Aim: To estimate absolute and relative incidence rates of hypoglycaemia when using once‐daily evening or morning regimens of insulin glargine (glargine) versus once‐daily evening NPH insulin (NPH) using individual patient data (IPD).
Materials and methods: Randomized controlled trials with accessible IPD and including white European people with type 2 diabetes (T2DM) using glargine or NPH once‐daily (with oral glucose‐lowering drugs) were identified. Two study pools were analysed: evening glargine versus evening NPH (pool 1); and morning glargine versus evening NPH (pool 2). The number‐needed‐to‐treat to avoid hypoglycaemia was calculated for glargine versus NPH.
Results: In study pool 1 (n = 2711), the risk of nocturnal hypoglycaemia was approximately halved with glargine compared with NPH [odds ratios (OR): 0.44–0.52, p < 0.001–0.047]. This led to a significant reduction in anytime risk of symptomatic hypoglycaemia [plasma glucose (PG) <3.9 mmol/l, OR: 0.64, p = 0.018; PG <2.0 mmol/l, OR: 0.51, p < 0.001]. In study pool 2 (n = 470), although a strong numerical reduction in all types of nocturnal hypoglycaemia was observed (OR: 0.16–0.64), statistical significance was reached only for symptomatic hypoglycaemia with PG <3.9 mmol/l (p < 0.001). Eight (pool 1) or five (pool 2) people with T2DM needed to use glargine rather than NPH to avoid one person from experiencing a nocturnal symptomatic hypoglycaemic event within a median of about 25 weeks of starting insulin.
Conclusions: This meta‐analysis of open‐label studies provides confidence that reductions of around 50% of risk for nocturnal hypoglycaemia can be achieved with using glargine instead of NPH.