Effectiveness and Safety of Insulin Glargine 300 U/ml in High-Risk Subgroups (Renal Impairment and Older Age ≥ 70 years) of Insulin-Naïve People with Type 2 Diabetes: A Post hoc Analysis of Real-World ATOS Study

Abstract

Introduction

This post hoc analysis of an A Toujeo^® Observational Study (ATOS) aims to evaluate the real-world effectiveness and safety of insulin glargine 300 U/ml (Gla-300) in high-risk subgroups of insulin-naïve people with type 2 diabetes (PwT2D) from multiple geographical regions (Asia, the Middle East, North Africa, Latin America, and Eastern Europe).

Methods

In these post hoc analyses of ATOS, a real-world, 12-month, prospective study included 4422 insulin-naïve adults (age ≥ 18 years) with type 2 diabetes (T2D) uncontrolled (HbA_1c > 7% and ≤ 11%) on one or more oral antidiabetic drugs (OADs) who initiated Gla-300 treatment as per routine practice. Primary and secondary endpoints were studied according to renal impairment (RI) status (without or with) and age group (</≥ 70 years).

Results

At baseline, participants with a history of RI (N = 581, 13.1%) and older participants (aged ≥ 70 years, N = 514, 11.6%) had a longer duration of diabetes and were more likely to present diabetic complications compared to without RI and younger participants (aged < 70 years). At month 6, the individualized HbA_1c target (as determined by their treating physician) was achieved in 27.5% of participants with RI compared to 24.8% of participants without RI whereas 32.3% of older participants achieved their individualized HbA_1c target compared to 24.2% of younger participants. In this post hoc analysis, Gla-300 treatment improved glycemic control with meaningful reductions in HbA_1c, fasting plasma glucose (FPG) and fasting self-monitored blood glucose (SMBG) across all subgroups. The incidence of hypoglycemia was low and changes in body weight were minimal across all subgroups.

Conclusions

In a real-world setting, the initiation of Gla-300 in insulin-naïve PwT2D uncontrolled on OADs resulted in improved glycemic control with a low incidence of hypoglycemia and minimal weight change in participants with a history of RI and in older participants.

Trial Registration

Clinicaltrials.gov number NCT03703869.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13300-025-01785-w.

Key Summary Points

Why carry out this study?

The global prevalence of diabetes is increasing, particularly concerning in high-risk populations such as individuals with chronic kidney disease (CKD) and older adults (aged ≥ 70 years).

Although randomized controlled trials (RCTs) have shown the efficacy and safety of insulin glargine 300 U/ml (Gla-300) in high-risk populations (people with a history of renal impairment [RI] and older adults), the clinical benefits observed in RCTs have been confirmed by only few real-world studies.

These post hoc analyses of an ATOS study aim to assess the effectiveness and safety of Gla-300 in insulin-naïve people with or without a history of RI and in younger or older age (</≥ 70 years) people with type 2 diabetes (PwT2D).

What was learned from the study?

Findings from these post hoc analyses of the ATOS study suggest that the initiation of Gla-300 resulted in improved glycemic control with a low risk of hypoglycemia and minimal weight change in insulin-naïve participants with a history of RI and in older adults with type 2 diabetes (T2D) who are at a high risk of hypoglycemia.

These findings align with data from other RCTs’ and real-world evidence (RWE) studies on Gla-300, supporting its suitability as a therapeutic option for these growing vulnerable populations from multiple geographical regions.

Introduction

The rise in the prevalence of diabetes is alarming globally, especially in high-risk populations, such as people with chronic kidney disease (CKD) and older adults (aged ≥ 70 years). Moreover, the prevalence in these high-risk populations associated with an increased risk of diabetes-related complications has increased and is a major health concern worldwide [1–3]. CKD is observed in approximately 30–40% of people with type 2 diabetes (PwT2D) and is linked to a higher risk of hypoglycemia and cardiovascular morbidity and mortality [4, 5]. Moreover, older adults (aged ≥ 65 years) with T2D have been reported to have a greater risk of hypoglycemia, cognitive impairment, depression, urinary incontinence, persistent pain, frailty and an increased risk of mortality compared to their peers who do not have diabetes [3, 6]. Consequently, the pharmacological management of T2D in these high-risk populations can be challenging for physicians due to heterogeneity in health conditions, co-morbidities, renal impairment (RI) and a higher risk of hypoglycemia, leading to poor glycemic control [2, 3, 7, 8]. Thus, there is a need for effective glucose-lowering treatments that will improve glycemic control with a low risk of hypoglycemia in these vulnerable populations to achieve a good quality of life and avoid diabetes-related complications.

The safety and effectiveness of certain glucose-lowering medications may be reduced in people with CKD and diabetes, underscoring the significance of insulin therapy as a viable treatment option [2, 5, 9]. Because some glucose-lowering agents, such as metformin, sulphonylureas and α-glucosidase inhibitors should be avoided in advanced CKD due to a higher risk of hypoglycemia or other adverse events (AEs) [2, 10, 11]. Sodium glucose cotransporter-2 inhibitors (SGLT-2i) are the first-line therapy for CKD in PwT2D; however, these inhibitors are likely to have reduced glucose-lowering efficacy when the estimated glomerular filtration rate (eGFR) is < 45 ml/min/1.73 m² [9]. The Kidney Disease: Improving Global Outcomes (KDIGO) practice guideline recommends glucagon-like peptide-1 receptor agonists (GLP-1 RAs) for CKD in PwT2D who have not achieved individualized glycemic targets despite the use of SGLT-2i treatment. However, access to the currently available GLP-1 RAs may be limited in most countries due to high prices and these agents can also cause weight loss, which is undesirable in some people with CKD [2, 5, 9, 12]. The recent KDIGO and American Diabetes Association (ADA) guidelines also recommend that insulin therapy may be a suitable therapeutic option for achieving optimal glycemic control in those with impaired renal function [5, 13]. Likewise, the ADA [3] and Endocrine Society [14] guidelines have recommended use of insulin therapy with less stringent glycemic targets of < 7.5% and < 8.0% in older PwT2D with few or multiple co-morbidities. Moreover, insulin therapy in older adults should be tailored according to personalized glycemic targets and modified based on co-existing chronic illnesses, cognitive function, and functional status [3].

Second-generation basal insulins (BIs) including insulin glargine 300 U/ml (Gla-300), have shown a more stable and prolonged pharmacokinetic and pharmacodynamic (PK/PD) profile, a longer duration of action and a reduced risk of hypoglycemia than the first-generation BI analogue, insulin glargine 100 U/ml (Gla‐100) [15, 16]. Previously, the efficacy and safety of Gla-300 in high-risk populations (people with a history of RI and older age) have been demonstrated in randomized clinical trials (RCTs), such as EDITION [17, 18] and BRIGHT [19, 20]. The clinical benefits of Gla-300 observed in RCTs has been confirmed in a few number of real-world studies [21–25]; however, the effectiveness and safety of Gla-300 in high-risk populations in a real-world setting remain less explored.

ATOS was a prospective, 12-month observational study that investigated the real-world effectiveness and safety of Gla-300 in insulin-naïve PwT2D living in 18 countries from Asia, the Middle East, North Africa, Latin America, and Eastern Europe. Overall, Gla-300 use was associated with improved glycemic control and low rates of reported hypoglycemia in the overall ATOS real-world study population, with minimal change in daily insulin dose and few AEs related to the study treatment [26]. The results from the post hoc analyses of the ATOS study are presented herein, which aim to evaluate the effectiveness and safety of Gla-300 in insulin-naïve people without or with a history of RI and in younger or older age (</≥ 70 years) PwT2D.

Methods

Study Design and Participants

The study design and methods of the ATOS study have been reported previously [26]. In brief, ATOS was a 12-month prospective, observational, international, multicenter study assessing the real-world clinical effectiveness and safety of the long-acting BI analogue, Gla-300, after oral antidiabetic drugs (OADs) failure in insulin-naïve PwT2D across 18 countries in multiple geographical regions. Of the 4550 participants included, 4422 were eligible for assessment. The study began in March 2018 and was completed in February 2020. In the present study, two post hoc analyses investigated the effectiveness and safety of Gla-300 in insulin-naïve PwT2D according to RI status, defined as ‘with RI’ at the time of study entry if the patient had reported at least one of the following medical history terms: RI, nephropathy, or CKD and ‘without RI’ if not and as per age groups: younger (< 70 years) and older (≥ 70 years). The eGFR data were not available for both without and with RI participants.

The main inclusion criteria for the ATOS study were insulin-naïve adults (aged ≥ 18 years) with T2D, uncontrolled on ≥ 1 OAD, with an HbA_1c > 7% and ≤ 11%. All eligible participants received once-daily Gla-300 as per physician’s decision and the time of injection was reported at every visit. The individual participant’s HbA_1c goal was set by the treating physician and titration was performed using the locally applicable titration algorithm. All participants provided written informed consent. The study was conducted in accordance with the guidelines for Good Epidemiology Practice and the Declaration of Helsinki, and the study is also registered at the US National Library of Medicine (Clinicaltrials.gov, NCT03703869). The study protocol was approved by the Institutional Review Board/Institutional Ethics Committee at each site according to local practice (Table S1).

Study Endpoints and Data Collection

Data were collected at baseline and at 3, 6, and 12 months of treatment with Gla-300 and the observation closest to the study schedule was to be recorded as a study visit. The primary efficacy endpoint was the percentage of participants achieving their predefined individualized HbA_1c goal (as determined by their treating physician) at month 6. If the predefined individualized goal was not specified at baseline, a general HbA1c goal of < 7.0% was used. The key secondary efficacy endpoints included changes in HbA_1c, fasting plasma glucose (FPG), fasting self-monitored blood glucose (SMBG), Gla-300 dose and body weight from baseline to months 3, 6, and 12. Safety endpoints included incidence and event rates of documented symptomatic hypoglycemia (blood glucose ≤ 3.9 mmol/l [≤ 70 mg/dl] and < 3.0 mmol/l [< 54 mg/dl]) and severe hypoglycemia (an event requiring assistance of another person to actively administer carbohydrate, glucagon, or other resuscitative actions) as well as the incidence of AEs.

Statistical Analysis

Efficacy analyses were undertaken in the evaluable population (all eligible participants with an HbA_1c assessment at month 6). Demographics, baseline characteristics and safety analysis were undertaken in the eligible population (all included participants who had signed informed consent, met the inclusion/exclusion criteria and who had initiated Gla-300 within ± 31 days from the start of study). Descriptive statistics of quantitative effectiveness and safety parameters were used. The mean changes from baseline in HbA_1c, FPG, fasting SMBG, and body weight were assessed using a mixed model for repeated measurements (MMRM) approach and the baseline-adjusted least square (LS) mean estimates and 95% confidence intervals (CIs) for the changes at months 6 and 12 were reported. The mean changes from baseline in Gla-300 dose, AEs, and hypoglycemia were assessed descriptively. All statistical analyses were performed using SAS version 9.4 or higher.

Results

Participant Disposition and Baseline Characteristics

In these post hoc analyses of the ATOS study (by RI status and age groups), the data are presented for eligible participants without (N = 3841) or with a history of RI (N = 581) and participants aged < 70 years (N = 3908) or ≥ 70 years (N = 514). Participants with a history of RI and those aged ≥ 70 years had a longer duration of diabetes and were more likely to have a history of diabetic neuropathy, retinopathy, microalbuminuria, macroalbuminuria or advanced kidney disease than the without RI and younger participant subgroups, respectively (Table 1). In both post hoc analyses, baseline HbA_1c, FPG and fasting SMBG were comparable across all subgroups. In participants with a history of RI and older adults, the mean physician-set individualized HbA_1c target was 7.2 ± 0.4% with the majority (50–60%) of participants having an HbA_1c target of ≥ 7 to < 7.5% as determined by their treating physicians (Table 1). At baseline, metformin was the most commonly used OAD, followed by sulphonylureas, dipeptidyl peptidase-4 inhibitors and SGLT-2i across all subgroups (Table 1).

Table 1.

Baseline characteristics (eligible population)

Baseline characteristics	RI status		Age category
	Without RI (N = 3841)	With RIe (N = 581)	Age < 70 years (N = 3908)	Age ≥ 70 years (N = 514)
Age, years	56.4 ± 10.7	62.2 ± 10.3	55 ± 9.4	74 ± 4.0
Male, n (%)	1879 (48.9)	252 (43.4)	1936 (49.5)	195 (37.9)
Female, n (%)	1962 (51.1)	329 (56.6)	1972 (50.5)	319 (62.1)
Body weight, kg	80.1 ± 16.1	84.1 ± 17.5	81.2 ± 16.4	76.8 ± 15.0
BMI, kg/m2	29.2 ± 5.2	30.7 ± 5.6	29.4 ± 5.4	28.9 ± 5.0
< 25, n (%)	695 (21.3)	65 (13.2)	665 (20.0)	95 (22.1)
25–30, n (%)	1326 (40.6)	183 (37.2)	1337 (40.2)	172 (40.1)
30–35, n (%)	829 (25.4)	142 (28.9)	858 (25.8)	113 (26.3)
≥ 35, n (%)	417 (12.8)	102 (20.7)	470 (14.1)	49 (11.4)
Duration of diabetes, years, median (Q1:Q3)	9 (5:13)	12 (8:17)	9 (5:13)	13 (9:19)
1–5 years, n (%)	763 (19.9)	53 (9.1)	779 (19.9)	37 (7.2)
5–10 years, n (%)	1438 (37.4)	157 (27.0)	1480 (37.9)	115 (22.4)
≥ 10 years, n (%)	1640 (42.7)	371 (63.9)	1649 (42.2)	362 (70.4)
Any diabetes complication and comorbidity history, n (%)	2639 (68.7)	580 (99.8)	2771 (70.9)	448 (87.2)
Diabetic neuropathy, n (%)	1268 (33.0)	418 (71.9)	1408 (36.0)	278 (54.1)
Diabetic retinopathy, n (%)	497 (12.9)	262 (45.1)	616 (15.8)	143 (27.8)
Microalbuminuria, n (%)	0 (0)	404 (69.5)	317 (8.1)	87 (16.9)
Macroalbuminuria, n (%)	0 (0)	102 (17.6)	81 (2.1)	21 (4.1)
Advanced kidney disease, n (%)	0 (0)	68 (11.7)	43 (1.1)	25 (4.9)
End-stage renal failure, n (%)	0 (0)	3 (0.5)	0 (0.0)	3 (0.6)
HbA1c %a	9.3 ± 1.0	9.3 ± 0.9	9.3 ± 1.0	9.2 ± 1.0
HbA1c %, n (%)
≥ 7 to < 7.5	82 (2.1)	12 (2.1)	80 (2.0)	14 (2.7)
≥ 7.5 to < 8	305 (7.9)	37 (6.4)	298 (7.6)	44 (8.6)
≥ 8 to < 9	1181 (30.7)	167 (28.7)	1192 (30.5)	156 (30.4)
≥ 9 to < 10	1152 (30.0)	199 (34.3)	1171 (30.0)	180 (35.0)
≥ 10	1121 (29.2)	166 (28.6)	1167 (29.9)	120 (23.3)
Pre-defined, physician-set individualized target HbA1c %	7.0 ± 0.4	7.2 ± 0.4	7.0 ± 0.3	7.2 ± 0.4
Pre-defined, physician-set individualized target HbA1c %, n (%)b
< 7	562 (14.6)	42 (7.2)	572 (14.6)	32 (6.2)
≥ 7 to < 7.5	2764 (72.0)	351 (60.4)	2847 (72.9)	268 (52.1)
≥ 7.5 to < 8	377 (9.8)	145 (25.0)	363 (9.3)	159 (30.9)
≥ 8	138 (3.6)	43 (7.4)	126 (3.2)	55 (10.7)
FPG, mmol/l	11.0 ± 3.1	11.1 ± 2.8	11.0 ± 3.1	11.1 ± 2.8
Fasting SMBG, mmol/l	10.8 ± 2.6	10.8 ± 2.4	10.8 ± 2.6	10.7 ± 2.3
Basal insulin dose, U/kg	0.18 ± 0.09	0.19 ± 0.10	0.19 ± 0.09	0.18 ± 0.09
Duration of OAD treatment, median (Q1:Q3)c	8 (5:13)	11 (8:16)	8 (5:12)	12 (9:18)
OAD use at baseline, n (%)d
1 OAD use	674 (17.5)	93 (16.0)	664 (17.0)	103 (20.0)
2 OAD use	1731 (45.1)	285 (49.1)	1778 (45.5)	238 (46.3)
≥ 3 OAD use	1436 (37.4)	203 (34.9)	1466 (37.5)	173 (33.7)
OAD at baseline, n (%)
Biguanides	3455 (90.0)	477 (82.1)	3503 (89.6)	429 (83.5)
Sulfonylureas	2763 (71.9)	465 (80.0)	2832 (72.5)	396 (77.0)
DPP-4 inhibitors	1667 (43.4)	258 (44.4)	1715 (43.9)	210 (40.9)
SGLT-2 inhibitors	542 (14.1)	88 (15.1)	563 (14.4)	67 (13.0)
Alpha-glucosidase inhibitors	254 (6.6)	10 (1.7)	242 (6.2)	22 (4.3)
Thiazolidinediones	194 (5.1)	22 (3.8)	199 (5.1)	17 (3.3)
Glinides	34 (0.9)	19 (3.3)	41 (1.0)	12 (2.3)

The values are mean ± SD unless otherwise indicated

BMI body mass index, DPP-4 dipeptidyl peptidase 4, FPG fasting plasma glucose, Gla-300 insulin glargine 300 U/ml, HbA_1c glycated hemoglobin, OAD oral antidiabetic drug, RI renal impairment, SD standard deviation, SGLT-2 sodium-glucose cotransporter-2, SMBG self-monitored blood glucose

^aIf there was no available value prior to Gla-300 initiation, baseline HbA1c value was defined as the first available value up to 2 weeks after first administration of Gla-300

^bPhysician-set individualized HbA1c goal

^cDuration calculated based on the participants who had reported at least one OAD

^dOAD use at baseline—medications taken within 6 months of screening

^eRI subgroup was defined as "With RI" if at least one of the following medical history terms and had been reported: renal impairment, nephropathy or chronic kidney disease; and "Without RI" if not. eGFR data were not available

Effectiveness Outcomes

Post hoc analyses by RI status: After 6 months of Gla-300 treatment, the predefined individualized HbA_1c target was achieved in 27.5% (95% CI 23.7–31.6) of participants with RI compared to 24.8% (95% CI 23.4–26.3) of participants without RI (Fig. 1A). In participants with RI, the mean ± SD HbA_1c reduced from 9.32 ± 0.97% at baseline to 7.82 ± 1.04% and 7.47 ± 0.86% at months 6 and 12, respectively, whereas in participants without RI, HbA_1c reduced from 9.27 ± 1.00% to 7.76 ± 1.06% and 7.37 ± 0.98%, respectively. HbA_1c reductions from baseline to months 6 and 12 were similar between participants without or with RI (Fig. 1B).

Fig. 1.

Percentage of participants reaching the predefined individualized HbA_1c target and mean HbA_1c reductions from baseline to months 6 and 12 (evaluable population); A and B, participants without RI and with RI; C and D, participants aged < 70 and ≥ 70 years. The efficacy analysis were undertaken in the evaluable population: all eligible participants with an HbA_1c assessment at month 6 (without RI: N = 3419; with RI: N = 512) and at month 12 (without RI: N = 3259; with RI: N = 489); all eligible participants with an HbA_1c assessment at month 6 (< 70 years: N = 3470; ≥ 70 years: N = 461) and at month 12 (< 70 years: N = 3314; ≥ 70 years: N = 434). ^#LS mean change assessed using a MMRM approach. CI confidence interval, HbA_1c glycated hemoglobin, LS least square, MMRM mixed model for repeated measurements, RI renal impairment, SD standard deviation

Post hoc analyses by age groups: After 6 months of Gla-300 treatment, a total of 32.3% (95% CI 28.1–36.8) of older participants (aged ≥ 70 years) achieved their individualized HbA_1c target compared to 24.2% (95% CI 22.8–25.7) of younger participants (aged < 70 years) (Fig. 1C). Moreover, a greater proportion of participants achieved HbA_1c targets of < 7.5% and < 8.0%, in both older-age and younger-age participants at months 6 and 12 (Supplementary Fig. S1). HbA_1c reductions from baseline to months 6 and 12 were similar in both age subgroups (< / ≥ 70 years). In older-age participants, HbA_1c reduced from 9.20 ± 0.94% at baseline to 7.73 ± 0.87% and 7.45 ± 0.89% at months 6 and 12, respectively, whereas in younger-age participants, HbA_1c reduced from 9.28 ± 1.00% at baseline to 7.78 ± 1.08% and 7.37 ± 0.97%, respectively (Fig. 1D).

The overall improvement in glycemic control was observed with meaningful reductions in both the mean FPG and fasting SMBG in all subgroups. The mean changes in FPG and fasting SMBG from baseline to months 6 and 12 were similar in both with and without RI subgroups as well as in both age subgroups (</≥ 70 years) (Fig. 2A–D).

Fig. 2.

Mean change in FPG and SMBG levels from baseline to months 6 and 12 (evaluable population); A and B participants without RI and with RI; C and D participants aged < 70 and ≥ 70 years. The efficacy analysis were undertaken in the evaluable population: all eligible participants with an HbA_1c assessment at month 6 (without RI: N = 3419; with RI: N = 512) and at month 12 (without RI: N = 3259; with RI: N = 489); all eligible participants with an HbA_1c assessment at month 6 (< 70 years: N = 3470; ≥ 70 years: N = 461) and at month 12 (< 70 years: N = 3314; ≥ 70 years: N = 434). ^#LS mean change assessed using a MMRM approach. CI confidence interval, FPG fasting plasma glucose, LS least square, MMRM mixed model for repeated measurements, RI renal impairment, SD standard deviation, SMBG self-monitored blood glucose

Hypoglycemia

The percentage of participants with documented symptomatic hypoglycemic events were low in both participants with RI (≤ 3.9 mmol/l: 1.89% and < 3.0 mmol/l: 0.34%) and without RI (≤ 3.9 mmol/l: 1.17% and < 3.0 mmol/l: 0.18%) at month 12 (Table 2). Similarly, in the analysis by age groups, the overall incidence of documented symptomatic (≤ 3.9 and < 3.0 mmol/l) or severe hypoglycemic events were low in both age subgroups (</≥ 70 years).

Table 2.

Incidence and event rates of hypoglycemia (eligible population)

	RI status				Age category
Hypoglycemia event	Without RI (N = 3841)		With RI (N = 581)		Age < 70 years (N = 3908)		Age ≥ 70 years (N = 514)
	All hypoglycemia	Nocturnal hypoglycemia	All hypoglycemia	Nocturnal hypoglycemia	All hypoglycemia	Nocturnal hypoglycemia	All hypoglycemia	Nocturnal hypoglycemia
	n (%) [event rate PPY]	n (%) [event rate PPY]	n (%) [event rate PPY]	n (%) [event rate PPY]	n (%) [event rate PPY]	n (%) [event rate PPY]	n (%) [event rate PPY]	n (%) [event rate PPY]
Documented symptomatic (≤ 3.9 mmol/l or ≤ 70 mg/dl)
Month 6a	28 (0.73) [0.021]	7 (0.18) [0.004]	10 (1.72) [0.066]	1 (0.17) [0.010]	32 (0.82) [0.027]	8 (0.20) [0.005]	6 (1.17) [0.031]	0 (0.00) [0.000]
Month 12b	45 (1.17) [0.017]	10 (0.26) [0.003]	11 (1.89) [0.050]	1 (0.17) [0.006]	44 (1.13) [0.020]	11 (0.28) [0.003]	12 (2.33) [0.031]	0 (0.00) [0.000]
Documented symptomatic (≤ 3.0 mmol/l or ≤ 54 mg/dl)
Month 6	3 (0.08) [0.002]	1 (0.03) [0.001]	2 (0.34) [0.007]	1 (0.17) [0.003]	5 (0.13) [0.003]	2 (0.05) [0.001]	0 (0.00) [0.000]	0 (0.00) [0.000]
Month 12	7 (0.18) [0.002]	1 (0.03) [0.000]	2 (0.34) [0.004]	1 (0.17) [0.002]	6 (0.15) [0.002]	2 (0.05) [0.001]	3 (0.58) [0.006]	0 (0.00) [0.000]
Severe (third-party assistance needed)
Month 6	2 (0.05) [0.001]	1 (0.03) [0.001]	3 (0.52) [0.014]	0 (0.00) [0.000]	3 (0.08) [0.002]	1 (0.03) [0.001]	2 (0.39) [0.012]	0 (0.00) [0.000]
Month 12	3 (0.08) [0.001]	1 (0.03) [0.000]	3 (0.52) [0.007]	0 (0.00) [0.000]	4 (0.10) [0.001]	1 (0.03) [0.000]	2 (0.39) [0.006]	0 (0.00) [0.000]

The safety analysis were undertaken in the eligible population (without RI: N = 3841; with RI: N = 581; < 70 years: N = 3908; ≥ 70 years: N = 514 who met the inclusion/exclusion criteria and started Gla-300 ± 31 days from the start of the study

n (%): number and percentage of participants with at least one hypoglycemia event

Gla-300 insulin glargine 300 U/ml, PPY per patient-year, RI renal impairment

^a6-month period was defined from the first treatment administration to month 6 visit or treatment discontinuation, whichever occurred first

^b12-month period was defined from the first treatment administration to month 12 visit or treatment discontinuation, whichever occurred first

Insulin Dose and Body Weight

An increase was observed in Gla-300 dose from baseline to month 12 by 0.12 U/kg and 0.10 U/kg in participants with RI and without RI, respectively (Fig. 3A) and by 0.11 U/kg and 0.10 U/kg in the older-age and younger-age subgroups, respectively (Fig. 3C). Changes in mean body weight from baseline to months 6 and 12 were minimal in both RI subgroups (with and without) and in both age subgroups (</≥ 70 years) (Fig. 3B and D).

Fig. 3.

Mean change in daily insulin dose and body weight from baseline to months 6 and 12 (eligible population); A and B participants without RI and with RI; C and D participants aged < 70 and ≥ 70 years. The analysis were undertaken in the eligible population who met the inclusion/exclusion criteria and started Gla-300 ± 31 days from the start of study; Mean change in daily insulin dose at month 6 (without RI: N = 3385; with RI: N = 491 and < 70 years: N = 3426; ≥ 70 years: N = 450) and at month 12 (without RI: N = 3282; with RI: N = 486 and < 70 years N = 3329; ≥ 70 years: N = 439) and mean change in body weight at month 6 (without RI: N = 3436; with RI: N = 507 < 70 years: N = 3478; ≥ 70 years: N = 465) and at month 12 (without RI: N = 3352; with RI: N = 498 and < 70 years: N = 3404; ≥ 70 years: N = 446). ^†Dose change from baseline is presented as mean change from baseline. ^#LS mean change assessed using a MMRM approach. CI confidence interval, Gla-300 insulin glargine 300 U/ml, LS least square, MMRM mixed model for repeated measurements, SD standard deviation

Adverse Events

The percentage of participants experiencing treatment-emergent adverse events (TEAEs) was 9.8% in RI and 5.9% in the without RI subgroups (Table 3). The incidence of TEAEs was low in both age ≥ 70 years (7.6%) and in the < 70 years (6.2%) subgroups. The most common TEAEs were infections and infestations, metabolism and nutrition disorders, nervous system disorders, gastrointestinal disorders or musculoskeletal and connective tissue disorders across all subgroups. Serious TEAEs were reported by 20 (3.4%) participants with RI and 37 (1.0%) participants in the without RI subgroups. Similarly, the occurrence of serious TEAEs were reported by 12 (2.3%) older-age participants and 45 (1.2%) younger-age participants. None of the participants in the RI or older-age subgroup had a TEAE leading to treatment discontinuation (Table 3).

Table 3.

Overview of treatment-emergent adverse events (eligible population)

	RI status		Age category
TEAE, n (%)a	Without RI (N = 3841)	With RI (N = 581)	Age < 70 years (N = 3908)	Age ≥ 70 years (N = 514)
Any TEAE	226 (5.9)	57 (9.8)	244 (6.2)	39 (7.6)
Infections and infestations	69 (1.8)	15 (2.6)	72 (1.8)	12 (2.3)
Nervous system disorders	27 (0.7)	7 (1.2)	27 (0.7)	7 (1.4)
Metabolism and nutrition disorders	33 (0.9)	5 (0.9)	34 (0.9)	4 (0.8)
Musculoskeletal and connective tissue disorders	26 (0.7)	12 (2.1)	34 (0.9)	4 (0.8)
Gastrointestinal disorders	33 (0.9)	5 (0.9)	35 (0.9)	3 (0.6)
Injury, poisoning and procedural complications	15 (0.4)	5 (0.9)	16 (0.4)	4 (0.8)
General disorders and administration site conditions	19 (0.5)	7 (1.2)	22 (0.6)	4 (0.8)
Cardiac disorders	11 (0.3)	6 (1.0)	17 (0.4)	0
Respiratory, thoracic and mediastinal disorders	16 (0.4)	2 (0.3)	15 (0.4)	3 (0.6)
Skin and subcutaneous tissue disorders	12 (0.3)	3 (0.5)	14 (0.4)	1 (0.2)
Any serious TEAE	37 (1.0)	20 (3.4)	45 (1.2)	12 (2.3)
Cardiac disorders	7 (0.2)	6 (1.0)	13 (0.3)	0
Nervous system disorders	4 (0.1)	2 (0.3)	2 (0.1)	4 (0.8)
Any related TEAE	10 (0.3)	2 (0.3)	8 (0.2)	4 (0.8)
Metabolism and nutrition disorders	3 (0.1)	0	1 (0.0)	2 (0.4)
Nervous system disorders	0	2 (0.3)	0	2 (0.4)
Any serious related TEAE	0	1 (0.2)	0	1 (0.2)
Hypoglycemic coma	0	1 (0.2)	0	1 (0.2)
Any TEAE leading to premature treatment discontinuation	8 (0.2)	0	8 (0.2)	0
Diabetic foot infection	1 (0.0)	0	1 (0.0)	0
Anxiety	1 (0.0)	0	1 (0.0)	0
Any TEAE leading to deathb	5 (0.1)	5 (0.9)	7 (0.2)	3 (0.6)
Cardiac disorders	1 (0.0)	3 (0.5)	4 (0.1)	0
Neoplasms benign, malignant and unspecified	2 (0.1)	1 (0.2)	2 (0.1)	1 (0.2)

The safety analysis was undertaken in the eligible population (without RI: N = 3841; with RI: N = 581; < 70 years: N = 3908; ≥ 70 years: N = 514) who met the inclusion/exclusion criteria and started Gla-300 ± 31 days from the start of the study. n (%): number and percentage of participants with at least one TEAE

Gla-300 insulin glargine 300 U/ml, RI renal impairment, TEAE treatment-emergent adverse event

^aTEAE observation period is recorded from the 1st administration to the last administration

^bNone of the fatal TEAEs were considered related to Gla-300 treatment

Discussion

People with CKD and older adults with T2D are reported to have a greater risk of cardiovascular disease, an increased risk of hypoglycemia and mortality resulting in an increased economic burden on healthcare resources [2, 3]. The majority of high-risk population fails to achieve optimal glycemic targets mainly due to an increased risk of hypoglycemia and the limitations of existing anti-diabetic therapies. Insulin therapy specifically long-acting BI analogs can be considered a suitable treatment option for the management of diabetes in high-risk populations [2, 3, 9]. However, there are limited studies on the use of BIs in these highly prevalent vulnerable populations. Therefore, in the present post hoc analyses of the ATOS study, we investigated the effectiveness and safety of Gla-300 in insulin-naïve PwT2D grouped by RI status (without/with) and age (younger/older). In both post hoc analyses, improvements were observed in the mean HbA_1c, FPG and fasting SMBG with low incidence and event rates of hypoglycemia and minimal body weight changes. The high baseline HbA_1c levels with a longer duration of diabetes suggest delayed insulin initiation in these high-risk participants in a real-world setting. Our results demonstrate that the effectiveness and safety of Gla-300 reported in the overall ATOS study [26] population are maintained in these high-risk subgroups who had a greater clinical risk and the highest rate of co-morbidities.

The results from the present study support the findings reported by other RCTs [17–20] or real-world studies [21–23] on Gla-300 use in high-risk populations in other geographical regions. The DELIVER real-world retrospective studies (DELIVER D+, DELIVER HIGH RISK and DELIVER 3) demonstrated the effectiveness and safety of Gla-300 in high-risk populations (in adults aged ≥ 65 years and with moderate/severe RI) [21–23]. Furthermore, Gla-300 showed comparable glycemic improvements with a lower risk of hypoglycemia than Gla-100 in mild-to-moderate RI participants and in older adults (aged ≥ 65 years) with T2D in a patient-level meta-analysis using pooled 6-month data from EDITION 1, 2 and 3 trials [17, 18]. Similarly, in a subanalysis of the BRIGHT trial, greater HbA_1c reductions were observed with Gla-300 without an increase in the risk of hypoglycemia than insulin degludec 100 U/ml (IDeg-100) in insulin-naïve people with impaired renal function and older adults (aged ≥ 70 years) with T2D [19, 20]. In the present post hoc analyses, the individualized HbA_1c target achievement at month 6 in RI and older-age participants was similar or higher than that observed in the overall ATOS study [26] and DELIVER studies [21–23] while it was lower than that observed in the meta-analysis of the EDITION trials and subanalysis of the BRIGHT trial [17–20]. The high-risk subgroups in the present post hoc analyses had a higher baseline HbA_1c (RI: 9.3% and aged ≥ 70 years: 9.2%) than that reported in the aforementioned RCTs [17–20], which might have contributed to the lower target achievement observed. As expected, compared to RCTs that commonly use more intensive treat-to-target titration algorithms, the Gla-300 dose titration was performed in a real-world setting according to locally applicable titration. The observed mean daily BI dose at 6 months was lower in the ATOS study than that in the EDITION and BRIGHT trials, suggesting that more intensive titration of Gla-300 in clinical practice could help more PwT2D achieve their glycemic targets. In addition, the failure to reach target HbA1c levels in both participants without RI and those aged < 70 years old may be due to higher baseline HbA1c values (9.3%) and suboptimal titration of Gla-300 dose (0.27 U/kg/day at month 6 and 0.29 U/kg/day at month 12) compared to aforementioned trials. This indicates that better glycemic control could be achieved through more regular follow-up to enable optimal BI titration, along with enhanced treatment compliance and surveillance program.

The HbA_1c reductions observed in RI and older-age participants at months 6 and 12 were comparable to those observed in the overall ATOS study [26]. However, the HbA1c reductions observed at month 6 in both RI and older-age participants were greater than those reported in the DELIVER studies [21–23] probably because of differences in participants characteristics (mean age, body mass index [BMI] or severity of co-morbidity) in the aforementioned real-world studies as compared to the present ATOS post hoc analyses. In line with changes in HbA_1c, the LS mean changes in both FPG and fasting SMBG from baseline to month 12 in RI participants were comparable to those observed in the overall ATOS study [26]. Glycemic control as measured by FPG and fasting SMBG improved with meaningful reductions from baseline to months 6 and 12 and was similar in the with and without RI subgroups. Similar results were observed in the older-age post hoc analysis. Although the effectiveness of Gla-300 using FPG and fasting SMBG is of high importance, future studies are warranted to further confirm its effectiveness using continuous glucose monitoring metrics, given the reduced reliability of HbA_1c in RI participants [27].

The observed improvement in glycemic control at month 6 in RI participants was achieved with a mean BI dose (0.28 U/kg/day), which was similar to the mean BI dose reported in the global ATOS study (0.27 U/kg/day) [26]. However, it was lower than that observed in the EDITION (0.81 U/kg/day) [17] and BRIGHT (0.47 U/kg/day) trials [19]. Similar results were observed in the older-age post hoc analysis. Moreover, in the present post hoc analyses, the incidence of hypoglycemia was low across all subgroups with Gla-300 treatment supporting that further insulin titration even in high-risk subgroups could be achieved safely, to help more people reach their glycemic target. The incidence of documented symptomatic and severe hypoglycemia (at any time of day and nocturnal) reported over 12 months with Gla-300 treatment was low across all subgroups. Additionally, the nocturnal hypoglycemic events were low in RI participants and were not observed in older-age participants. This finding is consistent with aforementioned RCTs [17–20] and real-world studies [25, 26, 28]. Similarly, the real-world LIGHTNING study showed that the predicated rates of severe hypoglycemia were lower with Gla-300 than first-generation BI analogs in insulin-naïve older adults (aged ≥ 65 or ≥ 75 years) [29]. In the SENIOR study (aged 65–75 and ≥ 75 years), Gla-300 showed comparable glycemic improvements with a lower risk of hypoglycemia in older adults compared to Gla-100 [30]. This low incidence of hypoglycemia seen in the high-risk participants could be attributed in part to the flatter time-action profile of Gla-300 [31] and to the less stringent Gla-300 dose titration.

In the present subgroup analysis by RI status, minimal changes in body weight were observed at months 6 and 12 in participants with RI (the LS mean changes were − 0.2 kg and − 0.3 kg), which is in-line with the overall ATOS study (− 0.0 kg and − 0.1 kg) [26] and with the meta-analysis of EDITION trials (0.14 kg at month 6) [17]. Similarly, in the subgroup analysis by age groups, minimal changes in body weight were observed at months 6 and 12 in older-age participants (0.4 kg and 0.5 kg, respectively) which is in-line with the overall ATOS study (− 0.0 kg and − 0.1 kg) [26] and with the meta-analysis of EDITION trials (aged ≥ 65 years: 0.4 kg at month 6) [18]. Commonly, weight loss is not recommended in older and frail people with diabetes. In the present analyses, weight gain (0.4 kg and 0.5 kg) was observed in older-age participants which could be beneficial (anabolic effects of insulin) [32]. Moreover, Gla-300 was well tolerated with the low incidence of AEs and no unexpected safety concerns were reported in high-risk participants in either of the post hoc analyses.

The strengths of the present analyses of ATOS include its prospective real-world study design that provides valuable information on the effectiveness and safety of Gla-300 in these high-risk populations in a routine clinical practice setting across wider geographical regions. Moreover, the study included a substantial number of participants with a history of RI (N = 581) and older adults (aged ≥ 70 years, N = 514), populations that are often poorly represented in RCTs. The limitations of this study include its post hoc nature, the lack of eGFR data, the exclusion of participants with severe RI (CKD stage 5), the lack of a comparator arm and the lack of accurate glycemic control data, as CKD anemia can influence HbA_1c levels. Although the new-user design ensured that the results could be attributed mainly to Gla-300 treatment, there was a possibility that some of the OADs used during the study might have contributed to the ‘study effect’. Owing to the observational nature of the study, there were few follow-up visits and hypoglycemic events were self-reported by participants. Therefore, under-reporting of hypoglycemia events may have occurred due to participants’ recall bias and could result in an underestimation of its true incidence [33]. As the present data were collected as a part of routine clinical practice, the potential impact of selection bias (as the participants were enrolled voluntarily) and other confounding factors (age, BMI, diabetes-related complications, co-morbidity history and other medications) cannot be ruled out.

Conclusions

In conclusion, the results observed from these post hoc analyses of the ATOS study suggest that the initiation of Gla-300 is effective in achieving pre-defined HbA_1c goals and improvement of glycemic control with a low risk of hypoglycemia and minimal weight change in insulin-naïve participants with a history of RI and in older adults with T2D who are at high-risk of hypoglycemia. The results observed in the present post hoc analyses are consistent with the data from other RCTs’ subanalysis and RWE studies on Gla-300 and further support the use of Gla-300 as a suitable therapeutic option in these growing vulnerable populations with a history of RI and in older adults living with T2D.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

All the named authors meet the International Committee of Medical Journal Editors (ICMJE) criteria for authorship for this article, take complete responsibility for the integrity of the data and accuracy of the data analysis and have given their approval for this version to be published. Conceptualization and methodology: Maria Aileen Mabunay, Lydie Melas-melt, Valerie Pilorget; Investigation: Amir Tirosh, Niaz Khan, Hernando Vargas-Uricoechea, Gagik Galstyan, Abdul Rahman Al Shaikh, Brij Mohan Makkar; formal analysis: Maria Aileen Mabunay, Lydie Melas-melt, Valerie Pilorget; Janaka Karalliedde: Critical review of the work for important intellectual content. All the authors participated in the interpretation of the data; the writing, reviewing, editing of the manuscript, and approved the final draft for submission.

Funding

This study and the journal’s Rapid Service Fee were funded by Sanofi, Paris, France.

Data Availability

Qualified researchers may request access to patient-level data and related documents including the clinical study report, study protocol with any amendments, blank case report form, statistical analysis plan, and dataset specifications. Patient-level data will be anonymized, and study documents will be redacted to protect the privacy of trial participants. Further details on data sharing criteria of Sanofi, eligible studies and process for requesting access can be found at: https://www.vivli.org

Declarations

Conflict of Interest

Amir Tirosh: Member of the ATOS Steering Committee and have received honoraria in relation to the ATOS study; He was a member of the National Diabetes Council; Head of the Diabetes Committee of the Israeli Endocrine Society; received honoraria from Sanofi; grant from Medtronic; consulting fees from Sanofi, NovoNordisk, MSD, Merck, AstraZeneca, Medtronic, Dreamed Diabetes; and received personal fees for participation in advisory boards from Sanofi, Novo-Nordisk, MSD, Merck, AstraZeneca, Bayer. Niaz Khan: Member of the ATOS Steering Committee and have received honoraria in relation to the ATOS study. Hernando Vargas-Uricoechea: Member of the ATOS Steering Committee and have received honoraria in relation to the ATOS study; received honoraria from Sanofi and Abbott; personal fees for participation in advisory boards and funding for attending meetings/travel from Sanofi. Gagik Galstyan: Member of the ATOS Steering Committee and have received honoraria in relation to the ATOS study. Abdul Rahman Al Shaikh and Brij Mohan Makkar: An experts, investigator for the ATOS and have received honoraria in relation to the ATOS study. Lydie Melas-melt: Employee of IVIDATA Life Sciences, Levallois-Perret, France, contracted by Sanofi. Janaka Karalliedde: Received research grants from Astra Zeneca, institution fee from Sanofi and speaker and advisory board fees from Boehringer Ingelheim, Lilly, Astra Zeneca and Daichi Sankyo. Maria Aileen Mabunay: Employee of Sanofi and may hold shares and/or stock options in the company. Valerie Pilorget was an employee of Sanofi at the time of study conduct and may hold stocks/shares in the company.

Ethical Approval

ATOS study was conducted in accordance with the guidelines for Good Epidemiology Practice and principles laid down in the 1964 Declaration of Helsinki by the 18th World Medical Assembly and its later amendments. The study is registered at the U.S. National Library of Medicine (Clinicaltrials.gov Number, NCT03703869). The study protocol was reviewed and approved by the Independent Interdisciplinary Ethics Committee on Ethical Review for Clinical Studies in accordance with the local regulations in each participating country/study center. All participants provided written informed consent.