On 3/27/14 I sent the following letter to Shirley Wang pointing out the significant error she made in reporting the results of a new study that looked at the relationship between protein intake and mortality from cardiovascular disease and cancer. The main point of the article was that a recent study showed an increased risk of dying from cancer with high versus low levels of IGF-1 in normal aging adults. She then drew the conclusion that taking HGH for "anti-aging" could have the opposite effect. The problem with this is that the article said no such thing! It showed that people with higher protein intake, not IGF-1 level, who are 50-65 had a higher risk of dying than those with relatively lower protein intake. In the letter below I simply ask her to make this correction.
Ms. Wang:
I read your WSJ article on growth hormone and aging with interest. I am curious to know the citations supporting your comment “Studies show that lowering IGF-1 by 50% decreases cancer risk significantly and that increased levels of IGF-1 are linked with higher cancer risk.”
While I am aware there are observational studies correlating higher IGF-1 levels with increases in common cancers in humans, the increase is almost completely nullified by controlling for the pro-apoptotic molecule IGFBP-3 which is increased in parallel with IGF-1 when GHRT is given. I am not aware of human studies that show a decreased incidence of cancer in humans in whom IGF-1 levels have been experimentally lowered. If you are talking about animal models, then that should be made clearer to your readers. These transgenic rodent models generally have extremely low or high IGF-1 levels from birth so that body size is considerably lower or higher than normal aging rodents. The only studies that give GH to normal aging rodents show no deleterious effects and rejuvenation of aging gene expression for metabolism (Tollet-Egnell 2004)
I have read the March 2014 Cell Metabolism paper from Dr. Longo’s group you appear to be quoting, and cannot find support for your statement, “They found that in those aged 65 and older, people with higher IGF-1 levels showed a fourfold increased risk for cancer and 75% increase in overall mortality compared with those with lower levels.”
On page 408 they do state the following: ” Among those ages 50–65, higher protein levels were linked to significantly increased risks of all-cause and cancer mortality (Table 1). In this age range, subjects in the high protein group had a 74% increase in their relative risk of all-cause mortality (HR: 1.74; 95% CI: 1.02–2.97) and were more than four times as likely to die of cancer (HR: 4.33; 95% CI: 1.96–9.56) when compared to those in the low protein group.” But this statement is about protein intake, NOT IGF-1 levels. They did look to see IGF-1 level interacted with protein intake to increase or decrease all-cause mortality or cancer mortality and only found an interaction in the 50-65 year old group.
While Dr. Barzilai’s group did report some interesting findings about lower IGF-1 levels predicting incremental survival in exceptionally aged females (but not males), it seems to me that this is hardly strong enough evidence to support the statement that growth hormone replacement studies in older adults would not be ethical.
Best regards,
Joseph M. Raffaele, MD
I didn’t get a response and sent the following email when a half-correction was made.
Ms. Wang:
I see that you have corrected that age group in the statement about IGF-1 and mortality risk, but you have not corrected the substitution of IGF-1 for protein. This is an egregious error that really should be fixed. I see that it is probably the supplemental table 4 that you are referring to but the study did not show any increased risk of all-cause, cancer, or cardiovascular mortality by IGF-1 level alone and that should be retracted. What it showed was that within the younger age group, those with higher protein levels had their cancer mortality risk further increased as IGF-1 level increased.
The authors overstate their conclusions that IGF-1 is causal as they have no data to support this. The mouse data they offer is interesting, but mice have much higher cancer incidence/mortality than humans and the cardio-protective effect of GH through IGF-1 is likely to outweigh any potential increase in cancer risk.
Best regards,
Joseph M. Raffaele, MD
It is a shame that we cannot get a response or have this corrected!
3/31 update: Ms. Wang responds and I comment on her response.
Hi Dr. Raffaele, thank you very much for your email. I have checked your concerns with one of the authors and he says that the way that I have worded my statement – as a correlation, not as a causal statement – is accurate. If you believe the authors overstate their conclusion that may be a point best brought up with them directly, but from the perspective of communicating their point, I stand by my statement because the author said it was accurate.”
I sent the following response:
Ms. Wang:
Thank you very much for responding to my email. I understand that you’d prefer not to be in the middle of a medical dispute, but by being the reporter and not putting quotations around this statement, you are placing yourself in a position of responsibility. This is not a direct quote from Dr. Longo or any of the other authors, but your paraphrasing of the major conclusion of the paper. I am pointing out that the paper demonstrated no such thing regardless of whether the author believes your interpretation is “accurate.” Would the WSJ editorial staff think it is accurate to report that a company had been profitable over the last quarter because the CFO said so when the financial statements themselves didn’t support the claim? One must go to the primary report to be accurate, in this case the paper itself, to see if what the researcher is saying about his work is accurate.
A primary example of this is the one you bring up in your article. The authors of the WHI paper initially said there was an increase in breast cancer when in fact there wasn’t after proper statistical adjudication took place. Moreover, there was never an increase in breast cancer in the estrogen alone group: in fact, they found a decrease in breast cancer risk, yet the lay press ( except your colleague, Melinda Beck) continues to incorrectly report the findings to the detriment of millions of women. So, the WHI is not a cautionary tale, but rather another example of researchers with a particular bias misinterpreting their data and the lay press blithely accepting their pronouncements without digging deeper into the material.
I cannot see how your statement (even after the age correction) “They found that in those ages 50 to 65, people with higher IGF-1 levels showed a fourfold increased risk for cancer and 75% increase in overall mortality compared with those with lower levels” is anything but a misrepresentation of what the paper clearly stated.
“Among those ages 50–65, higher protein levels were linked to significantly increased risks of all-cause and cancer mortality (Table 1). In this age range, subjects in the high protein group had a 74% increase in their relative risk of all-cause mortality (HR: 1.74; 95% CI: 1.02–2.97) and were more than four times as likely to die of cancer (HR: 4.33; 95% CI: 1.96–9.56) when compared to those in the low protein group.” Simply put, ‘IGF-1’ was mistakenly substituted for ‘protein.’
What you didn’t report (perhaps because it was buried in the supplemental material online) is that in this same 50-65 year old age group, those in the moderate protein intake group actually had a trend toward a decrease in all cause mortality as their IGF-1 levels increased. The real take home message is that higher IGF-1 levels are only detrimental if you consume a diet high in animal protein. When they looked at plant-based protein, these interactions were no longer evident.
The findings from these papers add to our understanding of aging, diet, and hormones, but they are considerably more nuanced than what your article conveys and don’t support the inflammatory statements of Dr. Longo.
Comments