Anaerobic threshold: pitfalls and limitations

Cardiopulmonary exercise testing (CPET) is a low risk, non-invasive investigation that allows accurate, dynamic assessment of cardiac and pulmonary performance during exercise. Pulmonary, cardiovascular, endocrine, neuropsychiatric, and skeletal muscle systems are all required to be functioning in a coordinated fashion in a stressed patient. Most pre-operative investigations assess one system when it is at rest (e.g. echocardiogram) [1]. By measuring dynamic gas exchange during graded exercise, CPET can identify potential deficiencies within these systems. These are often not adequately reflected in the indices of resting lung and cardiac function [2]. Traditional methods that assess maximum exercise capacity such as exercise time or maximum oxygen uptake are influenced by the patient’s motivation when performing a maximum exercise test or by non-cardiorespiratory symptoms [3]. These can lead to premature termination of the test. When performing graded exercise, the anaerobic threshold (AT) is an estimate of the onset of metabolic acidosis due to an oxygen supply: demand imbalance. Numerically, it is presented as the rate of oxygen consumption (VO2) at which acid starts to accumulate. This is seldom invasively measured but is usually taken as the time at which the patient starts to exhale increasing amounts of carbon dioxide (VeCO2) to compensate for a build up of lactic acid and resulting metabolic acidosis. It should always be expressed as a percentage of the maximum oxygen consumption (VO2 max) [4]. The AT occurs at about 50–60% of VO2 max in normal individuals but has a range of 30–80%. It is a well defined physiological point that can be measured in most patients. A low pre-operative AT has been shown to correlate with postoperative mortality [2]. Although there is no clear agreement on the best single predictor of outcome, thresholds of peak oxygen consumption (VO2 peak) of 15 ml O2.kg−1.min−1 and an oxygen consumption at the anaerobic threshold (AT) of 11 ml O2.kg−1.min−1 have been shown to discriminate between higher and lower risk patients in non-cardiac surgery [2]. This information has been used to optimise the use of critical care beds or to guide the choice of therapy. However, if the information is going to be used to recommend against radical but potentially life saving treatments because of the perceived risk, then we have got to be very confident in the accuracy of these data. Several factors could introduce error in AT measurement during the CPET. These include genuine alterations in level of fitness, the reliability of the CPET equipment and inter and intra-observer variations in interpretation of CPET results. Variation in protocols and changes in performance caused by a learning effect with repeated testing can also affect results. A learning effect is defined as an increased skill or improved performance caused by previous exposure to the exercise [5]. This is well documented with treadmill testing, although there is little evidence relating to cycle-based CPET in clinical populations [6]. The linking of VO2 max and AT raises the first problem with its use for preoperative risk assessment [7]. VO2 max is related to age and many patients over 80 years will have a predicted VO2 max of 20 ml.kg−1.min−1 or less. It is easy to see that they may have an AT of < 11 ml.kg−1.min−1 and still be ‘normal’. Similarly an AT of 11 ml.kg−1.min−1 in a 20 year-old male would be far more significant than the same result in an 80 year-old. Both of these numbers cannot carry the same degree of risk. Patient medication and specifically beta blockers often affect the results of a CPET. Many patients with impaired left ventricular function are heart rate dependent to increase oxygen delivery. This poses a difficult question for patients who have a low AT who are on beta blockers. Ischaemia may not be obvious during the test because of the blunted heart rate response to exercise. However, is it correct to continue beta blockers if under stress they are unable to adequately increase oxygen delivery? Looking at both the oxygen pulse and the continuous ECG may help to answer this question. The oxygen pulse is a non-invasive estimate of the stroke volume. It is a ratio of the VO2 to heart rate (HR) and is a reflection of the amount of oxygen extracted per heart beat. If the VO2 is largely dependent on cardiac output then if the heart rate is fixed but the VO2 continues to rise, then this must be mostly due to an increase in stroke volume [8]. Conversely if the HR is fixed and the VO2 remains low during exercise then the heart is not able to increase stroke volume and is rate dependent to meet any increase in oxygen demand. Part of the CPET test is a continuous12-lead ECG with real time ST segment analysis. In theory, if ischaemia is detected after the AT, then this is less concerning than if it occurs before the AT. This is especially important if the AT is low. Improving the ischaemia may improve the cardiac function and in so doing the AT. In patients on beta blockers with a low AT, CPET allows a safe assessment of their cardiac response during stress when beta blockers are stopped [8]. If the AT improves and little or no ischaemia is found, then this would seem a safer way forward. On the other hand if ischaemia is detected and the AT remains low, either the stress of the operation and recovery need to be modified or coronary perfusion will have to be improved. No studies are available to support the use of pre-operative CPET testing as a guide to selecting patients who would benefit from coronary revascularisation. However, if we believe that a low AT is a strong predictor of postoperative outcome and coronary ischaemia is the primary limiting factor to improving oxygen delivery in the patient, then intervention of some form seems logical. In practice some of these patients will have signs and symptoms that warrant intervention in their own right. Considering that at least 24 variables are measured during a CPET, it would seem unwise to place too much emphasis on only one of these variables. It is analogous to only assessing the ejection fraction on an echocardiogram and then trying to make a detailed assessment of cardiac function. The AT needs to be reported in the context of the other results so that a more accurate assessment can be made of not only prognosis but possible therapeutic interventions. A standardised reporting format for pre-operative patients would help to reduce error. We feel that at the very minimum the report should comment on: The patient’s cooperation and effort. The AT (absolute number and as a percentage of VO2max). Peak and predicted maximum VO2. Cardiac ischaemia; if it was detected, how it was measured and when it occurred. A summary of what the team feel the cardiorespiratory risk for the patient is and if they feel any intervention or change in medication would help. At present, no national standard exists for a pre-operative CPET test. Failure to contextualise a patient’s report or selective reporting of certain variables will dangerously bias the test result. If these tests are to be widely used then it is important that not only the clinicians doing the test but also the clinicians receiving the report have a detailed understanding of the data and its limitations. If you do not know what an oxygen pulse is, you are unlikely to ask for it. However, if it is given and interpreted then this is far more helpful. We also feel that it is wrong to separate the pre-operative CPET from the clinicians performing the anaesthetic. Ideally, this service should be run by a combination of anaesthetists involved with high risk cases and cardiologists. This allows an individual plan to be made to meet specific needs. If we do not put the time in before the operation we will put the time in afterwards dealing with the complications. Tests that measure multiple variables and that depend on both patient and technician for accurate results, have a strong possibility of error. Poor patient effort or preparation, incorrect machine calibration and sampling error or gas leak can all introduce significant error. These all occur before we attempt to interpret the test. The interpretation also has considerable scope for inter-observer variability. No degree of accurate interpretation will make up for a poor test and vice-versa. Centres that perform a higher number of tests are likely to have more reproducible results, and our local experience is that at least 30 trial runs are required before attempting to introduce a clinical service and this is with experienced cardiac physiologists running the tests. No discussion would be complete without discussing the cost of this test. CPET machines cost about £20 000 and service contracts about £3 000 per year. In addition to this, cardiac physiologist and consultant time needs to be considered. If a unit performs three tests per week then each test will cost approximately £150. The fewer tests that a unit performs, the more expensive each test is. If a unit only does 25 tests a year then each test costs approximately £440. It would seem logical to have pre operative CPET at larger hospitals and for smaller hospitals to refer patients to these centres [9]. This seems the best way of producing cost effective, reproducible results. In summary, AT is a useful tool if it is accurate and presented with supporting data. It is an expensive but safe test. However there are with many areas for potential error. It is important that clinicians know the limitations and pitfalls of the AT if they are going to use it to inform clinical decisions.