Introduction

1. Traditionally, standard management of DKA includes initiation of regular insulin infusions (with or without use of technology such as EndoTool)in the ED and continued in the ICU. Despite being labor intensive, this method is necessary for when patients present with critical illness*.
2. Subcutaneous insulin has shown efficacy in the non-ICU setting for mild to moderate diabetic ketoacidosis treatment.
3. To date, several randomized studies have evaluated subcutaneous insulin lispro injections every 1-2 h vs. continuous IV regular insulin for treatment of mild or moderate DKA.
4. Studies have shown similar mean time to resolution of DKA in both treatment groups (lispro 10-14.8h; regular insulin 11-13.2h), mean time to resolution of hyperglycemia, total insulin doses required, number of hospitalization days and number of hypoglycemic episodes  in both treatment groups.
5. Overall, reports on the basis of low to very low evidence, indicate therapy is neither advantageous nor disadvantageous with either treatment modality.

*severe diabetic ketoacidosis (pH<7, serum bicarb<10, serum ketones> 8mmol/L, AG>12, and/or stupor/coma), hypotension, anasarca, pregnancy, CHF exacerbation, ACS, ESRD, liver failure or cirrhosis, obesity, and etc.

	Insulin Lispro/Aspart – SC	Insulin Regular – IV	Insulin Regular – SC
Dose	0.15 units/kg/dose q 2-3h until resolution of metabolic acidosis;	0.1 – 0.14 units/kg/h	Continuous IV is preferred due to PK profile
Administration	Avoid cold injections Can inject into the thighs, arms, or abdomen; Novolog (aspart) can also be injected into buttocks	Do not use if viscous or cloudy	Avoid cold injections Can inject into the thighs, arms, buttocks, or abdomen
PK/PD	Onset: 10-20 min Peak: 30-90 minDuration of action (DOA): 3-5 hhalf-life: 1 h	Onset: 15 min   Peak: 3 h    DOA: 6-8 h   half-life: 0.5-1 h (dose-dependent)	Onset: 1-2 h, Peak: 0.8-2 h  DOA: 4-12 h    half-life: 1.5h
Adverse Effects	>10%: cough, headache, hypoglycemia, infection, injection site reaction, pharyngitis, rhinitis	>10%: headache, hypoglycemia; unknown frequency, severe ADR: anaphylaxis, angioedema, bronchospasm Same as IV- Insulin R
Drug Interactions and warnings	Major DDIs: Chloroquine; glimepiride; rosiglitazone; macimorelin; metformin; pramlintide	Major DDIs: Chloroquine; glimepiride; rosiglitazone; macimorelin; metformin; pramlintide Same as IV- Insulin R
Compatibility	N/A	Y-site incompatible w/ many drugs	N/A
Comments	Look alike sound alike drug	Look alike sound alike drug	Not a good option due to delayed onset of action and prolonged half-life6

---

Overview of Evidence

Author, year	Design/ sample size	Intervention & Comparison	Outcome
Umpierrez, 2004	Prospective, randomized, open trial (n=45)	to SC rapid-acting insulin (aspart/lispro) at different time intervals vs standard regular insulin IV protocol	Primary end point: duration of treatment until resolution of DKA   Secondary end points: total length of hospitalization, amount of insulin administration until resolution of hyperglycemia and DKA, and number of hypoglycemic events   No statistical differences in the mean duration of treatment correction of hyperglycemia (6.9 +/- 4, 6.1 +/- 4, and 7.1 +/- 5 h) or until resolution of ketoacidosis (10 +/- 3, 10.7 +/- 3, and 11 +/- 3 h) among treatment groups   No mortality and no differences in the length of hospital stay, total amount of insulin administration until resolution of hyperglycemia or ketoacidosis, or the number of hypoglycemic events
Andrade-Castellanos, 2016	Systematic review (n=201)	SC rapid-acting insulin vs. IV-regular	Primary end point: duration of treatment until resolution of DKA   In children with DKA, time to reach a BG< 250mg/dL there was no difference between insulin lispro and IV regular insulin   No difference comparing SC aspart vs IV regular insulin on time to resolution of DKA: MD -1 h (95% CI -3.2 to 1.2); P = 0.36; 30 participants; 1 trial.   No difference comparing SC rapid-acting insulin analogues versus IV regular insulin on hypoglycemic episodes: risk ratio (RR) 0.59 (95% CI 0.23 to 1.52); P = 0.28; 156 participants; 4 trials.   In summary, there was no advantages or disadvantages between SC rapid-acting insulin analogues and IV regular insulin in treatment of mild or moderate DKA.
Razavi, 2018	Prospective RCT (n=50)	(SC aspart @ 0.15 units/kg q2h in regular medical ward) vs control (IV insulin infusion @ 0.05-0.1 unit/kg/h until resolution of DKA in ICU)	Less mean total-dose of insulin in intervention (SC) than control (IV)SC insulin for treatment of moderate DKA resulted in faster recovery/shorter hospital stay
Ersoz, 2006	Prospective randomized, open trial (n=20)	Hourly SC lispro (initial 0.15 unit/kg bolus with IV regular, followed by 0.075 as hourly SC lispro) vs  IV regular (initial 0.15 unit/kg bolus with IV regular and continued infusion per protocol)	No difference in time to resolution of DKATreatment of mild and moderate DKA with SC insulin lispro is equally as effective and safe in comparison with IV regular insulin
Della Manna, 2005	RCT (n=60)	Randomized to SC lispro (0.15 units/kg) q2h vs continuous IV infusion regular (0.1 unit x kg x hr)	Capillary glucose level decreased by 2.9 and 2.6 mmol x l x h in the lispro and regular groups, respectively; however, blood glucose fluctuated at different time intervals.In IV regular insulin group, metabolic acidosis and ketosis resolved in the first 6h after capillary glucose reached 13.8 mmol/l, whereas in the lispro group, they resolved in the next 6h intervalBoth groups met recovery criteria without complicationsSC is cost-effective  and technically simplified procedure that precludes ICU admission

---

Conclusions

1. In patients with mild-to-moderate DKA, q1-2h insulin lispro is a feasible alternative to continuous IV regular insulin for avoiding ICU admissions of uncomplicated DKA cases.
2. Use this therapy approach when insulin drip is not feasible on a general medical floor or in the ED; however, must ensure adequate staffing as the treatment modality requires q1h-q2h monitoring.
3. Of note, dosing and monitoring q2-4h has become a popular and practical approach during the COVID-19 pandemic due to less contact frequency.
4. Pros: cost savings? (ICU care $$$); less time-consuming to setup, although time spent is more in q1h-q2h method
5. Cons: More injections

---

References

1. Micromedex [Electronic version].Greenwood Village, CO: Truven Health Analytics. Retrieved January 17, 2021, from http://www.micromedexsolutions.com/
2. https://care.diabetesjournals.org/content/27/8/1873
3. https://pubmed.ncbi.nlm.nih.gov/26798030/
4. https://pubmed.ncbi.nlm.nih.gov/26238182/
5. https://pubmed.ncbi.nlm.nih.gov/29797212/
6. https://pubmed.ncbi.nlm.nih.gov/16620355/
7. https://pubmed.ncbi.nlm.nih.gov/16043723/