Archive for the ‘Physiology’ Category

The longevity of the long distance runner, part 3: Cardiac Outcomes

January 10, 2016

After examining the anecdotal evidence provided by elderly marathoners and then grappling with some of the basic science underlying longevity in recent posts, it is time to attempt to draw some practical conclusions about what can be done to optimise longevity as a runner.   While the unfolding science of the molecular mechanisms by which our bodies respond to wear and tear offers intriguing prospects of identifying strategies for promoting heathy aging, there are at this stage more questions than answers. However we each live only once and if we want to optimise our chances of running fluently in old age, we must make the most of the evidence that is currently available. Despite the uncertainties, the evolving science provides a framework for weighing up the value of the lessons that might be drawn from the anecdotes.

But it is necessary to be aware of the pitfalls if we are too simplistic in our interpretation of the science. For example, at first sight the fact that catabolic hormones such as cortisol promote the break-down of body tissues to provide fuel for energy generation in stressful situations, suggests that avoiding sustained elevation of cortisol is likely to promote longevity. Indeed this conclusion might be valid in some circumstances, but it is not a universal rule. The strategy that is most successful for promoting longevity in animals is a calorie restricted diet. This works for creatures as diverse as worms, fish, rodents and dogs, and there is some evidence for health benefits in primates. The mechanisms by which it achieves its benefits include an increase in resistance to oxidative damage to tissues by virtue of more resilient mitochondrial membranes. However calorie restriction is stressful and promotes long term elevation of cortisol. It appears that animals actually need at a least a moderate level of ongoing stress to encourage heathy adaptation. The goal is achieving balance between stressful stimuli and adaptive responses.   The balance point depends on individual circumstances, and is likely to shift as we grow older.

The nature of training

The basic principle of athletic training is stressing the body in order to encourage it to grow stronger. One of the key mechanisms by which this is achieved is inflammation, a process by which damage to tissues generates a cascade of responses mediated by chemical messenger molecules that circulate in the blood stream, triggering repair and strengthening but also leaving a trail of debris.

There is abundant evidence that running increases life expectancy. In a 21 year follow-up study of elderly runners Chakravarty and colleagues at Stanford University found that continuing to run into the seventh and eighth decades has continuing benefits for both life expectancy and reduction of disability. Death rates were less in runners than in controls not only for cardiovascular causes, but also other causes, including cancer, neurological disorders and infections. Nonetheless, unsurprisingly, Chakravarty reported that although the increase in disability with age was substantially less in runners than in non-runners, the runners did nonetheless suffer increasing disability over the follow-up period. This is of course what would be expected if the processes by which training strengthens the body also leave a trail of debris.

If our goal is to increase not only life expectancy but also to achieve healthy aging and longevity as runners, we need to look more closely at the mechanisms by which running damages tissues. Ensuring longevity as a runner requires training in a manner that ensures that the accumulation of debris is minimised.

Healthy aging is a process affecting all parts of the body. Nonetheless, for the runner, the cardiovascular system, musculoskeletal system, nervous systems and endocrine systems are of special importance. There is a large body of evidence about how these systems age and about both the beneficial and the damaging effects of running on these systems.   I will examine the evidence regarding cardiovascular system in this post, and draw some tentative conclusions about how we might train to achieve healthy aging of the heart. In my next post I will examine the evidence regarding the musculo skeletal system, for which much more detailed information about cellular mechanism is available due to the feasibility of tissue biopsy. This will allow us to extend and consolidate the conclusions regarding optimal training for maximizing longevity as a runner.    In the final posts in this series I will turn my attention to the nervous and endocrine systems, and speculate about the way in which optimal training might impact upon health aging of those crucial systems.

Cardiovascular changes

Running produces both short term and long term changes in the cardiovascular system, some beneficial, some potentially harmful. I have discussed many of these changes in several previous blog posts (e.g. ‘The athletes heart’; ‘Inflammation, heart-rhythms, training-effects and overtraining’; ‘Endurance training and heart health, revisited’) and will present a brief overview here.

In the medium term (over time scale of weeks) regular training leads to an increase in blood volume. This increases venous return to the heart. The stretching of the heart muscle leads to more forceful contraction and a greater stroke volume. The cardiac output for a given heart rate is increased. Resting heart rate decreases and the heart rate required to run at a particular sub-maximal pace decreases.

In the longer term, the heart muscle is remodelled, with an increase in overall volume and in thickness of the ventricular walls. This condition is known as ‘athletes heart’. The mechanism is mediated by the intra-cellular signalling pathway that engages an enzyme known as Akt, which promotes growth of both muscle cells and capillaries. This is usually regarded as a benign physiological adaptation. The enlargement of the heart that accompanies pathological conditions such as high blood pressure or obstruction of the heart valves is also mediated by the Akt signalling pathway, but in contrast to the benign enlargement of the athlete’s heart, the Akt signalling is accompanied by inhibition of a growth factor required for the development of capillaries. Thus, in the athlete’s heart the enlargement is accompanied by adequate development of a blood supply to the heart muscle, whereas in pathological conditions the blood supply is usually inadequate.

However, in some athletes the enlargement might have adverse effects. There is compelling evidence that endurance runners with a long history of substantial training have an increased risk of disturbances of heart rhythm, including both ‘supra-ventricular’ disturbances such as atrial fibrillation, and potentially more lethal ventricular disturbances. The cause of these rhythm disturbances is not fully established but it is probable that the re-modelling of the heart muscle in a way that alters electrical conduction pathways plays a role.  It is likely that residual fibrosis at sites where damaged muscle has been repaired also plays a role by producing local irritability of the cardiac muscle cells leading them to fire spontaneously.

During intense prolonged exercise the strength of ventricular contraction, especially that of the right ventricle, is diminished, a condition known as Exercise-Induced Right Ventricular Dysfunction. If the exercise is sufficiently intense and prolonged, cardiac enzymes can be detected in the bloodstream, indicating a least temporary structural damage to heart muscle.

In a study of forty highly trained athletes competing in events ranging from marathon to iron-man triathlon, LaGerche and colleagues from Melbourne found transient weakening of the right ventricle immediately after the event. This was more severe the longer the duration of the event. The transient weakness returned near to normal within a week. However in 5 of the athletes, there was evidence of long term fibrosis of the ventricular septum, indicating chronic damage. Those with evidence suggesting long term damage had an average age of 43 and had been competing for an average of 20 years. Those without evidence of chronic damage had an average age of 35 and had been competing for an average of 8 years.   The evidence suggests that duration of endurance competition is a strong predictor of chronic damage.

Although an enlarged athlete’s heart usually has a much better blood supply than the enlarged heart associated with high blood pressure or obstruction of the heart valves, there is disconcerting but controversial evidence of excessive calcification of the arteries in at least some athletes, especially in males in who run marathons over period of many years. The mechanism is uncertain, though sustained inflammation is a plausible mechanism.

Effects of the amount and type of training

Although an overwhelming mass of evidence demonstrates that runners have a longer life expectancy and in particular, a lower risk of death from heart attack or heart failure than sedentary individuals, several large epidemiological studies raise the possibility that adverse health effects (especially cardiac events) tend to be a little more frequent in those who engage in a large amount of exercise than in those who exercise moderately. The US Aerobic Longitudinal Study examined the associations of running with all-cause and cardiovascular mortality risks in 55,137 adults, aged 18 to 100 years (mean age 44 years) over an average period of 15 years and found a marked decreased in both cardiac and all-cause mortality in runners compared with non-runners, but the reduction in mortality was a little less in those training 6 or more times per week compared with those training 1-5 times per week. The Copenhagen City Heart Study followed 1,098 healthy joggers and 3,950 healthy non-joggers for a period of 12 years and found that 1 to 2.4 hours of jogging per week was associated with the lowest mortality. These ‘moderate’ joggers had a mortality hazard ratio of 0.29 compared with sedentary non-joggers.

But closer look at the evidence reveals a potentially informative detail. In a study of heart health of over a million women, Miranda Armstrong and her co-investigators from Oxford  found that among obese women, those who did a large amount of exercise suffered more heart problems than those who did a moderate amount. However, in contrast, among the women who had a Body Mass Index less than 25, those doing a large amount of exercise had fewer heart problems than those doing a moderate amount of exercise.   This suggests that if there is a risk in doing a large amount of exercise, it is mainly confined to those for whom the exercise is excessively stressful due to other risk factors that shift the balance towards harm rather than benefit.

Although the evidence from the large epidemiological studies remains a topic of debate because of issues such as possible bias in participant selection and the relatively small numbers of individuals in the category who take a very large amount of exercise, I think the balance of evidence does indicate that at least some individuals who take a large amount of exercise do have an increased risk of death, including death form cardiac events, within a given time period.   In my opinion, the important question is what determines which individuals will be harmed by a large amount of exercise, and whether there are ways in which we can minimise the risk of harm.

There is evidence that adequate prior training can protect against damage.   Neilan and colleagues studied non-elite marathoners runners completing the Boston Marathon and reported that right ventricle weakness was more pronounced in those who had trained less than 35 miles per week compared with those who had trained more than 45 miles per week.   The logical conclusion from studies such as the Oxford study of obese female runners and Neilan’s study of marathoners is that running in a manner that exceeds the individual’s current ability to cope with the stress increases the risk of damage.   This in turn suggests that building up gradually in a manner that ensures that training sessions are never excessively stressful is likely to be the safest approach.

Furthermore, it is likely that lack of adequate prior training or obesity are not the only factors that impair the ability to cope with the stress of demanding training and racing. Following a very demanding marathon or ultra-marathon, the evidence of damage remains detectable for a period of weeks. It is plausible that demanding training when the heart is in a weakened state will compound the damage. It is widely accepted in practice that recovery following intense racing or heavy training is crucial, but unfortunately there is relatively little scientific evidence addressing the question of whether or not the adverse cardiac effects of intense exercise resolve during a recovery period, or conversely, whether the adverse effects are compounded by repeated bouts of exercise.   We must therefore turn to evidence from studies of rats.

Benito and colleagues exercised rats on a treadmill for 60 minutes at a quite vigorous pace of 60 cm/s (achieved after 2 weeks of progressive training) 5 days per week for a total of 4 weeks, 8 weeks or 16 weeks. For a rat, 16 weeks of life is roughly equivalent to 10 years for a human. During the first 8 weeks there was relatively little evidence of damage, but prominent signs of damage emerged between 8 and 16 weeks. After the 16 weeks of exercise, the rats exhibited hypertrophy of the left ventricle and also the reduced function of the right ventricle, similar to the findings reported in humans. Furthermore the rats had marked deposits of collagen in the right ventricle, and messenger RNA and protein expression characteristic of fibrosis in both atria and the right ventricle. The exercised rats had an increased susceptibility to induction of ventricular arrhythmias. A sub-group of the rats were examined after an 8 week recovery period following the 16 weeks of exercise. Although the increased weight of the heart had not fully returned to normal level, all of the fibrotic changes that had been observed after 16 weeks of exercise had returned to the normal level observed in sedentary control rats. Thus, at least in rats, the adverse potentially arrhythmigenic changes produced by intense exercise over a 16 week period appear to be reversible after an adequate recovery period.    Thus the best available scientific evidence does support the accepted principle that recovery following intense racing or heavy training is crucial.


Proposed cardiac outcomes of long-term training. The size of the ellipses indicates cardiac fitness at each stage; colour indicates balance between recovery (blue) and stress (red)

In summary, the evidence regarding the cardiovascular effects of running suggests the following guidelines for healthy aging and longevity as a runner:

  • Continuing to run regularly, at least into the seventh and eight decades decreases risk of death and disability.
  • Training volume should be built up gradually.
  • Adequate recovery after demanding events, such as a marathon (or indeed, even after heavy training sessions) is likely to be crucial.

My next post will examine the evidence regarding the effects of training on the musculoskeletal system, and will both consolidate and extend these conclusions.

Neuromuscular coordination for distance running

July 30, 2015

It is well established that countermove jump height (CMJ) is a good predictor of sprinting speed.  This is not surprising because CMJ performance depends on powerful type 2 muscle fibres and on the ability to coordinate the recruitment of these fibres.  In the CMJ, flexion of the hips and knees produces eccentric contraction of the corresponding extensors immediately prior to the explosive concentric contraction that propels the body upwards.  It is necessary to recruit the muscle fibres in a manner that harnesses the enhancement of power generated by the eccentric contraction.

The relationship between CMJ performance and distance running performance has been less thoroughly investigated.

In assessing endurance training, aerobic capacity and lactate threshold have been the main foci of attention, but other training-related variables also predict performance.  It is has been demonstrated that difference between elite athletes in volume of zone 1 training (comfortably below LT) predicts distance race performance (e.g. 10Km). In addition, it is fairly well established that a high values of the ‘stress hormone’ cortisol sustained across a period of months predicts poorer performance.

However, somewhat paradoxically, within an individual athlete, week to week variations in training volume and cortisol values make the opposite predictions.  In a study comparing seasons best and seasons worst performance in elite athletes, total training volume  was less but volume of zone 3 training  (appreciably above LT)  was greater in the week before the seasons best performance. Cortisol tended to be higher a week before the best performance, Countermove jump height was also higher in the week before the best performance.

This apparent paradox is consistent with the evidence that a taper should involve decrease volume but not a decrease in training intensity.  The fact that CMJ was higher before the season’s best performance suggest to me that zone 3 training in the week preceding the event promotes good neuromuscular coordination.

The importance of neuromuscular coordination is clearly illustrated by the clunkiness that triathletes experience during the bike to run transition.   The rapid gains in weight lifting performance  at the beginning of a lifting program are most likely due to improved recruitment of muscle fibres.  Conversely, fatigue impairs neuromuscular coordination, and measurement of postural sway has been proposed as a sensitive measure of impaired neuromuscular coordination arising from fatigue in footballers.

Overall, the evidence indicates that neuromuscular coordination is crucial for both athletic performance and injury minimization but it is rarely the focus of attention in endurance training.  While not a specific focus of attention, when we engage in routine warm up we do in fact achieve  short-term enhancement of neuromuscular coordination, and when we accumulate miles of training, we engage in long-term enhancement of neuromuscular coordination, but we rarely think of these activities as exercises in enhancing neuromuscular  coordination.  However, we are more likely to produce effective enhancement of  neuromuscular coordination if we plan our warm-up and training activities bearing neuromuscular coordination in mind.

The elements of coordination

Recruitment of the optimal number and type of muscle fibres:  because much of our training should be at an intensity less than racing intensity, we need pay attention to the need to ensure that we do retain the ability to recruit type 2 fibres as effectively as required at race pace.  Although the importance specificity in training is sometimes over-rated, at least some specificity is essential.  As discussed in my recent post on lactate shuttling, beneficial enhancement of the ability to handle the accumulation of the lactate can be achieved by a large volume of low intensity training.  However the danger of a program focussed too strongly on low intensity running is the development of a tendency to plod slowly under all circumstances.    It is therefore crucial to do at least some training at or near race pace, especially when fatigued as is likely to be the situation in the later stages of a race.  Progressive runs that achieve a pace at or even a little faster than race pace are likely to be beneficial

Recruitment of muscles in the optimal sequence:  The action of running entails a very complex combination of muscle contractions, requiring that the extensors and flexors at each of the major joints of the leg are recruited in a precisely timed sequence.

Speed of recruitment of muscle fibres:    With increasing age, deterioration in running speed is associated with loss of stride length; not cadence.    This is accompanied by a atrophy of muscles and loss of strength.  However as I found three years ago when engaged in  intense high-load weight lifting program for several months, I was able to increase my strength to the point where I could squat a heavier load than Mo Farah, but  my stride length did not increase appreciably, and my speed remained but a very pale shadow of Mo’s speed.    Speed depends on  power: the ability to exert force rapidly.  This requires effective, rapid recruitment of muscle fibres.  It is far harder to train power than speed, though there is some evidence that focussing on a rapid contraction during the concentric phase of a lift, at moderate load, can produce a worthwhile gain in power.

Implications for warming up For most of my training sessions, I employ a warm up procedure that includes 10 activities, beginning with simple movements designed to get all of the major joints  of the leg moving freely, and proceeds though a sequence in which  power output gradually increases.

Calf raises

Hip swings, (straight front to back; rotating from foot behind to opposite side in front.)

Body-weight  squats (aiming for hips below knees)

Single leg squats

Lunge, to front and side

High knees

High knees skipping

Hopping (fast, small hops)


Surges at race pace

Time for each is adjusted according to how my body is reacting, though typically each of the first 8 activities takes 20-60 seconds; the focus is on fluent action rather than effort.

Implications for injury minimization Recent studies, reviewed by Herman and colleagues, reveal  that in a variety of different sports, poor neurocognitive performance, either at baseline or in the aftermath of a concussion, is associated with elevated risk of musculoskeletal injury. It is probable that a thorough warm-up that  sharpens up neuromuscular coordination is a good way to minimise risk of injury.

Measuring neuromuscular coordination The CMJ is widely used  in various sports, especially team games such as football, to assess fitness.  However, it has three potential disadvantages as a measure of neuromuscular coordination for the distance runner: 1) it is not a ‘pure’ measure as performance depends on type 2 fibre strength in addition to coordination; 2) maximal performance is quite demanding and creates the risk of injury; 3) accurate measurement requires special equipment.

I have been experimenting with time taken to perform 20 line jumps as a test of coordination.  It does depend on other aspects of fitness such as muscle strength  to at least a small extent, but placing the emphasis on time rather than maximal power focuses attention on coordination rather than strength.  The risk of injury is small. At this stage, the utility of timed line-jump performance as a test remains speculative as I have not tested it systematically.  Typically, I find that my time for 20 line jumps decreases from 9.0 seconds after a few minutes of jogging to 7.5 seconds after the ‘neuromuscular’ warm up described above.  Time for 20 jumps increases dramatically after a long run.  Provided I can establish that the test yields consistent results when assessing deterioration in neuromuscular coordination associated with fatigue, I plan to use it to determine whether or not light weight shoes (Nike Free 3.0) result in greater deterioration in coordination during a long run, compared with more heavily padded shoes.

Conclusion It is almost certainly true that many of the activities that athletes and coaches have traditionally incorporated into warm-up and training achieve their benefit at least partly through enhancing neuromuscular coordination.  However by focussing on the more easily quantifiable physiological variables when planning and assessing training sessions, there is a risk that endurance athletes might fail to optimise training to achieve the required combination of aerobic capacity, strength and coordination.   Perhaps we should place more emphasis on a systematic approach to enhancing neuromuscular coordination during training, and on measuring it to assess the outcome of that training.

The lactate shuttle and endurance training

June 19, 2015

Races from 5000m to marathon are run at a pace that is strongly predicted by pace at lactate threshold because the accumulation of acidity is a major factor limiting muscle performance.  During the combustion of glucose to generate energy, the major source of acid is lactic acid.   Lactic acid is a compound of two ions: negatively charged lactate ions and positively charged hydrogen ions.  Under normal circumstances within the body, lactic acid dissociates into these two constituent parts, lactate and hydrogen ions, and it is the latter that create the acidity.   At least in part, the hydrogen ions are buffered (i.e. mopped-up by  other negative ions within tissue) but once this buffering capacity is saturated, hydrogen ions accumulation creates an acid environment that impairs the efficiency of muscle contraction.

However lactate itself can be used as fuel in various locations in the body, and as it is itself metabolised hydrogen ions are removed.    Thus, understanding the mechanisms by which lactate is transported around the body and the mechanisms by which it is itself metabolised provides the basis for rational planning of training.

The orthodox view

The scientific studies of Louis Pasteur, in the nineteenth and early twentieth century, and subsequently by Hans Krebs, AV Hill and others, early in the twentieth century uncovered the mechanisms of anaerobic and aerobic metabolism of glucose, and established the orthodox view that shaped the theory of training for distance training through the second half of the twentieth century. According to this orthodox view, glucose is initially metabolized by the process of glycolysis, to produce pyruvate,  This anaerobic transformation of glucose to pyuvate releases a modest amount of energy which is transferred into the high energy bonds of the energy-rich molecule, ATP, and ultimately can be used to fuel muscle contraction.  But in the presence of oxygen much more energy can be derived via aerobic metabolism of pyruvate. The pyruvate is transported into mitochondria and converted to acetyl-CoA by an enzyme complex known as the pyruvate dehydrogenase complex.  Acetyl-CoA then undergoes a series of chemical transformations catalysed by the aerobic enzymes within mitochondria.  The acetyl group is oxidized to carbon dioxide, releasing a relatively large amount of energy which is incorporated into ATP, thereby providing a substantial enhancement of the supply of energy for muscle contraction.   In contrast, when the rate of delivery of oxygen to muscle is inadequate, muscle contraction must be fueled via the modest energy yielded by conversion of glucose to pyuvate, and the pyruvate is converted to lactate, thereby generating potentially harmful acidity.

The lactate shuttle

The orthodox view is indeed substantially correct, but is misleading because a substantial proportion of pyruvate is converted to lactate even when oxygen supply is adequate to sustain aerobic metabolism in mitochondria.   The way in which the body deals with this lactate is of crucial importance to coaches and athletes.  It was not until the final years of the twentieth century that an adequate understating of the way in which lactate in handled in the body emerged.    The newly emerging understanding was based largely on the work of George Brooks of the University of California, Berkeley Campus, who developed the concept of the lactate shuttle.  Lactate shuttling refers to a group of processes by which lactate is transported within and between cells to locations where is undergoes metabolism.   There is still debate about the details of these mechanisms but the broad principles are now clear, and these principles have major implications for optimum training for distance running.

The first key point is that a proportion of the pyruvate generated in the first stage of glucose metabolism is converted to lactate in the cytosol (the fluid medium inside cells) of muscle fibres, across a very wide range of work-rates, extending from the low aerobic to anaerobic zones.  In the low and mid-aerobic range, the majority of the lactate is transported across various membranes to various different sites where it is metabolized, so there is very little observable accumulation of lactate in blood until the rate of generation of lactate rises near to the limit of the body’s ability to transport and utilize lactate.  Beyond this point, the concentration of lactate and hydrogen ions in blood rises rapidly (the Onset of Blood Lactate Accumulation, OBLA); respiration becomes very effortful, and the ability to maintain that pace is limited by the limited ability to tolerate acidity.

It is helpful to understand the various processes by which lactate is transported out of the cytosol of muscle cells and subsequently metabolized, in order to design a training program that is likely to enhance these processes.

There are four major pathways by which lactate is transported and metabolized:

  • Transport across the outer mitochondrial membrane to a site where lactate dehydrogenase converts lactate back to pyruvate which then undergoes aerobic metabolism within the same muscle cell. This process results in dissipation of lactate and acidity in the cell where it was created and hence is merely a mechanism that ensures that acid does not accumulate when oxygen supply is adequate to sustain aerobic metabolism.
  • Transport out of the muscle fibre where it was created, into nearby fibres where is can be transported across the outer mitochondrial membrane and metabolized. This mechanism makes it possible for lactate to be generated in type 2 muscle fibres, which are powerful but have relatively low aerobic capacity, and then metabolized in type 1 fibres which have high aerobic capacity.  Training ‘at a good aerobic pace’ as advocated by Lydiard is potentially a good strategy for developing this mechanism.  Although the training pace might be comfortably below OBLA, the ability to dissipate lactate and acidity will be enhanced, leading to a an increase in pace at OBLA and improved performance over distances from 1500m to marathon.
  • Transport out of the muscle fibre where it was created into the blood and thence to other organs, such as heart and brain where it can be metabolized to generate energy
  • Transport out of the muscle fibre where it was created into blood and thence to liver where it can be converted back to glucose by the process of gluconeogenesis. This mechanism is likely to help conserve glucose stores in a manner that is useful in long events such as the marathon.

It is important to note that all of these processes entail transport across a membrane or several membranes prior to metabolism.   Transport is an active process that is performed by specialised proteins known as monocarboxylate transporters (MCTs). There are several different type of MCT located in different types of membrane.   Like many proteins in the body, utilization encourages production of increased amounts of the protein.

Aerobic base-building

Low intensity running is not merely about developing the capillaries to deliver blood to muscle and mitochondrial enzymes that perform aerobic metabolism.  Because appreciable amounts of lactate are produced, transported and metabolized even at work-rates well below OBLA, low intensity training  helps build capacity to handle lactate.

Think of the flow of glucose into the energy metabolic pathway as being like water flowing from an inflow pipe into a sink. The inflow pipe is actually the anaerobic pathway (glycolysis).  A pool of interchangeable pyruvate and lactate tends to accumulate in the sink.  There are two ways out of the sink: transport out of the cell or down the plug hole into mitochondria where aerobic metabolism occurs.  In fibres in which the aerobic system has been well developed , the flow down the plughole can accommodate a large flow of glucose into the system without the sink overflowing.  Shuttling from type 2 fibres which have less well developed aerobic capacity, into nearby type 1 fibres also helps maintain the flow.  There is minimal accumulation of acid in either muscle or blood.  So purely aerobic development, which can be achieved by low intensity training, minimizes accumulation of acid at 10K and 5K pace.   Shuttling explains why runners who only do low intensity running during base building nonetheless usually find that 10K and 5K pace improve despite doing no training near race pace.

The important conclusion for the endurance athlete is that low intensity training is an effective way to develop an aerobic base which helps raise lactate threshold.  It enhances performance at all distances for 1500m to ultramarathon.

Interval training and fartlek

Lactate shuttling also provides an explanation for the effectiveness of interval training and fartlek.  A brief intense effort epoch above threshold generates a surge of lactic acid, and is followed by a recovery epoch during which lactic acid levels fall rapidly, before the sequence of effort and recovery is repeated.  Thus, large gradients in lactate concentration develop, resolve and develop again.  It is noteworthy that the drop in lactate during recovery is likely to be facilitated by low intensity running which maintains transport and utilization of lactate during recovery.  A high lactate gradient across a membrane makes high demand on the transport process and is likely to stimulate production of the relevant transporters, the MCTs.

It is plausible that this will be achieved most effectively by ‘cruise’ intervals of the type employed by Emil Zatopek, or by fartlek sessions in which intensity is high enough during the effort epochs to ensure substantial  production of lactate, while intensity during the recovery epochs is adequate to ensure transport and utilization of the lactate produced during the preceding effort epoch


The various pathways by which lactate is transported out of the muscle cells where it is created to locations where it can be usefully metabolized provide a set of mechanisms by which either low intensity training at a pace comfortably below lactate threshold, or by interval training in which lactate is generated in brief surges, can develop  the ability to cope with lactate production at race pace.    Thus training for sustained periods at or near race pace, which tends to be quite demanding by virtue of the sustained stress, has only a relatively limited role to play in training for endurance events.  This might be an explanation for the observation that many elite endurance athletes adopt a polarized approach to training, in which the majority of training is done at a pace comfortably below threshold, together with an appreciable minority of training at higher intensity during interval sessions and  a modest amount of sustained running near to threshold.

Charles Eugster: strength or power?

March 20, 2015

Perhaps the most enigmatic of the trainable capacities required for distance running is efficiency: the ability to generate as much speed as possible from a given amount of energy.  In endurance events, when the vast majority of energy is generated aerobically, it is the ability to extract as much speed as possible from a given rate of oxygen consumption.    Somewhat confusingly, it is usually reported as oxygen consumption required for a given speed and quantified in units of ml/Kg/Km, though this is actually the inverse of efficiency. It is also referred to as economy and is sometimes reported as speed at VO2max. A proportional increase in speed at VO2max will result in a similar proportional increase in speed at sub-maximal rates of energy consumption.

The first focus in planning a training program is on increasing the ability to consume oxygen (VO2max) by increasing blood supply to muscles and increasing mitochondrial enzymes, but for well-trained athletes, the scope for improving VO2max is small.   When generating energy at VO2 max, a substantial amount of metabolism is anaerobic and results in the build-up of lactic acid. Since accumulation of acid limits muscle power output, the second focus of endurance training is reducing lactic acid accumulation, either by increasing the ability to transport lactic acid out of muscle cells and metabolise it in other tissues, such as liver and heart, or by enhancing fat metabolism, which does not generate lactate.   However there is limit to what can be achieved by improving the capacity to minimise lactic acid accumulation.   Once VO2max has been maximised and the accumulation of lactic acid has been minimised, the focus of training must shift to efficiency.

As illustrated by Andrew Jones’ account of the physiological developments achieved by Paula Radcliffe in the decade prior to her phenomenal a marathon time of 2:15:25  in London in 2003, the crucial improvement was in efficiency.  Paula’s VO2max remain approximately constant at around 70 ml/Kg/min over the decade, but her estimated pace at VO2max increased by 15%.  Andrew Jones acknowledged that mechanism by which this increase in efficiency was achieved remains a mystery.

In my recent discussion of this issue, I distinguished between improvement in metabolic efficiency: the amount of work that can be achieved by muscle contraction per unit of energy consumed; and improvement in mechanical efficiency: the pace achieved for a given amount of work by the muscles.   Metabolic efficiency is strongly dependent on the relationship between rate at which muscle fibres shorten during contraction and optimum shortening rate for the type of fibre involved. In general during endurance running, the required of rate of shortening is relatively low. Type 1 fibres are more efficient than type 2 fibres when the rate of contraction is slow.  Therefore, one goal of training is increasing the proportion of work done by type 1 fibres.  This is achieved by low intensity training.  But the need for the powerful contraction provided by type 2 fibres cannot be completely neglected because of their role in achieving mechanical efficiency.

Maximising mechanical efficiency demands optimising the balance between the three major energy costs of running: getting airborne; overcoming braking and repositioning the swinging leg.  The cost of getting airborne can be reduced by spending more time on the ground, but that inevitably increases braking cost.  Braking cost can be reduced by increasing cadence, but the cost of repositioning the swing leg increases with cadence so there is a limit to cadence.   So one cannot escape the need to get airborne – indeed getting airborne is what distinguishes running from walking.  But getting airborne requires a powerful push against the ground.   The force-plate data of Peter Weyand and colleagues indicates that a major determinant of running speed is the ability to exert a strong push to lift off from stance.

For many distance runners, but especially women and elderly men, I think that the most likely factor limiting mechanical efficiency is lack of ability to exert an adequate push.    One of the striking features of Paula Radcliffe’s running in her heyday was her ability to get airborne.   The interesting question is how she developed this ability.  Following her disappointing performance in the 10,000m in Sydney in 2000, her physiotherapist, Gerard Hartmann identified her poor ability to generate the power required to step on and off a high box rapidly.   He recommended a course of plyometrics and weight training.  It is likely that this played a substantial part in her transformation from a leading athlete who was struggling to fulfil her potential into the greatest female marathoner the world has seen.  But was it the plyometrics or the weight training that played the greater part?

Charles Eugster

I was reminded of this question last week when Charles Eugster set a new world record for the MV95 200m in London, with a time of 55.38 seconds.  Like most people, my first reaction was ‘how amazing’ though I was not entirely surprised.  He had done a very entertaining TED talk a few years ago, in which he presented an enchanting picture of a 93 year old with a very positive outlook on life, and a strong message about the virtues of weight training for the elderly.   The story behind his MV95 200m record provokes some interesting speculations about how to improve the ability to get airborne.

Charles was the child of Swiss parents and had spent his childhood in London, before returning to Switzerland.  He had become a dentist despite his parents wish that he become a doctor.  One of his teachers at St Paul’s school had said: ‘Eugster, if I were to put your brain into a sparrow’s head, there would still be room for it to wobble’. He decided that he would be better advised to focus on learning about 32 teeth rather than all the organs of the body.

He had played sport all his life, and in a particular, rowed for his school, but in his eighties, decided that it was time to smarten up his body.  In his TED talk, he says it was all due to vanity. He claimed his ambition is to impress sexy young girls of 70 on the beach.  He had wanted to start on a six pack but his personal trainer, Sylvia Gattiker, former Swiss aerobics champion, said he needed to start with work on his glutes.  She stands no slacking, encouraging him along the lines: ‘…. and breathe out, no, that’s groaning…. breathe out’.

Under Sylvia’s guidance he has developed a spectacular body.  He has won numerous body building and strength awards.  At 89 he became world 80+ Strenflex champion.  Strenflex involves a set of exercises that are scored for form and number of repetitions in a given time.  It is a test of strength, endurance and flexibility.  About two years ago, he took up sprinting, and within less than a year, was British 100m and 200m Masters Athletics 90+ sprint champion.  So his world 200m record last week was amazing but scarcely surprising.

It is nonetheless intriguing to examine his style.

He runs with a high cadence but shuffling gait.  He does get airborne, but not for long.  I suspect this is part because Sylvia’s coaching emphasized reaching forwards with the swing leg, but it is a style quite characteristic of the elderly.   Ed Whitlock, who trained for his 2:54:45 marathon at age 71 by running up to 3 hours per day in a Toronto cemetery, deliberately cultivated as slow shuffle to protect his knees from the pounding during training, but during races, he is a delight to behold: He gets airborne in a manner reminiscent of Paula Radcliffe.

By virtue of his rapid cadence, Charles Eugster reduces both the costs of getting airborne and braking costs, but to my eye, he nonetheless incurs unnecessary braking costs.  His cadence appears is to be near the practicable limit, and any further improvement in efficiency would require spending a smaller proportion of the gait cycle on stance.   His energy cost would almost certainly be less if he got properly airborne, as younger sprinters do.  Impressive as his 200m time of 55.38 seconds is, it is almost certainly limited by a loss of the power required to get airborne.

The elderly lose strength but even more noticeably, they lose power: the ability to exert force rapidly.  It has been recognised for some years that the elderly can regain strength with remarkable effectiveness by lifting heavy weights.  Charles has clearly been very successful in achieving this.  It is noteworthy that in order to become a Strenflex champion, for which it is essential to be able to perform repetitions of various strength exercises rapidly, Charles must have retained some power.  Nonetheless, it appears that in training he has placed his main focus on shifting quite heavy loads, which is not a very effective way of re-building power.

Building power by explosive contractions

There is a substantial body of evidence indicating that the most effective way to build power is to perform muscle contractions  at the rate that maximises power.  Power is the rate of doing work.  Work is the product of force by distance, so power is the product of force generated by distance the load is moved divided by the time taken, or on other words, the product of force by speed of contraction.  Speed of contraction increases as load decreases, and for young adults, peak power is typically generated at about 30% of the maximum force that can be generated:  that is 30% of one repetition maximum  (1RM).   The elderly have less scope for increasing speed of contraction, and maximum power is usually generated at a somewhat great fraction of maximum load.  Typically maximum power is generated at 60% of 1RM.

Because the risk of muscle damage is greatest during eccentric contraction when a muscle is stretched while under load, the safest way to build power is to perform the eccentric phase of the exercise slowly and then execute the concentric phase explosively with maximum  possible speed.  Encouragingly, several recent randomised controlled trials have demonstrated that high velocity power training is feasible, well tolerated, and is effective in increasing leg muscle power in the elderly.    De Vos and colleagues randomly assigned 112 healthy older adults (aged 69 +/- 6 years) to explosive resistance training at one of three intensities (20%; 50%; or 80% of 1RM) for 3 sets of 8 rapid concentric contractions , for 8-12 weeks, or to a non-training control group. Peak power increased significantly by about 15% in all three groups doing the explosive exercises, compared with 3% in the control group.   In comparison with the groups using 20% and 50% of 1RM, the group using 80% of 1RM exhibited a greater increase in strength and also a greater increase in muscle endurance (the number of repetitions that could be performed with a load of 90% of 1RM).  It is also noteworthy that injuries were rare and relatively minor during the explosive sessions.  In fact there were more injuries during the testing of 1RM than during the explosive sessions, suggesting that explosive concentric contractions at moderate load are less risky that slow eccentric contractions at very high load.

Optimising the explosive contraction

The greater increase of strength at higher load reported by de Vos was expected, while the greater endurance at higher load when assessed at 90% 1RM might largely reflect a bias towards maximal recruitment of the type 2 fibres during the endurance test that were maximally recruited during the explosive training.  However the similar gain in maximum power between the three different explosive training loads raises an interesting question about the possibility of biasing the training in favour of type 1 fibres.  At low load, during the eccentric training phase, the type 1 fibres will be preferentially recruited and hence experience gentle pre-tensioning.  During the explosive concentric contraction, it is likely that a wide-range of fibres would be recruited though the bias would usually be towards type 2 fibres.  However, if the type  1 fibres have been pre-tensioned during the eccentric phase, these fibres will tend to show relatively more enhancement of recruitment.  Thus, one might expect a somewhat greater benefit for type 1 fibres at low loads.

During the push off from stance during running (and during squats) much of the load is borne by the extensors of hip, knee and ankle.  Most of the relevant muscles cross two joints, flexing one while extending the other, and hence the fibres undergo a relatively short contraction during the triple extension.   Thus, even during quite fast running, the rate of contraction of the important muscles is relatively slow, and likely to be in the range where type 1 fibres are metabolically more efficient.  This suggests that explosive training with low load might be potentially of greater benefit for endurance athletes.

With regard to the type of exercise most useful for runners, any exercise that produces a triple extension is likely to be beneficial.  While deep squats do produce such an extension, the range of motion is different from that during running, whereas hang cleans require a range more similar to that of running.  However, it is trickier to do hang cleans safely.  It should also be noted that short, steep hill sprints and squat jumps are likely to be effective for increasing power.

The mechanism

The mechanism of the increase in power is not clear.  It is likely that improved fibre recruitment due to enhanced neuromuscular coordination plays a role.   If so, one might expect there to be a ceiling on the benefit that can be obtained over a prolonged period of training.  Nonetheless, on the principle that using a muscle prevents atrophy, there are likely to be long term benefits in ensuring that muscles can be recruited with peak efficiency.

My preliminary experiments

During the past year, during which I have focused more on increasing volume of training than on either strength training or sprinting, I have been dismayed to find that my sprinting speed has decreased yet more than in previous years.   Although I do not yet shuffle quite as noticeably as Charles Eugster (who is a little over a quarter century older than me) I too am forced to increase cadence to a very high level in order to produce speed.

About a year ago I had introduced plyometrics into my routine, but abandoned these as they appeared to be exacerbating the aches in my knees.  I am therefore eager to find a safe way to increase leg muscle power.  I have recently introduced sessions in which I do 3 sets of 8 explosive squats at 60% of 1RM. This has produced no DOMS and in fact leaves me feeling invigorated. I am even experimenting with doing 1 set of explosive squats at moderate load as part of my warm up for running sessions, with the expectation that this will enhance neuromuscular coordination.

It is too early to say whether or not this explosive resistance training has produced a worthwhile increase in ether sprinting speed or in efficiency at sub-maximal paces.  Nonetheless, the impressive gain in strength achieved by Charles Eugster though training with heavy loads has somewhat paradoxically inspired me to try a different approach with the aim not only of increasing strength but also recovering the power to get properly airborne.  But before I allow myself to become too dismissive of the approach employed by Charles Eugster, I should establish that I can run a 200m in 55 sec at age 75, let alone 95.

More reminiscences and the science of running injury free

January 5, 2015

The New Year is a time for looking forward, but in the past few days several things have also prompted me to take a backwards glance.   One of these things was the summary of the previous year’s statistics for my blog which Word-press compile each year. Two features intrigued me. First, was the very wide geographic range of the readership.  But even more intriguing was the fact that three of the most viewed pages in 2014 had been posted in 2010 or earlier.   It led me to wonder just how much my views on the topics of those posts have changed in the past five years.

Heart rate variability (HRV)

The most frequently viewed post was one of my 2010 posts on HRV. At that time I was intrigued by the possibility that regular HRV measurement might be an effective way to guide the adjustment of training load. In the preceding years, several studies by Kiviniemi and colleagues from Finland had provided preliminary evidence indicating that daily measurement of HRV was potentially useful for adjusting training load. In particular, a decrease in high frequency HRV was reported to be a useful indicator of the need for recovery. Although several recent studies have confirmed that HRV is related to the effects of training, there are inconsistencies in the findings, and I am not yet convinced that these studies provide compelling evidence that daily measurement provides a reliable basis for day-by-day adjustment of training load. Indeed a recent study by Plews and colleagues reports that daily HRV measurement is only weakly related to training status, though a weekly average was a quite strong predictor of 10K performance.

My own experience is that HRV, measured either at rest or immediately after exercise is not sufficiently consistent to provide a useful guide to adjusting training on a day to day basis. Nonetheless, I do regard a marked reduction in resting high frequency HRV as a pointer towards the need for better recovery when it occurs in association with other signs of stress. Conversely, unusually large amplitude HRV while running appears to be a fairly reliable sign of stress (as illustrated in  figure 3a from my post in 2010). However, the main reason I continue to record R-R traces during training runs is to monitor the frequency of occurrence of rogue rhythms – a common but inadequately understood phenomenon in elderly runners.

fig 3a: Expanded view of HR variability in the period 23 to 25 minutes, before (18 July) and during fatigue (31 August)

fig 3a: Expanded view of HR variability in the period 23 to 25 minutes, before (18 July) and during fatigue (31 August)


The second most frequently viewed post in 2014 was my 2010 discussion of the Dr Romanov’s Pose method of running. That post had drawn many comments at the time of the original posting and has attracted many thousands of views since 2010. Although I started the post with a list of the positive features of Pose, the main message was that Dr Romanov had seriously under-estimated ground reaction force (GRF) during running and as a result had drawn the erroneous conclusion that the net force acting on the runner’s body after mid-stance is directed forward and downwards. This led him to the erroneous conclusion that the centre of gravity falls after mid-stance. In contrast, a realistic estimate of GRF indicates that the net force after mid-stance is directed forwards and upwards. Dr Romanov’s mantra of ‘pose, fall, pull’ does not correspond to what happens. As a result, Dr Romanov seriously underestimates the crucial role of a well-timed push upwards and forwards off stance.

Five years later, I still regard this as the cardinal flaw of the Pose technique. The fact that my post continues to be widely read five years later indicates that the running community is still curious about Pose, but as far as I aware, serious discussion about running technique is now more focussed on how to achieve a well- timed and directed push after mid-stance.  My own opinion is that the most effective strategy is to rely on the cue provided by a brisk downward swing of the arm contralateral to the leg on stance.

Forces acting on a runner after mid-stance.

Forces acting on a runner after mid-stance.

Lydiard v Maffetone

The other ‘historical’ post among the five most viewed in 2014 was my comparison of Lydiard v Maffetone, posted in 2009. I have a great respect for both Lydiard and Maffetone, particularly on account of the emphasis they place on the importance of base-building in the annual training cycle. The purpose of my post was to compare their recommendations for paces during base-building. Lydiard was less precise in his prescription than Maffetone, but as far as can be estimated, both made very similar recommendation for long run pace. However, Maffetone is much more adamant in recommending avoidance of training in the upper aerobic zone during base building. In my post I discussed several of the plausible physiological mechanisms by which running in the upper aerobic zone might be counter-productive, but I concluded that none of the mechanisms was likely to impair aerobic development provided the majority of sessions were lower aerobic sessions. I therefore came down favouring of Lydiard over Maffetone

I still adhere to the main principles that I advocated in that post, though my opinion about optimum paces has changed over the past 5 years. As discussed in my recent post on the virtues of high intensity v high volume training I now favour a somewhat more polarised approach during base-building (and also during race specific preparation) than either Lydiard or Maffetone would have recommended. I think that it is more constructive to replace some of the ½ effort and ¾ effort runs recommended by Lydiard with short high intensity sessions, as I consider that the cost/ benefit ratio of high intensity sessions is greater than that of upper aerobic sessions. However I do not regard this issue as settled beyond doubt.

Candy for Running Nerds

Of all the comments I have received on my blog over the years, the one that most tickled my fancy was Eternal Fury’s remark a few years ago that my blog is a candy shop for running nerds. I assume that the candy to which EF referred was the detailed technical examination of the physics and physiology of running. Indeed I do enjoy looking into the biomechanics and physiology of running.

I believe that the best foundation for making sensible decisions about training and racing is a synthesis of practical observation of what works for others and one’s own experiences, interpreted in the light of what we know about how the body works. So I hope that the candy is both interesting entertainment for running nerds and informative for runners who simply aim to run as ‘fast and injury free’ as possible.

I was therefore very pleased two days ago when Steven Brewer posted a series of questions on my 2012 post ‘Is there a magic running cadence’. It was interesting to review the evidence that I had  presented in 2012, and to note that the evidence that has emerged since then has largely consolidated the picture from 2012. However as the discussion started to become even more arcane, Steven remarked ironically ‘The more I learn, the more I am discovering that running is an extremely complicated process. This makes me appreciate the simplicity of calculating the energy levels of an electron trapped in an infinitely deep well! ‘ At that point, Laurent Therond interjected ‘IMHO, running form is mostly innate, and what your body has come up with is likely to be what works best for you.’

In fact I agree strongly with Laurent. Running is largely innate and therefore we should largely trust our body to run in the manner that comes naturally. However, many runners suffer injury, and when this happens repeatedly we need to ask why.

So is it possible to distil the science of running into a few clear principles?    The short answer is no. Running involves every system of the body, ranging from the brain (and mind) to the toes, so it would be fat too simplistic to try to condense the essentials into a few principles. However, when I started this blog, one of my main pre-occupations was how to run ‘injury free’.   I believe that it is possible to distil the science of running injury free into a small number of basic principles – though as Einstein once remarked, we should seek explanations that are as simple as possible, but no more simple. So here is my response to Steven in which I have tried to distil the science of running injury free, in a manner that is as simple as possible but not too simple.

The distilled science of running ‘injury free’

In the past decade I think there has been an over-emphasis on   faults of style as the major cause of injury. I believe that subjecting the body to greater stress than it can cope with is the major causal factor in in many instances. Therefore one of the main strategies for avoiding injury is building up training load gradually, and at all times recognising when the body is near the limit of coping. This is not easy, though a variety of signs covering both overall well-being (mood; heart rate variables) and focal signs such as accumulating aches in connective tissues or local muscle spasm provide useful clues.

However, in some instances faulty technique plays a role. But there is a great deal of misleading information in circulation, due to over-simplistic analysis. The human body is very complex. Nonetheless, it does obey several well established principles, based on both Newtonian mechanics, and the principles governing the behaviour of biological tissues.

Newtonian mechanics

The first things to consider are the laws of Newtonian mechanics. We need to meet two goals: minimising energy expenditure and minimising stress on body tissues. Minimising energy expenditure requires the optimum adjustment of the three major energy costs: getting airborne; overcoming braking and the cyclical repositioning the limbs relative to the torso. Simplistic focus on just one of these leads to over-emphasis on a single requirement. For example, we can minimise the cost of getting airborne by reducing vertical oscillation, but that inevitably increases the cost of overcoming braking, at a given cadence.   We can minimise both braking costs and cost of getting airborne by increasing cadence, but that increases limb positioning costs.   Conversely, if cadence is too slow there is a strong temptation to increase stride length by reaching out with the swinging leg, thereby incurring braking costs that are inefficient and potentially damaging.

The need to balance the three main costs leads to 2 important conclusions: first, simplistic rules such as avoid vertical oscillation without considering other costs are misleading and potentially dangerous. In general, it is best to let the non-conscious brain do the calculation required to optimise the three energy costs, because the brain is usually very good at finding the most efficient solution.

However, as discussed in most post on running cadence, there appears to be one exception: most recreational runners select a cadence which appears to be slower than optimal for fast running. Maybe this is because we are naturally adapted for slow running over rough terrain where maintaining balance is a high priority; alternatively it might be that a sedentary lifestyle encourages our brains to be too cautious. So if there is one simple issue we should address consciously, it is making sure that cadence is not too slow (eg at least 180 steps/minute at 4 m/sec).

The second general conclusion arising from the need to balance the three costs, is that large forces will necessarily be exerted. A quite strong push against the ground is inevitable. Running is a dance with the devil – gravity. We therefore must build up the strength to cope with this, by gradual increase in training volume. Specific exercises, both plyometrics and resistance exercises, can help.

Associated with the large forces required to get airborne is the fact that the geometry of our hips knees and ankles results in quite large rotational effects around two, or sometimes all three axes, at these joints. Some of these effects are not easily envisaged, so there are certain injuries that do require a careful biomechanical analysis. Nonetheless I consider that gradual build-up of training load and trusting your non-conscious brain is the most effective way of minimising the risk of such problems. If you do decide that you need to adjust style consciously, it is safest to focus on a compact, brisk but relaxed arm action.   This tends to promote a compact efficient leg action as a consequence of the way the brain codes movement.

The physiology of injury

With regard to the physiological properties of body tissues there are several principles that are worth considering. The first is that running mobilises the catabolic hormones (adrenaline, cortisol) necessary to promote energy metabolism. Catabolic processes break down tissues that must subsequently be prepared by anabolic processes. The art of training is largely about balancing catabolism and anabolism – the most important requirement is ensure that recovery is adequate.

The second valuable principle is that repair of damage tissues involves an inflammatory process that often involves laying down of collagen fibres in a randomly ordered manner. However, the relevant muscle or tendon functions best when the fibres are aligned in the direction of the usual forces, so once the initial inflammation has settled, gentle active recovery helps re-model the tissues in the optimum manner.

Overall, the science of running is complex. Some problems are indeed more challenging than applying quantum mechanics to calculate the energy levels of an electron trapped in an infinitely deep well, but a few simple principles provide most of the guidance a runner needs to maximise the chance of running injury-free.

Running Efficiency

December 2, 2014

Most endurance athletes focus their training on attempting to increase their aerobic capacity (VO2max) and their endurance. Training increases aerobic capacity by increasing the ability to deliver oxygen to muscle fibres and by increasing the capacity of mitochondrial enzymes to generate energy by oxidation of fuel. However our maximum capacity to generate energy by oxidation of fuel appears to be limited by our genes and/or early development.

The traditional approach to improving endurance is the long run. This increases resilience of muscles, tendons and ligaments, and enhances the ability to metabolise fats.   Once we have trained our aerobic capacity to our limit, and we have developed sufficient endurance to sustain us for the duration of our target event, what scope is there for further improvement in performance? The remaining option is increasing efficiency: that is the effectiveness with which we can use energy to produce speed.


Figure 1 shows speed (in metres/min plotted) against rate of energy production (VO2 measured in ml/min/Kg) for three hypothetical athletes. The slope of each line represents the efficiency of each athlete.   The line with medium slope represents the average runner in the sample used by Jack Daniels to derive the VDOT charts in his book ‘Daniels Running Formula’. The steeper line represents an athlete 10% more efficient than average. The less steep line represents an athlete 10% less efficient than average. Greater efficiency indicates a greater speed for a given rate of energy production. Note that the lines are almost straight, indicating that efficiency is nearly constant across a wide range of paces, though there is a minor degree of flattening of each line at higher paces, indicating a somewhat lesser increase in pace for each additional unit of oxygen consumed.

Figure 1: The relationship between pace and aerobic energy production.  These lines are derived from the data used by Jack Daniels to derive his VDOT tables. The middle line (brown) is the data for an athlete who had the average efficiency from the sample studied by Daniels. The upper (blue) line represents an athlete who is 10% more efficient than average.  The lower represents an athlete who is 10% less efficient than average.

Figure 1: The relationship between pace and aerobic energy production. These lines are derived from the data used by Jack Daniels to derive his VDOT tables. The middle line (brown) is the data for an athlete who had the average efficiency from the sample studied by Daniels. The upper (blue) line represents an athlete who is 10% more efficient than average. The lower represents an athlete who is 10% less efficient than average.

If each of the three hypothetical athletes had a VO2mx of 72 ml/min/Kg (typical of an elite distance runner) the athlete with average efficiency would achieve a pace of 350 metres/min at VO2 max while the athlete who was 10% more efficient would achieve a pace of 385 m//min. At 80% of VO2 max (57.5 ml/min/Kg), the athlete with average efficiency would be expected to achieve a pace of 293 m/min while the athlete who was 10% more efficient would achieve 322 m/min. It should be noted that for an athlete with a lower VO2max, the pace at VO2 max and at any given percentage of VO2max will be less, but the relative gain in pace from an increase in efficiency will be similar. In other words, for two athletes with the same VO2max, a 10% improvement in efficiency would result in 10% faster times in races run at any given proportion of VO2max.

Might training produce an increase in efficiency of 10% or more?   The measurements of Paula Radcliffe performed by Andrew Jones for more than a decade provide clear evidence that the answer is yes. In fact, the data shows that Paula achieved a 15% increase in efficiency over the decade from 1993 to 2003*. Her VO2max remained virtually constant at around 70 ml/min/Kg over this period. Thus, a major factor in Paula’s phenomenal marathon record of 2:15:25 recorded in 2003 appears to be the remarkable improvement in efficiency. How might an athlete improve efficiency?  There are two possibilities: increasing biomechanical efficiency and increasing metabolic efficiency.

Biomechanical efficiency

There are three major energy costs of running:

  • overcoming the braking that occurs while on stance;
  • getting airborne;
  • swinging the leg forwards after lift-off from stance.

There is also the cost of unnecessary tension or movements of other body parts, but as is well illustrated by Emil Zatopek and Paula Radcliffe, who both achieved phenomenal performance despite unnecessary upper body movements, the cost of such movements is relatively small, and there is unlikely to be more than a slight gain from reducing them.

Minimizing the sum of the three major costs requires a balance between conflicting effects. At a given cadence (steps/min) braking cost increases as the cost of getting airborne decreases because less time in the air inevitably results in a larger proportion of time on the ground. Although braking only occurs when the point of support is in front of the centre of gravity, braking cost cannot be reduced merely by attempting to land with the foot under the body, because at constant speed the forward-directed impulse generated after mid-stance and the backward directed impulse generated by braking before mid-stance must be equal (after allowing for overcoming wind resistance).  For a give cadence, the cost of braking can only be reduced by spending more time airborne.

During a marathon, many runners spend an increasing proportion of the time on stance as the race progresses. This is likely to result in greater braking and reduced efficiency. It is noteworthy that the well-known picture of Paula Radcliffe at mile 14 on her way to victory in the 2007 New York marathon shows her getting well-airborne. This demonstrates that she had adequately developed reserves of the leg muscle power required to get airborne. Andrew Jones’ measurements demonstrated that her vertical jump performance increased from 29cm in 1996 to 38cm in 2003.

Paula Radcliffe airborne at mile 14 in the New York marathon, 2007.  Photo by Ed Costello, Brooklyn, NY,US

Paula Radcliffe airborne at mile 14 in the New York marathon, 2007. Photo by Ed Costello, Brooklyn, NY,US

The cost of getting airborne can be reduced by increasing cadence, because the body falls a lesser distance during a series so short hops than during a longer hops covering the same distance, simply because a freely falling body accelerates, thereby gaining greater speed the longer it is airborne. The optimum cadence increases with increasing speed, because if cadence does not increase with increasing speed, stride length would necessarily have to increase disproportionately, resulting in heavy costs of getting airborne and also braking. However, there is a limit to the gains that can be achieved by increasing cadence, because the cost of moving the swing leg forwards increases in proportion to cadence (as shown on the my calculations page).

Nonetheless, many recreational athletes have scope for increasing efficiency by increasing cadence. The study by Heiderscheit and colleagues indicates that a typical recreational runner might improve efficiency by decreasing both airborne costs and braking costs by increasing the self-selected cadence by up to 10% . This increase in cadence also reduces stress at the joints by virtue of the reduction in forces required to get airborne and overcome braking.   Heiderscheit reported that a 10% increase in step rate from a self-selected mean step rate of 172.6 ± 8.8 steps/min at a pace of 2.9 ± 0.5 m/s led to an almost 20% reduction in energy absorbed at hip, knee and ankle joints.

It is probable that Paul Radcliffe achieved optimum balance between the cost of getting airborne, braking and advancing the swing leg largely by virtue of fairly intense running, together with hopping drills and weight lifting.   While training near to race pace might optimise neuromuscular coordination, I suspect that the major requirement for optimising mechanical efficiency is adequate muscle power. Although I do not have direct evidence to prove it, I think it is plausible that a small amount of high intensity training will achieve as much gain in mechanical efficiency with less total wear and tear on the body compared with a larger volume of threshold training, simply because training near to maximal effort is more effective for improving muscle strength and power.

Metabolic efficiency

Metabolic efficiency of oxygen consumption is a measure of the amount of mechanical work (and hence speed) that can be achieved from the consumption of a given amount of oxygen. Several factors influence this. The most important is the fact that the efficiency of conversion of metabolic energy to mechanical energy during contraction of a muscle fibre is greatest when the speed of contraction is near the middle of the range of contraction speed that can be achieved by that fibre. When a fibre contracts too slowly it consumes energy developing tension that does little work. Fast twitch fibres have an optimum speed of contraction that several times faster than that of slow twitch fibres. But the speed of fibre shortening during distance running (and also cycling) is better matched to the optimum contraction speed of slow twitch fibres.

It is noteworthy that many of the large muscles that act at hip and knee cross both joints, flexing one while extending the other or vice versa. However during running hip and knee flex simultaneously or extend simultaneously. Consequently, the rate of change in length in these muscle during running is small. Thus type 1 fibres which are well suited to the isometric contractions required to maintain upright posture are also well suited to distance running during which contraction rate is slow.

In the case of cycling, there is direct evidence that efficiency of metabolic to mechanical conversion is greater in individuals who have a higher proportion of type 1 fibres. Although I do not know of any similar measurements in runners, it is very likely that runners with more highly developed type 1 fibres will be more efficient.

The most effective way to develop type 1 fibres is likely to be consistent high-volume training over a sustained period. It is likely that a major part of Paula Radcliffe’s improvement in efficiency was consistent training, with a gradual increase in training volume over a period of a decade. As discussed in my previous post, Paula did a lot of her training at a moderate or high intensity. It remains a matter of speculation as to whether she could have achieved similar phenomenal marathon performances with less damage to her body by a more polarised approach, in which a modest amount of high intensity running was accompanied by a larger proportion of low intensity running. Perhaps she could have achieved similar improvement in metabolic efficiency with a larger proportion of low intensity training over a longer period of time. My own view, based on Skoluda’s evidence that many distance runners have evidence of sustained high levels of the potentially harmful catabolic hormone, cortisol, is that for many athletes a polarised approach offers the best prospect of gradual improvement in metabolic efficiency and hence, the prospect of year-on-year improvement over many years.


Paula Radcliffe’s spectacular 15% increase in efficiency over a period of about a decade, despite an approximately constant VO2max, provides compelling evidence that a worthwhile enhancement of efficiency is possible. It is likely that a combination of high intensity training, hopping drills and weight lifting honed her biomechanical efficiency.  For many recreational athletes, biomechanical efficiency might also be improved by increasing their self-selected cadence by as much as 10%.  It should be noted that optimum cadence increases with speed.

It is also likely that a gradual improvement in metabolic efficiency over a period of more than a decade was also a major contributor to Paula’s improved efficiency.  It is likely that she achieved this by consistent training, with a gradual increase in volume over the years.  Whether or not she might have achieved a similar enhancement of efficiency with a less damaging, more polarised approach to training remains a matter for speculation. Nonetheless, in my opinion, for many athletes, a polarised approach is likely to offer the best prospect of gradual improvement in metabolic efficiency over a period of many years.


*A minor point to note is that Andrew Jones estimated speed at VO2max by assuming a linear increase in pace with increased VO2. This is likely to produce a small over-estimate of actual pace at VO2max, because in reality the curve flattens a little at high values of VO2. Nonetheless, provided all the measurements are made at a similar region of the curve, the error in estimate will be consistent across different measurements. It is pace at around 80% of VO2max that matters most to a distance runner.

Getting maximum benefit from low intensity training

October 3, 2014

In my recent posts I have discussed the evidence suggesting that if one’s goal is year-on-year improvement in marathon performance, the best approach is a polarised program including a large volume of low intensity training and a small volume of high intensity training. Since at least 80% of the training time during a polarised program is devoted to low intensity runs, it is worth considering how to derive the greatest benefit for these sessions.

First, I should make it clear that I believe in a periodized approach that includes a base-building phase and a race-specific phase. Both phases should be polarised, each embracing both low intensity and high intensity training. In the base-building phase, the goal is to build all of the basic physiological capacities required for marathoning. In this phase, the low intensity sessions play a crucial role in developing several of these capacities, but high intensity sessions also play a key role. In the race-specific phase a small proportion of the sessions are devoted explicitly to developing the mental and physical strength required for racing. This post will deal mainly with the base-building phase. Nonetheless, even in the race-specific phase it is crucial to maintain the capacities developed during base-building, so many of the principles apply to planning training in both phases. I will address the specific requirements of race-specific phase in greater detail in a future post.

In my post of 20th September I listed the five physiological capacities that need to be trained in preparation for a good marathon:

  • VO2max – this measures the maximum rate at which oxygen can be delivered to tissues and hence the maximum rate at which muscles can generate energy.
  • Speed at VO2max.
  • Pace at lactate threshold as a proportion of pace at VO2max. For a well-trained marathoner, race pace is near to lactate threshold.
  • Ability to conserve glycogen so that glucose supply is not exhausted before 26.2 miles.
  • Resilience of leg muscles to sustain pounding for the duration of the marathon with only minimal loss of power.

Each of these five capacities can be enhanced by several different types of training stimulus, so full development each of the variables requires a complementary mix of low and high intensity training. The low intensity session play an especially important part in increasing three of the five capacities: VO2max; conservation of glycogen; and developing resilience of leg muscles, so we should consider each of these in turn.

Enhancing VO2max

The two major ways in which low intensity training enhances VO2 max are by increasing the aerobic enzymes in mitochondria; and by enhancing the capillaries that deliver blood to muscle fibres. These physiological processes can be examined in muscle biopsies from athletes, but more detailed information can be obtained from the study of animals. Despite the large scale anatomical differences between the human and rodent musculo-skeletal systems, at the level of individual fibres the structure and functional of muscles are similar across the species so studies of the ways rodent muscle fibres respond to training are likely to be informative about the ways in which human fibres respond.

One of the major questions in planning a training schedule is whether it matters whether the training is done in multiple short session or fewer long sessions. The long slow run has been regarded as a core feature of marathon training for years, though the data for study of animals suggest that with regard to enhancing aerobic enzymes and capillaries, multiple short runs might be just as beneficial.

In a recent study. Malek and colleagues trained mice on a treadmill on 5 days a week for 8 weeks. One group of mice trained for 30 minutes continuously on each training day, while another group of mice trained for 3 periods of 10 minutes separated by a 2 hour rest. The intensity of training was equal.  The animals started at the rather slow pace of 7.5 metres per minute, and over the 8 weeks, increased the rate of working up to 60% of maximum power output. A control group of mice were placed on the treadmill but did no training. Compared to the untrained controls, both groups of trained mice exhibited similar major improvements in both speed and distance covered during an incremental treadmill tests administered at the beginning and end of the training period. The untrained controls exhibited a 1% increase in the distance covered in the incremental test after 8 weeks, while the group who trained continuously for 30 minutes exhibited an increase of 107% and the group who trained for 3×10 minutes exhibited an increase of 117%. . Furthermore, after training, the capillary density and the number of capillaries per muscle fibre in quadriceps muscle were approximately twice as great in the two training groups compared with the untrained controls. Similarly the amount of the aerobic enzyme, citrate synthase, in the plantaris muscle in both of the trained groups was about twice as great as in the controls. Thus, running performance, muscle capillary density and aerobic enzymes showed large, but similar, increases in both trained groups. This suggests that with regard to improving VO2max , three 10 minute training sessions produce very similar enhancement to one 30 minute training session.

On the other hand, Dudley’s well known studies, in which he trained rats at several different speeds for varying lengths of time per day showed that very long sessions did not produce greater enhancement of aerobic session that medium length sessions. The greatest gain in cytochrome C (a complex of aerobic enzymes) occurred in rats trained at a pace of 30 metres/min. Thirty metres/min is a medium pace for a rat; typically a rat can maintain 60 metres/min for about 10-15 minutes whereas it can sustain 30 metres/min for an hour or so with little difficulty. The rats that trained at 30 m/min showed a steady increase in enhancement of cytochrome C in red soleus muscle with increasing daily run time up to 60 minutes, but the there was only a slight further training effect in the group who trained for 90 minutes, suggesting no benefit with regard to aerobic enzymes beyond 90 minutes..

It is important to note that Dudley’s rats did not have much opportunity to adapt to the training load.   After an initial five day introductory period of running 5-10 minutes daily at approximately 30 m/min, they were allocated to the designated training and the training load was ramped up at the rate of 12 additional minutes each day, ensuring that the animals allocated to train for 90 minutes per day were training at full load by the end of the second week. It is possible that the rats allocated to 90 minutes per day at a pace of 30 m/min were not given adequate time to adapt to the training load. Inadequate adaption to the load would be expected to result in excessive release of cortisol which has a damaging catabolic effect on muscle.

Taken together, the findings from Malek’s study of mice and Dudley’s study of rats, indicate that there is little difference in the training benefit derived from a single continuous session of 30 minutes compared with three session of 10 minutes, while the benefits of increasing the length of sessions beyond 60 minutes are small in the absence of an adequate period to adapt to the training load. It is speculative, but I consider quite plausible that provided long run duration is built up gradually, that additional benefit will accrue beyond 90 minutes. Nonetheless, with regard to enhancing erobic capacity, the evidence suggests that multiple short runs are likely to be at least as effective as a similar total duration of long runs. This would suggest that for VO2max development, doubles might be at least as effective, and perhaps more effective in terms of time spent and stress sustained than single daily sessions. However, enhancing VO2max is not the only goal of training.

Enhancing conservation of glycogen.

Endurance training produces an increase in the proportion of energy derived from fat, across a wide range of intensity of exercise. Although it is well established that in both trained and untrained individuals, the proportion of energy derived for fat is less at paces above lactate threshold than at paces below threshold, nonetheless, even above threshold, trained athletes derive a larger proportion of total energy from fat than untrained individuals. [See the opinion piece by Coggan.]

While many studies have demonstrated that trained individuals derive a greater proportion of the energy required during exercise from fat, compared with untrained individuals, there have been only a few longitudinal studies that have demonstrated the efficacy of a specific endurance training schedule for enhancing fat metabolism. One particularly informative study was done by Henriksson. He subjected 6 cyclists to a training program in which only one leg was trained for 45 min/day at 70% of VO2max (estimated for one leg) for an average of three days per week for a period of 8 weeks.

Before and after the training period, a muscle biopsy was taken from quadriceps for determination the activity of the aerobic enzyme, succinate dehydrogenase. The subjects were also tested on different submaximal and maximal one-legged and two-legged workloads. Catheters were inserted into the femoral arteries and veins at the groin in both legs to allow measurement of blood oxygen and carbon dioxide levels. In the submaximal test, participants performed two-legged exercise for 1 hour at 67% of VO2 max.  In the trained leg, there was a substantial greater capacity of muscle to extract oxygen from blood, demonstrated by increased arterio-venous difference. The respiratory exchange ratio (RER) was 0.91 throughout the 1 hour in the trained leg. Respiratory exchange ratio is the ratio of the amount of CO2 produced to O2 consumed. It has a value of 1 when carbohydrate is the sole fuel and a value less than 1 if fat is included in the fuel mixture. Thus the value of 0.91 indicates that a substantial proportion of energy was obtained from fat. In contrast, in the untrained leg, RER after 10 minutes was 0.96 indicating that the majority of the energy was obtained from carbohydrate metabolism and even after 50 minutes, the RER in the untrained leg was 0.94. Thus the proportion of energy derived from fat had increased as might be expected if glycogen stores were being exhausted, but nonetheless even after 50 minutes the proportion of energy derived from fat in the untrained leg was less than in the trained leg.

Henriksson estimated that in the trained leg, 42% of energy was obtained from glycogen while in the untrained leg, 62% of energy was derived from glycogen. Thus, the increased utilization of fat in the trained leg resulted in appreciable conservation of glycogen. Furthermore, rate of lactate release in the untrained leg was between 2.5 and 3 mmol/min in the period from 10 to 30 minutes yet remained below 0.5 mmol/min in the trained leg, confirming the expectation arising from the fact the fat metabolism does not generate lactate.

It was also of interest to note that measurement of free fatty acids in the blood demonstrated that the increase fat metabolism came largely for increased consumption of triglycerides stored within muscle. The mechanism of this increase was not clear. The activity of aerobic enzyme succinate dehydrogenase increased in the trained leg. This increase in aerobic enzymes would lead to faster fat metabolism. However it is probable that an increase in enzymes involved directly in fat metabolism and also an increase in ability to transport fats into mitochondria played a part

In summary, the study by Henriksson established that moderate intensity exercise (70% VO2max) for 45 minutes 3 days per week produced a substantial enhancement of fat metabolism, thereby conserving glycogen and reducing the production of lactate.   He did not address the question of whether or not a greater benefit would have been obtained from longer duration of exercise. However the observation that even in the untrained leg the proportion of energy derived from fat increased appreciably by 50 minutes confirms the expectation that a long duration low intensity session would be more effective for enhancing fat metabolism than multiple short duration sessions.

Thus, it is likely that long runs lasting an hour or more are the most effective for increasing fat metabolism, but it is necessary to bear in mind that unless long run duration is increased gradually, there is a risk of increased release of cortisol. Similarly, training in a fasted stated would be expected to produce greater enhancement of fat metabolism but care should be taken to avoid excessive stress. As I suggested in my previous discussion of training in the fasted state, I suspect that the inconsistent results reported by different studies of training in the fasted state reflect differences in the degree to which there was adequate adaptation to training in a fasted state.

Resilience of leg muscles

Eccentric contraction at foot strike causes microscopic tearing of muscle fibres. This is likely to be the major factor in the production of Delayed Onset Muscle Soreness (DOMS). However, as virtually every athlete knows from direct experience, after DOMS, the muscle adapts rapidly to prevent subsequent damage if the same exercise is repeated.

Although many recreational athletes preparing for a marathon experience some DOMS early in the program, they quickly develop sufficient resilience to prevent serious DOMS in subsequent weeks. However, the observation that many experience a recurrence of severe DOMS in the aftermath of the race itself, indicates that they failed to develop adequate resilience to cope with the sustained pounding of the marathon.   While a few days of DOMS after the event might be of little consequence, the effect of microscopic damage during the event is potentially of much greater importance. Muscle damage during the marathon appears to be one of the major causes of slowing down in the second half of the race. Therefore, development of adequate resilience is a high priority.

During the race itself it is the combination of the duration of the stress and the intensity of stress that does the damage. It is therefore likely that a multi-facetted approach, involving both long duration sessions and some more intense running is required to build the required resilience. An approach that relies too heavily on extending the duration of intense exercise would be risky. Although complete recovery from DOMS is usual provided there is adequate subsequent opportunity for recovery, if there inadequate opportunity for recovery, the muscle might eventually lose its capacity to recover. It appears likely that the relatively rare Fatigued Athlete Myopathic Syndrome (FAMS) is the end stage of repeated microscopic muscle trauma without adequate opportunity for recovery. In cases of FAMS the athlete exhibits marked disruptions of muscle microstructure and suffers loss of the ability to tolerate further training.  On the other hand, repeated exposure to a small stress can protect against future larger stresses, so long slow runs are the safest way to establish the foundation for the resilience required to withstand the repeated pounding at marathon pace


The evidence provides very strong grounds for arguing that a substantial volume of low intensity training is an effective complement to a small volume of more intense training during marathon preparation. In particular, the low intensity training is a safe way of enhancing VO2 max; promoting the ability to conserve glycogen (and incidentally reducing the production of lactic acid); and lays the foundation for the muscle resilience required to avoid slowing in the later stages of the race.

This evidence also provides guidance regarding how best to schedule this training. The first point to note is that multiple short session are likely to effective in enhancing VO2max in a safe manner. However, at least some longer duration sessions are also required to optimise the ability to conserve glycogen via utilization of fat, and to develop the required resilience.

The traditional answer to these requirements is to incorporate a weekly long run of gradually increasing length in the training schedule. This traditional answer certainly has merit. However, it is possible that this strategy relies too heavily on the weekly long run. For example, building up the long run length from less than 10 miles to more than 20 miles within the course of a 12-16 week program might not provide adequate opportunity to adapt adequately to the demands of the long run. This has two consequences. First, the long run itself might leave the runner tired and aching, limiting the quality of training on subsequent days. Secondly, there is a substantial risk that one long run per week will prove inadequate for developing the resilience required to maintain pace for the full 26.2 miles.

An alternative to placing so much of the emphasis on a single weekly long run is to add small increments to the duration of several runs each week, thereby creating a more uniform build-up of training load throughout the week and avoiding the disruptive influence of a single long run. This is the approach pioneered with great success by Ed Whitlock. He gradually built the capacity to cope with three or four 3 hour slow runs each week, in preparation for his phenomenal 2:54:48 in the Toronto Waterfront marathon at age 73. Ed acknowledges that although this worked so well for him, such an approach might not work so well for others. However, in my view the evidence we have considered in this post provides reasonable grounds for expecting that Ed’s approach might work well for others.   The evidence that capillaries and aerobic enzymes can be developed effectively in multiple relatively short sessions and that appreciable improvement in fat metabolism can be achieved with session of 45 minutes duration, while on the other hand, long sessions without adequate foundation create risk of excessive release of damaging catabolic hormones, suggests that gradual build-up of the duration of multiple longish runs each week might be effective.

It is important to note that the two crucial features of Ed’s approach are the multiplicity of long runs each week and the gradual increase in length of these long runs. While it is true that by age 73 most of his long runs were of 3 hours duration, he had built up to that duration over 6 years. When he first set a single age world record for the marathon at age 68, the majority of his long runs were only two hours in duration. He did not measure the distance of these runs, but as far as I can estimate from his own description of these easy long runs, it is unlikely that in 2 hours he ran more than 14 miles. I think that the evidence we have considered above suggests that the important feature is the consistency throughout the week rather that the focus on a single very long run each week.

There is one respect in which I am inclined to recommend a difference from Ed’s approach. His high intensity sessions took the form of frequent races over shorter distances, and occasional fartlek sessions, but no specific preparation for maintaining marathon pace throughout the race. I think that in the final 12-16 weeks of preparation for a marathon, one of the weekly long runs should be replaced by a race-specific training session aimed at developing the required mental and physical strength for racing. However I will defer detailed discussion of ways of implementing this to a future post.

I am at present working on gradually increasing the duration of four runs each week.   I started with four 65 minute easy runs in the first week of this program, and after 6 weeks have increased the duration of these four longish runs to 105 minutes each. I am also doing at least one high intensity session per week and one or two other short sessions. I am carefully monitoring the rate of increase in duration of these four runs to ensure that there is no cumulative tiredness. It is still far too early to deliver judgment on this strategy. However I have now built up to a weekly volume that is almost as high as the highest I have achieved at any time in the past few years. Last year and also the year before, I had found that after a few weeks at this volume of training I was showing signs of accumulating exhaustion and therefore was obliged to cut back the volume of training. If I continue to make small advances in run duration without accumulation of tiredness over the next few weeks I will have the first indications that this strategy is working.

Grete Waitz and Paula Radcliffe: do they make the case against polarised training?

September 15, 2014

My interest in polarised training was piqued more than five years ago by the study of the effects of a five month polarised training program in sub-elite cross country runners by Esteve-Lanoa and colleagues from Madrid. They provided a fairly convincing demonstration that a polarized training program in which about 80% of the work is done at low intensity, is more effective that a program including a higher proportion of work in the mid-zone, near to lactate threshold.   A subsequent review article by Stephen Seiler and Espen Tǿnnessen published in Sportscience in November 2009 presented a quite compelling argument in favour of polarised training: a large amount of low intensity training, together with a small amount of high intensity training, and a minimal amount on the intervening grey zone around lactate threshold. That review confirmed the direction of my own thinking about endurance training, so I posted a positive commentary but included a cautionary note.

The scientific method is mankind’s most successful way of making and testing predications about the natural world, but individual scientists are not dispassionate observers. The strength of science comes from the debate between scientists.   In this debate, each individual scientist tends to be biased towards the evidence that supports his/her own hypothesis. In my comment on Seiler and Tǿnnessen’s review, I noted that Seiler was a co-investigator in the study by Esteve-Lanao. Furthermore he and Tǿnnessen had been selective in the evidence presented in their review article. They reported major improvements in both performance and in physiological variables such as VO2max, after a change to a training program including a higher proportion of low intensity training, in the case of two Norwegian athletes: pentathlete and runner, Øystein Sylta, and cyclist Knut Anders Fostervold, but made no mention of Norway’s greatest female marathon runner, Grete Waitz . Waitz won the New York marathon 9 times, was the silver medallist in the 1984 Olympics in Los Angeles, and won gold at the 1983 World Championships. She did a large amount of training in the grey sub-lactate threshold zone.

Since 2009, the evidence in favour of polarised training has become even stronger, supported by both experimental studies such as that of Stoggl and Sperlich, and further examination of the training of elite athletes. My own recent examination of the training of seven elite masters marathoners led me to conclude that those who employed markedly polarised training had the greatest longevity at the top of the world rankings. I had selected the seven on the basis of predefined criteria that I knew would be satisfied by both Ed Whitlock, who employs markedly polarised training, and Yoshihisa Hosaka, who does twice daily interval sessions. When I set the criteria I was not sure who else would meet the criteria and therefore had little way of knowing what training patterns would be represented in the sample. However, despite my intention to be as dispassionate as possible, I am aware that my own beliefs about training influenced my presentation of the evidence.

In a comment on the Fetch polarised training thread, I was challenged over my failure to examine the training of any female marathoners in my blog. In fact no female marathoners had met my predetermined selection criteria, though Miyo Ishigami of Japan came nearest to meeting these criteria. She set the W55-60 record with a time of 2:57:55 at age 55 in 1989 and remained near the top of the rankings up to age 75 when she recorded 4:27:42. That is the 7th fastest ever for a lady in the 75-80 age group. Furthermore, there is an additional problem in examining the training of female ‘masters’ marathoners: information about their training is less accessible.

However there is abundant information about two younger female marathoners who played key roles in the transformation of the women’s marathon over period of a quarter century: Grete Waitz who took more than 2 minutes of Christa Vahlensieck’s world record of 2:34:48 in her first marathon in New York in 1978 and subsequently broke her own record on 3 occasions; lowering it to 2:25:29 in London in 1983; and Paula Radcliffe, whose 2:15:25 London in 2003 remains unchallenged as the outstanding women’s marathon performance in history. Both are famous for the demanding nature of their training. An examination of their training offers the prospect of putting polarised training into a more balanced perspective.

How do sex differences in physiology affect marathon performance?

Before examining the training of these two individual athletes it is potentially informative to address the question of whether the optimum training for women should be different from that for men.   It might be predicted that the lesser muscular strength of women would be a lesser handicap in the marathon than in shorter events, but this is not borne out by evidence. The proportional difference of almost 10% between Paula Radcliffe’s record and Wilson Kipsang’s male marathon record is similar to the proportional difference between female and male world records across track events from 100m upwards. There no longer a strong reason to claim that cultural bias against women running long distances accounts for the handicap. The fact that Paula’s record has stood for over a decade despite prominent recognition and prize money for the women similar to that for the men in many of the major marathons, indicates that the differences are likely to be mainly due to physiological differences. It is not clear which of the physiological differences plays the greatest role. Perhaps the lesser power of both cardiac and skeletal muscles in females does matter in the marathon as it does in shorter events.  Thus there might as much, if not more, reason for females to train in a manner that promotes cardiac and skeletal muscle power.

This is borne out in a study of the training of qualifiers for US Olympic marathon trials by Karp. Across the entire sample, the men ran more miles in training, though interestingly the women who achieved times less than 2:40 had a similar training volume to the men.   But more intriguing was the observation that the women did a higher proportion of their training at marathon pace or faster.  The women did 32% and the men only 25% at marathon pace or faster.   Perhaps a fast marathon does require power and these elite or sub-elite women got to the Olympic trials as a result a large proportion of relatively more intense training. Overall, there is at present little evidence to indicate that women might do better with relatively less demanding training than men. Perhaps they might benefit from an even higher relative intensity of training, and/or increased focus on building strength.

Grete Waitz

Grete’s husband, Jack, persuaded her that a trip to New York for the marathon would be like a second honeymoon for them, despite the fact that she had never run more than 12 miles in training. At the finish she took her shoes off and threw them at him declaring ‘never again’. However she had just taken over two minutes off the world record time, and despite her protestations, the marathon bug had bitten, just as she was contemplating retirement from international competition.  Three years earlier, in 1975, she had broken the world record for the 3000m, but after a disappointing race in the 1978 European Championships in Prague, she was planning to return to her full-time job as a school teacher.   However, earlier that year, in Glasgow in March, she has won the world cross country championship, and I suspect that Jack, who was her mentor and coach at that time, had a premonition that she had the makings of a marathon runner.   If so, he was right. She went on to win the New York Marathon on nine occasions in 11 years, the London marathon twice and the World marathon championship in Helsinki in 1983.

She had made the world record her own in New York in 1978, and by April 1983, when she lowered it for the fourth time in London, it stood at 2:25:29, almost 10 minutes lower than it had been when she stood unassumingly on the starting line in New York 5 years previously. However the day after her record-breaking victory in London, over 3000 miles away in Boston Joan Benoit took two minutes off Grete’s time. The following year, on a hot day softened by Los Angeles’ morning fog, in the inaugural women’s Olympic marathon, Joan made a bold early break from the leading group. Grete prudently held back but Joan’s confidence was justified. She took the gold, leaving Grete with the silver.   The women’s marathon was now an established event. Despite being deprived of Olympic gold by a worthy challenger. Grete had perhaps done more than any other person to raise the women’s marathon to Olympic standard.

Grete continued to perform at leading international standard, with another victory in New York later that year, and again in 1985, 86 and 88. She won the London marathon for a second time in 1986, during which she achieved her own personal best time of 2:24:54, but she was never again to hold the world record.

By 1990, at age 36, she was beginning to fade. She was fourth in New York in a time only marginally faster than the world record had been before she ran her first marathon twelve years previously. Nonetheless, she returned to New York two years later to run side by side with Frank Lebow, founder of the New York marathon who was at that time in temporary remission from a lethal brain cancer. When they crossed the line together in a time of 5:32:35 and raised their entwined arms in celebration, she and he cemented their intertwined places in the annals of marathoning. When Grete herself died on cancer at age 57, in 2011, the flood of tributes from marathoners of all levels that accompanied the article in the New York Times reporting her untimely death, confirmed that this tenacious, determined but humble and gracious woman indeed merited one of the highest places in the Pantheon of the marathon.


The training of Grete Waitz

Her husband, Jack Waitz was her coach during the year leading up to her first marathon. Here is his account of her training:

I’d never coached Grete for this kind of race. She never did high mileage; 80 miles a week, that was more or less what she did. ….At that time I was working as an accountant for a newspaper, and Grete was a schoolteacher. We lived in the suburbs of Oslo, and her routine was to run in the morning at 5 or 5:30, then she had to take a bus to the subway, and then another bus to get to the junior high where she taught. Then in the afternoon the same thing back. So it was pretty tough. But with any workout she did, she always ran fast. Knut trained with Grete and never wanted to run in the mornings with her, because she took off like that [he snaps his fingers]. She kept a good pace all the time.

In the afternoon, she often ran with one of her two brothers, Arild and Jan. According to Arild:

‘Jan and I had been running on the track, the 800 meters and the 1500 meters, for many years, but because of Grete we started to do races on the road—the 10-K, and half-marathons and marathons. Training with her was very systematic: Jack in the morning and Jan and me in the evening. In the afternoon we were running between 12 and 15 kilometers. But Jan and I took shifts. We couldn’t do what she was doing every day. We had to rest; her training was hard.’  

According to Jan: ‘She was very disciplined. She normally ran every kilometer around 3:50. We call that slow distance running, but it was pretty fast.’

In the forward to her book, Run Your First Marathon, co-authored with Gloria Avebuch, Grete wrote about her own first marathon in New York: ‘Make no mistake, I was able to run and run well because of my strong track background (and my will power) ‘. From her book, and from the comments of her husband and brothers, it emerges that early in her career, when she was focussed on 1500m and 3000m, that her training included a substantial number of high intensity sessions. Subsequently during her marathon career she did a substantial volume at or near lactate threshold, much of it near marathon pace. According to Johan Kaggestad, her coach later in her career, even her long runs (of more than 30Km) were never slower than 4 min/Km.

It is noteworthy that Johan Kaggestad also coached Norway’s other legendary female marathon champion, Ingrid Kristansen. Kristansen’s training was both in higher volume and somewhat more polarised than Grete’s. She ran twice daily, covering 160-200 Km per week. Her longest run during marathon preparation was a two-and-a-half-hour run covering about 36km at a pace of 4.10-4.20 min/Km. Kristiansen set a world record of 2:21:06 in London in 1985, more than 4 minutes faster than Grete’s time in London in 1983.


Paula Radcliffe

Like Grete Waitz, Paul Radcliffe came to the marathon during the final stages of a successful career on the track and cross country. She had won the world junior cross country championship in Boston in 1992, and the senior championship in Ostend in 2001. On the track she had won the European cup at 5000m on three occasions; the Commonwealth Games 5000m in 2002 and the 10,000m European Championship that same year. However, unlike Grete who had arrived in New York in 1978 knowing virtually nothing about the marathon, Paula had won half marathon championships in Veracruz in 2000 and in Bristol in 2001, and was better prepared for her first marathon in London in 2002. On that cool but pleasant April day in 2002 when Khalid Khannouchi had forged ahead of Paul Tergat and Haile Gebreselassie along the Embankment with little more than a mile to run, to take 4 seconds off his own world record in the men’s event, the women’s event belonged to Paula alone. She had broken clear of the leading group by 15Km and continued to exert her dominance with a series of sub- 5:10 miles in the second half. She crossed the finish line in 2:18:55, only 8 seconds outside Catherine Ndereba’s world record.

Later that year, in Chicago she ran away from the field to finish powerfully in 2:17:18, taking almost a minute and a half of Ndereba’s record.   Then the following April in London she ran the most phenomenal marathon ever run by a women to establish a world record of 2:15:25 which survives to this day. Despite the best efforts of Kenyans including Mary Keitany, and Ethiopians including Tiki Gelana, the only other woman to record a time within 3 minutes of Radcliffes’s world record is Liliya Shobukhova of Russia,  Shobhukhova has been suspended for two years on account of a blood profile suggesting blood doping. The terms of her suspension include annulment of performances since October 2009.  This would include her time of 2:18:20 recorded in Chicago in 2011.  See the footnote below for furter details .

Since 2003, Radcliffe has won the London marathon for a second time, the New York marathon on 4 occasions, and the world championship in Helsinki in 2005.   However despite these triumphs, her marathon career has been dogged by injury and misadventure. She failed to complete the 2004 Olympic marathon in Athens due to stomach upset, possibility caused by medication for a recent leg muscle injury; she finished 23rd in Beijing in 2008 after struggling to regain fitness following a stress fracture of her leg, and she was forced to withdraw from the GB team in advance of the 2012 Olympic marathon due to surgery for a foot injury – an injury that had been mis-diagnosed in 1994,and finally, 18 years later, was repaired .   Her most impressive performance in recent years was third place in the 2011 Berlin marathon in a time of 2:23:46.  She hopes to return to London next year to lay some of the demons to rest.


The training of Paula Radcliffe

Paula has always trained with determination, but since the early days of her track and cross country career she has been prone to injury. Following a disappointing 4th place in the 10,000m in the Sydney Olympics in 2000 she underwent a through biomechanical assessment by physiotherapist, Gerald Hartmann. Hartmann described this assessment in an interview with sports journalist, Frank Greally, published in Running Times in 2004.   Hartmann not only directed his attention to the prominent bobbing movement of Paula’s head which he attempted to alleviate by exercises to strengthen her shoulders and neck, but he also identified a lack of power in her legs.   He had asked Paula to do 20 hops up and down from a 16 inch high box as fast as she could. Whereas Kelly Holmes, 800m and 1500m gold medallist in Athens, had achieved 20 hops on and off the same box in 12.5 seconds, Paula took 27 seconds on her first attempt. This led Hartmann to devise a program of plyometric exercises and heavy weight sessions.

The fruit of this strengthening program were clearly apparent in the report on Paula’s physiological development reported by Andrew Jones in International Journal of Sports Science & Coaching in 2006. Paula’s vertical jump performance increased from 29cm recorded in in 1996 to 38cm in 2003.   Furthermore, her speed at VO2max increased from 20.5 Km/hour in 1992 to 23.5 Km/hour in 2003. Although the marathon is typically run at VO2/VO2max in the range 83-85%, the increased speed at VO2max would be expected to produce a similar relative increase at marathon pace.  Paula’s speed at lactate threshold increased from 14–15 Km/hour in 1992–1994 to 17.5–18.5 Km/hour in 2000–2003. Similarly, her speed at lactate turn point increased from 16 Km/hr in 1992 to 20 Km/hr in 2003. Paula’s average pace in London in 2003 was 18.6 Km/hr, consistent with the expectation that a well-trained marathoner can maintain a pace very near to lactate threshold.

It is likely that the strengthening and improved biomechanics achieved by Hartman’s program played a substantial part in the increased speed at VO2max and hence in Paula’s phenomenal marathon in London in 2003. However it should also be noted that the proportional increase in speed at lactate threshold from 1992-1994 to 2000-2003 was approximately 24% whereas the increase in speed at VO2max over this period was only around 14%.   This suggests that she had also increased her capacity to metabolise lactate. This possibility is confirmed by the fact that she exhibited a lactate concentration of only 4-6mM at maximum speed during treadmill tests whereas most athletes exhibit maximal concentration around 8-12 mM.

It is also noteworthy that Paula had already had an exceptionally high VO2max of 70 ml/min/Kg at age 19 in 1992, and this did not increase appreciably over the eleven years to 2003. Thus she was endowed with a high VO2max, but this did not increase with her training.

Apart from the strengthening program, what changes in her training occurred during these eleven years? First, she increased her total training volume greatly. At age 18 she did 20-30 miles per week but by 2003 she ran between 120 and 160 miles a week when in full marathon training. According to Andrew Jones, she would never compromise training quality for quantity. If tired she would cancel a session rather than perform at a lower level. She typically did the steady state continuous running that made up a large proportion of her training at 3:20–3:40 per km, only 5-25 seconds slower than her marathon pace.

In summary, Paula was endowed with a very high VO2max, which remained unchanged by training. The gains from the eleven years of training that turned her from a world junior cross country champion into the world’s fastest female marathoner were an increased speed at VO2max, perhaps attributable to the improved strength and biomechanics, and an even greater proportional increase in speed at lactate threshold, implying increased capacity to metabolise lactate in addition to her improved strength and biomechanics.


Both Paula and Grete were endowed with exceptionally high VO2max. At age 19, Paula had a measured VO2max of 70; Grete’s 3000m world record of 8:34 at age 21 also corresponds to a VO2max of 70. Both focussed on track and cross country racing in their early twenties. Once they turned to the marathon, both continued to run fast during training, Both did a substantial proportion of their training at a pace that might be described as sub-lactate threshold – though nearer to the threshold in Paula’a case. Paula also did a much larger volume of training, and achieved a personal best about 9 minutes faster than Grete, consistent with the evolution of women’s marathoning for which Grete had laid the foundations.

It is probable that for both women the substantial amount of sub-lactate threshold running helped develop their ability to metabolise lactate. In Paula’s case, the measurements reported by Andrew Jones provide direct evidence that this was the case. Thus, it might reasonably be argued that a large amount of running in the grey zone around lactate threshold, the zone that is avoided in a polarised program, played a substantial part in their success.   Their success is a challenge to the claims for polarized training.

However, great as these two athletes were, one is left with the feeling that they could have been even greater. In the five years from 1978 to 1983, Grete made the women’s world record her own, lowering it by almost 10 minutes. Yet in the year that the women’s marathon became an Olympic event, she was eclipsed by Joan Benoit.   Grete still remained near the top of the rankings for another five years, but she scarcely improved.   In light of the evidence that polarised training is the most effective way forward for an athlete who has achieved a plateau, I can’t help wondering if she might have gone on to even greater achievements if she had included a larger amount of low intensity running in her training schedule.

In Paula’s case the sense of frustrated hope is even more overt.   Not only did injury rob her of opportunities for Olympic gold, but it robbed her of the chance to demonstrate where that stellar trajectory of improvement that she exhibited in 2002-2003 might have taken her.   Would a more polarised approach to training have taken less toll on her body?   Might it have allowed her to reach an even higher level of performance? In light of the evidence that polarised training is the most effective way of improving VO2max once an athlete has reached a plateau, is it even possible that she might have been able to increase her VO2max beyond the level she achieved in her teens.   However, further increase in VO2max would be of limited value unless she maintained her extraordinary capacity to metabolise lactate, and perhaps that capacity was dependent on maintaining a large amount of sub-lactate threshold running in her schedule. Is it possible to minimise accumulation of acidity by other potentially less damaging means? That will be the topic of my next post.


Note regarding Liliya Shobukhova

In a comment below, Thomas points out that Liliya Shobukova has been suspended for a doping infringement, and suggests that I should not provide information about her marathon times. I am strongly opposed to drug abuse in sport. Furthermore, when providing information relevant to the training or physiology of runners I try to be as accurate as possible. As far as I can establish, the facts regarding Shobukhova’s suspension are:

1) She has been suspended on account of irregularities of her blood profile that suggest blood doping. Blood doping is the practice of boosting the number of red blood cells in the bloodstream in order to enhance athletic performance. Some methods, such as high altitude training, are legal; other methods such as blood transfusion are illegal. I do not know any details in Shobukhova’s case.

2) The suspension applies from 24th Jan 2013 to 23rd Jan 2015. The terms include the annulment of any performances dating from October 2009.

3) In August 2014 it was reported that she plans to appeal against the suspension.

4) Her profile, including her time of 2:18:20 in the 2011 Chicago marathon is still listed on the IAAF web-site (as of 15 Sept 2014).

If Shobukhova’s performance in the 2011 Chicago marathon was achieved with the aid of illegal blood doping, this would serve to emphasise the outstanding character of Paula Radcliffe’s world record time.

Addendum 7th Nov 2015: The evidence regarding drug abuse in athletics continues to emerge in an alarming manner and the situation has become increasingly muddy.  By now it is clear that Shobukova is not going to appeal her suspension.  She has publically acknowledged that there were apparently corrupt payments in connection with her case.  The fact that her performance in Chicago had remained on the IAAF website counts for little, as senior IAAF officials are themselves under suspicion of complicity in corruption related to drug abuse.  Paula Radcliffe herself has been the subject of speculation because of anomalous blood tests indicating unusual production of red blood cells.  There has been debate about the way she opposed publication of this information, but there are plausible innocent explanations for her blood test results.

In summary, Shobukova’s suspension is official confirmation that she employed illegal means to increase her levels of red blood cells.  On the other hand Radcliffe possibly achieved similar effects by legal means, such as training at altitude, and sleeping in a low oxygen environment.  With regard to the physiology of marathon performance, elevation of red blood cell levels appears to be advantageous, and is likely to improve performance by a small margin.

With regard to the main theme of my post, I do not believe that unusual production of red blood cells, by whatever means, was the main factor in Radcliffe’s extra-ordinary performances.   With regard to Shobukova, her performance should be regarded as invalid as she was convicted of using illegal means to gain advantage, and she has not appealed that conviction.  If this information had been clear at the time, I would not have mentioned her in my post.  Nonetheless, at this stage, I think that the least confusing way to describe a muddy situation is to add this addendum to my post.  If further information about Radcliffe’s unusual blood results emerges, I will add a further addendum.

Update May 2019: In fact shortly after my previous addendum, the IAAF concluded in November 2015 that the accusations against Paula were based on gross interpretation of incomplete data. The UK Anti Doping agency also concluded that there was no case to answer.  It is possible that the first questionable blood score was due to faulty equipment, while the third questionable score was a result of high altitude training in Kenya.

Does training induce long term muscle damage?

August 31, 2014

A glance down the list of single-age world records for the marathon reveals that a few names occur on more than one occasion. In those instances the same name always occurs within a span of a few years, apart from the instance of Ed Whitlock whose name appears 11 times, but even these 11 appearances are clustered within the past 15 years. This pattern demonstrates that it is difficult to remain at the top for more than a few years, and suggests that the stress of training and racing required to get to the top might produce damage that limits the tenure at the top .

This proposal becomes even more plausible in light of the evidence that heavy training can produce a long lasting over-training syndrome, and also the controversial evidence regarding the reversed-J shaped relationship between training load and health outcomes, such that moderate training load enhances health but very heavy training might damage health.


The fatigued athlete myopathic syndrome

One variant of the over-training syndrome is the fatigued athlete myopathic syndrome, in which it appears that muscles have a limited capacity to recover from exercise. Although the pathophysiology of this condition remains enigmatic, one thought-provoking feature is the abnormal shortening of DNA telomeres in muscle reported by Collins and colleagues from Capetown. Telomeres are DNA caps at the end of chromosomes. They become shorter following the cell division that occurs across the life-span in order to replace worn-out tissue,  This implies that there is a limit to the number of cell divisions that can occur during a lifetime. The shortening of telomeres is regarded as a marker of the aging process. In muscle, repairing the short term damage induced by training, especially the disruption of muscle fibres produced by the eccentric contraction at foot-strike, is dependent of the division of satellite cells, a type of stem cell unique to muscle. The observation by Collins of shorter satellite cell telomeres in biopsies from the vastus lateralis muscle in athletes suffering from the fatigued athlete myopathic syndrome, compared with healthy asymptomatic age- and mileage-matched control endurance athletes, suggests that the origin of the myopathic symptoms might indeed be attributable to damage produced by training and/or racing.

However, a key issue is the observation that age and mile-matched control athletes without symptoms of fatigue had suffered less shortening of their telomeres. Thus, it does appear that some athletes do suffer damage that limits their running career, but this is not a universal consequence. Kadi and Posnet report that when satellite cells are heavily recruited to regenerate skeletal muscle in athletes, telomere length is either dramatically shortened or maintained, possibly even longer than in non-trained individuals.   What are the factors determining whether or not training results in abnormally shortened telomeres? The answer is unknown.


What about the telomeres of elite masters marathoners?

It is unlikely that the seven elite masters marathoners discussed in my two previous blog posts suffered excessive shortening of their telomeres, and even plausible that they maintained longer telomeres that the average non-trained individual. This is mere speculation, but the existence of a mechanism by which some, but not all, athletes suffer sustained muscle damage adds plausibility to the proposal that these seven athletes suffered less sustained training-induced muscle damage than the average athlete, contributing not only to their phenomenal marathon performances but also to their impressive 1500m times.

If so, was their resistance to sustained damage due to their natural predisposition to longevity or might it be attributed to their training schedules? As the seven followed a variety of different training schedules, it is unlikely that the type of training schedule was a major influence. Nonetheless, in my previous post, I discussed the evidence that the two who followed the most markedly polarised programs with a large amount of easy paced running and a small proportion of fast running (Ӧstbye and Whitlock), exhibited greater longevity at the top of the world rankings than the two who appear to have included a greater amount of training at tempo pace or faster (Turnbull and Hosaka).   But it should be noted that Hosaka is still only 65 and might yet upset this observation.



Overall, it is plausible that long term muscle damage induced by training does limit the running careers of some but not all athletes. It is likely that a natural predisposition to longevity helps protect against the damage. However, the training of elite masters marathoners provides a thought-provoking hint that a polarised program that minimises the stress associated with a large training volume might be the most effective way to train in order to achieve longevity of one’s maximal level of performance.

Lessons from enduring masters marathoners

August 30, 2014

In my post on August 26th I summarised the running careers and training of the seven elite marathoners who had set world masters marathon records at age 60 or greater and had remained high in the masters marathon rankings for over a decade. These seven enduring elite marathoners, John Gilmour, Eric Ӧstbye, John Keston, Derek Turnbull, Luciano Acquarone, Ed Whitlock and Yoshhisa Hosaka, shared several features. Apart from John Keston who had only began running at age 55, all had shown signs of athletic talent in their youth, but none were of international class at that stage. The primary feature that made them world champions was a reduced rate of decline in early middle age. Furthermore, apart from Erik Ӧstbye who rarely ran on the track, all maintained their speed, recording international level times for 1500m during middle age.  

However, while all seven trained consistently and raced with determination, they employed a range of quite different types of training, suggesting that their enduring success owed more to their natural predisposition to longevity than to the type of training they did. At first sight, it might seem that there is little that a less gifted runner might learn about how to train from these exceptional individuals.  

I do not think that is the case. On closer inspection, examination of the differences in the training of these elite athletes in the light of what fifty years of research has taught us about the physiology of training provides potentially useful pointers towards the most effective training strategies for minimising decline in performance during middle and old age.

The crucial limit imposed by aerobic capacity.

If we are to run our fastest possible marathon we need to train many different physiological capacities. These include the ability to conserve glucose so that our supply of easily accessible fuel does not run our before the end of the race; the ability to metabolise lactate so that we can maintain a metabolic rate at a little above 80% of our aerobic capacity (VO2max) for several hours without continual accumulation of acid in our blood; the ability of leg muscles to stand the pounding produced by more than 30,000 foot-strikes; we need to maximise our VO2max; and several other attributes such as an efficient running style. Maximising VO2max in itself requires maximization of several trainable physiological attributes, including aerobic enzymes in mitochondria; capillary blood supply to muscles; and cardiac stroke volume.   Among all these trainable capacities, VO2max is especially noteworthy because it plays a crucial limiting role.

One of the important contributions of the great coach, Jack Daniels, to running science, is the realization to there is a quality, VDOT, which is a measure of aerobic capacity and corresponds approximately to VO2max, that provides a fairly accurate prediction of a runner’s best performance at any distance in the range 1500m to marathon. The rationale underlying the prediction is based on the assumption that aerobic capacity, quantified by the value of VDOT, is the limiting factor that determines performance across the range of distances.   This assumption is controversial. An alternative view is that performance is limited by a ‘central governor’ in the brain that acts to protect us from harm. In fact the apparently competing claims for the role of aerobic capacity and the central governor as the factor limiting performance can be reconciled, but that is a story for another blog post. In practice, predictions based on VDOT are usually fairly accurate for well-trained runners. If a runner is adequately trained at 1500m, performance at 1500m provides good estimate of aerobic capacity as quantified by VDOT, and in turn this value of VDOT can be used to predict marathon performance, provided the runner is adequately trained for the marathon.

The implication of this is that any consistent training program can maximise all the other factors such as ability to conserve glucose; ability to metabolise lactate; and the ability of leg muscles to stand the repeated pounding, so that the limit is set by aerobic capacity. This does not mean that these other factor do not play an essential role in performance. It simply indicates that these factors can be trained to the level where they are no longer the limiting factor.   Aerobic capacity itself can also be trained but it remains the limiting factor.

Was aerobic capacity the limiting factor for the elite masters marathoners?

What relevance does the limiting role of aerobic capacity have to understanding the phenomenal success of the seven elite masters marathoners? The first question to ask is does Daniel’s VDOT formula work for them? Because it is reasonable to assume that they were adequately trained for the marathon distance, it is appropriate to ask how well their marathon performance predicted their performance at 1500m.

Marathon times and 1500m times recorded on occasions no more than approximately a year apart are available for all of the seven of elite elderly marathoners apart from Erik Ӧstbye, who rarely raced on the track. Table 1 lists age, marathon time, estimated VDOT based on marathon time, 15000m time predicted from VDOT; actual 1500m time recorded on a race; and percentage error of the estimate. It can be seen that the percentage error is typically around 1-2% and never exceeds 2.5%. Granted that race performance on any one occasion might differ from maximal by a few percent, it appears that for these six athletes, their performance across distances from 1500m to marathon was determined largely by their aerobic capacity. The conclusion receives further support for the observation that in the case of Ed Whitlock for whom we have a measured value of VO2max less than a year before the marathon I have used to estimate his VDOT, the estimated VDOT of 52.9 ml/min/Kg is within a tenth of a percentage point the measured VO2max of 52.8 ml/min/Kg.

Comparison of actual 1500m times with predicted times based on VDOT estimated from marathon time

Comparison of actual 1500m times with predicted times based on VDOT estimated from marathon time

The data suggests that other factors might have contributed a percentage point or two to the performances of the six elite athletes. While a percentage point or two might matter in a race, it is merely the icing on the cake when it comes to understanding why these elite athletes stood head and shoulders above average athletes of similar age.  Thus the crucial question in understanding what made these athletes elite is the question of what endowed them with values of VO2max about twice the values expected for the average man of similar age.

What determines aerobic capacity?

Aerobic capacity is trainable, but it is also shaped by genetic factors.   As I previously discussed (2nd August) in the case of Ed Whitlock, I think it is likely that his very high maximum heart rate contributed appreciably to his extraordinary VO2max. Maximum heart rate appears to be largely determined by genetic factors. Those of us with lesser genetic endowment have little hope of matching Ed’s performances. However, if we accept that consistent training can optimise the factors other than VO2max such that these other factors are no longer limiting, if we wish to maximise our distance running performance in middle age and sustain that level of performance (relative to WAVA norms) into old age, the major question is how can we maximise our aerobic capacity and how can we sustain a maximal value over a prolonged period?

To address these questions we can draw on several different strands of evidence. Two of the most important strands of evidence are the body of evidence regarding the nature of overtraining and the evidence regarding polarised training.


I have reviewed the evidence regarding over-training including the role of cortisol on several occasions previously (e.g. here and here). In summary, effective training achieves it benefit by stressing the body in a way that elicits an anabolic state that strengthens the body so that it can withstand similar stress more effectively in future.  However if there is inadequate recovery, there is sustained elevation of the stress hormone, cortisol, that obstructs the anabolic phase, while also creating a risk of chronic inflammation which promotes the replacement of healthy tissue by fibrous tissue.

Polarized training

The evidence regarding the relative merits of high volume training compared with high intensity training (reviewed in my post of 31st March) indicates that high volume and high intensity training are each effective in increasing aerobic capacity. However, accumulating evidence indicates that polarised training, consisting of a large volume of easy running and a small amount of high intensity running, produces the greatest increases in aerobic capacity in athletes who have already achieved a plateau of fitness, and offers the best prospect of long term improvement, year upon year.

A closer look at the training of the elite elderly marathoners

Bearing in mind the necessity for avoiding over-training via adequate recovery, and the evidence indicating that polarised training offers the best prospect of year on year improvement, it is worth a closer look at the training of the seven elite ancient marathoners.

It is noteworthy that two, Ed Whitlock and Erik Ӧstbye, adopted a markedly polarised program consisting of a very large amount of easy running together with a quite small amount of intense running, mainly in the form of shorter distance races, over a sustained period of years. It is also important to note that Whitlock emphasizes the importance of building up the volume of his training very gradually, and on minimising impact stress on his legs during his long runs. Ӧstbye set world marathon records in the age bands 40-50; 50-55; 55-60 and 60-65, and remained high in the rankings until age 70. Whitlock has dominated the world masters marathon rankings for the past 15 years, and holds the current single-age world records for 11 of those 15 years.

In contrast, Yoshihisa Hosaka has adopted a program that includes a higher proportion of running at marathon pace or faster. During his twice daily interval sessions he runs approximately 10 Km per day at marathon pace or faster. As might be expected, his quite demanding program has led to very impressive performances up to age 60. As discussed in more detail in my post of 18th August, over the period from age 45 to 60 his performances declined at about half the rate expected during that age period. However, since age 60 his performances have declined at a rate about 70% faster than expected on the basis of WAVA predictions (as illustrated in figure 1). It would be unwise to draw any definite conclusions this stage. It will be very interesting to see how he fares in his ambitious quest to capture Derek Turnbull’s M65-70 record in the Waterfront Marathon in Toronto this November. However, as reported in his interview with Brett Larner at the time of last year’s Waterfront Marathon, he is finding it increasingly hard to manage his demanding training and might even consider changing to switch to a program more like that of Ed Whitlock.

Figure 1: The decline in marathon performance of Whitlock, Turnbull and Hosaka.  Apart from a minor ‘stutter’ at age 70, Whitlock did not exhibit marked decline until age 80; Turnbull exhibited a similarly marked decline in his late 60’s ; Hosaka shows a trend towards an even earlier decline. The data point at age 64 represents his time in the 2013 Gold Coast marathon.

Figure 1: The decline in marathon performance of Whitlock, Turnbull and Hosaka. Apart from a minor ‘stutter’ at age 70, Whitlock did not exhibit marked decline until age near to 80; Turnbull exhibited a marked decline in his late 60’s ; Hosaka shows a trend towards an even earlier decline. The data point at age 64 represents his time in the 2013 Gold Coast marathon.

John Keston trained with fierce determination and great success from his later 50’s until age 70. In 1994, it appeared likely he would become he first 70 year old to run a marathon in less than three hours. However, despite running 20 marathons in the year, he failed to achieve his target. He almost achieved the three hour mark a year later, but then sadly suffered a series of accidents and injuries. After his return he adopted a training program based on about 2 hours of running every third day with a similar period spent walking (often in woodland) on each of the intervening days. Though he never quite recovered his former dominance, he nonetheless continued to perform at very high level for another decade. At age 82 his marathon time was 6th in the all-time rankings for the M80-85 category at that time.    

It is more difficult to draw conclusions concerning Luciano Acquarone, Derek Turnbull and John Glimour due to lack of detailed information about their training. John Gilmour experienced more than a life-time’s worth of stress during his three years as a prisoner of war, and as far as I can gather, did quite a lot of demanding tempo running during his training. His training can scarcely be described as polarised. He remained near the top of the world rankings over a 12 year period from age 59 to 71, and continued to run for a further two decades, even in his 90’s still running 5Km per day. He certainly qualifies as an enduring elite marathoner but his time at the top of the rankings scarcely matches that of Ed Whitlock who shows no sign of relinquishing that position after 15 years.

Derek Turnbull did not follow a planned training program, but his spontaneously selected paces were often quite fast. In his obituary, Roger Robinson described Turnbull’s training as a perfect balance of long runs, tempo, and fast work. Almost certainly this ‘balance’ included substantially more tempo training than a polarised program. Of the seven elite marathoners we have discussed, Turnbull had the shortest time at the top of the rankings, a period extending from his world record for the M60-65 category set at age 60 to his time 3:15 at 70, which at the time was 13th in the rankings. As shown in figure 1, the decline in his performance in his late 60’s was marked.


Thus, despite the fact that all seven of these elite elderly marathoners were superb runners, when one looks in detail at the differences in their training programs, the available evidence indicates that those who adopted a more polarised training regimen, with a large amount of easy running and a small proportion of intense running, achieved greater longevity at the top.  

In a small sample of exceptional individuals it is of course possible that other differences, possibly directly associated with their natural predisposition to longevity, accounted for the observed longer duration at the top of the rankings. However, one fragment of evidence provides intriguing support for the claim that the training regimen made a crucial the difference. Ed Whitlock’s progress did stutter slightly at age 70. His times of 2:51:02 and 2:52: 50 recorded in the Columbus Ohio marathon at age 68 and 69 respectively made it appear highly probable that he would become the first to break 3 hours at age 70. He did not manage it that year. Nonetheless, despite some problems with arthritis, he continued to build up the frequency of daily 3 hour runs in the following two years. At age 73, after doing 67 three hour training runs in the preceding 20 weeks, he ran the Toronto Waterfront Marathon in 2:54:49, surely the most impressive marathon time ever recorded by a masters athlete, and one might argue, no less impressive than the 2:03:23 recorded by the current world open record holder, Wilson Kipsang.

There is another point that should be re-emphasized. Although the striking feature of Ed’s training was the daily 3 hour runs, he did also do some fartlek style speed work, and raced frequently over shorter distances. I suspect that this small but significant amount of faster running helped maintain his world class performances over 1500m and in particular ensured that 1500m time in at age 73 showed only a 1 second decrement from performance at age 72, whereas WAVA would have predicted a 5 second decrement over the year. As shown in table 1, for all six of the athletes for whom 1500m times are available, the comparison of actual 1500m time with the prediction based on VDOT calculated from their marathon times demonstrates that all maintained their speed over the shorter distances. All six did at least a small amount of faster running during training. Thus, it is likely that at least a small amount of intense running was an important component of their training.

It should also be emphasized that drawing general conclusions from the experiences of a small number of exceptional individuals is fraught with danger. If Hosaka can arrest his recent decline while continuing the twice daily interval training he did in his early 60’s, it will be necessary to re-appraise my current conclusions. But the evidence so far does confirm what would be expected on the basis of the studies of training physiology performed in samples more representative of typical runners, over the past half century. The key lesson is that a large amount of easy running together with a small amount of faster running is the best strategy for sustained optimal performance. Furthermore, avoidance of cumulative stress is essential. Ed Whitlock’s example suggests that this is best achieved by very gradual build-up of the training volume.

In my post of August 26th I mentioned Tim Noakes’ hypothesis that an athlete can only expect to remain at his/her peak (relative to the WAVA age norms) for a few years on account of the damaging effect of the training and racing required to attain one’s peak . Although the seven individuals we have been considering are clearly exceptions to any such rule, they do offer interesting insights into the limits to the validity of the hypothesis. In my next post I will examine the evidence for and against that hypothesis.