Interval training: how to improve your fitness and increase your VO2max

Cyclists, swimmers, rowers, cross-country skiers, orienteers, triathletes and runners all engage in interval training in order to increase the amount of time they spend exercising at very high intensities. A runner scampering along without stopping at his/her current best 10k velocity might be able to sustain the pace for only 25 minutes-or-so during a workout; by breaking the effort down into eight-minute intervals, however, the same runner could often work at 10k velocity for a total of 32 minutes (four eight-minute intervals), thus boosting the 'quality' of the session (the time spent above lactate-threshold velocity) by 28%.Although critics carp that interval sessions are unrealistic and not specific to competition (few races feature the recovery periods which are characteristic of interval workouts), one can find little fault with the way they enhance the quality of training. Training at intensities above 90% VO2max is one of the most potent routes to fitness, and intervals allow athletes to enter this 'red-line' training zone consistently and productively - for augmented periods of time.

There is some debate about the origins of interval work, but it's likely that the Finnish runner Hannes Kolehmainen was the first elite athlete to employ intervals consistently within a comprehensive training program. Kolehmainen, an Olympic gold-medal winner in 1912, liked to perform intervals at race pace, and was known to use a workout consisting of 5-10 repetitions of 3:05 per 1000m - a tempo of 74 seconds per 400m, or 19.5k/hour, which was very close to his 10k race speed (1). Training at race pace has remained a useful tool for athletes in a variety of sports; it is believed to enhance metabolic efficiency and boost mental confidence at race-specific velocities.

Another great Finnish runner, Paavo Nurmi, who captured four gold medals at the 1924 Olympics and once set three world records within a 90-minute time span, pioneered the practice of using higher-than-competitive intensities during interval workouts. The sturdy Nurmi, who dominated distance running in the 1920s, ran 5000m in 14:36, or 70 seconds per 400m. (It's interesting to note that at that pace he would be over 700m shy of the finish line when the current world record-holder crossed the tape.) But he liked to reel off a succession of 400m intervals in 60 seconds each within the overall context of a 10-20k run on wooded trails (op cit). A few simple calculations reveal that, at 60-second per 400m pace (6.67m per second) Nurmi was running about 14% faster than his 5k race tempo at almost 110 percent vVO2max, an intensity still favoured by endurance athletes seeking to improve their maximal cycling, swimming, rowing, skiing or running speeds.

About 15 years after Nurmi was in his prime, a tree trimmer from Sweden began to employ a new form of interval training with great success. Gunder Hagg liked to alternate hard intervals of speedy running with periods of easy coasting along the forested trails of central Sweden. Coached by Gosse Holmer, the 'father' of fartlek training, Hagg liked to complete at least 10k of fartlek intervals per day (often broken down into two 5k 'pieces'), and the speedy work paid off: Gunder set ten world records at seven distances during 1942 (2).

Zatopek's amazing 100 x 400m reps

In spite of the great success of Kolehmainen, Nurmi, and Hagg, interval training didn't really take off until after World War II, and it was not a Finn but a notable Czechoslovakian athlete, Emil Zatopek, who perhaps did the best job of convincing the athletic world at large that intervals represented an unparalleled training technique. Zatopek used a somewhat unusual interval combination: he ran relatively short repeats, often no more than 400m in length, but instead of blazing through these relatively short interval distances at high speeds, he ran them at close to his lactate-threshold velocity, ie well below both 10k and 5k race speeds (3).

The hard-working Czech was known to rattle off 100(!) of these 400m reps per day, alternated with 200m recoveries. Scientists later reckoned (op cit) that Zatopek reached lactate-threshold speed at about 85% of his vVO2max (ie at a pace of about 72 seconds per 400m), which meant that on a 100-work-interval day he was completing about 40k of work intervals in just 120 minutes or so. (Before you become completely incredulous, don't forget that there were 39 'recoveries' injected into such running to make the 100 x 400 scheme more palatable.) Zatopek's pattern of churning out high volumes of moderate-quality (below 90% VO2max) training was echoed, albeit on a less grandiose scale, by the American runner Jim Ryun in the late 1960s and early 1970s.

Interval training made its way into a scientific journal for the first time in 1959, when German exercise scientist H Reindell and colleagues described specific interval training carried out by top athletes and formulated some basic rules for interval work (4,5). Sigfried Hermann, an athlete described in Reindell's book, specialised neither in Zatopek's 'large bag' of slow intervals or Nurmi's 'short stack' of scalding reps; he engaged in what could be called 'variable-pace' interval training. A 3:40 1500m runner, Hermann would - within a single workout - run four sets of 6 x 200m, with rests of 50-60 seconds between the 200m work intervals and eight minutes between sets. Sigfried completed the first set at a tempo of 30 seconds per 200 (or 98% of 1500m speed), the next two sets at almost-exact 1500m race velocity, and most of the final set at 28 seconds per 200 (about 105% of 1500m race tempo). The German whizzed through the very last interval of the last set in 25 seconds, which was 118% of his race velocity over 1500m.

This variable-pace interval training, performed above and below actual race velocity, was mirrored in the early 1980s by famed British coach Frank Horwill, the founder of the British Milers' Club and mentor of two-time World Cross Country silver medallist Tim Hutchings. Horwill, however, liked to use the different paces on separate days of training, a scheme he called 'multi-paced training'. Peter Coe later based Seb Coe's overall training on Horwill's multi-speed foundation. Variably-paced intervals were also employed with great success in the 1980s by Said Aouita of Morocco, who held world records for both the 1500 and 5000m and ran at speeds ranging from lactate-threshold velocity all the way up to 1500m competition speed within the same session, with interval distances ranging from 200-3000m.

Broad-intensity training builds speed and stamina

This 'broad-intensity' (ie variable-pace or multi-paced) interval training has become increasingly appealing, mostly because it is believed to build both speed and stamina, but also because it is known that athletes seldom move at a constant pace during their competitions, even during world-record efforts. Indeed, French scientists Veronique Billat and Jean-Pierre Koralsztein have shown that top performances in distance races actually consist of series of 'wavelets' - regular, well-defined periods of high intensity alternated with sequences of lower overall power (6). In theory, to enhance efficiency over the range of paces employed over a single race distance, one would have to 'rehearse' the tempos properly during training, a process most conveniently carried out during an interval workout.

Exercise scientists have, of course, tried to determine which work interval intensities, work interval durations and recovery durations are best for optimising fitness. This work began in earnest in the early 1960s, thanks to the efforts of Per Olof Astrand of Sweden. Working at the famed Karolinska Institute in Stockholm, Astrand developed what he called 'long' interval training, with work intervals lasting for three minutes at intensities of around 90-92% vVO2max and complete rest between intervals. In spite of the very easy recoveries and submaximal intensities, Astrand's athletes were able to 'hit' VO2max itself during the last repetitions of the overall workout, an event which excited Astrand considerably.

The Swede suggested that his three-minute interval workout represented one of the very best forms of training to improve VO2max, since cardiorespiratory parameters (cardiac output, stroke volume, and oxygen consumption rate) reached their maximums during the session (7). That principle - that intervals should be contrived to ensure the attainment of VO2max - remains firmly entrenched and is practically unassailable from a logical standpoint. However, today's exercise scientists and up-to-date coaches judge the value of interval workouts not just by whether VO2max is attained but also by how long it is sustained within the session.At about the same time that Per Olof was carrying out his investigations, his colleague E H Christensen proposed a quite different form of interval training - 10-second work intervals at the higher intensity of vVO2max itself, with 10 seconds of complete rest for recoveries. Somewhat surprisingly (given the short duration of the work intervals), the rate of oxygen con-sumption also 'struck' VO2max toward the end of this workout, and blood-lactate accumulations were low, which was believed to be a good thing (8). It was thought at the time that low lactate levels would be linked with low levels of fatigue, thus ensuring that an interval workout could be continued for a substantial period of time. Today we know that lactate does not cause fatigue and that high concentrations of lactate are sometimes desirable, since they stimulate muscle cells to 'learn' how to clear lactate from the blood, an effect which improves lactate threshold.

Christensen's group probed deeply into the effects of work interval duration on workout quality, particularly during rather short work intervals (op cit). They found that when athletes alternated 15 seconds of work at vVO2max with 15 seconds of complete recovery it was possible to sustain exercise for at least 30 minutes, with VO2max again being attained toward the end of the session and lactate levels remaining low (at just 2.3 mmol/litre). When recovery intervals dropped to 10 seconds (in conjunction with the 15-second work intervals), lactate levels began to mount (reaching 5-6 mmol/L), since muscles were getting less chance to 'make lactate disappear.' However, when both work and recovery intervals were pared down to five seconds, blood lactate dropped back to 2.5 mmol/L, and athletes reached only 81% of VO2max during the workout, despite the fact that they were running at the same speed used during the 10 and 15-second intervals. It was clear that the five-second work intervals were not long enough to significantly stoke either oxygen consumption or lactate production.

When longer work intervals are better

Later studies showed that when work interval intensities were pulled below the vVO2max preferred by the Christensen-Hedman-Saltin team, the total load on the cardiorespiratory system could usually be directly related to the length of the work interval. For example, when the Astrand-Christensen group analysed work intervals of 30 seconds, one minute, two minutes, and three minutes - all carried out at moderately high but not vVO2max intensities - they found that the shortest work intervals produced a submaximal load on the circulatory and respiratory systems (just 63% of VO2max) as well as low lactate levels (2 mmol/L). By contrast, the two and three-minute intervals eventually tipped VO2max to 100% (even though each interval by itself was not as 'hot' as vVO2max) and caused blood lactate to soar to 16.6 mmol/L. Naturally, the authors recommended that athletes should selectively use the longer intervals whenever work-interval intensity was below vVO2max (9).

What made the longer intervals better? Basically, Astrand and another noted Swedish exercise physiologist, Bengt Saltin, determined that when an athlete is moving along at below vVO2max but above about 90% vVO2max it usually takes him/her about two minutes to reach VO2max during the first work interval of a training session (10). When this intensity zone is used, he/she would tend to fall short of VO2max during the first work interval of his/her training session if he/she employed 30 to 60-second work intervals. Interestingly enough, the subsequent 30-second (or one-minute) recovery intervals might knock oxygen consumption down so appreciably that it would be difficult to attain VO2max over the course of the workout. On the other hand, if an athlete extended the work intervals to three minutes, VO2max would be attained during the first work interval, and oxygen consumption would be so high that the subsequent recovery intervals would have less opportunity to depress oxygen usage during follow-up work intervals. In practice, VO2max was usually reached in less than two minutes during 'later' work intervals within the overall session.

If long intervals are 'good' and short intervals 'bad', why did Astrand's original patterns of 10-10 (10 seconds of work and 10 seconds of recovery) and also 15-15 allow for the attainment of VO2max? The answer is that the work interval intensities were - at vVO2max itself - higher, and the recovery intervals short enough (15 seconds or less) to ensure that that oxygen consumption rates didn't dip too far during recovery, allowing consumption to climb slowly but steadily towards VO2max in the course of the workout. As you have certainly guessed by now, VO2max is more likely to be reached within an interval workout when work intervals are intensified and recovery intervals abbreviated.

Are there times, though, when recovery intervals should 'go the other way', ie be longer than the corresponding work interval? There are a couple of key arguments in support of this idea. First, coaches rightly point out that when recovery intervals are kept short, work interval quality tends to erode over the course of a workout as fatigue levels mount, with the final intervals of the session often several seconds slower than planned pace. Practising slow-downs is sub-optimal, the argument goes, and therefore it would be better to take longer recoveries in order to ensure that each work interval is of the highest-possible quality. Secondly, some coaches and exercise experts argue that if athletes want to improve their economy, they should carry out their work intervals with the most efficient motor units within their muscles. Trimming recovery, these coaches suggest, ensures that the most energy-efficient motor units will be fatigued and non-functional late in the workout, thus leading to a situation in which sub-optimal motor units are being trained.

Why not train sub-optimal motor units?

Let's address this second argument first. Believe it or not, there is some fatigue associated with racing, and if efficient motor units are threatened by an interval workout they will be in much more trouble in a racing situation, which ultimately means that the sub-optimal motor units will be called into play to save the day. So why not train those sub-optimal units?

We should point out, too, that although the existence of super-efficient motor units is plausible, they have never actually been identified in scientific research, nor has there been any verification that these super collections of cells tend to fatigue before their less efficient colleagues. This being true, it seems there's no need to worry too much about the save-the-efficient-unit argument.

On the other hand, losing work interval quality is something to fret about. If your planned workout calls for you to cycle four 5000m repeats on your bike in six minutes each, for example, and this training tempo is a reasonable one for you, but your second interval comes off in 6:10, your third in 6:20, and your fourth in 6:35, you'll need to make some adjustments. If you were just having a bad day, you wouldn't worry about it, but if you were rested and feeling OK before the session, then it would be a good idea to expand the recoveries to make the work intervals more feasible. It's possible that you simply bit off more than you could chew, with recoveries that were too short to start with. If you decided on two-minute recoveries, you were probably being over-ambitious and would be advised to begin with equal recoveries, even if the idea of six minutes of lazy cycling between work reps is somewhat abhorrent. You could then start carving away at the recoveries and keep reducing them until it's no longer possible to pull off the workout.

In the case outlined above and in similar situations, would you ever use recoveries which were longer than the work intervals? For example, if you were running very fast 400s on the track, would it make sense to take long recoveries in order to boost the prospects of hitting really good 400s? Some coaches recommend recovery/work ratios as high as 5:1 (five minutes of recovery for each minute of work) in a bid to make work intervals more productive.

Keep recoveries short to boost VO2max

If you are an endurance athlete perplexed about what to do, simply remember that if you are trying to improve VO2max, vVO2max and/or lactate threshold, you should attempt to keep recovery intervals as short as you reasonably can. Lengthening recoveries will tend to drive down average oxygen-consumption rates and mean levels of lactate production, effects which are counterproductive in terms of VO2max and lactate-threshold.

Let's say, for example, that a distance runner named Ben is doing a classic interval workout,10 x 400m, and has chosen to carry out the session at a goal 5k race tempo which is four seconds per 400m faster than his current 5k pace. Since his 5k pace is 75 seconds per 400m, his interval pace will be 71 seconds per 400m. That's very reasonable, but should he use equal recoveries (71 seconds), short recoveries (30-60 seconds), or long recoveries (5-6 minutes)? Of course, the long recoveries are attractive because they would drive down fatigue and help keep Ben on course with his planned pace during the work intervals, thus bolstering his economy at goal velocity.

The best advice, though, is that Ben should start with roughly equal - not long - recoveries and then try to shorten them a bit. By doing so, he'll produce considerably higher rates of oxygen use over the course of the workout than with the long recovery scenario, and blood-lactate profiles will also be better. Interestingly enough, he'll also be in good shape from an economy standpoint, as long as his pace doesn't drop off too much during those seventh, eighth, and ninth intervals. (We won't worry about the tenth one, since it is always - miraculously - the fastest interval of the whole workout.) True, if Ben hits several intervals slower than 78 seconds-or-so during the second half of the workout, it's time for him to either increase his motivation and mental focus or add a little bit of fat to his recoveries. As long as he can complete the intervals in close to the planned time, however, he should hang in there with equal recoveries - and then shorten them as fitness improves and the workout becomes more manageable.

It's true that if you don't care about vVO2max and lactate threshold and simply want to improve economy, you should go ahead and use 5:1 (which in Ben's case would mean six-minute recoveries for each 71-second work interval). This would be great for the 400m runner, who has few concerns about aerobic capacity and lactate threshold. However, distance athletes do care about those key variables, and casting aside the training stimuli which help optimise them is not usually a sound practice.

Although I have focused so far on the impact of interval training on VO2max, lactate threshold and economy, it's important to recognise that interval training can also have a strong influence on the development of strength and power. I have assumed so far that interval workouts consist only of running, cycling, rowing, swimming or skiing segments at various speeds, but of course they can also include strengthening exercises. The renowned running coach Percy Cerutty made great use of such muscle-bolstering activities, calling on runners like Herb Elliott and John Landy to carry out a variety of strengthening moves within the context of 'circuit' workouts, which also included hard-pressed runs up severe sand dunes. Elliott was never beaten in the mile or 1500m, ran a 3:59.9 mile at the age of 19 and a 3:54.5 mile just one year later, then in 1960 carried off four sub-four-minute miles in a three-week time span, just before winning the 1500m Olympic gold medal with a world-record time of 3:35.6!

Strength movements can boost race times

That's anecdotal evidence, course, but Finnish researcher Laina Paavolainen recently provided strong evidence that workouts which combine high-speed running intervals with explosive strengthening movements (hops, jumps, bounds, presses etc) can significantly improve 5k race times (11). In this study, runners who increased mileage from 45 to 70 miles per week failed to improve 5k times, while runners who remained at 45 miles but added explosive running and strength drills to their training bettered their 5k performances by around 30 seconds. In effect, the explosive group replaced 32% of the training volume of the 70-mile group with the explosive drills - almost exactly the percentage of training time which Cerutty had suggested reserving for strengthening work. The explosively trained runners improved running economy and overall power on a high-speed treadmill test, while the 70-mile runners were unable to do so. Interestingly enough, Paavolainen's group found that max running velocity was a good predictor of 5k time, as was footstrike time (the amount of time a runner spends in the 'stance phase' of the gait cycle).

One special interval training technique involves carrying out intervals in two completely different aerobic activities within a single workout. For example, triathletes frequently perform both high-speed bike and running intervals within a single session, and even 'almost-pure' runners sometimes venture onto a bike during - or at least on the day of - a running interval workout. The idea, of course, is that this accumulation of high-intensity aerobic work will have a greater-than-usual impact on VO2max and perhaps lactate threshold. This seems fairly sensible: for example, a 5k runner who has completed six 800m running intervals within a workout would baulk at the very idea of another leg-muscle-tearing running interval or - worse still - two or three more running intervals; but that same runner could clamber onto a bike after the sixth interval and knock off several 5-6-minute cycling intervals, without impact damage to the leg muscles and without impairing recovery.

So far research doesn't support the idea, though; in fact it suggests that such cross training is not a good way to try to boost VO2max (12). The research hasn't been carried out in an optimal way, however since it has really examined the effects of replacing one discipline with another - running training with biking, for example. Of course, when that happens the poor runner will make fewer gains in running capacity (adaptations to training are sport-specific, after all); the idea is to add intervals in a 'cross' sport to what one is already doing. Such additions have not been closely investigated but are anecdotally appealing.

If increasing the length of work intervals and reducing those of recovery intervals is a good idea, what about actual work interval intensity? Is there a certain speed which optimises fitness improvement? Should one carry out work intervals at vVO2max? At lactate-threshold speed? Halfway between vVO2max and lactate threshold? At race pace?

Do some workouts at race speed

These questions have been hotly debated by athletes, coaches and exercise scientists, and it is clear that it makes sense to carry out some interval workouts at - or around - actual race velocities. For example, a runner completing 1600m work intervals at her current 5k race pace will no doubt improve efficiency (economy) at this speed, making it more likely that she will be able to move up to higher velocities in future races. In a similar vein, a runner performing work intervals at goal 5k pace (perhaps four seconds per 400m faster than current 5k pace), will find it easier to actually run at goal speed in a race because of the resulting gains in efficiency and confidence.

However, there are times when no racing is taking place, so there are no current race times available to govern training. In addition, many exercise scientists argue that it makes more sense to train at intensities which are judged optimal for producing selected physiological responses (for example, the improvement of lactate-threshold speed), since these physiological improvements will ultimately determine overall performance.

One such intensity, for example, would be vVO2max (an athlete's rate of movement when VO2max is attained). vVO2max is an outstanding predictor of performance, and a moment of reflection reveals why this is so: an athlete might have an extremely impressive VO2max but still perform rather poorly if somewhat mediocre movement speeds caused him to utilise almost all of that considerable oxygen-processing capability. In other words, if the athlete were inefficient (ie using a lot of oxygen to sustain a mundane pace) the voluminous VO2max would be of little benefit.

Why high VO2max may be a false positive

By contrast, an athlete with a very high vVO2max can move very fast at her VO2max intensity and thus is obviously fairly efficient. In effect, the athlete enjoys the best of both worlds - a very high aerobic capacity and very great efficiency. vVO2max thus becomes a powerful predictor of performance, while VO2max and economy by themselves carry much less information and thus are less predictive. An athlete with great economy, for example, might have a poor VO2max and thus be unable to reach high speeds at VO2max; his terrific economy would thus give a 'false positive' test for good performance.
Since vVO2max is so tightly linked with success, it makes sense for endurance athletes to carry out interval workouts which have the greatest chance of optimising this variable. The noted French researcher Veronique Billat has been able to show that the best way to do this is by utilising vVO2max itself during training. Again, a moment's thought reveals why this is the case: by working at vVO2max, you improve neuromuscular coordination and efficiency while moving very fast. Most importantly, you are certain of attaining VO2max intensity within the training session, providing the optimal stimulus for VO2max to expand further.

But how do you determine your vVO2max? As I have pointed out in these pages before, you can do this quite easily: on a day when you are feeling great, simply run, cycle, swim, race-walk, row or ski as far as you can in six minutes and then compute your distance. For example, if you're a runner and you ran 2000m in six minutes, your vVO2max would be 2000/360 or 5.55m per second (72 seconds per 400m). If you ran 1600 meters in six minutes, your vVO2max would be 1600/360 = 4.44m per second (90 seconds per 400m).

In a recent study, Billat asked eight experienced runners to take part in four weeks of training which included one interval session per week at vVO2max. The athletes specialised in middle and long-distance running (1500m to half-marathon), their mean age was 24, and average VO2max was a respectable 71.2 ml/kg/min (13).

During the four-week period, the runners completed one vVO2max interval workout per week, consisting of five three-minute work intervals at vVO2max, with three-minute jog recoveries. The rest of the running during the week was easy, except for a 'lactate-threshold-improving' session, which consisted of two 20-minute work intervals at 85% of vVO2max, with a five-minute easy-jog recovery between the two intervals. Total distance covered by the runners per week was about 50 miles.

Billat's vVO2max and lactate-threshold intervals were simple - and simply devastating. After four weeks, vVO2max rose by 3% - from 20.5k/hour to 21.1k/hour. In addition, running economy improved by an astounding 6%, while heart rate at 70% VO2max dropped by 4%. Although lactate threshold held steady at 84% of vVO2max, since vVO2max was higher after four weeks, velocity at lactate threshold also increased. Almost all of the key physiological variables associated with performance had improved!

Note in particular the dramatic improvement in economy (6%) achieved by Billat's runners, an almost unheard-of increase in efficiency in well-trained competitors, especially within such a short time frame. The reason for this efficiency groundswell is that exercising at vVO2max increases leg-muscle strength and power, and enhanced strength tends to boost economy; since muscle cells are stronger, fewer need to b