Repaglinide/Bedtime NPH Insulin Is Comparable to Twice-Daily NPH Insulin

This trial was conducted in 12 sites in Southeast Asia and South Africa. Inclusion criteria included having type 2 diabetes, age ≥18 years, BMI ≤35 kg/m², HbA_1c (A1C) ≥7.5 and ≤13%, previous treatment with sulfonylurea and/or metformin therapy for at least 3 months before study onset (for subjects on sulfonylurea and metformin therapy, metformin dose <1,700 mg/day), and fasting C-peptide >0.33 nmol/l.

During the 3- to 4-week titration period, the initial dose was 1 mg repaglinide at each main meal, which increased to 2 mg (maximum of four meals per day) if fasting plasma glucose (FPG) was ≥7.0 mmol/l. Subjects with FPG ≥8.0 mmol/l were randomized (one to one) to 2 mg repaglinide and bedtime NPH insulin (repaglinide/NPH) or twice-daily NPH insulin (NPH).

Of the 211 subjects enrolled, 54 subjects withdrew during the titration period, mainly (n = 42) due to meeting the exclusion criterion of fasting blood glucose <7 or >15 mmol/l (titration period) or <8 mmol/l (randomization). Sixty-nine percent completed the trial as planned (71 in NPH and 74 in repaglinide/NPH). Twelve subjects withdrew (6 from each group). Two subjects (one from each group) withdrew due to adverse events; both events were judged to be unlikely related to trial products. The remaining 10 subjects withdrew due to noncompliance with protocol/medication, ineffective therapy, and withdrawal of consent.

At screening, the groups were comparable with respect to sex distribution, race, A1C, and FPG but not to age, duration of diabetes, and BMI. At the end of the titration period (baseline), mean A1C was higher in NPH (from 9.1 to 9.4%) but remained unchanged in the repaglinide/NPH group (9.2%). Subjects in the NPH group were older (57.8 ± 8.5 vs. 54.9 ± 11.1 years), had a longer duration of diabetes (9.9 ± 6.1 vs. 7.5 ± 4.7 years), but had a lower BMI (25.8 ± 4.1 vs. 26.7 ± 4.4 kg/m²) than subjects in the repaglinide/NPH group. In accordance with older age and longer duration of diabetes, the percentage of subjects with diabetes-related complications was slightly higher in the NPH group than in the repaglinide/NPH group.

At the end of the study, the treatment difference in A1C between repaglinide/NPH and NPH was 0.08 ± 0.22% (95% CI −0.35–0.51, P = 0.71) after baseline adjustment, whereas absolute values were 8.42 ± 1.40 and 8.54 ± 1.68%, respectively. The noninferiority criterion of A1C <0.6% was met. Further analysis on A1C using backwards elimination (factors included: center, A1C at baseline, diabetes duration, previous diabetes medication, baseline interaction with center, and baseline interaction with treatment) showed that A1C at baseline and previous diabetes medication were significant factors (P < 0.05). Adjusting for both factors, ANCOVA yielded similar result to the primary analysis, with a treatment difference of −0.020 ± 0.22% (−0.453 to 0.413, P = 0.93).

Likewise, mean FPG values were higher in the NPH group (from 10.9 to 12.9 mmol/l) than in the repaglinide/NPH group (from 10.7 to 12.0 mmol/l) at the end of titration. At study end, mean FPG was significantly lower in the NPH group than in the repaglinide/NPH group after adjusting for baseline (6.77 ± 1.0 vs. 8.01 ± 1.0 mmol/l, P = 0.003).

No major hypoglycaemic episodes (blood glucose <2.8 mmol/l) were reported in either group. The incidence of hypoglycemic episodes was lower in repaglinide/NPH subjects (76 events) than in NPH subjects (206 events), of which >90% were “symptomatic only” in both groups, which translates to a threefold lower risk of hypoglycemia (reciprocal of relative risk 0.35 [95% CI 0.14–0.78], P = 0.02). The risk of minor hypoglycemic episodes was not different between treatments (0.45 [0.13–1.31], P = 0.16).

The efficacy of repaglinide/NPH has been shown to be superior to NPH monotherapy (7,8) and sulfonylurea/NPH (9) but comparable to gliclazide/NPH (10) with similar hypoglycemia profiles. In this study, the efficacy of repaglinide/NPH was comparable to NPH (A1C), whereas the treatment effect on FPG was more pronounced in the NPH group than in the repaglinide/NPH group, which could be due to different injection times in the evening insulin doses between groups. Also, inherent biases in open-label trial design could favor the standard regimen with which doctors are familiar and comfortable. However, a significantly better hypoglycemia profile was achieved with repaglinide/NPH in this study, suggesting that similar glycemic control can be achieved with repaglinide/NPH but at significantly less risk of hypoglycemia. Since repaglinide was developed for use as a prandial glucose regulator, tighter glycemic control through better postprandial glucose control without increased hypoglycemia may be achieved with repaglinide/NPH, but the effects need to be assessed in further studies.

Chun-Chung Chow, Clive Cockram, Ronald Ma, Wing-Bun Chan, Risa Ozaki, Wing-Yee So (Hong Kong); Wan Mohamad Wan Bebakar, Mafauzy Mohamed, Malik Mumtaz, Ahmed B. Gulam Sarvar, Ab Aziz al-Safi bin Ismail (Malaysia); Ma Teresa Plata-Que, Pepito dela Pena, Elizabeth Paz-Pacheco, Marife Guerrero (Philippines); Hans Hendrik Snyman, Johanna Adriana Kok, Ernst Retief Snyman, Ray Moore, Hennie Nortje, Susan Lyn Brown, Angela Murphy, M.-T. van der Merwe (South Africa); Chien-Wen Chou (Taiwan), Ole Molskov Bech (Singapore).

The study was supported by Novo Nordisk.