Ve-VCO2 slope: Just to AT or all the way to the peak?

We’ve had some questions lately about some of our CPET guidelines. These questions were informational in nature not confrontational but they served to remind us that the reference values we use for CPET interpretation were developed and put in place at least ten years ago and it is past time they were reviewed. As a starting point I’ve been re-reading the ATS-ACCP and AHA statements on cardiopulmonary exercise testing. One sentence from the AHA statement concerning the Ve-VCO2 slope caught my eye. Specifically:

“… calculation of the Ve/VCO2 slope with all exercise data obtained from a progressive exercise test (initiation to peak effort) appears to provide additional clinical information compared with submaximal calculations (i.e. those that use linear data points before the steepening associated with ventilatory compensation for metabolic acidosis).”

Ve-VCO2 slope is calculated using a linear regression function and we have been calculating it using only the test data between the start of exercise and the anaerobic threshold. The AHA statement however says we should be calculating it using the data all the way up to peak exercise (the ATS/ACCP statement is mute on this point since it does not even discuss Ve-VCO2 slope other than as a graph). Because Ve-VCO2 slope is a key component in our assessment of CPET results it is important that we get this right.

Ve and VCO2 have a reasonably linear relationship up to the anaerobic threshold. After the anaerobic threshold ventilation is driven by acidosis as well as CO2. This means that the Ve-VCO2 slope tends to be steeper (greater change in Ve per unit of VCO2) after anaerobic threshold than it was before. A Ve-VCO2 slope calculated from the entire CPET will therefore have steeper slope than one calculated using just using rest to AT.

VeVCO2 Slopes

 

For decades the peak VO2 from an exercise test has been used to assess surgical risk and as a predictor of mortality and hospitalization. One particularly valid criticism in the use of peak VO2 is that it can be falsely reduced for reasons that have nothing to do with either cardiac or pulmonary status. Research has shown that the Ve-VCO2 slope is probably a better indicator than peak VO2 and since the Ve-VCO2 slope from rest to AT is linear this means that an accurate Ve-VCO2 slope can be obtained from even a submaximal exercise test. This is one of the reasons that I had thought the consensus was that Ve-VCO2 slope should be calculated only using data up until AT.

When I reviewed the literature referenced in the AHA statement I found several studies that indicate that although the start to AT (sub-maximal) Ve-VCO2 slope was a powerful predictor of an individual’s clinical outcome, the start to peak Ve-VCO2 slope was superior. The reasons for this are not completely clear but it was speculated that it is due to poorer cardiac function at higher levels of exercise which leads to greater acidosis and a steeper Ve-VCO2 slope after anaerobic threshold. At least one study indicated that the difference between the sub-maximal Ve-VCO2 slope and the Ve-VCO2 slope after AT was itself a significant predictor of outcome, and that the greater the difference, the poorer the outcome was.

It does seem to make sense therefore, to calculate Ve-VCO2 from start to peak exercise. One concern I have with this approach would be that the peak Ve-VCO2 slope, like peak VO2, is dependent on patient effort. For this reason it is somewhat unclear what its normal range is. The upper limit of normal for the sub-maximal Ve-VCO2 slope is considered to be 34. The peak Ve-VCO2 slope should always be greater than the sub-maximal Ve-VCO2 slope and for this reason it should probably not have the same upper limit of normal. The AHA statement is a bit vague on this point in that it says that a peak Ve-VCO2 slope less than 30 is normal and one greater than 40 is abnormal, leaving the actual upper limit of normal somewhat up in the air. This is something that needs further research since I have seen several studies using peak Ve-VCO2 slope that used an ULN of 34.

An additional concern is that the Ve-VCO2 slope is calculated using linear regression. This is certainly acceptable for a sub-maximal Ve-VCO2 slope but the Ve-VCO2 slope from an entire CPET is not linear. This leaves open to question exactly what the numerical value that is calculated for a peak Ve-VCO2 slope is really saying. It seems to me there are probably better ways to characterize the peak Ve-VCO2 slope other than linear regression but since I am not a statistician or mathematician I will have to wait for somebody else to address this issue.

For my lab it seems the best approach will be to continue to calculate the sub-maximal Ve-VCO2. Its upper limit of normal is reasonably well characterized and it is also well suited to assessing sub-maximal tests. In addition however, I will also start calculating the peak Ve-VCO2 slope and will use the AHA value of 40 as the upper limit of normal. This should also be a useful check on those times when the anaerobic threshold is indeterminate. I will be interested to see if there are any patients who have a normal sub-maximal Ve-VCO2 slope but an abnormal peak Ve-VCO2 slope, or if both slopes will always be abnormal. This is something I haven’t seen addressed in any studies.

Twenty-five years ago we only reported the Ve/VCO2 at AT but since then it has gotten more complicated. Now we report the Ve/VCO2 at AT, the lowest observed Ve/VCO2, the sub-maximal Ve-VCO2 slope and the peak Ve-VCO2 slope. I would like to pare these down but it’s difficult to determine which of these values, if any, should be dropped. There are proponents for each of these measurements but only a few studies that have compared each approach and these have been primarily limited to the ability to prognosticate survival or hospitalization rates and not particularly towards the causes of the differences. The relationship between Ve and VCO2 is multi-factorial and it is possible that each of these measurements is saying something slightly different about patient physiology. It’s also just as possible that they overlap each other and that a single measurement (peak Ve-VCO2 slope?) would be sufficient but at this moment the jury is still out.

References:

Arena R, Myers J, Aslam SS, Varughese EB, Peberdy MA. Technical considerations related to the minute ventilation/carbon dioxided output slope in patients with heart failure. Chest 2003; 124: 720-727.

ATS/ACCP Statement on cardiopulmonary exercise testing. Am J Resp Crit Care 2003; 167: 211-277.

Balady GJ; et al. Clinician’s guide to cardiopulmonary exercise testing in adults: A scientific statement from the American Heart Association. Circulation 2010; 122: 191-225.

Chua TP, Ponikowski P, Harrington D, Anker SD, Webb-Peploe K, Clark AL, Poole-Wilson PA, Coats AJS. Clinical correlates and prognostic significance of the ventilatory response to exercise in chronis heart failure. J Am Coll Cardiol 1997; 29: 1585-1590.

Corra U, Mezzani A, Bosimini E, Scapellato F, Imparato A, Giannuzzi P. Ventilatory response to exercise improves risk stratification in patients with chronic heart failure and intermediate functional capacity. Am Heart J 2002; 143: 418-426.

Ingle L, Goode K, Carroll S, Sloan R, Boyes C, Cleland JGF, Clark AL. Prognostic values of the Ve/VCO2 slope calculated from different time intervals in patients with suspected heart failure. Int J Cardiol 2007; 118: 350-355.

Koike A, Itoh H, Kato M, Sawada H, Aizawa T, Fu LT, Watanabe H. Prognostic powere of ventilatory responses during submaximal exercise in patients with chronic heart disease. Chest 2002; 121: 1581-1588.

Sun XG, Hansen JE, Garatachea N, Storer TW, Wasserman K. Ventilatory efficiency during exercise in healthy subjects. Am J Resp Crit Care Med 2002; 166: 1443-1448.

Creative Commons License
PFT Blog by Richard Johnston is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License

18 thoughts on “Ve-VCO2 slope: Just to AT or all the way to the peak?

    • Jenny –

      We continue to report the Ve/VCO2 at AT but do not use it in interpreting results. I think that if you’re going to use a the ratio between Ve and VCO2 the lowest observed Ve/VCO2 is more diagnostic. There is also good concordance generally between the Ve-VCO2 slope to AT and the lowest observed Ve/VCO2 so I tend to think they’re measuring something similar. There is less concordance between the Ve-VCO2 slope to AT and the Ve-VCO2 slope to peak exercise, however. When the Ve-VCO2 slope to AT is abnormal (i.e. >34) but the Ve-VCO2 slope to peak exercise isn’t (i.e. <40) I think this is mostly due to the patient being unable (or unwilling) to push themselves to a more maximal exercise level. Less commonly I've seen the Ve-VCO2 slope to AT be normal (i.e. < 34) and the Ve-VCO2 slope to peak exercise be abnormal (i.e. > 40). Sometimes this comes with indications of a cardiac limitation (low VO2 at AT, elevated chronotropic index, VO2 plateau, abnormal BP response etc. etc.) so it’s easy to state that a primary limitation is cardiovascular but when there aren’t any additional indications it’s far less clear and the best we can say is that it’s a non-specific indication of a cardiac limitation.

      Strictly speaking the Ve/VCO2 at AT should be the same as the Ve-VCO2 slope to AT and the intercept of the slope at a zero Ve. We don’t report the intercept because there are no clear guidelines about what is and is not abnormal. I am aware of only one research study that has looked at the Ve-VCO2 slope intercept (Whipp BJ. Control of the exercise hyperpnea: the unanswered question. Adv Exp Med Biol 2008;605:16-21) which the author relates to changes in Vd/Vt relative to Ve. I’ve tried to look at the Ve-VCO2 intercept and have so far been unable to correlate it either to diagnosis or to specific patterns within the CPET but I may well be missing something.

      Regards, Richard

      • Richards,

        Thank you for your text, it is very clarify.
        I would want know, to assess the Ve/VCO2 intercept, do you use the values during exercise (begining until peak) or during rest to peak of exercise?
        Beyond, do you make the calculate in the excel? Like Ve/VCO2 slope

        Thank you so much

        • Guilherme –

          I use both slopes, from start to AT and start to peak. I do not use any of the resting Ve and VCO2 values since I’ve found that any hyperventilation at rest skews the slopes. When a CPET is maximal both slopes tend to concur with each other, i.e. they’re both usually either normal or abnormal. But when a CPET is submaximal I place more value on the start to AT slope. Yes, I use Excel to calculate the slopes. Easiest way is to graph the Ve and VCO2 data points (scatter graph) and then insert a trend line (linear regression). I like this because by looking at the data points you can see how well they really fit the regression line (any data entry errors also tend to stand out), and once in a while I’ll reject the calculated Ve-VCO2 slope if the point scatter around the line is too large.

          Regards, Richard

  1. thank you for your reply. I have a new CPET software and am trying to pick VECO2 slope and intercept for a report. I just am not sure if it should be picked to AT or to peak? Is this discussion still unresolved?

    • HI Richard,

      I hope you might be able to help me – I am trying to find out how to calculate the VE/VCO2 slope value – can you confirm whether this has to be done by CPET software or whether it can be done using excel?

      • I use an Excel spreadsheet to calculate Ve-VCO2 slope and do it from exercise start (I’ve found that including any resting data skews the slope) to AT and to peak exercise. The easiest way is to have two columns; VCO2 and Ve, graph them using an X-Y format and then insert a linear regression trendline. After I enter all the data I adjust the data range to measure the different slopes.

    • Ana –

      I’ve never used SPSS so the answer has to be no. I will say I’m not a super big fan of the V-slope technique. I think it’s usually easier to look for a nadir in PetO2 and Ve/VO2. The problem with V-slope is determining which data points belong to the lower slope and which belong to the upper slope. Not hard at the extremes so to speak, but in the middle where the break point is located noise and variability in Ve and VCO2 can make it difficult to determine what belongs with what and this can make a big difference in the calculated slopes and the AT (VT). At least a couple of CPET systems I know of have software that can determine AT via V-slope but I think their accuracy is so-so at best but since I haven’t run across any articles that looked at the quality of automatically determined AT’s, that’s just my opinion.

      Regards, Richard

      • Sorry, bother u once more. But do u know any technique to determine VT on SPSS?
        And, what are the method u usually use do get it?
        Tks

        • Hi! I read another post yours. I guess I found it the answer. I found this blog randomly and it´s very interesting.

          I just still don´t know what software is the more reliable to do get to AT.

          Tks for the informations

          • Ana –

            I would suggest plotting PetO2 and Ve/VO2 versus time. You can use a spreadsheet for this since you don’t need anything special. AT occurs at the nadir (lowest values) of PetO2 and Ve/VO2. I’d point out however, that there is a certain amount of noise and variability in CPET data so you also want to see distinct increases in PetO2 and Ve/VO2 following AT as well. I’d also mention that PetCO2 reaches its highest value and Ve/VCO2 its lowest value about 30 seconds (+/- a bit) after the nadir in PetO2 and Ve/VO2. You don’t need fancy software to do this, it’s just a matter of recognizing the pattern.

            Regards, Richard

  2. Hello Richard,
    About VE/VCO2 intercept. I’m looking at the literature and it’s not clear to me if I have to skip the initial part of the exercise (eg when the patient is quiet and is adapting to the mask, no pedalling) because he can hyperventilatilate to look for the intercept point. I’m trying, and if I quit the beginning the intercept is always LOWER than if I keep the initial part. What is your opinion? There is nice graph about VE/VCO2 intercept and slpe in Am J Resp Crit Care Med 2011 183: 1302-10 (fig 1). Sue is discussing about keeping/cuttting the hyperventilating last part of the exercise in order to have a “real” VE/VCO2 slope but form the fuigure it doesn’t seem the intercept changes (and of course, I would say since the beginning of the exercise is the same). No talk about the beginning of the exercise. Thank you for any suggestion
    Chiara Torregiani

    • Chiara –

      I don’t use any of the baseline/rest data when calculating the Ve-VCO2 slope since like you I’ve found it tends to skew the Ve-VCO2 slope (regardless of whether the slope is from test start to AT or from test start to peak exercise). For this reason I only use data obtained during exercise. I have not found any guidelines for this but when I’ve seen a research article comment on what data they used to calculate the Ve-VCO2 slope (and they often don’t comment at all) they usually say they used the data from exercise as well.

      My experience however, is that it’s not the intercept but the slope that matters. I routinely calculate the intercept when I calculate the slopes but I’ve yet to see any particular correlation between the intercept and the primary factors limiting exercise whereas there tends to be one for slope. In addition, while researching Ve-VCO2 slopes I’ve seen little written about the intercept. As an example in that article Sue comments on the intercept but states that “because a graph of Ve versus VCO2 (Figure 1) has a positive intercept (2.4 6 2.4 L/min, mean 6 SD) on the Ve axis in 90% of normal subjects (6), although the magnitude of the intercept differs in patients with heart and lung disease.” but frustratingly he (and many other authors) does not say what the difference is.

      – Richard

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.