Introduction
Material and methods
Search and study eligibility
Data extraction and risk of bias assessment
Results
Study | Setting | Age years, mean (SD) | Female percent | Functional status at baseline mean (SD) | Comparison | Outcomes reported | Risk of Bias |
---|---|---|---|---|---|---|---|
Schwartz 2002 [33] n = 629 | Community-dwelling | 73.8 (5.0)a | 100 | Any difficulty in IADL: 59%a Grip strength: 17.9 kg (4.7)a Quadriceps strength: 55.4 kg (27.5)a 5‑chair test: 13.8 s (5.2)a Poor tandem stand: 60%a 2‑m tandem walk: 19.1 s (8.8)a Walking speed: 0.8 m/s (0.2)a | Insulin vs. no insulin | Falls | Serious |
Pham 2003 [30] n = 73 | Nursing and residential homes | 76.0 (7.9) | 63a | Katz ADL: 70% dependent | Insulin vs. oral drugs vs. no pharmacological antidiabetic treatment HbA1c (%): <6.5% vs. 6.5–8% vs. >8% | Hypoglycemia Depression ADL Mortality | Serious |
Wang 2011 [43] n = 119 | Community-dwelling | 76.3 (3.4) | 55 | SPPB: 7.6 (3.4) | HbA1c: poorer vs. better control class | Lower extremity function | Serious |
Yau 2012 [46] n = 367 | Community-dwelling with need for skilled help | 80 (9) | 67 | ADL-score ≤8: 66% | HbA1c (%): <7.0% vs. 7.0–7.9% vs. 8.0–8.9% vs. ≥9.0% | Functional decline/death | Serious |
Sardu 2014 [32] n = 32 | Hospital patients with cardiac resynchronization therapy | 81.7 (6.3) | 41 | 6‑min walk: 249 m (50.7) | Insulin vs. no insulin | 6‑min walk HbA1c NYHA class | Serious |