Diabetes Ther (2021) 12:1073–1097 https://doi.org/10.1007/s13300-021-01030-0 ORIGINAL RESEARCH Impact of Age on the Effectiveness and Safety of Insulin Glargine 300 U/mL: Results from the REALI European Pooled Data Analysis Riccardo C. Bonadonna . Didac Mauricio . Dirk Müller-Wieland . Nick Freemantle . Gregory Bigot . Celine Mauquoi . Alice Ciocca . Mireille Bonnemaire . Pierre Gourdy Received: January 12, 2021 / Accepted: February 10, 2021 / Published online: March 1, 2021 Ó The Author(s) 2021 ABSTRACT Methods: Pooled efficacy and safety data were collected from 8106 European patients with Introduction: Patients aged C 65 years con- uncontrolled T2DM who were initiated on or tinue to be underrepresented in clinical studies switched to Gla-300 injected once daily for related to type 2 diabetes mellitus (T2DM). 24 weeks. Patients were categorised into five age Accordingly, the REALI pooled analysis was subgroups: \ 50 (N = 727), 50–59 (N = 2030), performed to evaluate the effectiveness and 60–69 (N = 3054), 70–79 (N = 1847) safety of insulin glargine 300 U/mL (Gla-300) and C 80 years (N = 448). across different age subgroups, using data from Results: Mean baseline haemoglobin A1c 14 interventional and non-interventional (HbA1c) decreased linearly from the youngest studies. (9.10%) to the oldest (8.46%) age subgroup. Following Gla-300 initiation, there were similar HbA1c reductions across age groups, with a least R. C. Bonadonna (&) C. Mauquoi Azienda Ospedaliero–Universitaria di Parma, Parma, International Drug Development Institute (IDDI), Italy Louvain-la-Neuve, Belgium e-mail: riccardo.bonadonna@unipr.it A. Ciocca  M. Bonnemaire R. C. Bonadonna General Medicines, Sanofi, Paris, France Department of Medicine and Surgery, University of Parma, Parma, Italy P. Gourdy Endocrinology, Diabetology and Nutrition D. Mauricio Department, Toulouse University Hospital, Department of Endocrinology and Nutrition, Toulouse, France CIBERDEM, Institut de Recerca Hospital de la Santa Creu i Sant Pau, Barcelona, Spain P. Gourdy Institute of Metabolic and Cardiovascular Diseases, D. Müller-Wieland UMR1048 INSERM/UPS, Toulouse University, Department of Medicine I, University Hospital Toulouse, France Aachen, Aachen, Germany N. Freemantle Institute of Clinical Trials and Methodology, University College London, London, UK G. Bigot IVIDATA Group, Paris, France 1074 Diabetes Ther (2021) 12:1073–1097 squares mean (95% confidence interval) change What was learned from the study? in HbA1c from baseline to week 24 of - 1.09% (- 1.18 to - 1.00), - 1.08% (- 1.14 to - 1.03), Gla-300 therapy initiated in patients with - 1.12% (- 1.17 to - 1.07), - 1.18% (- 1.24 to uncontrolled T2DM is associated with - 1.12) and - 1.11% (- 1.23 to - 0.99) in clinically important and consistent the \ 50, 50–59, 60–69, 70–79 and C 80 years reductions in haemoglobin A1c and subgroups, respectively. The incidences and fasting plasma glucose levels across a wide event rates of reported hypoglycaemia were range of ages, with limited hypoglycaemia overall low. Compared to younger age sub- concerns. groups, lower incidences of symptomatic The findings of the REALI pooled analysis hypoglycaemia occurring at any time of the day indicate that Gla-300 might be a (5.9 vs. 7.6–9.4% for the younger subgroups) or suitable therapeutic option in elderly during the night (0.5 vs. 1.6–2.5%) were recor- patients who represent a vulnerable ded in patients aged C 80 years. By contrast, the population prone to hypoglycaemia. highest incidence of severe hypoglycaemia occurring any time of the day was reported in the subgroup aged C 80 years (1.1 vs. 0.1–0.6% for the younger age subgroups). Conclusion: Gla-300 initiated in patients with DIGITAL FEATURES uncontrolled T2DM provides glycaemic This article is published with digital features, improvement with a favourable safety profile including a summary slide, to facilitate under- across a wide range of ages. standing of the article. To view digital features for this article go to https://doi.org/10.6084/ Keywords: Age; Europe; Glycaemic control; m9.figshare.13808264. Hypoglycaemia; Insulin glargine 300 U/mL; Older adults; Pooled analysis; Type 2 diabetes INTRODUCTION Key Summary Points Type 2 diabetes mellitus (T2DM) is a growing public health burden in older adults. World- Why carry out this study? wide, an estimated 135.6 million people aged Treatment of uncontrolled type 2 diabetes 65–99 years had diabetes in 2019, and this mellitus (T2DM) in elderly patients is number is expected to increase to 276.2 million more challenging than in non-elderly in 2045 [1]. Management of T2DM in elderly patients because of age-related patients is challenging for clinicians due to the pathophysiological features, increased difficulty in individualising glycaemic targets prevalence of comorbidities, and treatment strategies, as well as to the pres- polypharmacy and difficulties in adhering ence of coexisting comorbidities, polypharmacy to complex self-care activities. and hypoglycaemic risk [2]. Indeed, older adults with T2DM have a higher risk of premature Considering the limited data focusing on death, functional disability, accelerated muscle T2DM care in older adults, the REALI loss, depression, cognitive dysfunction, renal pooled analysis was performed to evaluate impairment and cardiovascular diseases, such as the effectiveness and safety of insulin hypertension, coronary heart disease and glargine 300 U/mL (Gla-300) across stroke, than those without T2DM [3, 4]. At the different age subgroups, using data from same time, older adults with T2DM are more 14 interventional and non-interventional prone to hypoglycaemia than their younger studies reflecting clinical practice in counterparts for multiple reasons, such as different European countries. potential difficulties with food ingestion, Diabetes Ther (2021) 12:1073–1097 1075 insufficient adjustment of insulin dose, conditions [13, 14], despite having the highest decreased renal function leading to slower prevalence of diabetes of any age group (19.3%) clearance of drugs, reduced responses to coun- [1]. In a descriptive analysis of 2484 diabetes- terregulatory hormones, lower blood glucose related interventional trials, very few trials threshold for autonomic symptoms and higher (0.6%) selectively enrolled patients aged C blood glucose threshold for cognitive dysfunc- 65 years; specifically, patients [ 65 years of age tion [5, 6]. Furthermore, older adults with were excluded from 30.8% of studies and those T2DM may either have elderly onset disease aged [ 75 years were excluded from 54.9% of (diagnosed at age C 65 years) or long-standing studies [15]. Considering the limited data diabetes with onset in middle age or earlier focusing on diabetes care in older adults, the years, adding to the complexity of T2DM man- REALI pooled analysis was performed to evalu- agement in the elderly [7]. ate the effectiveness and safety of Gla-300 in Long-acting, once-daily basal insulin repre- different prespecified age subgroups, using data sents an effective and safe therapeutic option in from 14 interventional and non-interventional older patients with T2DM not achieving gly- studies reflecting clinical practice in different caemic targets on oral glucose-lowering agents European countries. and/or glucagon-like peptide-1 (GLP-1) receptor agonists [3, 4, 6, 8]. Given the complexity of multiple-dose insulin regimens for older METHODS patients, once-daily basal insulin injection therapy is preferred in most elderly frail patients Study Selection and Population due to its ease of use and the limited risk of hypoglycaemia and disease-related distress Eligible studies for the REALI pooled analysis [4, 6, 8–10]. Insulin glargine 300 U/mL (Gla- had to be prospective (interventional or obser- 300) is a second-generation, long-acting basal vational) studies performed in European coun- insulin analogue, given as a once-daily subcu- tries among patients with uncontrolled T2DM taneous injection, which has demonstrated (i.e. in whom glycaemic targets have not been comparable glycaemic control to that provided achieved) initiated on Gla-300 therapy, and had by insulin glargine 100 U/mL (Gla-100), with a to have a duration of at least 24 weeks and reduced risk of hypoglycaemia at any time of individual participant-level data available for the day and at night in a broad population of both efficacy and safety outcomes [16]. In the almost 2500 patients with T2DM enrolled in the present analysis, 14 trials conducted between EDITION phase III clinical trial programme [11]. June 2015 and December 2018 were pooled A post-hoc analysis of the EDITION 1, 2 and 3 (Table 1), ten of which are already published trials provided evidence of a comparable gly- [17–27]. The rationale, methodology and a caemic control and a reduced risk of nocturnal detailed description of the variables have been hypoglycaemia for Gla-300 versus Gla-100 in already provided in the published protocol of both patients aged \ 65 and C 65 years [12]. the REALI project [16]. SENIOR, a randomised controlled trial (RCT) In each study, Gla-300 was administered comparing Gla-300 to Gla-100 in individuals subcutaneously once daily, using a pre-filled with T2DM aged C 65 years, showed compara- insulin pen at the same time of the day ± 3 h if ble reductions in haemoglobin A1c (HbA1c) and needed. Two of the included studies (Take a lower risk of documented symptomatic Control [19] and ITAS [20]) were RCTs in which hypoglycaemia versus Gla-100, with a signifi- patients were allocated to a self- versus a cant benefit on hypoglycaemic risk observed in physician-managed titration of Gla-300, participants aged C 75 years [5]. whereas the other studies were single-arm, i.e. With the exception of dedicated studies like Gla-300 was administered using a physician-led SENIOR [5], patients aged C 65 years continue titration algorithm. All studies, except to be underrepresented in clinical studies rela- COBALTA [25], were performed in the ambula- ted to T2DM or to diabetes-associated tory care setting. 1076 Diabetes Ther (2021) 12:1073–1097 Table 1 Characteristics of the studies included in the REALI pooled analysis Study name Location(s)/ Study Key inclusion Sample Age Age, years period of study description criteria (N) range (mean – SD) (years) Non-interventional studies Toujeo-Neo Germany/August 52-week, Patients with T2DM 1213 24–85 64.6 ± 10.6 (ISRCTN number: 2015 to March observational, previously treated ISRCTN93674355) 2017 open-label, with any basal multicentre, insulin except Gla- prospective 300, and with an study to assess HbA1c C 7.5% real-world and B 10.0% and effectiveness a FPG [ 130 mg/ and safety of dL switching basal component of any BOTplus or any basal- bolus insulin regimen to Gla- 300 OPTIN-D [27] The Netherlands/ 24-week, Patients with T2DM 162 53–77 65.7 ± 6.9 October 2015 multicentre, previously treated to September prospective, with insulin 2017 open-label, (basal ± prandial observational insulin) cohort study to for C 6 months document prior to Gla-300 changes over initiation time in PROs (e.g. emotional wellbeing, adherence, sleep quality and duration) To-Goal (Data on Serbia/November 15-month, Patients with T2DM 367 36–82 62.8 ± 8.4 file) 2017 to prospective, previously treated October 2018 observational with insulin study to (basal ± prandial evaluate real-life insulin) without effectiveness OADs and safety of Gla-300 Diabetes Ther (2021) 12:1073–1097 1077 Table 1 continued Study name Location(s)/ Study Key inclusion Sample Age Age, years period of study description criteria (N) range (mean – SD) (years) TOP-2 [17] Germany, Austria 52-week, Patients with T2DM 1640 38–88 64.7 ± 10.1 and observational, previously treated Switzerland/ open-label, with any basal June 2015 to multicentre, insulin except Gla- December 2016 prospective 300, and with an study to HbA1c C 7.5% evaluate real- and B 10.0% world effectiveness and safety of Gla-300 in patients uncontrolled on their previous BOT Toujeo-BB [18] Hungary/March 24-week, single- Patients with T2DM 229 44–75 61.6 ± 8.6 2016 to April arm, non- previously treated 2017 interventional, with basal-bolus multicentre regimens study aimed to (NPH ? regular evaluate real- insulin), with an world HbA1c C 8.0% effectiveness of or C 3 Gla- hypoglycaemic 300 ? insulin events per month glulisine in requiring patients correction uncontrolled on their previous basal-bolus regimen 1078 Diabetes Ther (2021) 12:1073–1097 Table 1 continued Study name Location(s)/ Study Key inclusion Sample Age Age, years period of study description criteria (N) range (mean – SD) (years) Toujeo-1 [21, 22] Germany and 52-week, non- Insulin-naı̈ve 1547 37–88 64.6 ± 11.1 Switzerland/ interventional, patients with June 2015 to multinational, T2DM previously December 2017 multicentre, treated with prospective OADs, with an study to HbA1c C 7.5% evaluate real- and B 10.0% world effectiveness and safety of initiating a BOT regimen with Gla-300 TOPAZ [23] Czech Republic/ 24-week, Patients with T2DM 300 47–78 63.8 ± 7.5 May 2016 to multicentre, previously treated March 2018 prospective, with basal insulin observational, with or without single-arm OADs, and with study to an HbA1c [ 7.6% evaluate clinical or repeated effectiveness hypoglycaemia and safety of Gla-300 MAGE [24] Belgium/June 12-month, Patients with T2DM 93 45–77 63.7 ± 7.5 2016 to August multicentre, for [ 1 year, with 2018 prospective, an observational, HbA1c C 7.0% single-arm and B 10.0%, study to assess previously treated treatment with any basal satisfaction, insulin except Gla- efficacy, and 300 plus mealtime safety of Gla- insulin 300 in a real- world setting Diabetes Ther (2021) 12:1073–1097 1079 Table 1 continued Study name Location(s)/ Study Key inclusion Sample Age Age, years period of study description criteria (N) range (mean – SD) (years) GOAL-Ro (Data on Romania/ May 24-week, Insulin-naı̈ve 1048 45–72 60.2 ± 7.0 file) 2017 to June prospective, patients with 2018 observational T2DM, with an study to HbA1c C 7.0% evaluate real-life effectiveness and safety of Gla-300 To UPGRADE Bulgaria/July 24-week, Patients with T2DM 286 27–83 61.8 ± 9.5 (Data on file) 2017 to prospective, previously treated December 2018 multicentre, with non- NPH ± prandial interventional insulin or study to premixed insulin evaluate real-life with or without effectiveness OADs and safety of Gla-300 Interventional, single-arm studies COBALTA [25] Spain/June 2016 26-week, open- Hospitalised patients 112 53–88 72.3 ± 10 (EudraCT number: to July 2018 label, with T2DM who 2015–004,715-20) multicentre, were C 3 months single-arm on treatment with study to basal insulin with evaluate efficacy or without OADs, and safety of with an Gla-300 during HbA1c C 8.0% hospitalisation and B 10.0% and therapy intensification at discharge 1080 Diabetes Ther (2021) 12:1073–1097 Table 1 continued Study name Location(s)/ Study Key inclusion Sample Age Age, years period of study description criteria (N) range (mean – SD) (years) TRANSITION 2 France/ January 24-week, Patients with T2DM 193 46–78 62.3 ± 8.0 [26] 2016 to July multicentre, previously treated (ClinicalTrials.gov 2017 prospective, with basal insulin identifier: open-label, with or without NCT02967237) single-arm other antidiabetics, study to HbA1c [ 7.5%, evaluate efficacy and fasting and safety of SMPG [ 130 mg/ Gla-300 in dL (mean of last 3 patients with measures) suboptimal glucose control on another basal insulin for whom a therapeutic change was indicated Interventional, randomised controlled studies Take Control [19] Greece, Spain, 24-week, Patients with T2DM 631 41–81 63.9 ± 8.5 (EudraCT number: Czech multinational, for C 1 year, who 2015–001,626-42) Republic, multicentre, were Switzerland, open-label, for C 6 months Poland, randomised on treatment Denmark, (1:1), two-arm, with C 1 OAD, Slovenia, parallel-group with or without a Slovakia, study to basal insulin, and Croatia, UK/ compare with an February 2016 efficacy and HbA1c C 7.0% to June 2017 safety of self- and B 10.0% for versus patients taking physician- basal insulin, managed or C 7.5% titration of Gla- and B 11.0% for 300 insulin-naı̈ve patients Diabetes Ther (2021) 12:1073–1097 1081 Table 1 continued Study name Location(s)/ Study Key inclusion Sample Age Age, years period of study description criteria (N) range (mean – SD) (years) ITAS [20] Italy/September 24-week, Insulin-naı̈ve 359 45–79 64.3 ± 8.4 (EudraCT Number: 2015 to multicentre, patients with 2015–001,167-39) October 2017 open-label, T2DM randomised for C 1 year, with (1:1), parallel- an HbA1c C 7.5% group study to and B 10.0% on compare OADs efficacy and safety of self- versus physician- managed titration of Gla- 300 BOT Basal insulin-supported oral therapy, BOTplus basal insulin-supported oral therapy plus a single or double dose of prandial insulin, FPG fasting plasma glucose, Gla-300 insulin glargine 300 U/mL, HbA1c haemoglobin A1c, NPH neutral protamine Hagedorn, OADs oral antidiabetic drugs, PROs patient-reported outcomes, SD standard deviation, SMPG self- monitored plasma glucose, T2DM type 2 diabetes mellitus Patients included within the REALI analysis Assessment of Outcomes were either insulin-naı̈ve or previously treated with insulin (basal insulin ± prandial insulin) Glycaemic control was evaluated for age sub- with or without non-insulin anti-hyperglycaemic groups by determining the mean values of agents. Exclusion criteria common for all studies HbA1c and fasting plasma glucose (FPG) at ini- included in this analysis were: the presence of tiation of Gla-300 (baseline) and subsequently cancer, a diagnosis of type 1 diabetes, pregnancy at weeks 12 and 24 of Gla-300 therapy. Changes and/or breastfeeding, a history of alcohol or drug in HbA1c and FPG from baseline to weeks 12 abuse, the presence of any clinically relevant and 24, as well as the percentages of patients somatic or mental disease, stage 5 chronic kidney achieving the HbA1c targets of \ 7.0% disease, known hypersensitivity or intolerance to (\ 53 mmol/mol), \ 7.5% (58.5 mmol/mol) Gla-300 or any of its excipients and inability to and \ 8.0% (63.9 mmol/mol) at week 24 of Gla- self-measure blood glucose levels. 300 treatment were determined. All pooled studies were conducted according Safety endpoints included the percentage of to the principles of the Declaration of Helsinki patients with C 1 hypoglycaemic event; the and the requirements of Good Clinical Practice, event rate of hypoglycaemic events; and the and were approved by the relevant institutional changes in daily insulin dose (in U/day and in review boards/ethics committees. All partici- U/kg/day) and body weight from baseline to pants gave written informed consent. Before weeks 12 and 24 of Gla-300 treatment. Hypo- data pooling, all patient information was de- glycaemic events were reported according to identified. Consequently, no ethical approval their time of occurrence, during the night and was required for this pooled analysis. at any time of the day. The definitions of 1082 Diabetes Ther (2021) 12:1073–1097 hypoglycaemia were predetermined in the pre- pooled analysis, of whom 727 (9.0%) were sent pooled analysis. Severe hypoglycaemia was aged \ 50 years, 2030 (25.0%) were between the defined as any event requiring assistance from ages of 50 and 59 years, 3054 (37.7%) were another person to actively administer carbohy- between the ages of 60 and 69, 1847 (22.8%) drates or glucagon or take other corrective were between the ages of 70 and 79 years and actions. Symptomatic hypoglycaemia was 448 (5.5%) were aged C 80 years. Of the 8106 defined as an event during which typical participants, 7929 (97.8%) were treated with at symptoms of hypoglycaemia occurred (e.g. least one dose of Gla-300. sweating, hunger, shakiness, palpitations). Overall, baseline characteristics were fairly comparable across the 14 pooled studies. Nev- Statistical Analysis ertheless, some differences were noted; for instance, participants in the interventional, Patients were categorised into five age sub- single-arm COBALTA study [25] were of a sub- groups: \ 50, 50–59, 60–69, 70–79 stantially higher mean age (72.3 years) com- and C 80 years. Efficacy and safety outcomes pared with participants in the other studies were analysed using 10-year age strata to enable (mean age ranged from 60.2 to 65.7 years) a detailed description of characteristics over a (Table 1). wide range of ages. The baseline characteristics of the pooled Baseline characteristics were reported as study population (N = 8106) by 10-year age frequencies and percentages for categorical strata are summarised in Table 2. At baseline, variables and as the mean, standard deviation the mean (± SD) age of the overall study pop- (SD), median and first and third quartiles ulation was 63.8 ± 9.7 years, with a mean (± (Q1–Q3) for continuous variables. Mixed SD) body mass index (BMI) of 32.0 ± 5.4 kg/m2 models for repeated measures (MMRM) in and a median (Q1–Q3) T2DM duration of 10.0 HbA1c and FPG were analysed to produce least (6.0–15.0) years. Slightly more than half of the squares (LS) mean estimates and 95% confi- patients (4323; 53.3%) were men. dence intervals (CIs) for each age subgroup. All Comparison of the five age subgroups other endpoints were assessed descriptively. revealed that the proportion of women was the Hypoglycaemic event rates were calculated as highest in the subgroup aged C 80 years (60.5 the number of events per patient-year of vs. 44.6–48.5% for the younger age subgroups) exposure. and that both baseline body weight and BMI Given the exploratory nature of the investi- decreased linearly with increasing age, with gation, there was no statistical adjustment for patients \ 50 years having the highest mean multiple comparisons. Missing patient baseline BMI (33.5 vs. 29.9–32.6 kg/m2 for the older age characteristics and missing outcome data were subgroups). Similarly, in studies reporting esti- noted in some studies; no imputation of miss- mated glomerular filtration rate (eGFR) ing data was performed. All statistical tests were (N = 1698), baseline eGFR decreased with two-sided, with a p value of \ 0.05 considered increasing age, with almost all patients to be statistically significant. All analyses were aged C 70 years experiencing some degree of performed using SAS version 9.4 (SAS Institute renal impairment. As expected, the median Inc, Cary, NC, USA). diabetes duration of the 10-year age strata increased linearly from the youngest (6 years) to the oldest (14 years) age subgroup. RESULTS At baseline, most study participants were previously treated with insulin (4920/8106; Study Population 60.7%) and with Gla-100 in particular (1926/ 8106; 23.8%). In addition, 70.2% of the overall In total, 8106 participants from 20 European study population (N = 5687) was previously countries were included in this patient-level treated with at least one non-insulin anti-hy- perglycaemic treatment. Compared to their Diabetes Ther (2021) 12:1073–1097 1083 Table 2 Baseline characteristics by 10-year age strata Characteristic Age subroups (years) < 50 50–59 60–69 70–79 ‡ 80 (N = 727) (N = 2030) (N = 3054) (N = 1847) (N = 448) Age (years) Mean ± SD 45.5 ± 3.7 55.2 ± 2.8 64.5 ± 2.9 73.9 ± 2.8 82.5 ± 2.3 Median (Q1–Q3) 47.0 56.0 65.0 74.0 82.0 (44.0–48.0) (53.0–58.0) (62.0–67.0) (71.0–76.0) (81.0–84.0) Female, n (%) 334 (45.9) 906 (44.6) 1376 (45.1) 895 (48.5) 271 (60.5) Body weight (kg) 96.6 ± 18.8 93.4 ± 16.8 91.6 ± 16.5 87.0 ± 15.5 81.3 ± 13.1 2 Body mass index (kg/m ) 33.5 ± 6.2 32.6 ± 5.9 32.2 ± 5.1 31.0 ± 5.0 29.9 ± 4.3 2 eGFR (mL/min/1.73 m ) 101.8 ± 19.1 90.2 ± 23.2 82.8 ± 43.5 73.0 ± 24.7 61.9 ± 23.1 Diabetes duration (years) 6.0 8.0 11.0 13.0 14.0 (3.0–10.0) (5.0–13.0) (7.0–16.0) (9.0–19.0) (9.0–20.0) Previous insulin use, n (%) 371 (51.0) 1188 (58.5) 1864 (61.0) 1212 (65.6) 285 (63.6) a Prior basal insulin use, n (%) 336 (46.2) 1065 (52.5) 1681 (55.0) 1077 (58.3) 241 (53.8) Insulin glargine 100 U/mL 128 (38.1) 422 (39.6) 713 (42.4) 530 (49.2) 133 (55.2) NPH insulin 114 (33.9) 319 (30.0) 501 (29.8) 276 (25.6) 52 (21.6) Insulin detemir 48 (14.3) 162 (15.2) 231 (13.7) 142 (13.2) 24 (10.0) Insulin degludec 30 (8.9) 83 (7.8) 99 (5.9) 82 (7.6) 23 (9.5) Prior basal insulin dose (U/day) 38.3 ± 27.4 37.7 ± 24.2 36.8 ± 23.3 34.2 ± 22.6 28.7 ± 16.9 b Prior rapid-acting insulin use, n (%) 83 (11.4) 253 (12.5) 377 (12.3) 211 (11.4) 27 (6.0) Insulin aspart 9 (10.8) 52 (20.6) 99 (26.3) 59 (28.0) 2 (7.4) Insulin glulisine 6 (7.2) 30 (11.9) 29 (7.7) 22 (10.4) 2 (7.4) Insulin lispro 8 (9.6) 23 (9.1) 44 (11.7) 26 (12.3) 12 (44.4) Other insulin 48 (57.8) 99 (39.1) 134 (35.5) 72 (34.1) 9 (33.3) Previous non-insulin anti-hyperglycaemic 465 (64.0) 1420 (70.0) 2083 (68.2) 1395 (75.5) 324 (72.3) treatment, n (%)c Biguanides 368 (79.1) 1073 (75.6) 1585 (76.1) 988 (70.8) 170 (52.5) Dipeptidyl peptidase-4 inhibitors 144 (31.0) 413 (29.1) 643 (30.9) 487 (34.9) 138 (42.6) Sulphonylurea 100 (21.5) 264 (18.3) 532 (25.5) 351 (25.2) 54 (16.7) SGLT-2 inhibitors 103 (22.2) 296 (20.8) 316 (15.2) 168 (12.0) 53 (16.4) GLP-1 receptor agonists 45 (9.7) 117 (8.2) 145 (7.0) 54 (3.9) 3 (0.9) Patients with C 1 CV event or risk factor, 441 (60.7) 1494 (73.6) 2502 (81.9) 1532 (82.9) 386 (86.2) n (%) Hypertension 355 (48.8) 1295 (63.8) 2266 (74.2) 1406 (76.1) 365 (81.5) 1084 Diabetes Ther (2021) 12:1073–1097 Table 2 continued Characteristic Age subroups (years) < 50 50–59 60–69 70–79 ‡ 80 (N = 727) (N = 2030) (N = 3054) (N = 1847) (N = 448) Dyslipidaemia 243 (33.4) 796 (39.2) 1350 (44.2) 679 (36.8) 112 (25.0) Peripheral arterial disease 31 (4.3) 248 (12.2) 546 (17.9) 390 (21.1) 100 (22.3) Previous myocardial infarction 19 (2.6) 142 (7.0) 301 (9.9) 214 (11.6) 65 (14.5) Previous stroke 3 (0.4) 92 (4.5) 225 (7.4) 184 (10.0) 44 (9.8) Other ischaemic heart disease 15 (2.1) 140 (6.9) 336 (11.0) 210 (11.4) 31 (6.9) Patients with C 1 diabetic complication, 140 (19.3) 469 (23.1) 895 (29.3) 412 (22.3) 45 (10.0) n (%) Diabetic neuropathy 107 (14.7) 340 (16.7) 639 (20.9) 249 (13.5) 14 (3.1) Diabetic retinopathy 22 (3.0) 133 (6.6) 281 (9.2) 156 (8.4) 14 (3.1) Diabetic nephropathy 22 (3.0) 112 (5.5) 211 (6.9) 154 (8.3) 12 (2.7) HbA1c (%) 9.10 ± 1.37 8.90 ± 1.33 8.80 ± 2.08 8.49 ± 1.05 8.46 ± 0.94 FPG (mg/dL) 185.5 ± 59.8 183.5 ± 55.6 182.1 ± 54.4 177.6 ± 51.7 184.0 ± 56.1 Data are expressed as the mean ± SD or the median with first and third quartiles (Q1–Q3) in parentheses, unless otherwise indicated N refers to all patients from the pooled REALI database included in the age subgroup mentioned; means and percentages are calculated based on data available for each variable CV cardiovascular, eGFR estimated glomerular filtration rate, GLP-1 glucagon-like peptide-1, NPH neutral protamine Hagedorn, SGLT-2 sodium glucose co-transporter-2 a The total number of patients who were previously treated with basal insulin in each age subgroup was used as the denominator to calculate the percentages of patients who received prior insulin glargine, NPH, detemir, or degludec b The total number of patients who were previously treated with rapid-acting insulin in each age subgroup was used as the denominator to calculate the percentages of patients who received prior insulin aspart, glulisine, lispro, or other c The total number of patients who were previously treated with non-insulin anti-hyperglycaemic agents in each age subgroup was used as the denominator to calculate the percentages of patients in each drug class younger counterparts, patients aged C 70 years proportion of patients with at least one diabetic were more likely to be treated with Gla-100 at complication was the highest in the subgroup baseline (49.2–55.2% of those previously treated aged 60–69 years (29.3%) and the lowest in with basal insulin versus 38.1–42.4%, respec- the C 80 years subgroup (10.0%). tively). The use of biguanides, sodium glucose Mean (± SD) baseline HbA1c decreased lin- co-transporter-2 inhibitors, and GLP-1 receptor early from the youngest (9.10% ± 1.37) to the agonists decreased with increasing age, while oldest (8.46% ± 0.94) age subgroup. By con- the use of dipeptidyl peptidase-4 inhibitors trast, mean (± SD) baseline FPG values were increased with increasing age (Table 2). higher in the \ 50 (185.5 ± 59.8 mg/dL) The proportion of patients with at least one and C 80 years (184.0 ± 56.1 mg/dL) subgroups concomitant cardiovascular event or risk factor than in the other age subgroups (Table 2). was higher with increasing age. By contrast, the Diabetes Ther (2021) 12:1073–1097 1085 Glycaemic Control - 1.03% across the 10-year age strata. At 24 weeks, the LS mean (95% CI) change in The improvement in HbA1c levels from baseline HbA1c from baseline was - 1.09% (- 1.18 to to weeks 12 and 24 of Gla-300 therapy was - 1.00), - 1.08% (- 1.14 to - 1.03), - 1.12% notable in all five age subgroups, with compa- (- 1.17 to - 1.07), - 1.18% (- 1.24 to - 1.12) rable HbA1c reductions across subgroups and - 1.11% (- 1.23 to - 0.99) in the \ 50, (Table 3). At 12 weeks, the LS mean decrease in 50–59, 60–69, 70–79 and C 80 years subgroups, HbA1c from baseline ranged from - 0.88 to respectively. Table 3 Mean HbA1c and changes in HbA1c from baseline to weeks 12 and 24 of Gla-300 treatment by 10-year age subgroups HbA1c (in %) parameters Age subroups (years) < 50 50–59 60–69 70–79 ‡ 80 (N = 708) (N = 1988) (N = 2992) (N = 1804) (N = 437) HbA1c at baseline, n 564 1583 2424 1390 297 Mean ± SD 9.13 ± 1.40 8.92 ± 1.34 8.81 ± 1.33 8.51 ± 1.08 8.45 ± 0.95 HbA1c at week 12, n 236 670 1155 556 59 Mean ± SD 7.80 ± 1.06 7.91 ± 1.08 7.87 ± 1.11 7.72 ± 0.97 7.61 ± 1.01 Change from baseline to week 233 657 1128 548 57 12, n LS mean ± SE - 0.92 ± 0.06 - 0.88 ± 0.03 - 0.89 ± 0.03 - 0.94 ± 0.04 - 1.03 ± 0.11 (95% CI) (- 1.03; (- 0.95; (- 0.94; (- 1.01; (- 1.24; - 0.81) - 0.81) - 0.83) - 0.87) - 0.82) LS mean ± SE differencea - 0.04 ± 0.06 0.03 ± 0.06 - 0.02 ± 0.07 0.11 ± 0.12 (95% CI) - (- 0.09; 0.16) (- 0.09; 0.15) (- 0.15; 0.11) (- 0.35; 0.12) Associated p value - 0.546 0.604 0.739 0.355 HbA1c at week 24, n 581 1633 2504 1455 318 Mean ± SD 7.78 ± 1.20 7.77 ± 1.16 7.68 ± 1.12 7.63 ± 0.98 7.58 ± 0.95 Change from baseline to week 554 1544 2362 1363 291 24, n LS mean ± SE - 1.09 ± 0.05 - 1.08 ± 0.03 - 1.12 ± 0.02 - 1.18 ± 0.03 - 1.11 ± 0.06 (95% CI) (- 1.18; (- 1.14; (- 1.17; (- 1.24; (- 1.23; - 1.00) - 1.03) - 1.07) - 1.12) - 0.99) LS mean ± SE differencea - 0.01 ± 0.05 - 0.03 ± 0.05 - 0.09 ± 0.05 - 0.02 ± 0.08 (95% CI) - (- 0.09; 0.11) (- 0.13; 0.06) (- 0.19; 0.01) (- 0.17; 0.13) Associated p value – 0.893 0.506 0.087 0.800 n refers to the number of patients with available data CI confidence interval, LS least squares, SE standard error a For the difference between the subgroups, the reference is the subgroup aged \ 50 years 1086 Diabetes Ther (2021) 12:1073–1097 Diabetes Ther (2021) 12:1073–1097 1087 b Fig.1 Percentage of patients achieving glycated haemo- 300 therapy was marginal across all age sub- globin (HbA1c) targets at week 24 of insulin glargine 300 groups (Table 6). U/mL (Gla-300) treatment, by 10-year age strata DISCUSSION The improvement in HbA1c across the 10-year age strata translated into similar HbA1c Findings from the REALI pooled analysis target achievements at 24 weeks of Gla-300 demonstrated that treatment with Gla-300 ini- therapy (Fig. 1). Between 22.8 and 26.1% of tiated in patients with uncontrolled T2DM patients across the five age subgroups achieved improved glycaemic control with a favourable HbA1c \ 7.0% at 24 weeks, and between 43.3 safety profile across a wide range of ages. In and 49.0% achieved HbA1c \ 7.5%. daily practice, practitioners treating elderly In line with changes in HbA1c, the LS mean patients with uncontrolled T2DM may face change in FPG from baseline to weeks 12 and 24 more challenges than with younger patients, of Gla-300 therapy was also comparable across due to age-related deterioration in glucose tol- the 10-year age strata (Table 4). erance, a reduction in endogenous insulin secretion and difficulties in adhering to com- Safety plex self-care activities [3, 4, 28]. Aging may also modify the counterregulatory and symptomatic The incidence and rate of hypoglycaemic events responses to hypoglycaemia, which can lead to reported over the 24-week Gla-300 treatment less intense symptoms of hypoglycaemia, con- period were overall low, with the incidence of sequently increasing the risk of hypoglycaemia hypoglycaemia occurring at any time of day in elderly patients due to hypoglycaemia ranging from 8.9 to 12.4% and the incidence of unawareness [6, 29]. Finally, progressive renal nocturnal hypoglycaemia ranging from 0.7 to impairment, as well as insulin deficiency 3.1% across the five age subgroups (Table 5). requiring insulin therapy, may contribute to the Compared to the younger age subgroups, a higher risk of hypoglycaemia in older adults lower incidence of symptomatic hypoglycaemia [4, 6]. occurring at any time of the day or during the Although a key objective in older people night was recorded in patients aged C 80 years with uncontrolled T2DM is to minimise hypo- (5.9% vs. 7.6–9.4% or 0.5% vs. 1.6–2.5%, glycaemia, achieving appropriate glycaemic respectively). Very few severe hypoglycaemic goals remains important [3, 4, 12]. The Endo- episodes occurring at any time of the day were crine Society recommends individualised gly- reported, but the highest incidence was caemic targets ranging from C 7.0% observed in the subgroup aged C 80 years (1.1 (53 mmol/mol) to \ 8.5% (69 mmol/mol) in vs. 0.1–0.6% for the younger age subgroups). No older adults tailored to overall health (e.g. nocturnal severe hypoglycaemic events were number of comorbidities, degree of cognitive reported in the subgroup aged C 80 years impairment) and to management strategies (e.g. (Table 5). where medication that can cause hypogly- The daily dose of Gla-300 (expressed in both caemia is used) [3]. Similarly, according to the U/day and in U/kg/day) increased over the most recent clinical practice recommendations 24-week Gla-300 treatment period in all five age of the American Diabetes Association (ADA), subgroups (Table 6), but the mean increase was older adults who are otherwise healthy with few the highest in the subgroup aged \ 50 years coexisting chronic illnesses and intact cognitive (? 9.21 U/day and ? 0.10 U/kg/day at function and functional status should adhere to 24 weeks) and the lowest in the subgroup a HbA1c target \ 7.0–7.5%, while those with aged C 80 years (? 5.12 U/day and ? 0.07 multiple coexisting chronic illnesses, cognitive U/kg/day at 24 weeks). impairment or functional dependence should The mean (± SD) change in body weight have less stringent glycaemic goals (such as from baseline to week 12 and week 24 of Gla- HbA1c \ 8.0–8.5%) [4]. In the REALI pooled 1088 Diabetes Ther (2021) 12:1073–1097 analysis, approximately half of the patients, The clinically important and consistent including those in the older age subgroups, reductions from baseline in HbA1c levels that achieved target HbA1c values \ 7.5% at were noted in REALI in different age subgroups 24 weeks of Gla-300 therapy, and approxi- are supported by the results of real-world studies mately two-thirds achieved an HbA1c tar- and RCTs evaluating Gla-300 in patients get \ 8.0%, which is mostly consistent with the aged C 65 years with T2DM who were uncon- level of glycaemic control recommended by the trolled on their prior anti-hyperglycaemic regi- current clinical practice guidelines [3, 4, 8]. men [5, 12, 30, 31]. In DELIVER 3 [31], a propensity-matched, retrospective, cohort Table 4 Mean FPG and changes in FPG from baseline to weeks 12 and 24 of Gla-300 treatment by 10-year age subgroups FPG (mg/dL) parameters Age subroups (years) < 50 50–59 60–69 70–79 ‡ 80 (N = 708) (N = 1988) (N = 2992) (N = 1804) (N = 437) FPG at baseline, n 358 1140 1704 1164 277 Mean ± SD 187.4 ± 58.7 184.1 ± 54.8 181.9 ± 54.2 177.8 ± 51.5 185.0 ± 52.0 FPG at week 12 260 838 1249 879 209 Mean ± SD 143.5 ± 40.7 145.2 ± 46.0 142.1 ± 43.8 138.5 ± 39.4 137.8 ± 39.9 Change from baseline to week 226 734 1118 769 159 12, n LS mean ± SE - 36.9 ± 2.7 - 33.8 ± 1.6 - 36.5 ± 1.3 - 38.5 ± 1.6 - 44.1 ± 3.2 (95% CI) (- 42.2; (- 36.9; (- 39.1; (- 41.5; (- 50.5; - 31.6) - 30.7) - 33.9) - 35.4) - 37.8) LS mean ± SE differencea - 3.1 ± 3.0 0.4 ± 2.9 - 1.6 ± 3.0 - 7.3 ± 4.1 (95% CI) - (- 2.8; 9.0) (- 5.2; 6.0) (- 7.4; 4.2) (- 15.2; 0.7) Associated p value - 0.298 0.889 0.592 0.0752 FPG at week 24, n 389 1215 1811 1253 321 Mean ± SD 145.3 ± 46.0 146.9 ± 47.1 140.4 ± 42.4 142.4 ± 76.0 136.7 ± 40.1 Change from baseline to week 343 1076 1621 1094 259 24, n LS mean ± SE - 37.1 ± 3.0 - 35.0 ± 1.8 - 40.1 ± 1.5 - 38.2 ± 1.8 - 45.3 ± 3.4 (95% CI) (- 42.9; (- 38.4; (- 43.0; (- 41.7; (- 52.0; - 31.3) - 31.5) - 37.2) - 34.8) - 38.5) LS mean ± SE differencea - 2.1 ± 3.3 - 3.0 ± 3.2 - 1.1 ± 3.3 - 8.2 ± 4.4 (95% CI) - (- 4.3; 8.6) (- 9.2; 3.2) (- 7.6; 5.4) (- 16.8; 0.5) Associated p value - 0.516 0.341 0.738 0.064 n refers to the number of patients with available data a For the difference between the subgroups, the reference is the subgroup aged \ 50 years Diabetes Ther (2021) 12:1073–1097 1089 Table 5 Incidence and event rate of hypoglycaemic events, by 10-year age strata Incidence/event rate of Age subroups (years) hypoglycaemia < 50 (N = 708) 50–59 60–69 70–79 ‡ 80 (N = 1988) (N = 2992) (N = 1804) (N = 437) Total patient-year exposure 308.82 862.64 1313.51 770.21 174.16 Any time of the day hypoglycaemia Any hypoglycaemia Patients with C 1 event, n 63 (8.9) 179 (9.0) 332 (11.1) 224 (12.4) 44 (10.1) (%) Total number of events (event 277 (0.897) 790 (0.916) 1566 (1.192) 1128 (1.465) 317 (1.820) rate)a Symptomatic hypoglycaemiab Patients with C 1 event, n 56 (7.9) 151 (7.6) 274 (9.2) 170 (9.4) 26 (5.9) (%) Total number of events (event 243 (0.787) 614 (0.712) 1105 (0.841) 706 (0.917) 131 (0.752) rate) Severe hypoglycaemiac Patients with C 1 event, n 1 (0.1) 8 (0.4) 10 (0.3) 10 (0.6) 5 (1.1) (%) Total number of events (event 1 (0.003) 8 (0.009) 22 (0.017) 15 (0.019) 6 (0.034) rate) Nocturnal hypoglycaemia Any hypoglycaemia Patients with C 1 event, n 13 (1.8) 47 (2.4) 92 (3.1) 46 (2.5) 3 (0.7) (%) Total number of events (event 38 (0.123) 126 (0.146) 232 (0.177) 78 (0.101) 5 (0.029) rate) Symptomatic hypoglycaemiaa Patients with C 1 event, n 11 (1.6) 43 (2.2) 74 (2.5) 38 (2.) 2 (0.5) (%) Total number of events (event 36 (0.117) 99 (0.115) 161 (0.123) 50 (0.065) 3 (0.017) rate) Severe hypoglycaemia Patients with C 1 event, n 0 4 (0.2) 3 (0.1) 2 (0.1) 0 (%) 1090 Diabetes Ther (2021) 12:1073–1097 Table 5 continued Incidence/event rate of Age subroups (years) hypoglycaemia < 50 (N = 708) 50–59 60–69 70–79 ‡ 80 (N = 1988) (N = 2992) (N = 1804) (N = 437) Total number of events (event 0 4 (0.005) 8 (0.006) 2 (0.003) 0 rate) a Event rates, which are based on total patient-year exposure, are expressed as the number of events per year b Symptomatic hypoglycaemia was defined as an event during which typical symptoms of hypoglycaemia occurred (e.g. sweating, hunger, shakiness, palpitations) c Severe hypoglycaemia was defined as any event requiring assistance from another person to actively administer carbo- hydrates or glucagon, or take other corrective actions study examining clinical outcomes in 2352 compared with Gla-100 (1.5 vs. 10.4%; relative patients with T2DM aged C 65 years switching risk 0.33; 95% CI 0.12–0.88) [5]. Even though, from basal insulin to Gla-300 or to a first-gen- compared to the present analysis, the study eration basal insulin (insulin detemir or Gla- populations of SENIOR [5] and DELIVER 3 [30] 100) in real-world clinical practice, HbA1c were older (mean age of approx. 71 vs. 64 years reductions were comparable in both cohorts, in REALI), with a higher proportion of diabetic with significantly reduced hypoglycaemia inci- complications (approx. 50 vs. 24% in REALI), dences and event rates in the Gla-300 cohort the results of the REALI analysis using 10-year [31]. Compared to REALI, patients in the Gla- age strata support the results of the SENIOR RCT 300 cohort of DELIVER 3 (N = 1176) had higher [5] and the DELIVER three real-world analysis mean HbA1c levels at the 3- to 6-months fol- [30]. low-up assessment (8.12% from a baseline Overall, the REALI findings, along with the HbA1c of 8.60% vs. 7.58–7.68% at week 24 from data from the aforementioned studies a baseline of 8.45–8.81% in patients [5, 12, 30, 31], indicate that Gla-300 is a treat- aged C 60 years in REALI). HbA1c target ment option equally beneficial in both younger attainment was also lower in the Gla-300 cohort and older patients with T2DM, achieved of DELIVER 3 [30] compared to REALI (\ 7.0%: through a sustained glycaemic control which 19.3 vs. 25.0–26.1% in patients aged C 60 years contributes to minimising the risk of hypogly- in REALI; \ 8.0%: 50.9 vs. 63.5–69.8%). In the caemia. Although insulin therapy, particularly 26-week SENIOR RCT [5] conducted in 1014 intensive insulin therapy with basal insulin patients aged C 65 years with uncontrolled alone or with basal-bolus insulin, has been T2DM who received either Gla-300 (N = 508) or associated with weight gain in elderly patients Gla-100 (N = 506), mean HbA1c decreased from [3], Gla-300 therapy had a weight-neutral effect 8.20% at baseline to 7.31% at week 26 among in the present analysis across the evaluated age Gla-300-treated patients, with the proportion of subgroups. This represents a practical advantage Gla-300-treated patients reaching HbA1c targets for both patients aged \ 50 years who had the of \ 7.0% and \ 7.5% of 33.3 and 60.6%, highest mean baseline BMI (33.5 kg/m2) and respectively. Among patients aged C 75 years, those aged C 80 years who had a lower mean who formed approximately 20% of the SENIOR baseline BMI (29.9 kg/m2) and who were able to study population, similar reductions in mean maintain a stable body weight. HbA1c from baseline to week 26 were observed, Although elderly patients with T2DM are from 8.17 to 7.29% in Gla-300-treated patients known to have a greater risk of hypoglycaemia (N = 135), with a significantly lower incidence compared to younger ones, Gla-300 therapy was of documented symptomatic hypoglycaemia associated with overall low incidence of hypo- occurring at any time of the day with Gla-300 glycaemia in the present analysis across the Diabetes Ther (2021) 12:1073–1097 1091 Table 6 Changes in Gla-300 daily dose and body weight from baseline to weeks 12 and 24 of treatment, by 10-year age strata Gla-300 dose/body weight Age subroups (years) < 50 years 50–59 60–69 70–79 ‡ 80 (N = 708) (N = 1988) (N = 2992) (N = 1804) (N = 437) Gla-300 daily dose (U/day) Baseline 28.41 ± 21.67 28.58 ± 18.77 28.61 ± 19.06 25.80 ± 17.07 21.03 ± 13.86 Week 12 37.27 ± 24.10 35.54 ± 20.00 35.70 ± 20.59 30.90 ± 16.91 26.26 ± 16.65 Change from baseline to 8.68 ± 14.05 7.17 ± 12.14 7.10 ± 11.41 5.67 ± 10.15 5.86 ± 11.46 week 12 Week 24 38.45 ± 23.99 37.62 ± 21.78 36.60 ± 21.27 31.88 ± 19.73 26.36 ± 15.74 Change from baseline to 9.21 ± 13.82 8.73 ± 13.97 7.87 ± 13.53 6.03 ± 13.45 5.12 ± 10.61 week 24 Gla-300 daily dose (U/kg/day) Baseline 0.30 ± 0.25 0.30 ± 0.18 0.30 ± 0.18 0.29 ± 0.17 0.26 ± 0.17 Week 12 0.44 ± 0.26 0.42 ± 0.20 0.41 ± 0.20 0.37 ± 0.16 0.34 ± 0.20 Change from baseline to 0.12 ± 0.14 0.10 ± 0.14 0.09 ± 0.12 0.07 ± 0.11 0.03 ± 0.10 week 12 Week 24 0.40 ± 0.23 0.39 ± 0.20 0.39 ± 0.21 0.36 ± 0.20 0.33 ± 0.19 Change from baseline to 0.10 ± 0.13 0.09 ± 0.14 0.09 ± 0.15 0.08 ± 0.15 0.07 ± 0.12 week 24 Body weight (kg) Baseline 96.56 ± 18.90 93.45 ± 16.80 91.58 ± 16.48 86.95 ± 15.52 81.16 ± 12.98 Week 12 90.47 ± 14.95 90.51 ± 14.38 88.86 ± 14.58 85.29 ± 14.47 73.35 ± 10.44 Change from baseline to -0.06 ± 2.76 0.01 ± 3.14 0.04 ± 2.38 0.00 ± 2.24 0.02 ± 1.59 week 12 Week 24 97.15 ± 18.81 94.96 ± 17.49 92.27 ± 16.83 87.50 ± 15.36 81.95 ± 13.55 Change from baseline to 0.03 ± 5.01 0.12 ± 4.25 0.04 ± 3.67 -0.14 ± 3.82 -0.19 ± 3.33 week 24 All data are expressed as mean ± SD evaluated age subgroups. In addition to its social and medical support has been shown to evenly distributed and stable pharmacokinetic reduce disease-related distress and hypogly- exposure and pharmacodynamic profile [32], caemia risk without worsening glycaemic con- the simple, once-daily dosing regimen of Gla- trol [3, 4, 8]. A lower incidence and event rate of 300 may have contributed to this lower inci- symptomatic hypoglycaemia occurring during dence of hypoglycaemia. Indeed, simplification the night or at any time of the day were recor- of insulin regimens to match an individual’s ded in patients aged C 80 years compared to self-management abilities and their available younger subgroups. We assume that this is 1092 Diabetes Ther (2021) 12:1073–1097 likely to be related to the impact of aging on hypoglycaemia with Gla-300 across various counterregulatory and symptomatic responses, patient subgroups with high hypoglycaemia thereby reducing the intensity of hypogly- risk, including both insulin-naı̈ve patients caemia symptoms [29]. A small study from the aged C 65 (N = 20885) and C 75 years UK, which compared the responses to hypo- (N = 10325) and patients switching from glycaemia of young and elderly patients with- another basal insulin analogue aged C 65 (N = out diabetes, showed that autonomic and 15837) and C 75 years (N = 5654). In all sub- neuroglycopenic symptom scores were signifi- group analyses, Gla-300 was associated with cantly lower in the older group [33]. Another lower rates of severe hypoglycaemia compared small Canadian study similarly found dimin- to first-generation basal insulin analogues, such ished autonomic activation leading to attenua- as Gla-100 and insulin detemir, irrespective of tion of symptom intensity as a feature of aging, prior insulin therapy status [36]. Similarly, independent of any effects of diabetes [34]. In a another post-hoc analysis, investigating the more recent study from the USA among 40 association of baseline patient characteristics patients aged C 69 years with HbA1c val- with key outcomes reported from the EDITION ues [ 8.0%, 95 of the 102 (93.1%) hypogly- 1, 2 and 3 trials, found that the comparable caemic episodes recorded were unrecognised by glycaemic control of Gla-300 versus Gla-100 symptoms or by fingerstick glucose measure- with less hypoglycaemia seen in the EDITION ments performed four times a day [35]. The studies was observed, irrespective of age, body lower incidence of symptomatic hypoglycaemia mass index, age at T2DM onset or duration of occurring during the night or at any time of the T2DM [37]. In summary, the sustained gly- day that was reported in REALI patients caemic benefits of Gla-300 in the older adult aged C 80 years might also have been related to population, as well as its reduced risk of hypo- a more cautious use of Gla-300 in the oldest glycaemia compared to first-generation basal patients, as the Gla-300 dose change (expressed insulin analogues, support its use in older adults in U/kg/day) was approximately two-thirds the with T2DM. In addition, the flexibility and change observed in the youngest age subgroup. convenience of a once-daily injection of Gla- Lack of hypoglycaemic symptoms recogni- 300 is advantageous in this population who tion can render elderly patients more suscepti- may rely on caretakers to administer insulin ble to severe hypoglycaemia [6], as reflected in [38]. the present pooled analysis by the higher inci- Somewhat surprising was the inverse rela- dence of severe hypoglycaemia occurring at any tionship between baseline glycaemic status and time of the day in patients aged C 80 years age, as indicated by the lower baseline mean compared to younger ones. The incidence of HbA1c with increasing age. Nevertheless, sev- severe hypoglycaemia occuring at any time of eral cross-sectional studies have reported a the day remains, however, low in REALI similar relationship between baseline HbA1c (0.1–1.1% across age groups) and in the range of and age [39–41]. The reasons for such relation- that observed in the SENIOR RCT (0.8% for all ship need to be considered. The high baseline Gla-300 treated patients and 0% for patients HbA1c levels seen in the youngest age subgroup aged C 75 years) [5]. Overall, the low risk of may be related to the rapid changes in lifestyle hypoglycaemia with Gla-300 across a wide that expose people, including those with dia- range of ages is an important finding, particu- betes, to increased biological and behavioural larly for older adults with T2DM, given that risk factors [39, 41]. It has also been speculated clinical concern relating to hypoglycaemia and that older patients may have a different patho- its associated adverse events is often a barrier to physiological form of T2DM than younger ones, effective dose adjustment and attainment of as found in a data-driven cluster analysis con- target glycaemic control [32]. ducted among 8980 adults with newly diag- Using information from a U.S. electronic nosed diabetes in which four subgroups of health records database, the real-world LIGHT- T2DM were identified with significantly differ- NING study [36] predicted the rate of severe ent patient characteristics and risk of diabetic Diabetes Ther (2021) 12:1073–1097 1093 complications [42]. One of these subgroups was 1.4% of the pooled study population. The labelled as mild age-related diabetes; patients in inclusion of hospitalised patients does not this cluster are older, with modest metabolic influence the results of REALI, given their mar- derangements and a lower HbA1c at diagnosis ginal number in the pooled analysis. However, compared to patients in other clusters, such as we have also elected to pool the results of the severe autoimmune diabetes and severe insulin- Toujeo-Neo (ISRCTN number: deficient diabetes [42]. ISRCTN93674355) and Toujeo-BB [18] studies At baseline, more than two-thirds of the conducted among patients with T2DM previ- patients had been previously treated with at ously treated with basal-bolus regimens, since least one non-insulin anti-hyperglycaemic the main focus of REALI was to assess the treatment. Interestingly, in a post-hoc analysis effectiveness and safety of Gla-300 in a broad [43] of patient-level data from the EDITION 3 range of European patients with T2DM in daily RCT and de-identified data from the Clinfor- clinical practice settings. The inclusion of matics real-world claims database, Gla-300 patients on basal-bolus insulin regimens could therapy initiated in insulin-naı̈ve patients with have potentially impacted the results of REALI T2DM uncontrolled on oral antidiabetic drugs due to different hypoglycaemia risks than in (OADs) was associated with reductions in prior insulin-naı̈ve patients or in patients previously OAD therapy without compromising glycaemic treated with basal insulin therapy only. Several control, while preserving the hypoglycaemic strengths of this pooled analysis deserve to be benefit of Gla-300 versus Gla-100 [43]. The ADA noted, such as the inclusion of a large number [4] currently recommends simplification of of participants, including 2295 patients aged C treatment regimens in older patients with 70 years (28.3%), from several prospective T2DM to reduce the risk of hypoglycaemia and studies, thereby increasing the statistical power polypharmacy. Thus, since the post-hoc analy- of the analysis, which resulted in a more precise sis [43] of data from EDITION 3 and from the estimate of the therapeutic benefit and safety of Clinformatics real-world database suggests that Gla-300. In addition, the REALI pooled analysis patients treated with Gla-300 could step down applied standardised endpoint definitions to OAD use without jeopardising glycaemic con- reduce study-specific differences. Most impor- trol and with a reduced hypoglycaemia risk, tantly, the REALI analysis provides valuable these findings, in line with those of REALI, information regarding the safety and effective- could have important ramifications for clinical ness of Gla-300 in an older group of individuals decision-making in older T2DM populations who are often excluded or underrepresented in regarding regimen simplification. clinical trials and includes data from non-in- Limitations of the REALI pooled analysis terventional studies close to real-world clinical include the unbalanced and uncontrolled practice. number of patients across the different age subgroups and the lack of assessments of cog- nitive function, functional capacity or frailty. CONCLUSION REALI is also a post-hoc analysis, rather than a dedicated prospective trial in older individuals Gla-300 was found to be effective with a good with T2DM. Another limitation is the lack of safety profile in both younger and older patients comparative data with another basal insulin. with uncontrolled T2DM, indicating that it may Moreover, the results of REALI may not have be a suitable therapeutic option in elderly accounted for certain elderly individuals with patients who represent a vulnerable population T2DM, particularly in those aged C 80 years, that is prone to hypoglycaemia. who were not accessible to enrolment in inter- ventional or observational studies [44]. Fur- thermore, the REALI pooled analysis included the COBALTA study [25] conducted in 112 hospitalised patients, who represented less than 1094 Diabetes Ther (2021) 12:1073–1097 ACKNOWLEDGEMENTS Lilly, Merck Sharp & Dohme, Novo Nordisk and Sanofi. Nick Freemantle has received funding The authors thank the participants, staff, and for consulting, research or speaker fees from investigators of all pooled studies. Sanofi, Novo Nordisk, Allergan, Merck Sharp & Dohme, AstraZeneca, PCT, Ipsen, Takeda, Akcea and Accelovance. Gregory Bigot is an IVIDATA Funding. This study was funded by Sanofi employee for Sanofi. Celine Mauquoi is an IDDI (Paris, France). The article processing charges employee, and has acted as a biostatistics con- were also funded by Sanofi (Paris, France). tractor for Sanofi. Alice Ciocca and Mireille Medical Writing and Editorial Assis- Bonnemaire are Sanofi employees. Pierre tance. Medical writing support in accordance Gourdy has received advisory board and speaker with Good Publication Practice (GPP3) guideli- honoraria from Abbott, Amgen, AstraZeneca, nes (http://www.ismpp.org/gpp3) was provided Novo Nordisk, Boehringer Ingelheim, Eli Lilly, by Thomas Rohban, MD, and Magalie El Hajj, Merck Sharp & Dohme, Mundipharma, Sanofi PharmD, of Partner 4 Health (Paris, France), and and Servier. was funded by Sanofi. Compliance with Ethics Guidelines. This Authorship. All named authors meet the analysis did not involve primary data collection International Committee of Medical Journal by the authors, and all post-hoc analysed data Editors (ICMJE) criteria for authorship for this were anonymised; consequently, separate ethi- article, take responsibility for the integrity of cal approval was not required. Protocols for all the work as a whole, and have given their included studies were approved by the appro- approval for this version to be published. priate ethics committees, and the studies were conducted according to Good Clinical Practice Authorship Contributions. All authors con- and the Declaration of Helsinki. tributed to the project design and the analysis plan. Celine Mauquoi performed the statistical Data Availability. The datasets generated analysis of the data. All authors were involved during and/or analysed during the current in the interpretation of the data, writing and study are available from the corresponding reviewing drafts of the manuscript, and author on reasonable request. approved the final version for submission. Open Access. This article is licensed under a Prior Presentation. Preliminary results were Creative Commons Attribution-NonCommer- presented at the 55th Annual Meeting of the cial 4.0 International License, which permits European Association for the Study of Diabetes, any non-commercial use, sharing, adaptation, 16–20 September 2019, Barcelona, Spain. distribution and reproduction in any medium or format, as long as you give appropriate credit Disclosures. Riccardo C. Bonadonna has to the original author(s) and the source, provide received funding for consulting or speaker fees a link to the Creative Commons licence, and from Sanofi, Merck Sharp & Dohme, Bristol- indicate if changes were made. The images or Myers Squibb, AstraZeneca, Janssen, Eli Lilly other third party material in this article are and Johnson & Johnson. Didac Mauricio has included in the article’s Creative Commons received funding for consulting or speaker fees licence, unless indicated otherwise in a credit from Almirall, Ascensia, AstraZeneca, Boehrin- line to the material. If material is not included ger Ingelheim, GlaxoSmithKline, Eli Lilly, Fer- in the article’s Creative Commons licence and rer, Janssen, Menarini, Merck Sharp & Dohme, your intended use is not permitted by statutory Novartis, Novo Nordisk and Sanofi. Dirk Müller- regulation or exceeds the permitted use, you Wieland has received funding for consulting or will need to obtain permission directly from the speaker fees from Amarin, Amgen, AstraZeneca, copyright holder. To view a copy of this licence, Bayer, Boehringer Ingelheim, Daiichi-Sankyo, Diabetes Ther (2021) 12:1073–1097 1095 visit http://creativecommons.org/licenses/by- in people with type 2 diabetes. Diabetes Obes nc/4.0/. Metab. 2015;17(9):859–67. 12. Yale JF, Aroda VR, Charbonnel B, et al. Glycaemic control and hypoglycaemia risk with insulin glar- gine 300 U/mL versus glargine 100 U/mL: a patient- REFERENCES level meta-analysis examining older and younger adults with type 2 diabetes. Diabetes Metab. 2020;46(2):110–8. 1. International Diabetes Federation. IDF diabetes atlas: ninth edition 2019. 2019. https://www. 13. Kirkman MS, Briscoe VJ, Clark N, et al. Diabetes in diabetesatlas.org/upload/resources/2019/IDF_ older adults: a consensus report. J Am Geriatr Soc. Atlas_9th_Edition_2019.pdf. Accessed 08 Sep 2020. 2012;60(12):2342–56. 2. Longo M, Bellastella G, Maiorino MI, Meier JJ, 14. Dardano A, Penno G, Del Prato S, Miccoli R. Opti- Esposito K, Giugliano D. Diabetes and aging: from mal therapy of type 2 diabetes: a controversial treatment goals to pharmacologic therapy. Front challenge. Aging. 2014;6(3):187–206. Endocrinol (Lausanne). 2019;10:45. 15. Lakey WC, Barnard K, Batch BC, Chiswell K, Tas- 3. LeRoith D, Biessels GJ, Braithwaite SS, et al. Treat- neem A, Green JB. Are current clinical trials in ment of diabetes in older adults: an endocrine diabetes addressing important issues in diabetes society* clinical practice guideline. J Clin Endocri- care? Diabetologia. 2013;56(6):1226–35. nol Metab. 2019;104(5):1520–74. 16. Freemantle N, Bonadonna RC, Gourdy P, et al. 4. American Diabetes Association. 12. Older adults: Rationale and methodology for a European pooled standards of medical care in diabetes—2021. Dia- analysis of postmarketing interventional and betes Care. 2021;44(Suppl 1):S168–79. observational studies of insulin glargine 300 U/mL in diabetes: protocol of REALI project. BMJ Open. 5. Ritzel R, Harris SB, Baron H, et al. A randomized 2020;10(4):e033659. controlled trial comparing efficacy and safety of insulin glargine 300 Units/mL versus 100 units/ml 17. Wiesli P, Schories M. Improved glycemic control in older people with type 2 diabetes: results from with insulin glargine 300 U/mL (ToujeoÒ) in the SENIOR Study. Diabetes Care. 2018;41(8): patients with type 2 diabetes: real-world effective- 1672–80. ness in Switzerland. Diabetes Ther. 2018;9(6): 2325–34. 6. Sathyanarayanan A, Rabindranathnambi A, Muraleedharan V. Pharmacotherapy of type 2 dia- 18. Hidvegi T, Stella P. Effectiveness of insulin glargine betes mellitus in frail elderly patients. Br J Hosp U300 used as part of basal bolus therapy in people Med (Lond). 2019;80(11):C162–5. with T2DM—Toujeo 6 months real-world data from Hungary [Abstract]. Diabetes. 2018;67(Suppl 1): 7. Kalyani RR, Golden SH, Cefalu WT. Diabetes and 1049-P. aging: unique considerations and goals of care. Diabetes Care. 2017;40(4):440–3. 19. Russell-Jones D, Dauchy A, Delgado E, et al. Take Control: a randomized trial evaluating the efficacy 8. Meneilly GS, Knip A, Miller DB, Sherifali D, Tessier and safety of self- versus physician-managed titra- D, Zahedi A. Diabetes in older people. Can J Dia- tion of insulin glargine 300 U/mL in patients with betes. 2018;42(Suppl 1):S283–95. uncontrolled type 2 diabetes. Diabetes Obes Metab. 2019;21(7):1615–24. 9. Abdelhafiz AH, Sinclair AJ. Management of type 2 diabetes in older people. Diabetes Ther. 2013;4(1): 20. Bonadonna RC, Giaccari A, Buzzetti R, et al. Italian 13–26. Titration Approach Study (ITAS) with insulin glar- gine 300 U/mL in insulin-naı̈ve type 2 diabetes: 10. Munshi MN, Slyne C, Segal AR, Saul N, Lyons C, design and population. Nutr Metab Cardiovasc Dis. Weinger K. Simplification of insulin regimen in 2019;29(5):496–503. older adults and risk of hypoglycemia. JAMA Intern Med. 2016;176(7):1023–5. 21. Pfohl M, Jornayvaz FR, Fritsche A, et al. Effective- ness and safety of insulin glargine 300 U/mL in 11. Ritzel R, Roussel R, Bolli GB, et al. Patient-level insulin-naı̈ve patients with type 2 diabetes after meta-analysis of the EDITION 1, 2 and 3 studies: failure of oral therapy in a real-world setting. Dia- glycaemic control and hypoglycaemia with new betes Obes Metab. 2020;22(5):759–66. insulin glargine 300 U/ml versus glargine 100 U/ml 1096 Diabetes Ther (2021) 12:1073–1097 22. Thomann R, Zechmann S, Alexander-David N, Jor- 32. Owens DR. Pharmacokinetics and pharmacody- nayvaz FR. Real-world effectiveness of insulin glar- namics of insulin glargine 300 U/mL in the treat- gine 300 initiation in Switzerland. Diabetes Metab ment of diabetes and their clinical relevance. Expert Syndr Obes. 2020;13:2359–65. Opin Drug Metab Toxicol. 2016;12(8):977–87. 23. Prázný M, Flekač M, Jelı́nek P, Mašková J. Insulin 33. Brierley EJ, Broughton DL, James OF, Alberti KG. glargine 300 Units/mL effectiveness in patients Reduced awareness of hypoglycaemia in the elderly with T2DM uncontrolled by basal insulin in real-life despite an intact counter-regulatory response. QJM. settings in the Czech Republic. J Diabetes Mellitus. 1995;88(6):439–45. 2020;10(3):109–23. 34. Meneilly GS, Cheung E, Tuokko H. Counterregula- 24. Colin IM, Alexandre K, Bruhwyler J, Scheen A, tory hormone responses to hypoglycemia in the Verhaegen A. Patient-reported outcomes with elderly patient with diabetes. Diabetes. 1994;43(3): insulin glargine 300 U/mL in people with type 2 403–10. diabetes: the MAGE multicenter observational study. Diabetes Ther. 2020;11(8):1835–47. 35. Munshi MN, Segal AR, Suhl E, et al. Frequent hypoglycemia among elderly patients with poor 25. Perez A, Carrasco-Sánchez FJ, González C, et al. glycemic control. Arch Intern Med. 2011;171(4): Efficacy and safety of insulin glargine 300 U/mL 362–4. (Gla-300) during hospitalization and therapy intensification at discharge in patients with insuf- 36. Pettus J, Roussel R, Liz Zhou F, et al. Rates of ficiently controlled type 2 diabetes: results of the hypoglycemia predicted in patients with type 2 phase IV COBALTA trial. BMJ Open Diabetes Res diabetes on insulin glargine 300 U/ml versus first- Care. 2020;8(1):e001518. and second-generation basal insulin analogs: the real-world LIGHTNING study. Diabetes Ther. 26. Gourdy P, Bahloul A, Boultif Z, Gouet D, Guerci B. 2019;10(2):617–33. Efficacy and safety of switching patients inade- quately controlled on basal insulin to insulin glar- 37. Twigg SM, Escalada J, Stella P, et al. Association of gine 300 U/mL: the TRANSITION 2 study. Diabetes patient profile with glycemic control and hypo- Ther. 2020;11(1):147–59. glycemia with insulin glargine 300 U/mL in type 2 diabetes: a post hoc patient-level meta-analysis. 27. Wieringa TH, de Wit M, Twisk JW, Snoek FJ. Diabetes Ther. 2018;9(5):2043–53. Improved diabetes medication convenience and satisfaction in persons with type 2 diabetes after 38. Cheng AYY, Wong J, Freemantle N, Acharya SH, switching to insulin glargine 300 U/mL: results of Ekinci E. The safety and efficacy of second-genera- the observational OPTIN-D study. BMJ Open Dia- tion basal insulin analogues in adults with type 2 betes Res Care. 2018;6(1):e000548. diabetes at risk of hypoglycemia and use in other special populations: a narrative review. Diabetes 28. Basu R, Breda E, Oberg AL, et al. Mechanisms of the Ther. 2020;11(11):2555–93. age-associated deterioration in glucose tolerance: contribution of alterations in insulin secretion, 39. Gilliland SS, Carter JS, Skipper B, Acton KJ. HbA(1c) action, and clearance. Diabetes. 2003;52(7): levels among American Indian/Alaska Native 1738–48. adults. Diabetes Care. 2002;25(12):2178–83. 29. Alagiakrishnan K, Mereu L. Approach to managing 40. El-Kebbi IM, Cook CB, Ziemer DC, Miller CD, Gal- hypoglycemia in elderly patients with diabetes. lina DL, Phillips LS. Association of younger age with Postgrad Med. 2010;122(3):129–37. poor glycemic control and obesity in urban African Americans with type 2 diabetes. Arch Intern Med. 30. Munshi MN, Gill J, Chao J, Nikonova EV, Patel M. 2003;163(1):69–75. Insulin glargine 300 U/mL is associated with less weight gain while maintaining glycemic control 41. Al-Lawati JA, Barakat MN, Al-Maskari M, Elsayed and low risk of hypoglycemia compared with MK, Al-Lawati AM, Mohammed AJ. HbA1c levels insulin glargine 100 U/mL in an aging population among primary healthcare patients with type 2 with type 2 diabetes. Endocr Pract. 2018;24(2): diabetes mellitus in Oman. Oman Med J. 143–9. 2012;27(6):465–70. 31. Bailey TS, Wu J, Zhou FL, et al. Switching to insulin 42. Ahlqvist E, Storm P, Käräjämäki A, et al. Novel glargine 300 units/mL in real-world older patients subgroups of adult-onset diabetes and their associ- with type 2 diabetes (DELIVER 3). Diabetes Obes ation with outcomes: a data-driven cluster analysis Metab. 2019;21(11):2384–93. of six variables. Lancet Diabetes Endocrinol. 2018;6(5):361–9. Diabetes Ther (2021) 12:1073–1097 1097 43. Dailey G, Reid T, White J, et al. Improved glycaemic 44. Herrera AP, Snipes SA, King DW, Torres-Vigil I, control and lower hypoglycaemia risk with reduced Goldberg DS, Weinberg AD. Disparate inclusion of prior oral antidiabetes drug therapy in patients with older adults in clinical trials: priorities and oppor- type 2 diabetes treated with insulin glargine 300 tunities for policy and practice change. Am J Public U/mL. Endocrinol Diabetes Metab. 2018;1(4): Health. 2010;100(Suppl 1):S105–12. e00035.