Glycemic Control in Medical Inpatients with Type 2 Diabetes Mellitus Receiving Sliding Scale Insulin Regimens versus Routine Diabetes Medications: A Multicenter Randomized Controlled Trial

Thank you for your interest in our manuscript. You are completely correct - our study supported the lack of efficacy of the sliding scale of insulin in the hospitalized diabetic patient admitted for a comorbid condition, as was our hypothesis. Poor glycemic control in this population and setting is common, but it is unknown if improved short-term glycemic control during hospitalization results in improved outcomes in this group. The definition of 300 mg/dL was chosen based on the ADA definition of hyperglycemia at the time of study development, and was considered a clinically important value beyond which most clinicians would make an intervention. A clinical trial is needed to determine if short-term glycemic control can be improved using long or intermediate acting insulin used in this population and setting.

Competing interests: None declared

The article simply demonstrated the need for better diabetes care in the hospital. Of the type 2 diabetes patients studied one out of three had hyperglycemia. This just proves there needs to be long or intermediate acting insulin used in these patients. Your definition of hyperglycemia at 300mg/dl is not in line with the 200mg/dl used in other studies. The number of patients with hyperglycemia at 200mg/dl are probably considerably more in patients with and without sliding scale. Your article mearly supports the idea that the sliding scale is ineffective.

Competing interests:   None declared

This report flies in the face of my own personal experience. My interpretation of your findings is as follows: since this is unblindied, I would suspect that those in the sliding scale group tending to be more passive in thier adjustiments of standing insulin dosing, or even in the institution of standing onsulin in thoses on oral agents. The may have been attempted to overrely on the sliding scale. In contrast those in the no sliding scale arm were essentially forced to add or adjust the standing insulin since they had no alternative. We would all agree that a proper standing regimen is superior to a sliding scale. But your report does not convince me that adding a well thought out supplemental sliding scale in additon to an aggressive adjusted mutlidose basal/ bolus regimen does not enhance the response.

Competing interests:   None declared

Dear Dr. Papreck:

I hope our article has been helpful to your P&T committee and its decision making regarding this issue. When we wrote the grant, lispro wasn't used significantly in practice in the SSI regimen in any of our participating institutions, so we did not consider that option. Of course, practice has changed in that direction today, so the interpretation of our results should consider that difference. You are also correct that the SSI regimen "times" are not exactly according to our plan of q 6 hours, but more in tune with the hospital schedule/meal plan. As our overall goal was to mimic usual practice, we did not dictate the SSI times of administration, just the scale noted in the manuscript. Any of these issues could affect the rates of hypoglycemia (and I guess hyperglycemia too), but since our rates of hypoglycemia were low already, it's unlikely that these would change our results significantly.

Sincerely, Lori M. Dickerson MUSC Family Medicine

Competing interests:   None declared

The article by Dickerson et al reminds us that we continue to follow therapeutic traditions in our daily practices that have never been submitted to rigorous scientific evaluation. Despite our emphasis on evidence-based medical practice, many of our diagnostic and therapeutic interventions lack supporting scientific evidence. The authors are to be applauded for taking on this longstanding therapeutic tradition. Although this is not a "perfect" randomized clinical trial (convenience sample, lack of blinding of patients and clinicians), the study's findings are clinically believable. Although there is potential for patient selection bias and clinician measurement or reporting bias, I believe that these are unlikely in this multisite, clinical research study. It is important that this study be replicated using a consecutive patient sampling method and double-blinding. In the meantime, I will think twice before writing for sliding scale insulin in a hospitalized type 2 diabetic patient.

Competing interests:   None declared