The “Old” Oral Antidiabetics

Abstract

Treatment of patients with type 2 diabetes aims to avoid acute symptoms of hyperglycemia and to prevent macro- and microvascular complications. In recent years, the number of glucose-lowering drugs increased to unprecedented levels. The American Diabetes Association (ADA) lists seven drug classes of available glucose-lowering agents in the last edition of their standards of medical care in diabetes (American Diabetes Association, Diabetes Care 44(Suppl 1):S111–S124, 2021). All are proven to decrease HbA1c-levels or postprandial glucose excursions, but evidence on patient-relevant outcomes, such as cardiovascular mortality, amputations, or retinopathy, is sparse. Reduction of HbA1c-values is often used as a surrogate outcome measure to assess the efficacy of antidiabetic medication. However, its appropriateness has been disproven (Buhse et al., Novelties in diabetes, endocrine development, Karger, Basel, 2015; Bousageon, Br J Clin Pract 67:85–87, 2017). In the Action to Control Cardiovascular Risk in Diabetes (ACCORD) study (Gerstein et al., N Engl J Med 358: 2545–2559, 2008; Tian et al., Diabetes Care 43:1293–1299, 2020) and the Veterans Affairs Diabetes Trial (VADT) (Duckworth et al., N Engl J Med 360:129–139, 2009; Reaven et al., N Engl J Med 380:2215–2224, 2019), a rigid treatment regime with low HbA1c-targets did not result in better patient-relevant outcomes. Patients in the intervention arm of the ACCORD study even had a higher risk of mortality, and consequently, the study was terminated earlier (Gerstein et al., N Engl J Med 358: 2545–2559, 2008). Other drugs have been withdrawn from the market because of a negative benefit-risk ratio, for example, phenformin, which increased the risk of lactic acidosis or rosiglitazone that reduced HbA1c-values but increased cardiovascular risk (Wallach et al., BMJ 368:I7078, 2020). In recent years, pharmaceutical companies decided to withdraw several new antidiabetic agents from the German market, such as vildagliptin and canagliflozin, because no additional benefit over usual care could be demonstrated and therefore health insurances would not have covered additional costs.

Introduction

Treatment of patients with type 2 diabetes aims to avoid acute symptoms of hyperglycemia and to prevent macro- and microvascular complications. In recent years, the number of glucose-lowering drugs increased to unprecedented levels. The American Diabetes Association (ADA) lists seven drug classes of available glucose-lowering agents in the last edition of their standards of medical care in diabetes [1]. All are proven to decrease HbA1c-levels or postprandial glucose excursions, but evidence on patient-relevant outcomes, such as cardiovascular mortality, amputations, or retinopathy, is sparse. Reduction of HbA1c-values is often used as a surrogate outcome measure to assess the efficacy of antidiabetic medication. However, its appropriateness has been disproven [2, 3]. In the Action to Control Cardiovascular Risk in Diabetes (ACCORD) study [4, 5] and the Veterans Affairs Diabetes Trial (VADT) [6, 7], a rigid treatment regime with low HbA1c-targets did not result in better patient-relevant outcomes. Patients in the intervention arm of the ACCORD study even had a higher risk of mortality, and consequently, the study was terminated earlier [4]. Other drugs have been withdrawn from the market because of a negative benefit-risk ratio, for example, phenformin, which increased the risk of lactic acidosis or rosiglitazone that reduced HbA1c-values but increased cardiovascular risk [8]. In recent years, pharmaceutical companies decided to withdraw several new antidiabetic agents from the German market, such as vildagliptin and canagliflozin, because no additional benefit over usual care could be demonstrated and therefore health insurances would not have covered additional costs.

In 2012, the ADA and the European Association for the Study of Diabetes (EASD) recommended patient-centered care including shared decision-making (SDM) [9] and reasserted this position in further statements [10]. SDM is a particular form of communication between patients and their health care professionals. It focuses on the mutual exchange of information in order to involve patients in the decision-making process [11]. Therefore, patients need understandable information on probabilities of benefits and harms of treatment options [12,13,14]. The question to be answered is, what option is the best to prevent diabetes-related complications and yet in line with individual patient values and preferences? Supportive tools in that process are patient decision aids, which help patients to weigh up pros and cons of diabetes treatment [15, 16].

This chapter gives an overview of older classes of antidiabetic agents and their efficacy. It is based on a systematic inventory published in 2015 and updated in the first edition of The Diabetes Textbook [2]. Sulfonylureas (SU) and biguanides are the oldest classes of oral glucose-lowering agents. Later, thiazolidinediones (TZDs), alpha glucosidase inhibitors (AGIs), and meglitinides were approved. Table 34.1 shows the old drug classes and their compounds that are still available in the United States or Europe. Newer classes, such as sodium-glucose cotransporter 2 (SGLT-2) inhibitors and dipeptidyl peptidase-4 (DPP-4) inhibitors, will be described in other chapters of this book.

Table 34.1 Overview of older classes of antidiabetic agents^a

According to recent guidelines [1, 7, 14, 17], this chapter focuses on the efficacy of metformin and SU monotherapies compared with other monotherapies as well as comparisons of metformin-based combinations. At the end of this chapter, we give an example of our decision aid for patients with type 2 diabetes and how diabetes educators share evidence-based information with their patients [18,19,20].

Methods

We updated our search from April 2014 [2]. In a first step, we searched PubMed and the Cochrane library for systematic reviews and meta-analyses published from May 2014 to the end of November 2021. Systematic reviews were considered if they included randomized controlled trials on the efficacy of metformin, sulfonylureas, thiazolidinediones, meglitinides, or alpha-glucosidase inhibitors as monotherapy or combination of two or three drugs. There is a growing number of network analyses. They typically comprise indirect comparisons when there is no head-to-head comparison available. Network analyses are methodologically challenging and can lead to false results and interpretations if differences between studies were not adequately considered [21]. Treatment of type 2 diabetes is complex, and as a result, RCTs in meta-analyses are usually heterogeneous. We therefore excluded network meta-analyses. In addition, inclusion criteria, such as study duration, sample size, target group, and drug classes, vary between systematic reviews. Hence, following our previous methodological approach [2], we extracted RCTs from the reviews that fulfilled our inclusion criteria: (1) patient-relevant primary endpoint, that is, macro- and microvascular complications, cardiovascular mortality, total mortality, and quality of life; (2) intention-to-treat analysis; (3) follow-up of at least 24 weeks and adequate sample size; and (4) hard clinical endpoints that had to be reported. Finally, we searched for further studies and screened the websites of the National Institute for Health and Care Excellence (NICE), the Agency for Healthcare Research and Quality (AHRQ), and the German Institute for Quality and Efficiency in Health Care (IQWIG) for new reports and guidelines.

Results

The search update for systematic reviews and meta-analyses resulted in 516 records. Most of them were network analyses. Although the Food and Drug administration (FDA) and the European Medicines Agency (EMA) changed licensing regulations toward cardiovascular outcome trials for glucose-lowering drugs in 2008, reviews mainly focused on surrogate endpoints, such as HbA1c-level. We identified one systematic review on the efficacy of metformin compared to no intervention, placebo, or lifestyle intervention [22], a meta-analysis comparing metformin and SU as monotherapy [23], a systematic review and meta-analysis to assess the effectiveness and safety of the addition of metformin to standard insulin therapy in children with type 1 diabetes aged 6–19 years [24], and an umbrella review of systematic reviews with meta-analysis to assess the efficacy and safety of metformin [25]. Regarding sulfonylureas, we identified two randomized clinical trials comparing glimepiride with dipeptidyl peptidase -IV inhibitors (DPP-IVi): a multicenter, randomized controlled trial to compare the efficacy and safety of glimepiride with saxagliptin in patients with type 2 diabetes inadequately controlled with metformin [26] and a randomized controlled trial comparing the effect of treatment of glimepiride versus linagliptin on cardiovascular safety in patients with type 2 diabetes [27]. With respect to sodium-glucose co-transporter-2 inhibitors (SGLT2i), we identified a randomized controlled trial comparing the efficacy and safety of empagliflozin in patients with type 2 diabetes inadequately controlled on metformin [28] and a randomized controlled trial comparing the efficacy and safety of glimepiride versus ertugliflozin in patients with type 2 diabetes inadequately controlled with metformin [29]. Four trials evaluated the effects of TZDs [30,31,32,33]. Systematic reviews about the effectiveness and safety of alpha-glucosidase inhibitors and meglitinides could not be identified. Regarding overall comparisons, a systematic review assessed the efficacy of eight classes of diabetes medications including metformin, sulfonylureas, and alpha-glucosidase inhibitors [34]; two meta-analyses by the Agency for Healthcare Research and Quality (AHRQ) evaluating all available glucose lowering drugs [35, 36]; a meta-analysis involving 301 clinical trials to assess the efficacy and safety of all classes of oral and injectable antidiabetics, including insulin [37]; and a systematic review including 453 trials comparing the efficacy and safety of nine classes of antidiabetics including monotherapies, add-on to metformin-based therapies and monotherapies versus add-on to metformin therapies [38]. This report includes RCTs and observational studies on (1) comparisons of monotherapies (metformin, thiazolidinediones, sulfonylureas, DPP-4 inhibitors, SGLT-2 inhibitors, and GLP-1 receptor agonists), (2) comparisons of metformin alone and metformin-based combinations, and (3) comparisons of metformin-based combinations where the second drug was one of the monotherapies or insulin treatment. The evidence was graded separately for both study types. The AHRQ search update was performed through December 2016. Our search for more recent RCTs from January 2016 to November 2021 yielded 310 records.

Metformin

Metformin belongs to the class of biguanides. It is the only still licensed compound of its class after phenformin was withdrawn from the markets. In the University Group Diabetes Program (UGDP) [39, 40], the first large RCT that evaluated the efficacy of glucose lowering drugs on macro- and microvascular outcomes, phenformin, was associated with an increase of cardiac mortality. In contrast, metformin is internationally recommended as initial drug treatment for people with type 2 diabetes [1, 9, 17, 29, 30]. This is mainly based on the results of the UK Prospective Diabetes Study (UKPDS), published in 1998 [41]. About 4000 patients with newly diagnosed type 2 diabetes were enrolled in this RCT. The study objective was to assess the efficacy of intensive blood glucose-lowering therapy compared to conventional treatment (primarily with diet). Patients in the intensive treatment group were supposed to achieve a fasting plasma glucose level of less than 6 mmol/L. The fasting plasma glucose target of the conventional treatment arm was less than 15 mmol/L with no symptoms of hyperglycemia. Non-overweight patients were randomly assigned to intensive treatment with insulin, intensive treatment with sulfonylureas, or conventional therapy with diet. A subgroup of overweight patients had the additional possibility to be randomized to intensive treatment with metformin [41, 42]. A total of 342 patients were assigned to metformin and 411 patients to conventional control with diet [43].

The median HbA1c-level of the intensive treatment group with metformin was 7.4% during the 10 years of follow-up. The conventional group had a median HbA1c-level of 8.0%. Compared to conventional treatment, patients in the metformin monotherapy arm showed significant reductions in any diabetes-related endpoint, a composite endpoint comprising the following outcome measures: sudden death, death from hyperglycemia or hypoglycemia, fatal or nonfatal myocardial infarction, angina, heart failure, stroke, renal failure, amputation of at least one digit, vitreous hemorrhage, retinopathy requiring photocoagulation, blindness in one eye, or cataract extraction. Moreover, diabetes-related death, all-cause mortality, and myocardial infarction significantly decreased in the intensive treatment group with metformin.

Based on these results, metformin became the first-line drug for patients with type 2 diabetes who do not achieve their HbA1c target with diet and other lifestyle interventions alone. However, the results of the UKPDS have not yet been reproduced [2, 44]. The UKPDS was a study with an open-label design, which may lead to overestimated results. The protocol was changed during the study. The initially defined significance threshold of 1% was later changed to 5%. The significant difference in reduction of total mortality and myocardial infarction in the metformin group was above the threshold of 1% [44].

Antihypertensive treatment or statins may have a greater effect on mortality than metformin [45]. This may also explain the results of the UKPDS follow-up study [46], which reported significant reductions in total mortality and cardiovascular mortality for all intensive treatment groups 10 years after the main publication of the UKPDS results. Considering the high risks of bias of the UKPDS, the interpretation of the follow-up results as long-term effect of intensive early glucose control might be misleading [47]. In addition, only about one third of the initially randomized patients were analyzed in the follow-up study.

A meta-analysis that included 13 studies comparing metformin as monotherapy or add-on therapy to diet, placebo, or no treatment found no significant effects on all-cause mortality, cardiovascular mortality, or microvascular complications [48]. Of the included RCTs that assessed patient-relevant outcomes as the primary endpoint [41, 49,50,51], only UKPDS [41] showed a beneficial effect for treatment with metformin.

In the UKPDS, metformin monotherapy was also associated with a decrease in any diabetes-related endpoint and all-cause mortality compared to intensive treatment with sulfonylurea or insulin [41]. Data on metformin compared to SU alone were not reported in the UKPDS [23].

The study on the Prognosis and Effect of Antidiabetic Drugs on Type 2 Diabetes Mellitus With Coronary Artery Disease (SPREAD-DIMCAD) [52] compared metformin with the SU glipizide in 304 Chinese people with type 2 diabetes mellitus and coronary artery disease. The targeted HbA1c-level was less than 7% for both groups. The primary endpoint was recurrent cardiovascular events, a composite outcome measure comprising nonfatal myocardial infarction, nonfatal stroke, arterial revascularization, cardiovascular death, and all-cause mortality. The study results showed a significant reduction in this endpoint in favor of the metformin group. However, there is a substantial risk of bias that limits the validity of the study results. The study was retrospectively registered, and there is no study protocol published. Data from 5 years of follow-up were analyzed, but the study drug was only administered for 3 years. It was not reported whether the study treatment was maintained after this time.

A meta-analysis on the effects of SU monotherapy compared to metformin monotherapy did not find any differences between treatment groups regarding all-cause or cardiovascular mortality [23]. A potential benefit of SU over metformin was identified in nonfatal macrovascular outcomes, but definitions of that composite endpoint were heterogeneous. There were no data on microvascular outcomes for a meta-analysis. Results of that meta-analysis were mainly based on “A Diabetes Outcome Progression Trial” (ADOPT), a multicenter, randomized controlled, double-blind trial with 4 years of follow-up [53]. Patients with untreated diabetes were randomized to metformin, glibenclamide, or rosiglitazone. The primary endpoint was time to treatment failure, defined as fasting plasma glucose level of more than 180 mg per deciliter after 6 weeks at maximum tolerated dose of the study drug. As this is not a clinical hard endpoint, we excluded this trial from our overview. However, there was no difference regarding all-cause mortality or fatal myocardial infarction between the glibenclamide and metformin groups [23, 53].

Compared to sulfonylureas alone, the combination of metformin and sulfonylureas significantly increased death from any cause and diabetes-related death in overweight and non-overweight patients in the UKPDS [41]. The meta-analysis by Boussageon et al. [48] confirmed a significant increase in all-cause and cardiovascular mortality for metformin plus SU compared to metformin monotherapy. The results were mainly based on the UKPDS. After excluding this study, no group difference was seen in both endpoints.

The Hyperinsulinemia: The Outcome of Its Metabolic Effects (HOME) trial evaluated the efficacy of metformin in the Netherlands [51]. The RCT included 390 overweight and obese patients with type 2 diabetes. Metformin added to insulin therapy was compared to insulin monotherapy. After about 4 years, there was no difference between groups regarding cardiovascular and total mortality or microvascular outcomes (progression of retinopathy, nephropathy, and neuropathy) but a significant reduction in a combined macrovascular endpoint for patients with metformin plus insulin treatment. This composite endpoint included a total of 13 separate outcome measures, for example, myocardial infarction, heart failure, stroke, diabetic foot, percutaneous transluminal coronary angioplasty, nontraumatic amputation, and sudden death. Patients’ characteristics were unequally distributed between the study groups at baseline. For example, in the metformin group, there were fewer smokers (19% vs. 30%) and more patients with antihypertensive medication (47% vs. 39%). In addition, the number of non-completers differed between the metformin plus insulin study arm (n = 65) and the insulin alone arm (n = 48), mainly because of adverse events.

A systematic review that analyzed RCTs published until February 2017 [22] found similar results regarding the efficacy of metformin. The authors identified no more recent RCTs than earlier meta-analyses, but study selection was not completely transparent. The UKPDS [41] was included in the meta-analysis, but only the combination of metformin and SU compared to SU alone, not the comparison of metformin with diet or metformin monotherapy with SU. Moreover, the authors included the 10 years follow-up UKPDS in their analysis, and observational studies were excluded. In fact, the level of evidence of the UKPDS follow-up publication [46] is quite similar to observational studies due to the already mentioned risks of bias [44, 47].

Compared with other interventions, metformin does not increase the risk of mild or severe hypoglycemia. The main adverse events associated with metformin are gastrointestinal, especially diarrhea. There have been warnings of lactic acidosis due to metformin. A Cochrane review [54] and the AHRQ report [36] did not find an increased risk of lactic acidosis with the use of metformin. Up to 2016, metformin was not recommended for patients with moderate to severe kidney function. Following this advice to practicing physicians might be one reason for a low number of reported cases of lactic acidosis. In 2016, the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) changed their recommendations to allow the use of metformin in patients with moderately reduced kidney function (GFR = 30–59 mL/min) [55, 56]. The FDA explicitly recommends assessment of risks and benefits in patients with metformin whose GFR fall below 45 mL/min/1.73 m². Starting metformin in patients with eGFR between 30–45 mL/min/1.73 m² was not recommended [56].

Sulfonylureas

The first-generation SU tolbutamide and chlorpropamide were introduced in the 1950s. In the UGDP, tolbutamide increased mortality risk. Nonetheless, both substances were extensively used even after publication of the UGDP in many countries. Today, first-generation SU have been replaced by the second- and third-generation SU. SU are recommended as initial drug therapy if metformin is contraindicated or not tolerated by patients [1, 17, 18]. The comparative effects of SU to metformin are already described in the metformin part of this chapter. We additionally searched for systematic reviews and RCTs on the efficacy of SU as monotherapy compared to diet, placebo, or lifestyle interventions.

As in our original overview [2], the only RCT that met our inclusion criteria was the UKPDS [42]. In the UKPDS 33, the effects of intensive blood-glucose control with either SU or insulin were compared to conventional treatment. A total of 615 patients were assigned to glibenclamide, and 896 received conventional treatment, which comprised dietary advice. Over 10 years, median HbA1c values were 7.2% for glibenclamide and 7.9% for conventional therapy. More patients in the conventional treatment arm had reached the primary endpoint “any diabetes-related endpoint” and microvascular complications, but there were no significant effects on macrovascular outcomes. The effect on the microvascular outcome was mainly attributed to fewer cases of retinal photocoagulation [42]. Patients in the SU group gained more weight (1.7 kg) than patients in the conventional treatment group, and more patients receiving SU had major hypoglycemic events (1.4% vs. 0.7%) over 10 years (Table 34.2).

Table 34.2 Metformin and sulfonylurea, identified evidence on efficacy of single RCTs

Thiazolidinediones

Thiazolidinediones were introduced in the 1990s. The first agent of this class, troglitazone, was withdrawn from the market because of the increased risk of severe liver damage and toxicity. The remaining compounds, rosiglitazone and pioglitazone, were under selling restrictions or withdrawn in some countries due to safety issues. Meta-analyses showed an increased risk of myocardial infarction in patients who received rosiglitazone [57, 58]. One of the included studies was the RECORD trial with a mean follow-up of 5.5 years [59]. A total of 4447 patients who were treated with metformin or SU monotherapy were randomized to additional rosiglitazone or additional metformin/SU. Patients of the rosiglitazone group had a twofold greater risk of fatal and nonfatal heart failure compared to patients with metformin plus SU treatment. There was no difference between groups regarding the combined primary endpoint, cardiovascular death or cardiovascular hospitalization. Patients receiving rosiglitazone therapy reported significantly more bone fractures. Further adverse effects of rosiglitazone comprised weight gain and edema [59]. Findings from the ADOPT trial confirmed higher cardiovascular risks and other adverse effects associated with rosiglitazone [53]. The FDA restricted access to rosiglitazone, which was part of the Risk Evaluation and Mitigation Strategy (REMS), and the RECORD trial showed several risks of bias. The RECORD trial was an open-label trial with low statistical power. An unplanned interim analysis was conducted, which could have repealed blinding. Patients’ compliance to rosiglitazone was low. In December 2015, based on an independent review of the study, the FDA stated that REMS was no longer needed and that the benefits of rosiglitazone outweighed the risks. In their Standards of Medical Care in Diabetes, the American Diabetes Association recommended TZD as add-on therapy or monotherapy if metformin was contraindicated [58].

In our updated search, we identified a meta-analysis on the effect of pioglitazone on cardiovascular outcomes, which also included participants with pre-diabetes and insulin resistance [32]. The primary endpoint was major adverse cardiovascular events (MACE) comprising cardiovascular death, nonfatal myocardial infarction, and nonfatal stroke. The use of pioglitazone in patients with diabetes was associated with lower risks of MACE, and the incidence of myocardial infarction or stroke did not differ between the pioglitazone and control groups. Pioglitazone was also associated with an increased risk of heart failure, bone fracture, edema, weight gain, and hypoglycemia [32]. The largest included RCT was the PROactive trial [60]. Patients with type 2 diabetes and previous stroke were randomized to receive pioglitazone or placebo, and the mean study duration of the study was 34.5 months. Albeit there was a reduction in the combined endpoint, death from any cause, nonfatal MI, and stroke for patients randomized to receive pioglitazone, patients in this group had significantly higher risks of heart failure, edema, and weight gain in addition to a nonsignificant higher rate of bladder cancer [60]. In the meta-analysis, no significant differences were found in bladder or any cancer risk [32]. Another meta-analysis reported a significantly higher risk [31], but it was mainly based on the results of the PROactive trial. A systematic review on the effects of TZD on bone fractures confirmed an increased risk of fractures in women who use rosiglitazone or pioglitazone [30]. The National Institute for Clinical Excellence (NICE) recommends pioglitazone when metformin is contraindicated but explicitly points out the risks of adverse events (Table 34.3) [18].

Table 34.3 Thiazolidinedione, identified evidence from RCTs

Alpha Glucosidase Inhibitors and Meglitinides

The ADA and the EASD do not explicitly recommend the use of AGIs due to their modest effects, but they accept that AGIs “may be tried in specific situations” [9]. Two Cochrane reviews including patients with type 2 diabetes and patients with impaired glucose tolerance did not find significant effects of AGIs on mortality or morbidity [61, 62]. We did not include the STOP-NIDDM trial in this overview because of its high risk of bias, which was extensively discussed in the literature. The Acarbose Cardiovascular Evaluation Trial (ACE) [63] evaluated the efficacy of acarbose on cardiovascular death, nonfatal MI, and nonfatal stroke in patients with impaired glucose tolerance and coronary heart disease. This RCT was completed in April 2017, and the results showed that patients received acarbose achieved a statistically significant reduction on 18% in the relative risk of diabetes, without reduction in the risk of major adverse cardiovascular events (MACE).

Meglitinides same as SU belong to the drug class of insulin secretagogues. Compounds of this class are nateglinide and repaglinide. In contrast to SU, they are rapid-acting secretagogues. The ADA and the EASD stated that meglitinides may be used as an alternative to SU in patients with irregular meal schedules [10]. In case repaglinide is considered as alternative to metformin, the NICE guidance on Type 2 diabetes in adults suggests physicians to inform patients that there is no licensed non-metformin-based combination with repaglinide [15]. There is no evidence on effects regarding clinically relevant and long-term outcomes with the use of meglitinides [64].

Conclusion

In conclusion, the older classes of oral antidiabetic agents still play central roles in diabetes care, but evidence on macro- and microvascular risk is lacking or insufficient.

The applicability of the results of clinical trials is limited due to the short duration of the studies [35]. Most studies assess the efficacy of medications on intermediate outcomes rather than long-term hard clinical endpoints [34, 35, 38]. Intermediate outcomes or surrogates must be interpreted with caution. Medications decreasing HbA1c-values do not necessarily reduce morbidity or mortality. In some cases of withdrawn drugs, blood glucose levels decreased, while risks of hard clinical endpoints did not change or even increased. Whenever RCTs included patient-relevant endpoints, they were mostly assessed as secondary endpoints or adverse effects. Available studies were often too small to identify any differences between groups. Composite outcome measures, such as any diabetes-related endpoint or macrovascular complications, which usually comprise endpoints of varying importance and validity, are challenging to interpret and may lead to overinterpretation of single outcomes.

The authors of the AHRQ report [36] concluded that the efficacy of all diabetes medications regarding all-cause mortality, cardiovascular and cerebrovascular morbidity as well as retinopathy, nephropathy, and neuropathy is still uncertain. The report showed moderate strength of evidence that sulfonylurea monotherapy compared with metformin alone was associated with an increased risk of cardiovascular mortality. This result was mainly based on two RCTs: ADOPT with patients with newly diagnosed diabetes and SPREAD-DIMCAD, which included patients with coronary heart disease. In contrast, the meta-analysis by Madsen et al. [23] did not find any differences between SU and metformin monotherapy of total or cardiovascular mortality but a potential benefit of SU regarding nonfatal macrovascular outcomes. However, definition of the composite endpoint differed between studies [23].

In the AHRQ report [36], evidence on intermediate outcomes, such as HbA1c values, was graded as high, and effects on HbA1c values were comparable between most oral antidiabetic agents. Monotherapy comparisons of metformin with sulfonylurea and metformin with TZDs show similar effects with respect to reduction in HbA1c values [65]. Moreover, metformin monotherapy reduced body weight more than TZDs or SU, though the clinical relevance of these differences may be debatable. Metformin monotherapy shows greater weight reduction when compared with the combination of metformin and SU or metformin plus TZDs, respectively [36, 66]. In addition, metformin was favored over SU monotherapy, the combination of metformin and TZDs, and over the combination of metformin and SU regarding hypoglycemia [67]. The risk of hypoglycemia is higher for SU than for TZDs [36], albeit differences in the risk of hypoglycemia have been documented, probably explained by differences in chemical structure, pharmacogenetic and pharmacodynamic properties between sulfonylureas [68].

Despite there is only one RCT with a small sample size, which demonstrated an effect on hard clinical endpoints, metformin is internationally recommended as first-line drug for patients with type 2 diabetes. It is used as comparator for the evaluation of new medications although high-quality evidence on patient-relevant outcomes is missing. Thus, the role of metformin as “gold standard” is questionable. Despite a huge number of studies and a stunning total of 427 meta-analysis until 2021, evidence on metformin in observational studies generally does not seem reliable, due to substantial heterogeneity between studies, small-study effects, and excess significance, while evidence from randomized trials suggests only a few effects with strong evidence for additional benefits [25].

Shared Decision-Making

Even though there is no single perfect treatment of hyperglycemia in patients with type 2 diabetes, decisions about treatment policies and diabetes drug therapy are made for thousands of patients every day. For many decades, the dominant approach to making decisions about treatment in the medical encounter has been one of paternalism, but in recent years, this model has been challenged by doctors, patients, medical ethicists, and researchers who advocate more of a partnership relation between doctors and patients [69]. Shared decision-making is a personalized and patient-centered approach [70] described by Charles, Gafni, and Whelan in 1997 to help patients and clinicians to select the treatment that best fits individual patient needs, values, and preferences [71]. It is a special way of conversation between patients and healthcare professionals comprising various elements, such as clarifying the patient’s situation, noticing that there is more than one treatment option, information about benefits and harms of the treatment options, and weighing up the pros and cons considering patient values and expectations. Patient decision aids are tools to promote SDM. They are proved to improve patients’ knowledge about treatment options and about probabilities of benefits and adverse effects of each option. Moreover, they help patients to find the option that is most important to them [16]. Decision aids can be used to prepare patients for the consultation with their clinician or within consultations [15]. We have developed an evidence-based patient decision aid on the prevention of myocardial infarction and a corresponding group counselling session in which diabetes educators help patients to understand the information and to define and prioritize own treatment goals regarding statin uptake, smoking cessation, and HbA1c and blood pressure goals [19, 20]. The intervention (informed shared decision-making program; ISDM) was evaluated in a proof of concept RCT [19]. Patients of the ISDM group achieved higher levels of risk comprehension and realistic expectations about benefits and harms of treatment options. For the following cluster RCT with family practices, we added a structured SDM training for physicians and a patient-held documentation sheet to the intervention in order to optimize the consultation in terms of SDM [72]. The study results showed that the whole ISDM program could be successfully implemented in everyday practice. Patients and clinicians of the ISDM group pursued common treatment goals significantly more frequently than the control group [73].

Figure 34.1 displays a 100-stick figure pictogram and bar graphs to visualize probable effects of more or less intensified glucose control on the combined diabetes-related endpoint (UKPDS 34) as used in our patient decision aid and group teaching session [19, 20, 72].

Fig. 34.1

Blood sugar control and “any diabetes-related event”

Effects on “any diabetes related event” can be explained as follows:

The term “any diabetes related event” is a collective term for different complications of diabetes. It included death from hyperglycemia (high blood sugar) or hypoglycemia, heart attack, angina, heart failure, stroke, kidney failure, amputation, vitreous hemorrhage in the eye (bleeding from abnormal blood vessels in the eye, which can lead to blindness), damage to the retina, blindness of one or both eyes, or eye surgery for cataract.

In the following, you can see the results from the UKPDS [41]. This is a study that was performed in Great Britain and lasted 10 years.

Imagine two groups, each with 100 patients with type 2 diabetes followed for 10 years.

One group was treated intensively with medication to control blood sugar levels. Patients of that group achieved an average HbA1c of 7%.

The comparator (control group) was treated conventionally with less intensive medication and achieved an average HbA1c of 8%.

• In the group with conventional treatment, “any diabetes-related event” occurred in 46 of the 100 patients during the 10 years period.

• In the group with intensive control, “any diabetes-related event” occurred in 41 of the 100 patients during the 10 years period.

That means, intensive blood sugar control over 10 years prevented “any diabetes-related event” in 5 of 100 patients. The remaining 95 of 100 people had no benefit from the intensive treatment over a period of 10 years because they also experienced a diabetes-related event (41 patients) or because they would not have experienced any event even with conventional treatment (54 patients).

Intensively treated patients also experienced harm due to hypoglycemia. An additional 7 out of 100 people suffered severe hypoglycemia with intensive treatment compared to the comparator group over 10 years [41].

Communication of uncertainties is challenging. No one can say if one particular patient would benefit from intensive treatment. Presenting the data helps patients to weigh up the pros and cons making a decision, which meets personal preferences and values. Moreover, the effects of antihypertensive treatment and statin intake should be taken into consideration. For example, intensive blood pressure lowering over 8 years (achieved RR 145/82 mmHg) prevented “any diabetes-related event” in 16 out of 100 patients [74].

According to ADA and EASD recommendations [9, 10], clinicians should talk with patients about the pros and cons of medications to achieve individual treatment goals. In our ISDM program, diabetes educators explain benefits and harms of evidence-based options to prevent cardiovascular complications. They guide patients to estimate their individual heart attack risk and then calculate their risks with and without statin intake and to estimate comparable effects of hypertensions or blood glucose control [19, 72].

Since efficacy of single diabetes medications seems uncertain [36], information about antidiabetic agents can only focus on intermediate outcomes, such as weight change, HbA1c values, hypoglycemia, and other side effects. Montori’s research group developed and evaluated diabetes medication choice decision aid cards on intermediate effects to be used during the clinical encounter [75]. Patients had improved knowledge and were more involved in the decision-making process [75]. Another decision aid addressed statin choice to prevent myocardial infarction in patients with type 2 diabetes [76, 77]. There are also interactive and web-based decision aids that are supposed to foster shared decision-making and goal setting [78] and patient decision aids on special treatments, such as starting insulin [79].

Communication of quality of data is challenging. Patient decision aids are supposed to provide the best available evidence. However, sometimes, there is no good evidence, but patients have the right to know. Information on level of evidence is provided in guidelines and should be included in the patient information material.

Diabetes care is complex and has to be individualized. The level of evidence of antidiabetic agents on patient-relevant outcomes is low, and it has been shown that treatment of hypertension is more effective than treatment of blood glucose [42]. New therapeutic classes have enlarged the scope of antidiabetics, and their cardiovascular and renal benefits are advantageous by comparison with the “old antidiabetics” [80]. Beyond these achievements, the first antidiabetics continue to be part of the clinical armamentarium because of efficacy and costs. Last but also very important, involving patients in decision-making and making informed choices should be standard in the medical encounter.

Multiple Choice Questions

• 1. Which is the aim of the treatment of type 2 diabetes?

  • (a) Fasting blood glucose control

  • (b) Avoid acute symptoms of hyperglycemia and prevent macro- and microvascular complications

  • (c) Post-prandial blood glucose control

  • (d) Increase the use of medications

  • (e) Weight reduction and control

• 2. Rigid treatment regimes with low HbA1c targets:

  • (a) Have resulted in better patient-relevant outcomes

  • (b) Have produced equal patient-relevant outcomes

  • (c) Are associated with higher risks of mortality

  • (d) Improve health-related quality of life

  • (e) Reduce hospital admissions and costs

• 3. What was the argument to withdraw several new antidiabetic agents from the German market?

  • (a) No additional benefit over usual care could be demonstrated and health insurances would not have covered additional costs

  • (b) Higher costs compared with traditional medications

  • (c) Higher risk of hypoglycemia

  • (d) Unacceptable risk of nondiabetic ketoacidosis

  • (e) All of the above

• 4. According to the recent ADA and EASD recommendations, clinicians should not discuss with patients the pros and cons of medications to achieve individual treatment goals.

  • (a) False

  • (b) True

• 5. What is the mechanism of action of metformin?

  • (a) Reduction of insulin resistance in target cells through transcription of several genes involved in glucose and lipid metabolism

  • (b) Inhibition of alpha-glucosidase, delaying intestinal degradation of complex carbohydrates and prolonging post-prandial glucose absorption.

  • (c) Multiple sites of action, including increase of insulin sensitivity by increasing peripheral glucose uptake, decrease of intestinal glucose absorption, and decrease of hepatic glucose production

  • (d) Stimulation of insulin release in pancreatic beta cells. Decrease in hepatic clearance of insulin. Additional extra-pancreatic mechanisms

  • (e) Increase insulin sensitivity by skeletal muscle

• 6. What is the mechanism of action of glyburide?

  • (a) Reduction of insulin resistance in target cells through transcription of several genes involved in glucose and lipid metabolism

  • (b) Inhibition of alpha-glucosidase, delaying intestinal degradation of complex carbohydrates and prolonging post-prandial glucose absorption

  • (c) Increase of insulin sensitivity by increasing peripheral glucose uptake, decrease of intestinal glucose absorption, and decrease of hepatic glucose production

  • (d) Stimulation of insulin release in pancreatic beta cells. Decrease in hepatic clearance of insulin. Additional extra-pancreatic mechanisms.

  • (e) Increase insulin sensitivity by skeletal muscle

• 7. What is the mechanism of action of thiazolidinediones?

  • (a) Reduction of insulin resistance in target cells through transcription of several genes involved in glucose and lipid metabolism

  • (b) Inhibition of alpha-glucosidase, delaying intestinal degradation of complex carbohydrates and prolonging post-prandial glucose absorption

  • (c) Increase of insulin sensitivity by increasing peripheral glucose uptake, decrease of intestinal glucose absorption, and decrease of hepatic glucose production

  • (d) Stimulation of insulin release in pancreatic beta cells. Decrease in hepatic clearance of insulin. Additional extra-pancreatic mechanisms.

  • (e) Increase insulin sensitivity by skeletal muscle

• 8. What is the mechanism of action of alpha-glucosidase inhibitors?

  • (a) Reduction of insulin resistance in target cells through transcription of several genes involved in glucose and lipid metabolism

  • (b) Inhibition of alpha-glucosidase, delaying intestinal degradation of complex carbohydrates and prolonging post-prandial glucose absorption

  • (c) Increase of insulin sensitivity by increasing peripheral glucose uptake, decrease of intestinal glucose absorption, and decrease of hepatic glucose production

  • (d) Stimulation of insulin release in pancreatic beta cells. Decrease in hepatic clearance of insulin. Additional extra-pancreatic mechanisms

  • (e) Increase insulin sensitivity by skeletal muscle

• 9. The evidence sustaining that sulfonylurea monotherapy compared with metformin alone was associated with an increased risk of cardiovascular mortality comes from:

  • (a) Experiences of primary care practitioners

  • (b) Two clinical trials: ADOPT with patients with newly diagnosed diabetes and SPREAD DIMCAD, which included patients with coronary heart disease

  • (c) Pharmaco-vigilance reports

  • (d) Conclusions of consensus groups

  • (e) The results of the Diabetes Control and Complications Trial (DCCT)

• 10. Moderate strength of evidence suggest sulfonylurea monotherapy compared with metformin alone was associated with:

  • (a) Higher risk of cardiovascular mortality

  • (b) An increase in metabolic control

  • (c) Lower weight gain

  • (d) Reducing oxidative stress and pro-inflammatory molecules

  • (e) Lower risk of severe hypoglycemia