Abstract
Introduction
I Although we now try to sort out the meanings of the studies [from Umpierrez et al.
2
and van den Burghe et al.,
3
] it is my hope that the 60-yr-old patient admitted to the medicine service with pneumonia and a random admission plasma glucose of 230 mg/dl will not be metabolically ignored. Instead, a conscientious physician will provide an appropriate assessment and treat the patient's hyperglycemia.
As we start 2016, the field of inpatient diabetes management is a mixture of good news and ongoing challenges. First the good news: • We now have evidence that basal-bolus insulin therapy is superior to standard “sliding scale” for maintaining hospital glucose control.
4
• Tool kits and review articles are available that allow clinicians and hospitals to implement sound hospital glucose management.
5,6
• Endocrine and diabetes associations have official statements about inpatient hyperglycemia with defined glucose targets.
7,8
• Hospitals have access to order sets for proper insulin use that are easy to use and are associated with improved glycemic control.
9,10
• Intravenous insulin infusion protocols with specific guidelines on how to use them safely have been written.
11
• For intensive care units, nursing units and staff are happy to implement insulin infusion protocols, sometimes for weeks, in order to maintain good glycemic control.
12
As for challenges, the finding of increased mortality with intensive glucose control (target glucose <108 mg/dL) in the NICE-SUGAR study has refocused our attention on the very real risk of hypoglycemia in the hospital. 13
Upcoming Trends
Here are several areas of research that are particularly promising for 2016:
Hypoglycemia and the heart
The negative results from the ACCORD, ADVANCE, and VADT studies have cast a shadow on the benefits of strict glucose control in patients with long-standing type 2 diabetes. 14 In particular, these trials demonstrated a higher mortality rate in patients experiencing severe hypoglycemia. Proposed mechanisms for this association are increased inflammation, increased platelet activation, increased sympathoadrenal response, and increased cardiac workload. 15 One additional mechanism that has yet to be explored is whether or not hypoglycemia causes direct cardiac injury via reduced glucose availability. Intrinsic heart disease is associated with changes in myocyte substrate metabolism. For example, in congestive heart failure the heart loses the capacity to use free fatty acids as a fuel and relies mostly on glucose for energy. 16 This loss of metabolic flexibility may place hospitalized patients at particular risk for cardiac decompensation when the blood glucose level drops quickly. Testing this finding is challenging in acutely ill hospitalized patients, but research in this area may provide insight on how to better use insulin in the hospital. For example, intravenous insulin may serve as a metabolic bridge to preserve myocardium in patients with acute coronary syndrome awaiting coronary revascularization. Avoiding hypoglycemia may be the preferred glycemic goal over tight control for any patient with a myocardium that is dependent on glucose. Research to find a noninvasive method for identifying these patients would be useful.
Perioperative optimization of diabetes to reduce surgical-site infection
Surgical-site infections (SSI) continue to be a major driver of prolonged hospital stay and mortality. New rules by third-party payers are demanding that hospitals implement “care bundles” of specific interventions to reduce SSI rates. A recent review of care bundles to reduce SSI in patients admitted for colorectal surgery found that care bundles that included optimal glucose control were associated with a reduced rate of SSI. 17 As hospital payments are more closely coupled to outcomes, it appears that managing both preoperative and postoperative glucose will demand higher attention.
Epidemiology: insulin requirements and glucose levels as a biomarker for severity of illness
Hyperglycemia in a hospitalized patient is a marker for increased mortality. 2 Conversely, normalization of blood glucose and reduction in insulin requirements are likely markers for hospital improvement. For example, diabetes patients receiving a left ventricular assist device had a marked improvement in their diabetes control and reduced insulin dose 1 month after left ventricular assist device placement. 18 Similar to heart failure patients with the left ventricular assist device, a patient who is septic with systemic inflammatory response syndrome often have very high insulin requirements that normalize after he or she improves. Thus, insulin requirements and glucose together may be a biomarker for severity of illness.
Continuous glucose monitoring
Use of continuous glucose monitoring would free the medical staff to spend more time caring for the patient. Accurate continuous glucose monitoring could be coupled to a variable insulin infusion to produce a closed-loop glucose management system. 19 Current systems still have serious gaps in evidence for safety and efficacy and are currently not recommended for use in the hospital setting. 20 As devices improve in accuracy, additional studies are needed prior to acceptance in the hospital.
Use of “smart” applications and decision support tools to enhance glucose management
Several commercial software programs are currently available to guide insulin therapy, as discussed by Gianchandani and Umpierrez. 21 Additional programs continue to be published in this journal. 22 It is hoped that these programs will provide a user-friendly interface so that more patients will have the benefit of optimum insulin management.
Conclusions
We have made great strides in our understanding of how glucose impacts hospital outcomes over the past 14 years. It is hoped that continued progress over the next decade will improve safety and improve hospital outcomes for patients with hyperglycemia.
Footnotes
Author Disclosure Statement
S.C. serves on the Advisory Board for the Vaccines Division of GlaxoSmithKline.