Randomized controlled clinical trials that compared intensive to conventional treatment in T2D, such as the United Kingdom Prospective Diabetes Survey (UKPDS) and the Kumamoto Study, established glycemic targets of diabetes treatment associated to better long-term outcomes. Although these studies present different epidemiological data, clinical interventions and outcomes, all agree with the fact that the reduction of blood glucose is effective in decreasing microvascular and neuropathic chronic complications.

The UKPDS study, that evaluated 5.112 T2D patients during 20 years, demonstrated a decrease in chronic complications with intensive treatment, with a reduction of approximately 1 % in A1c levels (7.9–7.0 %), resulting in a 25 % reduction in the risk of microvascular complications and 16 % of acute myocardial infarction in 10 years [2], and also demonstrated benefits of the intensive blood pressure control [45].

Another study (STENO-2), with T2D in intensive care, with A1c goals of <6.5 %, blood pressure <130/80 mmHg, statin use (total cholesterol <180 mg/dL and LDL <100 mg/dL) and inhibitors of angiotensin-converting enzyme (in cases of persistent microalbuminuria), showed a 53 % reduction in the risk of cardiovascular events [46].

The DECODE study showed that postprandial hyperglycemia is an independent risk factor for mortality [47]. Mainly, the prevention of microvascular and macrovascular complications in T2D patients requires tight control of fasting, preprandial and postprandial blood glucose, A1c, lipids, and blood pressure. However, in older patients with previous cardiovascular disease and longstanding diabetes, the glycemic control may not be so strict, as demonstrated by the ACCORD study, which showed an increase in cardiovascular mortality and other causes in T2D patients in intensive care [48].

The ADA recommends the targets for T2D treatment, preprandial glucose between 90 and 130 mg/dL and A1c levels <7.0 % [49, 50]. However, the American Association of Clinical Endocrinologists (AACE) [51] and the International Diabetes Federation (IDF) [52] suggest A1c levels <6.5 %. To achieve these goals, the proper maintenance of fasting glucose, in addition to postprandial glucose through basal–bolus insulin, is an important factor in the treatment of subjects with T2D [53].

The pathophysiology of pharmacological treatment of subjects with T2D depends on various aspects contributing to hyperglycemia. The peripheral insulin resistance (adipocytes and skeletal muscle), presented in 85–90 % of the cases [53], the deficiency of the insulin production by the beta cell and the excessive hepatic glucose production caused by the insulin resistance, are all contributor factors to the onset of hyperglycemia. As shown in UKPDS, the persistence of elevated levels of A1c leads to progressive loss of beta cell function in secreting insulin [2].

The T2D treatment begins with monotherapy and subsequently a second drug is added and, if necessary, a third oral drug or insulin is implemented. However, the new consensus algorithm, proposed by both societies (ADA and EASD) recommend that insulin might be implemented earlier, shortly after oral monotherapy. This treatment initially begins by changes in lifestyle (diet, cholesterol control, weight reduction, pressure control, physical activity and tobacco control) associated to metformin [54]. If these interventions are not effective in reducing A1c levels <7.0 %, another approach must be implemented 3 months after the initiation of metformin. However, there is no consensus on the second drug used, being insulin, sulfonylurea, thiazolidinedione (pioglitazone), or DPP-4 inhibitors, some of the treatment options available.

Insulin could be indicated as a second drug added after metformin, when the patient presents intense clinical evidence (polyuria, polydipsia, polyphagia, weight loss) of decompensation and/or A1c ≥8.5 %. However, if the clinical evidence is less intense, and/or A1c levels <8.5 %, another oral drug can be added before using insulin.

To start insulin therapy in T2D patients, it is recommended to start with 10 U (or 0.2 U/kg of body weight) of a basal insulin (NPH, glargine or detemir) at bedtime, maintaining the oral antidiabetic agents already being used [55]. In a study comparing the use of glimepiride associated with insulin glargine in the morning or at bedtime or bedtime NPH insulin in 695 patients with T2D, A1c levels decreased 1.24 % with glargine in the morning, 0.96 % using glargine at bedtime and 0.84 % using NPH insulin at bedtime [56]. The improvement in A1c was more evident with glargine in the morning than at bedtime (p = 0.008) and than NPH at bedtime (p = 0.001).

The randomized Treat-to-Target study, which compared insulin glargine or human NPH insulin to oral therapy in subjects with T2D, showed similar results of A1c levels (6.96 % vs. 6.97 %) and fasting glucose (117 mg/dL vs. 120 mg/dL) [57]. The insulin adjustment was performed every 3 days, according to the blood glucose tests. In the LANMET study, with T2D patients with inadequate glycemic control, on use of oral antidiabetics (90 % with sulfonylureas associated to metformin) and without previous treatment with insulin, 110 subjects were randomized to receive insulin glargine at bedtime plus metformin or NPH insulin plus metformin for 36 weeks [58]. The initial dose of insulin was 10 U for those using previously only metformin and 20 U for those using metformin associated with sulfonylurea. The individuals randomized to receive insulin glargine showed lower fasting plasma glucose than NPH group (103.5 mg/dL vs. 107.3 mg/dL, p < 0.001) [58]. In this study, the insulin dose was increased by 2 U every 3 days if the fasting plasma glucose (FPG) ≥100 mg/dL within 3 days and 4 U if FPG ≥180 mg/dL in the same period, and successively. The patient was responsible for his insulin adjustment or self-titration, maintaining their optimal dose without endocrinologist visits and telephone calls.

In the UKPDS, patients with T2D receiving insulin therapy had lower A1c levels, but 1.0–2.0 % more patients receiving insulin reported at least one episode of severe hypoglycemia per year than those patients receiving other therapies. Intensive therapy, with oral medications or insulin, has been shown to increase the risk of episodes of hypoglycemia.

The GOAL study, involving 7.893 patients with T2D, uncontrolled on oral antidiabetic randomized to four treatment groups with long-acting insulin analog glargine, involving different forms of titration and different ways of measuring A1c, found significant reductions in A1c and blood glucose in all groups (p < 0.0001) [59]. However, the group with active titration showed a greater reduction in A1c than those with usual titration (1.5 % vs 1.3 %, p < 0.0001) and a greater proportion of patients achieved an A1c <7.0 % (38 % vs. 30 %, p < 0.0001) [59].

In patients with T2D using only one dose of NPH at bedtime, it is possible to use twice a day (morning and bedtime) when goals are not being achieved, before starting to use fast-acting insulin. When the basal insulin at bedtime or twice a day associated to oral antidiabetics is no longer sufficient to maintain A1c levels <7.0 %, it is necessary to intensify insulinization. This means starting with fast-acting insulin (human or analogues), maintaining the basal or biphasic insulin. When insulinization is intensified, oral insulin secretagogues (sulfonylurea, glinides) must be suspended, but metformin is maintained.

In T2D, a study using insulin glargine as basal insulin demonstrated that an fast-acting insulin could be started initially at the main meal, as it causes the greatest increase in postprandial glycemia, or results in the highest elevation of premeal glucose, then extending this insulin to the other meals, if necessary, to achieve A1c goal [60].

In a study with T2D uncontrolled patients at diagnosis, which compared the use of NPH insulin at bedtime plus regular insulin before meals and NPH plus lispro, the use of the insulin analog showed to be superior in achieving metabolic control, with suppression of glucagon secretion and reduction of the glucotoxicity [61]. Another study also observed that use of insulin lispro in patients with T2D, administered before meals, was more effective in reducing A1c than the use of metformin or NPH insulin at bedtime. In a study comparing the insulin analogues lispro and aspart, the results evidenced that both have the same effectiveness in controlling postprandial glucose excursions [62]. The administration of insulin aspart 15 min after the meal is as or more effective in controlling postprandial hyperglycemia than implementing the regular insulin before meals [28]. Greater predictability of action and lower glycemic variation with the use of these analogues have also been described in individuals with T2D [63]. Individuals with T2D treated for 26 weeks with the long-acting analog insulin detemir associated to aspart insulin before meals, showed glycemic control comparable to those treated with NPH and aspart. However, as demonstrated in other studies, there was less variability between individuals, and less weight gain with insulin detemir [64, 65].

Regarding the use of biphasic insulin in subjects with T2D, studies demonstrated that the biphasic analogues are more effective in reducing the postprandial hyperglycemia than those containing NPH and regular, without significant reduction of A1c [66, 67]. However, the improvement of postprandial hyperglycemia and glycemic variability reduction, despite A1c levels, may be important factors in reducing the risk of onset and progression of microvascular and macrovascular chronic complications. The use of biphasic insulins (premix) may be recommended for those patients with the greatest difficulty in assimilating the basal–bolus regimen, considered apparently to be more convenient because the doses are already previously divided on a fixed percentage of basal and fast-acting insulin analogues 70/30, 75/25, or 50/50.

When patients with T2D need to use basal–bolus regimen, as in T1D, meaning that these patients have severe insulinopenia, and glucose variability occurs frequently, continuous subcutaneous insulin infusion could be indicated.

Hyperglycemia is a common acute complication of critically ill patients, during hospitalization at intensive care units (ICUs). Some studies demonstrated that the presence of hyperglycemia, in particular severe hyperglycemia, is associated with increased morbidity and mortality in some patients, but some clinical and randomized trials evaluating the effects of tighter glucose on the mortality of critically ill patients showed conflicting data.

Despite these conflicting evidences, the American Diabetes Association, the American Association of Clinical Endocrinologists, Brazilian Diabetes Society, and other professional organizations recommend intensive insulin therapy as the standard of care for critically ill patients [68–70].

The main barrier to implement a tight glucose control in patients during hospitalization and critically ill, is the increased risk of severe hypoglycemia that has been described to be associated to increase mortality. Because of these aspects and risks or benefits, the tight glucose control is used infrequently by some clinicians. The Normoglycemia in Intensive Care Evaluation—Survival Using Glucose Algorithm Regulation (NICE-SUGAR) study, an international multicentre trial involving 6,104 patients, largest randomized clinical trial of intensive insulin therapy, evidenced that blood glucose target ≤180 mg/dL is associated to lower mortality than did a target of 81–108 mg/dL [71].

It is recommended that all clinicians assess all patients admitted to the hospital for a history of diabetes. When present, this diagnosis should be clearly identified in the medical record. All patients, independent of a prior diagnosis of diabetes, have laboratory blood glucose (BG) testing on hospital admission. The patients without a history of diabetes with BG greater than 7.8 mmol/L (140 mg/dL) should be monitored with bedside capillary BG testing for at least 24–48 h. Those with BG greater than 7.8 mmol/L require appropriate therapeutic intervention [68].

The inpatient care of individuals with diabetes and hyperglycemia is complex, involving multiple providers with varying degrees of expertise who are dispersed across many different areas of the hospital. Multidisciplinary local protocols should guide safe glycemic control, hypoglycemia prevention, and patient preparation for care transitions. Poor coordination of glucose monitoring, meal delivery, and insulin administration is a common barrier to optimal care of the patients during the hospitalization period.