Summary: In 542 Chinese adults with prediabetes, a 6-month structured anti-inflammatory and antioxidant diet-and-exercise programme (SAIDEs) improved glycaemic and inflammatory markers and lengthened the diabetes-free interval (log-rank p=0.001), yet time-dependent Cox regression found no significant reduction in diabetes incidence across the roughly 8-year follow-up. The intervention delayed onset rather than prevented diabetes.
PICO Summary
| Element | Detail |
|---|---|
| Population | 542 adults with prediabetes (raised FBG, 2-h OGTT, or HbA1c) in Huzhou, China; prospective allocation by block randomisation into four arms with assessments at 6 months and ~7.5-8 years. |
| Intervention | 6-month structured anti-inflammatory and antioxidant dietary plus exercise programme (SAIDEs); single combined arm (one of four randomised groups). |
| Comparison | Three other randomised arms: routine community intervention, diet-only, and exercise-only. |
| Outcome | Primary incidence: time-dependent Cox regression showed no significant effect of diet or exercise intervention on diabetes incidence over the ~8-year follow-up. Survival distribution differed across the four arms (log-rank chi-square=15.63, p=0.001), with the longest mean diabetes-free time in the SAIDEs arm. Short-term (6-month) gains: lower body weight, blood pressure, cholesterol and glucose, improved beta-cell function, and reduced IgE, hs-CRP, IL-6 and TNF-alpha with higher IL-4 and IL-10 (all p<0.001). No hazard ratios, 95% CIs, ARR or NNT for the incidence endpoint were reported. |
Expert Commentary
This trial is best read as a negative result on its headline question dressed in encouraging short-term data. A 6-month anti-inflammatory diet-and-exercise package clearly improved weight, glycaemia, beta-cell function and a panel of inflammatory markers at six months, and the SAIDEs arm enjoyed the longest diabetes-free interval, with the log-rank comparison reaching p=0.001. The pivotal finding, however, is that the time-dependent Cox regression detected no significant reduction in diabetes incidence over roughly eight years. Onset was postponed, not averted, which is a meaningfully weaker claim than prevention. The single largest limitation is the absence of any reported hazard ratio, confidence interval, absolute risk reduction or number needed to treat for the incidence endpoint, so the magnitude and precision of any benefit cannot be judged, and a positive log-rank test with a null Cox model is a fragile basis for practice change. The metabolic improvements may also reflect short-term adherence that wanes once supervision ends. Can I use this with my patients? Cautiously: the data reinforce advising a structured diet-and-exercise programme for motivated adults with prediabetes to buy time and improve intermediate markers, but they do not support promising durable diabetes prevention from a brief intervention. Longer, fully powered trials reporting absolute event data are needed before stronger claims are warranted. One author holds a Harvard-affiliated post and the work received non-US-government and NIH extramural support, which should be noted though it is unlikely to drive a null primary result.
References
Wang Z, Qian L, Shen JT, Wang B, Shen XH, Shi GP. Short-term structured dietary and exercise interventions delay diabetes onset in prediabetic patients: a prospective quasi-experimental study. Front Endocrinol (Lausanne). 2025;16:1413206. doi:10.3389/fendo.2025.1413206
