Bianca Hemmingsen, Maria-Inti Metzendorf, Bernd Richter

People with type 1 diabetes mellitus (T1DM) need treatment with insulin for survival. Whether any particular type of (ultra-)long-acting insulin provides benefit especially regarding risk of diabetes complications and hypoglycaemia is unknown. To compare the effects of long-term treatment with (ultra-)long-acting insulin analogues to NPH insulin (neutral protamine Hagedorn) or another (ultra-)long-acting insulin analogue in people with type 1 diabetes mellitus. We searched the Cochrane Central Register of Controlled Trials, MEDLINE, Scopus, ClinicalTrials.gov, the World Health Organization (WHO) International Clinical Trials Registry Platform and the reference lists of systematic reviews, articles and health technology assessment reports. We explored the US Food and Drug Administration (FDA) and European Medical Agency (EMA) web pages. We asked pharmaceutical companies, EMA and investigators for additional data and clinical study reports (CSRs). The date of the last search of all databases was 24 August 2020. We included randomised controlled trials (RCTs) with a duration of 24 weeks or more comparing one (ultra-)long-acting insulin to NPH insulin or another (ultra-)long-acting insulin in people with T1DM. Two review authors assessed risk of bias using the new Cochrane 'Risk of bias' 2 (RoB 2) tool and extracted data. Our main outcomes were all-cause mortality, health-related quality of life (QoL), severe hypoglycaemia, non-fatal myocardial infarction/stroke (NFMI/NFS), severe nocturnal hypoglycaemia, serious adverse events (SAEs) and glycosylated haemoglobin A1c (HbA1c). We used a random-effects model to perform meta-analyses and calculated risk ratios (RRs) and odds ratios (ORs) for dichotomous outcomes and mean differences (MDs) for continuous outcomes, using 95% confidence intervals (CIs) and 95% prediction intervals for effect estimates. We evaluated the certainty of the evidence applying the GRADE instrument. We included 26 RCTs. Two studies were unpublished. We obtained CSRs, clinical study synopses or both as well as medical reviews from regulatory agencies on 23 studies which contributed to better analysis of risk of bias and improved data extraction. A total of 8784 participants were randomised: 2428 participants were allocated to NPH insulin, 2889 participants to insulin detemir, 2095 participants to insulin glargine and 1372 participants to insulin degludec. Eight studies contributing 21% of all participants comprised children. The duration of the intervention varied from 24 weeks to 104 weeks. Insulin degludec versus NPH insulin: we identified no studies comparing insulin degludec with NPH insulin. Insulin detemir versus NPH insulin (9 RCTs): five deaths reported in two studies including adults occurred in the insulin detemir group (Peto OR 4.97, 95% CI 0.79 to 31.38; 9 studies, 3334 participants; moderate-certainty evidence). Three studies with 870 participants reported QoL showing no true beneficial or harmful effect for either intervention (low-certainty evidence). There was a reduction in severe hypoglycaemia in favour of insulin detemir: 171/2019 participants (8.5%) in the insulin detemir group compared with 138/1200 participants (11.5%) in the NPH insulin group experienced severe hypoglycaemia (RR 0.69, 95% CI 0.52 to 0.92; 8 studies, 3219 participants; moderate-certainty evidence). The 95% prediction interval ranged between 0.34 and 1.39. Only 1/331 participants in the insulin detemir group compared with 0/164 participants in the NPH insulin group experienced a NFMI (1 study, 495 participants; low-certainty evidence). No study reported NFS. A total of 165/2094 participants (7.9%) in the insulin detemir group compared with 102/1238 participants (8.2%) in the NPH insulin group experienced SAEs (RR 0.95, 95% CI 0.75 to 1.21; 9 studies, 3332 participants; moderate-certainty evidence). Severe nocturnal hypoglycaemia was observed in 70/1823 participants (3.8%) in the insulin detemir group compared with 60/1102 participants (5.4%) in the NPH insulin group (RR 0.67, 95% CI 0.39 to 1.17; 7 studies, 2925 participants; moderate-certainty evidence). The MD in HbA1c comparing insulin detemir with NPH insulin was 0.01%, 95% CI -0.1 to 0.1; 8 studies, 3122 participants; moderate-certainty evidence. Insulin glargine versus NPH insulin (9 RCTs): one adult died in the NPH insulin group (Peto OR 0.14, 95% CI 0.00 to 6.98; 8 studies, 2175 participants; moderate-certainty evidence). Four studies with 1013 participants reported QoL showing no true beneficial effect or harmful effect for either intervention (low-certainty evidence). Severe hypoglycaemia was observed in 122/1191 participants (10.2%) in the insulin glargine group compared with 145/1159 participants (12.5%) in the NPH insulin group (RR 0.84, 95% CI 0.67 to 1.04; 9 studies, 2350 participants; moderate-certainty evidence). No participant experienced a NFMI and one participant in the NPH insulin group experienced a NFS in the single study reporting this outcome (585 participants; low-certainty evidence). A total of 109/1131 participants (9.6%) in the insulin glargine group compared with 110/1098 participants (10.0%) in the NPH insulin group experienced SAEs (RR 1.08, 95% CI 0.63 to 1.84; 8 studies, 2229 participants; moderate-certainty evidence). Severe nocturnal hypoglycaemia was observed in 69/938 participants (7.4%) in the insulin glargine group compared with 83/955 participants (8.7%) in the NPH insulin group (RR 0.83, 95% CI 0.62 to 1.12; 6 studies, 1893 participants; moderate-certainty evidence). The MD in HbA1c comparing insulin glargine with NPH insulin was 0.02%, 95% CI -0.1 to 0.1; 9 studies, 2285 participants; moderate-certainty evidence. Insulin detemir versus insulin glargine (2 RCTs),insulin degludec versus insulin detemir (2 RCTs), insulin degludec versus insulin glargine (4 RCTs): there was no evidence of a clinically relevant difference for all main outcomes comparing (ultra-)long-acting insulin analogues with each other. For all outcomes none of the comparisons indicated differences in tests of interaction for children versus adults. Comparing insulin detemir with NPH insulin for T1DM showed lower risk of severe hypoglycaemia in favour of insulin detemir (moderate-certainty evidence). However, the 95% prediction interval indicated inconsistency in this finding. Both insulin detemir and insulin glargine compared with NPH insulin did not show benefits or harms for severe nocturnal hypoglycaemia. For all other main outcomes with overall low risk of bias and comparing insulin analogues with each other, there was no true beneficial or harmful effect for any intervention. Data on patient-important outcomes such as QoL, macrovascular and microvascular diabetic complications were sparse or missing. No clinically relevant differences were found between children and adults.