Older Adults With Type 2 Diabetes Need More Than HbA1c

HbA1c matters. So does whether you can climb the stairs.

Type 2 diabetes care in later life often centers on HbA1c, the blood marker that reflects average glucose over roughly three months. That target matters, but it cannot tell you whether someone is losing strength, walking more slowly, or becoming less able to live independently. Older adults with diabetes need a plan that treats glucose control and physical function as connected problems.

A 2024 network meta-analysis by Yu et al. makes the tension visible. Across 27 randomized trials involving 2,289 older adults with type 2 diabetes, health worker-led management ranked best for lowering HbA1c and fasting blood glucose. Exercise combined with cognitive training ranked best for muscle strength, gait speed, and dynamic balance. No tested intervention ranked at the top on both metabolic and functional outcomes.

That finding does not mean people must choose between lower glucose and stronger legs. It means the trials tested several partial solutions. A separate one-year randomized trial offers a more useful proof of concept: a structured lifestyle program that combined dietary restriction with multimodal exercise improved both sets of outcomes in the same older adults.

The best evidence for a dual-purpose program

In Celli et al.'s 2022 randomized controlled trial, 100 older adults with type 2 diabetes and multiple comorbidities received either an intensive lifestyle intervention or a healthy-lifestyle control program for one year. The intervention paired a calorie-restricted diet with multimodal exercise, initially delivered in a research facility and later transitioned to community centers and home.

The differences were large and statistically significant. HbA1c fell by 0.8% (±0.1) in the intensive-intervention group and rose by 0.1% (±0.1) in the control group (between-group P<0.001). Insulin sensitivity increased by 1.2 (±0.2) versus a decrease of 0.4 (±0.2), and body weight changed by −8.4 (±0.6) kg versus −0.3 (±0.6) kg; both comparisons had P<0.001. Visceral fat fell by 261 (±29) cm³ compared with 30 (±27) cm³ in the control group (P<0.001).

The physical results moved in the same direction. Physical Performance Test score increased by 2.9 (±0.6) points versus 0.1 (±0.4) points in the control group, while peak oxygen uptake rose by 2.2 (±0.3) mL/kg/min versus a decline of 1.2 (±0.2) mL/kg/min (both P<0.001). Grip strength, gait speed, and the SF-36 physical-component score also improved relative to control, with reported P values below 0.001. Daily insulin dose fell by 19.8 (±4.4) units in the intervention group.

Reduced insulin use is clinically relevant, but it does not prove that every older adult can or should reduce medication. Medication changes require monitoring, especially when food intake and exercise change together. The trial does show that a sustained, supported lifestyle program can improve glucose, fitness, and functional performance at once.

Why 12 weeks may be too short for HbA1c

Expectations should match the marker. A 2025 study of 43 older adults with type 2 diabetes, mean age 69.8 years, tested 12 weeks of moderate combined aerobic and resistance exercise. Fructosamine, which reflects glycemia over roughly two to three weeks, improved. HbA1c did not change significantly.

This result does not cancel the one-year trial. The programs differed in duration and support, and fructosamine and HbA1c capture different periods of glucose exposure. HbA1c is a slow-moving outcome; a 12-week intervention only partly overlaps its measurement window. The shorter study also provides weaker causal evidence than a controlled long-term trial if, as reported, its available description does not establish a formal control group. It is best read as a warning against judging an exercise plan by an HbA1c test too soon.

A program can improve near-term glycemia before HbA1c has enough time to show the change.

What belongs in the exercise prescription

The diabetes-specific network meta-analysis favors exercise plus cognitive training for strength, gait speed, and dynamic balance. Its ranking does not establish an exact workout prescription, and its individual networks may be small. Still, broader evidence points in the same direction: older people gain more functional value from a program that includes more than one exercise mode.

Two network meta-analyses in older adults with sarcopenia—not specifically diabetes—found multicomponent exercise performed best for strength and physical performance. Shen et al. reported the strongest improvements in handgrip strength and physical function with multicomponent exercise, while Zeng et al. similarly ranked combined or resistance-dominant approaches highly for handgrip strength, gait speed, and timed-up-and-go performance. Those findings support including aerobic work, resistance training, and balance or functional practice. They do not guarantee the same effect size in diabetes, where neuropathy, vascular disease, and medication burden can change both exercise tolerance and response.

Diet should not be treated as background decoration. A 2025 network meta-analysis of cardiovascular-health interventions found that exercise combined with dietary intervention outperformed either component alone for cardiovascular markers. That does not isolate how much diet versus exercise produced each result in the Celli trial, but it supports the trial's combined design.

Cognition and quality of life: promising, but indirect for diabetes

Physical function and cognition often decline together in older age. Evidence outside diabetes suggests exercise can help both. A 2025 meta-analysis of 42 randomized trials involving 4,740 older adults with cognitive or physical frailty found small-to-moderate improvements in global cognition and well-being with physical-activity interventions, including multicomponent, resistance, aerobic, and mindfulness-based programs (Chang et al.). A separate meta-analysis found that aerobic and resistance training improved cognitive outcomes in adults over 60 with mild cognitive impairment (Li et al.).

There are limits. These were not trials restricted to older adults with type 2 diabetes, so they cannot quantify cognitive benefit for that group. Results on quality of life also deserve careful reading: Yuan et al. found improvements in cognition and several physical-function measures in cognitively frail older adults, but the overall effect on quality of life was not significant. Exercise may improve life more than a questionnaire can capture, but the trial result remains the trial result.

Longer programs may matter here as well. A 2024 review of interventions lasting at least one year found that long-term exercise and multidomain programs reduced the risk of mild cognitive impairment and dementia progression in older adults without dementia (Reparaz-Escudero et al.). In the 18-month EXERT randomized trial, moderate-to-high-intensity aerobic exercise and a low-intensity stretching, balance, and range-of-motion comparison program did not differ on the primary global cognitive outcome, although exercise slowed Alzheimer's biomarker progression in sedentary older adults with amnestic mild cognitive impairment (Baker et al.). Biomarker change is not proof that exercise prevents Alzheimer's disease.

What the evidence supports—and what it does not

The strongest practical conclusion is modest but useful. For an older adult with type 2 diabetes, a long-term program that combines dietary support with aerobic, resistance, and functional exercise can improve HbA1c and physical performance simultaneously. The best direct trial lasted one year and used intensive supervision and behavioral support. An eight- or twelve-week referral may improve short-term glucose markers and fitness, but it should not be expected to reproduce the full HbA1c and function results.

The evidence does not identify one universally optimal exercise dose, prove that a home-based program produces the same results as the research-supported intervention, or show that diabetes drugs themselves cause sarcopenia or functional decline. The Yu meta-analysis also provides directional rankings, not a precision-engineered menu: 27 trials spread across seven intervention categories leave some network comparisons uncertain.

For patients and clinicians, the useful question is broader than "How do we lower HbA1c?" It is whether the plan also preserves strength, walking capacity, balance, and the ability to do daily life. The one-year trial suggests that treating those goals together can reduce HbA1c by 0.8% while improving physical performance—not by magic, but by sustained work with enough support to keep it going.