14 replies to this topic

[zoolander]

I was reading a paper recently and thought that I would post a formula that helps you convert dosages given in animal studies into a human equivalent dosage

FASEB J. 2007 Oct 17; [Epub ahead of print]

    Dose translation from animal to human studies revisited.
    Reagan-Shaw S, Nihal M, Ahmad N.

    *Department of Dermatology,Paul P. Carbone Comprehensive Cancer Center;Molecular and Environmental Toxicology Center, University of Wisconsin, Madison, Wisconsin, USA.

    As new drugs are developed, it is essential to appropriately translate the drug dosage from one animal species to another. A misunderstanding appears to exist regarding the appropriate method for allometric dose translations, especially when starting new animal or clinical studies. The need for education regarding appropriate translation is evident from the media response regarding some recent studies where authors have shown that resveratrol, a compound found in grapes and red wine, improves the health and life span of mice. Immediately after the online publication of these papers, the scientific community and popular press voiced concerns regarding the relevance of the dose of resveratrol used by the authors. The animal dose should not be extrapolated to a human equivalent dose (HED) by a simple conversion based on body weight, as was reported. For the more appropriate conversion of drug doses from animal studies to human studies, we suggest using the body surface area (BSA) normalization method. BSA correlates well across several mammalian species with several parameters of biology, including oxygen utilization, caloric expenditure, basal metabolism, blood volume, circulating plasma proteins, and renal function. We advocate the use of BSA as a factor when converting a dose for translation from animals to humans, especially for phase I and phase II clinical trials.-Reagan-Shaw, S., Nihal, M., Ahmad, N. Dose translation from animal to human studies revisited.

    PMID: 17942826 [PubMed - as supplied by publisher]

The formula

Human Equivalent Dose (mg/kg) = Animal dose (mg/kg) multiplied by Animal Km / Human Km

Where

Species and Km values based on body surface area
Adult Human 37 (60kg)
Child Human 25
Baboon 20
Dog 20
Monkey 12
Rabbit 12
Guinea Pig 8
Rat 6
Hamster 5
Mouse 3


The above paper gives an example of converting a dose of resveratrol (22.4 mg/kg) given to a mouse into the Huma Equivant dosage based on body surface area

To convert into HED

the formula again

Human Equivalent Dose (mg/kg) = Animal dose (mg/kg) multiplied by Animal Km / Human Km

Plugging in values

HED = 22.4 (mg/kg) x 3 (mouse Km) / 37 (Human Km)

The HED = 1.82 mg/kg

If you weighed 70kg then your equivalent dose would be 127.4mg

[craigb527]

What does a mouse actually weigh?

[billypc99]

Funny,

I just brought up the same issue and calculations yesterday in a different thread only to find out that neogenic posted a similar thread not too long ago. Maybe people are starting to take note!

[malbecman]

I think HED using BSA is a good approach, however, keep in mind it does not take into account differences in Phase I and Phase II metabolism between species.

So the HED for resveratrol may be 127mg but if all of our dose is rapidly converted to glucurindated or sulfated forms and cleared, the HED dose may not reach its target(s). Of course, there is the whole question of whether the metabolites are as bioactive as the parent compound (no one knows) and also the question of whether there is an active transport mechanism for resveratrol which could help to concentrate it at the necessary target(s). Too many unknowns at this point, IMHO.....

[zoolander]

average mouse weighs about 20g

[niner]

We have peak plasma levels in rodents and humans which suggest that humans are more effective conjugators of resveratrol than rodents. In other words, humans don't reach as high of a blood level of RSV for a given mg/kg dose, even if you account for the interspecies scaling. The HED computation applied to Auwerx's supermice at 400mg/kg would have a 70 kg human taking a little over 2 grams/day. Craigb257 said that he didn't really see the enhanced endurance effect until he was taking 10-12 grams a day. The enhanced endurance he reported was substantial, whereas other reports from people taking lower doses have been a lot more subtle.

The HEDs taken from Body Surface Area (BSA) are a good general rule when you lack other data, and are particularly useful in the earliest stages of human use when you are worried about possible toxicities, but I'm of the opinion that it's better to try to match plasma levels, once these are known in both species.

[apache]

How would you extrapolate rat study effective dosages to humans?

For example, let's say 100mg of chemical 'x' showed signs of effectiveness in rats. How would you go about adjusting this for humans?

My guess is that rats weigh about 300 grams. So you if a human weights 70 grams, the multiplication factor is 233. Therefore, 100mg * 233 = 23.3 grams??????????

Suggestions?

[niner]

How would you extrapolate rat study effective dosages to humans?

For example, let's say 100mg of chemical 'x' showed signs of effectiveness in rats. How would you go about adjusting this for humans?

My guess is that rats weigh about 300 grams. So you if a human weights 70 grams, the multiplication factor is 233. Therefore, 100mg * 233 = 23.3 grams??????????

Suggestions?

Doses are often given in mg/kg body weight. If the rat weighs 300g, that's 0.3kg; 100mg/0.3g = 333mg/kg.
If you translate this straight across to humans, 333mg/kg * 70kg = 23310mg = 23.3g So, yes, you have it correct as far as you go. There are other considerations when translating doses between species. Different species have different metabolic rates, both in terms of energy metabolism and (perhaps more importantly) xenobiotic metabolism; i.e., how fast they metabolize foreign molecules. There are formulae for such conversions based on body surface area and the relation of BSA to mass, but to be honest I don't trust them. They are only used in tox studies where the compound has never been in a human before, so they tend to be very conservative. I prefer to look at the pharmacokinetics of the compound in both the rat and the human, and see how the dose relates to the blood level in each case. If that data is unavailable, then the compound hasn't been studied enough for me to be taking it, so I tend to drop it there.

[rwac]

Additionally, human brains are a lot more metabolically active than rat brains.

So even blood level is not necessarily an accurate measure of how dosage scales to you.

[apache]

How would you extrapolate rat study effective dosages to humans?

For example, let's say 100mg of chemical 'x' showed signs of effectiveness in rats. How would you go about adjusting this for humans?

My guess is that rats weigh about 300 grams. So you if a human weights 70 grams, the multiplication factor is 233. Therefore, 100mg * 233 = 23.3 grams??????????

Suggestions?

Doses are often given in mg/kg body weight. If the rat weighs 300g, that's 0.3kg; 100mg/0.3g = 333mg/kg.
If you translate this straight across to humans, 333mg/kg * 70kg = 23310mg = 23.3g So, yes, you have it correct as far as you go. There are other considerations when translating doses between species. Different species have different metabolic rates, both in terms of energy metabolism and (perhaps more importantly) xenobiotic metabolism; i.e., how fast they metabolize foreign molecules. There are formulae for such conversions based on body surface area and the relation of BSA to mass, but to be honest I don't trust them. They are only used in tox studies where the compound has never been in a human before, so they tend to be very conservative. I prefer to look at the pharmacokinetics of the compound in both the rat and the human, and see how the dose relates to the blood level in each case. If that data is unavailable, then the compound hasn't been studied enough for me to be taking it, so I tend to drop it there.

Thanks for the detailed answer. This is a new arena to me. Could you please be more specific when you use the term "blood level"...blood level of what? The concentration of Chemical X in the blood over a period of time? In this case, this would be a necessary step beyond extrapolating data from rats to humans, and actually testing it in humans to give an idea of rate of metabolism, etc. Is this correct? Thank you.

[kismet]

There are formulae for such conversions based on body surface area and the relation of BSA to mass, but to be honest I don't trust them. They are only used in tox studies where the compound has never been in a human before, so they tend to be very conservative.

I don't think that is perfectly accurate. BSA scaling is supposed to give you the true, effective dose and often it does (or sometimes). AFAIK there's at least a safety of margin of 10 (sometimes much more; e.g. with immunotherapy... to prevent this from ever happening again) used in phase I studies after BSA adjustment (divided by 12.3 for mouse, 6.2 for rat).
IIRC bisphosphonate doses scale accordingly. Human data will be always superior (microdosing is a pretty neat way to get human data early on), but metabolic scaling is pretty important to interpret basic research if no such data is available.

Interesting paper:

FASEB J. 2008 Mar;22(3):659-61. Epub 2007 Oct 17.
Dose translation from animal to human studies revisited.
Reagan-Shaw S, Nihal M, Ahmad N.
http://www.fasebj.or...t/full/22/3/659

Edited by kismet, 04 October 2009 - 10:12 PM.

[Herbanaut]

Hi all, I will appreciate clarification of the following:

What is the actual denotation of the Km constant? What term can be used instead of it?

Thanks!

[DeadMeat]

Hi all, I will appreciate clarification of the following:

What is the actual denotation of the Km constant? What term can be used instead of it?

Thanks!

Welcome to Imminst/Longecity!

km = body weight(kg) / surface area(m²)

[chrono]

I just wanted to bump this thread It seems like a lot of people here are using this allometric scaling calculator these days. I think it's worth noting that "allometric scaling" does not denote a single practice: allometry is the study of the relation of body size to overall shape, and in this context, how to scale dosages between different organisms.

This calculator uses average body mass, absolute dose, and an exponent to compute the human equivalent dose. This article explains the reasoning for the exponent; it strikes me as a kind of unified theory of biological scaling, whose aims are far beyond the scope of clinical application we discuss here.

I'd like to suggest that people start using the method described in this thread, instead. This type of allometric scaling employs body surface area (BSA) in addition to average weight, which research has shown to be more accurate than body weight in scaling lethal dose and maximum tolerated dose, along with basal metabolism, caloric expenditure, oxygen utilization, blood volume, circulating plasma proteins, and renal function. I think this is preferable to using one curve which has been fit to single-celled organisms and elephants, however close it may sometimes be (seems to be better for some animals than for others).

There are still severe limitations to this method: the scaling does not predict plasma concentration of drugs as neatly, and does not take into account differences in first-pass metabolism and drug clearance (probably among other things). So it really only gets us in the ballpark, but I think BSA gives us a better starting point.

It's especially handy since it doesn't involve exponents and average weights, and is very easy. Here's some very quick shortcuts for calculating absolute dose for a 70kg human from the mg/kg dose in common lab animals based on BSA:

• (mouse mg/kg dose) x 5.675 = 70kg human dose
• (rat mg/kg dose) x 11.35 = 70kg human dose
• (rabbit/monkey mg/kg dose) x 22.7 = 70kg human dose
• (dog/baboon mg/kg dose) x 37.83 = 70kg human dose
• (child human mg/kg dose) x 47.3 = 70kg human dose

I'll stress again that these shortcut numbers are used to go from mg/kg to an absolute adult human dose; they increase toward 70, which would be the factor to use for an adult human mg/kg dose.


Here's the handy tables from the paper:

Edited by chrono, 12 October 2011 - 11:34 PM.

[Thell]

Sorry about digging up such an old thread, but I didn't find a more relevant recent one...

 

TL;DR:: I think using the following formula may scale better.

 

human dosage(mg/day) = animal dose(mg/kg/day) * sqrt( animal weight(kg) * human weight(kg) * human height(cm) / animal length/height(cm)  )

 

Please comment on if the logic and math are sound or not.

 

Thanks,

Thell

 

 

 

Missing from the table chrono cites, as well as from the FDA Guidance[1], is the method of calculating BSA m². It looks like there are quite a few different BSA formulas[2] and each has a height/length element. FDA Guidance, Appendix B Table 3: Conversion of Animal Doses to Human Equivalent Doses Based on Body Surface Area doesn't give any indication as to which is used, nor does it list any length measures to backtrack to the BSA m² formula.

 

The beginning of Appendix B Analysis of Body Weight Effects on HED Calculations gives two methods of calculating kₘ, one that relies on a species unique K value and one that does not. Since the species K values are not listed they need to be calculated using K=100*W^.33/km, but given that kₘ is representative of the Weight/BSAm² ratio and since there is very minor difference, the maximum being .5 for Human kₘ and the median and mean being 0.00000 and 0.08539 respectively, between them the simple ratio can be used. (Perhaps someone here has access to the actual K values?)

 

Since the BSA formula isn't given and the Mosteller BSA equation shows quite a few hits within the FDA domain, although this guidance review paper[3] makes some statements as to the guidance being biased toward BSA and some of the drawbacks thereof, I decided to use it for this formula.

 

So if it is accepted that

 

kₘ = W/BASm²

 

then using Mosteller BSA yields

 

kₘ = W / sqrt( H * W / 3600 )

 

and since scaling is done using

 

effective dose (mg/day) = W * dose (mg/kg/day) * kₘ₁ / kₘ₂

 

we can expand and simplify with subscript ₁ representing the from and subscript ₂ representing the to

 

ED (mg/day) = d * sqrt( W₁ * W₂ * H₂ / H₁ )

 

where d is the dose in mg/kg/day and W are in kg and H are in cm.

 

Verification...

- Reworking the example from chrono's referenced table ( Dose translation from animal to human studies revisited[5] ) where the authors scale the mouse dosage given in the paper Resveratrol Improves Health and Survival[6] to humans.

  - From the paper: 22.4 * 3 / 37 * 60 = 108.973

Baur's et al use C57BL/6NIA mice in their paper that the average 30g on a standard diet and on the high calorie with resveratrol are closer to an average of 40g. Substantially greater than the FDA Guidance reference of 20g. The length is not referred to in the author's paper nor could I find a reference for the average length so the Guidance length of 8.82cm is used.

  - Reworked: 22.4 * sqrt( 60 * .04 * 157.46 / 8.82 ) = 146.62

The human height was calculated by solving for cm from the FDA Guidance BASm².

 

 Now, lets see the scaling of going up to a 70kg human

 - 22.4 * 3 / 37 * 70 = 127

vs

 - 22.4 * sqrt( .04 * 131.66 * 70 / 8.82) = 144.817

What's interesting to note is that since the weight changed but the BSAm² didn't the height decreased! A more accurate depiction of a human weighing 70kg would be to use data from something like http://niremf.ifac.c...BOOK/tbl-54.htmto get the average height and then calculate...

 - 22.4 * sqrt( .04 * 70 * 175 / 8.82 ) = 166.96

 

 

 

My interest in this arose from studies, in particular the paper Nicotinamide riboside restores cognition through an upregulation of proliferator-activated receptor-γ coactivator 1α regulated β-secretase 1 degradation and mitochondrial gene expression in Alzheimer’s mouse models[7] where in section 2.4 the authors state that 250 mg/kg/day of NR was given via water (which they state is equivalent to 1300 mg/day in humans). Using weight and length data from a few studies with Tg2576 mice[4] the from weight to be used in the reworked formula is 30g and the length is 9.8cm.

 

Calculating to two humans and one dog the results come out to

 

- Guidance method:

Ref. human = 250 * 3 / 37 * 60= 1216 mg/day

Human 1 = 250 * 3 / 37 * 70.3 = 1425 mg/day

Human 2 = 250 * 3 / 37 * 90.7 = 1838 mg/day

Dog     = 250 * 3 / 20 * 10.4 =  390 mg/day

-Reworked method:

Human 1 = 250 * sqrt( 70.3 * .03 * 166.37 / 9.8 ) = 1495.9 mg/day

Human 2 = 250 * sqrt( 90.7 * .03 * 198.12 / 9.8 ) = 1854.2

Dog     = 250 * sqrt( 10.4 * .03 *  63.5  / 9.8 ) =  355.4

 

These seem like extremely high dosages, although I guess 250 mg/kg/day is mega-dosing for a tiny mouse! I do recall reading something about how the dosages have an effective range which I guess goes more towards a PK/PD model (which I know nothing about, yet)...

 

 

 

References:

 

 [1]:http://www.fda.gov/d...s/UCM078932.pdf

 [2]:http://en.wikipedia....rea#Calculation

 [3]:http://www.fda.gov/o...03/80064268.doc- Docket No. 02D - 0492: Draft Guidance for Industry and Reviewers

 [4]:http://www.alzforum....h-models/tg2576

 [5]:http://www.fasebj.or...2/3/659.full#T1

 [6]:http://www.med.upenn...nd_survival.pdf

 [7]:http://www.ncbi.nlm....nihms429137.pdf