Type 2 diabetes is a big health concern due to its high prevalence and morbi-mortality. Medical treatment has a growing complexity which is focus on patients’ clinical situations. This article contains a consensus statement about recommendations on medical treatment of type-2 diabetes from the Working Group of Diabetes, Obesity and Nutrition of Spanish Society of Internal Medicine. The aim of this consensus is to facilitate therapeutic decision-making to improve the diabetes patients care. The document prioritizes treatments with cardiovascular, especially heart failure, and real benefits.
La diabetes tipo 2 constituye un problema de salud de elevada prevalencia y morbimortalidad. El tratamiento médico tiene una complejidad creciente en relación con las diversas situaciones clínicas del paciente. Este articulo recoge un documento de consenso de las recomendaciones para el tratamiento médico de la diabetes tipo 2 del Grupo de Diabetes, Obesidad y Nutrición de la Sociedad Española de Medicina Interna. El objetivo principal de este articulo es facilitar la toma de decisiones terapéuticas para mejorar la atención de los pacientes con diabetes. El documento prioriza los tratamientos con beneficios cardiovasculares, especialmente la insuficiencia cardíaca, y renales.
Internal medicine departments in Spain were responsible for 705,000 hospital discharges per year, which represents 19% of all discharges, during the period from 2013 to 2016.1 It has been estimated that approximately 30%–35% of all patients hospitalized in internal medicine departments have diabetes mellitus as a secondary diagnosis.2,3
In the Spanish population, the prevalence of diabetes is 13.8% and carbohydrate metabolism abnormalities, including prediabetic states, is 30%.4 This population is elderly, has a high rate of comorbidities, and frequently presents with kidney failure and established cardiovascular disease (CVD).3,5
In order to provide guidelines on the choice of antidiabetic agents, the Diabetes, Obesity, and Nutrition Working Group of the Spanish Society of Internal Medicine has developed recommendations based on a critical review of the most up-to-date literature. These guidelines are not intended to substitute clinical judgment; they must be applied with excellent clinical care, making the necessary adjustments to treatment according to individual preferences, comorbidities, and other patient-related factors.
The main objective is the creation of a brief, practical document that encompasses recommendations for the pharmacological treatment of type 2 diabetes mellitus (DM2) which balances simplicity and scientific rigor in order to facilitate decision-making regarding treatment (Fig. 1).
Materials and methods
The document consists of two parts: the first includes recommendations based on patients’ most frequent clinical conditions: established CVD, heart failure (HF), diabetic nephropathy (DN), obesity and overweight, patients over 75 years of age, those at high risk of hypoglycemic episodes, and time since onset of diabetes greater than 10 years. The second part consists of recommendations based on the therapeutic target for glycosylated hemoglobin (HbA1c) that we wish to reach.
Based on the January 2019 recommendations from the American Diabetes Association,6 we conducted a literature search on PubMed and selected randomized clinical trials on cardiovascular (CV) safety, systematic reviews, and meta-analyses published in 2019 to support the changes made to the aforementioned recommendations. The keywords used were “type 2 diabetes” combined with “cardiovascular disease,” “heart failure,” “diabetic nephropathy,” “obesity,” “elderly,” “hypoglycemia,” “GLP1 receptor agonist,” “SGLT2 inhibitor,” “DPP4 inhibitor,” and “insulin.” The bibliography shows the main works that support recommendations prior to 2019.
To evaluate the level of evidence, American Diabetes Association standards were used with the following classification: A (clear evidence from randomized, controlled trials), B (evidence from observational studies), C (evidence from uncontrolled studies), and E (expert consensus or clinical experience).7 The corresponding letter appears in parentheses at the end of the main recommendations.
Once the algorithm was built and the initial explanations drafted, the original text was debated in successive revisions by a committee of experts who are members of the Diabetes, Obesity, and Nutrition Working Group of the Spanish Society of Internal Medicine until a final consensus on its content was reached.
Results and recommendationsMetformin continues to be the first pillar of antidiabetic treatment and must be recommended so long as there is no intolerance or contraindication. This recommendation is supported by the fact that 75 %–80 % of patients included in the different CV safety studies were administered metformin in combination with the antidiabetic drugs that were studied (A).
Cardiovascular disease/high cardiovascular riskMetformin is considered a safe drug from a CV point of view (C).8 Glucagon-like peptide-1 receptor agonists (GLP-1ras) and sodium-glucose cotransporter-2 inhibitors (SGLT-2is) are the drugs of choice in patients with established CVD and/or multiple CV risk factors, as they have been demonstrated to lead to a reduction in CV events (A) versus a placebo.
In the LEADER study, liraglutide decreased the main composite outcome variable of major adverse CV events (MACE): non-fatal stroke, non-fatal heart attack, and CV mortality (hazard ratio [HR] 0.87; 95% confidence interval [95%CI]: 0.78−0.97; p = .01). On the ungrouped analysis, the relative risk reduction (RRR) was 22% (p = .007) for CV mortality and 15% (p = .02) for total mortality.9 In the SUSTAIN-6 study, semaglutide also decreased MACE (HR 0.74; 95%CI: 0.58−0.95; p = .02), with a RRR of 39% (p = .04) for stroke.10 In the REWIND study, the reduction in MACE (HR 0.88; 95%CI: 0.79−0.99; p = .026) for dulaglutide was also as a result of a RRR of 24% (p = .017) for stroke. The rest of the events analyzed were not statistically significant.11
With respect to the characteristics of the study population, in the LEADER and SUSTAIN-6 studies, 81% and 83% of patients, respectively, had established CVD whereas in the REWIND study, 69% of patients did not have established CVD.
In regard to SGLT-2i, in the EMPA-REG-OUTCOME study, empagliflozin also demonstrated a reduction in MACE (HR 0.86; 95%CI: 0.74−0.99; p = .04). The disaggregated analysis showed a RRR of 38% (p < .001) in CV mortality and 32% (p < .001) in total mortality.12 In the CANVAS study, canagliflozin also decreased MACE (HR 0.86; 95%CI: 0.75−0.97; p = .02), though a disaggregated analysis of its components did not reach statistical significance.13 However, in the DECLARE study, dapagliflozin was not statistically significant with respect to superiority in the reduction of MACE.14 The populations in these studies differed: 100% of patients in the EMPA-REG-OUTCOME study and 65.5% in the CANVAS study presented with established CVD whereas only 40% of patients in the DECLARE study with dapagliflozin did.
In the PROactive study, pioglitazone showed a RRR in MACE of 16% (p = .027) as a secondary objective.15 In the prediabetic population, treatment with pioglitazone was associated with a RRR of 24% (p = .007) of the combined variable of stroke and/or heart attack (A).16
Dipeptidyl peptidase-4 inhibitors (DPP4i) demonstrated noninferiority with respect to standard antidiabetic treatment (A). Although vildagliptin did not have a CV safety study, a meta-analysis supports its neutrality (B).17 In the CAROLINA study, linagliptin showed neutrality versus glimepiride in incidence of MACE, although a greater number of hypoglycemic events were observed in the sulfonylurea group (A).18
With respect to basal insulin, insulin glargine showed CV safety in the ORIGIN study19 and degludec, compared to insulin glargine in the DEVOTE study, showed noninferiority (A).20
Heart failureSGLT-2is are the drugs of choice in patients with HF or who are at risk of developing it (A). In the aforementioned CV safety studies, empagliflozin (HR 0.65; 95%CI: 0.50−0.85; p = .002), canagliflozin (HR 0.67; 95%CI: 0.52−0.87; p < .001), and dapagliflozin (HR 0.73; 95%CI: 0.61−0.88; p = .0005) have demonstrated a significant decrease in hospitalizations due to HF.12–14 In the DAPA-HF study, dapagliflozin managed to reduce the primary composite outcome of CV death, hospitalizations due to HF, and/or emergency department visits due to HF in patients with a reduced ejection fraction (<40%) regardless of the presence of diabetes (HR 0.74; 95%CI: 0.65−0.85; p = .00001).21
GLP-1ras behave in a neutral manner with respect to hospitalizations due to HF, with the exception of albiglutide, which reduced them (HR 0.71; 95%CI: 0.53−0.94; p < .0001) (A).22
DPP-4is were also neutral, with the exception of saxagliptin, which was associated with an increase in hospital admissions due to HF (HR 1.27; 95%CI: 1.07–1.51; p = .007).23 In the CAROLINA study, linagliptin also showed neutrality versus glimepiride in hospitalizations due to HF (A).18
Basal insulin analogs were also demonstrated to be neutral with respect to HF in the aforementioned CV safety studies (A). Pioglitazone is contraindicated in these patients because it is associated with an increase in hospitalizations due to HF as a result of an increase in hydrosaline retention (A).
Diabetic nephropathySGLT-2is reduce the risk of needing dialysis or kidney transplant as well as death due to renal causes (HR 0.67; 95%CI: 0.52−0.86; p = .0019), progression to end-stage kidney disease (HR 0.65; 95%CI: 0.53−0.81; p < .0001), and acute kidney injury (HR 0.75; 95%CI: 0.66−0.85; p < .0001), with benefits that are consistent across all studies (A). In an analysis by glomerular filtration rate (GFR) subgroups, the nephroprotective effects are sustained, even with a GFR of 30–45 mL/min/1.73 m2 (C). Kidney protection is independent of the presence of microalbuminuria and the use of renin-angiotensin system blockers.24 Among the studies on SGLT-2is, only the CREDENCE25 study with canagliflozin 100 mg included a majority of patients with GFR < 60 mL/min/1.73 m2 (60%). These patients were a minority in the DECLARE-TIMI 58 (8%), CANVAS (20%), and EMPA-REG OUTCOME studies (26%).
The GLP-1ras liraglutide (HR 0.78; 95%CI: 0.67−0.92; p = .003), semaglutide (HR 0.64; 95%CI: 0.46−0.88; p = .006), and dulaglutide (HR 0.85; 95%CI: 0.77−0.93; p = .0004) reduced renal events (GFR deterioration, macroalbuminuria, and death due to renal causes). These benefits are mainly due to a reduction in macroalbuminuria (A).22
The remaining antidiabetic agents have not shown benefits in the prevention of kidney function deterioration. Metformin requires a dose reduction with GFR < 45 mL/min and DPP-4is, except for linagliptin, require dose adjustment depending on GFR. The use of pioglitazone is not recommended due to the risk of hydrosaline retention (A).
Metformin, sulfonylureas, and SGLT-2is are contraindicated for GFR <30 mL/min/1.73 m2. Liraglutide, semaglutide, and dulaglutide can be used with GFRs of up to 15 mL/min/1.73 m2 as their beneficial CV and renal effects persist (A). Basal insulin requires a 25% reduction for GFRs of 15−30 mL/min/1.73 m2 and 50% for GFRs <15 mL/min/1.73 m2. Although it can be used in patients with advanced DN, repaglinide can increase the risk of hypoglycemia.
Obesity and overweightObesity and overweight are related to an increase in CV events, HF, and overall mortality.26 GLP-1ras are the drugs of choice. The weight loss obtained using different drugs is not equal; greater to lesser weight loss is achieved using semaglutide, liraglutide, dulaglutide, exenatide, and lixisenatide, in that order (A).
SGLT-2is are also associated with a decrease in body weight, though of a lesser intensity and durability (A). The combined use of GLP-1ras and SGLT-2is is recommended to optimize weight loss objectives (E).
Pioglitazone, sulfonylureas, and insulin are associated with an increase in body weight.27 DPP-4is behave in a neutral manner (A). The role of antidiabetic drugs in the treatment of non-alcoholic fatty liver disease is currently under study; patients in treatment with GLP-1ras and/or SGLT-2is may benefit.28
Patients older than 75 years of ageElderly patients are more vulnerable to the deleterious effects of hypoglycemia and other side effects. Therefore, laxer glucose control objectives must be proposed. However, there is evidence of CV, renal, and HF benefits in elderly patients.
In the EMPA-REG OUTCOME study, the reduction in events in individuals older than 75 years was even greater than among the total population, both in terms of CV mortality (HR 0.55; 95%CI: 0.32−0.94), hospitalizations due to HF (HR 0.45; 95%CI: 0.22−0.89), and renal events (HR 0.54; 95%CI: 0.37−0.79).29 In the CANVAS study, patients older than 65 years of age also had beneficial outcomes in terms of MACE (HR 0.80; 95%CI: 0.67−0.95).13 In a combined analysis of patients older than 75 years of age, canagliflozin demonstrated efficacy and safety with an equal incidence of side effects as younger subjects.30 GLP-1ras also maintained their beneficial effects in elderly patients (C).31
In addition, in the DAPA-HF study, patients older than 75 years of age were those who most benefited in terms of a reduction in hospitalizations due to HF and/or CV mortality (HR 0.68; 95%CI: 0.53−0.88; p = .003) with fewer side effects than the placebo, especially volume depletion and kidney injury (C).32
In the care of elderly patients with DM2, a clinical evaluation must be performed and a functional, cognitive, and social approach should be taken. It is recommended to limit the intensity of antidiabetic treatment in patients with functional or cognitive decline and/or those with a poor prognosis for life. In these cases, the most important aspect is to avoid severe symptomatic hyperglycemia and, most of all, hypoglycemia, by using drugs that have shown low risk of triggering them and, in the event insulin is needed, prioritizing safer ones. For more details, refer to the consensus on treatment of DM2 among the elderly by the Spanish Society of Internal Medicine.33
HypoglycemiaNon-insulin-dependent drugs are to be prioritized due to their low risk of hypoglycemia (metformin, GLP-1ra, SGLT-2is, and DPP-4is) (E). Both insulin glargine U300 and degludec have a lower incidence of hypoglycemic episodes than glargine U100 (A).20,34
Diabetes with onset more than 10 years agoBoth SGLT-2is12–14 and GLP-1ras9–11 have demonstrated beneficial effects in CV safety studies in patients with DM2 with disease onset more than 10 years ago, in contrast to what has been observed with sulfonylureas, metformin, and insulin (A).35–38
HbA1c-dependent approachMetformin continues to be the foundation of antidiabetic treatment. The document recommends the number of drugs that should be used depending on initial HbA1c levels and/or the level we can propose reaching based on the established therapeutic objective (E).
GLP-1ras and/or SGLT-2is are recommended as a second option after metformin, thanks to the aforementioned CV and renal benefits (A). The algorithm recommends different antidiabetic drugs for metabolic control in an established order. Basal insulin can be used in all situations that require it if metabolic control is required, especially if low insulin reserves, glucose levels > 300 mg/dl, cardinal symptoms, or HbA1c >10% with symptoms of hyperglycemia are observed as well as when there is a contraindication or intolerance to other therapies (E).
Even patients with long-duration diabetes (> 10 years) can use GLP-1ras and/or SGLT-2is so long as they are asymptomatic (A).
DiscussionThis document presents the recommendations of the Diabetes, Obesity, and Nutrition Working Group of the Spanish Society of Internal Medicine for decision-making on the pharmacological treatment of DM2. Although it is mainly aimed at internists, it may also be of use to other attending physicians who care for patients with DM2. The high prevalence and exponentially increasing incidence of DM2 mean that any physician may be faced with patients with DM2 in daily clinical practice.
The first part of the algorithm must be read horizontally and vertically. That is, first, the patients’ clinical condition or conditions must be selected (horizontal) and then, the drug it is advisable to use in each situation must be evaluated (vertical). The physician must always evaluate drugs’ contraindications, the need to adjust the dose, and if the indication is approved on the technical datasheet. Furthermore, he or she must also avoid therapeutic inertia. This document makes some recommendations that are currently outside of indications on the technical datasheet, but the existence of solid evidence means that said indications may be revised in the short-term.
With respect to the recommendations to reduce CV events, our document does not include differences in the use of GLP-1ra among patients with or without established CVD, unlike other consensus documents. Although the REWIND study with dulaglutide only had 31% of patients with established CVD and, consequently, the evidence on primary prevention seems to be more consistent than with other GLP-1ras, the differences in criteria in the definition of established CVD make it difficult to establish firm recommendations. The SUSTAIN-6 (83% with CVD) and LEADER (81% with CVD) studies also included presence of vascular stenosis greater than 50% in the coronary, carotid, or peripheral arteries; moderate kidney disease; and HF with functional class II-III as established CVD. On the contrary, these patients were not considered as having established CVD in the REWIND study. Therefore, the percentage of patients in secondary prevention could have been lower in the LEADER and SUSTAIN-6 studies if the same criteria were used as in the REWIND study.39
Although the overall perception of SGLT-2is shows a class effect in the reduction of CV events, in the DECLARE study, dapagliflozin did not achieve a reduction in MACE in the total population, although it was observed in patients with a prior history of acute myocardial infarction.40 One of the explanations could lie in the lower number of patients with prior CVD and advanced DN included in the DECLARE-TIMI study compared to other studies with SGLT-2i.14
Taking into account their mechanisms of action, SGLT-2is could have a cardiorenal benefit through hemodynamic mechanisms whereas GLP-1ras could exercise antiarteriosclerotic action and as such, they would fit in more along the entire continuum of CV risk.
With respect to the reduction in hospitalizations due to HF, SGLT-2is showed a class effect. However, the percentage of patients with prior HF in the initial studies was small; thus, the evidence on the prevention of incident HF is more robust.41 To date, the only specific study on SGLT-2is in a population with HF is the DAPA-HF study, in which dapagliflozin showed a decrease in hospitalizations and/or CV mortality in patients with HF and reduced ejection fraction (< 40%), regardless of presence of diabetes.21
Despite the neutrality results on the CAROLINA study,18 the increase in hypoglycemic events in the glimepiride arm advises against the use of sulfonylureas in patients with established CVD or high CV risk, given that numerous clinical trials and observational studies have linked the use of these antidiabetic drugs with an increase in CV risk.42
With respect to renal effects, the current indications on the SGLT-2i technical datasheet allow for initiating treatment in patients with GFR >60 mL/min/1.73 m2 and maintains it, with a dose reduction, for GFRs of as low as 45 mL/min/1.73 m2. However, the protective renal and CV effects of SGLT-2is are independent of the degree of GFR and are maintained in lower GFR levels. The greatest amount of evidence to date has been provided by the CREDENCE study25 with canagliflozin 100 mg, which included patients with GFRs of up to 30 mL/min/1.73 m2, although these findings are not yet reflected in the technical datasheet. These results reveal an important aspect of renal and CV protection of patients with moderate-severe DN.
Although in addition to CV and renal benefits, GLP1-ras have been demonstrated to reduce body weight both in patients with obesity and overweight, its funding is only authorized for patients with a body mass index ≥30 kg/m2.
In the care of elderly patients with DM2, a clinical evaluation must be performed and functional, cognitive, and social approach should be taken. Patients who, regardless of their age, maintain an adequate functional level should benefit from the same treatments and glycemic control objectives as younger subjects; chronological age is not in and of itself a limitation to treatment.
The intensity of antidiabetic treatment should be limited in patients with functional or cognitive deterioration and/or those with a poor prognosis for life. In these cases, the most important aspect is to avoid severe symptomatic hyperglycemia and, most of all, hypoglycemia, using drugs that have shown low risk of triggering them and prioritizing those that are safer in patients who receive insulin.33
Classically, patients with long-duration DM2 have received insulin. However, new families of treatments provide a possible alternative for receiving therapies with CV and renal benefits.
With respect to the metabolic approach, it is very important to set an HbA1c target that is patient-facing. It is advisable to keep in mind that lifestyle measures and metformin continue to be the pillars of the therapeutic approach. The algorithm helps predict the need to use one, two, or three drugs to reach the objectives. Although classically, insulin was recommended in patients with HbA1c >10% at the time of diagnosis, GLP1-ras and/or SGLT-2is in combination with metformin in many cases allow for proper metabolic control without the need to use insulin or, at least, the use of insulin in a transitory manner. In patients with cardinal symptoms and/or onset of decompensated hyperglycemia, SGLT-2is must not be used due to risk of euglycemic diabetic ketoacidosis.
Identifying insulinopenic patients who require insulin substitution therapy is of utmost importance, as metabolic failure is predicted if we start other therapies. However, insulin therapy in a basal-bolus regimen and the intensification of rapid-acting insulin with multiple doses is not the intention of this consensus.
The goal of this document is to increase the proper prescription of antidiabetic drugs in order to improve prognosis and reduce complications in patients with DM2 in any area of the healthcare system.
The main strength of this document is its simplicity, clarity, and brevity, all of which make it a highly useful tool in clinical practice. On the other hand, its main limitation is the fact that, due to the aforementioned reasons, it cannot provide a response to very specific and/or special clinical circumstances, as it lacks the depth to do so. However, the balance between simplicity and scientific rigor favor its applicability and acceptance of the consensus among the scientific and healthcare community in our setting.
FundingNo funding was received for this study.
Conflicts of interestThe authors declare no conflicts of interest in the creation of and consensus on the recommendations.
Fernández-Rodríguez JM, Carretero-Gómez J, Carrasco-Sánchez FJ, Ena J, Gómez-Huelgas R, León D, Miramontes P, Casado PP, García de Lucas D, Pérez-Soto I, Pérez-Belmonte L, and Formiga F.
Please cite this article as: Carrasco-Sánchez FJ, Fernández-Rodríguez JM, Ena J, Gómez-Huelgas R, Carretero-Gómez J. Tratamiento médico de la diabetes mellitus tipo 2: recomendaciones del Grupo de Diabetes, Obesidad y Nutrición de la Sociedad Española de Medicina Interna. Rev Clin Esp. 2021;221:101–108.
