = 0.003), adipositas (p = 0.03) and chronic obstructive pulmonary disease (p = 0.05). Furthermore, this group had a tendency to more previous cardiovascular interventions, otherwise the two cohorts were comparable (Table 2).  Sixteen percent of all patients irrespective of their functional capacity underwent a cardiovascular intervention after the initial hospitalization during the follow-up period (36 patients with > 4MET; 9 patients with < 4MET). Four of twenty patients with an unknown status of MET underwent postoperative cardiovascular interventions (20%). There was no signi cant difference among these groups (p = 0.91).

MET and major events (stroke or myocardial infarction) after hospitalization
In the group of patients with > 4MET (220 patients), 3.6% (8 patients) developed a myocardial infarction (MI) after hospitalization and 1.8% (4 patients) had a stroke, which results in a cumulative rate of 5.5% (12 patients) with major events after hospitalization. Compared to the group with < 4MET (56 patients), 8.9% (5 patients) suffered a myocardial infarction and no stroke. Only one patient (5%) had a myocardial infarction in the group with unknown MET status (20 patients). Although, there was no signi cant difference among the groups (p = 0.40), there might be a slight tendency to more myocardial infarction in the group with < 4MET.
MET and in-hospital death Thirteen patients died, which results in an in-hospital mortality of 4.4% for all procedures and approaches. In the infrarenal group of 182 patients the in-hospital mortality was 3.8%. Seven patients died, two from cardiovascular events, ve from other events. Six of these seven patients were operated conventionally by laparotomy, one received an endovascular prosthesis. In-hospital mortality after juxtarenal approach (n = 84) was 4.8% (2 cardiovascular deaths, 2 other deaths). During follow up, there were 8 deaths in the infrarenal (3 cardiovascular and 5 other) and 2 deaths (both cardiovascular) in the juxtarenal group.
In the thoraco-abdominal (n = 13) and thoracic approach group (n = 17) respectively, one patient died from cardiovascular cause, which results in an in-hospital mortality of 7.7% and 5.9%. During follow up, there were no further deaths in both groups.
In the subgroup analysis of infrarenal patients there were no signi cant differences concerning inhospital mortality with regard to preoperative MET status (p = 0.99).

Sensitivity of MET status for perioperative cardiovascular risk assessment
One hundred patients with > 4MET received concomitant cardiac assessment before the operation. In two of these patients the result of the assessment was unknown, while 81 showed normal results and 17 had pathological ndings (= 98 patients with > 4MET and cardiac assessment). In the group with < 4MET, 48 patients underwent preoperative cardiac work-up. Out of these 36 had normal results, whereas 12 showed pathological results in stress echocardiography or coronary angiography. Thus, sensitivity of functional capacity assessment with MET was 41% with a low positive predictive value of 25%. In the group of > 4MET with preoperative cardiac work-up (n = 98), 17 patients showed pathological results. In 7 (7%) out of these this had therapeutic consequences. In the group of < 4MET (n = 48), 4 (8.3%) out of 12 patients received preoperative interventions as a result of abnormal ndings. Thus, the risk to miss a potential need for cardiac optimization in patients > 4MET was 7% (Fig. 3).

Discussion
The aim of this study was to investigate, if the preoperative MET status as a functional and cost-effective parameter may contribute to the avoidance of unnecessary preoperative testing and may identify cardiovascular risk in indicated cases. Since there is no prospective study according to consequent preoperative cardiac evaluation on the basis of MET status, we would like to provide these retrospective data as an initial starting point for further discussion.
In our study, 13 patients died, which resulted in an in-hospital mortality of 4.4% for the whole cohort. In the infrarenal group of 182 patients the in-hospital mortality was 3.8% with no statistical difference between open and endovascular repair. This is consistent with those of major prospective studies investigating infrarenal aortic repair 15,16 . This study could not demonstrate a statistical difference between patients achieving more or less of a functional capacity of 4 MET in terms of survival and overall complications. The sensitivity of MET assessment in terms of the detecting of pathologic cardiac ndings is with 41% quite low. This is most probably due to the aw of retrospective design and the rather low sample size in this study. Using the gures of the EVAR 2 trial the 30-day mortality in the endovascular group judged un t for surgery was 7.3%. The patient cohort in EVAR 2 was categorized as un t for surgery based on their past medical history of myocardial infarction, arrhythmias 17 . Irrespective of the operative method in our cohort the mortality in patients > 4MET was 4.3% and 6.1% in those with < 4MET. A sample size calculation taking into account the given numbers with a hypothesis of a better outcome of the > 4MET patients would claim a patients number of approximately 7300 patients to include into a prospective study, by planning the test in an one-sided design (i.e. >4MET is associated with lower mortality), 4000 patients would have to be included.
In 1999, Reilly et al. reported that the self-reported exercise tolerance is a valuable tool in order to predict in-hospital perioperative risk 18 . We adopted the MET concept gradually mainly for infrarenal and juxtarenal aortic repair. Apart from cost reduction this regimen is supposed to avoid unnecessary preoperative investigations, which is claimed by current guidelines 8,19 .
The most commonly used, and a simple method to categorize a patients' operative risk preoperatively is the ASA classi cation, which is easy to assess and is based on the past medical history and judgement of the assessing physician. This simple categorization correlates astonishingly well with postoperative outcome 19, 20, 21 . As cardiovascular morbidity is predominately expected in vascular patients it is not the only potential factor of comorbidity causing postoperative problems. For instance concomitant impairment of liver function in cirrhotic patients or chronic obstructive pulmonary disease (COPD) are independent risk factors for considerably increased postoperative mortality in general major surgery 21,22,23 . Taking this into account it is clear that the surgeons intuition gives a good estimate of the perioperative risk as there is evidence that the surgeons personal judgment on the patients' individual risk correlates reliably with postoperative outcome as well 24 .
So far there have been only few reports concerning validity of preoperative risk assessment in noncardiac surgery and none, to our knowledge, in patients undergoing aortic surgery. In patients undergoing thoracic surgery there is evidence that stair climbing capacity is a good predictor of mortality. In patients undergoing high-risk non-cardiac surgery the inability to climb at least two ights of stairs did not provide an increased risk of perioperative mortality but was associated with more cardio-pulmonary complications. However, the fraction was less than 10% 18,25,26 .
In order to give valid data according to e cacy and safety of the functional status assessment in patients undergoing aortic surgery prospective data in by far larger numbers would be required. However, it seems that even in high-risk operations such as aortic surgery the MET assessment of the patients gives a good estimate of individual physical tness and overall physiological reserves as our data showed that the group < 4MET had a signi cant higher percentage of diabetics, claudicans and smokers. Most probably due to small sample sizes our data could not show any difference in mortality or cardiovascular events in patients with more or less than 4 MET in patients undergoing aortic repair.

Conclusion
The concept of MET analysis in the preoperative assessment of patients undergoing aortic surgery is a standardized tool for preoperative assessment and acts as a surrogate marker of the perioperative patients' performance. However, it does not substitute cardio-pulmonary testing in indicated cases as it failed to indicate a signi cant threshold in patients undergoing aortic surgery.