Why we don't use strength-endurance anymore
From an article by Dr Andy Coggan, reproduced with his permission, entitled "Strength endurance" training: a physiologist's view
For at least 20 y, some coaches have advocated that cyclists include so-called “strength endurance” (SE) intervals as part of their training program. While the precise format of these intervals tends to vary, in general they consist of pedaling for extended periods (e.g., from 5 to 20 min) at a moderate to high intensity but at an abnormally low cadence (e.g., 45-75 rpm). More recently, this form of training has gained popularity among triathletes under the rubric of “big gear” training. Based on anecdotal evidence, the proponents of this form of training have faith that it improves performance, although there seems to be no consensus on the particular aspect of performance that it is thought to benefit.
It is difficult to argue with the anecdotal evidence in favor of SE training, primarily because it is just that: anecdotal. Ideally, however, coaching practices should be based not upon anecdotal evidence and empirical observations, but upon formal scientific research in the sports sciences, or at least the application of scientific knowledge to sport (1). The purpose of the present article is therefore to examine SE training from such a perspective, using a graphical means of analyzing powermeter data that I have termed quadrant analysis (QA) (see http://home.earthlink.net/~acoggan/quadrantanalysis). Specifically, QA was used to assess the two most plausible physiological mechanisms by which SE training could potentially be beneficial, i.e., 1) as a form of on-the-bike resistance training that results in muscular hypertrophy and thus possibly increases maximal neuromuscular power, or 2) as a means of enhancing the recruitment of type II (fast twitch) motor units during prolonged bouts of exercise, thus improving their resistance to fatigue via increases in mitochondrial content, capillarization, etc.
Training sessions analyzed
To provide the necessary data for the present discussion, I performed two prototypical SE workouts, i.e., 2 x 20 min at 250 W and 45 rpm and 5 x 5 min at 300 W and 45 rpm. These power outputs were only 90% of that which I would normally maintain for these durations in training, but in fact they were nearly maximal considering the markedly suboptimal cadence. The sessions were performed on a Velodyne trainer (http://www.velodynesports.com) operating in ergometer (i.e., constant power) mode, with data for power and cadence recorded every 1 s using a carefully calibrated SRM Professional powermeter. To mimic the inertial load encountered when performing such intervals up a moderately steep hill, I used a 53x12 gear combination, and I applied a “no gripping” rule as advocated by some coaches to minimize the recruitment of upper body musculature. In addition, for comparison purposes I have included data from three other trainings sessions, namely 1) six maximal standing start efforts performed on a flat road in a 53x12 gear combination, starting from a track stand and continuing until I reached ~90 rpm (which took ~15 s), 2) ten maximal seated accelerations in the same gear on the Velodyne, starting with the cranks/ wheel/flywheel stationary and continuing until I reached ~75 rpm (which took ~10 s when working against the inertial load provided by the Velodyne's 10 kg flywheel), and 3) 2 x 20 min of 15 s on, 15 off “microintervals” performed on the Velodyne at a power of ~500 W during each on period and ~50 W during each off period.
SE training to increase strength
One factor that may have contributed to the increasing popularity of SE training is the growing realization among cycling coaches that, due to the specificity principle, strength per se generally plays very little, if any, role in determining endurance cycling performance (i.e., power output), and that as a result, more traditional forms of resistance training (e.g., weight lifting) at best provide limited benefits (see http://home.earthlink.net/~acoggan/misc/id4.html for a more detailed discussion of this issue). Thus, in an attempt to provide a more specific form of on-the-bike resistance training and/or to “help transfer the gains made in the gym to the bicycle”, many coaches have apparently turned to SE training. Noticeably, however, it appears that few, if any, have actually considered whether the forces generated during such training are in fact sufficient to truly represent an overload, such that beneficial adaptations would be expected to result. Instead, it appears that that most have simply assumed that just because the cadence is lower, and hence the pedaling forces higher, than usual that there will an increase in muscular size and strength, and hence in, e.g., maximal power.
To address this issue, I used QA to compare the average effective pedal force (AEPF) during the two typical SE sessions described above to the maximum that I can generate at zero velocity (i.e., the isometric strength of my pedaling musculature) and at all velocities above zero. The latter data were obtained during testing using an inertial load ergometer as described by Martin et al. (2). As expected and as shown in Fig. 1, pedaling at 45 rpm during the SE workouts resulted in an AEPF that was approximately twice that required to pedal at the same power(s) but at my normal self-selected cadence of 85-90 rpm (i.e., the horizontal and vertical dashed lines defining the quadrants). Nevertheless, the AEPF during the two SE sessions still only required ~50% of my strength, i.e., equivalent to weight training at only 50% of one repetition-maximum (50% of 1 RM). Indeed, even when the decline in muscle force with increasing cadence is taken into account, the AEPF during the SE workouts was still less than two-thirds my velocity-specific maximum force. This result may at first seem surprising, especially to anyone who has performed such training and has felt that they were pushing very hard on the pedals. However, it is consistent with the fact that I was able to perform 1125-1800 “reps” (i.e., 5 x 5 or 2 x 20 min at 45 rpm) with this load. Furthermore, when you consider that peak force on the pedal is roughly twice the average force (i.e., AEPF), then I was not pushing any harder with my leg extensors than I do when, e.g., performing step-ups with only my body mass for resistance (i.e., 68 kg x 9.81 N/kg = 667 N). The SE sessions were therefore actually much more akin to climbing many, many flights of stairs than they were to traditional weight training, and as such would be unlikely to result in any significant hypertrophy and/or strength gains, at least in an athlete already performing other forms of training. (This is especially true since for most people the discrepancy between the AEPF that they could sustain during an SE interval and their maximal force generating capacity is likely to be even greater. This is so because my muscles generate less force than average at all velocities above zero, but have higher-than-average muscle fatigue resistance, due to a high percentage of type I muscle fibers.) Consistent with this conclusion, a recent study reported on the internet (3) found that training in this manner twice per week for 8 wk did not result in any increase in either the size (estimated using anthropometry) or force output (determined using isokinetic dynamometry) of the leg extensor muscles.
SE training to increase type II motor unit recruitment
Another potential mechanism by which SE training has been hypothesized to improve performance is via enhanced recruitment of type II (fast twitch) muscle fibers, which induces greater physiological adaptations in these fibers than training at a normal cadence. It is difficult to envision how this might occur, however, since even in endurance athletes not performing SE training type II fibers are sufficiently recruited and hence trained to cause essentially complete replacement of type IIb (IIx) fibers by type IIa and to result in a mitochondrial respiratory capacity nearly equal to that of the same athlete’s type I fibers (e.g., Ref. 4). Moreover, the basic premise that SE training results in markedly greater utilization of type II fibers may not be correct, as discussed below.
Again consider the QA plot of the two SE training sessions that I performed. As illustrated in this figure, when pedaling at 250-300 W at 45 rpm my AEPF and CPV clearly fell into quadrant II, i.e., relatively low velocity and relatively high force pedaling. Since the AEPF was greater than that resulting from pedaling at functional threshold power and self-selected cadence, this would imply that significant recruitment of type II motor units must have occurred. As previously discussed, however, my AEPF during the SE workouts was still well below my velocity-specific maxima (shown by the black line), which presumably reflects maximal recruitment of all motor units, i.e., both type I and type II. Thus, based on this observation alone it is clear that either a significant percentage of my type II muscle fibers were not recruited by SE training, and/or they were not recruited as frequently as is possible, at least during brief, maximal efforts. In other words, while the AEPF associated with pedaling at functional threshold power and self-selected cadence represents the approximate threshold force at which significant type II fiber recruitment appears to begin, as indicated in the original description of QA the exact force varies in a velocity-specific manner, i.e., is higher at lower CPV and is lower at higher CPV. Or, to state it another way: the AEPF threshold for type II motor unit recruitment is better represented by the blue-to-red background shading of the figure, with greatest utilization of such motor units expected to occur when points fall into the purple region. (Again, as described in the original article the separation of quadrants I/II from III/IV using a strictly horizontal line is really just a matter of convenience, since no “anchor point” other than functional threshold power/self selected cadence is usually available.) Viewed from this perspective, it is now much less clear that SE training results in significantly greater utilization of type II muscle fibers, since the light blue triangles and purple circles representing these two workouts actually fall well into the blue region. In fact, it can be argued that if SE training resulted in markedly greater reliance on type II fibers that it would not be possible to maintain the required AEPF, and hence power, for the required duration, since type II fibers are more dependent than type I fibers upon glycogenolysis for ATP production and hence tend to be more fatigable. Consistent with this interpretation, Alquist et al. (5) found that the fiber type recruitment pattern was markedly similar when subjects pedaled for 30 min at 85% of VO2max and either 50 or 100 rpm.
Alternatives to SE training
If reducing cadence by approximately one-half when pedaling at a submaximal power output is insufficient to result in a true overload on the muscular system, then what will? The answer to this question, of course, is pedaling at an even higher power output, e.g., performing maximal or near-maximal efforts, especially from an initially low cadence (what Dr. Martin has long referred to as “special force” training). This is clearly illustrated in Fig. 2, in which I have plotted the AEPF-CPV data for the six standing start efforts that I performed on the road and the 10 x ~10 s maximal accelerations that I performed on the Velodyne along with the results from the SE training sessions. (Note that for clarity I have not plotted data collected between efforts, e.g., when pedaling easily back to my starting point during the workout on the road.) As can be seen in this figure, in contrast to the clearly submaximal forces produced by (during) SE training, the AEPF during the standing start efforts and seated accelerations were markedly higher. In fact, the AEPF during the initial phase of the standing start efforts were significantly higher than my theoretical velocity-specific maxima as determined using the inertial load test. On the other hand, my AEPF during the seated accelerations tended to fall below that expected from this testing. While these differences could be partially due to the different measuring techniques used (i.e., on-bike SRM powermeter vs. computer-interfaced ergometer), they are probably mostly due to 1) the fact that I was out of the saddle during the standing start efforts, vs. seated at all times during the inertial load test (6), and 2) the fact that I was unable to reach optimal cadence before fatigue begain to develop during the seated accelerations. In any case, however, it should be clear that this form of training represents a far more significant load on the neuromuscular system, and hence is much more likely to result in an improvement in neuromuscular power, than SE training. Indeed, this past summer I was able to increase my maximal 5 s power - tested by performing a maximal acceleration in a small gear, thus mimicking the inertial load method - by 10% in just 4 wk by performing standing start efforts as described above on a weekly basis. Similarly, an elite female track cyclist increased her maximal 5 s power by 25% in just 8 wk by performing similar workouts several times per week when attempting to peak for an important competition.
While, e.g., standing starts, maximal seated accelerations (termed “power stomps” by Chris Carmichael), etc., are clearly superior to SE training when it comes to overloading the neuromuscular system, they are obviously too short to result in significant improvements in the metabolic fitness of type II fibers. As previously discussed, however, it is unclear whether any special sort of training is needed to maximize this physiological adaptation. Nonetheless, if specific training of these more-difficult-to-recruit motor units is for some reason deemed necessary or desirable, then one possible way of doing so might be by performing microintervals, i.e., very short (15 s or less) on/off efforts, for an extended period of time. The impact of training in this manner on the AEPF-CPV relationship is also shown in Fig. 2, which demonstrates that while in an absolute sense the AEPF during such intervals was not quite as high as that seen during SE training, during the on periods it was actually closer to my maximal force-velocity line (i.e., closer to the purple shaded region), due to the velocity-dependent reduction in force generating ability. In other words, such intervals are actually more likely to result in significant recruitment of type II motor units than SE training, because they demand a greater fraction of my (or anyone’s) maximal velocity-specific force. Of course, another advantage of microintervals over SE training is that microintervals are performed at a normal cadence, thus fulfilling the specificity principle with respect to muscle contractile velocity and minimizing any risk of injury due to an abnormally elevated force requirement at an abnormally slow cadence as is the case with SE training . To my knowledge, however, the physiological adaptations that result from performing microintervals on a regular basis has never been studied.
While SE training is popular among certain groups, peer-reviewed scientific studies demonstrating the superiority of this form of training over other, more traditional forms of training for cycling are lacking. When this question is approached from a first principles/physiological perspective, however, it seems unlikely that SE training would result in significant increases in muscle size, strength, or power, at least in a cyclist who is already performing significant amounts of training at a normal cadence. This is so because the forces generated, while higher than normal, are still too low to represent a significant overload to the pedaling musculature. On the other hand, the forces generated during maximal efforts of brief duration performed at a low (initial) cadence, e.g., standing starts, or seated “stomps” in a large gear, are roughly comparable to those typically encountered when, e.g., training with weights. Consequently, this type of training would seem far more likely to result in increases in muscle size and strength, and hence possibly in maximal neuromuscular power. The latter is especially true because such on-the-bike resistance training is highly specific to the joint angles, overall muscle recruitment patterns, etc., that occur during pedaling – only the velocity of muscle contraction differs from that of “normal” pedaling, and even that becomes comparable if such efforts are continued long enough for cadence to rise into the normal range.
In contrast to the failure of SE training to result in a significant overload in a neuromuscular sense, it is less clear whether such training enhances demand on the metabolic “machinery” of type II muscles fibers. Even if does, however, there seems to be no reason to believe that this form of training would be superior to training at a normal cadence, since the latter 1) already results in significant recruitment of type II motor units, at least if the intensity is high enough (i.e., ≥ functional threshold power, 2) better fulfills the specificity principle, and 3) minimizes the risk of injury. If, however, it is deemed advantageous to attempt to specifically overload these motor units, then other forms of training such as microintervals, LT intervals performed following pre-depletion/pre-fatigue of type I motor units, etc., seem likely to be at least as effective as SE training, while also being more specific to the actual demands of racing and less risky from an overuse injury perspective.
1. Rushall BS. Coaching development and the Second Law of Thermodynamics (or belief-based versus evidence based coaching development). Editorial published online at http://www-rohan.sdsu.edu/dept/coachsci/csa/thermo/thermo.htm
2. Martin JC, Wagner BM, Coyle EF. Inertial-load method determines maximal cycling power in a single exercise bout. Med Sci Sports Exerc 29:1505-1512, 1997.
4. Chi MM-Y, Hintz CS, Coyle EF, Martin WH III, Ivy JL, Nemeth PM, Holloszy JO, Lowry OH. WADE H. MARTIN III, JOHN L. Effects of detraining on enzymes of energy metabolism in individual human muscle
fibers. Am J Physiol 244: C276-C287, 1983.
5. Ahlquist LE, Bassett DR Jr, Sufit R, Nagle FJ, Thomas DP. The effect of pedaling frequency on glycogen depletion rates in type I and type II quadriceps muscle fibers during submaximal cycling exercise. Eur J Appl Physiol Occup Physiol 65:360-364, 1992.
6. Bertucci W, Taiar R, Grappe F. Differences between sprint tests under laboratory and actual cycling conditions. J Sports Med Phys Fitness 45:277-283, 2005.