INTRODUCTION — Type 2 diabetes is by far the most common type of diabetes in adults and is characterized by hyperglycemia and variable degrees of insulin deficiency and resistance. Its prevalence rises markedly with increasing weight.

Management of patients with type 2 diabetes mellitus includes education, evaluation for microvascular and macrovascular complications, achievement of target glycemia, treatment of cardiovascular risk factors, and avoidance of drugs that can aggravate abnormalities of glucose or lipid metabolism. Weight reduction, diet, and oral medication (typically metformin) can all be used initially to improve glycemic management, although the majority of patients with type 2 diabetes will require additional therapy over time to maintain glycemic targets. The therapeutic options for such patients include adding a second or third oral agent or an injectable agent, such as a glucagon-like peptide 1 (GLP-1) receptor agonist or insulin, or switching to insulin.

The role of insulin in achieving optimal glycemic control in nonpregnant patients with type 2 diabetes will be reviewed here. Options for initial therapy, options for the management of persistent hyperglycemia, and other therapeutic issues in diabetes management, such as the frequency of monitoring and evaluation for microvascular and macrovascular complications, are discussed separately. Insulin treatment of pregnant women with diabetes is also reviewed separately.

●(See "Initial management of hyperglycemia in adults with type 2 diabetes mellitus".)

●(See "Management of persistent hyperglycemia in type 2 diabetes mellitus".)

●(See "Overview of general medical care in nonpregnant adults with diabetes mellitus".)

●(See "Pregestational (preexisting) diabetes mellitus: Antenatal glycemic control", section on 'Insulin pharmacotherapy'.)

●(See "Gestational diabetes mellitus: Glycemic control and maternal prognosis", section on 'Insulin'.)

●(Related Pathway(s): Diabetes: Initial therapy for non-pregnant adults with type 2 DM.)

●(Related Pathway(s): Diabetes: Therapy for non-pregnant adults with type 2 DM and an inadequate response to metformin monotherapy.)

●(Related Pathway(s): Diabetes: Therapy for non-pregnant adults with type 2 DM and an inadequate response to sulfonylurea monotherapy.)

●(Related Pathway(s): Diabetes: Initiation and titration of insulin therapy in non-pregnant adults with type 2 DM.)

GENERAL PRINCIPLES

Insulin preparations — In type 2 diabetes, insulin is generally provided in three ways:

●As a basal supplement with an intermediate-acting to long-acting preparation (neutral protamine Hagedorn [NPH], glargine, detemir, or the very long-acting degludec) to suppress hepatic glucose production and maintain glucose levels at target in the fasting state

●As a pre-meal (prandial) bolus dose of short-acting (regular) or rapid-acting (lispro, aspart, glulisine) insulin to cover the extra requirements after food is absorbed

●As a pre-mixed combination of intermediate-acting and short-acting or rapid-acting insulin

The approximate time of onset, half-life, effective peak, and duration of action of the most commonly used insulins are reviewed in the table (table 1) and in more detail elsewhere. There is substantial intra-individual and inter-individual variability in onset and duration of insulin action related to the volume injected, the injection site, injection technique, and multiple other factors; therefore, insulin must be adjusted based on patient response. (See "General principles of insulin therapy in diabetes mellitus", section on 'Insulin preparations'.)

Physiologic insulin replacement — Insulin is secreted in a pulsatile manner; pulses occur under basal (unstimulated) conditions and in response to meals [1]. Basal insulin secretion represents approximately 50 percent of 24-hour insulin production, with the remainder accounted for by prandial (mealtime) excursions.

The term "intensive insulin therapy" has been used to describe complex regimens that include basal insulin (given as one to two daily injections of intermediate-acting or long-acting insulin) and superimposed injections of short-acting or rapid-acting insulins three or more times daily before meals to limit postprandial glycemic excursions. While intensive regimens are the preferred regimen for patients with type 1 diabetes using multiple daily insulin injections (rather than an insulin pump), they are used for patients with type 2 diabetes as well, especially for patients taking high doses of insulin and for those with insulin deficiency.

Disadvantages of insulin therapy — The major drawbacks associated with insulin therapy in type 2 diabetes are weight gain and hypoglycemia. These adverse effects should be reviewed at every visit in those taking insulin, and weight loss or weight maintenance should be emphasized. Glycemic targets may need to be relaxed to reduce the risk of hypoglycemia and/or insulin-associated weight gain. (See 'Troubleshooting' below.)

The safety of human insulin versus insulin analogs, including concerns about diabetic complications and risk of cancer, is reviewed in detail elsewhere. (See "General principles of insulin therapy in diabetes mellitus", section on 'Safety'.)

Monitoring glycemia

●Glycated hemoglobin (A1C) – A1C is the most widely used clinical test to monitor chronic glycemic management. Target A1C levels in patients with type 2 diabetes taking insulin should be tailored to the individual, balancing the reduction in microvascular complications (figure 1) with the risk of hypoglycemia and insulin-associated weight gain. A reasonable goal of therapy for most patients might be an A1C value ≤7.0 percent (using an assay aligned to the Diabetes Control and Complications Trial [DCCT] in which the upper limit of normal is 6.0 percent) (calculator 1). The A1C goal should be set somewhat higher for older patients, patients with comorbidities, and those with a limited life expectancy. (See "Glycemic control and vascular complications in type 2 diabetes mellitus", section on 'Choosing a glycemic target'.)

●Fasting blood glucose (FBG) – Measurements of instantaneous glucose levels (self-monitoring of blood glucose [with fingersticks and a glucose meter] and real-time continuous glucose monitoring [CGM]) are used to manage diabetes from hour to hour and from day to day, to aid in dose selection in insulin-treated patients, and for safety.

In general, for healthy young and middle-aged adults to achieve an A1C goal ≤7.0 percent, an FBG of 80 to 130 mg/dL (4.4 to 7.2 mmol/L) and a postprandial glucose (90 to 120 minutes after a meal) less than 180 mg/dL (10 mmol/L) are generally given as targets, but higher achieved levels may suffice [2,3]. In older patients, those with chronic kidney disease, or those with other risk factors for hypoglycemia, in whom the A1C goal is set higher, a higher FBG target (eg, 90 or 100 to 150 mg/dL [5 to 8.3 mmol/L]) may be used.

Cardiovascular effects — Compared with oral agents and independent of the level of glycemia achieved, insulin does not appear to prevent or increase major adverse cardiovascular events. In addition, the type of insulin does not affect cardiovascular outcomes.

Addition of insulin

●Oral agent(s) plus glargine versus oral agent(s) only – Combination oral agent and glargine therapy does not appear to reduce or increase cardiovascular outcomes compared with oral agent(s) only, as illustrated by the findings of the Outcome Reduction with Initial Glargine Intervention (ORIGIN) trial [4]. In this trial, over 12,500 patients with cardiovascular risk factors plus type 2 diabetes or prediabetes were randomly assigned to an evening dose of glargine or to standard care. Approximately 60 percent of the patients with prior diabetes were using oral glucose-lowering agents (predominantly metformin or sulfonylurea). The glargine was titrated to achieve an FBG level of <95 mg/dL (5.3 mmol/L). After a median follow-up of six years, the achieved median FBG levels were 94 and 123 mg/dL (5.2 and 6.8 mmol/L), respectively.

The rates of incident cardiovascular outcomes were similar in the glargine and standard care groups (2.94 and 2.85 per 100 person-years, respectively). Only 11 percent of the patients in the standard therapy group received insulin. A1C values were similar at baseline (6.4 percent) and at study end (6.2 and 6.5 percent). Approximately 60 percent of patients in both groups were treated with statins and 75 percent with angiotensin-converting enzyme (ACE) inhibitors or angiotensin II receptor blockers (ARBs).

In contrast to the findings in the ORIGIN trial, a retrospective cohort study of United States veterans taking metformin showed that the addition of insulin versus sulfonylurea was associated with an increased risk of a composite outcome of nonfatal cardiovascular outcomes and all-cause mortality [5]. Of the two outcomes included in the composite, only the rate of all-cause mortality was significantly increased (33.7 versus 22.7 events per 1000 person-years). Myocardial infarction (MI) and stroke rates were similar (10.2 and 11.9 events per 1000 person-years). The main limitation of this nonrandomized, retrospective study is the inadequate control for unmeasured confounders, including bias by indication, which could be responsible for the findings, despite attempts to minimize this. In comparison with the number of metformin and sulfonylurea users (39,990), there were very few metformin and insulin users (2948). The patients receiving insulin had higher A1C levels and more comorbidities. Thus, those treated with insulin were sicker and may have had greater insulin resistance and/or other factors to explain the higher rate of mortality.

Comparison of insulin types

●Insulin degludec versus insulin glargine – In a two-year noninferiority trial, cardiovascular events were similar in patients randomly assigned to insulin degludec or glargine [6]. In this masked trial, 7637 patients with type 2 diabetes (mean A1C 8.4 percent) with or at high risk for cardiovascular disease (CVD) were randomly assigned to insulin degludec or glargine once daily. At baseline, approximately 60 percent of the patients were treated with metformin, 29 percent with sulfonylurea, and 84 percent with insulin. The primary composite outcome (first occurrence of a death from cardiovascular causes, nonfatal MI, or nonfatal stroke) occurred in 8.5 and 9.3 percent of the patients receiving degludec and glargine, respectively. Glycemic management was similar throughout the trial; rates of severe and nocturnal hypoglycemia were lower in patients taking degludec, although the absolute rates of hypoglycemia were very low in both treatment groups. The insulin titration target in this trial was FBG of 71 to 90 mg/dL (4 to 5 mmol/L), with the option to raise the target to 90 to 126 mg/dL (5 to 7 mmol/L) for older patients or per investigator discretion.

●Basal versus prandial insulin – The type of insulin (basal or prandial) does not appear to affect cardiovascular outcomes, as illustrated by the findings of the Hyperglycemia and its Effect After Acute Myocardial Infarction on Cardiovascular Outcomes in Patients with Type 2 Diabetes Mellitus (HEART2D) trial, which was designed to evaluate the effects of prandial (lispro) or basal (NPH twice daily or insulin glargine once daily) insulin on cardiovascular outcomes in 1115 patients after MI [7]. At baseline, 50 percent of the patients were taking metformin, sulfonylureas, or both, whereas the remaining patients were treated with insulin monotherapy. At a mean follow-up of 2.7 years, the trial was stopped early due to lack of efficacy. There was no difference between the prandial and basal groups in the time to a subsequent cardiovascular event (cardiovascular death, nonfatal MI, nonfatal stroke, coronary revascularization, or hospitalization for an acute coronary syndrome).

INDICATIONS FOR INSULIN

Initial treatment — Although we suggest lifestyle intervention (diet, weight reduction, exercise) and metformin (in the absence of contraindications) for the initial treatment of type 2 diabetes, some patients are candidates for insulin as their initial therapy. (See "Initial management of hyperglycemia in adults with type 2 diabetes mellitus", section on 'Initial pharmacologic therapy'.)

Indications for initial treatment with insulin include the following:

●Severe hyperglycemia on presentation – Insulin may be indicated for initial treatment for some patients with type 2 diabetes, depending on the severity of the baseline metabolic disturbance [8,9].

•For patients presenting with symptomatic (eg, weight loss, polydipsia, polyuria) or severe hyperglycemia with ketonuria, insulin is indicated for initial treatment.

•For patients presenting with severe hyperglycemia (FBG >250 mg/dL [13.9 mmol/L], random glucose consistently >300 mg/dL [16.7 mmol/L], A1C >9 percent [74.9 to 85.8 mmol/mol]) but without ketonuria or spontaneous weight loss, in whom type 1 diabetes is not likely, insulin is an option (or a glucagon-like peptide 1 [GLP-1] receptor agonist), along with metformin, if no contraindications.

●Difficulty distinguishing type of diabetes – Insulin is indicated for initial treatment in patients in whom it is difficult to distinguish type 1 from type 2 diabetes.

Patients who are initially thought to have type 2 diabetes may actually have type 1 diabetes. Although the peak incidence of type 1 diabetes occurs around the time of puberty, approximately 42 percent of cases present after 30 years of age [10].

There are certain clinical features that, if present at any age, suggest the diagnosis of type 1 diabetes. These include marked and otherwise unexplained recent weight loss (irrespective of the initial weight), a short history with severe symptoms (polyuria, polydipsia), and the presence of moderate or greater ketonuria.

A personal or family history of autoimmune disease (eg, hypo- or hyperthyroidism, celiac sprue) without family history of type 2 diabetes, especially in a patient without a history of overweight, is also suggestive of type 1 diabetes. Diabetic ketoacidosis at first presentation, especially if severe, suggests that the patient has type 1 diabetes and will require lifelong insulin treatment, although there are exceptions to this general rule [11]. (See "Classification of diabetes mellitus and genetic diabetic syndromes" and "Syndromes of ketosis-prone diabetes mellitus".)

Some individuals with adult-onset type 1 diabetes may be indistinguishable clinically from a patient with type 2 diabetes at presentation but will slowly progress to insulin dependence. This is sometimes referred to as "latent autoimmune diabetes in adults" (LADA) (see "Classification of diabetes mellitus and genetic diabetic syndromes", section on 'Distinguishing type 1 from type 2 diabetes'). LADA can be distinguished from type 2 diabetes by the presence of pancreatic autoantibodies, such as glutamic acid decarboxylase antibodies. These individuals may respond poorly to oral hypoglycemic drug therapy, and the use of sulfonylureas as initial therapy may cause earlier insulin dependence [12,13]. The best initial treatment strategy for LADA is unclear. Studies are required to determine whether early treatment with insulin or use of immunomodulator therapy may prevent disease progression.

●Pancreatic insufficiency – Insulin is also indicated for patients with secondary diabetes due to pancreatic insufficiency, including from cystic fibrosis, chronic pancreatitis, or after pancreatectomy. (See "Cystic fibrosis-related diabetes mellitus", section on 'Insulin therapy' and "Chronic pancreatitis: Management", section on 'Endocrine insufficiency (pancreatogenic diabetes)'.)

●Other – Initial intensive insulin treatment for a brief period (two to four weeks) may be beneficial in patients with type 2 diabetes and may induce a remission that can last for a year or more [14-16]. By inducing near-normoglycemia with intensive insulin therapy, both endogenous insulin secretion and insulin sensitivity improve [14,16,17]. The improvement in insulin secretion is presumably due to the elimination of the deleterious effects of hyperglycemia on beta cell secretory function, and, in some patients, it results in better glycemic management that can then be maintained with diet and exercise for many months or even years thereafter [16].

This was illustrated in a randomized trial of short-term (two to five weeks) intensive insulin therapy versus oral agents (gliclazide and/or metformin) in 410 Chinese patients with newly diagnosed type 2 diabetes (mean FBG 202 mg/dL [11.2 mmol/L]) [18]. Insulin and oral agents were titrated to achieve an FBG <110 mg/dL (6.1 mmol/L). Therapy was discontinued two weeks after achieving glycemic targets. The patients were instructed to continue diet and exercise alone and were closely followed for recurrence of hyperglycemia (FBG >126 mg/dL [7 mmol/L] or postprandial glucose >180 mg/dL [10 mmol/L]).

A greater proportion of patients receiving insulin achieved glycemic goals in less time than those assigned to oral agents (over 90 percent within 4 to 5.6 days versus 84 percent in 9.3 days). Remission rates after one year were higher in the insulin groups (insulin pump or multiple daily injections) than in the oral hypoglycemic group (51 and 45 versus 27 percent). Patients in remission had lower initial FBG and A1C, and they achieved glycemic goals more quickly than those with recurrent hyperglycemia.

Initial, short-term intensive insulin treatment is not widely used, possibly due to the complexity of introducing intensive insulin treatment at diagnosis, patient reluctance, and even provider reluctance (given the challenges with initiating basal insulin in patients with longstanding diabetes).

Persistent hyperglycemia on oral agents — Oral agents become less effective as beta cell function declines (figure 2). The therapeutic options for patients who have persistent hyperglycemia with lifestyle intervention and metformin are to add a second oral or injectable agent, including addition of insulin as an option, or to switch to insulin. There is no consensus on which option is most effective. However, insulin (or a GLP-1 receptor agonist if the patient is not catabolic) is the preferred second-line medication for patients with A1C >9 percent or with persistent symptoms of hyperglycemia despite metformin titration. Insulin is always effective and is preferred in insulin-deficient, catabolic diabetes (eg, polyuria, polydipsia, weight loss). Patients should be counseled that initiating insulin does not represent a personal "failure" and that many patients with type 2 diabetes will eventually require exogenous insulin, due to decline in endogenous insulin production (table 2). (See "Management of persistent hyperglycemia in type 2 diabetes mellitus", section on 'Our approach'.)

DESIGNING AN INSULIN REGIMEN — We advise clinicians to familiarize themselves with a limited number of regimens (to facilitate addressing patient preferences, lifestyle [meal and activity patterns], etc) and use them consistently.

The glycemic differences achieved among different insulin regimens, assuming that they are adequately titrated, are modest. However, basal insulin is frequently suboptimally titrated in clinical practice [19]. The majority of the studies provide information on lowering of glycemia but do not provide information regarding the effects of various insulin regimens on microvascular or macrovascular complications or mortality. When differences in A1C levels have been found, they are sometimes offset by adverse events such as hypoglycemia.

Insulin initiation — For patients who are initiating insulin (in addition to oral agents, in place of oral agents, or as initial treatment), we suggest initiating basal, rather than prandial, insulin.

Basal insulin will improve nocturnal and fasting blood glucose (FBG), whereas prandial (pre-meal) bolus insulin will decrease postprandial glucose excursions. Whether a basal or a prandial strategy is more effective in improving microvascular complications remains uncertain. The type of insulin regimen (basal or prandial) does not appear to affect cardiovascular outcomes. (See 'Cardiovascular effects' above.)

Initiation of insulin therapy with a basal insulin has the advantage of convenience and simplicity in patients who are using insulin for the first time. Although basal and prandial insulin are similarly effective in improving A1C when insulin doses are aggressively titrated to achieve glycemic goals, basal insulin is associated with greater patient satisfaction and less frequent hypoglycemia [20,21]. As examples:

●In a randomized trial of once-daily insulin glargine versus prandial insulin lispro in 415 patients who were inadequately managed with metformin and a sulfonylurea, there were similar improvements in A1C (mean decrease of 1.7 and 1.9 percent, respectively) and target A1C concentrations between 6.5 and 7.0 percent were achieved by 27 and 30 percent of subjects, respectively [20]. Basal insulin was associated with greater patient satisfaction and less hypoglycemia.

●The Hyperglycemia and its Effect After Acute Myocardial Infarction on Cardiovascular Outcomes in Patients with Type 2 Diabetes Mellitus (HEART2D) trial was designed to evaluate the effects of prandial (lispro) or basal (NPH twice daily or insulin glargine once daily) insulin on cardiovascular outcomes in 1115 patients after myocardial infarction (MI) [7]. At baseline, 50 percent of the patients were taking metformin, sulfonylureas, or both, whereas the remaining patients were treated with insulin monotherapy. At a mean follow-up of 2.7 years, the trial was stopped early due to lack of efficacy. There was no difference between the prandial and basal groups in A1C (7.7 versus 7.8 percent). The cardiovascular outcomes are reviewed above. (See 'Cardiovascular effects' above.)

Patients are usually not eager to start insulin. Reasons for reluctance to start insulin should be explored and addressed (table 2).

Choice of basal insulin — A single daily dose of either insulin NPH or detemir given at bedtime or insulin glargine or degludec given in the morning or at bedtime is a reasonable initial regimen (table 1). In practice, payer coverage is often an important consideration in the selection of basal insulin. (See "General principles of insulin therapy in diabetes mellitus", section on 'Insulin preparations'.)

The basal insulin preparations do not differ significantly in glycemic efficacy [22,23]. Among basal insulin preparations, insulin glargine, detemir, and degludec may have less nocturnal hypoglycemia (but not always total hypoglycemia) compared with NPH, with the important disadvantage of higher cost. There does not appear to be any difference in hypoglycemia-related hospital admissions or emergency department visits. As examples:

●In meta-analyses of trials comparing once-daily insulin glargine or detemir with once-daily or twice-daily NPH insulin, there were similar improvements in A1C with all types of basal insulin [23-26]. However, in some of the meta-analyses, the rates of overall symptomatic and nocturnal hypoglycemia (while relatively infrequent with either basal insulin) were lower in patients treated with either insulin glargine or detemir compared with NPH [23-25,27,28].

●In a retrospective observational study using data from a large health care delivery system (>25,000 patients initiating basal insulin), there was no benefit of insulin analogs compared with NPH in reducing emergency department or hospital admissions for hypoglycemia (11.9 versus 8.8 events per 1000 person-years, respectively) despite slightly better glycemic management in the NPH group (achieved A1C 8.2 versus 7.9 percent with NPH, suggesting they were not treated with less aggressive doses) [29].

Insulin degludec appears to have similar glycemic efficacy as that of insulin glargine and, in some trials, a lower rate of hypoglycemia, especially if aiming for more stringent glycemic targets [6,30-34]. As an example, in a 65-week, double-blind, crossover trial, 721 adults with type 2 diabetes (mean A1C 7.6 percent) and at least one risk factor for hypoglycemia were randomly assigned to receive once-daily insulin degludec or insulin glargine for 32 weeks and then crossed over to the alternate insulin treatment for the next 32 weeks [35]. The rate of overall (185.6 versus 265.4 episodes per 100 patient-years of exposure) and nocturnal (55.2 versus 93.6 episodes) symptomatic hypoglycemia was lower with degludec (rate ratios 0.70, 95% CI 0.61-0.80 and 0.58, 95% CI 0.46-0.74, respectively). There was no difference in relatively rare severe hypoglycemia (nonsignificant reduction of 0.62 episodes per 100 patient-years with degludec). Overall glycemic management was similar (A1C 7 to 7.1 percent).

Although degludec significantly reduced overall and nocturnal hypoglycemia, the modest benefit (on average, one episode less every five years) must be balanced against its relatively higher cost and its briefer clinical experience. [36].

Initial dose — For patients with type 2 diabetes, the initial dose of insulin (whether in addition to oral agents, in place of oral agents, or as initial treatment) is similar [25,37,38]. Many algorithms have been published; one simple and conservative algorithm is presented here (algorithm 1) (related Pathway(s): Diabetes: Initiation and titration of insulin therapy in non-pregnant adults with type 2 DM). We start with bedtime NPH or detemir, taken at 10:00 PM if the person is testing his or her FBG at 7:00 or 8:00 AM, or bedtime or morning glargine or degludec. Since glargine and degludec can be administered any time of day, the timing of daily insulin glargine or degludec is based on patient preference to facilitate adherence.

The initial dose for NPH, detemir, glargine, or degludec is 0.2 units per kg (minimum 10 units, up to 15 to 20 units) daily. If FBG levels are very elevated (>250 mg/dL [13.9 mmol/L]), A1C is >8 percent, or if a patient is known to be very insulin resistant, initial doses of basal insulin can be higher (eg, 0.3 units per kg or up to 15 to 20 units daily as an initial dose). Subsequent modifications can be made according to daily measurement of FBG and every three-month measurement of A1C. (See 'Titrating dose' below.)

●Combination therapy – The rationale for combination oral hypoglycemic drug and insulin therapy is that using glucose-lowering medications with different mechanisms of action may achieve glucose targets while minimizing total insulin requirements and weight gain [39]. Metformin is often continued with the addition of insulin. Other agents including glucagon-like peptide 1 (GLP-1) receptor agonists, dipeptidyl peptidase-4 (DPP-4) inhibitors, and sodium-glucose co-transporter 2 (SGLT2) inhibitors can also be continued when insulin is added, although the putative advantages of doing so must be balanced against the downside of regimen complexity (polypharmacy) and increased cost. Sulfonylureas, meglitinides, and pioglitazone are usually tapered and stopped when starting insulin, especially prandial insulin, due to reduced efficacy in comparison with other combinations and to adverse effects [40]. However, there are some situations where one of these agents may be combined with insulin (eg, using a thiazolidinedione in a patient with lipodystrophy and severe insulin resistance). (See "Management of persistent hyperglycemia in type 2 diabetes mellitus", section on 'Insulin initiation and intensification' and "Lipodystrophic syndromes", section on 'Treatment of lipodystrophy'.)

●Insulin monotherapy – Switching to insulin monotherapy may be cheaper than combined oral agent-insulin therapy depending on the combination used (metformin is very inexpensive) but may result in slightly more weight gain, depending on the combination, and more episodes of hypoglycemia, few of which are severe [39]. The oral agent may be discontinued when insulin is initially added or after the patient is on an adequate dose of basal insulin. The former approach may be associated with elevations in glucose levels until the dose of injected insulin is sufficient to achieve metabolic control. (See "Interactive diabetes case 2: Switching from oral agents to insulin in type 2 diabetes".)

Titrating dose — The basal insulin regimen is adjusted based on FBG, A1C values, and frequently, bedtime glucose. For many patients with type 2 diabetes, basal insulin alone is often adequate for glycemic management because it reduces glucose toxicity, thereby increasing endogenous insulin secretion, which is sufficient for postprandial excursions. However, patients with type 2 diabetes and persistently elevated A1C despite fasting blood glucose in target range or people on very high doses of insulin may require pre-meal insulin, similar to treatment for type 1 diabetes. Pre-meal insulin regimens require self-monitoring of glucose levels in addition to fasting levels (ie, before the meals when rapid-acting insulin is used to help determine and subsequently adjust preprandial dosing).

Persistent fasting hyperglycemia — If the mean FBG is above target (typically 130 mg/dL [7.2 mmol/L]), an increase of 2 to 4 units in the basal insulin dose should be made periodically (approximately every three days) (algorithm 1) (related Pathway(s): Diabetes: Initiation and titration of insulin therapy in non-pregnant adults with type 2 DM) [37]. In this way, the basal insulin dose can be titrated over a period of several weeks or months. If FBG levels are very elevated (>250 mg/dL [13.9 mmol/L]) or if a patient is known to be very insulin resistant, titration can be more aggressive, with the dosing increment decreasing as the patient approaches the target blood glucose. Patients should be proactively counseled to reduce the daily insulin dose if hypoglycemia develops.

As patients approach the glycemic target, it may be helpful to check a mid-sleep blood glucose to rule out hypoglycemia (this can be done when the patient spontaneously wakes overnight).

Persistent elevation in A1C with FBG in target range — For patients with persistently elevated glycated hemoglobin (A1C) levels who have fasting blood glucose (FBG) levels in the target range (80 to 130 mg/dL [4.4 to 7.2 mmol/L], or higher, depending on the individualized glycemic target), we advise the patient to check fingerstick capillary glucose levels fasting, pre-lunch, pre-dinner, and before bed while the insulin regimen is further adjusted. Prandial insulin is often started as a single injection before the largest meal of the day, but many strategies are possible (algorithm 1) (related Pathway(s): Diabetes: Initiation and titration of insulin therapy in non-pregnant adults with type 2 DM). Short-term monitoring with continuous glucose monitoring for two weeks may also be helpful in insulin dose adjustment. (See "Glucose monitoring in the management of nonpregnant adults with diabetes mellitus".)

Among patients whose A1C values remain above the desired target despite insulin adjustments, diet and exercise patterns should be reviewed. We advise the patient to continue to check fingerstick capillary glucose levels fasting, pre-lunch, pre-dinner, and before bed while the regimen is further adjusted. In general, inadequate dosing, dietary indiscretion, and/or mismatch of food and insulin dose underlie the apparent failure of many insulin regimens. Total daily insulin doses typically exceed 65 to 100 units per day and may sometimes be much higher in patients with type 2 diabetes and obesity. When the dose is >80 units daily, absorption kinetics (which are proportional to volume injected) may be altered. In this setting, we administer basal insulin in divided doses twice daily. (See "General principles of insulin therapy in diabetes mellitus", section on 'Size of subcutaneous depot'.)

Pre-lunch glucose elevated — If pre-lunch glucose values are elevated, options include adding prandial insulin at breakfast (see 'Combining prandial and basal insulin' below), or if the patient is taking bedtime detemir or NPH, adding a second dose of detemir or NPH at breakfast. Dietary modification at breakfast may also be effective.

When the insulin preparation is detemir, glycemic management over a 24-hour period may be more stable in patients taking two doses daily (table 1) [41]. If the goal is management of persistent hyperglycemia with a regimen that is simple and inexpensive, then twice-daily NPH will be effective in many patients [38,42]. (See 'Choice of basal insulin' above.)

Pre-dinner or bedtime glucose elevated — If blood glucose values are elevated before dinner and/or bed, we typically add prandial insulin (see 'Combining prandial and basal insulin' below). Either short-acting (regular) or rapid-acting insulin can be given before lunch and/or dinner. It is often simpler to counsel smaller lunches, and then add a single injection of prandial insulin before dinner, which, for many people, is the largest meal of the day. If lunch is the largest meal of the day, the reverse approach may be used (eg, cover lunch with prandial insulin, eat a smaller dinner).

For patients taking bedtime NPH or detemir who have pre-dinner hyperglycemia, adding another dose of NPH or detemir insulin at breakfast is an alternative, particularly if pre-lunch glucose is also elevated. (See 'Pre-lunch glucose elevated' above.)

Combining prandial and basal insulin — For patients with type 2 diabetes who require prandial insulin, the goal is to adjust the dose of short-acting or rapid-acting insulin immediately prior to a meal, and therefore, we prefer to keep basal and pre-meal insulin injections separate and adjust them independently. Patients may draw up their pre-meal and NPH insulin in the same syringe prior to injection, whereas glargine, degludec, and detemir cannot be mixed with rapid-acting insulin. In the setting of multiple daily insulin injections, oral agents other than metformin are usually discontinued to reduce polypharmacy and cost. Oral or injectable GLP-1 receptor agonists may also be continued.

●Choice of prandial insulin – The choice of prandial insulin is based upon availability, patient preference, cost, and payer coverage. Compared with regular insulin, the newer rapid-acting insulins have a minor glycemic advantage especially in patients with type 1 diabetes. The ability to inject the rapid-acting insulins immediately before meals (as opposed to the 30 to 45 minutes before the meal with short-acting insulins) may improve adherence in all insulin-treated patients. There is little glycemic advantage of analog compared with regular insulins in type 2 diabetes [43,44]. This was illustrated in a meta-analysis of 10 randomized trials (involving 2751 patients with type 2 diabetes) that compared rapid-acting insulin analogs with regular insulin as pre-meal bolus doses [44]. No significant differences were seen in serum A1C concentrations or the number of hypoglycemic episodes.

●Pre-meal insulin dosing – The optimal dose of pre-meal insulin depends upon many factors, including current and target blood glucose values, carbohydrate content of the meal, and activity. A typical starting dose is approximately 4 to 6 units or 10 percent of basal insulin dose (algorithm 1). The dose can be increased every three days until the postprandial blood glucose target is achieved.

Prandial insulin dose increases depend on how much prandial insulin the patient is using. As a rule of thumb:

•≤10 units – Increase by 1 unit

•11 to 20 units – Increase by 2 units

•>20 units – Increase by 5 units (or more, depending on patient insulin resistance, meal size, and content)

A more complex method for adjusting pre-meal insulin is to match insulin delivery to the anticipated glucose excursion with meals. Many patients benefit from specific training in carbohydrate counting, which requires some arithmetical computations that some patients find difficult or burdensome. It is uncertain if there is a glycemic advantage to carbohydrate counting in patients with type 2 diabetes. As an example, in a 24-week, randomized trial in 277 adults with type 2 diabetes, there was no difference in A1C values when mealtime insulin adjustments were based upon a simple algorithm (according to previous weeks' blood glucose monitoring results) versus carbohydrate counting, using an insulin-to-carbohydrate ratio for each meal [45]. Either method is acceptable, and patient preference can guide the choice of method for pre-meal dosing. A simplified adaptation in which patients take a slightly higher dose of prandial insulin for high carbohydrate meals may yield many of the benefits of carbohydrate counting without the complexity and is more straightforward for most patients with type 2 diabetes. (See "Nutritional considerations in type 2 diabetes mellitus", section on 'Carbohydrate consistency'.)

For some patients, providing supplemental insulin in addition to prandial insulin is an option. In this framework, every prandial insulin dose can be considered to have two components: the part covering the meal (prandial) and the part correcting the hyperglycemia (supplemental insulin). For people with type 2 diabetes, 1 to 2 units for every 50 mg/dL above the target glucose may be added to the prandial insulin dose. Since the math can be confusing for many people, patients may be prescribed a "mealtime sliding scale" in which the recommended dose of insulin increases with the plasma glucose level. In this framework, the "mealtime sliding scale" includes the prandial and supplemental components. It is important to educate the patient that this combined scale could lead to insulin overdose if used as supplemental or correction insulin when the patient is not eating. Patients should be counseled regarding how to reduce the insulin dose for exercise or for procedures that require limiting food intake. (See "Perioperative management of blood glucose in adults with diabetes mellitus", section on 'Glucose management'.)

●Pre-mixed insulin – Some insulins are commercially available in a pre-mixed formulation. Most pre-mixed (biphasic) preparations contain an intermediate-acting insulin and either a short-acting or a rapid-acting insulin. We suggest not using pre-mixed insulin initially, because of limited flexibility in adjusting doses. However, pre-mixed insulin is a reasonable option for patients with type 2 diabetes who are doing well on a stable, fixed ratio, particularly those who eat a larger breakfast and dinner and a smaller lunch, or are able to modify their diets to match the kinetics of pre-mixed insulin. (See "General principles of insulin therapy in diabetes mellitus", section on 'Pre-mixed insulins'.)

Pre-mixed rapid-acting preparations offer little glycemic advantage compared with adequately titrated basal and bolus insulin. In an open-label trial, 708 patients with type 2 diabetes who were suboptimally managed with metformin and a sulfonylurea were randomly assigned to pre-mixed biphasic insulin aspart (twice daily), prandial insulin aspart (three times daily), or basal insulin detemir (once or twice daily), there was no difference in median A1C levels among the three groups (7.1, 6.8, and 6.9 percent, respectively), but significantly more patients in the basal and prandial groups achieved an A1C level ≤6.5 percent than in the pre-mixed biphasic group (43, 45, and 32 percent, respectively) [21]. The majority of all three treatment groups used a second type of insulin, per protocol, during the trial to obtain the stipulated glycemic goals. Patients in the basal group had the fewest episodes of hypoglycemia. In other trials, pre-mixed rapid-acting preparations were more often associated with minor hypoglycemia and weight gain than long-acting insulin or oral agents [46].

●Intensive insulin regimens – If intensive insulin therapy is chosen in a patient with type 2 diabetes, the pretreatment considerations, choice of regimen, and management issues are similar to those for patients with type 1 diabetes. For people who are very hyperglycemic, it is best to advance the regimen slowly to reduce the risk of hypoglycemia as glucotoxicity resolves, to allow a slow osmotic re-equilibration and to enable the patient to learn to use insulin safely.

Insulin pump therapy is used infrequently in patients with type 2 diabetes, but it may have a role in a select group of patients with poorly managed type 2 diabetes taking multiple daily injections or for people with limited dexterity when someone is available to help set up the device [47]. Intensive insulin and insulin pump therapy are reviewed in detail elsewhere. (See "Management of blood glucose in adults with type 1 diabetes mellitus" and "Continuous subcutaneous insulin infusion (insulin pump)", section on 'Types of insulin pumps'.)

Use of an intensive insulin regimen (similar to that used in type 1 diabetes) targeting tight glycemic targets results in higher serum insulin concentrations and better glycemic management than that achieved with either an oral drug or conventional insulin therapy alone [48]. A potential problem is the weight gain (averaging 8.7 kg in one study) that can occur with intensive regimens that achieve near-normoglycemia [49]. This weight gain may, in some instances, be counterproductive or result in partial noncompliance with therapy. (See 'Disadvantages of insulin therapy' above and "Nutritional considerations in type 2 diabetes mellitus".)

Conversion between basal insulin products

●Conversion from twice-daily to once-daily basal insulin – If the patient is taking NPH or detemir twice per day and prefers once-daily basal insulin dosing, a switch to glargine (U-100 or U-300) or degludec once daily may be safely done by reducing the total daily basal insulin dose by 10 to 20 percent and re-titrating based on blood glucose levels. If the patient is very hyperglycemic without any hypoglycemia and confirms adherence to the prescribed twice-daily dose, the equivalent total daily dose may be given as a once-daily long-acting insulin.

●Conversion between once-daily basal insulins – If a patient is switching from once-daily detemir to once-daily glargine (U-100 or U-300 glargine) or degludec, we also reduce the dose by 10 to 20 percent. If the patient is very hyperglycemic, the equivalent total daily dose may be used.

Basal insulin administered at bedtime suppresses nocturnal hepatic gluconeogenesis. The longer-acting basal insulins may have less effect overnight and more effect into the next day, which should not be harmful. However, we generally reduce the dose initially to avoid the risk of hypoglycemia, and then increase the dose as needed to maintain glycemic targets.

Use of automated insulin algorithms — In order to achieve glycemic goals, the dose of insulin must be titrated often. Most patients with type 2 diabetes adjust their insulin dose with clinician advice (rather than independently), but the adjustments are made less frequently than necessary, in part due to clinician time constraints. The use of automated insulin algorithms may facilitate more timely dose adjustments. A smart pen insulin delivery device (which uses aspart or lispro cartridges) is available for people who use multiple daily insulin injections. The smart pen provides decision support for dosing similar to a bolus calculator on an insulin pump. It also tracks dosing history, which facilitates monitoring of adherence [50].

In addition, handheld devices are under development that contain a glucose meter and use an insulin algorithm to determine each insulin dose, based on the glucose readings. The software analyzes glucose patterns and adjusts the insulin dose to meet patient needs. In a randomized trial in 181 insulin-requiring patients with type 2 diabetes (mean A1C 8.6 percent), the use of one such device along with clinician support reduced A1C compared with clinician support alone (from baseline to six months, -1 versus -0.3 percentage points) [51]. The frequency of hypoglycemia (glucose <54 mg/dL [3 mmol/L]) was similar in the two groups, whereas there was more weight gain in the intervention group (2.3 versus 0.7 percent above baseline). This type of device, and similar mobile phone applications, shows promise for improving glycemic management in patients with type 2 diabetes.

TROUBLESHOOTING

Hypoglycemia — An increased risk of hypoglycemia is a potential complication of insulin therapy. However, patients with type 2 diabetes experience much less frequent hypoglycemia than patients with type 1 diabetes [52].

Although basal insulin is associated with less hypoglycemia than prandial insulin (see 'Insulin initiation' above), hypoglycemia can occur when the dose of basal insulin is increased to cover meals. If the patient subsequently eats less than usual, hypoglycemia may occur. Alternatively, some patients develop daytime hypoglycemia on a dose of basal insulin that controls fasting blood glucose (FBG). Both of these scenarios lead to obligate snacking, which may fuel insulin-associated weight gain. This problem may be identified by asking about symptoms of hypoglycemia when meals are skipped or snacking to prevent hypoglycemia. Other potential triggers (eg, changes in diet or activity) should be identified. Patients who make significant dietary changes (eg, starting a ketogenic diet) may require substantial reductions in insulin dosing (eg, ≥50 percent reduction). (See "Interactive diabetes case 18: A 61-year-old man with type 2 diabetes and a recent change in diet (medical nutrition therapy)".)

●Nocturnal hypoglycemia – The dose of basal insulin should be reduced (bedtime dosing if taking twice daily) by 4 units or 10 percent, whichever is greater (algorithm 1) (related Pathway(s): Diabetes: Initiation and titration of insulin therapy in non-pregnant adults with type 2 DM). If the patient is taking bedtime NPH, an alternative is to switch to detemir, insulin glargine, or degludec at 80 to 90 percent of the current total daily dose.

Among basal insulin preparations, insulin glargine, detemir, and degludec may have some relatively modest clinical advantages over NPH when pursuing tight glycemic targets (less symptomatic and nocturnal hypoglycemia) with the important disadvantage of high cost. (See 'Choice of basal insulin' above.)

●Daytime hypoglycemia – If the patient is taking prandial insulin, the dose should be decreased at the appropriate meal time(s) (eg, reduce breakfast prandial insulin if hypoglycemia occurs between breakfast and lunch). If the hypoglycemia is not severe, a typical approach is to decrease the dose based on how much prandial insulin the patient is taking at the relevant mealtime:

•≤10 units – Decrease by 2 units

•11 to 20 units – Decrease by 4 units

•>20 units – Decrease by 6 to 10 units or 50 percent

If the hypoglycemia is serious or severe, it is prudent to reduce the dose substantially (by 20 to 50 percent) and repeat the titration or to discontinue the prandial insulin and reinitiate/re-titrate if needed.

The patient should be instructed on how to adjust the prandial dose for meal size and carbohydrate content. In addition, patients should be asked about the timing of their prandial insulin dose and have appropriate timing reinforced if it appears to be contributing to episodes of hypoglycemia (for example, if the patient is taking the insulin following a meal rather than prior to the meal).

If the patient is not taking prandial insulin, the dose of basal insulin should be reduced (by 4 units or 10 percent, whichever is greater) (algorithm 1). If hyperglycemia develops after meals (based on self-monitoring of blood glucose), prandial insulin should be added to cover mealtime excursions. (Related Pathway(s): Diabetes: Initiation and titration of insulin therapy in non-pregnant adults with type 2 DM.)

●Severe hypoglycemia – If the patient has a concerning hypoglycemic event, especially one requiring the assistance of another person to actively administer carbohydrate (severe hypoglycemia) and there are no apparent changes in diet or activity to account for hypoglycemia, it is prudent to reduce the dose substantially (eg, by 20 to 50 percent) and repeat the titration of the basal insulin (and/or discontinue the prandial insulin and reinitiate/titrate, if needed).

●Unexplained hypoglycemia – In general, if there is concerning hypoglycemia, either due to the severity of the event or the level of risk in the patient (eg, an older person who lives alone, a person who has variable diet or activity), it is prudent to err on the side of safety, sometimes with a large dose reduction or even stopping the insulin (in a patient who is not insulin deficient), since incrementally reducing the insulin dose may not eliminate the hypoglycemia. The insulin may be resumed and re-titrated as needed.

The management of hypoglycemia in patients with diabetes is reviewed in more detail separately. (See "Hypoglycemia in adults with diabetes mellitus".)

Insulin resistance — In severely insulin-resistant patients (eg, requiring more than 200 total units of insulin daily), concentrated insulins can be used to manage hyperglycemia. Concentrated insulin formulations permit equivalent dosing in a smaller volume than U-100 insulin and without the need for numerous injections to deliver high doses. The pharmacokinetics of each product should be reviewed and individual patient response should be monitored since the concentrated insulins are very potent and the kinetics are often different than expected. As an example, the pharmacologic profile of U-500 regular insulin is most similar to that of NPH. (See "General principles of insulin therapy in diabetes mellitus", section on 'U-500 regular insulin'.)

Glargine 300 units/mL is very similar to glargine but has a volume one-third of that for the same dose of glargine 100 units/mL. The pharmacokinetics are slightly different, with less of a peak and a longer duration of action. Consequently, glargine 300 units/mL is more similar to degludec than glargine 100 units/mL [53]. (See "General principles of insulin therapy in diabetes mellitus", section on 'Basal insulin analogs'.)

Insulin-associated weight gain — Patients initiating insulin therapy should be aware of the potential for weight gain, and major emphasis should be placed on diet and lifestyle modification to prevent it. It is important to educate patients about insulin dose reduction for anticipated increases in physical activity and with changes in diet to reduce the risk of hypoglycemia as well as weight gain.

Patients with type 2 diabetes, insulin resistance, and obesity are susceptible to insulin-associated weight gain. This can be due to continued dietary indiscretion, reduction in glycosuria with improved glycemic management, conscious or subconscious snacking to support an insulin dose that is too high, overtreatment of hypoglycemia, overly tight glycemic targets, or a combination of these factors. The resulting weight gain worsens insulin resistance and may prompt insulin dose escalation, leading to a vicious cycle.

The magnitude of the weight gain depends upon the intensity of regimen (dose and frequency of insulin) and the dietary pattern [54]. In the United Kingdom Prospective Diabetes Study (UKPDS), the average weight gain after 10 years of insulin therapy was approximately 7 kg for patients with type 2 diabetes, with the most rapid weight gain occurring when insulin was first initiated [55]. Less intensive therapy with either insulin or a sulfonylurea (which increases endogenous insulin secretion) was associated with a 3.5 to 4.8 kg weight gain at three years versus no change with metformin monotherapy [48]. In a subsequent trial, weight gain was greater with prandial than basal insulin (4.8 versus 3.1 kg) [7]; however, patients receiving prandial insulin also received a higher insulin dose, which could account for the small difference in weight gain. In other trials, pre-mixed rapid-acting preparations were more often associated with weight gain than long-acting insulin or oral agents [46].

It is not clear if the weight gain is important for diabetes complications. Microvascular complications were reduced with insulin monotherapy in the UKPDS despite weight gain [55]. Whether weight gain with insulin might adversely affect risk for cardiovascular disease (CVD) in type 2 diabetes is not clear. However, in type 1 diabetes, intensive insulin therapy in Diabetes Control and Complications Trial (DCCT)/Epidemiology of Diabetes Interventions and Complications (EDIC) resulted in a major reduction in CVD despite weight gain [56]. A subsequent analysis of the DCCT data revealed that the CVD benefit of intensive therapy was attenuated by weight gain [57]. (See "Initial management of hyperglycemia in adults with type 2 diabetes mellitus", section on 'Diabetes education'.)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Diabetes mellitus in adults" and "Society guideline links: Diabetes mellitus in children" and "Society guideline links: Blood glucose monitoring".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topics (see "Patient education: Using insulin (The Basics)")

●Beyond the Basics topics (see "Patient education: Type 2 diabetes: Insulin treatment (Beyond the Basics)")

SUMMARY AND RECOMMENDATIONS

●Initial treatment of type 2 diabetes should begin with dietary modification, weight reduction, and exercise, which may achieve target glycated hemoglobin (A1C) levels if compliance is optimal. Metformin therapy (in the absence of contraindications) may be initiated, concurrent with lifestyle intervention, at the time of diabetes diagnosis. (See "Initial management of hyperglycemia in adults with type 2 diabetes mellitus".)

●Insulin is indicated as initial treatment for some patients with type 2 diabetes, depending on the severity of the baseline metabolic disturbance (eg, insulin is always indicated for patients presenting with symptomatic [eg, weight loss] or severe hyperglycemia with ketonuria, or in patients in whom it is difficult to distinguish type 1 from type 2 diabetes). (See 'Initial treatment' above.)

●After a successful initial response to oral therapy, most patients have worsening glycemia over time (figure 2) and require additional therapy (add a second oral or injectable agent, including insulin, or switch to insulin). For patients with a A1C relatively far from goal (eg, >9 percent [74.9 mmol/mol]) while taking metformin, or with persistent symptoms of hyperglycemia, we suggest adding insulin or a glucagon-like peptide 1 (GLP-1) receptor agonist (Grade 2B). While basal insulin has historically been the preferred medication to add to metformin when A1C is elevated (even in the absence of catabolic features), GLP-1 receptor agonists are a reasonable option when type 1 diabetes has been excluded. (See 'Persistent hyperglycemia on oral agents' above and "Management of persistent hyperglycemia in type 2 diabetes mellitus", section on 'Our approach'.)

●For patients who are initiating insulin (in addition to oral agents, in place of oral agents, or as initial treatment), we suggest initiating basal, rather than prandial, insulin (Grade 2B). This is predominantly due to a lower risk of hypoglycemia and greater convenience and simplicity for patients who are using insulin for the first time (algorithm 1). (See 'Insulin initiation' above.) (Related Pathway(s): Diabetes: Initiation and titration of insulin therapy in non-pregnant adults with type 2 DM.)

Either insulin neutral protamine Hagedorn (NPH) or detemir given at bedtime or insulin glargine or degludec given in the morning or at bedtime is a reasonable initial regimen (table 1). (See 'Choice of basal insulin' above.)

●The initial dose for NPH, detemir, glargine, or degludec is 0.2 units per kg (minimum 10 units) subcutaneously daily. Subsequent modifications can be made according to FBG and A1C values, until the FBG is in the target range (80 to 130 mg/dL [3.9 to 7.2 mmol/L] in young patients and with higher FBG targets for older patients and those at risk of hypoglycemia) (algorithm 1). (See 'Initial dose' above.) (Related Pathway(s): Diabetes: Initiation and titration of insulin therapy in non-pregnant adults with type 2 DM.)

●Among patients who are taking insulin and have A1C values above the desired target with FBG levels in the target range, dietary and exercise patterns should be reviewed. We advise the patient to check fingerstick capillary glucose levels fasting, pre-lunch, pre-dinner, and before bed while the regimen is being adjusted. Prandial insulin is often started as a single injection before the largest meal of the day, but many strategies are possible. (algorithm 1). (See 'Titrating dose' above.) (Related Pathway(s): Diabetes: Initiation and titration of insulin therapy in non-pregnant adults with type 2 DM.)

●For patients with type 2 diabetes who require prandial insulin, either short- (regular) or rapid-acting insulin can be given. The ability to inject the rapid-acting insulins immediately before meals (as opposed to the 30 to 45 minutes before the meal recommended for short-acting insulins) may provide improved convenience for patients. In this setting of multiple daily insulin injections, oral agents other than metformin are usually discontinued to reduce polypharmacy and cost. Oral or injectable GLP-1 receptor agonists may also be continued. (See 'Combining prandial and basal insulin' above.)

●For patients with type 2 diabetes who require prandial insulin, we suggest not using pre-mixed insulin initially (Grade 2B). The goal is to adjust the dose of short-acting or rapid-acting insulin immediately prior to a meal, and therefore, we prefer to keep basal and pre-meal insulin injections separate and adjust them independently. However, pre-mixed insulin is a reasonable option for patients with type 2 diabetes who are doing well on a stable, fixed ratio, especially if the meal pattern matches the kinetics of the pre-mixed insulin (eg, large breakfast and dinner with small or low-carbohydrate lunch). (See 'Combining prandial and basal insulin' above and "General principles of insulin therapy in diabetes mellitus", section on 'Pre-mixed insulins'.)

●Monitoring for complications of insulin (including hypoglycemia and weight gain) and identifying and addressing the triggers, if present (missed meals, alcohol, and unanticipated exercise for the former and dietary indiscretion and obligate snacking for the latter) should be performed at every visit to permit insulin regimen adjustment. Provide patient education at every visit to minimize these adverse effects of insulin therapy. (See 'Troubleshooting' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges David McCulloch, MD, who contributed to an earlier version of this topic review.