Comparison of efficacy and safety of insulin degludec/liraglutide and insulin glargine U‐100/lixisenatide in individuals with type 2 diabetes mellitus using professional continuous glucose monitoring

ABSTRACT

Aim/Introduction

Insulin glargine U100/lixisenatide and insulin degludec/liraglutide are fixed‐ratio combinations containing basal insulin and a glucagon‐like peptide‐1 receptor agonist capable of reducing both fasting and postprandial blood glucose levels with a single formulation. This study aimed to compare the time in range (TIR) and the time below range (TBR) level 1 using professional continuous glucose monitoring and to establish criteria for the differential use of the fixed‐ratio combinations.

Materials and Methods

Thirty‐six outpatients with type 2 diabetes mellitus (24 men and 12 women; average age, 62.1 years) were randomly assigned to the groups. At 0 and 18 weeks, a device was worn to compare the TIR and TBR level 1. The correlation between the C‐peptide index at baseline and TIR at 18 weeks was assessed.

Results

The TIR and TBR level 1 showed no significant differences between the two groups. Both groups showed significant positive correlations between the C‐peptide index and the TIR (P = 0.002, r = 0.679; P = 0.002, r = 0.681, respectively). The changes in glycemic variability, therapeutic indices, and body mass index were not significantly different among the groups (P > 0.05). The receiver operating curve analysis revealed that the cut‐off values of the C‐peptide index to achieve TIR of >70% at 18 weeks were 1.258 (sensitivity, 77.8%; specificity, 100%) and 1.099 (sensitivity, 57.1%; specificity, 90.9%) in the insulin glargine U100/lixisenatide and insulin degludec/liraglutide groups, respectively.

Conclusions

A TIR of >70% was achieved for both fixed‐ratio combinations without significant differences.

A comparison of IGlarLixi and IDegLira showed that a time in range >70% was achieved with no significant difference. The time in range was significantly correlated with the C‐peptide index at the start of treatment. The decision to choose IGlarLixi or IDegLira treatment with reference to each person's C‐peptide index would be useful in preventing future complications of diabetes.

INTRODUCTION

Early and inappropriate treatment of type 2 diabetes mellitus and persistent hyperglycemia can lead to diabetic microangiopathy, such as diabetic retinopathy, nephropathy, and peripheral neuropathy, and an increased risk of death owing to complications of ischemic cardiovascular disease, such as angina pectoris, myocardial infarction, stroke, and macroangiopathy, including arteriosclerosis obliterans ¹ , ² . Inappropriate glycemic management is caused by clinical inertia, such as delays in initiating or intensifying treatment ³ . Conversely, early and rigorous intervention increases the likelihood of achieving HbA1c treatment targets in individuals with type 2 diabetes mellitus ⁴ . Treatment choices depend on the conditions of those with diabetes. Since Japanese people often develop diabetes due to decreased insulin secretion, compared with Westerners ⁵ , ⁶ , early insulin introduction occurs in many cases. Therefore, the consensus statement from the Japan Diabetes Society titled ‘A proposed algorithm for pharmacotherapy in people with type 2 diabetes’ recommends an initial assessment of the absolute or relative indications for insulin therapy ⁷ . In relative indications, initial treatment approaches include basal‐supported oral therapy (BOT) using basal insulin. According to clinical recommendations, microvascular complications may be prevented with an HbA1c <7.0%, a corresponding fasting glucose level of 80–130 mg/dL, and a 1–2 h postprandial glucose level < 180 mg/dL ⁸ . However, basal‐supported oral therapy does not meet three medical needs that present barriers to achieving this recommendation. First, 35.6% of Japanese individuals with type 2 diabetes mellitus who received BOT had HbA1c levels >7.0% due to residual postprandial hyperglycemia, even when BOT reduced the fasting glucose level to the target range ⁹ . Thus, BOT often fails to achieve the target HbA1c level. Second, increasing long‐acting insulin doses to intensify therapy lowers the HbA1c levels but increases the risk of hypoglycemia; therefore, physicians are unable to implement treatment intensification for 75.5% of individuals with type 2 diabetes mellitus ¹⁰ . Third, insulin treatment can lead to weight gain, which can increase cardiovascular risk ¹¹ .

The glucagon‐like peptide‐1 receptor agonist (GLP‐1 RA) is one solution to these unmet medical needs. GLP‐1 RA stimulates insulin secretion in a glucose‐dependent manner and inhibits glucagon secretion, thereby maintaining glucose homeostasis ¹² , allowing central appetite suppression and delayed gastric content excretion ¹³ , and reducing the risk of hypoglycemia and weight gain. GLP‐1 RA also does not meet two medical needs. First, GLP‐1 RA may not adequately reduce fasting blood glucose levels due to its glucose‐dependent action ¹⁴ . Second, GLP‐1 RA may not be fully effective unless endogenous insulin secretory function is maintained to some extent ¹⁵ . Thus, combining long‐acting insulin and GLP‐1 RA can effectively lower fasting and postprandial glucose levels. However, concerns remain regarding treatment complexity due to the increased number of injections, necessitating treatment simplification. In Japan, two types of fixed‐ratio combinations (FRCs) of basal insulin and GLP‐1 RA are available: insulin glargine U100/lixisenatide (IGlarLixi) (Soliqua®; Sanofi, Paris, France) and insulin degludec/liraglutide (IDegLira; Xultophy®; Novo Nordisk A/S, Bagsværd, Denmark). The authors compared IGlarLixi pre‐breakfast and insulin glulisine (IGlu) pre‐supper injection therapy with multiple daily injections using professional continuous glucose monitoring (proCGM) (Freestyle Libre Pro™; Abbott Diabetes Care, Alameda, CA). The time in range (TIR) ¹⁶ percent of readings and the time at 70–180 mg/dL were equivalent with IGlarLixi and IGlu therapies, compared with those of multiple daily injections (MDI). However, the time below range (TBR) level 1 ¹⁶ percent of readings and the time at 54–69 mg/dL were significantly lower with IGlarLixi and IGlu therapies than with MDI and IGlarLixi and IGlu therapies. A significant negative correlation was identified between C‐peptide immunoreactivity (CPR) values and IGlu dosage ¹⁷ . Furthermore, the authors compared IDegLira and insulin degludec/insulin aspart (IDegAsp) injected pre‐supper and reported that IDegLira had a significantly higher TIR than IDegAsp, suppressed postprandial glucose levels at breakfast and lunch, and had a significant negative correlation with a 24 h coefficient of variation (CV) of glycemic variability (GV) and C‐peptide index (CPI) ¹⁸ . Analysis of CGM data with a time in range of 70–180 mg/dL confirmed a TIR of 70% with an HbA1c level of 7.0% ¹⁹ ; thus, a time in range of >70% is recommended for treatment with CGM ⁸ . In our previous study, TIR accounted for very high percentages of IGlarLixi and IGlu treatment (93.1%) ¹⁷ and IDegLira treatment (86.3%) ¹⁸ . In actual clinical practice, diabetes injection therapy is often initiated in an outpatient setting. However, these studies were conducted in a specialized environment, where dietary and exercise therapy were carefully managed, and oral or injection therapies were consistently administered during inpatient care. No direct comparative reports are available for the introduction of injection therapy with IGlarLixi and IDegLira in outpatient clinical settings, and the ideal patient profile for each of these fixed‐ratio combinations is not clearly defined. In previous studies, the effectiveness and safety of fixed‐ratio combinations were suggested to be related to endogenous insulin secretion capacity ¹⁷ , ¹⁸ . Therefore, this study aimed to determine indicators for selecting the two fixed‐ratio combinations using proCGM to clarify the relationship between time in range and endogenous insulin secretion capacity. As a result, achieving a time in range of >70% may lead to an HbA1c level of <7.0%, contributing to preventing diabetes complications. This treatment outcome may allow individuals with diabetes to have the same life expectancy and quality of life as those without diabetes.

MATERIALS AND METHODS

Study design and participants

This randomized, non‐blinded, parallel‐group comparison study of individuals with type 2 diabetes mellitus was conducted from June 2022 to July 2023.

This study was conducted in accordance with the Declaration of Helsinki (1975, as revised in 2013). The participants were provided with explanations about the research's importance, objectives, methods, and other aspects. Informed consent was obtained from all participants. The study protocol was approved by the Ethics Committee of Minami Osaka Hospital (No. 2022–1) and registered with the University Hospital Medical Information Network (UMIN000047518).

We enrolled 36 participants (24 men and 12 women) undergoing outpatient treatment at Minami Osaka Hospital. These participants had been receiving dietary and exercise therapy, along with oral hypoglycemic agents, but had not achieved their diabetes management goals, requiring further treatment intensification. The selection and exclusion criteria are shown in Table S1.

The study protocol is shown in Figure 1. Participants were randomized in a ratio of 1:1 to the IGlarLixi treatment group (IGlarLixi group) or to the IDegLira treatment group (IDegLira group) using RESEARCH RANDOMIZER (www.randomizer.org). If participants had previously used oral hypoglycemic agents, their dosages were continued without change, except for persons taking dipeptidyl peptidase‐4 inhibitors (DPP‐4i). The proCGM (Freestyle Libre ProTM; Abbott Diabetes Care) was worn on the day of random allocation, and pre‐breakfast injection therapy was started after 1 week of wearing the proCGM device, with five doses of IGlarLixi and ten doses of IDegLira for the starting dose, which was adjustable at the physician's discretion. Table S2 shows the titration algorithms for IGlarLixi and IDegLira. Participants performed self‐titration every 2 weeks, using median preprandial glucose levels at breakfast on the last 3 days (days 12–14) to achieve a target glucose level of 100–130 mg/dL. The maximum daily doses were 20 for IGlarLixi and 50 for IDegLira. Both groups wore proCGM devices at 18 weeks (16–20 weeks) after the start of the study. The treatment effect was evaluated using data from days 3–5 of proCGM. The body weight was measured at outpatient visits during this study period. HbA1c, glycated albumin, serum CPR, and fasting glucose levels were measured at 0 and 18 weeks of the study.

Figure 1.

Study protocol. OHAs, oral hypoglycemic agents; IGlarLixi, insulin glargine and lixisenatide; IDegLira, insulin degludec and liraglutide; proCGM, professional continuous glucose monitoring; DPP‐4i, dipeptidyl peptidase‐4 inhibitor.

Outcome measures

Primary and secondary endpoints were determined by analyzing proCGM data for 3 days (days 3–5) of each treatment. The primary efficacy and safety endpoints were TIR and TBR level 1, respectively ¹⁶ .

Secondary endpoints included the time above range (TAR) (percent of readings and time at >180 mg/dL) ¹⁶ ; nocturnal (0:00–5:59) TBR level 1 ¹⁶ ; standard deviation (SD) of glycemic variability (GV) ²⁰ ; percentage coefficient of variation (%CV) of glycemic variability ²¹ for 24 h and 6:00–18:00; 24 h M‐value (target glucose level, 100 mg/dL) ²² ; mean amplitude of glycemic excursion (MAGE) ²² ; 24 h mean of daily difference (MODD) (average of the differences based on proCGM data for days 1–2 and 2–3 over 3 consecutive days) ²² ; and mean glucose levels at 24 h, 0:00–6:00, 6:00–18:00, and 18:00–24:00. The following parameters were assessed at the end of this study: body mass index (BMI), HbA1c level, glycated albumin, duration to achieve target glucose level and complete titration, and FRC dosage. The changes in BMI, HbA1c level, glycated albumin, fasting glucose levels, and endogenous insulin secretory capacity were compared between pre‐ and post‐treatment timepoints. The correlation between C‐peptide index at baseline and time in range at the end of the study was analyzed.

Statistical analysis

The data are shown as mean ± SD unless otherwise specified. The Shapiro–Wilk test was used to test the data normality. Comparison of the glycemic variability indices of IGlarLixi and IDegLira and the treatment effect was performed using Student's t‐test when the data were normally distributed, and the Wilcoxon rank sum test when the data were nonparametrically distributed. The χ ² test was used to test the frequency difference between the two groups. A paired t‐test was used to compare pre‐ and post‐treatment indices for each treatment group. A Pearson product–moment correlation test was used to determine the correlation coefficient between the two variables. A two‐tailed P‐value <0.05 was considered statistically significant. In this study, a power analysis was conducted to determine the appropriate sample size. Since there are no reports of the TIR for IGlarLixi in Japan, we used HbA1c levels, which correlate with the TIR. The results from phase III trials have shown HbA1c level reductions of −1.4% for IGlarLixi ²³ and − 2.4% for IDegLira ²⁴ . The two aforementioned studies attempted to produce a 1.0% difference in HbA1c levels between groups. The sample size was 36, with the SD set to 0.9% (α error, 0.05; power, 0.9). Data were analyzed using EZR v. 1.37 (Saitama Medical Center, Jichi Medical University, Saitama, Japan) ²⁵ .

RESULTS

Participant characteristics

Table 1 shows the participant characteristics. Ten men and eight women were assigned to the IGlarLixi group, and the IDegLira group comprised 14 men and 4 women. One participant in the IDegLira group had an adverse event of mild nausea, but all participants completed the study without dropping out. On average, the participants were aged 62.1 years, the BMI was 27.0 kg/m², HbA1c level was 8.7%, glycated albumin was 22.8%, C‐peptide immunoreactivity was 2.7 ng/mL, and C‐peptide index (fasting CPR × 100/fasting plasma glucose) was 1.5. At baseline, no significant differences were identified among all parameters between groups. The oral hypoglycemic agents used during the study did not differ significantly between groups.

Table 1.

Baseline characteristics of the study participants

	Overall (n = 36)	IGlarLixi group (n = 18)	IDegLira group (n = 18)	P‐value*
Age (years)	62.1 ± 12.7	63.2 ± 11.0	61.1 ± 14.4	0.633
Duration of diabetes (years)	10.3 ± 7.8	10.4 ± 6.8	10.2 ± 8.9	0.933
Sex, male, n (%)	24 (66.7)	10 (55.6)	14 (77.8)	0.289
Body weight (kg)	71.6 ± 15.7	67.4 ± 13.8	75.7 ± 16.8	0.118
BMI (kg/m2)	27.0 ± 4.7	26.2 ± 3.8	27.8 ± 5.4	0.306
HbA1c (%)	8.7 ± 1.1	8.5 ± 0.9	8.8 ± 1.2	0.418
GA (%)	22.8 ± 5.3	21.6 ± 5.3	24.0 ± 5.1	0.164
FBG (mg/dL)	198.4 ± 73.6	173.3 ± 52.6	223.6 ± 83.9	0.111
CPR (ng/mL)	2.7 ± 1.4	2.5 ± 1.2	2.8 ± 1.6	0.509
CPI	1.5 ± 0.8	1.6 ± 1.0	1.3 ± 0.5	0.266
eGFR (mL/min/1.73 m2)	71.5 ± 19.1	67.2 ± 20.1	75.7 ± 17.6	0.188
TG level (mg/dL)	161.5 ± 93.1	146.4 ± 57.1	176.6 ± 118.8	0.338
LDL‐C level (mg/dL)	95.7 ± 29.8	90.5 ± 26.8	100.9 ± 32.5	0.303
HDL‐C level (mg/dL)	50.0 ± 14.7	53.3 ± 18.6	46.8 ± 8.5	0.187
S‐albumin level (g/dL)	4.3 ± 0.4	4.4 ± 0.3	4.2 ± 0.5	0.126
DPP‐4 inhibitor (pretrial), n	19	11	8	0.504
Antihyperglycemic drugs
Metformin, n	24	12	12	1.000
Sulfonylurea, n	5	3	2	1.000
SGLT‐2 inhibitor, n	21	12	9	0.499
α‐Glucosidase inhibitor, n	1	1	0	1.000

Data are presented as mean ± SD. *Student's t‐test or χ ² test were used to compare data between the two groups. Antidiabetic drug dosages did not change throughout the study period. BMI, body mass index; CPI, C‐peptide index; CPR, C‐peptide immunoreactivity; DPP‐4, dipeptidyl peptidase‐4; eGFR, estimated glomerular filtration rate; FBG, fasting blood glucose; GA, glycated albumin; HbA1c, glycated hemoglobin; HDL‐C, high‐density lipoprotein cholesterol; IDegLira, insulin degludec and liraglutide; IGlarLixi, insulin glargine and lixisenatide; LDL‐C, low‐density lipoprotein cholesterol; SGLT‐2, sodium‐glucose cotransporter‐2; TG, triglyceride.

Comparison of efficacy and safety between the IGlarLixi group and IDegLira group

None of the proCGMs of the participants were removed, and 100% of the data were acquired. The mean glucose curves from day 3 proCGM at baseline and the end of the study are shown in Figure 2. Both groups had similar mean glucose curves at baseline and at the end of the study. Table 2 shows proCGM parameters of glucose variability and diabetes‐related factors in persons treated with IGlarLixi or IDegLira. No significant differences were identified between the IGlarLixi and IDegLira groups regarding TIR, the primary efficacy endpoint, and TBR level 1, the primary safety endpoint.

Figure 2.

Three‐day mean glycemic variability curve of the 18 participants in each of the IGlarLixi and IDegLira groups based on proCGM data. The black and gray lines indicate the IGlarLixi and IDegLira groups, respectively, and the solid and dashed lines indicate week 18 and 0, respectively. IDegLira, insulin degludec and liraglutide; IGlarLixi, insulin glargine and lixisenatide; proCGM, professional continuous glucose monitoring.

Table 2.

proCGM parameters of glucose variability and diabetes‐related factors in persons treated with IGlarLixi or IDegLira

	IGlarLixi group (n = 18)	IDegLira group (n = 18)	P‐value*
Percentage of time in target glucose range (70–180 mg/dL)	78.3 ± 15.3	75.7 ± 19.0	0.431
Percentage of time in target glucose range (70–180 mg/dL) >70%, n (%)	9 (50.0)	11 (61.1)	0.737
Percentage of time below target glucose range (<70 mg/dL)	1.8 ± 3.2	2.7 ± 4.5	0.207
Percentage of time above target glucose range (>180 mg/dL)	19.9 ± 15.0	21.6 ± 17.6	0.598
Percentage of nocturnal time below target glucose range (<70 mg/dL)	1.8 ± 3.2	2.7 ± 4.5	0.648
24 h SD of glycemic variability (mg/dL)	42.3 ± 17.1	45.9 ± 19.1	0.304
0600–1800 h SD of glycemic variability (mg/dL)	39.1 ± 16.4	43.0 ± 18.3	0.255
24 h CV of glycemic variability (%)	29.9 ± 9.7	32.7 ± 9.9	0.137
0600–1800 h CV of glycemic variability (%)	27.1 ± 9.4	30.1 ± 10.5	0.121
24 h M value (target glucose level: 100 mg/dL)	10.9 ± 7.7	12.3 ± 10.7	0.667
MAGE (mg/dL)	98.7 ± 36.3	109.9 ± 37.8	0.123
MODD in glucose level (mg/dL)	32.2 ± 13.0	40.5 ± 25.0	0.237
24 h mean glucose level (mg/dL)	139.1 ± 22.3	136.7 ± 27.3	0.623
0000–0600 h mean glucose level (mg/dL)	114.7 ± 31.2	112.8 ± 39.7	0.791
0600–1800 h mean glucose level (mg/dL)	143.0 ± 22.9	140.0 ± 27.7	0.580
1800–2,400 h mean glucose level (mg/dL)	156.4 ± 38.4	149.5 ± 47.2	0.417
Delta body weight (kg)	−2.5 ± 2.7	−1.8 ± 3.4	0.520
BMI (kg/m2)	25.2 ± 3.3	27.1 ± 5.3	0.203
Delta BMI (kg/m2)	−1.0 ± 1.0	−0.7 ± 1.2	0.433
HbA1c (%)	7.0 ± 0.9	7.2 ± 0.6	0.394
Delta HbA1c (%)	−1.5 ± 1.1	−1.6 ± 1.3	0.822
GA (%)	16.4 ± 2.6	17.8 ± 2.9	0.141
Delta GA (%)	−5.1 ± 5.0	−6.2 ± 4.9	0.516
Titration period (weeks)	8.1 ± 4.8	9.3 ± 4.8	0.491
FRC (doses/day)	10.2 ± 3.7	17.2 ± 8.0	0.002*

Data are presented as mean ± SD. Data between groups are compared by Student's t‐test or χ ² test. BMI, body mass index; CV, coefficient of variation; FRC, fixed‐ratio combination; GA, glycated albumin; HbA1c, glycated hemoglobin; IDegLira, insulin degludec and liraglutide; IGlarLixi, insulin glargine and lixisenatide; MAGE, mean amplitude of glycemic excursion; MODD, mean of daily difference; proCGM, professional continuous glucose monitoring; SD, standard deviation. *Indicates a statistically significant difference between groups.

Regarding secondary endpoints, no significant differences were identified between the IGlarLixi and IDegLira groups for TAR; nocturnal TBR level 1; SD and %CV of GV for 24 h and 6:00–18:00; 24 h M‐value; MAGE; MODD; mean glucose level at 24 h, 0:00–6:00, 6:00–18:00, and 18:00–24:00; BMI; HbA1c level; glycated albumin at the end of the study; changes in body weight; BMI, HbA1c, and glycated albumin between pre‐ and post‐treatment timepoints; and duration to achieve the target glucose level and complete titration. The mean FRC dose was significantly higher in the IDegLira group (17.2 doses) than in the IGlarLixi group (10.2 doses; P = 0.002). Table 3 shows pre‐ and post‐treatment changes in BMI, HbA1c level, GA level, fasting glucose levels, and endogenous insulin secretory capacity in the IGlarLixi and the IDegLira groups. Both groups showed significant reductions in BMI, HbA1c, GA, and fasting glucose levels after treatment (each P value <0.001/0.029, <0.001/<0.001, <0.001/0.001, 0.010/<0.001) and no significant differences in endogenous insulin secretory capacity.

Table 3.

Changes in BMI, HbA1c level, GA level, FBG, and endogenous insulin secretory capacity between pre‐ and post‐treatment timepoints in the IGlarLixi and the IDegLira groups

	Pre‐treatment	Post‐treatment	P‐value
IGlarLixi group (n = 18)
BMI (kg/m2)	26.2 ± 3.8	25.2 ± 3.3	<0.001*
HbA1c (%)	8.5 ± 0.9	7.0 ± 0.9	<0.001*
GA (%)	21.6 ± 5.3	16.4 ± 2.6	<0.001*
FBG (mg/dL)	173.3 ± 52.6	128.5 ± 33.2	0.010*
CPR (ng/mL)	2.5 ± 1.2	2.3 ± 1.1	0.617
CPI	1.6 ± 1.0	2.0 ± 1.2	0.222
IDegLira group (n = 18)
BMI (kg/m2)	27.8 ± 5.4	27.1 ± 5.3	0.029*
HbA1c (%)	8.8 ± 1.2	7.2 ± 0.6	<0.001*
GA (%)	24.0 ± 5.1	17.8 ± 2.9	<0.001*
FBG (mg/dL)	223.6 ± 83.9	133.4 ± 31.3	<0.001*
CPR (ng/mL)	2.8 ± 1.6	2.7 ± 2.3	0.862
CPI	1.3 ± 0.5	2.1 ± 1.7	0.058

Data are presented as the mean ± SD. Pre‐and post‐treatment measurements were compared by paired t‐tests. BMI, body mass index; CPI, C‐peptide index; CPR, C‐peptide immunoreactivity; FBG, fasting blood glucose; GA, glycated albumin; HbA1c, glycated hemoglobin; IDegLira, insulin degludec and liraglutide; IGlarLixi, insulin glargine and lixisenatide. *Indicates a statistically significant difference between time points.

Correlation between CPI at baseline and TIR at the end of the study

The correlation between CPI at baseline and TIR at the end of the study is shown in Figure 3. Both IGlarLixi and IDegLira groups showed a significant positive correlation between CPI and TIR (P = 0.002 and r = 0.679; P = 0.002 and r = 0.681).

Figure 3.

Correlation between proCGM‐derived time in range at week 18 and C‐peptide index at baseline. A Pearson product–moment correlation test was used to determine the correlation coefficients between the two variables shown in each graph. (a) IGlarLixi group. (b) IDegLira group. CPI, C‐peptide index; IDegLira, insulin degludec and liraglutide; IGlarLixi, insulin glargine and lixisenatide; proCGM, professional continuous glucose monitoring; TIR, time in range.

CPI cut‐off values to achieve a TIR of >70% in IDegLira and IGlarLixi groups

Studies have reported that a TIR of 70% is equivalent to an HbA1c level of 7.0% ¹⁹ ; therefore, a TIR of >70% is recommended for treatment with CGM ⁸ . In this study, positive correlations between baseline CPI and TIR at the end of the study were confirmed for both the IGlarLixi and IDegLira groups. Therefore, the baseline CPI required to achieve a TIR of >70% was determined using receiver operating characteristic (ROC) analysis. In the IGlarLixi group, the predictive ability was highest when the CPI cutoff was 1.258, sensitivity was 77.8%, specificity was 100%, and the area under the curve (AUC) for a TIR of >70% was 0.914 (95% confidence interval, 0.775–1.000). In the IDegLira group, the predictive ability was highest when the CPI value was 1.099, sensitivity was 57.1%, specificity was 90.9%, and AUC for a TIR of >70% was 0.714 (95% confidence interval, 0.432–0.997; Figure 4).

Figure 4.

ROC curve for proCGM‐derived TIR at week 18 of 70% in CPI at baseline in IGlarLixi and IDegLira. CPI, C‐peptide index; IDegLira, insulin degludec and liraglutide; IGlarLixi, insulin glargine and lixisenatide; proCGM, professional continuous glucose monitoring; ROC, receiver operating characteristic; TIR, time in range

DISCUSSION

In this study, the IGlarLixi and IDegLira groups did not differ significantly in the primary efficacy endpoint TIR and safety endpoint TBR level 1. Both FRCs achieved the recommended TIR of >70% and TBR level 1 <4% ⁸ . Since a TIR of 70% correlates with an HbA1c level of approximately 7.0% ¹⁹ , and an HbA1c level <7% is the goal for preventing microvascular complications ²⁶ , both FRCs would be expected to be equally effective in preventing diabetic microvascular complications. The TIR significantly correlates with diabetic retinopathy progression ²⁷ and increased microalbuminuria ²⁸ ; thus, TIR should be increased as much as possible. In this study, 50.0% and 61.1% of the IGlarLixi and IDegLira groups, respectively, achieved TIRs of >70%; however, intensified treatment should be considered for those with the recommended TBR level 1 <4% and TIR <70% without increasing hypoglycemic risk.

Since the %CV of GV reportedly correlates with hypoglycemia risk ²⁹ , a value <36% is recommended ⁸ . However, when using medications with a high hypoglycemia risk, such as insulin or sulfonylureas, a %CV of GV <33% is reportedly safer ¹⁶ . Since IGlarLixi and IDegLira are FRCs with insulin and GLP‐1 RA, a target of %CV of GV <33% is reasonable, and since both FRCs had a mean %CV of GV <33% in this study, the hypoglycemic risk would be similarly reduced.

Glycemic variability increases have been reported to lead to macro‐ and microvascular diabetes complications ³⁰ , ³¹ . In the Hoorn Diabetes Care System cohort, individuals with a higher GV had worse metabolic profiles and were at higher risk for macrovascular complications and death ³² . Therefore, GV indices such as 24 h M‐value, MAGE, and MODD were comparable between both FRCs in this study, without significant differences, suggesting that treatment with either FRC could result in comparable efficacy in preventing future complications in individuals with type 2 diabetes mellitus.

In a Phase III clinical trial of insulin‐naïve Japanese individuals with type 2 diabetes mellitus who had inadequate effects with existing treatments, treatment with IGlarLixi for 26 weeks resulted in a body weight increase of 0.26 kg and a decrease in the HbA1c level of −1.40% ²³ . Further, treatment with IDegLira for 52 weeks resulted in a 2.9 kg increase in body weight and a 2.42% decrease in HbA1c level ²⁴ . Both FRCs showed significant reductions with no group differences after 18 weeks of treatment: −2.5 kg in the IGlarLixi group and −1.8 kg in the IDegLira group. The Phase III clinical trial results showed a slight increase in body weight. However, real‐world clinical data demonstrated that the weight‐reduction effects attributed to delayed gastric emptying and appetite suppression in the central nervous system, expected from GLP‐1 RA ³³ , surpassed insulin weight‐gaining effects. The changes in HbA1c levels were − 1.5% in the IGlarLixi group and − 1.6% in the IDegLira group, and both FRCs showed significant reductions in HbA1c levels after 18 weeks, with no difference between groups. The discrepancy between the results from phase III clinical trials of HbA1c level reductions of −1.40% with IGlarLixi ²³ and −2.42% with IDegLira ²⁴ may explain the results of this study. In the phase III clinical trials, (1) HbA1c levels at baseline were 8.08% in the IGlarLixi group and 8.5% in the IDegLira group, and HbA1c levels were higher in the IDegLira group. (2) Target fasting glucose levels were 79.2–100.8 mg/dL in the IGlarLixi group and 72.0–90.0 mg/dL in the IDegLira group, with lower target fasting glucose levels in the IDegLira group. (3) Fasting glucose levels at the end of the trial were 125.6 mg/dL in the IGlarLixi group and 104.9 mg/dL in the IDegLira group, with the IDegLira group having a lower fasting glucose level. The IGlarLixi and IDegLira groups were treated using the same titration algorithm to determine the treatment dose, which resulted in similar fasting glucose levels; therefore, there was no difference in the HbA1c levels. Both FRCs are expected to have the same HbA1c level‐lowering effect in actual clinical practice. Both FRCs resulted in TIRs >70%, TBRs <4%, and weight reduction due to GLP‐1 RA, which may resolve the three unmet medical needs associated with BOT (inability to achieve target HbA1c, increased hypoglycemia risk with intensified treatment, and barriers to intensified treatment due to weight gain).

Although no reports exist on the titration period required for the induction of treatment with injectable diabetes medications, a meta‐analysis of seven phase III clinical trials with insulin degludec and insulin glargine defined 15 weeks as the period for physicians to actively titrate insulin using an insulin titration algorithm ³⁴ . In this study, the participants understood the algorithm and self‐titrated the FRC. Participants reached target glucose levels in 8.1 and 9.3 weeks in the IGlarLixi and IDegLira groups, respectively.

In a report ³⁵ comparing insulin self‐titration and titration by physicians, HbA1c and fasting glucose levels decreased significantly, and the frequencies of hypoglycemia were similar in both groups, but the degrees of decrease and daily insulin units were significantly higher in the self‐titration group. The Insulin Therapy Self‐Efficacy Scale score was significantly higher in the self‐titration group, and insulin self‐titration enhanced self‐efficacy without impairing psychological distress or treatment satisfaction. Self‐titration with fixed‐ratio combinations was shown to result in better glycemic management without severe hypoglycemia, which could be useful for the future introduction of injection therapy in the outpatient environment.

The fixed‐ratio combination daily doses were significantly lower in the IGlarLixi group than in the IDegLira group. These differences may be related to differences in basal insulin and GLP‐1 RA content ratios among the fixed‐ratio combinations. IGlarLixi contains 20 units of insulin glargine for a maximum dose of 20 μg lixisenatide, and IDegLira contains 50 units of insulin degludec for a maximum dose of 1.8 mg liraglutide. Thus, IGlarLixi contains a higher ratio of GLP‐1 RA to basal insulin than IdegLira. The average dose of IglarLixi was 10.2 doses, and that of IDegLira was 17.2 doses; it was significantly higher for IGlarLixi. However, the IGlarLixi group used approximately 10 units of insulin glargine and 10 μg of lixisenatide (one‐half of the maximum dose of lixisenatide), while the IDegLira group used approximately 17 units of insulin degludec and 0.6 mg of liraglutide (one‐third of the maximum dose of liraglutide). Since no significant difference was present among the doses due to the different characteristics of the two fixed‐ratio combinations (i.e., whether to use both basal insulin and GLP‐1 RA to produce an effect, as in IGlarLixi, or to use a sufficient amount of basal insulin and supplement its effect with GLP‐1 RA, as in IdegLira), the choice of fixed‐ratio combination is not simply determined by effectiveness but rather by the appropriateness for each individual with type 2 diabetes mellitus.

In this study, both fixed‐ratio combinations showed significant positive correlations between C‐peptide index at baseline and time in range at 18 weeks. The CPI is an index of endogenous insulin secretion capacity; since the GLP‐1 RA in FRC promotes insulin secretion in a glucose‐dependent manner, individuals with residual endogenous insulin secretion capacity suppress postprandial glucose elevation and the basal insulin results in lower overall glucose levels. A study using liraglutide reported that CPI values were significantly higher in the effective group (FPG of approximately 115 mg/dL) and the ineffective group (FPG of approximately 120 mg/dL), using fasting plasma glucose as an index ³⁶ . The relative response to incretin and insulin combination therapy correlates with fasting glucose, and the correction of fasting glucose with basal insulin increases the incretin effect ³⁷ . The HbA1c‐lowering effect of liraglutide has been reported to correlate with CPI, a measure of residual pancreatic beta‐cell function. ROC analysis of CPI to achieve an HbA1c level <7% after 1 year of combination therapy with basal insulin revealed a cut‐off value of 1.103 ³⁸ . In this study, regarding the correlation between a TIR of 70% and an HbA1c level of approximately 7.0% ¹⁹ , ROC analysis of CPI at baseline with a TIR of >70% after 18 weeks resulted in cut‐off values of 1.258 for IGlarLixi and 1.099 for IDegLira. The results suggest that IGlarLixi is effective during residual endogenous insulin secretion because it contains a high ratio of GLP‐1 RA, while IDegLira is effective when endogenous insulin secretion begins to decline because it contains a high ratio of basal insulin. A distinction between the use of both FRCs may be possible according to the C‐peptide index. These results suggest that it is difficult for persons with a low level of CPI <1.0 to achieve a TIR of >70% with FRC treatment; in such cases, treatment with multiple daily injections would be necessary.

This study had some limitations. We conducted a randomized controlled trial with a small sample size of 18 cases per group at a single center. Because the participants visited a diabetes specialty department at a hospital, there was a risk of selection bias. For more realistic results of clinical outcomes, multicenter studies, including general clinics, should be conducted with a larger sample size using a common protocol. Furthermore, because we conducted this study over 18 weeks, participants may have achieved a temporary improvement in glycemic management due to the expectant effect of introducing FRC. A study lasting longer than 1 year is necessary to investigate the benefits of the medication on long‐term changes in body weight and HbA1c level, as well as the safety of the medication concerning adverse events, including hypoglycemia risk and digestive symptoms.

In conclusion, a time in range of >70% was achieved with IGlarLixi and IDegLira treatment without significant difference, and the time in range was significantly correlated with the C‐peptide index at the start of treatment. The choice of IGlarLixi or IDegLira, based on each person's C‐peptide index, may prevent future complications of diabetes. We plan to investigate the persistence of the effect of a TIR of >70% after 1 year of each FRC selection and to examine whether the C‐peptide index trends allow for the retention of endogenous insulin secretory capacity.

FUNDING

No funding or sponsorship was received for this study or publication of this article.

DISCLOSURE

Y. Kawaguchi received lecture honoraria or speaker fees from Sanofi K.K., Novo Nordisk Pharma, Boehringer Ingelheim, Sumitomo Pharma Co., Ltd, Mochida Pharmaceutical Co. Ltd, and Kowa Company Ltd. The other authors have no conflict of interest to disclose.

Approval of the research protocol: The study protocol was approved by the Ethics Committee of Minami Osaka Hospital (No. 2022–1) and registered with the University Hospital Medical Information Network (UMIN000047518).

Informed consent: All participants provided informed consent before participating in the trial.

Approval date of Registry and the Registration No. of the study/trial: The date of approval is June 2, 2022. The approval number is 2022–1.

Animal studies: N/A.

Supporting information

Table S1. Selection and exclusion criteria

Table S2. IGlarLixi and IDegLira titration algorithms