Purpose: To systematically review the evidence on benefits and harms of (1) screening for abnormal blood glucose and type 2 diabetes and (2) interventions for prediabetes or type 2 diabetes that was screen detected or recently diagnosed for populations and settings relevant to primary care in the United States.
Data Sources: PubMed/MEDLINE, the Cochrane Library, and trial registries through September 10, 2019; reference lists of retrieved articles; outside experts; and reviewers, with surveillance of the literature through May 21, 2021.
Study Selection: English-language controlled studies evaluating screening for abnormal blood glucose or evaluating interventions for prediabetes or type 2 diabetes that was screen detected or recently diagnosed.
Data Extraction: One investigator extracted data and a second checked accuracy. Two reviewers independently rated quality for all included studies using predefined criteria.
Data Synthesis: This review included 89 publications. Two randomized, controlled trials (RCTs) (ADDITION-Cambridge and Ely; described in 5 articles; 25,120 participants) evaluated invitations to screening for diabetes with (1) a stepwise approach (starting with random glucose) or (2) oral glucose tolerance test every 5 years. The trials found no significant difference between screening and control groups for all-cause or cause-specific mortality at 10 years or self-reported cardiovascular disease (CVD) events or quality of life at 7 through 13 years, but the trials were missing data from most participants for outcomes other than mortality. For harms, the trials reported no significant differences between screening and control groups for anxiety, depression, worry, or self-reported health, but one reported a short-term increase in anxiety (at 6 weeks) among persons screened and diagnosed with diabetes mellitus (DM) versus those not diagnosed with DM (State-Trait Anxiety Inventory scores: 46.7 vs. 37.0; p=0.031).
For screen-detected diabetes, one trial (ADDITION-Europe, described in 8 articles, 3,057 participants) evaluated a multifactorial intervention aimed at controlling glucose, blood pressure, and cholesterol and found no difference over 5 years in the risk of all-cause mortality, cardiovascular-related mortality, cardiovascular events, or other health outcomes between intervention and routine care groups. A post hoc analysis at about 10-years followup similarly found that differences remained nonsignificant for the primary composite outcome and for all-cause mortality. For recently diagnosed (not screen-detected) diabetes, five RCTs (8 publications, 5,138 participants) were included. In the United Kingdom Prospective Diabetes Study, long-term health outcomes were improved with intensive glucose control with sulfonylureas or insulin: decreased risk for all-cause mortality (relative risk [RR], 0.87 [95% confidence interval {CI}, 0.79 to 0.96]), diabetes-related mortality (RR, 0.83 [95% CI, 0.73 to 0.96]), and myocardial infarction (RR, 0.85 [95% CI, 0.74 to 0.97]) over 20 years (10-year post-trial assessment) but not at shorter followups. For overweight people, intensive glucose control with metformin decreased the risk for all-cause mortality (RR, 0.64 [95% CI, 0.45 to 0.91]), diabetes-related mortality [RR, 0.58 [95% CI, 0.37 to 0.91]), and myocardial infarction (RR, 0.61 [95% CI, 0.41 to 0.89]) at the 10-year followup, and benefits were maintained longer term.
For prediabetes interventions, most trials reporting on health outcomes had insufficient duration of followup for long-term health outcomes, reported few events, and found no difference between groups. One trial of a 6-year lifestyle intervention conducted in China (Da Qing, n=576) reported lower all-cause mortality (28.1% vs. 38.4% [hazard ratio {HR}, 0.71 {95% CI, 0.51 to 0.99}]) and CVD-related mortality (11.9% vs. 19.6% [HR, 0.59 {95% CI, 0.36 to 0.96}]) for intervention groups than for controls at 23 years but not at earlier followups; 30-year followup also reported lower all-cause mortality (HR, 0.74 [95% CI, 0.61 to 0.89]) and CVD-related mortality (HR, 0.67 [95% CI, 0.48 to 0.94]) for intervention groups than for controls. Lifestyle interventions (most involving >360 minutes contact) for obese or overweight people with prediabetes were associated with reductions in the incidence of diabetes (23 RCTs, pooled RR, 0.78 [95% CI, 0.69 to 0.88]). Lifestyle interventions were also associated with reduced systolic blood pressure and diastolic blood pressure (pooled weighted mean difference [WMD] −1.7 mm hg [95% CI, −2.6 to −0.8] and −1.2 mm hg [95% CI, −2.0 to −0.4], respectively), weight (pooled WMD, −1.2 kg [95% CI, −1.6 to −0.74]), and body mass index (BMI) (pooled WMD, −0.54 kg/m2 [95% CI, −0.76 to −0.33]). For medications, metformin, thiazolidinediones (TZDs), and alpha glucosidase inhibitors (AGIs) were associated with a reduction in diabetes incidence (pooled RRs, 0.73 [0.64, 0.83], 0.50 [0.28, 0.92], and 0.64 [0.43, 0.96], respectively), but evidence for TZDs and AGIs was limited by imprecision, inconsistency, and risk of bias. Most trials of medications found no statistically significant association between hypoglycemic agents and changes in blood pressure or lipids, but they did find a reduction in weight and BMI for metformin, acarbose, or liraglutide, but TZDs were associated with weight gain (pooled WMD, 1.9 kg [95% CI, 0.8 to 3.1]).
Limitations: No trials assessed initial screening with A1c or fasting glucose and none assessed screening for prediabetes. For outcomes other than mortality, screening trials were missing data from most participants. Duration of followup was too short to assess health outcomes in most studies. A single trial evaluated interventions for screen-detected diabetes. The Da Qing trial conducted in China (n=576 participants enrolled in 1986) has not been replicated and was limited by at least medium risk of bias because of unclear randomization and allocation concealment methods and baseline differences likely to bias results in favor of the intervention. Harms were rarely assessed; none of the trials reported on labeling, harms from false-positive results, burden, inconvenience, or unnecessary testing and treatment.
Conclusions: Trials of screening for diabetes found no mortality benefit at 10 years but had insufficient data to assess other health outcomes. Evidence on harms of screening was scant. For people with screen-detected diabetes, one trial found no improvement in health outcomes over 5 to 10 years. For people with recently diagnosed (not screen-detected) diabetes, interventions improved health outcomes over 10 to 20 years. For obese or overweight people with prediabetes, interventions were associated with reduced incidence of diabetes and improvement in other intermediate outcomes, and limited evidence suggests that very high contact lifestyle interventions improve health outcomes after more than 20 years.