The International Diabetes Federation estimated that worldwide nearly 250 million people aged 20–79 years had diabetes in 2007 . Approximately 85–95% of patients in developed countries and an even higher proportion in developing countries have type 2 diabetes . Some 3.8 million people worldwide were estimated to have died from diabetes-related causes in 2007 . About 50–70% of people with diabetes die of cardiovascular disease (CVD) [2,3]. Control of blood glucose, generally measured in terms of glycated haemoglobin (HbA1c) levels, has been shown to reduce microvascular complications of type 2 diabetes, such as nephropathy and retinopathy, and (to a lesser extent), macrovascular events [4,5]. In view of these benefits, guidelines for the management of diabetes recommend tight control of blood glucose with a target HbA1c level of 6.5–7.5% [6–9] (approximately 48–58 mmol/mol on the standardized measurement of the International Federation of Clinical Chemistry [IFCC]) [10,11]. Control of other risk factors, such as hypertension and dyslipidaemia, is also indicated. In the United Kingdom Prospective Diabetes Study (UKPDS) dietary intervention maintained glycaemic control in approximately 25% of patients with type 2 diabetes 3 years after diagnosis, but became less effective with time . Pharmacological treatment eventually becomes necessary for most patients. Clinical trials and physician experience show that many patients do not reach their goal for glycaemic control. This review discusses this challenge and how patients can be helped to reach treatment goals, particularly in the light of new approaches to patient care and the availability of new therapeutic options.
How Many Patients Are Not Reaching Treatment Targets?
Although control of blood glucose in people with diabetes has improved in recent years, at least in the UK , treatment fails to help a substantial proportion of patients to achieve goals for glycaemic control and risk factors, such as total cholesterol (TC) and blood pressure (BP) (Table 1) [14–18]. Comparisons between studies should be made with caution because of differences between populations, targets and study techniques, but the results shown in Table 1 suggest that a target HbA1c <6.5% (≈48 mmol/mol) is appreciably more difficult to achieve than targets of 7.0% (≈53 mmol/mol) or 7.5% (≈58 mmol/mol).
|Study||Patient population (data source)||n||HbA1c||BP||TC|
|Target (%) (mmol/mol)||Above target (%)||Target (mmHg)||Above target (%)||Target (mmol/l)||Above target (%)|
|NHANES  (USA)||Adults, type 1 or 2 diabetes (NHANES study)||1334||<7.0% (53 mmol/mol)||43||—||—||—||—|
|Southeast London (longitudinal study) ||≥18 years, type 1 or 2 diabetes (regional primary care database)||4284||≤7.0% (53 mmol/mol)||63||<140/80||58||≤5.0||30|
|UK cross-sectional study ||Type 1 or 2 diabetes (national primary care database)||1 852 762||≤7.4% (57 mmol/mol)||40–43||≤145/85||29–31||≤5.0||27–30|
|UK National Audit ||Type 1 or 2 diabetes, children or adults (GP practices, hospital trusts and paediatric units)||656 000||≤7.5% (58 mmol/mol)||42||<160/110||10||≤5.0||27|
|<6.5% (48 mmol/mol)||77||≤135/75||73|
|Swedish National Register ||18–79 years, type 2 diabetes and coronary heart disease (national diabetes register)||1414*||<7.0% (53 mmol/mol)||46||≤130/80||60||<4.5||40|
What Is the Optimal Level of Glycaemic Goal?
Three recent studies—Action in Diabetes and Vascular Disease: Preterax and Diamicron Modified Release Controlled Evaluation (ADVANCE) , the Veterans Affairs Diabetes Trial (VADT)  and Action to Control Cardiovascular Risk in Diabetes (ACCORD) —compared intensive glycaemic control with standard therapy in high-risk patients with established type 2 diabetes. The mean HbA1c levels with intensive therapy were 6.5% (≈48 mmol/mol) , 6.9% (≈52 mmol/mol)  and 6.4% (≈46 mmol/mol) in ADVANCE, VADT and ACCORD, respectively . Compared with standard therapy, ADVANCE  and VADT  reported no significant reduction in macrovascular events. In ACCORD, intensive therapy was associated with significantly higher all-cause mortality than conventional therapy (5.0% vs. 4.0% over a mean follow-up of 3.5 years) . By contrast, the UKPDS reported after 10 years' follow-up that intensive therapy significantly (p ≤ 0.05) reduced microvascular disease and macrovascular events compared with conventional therapy in newly diagnosed patients . These benefits were seen long after the differences between treatment arms in HbA1c were lost (the ‘legacy effect'). The different outcomes in these studies may have been related to disease duration, the speed of HbA1c reduction or the regimens employed . The UKPDS enrolled newly diagnosed patients, whereas in ADVANCE, VADT and ACCORD the patients had type 2 diabetes of on average 7–11 years' duration. A recent meta-analysis concluded that intensive glycaemic control reduces coronary events in patients with type 2 diabetes compared with standard control treatment, with no effect on overall mortality . The latest guidelines from the UK National Institute for Health and Clinical Excellence (NICE) reflect this evidence and advocate a target HbA1c <6.5% (≈48 mmol/mol) for patients treated with a single oral agent and <7.5% (≈58 mmol/mol) for patients taking two or more oral drugs or needing insulin; these goals can be adapted to the individual patient .
Implications of the Pathology of Type 2 Diabetes
Progressive failure of insulin secretion from pancreatic β-cells plays a major role in the pathology of type 2 diabetes. An investigation in a subgroup of patients from the UKPDS indicated that participants had already lost approximately 50% of their β-cell secretory function (estimated by the homeostasis model assessment) at the time of diagnosis . β-Cell function in patients with type 2 diabetes continues to decline after diagnosis [24–26]. Insulin resistance, which increases as individuals progress from normal glucose tolerance through impaired glucose tolerance (IGT) to diabetes, also plays a part in the development of the condition [25,26]. Insulin resistance alone might not lead to overt diabetes provided that a compensatory increase in insulin secretion can be maintained . Failure of glucose control after an initial response to dietary therapy may be primarily related to the decline in β-cell function . Progression of β-cell dysfunction may thus explain why treatment becomes less effective with the passage of time. In the UKPDS, monotherapy maintained glycaemic control in only 25% of patients 9 years after diagnosis, and by this time, a substantial proportion of patients required the addition of insulin .
Adherence to Treatment in Patients with Type 2 Diabetes
Poor patient adherence is an important barrier to glycaemic control. Retrospective studies in people with type 2 diabetes reported adherence rates of 36–93% for oral agents and 62–64% for insulin . In Scotland, the Diabetes Audit and Research in Tayside Study examined adherence, glycaemic control and clinical outcomes in an unselected population of people with diabetes. Among people with type 2 diabetes, adherence was approximately 80% with metformin, 70% with insulin and (after the first quarter) 60% with statins [30–32]. The same study reported that only approximately 30–35% of patients took >90% of their prescribed dose of metformin or sulphonylurea . An increase in the daily number of tablets and the use of combination regimens rather than a single agent are generally associated with lower rates of adherence [29,33,34].
Patient Perceptions of Diabetes and its Treatment
The Diabetes Education and Self-Management for Ongoing and Newly Diagnosed (DESMOND) type 2 diabetes programme is a structured education programme for patients with type 2 diabetes in the UK. In this programme, only 54% of newly diagnosed patients considered diabetes a serious condition, 33% admitted that they did not understand the disease, 73% were unclear about symptoms and very few reported undertaking recommended levels of physical activity . Patient perceptions of diabetes predict metabolic control and psychological well-being . Issues with regard to insulin therapy include: anxiety about injections; fear of adverse events (AEs); and a belief that the need for insulin means that they have managed their condition poorly, that treatment has failed or the disease has progressed [37–39].
Psychological Issues in Diabetes
Psychological issues, including depression, are common in people with type 2 diabetes and are associated with poor adherence and self-care [40–43]. Disability resulting from diabetes is at least partly responsible for its association with depression . In the cross-national Diabetes Awareness Wishes and Needs (DAWN) study, 41% of people with type 2 diabetes had poor psychological well-being but only 12% reported receiving psychological treatment in the past 5 years . The majority of healthcare providers were aware of psychological morbidities in people with diabetes, but few referred their patients to psychological services and many reported lacking the resources to deal with these problems .
Issues in the Clinician–Patient Relationship
The DAWN study reported that 50–55% of general practitioners and nurses delay giving patients insulin therapy for as long as possible. This reluctance to prescribe insulin may be associated with low confidence in its efficacy: just over half of physicians and nurses agreed that insulin can have a positive impact on care . Clinicians who delay prescribing insulin also tend to delay oral therapy, suggesting a general reluctance to introduce pharmacological treatment . Delay of insulin therapy was significantly less common among specialists and in clinicians who believed insulin to be efficacious. Healthcare professionals' lack of time, fear of AEs or poor confidence in their own or patients' ability to manage insulin also discourage initiation of therapy . Some clinicians use inappropriate techniques in an attempt to encourage adherence, such as adopting a paternalistic attitude, frightening patients, threatening to refer them to hospital or using the possibility of insulin therapy as a threat .
The frequency of hypoglycaemia reported in the literature for patients with type 2 diabetes varies widely because of differences in study design, patient selection, definitions of hypoglycaemia and the treatments administered . The risk of hypoglycaemia depends on the glucose level attained and on the mechanism of glucose reduction. Agents that increase the plasma insulin concentration independently of the ambient glucose level, such as exogenous insulin and insulin secretagogues, are associated with higher rates of hypoglycaemia than metformin, acarbose or thiazolidinediones [46,47]. The HbA1c level attained with monotherapy is not an accurate predictor of hypoglycaemia. For example, in UKPDS 24, the HbA1c level was similar between treatment groups, but the annual rate of major hypoglycaemic episodes was 0.4% with metformin, 2.1–2.2% with insulin and 0.4–0.9% with sulphonylureas . In A Diabetes Outcome Progression Trial (ADOPT), the HbA1c level was significantly lower with rosiglitazone than in the other treatment arms, whereas serious hypoglycaemic events were self-reported by 0.1%, 0.1% and 0.6% of newly diagnosed patients randomized to monotherapy with rosiglitazone, metformin or glyburide, respectively . Intensification of therapy increases the risk of hypoglycaemia, depending on the status of the patient and the regimen used [19,20]. In a recent international survey, 38% of patients receiving metformin in combination with a thiazolidinedione or a sulphonylurea reported hypoglycaemic symptoms in the past year, and this was associated with lower treatment satisfaction and more barriers to adherence . ADVANCE , VADT , ACCORD  and UKPDS 33  reported significantly higher rates of hypoglycaemia with intensive than with conventional therapy: ADVANCE, UKPDS 33 and ACCORD also reported greater weight gain in the intensive arms (Table 2). In VADT, severe hypoglycaemia was predictive of the primary outcome and of CVD mortality . Exploratory analyses of ACCORD have not found an explanation for the excess mortality with intensive therapy, but it has been suggested that it may have been at least in part related to hypoglycaemia .
|Study||Mean HbA1c at study end||Hypoglycaemia||Mean weight change (kg/year)*|
|ACCORD ||7.5% (≈58 mmol/mol)#||6.4% (≈46 mmol/mol)#||Requiring any assistance||1.5%/year†||4.6%/year†||+0.1||+1.0|
|ADVANCE ||7.3% (≈56 mmol/mol)#||6.5% (≈48 mmol/mol)#||Blood glucose <2.8 mmol/l (50 mg/dl) and transient CNS dysfunction requiring help from another person||0.4%/year†||0.7%/year†||−0.2||−0.02|
|VADT ||8.4 (≈68 mmol/mol)#||6.9 (≈52 mmol/mol)#||With symptoms||3.8/pt-year‡||13.3/pt-year‡||+0.7||+1.5|
|UKPDS 33 ||7.9 (≈63 mmol/mol)#||7.0 (≈53 mmol/mol)#||Requiring third-party or medical assistance||0.7%||1.4% (glibenclamide) 1.8% (insulin)||+0.2 (glibenclamide)§ +0.4 (insulin)§|
Characteristics such as older age, longer duration of diabetes, presence of comorbidities and loss of residual insulin secretion can increase a patient's vulnerability to hypoglycaemia . Impairment of counter-regulatory responses may also predispose patients with type 2 diabetes to hypoglycaemia [46,47]. Some patients keep their glucose level higher than recommended in order to avoid hypoglycaemia  and concern about hypoglycaemia discourages some clinicians from prescribing insulin .
Weight gain associated with drug treatments for type 2 diabetes (or the fear of weight gain) may offset the benefits of therapy, discourage adherence, increase diabetes-related distress and deter clinicians from prescribing certain drugs [52–54]. Insulin therapy is consistently associated with an increase in body weight, typically approximately 2–3 kg over 4–12 months after initiation . The weight gain appears to be proportional to the insulin dose and the improvement in glycaemic control. Postulated mechanisms include conservation of calories by correction of glycosuria, anabolic actions of insulin and an increased calorie intake by patients in an attempt to avoid hypoglycaemia (‘defensive snacking’) . Insulin secretagogues, including sulphonylureas and glinides, are also associated with weight gain [51,53,55–57]. Acarbose appears to be weight neutral . Metformin is associated with weight neutrality or modest weight loss [49,55,59], possibly resulting in part from gastrointestinal AEs leading to a reduction in appetite . The thiazolidinediones are also associated with weight gain, possibly resulting from fluid retention [49,53,55,61]. Weight gain can occur with agents that increase plasma insulin levels and with drugs that do not (e.g. thiazolidinediones), indicating that mechanisms other than elevation of insulin levels can induce weight gain.
How Can Glycaemic Control Be Improved?
Approaches to improve glycaemic control include earlier detection of type 2 diabetes, structured education programmes, improved healthcare delivery and use of more recent therapies including incretin-based agents, thiazolidinediones and long-acting insulin analogues.
Earlier Detection of Type 2 Diabetes. Increased screening may detect people with diabetes earlier in their disease course and enable more of them to attain glycaemic control. In the UK, a vascular risk assessment and management programme has been proposed, based on screening everyone aged between 40 and 75 years in the community . An important issue is whether to lower the threshold of the test and to screen for IGT and type 2 diabetes together. A modelling analysis in the UK indicated that this would be more cost-effective than no screening or screening for diabetes alone . Efforts should be made to detect individuals with IGT, because lifestyle and pharmacological interventions can delay the development of type 2 diabetes in this population .
Structured Patient Education. The DESMOND programme is a recent example of structured patient education in the UK. Patients in the community with newly diagnosed type 2 diabetes received 6 h of education from a trained educator. At 12 months, structured education was associated with significantly greater improvements in smoking cessation and weight loss compared with usual care; there was no significant difference in HbA1c. The CVD risk score (calculated with the UKPDS risk engine) was reduced significantly more in the structured education group than in the usual care group . Another structured education programme in people with established type 2 diabetes (X-PERT) was shown to significantly improve glycaemic control, TC and body weight compared with a control group . Group-based education has also been reported to improve glycaemic control in patients with type 2 diabetes . Structured education is also recommended when insulin therapy is initiated in patients with type 2 diabetes . Patients with type 2 diabetes treated with insulin have significantly lower concerns about this agent than insulin-naïve patients , which suggests that once transferred to insulin, patients' perceptions of this treatment improve. Insulin pens improve treatment satisfaction, adherence and clinical outcomes compared with vial and syringe administration .
Improved Healthcare Delivery. Computerized recall systems for patients and general practitioners may improve glycaemic control . Self-monitoring of blood glucose has not been shown to significantly improve glycaemic control in people with type 2 diabetes managed with oral therapy , but it should be routinely offered to those treated with insulin . Electronic monitoring systems also improve treatment adherence . In the UK, a contract for general practitioners in the National Health Service offers payment by results if they deliver specified interventions, including measurement and management of HbA1c, in patients with diabetes. Comparison of electronic general practice records before and after introduction of this contract suggests that payment by results significantly improves the control of blood glucose, TC and BP (figure 1) [15,71].
Potential of Newer Agents to Improve Glycaemic Control
Recently introduced agents for glycaemic management in type 2 diabetes include incretin-based therapies, thiazolidinediones and basal analogue insulins.
Incretin-based Therapies. The incretin-based therapies fall into two types: dipeptidyl peptidase 4 (DPP-4) inhibitors (incretin enhancers) and glucagon-like peptide-1 (GLP-1) agonists (incretin mimetics). Both classes were developed after studies of the incretin response—the phenomenon by which an oral glucose load produces a greater stimulation of insulin secretion in healthy humans than intravenous administration of the same glucose load [72,73]. This response is principally mediated by the intestinal peptides GLP-1 and glucose-dependent insulinotropic polypeptide. Effects of GLP-1 include stimulation of glucose-dependent insulin secretion, inhibition of gastric emptying and reduction of food intake [72–74]. Circulating levels of this peptide decrease rapidly because of its inactivation by the enzyme DPP-4. The DPP-4 inhibitors are oral agents that protect GLP-1 from inactivation and thereby increase circulating levels of the peptide. The GLP-1 receptor agonists, administered by subcutaneous injection, stimulate GLP-1 receptors and are resistant to DPP-4 [72–74]. At present, three DPP-4 inhibitors (saxagliptin, sitagliptin and vildagliptin) and two GLP-1 agonists (liraglutide and exenatide) are licensed for use in patients with type 2 diabetes in Europe. More DPP-4 inhibitors (including alogliptin) and GLP-1 agonists (including taspoglutide and albiglutide) are under development .
DPP-4 Inhibitors. Compared with placebo, both sitagliptin and vildagliptin significantly reduce HbA1c (by approximately 0.5–1.0% [≈6–11 mmol/mol] compared with baseline) . Significant reductions in fasting plasma glucose (FPG) were also observed and the DPP-4 inhibitors increased the proportions of patients reaching HbA1c goals . Sitagliptin and vildagliptin have been reported to improve markers of β-cell function and to reduce systolic BP . DPP-4 inhibitors are generally well tolerated . Liver function test monitoring and caution in patients with renal impairment are required with vildagliptin . In clinical trials, the risk ratios for serious AEs did not differ significantly between the DPP-4 inhibitor and control groups . The incidence of hypoglycaemia is low when DPP-4 inhibitors are combined with metformin or a thiazolidinedione, but increases when they are combined with a sulphonylurea . The low incidence of hypoglycaemia in the absence of a sulphonylurea may be explained by the observation that the effects of GLP-1 on insulin secretion are glucose- dependent and the counter-regulatory release of glucagon in response to hypoglycaemia is preserved even in the presence of pharmacological concentrations of GLP-1 .
GLP-1 Agonists. Compared with placebo, exenatide and liraglutide significantly reduce HbA1c (by approximately 0.8–1.5% [≈9–14 mmol/mol] from baseline), FPG and postprandial glucose (PPG) levels, and significantly increase the proportions of patients reaching HbA1c goals [77,79–81]. Incretin mimetics also improve surrogate markers of β-cell function and reduce systolic BP by approximately 2–3 mmHg vs. baseline in most studies . The reduction in systolic BP with liraglutide is observed within 2 weeks of therapy initiation and it cannot be explained entirely by reductions in body weight . In contrast to treatment with DPP-4 inhibitors, which are weight neutral, exenatide and liraglutide reduce body weight by a mean of approximately 2–3 kg compared with baseline, possibly by inhibition of gastric emptying, increased feelings of satiety and reduced food intake . Metformin appears to increase plasma GLP-1 levels in patients with type 2 diabetes via inhibition of DPP-4 [83,84]; it is not clear whether this effect contributes to weight loss with metformin.
The most common AEs of GLP-1 agonists are gastrointestinal events, such as nausea and vomiting, which may result from delayed gastric emptying. Nausea occurs in approximately 45–51% of patients receiving exenatide and 7–40% of those receiving liraglutide . In clinical trials, approximately 2–11% of patients discontinued these agents because of nausea [79–81,85–87]. However, gastrointestinal AEs tend to decline with continued administration [79–81,85–87]. Rare cases of acute pancreatitis have been reported with exenatide and liraglutide but their clinical significance is unclear . The incidence of hypoglycaemia is low with liraglutide and exenatide when given with metformin or a thiazolidinedione, but increases if they are combined with a sulphonylurea [78,88,89]. Approximately 40–50% of patients treated with exenatide and 0–13% of those treated with liraglutide develop antibodies against the GLP-1 agonist . Approximately 6% of patients receiving exenatide develop high titres of antibodies; attenuation of the response during was seen in about half of these individuals in Phase III trials .
Thiazolidinediones. Thiazolidinediones modulate the peroxisome proliferator-activated receptor γ which regulates the transcription of genes involved in glucose and lipid regulation. They reduce blood glucose principally by increasing the sensitivity of muscle, fat and liver tissue to insulin [90,91], although they also improve markers of β-cell function . Pioglitazone and rosiglitazone appear to be as effective as metformin and sulphonylureas in the reduction of HbA1c and they improve glycaemic control when added to conventional treatments .
In the short term, the AEs associated with rosiglitazone and pioglitazone treatment include weight gain, fluid retention, peripheral oedema and expansion of plasma volume . Longer-term AEs include an increased risk of bone fractures in women [9,92]. In the Prospective Pioglitazone Clinical Trial in Macrovascular Events (the PROactive Study), pioglitazone was associated with an increase in hospital admissions for heart failure . The significance of this observation is unclear: pioglitazone was not associated with increased mortality in patients with serious heart failure in this trial  and studies have not indicated an increased myocardial infarction (MI) rate with pioglitazone . Pioglitazone appears to be associated with more favourable effects on lipid profiles than rosiglitazone .
Some meta-analyses reported a significantly increased risk of MI with rosiglitazone [96,97], but the absolute event rates were low and other analyses have not reproduced this finding . In the Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycaemia in Diabetes (RECORD) trial, 4447 patients with type 2 diabetes were randomized to rosiglitazone, added to metformin or a sulphonylurea, or to metformin plus a sulphonylurea . After a mean follow-up of 5.5 years, there was no significant difference between the two arms in the primary endpoint of cardiovascular hospitalization or cardiovascular death, or in the rate of MI, but heart failure was significantly (p = 0.001) more common with rosiglitazone .
Long-acting Insulin Analogues. Insulin detemir and insulin glargine are long-acting insulin analogues prepared by modifying human insulin to change its solubility. This allows slow release of insulin into the bloodstream from subcutaneous tissue and hence a longer duration of action, which more closely mimics natural basal insulin secretion. Long-acting insulins are as effective as neutral protamine Hagedorn (NPH) insulin in the reduction of HbA1c in patients with type 2 diabetes and are associated with lower rates of hypoglycaemia [9,100].
A German cohort study, submitted for publication in 2008, reported a dose-dependent increase in cancer risk for patients treated with insulin glargine compared with human insulin . This study had serious limitations, including short follow-up (median 1.3 years for insulin glargine), the failure to correct for confounders, including body mass index (BMI), smoking, social status and duration of diabetes, and the lack of data on the type of tumour [101,102]. In view of these limitations, the journal (Diabetologia) decided that publication should be dependent on other studies being performed . These were: a population-based study in Sweden ; a cohort study in Scotland ; and a primary care database study in the UK .
The Swedish study reported a higher incidence of breast cancer in users of insulin glargine monotherapy compared with users of other insulin monotherapy, but the number of cases was low (25 on insulin glargine and 183 on insulins other than glargine) and data on important possible confounders were missing . The Scottish study found higher rates of cancer overall in patients using insulin glargine monotherapy than in those using other insulin monotherapy, but the former group were older than other patients, with worse glycaemic control . The authors concluded that allocation bias probably influenced their results . The UK primary care database showed no increase in cancer rates with insulin analogues compared with human insulin . An excess risk was seen in people with type 2 diabetes receiving insulin or insulin secretagogues compared with those receiving metformin, but this study did not correct for key confounders of increased malignancy such as BMI or smoking .
An overview of all these studies drew attention to the limitations of observational studies and concluded that there is insufficient evidence that insulin causes cancer, and no evidence for an overall increase in cancer development associated with insulin glargine .
Recommendations for the Use of Newer Agents
Guidelines for the management of type 2 diabetes recommend metformin as the initial pharmacological therapy for most patients [7–9]. When first-line therapy fails, guidelines agree that treatment should be intensified and adapted to the needs of each patient, although they differ in the details of their recommendations [7–9]. Efficacy in glycaemic control is the first consideration when selecting therapy, but this must be balanced against the risks of weight gain and hypoglycaemia, which are important factors in patient adherence and quality of life. As discussed above, these risks are not predicted by the reduction in HbA1c (Table 3).
|Class||Reduction in HbA1c with monotherapy (%)*||Risk of hypoglycaemia||Effects on body weight|
|Metformin||1.0–2.0 ||Low [46,47]||Neutral or modest loss |
|Insulin||1.5–3.5 ||High [46,47]||Gain |
|Sulphonylureas||1.0–2.0 ||Moderate [46,47]||Gain |
|Thiazolidinediones||0.5–1.4 ||Low [46,47]||Gain |
|Glinides||0.5–1.5 ||Moderate [46,47]||Gain |
|α-Glucosidase inhibitors||0.5–0.8 ||Low [46,47]||Neutral |
|DPP-4 inhibitors||0.5–1.0 ||Low [78,89]||Neutral |
|GLP-1 agonists||0.8–1.1 ||Low [77,88]||Modest loss |
Table 4 summarizes treatment recommendations for DPP-4 inhibitors, GLP-1 agonists and thiazolidinediones from the joint American Diabetes Association and European Association for the Study of Diabetes (ADA/ EASD) , NICE  and the American Association of Clinical Endocrinologists (AACE) .
|Guideline||Added to metformin||Added to sulphonylurea||Added to metformin + sulphonylurea||Added to insulin (± other oral agents)|
|ADA/EASD||+ (pioglitazone)||−†||+ (pioglitazone)||−|
NICE suggests that a long-acting insulin analogue can be considered for selected patients, generally where a reduction in the frequency of injections or the avoidance of hypoglycaemia could be beneficial . The ADA/EASD guidelines include long-acting insulins as an alternative to intermediate-acting insulins, and acknowledge that they modestly reduce hypoglycaemia compared with NPH insulin . The AACE guidelines include long-acting insulins as an option when insulin is indicated .
The DPP-4 inhibitors and GLP-1 analogues are suitable for selected patients in primary care, although they are not currently licensed as monotherapy. Both classes are generally well tolerated, with a low risk of hypoglycaemia or weight gain. Use of these agents does not necessitate self-monitoring of blood glucose. Liraglutide and exenatide are supplied in dosing pens for subcutaneous injection: experience with insulin suggests that many patients can accept and use these devices.