Diabetes Care Volume 43, September 2020 1983 Metformin Should Not Be Used to Mayer B. Davidson Treat Prediabetes – Diabetes Care 2020;43:1983 1987 | https://doi.org/10.2337/dc19-2221 PERSPECTIVES IN CARE Based on the results of the Diabetes Prevention Program Outcomes Study (DPPOS), in which metformin significantly decreased the development of diabetes in individuals with baseline fasting plasma glucose (FPG) concentrations of 110–125vs. 100–109 mg/dL (6.1–6.9 vs. 5.6–6.0 mmol/L) and A1C levels 6.0–6.4% (42–46 mmol/ mol) vs. <6.0% and in women with a history of gestational diabetes mellitus, it has been suggested that metformin should be used to treat people with prediabetes. Since the association between prediabetes and cardiovascular disease is due to the associated nonglycemic risk factors in people with prediabetes, not to the slightly increased glycemia, the only reason to treat with metformin is to delay or prevent the development of diabetes. There are three reasons not to do so. First, approximately two-thirds of people with prediabetes do not develop diabetes, even after many years. Second, approximately one-third of people with prediabetes return to normal glucose regulation. Third, people who meet the glycemic criteria for prediabetes are not at risk for the microvascular complications of diabetes and thus metformin treatment will not affect this important outcome. Why put people who are not at risk for the microvascular complications of diabetes on a drug (possibly for the rest of their lives) that has no immediate advantage except to lower subdiabetes glycemia to even lower levels? Rather, individuals at the highest risk for developing diabetesdi.e., those with FPG concentrations of 110–125 mg/dL (6.1– 6.9 mmol/L) or A1C levels of 6.0–6.4% (42–46 mmol/mol) or women with a history of gestational diabetes mellitusdshould be followed closely and metformin im- mediately introduced only when they are diagnosed with diabetes. The Diabetes Prevention Program (DPP) studied the effect of an intensive lifestyle interventionandmetforminonthedevelopmentof diabetesinacohortof peoplewith an increased risk for diabetes (termed prediabetes). After a mean of 2.8 years of follow-up, 31% fewer metformin-treated individuals developed diabetes than in- dividuals in the control group (1). Eighty-six percent of members of the metformin and placebo groups agreed to be followed and entered the Diabetes Prevention Program Outcomes Study (DPPOS). The placebo was discontinued and metformin (850 mg Charles R. Drew University, Los Angeles, CA b.i.d.) was unmasked and continued. The 15-year follow-up results in the DPPOS Corresponding author: Mayer B. Davidson, metformin-treated group recently showed significantly less development of diabetes [email protected] in participants with higher baseline fasting plasma glucose (FPG) concentrations (110– Received 5 February 2020 and accepted 28 April 125 vs. 100–109 mg/dL [6.1–6.9 vs. 5.6–6.0 mmol/L]) (2), in those with A1C levels 6.0– 2020 6.4% (42–46 mmol/mol) vs. ,6.0%, and in women with a history of gestational © 2020 by the American Diabetes Association. Readers may use this article as long as the work is diabetes mellitus (2). An accompanying editorial (3) invited arguments discussing properly cited, the use is educational and not for whether people meeting the criteria for prediabetes should be treated with metformin. profit, and the work is not altered. More infor- Since 33.9% of the population over 18 years of age in the U.S., 84.1 million people, have mation is available at https://www.diabetesjournals prediabetes (4), use of metformin to treat them would increase drug costs considerably .org/content/license. for payers as well as for many individuals. This Perspective will argue against doing so. See accompanying article, p. 1988. 1984 Metformin for Prediabetes: No Diabetes Care Volume 43, September 2020 It is instructive to review the history of would greatly reduce the predictive dis- stated that there was insufficient evi- diagnosing prediabetes. Before 1979, crepancy between IGT and IFG for the dence to decide on A1C values ,6.5% there were six different criteria for di- subsequent development of diabetes. (48 mmol/mol) (25). agnosing diabetes. In that year, the Na- In 2008, an invited expert panel (IEP) The Diabetes Canada 2018 Clinical tional Diabetes Data Group (NDDG) recommended that diabetes could be Practice Guidelines recommended the published a single set of criteria for the diagnosed by an A1C level of $6.5% criteria for prediabetes as IFG concen- diagnosis (FPG $140 mg/dL [7.8 mmol/L] (48 mmol/mol) and also suggested trations of 110–125 mg/dL (6.1–6.9 mmol/L) or 2-h glucose concentration on an oral that values of 6.0–6.4% (42–46 mmol/ or A1C levels of 6.0–6.4% (42–46 mmol/ glucose tolerance test [OGTT] $200 mg/ mol) required close follow-up and testing mol) (26). dL [11.1 mmol/L]) based on three pro- (10). In response, the ADA, the European Although numerous studies have shown spective studies in subjects who had a Association for the Study of Diabetes, that glycemia is not an independent risk baseline OGTT and were evaluated for and the International Diabetes Feder- factor for CVD (14–21), it certainly is for diabetic retinopathy 3 to 8 years later (5). ation appointed an International Expert the development of diabetes. However, They also opined that individuals whose Committee that agreed with the invited there is no obvious threshold; the risk 2-hglucosevaluewas$140to199mg/dL expert panel regarding the diagnosis of starts to increase beginning at FPG con- (7.8 to 11.0 mmol/L) had impaired glu- diabetes (if the A1C level were con- centrations of 82–87 mg/dL (4.6–4.8 cose tolerance (IGT), which indicated an firmed) (11). However, that committee mmol/L) and progresses in a curvilinear increased risk for developing diabetes. also opined that because of the progres- fashion (27–29). For instance, the risk No FPG criterion for diagnosing predia- sive continuum of risk of increasing gly- with the WHO IFG criterion of 110– betes was offered. cemia below the diagnostic levels of 125 mg/dL (6.1–6.9 mmol/L) is 2.1- to The NDDG criteria for diagnosing di- diabetes for the subsequent develop- 11.3-fold higher than with the lower abetes were not equally sensitive. Al- ment of diabetes, it was inappropriate bound of the ADA IFG criterion of 100– though 95% of all persons with an FPG to define a specific prediabetes risk group. 109 mg/dL (5.6–6.0 mmol/L) (14,30,31). concentration $140 mg/dL (7.8 mmol/L) The ADA subsequently adopted the rec- Similarly, the risk with the A1C IEP crite- also had a 2-h glucose concentration ommended A1C level for diagnosing di- rion of 6.0–6.4% (42–46 mmol/mol) is 2.0- $200 mg/dL (11.1 mmol/L) on the OGTT, abetes but also included an A1C criterion to 6.5-fold higher than with the lower this level of concordance was not seen of 5.7–6.4% (39–46 mmol/mol) for pre- bound of the ADA A1C criterion of 5.7– with all persons who had a 2-h glucose diabetes (12). The lower bound of the 5.9% (39–41 mmol/mol) (14,31). concentration $200 mg/dL (11.1 mmol/L). prediabetes criteria was based on mod- Claims have been made that treating Only one-quarter to one-half of these eling the estimated composite risk of people with prediabetes with antihyper- individuals also had an FPG $140 mg/dL developing diabetes and cardiovascular glycemic drugs (metformin, thiazolidine- (7.8 mmol/L) (6). The American Diabetes disease (CVD) using cross-sectional data diones [TZD], a-glucosidase inhibitors, Association (ADA) convened an Expert from the 2005–2006 NHANES (13). How- glucagon-like peptide 1 agonists, basal Committee to address this imbalance (7). ever, the glycemia of prediabetes is not insulin) has delayed or even prevented BasedonananalysisbytheExpertCom- independently associated with CVD the development of diabetes. This is a mittee of the third National Health and (14–21). Furthermore, in people who ex- misinterpretation of the situation. These Nutrition Examination Survey (NHANES perience an acute coronary syndrome, the drugs have simply treated a level of gly- III) and several other published studies, outcomes (length of hospital stay, 28-day cemia lower than the diagnostic criteria for the new FPG criterion for diagnosing readmission rate, acute pulmonary edema, diabetes retarding its increase to the level diabetes was set at $126 mg/dL (7.0 12-month recurrent acute coronary syn- at which a diagnosis of diabetes would mmol/L), which yielded the same prev- drome, or mortality) are no different occur. After these drugs were discontin- alence of diabetes as did a 2-h glucose between those with prediabetes (A1C 5.7– ued, the prevalence of diabetes in treated value on the OGTT of $200 mg/dL (11.1 6.4% [39–46 mmol/mol]) or with A1C individuals mirrored that in the placebo mmol/L). Since no studies defining a normal levels ,5.7% (39 mmol/mol) (22). Rather, group. FPG concentration were known, the often- the association between prediabetes and An argument has been made that the stated normal glucose value of ,110 mg/dL CVD is due to the other risk factors for difference between the placebo and (6.1 mmol/L) used by clinical laboratories CVD that people meeting the glycemic metformin groups in the DPP only de- was adopted. The FPG range of 110–125 criteria for prediabetes also have.