Results
Enrolment occurred between February 2013 and February 2014, and the last participant completed the trial in August 2014. Figure 1 shows the flow of participants through the trial, including those who were lost to follow-up and discontinued the trial.
(Enlarge Image)
Figure 1.
Flow of participants through trial
Full Analysis Set Evaluation
In total 124 participants were randomised and 122 had at least one follow-up measurement and were included in the full analysis set used for the primary efficacy analysis. Table 1 shows the baseline characteristics of the participants. The HbA1c levels in the full analysis set were 74.6 mmol/mol (8.98%) for participants in the liraglutide group and 74.4 mmol/mol (8.96%) for those in the placebo group. Age, sex, body mass index, total daily insulin dose, number of insulin injections, duration of diabetes, and other participant characteristics were similar at baseline between the groups.
Table 2 shows the results of the predefined endpoints in the evaluation of the full analysis set. The HbA1c concentration was reduced by 16.9 mmol/mol (1.54%) in the liraglutide group compared with 4.57 mmol/mol (0.42%) in the placebo group (difference −12.3 mmol/mol, 95% confidence interval −15.8 to −8.8 and −1.13%, −1.45 to −0.81; P<0.001). More participants in the liraglutide group (42.9%) than in the placebo group (5.1%) reached an HbA1c concentration <53 mmol/mol (7.0%; P<0.001). A greater proportion of participants treated with liraglutide also reached HbA1c concentrations <58 mmol/mol (7.5%) and 64 mmol/mol (8.0%, Table 2).
Body weight was reduced in participants treated with liraglutide, by 3.8 kg, but not in participants treated with placebo (0.0 kg, difference −3.8 kg, 95% confidence interval −4.9 to −2.8; P<0.001). The daily insulin dose was reduced by 18.1 units with liraglutide and by 2.3 units with placebo (difference −15.8 units, 95% confidence interval −23.1 to −8.5; P<0.001). In the liraglutide group the total daily basal insulin dose was reduced by 6.8 units and total daily mealtime insulin dose by 11.2 units. In the placebo group the corresponding reductions were 0.5 units and 1.9 units.
Reductions in fasting and postprandial glucose levels were greater in the liraglutide group than in the placebo group (difference −1.5 and −2.0 mmol/L (−27 and −36 mg/dL), respectively; P<0.01 for both). Reductions in mean glucose levels and standard deviation estimated by masked continuous glucose monitoring were significantly greater in the liraglutide group than in the placebo group (−1.9 and −0.5 mmol/L (−34.2 and −9 mg/dL), respectively; P<0.001).
Neither group experienced severe hypoglycaemic events and no significant differences were found in symptomatic or asymptomatic non-severe hypoglycaemias, with values <4.0 or <3.0 mmol/L (<72 or <54 mg/dL) between the liraglutide and placebo groups (Table 2). Non-severe symptomatic hypoglycaemia <4.0 mmol/L was most common, with a mean of 1.29 events during follow-up in the liraglutide group and 1.24 in the placebo group (P=0.96).
Treatment satisfaction was significantly greater with liraglutide than with placebo, when estimated by both of the diabetes treatment satisfaction questionnaires (current and retrospective) (Table 2).
Systolic blood pressure decreased by 4.62 mm Hg in participants treated with liraglutide and increased by 0.86 mm Hg in participants treated with placebo (difference −5.47 mm Hg, 95% confidence interval −9.85 to −1.10, P=0.015). No significant difference was found for diastolic blood pressure, as was the case for levels of total cholesterol, low density lipoprotein cholesterol, high density lipoprotein, and triglycerides (Table 2).
Of the 64 participants in the liraglutide group, 57 (89.1%) received a dose of 1.8 mg, 3 (4.7%) 1.2 mg, and 4 (6.3%) 0.6 mg at the end of follow-up, whereas all 60 participants in the placebo group received 1.8 mg.
Withdrawals
Five participants withdrew overall: three in the liraglutide group and two in the placebo group.
Rescue Treatment
One participant (1.6%) in the liraglutide group and three (5%) in the placebo group received rescue treatment. Rescue treatment was initiated during the first 12 weeks in two participants in the placebo group and during the second half of the trial in one participant in each group.
Per Protocol Evaluations
Overall, 58 (90.6%) participants in the liraglutide group and 55 (91.7%) in the placebo group were included in the predefined per protocol population. Supplementary etable 1 http://www.bmj.com/content/bmj/suppl/2015/10/28/bmj.h5364.DC1/linm027803.ww1_default.pdf shows the reasons for excluding participants from the per protocol analysis. The difference in the change in HbA1c concentrations from baseline to 24 weeks between groups in favour of liraglutide was −12.4 mmol/mol (95% confidence interval −16.0 to −8.8 and −1.14%, −1.47 to −0.80; P<0.001). The same endpoints as in the intention to treat analysis became significant in favour of liraglutide in the per protocol analysis (see supplementary etable 2 http://www.bmj.com/content/bmj/suppl/2015/10/28/bmj.h5364.DC1/linm027803.ww1_default.pdf).
Time Evaluations of Change in HbA1c Levels, Weight, and Total Daily Insulin Dose
Figure 2 shows the change in HbA1c levels, weight, and total daily insulin dose for participants in the liraglutide and placebo groups at baseline and every six weeks. There was a significantly greater reduction in HbA1c concentration (−9.4 mmol/mol, 0.86%), weight (−2.4 kg), and insulin dose (−15.1 units) already at six weeks (P<0.001 for all) in participants treated with liraglutide compared with placebo.
(Enlarge Image)
Figure 2.
Change in HbA1c concentration, weight, and daily insulin dose by treatment group over time (mean and 95% confidence interval). IFCC= International Federation of Clinical Chemistry; LOCF=last observation carried forward
Adverse Events
During follow-up, gastrointestinal symptoms were experienced by 30 (46.9%) participants in the liraglutide group and eight (13.3%) in the placebo group. The corresponding numbers for nausea were 21 (32.8%) and 1 (1.7%) and for diarrhoea were 5 (7.8%) and 3 (5.0%). Gastrointestinal symptoms were mainly present at the start of the trial and at 24 weeks. Only two (3%) participants in the liraglutide group had nausea (fig 3). Overall, three (4.7%) participants in the liraglutide group and four (6.7%) in the placebo group had any serious adverse event (Table 3).
(Enlarge Image)
Figure 3.
Incidence of nausea (%) by treatment group over time (safety population)
Sensitivity Analysis
We performed a sensitivity analysis on all randomised participants (n=124)—that is, also including the two participants without any valid follow-up measurements who were not included in the primary efficacy analysis (full analysis set evaluation). All predefined endpoints were evaluated (see supplementary etable 3 http://www.bmj.com/content/bmj/suppl/2015/10/28/bmj.h5364.DC1/linm027803.ww1_default.pdf). The results remained unchanged; all endpoints that were highly significant in favour of liraglutide in the primary efficacy analysis remained so, with a greater reduction in HbA1c concentrations of 12.1 mmol/mol (1.1%), compared with participants in the placebo group.