A systematic review pooled 21 studies and 5,650 pregnant women, comparing continuous glucose monitoring with self-monitored blood glucose in pregnancies complicated by gestational diabetes. It mixed randomised trials with observational studies, published between 1980 and April 2026, and rated risk of bias as low to moderate.
The glycaemic differences were consistent but small. Mean glucose fell by 0.24 mmol/L (95% CI -0.41 to -0.06), the coefficient of variation by 0.78% (-1.50 to -0.06), the mean amplitude of glycaemic excursions by 0.22 mmol/L (-0.43 to -0.02), and time above 7.8 mmol/L by 2.19 percentage points (-3.78 to -0.60). Alongside this, monitoring was associated with 21% more medication use and reductions of 8% to 35% in caesarean delivery, macrosomia, neonatal hypoglycaemia and hyperbilirubinaemia.
The more-medication finding is the honest reading of how monitoring works in this setting: it finds excursions that fingerprick testing at fixed times misses, and clinicians then treat them. That is a feature, not a harm, but it means a unit adopting monitoring in gestational diabetes should expect its insulin and metformin initiation rate to rise, and should plan clinic capacity accordingly.
In Indian practice, where gestational diabetes prevalence is high and many women are diagnosed late, the case for monitoring in the higher-risk subset — early diagnosis, prior gestational diabetes, obesity, poor control on diet — is stronger than the case for using it in everyone. Cost still rules out universal use in most settings, so target it.
- Target monitoring at the higher-risk group first: diagnosis before 24 weeks, previous gestational diabetes, obesity, or fasting glucose already above target on diet
- Warn the woman before starting that more data often means starting medication sooner
- Plan clinic capacity for a higher medication initiation rate, not just the sensor cost
- Keep reviewing fasting and one-hour post-meal values — those remain the treatment triggers, not the sensor summary alone
- previous gestational diabetes → early monitoring in the next pregnancy → postpartum glucose test at 6-12 weeks → annual screening thereafter
The statistics, in plain English
Every glycaemic interval here excludes zero, but only just: the upper bound for mean glucose is -0.06 mmol/L, a difference no clinician would notice in an individual patient. The outcome reductions of 8% to 35% are the clinically important part, and they are also the weakest evidence in the paper, because observational studies were pooled with randomised trials for those endpoints. Women who receive monitoring in routine care differ from those who do not, and that difference alone can produce the outcome gap.
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 diabetes & endocrinology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free