1. Sulfonylurea
Drugs belonging to the sulfonylurea (SU) class are characterized by the stimulation of pancreatic β-cells to release endogenous insulin by closing ATP-sensitive potassium channels. Due to this mechanism of action stimulating endogenous insulins secretion, these agents are considered to be unsuitable for the use as add-on therapy in T1D. However extrapancreatic effects, hypothetically increasing the number of cellular insulin receptors, were speculated
[1].
Glibenclamide has been the only SU which was investigated in T1D and the studies were all of a small sample size
[2][3][4][5][6] with the largest trial including 40 participants
[5]. While some positive effects on HbA
1c, in particular in people with preserved insulin secretion, were observed, with a neutral effect on hypoglycemia, research in this field has petered out in the last decades and SUs do not play a role in people with T1D.
2. Metformin
Metformin remains the most widely used glucose-lowering agent in persons with T2D globally. Most T2D treatment guidelines recommend metformin as a first-line pharmacotherapy
[7]. The main arguments in favor of this drug are its low cost, positive safety profile, its efficacy in lowering glucose levels, its reported durability, and the positive cardiovascular data mainly derived from the UKPDS study
[8]. The glucose-lowering effect of metformin is primarily attributed to a decreased hepatic gluconeogenesis and increased insulin-stimulated peripheral glucose uptake, mainly in the skeletal muscle, by inhibition of the mitochondrial respiratory chain and activation of AMP-activated protein kinase
[9].
Metformin in T1D: In the context of metformin used in people with T1D, a recent meta-analysis summarizing eight placebo-controlled trials on the application of metformin in T1D provided evidence of significant weight loss (−2.4 kg; (95% CI −4.2 to −0.65)), a total daily insulin reduction (−1.36 IU; (−2.3 to −0.45)), and improvements in LDL cholesterol in the metformin arm. HbA
1c and fasting plasma glucose as well as the risk of hypoglycemia and DKA did not differ in the various groups. The metformin groups had a greater likelihood of experiencing adverse gastrointestinal effects and a higher rate of adverse event-related study discontinuations when compared to a placebo
[10].
While the majority of the studies included in the above-mentioned meta-analysis were of short duration, a 1-year intervention-based placebo-controlled trial of 100 individuals with poorly controlled T1D (HbA
1c ≥ 8.5%) revealed no benefits on glucose control, but similar advantages in terms of reducing body weight and the total daily insulin dose. Notably, severe hypoglycemia accompanied by the loss of consciousness occurred more frequently in the metformin group than in the placebo group (10 versus 2 events;
p < 0.05). Nearly half of the individuals allocated to the intervention arm reported gastrointestinal side effects, however, not significantly more than in the placebo group
[11]. The same study reported improvements in levels of proatherogenic lipid profiles, irrespective of previous background statin therapy
[12]. Evidence from a small randomized-controlled study also suggests fewer glycemic fluctuations in metformin users with T1D
[13].
A retrospective real-world study addressed the characteristics of metformin users and the effects of metformin on glycemic control during a 10-year follow-up as compared to a population of metformin-naïve patients. At baseline, the metformin group had a higher BMI and required higher insulin doses but had similar HbA
1c levels as metformin-naïve patients. Reductions in BMI and insulin doses were noted in the first few years, but the 10-year observation period revealed no sustained effect on HbA
1c, insulin doses, or BMI
[14].
The multicenter, placebo-controlled REMOVAL trial investigated the effect of metformin in a high-risk population of more than 400 persons with T1D, ≥40 years of age, and with three or more cardiovascular risk factors. While intermittent positive effects on HbA
1c and insulin dose were observed during the first few months of the study, these parameters did not differ from the placebo group at the end of the study after three years of metformin therapy. However, body weight and LDL cholesterol remained significantly lower three years after the initiation of the study. The progression of mean carotid intima media thickness was not significantly reduced with metformin and endothelial function as measured by reactive hyperemia index was not beneficially modified when compared to the placebo. Concerning microvascular complications, the REMOVAL program showed no difference in the progression of existing retinopathy or the development of new-onset retinopathy. Notably, metformin users experienced a significant increase in their glomerular filtration rates. However, data on albuminuria were not available and any substantial positive effect of metformin on microvascular disease in T1D remains speculative
[15]. Notably, in a recently published subgroup analysis of the REMOVAL trial, the progress in cIMT differed according to the smoking status of the participants. Never-smokers who took metformin had significantly slower progress in cIMT than those who used placebo, while previous or current smokers did show a similar extent of cIMT progression as placebo
[16]. This finding underlines the detrimental effect of smoking as a risk factor for atherosclerosis in T1D. It might also give rise to an individualized treatment approach: metformin might hinder the progression of atherosclerosis only in non-smokers with T1D. REMOVAL showed gastrointestinal side effects, compelling the premature discontinuation of the trial in 27% of the intervention group (and 12% of the placebo group). Besides, a vitamin B12 deficiency was noted more frequently in persons on metformin therapy (12 vs. 5% over three years), which might be particularly important in persons with T1D, as pernicious anemia, celiac disease, and diabetic gastroparesis are well-known accompanying conditions in T1D. The authors concluded that metformin does not improve glycemic control in long-standing T1D but does appear to play a role in cardiovascular risk management
[15].
3. Thiazolidinediones (TZD)
TZDs or glitazones are selective agonists of the peroxisome proliferator-activated receptor-gamma (PPARϒ), a nuclear receptor that alters the transcription of several genes involved in glucose and lipid metabolism. As an insulin sensitizer, TZDs reduce insulin resistance in adipose tissue, liver, and muscle cells by promoting endocrine signaling from adipocytes
[17]. In people with T2D and/or metabolic syndrome, TZDs were shown to improve insulin sensitivity, reduce plasma-free fatty acids and triglycerides, and increase high-density cholesterol (HDL-C). Furthermore, TZDs have been associated with positive effects on endothelial function, inflammation, and the procoagulatory state. Last but not least, TZDs improve both fasting and postprandial hyperglycemia, reduce HbA
1c, and have a sustainable positive effect on glycemia (good durability) in T2D
[18]. TZDs might have beneficial effects in preserving β-cell function in T2D, as suggested by the use of rosiglitazone (DREAM trial) and pioglitazone (ACT NOW trial), and thus slowing down the progression of prediabetes into diabetes
[19][20]. While in T2DM some positive effects on cardiovascular endpoints have been attributed to the use of pioglitazone (PROactive study), the incidence of hospitalization for heart failure was increased
[21]. Rosiglitazone has already been withdrawn from the market.
To date, only a handful of randomized controlled trials have addressed the effects of TZDs as add-on therapy to insulin in individuals with T1D.
Pioglitazone in T1D: In a randomized controlled trial including 60 peripubertal and lean adolescents with T1D who were diagnosed more than one year previously, treatment with 30 mg of pioglitazone once daily for 6 months led to a significant reduction of HbA
1c (−0.22 ± 0.29%) and improvements in postprandial plasma glucose in the intervention group. The number of participants who achieved an HbA
1c level ≤7% increased from 53 to 70% during pioglitazone treatment while it did not beneficially change in the placebo group. No differences were seen in body weight, hypoglycemic events, insulin requirements, or lipid profiles
[22]. In a small, randomized placebo-controlled pilot study comprising newly diagnosed children and adolescents (n = 15) with T1D, the authors explored the effect of once-daily pioglitazone on β-cell preservation by adding pioglitazone to insulin. After 24 weeks of treatment, pioglitazone did not preserve endogenous insulin secretion; insulin requirements and glycemic control were also not improved by pioglitazone therapy
[23]. It seems there is no trial has addressed the impact of TZDs on vascular protection in populations with T1D.
Another RCT examining 35 adolescents with suboptimally controlled T1D (HbA
1c 7.5–11%) and features of insulin resistance (insulin requirements >0.9 IU/kg/d) showed no improvement in glycemic control when pioglitazone was administered for six months and compared to a placebo. In fact, the intervention group experienced an increase in BMI when compared to placebo. Lipid parameters were similar in both groups
[24].
Rosiglitazone in T1D: Rosiglitazone given twice daily for 8 months in adults with T1D and a BMI ≥ 27 kg/m
2 resulted in improvements in HbA
1c. A higher proportion of study participants randomized to rosiglitazone achieved an HbA
1c level <6.5% (36 vs. 16%), while the prevalence of hypoglycemia was similar. When compared to baseline levels, weight gain was significantly evident in both groups (rosiglitazone and placebo)
[25].
4. Alpha-Glucosidase Inhibitors (AGIs)
AGIs (acarbose and miglitol) exert their glucose-lowering effect by inhibiting the absorption of carbohydrates from the small intestines, and thus mainly improve postprandial hyperglycemia. AGIs are associated with a modest decrease in body weight, a low risk of hypoglycemia, and positive effects on blood pressure and lipid profiles in people with T2D. While some studies conducted in T2D suggested a risk reduction for myocardial infarction or any cardiovascular event after treatment with acarbose over three years
[26][27], the properly powered ACE trial, which was conducted in an Asian population with IGT and coronary artery disease, did not reveal such cardiovascular benefits
[28]. AGIs are not included in most of the current guidelines for the treatment of T2D
[7].
To date, four randomized placebo-controlled trials (three for acarbose and one for miglitol) have assessed the efficacy and safety of AGIs for use in T1D
[29].
Acarbose in T1D: The most important placebo-controlled double-blind intervention study, in which acarbose was administered to 114 adults with T1D, showed a significant reduction in HbA
1c (−0.48%; placebo-subtracted), as well as fasting and postprandial hyperglycemia, but with no significant effect on insulin dose, risk of hypoglycemia, body weight, or lipid levels during 24 weeks of treatment. Eighty-four percent of the participants randomized to the acarbose group reported side effects, which were mainly confined to the gastrointestinal tract (49% for placebo)
[30]. With a number needed to harm of 13, acarbose was the agent with the highest side-effect-related study dropout rates when compared to other glucose-lowering agents (metformin, SGLT2-i, GLP1-RAs, SU, DPP4 inhibitors) used in T1D as shown in a meta-analysis
[29].
Miglitol in T1D: The only study which investigated miglitol showed a less pronounced concern regarding gastrointestinal side effects, but failed to report any advantages in glucose control, insulin dose, or lipids when compared to placebo
[31].
5. DPP4 Inhibitors
Dipeptidyl-peptidase-4 inhibitors (DPP4-i) inhibit the degradation of the incretin hormones glucagon-like peptide 1 (GLP1) and glucose-dependent insulinotropic peptide (GIP), leading to a stimulation of meal-dependent endogenous insulin secretion, resulting in reduced fasting and postprandial glucose levels. DPP4-i are approved for the treatment of T2D and are characterized by weight neutrality, a low risk of hypoglycemia, and general tolerability. Cardiovascular outcome trials conducted for several DPP4-i have confirmed their cardiovascular safety, but have identified no benefit in terms of reducing major cardiovascular events
[32][33][34][35][36]. The use of saxagliptin was associated with a higher risk of hospitalization for heart failure, as shown in the SAVOR-TIMI 53 study
[32].
The rationale for using DPP4i in T1D is based on the reduction of postprandial glucagon secretion, exerting a positive impact on glucose control. On the other hand, DPP4i are believed to possess immunomodulatory and other pleiotropic features, potentially improving insulitis and delay β-cell destruction in newly diagnosed persons with T1D who are still able to produce endogenous insulin
[37].
Sitagliptin in T1D: In a crossover pilot study, eight weeks of treatment with sitagliptin 100 mg daily in 20 adults with T1D reduced postprandial and 24-h glycemia and reduced prandial insulin requirements significantly. The time spent in hypoglycemia was similar in both trial arms
[38]. The same research group published a larger and longer placebo-controlled trial (141 participants) which investigated the use of sitagliptin on postprandial glucagon levels and parameters of glycemic control. The area under the curve of post-meal glucagon was comparable for sitagliptin and placebo. HbA
1c and insulin doses also did not differ significantly between groups. Additionally, the subgroup of C-peptide-positive participants experienced no beneficial changes in the aforementioned parameters
[39].
A small study comprising 18 persons with T1D and preserved insulin secretion investigated the effect of sitagliptin or exenatide plus insulin or insulin plus placebo for one year. Insulin requirements were significantly lower in those treated with sitagliptin or exenatide. However, stimulated C-peptide was not preserved when the participants were given sitagliptin or exenatide
[40]. Schopman et al. performed a placebo-controlled study and investigated the effect of sitagliptin on counter-regulatory and incretin hormones in response to hypoglycemic clamp experiments. After six weeks of treatment with sitagliptin, GIP and GLP1 levels were significantly increased in response to hypoglycemia. No such effects were registered for glucagon or adrenergic counter-regulatory hormones
[41].
Vildagliptin in T1D: An RCT of similar design investigated vildagliptin (2 × 50 mg for 4 weeks) in C-peptide-negative persons with T1D. While glucagon levels in the vildagliptin group were more intensively suppressed during a meal challenge when compared to placebo, glucagon levels in response to a clamp-induced hypoglycemic episode were not significantly altered. Other counter-regulatory hormones were not significantly influenced in response to vildagliptin
[42].
Saxagliptin in T1D: Treatment with saxagliptin over 12 weeks did not yield benefits in regard of glucose variability, hypoglycemic frequency, or awareness, and also did not improve cognitive function or hormonal counter-regulation in response to hypoglycemia. In addition, treatment with saxagliptin in T1D did not positively influence HbA
1c, insulin doses, or body weight
[43].
6. GLP1-Receptor Agonists
Glucagon-like peptide-1 receptor agonists (GLP1-RA) exert their main effect by stimulating glucose-dependent insulin secretion from β-cells, and were shown to slacken gastric emptying, inhibit post-meal glucagon release, and reduce food intake
[44]. Consequently, GLP1-RAs represent a drug class that is able to significantly reduce body weight. In T2D, CVOTs investigating the GLP1-RAs liraglutide, semaglutide, albiglutide, and dulaglutide have proven efficacy in reducing cardiovascular events
[45][46][47][48]. While most of the data presently available were generated in people with obesity and T2D, also some data are available from RCTs performed in T1D.
Liraglutide: The ADJUNCT program investigated the effect of liraglutide administered in three doses (0.6, 1.2, and 1.8 mg) in two phase three trials comprising more than 2.000 persons with T1D
[49][50]. Both studies included a wide range of persons with T1D, differing in body weight and diabetes duration. The duration of the ADJUNCT ONE trial was 52 weeks, and that of ADJUNCT TWO was 26 weeks. Both studies revealed significant reductions in HbA
1c, insulin doses, and body weight, but higher rates of hypoglycemia and ketosis (ketone level >1.5 mmol/L), especially in the higher-dosed groups
[49][50].
A post hoc analysis of the studies revealed that the efficacy of liraglutide in reducing HbA
1c did not depend on baseline body weight and glycemic control subgroups. However, residual β-cell function was associated with a significant impact on HbA
1c reduction and the risk of hypoglycemia
[51]. Adverse events were higher in the liraglutide groups and were mainly related to the gastrointestinal system. Higher doses of liraglutide (1.2 and 1.8 mg), which were administered over 12 weeks in C-peptide-negative and overweight persons with T1D, were associated with a modest reduction in mean glucose levels and insulin doses, as well as significant weight loss
[52].
The NewLira study indicated that liraglutide 1.8 mg, administered for 52 weeks, preserved insulin secretion in adults with newly diagnosed T1D, as shown by sustained stimulated C-peptide secretion and lower insulin requirements when compared to placebo
[53]. Liraglutide 1.8 mg had no beneficial effects on blood pressure or IMT when administered over 24 weeks in 100 overweight persons with T1D but did increase heart rate, a common observation with this drug class
[54]. In contrast, another study comprising overweight persons with T1D reported the benefits of liraglutide 1.8 mg in terms of reducing systolic blood pressure and markers of obesity, whereas lipids were not influenced by the treatment
[55].
A recently published multicenter RCT investigated whether the combination of anti-interleukin (IL)-21 antibody combined with liraglutide would better enable ß-cell survival when compared to IL-21 or liraglutide alone or placebo in recently diagnosed individuals with T1D. After 54 weeks of treatment, the meal-stimulated C-peptide was significantly higher in the combination group but not in the liraglutide or IL-21 group alone when compared to placebo. Body weight was significantly decreased by all active treatment groups. Hypoglycemic events occurred significantly less in the liraglutide arm when compared to placebo
[56].
Exenatide: The MAG1C RCT investigated the effects of the short-acting GLP1-RA exenatide, administered three times daily over a period of 26 weeks in 108 individuals with T1D who were randomly assigned to receive either exenatide or placebo. While exenatide caused significant reductions in body weight, it had no significant effects on HbA
1c or reductions in insulin doses when compared to placebo. Hypoglycemia was observed at the placebo level. No changes in blood pressure or lipids were seen. Study discontinuation rates due to adverse events were higher in the exenatide arm
[57][58].
Albiglutide: In a study including 67 participants with newly diagnosed T1D, albiglutide administered once a week for one year had no appreciable effect on β-cell function, HbA
1c and weight in newly diagnosed individuals with T1D when compared to placebo
[59]. Anyway, albiglutide was withdrawn from the market in 2018.
Further GLP1-RAs used in T1D are being investigated in the DIAMOND GLP1 (NCT03668470) study (dulaglutide versus placebo), and the TTT1 (NCT03899402) trial (subcutaneous semaglutide, dapagliflozin, or combination versus placebo). The results of these RCTs are not published yet.
7. SGLT1/2 Inhibitors
Sodium–glucose linked transporter-2 inhibitors reduce renal tubular glucose reabsorption, leading to glucosuria, and therefore reduce blood glucose in an insulin-independent manner. Further pleiotropic effects include reductions in blood pressure and weight
[60]. A large body of evidence from CVOTs in T2D has demonstrated cardiovascular and renal benefits when SGLT2-i were investigated, as well as benefits in people with heart failure
[61][62][63][64][65][66][67].
SGLT2-I also showed promising results in persons with T1D. To date, RCT data have been published for nearly all existing SGLT2-i being used in T1D (canagliflozin, dapagliflozin, empagliflozin, ipragliflozin, sotagliflozin).
Empagliflozin in T1D: The EASE program, which included three placebo-controlled RCTs (in total 1707 participants), focused on the efficacy and safety of empagliflozin in different doses during 52 weeks of treatment. Improvements in HbA
1c, time in range, body weight, insulin dose, and systolic blood pressure were seen in all of the dosage arms when compared to placebo. The frequency of severe hypoglycemia was similar in the intervention and placebo arms. However, the rate of DKA was significantly higher with the two higher doses of empagliflozin (10 and 25 mg) when compared to placebo (4.3% with 10 mg, 3.3% with 25 mg, and 1.2% with placebo), while no significant increase was observed with the 2.5 mg dose (0.8%)
[68]. Lunder et al. explored arterial function in an RCT comprising 40 participants with T1D; empagliflozin and metformin alone as well as the combination of both were tested against placebo with regards to changes in arterial function. After 12 weeks of treatment, the combination of empagliflozin and metformin improved arterial stiffness to a greater extent than metformin alone and placebo, leading the authors to conclude that empagliflozin potentially provides similar cardiovascular protection, as is a known function of RCTs in patients with T2D
[69]. A recently published free-living, placebo-controlled, cross-over study of 8 weeks duration investigating 5 mg of empagliflozin as an adjunct to advanced continuous subcutaneous insulin infusion (CSII) therapy, using either an automated insulin delivery (AID) system or a predictive low glucose suspend (PLGS) system in 39 individuals with T1D, demonstrated a significantly improved time in glycemic range in both trial arms (AID: time in glycemic range was 81% with empagliflozin versus 71% without empagliflozin; PLGS: 80% versus 63%). The risk of hypoglycemia was comparable; DKA occurred in one participant receiving empagliflozin and this was associated with a nonfunctioning insulin pump
[70].
Dapagliflozin in T1D: The DEPICT studies examined the efficacy and safety of add-on therapy with the SGLT2-i dapagliflozin in persons with poorly controlled T1D over 52 weeks (DEPICT-1 study; 833 participants) and 24 weeks (DEPICT-2; 1465 participants) of treatment. Two different doses of dapagliflozin (5 and 10 mg) improved HbA
1c, body weight, insulin requirements, and blood pressure. Both doses of dapagliflozin were associated with a higher risk of DKA when compared to placebo
[71][72]. A post hoc analysis of the DEPICT study also showed significant reductions in albuminuria
[73]. A recently published meta-analysis revealed a higher risk of overall adverse events and serious adverse events in the dapagliflozin-treated participants. Although there was a numerical increase in DKA, the risk ratio did not reach statistical significance, which can be attributed to a limited number of DKA events
[74].
Canagliflozin in T1D: Rodbard and Henry, who performed research on RCTs by investigating the effects of canagliflozin administered in two doses versus placebo on T1D in trials with a duration of 18 weeks reported similar outcomes in the indices of glycemic efficacy and body weight, as well as reductions in insulin doses and a higher risk of DKA
[75][76]. In addition, Rodbard reported improvements in patient satisfaction on the DTSQ questionnaires concerning treatment with canagliflozin
[77].
Ipragliflozin in T1D: The SGLT2-i ipragliflozin, which is only available in Japan, yielded similar efficacy data and no higher prevalence of DKA. However, the number of participants was rather small in both available studies investigating ipragliflozin
[78][79].
Sotagliflozin in T1D: Sotagliflozin is a dual inhibitor of SGLT1 and SGLT2. Hence, while SGLT2 is predominantly present in the renal tubules, SGLT1 is involved in intestinal glucose absorption
[80]. A multicenter study comprising more than 1400 persons randomly assigned to receive sotagliflozin 400 mg or placebo over 24 weeks showed that the primary endpoint (HbA
1c < 7% at week 24 without the occurrence of severe hypoglycemia or DKA) was achieved in more sotagliflozin users compared to placebo users (28.6 vs. 15.2%). In addition, persons receiving sotagliflozin showed significant improvements in body weight, blood pressure, and insulin doses when compared to placebo. Severe hypoglycemia (<55 mg/dL) was less frequent in the sotagliflozin group while the prevalence of DKA was increased with sotagliflozin
[81]. The inTandem1 (United States and Canada) and inTandem2 (Europe and Israel) studies investigated the effect of two different doses (200 and 400 mg) of sotagliflozin versus placebo in adults with T1D after a run-in phase of six weeks of insulin therapy optimization. Following 24 weeks of treatment, the percentages of study participants with baseline HbA
1c ≥ 7% achieving an HbA
1c level <7% were 15.7, 27.2, and 40.3% when placebo or sotagliflozin 200 or 400 mg, respectively, were used (inTandem1). After 52 weeks of treatment, fasting plasma glucose, weight, and insulin dose remained significantly better in the treatment groups. Treatment satisfaction scores were significantly higher in persons taking sotagliflozin. The risk of DKA was higher in persons randomized to sotagliflozin, and the risk of severe hypoglycemia was reduced in the sotagliflozin groups
[82][83]. A post hoc analysis of the inTandem program revealed a more favorable efficacy and safety profile when sotagliflozin was used in persons with a BMI > 27 kg/m
2 compared to those with a BMI < 27 kg/m
2 [84]. Interestingly, in a recently published small study including 85 participants with not-well-controlled T1D and a younger age (18–30 years) receiving sotagliflozin 400 mg or placebo, the risk of DKA was not increased in the sotagliflozin group. This might be justified by the meanwhile-gained knowledge on DKA risk with sotagliflozin and an improved educational procedure prior to therapy initiation
[85].