Lower rates of hypoglycaemia in older individuals with type 2 diabetes using insulin degludec versus insulin glargine U100: Results from SWITCH 2

Aim This study aimed to investigate the safety of insulin degludec (degludec) in relation to age and risk of hypoglycaemia post hoc in individuals with type 2 diabetes (T2D) (SWITCH 2 trial). Methods In this crossover study, individuals with T2D who were at risk of hypoglycaemia were randomized to double‐blind treatment with degludec or insulin glargine 100 units/mL (glargine U100) ± oral antidiabetic drugs. After 32 weeks, patients crossed over to the other treatment. Primary endpoint was number of overall severe (positively adjudicated) or glucose‐confirmed (plasma glucose <56 mg/dL; 3.1 mmol/L) symptomatic hypoglycaemia events during the two 16‐week maintenance periods. Results For individuals ≤65 (n = 450) and >65 (n = 270) years, baseline median (range) duration of diabetes was 12 (1–40) vs 15 (1–54) years, mean HbA1c was 7.7% vs 7.4% and mean estimated glomerular filtration rate was 87.0 vs 63.7 mL/min/1.73 m2, respectively. No significant differences in HbA1c reduction were seen in individuals ≤65 or >65 years. During both maintenance periods, treatment with degludec lowered rates of hypoglycaemia (overall/nocturnal symptomatic) vs those with glargine U100 in individuals ≤65 (31% vs 43%) and >65 (30% vs 41%) years. With degludec and glargine U100, respectively, six vs nine severe hypoglycaemic events occurred in individuals ≤65 years and four vs eight events occurred in those >65 years. Adverse event rates were 3.2 and 3.3 events/patient‐year for individuals ≤65 years and were 3.5 and 4.1 events/patient‐year for individuals >65 years with degludec and glargine U100, respectively. Conclusion Treatment with degludec was safe and effective, with a frequency of hypoglycaemia lower than that with glargine U100 in both younger and older individuals (>65 years) with T2D.


| INTRODUCTION
Type 2 diabetes (T2D) is a chronic, progressive disease that frequently necessitates treatment with basal insulin to maintain adequate glycaemic control. 1,2 In an ageing population with increasing longevity, the global burden of diabetes in individuals 65 years of age or older is projected to increase from 122.8 million in 2017 to 253.4 million in 2045. 3 Hypoglycaemia, primarily associated with diabetes therapies, particularly insulin, is common in T2D, 4,5 and increases with age and longer duration of diabetes. 6 Non-severe episodes are associated with increased utilization of healthcare services and loss of work time, [7][8][9] as well as impairment of quality of life; prevention of these episodes is therefore important. Severe hypoglycaemia is of even greater concern, as it has been shown to be associated with increased risk of cardiovascular events and mortality. 10,11 As with non-severe events, severe episodes can increase utilization of healthcare resources, with adverse economic consequences. 12 Delay in intensifying treatment with insulin for many individuals with T2D is common, with fear of hypoglycaemia among patients and healthcare providers an important contributing factor. 13 The problem of clinical inertia may be magnified in the context of older individuals. 14 Treatment delay could place older individuals at greater risk of the microvascular and macrovascular complications of diabetes, as they often have less organ reserve and more comorbid conditions than younger individuals. 15 Basal insulin analogues, now in widespread use, have advantages over human insulin in reducing the risk of hypoglycaemia because of better pharmacokinetic/pharmacodynamic (PK/PD) profiles. 16 The basal insulin analogue insulin degludec (degludec) has a PK/PD profile with an ultra-long duration of action, 17 and these PK/PD properties have been shown to be preserved in elderly individuals. 18 In type 1 diabetes, degludec has a four-times lower PD variability than insulin glargine 100 units/mL (glargine U100) (AUC GIR 0-24h,SS , coefficient of variation, 20% vs 82%). 19 The performance of degludec vs glargine U100 has been studied in a large clinical development programme in which degludec was associated with a lower rate of hypoglycaemia, with rate reductions of 17%-86%, compared with glargine U100. 20 In a dedicated cardiovascular outcomes trial (DEVOTE), a statistically significant 40% lower rate of adjudicated severe hypoglycaemia was observed in individuals using degludec compared with those using glargine U100. 21 Most randomized trials of insulins include very few older individuals, a population that is heterogeneous with respect to prevalence and severity of comorbidity, frailty and overall health. 22 Consequently, little information exists concerning the performance of basal insulin analogues in older individuals with diabetes, particularly with respect to hypoglycaemia. A pre-planned meta-analysis of seven trials comparing degludec and glargine U100 in older patients (≥65 years) indicated that degludec had a 24% lower rate of overall confirmed hypoglycaemia vs glargine U100, and a 36% lower rate of confirmed nocturnal hypoglycaemia. 23 Secondary analysis of the DEVOTE trial, which compared degludec with glargine U100, showed that degludec was associated with a lower rate of hypoglycaemia than glargine U100, regardless of age. 24 A post hoc analysis of data from the SWITCH 2 trial 25 has been utilized in the present report to explore whether older (>65 years) individuals with T2D responded similarly to younger individuals, with respect to the definitions of hypoglycaemia used in the primary analysis and other safety parameters when comparing degludec with glargine U100.

| METHODS
The detailed design of the SWITCH 2 trial and results of the primary analysis have been published. 19 Briefly, SWITCH 2 was a randomized, double-blind, treat-to-target, two-period crossover trial in which adults (≥18 years) with T2D currently undergoing treatment with basal insulin, with or without oral antidiabetic drugs, were randomized 1:1 to receive degludec (Novo Nordisk, Bagsvaerd, Denmark) once daily and glargine U100 (Sanofi, Paris, France) once daily, in a randomized sequence by period. Participants were also randomized 1:1 within each sequence to a morning dose, between awakening and breakfast, or an evening dose, between the main meal and bedtime. To maintain blinding, both insulins were administered using vial and syringe In this post hoc analysis, data from the primary trial 25 were examined according to age category at baseline; the younger group comprised individuals 65 years of age or younger and the older group comprised individuals above 65 years of age. Statistical analysis was similar to that used in the primary trial. Briefly, a Poisson model with individuals as random effect, with treatment, period, sequence and dosing time as fixed effects, and with logarithm of the observation time as offset was used to estimate the rate ratio for each classification of hypoglycaemia during the maintenance period. 25 Age group was added to the model as a fixed class variable to facilitate age comparisons pooled across treatments.
Among younger participants, the distribution according to sex was similar (49.8% female); however, among older participants, there were fewer females than males (42.2% vs 57.8%). At baseline, younger participants tended to be heavier (mean body mass index, 32.8 [5.8]  and older participants, respectively). With respect to any age-related differences in inclusion criteria relevant to the risk of hypoglycaemia, a larger proportion of older participants had been treated with insulin for more than 5 years (52.6% vs 47.6% of older and younger participants, respectively). However, fewer older participants had experienced at least one severe hypoglycaemia event during the previous year (14.4% vs 17.6% of older and younger participants, respectively).
Completion rates were comparable for both younger and older participants, and comparable for both treatments, ranging from 89% to 91%.

| Comparisons between age groups in the pooled population
The cumulative number of hypoglycaemic events, by age group, during the two 16-week maintenance periods is shown in Figure 1

| Comparisons by treatment within age group
The observed rate of severe or blood glucose (BG)-confirmed symptomatic hypoglycaemia was lower for degludec compared with glargine U100 in older as well as in younger participants during the maintenance period (younger group, 184 vs 263 events/100 patientyears of exposure [PYE] for degludec and glargine U100, respectively; older group, 188 vs 269 events/100 PYE for degludec and glargine U100, respectively) ( Table 2), with an estimated 31% and 30% lower rates of severe or BG-confirmed symptomatic hypoglycaemic events with degludec compared with glargine U100, in the younger group  Serious AEs by preferred term and system organ class occurring ≥5% in either treatment arm for both age groups are listed in Table S1.

| Adverse events
These were mostly upper respiratory in nature. There were seven fatal events in total: two in patients treated with degludec (none in the younger group; one associated with sudden cardiovascular death and one caused by stroke in the older group) and five in patients treated with glargine U100 (one secondary to acute myocardial infarction and one caused by sepsis in the younger group; two associated with malignancy and one caused by sepsis in the older group).

| DISCUSSION
Results of the randomized, double-blind, crossover trial, SWITCH 2, were examined to assess the effect of age on hypoglycaemia risk, comparing degludec with glargine U100. The SWITCH 2 trial was powered to evaluate the superiority of degludec vs glargine U100 with respect to overall symptomatic hypoglycaemia. During the maintenance period, treatment with degludec was associated with statistically significantly lower rates  24 In contrast to the SWITCH 2 trial, the core trials in the meta-analysis were powered to detect differences in HbA1c. Nevertheless, these similarly lower rates across trials with heterogeneous patient populations, ranging from insulin-naïve to basalbolus users, with or without an increased risk of hypoglycaemia at baseline, support the overall benefit of treatment with degludec compared with glargine U100.
The number of severe hypoglycaemic events reported in the present analysis were not statistically different by treatment for either age group, probably because the overall number of events was very low. As this was a treat-to-target trial, as expected, no treatment differences were observed in change in HbA1c from baseline in either the older or the younger participants. Younger individuals required a higher mean insulin dose (U/kg) compared with older individuals throughout the trial, which may have been related to greater insulin resistance, in line with the tendency of a higher body mass index among younger individuals.
Older individuals with T2D are generally at increased risk of, and more vulnerable to, hypoglycaemia than younger individuals, for several reasons, including impaired renal function, 26 reduced ability to recognize and respond to hypoglycaemia, 27 and altered physiological responses to low glucose levels. 28 Symptoms of hypoglycaemia become less intense and their symptom profile changes with increasing age, 29 with symptomatic responses manifesting only at BG levels lower than those in younger individuals, leaving less time to recognize and respond to them. 30,31 Thus, when hypoglycaemia develops in an older individual with T2D, it might not be identified or reported, which may explain the absence of a significantly higher rate of overall hypoglycaemia in older individuals, as was observed in the present study.
Reduced awareness of hypoglycaemia with advancing age may increase the risk that an event progresses in severity and results in more severe events in the older age group. There was a numerically greater rate of severe hypoglycaemia in older individuals compared with younger individuals (RR, 1.38) in the current trial, although the trend was not statistically significant. Severe hypoglycaemia is generally much more common in real-world populations than in randomized clinical trials. 4,5,32 In addition, continuous glucose monitoring has demonstrated that many episodes are unrecognized and under-reported. 33 Furthermore, it has been estimated that only 5% of self-reported F I G U R E 2 Hypoglycaemic events for younger (≤65 years) and older (>65 years) individuals, by treatment group Abbreviations: BG, blood glucose; CI, confidence interval. Values are treatment ratios (insulin degludec/insulin glargine U100) for the two 16-week maintenance periods. P-values derived using a Poisson model with logarithm of exposure time (100 years) as offset; estimates adjusted for treatment period, period sequence and dosing time as fixed effects, and subjects as a random effect collected. It would have been valuable to assess whether frail patients were at higher risk of the differing severities of hypoglycaemia. In this study, a larger percentage (57.8%) of individuals in the older age group were male; this may affect the generalizability of results to the older adult population, which tends to have a greater proportion of females.
Significant strengths of the SWITCH 2 trial 25 include the doubleblinded, treat-to-target design. A crossover design allowed participants to serve as their own controls when comparing treatment efficacy. In terms of assessing safety, the studies were powered with hypoglycaemia as the primary endpoint, as opposed to HbA1c, which was used in other trials that were being conducted for regulatory purposes. Furthermore, severe as well as BG-confirmed symptomatic hypoglycaemia events were included, and all severe episodes were confirmed by adjudication. Notwithstanding the high proportion of older males in the trial, both the inclusion criteria for hypoglycaemia risk and the inclusion of older individuals in this study provide valuable insight into a population seen in real-world practice.
To conclude, in patients with T2D, older and younger patients were at similar risk of overall symptomatic hypoglycaemia or nocturnal symptomatic hypoglycaemia, but older patients showed a tendency toward higher risk of severe hypoglycaemia. Treatment with degludec led to similar reductions in HbA1c and a similar adverseevent profile, with a lower risk of hypoglycaemia than treatment with glargine U100, both in older and younger individuals with T2D.