Pharmacokinetic and pharmacodynamic differences of new generation, longer-acting basal insulins: potential implications for clinical practice in type 2 diabetes

ABSTRACT

The treatment of type 2 diabetes (T2D) is often complicated by factors such as patient co-morbidities, complex drug–drug interactions, and management of adverse events. In addition, some of these factors are highly dependent on the nature of the treatment regimen and the molecular and physical properties of the drugs being used to treat patients with this disease. This calls for a better understanding of how the properties of individual drugs affect the overall outcome for patients with diabetes.

Clinical pharmacology studies to assess the pharmacokinetic (PK) and pharmacodynamic (PD) characteristics of new diabetes drugs play an important role in advancing our understanding of the interactions between a drug and the human body. Specific PK and PD techniques such as the glucose clamp test can be applied to assess the properties of drugs used for the treatment of diabetes.

Basal insulin analogs are a common treatment option for the maintenance of glycemic control in patients with T2D. These drugs work by mimicking endogenous insulin secretion within the body and provide stable and prolonged insulin action to achieve optimal glucose levels.

Insulin glargine 300 U/mL (Gla-300) and insulin degludec (IDeg) 100 U/mL and 200 U/mL represent a new generation of longer-acting basal insulins. These drugs demonstrate improved PK and PD properties compared with previous basal insulins, allowing them to more closely mimic physiological basal insulin secretion.

Here we review the methods used to evaluate the PK and PD profiles of Gla-300 and IDeg and describe studies that have investigated the PK/PD properties of these drugs in type 1 diabetes. The aim of this review is to inform primary care physicians of the value and limitations of data from clinical pharmacology studies when prescribing these agents for the management of T2D.

KEYWORDS:

• Pharmacokinetics
• pharmacodynamics
• basal insulin
• diabetes mellitus
• glucose clamp technique

1. Introduction

Basal insulin analogs are the backbone of treatment in type 1 diabetes (T1D) and type 2 diabetes (T2D). They were designed to better mimic the physiological profile of endogenous basal insulin secretion and to provide prolonged, steady insulin action to control plasma glucose (PG) levels in the fasting state, between meals, and overnight. As such, they provide important advances over regular human insulin and neutral protamine Hagedorn (NPH) insulin in terms of duration of action and physiological effect. Although first-generation basal insulin analogs (insulin glargine 100 U/mL [Gla-100] and insulin detemir [IDet]) come close to providing the evenly distributed, predictable, and prolonged action needed to manage fasting hyperglycemia while minimizing the risk of hypoglycemia, their duration of action does not always fully reach or exceed 24 hours. However, hypoglycemia remains a concern with these agents [1,2], particularly nocturnal hypoglycemia. Recently, a new generation of longer-acting basal insulins (insulin glargine 300 U/mL [Gla-300] and insulin degludec [IDeg] 100 U/mL and 200 U/mL), has been developed. These new drugs have pharmacokinetic (PK) and pharmacodynamic (PD) properties that more closely mimic physiological basal insulin secretion than their predecessors, and their duration of action exceeds 24 hours, which provides patients with more predictable coverage and translates to less frequent hypoglycemia and greater dosing flexibility [2].

In human physiology, basal insulin is secreted in small pulses by β cells in the pancreas at a continuous rate to maintain glucose homeostasis [3,4], allowing sufficient glucose for metabolism but not so much that hypoglycemia occurs. To complement basal insulin secretion, β cells release a spike of additional insulin to control glucose after a meal. On a graph, endogenous basal insulin secretion resembles a mostly flat line, interspersed with spikes of prandial insulin secretion [3]. T2D develops when the amount of prandial (earlier deficiency) and basal (later deficiency) insulin secretion is inadequate for the body’s needs, resulting in hyperglycemia [5]. T2D is also characterized by insulin resistance, primarily in the liver but also in peripheral tissues. Over time, β cells become progressively less functional and eventually cannot produce enough insulin to maintain homeostasis. In T1D, β cells are destroyed by an autoimmune process, leaving patients without the ability to secrete any insulin [6].

Clinical pharmacology studies designed to characterize the PK and PD properties of new insulins form the basis of our understanding of how longer-acting basal insulin analogs should be used to target clinical outcomes in patients with diabetes, and they can be utilized to establish dosing intervals for these insulins in clinical practice. These concepts, while related, each provide a specific type of information about a drug and its actions. PK can be summarized as measuring the way in which the human body processes a drug; its values are time course of drug absorption, distribution into tissues, metabolism, and eventual excretion. PD can be summarized as what the drug does to the body in terms of the relationship between drug concentration at the site of action and its effect on duration of action, degree of therapeutic effect, and likelihood of adverse effects (AEs) [7]. However, data from these studies are often not straightforward, and it can be difficult to translate the information from clinical pharmacology studies into clinical practice.

This article reviews the PK and PD characteristics and concepts that are important in the evaluation of basal insulins, and explains how they apply to the new longer-acting basal insulin analogs Gla-300 and IDeg, as demonstrated in recent studies. Our goal is to put the data from these studies into perspective for the primary care physician (PCP) by explaining both their value and their limitations and discussing how such may translate into the use of these agents use in clinical practice.

2. PK and PD methods

One issue with interpreting PK/PD studies in clinical practice is that different study design standards are often used, resulting in different interpretations of PK/PD data. This can make comparing and assessing data complicated and lead to misinterpretation. Table 1 summarizes the key variables that describe the PK/PD profiles of basal insulin analogs.

Table 1. Essential PK/PD definitions for PCPs (adapted from Goldman 2017 [9]).

Parameter	Definition/Clinical Relevance
Pharmacokinetic (PK) studies	PK studies describe how the body influences the drug’s absorption, distribution, metabolism, and excretion.
Pharmacodynamic (PD) studies	PD studies describe the actions the drug exerts on the body, in terms of binding to the target site and its therapeutic and adverse effects.
Euglycemic glucose clamp	The euglycemic glucose clamp is used to measure the body’s sensitivity to insulin injected i.v., i.m. or s.c. When applied to study long-acting insulin analogs, a dose of insulin is injected s.c. and its ability to lower PG over time is assessed by infusing glucose i.v. at varying rates (called the glucose infusion rate [GIR]), as needed to ‘clamp’ PG at a set target (PG 90−100 mg/dL, euglycemic) level. The test is performed over a set time period of fasting, which for long-acting insulins is at least 24 hours. Thus, the GIR needed to maintain euglycemia is a measure of basal insulin activity, and the dynamic changes of GIR reflect parallel changes of insulin activity over time.
Glucose infusion rate (GIR)	The rate of glucose infused over time to maintain PG at euglycemia over time. GIR evaluates the glucose-lowering effect of insulin over time. A flatter curve vs. a peak suggests a more constant glucose-lowering effect and therefore a lower risk of hypoglycemia.
Area under the GIR curve (AUCGIR)	The total amount of glucose infused over a time period indicates total insulin effect/glucose-lowering effect over the time of experiment. Variance in AUCGIR over different days, or within the 24 hours, in a patient can give an indication of between-day and within-day variability in insulin effect.
Area under the insulin curve (AUCINS)	Provides a measurement of the total amount of insulin absorbed that has entered circulation. Allows evaluation of bioavailability.
Elimination or terminal half-life (t1/2)	After the insulin has been maximally absorbed, the time taken for the insulin concentration to drop by 50%. Gives an indication of required dosing intervals (shorter half-life, more rapid elimination, therefore requires more frequent dosing) and of the time taken to reach steady state.

i.m.: intramuscular; i.v.: intravenous; s.c.: subcutaneous; PG: plasma glucose.

The euglycemic clamp method is the established standard for evaluating the PK/PD of basal insulin and is required by regulatory agencies for all new insulin preparations [8−10]. In this test, a dose of insulin is injected subcutaneously at a given time (T0), and the effects are evaluated over a set period of time (at least 24 hours for long-acting insulins). The ability of the insulin to lower PG is assessed by the amount of intravenously infused glucose at varying rates (called the glucose infusion rate, or GIR) over time as needed to keep (‘clamp’) the PG at a set target, typically 90–100 mg/dL (euglycemic level) [8,10]. Euglycemic clamp studies can be performed manually, with GIR adjusted at set time intervals, related to PG measurement, or using an automated system that continuously measures blood glucose and adjusts the GIR in pre-defined intervals [9,11]. The resulting curves, called GIR profiles, are examined in either a ‘raw’ or ‘smoothed’ form. ‘Smoothing’ is required to minimize GIR oscillations of the automated system. Different ‘smoothing’ programs may generate GIR profiles with different shapes and different data calculated from them [9]. The clamp studies are standardized, and in order to isolate the insulin effect based on the amount of glucose infused over time, they require individuals to fast prior to and during the test and to avoid certain foods or exercise in the days before the procedure. The profiles produced by glucose clamp studies give an overall illustration of how an insulin acts (its PD) and how the levels of an insulin change over time (its PK). Additional PK/PD variables can be used to further investigate patterns to guide clinical studies and compare agents (Figure 1). To accurately assess basal insulin analogs, the clamp studies are better performed with the investigated insulin at steady state (i.e. after several days of its use, not after the first injection).

Figure 1. (a) Plasma-free insulin, (b) initial rates of intravenous insulin infusion and (c) plasma glucose after s.c. injection of insulin detemir or insulin glargine 100 U/mL. Adapted with permission from Porcellati 2007 [12] © 2007 the American Diabetes Association. Copyright and all rights reserved. Material from this publication has been used with the permission of the American Diabetes Association.

Gla-100: insulin glargine 100 U/mL; IDet: insulin detemir; s.c.: subcutaneous; SE: standard error.

The most accurate assessment of basal insulin analogs is achieved when these tests are carried out in patients with T1D as these patients lack endogenous insulin secretion, which can add its own PK/PD effects to those of the injected insulin, thus confounding final interpretation [9]. This is why healthy individuals should not be included in PK/PD insulin studies and why PK/PD results differ between patients with T1D and T2D in proportion to residual β-cell function [8,9,11].

PK data represent the most specific characteristics of a given insulin preparation. PK drives PD, which is determined by additional factors that are independent from the injected insulin such as insulin sensitivity. The most important PD variables for basal insulins are those that reflect how closely the treatment mimics endogenous insulin secretion, with a lack of peaks or troughs and an even distribution of activity over the dosing period (Table 1). For insulin dosed once-daily, a duration of action that extends beyond 24 hours is an advantage, providing a low peak-to-trough ratio once steady state has been achieved [13]. A more detailed analysis of PD parameters can be used to determine the proportion of the total 24-hour area under the GIR curve (AUC_GIR) at different 6-hour time intervals, which ideally should be equivalent to 25% of the total 24-hour GIR effect to demonstrate a perfectly even distribution of action over a 24-hour dosing period [9].

Another consideration for basal insulin administered once daily is within-day variability. Within-day variability is the increase/decrease of PK/PD parameters over the 24-hour period between doses. The higher the within-day variability, the higher the risk for hyperglycemia/hypoglycemia. Low day-to-day variability is also an important factor, as it means that identical doses taken by the same patient on different days should result in predictable effects.

Although valuable, glucose clamp studies have a number of limitations that should be taken into consideration. As highly standardized tests, glucose clamp studies do not necessarily reflect ‘real-life’ dosing in patients with diabetes [9]. Furthermore, the tests are also performed in small numbers of selected patients, and so cannot represent the wide diversity seen in clinical practice. In addition, results are reported as mean values, and while this can provide a general picture of the expected clinical effects for the majority of patients, they do not necessarily apply to any individual patient.

While glucose clamp studies are the method of choice to provide valuable data when planning clinical trials, to date there is no reliable method of directly translating GIR results to clinical outcomes for ‘real patients’ [9,11]. Glucose clamp studies show the potential of the insulin, but they cannot predict its absolute clinical effects; therefore the true clinical relevance of the insulin needs to be verified in properly designed (head-to-head) clinical trials [9]. In clinical pharmacology studies, these differences emphasize the need for individualization of treatment for all patients in a clinical care setting.

3. PK and PD studies of longer-acting basal insulins

Recently, two second-generation formulations of basal insulin have become available that aim to improve on the PK/PD profiles of earlier basal insulins and, potentially, improve clinical outcomes. Gla-300 contains the same molecule as the well-established Gla-100, but in a formulation that delivers the same amount of insulin in one-third of the injection volume [14]. By delivering the same dose in a smaller volume, an injection of Gla-300 results in a smaller, more compact subcutaneous depot with a reduced surface area, which slows the re-dissolution rate and produces a more prolonged and constant release of insulin into the bloodstream [15]. IDeg has a different molecular structure that forms multi-hexamer chains and reversibly binds to serum albumin, allowing a slower release of IDeg into the systemic circulation [13,16]. IDeg is available in 100 U/mL and 200 U/mL concentrations; the 200 U/mL formulation allows for reduction of injection volumes with no appreciable change in the PK/PD profile [17].

The PK/PD profiles of these longer-acting basal insulin analogs have been investigated in patients with T1D using Gla-100 as a comparator. Both Gla-300 and IDeg have also been compared with Gla-100 in clinical trials in T1D and T2D. In addition, data from head-to-head PK/PD trials in T1D are now becoming available [18,19], and head-to-head clinical trials are underway in patients with T2D (Table 2). So far, no head-to-head clinical trial comparing Gla-300 and IDeg are available in T1D, although it is clear that the new longer-acting insulins, with a flatter pharmacokinetic/pharmacodynamic profile and a duration of action beyond 24 hours, as compared to Gla-100 and IDet, are well suited for participants with the greatest deficiency in basal insulin secretion, i.e. patients with T1D.

Table 2. A comparison of clinical pharmacology studies and clinical trials conducted using basal insulin analogs.

	Clinical Pharmacology Studies	Safety/Efficacy Studies
Gla-100 vs. Gla-300	PK/PD study comparing Gla-300 with Gla-100 which demonstrated a more evenly distributed PK/PD profile and longer duration of action with Gla-300 compared with Gla-100 [15].	EDITION 1, phase 3a study, which demonstrated that Gla-300 was equally effective compared with Gla-100, with lower nocturnal hypoglycemia risk in patients with T2D on basal insulin and mealtime insulin [27]. EDITION 2, phase 3a study, which showed that Gla-300 was equally effective compared with Gla-100 at reducing A1C, with lower hypoglycemia risk during the night and day in patients with T2D on basal insulin and oral antihyperglycemic drugs [20]. EDITION 3, phase 3a study showing that Gla-300 was equally as effective at lowering A1C compared with Gla-100 with lower hypoglycemia risk in insulin-naive patients with T2D on oral antidiabetic drugs [28].
IDeg 100 U/mL vs. IDeg 200 U/mL	PK/PD study comparing IDeg 100 U/mL with IDeg 200 U/mL, which showed that the change in concentration of IDeg does not affect PK/PD parameters, implying that either concentration can be used in clinical practice [17].	Treat-to-target trial, showing that the 200 and 100 U/mL formulations of IDeg are comparable in achieving glycemic control and demonstrate low rates of nocturnal hypoglycemia in patients with T2D on basal insulin and oral antihyperglycemic drugs [21].
Gla-100 vs. IDeg	PD study comparing Gla-100 and IDeg 100 U/mL, which showed that, at steady state, IDeg demonstrated a lower variability in glucose-lowering effect compared with Glar-100 [22]. PK/PD study comparing Gla-100 and IDeg, which showed that IDeg treatment results in a more stable glucose-lowering effect over 24 hours compared with treatment with Gla-100 [37].	DEVOTE, phase 3 clinical trial in patients with T2D at high risk of cardiovascular events comparing Gla-100 with IDeg, which showed that IDeg did not result in an increase in major cardiovascular events compared with Gla-100 [47]. BEGIN FLEX, phase 3 trial comparing the safety and efficacy of flexible dosing of IDeg vs. IDeg dosed at the same time each day and Gla-100 dosed at the same time each day in patients with T2D. This trial found that flexible dosing of IDeg was just as effective at reducing A1C, with no significant differences in hypoglycemia compared with Gla-100 dosed at the same time each day [40]. SWITCH 1, clinical trial comparing Gla-100 with IDeg, which showed that IDeg treatment in patients with T1D and at least one risk factor for hypoglycemia resulted in a reduced rate of overall symptomatic hypoglycemia episodes compared with Gla-100 [45]. SWITCH 2, clinical trial comparing Gla-100 with IDeg, which showed that IDeg treatment in patients with T2D and at least one risk factor for hypoglycemia resulted in a reduced rate of overall symptomatic hypoglycemia compared with Gla-100 [46].
Gla-300 vs. IDeg	PD study comparing Gla-300 and IDeg 200 U/mL, which showed that IDeg exhibits a lower day-to-day and within-day variability than Gla-300 and has a more stable glucose-lowering effect [18]. PK/PD study comparing Gla-300 and IDeg 100 U/mL, which showed that Gla-300 has lower within-day variability compared with IDeg 100 U/mL [19].	Open-label clinical trials of Gla-300 and IDeg are completed or ongoing in patients with T2D inadequately controlled with GLP-1 RAs (NCT02738151) or with basal insulin (NCT03078478).

A1C: glycated hemoglobin A_1c; Gla-100: insulin glargine 100 U/mL; Gla-300: insulin glargine 300 U/mL;  GLP-1 RA: glucagon-like peptide-1 receptor agonist; IDeg: insulin degulec; PD: pharmacodynamic; PK: pharmacokinetic; T2D: type 2 diabetes.

3.1. Individual studies: Gla-300

Compared with Gla-100, the PK/PD characteristics of Gla-300 demonstrate a more even profile and a longer, more stable duration of action, with glycemic control extending up to 36 hours [15,23,24]. In a randomized, double-blind, crossover, euglycemic clamp study in people with T1D, a daily dose of 0.4 U/kg of Gla-300 at steady state resulted in a more even and longer duration of insulin exposure compared with Gla-100 (PK component) (Figure 2(a)) [15]. In this same study, the GIR profile of Gla-300 was stable, constant, and non-peaking. The activity profile over 24 hours was flat and more evenly distributed vs. Gla-100, remaining effective in excess of 24 hours and to the end of the 36-hour clamp period (PD component) (Figure 2(b)) [15]. In a separate study, Gla-300 was also shown to have low within-day and between-day variability when given over the course of two 24-hour time periods separated by a 7- to 21-day washout period [25]. In a recently published study, a 24-hour euglycemic glucose clamp technique was applied using individualized, and not fixed, doses of Gla-100 and Gla-300 (0.28 ± 0.07 U/kg for Gla-100 and 0.35 ± 0.08 U/kg for Gla-300) under experimental conditions similar to clinical practice, where patients with T1D on a basal-bolus regimen need different doses of the two insulins, after three months of optimal titration of both basal insulins (and prandial insulin) at steady-state conditions. This randomized, crossover, single-blind study showed that, with clinical higher doses of Gla-300 vs. Gla-100 administered in the evening, Gla-300 has a flatter PK/PD profile and longer duration of action, and is expected to decrease the risk of nocturnal hypoglycemia in T1D in the clinical setting since the suppression of hepatic glucose production at night (from 0 to 6 hours) is lower than suppression in the afternoon (from 18 to 24 hours). In addition, Gla-300 showed lower within-day variability of PK/PD parameters vs. Gla-100 [25]. With a fixed dose of 0.4 U/kg, PK data showed an approximately 17% relative reduction in bioavailability and 27% lower PD activity for Gla-300 compared with Gla-100 [15,22], which has potential clinical implications for patients switching to Gla-300 from Gla-100. In fact, the two are not directly interchangeable by dose, and patients using Gla-300 require larger insulin doses as demonstrated in the EDITION series of clinical trials in T2D [26−28] as well as in T1D [29,30]. This is better demonstrated by the study of Porcellati et al. [25], which showed PK/PD bioequivalence of the 25% higher clinical doses of Gla-300 vs. Gla-100, as confirmed by clinical trials in T1D (Figure 2(c,d)) [29,30].

Figure 2. Steady-state profiles of mean insulin (a) and smoothed body-weight–standardized GIR for Gla-300 0.4 units/kg vs. Gla-100 0.4 units/kg (b), and 24-hour profiles of clinical doses of Gla-300 vs. Gla-100 (c) and GIR (d) after administration under steady-state conditions. Reproduced with permission from Becker 2015 [15] © 2015 American Diabetes Association; and from Porcellati 2018 [25] © 2018 American Diabetes Association. Copyright and all rights reserved. Material from these publications has been used with the permission of the American Diabetes Association.

GIR: glucose infusion rate; Gla-100: insulin glargine 100 U/mL; Gla-300: insulin glargine 300 U/mL; INS: insulin; LLOQ: lower limit of quantification; s.c: subcutaneous.

The manner in which these PK/PD differences translate into clinical outcomes was demonstrated in comparisons of Gla-100 and Gla-300 for patients with T2D inadequately controlled while using basal and mealtime insulin (EDITION 1), basal insulin and oral agents (EDITION 2) or oral agents (EDITION 3) [26−28]. The primary outcome measure of these trials was change from baseline to week 24 in glycated hemoglobin A_1c (A1C), and each trial was designed to show that Gla-300 was at least as effective as Gla-100 in improving glycemic control, an outcome that was met in each trial (Table 2).

Consistent with PK/PD data that showed more stable and consistent activity with low variability, EDITION 1, 2 and 3 further demonstrated that Gla-300 was associated with less hypoglycemia, particularly at night, and less weight gain in EDITION 2 and 3 [26−28]. The pattern of glycemic control with less hypoglycemia and less weight gain with Gla-300 was sustained at 12 months [31–33]. Subsequent patient-level meta-analyses of data from the 6-month EDITION 1, 2, and 3 trials confirmed that, compared with Gla-100, Gla-300 was associated with a lower rate of hypoglycemia at any time of day and nocturnal hypoglycemia and a lower rate of nocturnal hypoglycemia in the 12-month extension trials [34,35].

As expected from the earlier PK/PD studies, the insulin dose at 6 months was 12% higher with Gla-300 than with Gla-100 [34]. Insulin doses at 6 months with Gla-300 and Gla-100, respectively, were: 0.97 U/kg/day and 0.88 U/kg/day (103 U/day and 94 U/day) in EDITION 1; 0.92 U/kg/day and 0.84 U/kg/day) (91 U/day and 82 U/day) in EDITION 2; and 0.62 U/kg/day and 0.53 U/kg/day (59.4 U/day and 52.0 U/day) in EDITION 3 [26−28]. Despite the higher dose of Gla-300, hypoglycemia incidence and weight gain did not increase; this is significant, as these AEs are important limiting factors in the glycemic management of T2D [35,36]. Patients switching from Gla-100 to Gla-300 should start with the same number of units to minimize risk of hypoglycemia, but they are likely to need a higher relative insulin dose on titration [35].

3.2. Individual studies: IDeg

The PK/PD properties of IDeg, compared with those of Gla-100, have been investigated in a number of studies in both T1D and T2D as well as in a range of patient populations (Table 2) [16]. Glucose clamp studies in patients with T1D showed that IDeg 100 U/mL has a flatter, more stable profile than Gla-100 and that it provides more even exposure and action over the 24-hour dosing period with less within participant variability [37]. Although most studies were performed using IDeg 100 U/mL, a subsequent crossover study in patients with T1D demonstrated that the PK/PD characteristics of IDeg at both the 100 U/mL and 200 U/mL concentrations were similar, although some differences may be noted [17].

The BEGIN series of phase 3 clinical trials compared IDeg with Gla-100 in patients with T2D inadequately controlled on oral agents (BEGIN ONCE LONG, using IDeg 100, and BEGIN LOW VOLUME, using IDeg 200) or previously using either oral agents or basal insulin (BEGIN FLEX, using IDeg 100). Each of these trials met the primary outcome, which was to show that IDeg was non-inferior to Gla-100 in terms of A1C reduction. In addition to similar A1C reductions, the administration of IDeg vs. Gla-100 was associated with similar overall rates of overall confirmed hypoglycemia and similar or lower rates of nocturnal hypoglycemia [38–40]. Subsequently, a pre-planned meta-analysis of hypoglycemia frequency across all of the BEGIN studies demonstrated significantly lower rates of overall confirmed and nocturnal confirmed hypoglycemia episodes with IDeg vs. Gla-100, particularly during the maintenance phase of the studies [41].

In general, when the same doses of Gla-300 and IDeg are compared, the results tend to favor Gla-300 in terms of hypoglycemia. A crossover study in patients with T2D that compared the two basal insulins, using continuous glucose monitoring, after titration of basal insulin doses to a glucose target of 100–130 mg/dL, showed a lower mean percentage of time in hypoglycemia, as well as severe and nocturnal hypoglycemia, for Gla-300 [42]. In addition, in a second crossover study using flash glucose monitoring, use of Gla-300 resulted in lower nocturnal hypoglycemia [43]. However, a comparative analysis of patients switching from Gla-100 or IDet to longer-acting insulins showed similar incidence of hypoglycemia for Gla-300 and IDeg, but a lower rate of events associated with hospitalizations and/or admissions to the emergency department [44].

3.3. Head-to-head studies: Gla-300 vs. IDeg

The PK and PD profiles at steady state of Gla-300 and IDeg have been compared in two euglycemic glucose clamp studies in patients with T1D (Table 2) [18,19].

In the first study, which was a PD-only study, the mean value for ‘between-day’ fluctuations in GIR in individual patients was estimated from three separate 24-hour clamp studies performed 3 days apart (days 6, 9 and 12) [18]. Importantly, the within-day variability measured in this study was a post-hoc analysis and involved estimation of relative, rather than absolute, fluctuation. The study used a crossover design, so that patients received 0.4 U/kg of either Gla-300 or IDeg 200 U/mL, administered in the evening for 12 days, then switched to the other basal insulin analog after a washout period. In this study, IDeg was associated with less between-day and within-day glycemic variability than Gla-300. In addition, the mean smoothed 24-hour GIR profiles were relatively similar, although IDeg exhibited a transient activity increase at about 12 hours where Gla-300 had a small initial activity increase shortly after dosing (Figure 3) [18]. This study also reported significantly less day-to-day variability in terms of AUC_GIR during one dosing interval (0–24 hours), again using a mean value from three separate tests [18].

Figure 3. Pharmacodynamic profiles: 24-hour GIR profiles and distribution of glucose-lowering effect at steady state for IDeg (a and c) and Gla-300 (b and d). Shaded bands represent the standard error of mean in (a) and (b). (c) and (d) present AUC_GIR for each 6-hour interval as a percentage of AUC_GIR,τ,SS. *P < .0001 compared with the 0- to 6-hour and 18- to 24-hour intervals. Adapted from Heise and Norskov 2017 [18].

AUC: area under the curve; IDeg, insulin degludec; GIR: glucose infusion rate; Gla-300: insulin glargine 300 U/mL.

The second study used morning dosing and compared Gla-300 and IDeg 100 U/mL, using the euglycemic clamp method in patients with T1D (Figure 4) [19]. This was a crossover study, in which two cohorts of 24 patients were randomly assigned to either Gla-300 or IDeg 100 U/mL at a dose of 0.4 or 0.6 U/kg for 8 days. The clamp test was performed during the last treatment day and, following a washout period, patients crossed over to the other basal insulin for 8 days and a second clamp test. The primary endpoint of this study was within-day variability, as calculated by fluctuations in the smoothed GIR curve. Within-day variability over the dosing period was significantly lower with Gla-300 than with IDeg 100 U/mL at a dose of 0.4 U/mL (Figure 4) and, although lower with Gla-300, it was not significantly different at the 0.6 U/mL dose. Overall, Gla-300 provided a more stable and more evenly distributed PK/PD profile compared with IDeg-100. Between hours 2 and 16, insulin exposure with Gla-300 remained flat before dropping off, whereas insulin exposure to IDeg 100 U/mL increased through hour 10 and then began to drop off [19].

Figure 4. Mean GIR profiles of Gla-300 and IDeg 100 U/mL at the 0.4 U/kg/day dose level in steady state (h) (a), distribution of activity (b) and exposure (c) of Gla-300 and IDeg 100 U/kg over 24 hours. Dashed gray lines represent the ideal constant activity of a basal insulin over 24 hours, with the same 25% of activity distributed in each of the four 6-hour time periods. Adapted from Bailey 2018 [19] © 2017 The Authors. Elsevier Masson SAS. All rights reserved.

AUC: area under the curve; CI: confidence interval; GIR: glucose infusion rate; Gla-300: insulin glargine 300 U/mL; IDeg: insulin degludec; INS: serum insulin concentration.

3.4. Interpreting head-to-head PK/PD studies

Some differences in the way these studies were designed, aside from the evening and morning dosing, may be reflected in their findings. In the first study, only PD was investigated and insulin levels were not measured or reported [18]. This study assessed PD on three different days, separated by two-day intervals and, as such, principally assessed between-day differences in PD variability averaged over three measurements and without separate assessment of different insulin doses. The within-day fluctuations reported in this study were based on derived, rather than directly measured parameters [18]. By contrast, in the second study, both PK and PD were measured directly, allowing for reliable assessment of the correlation between these two parameters [19]. In this study, within-day variability of PK and PD was also derived from direct measurements. The fact that one study used IDeg 100 U/mL and the other used IDeg 200 U/mL may be considered, although previous studies suggest that the PK and PD of the two different concentrations of IDeg are similar [17]. However, in the first study, the clamp did not achieve euglycemia in several circumstances, especially with Gla-300, and PG increased well over the target of 100 mg/dL [18]. As a consequence, GIR went to zero, resulting in greater calculated variability of Gla-300 vs. IDeg 200 U/mL, an artifact due to sequential increase and decrease of PG from and to euglycemia (Figure 5). This is likely the result of non-bioequivalence of identical doses of Gla-300 and IDeg, especially with evening dosing, which resulted in lower PD activity with Gla-300. The question of day-to-day and within-day variability requires additional studies where Gla-300 and IDeg are given at bioequivalent doses, resulting in the same PD effects.

Figure 5. Individual blood glucose profiles for patients treated with IDeg (a) and Gla-300 (b). Red line = mean BG in each treatment group; number of patients = 57; BG target in the clamp was 5.5 mmol/L (100 mg/dL). Adapted from Heise and Norskov 2017 [18].

BG: blood glucose; Gla-300: insulin glargine 300 U/mL; IDeg: insulin degludec.

What meaning do these data have for clinical practice? IDeg and Gla-300 are molecularly different and can be expected to have different PK/PD profiles, which would typically be detected in head-to-head glucose clamp studies in T1D. When more than one head-to-head study is conducted on the same insulin product and those studies yield inconsistent data, this naturally creates a dilemma for the clinician when trying to translate the results into clinical practice. The head-to-head studies reviewed here are a textbook example for how glucose clamp studies following different design principles can create a confusing picture for the practicing clinician, rather than contributing to sound clinical decision-making. Inconsistencies in dosing regimens, selection of dosing periods and assessment periods as well as in the choice of analytical approaches applied to the data can be assumed to have resulted in creating a clinically confusing picture. To date, clinical outcome studies have compared each of these with Gla-100 and found that both the longer-acting basal insulin analogs provide benefits in terms of risk reduction of hypoglycemia and weight changes, although the exact benefits may differ somewhat between agents.

For instance, the SWITCH-1 study included patients with T1D and at least one risk factor for hypoglycemia who were treated with IDeg or Gla-100 and then crossed over to the other basal insulin. Patients monitored their blood glucose levels daily before the main meals (morning, midday and evening) and at bedtime and adjusted their basal insulin dose weekly based on the lowest of three pre-breakfast values, with a fasting target of 71–90 mg/dL, during the initial 16 weeks of the titration period. This was followed by a 16-week maintenance period for comparison of hypoglycemia outcomes. The results showed reduced rates of overall symptomatic hypoglycemia and nocturnal symptomatic hypoglycemia in the 16-week maintenance period with IDeg compared with Gla-100 [45]. The SWITCH-2 crossover study, which followed a similar design, evaluated patients with T2D and at least one risk factor for hypoglycemia. Significant reductions in overall and nocturnal symptomatic hypoglycemia were observed for IDeg compared with Gla-100 [46]. These findings were supported by data from the DEVOTE study, which was conducted to investigate the incidence of major cardiovascular events in patients treated with IDeg or Gla-100, with severe hypoglycemia events being a secondary endpoint. As in the SWITCH studies, patients titrated the basal insulin dose weekly, based on the lowest of three pre-breakfast values measured two days before and on the day of the dose adjustment, to reach a target of 71–90 mg/dL. The DEVOTE study showed that patients treated with IDeg experienced less severe hypoglycemia events compared with Gla-100-treated patients [47].

A clinical head-to-head study of Gla-300 and IDeg is ongoing in patients with T2D inadequately controlled with basal insulin (NCT03078478). Data from a second head-to-head study in patients inadequately controlled with oral antihyperglycemic drugs/glucagon-like peptide-1 receptor agonists (GLP-1 RAs) (BRIGHT study; NCT02738151) have been recently published; similar effects on A1C were seen with Gla-300 and IDeg, but Gla-300 use resulted in lower hypoglycemia incidence and event rates at any time over 24 hours, and lower event rate during the night, between 0:00 and 6:00 am, during the initial 12-week titration period [48]. Furthermore, analysis of real-world data has shown that older adults switching to Gla-300 have decreased hypoglycemia, with equivalent glycemic control, compared with patients who switched to other basal insulins [49]. Gla-300 will be further evaluated in a randomized, controlled, real-world trial, investigating the effects of Gla-300 compared with other basal insulin therapy in insulin-naive patients (NCT02451137) [50].

Both SWITCH studies and the DEVOTE study were designed as double-blinded studies, whereas the completed BRIGHT study and the ongoing head-to-head study comparing Gla-300 with IDeg have an open-label design. However, it must be noted that concentrated insulin formulations are associated with smaller injection volumes and increased rates of injection-related side effects, making it difficult to maintain full blinding.

4. Stacking, steady-state and titration, and overbasalization: lessons from PK/PD studies

There has been some concern expressed that daily injections of longer-acting basal insulin analogs that remain in the circulation for more than 24 hours might result in a phenomenon known as ‘insulin stacking.’ This concept is derived from the accumulation of insulin observed when additional doses of prandial insulin are given before the previous dose has been completely absorbed, resulting in overlapping insulin effects and an increased risk of hypoglycemia [51]. However, in contrast to the goals of prandial insulin therapy – to normalize postprandial hyperglycemia following glucose ingestion – the goal of basal insulin therapy is to achieve steady-state levels with minimal peaks and troughs, thus mimicking endogenous basal insulin secretion. Insulin stacking does not occur in clinical use of basal insulins such as Gla-100 and IDet because their duration of action is not greater than 24 hours (Gla-100) or is well under 24 hours (IDet) [10].

Although basal insulin therapy is not associated with insulin stacking, it is important to know how many days are required for the insulin to reach steady state and to avoid increasing the dose before steady state is achieved [51]. The achievement of steady state has implications for dose titration, as dose adjustments with longer-acting basal insulins should be done no more frequently than every 3 days or once or twice weekly to allow the full clinical effect to be assessed. A premature dose increase may lead to too much circulating insulin and increase hypoglycemia risk. Steady-state concentrations are reached in at least 5 days with Gla-300 and 2−3 days with IDeg [51,52].

A further benefit of longer-acting basal insulin analogs such as Gla-300 and IDeg is that they allow patients some flexibility in dosing without compromising glycemic control. Although patients should still take insulin daily, if they take their insulin a bit earlier or later than usual it should not affect their glycemic control. This was demonstrated in a study with Gla-300, where patients with T2D either maintained a strict 24-hour dosing schedule or took Gla-300 up to 3 hours earlier or later than the day before. Flexible dosing had no negative effect on glycemic control, hypoglycemia risk, or AEs [53]. Similarly, a study in Japanese patients with T2D treated with IDeg showed no negative effects in terms of A1C levels in patients who followed a flexible dosing schedule (±8 hours from an agreed dosing time) rather than at the same time each day [54]. In agreement with these findings, the prescribing information for IDeg states that daily injections may be given at any time of day in adults, as long as 8 hours elapse between consecutive doses [55].

5. Conclusions

Studies designed to investigate the PK/PD properties of basal insulins are conducted in patients with T1D and can give valuable insights into potential clinical features. However, due to their designs in T1D, they are not well suited to support robust conclusions about the most appropriate clinical utility, long-term clinical features or clinical risk/benefit scenarios of specific basal insulin formulations in T2D. This is evident from the conflicting conclusions of studies that investigated the same insulins using different study design features and analytical approaches. Since these studies have been carried out in patients who have T1D and lack endogenous insulin, the results cannot therefore be immediately translated into T2D, especially when insulin secretion is maintained to some extent. This further limits the general interpretability of these studies, meaning their conclusions may not be applicable to all patients with diabetes. It should be noted that, in contrast to the T1D population who is frequently investigated in clinical pharmacology studies, more than 95% of patients in clinical practice have T2D. The action of basal insulin therapy in patients with T2D can be altered by endogenous insulin as well as other metabolic characteristics of the disease, e.g. excessive insulin resistance. Furthermore, the standard euglycemic clamp studies used in PK/PD studies typically do not account for the specific insulin needs of individual patients and also require significant periods of fasting, meaning they do not reflect a real-world setting. This is important when determining the individual insulin requirements for patients with diabetes who require basal insulin, and when assessing individual treatment responses, clinical efficacy and safety profiles.

The strength of clinical pharmacology studies and the PK/PD data generated is that they provide PCPs with information of how basal insulin formulations may affect their patients and what clinical outcomes they might observe in a real-world setting. For example, Gla-100 and IDet offer PK/PD advantages over NPH, and Gla-300 and IDeg offer additional benefits over Gla-100 and IDet in terms of duration of action, and more physiological PK profiles that translate into lower hypoglycemia risk and increased dosing flexibility. Therefore, while these clinical pharmacology studies can inform treatment decisions when PCPs and their patients discuss options for basal insulin therapy, they cannot replace clinical outcome studies designed to better reflect the real-world environments of patients and clinicians. In fact, clinical pharmacology studies and the PK/PD profiles obtained from those studies frequently provide information that would allow for better-designed, large-scale clinical studies. This would then help to further and more robustly differentiate relevant clinical features and outcomes of novel basal insulins. It will be interesting to see how PK/PD findings with Gla-300 and IDeg, and the conclusions drawn from them, are reflected in comparative clinical data from the ongoing head-to-head study in patients on basal insulin therapy and if these confirm the preliminary results of the BRIGHT study in insulin-naive patients.

Declaration of interest

Marcus Hompesch is an employee and shareholder at ProSciento. Dhiren K. Patel is a member of the Speakers Bureau for AstraZeneca, Boehringer Ingelheim, Mannkind Corporation, Merck, Novo Nordisk, and Sanofi, and a consultant and member of the Advisory Board for Eli Lilly, Merck, Novo Nordisk, and Sanofi. James R. LaSalle is an advisor and speaker for Novo Nordisk and is an advisor and on the Editorial Board for Sanofi. Geremia B. Bolli receives honoraria for consulting and lecturing from Sanofi and Menarini. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties. A peer reviewer on this manuscript is a consultant and speaker for Sanofi and Novo Nordisk. Peer reviewers on this manuscript have no other relevant financial relationships or otherwise to disclose.

Acknowledgments

The contents of the paper and the opinions expressed within are those of the authors, and it was the decision of the authors to submit the manuscript for publication. The authors would like to thank Dr. Francesca Porcellati for her critical review of the manuscript.

Additional information

Funding

The authors received writing/editorial support in the preparation of this manuscript, which was provided by Ila Karve, PhD, Grace Richmond, PhD, and Rasila Vaghjiani, PhD, of Excerpta Medica, and funded by Sanofi US, Inc.