Is Metformin the Fountain of Youth?
Is metformin the magic cure to slow the aging process?
Is metformin the Fountain of Youth? That is what Jill Crandall, MD, and Nir Barzilai, MD, of the Albert Einstein College of Medicine want to find out in their proposed study titled “Targeting/Taming Aging With Metformin” (TAME).
As a pharmacist, it’s my job to understand not just the benefits of a drug, but it’s limitations, as well. So, when I heard that metformin was at the center of the latest attention-grabbing headline, I was skeptical, but I wanted to learn more.
Metformin is an oral hypoglycemic agent that is considered the first-line oral treatment for type 2 diabetes.1 The United Kingdom Prospective Diabetes Study demonstrated that overweight patients on metformin had a lower risk of diabetes-related complications, less weight gain, and fewer hypoglycemic episodes compared with insulin and sulphonylureas.2 This finding, coupled with the drug’s safety record, has made metformin one of the most widely prescribed medications in the world.
In a recent interview about the TAME study, Dr. Barzilai explained how he hopes to examine the effects of metformin not in the context of targeting a specific disease, but of aging itself, the process that would make a person susceptible to disease.3 Regardless of whether metformin could be a magic bullet to slow aging, Dr. Barzilai stated that the true goal of the TAME study is to convince the FDA that aging is a legitimate target for research and investigation, which could lead to more investment in this type of research and potentially more future breakthroughs in this field.
Dr. Barzilai pointed to a number of studies that support the beneficial effects of metformin outside of those that are traditionally understood. He specifically highlighted 3 studies that have contributed to his concept of conducting a well-designed, FDA-supported study in humans to test metformin’s effects on aging.
Here are my summaries of those studies:
1. “Metformin retards aging in C. elegans by altering microbial folate and methionine metabolism.”4
The mechanism by which metformin extends lifespan in C. elegans is mediated by co-cultured E. coli where metformin inhibits bacterial folate and methionine metabolism.
This is a fascinating study to read for a number of reasons. The first is that it is a well-designed medical biology study that used worms as a human proxy to determine metformin's effect on lifespan via alteration of E. coli with the idea of possibly representing effects on the human intestinal microbiota. Through numerous experiments, the investigators attempted to elucidate the mechanism and circumstances under which exposure to metformin significantly increases the lifespan of worms.
The results showed that in the presence of E. coli, exposure to metformin at 25, 50, and 100 mM increased the lifespan of the worms by 18% (p<0.001), 36% (p<0.001), and 3% (p=0.002), respectively, in comparison to those with no metformin exposure. The study continued to examine the effects of alterations in the worm population, the E. coli colony, and drug exposure on worm lifespan, and the authors reached a conclusion that metformin’s effects on worm lifespan are caused by a complex effect on metabolism in the worm mediated by the co-cultured E. coli.
One finding with potential external significance was that when the worms were supplemented with glucose, metformin exposure did not extend worm lifespan.
From my point of view, this study has excellent internal validity, as the investigators were able to control a majority of factors that could affect the study populations. The authors did a great job at eliminating many potential extraneous causes for the effect through a number of novel experiments, and the results seem scientifically sound.
Of course, if we’re talking about metformin’s effect in humans, a study on worms has very poor external validity. Humans are many orders of magnitude more complex than worms cultured with a single strain of E. coli.
Nonetheless, the results are intriguing and do provide a better understanding of the basic biochemical effects of metformin that could be studied further.
2. “Metformin improves healthspan and lifespan of mice.”5
Middle-aged male mice given a standardized diet and supplemented with metformin for the rest of their lives will have increased health span and longevity in comparison to controls.
The investigators behind this study wanted to determine the effect of metformin on mouse lifespan and health span. Two genetic populations of healthy mice (C57BL/6 and B6C3F1) were given metformin 0.1% and 1% (w/w) supplemented into their otherwise-standardized diet. Metformin supplementation was started at 1 year of life and continued for the duration of the mouse’s life.
The results show that diet supplementation with 0.1% metformin led to a 5.83% extension in mean lifespan (2=5.46, p=0.02) in the C57BL/6 population. The B6C3F1 population demonstrated a trend towards increased mean lifespan that did not reach statistical significance (4.15% increase, 2=3.43, p=0.064).
C57BL/6 mice given 1% metformin showed a decreased lifespan of 14.4% (2=51,70, p<0.001), potentially caused by metformin toxicity. The investigators also looked at a number of indicators including weight, food intake, energy expenditure, physical health, and glucose tolerance and found results with varying statistical significance.
This study seemed to be designed well, at least from the standpoint of assessing the effect of metformin on the lifespan of mice. The methods were clearly delineated, and from my standpoint, confounding factors were well controlled using good standard laboratory practice.
Regarding external validity, this study used mammals as subjects, but the mouse model is by no means a perfect surrogate for assessing response in humans. Oftentimes, effects in simple models are not seen in more complex organisms, as is common when translating animal studies to humans.
Despite this, the effect of metformin on extending mouse lifespan in the C57BL/6 population of mice was demonstrated in this study, but was not reproduced to the same extent in the second population of mice.
3. “Can people with type 2 diabetes live longer than those without? A comparison of mortality in people initiated with metformin or sulphonylurea monotherapy and matched, non-diabetic controls.”6
Compare the risk of all-cause mortality associated with first-line metformin therapy or sulfonylurea therapy with that of matched non-diabetic controls.
This observational study used the Clinical Practice Research Datalink (CPRD) to attain a large volume of retrospective patient data aggregated from primary care providers in the United Kingdom. The CPRD is an anonymized database of UK public health records managed by the UK National Health Service that contains data from over 13 million patients.
By comparing survival in diabetic patients to non-diabetic matched controls, the investigators hoped to determine whether previously identified benefits of metformin would be seen independent of the detrimental effects of sulfonylureas.
Patients were included if they were diagnosed with type 2 diabetes and were initiated on either metformin or a sulfonylurea as first-line treatment and received a minimum 180 days of treatment. Patients were followed from the first metformin or sulfonylurea prescription until death or censorship (5 years plus 180 days).
Each case was matched to a control patient according to age at baseline, gender, general practice, cancer status, and smoking status. Matched controls were also indexed on the same date. Patients were excluded only if they had any record of secondary diabetes.
Numerous statistical tools were used to compare baseline characteristics and assess outcomes. Subgroup analyses were also conducted to allow for the impact of concomitant cardioprotective medication to be evaluated.
Comparing the metformin group to its control group, metformin patients had a higher body mass index, more contact with their doctor, more previous cardiac events, and were on more cardioprotective medications. Comparing the sulfonylurea group to its control group, sulfonylurea patients had more comorbidities, doctor contact, previous cardiac events, and were on more cardioprotective drugs.
The study identified 78,241 patients treated with metformin and 12,222 treated with sulfonylurea, and the same number of non-diabetic patients were matched to these cases for comparison.
In crude event rates, in metformin versus controls, there was a non-significant decrease in death rate of 14.4 versus 15.2 deaths per 1000 person-years, respectively (p=0.054). In sulfonylurea versus controls, there was a significant increase in death rate of 50.9 versus 28.7 deaths per 1000 person-years, respectively (p<0.001).
The KM curves show a small yet significant difference between metformin and control survival (p=0.037). Comparing the 2 control group KM curves shows an equal pattern of survival (p=0.879).
The final model adjusted for medication used for cardiovascular disease prophylaxis prior to index date showed a 15% increase in survival in the metformin group versus control subjects [Survival Time Ratio (STR)=0.85, CI 0.81-0.91] and a 38% decrease in survival in patients on metformin versus sulfonylureas (STR=0.62, CI 0.58-0.66).
As an observational study, a large group with little exclusion is a realistic representation of a population if the treatment being considered may be used across a wide and varied group. Broad inclusion of data also allows for detailed subgroup analyses later.
Matching of controls, however, was done on a very limited number of factors, and it is concerning that other medications were not considered in matching. Even in post-hoc analysis, only cardioprotective drugs were included in the analysis.
Selection of analytic methods for the various data types seemed appropriate to identify differences between group baseline characteristics and outcomes. The adjusted outcome presented as STR was an interesting choice and did add some complexity to understanding this outcome.
As expected with so few matching factors for the control group, there is a large amount of variability between the comparison groups. The large sample size of the study should have helped prevent skewing of results by outliers; however, it is difficult to attribute any benefit to the intervention when there are so many potential confounders unaccounted for.
Despite all of the limitations with this type of study, the results are still intriguing with a few caveats. The small survival benefit seen with metformin in comparison to controls was bordering on significance in 2 of the 3 analyses. Combine that with the numerous potential confounders limiting the effect that can actually be attributed to metformin.
In the final model, the authors made an effort with the available data to adjust for known confounders, and they were able to separate out a larger effect for metformin. They presented a final overall effect on STR as being significant; however, looking at each individual factor, a number of them (age <=53, never smoked, prior lipid-lowering therapy, prior antihypertensive therapy) crosses the line into nonsignificance.
Throughout the study, comparisons were also made between the metformin and sulfonylurea groups; however, the authors stated that this study was not designed to identify differences between metformin and sulfonylureas. The decreased survival seen with sulfonylureas versus controls is concerning, but again, these patients had worse baseline characteristics.
For what it is, this is a well-designed observational study, and the investigators made the best of the limitations imposed by this type of foundation. They did a good job at identifying limitations and outright stated that the study results only show that diabetic patients on metformin can expect their survival to be at least as good as the non-diabetic population.
If anything, this study raises more questions than answers, but the results do support the concept that more well-designed studies may be worth conducting to further elucidate metformin’s role in independently increasing survival in humans.
After reading each of these foundation studies, I’ll admit that I am as skeptical as ever. Blowing off the concept of treating aging with metformin as another pie-in-the-sky attempt at a miracle cure is the easy reaction that I’m sure many in the scientific community will have.
Looking at these 3 studies, however, it is difficult to completely ignore the results. Many attention-grabbing medical headlines are foundationless, but this one seems to be routed in actual science.
If the TAME study gets off the ground and the FDA gets on board, then the concept of directly targeting aging as the foundation for disease may be the next real breakthrough in the treatment of the human condition.
- Herman WH. Response to Comment on Inzucchi et al. Management of Hyperglycemia in Type 2 Diabetes, 2015: A Patient-Centered Approach. Update to a Position Statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care 2015;38:140-149. Diabetes Care. 2015;38(9):e143.
- Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes (UKPDS 34). UK Prospective Diabetes Study (UKPDS) Group. Lancet. 1998;352(9131):854-65.
- Dr. Nir Barzilai on the TAME Study. The Healthspan Campaign Newsletter. March 2015. http://www.healthspancampaign.org/2015/04/28/dr-nir-barzilai-on-the-tame-study/. Accessed online November 30, 2015.
- Cabreiro F, Au C, Leung KY, et al. Metformin retards aging in C. elegans by altering microbial folate and methionine metabolism. Cell. 2013;153(1):228-39.
- Martin-montalvo A, Mercken EM, Mitchell SJ, et al. Metformin improves healthspan and lifespan in mice. Nat Commun. 2013;4:2192.
- Bannister CA, Holden SE, Jenkins-jones S, et al. Can people with type 2 diabetes live longer than those without? A comparison of mortality in people initiated with metformin or sulphonylurea monotherapy and matched, non-diabetic controls. Diabetes Obes Metab. 2014;16(11):1165-73.