This interview is part of The Analytical Scientist’s feature exploring how analytical technologies are changing the science of sport – from metabolomics, microsampling, and wearables to anti-doping, recovery, muscle growth, and precision medicine.
As part of our feature exploring how analytical chemistry is changing the science of sport, we wanted to find out how analytical techniques are currently being used in hypertrophy research, whether that might change in the future, and, from the point of view of an athlete or coach, what chemical measurements would help lifters most.
Here, we put these questions to Eric Helms, who is: Co-director of the Sport Research Institute New Zealand (SPRINZ) and Senior Research Fellow at Auckland University of Technology – where he focuses on the science and practice of training and nutrition for physique and strength athletes – as well as a professional natural bodybuilder, coach, and all-round physical culture enthusiast.
How much of a role do analytical techniques play in hypertrophy research?
There is a little bit, but it is less directly about measuring hypertrophy and more about understanding potential mechanisms.
Someone like Mike Roberts at Auburn might look at epigenetic responses. Others might look at autocrine and paracrine responses – things like IGF-1 or MGF – or local signaling hormones. There is also work looking at systemic hormones, such as acute changes in testosterone, using assays.
With NIRS, you are looking at oxygenated and deoxygenated hemoglobin through changes in near-infrared spectroscopy. But the challenge is that this is a couple of steps removed from hypertrophy. First, it is not measuring hypertrophy directly. Then there is the question of why we are measuring deoxygenated hemoglobin in the first place. It is a proxy for, or arguably a measure of, “metabolic stress.”
So then the next question is: why measure metabolic stress? There have been hypotheses that metabolic stress could contribute to hypertrophy in certain contexts, but that is debated in the field. Its role probably falls into one of three categories. One is that it directly drives hypertrophy, but to a lesser degree than mechanical tension. Two is that it acts as an indirect contributor, where the metabolic status of the muscle changes motor unit recruitment – which may help explain why high-rep training to failure can produce similar growth to heavy, low-rep training or moderate-rep training. The third possibility is that metabolic stress is just correlated with hypertrophy-related processes, but does not actually contribute to or cause anything.
That is why I am skeptical about the utility of NIRS for hypertrophy specifically. You are measuring a questionable mechanism, and only one aspect of that mechanism, which may or may not be causative. That does not mean measuring metabolic stress is not potentially useful in other contexts. Metabolic stress is really a construct. What it reflects is reliance on anaerobic glycolysis.
When you are doing sets of four to six or four to eight reps, with long rest periods and low training density, you are primarily relying on existing ATP in the cell and the phosphocreatine system that recycles it. As you move into higher-rep training, higher-density training, supersets for the same muscle group, drop sets, sets of 15, 20, or 30 reps – all things that can still induce hypertrophy – you start to see a buildup of metabolites and fuel sources such as lactate that help keep the muscle contracting.
Those metabolites correlate strongly with the work being done. Because of how lactate is produced and recycled, lactate also rises alongside growth hormone. We have had a long-standing interest in those connections. I think the term “growth hormone” has probably led to too much focus on its growth-related role. If you look at the so-called hormone hypothesis – developed by William Kraemer in the 1980s and 1990s, and largely abandoned around 20 years ago – the idea was that transient increases in growth hormone, and to a lesser degree other hormones, were causative for hypertrophy.
When I started as a trainer 20 years ago, guidelines often recommended restricted rest intervals, moderate reps, and compound movements to drive up anabolic hormones, because that was thought to be one of the mechanisms driving hypertrophy. But those relationships turned out to be largely correlational. When people started doing controlled studies, finding ways to increase those hormones without changing other hypertrophy-related factors, it did not seem to have much impact.
So the acute hormone hypothesis has largely been abandoned. But the link between growth hormone, lactate, metabolic stress, and hypertrophy held on for longer. Now I think we are gradually chipping away at any direct causative role of metabolic stress as a mechanistic trigger for hypertrophy.
It may contribute on some small level, but in the grand scheme of things – at least in adults with normal physiological hormone ranges – hypertrophy seems to come down primarily to mechanical tension. That is not to oversimplify it. There are different types of tension, and stretch-mediated hypertrophy may even involve a different signaling pathway from traditional resistance training. There are multiple mechanosensors that sense tension. But studies looking at signaling pathways and epigenetic responses to high reps and low reps, including work by Mike Roberts, show quite similar patterns. You also get similar muscle proteins and similar whole-muscle hypertrophy.
All of that is to say that some spectroscopy techniques are not necessarily measuring something directly related to hypertrophy. They may be useful more broadly in sports science, because energy system use obviously matters. If you are working with an 800-meter runner, for example, it may be useful to know at what point they reach certain thresholds of deoxygenated hemoglobin, how long they can maintain them, or what pH their muscles can tolerate compared with someone untrained. That is not my area of expertise, but I can see an application.
Some cool work Takahiro Itagaki (a PhD student in our group) is doing – we have collected the data and plan to publish it – looks at subjective ratings of pump. We ask: what is someone’s actual feeling of pump on a Likert scale? Then we compare that with immediate ultrasound responses, which let us measure the acute hyperemia and hypohydration response in the muscle. In other words, what is the transient increase in muscle size, measured by muscle thickness, immediately after exercise?
We are also looking at the NIRS response – deoxygenated hemoglobin, total hemoglobin – and lactate, typically measured through a finger prick or earlobe sample. That gives blood lactate, not necessarily muscle lactate, which would be harder to measure and would probably require a biopsy. Then we are looking at the relationships between all of those measures using statistical modeling, including reliability and agreement.
We are interested in whether people are experiencing and measuring the same thing. A huge part of the supplement industry is built around blood-flow-enhancing supplements, based on the idea that the pump is important. That goes back to things like arginine, which has largely been discarded because it does not seem to reliably affect blood flow, and then supplements like citrulline malate or nitrate, which are supposed blood-flow enhancers.
But I think an important distinction is that enhancing blood flow and experiencing a pump are not necessarily the same thing. The most extreme pump you can experience is probably from blood flow restriction training. It is blood pooling – blood getting in, but not getting out – along with the accumulation of metabolites that create that burning effect. The muscle expands, and between reps or rest periods it does not get much opportunity to clear blood because it is cuffed.
Even normal training has an occlusive effect. That is why you feel more of a pump during high-rep sets, drop sets, or other approaches where the muscle does not get much chance to clear blood because it is contracting. When a muscle contracts, it compresses the veins and arteries.
People experience a pump when blood is getting in but not getting out efficiently. Blood-flow-enhancing supplements, by contrast, typically improve both inflow and outflow. Turnover is better. So the potential benefits of enhancing blood flow may be real – better fuel delivery, clearance, and recovery – but that is not necessarily the same as experiencing a pump.
The pump probably reflects things like an increase in ultrasound-measured muscle thickness, because the muscle is larger acutely, and perhaps deoxygenated hemoglobin, because that is a proxy for metabolic byproducts from NIRS. But whether any of that relates to long-term hypertrophy is the real question.
So it is an interesting study. We are assessing someone’s subjective ability to gauge metabolic stress and asking how closely, and how reliably, that maps onto constructs such as circulating lactate, immediate increases in muscle thickness, and changes in hemoglobin or deoxygenated hemoglobin in the muscle. Whether any of that predicts hypertrophy remains questionable – but it is a neat study.
From a practical perspective – for coaches, athletes, and well-funded sports programs – what chemical information is actually useful?
I don’t want to be too myopic, because I think we’ve operated for a long time – probably since the 1990s – with access to some of these kinds of data, but coaches have often concluded that they are not as useful as simply watching what is happening in training and asking athletes how they feel using validated questionnaires.
That said, the data are getting better, and wearables are improving. Even in the last five years, wearables have fundamentally changed in terms of accuracy and reliability. Now the bigger issue is that the market and manufacturers are moving faster than the science. Sometimes they are accurately and reliably measuring things and mapping algorithms onto outcomes that have not really been proven.
A good example is continuous blood glucose monitors. Yes, we can measure glucose continuously, but why? It is totally normal to have glucose excursions, so you can end up measuring something that does not matter very well. I think there is a lot of that going on currently.
What you typically see in high-performance sport is measurement of things that practitioners believe may have an impact on recovery. For example, repeated hormone draws to look at testosterone changes. That can be heavily misinterpreted, but in the right hands, someone might use it more as a marker of recovery or as a prompt to ask more questions, rather than as something with a direct impact on hypertrophy within normal physiological ranges. In those ranges, testosterone is not very predictive.
Right now, none of the physiological measures of hormones or metabolic byproducts are especially useful for hypertrophy, because acute responses to training do not reliably map onto long-term hypertrophy. They are hard to measure, and people – especially commercial app developers – do not necessarily have a good understanding of whether they are even causative factors.
There is still debate among scientists. I could sit with Brad Schoenfeld, Mike Roberts, and Stuart Phillips, and we would probably have minor disagreements about some of these edge cases: the roles of hormones, signaling factors, and related mechanisms in hypertrophy. That shows we are at the edge of our current mechanistic knowledge.
Our ability to measure some of these things has moved beyond our mechanistic understanding. That is why we are in this strange space where hormone coaches and biohackers can measure these things, but many of the measurements may not actually matter – continuous glucose monitoring is a good example. Our technology has advanced past our theoretical understanding of the mechanisms, at least for this physiological outcome, which is a weird place to be.
From a more speculative point of view, if we did eventually identify a useful set of biomarkers for fatigue, would that help coaches or athletes?
Yes, but fatigue for resistance training, mental or psychological fatigue, and fatigue in something like endurance sport are all different and interrelated.
Systemic fatigue is important for resistance training because our muscles are connected to our brains, our body, our digestive tract, and everything else. But the surprising thing is that we have never really observed a case of overtraining syndrome – defined by an inability to perform despite training, when that decline cannot be explained by anything else – purely due to resistance training in the literature. It is almost always related to under-eating at the same time, or concurrent training.
Even among endurance athletes and team-sport athletes, only a small proportion of potential overtraining syndrome cases – where you see a sustained reduction in performance despite consistent training, and it does not resolve within months – end up being true overtraining syndrome. In many cases, it turns out to be something else.
So while I do think measuring fatigue could be useful, in many cases it is not necessarily the bottleneck for hypertrophy, despite some claims to the contrary.
What might be more helpful for hypertrophy would be some quick measurement of muscle volume or thickness. We are getting quite good at image capture. You could imagine standing in the same spot, setting your phone up from a few angles, moving through space, spinning, and having the phone create a 3D model that captures circumferences. There are already some impressive efforts to turn phones into 3D imaging systems. You could track whether your waist stays relatively stable while your arm or thigh circumference increases over time, which would be a good sign.
If we were talking in a Star Trek-style hypothetical, then something that measured how many grams of protein are being accreted or delivered to your muscles on a regular basis would be fascinating. At-home muscle protein synthesis would be very interesting: which workouts produced higher or lower levels?
Right now, muscle protein synthesis – at least in the short term – has to be manipulated a lot before it predicts hypertrophy. Gross muscle protein synthesis can increase because of muscle regeneration and damage, which does not necessarily mean hypertrophy. You might see an acute response, but then the person does not eat enough protein for the next four days, starts dieting, or does not train again. So many other factors predict long-term hypertrophy.
But if you had a continuous analysis of muscle protein synthesis, and you also knew someone’s maximum rate and rate of breakdown, then you could know more about how effective the training was. Hypothetically, there are things we could measure, but it would almost require a “God machine” measuring all the muscles in your body. A constant measure of mTOR activation would also be interesting. In theory, it could tell us whether we were in an anabolic state and whether the training was effective.
But hypertrophy is difficult because it is a physiological process rather than a performance adaptation.
Measuring performance and the determinants of performance is easier than measuring the physiological change in muscle mass. Endurance, power, and strength all have clearer outputs and determinants. Hypertrophy influences those outputs, but it is not always the final outcome itself.
Unless you take a muscle biopsy before and after training, it seems very difficult to know whether a given intervention has actually worked. Is that a fair way to think about the problem?
Yes, and I would add that you can become increasingly accurate and less invasive, but it is still a snapshot. Hypertrophy is an ongoing process rather than a performance outcome at a single point in time.
In sport, you might want to know whether someone is going to run a personal best in a 10-meter sprint tomorrow. You may be able to predict that with reasonably high accuracy based on prior data. But with hypertrophy, there is not really a “peak” in the same way. It is more like blood pressure or another physiological marker. You do not see big acute ups and downs.
With hypertrophy, you are usually seeing changes in grams over weeks, unless you are looking at beginners, muscle wasting, immobilization, or astronauts. In most cases, the people who are most interested in optimizing hypertrophy are already doing enough to be bigger. The question becomes: how do I go from gaining 80 percent of the maximum amount of muscle I could gain to 85 or 90 percent, when the rate of growth is already relatively slow?
So yes, you can get the precision to measure some of these things, but hypertrophy is a continuous process. The time scales involved make acute measurements almost invalid – not because they are inaccurate, but because they may not tell you what you need to know.
If I could get a full profile of everything happening in the blood, or even in a very small amount of muscle tissue through a microbiopsy, the question would still be: what if I do not eat tomorrow? Those measurements would need to be continuous, representative, accurate, and actionable. Those are high bars.
You would need some kind of algorithmic assessment using multiple measurements, followed by good interpretation. Unless the measurement gets very close to the thing we actually care about – muscle size – there is a risk of interpreting it incorrectly if your understanding of what influences hypertrophy is flawed. If you are convinced that fatigue or hormones are the most important factors, when I would argue they probably are not, then you may make changes based on an assumption rather than on the actual driver.
So if there were a move toward more useful measurements that could change end-user decisions, I think it would involve increasingly accurate remote anthropometry: the ability to get a muscle-thickness measurement from your phone, to use imaging software to estimate reliable circumferences or body fat, and to gauge whether you are training as hard as you think you are.
We are not far from some of those things, but they are not necessarily what analytical scientists are working on.
Overall, how well do we understand hypertrophy?
It might be surprising to some readers to realize that we still do not fully understand how hypertrophy occurs, even though we have drugs that can induce it and are developing more drugs that can induce it. But those drugs are often working through different mechanisms than lifting weights.
But if you were to ask the most mechanistically focused hypertrophy researcher why trained lifters eventually stop growing, the honest answer would still be somewhat unclear. Is it an energy supply issue? Is it related to myonuclear domain theory? What is the actual limiting factor? And if you then ask why anabolic steroids allow someone to surpass that limit, the answer becomes even blurrier. We can say the system is being supersaturated, but there is still a lot of uncertainty around the actual mechanisms of hypertrophy.
As I mentioned earlier, we have identified some mechanosensors that we think stimulate hypertrophy, and we have some understanding at the signaling level. But if you really try to pin all of this down, there are still disconnects between the most mechanistic understanding, the training variables, and the observations from applied research.
What we are trying to do is connect the kind of data produced by someone like Stu Phillips or Mike Roberts with the kind of data produced by people like me or Brad Schoenfeld. In the middle, you often see papers saying, from the mechanistic side, “We looked at these mechanisms, and this might apply to these training studies.” Then, from the training-study side, we say, “If we think these mechanisms operate in this way, then we should observe this when we do a training study.” And sometimes we sort of do, and sometimes we sort of do not – and then we try to explain why.
All of that is confounded by the fact that we are usually trying to measure small changes over short periods of time, often in small sample sizes. A lot of the advances we have made recently have come from having 30 years of accumulated data, and from the field evolving to the point where we now have the statistical ability to do better meta-science.
It is a challenging problem because, in sports science and sports nutrition, we are inherently interested in small changes in people who do not change very much. And sports science is not funded like sarcopenia research or broader health research. We do not get to run large multi-site trials. Some medical drug trials have more participants than our largest meta-analyses.
In some cases, it may not be realistic to expect science to provide individual-level answers. What we can do is describe relationships, general responses, and mechanisms, and then pass the baton to the sports scientist or coach at that second level I was talking about, where monitoring may have more utility.
But again, unless we are in Star Trek territory, the practical value will probably come from leveraging the relationships we develop in science. We might understand, more clearly, the relationships between volume, frequency, proximity to failure, fatigue, perceived exertion, and mental fatigue. We might develop a broader theory of how much training someone should do, when they should stop, which levers they should pull, and what they need to watch for.
Then we can monitor some of those inputs and look at outputs such as anthropometry. There is still a black box in the middle: the actual physiological variables driving hypertrophy, and the process of building muscle itself. We do not really have that in a practical setting yet. What we can do is measure reliable changes over reasonable time frames, and quantify training inputs more accurately.
But I think we are still probably decades away from being able to get inside that black box in a practical setting.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
