Metformin for prevention or delay of type 2 diabetes mellitus and its associated complications in persons at increased risk for the development of type 2 diabetes mellitus.

The projected rise in the incidence of type 2 diabetes mellitus (T2DM) could develop into a substantial health problem worldwide. Whether metformin can prevent or delay T2DM and its complications in people with increased risk of developing T2DM is unknown. To assess the effects of metformin for the prevention or delay of T2DM and its associated complications in persons at increased risk for the T2DM. We searched the Cochrane Central Register of Controlled Trials, MEDLINE, Scopus, ClinicalTrials.gov, the World Health Organization (WHO) International Clinical Trials Registry Platform and the reference lists of systematic reviews, articles and health technology assessment reports. We asked investigators of the included trials for information about additional trials. The date of the last search of all databases was March 2019. We included randomised controlled trials (RCTs) with a duration of one year or more comparing metformin with any pharmacological glucose-lowering intervention, behaviour-changing intervention, placebo or standard care in people with impaired glucose tolerance, impaired fasting glucose, moderately elevated glycosylated haemoglobin A1c (HbA1c) or combinations of these. Two review authors read all abstracts and full-text articles and records, assessed risk of bias and extracted outcome data independently. We used a random-effects model to perform meta-analysis and calculated risk ratios (RRs) for dichotomous outcomes and mean differences (MDs) for continuous outcomes, using 95% confidence intervals (CIs) for effect estimates. We assessed the certainty of the evidence using GRADE. We included 20 RCTs randomising 6774 participants. One trial contributed 48% of all participants. The duration of intervention in the trials varied from one to five years. We judged none of the trials to be at low risk of bias in all 'Risk of bias' domains. Our main outcome measures were all-cause mortality, incidence of T2DM, serious adverse events (SAEs), cardiovascular mortality, non-fatal myocardial infarction or stroke, health-related quality of life and socioeconomic effects.The following comparisons mostly reported only a fraction of our main outcome set. Fifteen RCTs compared metformin with diet and exercise with or without placebo: all-cause mortality was 7/1353 versus 7/1480 (RR 1.11, 95% CI 0.41 to 3.01; P = 0.83; 2833 participants, 5 trials; very low-quality evidence); incidence of T2DM was 324/1751 versus 529/1881 participants (RR 0.50, 95% CI 0.38 to 0.65; P < 0.001; 3632 participants, 12 trials; moderate-quality evidence); the reporting of SAEs was insufficient and diverse and meta-analysis could not be performed (reported numbers were 4/118 versus 2/191; 309 participants; 4 trials; very low-quality evidence); cardiovascular mortality was 1/1073 versus 4/1082 (2416 participants; 2 trials; very low-quality evidence). One trial reported no clear difference in health-related quality of life after 3.2 years of follow-up (very low-quality evidence). Two trials estimated the direct medical costs (DMC) per participant for metformin varying from $220 to $1177 versus $61 to $184 in the comparator group (2416 participants; 2 trials; low-quality evidence). Eight RCTs compared metformin with intensive diet and exercise: all-cause mortality was 7/1278 versus 4/1272 (RR 1.61, 95% CI 0.50 to 5.23; P = 0.43; 2550 participants, 4 trials; very low-quality evidence); incidence of T2DM was 304/1455 versus 251/1505 (RR 0.80, 95% CI 0.47 to 1.37; P = 0.42; 2960 participants, 7 trials; moderate-quality evidence); the reporting of SAEs was sparse and meta-analysis could not be performed (one trial reported 1/44 in the metformin group versus 0/36 in the intensive exercise and diet group with SAEs). One trial reported that 1/1073 participants in the metformin group compared with 2/1079 participants in the comparator group died from cardiovascular causes. One trial reported that no participant died due to cardiovascular causes (very low-quality evidence). Two trials estimated the DMC per participant for metformin varying from $220 to $1177 versus $225 to $3628 in the comparator group (2400 participants; 2 trials; very low-quality evidence). Three RCTs compared metformin with acarbose: all-cause mortality was 1/44 versus 0/45 (89 participants; 1 trial; very low-quality evidence); incidence of T2DM was 12/147 versus 7/148 (RR 1.72, 95% CI 0.72 to 4.14; P = 0.22; 295 participants; 3 trials; low-quality evidence); SAEs were 1/51 versus 2/50 (101 participants; 1 trial; very low-quality evidence). Three RCTs compared metformin with thiazolidinediones: incidence of T2DM was 9/161 versus 9/159 (RR 0.99, 95% CI 0.41 to 2.40; P = 0.98; 320 participants; 3 trials; low-quality evidence). SAEs were 3/45 versus 0/41 (86 participants; 1 trial; very low-quality evidence). Three RCTs compared metformin plus intensive diet and exercise with identical intensive diet and exercise: all-cause mortality was 1/121 versus 1/120 participants (450 participants; 2 trials; very low-quality evidence); incidence of T2DM was 48/166 versus 53/166 (RR 0.55, 95% CI 0.10 to 2.92; P = 0.49; 332 participants; 2 trials; very low-quality evidence). One trial estimated the DMC of metformin plus intensive diet and exercise to be $270 per participant compared with $225 in the comparator group (94 participants; 1 trial; very-low quality evidence). One trial in 45 participants compared metformin with a sulphonylurea. The trial reported no patient-important outcomes. For all comparisons there were no data on non-fatal myocardial infarction, non-fatal stroke or microvascular complications. We identified 11 ongoing trials which potentially could provide data of interest for this review. These trials will add a total of 17,853 participants in future updates of this review. Metformin compared with placebo or diet and exercise reduced or delayed the risk of T2DM in people at increased risk for the development of T2DM (moderate-quality evidence). However, metformin compared to intensive diet and exercise did not reduce or delay the risk of T2DM (moderate-quality evidence). Likewise, the combination of metformin and intensive diet and exercise compared to intensive diet and exercise only neither showed an advantage or disadvantage regarding the development of T2DM (very low-quality evidence). Data on patient-important outcomes such as mortality, macrovascular and microvascular diabetic complications and health-related quality of life were sparse or missing.