I received some very good questions about a fundamental axiom underlying my Muscle-Specific Hypertrophy articles, namely that slow twitch fibers have a higher optimum volume for muscle growth than fast twitch fibers. Is that really true?
Note to readers: If you already believe this to be true, you can stop reading here, because I won't show you anything new. Although if you're interested in how to interpret findings from empirical studies, you may find the information below to be of use.
Update 11 Nov 2011: This article also found preferential slow-twitch fiber hypertrophy for low-intensity training and vice versa.
Physiological effects of using the low intensity strength training without relaxation in single-joint and multi-joint movements. Netreba AI, Popov DV, Liubaeva EV, Bravy? IaR, Prostova AB, Lemesheva IuS, Vinogradova OL. Ross Fiziol Zh Im I M Sechenova. 2007 Jan;93(1):27-38.
The classic rationale for this is as follows: Muscle fibers are recruited according to the Size Principle, meaning faster twitch fibers are only called upon when the resistance becomes too heavy for the slower twitch fibers. Muscle fibers need to be fatigued, not just stimulated, in order to grow (see Zatsiorsky's work). Therefore, fast twitch fibers need a higher intensity to grow than slow twitch fibers. As intensity is inversely correlated with volume, slow twitch fibers require a higher volume for optimal muscle growth than fast twitch fibers.
Theory is always interesting and all, but empirical data trumps everything and the review by Fry (2004) from my references was noted by some readers (props to markrot and HiThere111 for being sceptical) to be at odds with the above axiom. Look at the graphs from that paper below.
You may conclude from it that an intensity of 80-85% 1RM is optimal for both slow and fast twitch fibers. However, there are a few very important limitations of this paper that need to be addressed before interpreting these findings.
1. These graphs only show intensity, while many of the studies in this meta-analysis control for volume. Let's compare 2 study protocols for the leg press: 10×3 (~6RM) and 3×10 (~12RM). Of course 10×3 will yield superior results: it uses more weight and equal volume. As the meta-analysis only looks at intensity and hypertrophy, it concludes 6RM > 12RM. That obviously cannot be concluded irrespective of other training variables. For example, 1×3 at 6RM won't do much for you, but DC training at 12RM definitely will. A more interesting study would be the effects on fiber-specific hypertrophy of 3×3, 3×7 and 3×11, all to failure with full rest in between sets.
In short, because the meta-analysis did not exclude studies controlling for volume, it cannot be interpreted in full (or at all).
2. The interpretability limitations due to controlling for volume and other training variables can be further shown by looking at the graphs again without being biased by the fit lines. Look at the graph on the right. Ignore the fit line for a moment. If the 4 studies on the right were excluded, the fit line through the remaining points would have been virtually vertical. This would be interpreted as the same intensity resulting in highly variable results and this makes sense, considering intensity is only a single part of a program. If the 2 studies on the top were excluded, the line would be almost horizontal. This would be interpreted as there being no effect of intensity on hypertrophy at all in type I fibers. Indeed, only 18% of the variance in the data could be explained by the fit line. Of course you cannot go around excluding studies for no reason (I happen to have a Minor in Methodology & Statistics), but the 2 studies at the top are rather suspect. Their listed muscle growths are high even for type II fibers. Since excluding those would result in an almost horizontal fit line, drawing the current linear fit line would probably net you an F in a statistics course.
In short, the relation between intensity and type I fiber hypertrophy is, as far as it can be interpreted, weak: weaker than 18% of explained variance (less than half that of the type II fibers).
This is all not to say that these graphs have no meaning. They do, although it is not readily apparent. Consider the following more points.
3. The used studies were all conducted on the vastus lateralis of the quadriceps of untrained (!) persons, which means the intensities are strongly positively biased. There's no way a beginner can display a true 1RM. They're too neurologically inefficient and unaccustomed to using their type II fibers. Hell, practically all these studies only used leg extensions and leg presses. No one can display a true 1RM on a leg extension, because you have to start from a dead stop and can't use any stretch reflex. And untrained persons, some of them females, displaying a true 1RM the first time ever on a leg press? It takes brass balls to do that.
What does this mean? It means the intensities listed in the graph are overstated. 80% is more like a ~15RM in most of these studies, which is evident if you look at the protocol results and see they can do 10×8 with 80%, only reaching failure on the last set. Try GVT and you'll see that's impossible.
My Conclusions of Fry (2004)'s Meta-Analysis: Without taking volume into account, which makes sense for beginners who can still recover from most programs, there does seem to be an optimum intensity at about 80% 1RM for the type I fibers, whatever amount of reps this corresponded with. Going all the way up to 95% 1RM doesn't provide additional muscle gains. However, the relation between intensity and hypertrophy is weak, which means it doesn't matter that much which intensity you use. This also makes sense, as type I fibers are easily activated and hard to fatigue. Intensity is less important than volume for them.
The relation between intensity and hypertrophy of type II fibers is far stronger and its optimum is probably not at 80% of 1RM, not even at 8RM. I think it's higher and the meta-analysis cannot provide any information on this matter, as the study with the highest used intensity was 80% of 1RM.
As such, these results do not falsify the traditional axiom that type I fibers require a higher volume for optimal growth than type II fibers.
In fact, other information in this same article indicates the axiom is true. Weightlifters and powerlifters have muscle fiber areas composed of roughly 2/3 type II fibers and 1/3 type I fibers. In bodybuilders it's the other way around: 2/3 of their muscles' cross-sectional area comprises slow twitch and only 1/3 of it is fast twitch. Now, correlation does not equal causation, so there are basically two explanations for this: nature or nurture. Maybe people with specific fiber type compositions just make better bodybuilders and other people make better power/weightlifters. That doesn't make sense though. Type II fibers are not just stronger: they also grow much faster and are far bigger. Nurture seems like a better explanation. The differences in preferential fiber type hypertrophy is due to the rep ranges they use. Using mainly loads >85% 1RM is great for the fast twitch fibers, but it leaves the slow twitch fibers understimulated. Bodybuilders maximize muscle mass in both fibers.
As for the optimal training parameters for hypertrophy, including intensity, I will discuss this in a future article series and I promise you I will do a better job than Fry.
p.s. It's worth noting that muscle fiber type composition and intensity are interrelated. For the soleus (super slow twitch dominant), you can expect to be able to do a whole lot of reps at 80% 1RM, whereas for the hamstrings (super fast twitch dominant) you most likely can't do more than 7. Go try it yourself on seated calf raises and leg curls if you don't believe me.