- Design
- Systematic review and meta-analysis of 8 RCTs
- Population
- 483 critically ill and 137 long-term care patients with hyperglycaemia
- Primary outcome
- Glycaemic control
- Effect
- ICU mean glucose −0.54 mmol/L (−0.78 to −0.31); long-term care HbA1c −0.92%
A systematic review in BMJ Open Diabetes Research & Care (9 September) included eight randomised trials comparing diabetes-specific tube-feeding formulas with standard formulas in patients with hyperglycaemia: five in critical care (483 patients) and three in long-term care (137).
In critically ill patients, diabetes-specific formulas lowered mean glucose by 0.54 mmol/L (95% CI 0.31 to 0.78; I² 9%, moderate certainty) and reduced glucose variability, though those estimates were highly heterogeneous (I² 98%). In long-term care, HbA1c fell by 0.92% (low certainty). Effects on infections, length of stay and mortality could not be established.
A half-millimole fall in mean glucose is small. The formulas may make glucose control smoother and reduce insulin requirements, but they have not been shown to change outcomes that matter to patients.
- Consider a diabetes-specific formula when enteral feeding makes glucose hard to control
- Keep insulin protocols and glucose targets unchanged
- Watch for hypoglycaemia if feeds are interrupted
- Do not expect outcome benefit beyond glucose control
Why it matters
It sizes a common product choice honestly: a small glucose effect without evidence on outcomes.
Don't overread it
The review found no evidence that these formulas reduce infections, ICU stay or death.
The statistics, in plain English
The mean glucose difference has low heterogeneity (I² 9%), so it is consistent across trials. The variability measures have I² of 98%, meaning trials disagreed so much that the pooled figure is unreliable. With only eight small trials, patient outcomes could not be assessed.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for emergency & critical care, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free