Diabetes Care Volume 42, April 2019 499

COMMENTARY
William T. Cefalu1 and Matthew C. Riddle2
More Evidence for a Prevention-
Related Indication for Metformin:
Let the Arguments Resume!
Diabetes Care 2019;42:499–501 | https://doi.org/10.2337/dci18-0062

Downloaded from http://diabetesjournals.org/care/article-pdf/42/4/499/550658/dci180062.pdf by guest on 16 February 2025

Global estimates of the number of in- (3). The growing personal burdens of di- lifestyle efforts, for preventing emer-
dividuals with diabetes are staggering, abetes are not acceptable given our re- gence of overt type 2 diabetes. In par-
and the prevalence is reported to be search advances, and the costs of caring ticular, there is clearly a need to develop
increasing in most countries. In 2017, for diabetes are simply not sustainable! evidence-based policies on the possible
it was estimated that approximately Thus, preventive strategies on a global role of metformin in this setting.
425 million adults (20–79 years of age) scale remain a priority. The Diabetes Prevention Program
were living with diabetes and that half Clinical studies have provided exten- (DPP) and its ongoing Diabetes Preven-
were undiagnosed (1). This number with sive and growing evidence for the effi- tion Program Outcomes Study (DPPOS)
diabetes is projected to increase to cacy of preventive strategies (4–8). There has contributed greatly to this discus-
629 million individuals by 2045 (1). An is convincing support for using lifestyle sion. The DPP enrolled a population of
overwhelming majority of cases are at- modification (focused on healthy eating, 3,234 high-risk individuals, with random-
tributable to type 2 diabetes. In the U.S., weight loss, and enhanced physical ac- ized comparison of lifestyle modification,
the latest projections identify over 30 mil- tivity) as the cornerstone of preventive metformin, or placebo. The metformin
lion Americans living with diabetes (2). therapy. We are encouraged by early group was assigned to take metformin
The numbers at risk for developing results from lifestyle-based prevention 850 mg twice daily. An initial analysis
diabetes are also of great concern. strategies now being applied in real- after approximately 3 years of treatment
Worldwide, 352 million people are world and community settings and in demonstrated conclusively that both ac-
thought to be at risk, with 84 million high-risk ethnic groups (9,10). However, tive interventions were effective in de-
of those individuals in the U.S. (1,2). it is clear that translation of lifestyle laying the time to a new diagnosis of
The global diabetes burden imposes intervention is not always easy or effec- diabetes. At that time, progression to
great personal costs to each individual tive. The results (notably, degree of diabetes was reduced by 58% with life-
and to society in general. Beyond the weight loss and delay in diabetes de- style and 31% with metformin, relative
overall reduction of quality of life and velopment) achieved in clinical trials and to placebo. A large part of the DPP
productivity due to complications of di- in well-structured academic centers are population has been followed subse-
abetes, in 2017 diabetes was estimated not always replicated in real-world set- quently in DPPOS, allowing refinement
to result in 4 million deaths worldwide tings. Thus, key challenges for widespread and extension of the early observations.
(1). In addition, the International Diabe- implementation of preventive strategies re- Participants previously assigned to met-
tes Federation estimates that diabetes main despite considerable progress (8–14). formin were advised to continue taking it.
caused at least $727 billion (USD) in A major and lingering concern is how to Notably, later analyses have shown that
health expenditure in 2017d12% of maintain lifestyle changes and their metformin continues to protect against
total spending on adults (1). In the beneficial effects over extended periods progression to overt diabetes over a
U.S., the estimated total costs of diag- of time on a community basis. Fortu- long-term follow-up, especially in cer-
nosed diabetes have risen from $245 nately, additional evidence also supports tain subgroups (15), and have demon-
billion in 2012 to $327 billion in the consideration of various pharma- strated that the group randomized to
2017, a 26% increase over a 5-year period cologic therapies, in conjunction with metformin had reductions of some

1
American Diabetes Association, Arlington, VA
2
Oregon Health & Science University, Portland, OR
Corresponding author: William T. Cefalu, wcefalu@diabetes.org
© 2019 by the American Diabetes Association. Readers may use this article as long as the work is properly cited, the use is educational and not for profit,
and the work is not altered. More information is available at http://www.diabetesjournals.org/content/license.
See accompanying article, p. 601.
500 Commentary Diabetes Care Volume 42, April 2019

cardiovascular risk factors (16). A report compares favorably with lifestyle in its findings show substantial effectiveness
after 10 years of follow-up showed that ability to delay progression of hypergly- of metformin in both women and men
protection against progression from dys- cemia. The initial report after 3 years of and over a wide range of age, adiposity,
glycemia to overt diabetes was attained randomized treatment, using glucose and ethnicity.
with both lifestyle modification and met- values as the end point, found a 31% Overall, the current report from the
formin when HbA1c was used as a mea- placebo-adjusted reduction of risk with DPP/DPPOS group provides strong sup-
sure of outcome, as with fasting or metformin and 58% with lifestyle. The port for further discussion of using met-
postchallenge glucose in prior analyses 10-year report using an HbA1c end point formin early in the evolution from
(17). measurement suggested incidence was dysglycemia to type 2 diabetes. It sup-
Influenced in part by these findings, reduced by 38% by metformin and 29% ports the view that HbA1c values higher
the American Diabetes Association (ADA) by lifestyle at this later time (17). The than 6.0% but not yet 6.5% or higher
has been recommending (as outlined in current analysis using HbA1c suggests should prompt reconsideration of treat-
its 2019 Standards of Medical Care in 36% lower risk with metformin versus ment strategies, potentially including
Diabetes) that “Metformin therapy for placebo after 15 years (19). This was a pharmacotherapy. It also supports fur-

Downloaded from http://diabetesjournals.org/care/article-pdf/42/4/499/550658/dci180062.pdf by guest on 16 February 2025

prevention of type 2 diabetes should be post hoc analysis, by necessity using a ther discussion and heightened aware-
considered in those with prediabetes, subgroup of the overall DPP population, ness of the need for postpartum
especially for those with BMI $35 kg/m2, and therefore should be interpreted with screening of women with prior gesta-
those aged ,60 years, and women caution as suggested by the authors. tional diabetes mellitus. This very high-
with prior gestational diabetes mellitus” However, it offers reassurance that met- risk subgroup of women, typically in the
(18). In this issue of Diabetes Care is a new formin can provide a quantitatively im- 20–40 year age range, deserves special
report from the DPP/DPPOS investigators portant reduction of risk of progression consideration for more vigorous preven-
that provides additional guidance on this over this longer interval. It can be de- tive therapy to delay or prevent appear-
clinical question, especially with regard to bated whether early therapeutic inter- ance of type 2 diabetes at an early age.
the most appropriate candidates for use vention in the DPP population should While economic factors are beyond the
of metformin to slow long-term progres- be regarded as “prevention” in the strict scope of the present publication, these
sion to overt diabetes (19). sense of this term or as an early in- findings call for consideration of poten-
In this article, the Diabetes Prevention tervention to limit dysglycemia over tial cost savings derived from using met-
Program Research Group provides evi- time, across a continuum of values. Nev- formin for diabetes prevention (20).
dence that extends prior reports in sev- ertheless, metformin’s ability to maintain It is also worth noting that, beyond the
eral important ways. First, these analyses HbA1c below 6.5% for many individuals clinical trial results already discussed,
include data for an average follow-up of over 15 years must be considered a our understanding of metformin has
15 years. Second, the separate analyses clinically relevant observation. expanded in other ways. The remark-
are done using both glucose measure- Other important observations con- ably improved cardiovascular outcomes
ments and also HbA1c measurements cerned the subgroups of participants among newly diagnosed patients in the
(with individuals with HbA1c $6.5% at with greatest benefit from metformin. UK Prospective Diabetes Study (UKPDS)
baseline omitted from the analysis) for Metformin was more effective for par- who were assigned to metformin are
defining progression to diabetes. Third, ticipants whose fasting plasma glucose or well known and have led to study of
statistical assessment of progression to HbA1c was closer to diagnostic thresholds its mechanisms of action. Although these
diabetes was done both by a method already. Confirmation that higher base- are still not well understood, recent
describing relative risk (proportional haz- line glycemic measures suggest greater reports suggest that beyond effects on
ard) and one based on rates (showing potential benefit from pharmacologic the liver there are direct effects on the
absolute differences). Fourth, subgroups intervention comes as no surprise, as intestinal epithelium leading to neural
by age, sex, ethnicity, BMI, prior history we recognize that there is a continuum and hormonal signaling to the central
of gestational diabetes mellitus, and of risk for values in the prediabetic range. nervous system and elsewhere (21,22). In
baseline fasting glucose and HbA1c Nevertheless, it provides a starting point the setting of dysglycemia or early type 2
were systematically compared. A major for discussion regarding objective ways diabetes, metformin treatment may be
strength of this comprehensive analysis to identify individuals for whom phar- as effective as gastric banding with re-
is that the longer period of follow-up macologic intervention might be most spect to effects on glycemic control and
provides larger numbers of events and desirable. The other subgroup with b-cell function (23). Data from the DPP/
thus greater statistical power. Another is prominent benefit from metformin was, DPPOS cohort has provided evidence for
the inclusion of information about gly- as in prior analyses, women with certain microvascular and cardiac effects
cemic progression that is diagnosed by a history of gestational diabetes melli- (16,24,25), but more information on these
measurement of HbA1c, a procedure now tus. The greatest advantage of using outcomes is needed. The risk of lactic
arguably more common, less burden- metformin in this subgroup was shown acidosis as an adverse consequence of
some, and less time-consuming than use using glucose measurements as the mea- metformin therapy does not seem as
of glucose tolerance testing. sure of outcome. Unlike prior reports great as has been feared, especially
Several findings are of particular from the DPP, this analysis of 15-year when dosage is reduced when renal
interest. First, the analysis using data did not find that younger age and impairment is present (26). Lower doses
HbA1c $6.5% as a measure of outcome greater adiposity predict greater benefit of metformin with delayed absorption
further supports the view that metformin from metformin. Instead, the present may limit both blood levels and side
care.diabetesjournals.org Cefalu and Riddle 501

effects while retaining clinical actions Adocia, AstraZeneca, Dance, Elcelyx, Eli Lilly, supported-affordable-care-act-saves-money
(21,22). GlaxoSmithKline, Novo Nordisk, Sanofi, and Ther- .html. Accessed 19 January 2019
acos. These potential dualities of interest have 13. Fradkin JE, Roberts BT, Rodgers GP. What’s
This new information, too, deserves been reviewed and managed by Oregon Health & preventing us from preventing type 2 diabetes? N
further consideration. Given the global Science University. No other potential conflicts of Engl J Med 2012;367:1177–1179
burden of diabetes, it is clear we are at an interest relevant to this article were reported. 14. Kahn R, Davidson MB. The reality of type 2
important crossroad for diabetes pre- diabetes prevention. Diabetes Care 2014;37:
vention. There is a great need to evaluate References 943–949
1. International Diabetes Federation. IDF Dia- 15. Aroda VR, Christophi CA, Edelstein SL, et al.;
all the lessons learned and to review all Diabetes Prevention Program Research Group.
the incredible amount of data now avail- betes Atlas, 8th edition [Internet], 2017. Avail-
able from www.diabetesatlas.org. Accessed 18 The effect of lifestyle intervention and metfor-
able in order to take logical next steps as a January 2019 min on preventing or delaying diabetes among
society. We realize that lifestyle modifi- 2. Centers for Disease Control and Prevention. women with and without gestational diabetes:
cation remains the cornerstone of diabe- National Diabetes Statistics Report, 2017: Esti- the Diabetes Prevention Program outcomes
mates of Diabetes and Its Burden in the United study 10-year follow-up. J Clin Endocrinol Metab
tes prevention, yet long-term adherence 2015;100:1646–1653
States [Internet]. U.S. Department of Health
is difficult and implementation can be 16. Aroda VR, Knowler WC, Crandall JP, et al.;
and Human Services, 2017. Available from

Downloaded from http://diabetesjournals.org/care/article-pdf/42/4/499/550658/dci180062.pdf by guest on 16 February 2025

labor intensive and costly and may face https://www.cdc.gov/diabetes/pdfs/data/statistics/ Diabetes Prevention Program Research Group.
obstacles in real-world settings. There- national-diabetes-statistics-report.pdf. Accessed 18 Metformin for diabetes prevention: insights
fore, in order to be successful for diabetes January 2019 gained from the Diabetes Prevention Pro-
3. American Diabetes Association. Economic gram/Diabetes Prevention Program Outcomes
prevention efforts, we need to continue to Study. Diabetologia 2017;60:1601–1611
costs of diabetes in the U.S. in 2017. Diabetes
address these hurdles for successful life- 17. Diabetes Prevention Program Research Group.
Care 2018;41:917–928
style modification in a real-world setting. 4. Pan XR, Li GW, Hu YH, et al. Effects of diet and HbA1c as a predictor of diabetes and as an outcome
However, there is continued interest in exercise in preventing NIDDM in people with in the diabetes prevention program: a randomized
preventive pharmacotherapy as an ad- impaired glucose tolerance. The Da Qing IGT and clinical trial. Diabetes Care 2015;38:51–58
Diabetes Study. Diabetes Care 1997;20:537–544 18. American Diabetes Association. 3. Preven-
junct to lifestyle modification, and met- tion or delay of type 2 diabetes: Standards of
5. Tuomilehto J, Lindström J, Eriksson JG, et al.;
formin continues to be the leading Finnish Diabetes Prevention Study Group. Pre- Medical Care in Diabetesd2019. Diabetes Care
candidate. Its long-term efficacy and vention of type 2 diabetes mellitus by changes in 2019;42(Suppl. 1):S29–S33
safety are well established. Although lifestyle among subjects with impaired glucose 19. Diabetes Prevention Progam Research
the ADA guidelines have, since 2008, tolerance. N Engl J Med 2001;344:1343–1350 Group. Long-term effects of metformin on di-
6. Knowler WC, Barrett-Connor E, Fowler SE, abetes prevention: identification of subgroups
endorsed consideration of metformin that benefited most in the Diabetes Prevention
et al.; Diabetes Prevention Program Research
for individuals who are at very high risk Group. Reduction in the incidence of type 2 Program and Diabetes Prevention Program Out-
for diabetes, further evolution of these diabetes with lifestyle intervention or metfor- comes Study. Diabetes Care 2019;42:601–608
guidelines may be needed. The latest min. N Engl J Med 2002;346:393–403 20. Herman WH. The cost-effectiveness of di-
report from the DPP/DPPOS investigators 7. Ramachandran A, Snehalatha C, Mary S, abetes prevention: results from the Diabetes
Mukesh B, Bhaskar AD, Vijay V; Indian Diabetes Prevention Program and the Diabetes Prevention
further supports the rationale for metfor- Program Outcomes Study. Clin Diabetes Endo-
Prevention Programme. The Indian Diabetes Pre-
min as the most likely candidate for pre- vention Programme shows that lifestyle modifi- crinol 2015;1:9
ventive strategies. More importantly, the cation and metformin prevention type 2 diabetes 21. Rena G, Hardie DG, Pearson ER. The mech-
data reported in this issue of Diabetes in Asian Indian subjects with impaired glucose anisms of action of metformin. Diabetologia
Care help to identify those individuals who tolerance (IDPP-1). Diabetologia 2006;49:289–297 2017;60:1577–1585
8. Cefalu WT, Buse JB, Tuomilehto J, et al. 22. Buse JB, DeFronzo RA, Rosenstock J, et al.
would benefit most from this interven- The primary glucose-lowering effect of metfor-
Update and next steps for real-world translation
tion. It is important to note that ADA’s of interventions for type 2 diabetes prevention: min resides in the gut, not the circulation: results
Standards of Care does not recommend reflections from a Diabetes Care Editors’ Expert from short-term pharmacokinetic and 12-week
use of other medications for diabetes Forum. Diabetes Care 2016;39:1186–1201 dose-ranging studies. Diabetes Care 2016;39:
prevention aside from metformin. 9. Hall DL, Lattie EG, McCalla JR, Saab PG. 198–205
Translation of the Diabetes Prevention Program 23. Xiang AH, Trigo E, Martinez M, et al.; RISE
Despite a clear need for better means Consortium. Impact of gastric banding versus
to ethnic communities in the United States. J
of preventing diabetes, the lack of a Immigr Minor Health 2016;18:479–489 metformin on b-cell function in adults with
formal prevention-related indication for 10. Jiang L, Manson SM, Beals J, et al.; Special impaired glucose tolerance or mild type 2 di-
metformin is and will remain a signifi- Diabetes Program for Indians Diabetes Preven- abetes. Diabetes Care 2018;41:2544–2551
cant barrier to more widespread use. Per- tion Demonstration Project. Translating the Di- 24. Diabetes Prevention Program Research Group.
abetes Prevention Program into American Indian Long-term effects of lifestyle intervention or met-
haps we are now ready to surmount this and Alaska Native communities: results from formin on diabetes development and microvascular
hurdle. We fully expect that, in the near the Special Diabetes Program for Indians complications over 15-year follow-up: the Di-
future, discussion will resume as to when, Diabetes Prevention demonstration project. abetes Prevention Program Outcomes Study.
for whom, and how metformin should Diabetes Care 2013;36:2027–2034 Lancet Diabetes Endocrinol 2015;3:866–875
be deployed as preventive therapy. Let 11. Albright AL, Gregg EW. Preventing type 2 25. Goldberg RB, Aroda VR, Bluemke DA, et al.;
diabetes in communities across the U.S.: the Diabetes Prevention Program Research Group.
the arguments begin! National Diabetes Prevention Program. Am J Effect of long-term metformin and lifestyle in the
Prev Med 2013;44(Suppl. 4):S346–S351 Diabetes Prevention Program and its Outcome
12. U.S. Department of Health and Human Ser- Study on coronary artery calcium. Circulation
Funding and Duality of Interest. No commer- vice. Independent experts confirm that diabetes 2017;136:52–64
cial sponsors contributed support. W.T.C. is an prevention model supported by the Affordable 26. Lalau J-D, Kajbaf F, Bennis Y, Hurtel-Lemaire
employee of the ADA. M.C.R. reports receiving Care Act saves money and improves health A-S, Belpaire F, De Broe ME. Metformin treat-
research grant support through Oregon Health & [Internet], 2016. Available from http://www ment in patients with type 2 diabetes and chronic
Science University from Novo Nordisk and As- .hhs.gov/about/news/2016/03/23/independent- kidney disease stages 3A, 3B, or 4. Diabetes Care
traZeneca and honoraria for consulting from experts-confirm-diabetes-prevention-model- 2018;41:547–553