The use of cardiopulmonary exercise testing (CPET) as a preoperative risk stratification tool for a range of non-cardiopulmonary surgery is increasing. The utility of CPET in this role is dependent on the technology being able to identify accurately and reliably those patients at increased risk of perioperative events when compared with existing risk stratification tools. This article identifies and reviews systematically the current literature regarding the use of CPET as a preoperative tool for stratifying risk in major non-cardiopulmonary surgery. Specifically, it focuses on evaluating the capacity of CPET variables to predict the risk of postoperative complications and mortality in comparison to other methods of risk assessment. Furthermore, the potential for combining results from CPET and non-CPET methods of risk prediction to enhance the capacity to identify high risk patients is considered. The review indicates that CPET can identify patients at increased risk of adverse perioperative outcomes. However, the selection of variables and threshold values to indicate high risk differ for different surgical procedures and underlying conditions. Furthermore, the available data suggest that CPET variables outperform alternative methods of preoperative risk stratification. Several studies also identify that CPET variables may be used in combination with non-CPET variables to increase perioperative risk prediction accuracy. These findings illustrate that CPET has the capacity to identify patients at increased risk of adverse outcome before a range of non-cardiopulmonary surgical procedures. Further research is required to optimise its use, potentially by combining CPET results with alternative methods of risk stratification.
- Cardiopulmonary exercise testing (CPET)
- perioperative risk evaluation
- non-cardiopulmonary surgery
- exercise testing
- perioperative risk
- adult anaesthesia
- adult intensive & critical care
- preventive medicine
- adult surgery
Statistics from Altmetric.com
- Cardiopulmonary exercise testing (CPET)
- perioperative risk evaluation
- non-cardiopulmonary surgery
- exercise testing
- perioperative risk
- adult anaesthesia
- adult intensive & critical care
- preventive medicine
- adult surgery
Cardiopulmonary exercise testing (CPET) is a clinical tool used to measure the performance of the cardiorespiratory system and assess an individual's functional capacity (fitness level). CPET is arguably the most objective and comprehensive method of achieving this aim. Alternative tests such as the 6 min walk,1 shuttle walking,2 and stair climbing3 have also been used to evaluate functional capacity. However, the relationship of these measures to CPET variables seems inconsistent and at present CPET remains the gold standard, with no effective screening tool described.2 CPET is used to aid clinical decision making relating to a range of pathologies. Some of the uses of CPET include: providing a differential diagnosis of exercise intolerance and unexplained dyspnoea,4 evaluating the severity of heart failure and the impairment caused by COPD,4 and as pre-operative risk stratification tool used before cardiopulmonary and non-cardiopulmonary surgery.4 5 It is this latter use of CPET—its role in evaluating perioperative risk in non-cardiopulmonary surgery—which forms the topic of the current review.
Almost two decades ago, Older and colleagues identified an association between low functional capacity determined by CPET, and poor patient outcome following non-cardiopulmonary surgery.5 Based on these findings, and others, CPET has been increasingly adopted as a preoperative risk stratification tool, for a range of non-cardiopulmonary surgery. National Health Service (NHS) Trusts throughout the UK are employing CPET to assist clinical risk profiling, and to better inform perioperative management. Information gained from CPET is used to influence the choice of operation (full planned procedure, palliative procedure, or no operation), choice of perioperative management strategy (eg, intraoperative fluid therapy) and, most commonly, choice of postoperative care (level 1, 2 or 3 care).6
The increased implementation of CPET for the prediction of risk in the perioperative period raises several questions. Can CPET predict the risk of perioperative complications and/or mortality in individuals undergoing major general surgical procedures? If so, which CPET variables predict risk with the greatest precision? What are the optimal threshold levels for each of these variables? Do the variables and threshold values vary for different types of surgery? How do CPET derived variables compare with alternative preoperative methods of risk stratification, such as clinical risk scores (eg, Revised Cardiac Risk Index (RCRI))?7 Finally, does CPET provide an incremental improvement to perioperative risk stratification when added to other risk prediction methods?
The aim of this article is to review the current literature on CPET as a preoperative risk stratification tool for general surgery. Specifically, we aim to assess the ability of CPET to predict outcome, consider its value relative to other methods of preoperative assessment, and evaluate whether CPET results can be combined with other non-CPET variables to increase the capacity to predict outcome. In addition, for those less familiar with CPET, we begin this review by briefly describing the conduct of a CPET test.
Conducting a CPET
CPET is a graded exercise challenge which, through non-invasive measurements of gas exchange at the mouth, ECG trace observation, and analysis of heart rate, blood pressure, and peripheral oxygen saturation, allows the cardiovascular and respiratory systems to be studied under the controlled ‘stress’ of exercise. This allows inspection of the integrated oxygen delivery system when the demand for oxygen is high and the system is required to function near to its maximum capacity. Despite requiring a moderate to high level of exertion, CPET is well tolerated by patients8 9 and is safe to conduct on most patient cohorts.10
Before the start of a CPET several procedures should be conducted: the breath-by-breath gas analyser must be calibrated, a medical history should be taken, and basic demographic and anthropometric data should be obtained (ie, age, gender, height, and body mass). Other pre-test measures can be conducted to aid CPET analysis, such as measuring haemoglobin concentration and conducting simple spirometric tests to determine maximal voluntary ventilation. After these procedures are completed the patient is connected to various pieces of monitoring equipment (ie, metabolic cart, pulse oximeter, sphygmomanometer, and ECG machine) in preparation for the exercise test.
In the assessment of postoperative risk following general surgery, CPET is usually conducted on a cycle ergometer and the test protocol normally includes four phases. Typically, an initial rest phase (approximately 3 min) is employed to establish baseline values, followed by an unloaded cycling (0 W) phase to allow the patient to become familiar with the cycling motion and to reduce the influence of the lag present between increased work rate (WR) and the oxygen uptake () response.11 Following this, the incremental exercise phase begins. A ramp protocol is commonly used, during which the set WR is increased linearly with time, with a corresponding increase in the intensity of the exercise. The criteria for test termination differs between laboratories; in some, the test is terminated by the patient at volitional exhaustion, while others perform a submaximal test and stop exercise when a particular criterion has been met, such as a respiratory exchange ratio (RER) above 1.9 Upon test completion, a recovery period of low intensity exercise should be performed to encourage maintenance of venous return, thus reducing the risk of brisk blood pressure reduction and associated light-headedness.11 The patient must be observed throughout recovery until physiological variables, including heart rate, blood pressure, ventilation, and oxygen saturation, are close to baseline levels and any exercise induced ECG changes have resolved. Close inspection of the patient and physiological data should take place throughout all the phases of the CPET. The test should be stopped if the patient experiences any adverse symptoms (eg, chest pain, dizziness, or severe breathlessness) or if the physiological data indicate a potential adverse event (eg, ECG abnormalities, substantial blood pressure changes).10
Of the many physiological variables determined during CPET, in the preoperative assessment setting, three have been shown to help identify high risk patients: peak/max,12 anaerobic threshold (AT),5 8 9 13 and ventilatory equivalent for carbon dioxide .9 14 Consequently, these three variables are used most frequently to stratify risk for non-cardiopulmonary surgery.6 For more detailed descriptions of CPET protocols and physiology the reader is directed to Wasserman et al11 and the American Thoracic Society/American College of Chest Physicians statement on CPET.10
Evidence base for CPET as a risk stratification tool
Methods of review
To identify the papers to be included in this review we conducted a systematic search using the approach of Smith et al.15 We used the same web based archives (Pubmed and ISI Web of Knowledge) and search terms (‘CPET/surgery’, ‘CPEX/surgery’, ‘cardiopulmonary/exercise testing/surgery’, and ‘max/surgery’) as Smith and colleagues, and consistent with their approach, we checked reference lists of eligible papers for additional candidate papers.15 All searches were conducted on the 20 December 2010.
To meet the inclusion criteria of this review, the principal focus of a primary research paper had to be the association between CPET variables and postoperative outcome following major non-cardiopulmonary surgery. Papers were excluded based on the following criteria: the operation was minor or cardiopulmonary, surgical outcome data were not available, or the paper was not a full research article (eg, conference abstract or letter to the editor). The search criteria identified 948 manuscripts; of these, 12 met the predefined criteria (11 identified from electronic database search and one article12 identified through reference list searches). Study characteristics of included studies are presented in table 1.
For all eligible manuscripts we extracted data on the associations between all measured preoperative risk stratification variables (CPET and non-CPET) and measures of perioperative outcome (mortality and markers of complications). The 12 studies included 2275 patients; this consisted of 1726 patients undergoing mixed intra-abdominal surgical procedures from five studies, 330 patients undergoing upper gastrointestinal (GI) surgery from four studies, 160 patients included in two studies who were having an abdominal aortic aneurysm (AAA) repair, and 59 patients having a hepatic transplant who were from a single study. For one full research article,17 only the abstract was available in English. The information contained in the abstract was deemed sufficient for the study to be included in analysis.
Each study was assessed to determine:
Which, if any, CPET variables were associated with measures of postoperative outcome, and to establish the strength of the associations.
Whether markers derived from CPET had the capacity to predict those at increased risk of morbidity or mortality following surgery.
How CPET variables compared with other risk stratification measures at predicting those at greater risk of poor outcome.
If, using multivariate analysis, CPET measures combined with other variables being used to assess perioperative risk to improve the capacity to predict surgical outcome.
Review of included studies
The 12 eligible studies, categorised by type of surgery, are reviewed below.
Mixed intra-abdominal surgical procedures
The results from studies conducted on patients undergoing mixed intra-abdominal surgical procedures demonstrate an association between CPET variables and postoperative mortality5 13 and morbidity (table 2).8 Two investigations by Older and colleagues initially identified that a low AT was associated with an increased mortality rate in elderly patients following general surgery.5 13 In the first of these studies, patients with an AT <11 ml/kg/min had a higher rate of cardiovascular mortality (18%) than those with an AT ≥11 ml/kg/min (0.8%, p <0.001).5 Furthermore, an AT <11 ml/kg/min, in combination with exercise induced ischaemia, was associated with the highest rate of cardiovascular mortality (42%, p <0.001). These findings are supported by a large (n=548) interventional study conducted by the same group, where postoperative care level was determined by surgery type and CPET results, with an AT <11 ml/kg/min signifying high risk and mandating triage to intensive therapy unit (ITU) care.13 Those identified as high risk due to poor CPET results and/or surgery type had the highest rate of cardiovascular mortality (n=7/153, 4.6%), when compared to the rest of the study population (n=2/395, 0.5%). The study is likely to underestimate the capacity for CPET to identify high risk patients, as, by the nature of the study design, those deemed at greatest risk received the highest level of care, which would be expected to reduce mortality in this group. In addition, in this second study, only descriptive statistics were used to evaluate the data, therefore the probability of the results occurring due to chance alone are not known. Despite this limitation, these studies indicate that results from CPET, particularly AT and exercise induced ischaemia, are associated with mortality rate following major intra-abdominal surgery.
In a recent retrospective study of 843 patients, Wilson et al showed that an AT <11 ml/kg/min was associated with and able to predict hospital mortality (AUC 0.68, 95% CI 0.59 to 0.76) with a sensitivity of 88% and a specificity of 47%.9 A ≥34 was also a predictor of in-hospital death (AUC 0.69, 95% CI 0.55 to 0.82), with the ‘cut-off’ of 34 giving a sensitivity of 88% and a specificity of 47%.
Snowdon et al found that both AT and peak predicted complications in 116 patients following general surgery.8 The optimal AT to distinguish those at increased risk of postoperative complications was found to be 10.1 ml/kg/min, which is notably lower than the AT <11 ml/kg/min used previously.13 The threshold of 10.1 ml/kg/min predicted complications (AUC 0.85, 95% CI 0.669 to 0.889; p<0.001) with good sensitivity (88%) and specificity (79%).
In a smaller study, Hightower et al reported that an AT <75% of the predicted value was found to predict those at increased risk of complications (AUC 0.72; sensitivity 88%; specificity 79%; p=0.016) in 32 patients undergoing major intra-abdominal surgery.16
A particular strength of the studies by Snowden and Hightower is that clinicians were blinded to the CPET results, ensuring they did not influence patient care or data collection.8 16 This should give a more accurate reflection of the true magnitude of the association between CPET variables and outcome by removing the effect of confounding due to clinicians acting on CPET variables and influencing outcome (eg, choice of level of postoperative care received by the patient).22
These studies show that, for intra-abdominal surgery, the AT is associated with postoperative outcome and has the capacity to predict morbidity and mortality with a reasonable degree of accuracy.8 9 In general, the associations between CPET variables and outcome in mixed intra-abdominal surgery were higher than for other methods of risk prediction. In studies in which CPET was found to be a moderately strong predictor of outcome, the RCRI was not associated with postoperative complications or all cause mortality.8 9 Similarly, the AT was also found to be a better predictor of postoperative complications than the Veterans Specific Activity Questionnaire (VSAQ), which estimates functional capacity.8 Furthermore, multivariate regression modelling showed that the ability to predict those at risk of having postoperative complications was highest when the AT and VSAQ were used in combination (AUC 0.894).8
Upper GI surgery
The first two studies to investigate the association between CPET derived variables and outcome following upper GI surgery were conducted by the same group in Japan.17 18 In their initial study, Nagamatsu et al analysed data from 52 patients who had a right thoracolaparotomy for thoracic oesophageal cancer, and observed significant differences in max and AT (both normalised to body surface area) between patients with and without postoperative cardiopulmonary complications.17 In a follow-up study, they retrospectively analysed data from 91 patients who had undergone an oesophagectomy with three field lymphadenectomy for squamous cell carcinoma and preoperative CPET.18 Consistent with their original study, max values (normalised to body surface area) were significantly lower in the cohort of patients that experienced cardiopulmonary complications (789 ml/min/m2) than in those without complications (966 ml/min/m2) (p<0.001). However, no association was observed between complications and AT. Further analysis revealed that a max of 800 ml/min/m2 was the optimal threshold to discriminate those at high risk of postoperative cardiopulmonary morbidity. Below this threshold, each 100 ml/min/m2 decrease in max was accompanied by an increase in patient complication rate. In addition, these studies indicated that the majority of resting lung function variables were not associated with the subsequent occurrence of postoperative complications for either surgery type.
The capacity of CPET markers to predict patients at increased risk of postoperative complications following upper GI surgery was also studied in 78 patients undergoing an oesophagectomy.19 In line with previous upper GI studies, the AT did not differ between those with and without cardiopulmonary complications. However, patients with postoperative cardiopulmonary complications had a significantly lower peak (19.2±5.1 ml/kg/min) than those without complications (21.4±4.8 ml/kg/min) (p=0.04). However, the 95% CI constructed around the mean difference between the two groups indicates that no difference in peak may actually be present (mean difference 2.3 ml/kg/min, 95% CI −0.06 to 4.5 ml/kg/min). Consistent with the borderline association between peak and outcome, prediction of those at risk of postoperative cardiopulmonary complications was poor (AUC 0.63, 95% CI 0.50 to 0.76). This study indicates that an association may exist between peak and outcome following an oesophagectomy, although its capacity to predict those at increased risk is low. However, inconsistencies in results from statistical analysis preclude firm conclusions being drawn.
Considering the data from these three studies together, peak appears to be weakly associated with outcome following oesophagectomy. However, these studies all had small samples (n=52–91 patients) and were likely to be underpowered to detect associations between CPET variables and outcome markers. Therefore, the clinical utility of peak, or other CPET variables, for prediction of postoperative outcome in this type of surgery is uncertain. Further research with adequate methodology (clinicians blinded to CPET results) and larger patient numbers would be required to resolve this uncertainty.
The association between CPET variables and outcome following upper GI surgery was also investigated in 109 obese patients (mean body mass index (BMI) 48.1 kg/m2) undergoing laparoscopic Roux-en-Y gastric bypass surgery as a bariatric procedure.12 Primary composite complications were death, unstable angina, myocardial infarction, deep vein thrombosis, pulmonary embolus, renal failure, and stroke; secondary outcomes were hospital length of stay (LOS) and readmission. The population was divided into tertiles according to peak values achieved during treadmill exercise. The rate of complications was significantly higher for the first tertile (n=6, 16.2%) compared to the second and third tertiles (for both; n =1, 2.8%) (p=0.03). Similarly, hospital LOS was longer for the first tertile (3.8 days) than the second (2.9 days) and third tertiles (2.8 days) (p=0.005). Of the CPET variables that were measured, peak had the greatest capacity to predict those at risk with an AUC of 0.77. The threshold peak value of 15.8 ml/kg/min, which was the upper boundary of the lowest tertile, had a sensitivity of 75.0% and a specificity of 73.3% to predict the occurrence of postoperative complications. These results indicate that for a morbidly obese population having bariatric surgery, a peak ≤15.8 ml/kg/min has a reasonable capacity to predict those at increased risk of postoperative complications and longer hospital LOS.
A small study (39 patients) originally highlighted the potential use of CPET for identifying high risk individuals before undergoing an AAA repair.20 In this study, a peak <20 ml/kg/min was tentatively proposed as a marker to identify patients at increased perioperative risk, despite no differences in peak being present between patients with or without postoperative complications. Following this, Carlisle and Swart retrospectively studied the association between four CPET markers ( peak, AT, , and ), four other risk stratification methods (RCRI, POSSUM, Simplified Acute Physiology Score (SAPS) II, and the Acute Physiology And Chronic Health Evaluation (APACHE) II) and all cause mortality following AAA repair.14 The study differs from other CPET articles in non-cardiopulmonary surgery, in that the mortality was not only measured in the initial postoperative period, but also up to a median of 35 months post-surgery, termed mid-term. Of the 130 patients studied, a total of 29 (22.3%) had died by the time of the last follow-up, 14 (10.8%) doing so in hospital within 30 days of surgery. All four measured CPET variables were correlated with mid-term survival, as were the other four risk stratification methods, although to a lesser degree. The had the strongest association with rate of mortality at 30 days and at mid-term, with a HR for mortality of 1.14 (95% CI 1.08 to 1.20; p<0.001) for the latter. Of the perioperative risk scores, the RCRI had the strongest association with survival with a HR of 1.86 (95% CI 1.25 to 2.78; p=0.002). A value ≥42 and RCRI >1 were found to be the optimal thresholds to distinguish patients at increased risk of mortality. Kaplan–Meier survival curves demonstrated a substantially lower rate of survival over a 3 year period in the group with both a ≥42 and a RCRI >1.
Based on the results from Carlise and Swart, is the optimal CPET variable to indicate high risk patients undergoing an AAA repair.14 However, analysis was not conducted to identify the classification accuracy of this variable; therefore it is not possible to know the extent to which it can predict those at greater risk of mortality following surgery.
Hepatic transplantation surgery
Fifty-nine patients underwent CPET, resting pulmonary function testing, and resting cardiovascular function testing, to assess whether variables derived from these tests were associated with 100 day mortality following hepatic surgery.21 Six (10.2%) of the 59 patients studied died within 100 days of undergoing surgery. Patients who survived did not differ from those who died in terms of age, gender, severity of liver disease, resting pulmonary function, or resting cardiovascular function, including left ventricular ejection fraction (61±4 vs 62±4, respectively; p=0.48). However, the non-survivor group had a greater proportion of patients with a peak <60% of their predicted (4/6 patients, 66.6%) when compared to the survivor group (12/53 patients, 22.6%) (p=0.04). The frequency of patients with an AT <50% of the predicted peak value was also higher in non-survivors (5/6 patients, 83.3%) than in survivors (18/53 patients, 34.0%) (p=0.03). In addition, a peak <60% of predicted and an AT <50% of the predicted peak was associated with a mortality rate of 25%. This study shows that markers of cardiovascular reserve ( peak and AT), determined by CPET, are associated with mortality following hepatic surgery, whereas resting cardiac and pulmonary function are not. Evaluation of the capacity of these variables to predict morbidity and mortality following liver transplant surgery is awaited.
The majority of the available literature indicates that CPET markers are associated with surgical outcome following non-cardiopulmonary surgery, and that for many types of surgery CPET has the capacity to identify high risk patients. However, the optimal predictor of high risk appears to differ between surgery types, with AT shown to be the best indicator of higher risk patients for major intra-abdominal surgery, peak for bariatric surgery, and for AAA repair surgery (table 2).8 9 12 14 Furthermore, it appears that CPET is of limited use for risk stratification before an oesophagectomy.19 Variation in the type and strength of the association between CPET variable and outcome in different types of surgery is likely to be due to variation in the relative importance of different factors predisposing to adverse outcome. For example, the relative contributions of surgical technique, perioperative care, and patient physiological reserve are likely to vary between different procedures. Furthermore, the current literature relating CPET variables to outcome is inconsistent in methods of analysis and presentation of data, precluding firm conclusions with respect to choice of variable and optimal cut points for identifying high risk patients for specific procedures.
Articles directly comparing the capacity of CPET variables to predict perioperative risk against other methods of risk assessment are scarce. However, data that are available indicate that CPET variables outperform methods that estimate functional capacity such as the VSAQ,8 markers of resting cardiac and respiratory function,17 18 and risk indices such as the ASA16 and the RCRI.8 9 For mixed intra-abdominal surgery, the ability of AT to predict all cause mortality (AUC 0.68)9 was higher than that reported for the RCRI in a recent systematic review (AUC 0.62).23
Given that many, often unrelated, factors cause postoperative morbidity, it is clear that CPET alone has a limited capacity to stratify perioperative risk. To maximise our ability to identify high risk patients before surgery, it is likely that CPET results will ultimately be combined with alternative methods of risk stratification which identify other aspects of perioperative risk. Indeed, studies included in the current review identified that CPET variables could be used in conjunction with the RCRI and the VSAQ to enhance the capacity to predict perioperative risk.8 14 However, the capacity of these combinations of variables to stratify risk has not been validated using different populations and therefore their use is not warranted in the clinical setting at this time. In addition, despite the intriguing data on prediction of postoperative outcome using plasma biomarkers (eg, N-terminal pro-brain natriuretic peptide),24 the relative predictive performance of this approach and any incremental improvement in prediction in association with other approaches remains unknown.
Strengths and limitations of the review
The strengths of this review are that it provides a current review of the CPET literature, which was conducted in a systematic manner, and that it analysed the predictive capacity of CPET variables in comparison to non-CPET markers to establish which preoperative risk stratification tool is most precise.
The current review was performed in systematic fashion, ensuring that all studies identified using the specific search terms were analysed in the same way, reducing bias in the inclusion and interpretation processes. In addition, the review included data from a fairly large number of patients (n=2275) taken from 12 studies. When articles were reviewed a clear distinction was made between those that identified an association between a preoperative assessment variable and outcome, and those that identified that a preoperative test had the capacity to predict outcome, with the latter giving a better indication of the practical use of CPET in the clinical setting. Final strengths of the review are that CPET variables were compared with other non-CPET markers to evaluate if CPET compares favourably with other risk stratifications tools, and CPET and non-CPET markers were combined to assess whether it improves the capacity to predict surgical outcome.
The limitations of the review are that it included studies with study designs and analysis methods which are not ideal, including those that are small, single centred, and unblinded, and those that did not analyse the predictive capacity of CPET variables.
The review included some small studies, which did not included a priori power calculations, and are likely to be underpowered. Furthermore, all the studies included in the review were conducted at a single centre, therefore it is uncertain what the external validity of these studies is (generalisability). Furthermore, the majority of these observational studies (9/11 studies) were unblinded, which increases the risk of confounding by indication.22 The effect of this will be to reduce the strength of association between predictor variable (CPET) and outcome, in relation to the true association. The review was also limited by the analyses used in several of the studies, often analysing only the association between CPET variable and outcome, with many not exploring the value of multivariate risk models. The limitations of this review reflect the limitations of the literature base regarding the use of CPET in the preoperative evaluation for non-cardiac surgery, which will improve as more high quality studies are published.
Recommendations for current practice
Risk prediction using CPET derived variables (and other formal methods of risk evaluation) should always be evaluated in the context of the overall clinical picture. In general, a prudent approach is likely to result in optimal patient outcomes, but this needs to be balanced against the increased resource usage (eg, intensive care unit bed days).
Based on the reviewed literature the following guidance can be given:
An AT of 11 ml/kg/min appears an acceptable threshold to use in a clinical setting to indicate increased perioperative risk following major intra-abdominal surgery, despite some indication from a recent study that a lower threshold may be optimal.8
Based on findings by Carlise and Swart, a ≥42 should be used to help identify high risk patients undergoing AAA repair surgery.14
The peak may also be used to help stratify risk, particularly for bariatric surgery.12
A recent evaluation of the conduct and interpretation of CPET throughout the UK showed that interpretation of CPET results throughout NHS trusts is in line with current empirical findings, with AT, peak, and the relationship being central to interpretation, and other CPET variables being used in addition to these.6
To attain greater understanding of the role that CPET has as a predictor of postoperative complications and mortality following non-cardiopulmonary surgery, further studies need to be performed. Future studies should endeavour to use more robust study design (blinded studies with an appropriate sample size) and analysis should include evaluation of the predictive capacity of CPET variables. Future investigations should also compare CPET with alternative preoperative assessment approaches, such as clinical risk score indices, resting physiological measurements, and plasma biomarkers, in order to assess the incremental benefit of combining these approaches to establish the optimal method to predict risk. Studies investigating the use of CPET in combination with plasma biomarkers are particularly warranted, as both appear to have a high capacity to predict outcome independently,8 25 and are likely to identify different factors that contribute to perioperative risk. Furthermore, this combination has not been evaluated to date.
Finally, due to the infrequency of mortality in most general surgical procedures, unless study population numbers are high, studies must use postoperative complication rates or other markers of morbidity, to assess the capacity of CPET to predict outcome. Other, potentially interesting areas of research are: (1) Identifying whether other CPET variables influence postoperative morbidity and/or mortality, such as the O2 pulse, oxygen uptake efficiency slope, presence of oscillating breathing pattern, and /work rate slope. (2) Verifying and optimising threshold values for CPET variables and deriving multivariate models used to predict risk. (3) Exploring the variation in the optimal CPET marker for risk prediction across surgical types. (4) Investigating whether training programmes aimed at increasing CPET measured variables improve outcome following surgery. (5) Investigating whether alternative methods of normalising CPET values, other than by weight (ml/kg/min), alters the precision of outcome prediction. For example, the approach used by Epstein et al,21 where patient's peak and AT were presented as a percentage of that predicted for their weight, height, age, and gender, is of interest as this allows a patient's results to be compared with others of the same demographic group.
Physical fitness, defined by cardiopulmonary exercise testing (CPET) derived variables, is associated with outcome following non-cardiopulmonary surgery: less fit patients have a higher incidence of morbidity and mortality.
Predicted risk from CPET derived data should be considered in the context of the overall clinical picture and other risk prediction methods (eg, clinical risk scores).
The best CPET variable(s) to predict the risk of adverse patient outcome around the time of surgery differ(s) between surgery types.
Predicted risk may guide choice of procedure, perioperative interventions, and postoperative care environment.
Further studies, with robust study design, are required to improve understanding of the role of CPET in perioperative risk stratification and decision guidance.
Current research questions
Can choice of CPET variables and threshold values be refined to improve the precision of risk prediction?
Can CPET markers be combined with alternative preoperative assessment approaches to improve the precision of risk prediction?
Can training programmes be used to increase cardiorespiratory fitness and improve outcome following surgery?
Older P, Smith R, Courtney P, et al. Preoperative evaluation of cardiac failure and ischaemia in elderly patients by cardiopulmonary exercise testing. Chest 1993;104:701–4.
Older P, Hall A, Hader R. Cardiopulmonary exercise testing as a screening test for perioperative management of major surgery in the elderly. Chest 1999;116:355–62.
Smith TB, Stonell C, Purkayastha S, et al. Cardiopulmonary exercise testing as a risk assessment method in non cardio-pulmonary surgery: a systematic review. Anaesthesia 2009;64:883–93.
Snowden CP, Prentis JM, Anderson HL, et al. Submaximal cardiopulmonary exercise testing predicts complications and hospital length of stay in patients undergoing major elective surgery. Ann Surg 2010;251:535–41.
Wilson RJ, Davies S, Yates D, et al. Impaired functional capacity is associated with all-cause mortality after major elective intra-abdominal surgery. Br J Anaesth 2010;105:297–303.
Questions (true (T)/false (F): answers after the references)
Many patients awaiting major intra-abdominal surgery find it difficult to complete CPET
The majority of studies investigating the utility of CPET for preoperative risk assessment before non-cardiopulmonary surgery are adequately blinded
A high relationship is associated with poor outcome following AAA repair surgery
The AT is the optimal CPET variable for stratifying perioperative risk for all types of general surgery
CPET is particularly useful for identifying high risk patients awaiting an oesophagectomy
Funding PH is supported at the Portex Unit at the UCL institute of Child Health by unrestricted research funding from Smiths Medical Ltd. Some of this work was undertaken at University College Hospitals - University College London Comprehensive Biomedical Research Centre which received a portion of funding from the UK Department of Health National Institute for Health Research Biomedical Research Centres funding scheme. MG holds the Royal College of Anaesthetists BOC Research Grant awarded by the National Institute of Academic Anaesthesia.
Competing interests None.
Provenance and peer review Commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.