Inpatient Treatment of Type 2 Diabetes

The prevalence of diabetes mellitus in Germany is 9.3% in women and 8.2% in men, according to the latest health statistics of the Robert Koch Institute (RKI) (1). Diabetes is more common in hospitalized patients than in the general population, with a prevalence of about 30% (2). The treatment of diabetes mellitus in the hospital is thus an everyday issue. The existing guidelines of the medical specialty societies cannot be directly implemented in the hospital, as they primarily concern the long-term treatment of patients with diabetes in early stages and without any major accompanying illnesses (3, 4). Most hospitalized patients with diabetes are in advanced stages of the disease, and many have accompanying illnesses and sequelae of diabetes that must be taken into account in therapeutic decision-making (5). A further difficulty arises from the ever shorter times that patients spend in the hospital (in 2009, the mean length of stay was only 8.0 days) (5). Long-term pharmacotherapy is practically impossible during such a short period of hospitalization, to say nothing of treatment based on changes of lifestyle.

The major goals of treatment during a patient’s stay in the hospital include

• the safe regulation of the blood glucose concentration with moderate target values,
• the strict avoidance of hypoglycemia, and
• the initiation of long-term antidiabetic treatment or the optimization of existing treatment.

The aim of this article is to provide an overview of the purpose and scope of medical treatment options for diabetes in multimorbid hospitalized patients in the light of published guidelines from Germany and abroad, Cochrane reviews, a selective Medline search, and manufacturer-provided information on individual antidiabetic drugs.

The evidence for each of the main statements of this article will be rated with the GRADE system for evaluating evidence and recommendations in guidelines (GRADE, Grades of Recommendation, Development and Evaluation). In this system, the evidence for desired and undesired effects of treatment is designated as being of high, intermediate, low, or very low quality, while recommendations are designated as either strong or weak (6). The evidence ratings for the main recommendations given here are shown in Table 1.

Table 1

Main recommendations for the treatment of diabetes mellitus in hospitalized patients, with evidence grades

Enlarge All figures

Glycemic targets

The German Diabetes Association (Deutsche Diabetesgesellschaft, DDG) recommends an HbA_1c value less than or equal to 6.5% as a long-term target in the outpatient treatment of type 2 diabetes. This target should be reached in three to six months. The importance of avoiding hypoglycemic episodes is explicitly mentioned; weight gain should also be avoided. An HbA_1c value up to 7% is acceptable if lower values can only be attained at the risk of side effects (4). The long latency of the HbA_1c value disqualifies it from use for blood-sugar monitoring in inpatients.

Intensified insulin regimens for glycemic control with near-normal target values have been tested for use on hospitalized patients in various situations. A basic distinction should be drawn between acutely ill patients in intensive care units and clinically stable patients on regular hospital wards.

Early trials on critically ill patients on a surgical intensive care unit (ICU) suggested a significant survival advantage for intensified insulin therapy with blood glucose targets of 80−110 mg/dL, rather than standard therapy with targets of 180−200 mg/dL (e1). These findings, however, could not be replicated in later, more extensive trials (e2, e3). A current consensus paper of the American Diabetes Association (ADA) therefore recommends that, in critically ill patients, insulin therapy should be initiated if the blood glucose concentration exceeds 180 mg/dL, with a target value of 140 to 180 mg/dL (3). The recommended target values for clinically stable patients are below 140 mg/dL preprandially and below 180 mg/dL postprandially (3). Lower targets can also be aimed at in individual patients, if the risk of hypoglycemia is considered to be low (3).

Pharmacotherapy

Although the ADA recommends transient insulin treatment for most hospitalized patients (3), It would nonetheless appear reasonable to continue the patient’s home regimen in the hospital if the hospital stay is expected to be short and without any metabolic derangement (e4, e5). The main oral antidiabetic drugs and their dosages and routes of elimination are listed in Table 2.

Table 2

Available drugs for the treatment of type 2 diabetes (as of 10 January 2012) (7)

Enlarge All figures

A number of situations will now be discussed in which the antidiabetic regimen must be adjusted during hospitalization.

Fasting periods

After a period of fasting (“nil per os,” NPO), oral antidiabetic drugs should be resumed only when the patient eats the first normal-sized meal. Special recommendations are in effect for metformin and sulfonylureas. Metformin, according to the manufacturer’s instructions, should be stopped 48 hours before the intended procedure and resumed only when the patient is reliably tolerating oral food intake, as long as there are no other contraindications (4, 7). When a patient who has been taking metformin requires emergency surgery, the metabolic situation must be monitored more closely than usual in the perioperative period (e6). If periods of fasting are already envisioned at the beginning of the patient’s hospital stay, temporarily stopping treatment with sulfonylureas (SU) is recommended because of the increased risk of hypoglycemia (8). As some SUs have a longer-lasting effect (up to 72 hours), the blood glucose concentration should be checked frequently. When oral antidiabetic agents are temporarily discontinued, insulin can be given: the blood glucose concentration can be controlled either with subcutaneous insulin injections or with intravenous insulin through a perfusor pump, depending on the length and invasiveness of the procedure (8). Hourly blood glucose determinations are recommended when a perfusor is used. The simultaneous infusion of a solution containing glucose seems reasonable, in order to assure adequate delivery of nutritional substrate to the cells (8). Detailed recommendations on the use of intravenous insulin infusion regimens are summarized in Table 3 and in references (8) and (9).

Table 3

Insulin therapy recommendations for inpatients with type 2 diabetes in special clinical situations; modified from references (8, 24) Neutral

Enlarge All figures

Radiological contrast studies

As the use of iodinated radiological contrast media can temporarily lower the glomerular filtration rate, potentially inducing renal failure, metformin should be discontinued at least 24 hours before the radiological study and restarted no sooner than 48 hours afterward, and only if laboratory tests confirm normal renal function. Emergency studies can be performed even if the patient has been taking metformin (e7). It would also seem reasonable to discontinue sulfonylureas temporarily before contrast studies, as these drugs are mainly eliminated through the kidneys.

High-dose glucocorticoid treatment

The insulin requirement increases during high-dose glucocorticoid treatment; in particular, the insulin dose must be adjusted postprandially and 8 to 12 hours after glucocorticoid administration. More detailed recommendations are given in Table 3 and in references (8) and (9).

Coronary heart disease

The effect of near-normal glycemic control on the microvascular complications of type 2 diabetes (e.g., nephropathy and retinopathy) has been documented in large-scale studies (e8, e9). The data on macrovascular complications are inconsistent. Intensified blood-glucose reduction had no significant effect on the cardiovascular endpoints of the UKPD trial (e8), nor did intensified antihyperglycemic treatment significantly lower the frequency of macrovascular endpoints in the ADVANCE trial (e10). The ACCORD trial actually revealed increased mortality among the more intensively treated patients (e11). In the VADT trial, the two treatment groups did not differ in the frequency of cardiovascular events (e12). Meta-analyses of the large-scale endpoint trials have shown a significant advantage for more intensive glycemic control only with respect to non-fatal myocardial infarction (10). Further analysis indicates that excessive reduction of the HbA_1c concentration can be dangerous, particularly for elderly patients who have had diabetes for many years and for patients who have had a prior coronary event. Hypoglycemia is thought to induce catecholamine-mediated arrhythmias and other ECG changes (11). Thus, strict avoidance of hypoglycemia seems advisable in patients with coronary heart disease (CHD). Sulfonylureas and glinides should be used with caution in view of the danger of hypoglycemia and of potential interactions with cardiac sulfonylurea receptors (e13, e14). With regard to metformin, a subgroup analysis in the UKPD trial revealed a significant reduction of cardiovascular endpoints among overweight patients (12). For patients with acute coronary ischemia, temporary discontinuation of oral hypoglycemic treatment is recommended, in view of the potential need for a radiological contrast study and the increased risk of lactic acidosis.

Dipeptidyl-peptidase-4 (DPP-4) inhibitors and glucagon-like-peptide-1 (GLP-1) analogues can be used for glycemic control with practically no risk of hypoglycemia (e15, e16). Small-scale clinical trials have shown improved cardiac performance in CHD patients after the administration of a DPP-4 inhibitor or during the infusion of native GLP-1 (e17, e18). Preliminary safety studies of the currently available DPP-4 inhibitors also suggest that these may lower the frequency of cardiovascular events (Lamanna C, et al. Dipeptidyl peptidase-4 inhibitors and cardiovascular events: a protective effect? European Association for the Study of Diabetes 2011 Meeting; Lisbon, Portugal. Abstract 244). These results still need to be validated in endpoint trials, and thus no recommendation can yet be given with regard to the use of these drugs in patients with CHD. In case of an acute coronary event, the recommendation is to stop all oral antidiabetic medication temporarily and switch to subcutaneous insulin administration (8). Current data do not support any general recommendation for intravenous therapy with insulin, glucose, and potassium (e19, e20).

Congestive heart failure

Congestive heart failure is at least twice as prevalent among patients with diabetes mellitus as in the normal population (13). Diastolic dysfunction frequently arises at an early stage of diabetes, often as early as the stage of reduced glucose tolerance (e21). Metformin is contraindicated for antidiabetic treatment in patients with advanced congestive heart failure (New York Heart Association [NYHA] stages 3 and 4) (7), because the inadequate pumping function of the heart promotes the accumulation of acid metabolites, leading to the risk of lactic acidosis (4). On the other hand, retrospective trials have shown a long-term survival advantage from metformin treatment in diabetics with congestive heart failure (14). Thus, metformin is recommended for use by diabetics with NYHA stage 1 and 2 congestive heart failure. As for sulfonylureas, glinides, and insulin, there are no general restrictions on the use of these drugs in diabetics with congestive heart failure. The physician should be mindful of the risk of hypoglycemic episodes and of fluid retention, which is occasionally seen under treatment with insulin (e22). Pioglitazone has been found in many trials to elevate the incidence of cardiac decompensation and is therefore contraindicated in patients with congestive heart failure (in any NYHA stage) (e23). GLP-1 analogues and DPP-4 inhibitors seem to have positive net effects on myocardial function (e16, e17) and may, therefore, be a good option for patients with congestive heart failure, although prospective endpoint trials on this question are entirely lacking. In cases of acute cardiac decompensation, temporary insulin treatment would appear to be the best option for in-hospital glycemic control (8); here, too, there is a lack of evidence from large-scale randomized trials and longitudinal studies.

Renal failure

Twenty to forty percent of patients with type 2 diabetes develop renal failure in the course of their disease (15). Renal failure is classified by chronic kidney disease (CKD) stages, as shown in Table 4 (e24). As most oral antidiabetic drugs are eliminated through the kidneys, there are constraints on their use in patients with impaired renal function.

Table 4

Stages of Chronic Kidney Disease (CKD) (modified from e24)

Enlarge All figures

Metformin is contraindicated because of potential accumulation if the creatinine clearance is lower than 60 mL/min (4, 7, 16) but can be given to patients in CKD stages 1 and 2 as long as their renal function is monitored by laboratory testing two to four times a year. A Cochrane review found no significant increase in the frequency of lactic acidosis under treatment with metformin (17). Some experts have suggested further liberalization of the glomerular-filtration-rate (GFR) cut-off value for metformin (e.g., to 45 mL/min), but this cannot be generally recommended, in view of the medicolegal consequences in the rare event that lactic acidosis does develop (e25).

The renally eliminated sulfonylureas can accumulate and cause dangerous hypoglycemia in patients with renal failure. This is especially true of the long-acting drugs glibenclamide and glimepiride; hypoglycemia can even arise several days after these drugs are discontinued. Thus, sulfonylureas are contraindicated in CKD stages 4 and 5 (4, 7, 16). The same contraindication applies for gliquidone, a sulfonylurea often given to patients with impaired renal function (7).

Glinides, on the other hand, are eliminated almost exclusively through the liver and are thus, after dose adjustment, suitable for use in advanced CKD stages (7).

Incretin-mimetic drugs and DPP-4 inhibitors can be used without any special restrictions by patients with renal failure in CKD stage 1 or 2 (7); exenatide can be used at a reduced dose in stage 3 as well (7), saxagliptin up to stage 4 according to current data (e26), sitagliptin and vildagliptin up to stage 5 (7, 25). Restrictions on the use of oral antidiabetic drugs are summarized in Table 5.

Table 5

Restrictions on the use of oral antidiabetic drugs in patients with renal failure (RF) or hepatic insufficiency (HI), as of 10 January 2012 (7, 25)

Enlarge All figures

Insulin treatment is generally recommended for patients with advanced renal failure. It should be borne in mind, however, that the effect of insulin can be prolonged in renal failure because of reduced insulin clearance, while patients with renal failure may simultaneously have increased resistance to insulin (18). The insulin dose usually has to be lowered in the setting of renal failure (19), but there are no uniform recommendations on this matter, and the dosage must be carefully titrated. Treatment with short-acting insulin may be advantageous for patients with renal failure (e27).

Hepatic disease

Although diabetes is more common in patients with impaired liver function, there are hardly any data on the treatment of diabetes in such patients. Thus, antidiabetic drugs are chosen mainly on the basis of their side-effect profiles. No more than a few cases of hepatotoxic side effects have been reported for metformin (e28), but the risk of lactic acidosis may be higher in the setting of impaired liver function, and metformin is therefore contraindicated for patients with advanced liver failure (4, 7, 16). On the other hand, in the much more common cases of hepatopathy without liver failure, metformin actually appears to have a beneficial effect on hepatic metabolism (e29–e32). As regards treatment with sulfonylureas, glinides, and insulin, one should bear in mind that there is an increased risk of hypoglycemia because of diminished hepatic gluconeogenesis. Moreover, glinides are eliminated almost exclusively through the liver and are thus contraindicated in the setting of impaired liver function (4, 7). Sulfonylureas are contraindicated in advanced liver failure (4, 7). Data are currently lacking on antidiabetic treatment with incretin-mimetic drugs and DPP-4 inhibitors in the setting of coexistent liver disease. Patients with moderately impaired liver function can take sitagliptin or exenatide (e33, e34). Liraglutide is considered to be contraindicated in the setting of impaired liver function because of a lack of data from clinical studies (7). In clinical studies of vildagliptin, high doses caused elevation of hepatic transaminases in a small number of patients; thus, this drug is not recommended for patients with impaired liver function (7). All oral antidiabetic agents are contraindicated in patients with advanced liver disease. Restrictions on the use of oral antidiabetic drugs are summarized in Table 5. As can be seen, for many patients with advanced liver disease, insulin treatment is the only available option for glycemic control. No systematic studies have yet been devoted to the question whether a prandial or basal insulin regimen is better in this setting.

Overview

In view of the many restrictions on the use of oral antidiabetic drugs that have been described in the foregoing paragraphs, temporary insulin treatment remains the only practical means of glycemic control for many hospitalized patients with type 2 diabetes. To this end, sliding scales can be used, supplemented by the additional administration of a fixed number of units of insulin with the main meals of the day and/or long-acting insulin analogues given once or twice daily (9, 16). These treatment options are summarized in the Figure. Detailed recommendations on the concrete implementation of temporary insulin treatment can be found, for example, in reference (16) (in German). In a prospective, randomized multicenter trial, base-bolus therapy was found to provide better glycemic control than a sliding scale (20). For long-term treatment, the once-daily administration of a long-acting insulin analogue seems to lead to fewer hypoglycemic episodes than the administration of short-acting insulin three times a day, while yielding comparable HbA_1c values (21, 22). In a recent trial, a combination of rapid- and long-acting insulin analogues was found to have no therapeutic advantage over neutral protamine Hagedorn (NPH) human insulin in the inpatient setting (23).

Figure

Sliding scales as an option for the temporary treatment of hospitalized patients with insulin

Enlarge All figures

Antidiabetic treatment in hospitalized patients thus differs in many respects from the treatment of outpatients. The special accompanying circumstances and comorbidities necessitate flexible, individualized glycemic management in the hospital.

Conflict of interest statement
Professor Meier has received lecture honoraria and consulting fees from the following companies: Astra Zeneca, Berlin-Chemie, BMS, Boehringer-Ingelheim, Eli Lilly, MSD, NovoNordisk, Novartis, Roche, and Sanofi-Aventis. He has received reimbursement of congress participation fees and travel expenses from MSD, NovoNordisk, and Sanofi-Aventis. MSD, NovoNordisk, and Sanofi-Aventis support research projects that he initiated.

Dr. Breuer has received reimbursement of congress participation fees and travel expenses from Sanofi-Aventis and NovoNordisk.

Manuscript submitted on 30 June 2011, revised version accepted on 12 January 2012.

Translated from the original German by Ethan Taub, M.D.

Corresponding author
Prof. Dr. med. Juris J. Meier
Abteilung für Diabetologie
Medizinische Klinik I
Universitätsklinikum St. Josef-Hospital
Klinikum der Ruhr-Universität Bochum
Gudrunstr. 56
44791 Bochum, Germany
juris.meier@rub.de

@For eReferences please refer to:
www.aerzteblatt-international.de/ref2612