This is from another forum, I'm curious what your opinions are.
-------------------------------------------------

"It is recommended in Joe Friel's Training Bible...

My partner and I have both "tested" it this last week to make sure it doesn't upset our stomachs. We both seem fine.

We are going to try loading with it - the 3 day regimen (4 grams per day) in a week and a half just prior to a TT race in a couple of weeks.

It is meant to help the haemoglobin release oxygen in the muscle, which effectively means that the muscles are able to operate at aerobically at higher speeds and power outputs than usual.

Some studies have shown that it effectively raises the lactate threshold by up to 10%. The gains from Sodium Phosphate loading are supposedly still there after a week."

Never heard of this..

lots of google 'information'.

http://www.uvm.edu/~uvmhpl/research.html#KM

Effects of Sodium Phosphate Supplementation on Maximal Oxygen Consumption and Blood Lactate
KM Brennan

The use of sodium phosphate (NaPO4) supplementation to improve endurance performance has long been proposed without substantial evidence. There are several investigations of NaPO4 supplementation in the literature, however the findings are equivocal. Phosphate loading has been proposed to alter metabolic and cardiovascular functions including; an attenuated lactate threshold, an increased availability of inorganic phosphate for creatine phosphate synthesis, enhanced myocardial and cardiovascular efficiency, and an improved aerobic capacity. To test the theory that NaPO4 could improve aerobic performance the current study was undertaken. 12 previously trained male subjects (Mean VO2max 60.6 ± 4.4 ml?kg-1?min-1) participated in a double blind crossover experiment to determine the effects of a commercial sodium phosphate supplement on maximal oxygen consumption (VO2max) and blood lactate. Subjects performed an incremental VO2max test on a cycle ergometer using 3-minute stages. Lactate concentrations were determined from fingerstick blood samples at the end of each stage and immediately analyzed as whole blood. After the completion of pre-supplement testing, subjects were randomly assigned either a sodium phosphate or a placebo supplement. The supplementation regimen consisted of 1000mg of either dibasic sodium phosphate or placebo four times a day for 4 days. On day 5 subjects returned to the lab to repeat the testing protocol. Repeated measures analysis of variance (P< 0.05) indicated no significant effects of sodium phosphate supplementation in contrast to the placebo treatment for VO2max or blood lactate. In conclusion, our data suggest that NaPO4 is an ineffective supplement when used to enhance endurance performance.



This is from another forum, I'm curious what your opinions are.
-------------------------------------------------

"It is recommended in Joe Friel's Training Bible...

My partner and I have both "tested" it this last week to make sure it doesn't upset our stomachs. We both seem fine.

We are going to try loading with it - the 3 day regimen (4 grams per day) in a week and a half just prior to a TT race in a couple of weeks.

It is meant to help the haemoglobin release oxygen in the muscle, which effectively means that the muscles are able to operate at aerobically at higher speeds and power outputs than usual.

Some studies have shown that it effectively raises the lactate threshold by up to 10%. The gains from Sodium Phosphate loading are supposedly still there after a week."

Some studies have shown that it effectively raises the lactate threshold by up to 10%. The gains from Sodium Phosphate loading are supposedly still there after a week."

Despite the lack of VALID studies showing any benefit, this comes up every few years. My recommendation for those wanting more phosphate would be to drink Coke, which has phosphoric acid and therefore is a good source of phosphate. :)

Despite the lack of VALID studies showing any benefit, this comes up every few years. My recommendation for those wanting more phosphate would be to drink Coke, which has phosphoric acid and therefore is a good source of phosphate. :)

I've also heard that taking a couple of antacid tabs (ie Tums) will help keep LA levels down during hard efforts. Might be BS, but I've heard people say this before.

I like Coke.

Although a couple weeks ago I pulled the "not very flat coke in the water bottle on the seat tube and then forget and close the bottle" trick.

Which means your whole underside gets a Coke shower at the first bump.

Despite the lack of VALID studies showing any benefit, this comes up every few years. My recommendation for those wanting more phosphate would be to drink Coke, which has phosphoric acid and therefore is a good source of phosphate. :)

There are a couple of studies out there showing that Sodium Bicarbonate (baking soda) lowers plasma pH and leads to improved cycling performance. I'm surprised someone in the supplement industry hasn't jumped on this but when you consider the lack of a scientific basis for most things in the supplement industry and how cheap baking soda is, it's not surprising :)

I thought that you had to sniff coke to get more energy.

One of the central limitations of sports medicine research is generally these studies are done on very small samples. Because of this they have very little power, ie. their conclusions are really not statistically significant. When you look at any given topic there are multiple small studies with differing conclusions. Sodium phosphate is typical, there are studies on either side; benefit, no benefit.
To be fair these studies are generally poorly funded and technically difficult to perform. It would be very, very difficult to conduct the large placebo based randomized trials everyone in medicine loves.
One word of warning though, look at who funds these different projects. Often these research studies are conducted by researchers or labs with a vested intereest in the product they are assessing. Accelerade is certainly a good example of bad science conducted by biased researchers.
My take on phosphate is it probably won't hurt so give it a shot if you want. Doubt it does much of anything though.

"Because of this they have very little power, ie. their conclusions are really not statistically significant."

I'm no statistician but I don't believe that is an accurate statement. If something is statistically significant, well then it's statistically significant. Statistical "power" is the ability to detect a difference. By definition, if you have a statistically significant difference then you have sufficient power. However, if you fail to detect a difference you could be under-powered, and more subjects could lead to a statisically significant difference.

By not "really statistically significant" you may mean "meaningful"? Although I believe you are still off base. A study that has low power (due to low # of subjects or large variability in the measure of interest) is less likely to lead to a statistical difference but if it does, the effect size is likely to be large and therefore meaningful. OTOH, a study with a large number of subjects can more readily yield a statistical difference despite a small effect size, and therefore might not be really all that meaningful.

"When you look at any given topic there are multiple small studies with differing conclusions. Sodium phosphate is typical, there are studies on either side; benefit, no benefit."

Often it is true that different studies examining the same factor of interest reveal mixed results. I'd be more inclined to attribute these differences to methodology (different study populations, different doses, different outcome measures, etc.) rather than low sample numbers.

I've tried it (Hammer Nutrition Race Day Boost) prior to my bigger races over the last two seasons with success. I felt stronger during the early stages of my races (24 hour solo mountain bike races). Obviously this is a very subjective since the whole point is peak on these days anyway and I'm amped to be racing and blah, blah blah....., so give it a try and see.

Eddie O

Sodium Phosphate can be found in smoked/cured meats, if one is looking for a source in their diet. I don't know how many milligrams per ounce of meat one would get. I do recall reading in the past that phosphates are important for bone, teeth and skin cells. Lean smoked turkey breast, roast beef, ham, jerkey, etc. could be incorporated into one's diet.

Deek

I tried it a couple of years ago. 3 times in one season, and once the following season. The stuff don't do crap. I felt no different, I was no faster, I was $50 poorer, and my blood pressure was significantly higher.
Waste of cash. Save your money and buy better tires, they will make you faster for longer.

I'm no statistician but I don't believe that is an accurate statement. If something is statistically significant, well then it's statistically significant. Statistical "power" is the ability to detect a difference. By definition, if you have a statistically significant difference then you have sufficient power. However, if you fail to detect a difference you could be under-powered, and more subjects could lead to a statisically significant difference.

Power is a measure of the ability to extrapolate a study's results to the entire population. Statistical significance refers to whether or not the data you obtained for the 2 groups is actually different. It says nothing about the ability to detect a difference. Statistical difference is used to describe the data, power describes the study.

So what the original poster said is exactly right. It is very possible to have a study produce a 'statistically significant' difference but lack the power to say the difference is real.

Scott

P.S. Power, just like SS, is not binary. The convention for 'statistical significance' is a 95% confidence interval, i.e. there is a 95% chance that the difference you detected is not due to a statictical anomoly. Power is also a gradient. You decide what power you want to achieve when determining your study size (or, more comonly, scientists ignore this step and just go ahead regardless =).

"Power is a measure of the ability to extrapolate a study's results to the entire population."

Like I said I'm not a statistician but that's not my understanding of statistical power (or perhaps I'm not understanding your meaning). My understanding is just as it is described here:

http://en.wikipedia.org/wiki/Statistical_power

That is, power is the ability to detect a difference in your study population assuming there is a real difference in the population as whole.

"Statistical significance refers to whether or not the data you obtained for the 2 groups is actually different. It says nothing about the ability to detect a difference."

Yes I hope I didn't convey that but I was trying to convey that power does influence your ability to detect a difference.

"Statistical difference is used to describe the data, power describes the study."

Again perhaps I don't get your point but that's not how I think of power.

"So what the original poster said is exactly right. It is very possible to have a study produce a 'statistically significant' difference but lack the power to say the difference is real."

I guess I don't see that from my understanding of power which again is consistent with the Wikipedia definition. My understanding is that if you detect a statistical difference than you have sufficient power by definition (that is, you can't make a Type II error as you don't have a "false negative" you in fact have a positive finding).

"You decide what power you want to achieve when determining your study size (or, more comonly, scientists ignore this step and just go ahead regardless =)"

If I understand you that is the context I've seen it used (i.e. to estimate study size) but I still fail to see how one can achieve a statistically significant difference between your study groups and due to lack of power not have it apply to the population as a whole?

You can use a lot of difference statistical tools when looking at two groups and seeing if differences between the two are due to chance alone or are true differences. Bottom line is that when one says it is statistically significant it means that the observed differences are "true" and less likely to be due to chance based on the statistical tool you use.
The power of a study is a calculation of how many subjects you need to have in order to perceive a difference if one existed. This is a bit of a chicken and the egg problem because you need to know the prevelence of something before you can design a study to measure it. For example if there was a 90% difference in performance you would need fewer subjects than if there was only a 10% difference. So even if you did find a statistically significant difference if your power calculations were off then you cannot be sure your statistical test was accurate. Typically the way you solve the chicken and the egg problem is you make a rough guess, overestimate and then once you have your data you retrospectively do your power calculations. Hopefully you had enough subjects in your study.
The bottom line is that many of these studies suffer both from low power as well as low statistical significance. Often the p-values (a term used in one of the forementioned statistical tools) are either not reported or are not significant. I think my original language was a bit hurried and not as accurate as it could have been.
The differences may certainly be due to study design, but they also suffer from underpowered studies with low statistical significance. It's hard to know which is true. But if there really was a "significant" difference you would expect to see a consistent result regardless of study design.
Enough of that. Now lets talk about bikes.

"So even if you did find a statistically significant difference if your power calculations were off then you cannot be sure your statistical test was accurate. Typically the way you solve the chicken and the egg problem is you make a rough guess, overestimate and then once you have your data you retrospectively do your power calculations. Hopefully you had enough subjects in your study."

I'm really not being pedantic but as I understand it power is primarily related to the chance of Type II error (i.e. false negative). I don't see how it is related to a Type I error due to low subject numbers (i.e. false positive) which was the implication in the original post. In fact, just the opposite is true. When you have large numbers of subjects you could get a statistically significant difference in your groups that is not reflective of the population (i.e. you've made a type I error) because you're over-powered. So in a small sample study, which is likely to be under-powered, if you find a statistically significant difference in your groups it is highly likely to be generalizable to the population as a whole, no?


"The bottom line is that many of these studies suffer both from low power as well as low statistical significance. Often the p-values (a term used in one of the forementioned statistical tools) are either not reported or are not significant."

Well I was referring to peer-reviewed studies in scientific journals. I can't recall an article ever not reporting statistical significance of being at least p < .05. True, power is almost never reported but again if I understand it, low power is not a problem if you detect a difference in your groups.

"But if there really was a "significant" difference you would expect to see a consistent result regardless of study design."

I don't know how that could be true. For example, the most common assumption people make is if a study shows substance X improves performance then all they look for is substance X. They fail to take into account the dosage size of substance X which will always matter. Similarly, performance is often measured via different means (peak sprint power, repeated sprint power, time to exhaustion, increased power over a given time interval, etc.) or even more confounding all they look at is some physiologic measure (like lactate threshold or lowered blood pH) that often corresponds to performance, yet they never actually measure performance. I could go on, but in my opinion mixed results from supplement studies probably almost always come down to these factors.

You really dig this stuff don't you?
Power has three factors 1)Amount of error you are willing to accept 2)SIze of the difference you are trying to detect 3)Sample size.
While power is most directly related to beta error, you can reduce both alpha and beta error by enlarging the group. The idea of a study being overpowered and not applicable to a general population is just not correct.
I have in fact seen a number of sports medicine articles who report positive findings with p values in excess of 0.05 whose confidence intervals cross 1.
Doing a study of twelve althletes and detecting a difference of 10% and reporting this as statistically significant is an underpowered study. And a bad conclusion.
Can we talk about bikes now?
What about wheels...the tour? Lance, Landis....something else please?

Amphetamines work much better. They will knock 3-5 minutes off your best TT time.

"You really dig this stuff don't you?"

No if I really dug it, I'd understand it. But as a scientist I need to have at least a working knowlege of stats. What's been posted here is seemingly at odds with my understanding of power.

"Power has three factors 1)Amount of error you are willing to accept 2)SIze of the difference you are trying to detect 3)Sample size."

Yes, but the "amount of error" (as I understand it) is type II error, that is you fail to detect a difference when one really exists. How does power relate to type I error, that is you detect a difference but it doesn't really exist? I'm failing to see that but that is the crux of your claim, no?

"While power is most directly related to beta error, you can reduce both alpha and beta error by enlarging the group. The idea of a study being overpowered and not applicable to a general population is just not correct."

Yes, rereading my previous post I see that I completely typed the wrong thing. It's not that a study that is "over-powered" increases the chances of an Type I error it's that you can get a signficance difference yet your effect size is small and therefore might not be all that meaningul.

"Doing a study of twelve althletes and detecting a difference of 10% and reporting this as statistically significant is an underpowered study. And a bad conclusion."

I'm not disputing you, but again if I understand you you're saying that an underpowered study can lead to Type I errors as well as Type II errors? Or are you saying something entirely different?

Last post on the issue, I promise.

Alpha and beta error are two sides of the same coin. The more your study is geared to reduce alpha error the more beta error you will potentially generate.
For example I have ten patients and I want to know if you are having a heart attack, so I find a nice lab test like troponin and use that to assess if you are having a heart attack. Tropoinin is a cardiac protein that goes up if you are having a heart attack. Now there is a range of possiblities...say I can pick 1 or 0.5 as the magic number. If I was to pick 1 then the chances of a person having a heart attack rather than a false positive is going to get better...my confidence that a positive test actually means you are having a heart attack goes up. But that also means that if the result is 0.7 then my confidence that it is not a heart attack goes down. Now say I pick 0.5 as the cutoff. If you have a level of 0.4 my confidence that you are not having a heart attack goes up, but likewise if the level is 0.6 my confidence that you are having one goes down. So alpha and beta are inversely related the better you are at one the worse you are at the other.
Power depends on the prevalence of disease. If only 0.1% of the general population gets a heart attack and I study a random set of ten then the chance of me having beta error skyrockets. The chances of me finding someone with a heart attack are decidedly low. So even if my results are statistically significant, I did not look at enough patients to actually know if their troponins elevate when they are having a heart attack. But if my sample size is correct than I can both reduce alpha and beta error. I have 10,000 patients and now I have enough people with a heart attack to know what their troponins do in the setting of a heart attack. So now I can reduce alpha error as well. Say one out of ten patients will have an elevated tropoinin in the setting of kidney failure. If I only have one patient with a heart attack in my study then I can't be as confident that their elevated troponin is related to kidney failure than having a heart attack. So while power is most intimately connected with beta error it is also very related to alpha as well.
And thats the omega

Thanks for all the responses - I'm the one Snap quoted from another forum

I was hoping the members there would give me the kind of feedback thats happening here, but that conversation has not really taken hold.

The race is in 8 days and I'm probably gonna try the loading 3 days out (its too late for Friel's suggestion of loading from 19 days out anyway)

Zoikz and Dwayne, where your discussion started kind've hints at my dilemma too - is it worth me taking it when I may not even know if it worked? I have been doing TT specific training and know I have improved (for example, I did and Aerobic Time Trial last week and shaved almost 2 minutes off my previous PB).

So on Saturday week, if I can lift my performance like I did for the ATT (and I will be racing at/above my LT not just below like the ATT requires) then will my good time (I hope) be about specific useful training, or about NaPO4?

I just don't know.

Wish I had a power tap...

Amphetamines work much better. They will knock 3-5 minutes off your best TT time.

Caffeine works just aswell and its legal...too hard to ban.

Caffeine works just aswell and its legal...too hard to ban.

Actually there is a limit on caffeine as well. It's pretty easy to ban most anything....the hard part is enforcing it.

Eddie O

There's a limit on caffiene?

(Not that I race at a level where I will be tested!)

Staring loading with NaPO4 today by the way...

There's a limit on caffiene?

(Not that I race at a level where I will be tested!)

Staring loading with NaPO4 today by the way...
There used to be a limit on caffeine, not anymore. Guess they found out the ill effects of heavy use outweighs the positive ones.

This is the active ingredient in bowel cleansing Fleet Phosphosoda used for colonoscopy prep. If you know someone who'se done this, they'll gladly share the horror which is prep day.

http://www.phosphosoda.com/consumers/products/phosphosoda.aspx

I think this risk FAR outweighs any potential performance benefit!