HumAn-1 was an open-label randomised trial at one site in Bangladesh and two in Tanzania. Four hundred people aged 7 to 25 with type 1 diabetes were randomised to insulin glargine or to continue usual care, which meant human isophane insulin or premixed 70/30. The two co-primary outcomes were measured on blinded continuous glucose monitors at 6 months: time in very low range (below 3 mmol/L) and time in target range (3.9 to 10.0 mmol/L).
Neither moved. Time in very low range was 3.6% with glargine and 3.4% with usual care, adjusted mean difference 0.22% (97.5% CI -0.83 to 1.27, p=0.63). Time in target range was 40.5% versus 38.1%, adjusted mean difference 0.55% (97.5% CI -2.78 to 3.89, p=0.71). Serious adverse events were uncommon in both arms: five participants in the glargine group, 13 in usual care.
The number that should hold your attention is not the difference but the level. Both groups spent about 40% of the day in range. That is a long way from the 70% target, and it did not improve when the basal insulin was switched to a modern analogue. What limits glycaemia in these settings is supply reliability, monitoring access, education and food security, not the molecule. Substituting a costlier insulin without fixing the rest buys very little.
This matters directly in Indian practice, where cost drives basal insulin choice constantly and human insulin remains widely used in public and low-cost settings. This trial does not say analogues are useless — the individual patient with recurrent nocturnal hypoglycaemia on isophane still has a case. It says that as a population-level substitution, in a system without reliable monitoring, it does not deliver.
- Do not assume switching from isophane to glargine will fix poor control; check monitoring frequency, insulin storage and dose adjustment skills first
- Reserve the analogue for a documented problem it can solve — recurrent nocturnal hypoglycaemia, or an inability to take a second daily injection
- Where money is limited, spending it on glucose testing supplies and structured education is likely to buy more time in range than a basal insulin upgrade
- Record time in range or an equivalent measure at every review, so that any switch has a before-and-after to justify it
- Ask directly about insulin storage in hot months — degraded insulin looks exactly like treatment failure
The statistics, in plain English
Both confidence intervals comfortably include zero difference, and they are narrow enough to be informative rather than merely inconclusive: the trial can rule out anything bigger than about a 1.3 percentage point increase in time below 3 mmol/L, and about a 3.9 point gain in time in range. A gain of under 4 percentage points of time in range is roughly an hour a day — real, but not what a clinician expects when they pay several times more for an insulin. The wider point is that the trial was powered to detect a meaningful benefit and found none.
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