You probably already know most of this, but for a quick refresher:
Standard practice is to discontinue all outpatient anti-hyperglycemics, though this is changing and you should consider continuing / restarting oral agents if the patient is stable and eating
Metformin = risk of lactic acidosis, CHF exacerbation (though evidence is poor)
Sulfonylureas = risk of hypoglycemia
TZDs = risk of CHF
GLP-1 agonists = risk of GI symptoms, pancreatitis
SGLT2 inhibitors = risk of euglycemic DKA, genitourinary infections (but NOT UTI), hypovolemia. We often restart these during admission however once patient is improving
* Goal blood glucose while inpatient = 140-180 mg/dL per NICE-SUGAR (2009)
Was done in ICU level patients but now kind of extrapolated to all inpatients
You should add the "endocrine" or "glycemic management" tab to your patient summary if you haven't already (if you're on Epic). Your institution should probably have something similar along these lines. It gives you a very clear look at their glucoses and number of units of insulin given during the hospitalization. Example
If fasting AM glucose is elevated, you need to increase their basal insulin.
If premeal and bedtime glucose are elevated, you need to increase their mealtime insulin. Will discuss more below.
If they ARE NOT on insulin outpatient:
0.2-0.3 units/kg/day for patients aged >70 and/or eGFR <60 mL/min or otherwise at risk for hypoglycemia
0.4 units/kg/day for patients with BG 140-200 mg/dL
0.5 units/kg/day for patients 201-400 mg/dL and insulin-resistant patients
0.6 units/kg/day for steroid-induced hyperglycemia
These numbers typically underestimate how much the patient will need, recommendations in Table 1 of Endocrine Society
Once you have calculated the total daily insulin dose, remember to split it to 50% basal glargine at night and 50% divided up three times a day with meals. Eg, if patient needs 60 units of insulin daily, they should get 30 units glargine qhs and 10 units aspart tid ac. Basal-bolus is the way to go per RABBIT-2 (2007)
I have always felt nervous "blindly" giving diabetic patients high doses of insulin when their insulin requirement is not known, however this study is very reassuring that the risk of hypoglycemia is pretty low with these ranges.
To be fair, if they do have AKI or there is any strong worry of precipitating hypoglycemia, I would err on the side of caution and just give sliding scale insulin for the first 24 hours. Then you can calculate their total daily insulin requirement and change them to basal-bolus the following day.
If they ARE on insulin outpatient:
Usually, take 75-80% of their insulin dose and start with that while they're in the hospital
* Sliding scale alone is wrong 99% of the time, but basal (0.2-0.3 units/kg) + correction is a reasonable starting regimen. Basal-bolus is the best
* Put everyone on sliding scale insulin (sensitive, moderate, or resistant) on TOP of basal-bolus
* Remember to add hypoglycemia protocol so nurses can give juices/D50/glucagon if needed for hypoglycemia
Titrating up while inpatient:
Oftentimes while your patients are admitted, you will notice their sugars consistently hanging out around the 200s or something, and you will need to adjust their total insulin dose.
If morning glucose is 180-200, increase basal dose by 10%. If 200-300, increase by 20%. If >300, increase by 30%
If premeal glucose is 180-200, increase nutritional dose by 10%. If >200, increase by 20%. >300, increase by 30%
If ALL glucose measurements are elevated, then just calculate their total daily dose and add 10-20% and split it again
* If your patient is going NPO, discontinue mealtime short-acting insulin and check blood glucose q4h. Continue their long-acting basal insulin
* Rule of 1500 / 1800 = you can estimate how much 1 unit of insulin would be expected to drop a patient's glucose by using the calculating 1800 / TDD of insulin. For example, if a patient is taking 60 units of insulin a day, then that means 1 unit of insulin drops their sugar by about 1800 / 60 = 30 mg/dL.
