Welcome to the Science of Sport, where we bring you the second, third, and fourth level of analysis you will not find anywhere else.

Be it doping in sport, hot topics like Caster Semenya or Oscar Pistorius, or the dehydration myth, we try to translate the science behind sports and sports performance.

Consider a donation if you like what you see here!

Did you know?
We published The Runner's Body in May 2009. With an average 4.4/5 stars on Amazon.com, it has been receiving positive reviews from runners and non-runners alike.

Available for the Kindle and also in the traditional paper back. It will make a great gift for the runners you know, and helps support our work here on The Science of Sport.

Sunday, June 24, 2007

Drugs work - but by how much? A look at doping and performance improvements

Performance enhancing substances - do they work? And by how much? It's a relevant question, and I thought it would be good to have a look at the answers. I was going to write an article on the Ethiopian dominance of long distance running in the last ten years, but we've been receiving such thoughtful comments and questions from people regarding our posts on drug use in cycling that I thought it might be better to continue on this theme and look at just what a scientist makes of the prevalence of doping in sport. I also said I'd look at altitude training, and I will, but I'll do that tomorrow. For now, we keep going on doping.

Yesterday LuckyLab asked how much of a difference doping actually made? And unfortunately, it's quite difficult to get hold of that information! As you can imagine, it's difficult enough to find out who is doping and with what, never mind trying to track their improvements in a controlled manner! And make no mistake - this information exists, because scientists are giving out the drugs, and as scientists, they are compelled to measure changes and improvements (it's our curse as scientists!). But this information rarely makes it out into the public eye, except once....

As far as I'm aware, the only paper published in a reputable journal is one by Franke and Berendonk. This paper was published in a journal called Clinical Chemistry in 1997, and this post is a breakdown of what it showed...

To begin with, the paper by Franke and Berendonk opens with the following sentences:

"One of the largest pharmacological experiments in history has been running for more than three decades, namely, the administration of drugs to athletes to enhance performance in many different kinds of sports"

Another quote from the article's abstract (or introduction summary):

Top-secret doctoral theses, scientific reports, progress reports of grants, proceedings from symposia of experts, and reports of physicians and scientists who served as unofficial collaborators for the Ministry for State Security (“Stasi”) reveal that from 1966 on, hundreds of physicians and scientists, including top-ranking professors, performed doping research and administered prescription drugs as well as unapproved experimental drug preparations. Several thousand athletes were treated with androgens every year, including minors of each sex. Special emphasis was placed on administering androgens to women and adolescent girls because this practice proved to be particularly effective for sports performance.

Scary stuff. I won't go into massive detail in this post on the background of the study. But very briefly, Werner Franke (the author of the research paper) is one of the world authorities on doping in sport. He has a personal interest, because his wife, Brigette Berendonk, was a German Olympic shot put and discus thrower, and was on the programmes he writes about.

And their paper is all about the hormonal doping and androgenization (which means making males out of females) of East German athletes. This is still topical, because the case of the current Tour de France champion, involves a similar product - testosterone. In this research article, they write about a number of anabolic agents, notably Dianabol and Turinabol.

But let's get stuck into it. And we'll use the example they give of a German Female shot-put athlete. This athlete had been training for 14 years before she was put onto the programme, so certainly, her improvement in results cannot be explained by a response to training, as it might be if she was still a junior athlete.

So the Figure below shows the improvement in performance in this woman shot putter during a 5 month period in 1968. The blue block on the X-axis shows when she was given the drug - Turinabol, in doses of 14 tablets per week (10 mg per day, each tablet = 5mg). You'll see how she gains muscle strength and improves her performances dramatically in only a few weeks - we're talking 9% over the course of about 10 weeks of drug use! But it gets better (or worse, depending on your point of view!)

Jump ahead to 1969, where the same woman went through very much the same procedure. The Figure below shows the performances over a similar five month period. This time, the drug was administered in slightly higher doses, and in three separate cycles. The result is the same. Her performance improved by an amazing 17% from the start to the end of the season, with a new world record of 20.10 m the culmination.

There was more to come. In 1970, the same athlete was again systematically doped, but with even higher doses. She once again broke the world record, and this time, broke the indoor world record as well. She ended with a personal best of 20.22 m, which was a fully 2m further than she had managed to achieve on 14 years of training BEFORE starting on the doping programme! This is an 11% improvement! But what is particularly important is to notice how her performances when NOT ON THE DRUG are so 'poor' - she starts each season a full 3 m off what she will reach at the end, with the aid of the steroid.

The paper by Franke further breaks down the recorded improvements in performance when using anabolic agents. The table below shows some of these improvements:

Not exactly and all-encompassing and comprehensive list, but it illustrates two points. Firstly, we don't know enough about the effects of these drugs on performance. That's pretty obvious. But secondly, what we do know tells us that they work really well - 7 seconds on a 1500 m time is enormous. So is 4 seconds on a 400 m. Even 2 seconds in 400 m is a lifetime. So if you look at the current world record in the 400 m event, it stands at 47.60 (Marita Koch). No one has managed to even threaten it since the 1980's, and it's quite conceivable, assuming the table above is correct, that this 47.60 is worth about, oh, 50.60. And suddenly then, today's runners are comparable!

So, we can pretty much see that doping with anabolics does work. If only there was similar evidence of the effect of EPO. There is some information on EPO, however, and it comes from Marco Pantani, and is reported in a book by Matt Rendell called The Death of Marco Pantani and then reported in an article by Tim Noakes published in the International Journal of Sports Science and Coaching in 2006.

In this book, Rendell exposes results of rider's hematocrit levels during the course of three years - 1993 to 1995. Remember that the hematocrit levels are an indication (albeit indirect) of the use of blood doping or EPO. A normal individual would have a hematocrit between about 40 and 43%, while a trained athlete could see this get up to about 48%, possibly even 50% if that person has done altitude training. However, it is also important to remember that when a cyclist is training hard, his hematocrit levels will fall. Therefore, in the peak of competition, it would not be unusual to see relatively lower hematocrit levels, in even the elite athlete population.

So what does the data say? The graph below is a rough summary of the information from Rendell's book. It shows the hematocrit of riders who were part of a state run programme by the University of Ferrara in Italy. I have highlighted in red the periods where cyclists would be competing in the Tour de France or Vuelta. The bars in green are the "off season" where cyclists are not racing and blue are mid-season. Remember, it would be normal to see that the hematocrit is highest during the rest period, and lowest during the racing season. Yet it is quite obvious from this graph that it is the other way around, and that is something that would presumably only be possible if the athlete was using a product such as EPO.

I have also shown on the graph the values of one Marco Pantani (yellow triangles), and you can see how in 1995, his hematocrit reached an astonishing 64%! This kind of variation just cannot be explained by anything other than pharmaceutical use, in my opinion.

In his book, Rendell concludes that "There is incontrovertible evidence that Marco’s entire career was based on r-EPO abuse, which was both effective and, until 2001, undetectable by tests used in professional cycling. Is it reasonable to suppose that the most successful period of his career, from 1998 until 5 June 1999, depended on anything else?

THe bottom line then, is that doping seems to work VERY WELL. It's still difficult, if not impossible, to put an exact number to the benefit, though the data of Franke et al give a pretty good indication that it's at least 15% in those power based sports. It may be slightly less for endurance based sports, like cycling. But as i wrote yesterday, there's substantial evidence that hormone levels, particularly testosterone fall during the course of a race like the Tour. And so if Floyd Landis and others are using the drug, the benefit would come from defending this drop-off, which promotes recovery and hence allows them to maintain their form throughout the Tour. In theory then, the systematic use of drugs will have a large effect in cycling, not because it acutely boosts performance, but because it allows it to be maintained. Think of your own training - you always have good days and bad days. But what if I said that by using a drug, like testosterone, you could drastically reduce the number of bad days - that is what these drugs will do for cyclists.

So that's it - a heavy, quite technical post. Well done for making it this far! As usual, I welcome any feedback, it always stimulates me for future posts! That's it for now as far as doping goes, though I'm sure this is one I'll return to in the future. Next time, I'll hopefully get round to that altitude training article, and the one on the Ethiopian athletes!

See you soon!
Ross T


Andrew said...

Ross, I'm learning so much here. I can't remember if I've asked this before, so if I have, excuse me. What are the effects of the legal supplements, the one you can buy in store for example? I've heard that many of these, while legal, actually contain banned substances. Is this true?

Ross Tucker and Jonathan Dugas said...

Hey Andrew

You did ask, and I haven't forgotten - it's a post that I would like to do very much. I'll probably only get around to doing it next weekend, but to give you the short version, yes, there is quite a lot of evidence that these products do get contaminated. So what the ingredient label says is often not what is inside!

There are a few reasons for this - usually, it's because the same company manufacturers and range of products and there's simple contamination between products.

But I'll put a comprehensive post on this in the coming week. There is so much stuff to write - yesterday, I planned two articles one on drugs and one on Ethiopian athletes, and I've ended up doing 5 on drugs, with more planned!

I really wish I could quit my job and make a living off doing this kind of writing! If anyone out there knows how I could do this, let me know!

But thanks for your interest and I'm pleased you are enjoying it and learning! That article is on the way!


Andrew said...

Hehe, thanks Ross. I'm in no rush at all, you have no obligation towards me, I'm just enjoying reading everything on a rainy sunday afternoon. So get to it when you have time.
As for doing it as a job, just keep plugging away, you never know what opportunities might open up. You're onto something here, that's for sure.

LuckyLab said...

The really hard thing about the East German science is there is no true control (at least not as listed). I would expect an athlete to have improvements in performance over the course of the season so the real question is: Is that really a 17% improvement? Where would she have finished the season without dope? The difference there is the actual performance increase.

As you note, the influence of dope on short duration, explosive-type events is much greater than endurance events. I'm still longing for hard numbers but as you know scientific studies of the effects of doping are what we would consider a little unethical.

Thanks for the sources.

pensum said...

Lucky Lab has put his thumb on my same concern. However in this particular case it would appear simple to provide at least a circumstantial control as they should have the performance records of the female shot putter for the 14 years she trained and competed before doping. from this data one could see what her "normal" improvement over a season would be.
Nonetheless Mario's hemocrit levels seem to be both revealing and rather shocking.
thanks for all the great insight, this whole Landis affair has brought to light a lot of interesting facts, and it's great to get some rational explanations.

Anonymous said...

Oi, achei teu blog pelo google tá bem interessante gostei desse post. Quando der dá uma passada pelo meu blog, é sobre camisetas personalizadas, mostra passo a passo como criar uma camiseta personalizada bem maneira. Até mais.

Ross Tucker and Jonathan Dugas said...

Hi LuckyLab and Pensum

I hear your points, and you're right, but the 17% reflected is the IMPROVEMENT on the athelte's Personal Best from her previous 14 seasons. That's why they chose to use that particular athlete - 14 years of training without drugs had given her a PB of around 18 m - that's the performance you see on the graph from 1968, actually.

So after 14 years without drugs getting to 18 m, one year with drugs, and she was hitting 20 m! That's the difference made by drugs, and I'm quite confident that they would control for natural improvements. In fact, that's why Franke chose this athlete.

I'm not sure where the table comes from - the improvements by event are listed in the Franke paper, but unsubstantiated by the figures or records. There may be major ethical issues at play here though (apart from the obvious drug use!) because it's quite simple to work out who these athletes are, given their performances.

I also wish we could do a study on the performance effects of an EPO or blood doping trial. I have no doubt that this data exists, perhaps some day it will surface. But even a 5% improvement (which I don't think is unreasonable) is massive - that's the difference between winning on Alpe d'Huez and finishing about 90 seconds down.

Thanks again!!!


But thanks for your comments,

Anonymous said...


I am a high school student doing a science project on blood doping and other performance enhancing techniques in sport - please can i use some of your graphs in my project if i credit your blog as the source?

Ross Tucker and Jonathan Dugas said...

Hi Anonymous, and thanks for visiting our blog.

We do not have any problems with you using the figures on the blog. In fact, much as any image on the internet, you can download these images.

How you credit these images is between you and your instructor(s), but thank you for asking us in advance.

Good luck with your project, and we would really be interested in seeing the finished product. You can email it to:


Kind Regards,

Francois said...

I have a question of clarification about one of the statements you make in this very interesting post.

You say when discussing the Pantani chart: "Remember, it would be normal to see that the hematocrit is highest during the rest period, and lowest during the racing season."

Why is that?

Would elite athletes not want to aim for a higher hematocrit rate as they get closer to their major competitions? Is it not why a lot of them train in altitude or use hyperbar chamber?

Moreover, it seems that it is easier to keep a high hematocrit rate for cyclists and swimmers than it is for runners as high-impact sports are detrimetal for "richer" red blood cells".

This would explain why Ironman elite athletes perform their most difficult running week not 6 to 8 weeks before their Ironman race before rather 3 to 5 months prior to their big competitions.

Thanks again for this interesting article.


Ross Tucker and Jonathan Dugas said...

Hey Francois

Thanks a lot for the visit and the email, great to hear from you!

To respond to your question:

The reason one would actually expect the hematocrit to be low during the racing season is that when the cyclists are racing every day, their bodies are in what is referred to as a 'catabolic' state. This means that they are in a state of breakdown, not building up. So the stresses of 5 or 6 hours a day of hard racing would cause the hematocrit to drop.

There are studies showing this - the levels of certain hormones, like cortisol, are increased, while others, like testosterone, are decreased. Cortisol is a hormone that is catabolic, whereas testosterone is anabolic. So for the body to be producing red blood cells, you need to be in that anabolic state - this happens when your energy balance is normal, the stress response is minimal, and so for a Tour cyclist in peak racing to prevent the drop in hematocrit is very unusual indeed.

Of course, they WANT to have a high hematocrit, because that's where they derive the performance benefit. The problem is, physiologically, the stress of racing is driving it the other way - down! So when you look at a rider's hematocrit levels, and you see normal levels in the peak of the season, that should cause suspicion right away. To then see the HIGHEST levels during the racing season, when the stresses are the highest - that's a sure sign of a doping case!

I hope that clarifies it. You are quite correct in suggesting that the cyclists would AIM for higher hematocrit, but actually achieving this is not possible without doping, because everything in the physiology is pushing it the other way.

Thanks for the question!


Francois said...

Thank you very much for your excellent answer. Just one question of clarification. Are you saying that it is not possible for an elite rider to have a high hematocrit (let say 45 to 50)at its peak?

For example, I read somewhere that Armstrong's peak was the FIRST day of the Tour de France. He was not riding a lot in the weeks and months prior to the Tour such that he would be fully rested on Day 1.

That would imply that he had a relatively long tapering period. And this would also imply that his body was in an "anabolic" state (rather than in a "catabolic" state, as your analysis would imply). (or maybe I just did not well understand your comment properly).

However, if my Armstrong example is a good representation of what really happens at the elite level, his hematocrit should have been relatively high at day 1 and (normally!) gradually decline during the Tour.

So my point is really: Does it make sense to conceive that you can be at your peak AND have a rate of hematocrit that is higher that the one during the rest season?

Some people tell me that all you need is a well-conceived training program and an appropriate tapering (supplemented by high altitude training, if need be).

Thanks again. As usual, I am learning so much reading your material.


Ross Tucker and Jonathan Dugas said...

Hi again Francois

No problem, thanks for the reply.

My opinion on the hematocrit issue, borne out by some evidence, is that when the athlete is in the serious racing phase of their season, it would be very unusual to have a hematocrit higher than 45%. Remember, "normal" is between 42 and 45, 45 to 50 is very high and often an indication of a doping case (here's where things get tricky - altitude training, genetic differences, and you can quite easily get up to the mid-40's, even 50.

Now if you take Armstrong, he's aiming to hit the Tour in peak condition - this means a hematocrit as high as possible. Quite what that means is anyone's guess - let's assume a hematocrit of 45%. Problem is, he's still training incredibly hard in the weeks leading up to the Tour. He always rode in either the Dauphine or the Tour of Switzerland, which ends about 3 weeks before the Tour. Those races, a week long, are highly demanding, even though he's riding them below maximal levels.

So even though he's aiming for peak, he's still riding sufficiently long and hard to be in that position, where a high hematocrit is very difficult to attain "naturally". The same goes for everyone else, incidentally. Then, over the course of the race, it would come down even more.

So what you typically see in a profile of hematocrit over a year is about a 4 or 5% swing - a peak of maybe 45%, which usually happens right at the start of the season in March/April, followed by a decline during the season, regardless of how hard they are training (or easy, as the case may be - because they're still racing in between), and then a pick up after the recovery period at the end of the year, building up to that peak. So you have this natural cycle.

Then, if you introduce some doping into the picture, you can alter that pattern completely, and the next thing you know, you have an 'upside down' curve like Pantani, where the highest values are in the racing season! That's only possible through artificial means.

It's been said that the most your hematocrit should vary in a given period is about 4 to 5%, but the professional cyclists are swinging it by between 10 and 15% - a low of maybe 40% and a peak of 55%. Obviously, we now have tighter controls, and so they probably hit 49.9% using blood doping (which is not detectable) and then ride like that.

Just one last thing on the altitude - training at altitude in the final month before the Tour would be a bad idea, because you can't train as hard at altitude. So even though you gain in terms of oxygen carrying capacity, you lose out in other areas. For that reason, the altitude training stint should either end at least a month before the Tour, or you should do your taper at altitude when you don't need to train especially hard.


Ross Tucker and Jonathan Dugas said...

Hi Francois, and thanks for the great questions. Please keep them coming!

I am going to chime in here, and add that even with altitude training, hypoxic tents, and other "legal" and "natural" ways to elevate your Hct, we must always remember that physiological response to these stimuli is not infinite. In other words, if you spend six months at altitude, you Hct goes up a bit, but if you spend 12 months or even 18 months the effect is not two or three times as large.

As is the nature of a homeostatic system, the body will maintain the Hct with a desirable range. So going to altitude will increase it, yes, as you respond to the stimulus of hypoxia.

However, rarely does the Hct rise above 50% even in athletes who live at altitude. The body will enhance its ability to move oxygen around by increasing the Hct, but only up to a certain point, for the additional strain placed on the heart with increasing Hct values is relevant.

Therefore the body will increase Hct, but only up to a point that all of its systems can remain happy with it. Again, it will not increase its Hct indefinitely so that you have a Hct of 60%. the body will regulate the response so that it does not get that high.

Finally, we must not forget that the measure of Hct is one that is very easily manipulated. In fact, there is honestly no excuse for getting caught breaking this rule, as all the athlete must do is rapidly infuse saline so that the ratio of cells to fluid decreases, therefore lowering the Hct. So it is relatively meaningless when we see a cyclist report with an Hct value of 49.9%. . .and it is entirely likely that they have simply manipulated their value so that it is at the highest legal limit.

Thanks again for the questions and comments!

Kind Regards,

Francois said...

Thanks for those enlightening comments! I need to think a bit before replying.

In the meantime, I am curious, would you have a chart showing the annual typical profile of hematocrit for an competitive athlete? (i.e. high at the beginning at the season, then declining and then picking up again during the recovery period)? All the ones I have found so far are cross section (hematocrit taken at a given time in a given year, spread over a number of years). I would rather have one that shows your annual profile for a typical year.

I am searching the net in scholar articles but can't find one...

Thanks again!


Francois said...

Actually, I found this document which contains interesting charts.


Francois said...

In the link above, they claim in the explanatory notes that "normal individual variation in hematocrit and haemoglobin is 15% at the most (e.g. 0.42-0.48 in hematocrit)".

This would be significantly higher than what you would consider as "normal', i.e, 5%.

(don't forget that I was just trying to find an annual profile of hematocrit (within a typical year)! I am not trying to sidetrack the discussion!!)


Ross Tucker and Jonathan Dugas said...

Hi Francois

Just to answer that last question - I'll get to the earlier one tomorrow morning.

But with regards to the natural variation - we're both right.

What the link you found says is that a change from 42% to 48% is a change of 15%. This is the classic example that often catches people in finance, because hematocrit is already a percentage!!!

Therefore, if you say that 42% to 48% is already a percentage, then my position would be that it's a 6% change! However, what the link you found says is 15%, because 6/42 is about 15%. This is also true, it says the same thing, just a little differently, if that makes sense?

So to clarify then, what I am saying is that a normal change would be something from 42% to 48%, but what you see with pro cyclists is a range from 42 to the mid-50's. According to the link's method, a change from 42 to 55% would be equal to about 31%, whereas I would say it's only 13%.

Does that makes sense?

Will answer the other one tomorrow!


Francois said...

I understand your point very well because even if my community (I'm an economist), people make that mistake very often!

Just to be cristal clear though, I would express a variation from 42%to 48% as:

1) an increase of about 15%;

2) an increase of 6 percentage points.

Both are correct, they just use different units.

Hope that it helps!


Francois said...


Don’t forget that you owe me an answer! :-)

Just kidding! But I would really like to wrap up our discussion on hematocrit.

Have a great week-end.