Metformin, a biguanide, primarily decreases hepatic glucose production by inhibiting gluconeogenesis. It enhances peripheral insulin sensitivity in the skeletal muscle by increasing insulin receptor tyrosine kinase activity and glucose transporter (GLUT)-4 translocation to the cell membrane (36). Metformin also improves beta-cell responsiveness to a glucose load through correction of glucotoxicity (37).

The Diabetes Prevention Program (DPP) study was a multi-center trial which enrolled 3,234 subjects with prediabetes and randomized them to either intensive lifestyle intervention (intended to achieve 7% body weight loss), metformin (850mg twice daily) or standard lifestyle recommendations (38). Participants in the study were overweight or obese (mean BMI 34 kg/m²), mostly middle-aged adults with IGT or impaired fasting glucose (IFG) with values between 95-126 mg/dL. In the ensuing 2.8 years, T2DM incidence was reduced by 58% with lifestyle intervention and by 31% with metformin compared to placebo (6). It is interesting to note that genome-wide association studies in the DPP cohort revealed novel ethnic-specific associations with metformin response and may have implications for individualized therapy (39).

A secondary analysis of the DPP study was performed in subjects with history of gestational diabetes (GDM). Women with history of GDM were compared with women with previous live birth but without GDM history. It was found that progression to T2DM is more common among women with history of GDM compared with those without it, despite equivalent degrees of IGT at baseline. Both lifestyle modifications (LSM) (53% reduction in T2DM incidence) and metformin (50% reduction in T2DM incidence) were highly effective in delaying T2DM in women with IGT and history of GDM (40).

Eighty-eight percent of the surviving DPP cohort enrolled in a long-term follow up study, the Diabetes Prevention Program Outcomes Study (DPPOS). During the DPPOS, unmasked metformin was continued as a study intervention in the original metformin group. T2DM incidence was reduced by 34% in the lifestyle group and 18% in the metformin group compared to placebo in the following 10 years (7). Significant reduction in T2DM incidence persists in the metformin group at 15-year (relative risk reduction/RRR 18%; p=0. 001) (8) and 22-year follow up (RRR 18%) (9).

Similarly, The Indian Diabetes Prevention Program (IDPP) randomized 531 Asian Indian subjects with IGT (mean BMI 25.8 kg/m2) to 4 groups: LSM or metformin (250-500mg twice daily) or LSM plus metformin or control group (standard healthcare advice). The median follow-up period was 30 months. The 3-year cumulative incidences of diabetes were 39.3%, 40.5%, 39.5% and 55.0%, respectively. The RRR was 26.4% with metformin (95% CI 19.1–35.1, p=0.029) and 28.2% with LSM plus metformin (95% CI 20.3–37.0, p=0.022), as compared to the control group (10).

The Chinese Diabetes Prevention Program (CDPP) evaluated the preventive effect of lifestyle intervention with diet and exercise, acarbose, and metformin on T2DM progression in 321 subjects with IGT. Annual T2DM incidence was 11.6, 8.2, 2.0, and 4.1% in the control, lifestyle intervention, acarbose, and metformin groups, respectively (11).

The Early Diabetes Intervention Trial (EDIT) analyzed the effects of metformin and acarbose in T2DM prevention in 631 subjects with IFG. At three years, there was an 8% risk reduction with acarbose and 37% with metformin, compared to placebo (12) but there was no difference in the relative risk for T2DM at six-year of follow-up. For patients with IGT at baseline, the RRR was significant with acarbose (0.66) but not with metformin (1.09), suggesting different effects of different therapies in subjects with IGT and IFG (13).

A study by Begum et al. (2009) showed that occurrence of GDM among women with polycystic ovarian syndrome (PCOS) was significantly lower in the metformin treatment group with only one subject (3.44%) vs nine of the 30 pregnancies (30%) without metformin (14). Similarly, Glueck et al. (2002) reported the odds ratio for GDM in women with metformin vs those without metformin was 0.093 (95% CI: 0.011 to 0.795) (15).

Recent studies showed promising effects of metformin in combination with glucagon-like peptide-1 (GLP-1) receptor agonists (16) and dipeptidyl peptidase 4 (DPP-4) inhibitors (17, 18) for T2DM prevention.

According to the American Diabetes Association (41), metformin should be considered for T2DM prevention in adults with BMI ≥35 kg/m2, age ≤ 60 years, higher fasting plasma glucose (≥ 110 mg/dL), and higher A1C (≥ 6.0%), and in women with prior GDM.

Thiazolidinediones (TZDs) are insulin sensitizers. They activate gamma isoform of peroxisome proliferator-activated receptor (PPAR γ), enhance glucose uptake by skeletal muscles and adipocytes, improve insulin sensitivity and consequently improve pancreatic beta-cell function (42, 43).

The Troglitazone in Prevention of Diabetes (TRIPOD) study (19) compared Troglitazone with placebo in 266 nondiabetic Hispanic women (mean age 34.6 years; mean BMI 30.5 kg/m²) with previous GDM, about 70% of whom had IGT at entry into the trial. Troglitazone treated group had a 55% reduction in the incidence of T2DM over 2.5 years. The drug was recalled before the planned study-end because of reports of hepatic failure.

The Pioglitazone in Prevention of Diabetes (PIPOD) study (20) was an open-label follow-up of 89 women from TRIPOD who had not developed diabetes (A1C <7%). These participants showed an average rate of diabetes of 4.6% per year during treatment with pioglitazone for three years, which was much lower compared with the rate of 12.1% per year that was observed during placebo treatment in the TRIPOD study.

Actos Now for the Prevention of Diabetes (ACT NOW) study (21) was an RCT conducted to examine the effectiveness of pioglitazone in preventing T2DM among 602 subjects (mean age 52 years; mean BMI 34 kg/m²⁾ with IGT. The annual rate of T2DM was 7.6% in placebo-treated vs. 2.1% in pioglitazone treated subjects (HR 0.28, p<0.0001) over 2.4 years.

Concerns over adverse effects of TZDs has dampened the enthusiasm to use pioglitazone for T2DM prevention. Undesirable effects of TZDs include fluid retention, increased risk of heart failure, weight gain and loss of bone density increasing fracture risk.

These drugs act by competitively inhibiting alpha- glucosidase enzyme and decreasing carbohydrate absorption from the small intestine, thus reducing postprandial glucose levels (44).

The Study TO Prevent Non-Insulin Dependent Diabetes Mellitus (STOP-NIDDM) (24) evaluated the effects of acarbose in delaying progression of IGT to T2DM in 1429 subjects. Over the 3.3-year follow-up period, there was a 25% reduction in the incidence of T2DM in the acarbose group compared to placebo. Furthermore, acarbose significantly increased the regression of IGT to normal glucose tolerance (HR 1.42, 95% CI 1.24–1.62; p<0·0001). About one-quarter of the cohort did not complete the study, and this was attributed to the gastrointestinal side effects of acarbose.

Kwawamori et al., randomized 1780 Japanese subjects with IGT to either receiving voglibose 0.6 mg/day or placebo. At the end of the study (11 months; ended early due to efficacy) the T2DM hazard ratio (voglibose vs. placebo) was 0.60 (95% CI 0.43–0.82) (25).

The SGLT2 is expressed in the proximal tubule and mediates reabsorption of approximately 90 percent of the filtered glucose load to the kidneys. SGLT2 inhibitors promote renal excretion of glucose and thereby modestly lower elevated blood glucose levels (45).

Pre-specified pooled analysis (26) of the DAPA-CKD and DAPA-HF trials suggests that dapagliflozin (10 mg daily) may reduce new-onset T2DM compared with placebo (HR 0·67 [95% CI 0·51–0·88]). In pooled analysis of selected 4,003 participants who had no previous diagnosis of T2DM (mean age 63 years, mean A1C 5.7%), dapagliflozin reduced new-onset T2DM incidence (defined as having A1C ≥6.5% on two consecutive follow-up visits, or a clinical diagnosis of T2DM that led to initiation of a glucose-lowering agent) by approximately one-third. The overall incidence of new-onset T2DM was 2.6 events per 100 patient-years in the dapagliflozin group vs 3.9 events per 100 patient-years in the placebo group (HR 0.67; 95% CI 0.51–0.88; p=0.0040). Treatment was predominantly beneficial in participants with prediabetes. Possible mechanisms include protection of pancreatic beta-cells from glucotoxicity, weight loss, and improvement in hepatic insulin sensitivity. Improvement in CKD and heart failure may have contributed to insulin sensitivity (26). A prospective RCTs is needed to confirm whether dapagliflozin truly prevents T2DM.

Although empagliflozin did not demonstrate significant benefit for T2DM prevention among patients with heart failure and prediabetes in EMPEROR-Preserved study (46) (HR: 0.84; 95% CI 0.65–1.07) and EMPEROR-Reduced study (47) (HR: 0.86; 95% CI 0.62–1.19), the hazard ratios were consistent with potential benefit. A recent meta-analysis of four RCTs showed that SGLT2 inhibitors (empagliflozin and dapagliflozin) were significantly associated with a lower risk of new-onset diabetes (relative risk, 0.79; 95% CI, 0.68-0.93) (48). This meta-analysis included 5655 participants with pre-diabetes. The relative risks of new-onset diabetes in dapagliflozin and empagliflozin were 0.68 (95% CI, 0.52-0.89) and 0.87 (95% CI, 0.72-1.04), respectively.

Orlistat acts by reversibly inhibiting gastric and pancreatic lipases (49). It also increases postprandial glucagon-like peptide 1 (GLP-1) levels. In a pooled analysis by Heymsfield et al. (50) orlistat compared to placebo, reduced 2-year cumulative diabetes incidence by 61% (7.6% in the placebo group vs. 3.0% in the orlistat group) among those with IGT. However, due to the side effects of orlistat, only 69% of the participated subjects completed the study.

Xenical in the Prevention of Diabetes in Obese Subjects (XENDOS) study (27) was a 4-year, double-blind, prospective study where 3,305 obese subjects with prediabetes were randomized to either lifestyle intervention plus orlistat 120 mg or plus placebo. It showed that the cumulative incidence of T2DM was 9.0% with placebo and 6.2% with orlistat, corresponding to a risk reduction of 37.3% (p=0.0032). Independent of orlistat or placebo treatment, the relative risk of developing T2DM was greater in patients with IGT, men, older subjects, and subjects with a higher BMI.

Phentermine is a sympathomimetic drug, which stimulates norepinephrine release in the hypothalamus, suppresses appetite and increases satiety. Topiramate’s action is not well understood but hypothesized to work on alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors and kainite receptors to reduce food cravings, and on Gamma-aminobutyric acid (GABA) receptors to increase energy expenditure (51).

The CONQUER study evaluated the efficacy of phentermine/topiramate combination as an adjunct to LSM for weight loss and metabolic risk reduction in individuals who are overweight and obese, with two or more risk factors (28). They enrolled 84% of subjects without T2DM at baseline. In this population, development of T2DM was less in phentermine 15 mg/topiramate ER 92 mg group compared to placebo after 56 weeks of intervention (1.7% versus 3.6%; HR, 0.47; 95% CI, 0.25–0.88).

In the long-term follow up study, SEQUEL, a total of 78.1% of subjects in the original CONQUER study continued to take blinded medication over 108 weeks. The annual incidence rates for progression to T2DM were 0.9% in high dose phentermine/topiramate group and 3.7% in the placebo group (P=0.008) (52). These studies were not performed to assess prevention of T2DM as a primary outcome.

A subgroup analysis of the CONQUER study including subjects with prediabetes and/or metabolic syndrome at baseline showed that the annual incidence of T2DM was 1.3% for high dose phentermine/topiramate and 6.1% in placebo (53).

Bupropion is a norepinephrine and dopamine reuptake inhibitor which stimulates proopiomelanocortin (POMC) neurons in the hypothalamus, with a downstream effect of increased satiety. Naltrexone prevents rebound inhibition of POMC neurons by β-endorphin, and synergistically works with bupropion to increase satiety (51).

There were four Contrave Obesity Research (COR) studies looked at the weight loss effects of Naltrexone/Bupropion (NB): the COR-I, COR-II, COR-BMOD (behavior modification), COR-Diabetes (54–57). None of these studies were designed to analyze progression of IGT to T2DM, but COR-I study (54) showed significant decrease in fasting plasma glucose in response to naltrexone SR 32 mg/bupropion SR 16 mg combination treatment. COR-Diabetes study (57) including subjects with T2DM reported improvements in glucose homeostasis, resulting in A1C reduction of 0.6% in the NB group, and the A1C goal of ¾7% was achieved in 44% in NB group vs 26% in placebo group.

Further longer-term studies are needed to establish the effect of this medication in T2DM prevention.

Glucagon-like peptide 1 receptor agonists (GLP-1 RA) mimic the action of endogenous GLP-1 in enhancing glucose-dependent insulin secretion and suppress glucagon production from pancreatic alpha cells and thus used to treat T2DM. Apart from glycemic control, they are also thought to reduce neuroinflammation, promote nerve growth, improve cardiac function, suppress appetite, delay gastric emptying, regulate lipid metabolism and reduce fat deposition (58).

In a study (30) of 152 obese patients (mean age 46 years; mean BMI 39.6 kg/m2) with and without prediabetes (IGT or IFG), blood glucose was normalized in 77% and 56% after exenatide and placebo, respectively. There was a significantly higher weight loss with exenatide of 5.1 kg vs 1.6 kg with placebo.

In the SCALE study Obesity and Prediabetes trial (31), 2,254 adults with prediabetes and BMI ≥30 kg/m² or ≥27 kg/m² with comorbidities (mean age 47 years, mean BMI 39 kg/m²) were randomized to receive liraglutide 3mg/day vs a matched placebo, as an adjunct to LSM. Time to onset of diabetes in the treatment group was 2.7 times longer with liraglutide compared to placebo (95% CI 1·9 to 3·9, p<0·0001; HR 0.21, 95% CI 0.13–0.34). However, about half the participants withdrew from the study. Based on an analysis imputing the missing data, T2DM incidence was reduced by 66% (HR 0.34, 95% CI 0.22–0.53) with liraglutide vs 36% with placebo.

Recently, the effect of semaglutide in T2DM prevention was assessed in comparison to placebo in overweight or obese individuals in the Semaglutide Treatment Effect in People with Obesity (STEP) studies. In a post-hoc analysis (32) of the STEP 1 (33) (68 weeks) and STEP 4 (34) (20-week run-in on semaglutide, 48-week randomized withdrawal) trials, it was shown that semaglutide 2.4 mg could reduce the 10-year risk of progression to T2DM by 61%, regardless of the initial glycemic status (vs 13% reduction in the placebo group (p<0.01). In this analysis, the 10-year risk of T2DM was calculated using Cardiometabolic Disease Staging (CMDS). Most of the risk reduction occurred during the initial weeks (0-20 weeks), from 21% to 11%. Risk scores further decreased with continued semaglutide (weeks 20-68) but increased with switch to placebo (32% reduction vs. 41% increase, p<0.01).

Tirzepatide, the dual GIP/GLP-1 receptor agonist was recently approved by the FDA in May 2022 for treatment of T2DM. In the 72-week SURMOUNT-1 trial (35), 40.6% of the patients had prediabetes at baseline. At the end of the trial, 95.3% of those participants reverted back to normoglycemia, as compared with 61.9% of participants in the placebo group. There was also a notable decrease in the fasting insulin levels in the treatment groups. This is probably attributed to the significant weight loss that was seen in association with tirzepatide (15-20.9% weight loss), however future dedicated studies are still needed to confirm the role of tirzepatide for T2DM prevention.

Various trials (23, 59–65) suggested that RAAS inhibition may reduce the incidence of new onset T2DM in patients with or without hypertension or at high risk of T2DM. The risk reduction was explained by hemodynamic effects such as improved delivery of insulin and glucose to the peripheral skeletal muscle, non-hemodynamic effects, including direct effects on glucose transport and insulin signaling pathways, which collectively decrease insulin resistance.

A meta-analysis (66) of new-onset T2DM in select comparative outcome trials involving the use of RAAS blockade vs non-RAS blockade showed a lower risk ratio (RR 0.78; 95% CI 0.74-0.88). However, most of these trials did not include T2DM incidence as the pre-specified end point.

The Vitamin D and Type 2 Diabetes (D2d) Study (67), randomly assigned 2,423 adults with prediabetes to receive 4,000 units of vitamin D3 per day or placebo, regardless of their baseline vitamin D status. After a median follow-up of 2.5 years, T2DM occurred in 293 participants in the vitamin D group, and 323 in the placebo group (9.39 and 10.66 events per 100 person-years, respectively). The hazard ratio for vitamin D compared with placebo was 0.88 (95% CI, 0.75 to 1.04; p=0.12), failing to show a significant decrease in T2DM incidence among the study population.

However, subsequent meta-analyses of studies in patients with pre-diabetes demonstrated that vitamin D supplementation at moderate to high doses significantly reduced the incidence risk of T2DM (68, 69). A secondary analysis of the D2d study showed that participants who maintained higher intratrial serum vitamin D levels during follow-up had a reduced risk of diabetes (70).

Men who are overweight frequently have low serum testosterone (T) concentrations, which is in turn associated with increased risk of T2DM (71).

T4DM (72) was a randomized, placebo-controlled trial, which included 1,007 men who were enrolled in lifestyle intervention program and randomly assigned to receive T or placebo. At 2 years, 21% of the subjects in the placebo group failed OGTT vs 12% in the T group (RR 0.59, CI 0.43-0.80, p=0.007). Treatment effect was independent of baseline serum testosterone level. So far, testosterone use is not approved for T2DM prevention or treatment.