Insulin analogues, such as insulin glargine, detemir and degludec, are structurally altered human insulins that mimic the pharmacokinetic properties of endogenous insulin more closely than intermediate-acting insulins. Because of the long duration of action and the less pronounced insulin peak, long-acting insulin analogues have less risk of hypoglycaemia especially nocturnal hypoglycaemia. The evidence of this benefit was conflicting in earlier clinical trials [15,16,17,18,19,20,21,22]. However, most of these earlier studies predominantly included patients under the age of 60 years, which caused it to be less powered in detecting the efficacy and safety of long-acting insulin analogues in older age groups who are at increased risk of hypoglycaemia and its severe consequences than younger people. Through our literature search and following the application of the exclusion criteria, five studies investigated the safety of long-acting insulin analogues in older people with diabetes and were included in this manuscript. The recent studies that included older people with type 2 diabetes have shown some benefits of the new long-acting insulin analogues, compared to the older human insulins (Table 1). Fujimoto et al. showed that twice-daily insulin degludec/insulin aspart to improve daily glucose level variability, morning and evening glucose control and quality of life (QOL) in 22 Japanese men, with a mean (SD) age of 68.0 (9.9) years, previously treated with premixed insulin [23]. However, there was no significant difference in the incidence of hypoglycaemia before and after insulin switching. The total and therapy-related QOL feeling scores favoured insulin degludec/insulin aspart; whereas social, physical and daily activities scores were not significantly different. The flexibility of injection timing and glycaemic control may explain the improvement in the total and therapy-related feeling subscores in the QOL questionnaire. However, this study was limited by the small sample size and the short duration of follow-up, which may suggest that the switch in the insulin regimen might not explain all the changes in the endpoints, and other factors, such as lifestyle changes and physicians’ motivations, might have contributed to the results. Another limitation was that the incidence in hypoglycaemia may have not been accurate, because the frequency of this event was calculated based on self-measured blood glucose levels or patients’ symptoms. Lipska et al., in their large retrospective observational study of 22,489 patients with type 2 diabetes, found that the initiation of a basal insulin analogue (glargine or detemir) was not associated with a reduced risk of hypoglycaemia-related emergency department (ED) visits or hospital admissions compared with NPH insulin. Glycaemic control was similar in both groups after one year of follow-up [24]. However, the population included in this study were relatively young, with a mean (SD) age of 60.2 (11.8) years. Previous studies using the national registries in Finland that included participants of similar ages to those presented in Lipska et al.’s study showed a significantly increased risk of hospitalisation related to severe hypoglycaemia with the use of NPH insulin compared with insulin detemir or glargine [25,26]. In addition, although Lipska et al.’s was a large study, only 1928 participants of the total 25,489 used insulin analogues, and despite matching on the propensity score quintiles, some differences between the two groups remained, suggesting that the study did not fully adjust for the confounding factors. Recently, Bradley et al. showed that the initiation of long-acting insulin analogues was associated with a lower risk of ED visits or hospitalisations for hypoglycaemia compared with NPH insulin in older patients (≥65 years) with type 2 diabetes in Medicare beneficiaries [27]. The strength of this study was the large sample size of 575,008 patients with type 2 diabetes, of an older age, with a mean (SD) of 74.9 (6.7) years, and the fact that a large proportion of patients were treated with insulin glargine (407,018 patients) or insulin detemir (141,588 patients). The hazard ratio (HR) for hypoglycaemia was 0.71, 95% confidence interval (CI) 0.63 to 0.80 for glargine vs. NPH insulin, and 0.72, 0.63 to 0.82 for detemir vs. NPH insulin. The older ages of the participants in this study compared to the study conducted by Lipska et al., suggest that age may have contributed to the disparity between the two studies [24]. In the post hoc analysis, Bradley et al. observed that in participants aged 65–68 years; the use of glargine or detemir was not associated with ED visits or hospitalisations for hypoglycaemia compared with NPH insulin [27]. However, in older participants (69–87 years of age), the use of long-acting analogues was associated with a reduced risk of hypoglycaemia compared with NPH insulin. Betônico et al. demonstrated better glycaemic control and fewer nocturnal hypoglycaemia in 34 patients, mean (SD) age 63.0 (7.0) years, using insulin glargine compared with 16 patients, with a mean (SD) age of 60.0 (8.7) years, using NPH insulin [28]. The importance of this study was that it included patients with chronic kidney disease (CKD) stages 3 and 4, which is more common in older people. CKD is associated with a slower insulin degradation, increasing its duration of action that might increase the risk of hypoglycaemia [29]. However, because the insulin analogue has no peak action, it showed less risk of hypoglycaemia in this population. This is clinically relevant as, with the progression of CKD, most hypoglycaemic medications need dose reductions, and the adjustment of these medications, in the face of renal impairment, may not be enough to keep diabetes under control, and therefore insulin is the most effective therapy in this situation [30]. Özçelik et al. showed that the switch from premixed and intensive insulin to twice daily degludec/aspart insulin was associated with a significant reduction in the daily insulin dose requirement and the incidence of hypoglycaemia [31]. The use of premixed and intensive insulin is a complex regimen and may not be an easy option for daily life in older people with diabetes; therefore, the switch to degludec/aspart insulin may be a less complex regimen, as demonstrated in this study and previous studies [32]. Figure 1 illustrates the advantage of the physiological, clinical and therapeutic properties of insulin in the AM frail phenotype.

The potential effect on body weight should be considered when prescribing hypoglycaemic agents in frail older people with type 2 diabetes. For example, the use of weight limiting agents, such GLP-1RA and SGLT-2 inhibitors in the AM phenotype, are inappropriate due to the increased risk of further weight loss, dehydration, hypotension and increased risk of falls. Acarabose is associated with weight loss, significant gastrointestinal side effects and is less tolerated. Insulin secretagogues, such as sulfonylureas or glinides, although they have the advantage of desirable weight gain in the malnourished frail phenotype, are unsafe due to their high risk of hypoglycaemia. This population is also likely to have a high prevalence of dementia, which may be associated with erratic eating patterns, and the use of insulin secretagogues may significantly increase their risk of hypoglycaemia. Metformin may not be a suitable choice for many patients who have renal impairments. Additionally, pioglitazone is associated with the increased risk of lower-limb oedema, volume overload and exacerbation of congestive cardiac failure. Insulin has always been perceived as a last resort hypoglycaemic therapy after oral agents due to the associated side effects, such as the increased risk of hypoglycaemia, undesirable weight gain, inconvenience of frequent injections and the burden of blood glucose monitoring. However, in the AM phenotype of frailty, insulin may be a preferred early stage therapy. This phenotype is characterised by anorexia and significant weight loss. As a result, this phenotype has less insulin resistance and is likely to be more responsive to insulin therapy, in comparison to the SO phenotype that is characterised by increased insulin resistance [6]. Insulin-related weight gain is an advantage in this frailty phenotype. It may also have the potential to improve muscle mass and muscle function independent of glycaemic control. Therefore, in the milder form of the AM phenotype, such as people who are still compliant with oral therapy and nutrition, metformin, dipeptidyl peptidase-4 (DPP-4) inhibitors or glitazones can be used as first-line therapy, mainly due to their lower risk of hypoglycaemia. However, in patients with severe malnutrition and those less compliant with oral medications, insulin could be the first line of therapy. Insulin therapy has been shown to produce a sustained improvement in the well-being of older people [33]. Insulin-associated side effects, such as the inconvenience of frequent injections, blood glucose monitoring and the increased risk of hypoglycaemia, should be considered. The new insulin analogues appear as potentially favourable therapy in the AM frail phenotype due to the low risk of hypoglycaemia and the convenience of a once daily injection. In the SO phenotype, insulin therapy remains a last resort choice due to the significantly increased insulin resistance and undesirable weight gain in this phenotype. Metformin is the preferred first-line agent due to its cardiovascular benefits, weight-neutral effects and a potential positive effect on frailty [34,35]. GLP-1RA and SGLT-2 should be considered as a second-line, or first choice in patients not tolerant to metformin, due to their advantage of inducing significant weight loss and their cardio-renal protective effects [36]. Dipeptidyl peptidase-4 (DPP-4) is well tolerated with a low risk of hypoglycaemia or weight gain. Acarabose can be considered as an add-on therapy, if well tolerated. Although it can cause diarrhoea, it may have some cardiovascular benefits, low risk of hypoglycaemia and it promotes weight loss [37]. Insulin secretagogues and glitazones should be avoided in this frailty phenotype due to their increased risk of further weight gain (Figure 2).