Keto & Cancer -- False Hopes?

I have known a number of people personally, including family members, that have had cancer.  I remember a high school friend's mother dying of lung cancer.  I remember a high school class and teammate being diagnosed with pituitary cancer.  Obviously none of these people are me, but I've had a taste from a fairly young age of the disease and what people go through "up close and personal" from early on.  

Although no medical treatment is fool proof or works 100% of the time, when, in my 20's, I was working alongside folks with kids in day care, I contracted strep throat quite a few times.  I would go to the doc, they would do the swipe test, and within a short time I would walk out with my prescription for an antibiotic, I'd fill the script, take it, and the infection would go away.

Cancer is not like that.  It's different in different organs, there are different types in the same organs/tissues, there's metastasis, different growth rates, responsiveness to hormones, etc.  Indeed dysregulated cell division = cancer seems to be where any similarities seem to end.  I took a course specifically on cancer one summer in college.  That was the 80's.  We know soooo much more now.  Contrary to what the anti-conventional medicine establishment will tell you, great advances have been made.  Do people still get cancer?  Of course.  Do some die?  Of course.  Do more live longer and/or with less suffering than used to?  It would be hard to argue against a yes there as well.

Yet even if a doctor has had amazing success with a high percentage of his patients using a particular therapy for your type of cancer, they can make nowhere near the "guarantee" that the doctor who prescribed me antibiotics for my strep.  And so you have people like my mother, who 21 years ago had her doctor find a lump, underwent a lumpectomy, a course of chemo and some radiation, and has been cancer free ever since.  And then you have my former neighbor (because we moved) who is now undergoing her fourth bout with recurring lung cancer in about the same time frame, but still fighting the good fight.  And you sadly have my friend's mother who was lively and healthy one Labor Day and gone by Christmas.   Cancer is more multifactorial and unpredictable than obesity.  It's way more of a bitch.  It's way more exasperatingly complicated and beyond knowing completely.

I firmly believe that diet and lifestyle influence health.  I firmly believe there are limits to the degrees to which they do, however.  If anyone reaches the age of, say, 25 and they haven't met someone with a shitty lifestyle who is 80 years old and going strong, and another person who has walked the straight and narrow health path who died young anyway, you probably need to get out more.  The examples are all around us, even if you haven't met such a person directly.   If that isn't a reality gut check ... well ... it ought to be. 

I didn't mean for this to be quite so long a set up, but lately it seems that the volume has been turned waaaaayyyyyy up blaming carbs and sugar for everything.  I'm not going to address the wrinkles, gray hair, diabetes, obesity or whatever other aging process or ailment they attribute to sugar today.  No, I'm going to deal with the sugar causes cancer thing.  Gary Taubes has stated that if he gets cancer he'll really be pissed off because he is convinced carb are the cause, but especially sugar.  It was to be his next book for years now.  Wonder what happened to that?  As with just about every other metabolic thing he takes out of context, the role of glucose and insulin in cancer was ripe for the picking though I'm not exactly sure how fructose figures in.   

From:  Cancer Cell Metabolism: Warburg and Beyond 

It is hard to begin a discussion of cancer cell metabolism without first mentioning Otto Warburg. A pioneer in the study of respiration, Warburg made a striking discovery in the 1920s. He found that, even in the presence of ample oxygen, cancer cells prefer to metabolize glucose by glycolysis, a seeming paradox as glycolysis, when compared to oxidative phosphorylation, is a less efficient pathway for producing ATP (Warburg, 1956). The Warburg effect has since been demonstrated in different types of tumors and the concomitant increase in glucose uptake has been exploited clinically for the detection of tumors by fluorodeoxyglucose positron emission tomography (FDG-PET).  Although aerobic glycolysis has now been generally accepted as a metabolic hallmark of cancer, its causal relationship with cancer progression is still unclear.  In this Essay, we discuss the possible drivers, advantages, and potential liabilities of the altered metabolism of cancer cells (Figure 1). Although our emphasis on the Warburg effect reflects the focus of the field, we would also like to encourage a broader approach to the study of cancer metabolism that takes into account the contributions of all interconnected small molecule pathways of the cell.

Again ... this isn't a post about cancer cell metabolism per se.  The thing about conventional chemotherapy and radiation is that these target cell division.  Few normal cells divide much in the formed human ... even the infant, though there are more obviously until adulthood.  The defining characteristic of cancers are dysregulated cell growth and de-differentiation.  Normally your cells each express a certain phenotype -- that is they have certain shapes and functions that make them skin cells, muscle cells, fat cells, etc.  Many of these cells share common precursor cell lines and differentiate (develop into) specialized adult functional cells.  These cells recognize other cells and act in concert and know when to stop growing and dividing.  Cancer cells, in addition to going haywire in the growing and dividing department also tend to lose their cellular identity of "whatever" cells they are supposed to be.  Thus not only are they hogging nutrition from other cells they aren't doing the job of the cells they are overrunning and replacing.

The low carb hypothesis, based onWarburg, is that through diet one can starve the cancer of glucose, it will not progress.  Enter Eugene Fine, who presented at the recent American Society of Bariatric Physicians CONference in the Nutrition & Metabolism Society Symposium portion.  While cancer risk is heightened by obesity, the relevance of this research at such a conference seems suspect to me.  But the only research on cancer and keto diets that I could find published by Fine is this study:  Targeting insulin inhibition as a metabolic therapy in advanced cancer: A pilot safety and feasibility dietary trial in 10 patients (full text PDF is available here at time of writing this).  It is of note that the Atkins Foundation funded this study in part.

He discussed this in the talk at the Ancestral Health Symposium 2012.

A couple points:

• Not all cancers are dependent solely on glucose
• Physiologic BG doesn't vary much so cancers need to be good at getting glucose "which they are"  (this article on GLUT transporters might be of interest here)
• Metformin upregulates AMPK (you might find this review on AMPK and obesity and diabetes interesting)
• Lots of signalling going on (you don't say!)

• One trial targeting ↓mTOR only worked in 4 of 55 subjects because hyperglycemia ensued.  Coupled with metformin to counter this in a small pilot, efficacy improved to 6 of 10.  
• All of these are "downstream of the insulin receptor" (I may try and follow up on this at some point.  Interestingly the problems are associated with "insulin resistance" which is quite different from insulin secretion as I've blogged on here frequently.)

• Why not reduce insulin secretion with diet?  
• Low carb diets are safe
• Our evolutionary heritage is low carb 
• Mouse tumor models and cell culture lines it's been shown that insulin suppression have shown promise.
• In ONE study on TWO children with brain cancer, ONE child continued the diet for one year without disease progression.    
• So Fine and colleagues did a study ...

The study involved 10 advanced cancer patients who have had chemo previously, most multiple times.  We're dealing with some pretty sick people here, so must keep that in mind, but on the other hand I imagine this is the only sort of population in which such a study could be done.  The shortest duration of diagnosis was 2 years ad extended up to 14 (see paper) with most being around 5 years into their cancer diagnosis.  The study lasted only 4 weeks.  

Now ... you'll have to click on the table below to enlarge to read it.  I compiled this from merging various details from the tables. 

The first thing that bothered me about Fine et.al.'s analysis was that they lumped SD = stable disease with PR = partial remission.  It would appear from the caption on Table 3 in the paper that SD (NC) NC = "no change" = <10 change.="" nbsp="" p="">

Of note:

• 1 Subject experienced 16% remission
• 4 Subjects experienced no change 
• 1 Subject experienced < 25% progression 
• 2 Subjects experienced >25% progression
• 2 Subjects experienced >30% progression

This amounts to 50% experiencing more than 10% progression of disease during the ketogenic diet.  Only one subject experienced remission.  Why didn't they just report the % change in the PET scan?  This raises suspicions, and I can only surmise that none of the SD(NC) subjects experienced nominal remission but rather these all had 0 to 10% progression.  Also, if 16% is designated as partial remission, then why isn't < 25% designated as a progression?   Given the breakdowns, I suspect that this subject had between 20 and 25% progression or the change would have been designated as < 20% or < 15%.

In view of this they are essentially comparing greater vs. lesser averaged progression of the disease here, but didn't even break that up into 50/50, rather 60% less progressed vs. 40% more.  

Then there's the ketosis.  Expressed as a ratio it is more impressive than the actual levels, although most were clearly in ketosis.  As a ratio two subjects exceeded 20-fold, but "extreme" ketosis levels over 1.5 mmolar were only attained by one subject according to Fig.1.  Unfortunately, we are not provided with the absolute BHB levels before and after the diet.  I cannot understand why since they make a big deal about ketosis being a key factor in their discussion and they provided so much ancillary data such as creatine levels (that I omitted from my table).  

Breaking up 60/40 and using the BHB ratio, we get "A" below.  "B" is described as such in the legend:  Ketonemia versus insulinemia: the lowest insulinemia correlated with the highest ketonemia levels, as physiologically expected. Uniquely colored symbols represent values for each patient (numbered as in Table 3).

I can only presume the missing numbers 1&2 are over-ridden by data points, but can YOU figure out what's going on there?  I see far more than 10 points.  I would also note that this must be fasting insulin which everyone should know by now is unrelated to postprandial insulin.  Where did #10 up there come from?  BHB ratio over 30?  This is very confusing, and I can only speculate that it is deliberately so.

Rather than breaking the small number of data points into arbitrary groups, how about instead let's look at a few more scatter plots.  I took the liberty of playing with my spreadsheet and generating a few.  Note that the x-axis of all of these is % progression where I generously plotted the < 25% point at 15% progression and the progression values at exactly 25 and 30% though they are obviously greater.

While one could envision a very loose negative correlation between ketosis and disease progression, this is not nearly -- not even close! -- as impressive as when that dot up there between 10 and 20 is "itemized" out.  How about glucose?

I'd say that if any trend can be established in the Final FBG vs. DP, we are essentially flat lined, and if there's anything to the somewhat negative correlation on with Change in FBG, it is not in the direction desired.  Interestingly only 4 (two overlap) subjects had a reduced FBG.  It's really in a range of meaningless anyway if you ask me!  So glucose is a bust.  Did you notice the calorie reduction?  I mean you have only two subjects over "starvation" levels, 7 under 1200 cal/day with another under 1300 cal/day.  But let's look at these as scatter plots vs. the "wash" presented by the authors in grouping them.

Now, the trends aren't very strong and they are somewhat in opposition.  Uh oh, CICO is in trouble.  Yeah right.

Calorie intake vs. progression flatlines.  Important?  Nope.  Stick with me, I'm going somewhere here.  How about the macros?  Wow, the one who went into minor remission ate the highest percent carbs.   

But wait, on absolute intake ...     

Oh nevermind, the fewer carbs the more the progression.  Weak, but hey, I'm only following the lead of Fine et.al. in pointing such out.  How about fat?   Phew.  At least this flat lined.  High fat for the win!! (I would note that most were consuming rather low fat levels.  Not sure where y-axis went but these start at 0 grams and are 20g per division up to 180g max.)

Protein flatlined as well, what with all the mTOR stuff, as did carb + protein.

How about insulin levels or change in insulin levels?  Here are those plots.

Nope, this time there's no anomaly in my graphing here.  There is NO DATA.  I looked and looked ... I couldn't find the data on insulin secretion.  Not basal, not postprandial, not any attempt at a 24 hour AUC.  No c-peptide (insulin secretion).  No raw data.  None.  But ... we need to get something out of this! Before we go on, let me remind you of the title of this study/paper:

Targeting insulin inhibition as a metabolic therapy in advanced cancer.

What the H-E-double toothpicks is going on over at the Elsevier Nutrition Journal?  Who is peer reviewing this stuff?  I cannot imagine that something so basic as this would pass even a cursory peer review.  I'm reminded that the debacle of thermodynamic mangling that also bears the name Fine, and includes the name Feinman, was also published in this journal.  In the abstract they even state:

Insulin inhibition, effected by a supervised carbohydrate dietary restriction (5% of total kilocalories), was monitored for macronutrient intake, body weight, serum electrolytes, b-hydroxybutyrate, insulin, and insulin-like growth factors-1 and -2.

It's pretty unconscionable that both BHB and insulin levels were not reported before and after.  This is needlessly suspicion-raising if there's some innocent reasoning for the omission!!  Instead we get:

How about give us the insulin levels (not necessarily synonymous with insulin secretion, by the way)?  They were measured.  Incidentally while the relationship with ketones was an inverse one, the relationship with glucose levels is a direct one, albeit not extremely "significant" physiologically.

"Decreases in insulin by 75% to 90% compared with baseline values were seen only in patients with a 10- to 35-fold increase in ketosis. " 

I've labeled the subjects now on the keto-ratio plot that experienced 10-35 fold increase in ketosis.  I'm still not getting where the 35 is coming from (?? help!!) but using the data provided, as with the distribution of ketosis in terms of progression, so, too, we see varying disease progression amongst those in which we are to presume they saw the greatest decrease in insulin (basal?) levels.

I would encourage anyone interested further to read the paper themselves.  The authors do mention the calorie restriction and how cancer risk is reduced by calorie restriction in the 30-40% range.  I'm surprised there's no mention of the absolute calorie levels being extremely low.  It would seem to me that such low intakes were worthy of mention and even alarm given the wasting (being cachectic) that can occur in cancer.  Well, they do mention this in limitations:

... not all patients with advanced cancer would be appropriate for this approach because of comorbid medical conditions or general frailty, and these results cannot be extrapolated to patients who are cachectic without further study.

... It is important to note that all 10 study participants spontaneously decreased their caloric intakes, nine of whom lost weight, despite our best efforts to maintain a stable weight by encouraging increased food consumption. Participants showed a mean 35% caloric deficit and a 4% weight loss, raising the question of whether caloric restriction played a role in our findings. 

As they wrap up:

The extent of the metabolic response of subjects was consistent with the expected effects of insulin inhibition with our hypothesis ...

Sorry, I'm not seeing that.  They may not be inconsistent, but absent the simple straightforward measurement and analysis of insulin secretion, and given the scattered responses seen when "itemized" rather than lumped in arbitrary groupings, this statement is irresponsible.

It is essential to unravel the mechanisms of CHO restriction through further in vitro and in vivo investigations and to clarify the extent to which caloric restriction and CHO restriction are related or independent effects.

Funny, those accusing others of grant whoring are oddly silent ...

Insulin inhibition effected by dietary CHO restriction was found safe and feasible in 10 patients with advanced cancer. The three-fold higher ketosis, demonstrated in patients with SD or PR compared with those with continued PD, must be interpreted cautiously in this small pilot study. 

I'll say.