The Modern Role of Basal Insulin in Advancing Therapy in People With Type 2 Diabetes Free

It might appear redundant to discuss insulin therapy any further, given the many recent articles and reviews marking the 100-year anniversary of its first successful use in humans in Toronto in 1922. Insulin is one of the most important lifesaving innovations in medicine, alongside general anesthesia and antibiotics (1); no one would challenge its obligatory use in type 1 diabetes mellitus (T1DM) (2). However, the role and types of insulin in managing type 2 diabetes mellitus (T2DM) have long been debated. Today this debate has intensified with the emergence of newer noninsulin medications, notably the glucagon-like peptide 1 receptor agonists (GLP-1RA) and their derivatives but also the SGLT2i preparations (3).

Basal insulin (BI) has long been the preferred first-line injectable option for T2DM in Western countries, with the introduction of insulin glargine 25 years ago, to be used when glucose control became inadequate with lifestyle modification and other oral glucose-lowering medications (4). However, since 2018, the role of BI has been reassessed, and BI is considered later in the treatment pathway, following evidence that GLP-1RA and derivatives (3,5), as well as SGLT2i, offer some advantages such as a lower risk of hypoglycemia, limited dose titration, and weight loss. Importantly, there is also direct evidence (though differing between classes) of early protection against elements of atherosclerotic cardiovascular disease (ASCVD), chronic kidney disease (CKD), and heart failure in certain groups at risk (3,6,7).

The choice of therapy needs to include consideration of the pathophysiological status of each individual (8), reflecting the well-recognized heterogeneity of T2DM phenotypes (9). The benefits of BI, extensively demonstrated over the years, should not be forgotten (10). It is therefore important to harmonize the exciting advantages of newer therapeutic innovations with the well-established benefits of BI.

The scope of this review includes three objectives: first, to briefly review the key role of BI secretion in the physiology of glucose regulation and its disruption in T2DM; second, to discuss how BI replacement can restore a more normal glycemic pattern and review the development of BI formulations for this purpose; and third, to explore the future role of BI in advancing an individual’s therapy in the current era of GLP-1RA and SGLT2i and in combination with these newer medications.

Insulin is continuously secreted by pancreatic islet β-cells into the portal circulation, playing a dominant role in maintaining glucose homeostasis (11). The insulin secreted overnight and between meals, referred to as BI, is released at a low rate (∼0.5–1.0 units/h) in response to the prevailing plasma glucose (PG) levels. During mealtimes, several units of insulin are secreted within a few minutes (“prandial insulin”) and result in peak peripheral plasma concentrations of ∼400–600 pmol/L in response to incretins, PG, and neural signals.

BI also modulates glucagon release from neighboring α-cells in the pancreatic islets (12). The ensuing portal insulin-to-glucagon molar ratio, rather than the individual secretion rates of insulin and glucagon, predicts the rate of release of glucose from the liver (endogenous glucose production [EGP]), the primary determinant of basal PG. During its first pass through the liver, ∼50% of BI is extracted, and therefore insulin circulates in peripheral plasma at a much lower concentration (∼30–70 pmol/L), 2.5–3.0 times lower than portal concentrations, with a nearly flat action profile and low variability (coefficient of variation <5%) (11). Because of its low concentration, BI primarily exerts its influence in the more insulin-sensitive organs, notably, adipose tissue and the liver, with minimal impact on muscle glucose utilization. Prandial insulin excursions, however, are responsible for stimulating glucose uptake in muscle (13).

BI also restrains lipolysis, reducing the fatty acid flux to the liver, and exerts a finely tuned restraining action on EGP minute by minute, thus maintaining PG within a strict normal range during fasting and between meals (14) (Fig. 1).

Insulin also possesses extrametabolic effects, including anabolic, anti-inflammatory, antiatherosclerotic, platelet inhibitory, and immunomodulatory properties (15). These effects are both direct and indirect, with the latter resulting from the removal of gluco- and lipotoxicity (16). Peripheral insulin concentrations enhance endothelium-dependent and endothelium-independent vasodilation (17,18), reverse insulin resistance in T2DM by improving both insulin action and secretion (19,20), and reduce hepatic fat content (21).

In T2DM, the islet β-cells fail to secrete sufficient insulin to meet physiological needs. The reduced portal insulin-to-glucagon molar ratio leads to insufficiently restrained EGP, with consequent increase in PG. Because glucose sensitivity of the β-cell is also reduced in T2DM (22), the increase in insulin secretion response to elevated glucose is nevertheless inadequate to fully mitigate the increase in EGP. The initially compensatory mechanism of hyperinsulinemia eventually fails when insulin sensitivity in the liver and other tissues is also reduced (insulin resistance), as is typically the case in T2DM, leading to the progression of hyperglycemia (13,23–25).

The earliest detectable dysfunction in the natural history of T2DM is usually loss of the early phase of prandial insulin secretion (25). However, BI deficiency is also present in the initial stages of T2DM, with increased EGP and fasting PG (FPG) levels most evident in the predawn hours during nocturnal fasting (Fig. 1) and accompanied by accelerated lipolysis (23). As insulin deficiency progresses with time, both basal and postprandial hyperglycemia worsen (10,25).

Replacement of the deficient hormone, BI, in people with T2DM addresses the primary mechanism of dysglycemia, namely, excessive EGP, primarily at dawn, and thus contributes to the lowering of FPG (Fig. 1). Replacement of BI with intravenous insulin in the fasting state to reduce PG to nondiabetes levels increases plasma insulin concentration by ∼25%, with glucose concentration changes attributed primarily to its antilipolytic (26) and hepato-specific properties. There is no evidence of stimulation of muscle glucose uptake (27) (Fig. 2). This partly explains the lower propensity of BI replacement to induce hypoglycemia in contrast to prandial insulin, which promotes muscle glucose uptake in addition to suppressing EGP (13).

Similarly, supplementing BI deficiency with subcutaneously injected long-acting insulin preparations increases the portal insulin-to-glucagon molar ratio, suppresses lipolysis, and reduces EGP (13,26), resulting in a decrease in FPG. In addition, BI suppresses proteolysis favoring protein synthesis and an anabolic state (28). The better glycemic control and suppression of lipolysis induced by BI reduce glucose and lipid toxicity (16), lower the stress on the β-cell, and lower liver and muscle insulin resistance, thus promoting remission of T2DM in people with new- or recent-onset disease (29).

As insulin deficiency in T2DM progresses, treatment with BI alone may no longer be sufficient to achieve target HbA_1c, despite continuing optimal titration (30). Therefore, in such cases, specific targeting of postprandial hyperglycemia is indicated.

During the decade following the introduction of insulin in 1922, early therapy relied on unmodified (“regular,” “soluble”) insulin extracted from animal pancreas and prepared in an impure solution, with a relatively short duration of action. Consequently, many attempts were made to prolong exogenous insulin action by delaying its absorption from subcutaneous tissue (reviewed in Owens et al. [31] and Bolli et al. [32]). These efforts culminated in the mid-1930s with the development of protamine zinc insulin, followed by NPH insulin in the 1940s and then the zinc-complexed family of short-, medium-, and long-acting lente insulins in the 1950s (31–33).

However, it was only in the second half of the 20th century that the concept of using BI in the treatment of T2DM was introduced (24) and adopted into clinical practice (10,34–38). These early BI preparations, either animal or human insulin, failed to achieve the flat action profile needed to mimic normal BI levels over a 24-h period (39) and, further, were highly inconsistent from day to day. When admixed, lente and ultralente insulin reduced the effect of regular insulin (30,31). NPH injection at bedtime was widely used from the 1980s for BI replacement, with a peak of action 4–6 h after injection followed by a gradual decline over 12–24 h (40). Increasing the dose to prolong its time of action only resulted in a higher risk of hypoglycemia. Consequently, NPH or lente insulin was administered twice daily (or more often), in an attempt to smooth the peak action and prolong duration of action (41). Furthermore, as NPH and lente insulin are insoluble preparations, they required appropriate resuspension of the crystalline complexes (which had rapidly settled on storage) to avoid high day-to-day variability. This required the gentle tipping up and down of the insulin vial or pen 15–20 times prior to dosing (42,43). These challenges made it difficult to consistently attain near-normal fasting glucose levels, and aggressive titration often led to nighttime hypoglycemia, of concern to both users and health care professionals. This dilemma led to research into new classes of longer-acting insulin formulations with an action profile closer to physiology (39).

The first generation of long-acting insulin analogs became commercially available in the early 21st century, with the introduction of insulin glargine 100 units/mL (Gla-100) and insulin detemir. Insulin glargine was the first long-acting insulin with both a reasonable pharmacokinetic profile (half-time of absorption 12 h) and more consistent bioavailability. The more physiological pharmacokinetic and pharmacodynamic characteristics of these long-acting insulin analogs (39) translated into clinical benefits, particularly a lower risk for nocturnal hypoglycemia in T1DM (44), and in T2DM, as first demonstrated in the landmark Treat-to-Target Trial (45). Figure 2 illustrates the more physiological suppression of EGP overnight with Gla-100 compared with NPH insulin when given in the late evening, explaining the lower risk for nocturnal hypoglycemia (44,45).

Because the long-acting analogs are presented as solutions, in contrast to the suspensions of NPH and lente, the day-to-day variability of FPG will also be reduced in some users (45). Gla-100, administered once daily, has been shown to be more advantageous in T2DM than prandial insulin given multiple times a day, with a lower risk for hypoglycemia and less weight gain for similar glycemic control (46). This observation supports the recommendation to begin insulin therapy in T2DM with BI rather than prandial (or premix) preparations (4,47).

The “second-generation” long-acting insulin analogs, insulin degludec (IDeg) and insulin glargine 300 units/mL (Gla-300) were approved in 2013 and 2015. These analogs provide a longer and flatter action profile in comparison with their predecessors, explaining their lower risk for hypoglycemia in clinical trials (31,32). IDeg and Gla-300 insulin differ significantly in their primary structure and mechanism for retarding their action after injection (31,32), with a different glucose-lowering effect on a unit-to-unit basis after subcutaneous dosing (48,49). The reduced glucose-lowering effect per unit of Gla-300, coupled with its lower within-day variability (48), may explain the lower risk for hypoglycemia during the phase of initial titration of this insulin analog (50–52).

The once-weekly insulin icodec formulation was approved for use in T1DM and T2DM by the European Medicines Agency in May 2024. However, its approval by the U.S. Food and Drug Administration is pending resolution of issues regarding the manufacturing process and its use in T1DM (53). Icodec is an acylated insulin analog that has a reduced affinity for the insulin receptor and reduced insulin receptor–mediated clearance and enzymatic degradation (54). This reduces the frequency of injection to once weekly, which can improve adherence and convenience. A second weekly BI formulation, basal insulin Fc (BIF) (or insulin efsitora alfa), with a more prolonged time-action profile, is currently in phase 3 trials. Both insulin icodec and BIF have demonstrated noninferiority to BI Gla-100 or IDeg administered daily in controlling HbA_1c in T2DM (54,55). While the risk of hypoglycemia with these newer BIs is generally low in absolute terms, the risk is higher in comparison with a once-daily BI (54,55). The more elevated risk for hypoglycemia, also observed among individuals with T1DM (56,57), suggests that caution is needed, along with additional studies in the subgroup with insulinopenic T2DM more prone to hypoglycemia (58). These long-acting insulins can be used alone or coadministered with other glucose-lowering agents including GLP-1RA and SGLT2i as free- or fixed-dose (59) combinations to improve glycemic control and reduce side effects of BI such as hypoglycemia and weight gain.

Table 1 lists the clinically desirable features of BI therapy, as well as some limitations and ways to mitigate them. BIs have been available for decades and are arguably the best studied medications for management of hyperglycemia. Although replacement of BI does not directly improve β-cell function (60), as do thiazolidinediones (61), it does benefit prandial insulin responses by reducing glucotoxicity and lipotoxicity (16). Replacement of BI in people without T2DM, but at risk of developing T2DM, reduces the risk of progression to T2DM (62).

Use of BI is effective in essentially all people with T2DM, and its therapeutic power equals or exceeds that of other therapies (63). BI can be used in combination with all glucose-lowering therapies. Dosing can be continually titrated to match the requirements of anyone with T2DM and can be adjusted for changing needs over time. Beside reliably improving glycemic control, BI has favorable effects on lipid metabolism, vascular function, and insulin secretion and action. In people who cannot maintain the target HbA_1c on other noninsulin therapies, BI may be the only option to improve glycemic control. When this occurs in those with sufficient life expectancy, the evidence of vascular protection from the UK Prospective Diabetes Study (UKPDS) and Diabetes Control and Complications Trial (DCCT) extension studies suggests that BI-maintained glucose control will reduce ASCVD/CKD risk (64–66) .

The leading barriers to introduction and escalation of dosage of BI are the need for injection, the burden for continued dose adjustment based on self-monitoring over many months, the risk of hypoglycemia, and the tendency to weight gain. Injections of insulin are less well accepted than injections of incretins because of the need for continuing dose titration and supervision, and perhaps the stigma that the need for insulin marks a terminal stage in the disease. Over recent decades each of these difficulties has been ameliorated, if not removed, by improvement in insulin delivery and management thereof, including pen injectors, easier protocols and systems of self-monitoring, and better patient education protocols and advice. Self-glucose testing devices are now more user-friendly, and protocols for starting insulin rely on once-daily testing only. For when a patient is established on insulin, and particularly after progression to multiple injection regimens, continuous glucose monitoring is now widely available and can significantly reduce the risk of hypoglycemia where that is a problem.

Modern insulin formulations also provide very flat and reproducible profiles of action that further limit the risk of hypoglycemia. Despite this progress, hypoglycemia continues to be the most important side effect of treatment of diabetes with any type of insulin preparation, affecting perhaps 20% of BI users in any year. This can be secondary to erratic insulin administration, erratic subcutaneous absorption, and/or erratic lifestyle. It can have negative consequences on daily life, and on the risk of unawareness of hypoglycemia (67), already raised in older people. However, with proper dose titration severe hypoglycemia is very uncommon (50,62,68) and the risk of nonsevere clinically significant hypoglycemia can be greatly mitigated. The risk of hypoglycemia with BI is considerably lower than with prandial insulin (46) or mealtime plus basal or premixed insulin (47) because BI will not usually oversuppress EGP or increase muscle glucose utilization (Fig. 2). In people with longer diabetes duration and low values of C-peptide who are more prone to hypoglycemia (58), BI titration should be based on more conservative algorithms that prioritize prevention of hypoglycemia (69,70)—and not on the more aggressive algorithms used in some phase 3 clinical trials (71). Whenever possible, BI should be combined with other medications (GLP-1RA and/or SGLT2i) to lower the BI dose and risk of hypoglycemia. Where necessary to gain control of prandial glucose excursions, mealtime insulins will be necessary if these alternatives are not sufficiently effective.

As with the risk of hypoglycemia, weight gain with BI occurs but it is uncommonly a significant problem. Weight gain may be more important when BI is begun late in the scenario of long-term severe hyperglycemia and marked glycosuria with a negative energy balance (catabolic status). Under these conditions BI reduces glycosuria and calorie loss as result of lower PG, reduces lipolysis, and favors restoration of protein metabolism and energy stores (anabolic status with positive energy balance), resulting in weight gain proportional to the degree of initial hyperglycemia (72). This increase of body weight may be seen as healthy recovery of the lost weight, associated with better insulin sensitivity (21), though it may not be perceived as such for those already overweight. However, if BI is started with HbA_1c no higher than 7.0% (53 mmol/mol), when glycosuria, calorie loss, and catabolic status are negligible, weight gain will be modest or even absent (37,62,73). These observations add to the arguments favoring a timely initiation of BI in the natural history of T2DM. Finally, combining BI with oral or injectable agents that limit weight gain, such as metformin, GLP-1RA, or SGLT2i, reduces the weight effects of marked improvements of glucose control.

In part due to the association between serum concentration of endogenous insulin secretion and overweight, there has been long-standing concern about a possible increase in risk of cardiovascular (CV) risk and cancer following introduction of exogenous insulin in T2DM (74). Recently, large epidemiologic analyses have provided reassurance on this point, generally showing that cardiac and neoplastic risk is linked more strongly to the underlying adiposity of individuals per se rather than neo-generated peripheral adipose tissue following exogenous insulin (65,75).

More recent data from Outcome Reduction With Initial Glargine Intervention (ORIGIN Trial) (62) and Glycemia Reduction Approaches in Diabetes: A Comparative Effectiveness Study (GRADE) (73) provide important reassurance on hypoglycemia and CV outcomes in using BI in T2DM. In the ORIGIN Trial investigators studied for >5 years ∼12,000 individuals with high CV risk and dysglycemia or short-term T2DM, comparing BI with conventional oral agent stepped therapy (62). Good glycemic control (mean HbA_1c ∼6.5% [∼47 mmol/mol]) was maintained in both arms. As expected, more hypoglycemia was noted in the insulin group, although uncommon in absolute terms (68). There was no excess of CV-related outcomes with insulin, while a substudy of carotid intima-media thickness suggests a potential for CV protection (76). Among those with prediabetes, fewer developed diabetes when taking the insulin. The mean body weight difference at 5 years was increase of 1.5 kg. Notably, concerns that insulin, and particularly insulin glargine (77), might increase cancer risk (78) have been clearly refuted (79–83), which was confirmed by the ORIGIN Trial results, which showed no relationship (62).

In GRADE Gla-100 was compared with liraglutide, glimepiride, and sitagliptin, all as second-line therapy after metformin in T2DM (starting mean baseline HbA_1c 7.5% [58 mmol/mol]) (73). There was approximately equivalent maintenance of glycemic control with glargine and liraglutide, these being a little better than with glimepiride or sitagliptin. After 4 years weight declined by 3.5 kg with liraglutide and 0.6 kg with glargine. Severe hypoglycemia occurred in 1.0% of participants assigned to liraglutide vs. 1.3% of those allocated to Gla-100, while discontinuation rate was 23% with liraglutide and 14% with glargine. Overall, Gla-100 was as effective as its GLP-1RA comparator, and with no weight gain on average. The discontinuation rates suggest somewhat better tolerance, preference, and/or continued efficacy of the insulin therapy versus the GLP-1RA.

While these randomized studies cannot be fully reflective of experience in routine medical practice, the findings do suggest that use of modern BI early in the course of T2DM is as effective and safe and tolerated as well as other therapy options. Reluctance to start and continue BI treatment may continue to be a barrier, at least for individuals who are unaware of current experience with BI, but the evidence should reduce the lingering stigma of this therapy.

There are clinical settings where insulin (and BI) is the dominant preferred treatment approach, notably in and after metabolic emergencies (3,84) (Table 2). BI is also a therapy of choice in conditions of more marked insulin deficiency (58) and, notably, in pancreatic diabetes, and LADA (85,86). BI has a long track record in pregnancy, either alone or as part of a mealtime plus basal regimen. Insulin (BI and/or short acting) is useful to mitigate acute hyperglycemia on sick days, during cycles of steroid treatment, and with metabolic stress (including myocardial infarction and major surgical interventions), trauma, and/or hospitalization. BI has the best-established safety record where comorbidities exist, such as organ failure or cancer, the latter notably where chemotherapy interacts with insulin sensitivity. But the biggest need remains for BI therapy when all other measures cannot give acceptable glucose control, a situation that generally occurs in most people with T2DM in time, given their improving life expectancy.

There are also settings where BI is the preferred first injectable agent, including symptomatic hyperglycemia at diagnosis of diabetes, when initial remission is targeted by intensive treatment (29,87–89), or when a diagnosis of T1DM is possible based on the presentation of the patient. Although any treatment of T2DM may induce remission in the right scenario, BI allows improvement in β-cell function and insulin secretion and action by ameliorating gluco- and lipotoxicity with appropriate dose titration more quickly and more effectively than other therapies (90). Intermittent add-on of BI may help in maintaining glycemic control and is safe (91). BI treatment at diabetes diagnosis, often followed by remission of the condition, may be helpful also for taking the opportunity to experience insulin treatment. In the case of subsequent recurrence of diabetes and need for insulin, the experience will often improve acceptance of BI going forward, due to previous positive outcomes.

In the nonnegligible minority of people without obesity or even lean people with markedly insulin-deficient T2DM, primarily but not exclusively in Asia and Japan, BI can be the preferred choice, in particular as their limited response to incretins may fail to control HbA_1c (92).

As well as being useful in people in whom target glucose cannot be maintained with other therapies, add-on or replacement with BI can be indicated to reduce dosing or to replace other agents, thus mitigating or eliminating relevant side effects and possible risks. These might include diarrhea and vitamin B₁₂ deficiency with metformin, hypoglycemia with sulfonylureas, fluid retention and/or heart failure and possible macular edema with thiazolidinediones, and pancreatitis with some incretins. Other side effects include gastrointestinal symptoms with GLP-1RA (93) and genitourinary tract infections and rarely ketoacidosis with the SGLT2i (94). Finally, animal studies suggest the possibility of thyroid medullary carcinoma and reproductive toxicity with the GLP-1RA. Although some authorities suggest that incretins are not indicated in the presence of retinopathy (95), clarification regarding this concern may come from the FOCUS trial in 2026/2027.

Finally, BI should be considered as add-on in those people who lose weight and lean body mass progressively to an undesirably low BMI, either as part of the natural history of the disease (catabolic state of insulin deficiency) or as a result of treatment with GLP-1RA. Add-on of a low dose of BI to recover weight and anabolic status should be beneficial, notably in the increasingly large population of older people.

As with other medical conditions with multiple pathogenetic contributions, management of T2DM with more than one intervention can lead to greater efficacy and, in some circumstances, improved tolerance and safety. BI can be used in combination with any other glucose-lowering medication, although caution is needed with sulfonylureas (risk for hypoglycemia) (96) and thiazolidinediones (risk of fluid retention and heart failure) (4). In people with T2DM and more marked insulin deficiency (fasting C-peptide <0.40 nmol/L) (58), BI may mitigate the risk of euglycemic ketoacidosis when used in combination with SGLT2i. With the advantages of the newer medication classes of GLP-1RA and SGLT2i, these are replacing some other orals agents as second- and even first-line treatment (3), meaning logically that BI will often be combined with these if glucose control is still unacceptable. Additional opportunities are for starting with combination therapies, included fixed-ratio preparations, or for adding these agents to BI where that has already been started for other reasons.

In physiology BI regulates fasting and interprandial PG concentrations by modulating basal EGP, whereas prandial insulin, regulated dominantly by incretins, suppresses EGP completely and stimulates muscle glucose uptake during and after meals. Concordantly, in people with T2DM the combination of BI with GLP-1RA improves the 24-h PG via these same integrated mechanisms (plus decreased substrate load) more than the administration of one of them alone, as shown in Fig. 3. Improved tolerability and ASCVD protection are a bonus. BI should be added to a preexisting treatment with GLP-1RA whenever FPG and HbA_1c do not decrease to target, or when the target is no longer maintained over time after initial success. Vice versa, a GLP-1RA should be added to BI treatment when postprandial PG and HbA_1c increase above the target despite well-controlled FPG with BI. BI plus incretins may be so efficacious in controlling postprandial hyperglycemia that it is possible to withdraw or reduce any dose of prandial insulin boluses in approximately half of people with T2DM on mealtime plus BI therapy, and with better secondary outcomes (97,98) (Fig. 3).

The combination of BI plus incretins can be implemented as separate injections (daily BI and weekly GLP-1RA, or both weekly) or as fixed-ratio coformulations (59,99). Separate injection allows more flexibility in dose titration of BI, while coformulations provide more convenience and acceptability.

BI may be used together with SGLT2i (100,101). When added to BI, SGLT2i reduce PG over the 24 h, promoting glycosuria with a glucose-dependent mechanism, with no increase of insulin dose or even a modest decrease. Consequently, there is a lower risk of hypoglycemia, a loss of body weight (due to the glucosuria-mediated loss of calories), and reduced risk of cardiorenal problems including heart failure (3). However, caution is advised because the risk of ketoacidosis is higher for those with more marked insulin deficiency, as marked by the need for insulin supplementary therapy and thinner body phenotype. BI can be added to an SGLT2i to reduce adverse events and improve glucose control (100,101) (Fig. 3).

No studies are available on the possible combination of the three classes of BI, GLP-1RA, and SGLT2i. However, given their different and complementary mechanisms of action, logically the triple association is rational. It might be useful to add on an SGLT2i to the combination BI plus incretin (or incretin to BI plus SGLT2i) to potentiate the loss of weight while decreasing HbA_1c and to ensure a more robust protection from ASCVD/CKD risk. The order in which any such combination is achieved will depend on an individual’s predominant phenotype and clinical profile, such as obesity, risk of heart failure/CKD, or insulin deficiency.

Traditionally, rapid-acting insulin at mealtimes is added to BI when BI alone can no longer keep HbA_1c at target despite FPG at target (inappropriately referred to as “BI failure”). However, beginning rapid-acting insulin is not now recommended in this circumstance unless the options of incretins and/or SGLT2i have already been trialed (3).

The efficacy and safety of BI will be best effected when the dose is carefully titrated to optimize FPG to the target with no or minimal hypoglycemia (45,50). Several titration options are available (45,50,69,70). However, a systematic and personalized scheme is needed for each individual to best meet his or her individual characteristics. In insulin-naive people the initial dosing is best kept low (0.1 units/kg), except in the case of FPG >200 mg/dL and/or obesity with insulin resistance (up to 0.2 units/kg). When FPG is <180 mg/dL and/or for people at higher risk for hypoglycemia (58) conservative algorithms should be preferred over those that are more aggressive (71) to minimize events of hypoglycemia, with no more than 2 units of insulin dose change per week (69,70). In general, weekly titration is sufficient, but in selected settings twice-weekly decisions may be appropriate. Table 3 shows a simple and practical algorithm for starting and titrating BI, one that can be managed by insulin users themselves with limited supervision by medical personnel (69,70). The target FPG should preferably be in a safe range (100–120 mg/dL [5.5–6.6 mmol/L])—not lower, as the high variability of FPG will then give rise to a significant rate of hypoglycemia (50,69). This range is in the center of the preprandial desirable level suggested in the current American Diabetes Association Standards of Care in Diabetes for PG in adults (80–130 mg/dL) (102) but will mitigate the higher hypoglycemic risk with insulin, and reduce the hyperglycemic risk from prandial glucose excursions later in the day. The final BI dose may vary considerably from 0.2 to 0.3 units/kg to hundreds of units per day, requiring bigger steps and longer titration (81). Visits in person may then be necessary only every 3–4 months, but other communication with the diabetes center is desirable every 1–2 weeks to maintain titration momentum. However, titration of BI is an unending process, since insulin sensitivity may vary in the short term, while in the longer term endogenous insulin deficiency will progress.

In physiology the continuous secretion of insulin in the fasting and interprandial state (BI) plays a key role in the admirable minute-to-minute regulation of glucose, lipid, and protein metabolism homeostasis. In T2DM the deficiency of BI is present from an early phase and often progresses over the years, leading to unrestrained EGP and hyperglycemia, often aggravated by hepatic and extrahepatic insulin resistance. Replacement of BI with appropriate titration is pivotal to fix this pathophysiological defect of T2DM.

In insulin-naive people, replacement of BI only is always effective in improving the metabolic abnormalities and targeting the FPG and is effective in lowering HbA_1c to <7.0% (<53 mmol/mol) in ∼50%. In people on noninsulin therapies, add-on of BI is always beneficial and often necessary to reach glycemic targets and mitigate the side effects of the other therapies. BI is never completely contraindicated and has no significant side effects, with the notable exception of hypoglycemia, which may occur in particular with improper use or overaggressive titration.

Today the availability of incretins and SGLT2i medications, with innovative clinical profiles and unprecedented protection of people with cardiorenal risk, has reduced the need for and the appeal of BI as first-line add-on treatment after metformin and/or other medications (3). However, the different and complementary mechanisms of action encourage combination of the new therapies with BI in a flexible manner tailored to the needs of the individual, given the large variety of clinical circumstances in the heterogeneous population of T2DM. BI is needed especially with advanced insulin deficiency (58), and when comorbidities and medical management complexity are present. Depending on the clinical situation, BI may still be the first-line therapy to be followed by the option of add-on of a GLP-1RA/an SGLT2i or more often be the second-line should these drugs fail regarding reaching the HbA_1c target for whatever reason including side effects (92). In the large subgroup of people with T2DM who need BI to keep HbA_1c to target, BI exerts a unique long-term micro- and macrovascular protection so far not generally recognized (“hidden ASCVD protection of BI”) (64,65).

In diabetes management, we need to move from the Manichean view of the past choosing one intervention as opposed to another and, rather, choose a strategy tailored to each individual of a convenient sequence of medications and/or multiple combinations for the different stages of T2DM. BI is not the magic bullet of treatment of all people with T2DM, but timely supplementation of BI is often preferable at onset of disease when BI secretion is already deficient. It usually becomes necessary over the years, and often is more helpful in combination with a GLP-1RA and/or an SGLT2i.

Acknowledgments. M.C.R. is an editor of Diabetes Care but was not involved in any of the decisions regarding review of the manuscript or its acceptance.

Funding. This work was funded by the University of Perugia, without any additional funding from the public, commercial, or not-for-profit sectors. M.C.R. is partly supported by the Rose Hastings and Russell Standley Memorial Trusts for Diabetes Research.

Duality of Interest. G.B.B. has received honoraria for lecturing from Sanofi. P.D.H. receives research support from Sanofi and has consulted for Sanofi, Novo Nordisk, and Eli Lilly and manufacturers of biosimilar insulins. F.P. receives clinical trial funding and advisory board and lecture fees from Abbott, AstraZeneca, Lilly, Novo Nordisk, and Sanofi. H.C.G. holds the McMaster-Sanofi Population Health Institute Chair in Diabetes Research and Care. He reports research grants from Eli Lilly, Novo Nordisk, and Hanmi Pharmaceutical; continuing education grants from Eli Lilly, Abbott, Sanofi, Novo Nordisk, and Boehringer Ingelheim; honoraria for speaking from AstraZeneca, Eli Lilly, Novo Nordisk, Zuellig Pharma, and Jiangsu Hanson; and consulting fees from Abbott, Bayer, Eli Lilly, Novo Nordisk, Pfizer, Sanofi, and Hanmi Pharmaceutical. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. G.B.B. wrote the first draft of the manuscript. All other authors equally contributed in rewriting and editing the text. All authors approved the final text.

Handling Editors. The journal editors responsible for overseeing the review of the manuscript were Steven E. Kahn and Vanita R. Aroda.