Italian guidelines for the treatment of type 2 diabetes

Aim: This guideline is aimed at providing a reference for the pharmacological and non-pharmacological treatment of type 2 diabetes in adults. Methods : These recommendations apply to outpatients, either in primary care or at specialist referral. Prior cardiovascular events, heart failure, renal disease, hypoglycemic risk and other conditions affecting life expectancy have been considered as factors capable of modifying treatment strategies. The following areas have been assessed: therapeutic goals, nutritional therapy, physical exercise, educational programs, pharmacological treatment, glucose monitoring. This guideline has been developed following the methods described in the Manual of the National Guideline System (http:// www. snlg- iss. it). For each question, the panel nominated by the Società Italiana di Diabetologia (SID) and Associazione Medici Diabetologi (AMD) identified potentially relevant outcomes, which were then rated for their impact on therapeutic choices. Only outcomes classified as “critical” were considered in the systematic review of evidence and in the formulation of recommendations. Results: The present guideline contains recommendations on the following clinical aspects of type 2 diabetes: 1) treatment

1. Economic interests, i.e., financial relationships with organizations involved in products or services relevant for the subject of the guideline, including any direct payment for services, property shares, stock options, and patents or copyright royalties).
Economic interests can be either: a) personal economic interest, i.e., related to a personal financial benefit; b) familial economic interest, i.e., related to the income of family members; c) institutional economic interests, i.e., related to benefits for the institution in which the subject works.
2. Intellectual interests, i.e., benefits for career advancement and social status.
Both economic and intellectual interests can be specific (i.e., directly related to the subject of the guideline) or aspecific (when they are not related to the content of the guideline).
Any reported potentially conflicting interest is classified as: • Level 1 (minimal or not relevant): no action needed • Level 2 (potentially relevant): this can be managed either with • full participation to the development of the guideline with public disclosure of the conflict of interest at the end of the recommendation related to the interest; • exclusion of the subject with the competing interest form the discussion of those recommendations possibly influenced by the competing interest.
• Level 3 (relevant): this can be managed with the exclusion of the subject with the competing interest from the discussion of possibly affected recommendation, or with the total exclusion of the subject with competing interest from the elaboration of the guideline.

DECLARATION OF POTENTIAL CONFLICTS OF INTEREST
All members of the panel and of the evidence review team compiled annually a declaration of potential conflicts of interest, which were collectively discussed to determine their relevance. In all cases, the reported conflicts were considered minimal or irrelevant (Level 1); therefore, all components of the panel and of the evidence review team participated to the elaboration of all recommendations.
Panel members: Edoardo Mannucci received fees for training activities from Mundipharma and speaking fees from Abbott, Eli Lilly e Novo Nordisk; Riccardo Candido received consulting fees from Boehringer Ingelheim, Eli Lilly, Merck, Menarini and Roche, and speaking fees from Abbott, Eli Lilly, Mundipharma, Novo Nordisk and Sanofi; Andrea Giaccari received consulting fees from Abbott, AstraZeneca, Boehringer Ingelheim, Eli Lilly, Merck, Mundipharma, Novo Nordisk e Sanofi, and his Institution received research grants from Amgen and AstraZeneca; Gerardo Medea received consulting fees from AstraZeneca and Grunenthal; Basilio Pintaudi received consulting and/or speaking fees from Eli Lilly e Novo Nordisk; and Giovanni Targher received consulting fees from Novartis; Giuseppe Turchetti received speaking fees from Eli Lilly, and his Institution received research grants from Merck. Lina Delle Monache, Marco Gallo, Maria Luisa Masini, Angela Mazzone and Marina Trento have no interest to declare.
Evidence review team members: Matteo Monami receives speaking fees from Sanofi; Valentina Lorenzoni has no interest to declare.
External reviewers: Gian Paolo Fadini received research grants from Mundipharma, consulting fees from Abbott, Boehringer, Novo Nordisk and Lilly, and speaking fees from Abbott, Novo Nordisk, Sanofi, Boehringer e AstraZeneca; Gianluca Perseghin received consulting fees from AstraZeneca, Boehringer Ingelheim, Eli Lilly, Merck, Novo Nordisk, PicDare; and Antonio Nicolucci received research grants from Sanofi and Novo Nordisk.

FINANCIAL SUPPORT
No external financial support was collected for the development of this guideline. Travel expenses for panel meeting were paid for by Società Italiana di Diabetologia. Members of panel and evidence review team did not receive any payment for their work in developing the guideline.

AIMS OF THE GUIDELINE
Type 2 diabetes is the most common form of diabetes; its prevalence is rapidly increasing, with a relevant impact on public health. People with type 2 diabetes (over 3 million in Italy) show increased risks of hospitalization, disability and mortality with a yearly cost exceeding 20 billion Euros3.
In Italy, the care of patients with type 2 diabetes is provided by a capillary network of specialist clinics and general practitioners, which warrants a good quality of healthcare. However, some areas still need to be improved: A fraction of patients does not reach therapeutic targets and the management of pharmacological therapy is widely heterogeneous. This heterogeneity is partly determined by the fast development of therapeutic options and clinical evidences; the timely synthesis of those evidences in the format of clinical recommendations and their dissemination among physicians is objectively difficult. The two main dialectological societies in Italy formulated joint guidelines on the management of diabetes in 20,184, without participation of other healthcare professionals involved in the care of diabetes. In addition, other guidelines 5-7 formulated in different organizational contexts are often used by Italian healthcare providers.
This guideline is aimed at providing a reference for pharmacological and non-pharmacological treatment of type 2 diabetes in adults (age of 18 years or more).
Recommendations are designed as indications for healthcare professionals in charge of diabetes treatment, primarily based on clinical needs of people with diabetes and considering the existing organization of healthcare. These recommendations apply to outpatients, either in primary care or at specialist referral. Prior cardiovascular events, heart failure, renal disease, hypoglycemic risk and other conditions affecting life expectancy will be considered as factors capable of modifying treatment strategies.
The following areas will be assessed: therapeutic goals, nutritional therapy, physical exercise, educational programs, pharmacological treatment, glucose monitoring. All the interventions considered are usually reimbursed, with some regional differences for glucose monitoring devices and nutritional therapy. Recommendations will be formulated on the basis of available evidence, independent of current reimbursement policies.
The guideline is directed to physicians, nurses, dietitians and educators working in Diabetes specialist clinics; general practioners; nurses and dietitian working in territorial services or private offices; patients with diabetes. During the development of the guideline, available resources will be considered, verifying the effects of each recommendation on the organization of care and collecting cost-efficacy and cost-utility data whenever possible.
The implementation of the guideline will be pursued through their dissemination, performed by: 1) Scientific societies, using their websites and official journals and organizing specific activities of continuous medical education; 2) regional healthcare systems.

METHODS FOR GUIDELINE DEVELOPMENT
The guideline was developed following the methods described in the Manual of the National Guideline System (http:// www. snlg-iss. it).

Clinical questions
Each recommendation answers a clinical question, formulated by the panel using the PICOS framework.

Selection of outcomes
For each question, the panel identified potentially relevant outcomes, which were then rated for their impact on therapeutic choices using a 9-point scale: 1 3 0-3 points: outcomes of limited relevance 4-6 points: important, but not critical outcomes 7-9 points: critical outcomes.
Only outcomes classified as "critical" were considered in the systematic review of evidences and in the formulation of recommendations. A complete list of outcomes with their scores, for each recommendation, is reported in Appendix. Evidence review and assessment of quality of evidence A systematic review for critical outcomes for each question was performed on the following databases: Specific search strategies were used for each database, as specified in each chapter of Appendix. Searches for pharmacoeconomic studies were limited to the last 10 years, whereas no time limits were imposed for all the other searches. Only items in English were considered. References of retrieved items were searched for further studies meeting inclusion criteria.
The systematic review was performed through the following steps: 1. Selection of potentially eligible studies obtained with the initial search, on the basis of title and abstract, for retrieval as full text; 2. Identification among retrieved full-text items of relevant studies, on the basis of a priori inclusion and exclusion criteria; 3. Critical assessment of the risk of bias using validated instruments (i.e., AMSTAR 2 8 for systematic reviews and the Cochrane collaboration tool 9 for randomized trials). 4. Extraction of the main characteristics of selected studies (population enrolled, considered outcomes, results), summarized in tables. 5. Quantitative synthesis for each outcome, calculating MH-OR for categorical outcomes and WMD for continuous variables, both with 95% confidence intervals. The main analysis was always performed with random effects models, whereas fixed effects models, when used, were considered only for sensitivity analyses; 6. Assessment of heterogeneity (I 2 ) and of publication bias (Funnel plot); 7. The overall quality and strength of available evidence for outcomes selected by the panel were rated using the GRADE 10 criteria. 8. Synthesis of results, using the GRADEPro Guideline Development tool (https:// grade pro. org), with the frameworks EtD 11 , which summarize results of systematic reviews for problem priority, desired and undesired effects of treatments, strength of available evidence, values and preferences of stakeholders, economic resources needed, equity, acceptability and feasibility of interventions.
For pharmacoeconomic studies, relevant records were selected on the basis of title and abstract for full text retrieval. Due to the geographical and methodological heterogeneity of retrieved studies, no formal meta-analysis was performed; methods and results were summarized in tables, including type of analysis, context, year(s) to which costs were referred, efficacy, cost-efficacy and cost-utility, main conclusions.

Development of recommendations
The guideline panel examined and discussed, for each clinical question, EtD frameworks, tables of evidence and summaries of results (forest plots of meta-analyses). Recommendations were formulated on the basis of results of available studies and quality of evidence. Disagreements were resolved through collective discussion.

External review
The panel identified three external reviewers, chosen among Italian healthcare professionals with a specific experience of clinical research in diabetes, with known methodological skills, who had published at least 150 peer-reviewed original articles on International medical journals and who had a h-index of at least 40. Members of the guideline panel and evidence review team, and current members of the Board of SID or AMD, were excluded.
External reviewers received a draft version of the guideline and provided their observations to the panel. The panel collectively discussed the points raised by the external reviewers, elaborating the amendments to the guideline and the response to reviewers.

Guideline update
Systematic reviews will be updated, using the same search strings, once every year, starting from the date of final approval of the guideline. The evidence review team and the guideline panel will verify whether new evidences will modify the risk/benefit ratio or the overall quality of evidences to the extent of modifying the formulation of a recommendation, of its strength or of the quality of evidence.
Once every year, the guideline panel will verify the need to modify, update, add or remove clinical questions, and the opportunity of modifying the outcomes of interest and their relative relevance. In case of changes in clinical questions and/or critical outcomes, the whole process of evidence review and development of recommendation will be performed anew.

Quality of evidence
HIGH: Highly reliable results. It is very unlikely that further studies modify the confidence in estimated effects. MODERATE: Moderately reliable results. It is possible that further studies modify the confidence in estimated effects. LOW: Results are still uncertain. Further research is needed for a reliable assessment of positive and negative effects of the intervention. VERY LOW: Available data are not reliable, and estimates of effects should be considered with caution.

Strength of recommendations
Strong recommendation • for clinicians: the majority of patients must receive the recommended intervention; • for patients: almost all properly informed patients follow the recommendation and only a small fraction choses different options; • for policy makers: the recommendation can be used for planning the use of available resources.

Weak recommendation
• for clinicians: the final choice should include a careful consideration of patients' values and preferences; • for patients: the majority of properly informed patients follow the recommendation, but a minority choses different options; • for policy makers: a discussion involving stakeholder should be developed.

Treatment targets
Strength of the recommendation: strong. Quality of evidence: low. 5. Pharmacological treatment 5.1 We recommend the use of metformin as first-line long-term treatment in patients with type 2 diabetes, without previous cardiovascular events. SGLT-2 inhibitors or GLP-1 receptor agonists are recommended as second-line treatments. Pioglitazone, DPP-4 inhibitors, acarbose and insulin should be considered as third-line treatments (Figure 1).
Strength of the recommendation: strong. Quality of evidence: moderate. 5.2.1. We recommend the use of metformin, SGLT-2 inhibitors or GLP-1 receptor agonists as first-line long-term treatment in patients with type 2 diabetes with previous cardiovascular events and without heart failure. DPP-4 inhibitors, pioglitazone, acarbose and insulin should be considered as second-line treatments ( Figure 1).
Strength of the recommendation: strong. Quality of evidence: moderate. 5.2.2. We recommend the use of SGLT-2 inhibitors as first-line long-term treatment in patients with type 2 diabetes with previous heart failure. GLP-1 receptor agonists and metformin should be considered as secondline treatments. DPP-4 inhibitors, acarbose and insulin should be considered as third-line treatments (Figure 1).
Strength of the recommendation: strong. Quality of evidence: moderate. 5.3. We recommend the use of basal insulin analogues, instead of NPH, for all patients with type 2 diabetes needing treatment with basal insulin.
Strength of the recommendation: strong. Quality of evidence: very low.

We suggest the use of prandial insulin analogues for patients with type 2 diabetes needing treatment with prandial insulin.
Strength of the recommendation: weak. Quality of evidence: very low. 5.5. The routine use of continuous subcutaneous insulin infusion in inadequately controlled patients with type 2 diabetes is not recommended.
Strength of the recommendation: weak. Quality of evidence: very low.
6. Glycemic monitoring 6.1 We suggest to structure (with a pre-defined scheme of required tests) capillary blood glucose self-monitoring in the treatment of type 2 diabetes.
Strength of the recommendation: weak. Quality of evidence: very low.
6.2. We do not suggest a continuous glucose monitoring (continuous or on demand) rather than self-monitoring blood glucose in patients with type 2 diabetes on basal-bolus insulin therapy.
Strength of the recommendation: weak. Quality of evidence: very low.

THERAPEUTIC TARGETS 1.1 HbA1c target in patients treated with drugs inducing hypoglycemia
Question: Which is the target HbA1c in patients with type 2 diabetes who are not treated with drugs capable of inducing hypoglycemia (insulin, sulfonylureas, glinides)?  Justification. Several randomized trials show that the intensification of glucose control prevents long-term complications of diabetes, suggesting the need to reach and maintain HbA1c levels below 58 mmol/mol (7.5%). Lower targets (i.e., HbA1c < 48 mmol/mol or 6.5%) further reduce the risk of microvascular complications, but not of cardiovascular disease or mortality; however, a very strict glycemic control increases the risk of severe hypoglycemia, with an unfavorable risk/benefit ratio. For this reason, the most convenient HbA1c range for patients treated with drugs capable of inducing hypoglycemia is between 69 and 58 mmol/mol (6.6-7.5%). Higher targets can be considered for patients aged > 75 years or with reduced life expectancy because of comorbidities.

Subgroup considerations.
There are no available data from randomized trials on the safety and efficacy of intensification of glucose control in patients aged > 75 years; in addition, benefits of long-term glucose control are evident only after 2 years of treatment. This could motivate higher HbA1c targets in patients aged > 75 years or with reduced life expectancy because of comorbidities.
Implementation. Specific programs for continuous medical education should be planned, to increase the awareness of healthcare professionals of the benefits of adequate glycemic control and the risks associated with very low HbA1c values in patients treated with hypoglycemia-inducing drugs.
Assessment and monitoring. Adherence to this guideline can be assessed by estimating the proportion of patients at HbA1c target in existing databases.

Assessment Problem
Is the problem a priority? Judgment Research evidence Additional considerations Yes The reduction of HbA1c levels in type 2 diabetes is associated with a lower risk of macroand microvascular complications and mortality 12,13 . However, there is a wide heterogeneity of results obtained with different strategies, in particular when using treatments associated or not with hypoglycemic risk 12

RECOMMENDATION (1.2.1): A target HbA1c below 53 mmol/mol (7%) is recommended for patients with type 2 diabetes not treated with drugs capable of inducing hypoglycemia.
Strength of the recommendation: strong. Quality of evidence: low.
Justification. Several randomized trials show that the intensification of glucose control prevents long-term complications of diabetes, suggesting the need to reach and maintain HbA1c levels below 53 mmol/mol (7.0%). In particular, accurate glycemic control appears to reduce the risk of cardiovascular disease, with a variable cost/benefit ratio.
Subgroup considerations. There are no available data from randomized trials on the safety and efficacy of intensification of glucose control in patients aged > 75 years; in addition, benefits of long-term glucose control are evident only after 2 years of treatment. This could motivate higher HbA1c targets in patients aged > 75 years or with reduced life expectancy because of comorbidities.
Implementation. Specific programs for continuous medical education should be planned, to increase the awareness of healthcare professionals of the benefits of adequate glycemic control.
Assessment and monitoring. Adherence to this guideline can be assessed by estimating the proportion of patients at HbA1c target in existing databases 1,2 .

RECOMMENDATION (1.2.2):
A target HbA1c of 48 mmol/mol (6.5%) or lower is suggested for patients with type 2 diabetes treated with drugs that are not capable of inducing hypoglycemia.
Strength of the recommendation: strong. Quality of evidence: low.
Justification. No randomized trials assessed the effect of reaching and maintaining HbA1c ≤ 48 mmol/mol with drugs not capable of inducing hypoglycemia. Conversely, trials with hypoglycemia-inducing drugs show that the reduction of HbA1c below 48 mmol/mol prevents microvascular complications of diabetes. Pharmacoeconomic studies suggest that the achievement of this target, when obtained with drugs that do not induce hypoglycemia, reduces the need for hospitalization for diabetic complications, thus reducing overall health expenditure.
Subgroup considerations. There are no available data from randomized trials on the safety and efficacy of intensification of glucose control in patients aged > 75 years; in addition, benefits of long-term glucose control are evident only after 2 years of treatment. This could motivate higher HbA1c targets in patients aged > 75 years or with reduced life expectancy because of comorbidities.
Implementation. Specific programs for continuous medical education should be planned, to increase the awareness of healthcare professionals of the benefits of adequate glycemic control.
Assessment and monitoring. Adherence to this guideline can be assessed by estimating the proportion of patients at HbA1c target in existing databases 19,20 . Assessment for HbA1c < 53 mmol/mol (7%) Problem Is the problem a priority? Judgment Research evidence Additional considerations Yes The reduction of HbA1c levels in type 2 diabetes is associated with a lower risk of macroand microvascular complications and mortality 12,13 . However, there is a wide heterogeneity of results obtained with different strategies, in particular when using treatments associated or not with hypoglycemic risk 12 Justification. A small number of available trials, with methodological limitations and with relatively small sample size, show small but significant improvements in glycemic control and body weight with structured Medical Nutrition Therapy (MNT, composed of nutritional assessment, diagnosis, intervention and monitoring) when compared to unstructured nutritional advice. The low quality of evidence and the methodological biases of available studies limit the strength of this recommendation. Economic resources needed for implementation are negligible since unstructured nutritional advice is also time-consuming.
Subgroup considerations. There are no available data from randomized trials on the safety and efficacy of MNT in patients aged > 75 years; in addition, patients with mental disorders and/or cognitive impairment could receive greater benefits from a traditional prescription of a diet, provided to the caregiver(s).
Implementation. The awareness of healthcare professionals of the benefits of MNT could be increased by specific educational programs. The inclusion of MNT among indicators of the quality of care for diabetes could be of help in increasing adherence to this recommendation.
Assessment and monitoring. Justification. Few studies with methodological biases and a small number of included patients show small, but significant advantages on glycemic control of a balanced (Mediterranean) diet, when compared to a low-carbohydrate diet. The low quality of evidence and the methodological biases of available studies limit the strength of this recommendation. Economic resources needed for implementation are assumed as negligible, although no specific pharmacoeconomic studies were retrieved.
Subgroup considerations. No data are available on the long-term renal safety of low-carbohydrate diets. Patients with renal impairment are usually excluded from clinical trials.
Implementation. The awareness of healthcare professionals of the advantages of a balanced diet could be increased by specific educational programs.
Assessment and monitoring. The monitoring of this recommendation is problematic.
Research priorities. Further trials with good methodological quality comparing balanced and low-carbohydrate diets and assessing renal function among predefined outcomes are needed, to increase the strength of this recommendation. Assessment Problem Is the problem a priority? Judgment Research evidence Additional considerations Justification. Several epidemiological studies showed beneficial effects of physical exercise on health outcomes, including the reduction of HbA1c and body weight, with no side effects and relevant costs, in type 2 diabetes 29 . The quality of available evidence is sufficient for drawing a recommendation, but some methodological flaws and the scarce number of patients included in the available studies downgrade the strength of this guideline.
Subgroup considerations. There are no available data from randomized trials on the safety and efficacy of physical exercise in elderly patients.
Implementation. The awareness of healthcare professionals of the benefits of physical exercise could be increased by specific educational programs. The inclusion of physical exercise among indicators of the quality of care for diabetes could be of help in increasing adherence to this recommendation.
Assessment and monitoring. The monitoring of this recommendation is problematic. The recommendation of practicing physical exercise can be added during the routine visits

Aerobic physical exercise and duration
Question: Which is the minimum recommended duration of aerobic physical exercise for diabetes control in patients with type 2 diabetes?

RECOMMENDATION:
We suggest combined (aerobic and resistance) training, rather than aerobic training alone, for the treatment of type 2 diabetes. Strength of the recommendation: weak. Quality of evidence: low.
The preference for combined aerobic and resistance training was based on the greater reduction of HbA1c reported in available trials. The small between-group difference in HbA1c and the small sample size limit the strength of this recommendation. No issues of sustainability or equity were identified. The quality of available evidence is poor because of the limited sample size and of some methodological issues in clinical trials. Subgroup considerations. Some subpopulations of patients with type 2 diabetes (e.g., advanced age, heart failure, etc.) could benefit more from other modalities of physical exercise different from aerobic training.
Implementation. The medical community should be made aware of the potential advantages of combined aerobic/anaerobic training through CME programs dedicated to non-pharmacological treatments of type 2 diabetes.
Assessment and monitoring. The monitoring of adherence to guidelines on recommendations regarding nonpharmacological interventions and lifestyle behavior is problematic. Justification. The preference for grouped-based educational programs is based on the possible better glycemic control, weight loss, quality of life and reduced costs. The quality of available evidence is poor because of the limited sample size and of some methodological issues in clinical trials, thus reducing the strength of this recommendation.

Assessment
Subgroup considerations. Few available data on elderly patients do not allow to assess the efficacy of the structured educational therapy in the advanced decades. Patients with psychiatric disorders or cognitive impairment could benefit more from traditional education often managed by caregivers.
Implementation. The medical community should be made aware of the potential advantages of structured educational therapy through CME programs dedicated to nonpharmacological treatments of type 2 diabetes.
Assessment and monitoring.

Group-and individual-based educational therapy
Question: Should group-based educational therapy be preferable in comparison with individual therapy for diabetes control in patients with type 2 diabetes?

Population
People with type 2 diabetes Intervention Group-based educational therapy

RECOMMENDATION:
We suggest grouped-based educational programs, rather than individual, for the treatment of type 2 diabetes. Strength of the recommendation: weak. Quality of evidence: very low.
Justification. The preference for grouped-based educational programs is based on the possible better quality of life and reduced costs. There is no effect on HbA1c, thus limiting the strength of this recommendation.
Subgroup considerations. The possibility that some subgroup of patients can have some advantages on glucose control cannot be completely ruled out. Group-based therapy could determine better glycemic control in programs with longer duration and in non-insulin-treated patients with lower baseline HbA1c levels. Conversely, available clinical trials do not include very old patients, those with cognitive impairment and those with major psychiatric conditions.
Implementation. The medical community should be made aware of the potential advantages of a macronutrientbalanced diet through CME programs dedicated to nonpharmacological treatments of type 2 diabetes.
Assessment and monitoring.

PHARMACOLOGICAL THERAPY 5.1 Glucose-lowering therapy in patients with type 2 diabetes and no previous cardiovascular events
Which glucose-lowering agents should be considered as first-, second-and third-line therapy for glycemic control in patients with type 2 diabetes and no previous cardiovascular events?

Population
People with type 2 diabetes Intervention Glucose-lowering therapy Comparison Glucose-lowering therapy Outcome HbA1c, hypoglycemia, medium/long-term adherence, mortality; major cardiovascular events Setting Outpatient

RECOMMENDATION: We recommend the use of metformin as a first-line longterm treatment in patients with type 2 diabetes without previous cardiovascular events. SGLT-2 inhibitors or GLP-1 receptor agonists are recommended as secondline treatments. Pioglitazone, DPP-4 inhibitors, acarbose and insulin should be considered as third-line treatments.
Strength of the recommendation: strong. Quality of evidence: low. Justification. A major body of evidence from randomized controlled trials supports the use of metformin, SGLT-2 inhibitors, or GLP-1 receptor agonists as first-line treatment in patients with type 2 diabetes due to relevant efficacy in reducing HbA1c without increasing the risk of hypoglycemia and less risk of MACE and all-cause mortality. Moreover, GLP-1 receptor agonists and SGLT-2 inhibitors also have beneficial effects on body weight. Insulin secretagogues have shown a lower efficacy in reducing HbA1c with a higher risk of hypoglycemia in comparison with metformin; in addition, a higher mortality rate was observed in comparison with other glucose-lowering agents/placebo, and therefore, their use should be avoided for the treatment of type 2 diabetes. The quality of available evidence is generally satisfactory. Several good-quality pharmacoeconomic studies showed that metformin has the lowest direct costs in comparison with other classes of glucose-lowering agents which have similar clinical effects.
Subgroup considerations. This recommendation provides more than one option for both second-and third-line therapy. The choice among available options can be affected by patients' characteristics such as age, renal failure, body weight, duration of diabetes, comorbid conditions, diabetic complications, etc., or by clinical conditions (e.g., high degree of hyperglycemia) based on clinicians' Judgment.
Implementation. Sulfonylureas should not be added to ongoing therapy; existing treatments with sulfonylureas should be progressively deprescribed or substitutes with other therapies irrespective of glycemic control.
The whole medical community should be made aware of this recommendation to homogenize the therapy for type 2 diabetes in line with evidence-based medicine. Continuous medical education programs are needed to implement the knowledge of physicians in this respect.
Assessment and monitoring. The monitoring of adherence to guidelines on the pharmacological treatment of type 2 diabetes can be implemented through the consultation of existing databases 7,8

RECOMMENDATION:
We recommend the use of metformin, SGLT-2 inhibitors, or GLP-1 receptor agonists as first-line long-term treatment in patients with type 2 diabetes with previous cardiovascular events and without heart failure. DPP-4 inhibitors, pioglitazone, acarbose, and insulin should be considered as second-line treatments.
Strength of the recommendation: strong. Quality of evidence: moderate.

Justification.
A major body of evidence from randomized controlled trials supports the use of metformin, SGLT-2 inhibitors, or GLP-1 receptor agonists as first-line treatment in patients with type 2 diabetes due to relevant efficacy in reducing HbA1c without increasing the risk of hypoglycemia and less risk of MACE and all-cause mortality. In particular, SGLT-2 inhibitors in comparison with metformin and GLP-1 receptor agonists, have favorable effects on the risk of hospitalization for heart failure. Moreover, GLP-1 receptor agonists and SGLT-2 inhibitors also have beneficial effects on body weight. Insulin secretagogues have shown a lower efficacy in reducing HbA1c with a higher risk of hypoglycemia in comparison with metformin; in addition, a higher mortality rate was observed in comparison with other glucose-lowering agents/placebo, and therefore, their use should be avoided for the treatment of type 2 diabetes. The quality of available evidence is generally satisfactory. Several good-quality pharmacoeconomic studies showed that metformin has the lowest direct costs in comparison with other classes of glucose-lowering agents; moreover, metformin and SGLT-2 inhibitors and, to a lesser extent, GLP-1 receptor agonists have a good cost-effective ratio.
Subgroup considerations. This recommendation provides more than one option for both second-and third-line therapy. The choice among available options can be affected by patients' characteristics such as age, renal failure, body weight, duration of diabetes, comorbid conditions, diabetic complications, etc., or by clinical conditions (e.g., high degree of hyperglycemia) based on clinicians' Judgment.
Implementation. Sulfonylureas should not be added to ongoing therapy; existing treatments with sulfonylureas should be progressively deprescribed or substitutes with other therapies irrespective of glycemic control. The whole medical community should be made aware of this recommendation to homogenize the therapy for type 2 diabetes in line with evidence-based medicine. Continuous medical education programs are needed to implement the knowledge of physicians in this respect.
Assessment and monitoring. The monitoring of adherence to guidelines on the pharmacological treatment of type 2 diabetes can be implemented through the consultation of existing databases.

Question #2
Which glucose-lowering agents should be considered as first-, second-and third-line therapy for glycemic control in patients with type 2 diabetes and previous heart failure?

Population
People with type 2 diabetes Intervention Glucose-lowering therapy Comparison Glucose-lowering therapy

RECOMMENDATION:
We recommend the use of SGLT-2 inhibitors as first-line long-term treatment in patients with type 2 diabetes with previous heart failure. GLP-1 receptor agonists and metformin should be considered as second-line treatments.

DPP-4 inhibitors, acarbose and insulin should be considered as third-line treatments.
Strength of the recommendation: strong. Quality of evidence: moderate.
Justification. A major body of evidence from randomized controlled trials supports the use of metformin, SGLT-2 inhibitors, or GLP-1 receptor agonists as first-line treatment in patients with type 2 diabetes due to relevant efficacy in reducing HbA1c without increasing the risk of hypoglycemia and less risk of MACE and all-cause mortality. In particular, SGLT-2 inhibitors in comparison with metformin and GLP-1 receptor agonists, have favorable effects on the risk of hospitalization for heart failure. Moreover, GLP-1 receptor agonists and SGLT-2 inhibitors also have beneficial effects on body weight. Insulin secretagogues have shown a lower efficacy in reducing HbA1c with a higher risk of hypoglycemia in comparison with metformin; in addition, a higher mortality rate was observed in comparison with other glucose-lowering agents/placebo, and therefore, their use should be avoided for the treatment of type 2 diabetes. The quality of available evidence is generally satisfactory. Several good-quality pharmacoeconomic studies showed that metformin has the lowest direct costs in comparison with other classes of glucose-lowering agents; moreover, metformin and SGLT-2 inhibitors and, to a lesser extent, GLP-1 receptor agonists have a good cost-effective ratio.
Subgroup considerations. This recommendation provides more than one option for both second-and third-line therapy. The choice among available options can be affected by patients' characteristics such as age, renal failure, body weight, duration of diabetes, comorbid conditions, diabetic complications, etc., or by clinical conditions (e.g., high degree of hyperglycemia) based on clinicians' Judgment. Metformin can be used only in patients with NYHA < III. Saxagliptin should be avoided due to the high risk of hospitalization for heart failure.
Implementation. Sulfonylureas should not be added to ongoing therapy; existing treatments with sulfonylureas should be progressively deprescribed or substitutes with other therapies irrespective of glycemic control. The whole medical community should be made aware of this recommendation to homogenize the therapy for type 2 diabetes in line with evidence-based medicine. Continuous medical education programs are needed to implement the knowledge of physicians with this respect.
Assessment and monitoring. The monitoring of adherence to guidelines on the pharmacological treatment of type 2 diabetes can be implemented through the consultation of existing databases.

Assessment (both for questions #1 and #2)
Problem Is the problem a priority? Judgment Research evidence Additional considerations

RECOMMENDATION:
We recommend the use of basal insulin analogues, instead of NPH, for all patients with type 2 diabetes needing treatment with basal insulin. Strength of the recommendation: strong. Quality of evidence: very low.
Justification. A major body of evidence from randomized controlled trials supports the use of basal insulin analogues due to less risk of total and nocturnal hypoglycemia, with a trend toward reduction of severe hypoglycemia. Despite the treat-to-target design of the majority of RCT, a modest positive effect on HbA1c and FPG was observed (detemir e glargine U100). There are no available trials comparing newer basal insulin analogue formulations with NPH insulin. However, comparisons between glargine U100 and the newer formulations of insulin (degludec and glargine U300) show similar, and for same endpoints, more favorable effects for these latter two insulin formulations. Therefore, the recommendation to use basal insulin analogues, instead of NPH insulin, can be extended also to degludec and glargine U300.
The quality of available evidence is generally low, particularly due to the open-label design of the majority of the included trials and to the presence of heterogeneity.
Pharmacoeconomic studies showed that direct costs of drugs is generally increased with newer formulations despite the cost-effectiveness ratio generally suggest good value for money because of the implication in terms of both QALY and the effects on the risk of events, weight gain etc.; the availability of biosimilars contains the cost of out-of-patent insulin analogues.
Subgroup considerations. No available evidence in patients aged over 75 years.
Implementation. Long-acting analogues are already the standard of care. The prescription of NPH insulin should be strongly discouraged, with specific educational program for non-specialists, recommending its substitution with longacting analogues.
Assessment and monitoring. The monitoring of adherence to guidelines on pharmacological treatment of type 2 diabetes can be implemented through the consultation of existing databases. control. The quality of available evidence is low, and the limited sample size and some methodological issues in clinical trials downgrade the strength of the evidence. There is no expected difference in required resources.

Assessment
Subgroup considerations. There are few available data from randomized trials on the safety and efficacy of structured glucose in elderly patients. Patients with psychiatric disorders and cognitive impairment could benefit more from traditional educational prescription, often managed by caregivers.
Implementation. The awareness of healthcare professionals of the benefits of structured glucose monitoring could be increased by specific educational programs. The inclusion of structured glucose monitoring among indicators of the quality of care for diabetes could be of help in increasing adherence to this recommendation.
Assessment and monitoring. The monitoring of this recommendation is problematic. Assessment