Chapter Headings
People with type 2 diabetes form a heterogeneous group. Consequently, treatment regimens and therapeutic targets should be individualized. The treatment of type 2 diabetes involves a multi-pronged approach that aims to treat and prevent symptoms of hyperglycemia, such as dehydration, fatigue, polyuria, infections and hyperosmolar states; and to reduce the risks of cardiovascular (CV) and microvascular complications (1). This includes healthy behaviour interventions (see Reducing the Risk of Diabetes chapter, p. S20; Cardiovascular Protection in People with Diabetes chapter, p. S162) and antihyperglycemic medications. This chapter provides updated recommendations for the approach to antihyperglycemic therapy and selection of pharmaceutical agents. The number of available antihyperglycemic agents is ever expanding, requiring the health-care provider to consider many of the following factors when choosing medications: degree of hyperglycemia, medication efficacy for reducing diabetes complications (microvascular and/or CV) and lowering glucose, medication effects on the risk of hypoglycemia, body weight, other side effects, concomitant medical conditions, ability to adhere to regimen, broader health and social needs, affordability of medications, and patient values and preferences. Recommendations in this chapter are based on a rigorous and careful review of the evidence regarding the efficacy and adverse effects of available medications on clinically important outcomes.
Table 1 Antihyperglycemic agents for use in type 2 diabetes |
|||||||
---|---|---|---|---|---|---|---|
A1C, glycated hemoglobin; BG, blood glucose; CrCl, creatinine clearance; CV, cardiovascular; CVD, cardiovascular disease; eGFR, estimated glomerular filtration rate; GI, gastrointestinal; HDL-C, high density lipoprotein cholesterol; LDL-C, low density lipoprotein cholesterol; MI; myocardial infarct. |
|||||||
Class and mechanism of action | Drug | Cost | A1C lowering* | Hypoglycemia | Weight | Effect on primary CVD outcomes | Other therapeutic considerations |
First Line | |||||||
Biguanide: Enhances insulin sensitivity in liver and peripheral tissues by activation of AMP-activated protein kinase | Metformin Metformin extended-release |
$ | Approx. 1.0† |
Negligible risk as monotherapy | Neutral | Reduction in myocardial infarction in overweight individuals |
|
Second Line | |||||||
Incretin: Increases glucose-dependent insulin release, slows gastric emptying, inhibits glucagon release | DPP-4 inhibitors Alogliptin Linagliptin Saxagliptin Sitagliptin |
$$$ | 0.5 to 0.7 | Negligible risk as monotherapy | Neutral | Neutral (for alogliptin, saxagliptin and sitagliptin) |
|
GLP-1 receptor agonists** Short-acting Exenatide Lixisenatide Longer-acting Dulaglutide Exenatide extended-release Liraglutide |
$$$$ | 1.0 | Negligible risk as monotherapy | Loss of 1.6 to 3 kg | Reduction in MACE‡ and CV death in participants with clinical CVD (for liraglutide) Neutral (for exenatide ER, lixisenatide) |
|
|
SGLT-2 inhibitors: Inhibits SGLT-2 transport protein to prevent glucose reabsorption by the kidney | Canagliflozin Dapagliflozin Empagliflozin |
$$$ | 0.4 to 0.7 | Negligible risk as monotherapy | Loss of 2 to 3 kg | Reduction in MACE‡ (empagliflozin and canagliflozin) and CV death (empagliflozin) in participants with clinical CVD |
|
Alpha-glucosidase inhibitor: Inhibits pancreatic α-amylase and intestinal α-glucosidase | Acarbose | $$ | 0.7 to 0.8§ | Negligible risk as monotherapy | Neutral | — |
|
Insulin: Activates insulin receptors to regulate metabolism of carbohydrate, fat, and protein | Bolus (prandial) Insulins Rapid-acting analogues Aspart Aspart (faster-acting) Glulisine Lispro U-100 Lispro U-200 Short-acting Regular Basal Insulins Intermediate-acting NPH Long-acting analogues Degludec U-100 Degludec U-200 Detemir Glargine U-100 Glargine U-100 (biosimilar) Glargine U-300 Premixed Insulins Premixed regular-NPH Biphasic insulin aspart Lispro/lispro protamine suspension |
$ to $$$$ | 0.9 to 1.2 or more | Significant risk | Gain of 4 to 5 kg Gain of 0 to 0.4 kg for long-acting analogue alone |
Neutral (for glargine and degludec) |
|
Insulin secretagogue: Activates sulfonylurea receptor on β-cell to stimulate endogenous insulin secretion | Sulfonylureas Gliclazide Gliclazide modified-release |
$ | 0.7 to 1.3 | Minimal/ moderate risk |
Gain of 1.5 to 2.5 kg | — |
|
Glimepiride | Moderate risk | ||||||
Glyburide | Moderate risk | ||||||
(note: chlorpropamide and tolbutamide are still available in Canada, but rarely used) | |||||||
Meglitinides Repaglinide |
$$ | 0.7 to 1.1 | Minimal/ moderate risk | Gain of 0.7 to 1.8 kg | — | ||
Thiazolidinedione (TZD): Enhances insulin sensitivity in peripheral tissues and liver by activation of peroxisome proliferator- activated receptor gamma receptors | Pioglitazone Rosiglitazone |
$$$ | 0.8 to 0.9 | Negligible risk as monotherapy | Gain of 2.5 to 5 kg | Neutral (for pioglitazone) |
|
Weight loss agent: Inhibits lipase | Orlistat | $$$ | 0.2 to 0.4 | Negligible risk as monotherapy | Loss of 3 to 4 kg | — |
|
Individuals presenting with newly diagnosed type 2 diabetes require a multifaceted treatment plan. This includes diabetes education by an interprofessional team (see Self-Management Education and Support chapter, p. S36), healthy behaviour interventions (diet and physical activity, smoking cessation) with a target of 5% to 10% weight loss for overweight individuals (see Weight Management in Diabetes chapter, p. S124; Cardiovascular Protection in People with Diabetes chapter, p. S162), and screening for complications. It should be emphasized to people with type 2 diabetes that healthy behaviour interventions and weight loss can lead to withdrawal of antihyperglycemic medication and even remission of type 2 diabetes in some cases (2). The Look AHEAD (Action for Health in Diabetes) trial showed that an intensive healthy behaviour intervention resulted in a significantly greater weight loss and likelihood of diabetes remission after 1 year compared to standard care, with the greatest benefit seen in persons with new-onset type 2 diabetes (21.2% remission rate) (2). Antihyperglycemic therapy with metformin may also be initiated at diagnosis, depending on the current and target glycated hemoglobin (A1C).
The treatment of hyperglycemia should begin with the establishment of a target A1C which, in most cases, will be ≤7.0% as this has been shown to reduce long-term microvascular complications in newly diagnosed people with type 2 diabetes (3). A1C targets may be higher (up to 8.5%) if the benefits of intensive glycemic control are unlikely to outweigh the risks and burden, such as in individuals with limited life expectancy, high risk of hypoglycemia, multimorbidity, or based on the values and preferences of the person with diabetes (see Targets for Glycemic Control chapter, p. S42 for recommendations). It should be emphasized to people with type 2 diabetes that reductions in A1C levels are associated with better outcomes even if recommended glycemic targets cannot be reached, and inability to achieve A1C target should not be considered a treatment failure (3,4).
If the A1C level at diagnosis is less than 1.5% above target and the person with type 2 diabetes lacks metabolic decompensation and/or symptoms of hyperglycemia, the first line of treatment should be healthy behaviour interventions (see Reducing the Risk of Diabetes chapter, p. S20). If healthy behaviour interventions are insufficient to achieve target A1C levels within 3 months, they should be combined with antihyperglycemic medications. In the face of significant hyperglycemia (i.e. A1C >1.5% above target), pharmacotherapy is usually required at diagnosis concurrent with healthy behaviour interventions. People who have evidence of metabolic decompensation (e.g. marked hyperglycemia, ketosis or unintentional weight loss) and/or symptomatic hyperglycemia should be started immediately on insulin, regardless of A1C level. Insulin may later be tapered or discontinued once stability is achieved.
In general, A1C will decrease by about 0.5% to 1.5% with monotherapy, varying with the specific agent used and the baseline A1C level. By and large, the higher the baseline A1C, the greater the A1C reduction seen for each given agent. The maximum effect of noninsulin antihyperglycemic agent monotherapy is observed by 3 to 6 months (5,6).
Initial combination therapy (with or without insulin) may be required in settings of more severe hyperglycemia and/or metabolic decompensation to provide a more rapid and larger decrease in A1C (7–11). Evidence indicates that initial combination of metformin with another agent is associated with an additional mean 0.4% to 1.0% reduction in A1C and a relative 40% higher chance of achieving A1C <7.0% after 6 months compared to metformin alone (7–9,12).
The initial use of combinations of submaximal doses of antihyperglycemic agents produces more rapid and improved glycemic control and fewer side effects compared to monotherapy at maximal doses (13–17).
Table 1
Insulin may be used at diagnosis in individuals with marked hyperglycemia and can also be used temporarily during illness, pregnancy, stress or for a medical procedure or surgery. The use of intensive insulin therapy may lead to partial recovery of beta cell function when used in people with metabolic decompensation, and studies suggest that early insulin treatment may induce remission in people with newly diagnosed type 2 diabetes (28,29–31). Trials of this approach are ongoing.
The natural history of type 2 diabetes is that of ongoing beta cell function decline, so blood glucose (BG) levels often increase over time even with excellent adherence to healthy behaviours and therapeutic regimens (32). Treatment must be responsive as therapeutic requirements may increase with longer duration of disease. If A1C target is not achieved or maintained with current pharmacotherapy, treatment intensification is often required. A review of potential precipitants of increasing A1C (e.g. infection, ischemia) and medication adherence should first be conducted, and current therapy may need to be modified if there are significant barriers to adherence. Dose adjustments and/or additional antihyperglycemic medications should be instituted to achieve A1C target within 3 to 6 months, to avoid clinical inertia and manage ongoing disease progression (33). Healthy behaviour interventions, including nutritional therapy and physical activity, should continue to be optimized while pharmacotherapy is being intensified. Metformin should be continued with other agents unless contraindicated.
In general, when combining antihyperglycemic agents with or without insulin, classes of agents that have different mechanisms of action should be used. Simultaneous use of agents within the same class and/or from different classes but with similar mechanisms of action (e.g. sulfonylureas and meglitinides or DPP-4 inhibitors and GLP-1 receptor agonists) is currently untested, may be less effective at improving glycemia and is not recommended at this time. Table 1 identifies the mechanism of action for all classes of antihyperglycemic agents to aid the reader in avoiding the selection of agents with overlapping mechanisms.
Figure 1
Management of hyperglycemia in type 2 diabetes.
A1C, glycated hemoglobin; CHF, congestive heart failure; CV, cardiovascular; CVD, cardiovascular disease; DKA, diabetic ketoacidosis; eGFR, estimated glomerular filtration rate; HHS, hyperosmolar hyperglycemic state.
In deciding upon which agent to add after metformin, there must be consideration of both short-term effects on glycemic control and long-term effects on clinical complications. Agents with evidence demonstrating the ability to not only lower glucose levels but also reduce the longer-term risk of microvascular and/or CV complications should be prioritized. While intensive glycemic control with a variety of agents is associated with a reduction in microvascular complications (3) and possibly CV complications (34) (see Targets for Glycemic Control chapter, p. S42), Table 1 highlights agent-specific effects on CV or microvascular complications (e.g. CKD) based on trials where glycemic differences between treatment arms were minimized.
The effect of exogenous insulin on the risk of CV complications has been shown to be neutral (35,36). The Outcome Reduction with Initial Glargine Intervention (ORIGIN) trial studied the use of basal insulin titrated to a FBG of <5.3 mmol/L in people at high CV risk with prediabetes or early type 2 diabetes over 6 years. There was a neutral effect on CV outcomes and cancer, and a slight increase in hypoglycemia and weight (36,37).
Earlier trials evaluated effects of thiazolidinediones on CV events. Meta-analyses of smaller studies suggested possible higher risk of myocardial infarction (MI) with rosiglitazone (38,39); however, CV events were not significantly increased in a larger randomized clinical trial (40,41). Conversely, the evidence for pioglitazone suggests a possible reduced risk of CV events, but the primary CV outcome was neutral (42,43). While these agents have comparable glucose-lowering effects to other drugs, the edema, weight gain, risk of congestive heart failure (CHF) (44), increased risk of fractures (45,46) and inconsistent data regarding MI risk with rosiglitazone (38–40) and bladder cancer risk with pioglitazone significantly limit the clinical utility of this drug class (47,48).
Based on controversies regarding rosiglitazone, in 2008, the United States Food and Drug Administration (FDA) required that all new antidiabetic therapies undergo evaluation for CV safety at the time of approval. Subsequently, several industry-sponsored placebo-controlled trials were initiated to evaluate CV outcomes of drugs from 3 newer classes: DPP-4 inhibitors, GLP-1 receptor agonists and SGLT2 inhibitors (see Table 2
Three DPP-4 inhibitor trials have been completed (Table 2). None have shown inferiority or superiority compared to placebo for the risk of major CV events (49,50). Saxagliptin was associated with an increased incidence of hospitalization for heart failure (50) that has yet to be fully explained and, therefore, this agent is not recommended in people with a history of CHF, especially in people who also have renal impairment and/or history of MI. There was a non-statistically significant increase in hospitalizations for CHF with alogliptin in the Examination of Cardiovascular Outcomes with Alogliptin versus Standard of Care (EXAMINE) trial (49) and there is limited experience treating people with a history of CHF with linagliptin; therefore, these agents should be used with caution in that setting. Moreover, a secondary analysis of the data suggested a possibly higher relative risk of unstable angina and all-cause mortality with saxagliptin in those under 65 years (51). The significance of these findings is unclear and further studies are needed. The GLP-1 receptor agonist, lixisenatide, was also shown to be non-inferior to placebo after a median 2.1 years of follow up (52).
Figure 2
Antihyperglycemic medications and renal function. Based on product monograph precautions.
CKD, chronic kidney disease; CV, cardiovascular; GFR, glomerular filtration rate; TZD, thiazolidinedione.
Three approved and one unapproved antihyperglycemic agent, thus far, have shown benefit in reducing major CV outcomes in individuals with clinical CVD, the SGLT2 inhibitors empagliflozin (53) and canagliflozin (54), and the GLP-1 receptor agonists liraglutide (55) and semaglutide (56). The Empagliflozin Cardiovascular Outcome Event Trial (EMPA-REG OUTCOME) included 7,020 people with type 2 diabetes and clinical CVD (defined by ≥1 of the following: MI >2 months prior, multivessel CAD, single-vessel CAD with positive stress test or unstable angina hospitalization in prior year, unstable angina >2 months prior and evidence of CAD, stroke >2 months prior, occlusive peripheral artery disease), most of whom (78%) were already on antihyperglycemic therapy and 82% had diabetes for more than 5 years. Those treated with empagliflozin had significantly fewer CV events (CV death, nonfatal MI, nonfatal stroke) compared to placebo-treated participants after a median 3.1 years follow up (10.5% vs. 12.1%, hazard ratio [HR], 0.86, p<0.001 for noninferiority, p=0.04 for superiority), which was driven by a significant decrease in CV mortality as nonfatal events were not significantly reduced. In a secondary analysis, empagliflozin was associated with a significant reduction in hospitalizations for CHF (4.1 vs. 2.7%, HR 0.65, 95% confidence interval [CI] 0.50–0.85) (53,57). Recent meta-analyses of SGLT2 inhibitors confirmed a significant benefit of this class of agents on major CV outcomes, which was largely driven by EMPA-REG OUTCOME results (58–60).
The CANagliflozin cardioVascular Assessment Study (CANVAS) program, which integrated findings from 2 placebo-controlled trials (CANVAS and CANVAS-R), evaluated the CV effects of canagliflozin (54). The trials enrolled 10,142 participants (4,330 in CANVAS and 5,812 in CANVAS-R) with type 2 diabetes (mean duration 13.5 years), who were aged 30 years or older with symptomatic CVD (symptomatic atherosclerotic vascular disease (coronary, cerebrovascular or peripheral) (66%) or 50 years or older with at least 2 CV risk factors (duration of diabetes ≥10 years, systolic BP >140 mmHg while on ≥1 antihypertensive agent, current smoker, microalbuminuria, macroalbuminuria or HDL cholesterol <1.0 mmol/L) (34%). Over a median follow up of 2.4 years, significantly fewer persons randomized to canagliflozin than placebo had the primary outcome of CV death, nonfatal MI or nonfatal stroke (26.9 vs. 31.5 per 1,000 person-years respectively; HR 0.86, 95% CI 0.75–0.97, p<0.001 for noninferiority and p=0.02 for superiority). There were no statistical differences in the individual components of the composite outcome. There was a reduction in hospitalization for heart failure and in several adverse renal outcomes; however, these were considered exploratory outcomes due to pre-specified rules of evidence hierarchy. While one-third of participants did not have CVD, a significant decrease in the primary endpoint was only found in those with CVD. Therefore, as with other CV outcome trials, these results largely apply to people with type 2 diabetes requiring add-on antihyperglycemic therapy who have established clinical CVD. Canagliflozin was also associated with an increase in fracture rates (HR 1.26, 95% CI 1.04–1.52), and higher rates of genital infections and volume depletion. Importantly, canagliflozin was associated with doubling in the risk of lower extremity amputation (HR 1.97, 95% CI 1.41–2.75). This risk was strongest in participants with a prior amputation. Canagliflozin should, therefore, be avoided in people with a prior amputation, as the harms appear to be greater than the benefits in that population.
The Liraglutide Effect and Action in Diabetes: Evaluation of Cardiovascular Outcome Results (LEADER) trial enrolled 9,340 participants with longstanding type 2 diabetes (median duration 12.8 years) and 88% were on antihyperglycemic therapy at baseline (55). The majority of included participants (81%) were ≥50 years of age on pre-existing antihyperglycemic therapy with at least 1 CV condition (coronary heart disease [CHD], cerebrovascular disease, peripheral arterial disease, CHF or stage 3 or higher CKD). Over a median follow up of 3.8 years, fewer participants in the liraglutide arm compared to placebo had the primary endpoint of CV death, nonfatal MI or nonfatal stroke (13% vs. 14.9%, respectively; HR 0.87, 95% CI 0.78–0.97), fulfilling the statistical criteria for both noninferiority (p<0.001) and superiority (p=0.01). While the LEADER trial included some people with CV risk factors only, over 80% of participants had cardiovascular disease (CVD) and only 10.5% of the primary events occurred in those without clinical disease. Therefore results are most applicable to people with type 2 diabetes with clinical CVD requiring add-on antihyperglycemic therapy.
The Trial to Evaluate Cardiovascular and Other Long-term Outcomes with Semaglutide in Subjects with Type 2 Diabetes (SUSTAIN-6) enrolled 3,297 participants with a mean duration of type 2 diabetes of 13.9 years (56). At baseline, 98% were on antihyperglycemic therapy and 83% had established CVD or stage 3 or higher CKD. After a median follow up of 2.1 years, the primary composite outcome of CV death, nonfatal MI or nonfatal stroke occurred in 6.6% of participants treated with semaglutide and 8.9% of participants treated with placebo (HR 0.74, 95% CI 0.58–0.95), fulfilling statistical criteria for noninferiority (p<0.001); a non- pre-specified test for superiority was also significant (p=0.02). There was, however, a higher rate of diabetic retinopathy complications in the semaglutide group compared to placebo group (3.0% vs. 1.8%, HR 1.76; 95% CI 1.11–2.78; p=0.02). It is unclear at this time if there is a direct effect of semaglutide or other explanations for this unexpected difference in retinopathy complication rates, although the risk appeared greatest in individuals with pre-existing retinopathy and rapid lowering of A1C.
All 4 trials reported lower rates of kidney disease progression in the treated groups compared to placebo (53,55,56). It should also be noted that the majority of people in these trials had pre-existing CVD and required add-on antihyperglycemic therapy. In addition, because these were placebo-controlled trials, no conclusions can be made about how the cardioprotective properties of empagliflozin, canagliflozin, liraglutide and semaglutide compare to those of other agents. CV outcome trials for other agents are expected to be completed by 2019; therefore, based on evidence to date, a GLP-1 receptor agonist or SGLT2 inhibitor with demonstrated CV outcome benefit should be considered as initial add-on therapy for people with pre-existing type 2 diabetes and clinical CV disease who have not achieved target A1C on existing treatment to reduce CV risk.
A careful review of the methods and findings from these trials was conducted by an independent committee. While primary analyses results were similar for canagliflozin, empagliflozin and liraglutide, it was concluded that the strength of evidence for CV benefit was weaker for canagliflozin than for the other agents. This conclusion was based on three factors. First, in 2012 an interim analysis of the CANVAS study for medication approval necessitated unblinding of study data. A decision was then made to combine this study with the CANVAS-R study, presumably to provide greater power for CV outcomes. The interim unblinding and protocol revision were viewed as potential threats to internal validity, thereby weakening the strength of evidence for benefit. Second, while canagliflozin was associated with a significant decrease in the composite MACE outcome, there was no significant benefit on individual outcomes, such as all-cause or CV mortality. Third, the findings of increased risk of fractures and amputations with canagliflozin treatment in the context of a noninferiority design where the comparator is placebo was particularly concerning, indicating that harms may outweigh benefits. For these reasons, the committee decided that the uncertainty regarding benefits should be acknowledged with a lower grade of recommendation for canagliflozin than for other agents with demonstrated CV benefit.
Table 2 Major clinical outcome trial characteristics for antihyperglycemic agents |
||||||||||
---|---|---|---|---|---|---|---|---|---|---|
2xScr, doubling of serum creatinine; ≥50%⇓eGFR, minimum 50% decline in estimated glomerular filtration rate; CV Death, death from cardiovascular causes; ESRD, end stage renal disease; HF hosp, hospitalization for heart failure; LL Amp, lower limb amputation; MACE, major adverse cardiovascular event (cardiovascular death, nonfatal myocardial infarction or nonfatal stroke); MACE + UA, MACE plus hospitalization for unstable angina; Prog Alb, progression of albuminuria; UL, upper limit of 95% confidence interval. |
||||||||||
A1C (%) | ||||||||||
Study | Clinicaltrials.gov | Agent (Dose) (n) | Age (yrs) | Men | DM (yrs) | Start | End | Follow up (yrs) | Completed | Results* |
Dipeptidyl Peptidase-4 Inhibitors | ||||||||||
EXAMINE (49,145) | NCT00968708 | Alogliptin (25 or 12.5 mg) (n=2,701) | 61.0† | 68% | 7.1† | 8.0 (±1.1) | -0.33 | 1.5† | n=2,692 (99%)‡ | MACE: 0.96 (UL 1.16) |
Placebo (n=2,679) | 7.3† | +0.03 | n=2,663 (99%)‡ | HF hosp: 1.07 (0.79–1.46) | ||||||
CARMELINA§ | NCT01897532 | Linagliptin (5 mg) (n=4,150 estimated) | Estimated completion in 2018 | MACE + UA | ||||||
Placebo (n=4,150 estimated) | ||||||||||
CAROLINA (143) | NCT01243424 | Linagliptin (5 mg) (n=unknown) | 64 | 60% | 6.2 | 7.2 | Estimated completion in 2019 | MACE + UA | ||
Glimepiride (1–4 mg) (n=unknown) (total enrolled n=6,051) | ||||||||||
SAVOR-TIMI 53 (50) | NCT01107886 | Saxagliptin (5 or 2.5 mg) (n=8,280) | 65.1 | 67% | 10.3† | 8.0 (±1.4) | 7.7 | 2.1† | n=8,078 (97%) | MACE: 1.00 (0.89–1.12) |
Placebo (n=8,212) | 65.0 | 7.9 | n=7,998 (97%) | HF hosp: 1.27 (1.07–1.51) | ||||||
TECOS (144) | NCT00790205 | Sitagliptin (100 or 50 mg) (n=7,332) | 65.4 | 71% | 11.6 | 7.2 (±0.5) | 0.29 lower than placebo | 3.0† | n=6,972 (95%) | MACE + UA: 0.98 (0.88–1.09) |
Placebo (n=7,339) | 65.5 | 70% | n=6,905 (94%) | HF hosp: 1.00 (0.83–1.20) | ||||||
GLP-1 receptor agonists | ||||||||||
HARMONY Outcomes§ | NCT02465515 | Albiglutide (30 or 50 mg) (n=unknown) estimated enrolment 9,400 | Estimated completion in 2018 | MACE | ||||||
Placebo (n=unknown) | ||||||||||
REWIND§ | NCT01394952 | Dulaglutide(1.5 mg) (n=unknown) total enrolled n=9,622 | Estimated completion in 2018 | MACE | ||||||
Placebo (n=unknown) | ||||||||||
EXSCEL (146) | NCT01144338 | Exenatide (2 mg) (n=7,356) | 60.2† | 62% | 12.0† | 8.0† | 0.53 lower than placebo | 3.8† | n=7,094 (96%) | MACE: 0.91 (0.83–1.00) CV death: 0.88 (0.76–1.02) |
Placebo (n=7,396) | 60.2† | 62% | n=7,093 (96%) | HF hosp: 0.94 (0.78–1.13) | ||||||
FREEDOM-CVO§ | NCT01455896 | ITCA 650 (Exenatide in DUROS) (60 µg) (n=unknown) estimated enrolment n=4,000 | Study completed April 2016 | MACE + UA: results not released yet | ||||||
Placebo (n=unknown) | ||||||||||
LEADER (55) | NCT01179048 | Liraglutide (1.8 mg) (n=4,668) | 64.2 | 65% | 12.8 | 8.7 (±1.5) | 0.40 lower than placebo | 3.8† | n=4,529 (97%) | MACE: 0.87 (0.78–0.97) CV death: 0.78 (0.66–0.93) |
Placebo (n=4,672) | 64.4 | 64% | n=4,513 (97%) | HF hosp: 0.87 (0.73–1.05) | ||||||
ELIXA (52) | NCT01147250 | Lixisenatide (20 µg) (n=3,034) | 59.9 | 70% | 9.2 | 7.7 | 0.27 lower than placebo | 2.1† | n=2,922 (96%) | MACE + UA: 1.02 (0.89–1.17) |
Placebo (n=3,034) | 60.6 | 69% | 9.4 | 7.6 | n=2,916 (96%) | HF hosp: 0.96 (0.75–1.23) | ||||
PIONEER 6§ | NCT02692716 | Semaglutide (not stated) (unknown) estimated enrolment n=3,176 | Estimated completion in 2018 | MACE | ||||||
Placebo (n=unknown) | ||||||||||
SUSTAIN 6 (56) | NCT01720446 | Semaglutide (0.5 mg) (n=826) | 64.6 | 60% | 14.3 | 8.7 | -1.1 | 2.1† | 1,623 (99%) | MACE: 0.74 (0.58–0.95) |
Semaglutide (1.0 mg) (n=822) | 64.7 | 63% | 14.1 | 8.7 | -1.4 | HF hosp: 1.11 (0.77–1.61) | ||||
Placebo (n=1,649) | 64.6 | 60% | 13.6 | 8.7 | -0.4 | n=1,609 (98%) | Retinopathy: 1.76 (1.11–2.78) | |||
Sodium-glucose co-transporter-2 inhibitors | ||||||||||
CANVAS (54) | NCT01032629 | Canagliflozin (100 mg) (n=1,445) | 62.4 | 66% | 13.4 | 8.2 (±0.9) | 5.7 | MACE: 0.88 (0.75–1.03) | ||
Canagliflozin (300 mg) (1,444) | HF hosp: 0.77 (0.55–1.08) | |||||||||
Placebo (1,444) | ||||||||||
CANVAS-R (54) | NCT01989754 | Canagliflozin (300 mg) (n=2,907) | 64.0 | 63% | 13.7 | 8.3 (±1.0) | 2.1 | Prog Alb: 0.64 (0.57–0.73) | ||
Placebo (n=2,905) | MACE: 0.82 (0.66–1.01) | |||||||||
CANVAS Program (54) | Canagliflozin (100 or 300 mg) (n=5,795) | 63.2 | 65% | 13.5 | 8.2 (±0.9) | 0.58 lower than placebo | 3.6 | n=9,734 (96%) | MACE: 0.86 (0.75–0.97) | |
Placebo (n=4,347) | 63.4 | 63% | 13.7 | 8.2 (±0.9) | Prog Alb: 0.73 (0.67–0.79) HF hosp: 0.67 (0.52–0.87) LL amp: 1.97 (1.41–2.75) |
|||||
CREDENCE§ | NCT02065791 | Canagliflozin (100 mg) (n=unknown) estimated enrolment n=4,200 | Estimated completion in 2019 | ESRD, 2xSCr, renal or CV death | ||||||
Placebo (unknown) | MACE + HF + UA | |||||||||
Dapa-CKD§ | NCT03036150 | Dapagliflozin (5 or 10 mg) (n=unknown) estimated enrollment n=4,000 | Estimated completion in 2020 | ≥50% ⇓ eGFR, ESRD, renal or CV death | ||||||
Placebo (n=unknown) | ||||||||||
Dapa-HF§ | NCT03036124 | Dapagliflozin (5 or 10 mg) (n=unknown) estimated enrolment n=4,500 | Estimated completion in 2019 | CV death or HF hosp | ||||||
Placebo (n=unknown) | ||||||||||
DECLARE-TIMI 58§ | NCT01730534 | Dapagliflozin (10 mg) (n=unknown) total enrolled n=17,276 | Estimated completion in 2019 | MACE | ||||||
Placebo (n=unknown) | ||||||||||
EMPA-REG Outcome (53,57) | NCT01131676 | Empagliflozin 10 mg (n=2,345) | 63.0 | 71% | 57% had diabetes >10 yrs | ~8.0 | 0.24 lower | 3.1† | n=2,264 (97%) | MACE: 0.86 (0.74–0.99) CV death: 0.62 (0.49–0.77) |
Empagliflozin (25 mg) (n=2,342) | 63.2 | 72% | 0.36 lower | n=2,279 (97%) | HF hosp: 0.65 (0.50–0.85) | |||||
Placebo (n=2,333) | 63.2 | 72% | @206 wks | n=2,266 (97%) | ||||||
EMPEROR-Preserved§ | NCT03057951 | Empagliflozin (10 mg) (n=unknown) estimated enrollment n=4,126 | Estimated completion in 2020 | CV death or HF hosp | ||||||
Placebo (n=unknown) | ||||||||||
EMPEROR-Reduced§ | NCT03057977 | Empagliflozin (not stated) (n=unknown) estimated enrolment n=2,850 | Estimated completion in 2020 | CV death or HF hosp | ||||||
Placebo (n=unknown) | ||||||||||
VERTIS CV§ | NCT01986881 | Ertugliflozin (15 mg) (n=4,000 estimated) | Estimated completion in 2019 | MACE | ||||||
Placebo (n=4,000 estimated) |
In the absence of evidence for long-term clinical benefit, agents effective at A1C lowering should be considered in terms of both the degree of baseline hyperglycemia needing correction, and any heightened concerns regarding hypoglycemia (e.g. elderly people or those with renal or hepatic dysfunction) (see Diabetes in Older People chapter, p. S283). While most medications added to metformin lower A1C to a similar extent, insulin and insulin secretagogues are associated with higher rates of hypoglycemia than other agents (21,23,24,61). Insulin treatment is recommended for people with metabolic decompensation and/or symptomatic hyperglycemia. In those who are stable, other agent-specific advantages and disadvantages should be weighed as treatment is individualized to best suit the patient's needs and preferences. Each of the agents listed in Table 1 and Figure 1 has advantages and disadvantages to consider. Figure 2 illustrates the basis on which agent selection is influenced by renal function as dictated by product monograph precautions.
Recent meta-analyses have summarized head-to-head comparisons of metformin-based combinations (19,24,62,63). Combinations of metformin with a sulfonylurea, a thiazolidinedione (TZD), an SGLT2 inhibitor and a DPP-4 inhibitor have comparable A1C-lowering effects (19,24,62–66), while the combination of metformin with a GLP-1 receptor agonist reduced A1C more than combination with a DPP-4 inhibitor. TZDs, insulin and sulfonylureas are associated with the most weight gain (1.5 to 5.0 kg) when added to metformin, whereas GLP-1 receptor agonists and SGLT2 inhibitors are associated with weight loss. Hypoglycemia risk is also lower with TZDs, DPP-4 inhibitors, SGLT2 inhibitors and GLP-1 receptor agonists compared to sulfonylureas and insulin (19,24,62–65,67,68). Network meta-analyses that indirectly compared the net benefits of second- and third-line treatment options have found similar results (21,23,24,69–71). Evidence on comparative effectiveness of acarbose and orlistat is limited, although they are associated with a low risk of hypoglycemia and weight gain. Based on these findings, people on metformin monotherapy requiring treatment intensification and their providers may prefer an incretin agent (DPP-4 inhibitor or GLP-1 receptor agonist), and/or SGLT2 inhibitor to other agents if there are no contraindications and affordability and access are not barriers, as they will improve glycemic control with a low risk of hypoglycemia and weight gain. These agents should be considered before an insulin secretagogue (sulfonylurea or meglitinide) or insulin as add-on therapy in people with a high risk of hypoglycemia (such as elderly people or those with impaired renal function) and/or obesity. The safety of incretin agents, SGLT2 inhibitors and TZDs in pregnancy is unknown; therefore, these agents should be avoided or discontinued in women who are pregnant or planning a pregnancy (see Diabetes and Pregnancy chapter, p. S255).
If a sulfonylurea is added to metformin, gliclazide should be considered as first choice as it is associated with a lower risk of hypoglycemia (67,72), CV events and mortality relative to other sulfonylureas (73). Glimepiride is also associated with a lower risk of CV events and mortality (73), but has a similar rate of hypoglycemia (67,72) compared to other sulfonylureas.
For people already taking metformin and a sulfonylurea, the addition of either a DPP-4 inhibitor, a GLP-1 receptor agonist or SGLT2 inhibitor may be considered as they are associated with effective A1C lowering with less hypoglycemia than insulin or TZDs (21,69,70,74,75); GLP-1 receptor agonists and SGLT2 inhibitors are also associated with weight loss (70,71) (see Weight Management in Diabetes chapter, p. S124). Concurrent addition of 2 antihyperglycemic agents (+/- insulin) to metformin therapy may be considered in settings of more severe hyperglycemia. For instance, the combination of a DPP-4 inhibitor or a GLP-1 receptor agonist and an SGLT2 inhibitor added to metformin has been shown to be as safe and more efficacious at lowering A1C after 24 weeks than either agent alone (76,77).
SGLT2 inhibitors and GLP-1 receptor agonists added to metformin have also been shown to reduce systolic BP compared to metformin alone, and add-on of SGLT2 inhibitors reduce systolic BP more than add-on of sulfonylureas or DPP-4 inhibitors (19).
A combination of noninsulin antihyperglycemic agents and insulin often effectively controls glucose levels. Insulin treatment includes long-acting or intermediate-acting insulin analogue injections once or twice daily for basal glycemic control, and bolus injections at mealtimes for prandial glycemic control. Adding insulin to noninsulin antihyperglycemic agent(s) may result in better glycemic control with a smaller dose of insulin (78), and may induce less weight gain and less hypoglycemia than that seen when non-insulin antihyperglycemic agents are stopped and insulin is used alone (79,80). A single injection of an intermediate-acting (NPH) (81) or long-acting insulin analogue (insulin glargine U-100, insulin glargine U-300, insulin detemir or insulin degludec) (82–84) may be added. The addition of bedtime insulin to metformin therapy leads to less weight gain than insulin plus a sulfonylurea or twice-daily NPH insulin (85). When insulin is used in type 2 diabetes, the insulin regimen should be tailored to achieve good metabolic control while trying to avoid hypoglycemia. With intensive glycemic control, there is an increased risk of hypoglycemia, but this risk is lower in people with type 2 diabetes than in those with type 1 diabetes. The mode of insulin administration (continuous subcutaneous infusion vs. injections), the number of insulin injections (1 to 4 per day) and the timing of injections may vary depending on each individual's situation (86).
As type 2 diabetes progresses, insulin requirements will likely increase and higher doses of basal insulin (intermediate-acting or long-acting analogues) may be needed. DPP-4 inhibitors, GLP-1 receptor agonists and SGLT2 inhibitors have been shown to be efficacious at further lowering glucose levels when combined with insulin therapy (87–98). A meta-analysis determined that the addition of a GLP-1 receptor agonist to basal insulin regimens results in greater A1C reduction, more weight loss and less hypoglycemia compared to the addition of bolus insulin (99). A GLP-1 receptor agonist should, therefore, be considered before bolus insulin as add-on therapy in people on basal insulin (with or without other agents) who require antihyperglycemic treatment intensification if there are not barriers to affordability or access.
If glycemic control is suboptimal on treatment regimens that include basal insulin with other agents, bolus insulin at mealtimes (short- or rapid-acting analogues) may be added. Generally, once bolus insulin is introduced into a treatment regimen, either as a separate mealtime bolus or as part of a premixed containing regimen, insulin secretagogues, such as sulfonylureas and meglitinides, should be discontinued. Concomitant therapy with metformin and, if applicable, a GLP-1 receptor agonist, DPP-4 inhibitor or SGLT2 inhibitor should be continued with regimens containing bolus insulin unless contraindicated, to allow for improved glycemic control with less risk of weight gain and hypoglycemia (100).
The reduction in A1C achieved with insulin therapy depends on the dose and number of injections per day (101). A meta-analysis of 12 articles compared basal-bolus and biphasic insulin regimens, and found that both approaches are equally efficacious at lowering A1C, with comparable effects on hypoglycemia risk and weight—although basal-bolus regimens were modestly more efficacious in people with type 2 diabetes already on insulin (102). Bolus insulin should be initiated using a stepwise approach (starting with 1 injection at the largest meal and additional mealtime injections at 3-month intervals if needed), as it was shown to be as efficacious at A1C lowering as a full basal-bolus regimen, and is associated with less hypoglycemia and greater patient satisfaction after 1 year (103).
Lower rates of hypoglycemia have been observed in some studies of individuals with type 2 diabetes treated with rapid-acting insulin analogues (insulin aspart, insulin lispro, insulin glulisine) compared to those treated with short-acting (regular) insulin (104–106). Use of long-acting basal insulin analogues (insulin detemir, insulin glargine, insulin degludec) in those already on antihyperglycemic agents reduces the relative risk of symptomatic and nocturnal hypoglycemia compared to treatment with NPH insulin (83,104,107–112). Meta-analyses indicate a relative reduction of 0.89 (95% Cl 0.83–0.96) and 0.63 (95% Cl 0.51–0.77) for symptomatic and nocturnal hypoglycemia respectively (112); and rates of 26% vs. 34% and 13% vs. 22% for at least one symptomatic and nocturnal hypoglycemic event with an analogue vs. NPH (111). Insulin degludec has been associated with lower rates of overall and nocturnal hypoglycemia compared to glargine U-100 (82,84,113). The Randomised, Double Blind, Cross-over Trial Comparing the Safety and Efficacy of Insulin Degludec and Insulin Glargine, With or Without OADs in Subjects With Type 2 Diabetes (SWITCH 2) trial randomized patients with type 2 diabetes and at least one risk factor for hypoglycemia (history of hypoglycemia, >5 years of insulin therapy, hypoglycemia unawareness or moderate chronic renal failure) to insulin degludec or glargine U-100. After 32 weeks of treatment, insulin degludec was associated with a significantly lower rate of the primary endpoint of overall symptomatic hypoglycemic episodes (rate ratio 0.70, 95% CI 0.61–0.80). The proportions of patients with hypoglycemic episodes were 9.7% and 14.7% for insulin degludec and glargine U-100, respectively (114). The Trial Comparing Cardiovascular Safety of Insulin Degludec versus Insulin Glargine in Patients with Type 2 Diabetes at High Risk of Cardiovascular Events (DEVOTE) randomized patients with type 2 diabetes at high risk of CV disease to insulin degludec or glargine U-100, and found no difference in the primary outcome of CV events but a significant decrease in severe hypoglycemia with degludec (4.6%) compared to glargine U-100 (6.6%; odds ratio, 0.73; p<0.001 for superiority) (84). Insulin degludec may thus be considered over glargine U-100 in patients at high risk of hypoglycemia and/or CV disease. There is also some evidence of lower hypoglycemia rates with glargine U-300 compared to glargine U-100 (115) and may also be considered over glargine U-100 if reducing hypoglycemia is a priority (116). Efficacy and rates of hypoglycemia are similar between glargine U-100 and detemir (117).
Aside from effects of some antihyperglycemic agents on the occurrence of hypoglycemia and weight, there are adverse effects unique to each agent (Table 1). Gastrointestinal side effects are more common with metformin, alpha glucosidase inhibitors, GLP-1 receptor agonists and orlistat than with other agents. Metformin can cause diarrhea, which tends to resolve over time and is minimized with starting at a low dose and subsequent slow titration of the dosage. Extended-release metformin can also be used to improve tolerability in individuals experiencing gastrointestinal side effects with immediate-release metformin (118–121). Metformin is also associated with an approximate 2-fold increased incidence of vitamin B12 deficiency (122–124), and vitamin B12 levels should be measured periodically in people taking metformin or with signs or symptoms of deficiency (such as impaired proprioception or peripheral neuropathy). GLP-1 receptor agonists and, less commonly, DPP-4 inhibitors can cause nausea and GLP-1 receptor agonists can also cause diarrhea. A meta-analysis comparing the risk of congestive heart failure between antihyperglycemic therapies found an increased risk with TZDs and DPP-4 inhibitors (driven by higher risk with saxagliptin) (44), although another meta-analysis (125) and a large observational study of over one million participants (126) failed to find an increased risk of heart failure with DPP-4 inhibitors compared to other agents. TZDs are also associated with a 47% increased risk of fractures compared to other agents that is predominantly seen in women (127). Reports of acute pancreatitis have been noted with DPP-4 inhibitors and GLP-1 receptor agonists. A small significant increase in pancreatitis but not pancreatic cancer was seen with DPP4-inhibitors in a meta-analysis of 3 large randomized controlled trials of over 20,000 participants (128). However, a recent large Canadian observational study of over 1.5 million people did not confirm a higher risk of pancreatitis with incretin-based therapies compared to other agents (129). SGLT2 inhibitors are associated with a 3- to 4-fold increased risk of genital mycotic infections (19,69,95), as well as higher rates of urinary tract infections, volume depletion, rare acute kidney injury and rare DKA (130,131). Canagliflozin treatment is associated with an increased risk of fractures (54,132) and a twofold increased risk of amputations (54). In a retrospective analysis, empagliflozin was not associated with an increased risk of amputations in the EMPA-REG trial (133). There is evidence of a higher risk of bladder cancer with pioglitazone in some studies (47,48) but not others (134–136), and some reports of increased bladder cancer risk with dapagliflozin (137). GLP-1 receptor agonists have been shown to promote the development of pancreatic and medullary thyroid cancer in rodents, but an increased risk has not been seen in humans (138). Semaglutide was associated with a higher risk of retinopathy in SUSTAIN-6 (see above) (56). Earlier epidemiological evidence suggesting a possible link between insulin glargine and cancer has not been substantiated in review of clinical trial data for either glargine or detemir (36,139,140).
Insulin glargine U-300 may be considered over insulin glargine U-100 to reduce overall and nocturnal hypoglycemia [Grade C, Level 3 (116)].
Abbreviations
A1C, glycated hemoglobin; BG, blood glucose; BP, blood pressure; CHF, congestive heart failure; CHD, coronary heart disease; CI, confidence interval; CV, cardiovascular; CVD, cardiovascular disease; DKA, diabetic ketoacidosis; HR, hazard ratio; MI; myocardial infarct; NPH, neutral protamine Hagedorn; TZD, thiazolidinedione.
Literature Review Flow Diagram for Chapter 13: Pharmacologic Glycemic Management of Type 2 Diabetes in Adults
*Excluded based on: population, intervention/exposure, comparator/control or study design.
From: Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(6): e1000097. doi:10.1371/journal.pmed1000097 (147).
For more information, visit www.prisma-statement.org.
Dr. Goldenberg reports personal fees from Abbott, AstraZeneca, Boehringer Ingelheim, Eli Lilly, Janssen, Merck, Novo Nordisk, Sanofi, and Servier, outside the submitted work. Dr. MacCallum reports personal fees from Janssen and Novo Nordisk, outside the submitted work. No other author has anything to disclose.
*The Canadian Diabetes Association is the registered owner of the name Diabetes Canada. All content on guidelines.diabetes.ca, CPG Apps and in our online store remains exactly the same. For questions, contact communications@diabetes.ca.