Resistance training improved HbA1c versus control in DARE but did not reach statistical significance in HART-D.
This guide compares three type 2 diabetes trial designs. Read the control condition and time commitment alongside each result.
Illustration note: The exercise scenes are illustrations, not verified study participants, prescribed sessions or measured glucose responses.
The HART-D Comparisons
HbA1c reflects average glucose over about three months.
HART-D assigned 262 sedentary adults with type 2 diabetes to aerobic, resistance, combined exercise or nonexercise control for nine months. Active programmes had approximately similar weekly time commitments.
| Exercise | HbA1c difference | 95% confidence interval |
|---|---|---|
| Aerobic | -0.24 | -0.55 to +0.07 |
| Resistance | -0.16 | -0.46 to +0.15 |
| Combined | -0.34 | -0.64 to -0.03 |
The combined comparison was statistically significant; the single-mode comparisons were not. Negative values indicate lower HbA1c than control.
Interpreting HART-D
A significant comparison with control does not prove superiority over another active group. Nonsignificance does not establish zero effect.
For the broader timing question, see our lifting and blood-sugar evidence guide. This page focuses on the comparison each programme tested.
Two Other Trial Designs
| Study | Design | Finding and limitation |
|---|---|---|
| DARE, 2007 | 251 adults aged 39 to 70 with type 2 diabetes; 22 weeks of exercise, with HbA1c measured at six months. Three exercise groups and a nonexercise control. | Combined training improved HbA1c more than either single mode, but involved more total exercise time. The benefit cannot isolate exercise mix from exercise amount. |
| IDES, 2010 | 606 sedentary participants with type 2 diabetes and metabolic syndrome in 22 clinics; 12 months of supervised aerobic and resistance exercise plus counseling versus counseling alone. | HbA1c difference: -0.30 percentage points, 95% confidence interval -0.49 to -0.10. With no single-mode arms, this trial cannot rank them against combined training. |
A Comparison Record
Use this original reading record for one exercise headline. It organizes evidence, not a workout or a clinical decision.
- Population: Who was enrolled, and who was excluded?
- Comparison: What did each group actually receive?
- Amount and time: Did both programmes require similar time, and when was the outcome measured?
- Result: Record the estimate and uncertainty before writing a conclusion.
What These Results Leave Open
Group HbA1c changes over months do not predict an individual’s glucose reading after one workout.
DARE used a run-in period and exercise-test or cardiologist clearance. These enrollment conditions limit generalization; they are not a universal screening rule.
Keep Personal Care Separate
Ask your diabetes-care team how exercise, medication and monitoring should fit your circumstances. This article cannot select treatment changes or an individual exercise progression.
A Practical Takeaway
Save the source link with your completed comparison record. Note which unanswered question you would ask before using the headline in a personal decision.
Want more like this? Subscribe to WorkoutHealthy Insider for practical explanations behind popular fitness claims.
Sources and Limits
Church et al. (2010). JAMA 304:2253-2262. HART-D abstract.
Sigal et al. (2007). Annals of Internal Medicine 147:357-369. DARE abstract.
Balducci et al. (2010). Archives of Internal Medicine 170:1794-1803. IDES abstract.
NIDDK. The A1C Test and Diabetes.
ADA Professional Practice Committee (2026). Diabetes Care 49(Suppl 1):S89-S131. Selected care guidance.
These summaries use received abstracts and a previously reviewed HART-D full report.
Medical disclaimer: This is general research education, not a personal diabetes-management plan. Follow your diabetes-care team’s exercise, medication and monitoring instructions.
