LONGECITY

Hi everyone,

I figure I will start this showing some data regarding dosages in studies.

I will update this picture as I get more info from the various studies done using Resveratrol. Please remember that this is just a direct conversion using math, no species related changes have been performed on this. Since Bioavailability is an issue in humans, it may take more RSV for humans if RSV is taken by mouth. If you have dosages from other studies or corrections, download the excel and email it to me with them. I will post them here.

Now let me start with a couple popular ones:

Excel Sheet With Animal Dosages

Here is what the excel sheet currently shows:


This sheet states what the equivalent dose for a human weighing 130lbs would be IF the studies directly correlated to humans.
This is just for reference, don't use it to calculate your own dosage, we know it's not accurate... but certainly interesting.

Thanks
Anthony Loera
Edited by Anthony_Loera, 21 September 2007 - 04:46 PM.

Interesting topic. Does anyone know-I've been meaning to comb through the literature- have people looked at the levels of resveratrol in the blood of the animals used in the studies where a benefit has been observed?

Interesting that you chose not to include the 1/6 dosage scaling that has been thrown around a lot.

geo12the,

here is some info lifted off the literature... (excuse my spelling if I get a typo on these..)

(high abosption but low bioavailability of oral resveratrol in humans, 2004)
(using 25mg per 70kg)Humans -Serum Peak <22 nM

(metabolism and disposition of Resveratrol in Rats, extent of absorption, glucoronidation, and enterophatic circulation, 2002)
(using 50mg/kg of bodyweight) Rats -Serum Peak <6.6 µM <- This is supposed to be 'micro' but its not showing up right.

(Pharmacokinetics in mice and growth inhibitory properties of putative cancer chemopreventative agent resveratrol and synthetic analogue trans, 2004)
(using 240mg/kg of bodyweight) Mouse -Serum Peak <32 µM <- This is supposed to be 'micro' but its not showing up right.

Here's a chart for an easy understanding...(gotta love the wiki!)


Anthony Loera

Fish study added... thanks Malbec !!!!

(high abosption but low bioavailability of oral resveratrol in humans, 2004)
(using 25mg per 70kg)Humans -Serum Peak <22 nM

metabolism and disposition of Resveratrol in Rats, extent of absorption, glucoronidation, and enterophatic circulation, 2002)
(using 50mg/kg of bodyweight) Rats -Serum Peak <6.6 µM

(Pharmacokinetics in mice and growth inhibitory properties of putative cancer chemopreventative agent resveratrol and synthetic analogue trans, 2004)
(using 240mg/kg of bodyweight) Mouse -Serum Peak <32 µM

Using these numbers, plus the study below, where humans given 71mg/kg (assuming 70kg body wt. with 5gm oral resv) had a serum peak of 2.4 µM, we can look at the serum levels obtained per 1mg/kg dose.

6.6uM/50mg kg(-1) = 0.132 (rat)
32 uM/240mg kg(-1) = 0.133 (mouse)
0.022 uM/0.357 mg kg(-1) = 0.0616 (human low dose)
2.4 uM/71mg kg(-1) = 0.0338 (human high dose)

I consider the human high dose number to be more reliable due to the larger values involved, Boocock's involvement in development of resveratrol quantitation methodology, and the fact that the 22 nM value (from Walle et al., Drug Metab Dispos. 2004 Dec;32(12):1377-82.) was given as an upper limit. I wish that I had the full paper from Boocock et al., because then I'd feel comfortable comparing the high and low dose numbers to see if there is a "swamping" effect. Rat and mouse are certainly consistent here, though they are both at high dose levels. For the time being I'll just compare the rat and human high dose numbers:

0.132/0.0338 = 3.9
Edit 7-09-08: This ratio is too high because the rat and mouse data above uses formulations that enhance bioavailability, while the human data does not. A more correct ratio is 1.3. For more explanation, see this post.

So it looks like the factor of six that is sometimes used to compare doses between mice and men works in reverse in the case of resveratrol. We don't need one sixth of the dose, but rather something on the order of two to four times one half to two times as much, if we wish to reach plasma levels comparable to the mice.

Cancer Epidemiol Biomarkers Prev. 2007 Jun;16(6):1246-52.
Phase I dose escalation pharmacokinetic study in healthy volunteers of resveratrol, a potential cancer chemopreventive agent.
Boocock DJ, Faust GE, Patel KR, Schinas AM, Brown VA, Ducharme MP, Booth TD, Crowell JA, Perloff M, Gescher AJ, Steward WP, Brenner DE.
The red grape constituent resveratrol possesses cancer chemopreventive properties in rodents. The hypothesis was tested that, in healthy humans, p.o. administration of resveratrol is safe and results in measurable plasma levels of resveratrol. A phase I study of oral resveratrol (single doses of 0.5, 1, 2.5, or 5 g) was conducted in 10 healthy volunteers per dose level. Resveratrol and its metabolites were identified in plasma and urine by high-performance liquid chromatography-tandem mass spectrometry and quantitated by high-performance liquid chromatography-UV. Consumption of resveratrol did not cause serious adverse events. Resveratrol and six metabolites were recovered from plasma and urine. Peak plasma levels of resveratrol at the highest dose were 539 +/- 384 ng/mL (2.4 mumol/L, mean +/- SD; n = 10), which occurred 1.5 h post-dose. Peak levels of two monoglucuronides and resveratrol-3-sulfate were 3- to 8-fold higher. The area under the plasma concentration curve (AUC) values for resveratrol-3-sulfate and resveratrol monoglucuronides were up to 23 times greater than those of resveratrol. Urinary excretion of resveratrol and its metabolites was rapid, with 77% of all urinary agent-derived species excreted within 4 h after the lowest dose. Cancer chemopreventive effects of resveratrol in cells in vitro require levels of at least 5 mumol/L. The results presented here intimate that consumption of high-dose resveratrol might be insufficient to elicit systemic levels commensurate with cancer chemopreventive efficacy. However, the high systemic levels of resveratrol conjugate metabolites suggest that their cancer chemopreventive properties warrant investigation. (Cancer Epidemiol Biomarkers Prev 2007;16(6):1246-52).

PMID: 17548692

Edited by niner, 09 July 2008 - 05:45 AM.

The top post...130#...what humans are we talking about? I agree for going for healthy assertions of somatotypes, but wow. Even using dietetic calculations, that is the Ideal Body Weight for a 5'6" female and IBW for a 5'4" Male. The 70kg number (154#) is a bit more realistic and would be the IBW for a 5'8" male. Personally, none of these check out for me when I am 6'3" 211# @ 9% BF. I think I'd have to reassess all these calculations for my intents and purposes.

Random thought, I know, but these numbers threw me a bit.

Good point, neogenic. 70kg has been "the standard of the industry" for a long time. Maybe too long... I've seen it used in these kinds of calculations forever, but you're right; I don't think it's in the middle of the curve of weight distributions today. I'm 5' 10.5", 68.2kg, and not emaciated, but I'm not bulked out either. Anthony, how did you pick 130 lbs?

It's an Excel sheet guys... If you type in the weight it will calculate the rest...
I'll change it to the equivalent over the weekend if you wish.

thanks
A

What conversion was used for metabolic differences between Rodent and MAN? I believe I've seen 7:1 or 8:1. I can't remember the source of the data, but it certainly lends useful in many studies we discuss. If this wasn't used than the gram amounts for the human model are significantly overcalculated.

Example of metabolic rate conversion (formula can be put in the spreadsheet, too).
Say the small animal weighs 1kg, and the human 81kg. The rate is then (81/1)^0.25=3 (the small animal needs 3 times as much stuff /kg, as its metabolism is roughly 3 times faster)

So it looks like the factor of six that is sometimes used to compare doses between mice and men works in reverse in the case of resveratrol. We don't need one sixth of the dose, but rather something on the order of two to four times as much, if we wish to reach plasma levels comparable to the mice.

Niner, your math looks right (a factor of 4 times /kg in our favor to accomplish similar-to-rodent concentrations), but goes counter metabolic dogma. Why should resv be so special? Any other examples of reverse metabolic direction you've come across?
Is it possible we just don't have enough experimental data to form a definitive conclusion?

What's wrong with 130 lbs? (at 5'9")? Works for me!

What's wrong with 130 lbs? (at 5'9")?  Works for me! 

I don't know if that's longevity if you're hitting that kind of CR. Anorexics may live forever if that's the case.

To convert animal dose in mg/kg to HED in mg/kg, you divide the dose of the animal dose in mg/kg by 12.3 for a mouse. Rat is 6.2. Guinea pig is 4.6. Monkey is 3.1. I did some searching. These numbers need to be reflected in the calculations above.

http://www.fda.gov/cber/gdlns/dose.htm

I guess the 9 or 8:1, that came to mind was a hybrid number of mice and rats.

For the fish study above, my guess is (since its not listed in the chart fom the link I give) it'd be around where a rat is for conversion.

Niner, your math looks right (a factor of 4 times /kg in our favor to accomplish similar-to-rodent concentrations), but goes counter metabolic dogma. Why should resv be so special? Any other examples of reverse metabolic direction you've come across?
Is it possible we just don't have enough experimental data to form a definitive conclusion?

I seem to recall that rats are heavier on oxidative metabolism than humans, but resveratrol metabolism in humans is dominated by conjugation, not oxidation. That's one hypothesis, anyway, though not all that satisfying, since just because they are heavier on oxidative metabolism doesn't mean they are weaker on conjugative. My first stab at it was just to equalize plasma levels, but that might not be right. Humans might not need equivalent plasma levels to get the same pharmacological effect as a rat. To compute that, we should probably consider the volumes of the relevant tissue compartments- muscle and neural, for example. These may well scale better with body surface area, so perhaps the body surface area scaling factors that neogenic gave should still be used. My inclination would be to still use the plasma concentration data, since that eliminates unknowns regarding absorption and metabolism, but to apply the scaling factors to that. Do you use 12, from the mouse, or 6, from the rat? We also have a factor of two difference in the human Cmax/mg kg^-1. Depending on which two factors you select from these four, you would have a dose scaling of anywhere from 1/6 to 2/3 of the rodent dose. (From 2/12 and 4/6, respectively.) That is certainly encouraging, but it seems to be relying on an assumption that resveratrol from plasma is being concentrated in the relevant tissues. Resveratrol has a logP of about 3, so it should partition preferentially into hydrophobic media, like a cell membrane. I don't know what the hydrophobicity of the environment directly around the target protein is, but I'd have to guess that it was aqueous. The fact remains, the target protein needs to be exposed to a concentration around the ED50 for binding in order to elicit a response. That inescapable fact leaves me feeling a little uncomfortable about the scaling...

When's the Sirtris human data supposed to hit the street? Late '08? If they publish enough of what they learn, that should tell us something about the relationship between blood levels and pharmacologic response in humans.

edit: added inconclusive comments on logP
edit2: removed stray word...
Edited by niner, 24 September 2007 - 04:54 AM.

The serum-levels in humans could also be largely irrelevant as a measure of bioavailability. Consider:

Biochem Pharmacol. 2004 Sep 15;68(6):1113-8.Click here to read Links
    Transport of resveratrol, a cancer chemopreventive agent, to cellular targets: plasmatic protein binding and cell uptake.
    Jannin B, Menzel M, Berlot JP, Delmas D, Lançon A, Latruffe N.

    Laboratoire de Biologie Moléculaire et Cellulaire, Université de Bourgogne, 6 boulevard Gabriel, 21000 Dijon, France.

    Resveratrol produced by several plants, berries and fruits, including grapes, is one of the best known natural food microcomponents with potent chemopreventive properties towards the most severe contemporary human diseases: cardiovascular sickness, cancer and neurodegenerative pathologies. Demonstration of its mechanism of action also implies the elucidation of the steps of bioavailability and bioabsorption in cells and tissues. In order to estimate the relationships between the amounts of resveratrol taken up by food or drink intake, and the several possible benefits illustrated from in vitro/in vivo experiments and from epidemiological studies, it is essential to demonstrate step by step the route of resveratrol from plasma to the cell active site. In plasma, resveratrol was shown to interact with lipoproteins. This commentary also contains previously unpublished results about interactions between resveratrol and albumin and the enhancement of this binding in presence of fatty acids. We have previously described that resveratrol uptake by hepatic cells involves two processes--a passive one and a carrier-mediated one. Thanks to this last process, resveratrol, while tightly bound to blood proteins, could be largely delivered to body tissues. The intracellular proteic targets of resveratrol remain to be identified.

    PMID: 15313407 [PubMed - indexed for MEDLINE]

Niner, your math looks right (a factor of 4 times /kg in our favor to accomplish similar-to-rodent concentrations), but goes counter metabolic dogma. Why should resv be so special? [...]

I seem to recall that rats are heavier on oxidative metabolism than humans, but resveratrol metabolism in humans is dominated by conjugation, not oxidation. That's one hypothesis, anyway [...] My first stab at it was just to equalize plasma levels, but that might not be right. Humans might not need equivalent plasma levels to get the same pharmacological effect as a rat. [...] My inclination would be to still use the plasma concentration data, since that eliminates unknowns regarding absorption and metabolism, but to apply the scaling factors to that. [... T]he target protein needs to be exposed to a concentration around the ED50 for binding in order to elicit a response. That inescapable fact doesn't leaves me feeling a little uncomfortable about the scaling...

You don't think the FDA's proposal for all medications and trial in stages using the animal model for Human Equivalency Dose is relevant that I listed? It's based on truckloads of data. I understand your point, but many drugs work by a variety of mechanisms and this is what is the gold standard when moving to HED...which is the point of the graph listed above. I wish I could just post the table in here, but if you go to the link and scroll down a bit...it's there.
Edited by Michael, 24 July 2009 - 03:25 PM.
Trim quotes

You don't think the FDA's proposal for all medications and trial in stages using the animal model for Human Equivalency Dose is relevant that I listed? It's based on truckloads of data. I understand your point, but many drugs work by a variety of mechanisms and this is what is the gold standard when moving to HED...which is the point of the graph listed above. I wish I could just post the table in here, but if you go to the link and scroll down a bit...it's there.

Well, it's for first uses in humans before you have any data on blood levels. If you're willing to accept a blood level that's lower than the IC50 (or ED50) for the target, then you have to rely on the drug being concentrated at the target via active transport or favorable partitioning. The paper that Maxwatt posted suggests the existence of an active transport route, so maybe we don't need as high a blood level as the in vitro binding data would suggest. Such a transport system, if it exists, may well be active in rats as well.

From the FDA site link that was pointed out:

If the parent drug is measured in the plasma at multiple times and fits the range of toxic dose for two or more animal species, it may be possible to develop a pharmacokinetic model predicting human doses and concentrations and draw inferences about human safe plasma levels in the absence of prior human data. While quantitative modeling for this purpose may be straightforward, the following points suggest this approach may present a number of difficulties when evaluating estimates of a safe starting dose. Generally, at the time of IND initiation, there are a number of unknowns regarding animal toxicity and comparability of human and animal pharmacokinetics and metabolism: (1) human bioavailability and metabolism may differ significantly from that of animals; (2) mechanisms of toxicity may not be known (i.e., toxic accumulation in a peripheral compartment; and/or (3) toxicity may be due to an unidentified metabolite, not parent drug. Thus, to rely on pharmacokinetic models (based on parent drug in plasma) to gauge starting doses would require multiple untested assumptions. Modeling may be used with greatest validity to estimate human starting doses in special cases where few underlying assumptions would be necessary. Such cases are exemplified by large molecular weight proteins (like humanized monoclonal antibodies), which are intravenously administered, are removed from circulation by endocytosis rather than metabolizism, have immediate and detectable effects on blood cells, and have a volume of distribution limited to the plasma volume. Here, allometric, pharmacokinetic, and pharmacodynamic models have been useful in identifying the human mg/kg dose that would be predicted to correlate with safe drug plasma levels in nonhuman primates. Even in these cases, uncertainties (such as differences between human and chimpanzee receptor sensitivity or density) have been shown to affect human pharmacologic or toxicologic outcomes, and the use of safety factors as described in this document is still warranted.

I updated the excel sheet with some suggested calculations.

let me know if I fat fingered something, and need to edit it again.

A

Thanks for putting this together Anthony. Interesting indeed.

Wow, 28g/day t-resveratrol to get the equivalent rat dosage and double endurance (yes, approximately and hypothetically, I know). About 7 daily teaspoons of 99% resveratrol. Quite a lot, no?

I guess we need to wait for more results ...

Stephen

The FDA conversion factors are conservatively based off of body surface area instead of volumetric-metabolic scaling, which the guide admits is inaccurate and will result in conservatively low human doses.

See:
"Despite the subsequent analyses showing that the MTDs for this set of drugs scale best between species when doses are normalized to W0.75 rather than W0.67 (inherent in body surface area normalization), normalization to body surface area has remained a widespread practice for estimating an HED based on an animal dose. "

General rule of thumb puts mice around 30g, rats around 325g, 3kg rabbit?, humans 70kg.

The formula to get a mg/kg conversion factor would then be (animal weight/human weight)^(0.75)*humanweight/animalweight = 7.3 for a mouse, 3.8 for a rat, and I think about 1.85 for a rabbit. Anybody know what the fish in the fish study weighed, or better yet, how many calories per gram of fish food?

Thanks

Hormesis,

I have the fish paper and they are purposely obtuse about what the fish weighed. They do say that the fish consumed 50 mgs food per gram of fish weight per day (and the resveratrol concentration in the food was 120 micrograms/gram food). That is how I was able to scale to a human dose by using the human weight in the Excel spreadsheet.

These killifish are pretty tiny, I am guessing a weight of ~10-15 grams. Funny, I did a Google on them and almost all the hits come back referencing this study with resveratrol in one way or another.

The FDA conversion factors are conservatively based off of body surface area instead of volumetric-metabolic scaling [...]

The formula to get a mg/kg conversion factor would then be (animal weight/human weight)^(0.75)*humanweight/animalweight = 7.3 for a mouse, 3.8 for a rat, and I think about 1.85 for a rabbit. Anybody know what the fish in the fish study weighed, or better yet, how many calories per gram of fish food?

Edited by Michael, 24 July 2009 - 03:27 PM.
Trim quotes

Tables 4&5 get in to using body weights and how they relate to the surface area with analyses and algorithms. It does go a bit deeper than just "surface area".
http://www.fda.gov/cber/gdlns/dose.htm

Tables 4&5 are still using the b=0.33 exponent for body surface normalization. You'ld need to use 0.25 for metabolic normalization.

I did a search on killifish and found one site quoting 0.1g for 3-5cm fish which sounds awfully low.
http://www.uoguelph....es/fishing.html
5.6g for 7cm fish -which sounds about right
http://www.pubmedcen...gi?artid=434418

So assuming 5.6g, the'ld be eating about 5% of their body weight per day which sounds about right.

Assuming a 5.6g fish, the scaling factor would be 10.57

I took Anthony's t-res 99% pure about a tablespoon worth and still get bad diarrhea -- any idea why? Supposed to be no emodin.

Is there something else in pure t-res that causes this? It's consistent with anthony's -- but I don't get it at all with Longivenix's res which I also take -- but at smaller dosages admitedly.

I can't take large doses if this keeps up unless the res is low quality.

A tablespoon would be about 12g. I've never heard of anyone on a dose that high.

Stephen

Edit: I should have been using 1/2 teaspoon per 1 g, so that would come out to 6 g total -- sorry.
Edited by stephen_b, 26 September 2007 - 01:06 PM.