More on Water Weight & Insulin

There's an interesting related conversation on this topic going on over at Paul Jaminet's site:

Water Weight: Does It Change When Changing Diets? Does It Matter?

For those not familiar with the Perfect Health diet, one component is to get ~400 cal in "safe starches" -- around 100g.  For those transitioning to their diet from a low carb diet, some have experienced a not-unexpected weight gain.  I've added some thoughts to the comments there and -- can't be sure and some is just a hunch -- but I think any weight bounce is probably more carb associated, predominantly glycogen repletion.  It may well be due to replenishing other "carb" molecules with their associated water as well.  My guess is that for the most part these changes show up on the scale more than in dimensions.  That was my experience during the almost 3 years of my "low carb cheating" plan.  I can't know for sure for the first year and a half or so of that, because I never weighed, but I didn't gain size during my cheats (pants still fit fine) and later after I weighed I sometimes saw a few pound gain, but again this didn't effect the fit of the skinny-jeans much if at all.  

However my current interest/focus is building on this recent post of mine:

Insulin, Weight Loss & Water Weight

In that post I referenced two studies that I'll repeat link to here:

Changes in abdominal subcutaneous fat water content with rapid weight loss and long-term weight maintenance in abdominally obese men and women - This one discusses how following rapid weight loss and improvement in insulin sensitivity of the fat tissue, fat mass increases due to increased associated water resulting from improved blood flow.

Disparate Hydration  in Adipose and Lean Tissue Require a New Model for Body Water Distribution in Man.  This one discusses various models of determining total body water content, and the contribution of adipose tissue often ignored for lean populations.  It also discusses the extra and intracellular distribution of water in adipose tissue and lean tissue.

So, adipose tissue is approximately 75% lipid and 14% associated water of which 11% is extracellular (between the cells) and 3% is intracellular (within the cells).  The remaining mass is presumably relatively constant comprised of the cell walls, connective tissue, organelles, etc.  The ECW:ICW ratio is 3.5:1

Non-adipose tissue is approximately 79% water, and in papers I've read on intramyocellular lipids (IMCL), lipid comprises between 1% "normally" and 2% when "elevated".   This leaves the rest comprising roughly 20%.  Interestingly, although by all indications AT water distribution is similar in lean vs. obese, it varies for AFM (adipose free mass).  In lean ECW:ICW = 0.42 while for obese it ranged from 0.49 to 0.99 averaging 0.72.  The obese tend to have higher levels of IMCL, the ratio of ECW:ICW correlated with degree of obesity, so it seems that increased IMCL content results in a water shift from inside to outside of the non-adipose tissue cells.  I am not sure the implications of this, but perhaps we need not even really bother because the associated water weight was not significantly different.  IOW, even if we double the lipid content of the cells, the mass of AFM will remain fairly constant.

So, a little math.   
Let’s say you have 75 g lipid in your fat cells, this represents 100 g adipose tissue mass (lipid content is 75% of total fat mass), and of that 100 g AT, 14% is water (14 g), 11% being extracellular (11 g), 3% intracellular (3 g).   So:  75 g lipid is associated with 14 g water in adipose tissue.  Or for every gram lipid released from an adipocyte, you "lose" almost 1.2 g "total weight" and "fat mass".

If the lipid is oxidized we can assume the water is eventually excreted.  But with reduced insulin (as in that DZ study) more lipid is released.  That study showed, however, that more was not oxidized.  So, it was taken up by AFM where the associated water is inconsequential.  Now for every gram lipid released you retain that gram of lipid, but "lose" almost 0.2 g "water weight".  

How does this match up for DZ vs. placebo in that study?  Well I can't do a total mathematical analysis, but insulin reduction did lead to greater FA release with a concurrent "freebie" loss of ~20% of the released mass.  That wouldn't seem to explain double the weight loss, but perhaps the "dehydrating" of the non-adipose cells could also lead to more losses.  And I'm not at all sure we could measure IMCL and extrapolate that to total AFM lipid mass.  Lastly, the hyperinsulinemic placebo group may have seen increased water content in AT as insulin sensitivity improved.    I would note that for whatever reason, the "after" placebo insulin levels still exceeded the "before" DZ group levels, despite being reduced by approximately half the absolute amount achieved by DZ treatment.  What this means?  Dunno.

Yes, dear readers, lots of ifs and perhaps here.  But I'll leave you with this.  If someone is 100 lbs overweight, they probably have about 80 lbs excess fat.  Of the 80 lbs excess AT, that's 60 lbs "fat" and about 11 lbs water.  Hormones, like insulin, CAN and DO influence the distribution of lipid stores in our tissues.  Lower insulin and higher fat consumption and circulating FFA's do shift some stores from AT to AFM with a resulting loss of associated water weight.

So ... there you have it folks.  A non-metabolic advantage!

It remains to be seen (and herein lie my concerns) whether:
  • A shift of lipid stores to ectopic (non-adipose) tissues is preferable, deleterious or innocuous
  • The "dehydrated state" is preferable, deleterious or innocuous
And it does appear, based on the weight loss trajectories of long term studies as well as anecdotal evidence, that once weight loss has occurred and the body stabilizes over a period of months and years, if this repartitioning persists or re-adjusts.  Also, since I can take a diuretic pill and shed water weight in a matter of hours, how much does this impact our total body water content in the long haul?


Paul Jaminet said…
Hi CarbSane,

That's a very interesting analysis. However I'm a bit confused by some of the numbers.

If water is 79% of adipose-free mass then why when fatty acids move to non-adipose tissue is the change in water weight inconsequential? You'd think water weight would increase as lipid mass increased.

The assumption of no change in AFM seems inconsistent with the hypothesis that fatty acids are migrating from AT to AFM with no change in fatty acid oxidation.

The idea that intramyocellular lipids increase extracellular water (and presumably glycative modification of membrane proteins) is new to me.

Like you I haven't seen evidence whether these water weight changes are preferable, deleterious, or innocuous. :) I think it could be a big issue!
CarbSane said…
I'll probably do one more post on this - when my head starts spinning who knows what comes out ;)

I have some thoughts but need a bit to gel them :D
LeonRover said…
WEIGHT Watching rather than tape-MEASURE watching: result of using a wrong "marker", like LDL (or insulin.

I put it down as another consequence of The Ideology of Nutritionism (Gyorgy Scrinis), and include the idea of the "Balanced Diet".

The correlation of the tape measure with fat weight or %fat is extra-ordinarily in-exact. The LEAST bad methods use calipers; while %H2O a la Tanita is a hoot!
mem said…
For me, experientially, there is no question that it changes. And it has always been this way for me, since I was a young woman. If I eat the SAD, I have varying amounts of edema at the ankles, and up to my lower shins. No, it is no usually huge, but if I press the tissue firmly it pits, quite obviously. And it comes within 8 hours or so of eating a standard carb meal and is worse if I've really gone high carb. Within 48 hours or less, when i return to VLC or LC, it completely disappears. Oh, and it has always affected my hands (fingers) as well. No, I don't look like the pillbury dough girl, but it is quite obviously there and if I get on the scale, it shows there too. Among much more important benefits, it was yet another that made LC the WOE for me. I've been a steady LC since 2000. I first did Atkins in the 70's for a couple years and should have never stopped!
CarbSane said…
@Leon, in the DZ study they did make the following statement:

the decrease in BF was significantly greater in the DZ-treated group, with significantly greater body water and FFM/BF ratio than in the placebo group. However, it should be pointed out that the use of bioelectrical impedance has methodological limitations during periods of changing fluid states and possible DZ-induced edema. Therefore, the estimated FFM and body water data should be interpreted cautiously.

I tend to agree with those who say "so what" when someone says "oh it's just water weight" past the glycogen associated water. One thing I noticed my first stint on Atkins was that my cellulite all but disappeared (TMI probably). I also have become denser and denser following 3 cycles of LC to where I'm now several pants sizes smaller than when I've weighed this much in the past -- we're talking 8/10 vs. 14/16! Yet my %BF is "off the chart" on one of the hand held ones. Things that make ya go hmmmmmm....

Weight is a very poor measure of "bodily state", and my only concern over it at this point in my life is that by BMI I am considered just south of obese so my medical chart doesn't look so hot.

@mem: Your experience is not unlike what I've seen many report. Obviously if your body reacts this way, carbs may not be good for you. I would wonder, however, if its a reaction to some carb (wheat?) that you're intolerant of. What was odd in the DZ study is that pharmaceutically lowered insulin led to edema in some which would be the opposite of higher insulin levels when consuming carbs. Things that make you go hmmmmmm...

Welcome to the Asylum!