Design: Systematic review and network meta-analysis of randomised trials
Population: 27 RCTs, 965 adults with type 2 diabetes
Primary outcome: Glycaemic, lipid, body composition and blood pressure measures
Effect: Both orders beat control on HbA1c and lipids; no significant difference between orders

A network meta-analysis in Endocrine pooled 27 randomised trials, 965 adults with type 2 diabetes in total, comparing combined training done aerobic-then-resistance with resistance-then-aerobic, and with no exercise.

Both sequences improved body-fat percentage, HbA1c, fasting glucose, total cholesterol, triglycerides and LDL cholesterol compared with control. Rankings placed resistance-first ahead for glycaemia and body composition, and aerobic-first ahead for blood pressure. But the direct comparisons between the two orders were not statistically significant, and the authors rated certainty as low.

The useful message for clinic is the one that was already true: combined training works, and the order is a matter of preference. Telling a patient that one order is better would be reading more into ranking probabilities than they can carry.

Recommend combined aerobic and resistance training; either order is reasonable.
Let the patient pick the sequence they are most likely to keep doing.
Rankings from a network meta-analysis are not head-to-head evidence of superiority.
Trials here were small, with 965 participants across 27 studies.

Why it matters

It removes a detail patients sometimes worry about and keeps the focus on doing both kinds of exercise.

The statistics, in plain English

SUCRA rankings give the probability that an option is best, but they can favour one option even when the direct comparison shows no significant difference. Here the head-to-head comparisons between the two orders were not significant, and certainty was rated low, so the rankings are hypotheses rather than findings.