The CHALLENGE trial ignited social media with its great story of exercise as a cancer therapeutic. But liking a conclusion is not a reason to stop thinking.
I have specialized in sports medicine in France for 38 years. I have noticed that my patients who have cancer are quite few in number, compared to the size of the population I have treated over all these years. I should point out that 90% of them are recreational athletes, even if they compete, and 10% are high-level athletes. I have also noticed over all these years that the specialists who treat them always say the same thing: they are always surprised, because these patients tolerate treatments better, recover better, more quickly, they seem to heal more and relapse less than their usual sedentary population. There is a fairly extensive literature showing that athletes have fewer cancers than sedentary people. But I do not know if there are studies showing that those who have cancer have the characteristics I mentioned above, significantly more than sedentary people. However, it is permissible to think so. We always come back to the same problem: it seems to me that primary prevention through regular physical activity, a balanced diet and a good psycho-social balance will always be more effective and easier to implement (or less difficult) than secondary or tertiary prevention in people who are already ill and, a fortiori, undergoing one or more treatments that are difficult to bear...
It's a simple concept.... exercise increases mitochondrial content. Improved mitochondrial content is global in the body and not restricted to muscle. Immune function should improve collateral to lower insulin levels. The biologic plausibility of improved outcomes is clear. It is not definitive by any means but seminal nevertheless.
Why would one argue against such a sensible intervention for any of our patients. It does no harm and by any measure of biology beneficial.
The continued absence of physicians to recognize that focusing on cellular metabolism as a primary intervention versus drugs is why we are all unhealthier. Chemotherapy works in conjunction with the immune system, the latter of which, kills the cancer and not the drug. The excitement around immunotherapies for cancer is simply it's the pt's immune system being activated directly against the oncologic event versus indirect effects of standard drugs. I believe we should be celebrating this study as a first step towards evaluating how to treat cancer patients beyond just giving chemotherapy after chemotherapy.
I think it’s more important to focus on what this trial accomplished.
CHALLENGE was a tough study to run. Long follow-up, behavioral intervention, limited funding, and a population that’s hard to engage over years. Despite all that, the team pulled off a large, international randomized trial that showed a meaningful benefit in recurrence and possibly survival.
Of course there are limitations. Maybe the effect size is inflated. Maybe adherence diluted some of the signal. But to me, this still reads as a positive trial. The direction of benefit was consistent, the outcomes were clinically relevant, and the intervention was safe and well tolerated.
As an oncologist, I spend a lot of time talking about toxic therapies with marginal gains. Here’s something that improves quality of life, reduces recurrence, and might help people live longer. We should be asking how to make this part of routine care, not looking for reasons to dismiss it.
This was a hard study. And an important one. Let’s build on it.
Med student ignorance here... But could reverse causality be a factor? With the elephant in the room being the massive amount of intensive wellbeing support the study arm is getting... Positive feedback loops all round... The actual role of exercise being moot. I presume there are 5 yr randomised controlled trials on intensive positive wellbeing initiatives in cancer patients??? Can't see drug companies running them...
The major point about lack of objective changes and hence plausibility is not really valid. They showed an improvement of 2ml/kg/min in cardiorespiratory fitness and 22 m for the 30 m walk. All these are clinically significant (and statistically), especially considering the low function of the population. A decrease in body weight or waist circumference is obviously not plausible considering the small dose and absence of diet intervention. If they had measured body compositon, I believe they would have noticed a decrease in body fat, especially visceral fat which typically inflammation and insulin resistance.
I wonder, did the analysis screen and account for depression, or lack of it, as a confounding variable? Patients suffering from depression and anxiety after even successful cancer treatment would be less likely to join a study like this or to comply with a protracted, challenging fitness program. And depression itself might contribute to all-cause mortality through various channels.
Interesting study. But I agree with Dr. JMM. For me the concerns are the implausible effect size and implausible rapidity of onset of benefit, as well as the impractical/impossible nature of trying to reproduce the protocol in a practice setting. And it’s also a bit curious that the amount of exercise required to reduce some fairly hard endpoints in cancer was insufficient to produce much difference in measures of fitness.
But I think the greater short term risk is wellness/ “health” types misusing the data to encourage people to shun chemo etc and “just do some exercise”.
This trial and its outcome IS revolutionary. It’s revolutionary because it reveals that there are molecular mechanisms in cancer progression and in survival that have long been unexplored, and we need to identify their origins.
It’s revolutionary because it goes to the fundamental basis for what patients are seeking. Holistic approaches that allow them to lead a productive and meaningful life with a lethal disease rather than succumbing to it.
The slowed accrual, I believe, is really a result of deprioritizing of its effects since we ALL, over the last 7-8 decades, have been trained to build from molecular mechanisms of disease action up as opposed to working from enhanced survival down. Thus, it’s a welcome change back to that classic paradigm of clinical science.
In fact, perhaps all hypothesis generation should begin AT the BEDSIDE, and culminate in bench and cellular mechanisms.
This NEJM report also recognizes that targeted drugs aren’t the only way to both make patients feel better with a lethal disease but also enhance their survival. Thus, it shifts, or rather readjusts, the fulcrum of medical interventions towards making strides back towards survival, as opposed to surrogate molecular measures that often don’t pan out as predictive when working our way up from molecular mechanisms to clinical outcomes.
You're on it! Everyone forgets basic cellular physiology let alone molecular intra-cellular operations. Exercise leads to a 50% increase in mitochondrial volume in 3 weeks. It's global and not specific to muscle. Thus in turn, a myriad of anti-neoplastic effects could occur ... without loosing weight, or building significantly more muscle mass, etc.
This study doesn't prove anything yet. It should challenge scientists to think more about impacting cellular activity through other means than just meds.
"Due to slow recruitment and a slower-than-expected event rate they changed to a five-year analysis. Yet they still had far less than the expected primary endpoint events (224 vs 380). This reduces statistical power and raises the possibility of false positive findings—which is consistent with the biological implausibility."
But low power increases risk of false positive results, not false negative ?
Im not a doctor but cant see how this study was proof of anything. I can see so many ways why it would be invalid.This would have to be so tightly controlled and have many more subjects for a more true curve. There are (as of course you doctors dont need to be told by me) so many factors contributing to whether or not a patient will relapse or stay cancer free. As soon as my un trained mind thought about the correlation between cancer relapse and exercise, I went into scoff mode. Sure a more healthy body is a plus so excercise can put you in better shape. Just from a statistical view, the number tracked and the methods used were a big disappointment after the potential of the subject.
John thank you for your assessment on this story. I saw a lot of the same things when I looked at the paper. There was a lot of fuss about it, but ultimately the effect was disappointing IMO. When I saw that it was mostly geared toward health education etc I couldn’t help but feel there is some fudging going on here. The other obvious thing is cancer prevention altogether…why don’t we take two cohorts matched for age, BMI etc one that exercises to a certain level and one that doesn’t exercise at all and just track cancer diagnosis. Because if exercise isnt preventative…should we expect it to be a “treatment?” I agree the results are promising but lots of lurking stuff there. I think the exercise effect could likely be higher but would love to see more prevention studies.
You have some thoughtful critiques here. Regarding the points you raise:
1. Implausibility. This statement is made without looking at available data. There is observational data on exercise after cancer which has relatively large effects on survival. There are also very few randomised trials of exercise with long follow up, so we don't really know.
2. Lack of objective exercise effects. There were some differences in VO2max for example, arguably the most important measure.
3. 'Early' separation in survival curves at 12 months. This is an arbitrary distinction and not a strong criticism. Colorectal cancer tends to relapse quite early at around 15 - 18 months. If exercise is going to have an effect then 'early' separation like this is to be expected.
4. Poor adherence. I think that the adherence was quite good for an exercise study in inactive 60 year olds. Not sure what the argument is here - because adherence was poor, then the effect is smaller than it appears?
5. This is valid, but lower than expected event rate is endemic in oncology trials. The fact that the OS separation is more obvious after 3 years reflects the survival kinetics of colon cancer.
6. I'm struggling to think how more contact with the study team and physical therapists would prevent cancer recurrence. Consider that on the one hand, you are saying it is implausible that exercise has such a large effect, while on the other hand trying to claim that increased attention has a large effect.
7. Some misapprehensions here. I don't think anyone believes that the specific exercise intervention in this study is what made the difference, aside from the fact they tried quite hard. Given everything we know about exercise already, a reasonable takeaway is "exercise makes a difference, do more than you are doing now" if you are receiving curative intent cancer therapy. The patients were younger than the median age of diagnosis of colon cancer, but the benefits of exercise if anything increase in older patients. It is wrong to suggest this doesn't have broad applicability.
I dispute that the cost of accepting this protocol would be 'massive'. No clinical trial protocol is implemented verbatim in the real world. Again, it is not the specific protocol that matters. A real world implementation would largely be getting people to exercise to the already recommended level, which doesn't seem overly onerous or expensive. I have little doubt that such an intervention would be cost effective. Cancer therapy gets more and more expensive, and people are also getting diagnosed at younger ages, so the marginal cost of each recurrence only increases over time. Additionally, exercise prevents other diseases which further improves the cost benefit analysis. And in any case, even if no attempt is made to implement formal exercise programs, on an individual level it is important that a patient knows they can do more to reduce their risk of recurrence.
What does concern me about this study is the potential for informative censoring - ie. people who dropped out of the exercise arm were also more likely to have a recurrence due to being generally less well or more frail. I still think that is quite unlikely to eliminate the observed benefits of exercise, however. Timothée Olivier looks at this on X (https://x.com/Timothee_MD/status/1929411436272382303), and the curves still separate.
I also wonder, as an 8th point: would simply the intensity of the in-person, mandatory intervention have its own effect? That's a lot of attention from other human beings. What other kinds of conversations happened during those interactions? Is there any data on the impact of social interactions on outcomes?
Personally, I am not as bothered the lack of between-group differences in "objective" metrics like body weight, waist circumference, etc (although the exercise group did have > VO2max). It's totally plausible to me that the benefit of exercise is beyond these measures. But I agree that the lack of adherence to the protocol makes these findings pretty implausible. The self-reported MET increase was roughly half of their goal (6.3 vs a goal of 10) and that finding is very likely inflated due to observation bias. Also, only 20% of the exercise group did the recommended supervised exercise session in the first 6 months -what?? With that kind of adherence it is simply amazing to see an effect size this large.
You just wonder if being able to exercise was a marker for a better immune system and not a cause of a better immune system
My dad is 97 in assisted-living and there's only a few people his age but the ones that are all like him mentally intact. And that means he's less likely to fall was more likely to show up for the doctor, etc..
I have specialized in sports medicine in France for 38 years. I have noticed that my patients who have cancer are quite few in number, compared to the size of the population I have treated over all these years. I should point out that 90% of them are recreational athletes, even if they compete, and 10% are high-level athletes. I have also noticed over all these years that the specialists who treat them always say the same thing: they are always surprised, because these patients tolerate treatments better, recover better, more quickly, they seem to heal more and relapse less than their usual sedentary population. There is a fairly extensive literature showing that athletes have fewer cancers than sedentary people. But I do not know if there are studies showing that those who have cancer have the characteristics I mentioned above, significantly more than sedentary people. However, it is permissible to think so. We always come back to the same problem: it seems to me that primary prevention through regular physical activity, a balanced diet and a good psycho-social balance will always be more effective and easier to implement (or less difficult) than secondary or tertiary prevention in people who are already ill and, a fortiori, undergoing one or more treatments that are difficult to bear...
It's a simple concept.... exercise increases mitochondrial content. Improved mitochondrial content is global in the body and not restricted to muscle. Immune function should improve collateral to lower insulin levels. The biologic plausibility of improved outcomes is clear. It is not definitive by any means but seminal nevertheless.
Why would one argue against such a sensible intervention for any of our patients. It does no harm and by any measure of biology beneficial.
The continued absence of physicians to recognize that focusing on cellular metabolism as a primary intervention versus drugs is why we are all unhealthier. Chemotherapy works in conjunction with the immune system, the latter of which, kills the cancer and not the drug. The excitement around immunotherapies for cancer is simply it's the pt's immune system being activated directly against the oncologic event versus indirect effects of standard drugs. I believe we should be celebrating this study as a first step towards evaluating how to treat cancer patients beyond just giving chemotherapy after chemotherapy.
Noel R Williams MD
I think it’s more important to focus on what this trial accomplished.
CHALLENGE was a tough study to run. Long follow-up, behavioral intervention, limited funding, and a population that’s hard to engage over years. Despite all that, the team pulled off a large, international randomized trial that showed a meaningful benefit in recurrence and possibly survival.
Of course there are limitations. Maybe the effect size is inflated. Maybe adherence diluted some of the signal. But to me, this still reads as a positive trial. The direction of benefit was consistent, the outcomes were clinically relevant, and the intervention was safe and well tolerated.
As an oncologist, I spend a lot of time talking about toxic therapies with marginal gains. Here’s something that improves quality of life, reduces recurrence, and might help people live longer. We should be asking how to make this part of routine care, not looking for reasons to dismiss it.
This was a hard study. And an important one. Let’s build on it.
Med student ignorance here... But could reverse causality be a factor? With the elephant in the room being the massive amount of intensive wellbeing support the study arm is getting... Positive feedback loops all round... The actual role of exercise being moot. I presume there are 5 yr randomised controlled trials on intensive positive wellbeing initiatives in cancer patients??? Can't see drug companies running them...
Thank you John for the good points!
The major point about lack of objective changes and hence plausibility is not really valid. They showed an improvement of 2ml/kg/min in cardiorespiratory fitness and 22 m for the 30 m walk. All these are clinically significant (and statistically), especially considering the low function of the population. A decrease in body weight or waist circumference is obviously not plausible considering the small dose and absence of diet intervention. If they had measured body compositon, I believe they would have noticed a decrease in body fat, especially visceral fat which typically inflammation and insulin resistance.
I wonder, did the analysis screen and account for depression, or lack of it, as a confounding variable? Patients suffering from depression and anxiety after even successful cancer treatment would be less likely to join a study like this or to comply with a protracted, challenging fitness program. And depression itself might contribute to all-cause mortality through various channels.
Interesting study. But I agree with Dr. JMM. For me the concerns are the implausible effect size and implausible rapidity of onset of benefit, as well as the impractical/impossible nature of trying to reproduce the protocol in a practice setting. And it’s also a bit curious that the amount of exercise required to reduce some fairly hard endpoints in cancer was insufficient to produce much difference in measures of fitness.
But I think the greater short term risk is wellness/ “health” types misusing the data to encourage people to shun chemo etc and “just do some exercise”.
This trial and its outcome IS revolutionary. It’s revolutionary because it reveals that there are molecular mechanisms in cancer progression and in survival that have long been unexplored, and we need to identify their origins.
It’s revolutionary because it goes to the fundamental basis for what patients are seeking. Holistic approaches that allow them to lead a productive and meaningful life with a lethal disease rather than succumbing to it.
The slowed accrual, I believe, is really a result of deprioritizing of its effects since we ALL, over the last 7-8 decades, have been trained to build from molecular mechanisms of disease action up as opposed to working from enhanced survival down. Thus, it’s a welcome change back to that classic paradigm of clinical science.
In fact, perhaps all hypothesis generation should begin AT the BEDSIDE, and culminate in bench and cellular mechanisms.
This NEJM report also recognizes that targeted drugs aren’t the only way to both make patients feel better with a lethal disease but also enhance their survival. Thus, it shifts, or rather readjusts, the fulcrum of medical interventions towards making strides back towards survival, as opposed to surrogate molecular measures that often don’t pan out as predictive when working our way up from molecular mechanisms to clinical outcomes.
Raj,
You're on it! Everyone forgets basic cellular physiology let alone molecular intra-cellular operations. Exercise leads to a 50% increase in mitochondrial volume in 3 weeks. It's global and not specific to muscle. Thus in turn, a myriad of anti-neoplastic effects could occur ... without loosing weight, or building significantly more muscle mass, etc.
This study doesn't prove anything yet. It should challenge scientists to think more about impacting cellular activity through other means than just meds.
"Due to slow recruitment and a slower-than-expected event rate they changed to a five-year analysis. Yet they still had far less than the expected primary endpoint events (224 vs 380). This reduces statistical power and raises the possibility of false positive findings—which is consistent with the biological implausibility."
But low power increases risk of false positive results, not false negative ?
Or I'm mistaken ?
Im not a doctor but cant see how this study was proof of anything. I can see so many ways why it would be invalid.This would have to be so tightly controlled and have many more subjects for a more true curve. There are (as of course you doctors dont need to be told by me) so many factors contributing to whether or not a patient will relapse or stay cancer free. As soon as my un trained mind thought about the correlation between cancer relapse and exercise, I went into scoff mode. Sure a more healthy body is a plus so excercise can put you in better shape. Just from a statistical view, the number tracked and the methods used were a big disappointment after the potential of the subject.
John thank you for your assessment on this story. I saw a lot of the same things when I looked at the paper. There was a lot of fuss about it, but ultimately the effect was disappointing IMO. When I saw that it was mostly geared toward health education etc I couldn’t help but feel there is some fudging going on here. The other obvious thing is cancer prevention altogether…why don’t we take two cohorts matched for age, BMI etc one that exercises to a certain level and one that doesn’t exercise at all and just track cancer diagnosis. Because if exercise isnt preventative…should we expect it to be a “treatment?” I agree the results are promising but lots of lurking stuff there. I think the exercise effect could likely be higher but would love to see more prevention studies.
You have some thoughtful critiques here. Regarding the points you raise:
1. Implausibility. This statement is made without looking at available data. There is observational data on exercise after cancer which has relatively large effects on survival. There are also very few randomised trials of exercise with long follow up, so we don't really know.
2. Lack of objective exercise effects. There were some differences in VO2max for example, arguably the most important measure.
3. 'Early' separation in survival curves at 12 months. This is an arbitrary distinction and not a strong criticism. Colorectal cancer tends to relapse quite early at around 15 - 18 months. If exercise is going to have an effect then 'early' separation like this is to be expected.
4. Poor adherence. I think that the adherence was quite good for an exercise study in inactive 60 year olds. Not sure what the argument is here - because adherence was poor, then the effect is smaller than it appears?
5. This is valid, but lower than expected event rate is endemic in oncology trials. The fact that the OS separation is more obvious after 3 years reflects the survival kinetics of colon cancer.
6. I'm struggling to think how more contact with the study team and physical therapists would prevent cancer recurrence. Consider that on the one hand, you are saying it is implausible that exercise has such a large effect, while on the other hand trying to claim that increased attention has a large effect.
7. Some misapprehensions here. I don't think anyone believes that the specific exercise intervention in this study is what made the difference, aside from the fact they tried quite hard. Given everything we know about exercise already, a reasonable takeaway is "exercise makes a difference, do more than you are doing now" if you are receiving curative intent cancer therapy. The patients were younger than the median age of diagnosis of colon cancer, but the benefits of exercise if anything increase in older patients. It is wrong to suggest this doesn't have broad applicability.
I dispute that the cost of accepting this protocol would be 'massive'. No clinical trial protocol is implemented verbatim in the real world. Again, it is not the specific protocol that matters. A real world implementation would largely be getting people to exercise to the already recommended level, which doesn't seem overly onerous or expensive. I have little doubt that such an intervention would be cost effective. Cancer therapy gets more and more expensive, and people are also getting diagnosed at younger ages, so the marginal cost of each recurrence only increases over time. Additionally, exercise prevents other diseases which further improves the cost benefit analysis. And in any case, even if no attempt is made to implement formal exercise programs, on an individual level it is important that a patient knows they can do more to reduce their risk of recurrence.
What does concern me about this study is the potential for informative censoring - ie. people who dropped out of the exercise arm were also more likely to have a recurrence due to being generally less well or more frail. I still think that is quite unlikely to eliminate the observed benefits of exercise, however. Timothée Olivier looks at this on X (https://x.com/Timothee_MD/status/1929411436272382303), and the curves still separate.
I also wonder, as an 8th point: would simply the intensity of the in-person, mandatory intervention have its own effect? That's a lot of attention from other human beings. What other kinds of conversations happened during those interactions? Is there any data on the impact of social interactions on outcomes?
Thanks. Reminds me of the Tyler Cowen YouTube lecture "Be suspicious of stories".
Personally, I am not as bothered the lack of between-group differences in "objective" metrics like body weight, waist circumference, etc (although the exercise group did have > VO2max). It's totally plausible to me that the benefit of exercise is beyond these measures. But I agree that the lack of adherence to the protocol makes these findings pretty implausible. The self-reported MET increase was roughly half of their goal (6.3 vs a goal of 10) and that finding is very likely inflated due to observation bias. Also, only 20% of the exercise group did the recommended supervised exercise session in the first 6 months -what?? With that kind of adherence it is simply amazing to see an effect size this large.
You just wonder if being able to exercise was a marker for a better immune system and not a cause of a better immune system
My dad is 97 in assisted-living and there's only a few people his age but the ones that are all like him mentally intact. And that means he's less likely to fall was more likely to show up for the doctor, etc..