Diazoxide, Insulin & Obesity

In his recent post on insulin and obesity, Fat Tissue Insulin Sensitivity and Obesity , Stephan brings up one of those *nagging* studies that have been touted by TWICHOO advocates for quite some time.

Back in May 2008, Peter/Hyperlipid posted Weight loss when it's hard 2. Diazoxide.  This discusses the following study:  Beneficial Effect of Diazoxide in Obese Hyperinsulinemic Adults.

What is diazoxide?  It is a drug that reduces insulin secretion.  As Peter writes:
You can simply reduce insulin secretion using diazoxide. Find enough obese people willing to put up with the hunger generating regime supplied by Slimfast and semi starve them for 8 weeks. Half can have a placebo, half get diazoxide in addition to Slimfast starvation. Figure 1 in the results is where you want to look. The idle porkers starving on Slimfast plus diazoxide lost significantly more weight in 8 weeks than the idle porkers starving on Slimfast alone.

Figure 1 there sure looks like if we can just lower that insulin, that is the answer!  Interestingly, nobody seemed all that concerned about investigator bias when reading the following in the abstract:
Assuming that hyperinsulinemia plays a major role in the development of human obesity, then its reversal should have therapeutic potential.
So I would also ask,  if the above statement is true, then why aren't pharmaceutical companies clamoring to develop drugs that suppress basal insulin secretion in the hyperinsulinemic?  Could it be that diazoxide has been around since the mid 90's and people taking it have not been miraculously reverted to lean mean fighting machines?  Perhaps it's just the side effects that has prevented the obese from learning this "secret" and flocking to their MD's for a script?  Again, Big Pharma would be all over this like butter on a certain low carber's veggies -- to tweak and find the formulation that minimized side effects to within acceptable levels.  Let's face it.  If millions are willing to risk sure-to-be embarrassing fecal accidents to realize the somewhat minor benefits of Orlistat (Alli), they will endure what needs be for this miracle!  

I had blogged on this a while back:  Insulin, Weight Loss & Water Weight.  The arguments put forth in that post boil down to the following:
  • The differential in fat mass and total weight loss between the DZ v. placebo groups was not accompanied by changes in REE or either glucose or fat oxidation rates.  The DZ group did not burn more fat.
  • Therefore, the fatty acids were still in the DZ group
  • Water weight associated with where the fatty acids were stored may explain these results.  
Points 1&2 are indisputable.   Although they did not report the substrate use rates and this does seem a little suspicious ... perhaps b/c these results trended towards increased glucose oxidation? < /rank speculation > Point 3 is just a "thinking outside the box" theory of mine based on information outlined in that prior post, that I'll not rehash here.  

The theory on how low carb diets work is:
 reduced insulin  increased lipolysis/mobilization  increased fat burning  weight/fat loss

Let's call this TWICHOOAROO (OMG she's getting out of hand with these acronyms.  Someone please stop her!! -- the AROO stands for And Reversal Of Obesity).  This study is yet another touted by low carbers as appearing to support the above.  Pretty convincingly at that ... until the authors point out specifically that the bolded blue arrow in the progression above was not observed.  As such, I'll add Alemzadah et.al.to my list of studies, headed by Grey & Kipnis, purported to support the TWICHOOAROO that actually refute it.  As I've said here many times before, lipolysis is not fat burning.  It is the first step, while beta-oxidation in the cells is the actual fat burning.  As Gary himself stated in that snippet of his interview (FF past 40 min mark) of a little over a year ago I blogged on here, we are constantly, thin or fat, releasing an excess of NEFA from our fat cells, how much gets utilized for energy is dictated by our needs.  Lipolysis is NOT rate limiting.  It doesn't govern weight loss.

So, back to our study.  If not increased metabolism, how to explain the results here?  As postulated previously, it is plausible that, although %water weight apparently increased for the DZ group, and some developed edema, these subjects could still have lost water weight in the form of extracellular fluid in the fat tissue as their NEFA were released and stored elsewhere (or remained in circulation).   Basically a shift of fatty acids from fat cells (where they are associated with an extracellular/intracellular fluid balance of 3.5X) to muscle cells (EC/IC <1) likely resulted in the release of more extracellular fluid from the DZ group (some excreted, some pooled).

Or ... the results could be nothing more than an aberration from the method used to measure body composition:  This little caveat from the study was glossed over by enthused low carb advocates:
Also, 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.
Now yes there appears to be a newcomer to policing the blogosphere running about trying to strike up a chorus of Kumbaya! and chastising any one blogger who dares address the writings of another, especially critically (praise and kudos are allowed).  But,  at the risk of running afoul of Low Carb Internet Kindergarten Cop III, I'm going to address Peter's post directly.  First, folks often criticize my demeanor and writing style.  Which is certainly fair, but the double standard is breathtaking at times.  Take for example how Peter mocks any who believe that the results of this study must be explained by the laws of thermodynamics.  This was a free-living study, not a metabolic ward study so he sarcastically addresses cheating, underreporting and exercise by proposing that it would have to be the DZ group who exercised more despite admonitions to maintain normal activity, and the placebos were haunting the donut shops daily.
Or maybe, just maybe, you could just accept that in the real world, outside a meatball ward, sorry, metabolic ward, the level of insulin in your blood stream influences your rate of weight loss, specifically your rate of fat loss.
That's right folks, in the "real world" insulin possesses magical properties.  In the meat ball ward (populated by social misfits per Eades), rather in the metabolic ward, where other variables can be tightly controlled for and monitored, the magical properties vanish.

Secondly, let's address his overall analysis of the study.  Perhaps Peter got a bit carried away trying to be sarcastic about all of this, but in doing so he misrepresents, rather badly IMO, the study design.  He makes it sound as if all of the subjects were on a "Slimfast" diet for 8 weeks.  Either he misread, or he misrepresented this, as well as glossing over issues with the study design, how dietary intake was assessed specifically.  Here was the study design:
  • To establish baseline caloric intake of the subjects prior to the study, a single 24 hour dietary recall was used.  I don't suppose that with a sample size as small as 12 subjects per group, such an assessment could have thrown things off just a bit?  Recall in Grey & Kipnis that caloric intake for the women in that study varied by 2100 calories between subjects (from 2100 to 4200).  Might one be just a tad bit suspicious of under-reporting in this study when both groups reported an average of less than 2000 calories and a standard error under 200 calories?   It would certainly not be unfathomable for one or even two gross underestimators to be randomized to the DZ group.
  • On the graph, the week labeled "lead in period" is actually the first week of the study where all subjects were switched to a hypocaloric meal replacement diet, "designed to exclude individuals who were unlikely to be compliant 
  • and to ensure 
  • stable body weight before randomization".  On
  • e problem with this that I see, is that hypocaloric (weight loss) and stable body weight seem to be mutually exclusive goals of this phase.  They did indeed lose a bit of weight during this week. 
  • The lead in week supplied 1260 kcal/day for women, and 1570 kcal/day for men.  For the first 6 days this was consumed in the form of shakes and bars.  On the 7th day, they were switched to an Optimeal  plan.  Thus the subsequent 8 weeks of the study the subjects were not starving on Slimfast, a more homogeneous and easier way to control intake, they were semi-starving on a presumably far more varied diet of prepackaged "real" foods.  
  • How was intake for the 8 weeks assessed?  Well, on the aforementioned "day 7" they were instructed to keep a food diary "to ensure compliance with their meal plan and were requested not to change exercise or activity level."   
To recap:  A single 24-hour recall was used to assess baseline caloric intake -- an assessment method that is fraught with error.  Then everyone got a liquid diet and bars for 6 days.  And on the 7th day they got to choose between some regular foods.  Based on a single day's choices and accounting of that, we are to presume all subjects maintained the prescribed caloric intake for the next 8 weeks.  OK.  Welcome to the real world of science I suppose.  Seriously, if you really wanted to slam scientists it would be for laziness in study design where cost would not be a major factor.  Why not have subjects keep food diaries for a week prior to the study.  Not perfect, but averaging several days of food logging has got to trump a single 24 hour dietary recall for accuracy!  How about having subjects log food on the 7th day of each week of the study to assure compliance?  It's this sort of thing that drives me nuts about these studies.  But you see, if you get the results supporting your beliefs, the transgressions are easier to overlook.

Now clearly by this statement of Peter's:
BTW The folks in this study are clearly pathological liars too. Any trainer being told that a fatty can maintain a BMI >30 on 1900kcal per day will be snorting in their whey protein shake. You know calories in calories out, yawn.
... we see that he's too caught up guffawing over his witty writing to realize that, yeah, this caloric intake for a group averaging 250 lbs, BMI around 40, does seem almost unbelievably low.  But that is what it is and open to interpretation I suppose.  What is not open to interpretation is where he takes the post next:
... I think both groups ate just 1570kcal/d and skipped the gym.  The diazoxide group used all of those calories PLUS 1660kcal/d of their own fatty tissue.  The placebo group appeared to use that 1570kcal plus 600kcal of their own fat, but this is an average over time. They lost more than this initially per day, but by the end of the study they didn't look to be loosing any weight at all on those 1570kcal.  By week eight they USED 1570kcal/d and that was it. 
That amounts to the DZ group having daily caloric expenditures of  3230 calories!   OK, I'm about to nit pick a bit first, but if you're going to so the sort of analysis that Peter does at least make an attempt at accuracy.  Each group consisted of 10 women and 2 men so the average caloric intake was 1312 kcal/day.  So looking at the DZ group, we have the following "facts":
  • At baseline they consumed 1977 kcal/day
  • For 8 weeks they "consumed" 1660 kcal/day of their own fat tissue 
  • For 8 weeks their (corrected) dietary consumption was 1312 kcal/day.  
  • Thus, for 8 weeks their total daily expenditure was 2972 kcal/day. 
Something is not adding up for me here.  Because Peter is expecting his readers to believe that by dropping insulin levels, in the face of calorie restriction that if anything will tend to reduce metabolic rate over time,  our 12 diazoxide folks experienced a 50% increase in energy expenditure over baseline.  [ (2972-1977)/1977]  That's pretty darned amazing eh?  All of this with the investigators reporting:
There were no significant changes in the REE values of either group. Substrate use, as derived from indirect calorimetry, did not reveal any significant change in carbohydrate or fat metabolism in either group (data not shown).
Let's do a bit more math.  Let's say that at baseline the subjects were eating the usual diet of obese people, roughly 48% carb, 38% fat and 14% protein (averaged from G&K).  And let's say they were weight stable obese at 1977 cal/day.  This means that before the study they are getting roughly 950 cals from carb and 750 cals from fat.  And I will cop to laziness in figuring out the macro breakdown of the study diet, but let's say it's a relatively standard LF formulation (65C/20F/15P), our subjects are now getting roughly 850 cals from carb and 260 from dietary fat.  Add in the whopping 1660 cal/day they are cannibalizing from fat stores and the DZ group is "consuming" a diet that is 29% carb, 65% fat and 6% protein.  According to everything we're supposed to believe about insulin altering substrate usage and such ... even with this astronomical shift in what Peter contends the body tissues of the DZ group consumed ... how can this jive with the above quoted finding in the study??

It can't.  For the subjects to burn 50% more calories than baseline, you would have to see a significant uptick in REE.  It was virtually a wash.  Not possible were Peter's scenario to be even remotely plausible.  And going from carb/fat ratio of roughly 1.27 at baseline to 0.44 we would expect some demonstrable difference in substrate usage/oxidation/etc. -- especially if these folks were really burning 1660 cal/day of their own body fat!  Yet we do not.

There's one last slight confounder here as pertains insulin's possible role.  The placebo controls started out considerably more hyperinsulinemic than the DZ group.  They went from 204 to 174 pmol/L while DZ had double the absolute reduction in insulin levels from 168 to 108 pmol/L.  IF insulin is impeding lipolysis is impeding fat burning, then the placebo group started out with a distinct disadvantage.   I don't think this is all that important, but since this is what the study was looking ...

In wrapping it up the investigators say this:
This preliminary study demonstrates that DZ-induced attenuation of hyperinsulinemia in obese individuals results in a significant decrease in BF, presumably caused by a decrease in lipogenesis.
By this I think they mean esterification/fat deposition in fat cells.  Not increased mobilization oddly enough.  Which means that fat went somewhere else in the subjects' bodies because it didn't get oxidized.  I suspect some remained in circulation, the rest found its way into lipid droplets in ectopic tissues.

Back to Big Pharma and the implications of studies like this.  If there were anything to this we'd have seen it by now.  It's been over a decade since these results were published in 1998, and this was a drug already commercially available -- one need not even have subjected themselves for "lab rat duty" in a clinical trial to  access it.   It was already FDA approved and we had these results amidst the burgeoning obesity epidemic.  Don't you think we would see more of it in use to combat obesity?  The comments section is interesting in this regard!


Duffy Pratt said…
Good post. I don't agree with your assessment that the big drug companies would be all over this if it worked, for two reasons. First, nothing would prevent them from trying to sell something like this even if it didn't work. That's what they have done with round after round of new blood pressure meds that end up not working any better than your basic diuretic.

More importantly, however, diazoxide is already generic. That means the big money pot has already vanished. So, even if it were effective, that doesn't mean that the drug companies would be all over it, unless of course they could tweak it to secure a monopoly.
Sanjeev said…
> in the "real world" insulin possesses magical properties

click or manually:

Duffy Pratt said…
Nice Sanjeev. Here's my favorite variation on the same theme:

I've been on Weight Watchers (successfully) for a while now, and agree with what you said about inaccuracies in tracking your food intake. I'm avid (okay, fanatic) about doing it everyday, and if I don't stay on top of it throughout the day - ie. if I had to do it from recall over the last 24 hours, I would definitely get it wrong & forget stuff, ie. underreport (and I'm in practice). Also, I have a question: when you mentioned that the fat on the people in this study ended up where it otherwise really shouldn't be, in ectopic tissue, is this similar to what happens after liposuction when people gain all the fat mass back? Isn't that incredibly dangerous to one's health? Weight Watchers may be a bit old fashioned (and no, I definitely don't believe their schtick that you can eat "anything" you want - I think that's incorrect & agree with Stephan G about food reward), but at least it keeps me on track eating real food in calorically-sensible quantities...and lets my insulin work its "magic".
John said…
Regarding what Duffy said, do we really have to resort to arguments like that?...C'mon, that won't sway anyone.

I don't see how your two "indisputable" points are so. This whole thing is a little dicey and revolves around a statement for which "data not shown." I don't know what to think about that since, for REE in table 2, as far as I know, 1926-1911 does not equal -65. What's going on?

You said, "According to everything we're supposed to believe about insulin altering substrate usage and such ... "

Are you saying that insulin does not affect substrate usage?

What is preventing us from hypothesizing the Diazoxide group ate fewer calories over the weeks? This is a strong point with many of the insulin hypothesis proponents.
Frank said…
I have not looked at both paper in depth but what does Peter says of that paper

No effect of inhibition of insulin secretion by diazoxide on weight loss in hyperinsulinaemic obese subjects during an 8-week weight-loss diet. (2007)

CarbSane said…
Hi john: I've had quite a few here so not sure if you've posted before. I don't recall a "small j" john with a picture before, so if you're new here ... welcome to the Asylum!

Here's what they said about REE and substrate oxidation:

Resting energy expenditure (REE) was measured by indirect calorimetry, using a Deltatrac Metabolic Monitor, after an overnight 12-h fast, with subjects lying supine for a period of 30 min. Urine was collected over the corresponding 24 h, for measurement of total nitrogen and determination of substrate use, before and after the study.

Perhaps not the best wording on my part, but I do not consider those two results in "dispute". I speculated as to why they didn't include the ox data ... there are length restrictions for many publications so it may have been that. I was only being a little silly wondering if they perhaps even favored glucose. My water weight explanation is speculation.

Yes there's an error on the REE, but it is inconsequential. The title and several points made by the authors seem to convey a bias towards diazoxide being effective for weight loss, but if there was actually increased fat burning going on it would have been reflected in REE.

Eating less? If Peter's cals are correct, the DZ group would have had to eat like 300 cal/day spontaneously. Not likely. It might be possible that more DZ group stuck to the diet more consistently.

Insulin IS the substrate cop, but postprandially that's dictated by the composition of meals, not basal levels per se. If insulin were doing it's job locking fat away, then when it's released in greater numbers increased substrate should increase substrate usage if basal insulin per se were the determinant factor.

I'm going to be bumping an old post of mine from back when I had like 2 readers about energy status dictating substrate usage not carb content of diet when I get to Parts II & III on the formerly obese series.

@Frank: I just got the full text on that one, I'd seen it before but forgot about it.
CarbSane said…
@john: I would add that the study Frank linked to, even in the abstract, states DZ didn't alter appetite in that study.
John said…
Yea, I think Frank's is the same one that Stephan referenced.
Tonus said…
@Duffy: "More importantly, however, diazoxide is already generic. That means the big money pot has already vanished."

Not really. The big money pot would still be there if diazoxide could be effectively marketed as a miracle weight loss drug. It would simply be a bit more spread out. If there were any truth to its weight-controlling potential, the market for it would be very large, and I can't imagine that NO ONE would be willing to jump on it out of a concern that they wouldn't be able to monopolize it.

As you point out, drug firms will work hard to promote drugs that are ineffective (and sometimes dangerous). If this drug worked, I think that they'd be marketing it for all that they are worth.
Frosty said…
I wish I knew who the various Low Carb Internet Kindergarten Cops were so I could join in the lulz. :(
CarbSane said…
Hi Frosty, LCIKC1 is the notorious Fred Hahn. No. 2 was Kurt Harris, though he seems to have relinquished his badge and therefore I am moving on from that. I'm pretty sure I've identified these here a few times, but that doesn't mean every reader should have read or recalled that. Still, this requires a 3 for the newbie, someone who commented various places named "Steve". He earned that title for this comment. I don't know who this is but I have an idea of a relatively obscure blogger who fits the bill. :-)
Tonus said…
I thought that guy was Rodney King!
M. said…
An example of Fred Hahn in all of his glory in the comments at http://blog.cholesterol-and-health.com/2011/08/against-dietary-dogmatism.html
CarbSane said…
Another thought re:generic. Aspirin has been generic for decades, and over the counter. Hasn't stopped the drug companies from selling it in different forms at inflated prices. And people buy it even though there are "no name" and "store brands" that are likely made right alongside them in the same facilities sold for a fraction of the cost.