Insulin protocols for managing DKA

Several insulin therapy protocols for management of DKA in dogs and cats have been reported. All of them work and the most important thing is that the patient with DKA gets prompt and appropriate insulin therapy. Clinicians should use the protocol they are most comfortable using. Perhaps even more importantly, because DKA is a multi-day (and therefore multi-doctor) management disease, the team of doctors who manage inpatients should become familiar with more than one protocol, and have a group decision about which protocol(s) you will use in your hospital. This decision involves consideration of what you are comfortable using, what seems to be best for your patient population, what is cost-effective in your hospital system, and what will be most feasible and reliable based on your level of staffing. Switching protocols every shift to suit the doctor on duty is NOT in the patient’s best interest. Having a plan before these cases occur will be better for the patient and less stressful for the team in the long-run. 

There are four basic ‘formulas’ for insulin administration in DKA:

    1. Intermittent IM injection of short-acting insulin
    2. Intravenous CRI of short-acting insulin
    3. Combination therapy with IM injection of short-acting insulin plus CRI of short acting insulin
    4. Combination therapy with IM injection or CRI of short-acting insulin plus SQ injection of long-acting insulin

The rationale for IV insulin during DKA is that because of poor perfusion and consequently variable absorption of drug from the SQ or IM space IV CRI of insulin allows for reliable and rapidly titratable dosing. 

The rationale for co-administration of SQ long-acting insulin along with IV or IM short-acting insulin is that this combination more closely mimics the activity of the functional pancreas.

Based on the information available right now, the only rationale for using lispro or aspart as a CRI would be if regular insulin is discontinued from the market, or became very expensive. There IS a physiologic rationale to use lispro insulin if giving intermittent IM or SQ injections rather than CRI (it dissociates faster and might be more reliably absorbed in altered perfusion states) 

The currently published protocols for insulin therapy for DKA are outlined below. Readers are encouraged to review the primary publications to decide which protocols may work best for individual patients. 

  1. Low dose intramuscular insulin therapy for diabetic ketoacidosis in dogs. Chastain CB, Nichols CE. JAVMA 1981. (PMID: 6790505)
    • Dogs 10kg or less: 
      • Initial 2 units regular insulin IM, followed by hourly 1 unit regular insulin IM
    • Dogs >10kg
      • Initial 0.25U/kg regular insulin IM, followed by hourly 0.1U/kg regular insulin IM
    • Glucose monitored hourly
  1. Treatment of diabetic ketoacidosis in dogs by continuous low-dose intravenous infusion of insulin. Macintire. JAVMA 1993 (PMID: 8496083)
    • Regular insulin
    • 2.2U/kg in 240mL of 0.9% NaCl (0.45% NaCl with dextrose once BG drops) starting at 10cc/hour
    • Patients rarely got the full prescribed daily dose due to rate adjustments to account for BG
    • CRI protocol was as follows:
Blood glucose concentration (mg/dL) Solution to fulfill fluid requirements Rate of insulin solution (mL/h) Dose of administered insulin (U/kg/h)
>250 0.9%NaCl 10 0.09
200-250 0.45% NaCl + 2.5% dextrose 7 0.064
150-199 0.45% NaCl + 2.5% dextrose 5 0.045
100-149 0.45% NaCl + 5% dextrose 5 0.045
<100 0.45% NaCl + 5% dextrose Stop infusion Stop infusion
  1. Comparison of regular insulin fusion doses in critically ill diabetic cats: 29 cases (1999-2007). Claus et al. JVECC 2010. (PMID: 20955302). 
    • Retrospective assessment of 3 groups, all got regular insulin
      • CRI 1.1U/kg/d (group 1)
      • CRI 2.2U/kg/d (group 2)
      • CRI started at 1.1U/kg/d and escalated to 2.2U/kg/d or more (group 3)
      • All CRIs were made as 240cc 0.9% NaCl with the prescribed daily insulin dose added
    • BG q2h, bicarbonate q4-12h, phos and Mg once daily, ketones daily with urine strip test (either plasma or urine was tested)
    • The CRI protocol loosely followed Macintire but exact values for alteration of treatment were at clinician’s discretion
    • Cats in group 1 were found to be healthier than those in group 2, unclear due to retrospective nature if this played a role in how insulin was prescribed
    • Cats in group 1 took longer to reach target BG and become ketone negative than group 2; no difference in duration of hospitalization
    • Majority of cats were hyperosmolar at presentation, no relationship between presenting osm and declining mentation status in this study regardless of prescribed dose of insulin
    • Cats rarely received the prescribed daily dose due to rate adjustments based on the sliding scale (loosely following Macintire)
  1. Use of lispro insulin for treatment of diabetic ketoacidosis in dogs. Sears et al. JVECC 2012. (PMID: 22390184)
    • Lispro insulin is a short-acting insulin
    • Randomized but not blinded, compared CRI of lispro to CRI of regular insulin
    • 2.2U/kg of regular or lispro insulin into 240cc 0.9% saline (later 0.45% with dextrose)
    • Monitored glucose q2h, other stuff at various intervals 
    • Bottom line: basically showed it is similar to regular insulin, might be a slight trend toward faster resolution of combo hyperglycemia, acidosis, and ketones but too small of a group to tell if it’s real or not 
    • CRI protocol was as follows (Same as Macintire but using lispro in the experimental group):
Blood glucose concentration (mg/dL) Solution to fulfill fluid requirements Rate of insulin solution (mL/h) Dose of administered insulin (U/kg/h)
>250 0.9%NaCl 10 0.09
200-250 0.45% NaCl + 2.5% dextrose 7 0.064
150-199 0.45% NaCl + 2.5% dextrose 5 0.045
100-149 0.45% NaCl + 5% dextrose 5 0.045
<100 0.45% NaCl + 5% dextrose Stop infusion Stop infusion
  1. Intramuscular glargine with or without concurrent subcutaneous administration for treatment of feline diabetic ketoacidosis. Marshall et al. JVECC 2013. (PMID: 23530935)
    • In humans, glargine IV and IM behaves similar to regular insulin IV – rationale for study: if glargine IM and IV behaves in cats similarly to humans, this might offer an alternate therapy for DKA in cats
    • Study design unclear – states “the study involved a retrospective case series of primary accession and referred cases presented to an exclusively feline veterinary clinic. Included cases had a diagnosis of DKA and were initially stabilized with IM glargine. Cases diagnosed with DKA and treated with IM glargine, with or without concurrent SC glargine between November 2005 and November 2008 were eligible for the study. Informed consent was obtained prior to enrollment in the study.”
    • No randomization of cases 
    • 15 cats, all got IM glargine. 12/15 also got SC glargine
    • 14 of 15 cats got 1U IM glargine as initial dose, 1/15 got 2U IM glargine – “degree of hyperglycemia did not influence initial IM dose.”
    • Initial SC dose was 1-3U/cat
    • BG checked q2-4h, subsequent dose adjusted with goal of lowering BG 2-3mmol/L/h (36-54mg/dL/h) Until reaching 10-14mmoL/L (180-252mg/dL)
    • “A prescribed protocol was not strictly adhered to with repeated 0,5-1U IM injections given as required (between 2 and 22h) and SC injections administered every 12 hours or longer to maintain BG concentration between 10 and 14mmol/L
    • Huge variation in timing and dose administered
    • No overnight monitoring or injections were given
    • If BG <180mg/dL (10mmol/L) 1g/kg of 50% glucose IV over 5 minutes followed by CRI of 2.5% solution (no rate stated)
    • 15 of 15 survived to discharge (median 4 days, range 2-5d)
    • 13/15 survived to 14 days post-discharge (Definition of ‘success’ was to survive 14 days and not have recurrence of DKA – no comment about recurrence seen?)
  1. Retrospective evaluation of continuous rate infusion of regular insulin intravenously for the management of feline diabetic ketoacidosis. Bollinger and Moore. CVJ 2015. (PMID 25565711)
    • The cats all got a CRI that had 1.1U/kg regular insulin in 250cc of 0.9% saline 
    • The CRI protocol referenced is Macintire 1993 JAVMA (see above)
    • Rate adjusted q2h
    • 50% mortality (much higher than all the other CRI publications) – authors list possible causes as being late referrals to tertiary center and thus much sicker patients, or possibly an inadequate dose of insulin
  1. A pilot study comparing a protocol using intermittent administration of glargine and regular insulin to a continuous rate infusion of regular insulin in cats with naturally occurring diabetic ketoacidosis. Gallagher et al. JVECC 2015. (PMID: 25546713)
    • Randomized cats using randomization table 
    • Glucose monitored q 2-4h
    • Ketones monitored q8h
    • Protocols below. Results (keep in mind it’s a pilot study):
      • Clinically significant and statistically significant shorter time to discharge, time to resolution of acidosis, stat sig shorter time to resolution of hyperglycemia and ketonemia. 
    • SQ/IM group glargine 0.25U/kg SQ q12h AND 1U regular insulin IM up to q6h if BG >250mg/dL. No additional insulin at other time points regardless of BG
      • Dextrose supplemented:
        • BG <250mg/dL = 2.5% in IV fluids (no rate given) 
        • BG 80-249mg/dL = 5% in IV fluids (no rate given)
        • BG <80mg/dL = 5% in IV fluids (no rate given) PLUS 0.5mg/kg bolus 50% dextrose
    • CRI 1U/kg diluted in 240mL 0.9% NaCl, protocol as follows 
Blood glucose concentration (mg/dL) Solution to fulfill fluid requirements Rate of insulin solution (mL/h) Dose of administered insulin (U/kg/h)
>500 0.9%NaCl 20 0.09
400-500 0.9%NaCl 15 0.069
250-399 0.9%NaCl 10 0.045
80-249 0.9% NaCl + 2.5% dextrose 0-5 0-0.023
<80 0.9% NaCl + 5% dextrose + bolus IV 0.5mL/kg 50% dextrose Stop insulin 0
  1. Use of intravenous insulin aspart for treatment of naturally occurring diabetic ketoacidosis in dogs. Walsh et al. JCVECC 2016 (PMID: 26379102)
    • Insulin aspart is a short-acting insulin, used as a CRI
    • 2.2U/kg into 240mL 0.9% NaCl (later 0.45% saline with dextrose) and start at 10cc/h (same as the regular insulin CRI described by Macintire) 
    • Glucose monitored every 2 hours, ketones (beta-hydroxybutyrate) every 4-6 hours, venous blood gas, sodium, potassium magnesium, lactate monitored every 4-6 hours
    • their study protocol for insulin therapy was as follows (same as reported in Macintire and Sears, just different insulin):
Blood glucose concentration (mg/dL) Solution to fulfill fluid requirements Rate of insulin solution (mL/h) Dose of administered insulin (U/kg/h)
>250 0.9%NaCl 10 0.09
200-250 0.45% NaCl + 2.5% dextrose 7 0.064
150-199 0.45% NaCl + 2.5% dextrose 5 0.045
100-149 0.45% NaCl + 5% dextrose 5 0.045
<100 0.45% NaCl + 5% dextrose Stop infusion Stop infusion
  1. Use of lispro insulin for treatment of diabetic ketoacidosis in cats. Malerba et al. J Feline Med Sure 2018. (PMID: 29513157)
    • Compared lispro CRI to regular insulin CRI
    • Based on drug kinetics, lispro is probably absorbed and active from the SQ space faster than regular insulin, but probably no difference between them when given IV
    • Glucose measured q1h in first 24h the q2-3h for duration of insulin CRI
    • Beta-hydroxybutyrate measured q4h
    • Historic controls compared to prospective lispro group
    • Bottom line, no big differences in time to resolution of DKA between groups
    • CRI protocol: 1.1U/kg of regular or lispro insulin added to 48mL 0.9% NaCl or or Ringer’s solution
Blood glucose concentration (mg/dL) Solution to fulfill fluid requirements Rate of insulin solution (mL/h) Dose of administered insulin (U/kg/h)
>250 0.9%NaCl or Ringer’s 2 0.045
200-250 + 2.5% dextrose 1.5 0.034
150-199 + 2.5% dextrose 1.5 0.034
100-149 + 5% dextrose 1 0.023
<100 5% dextrose Stop insulin Dextrose only
  1. Lispro insulin and electrolyte supplementation for treatment of diabetic ketoacidosis in cats. Anderson et al. JVIM 2019. (PMID: 31134702)
      • Comparison of 6 cats getting lispro CRI vs 6 cats getting regular insulin CRI
      • Doctors blinded to insulin type, technicians not blinded
      • All patients received 0.9% NaCl for resuscitation and ongoing fluids, supplemented with KCl or KPhos
      • No basal insulin dosing used in the study
      • Blood glucose normalized faster (statistically) in the lispro group (median 7h, range 2-10h) vs the regular insulin group (median 12.5h, range 8-20h), however time to resolution of ketosis and acidosis did not differ between groups. Authors suggest comparison of a much larger group is necessary to determine if the faster normalization of BG with lispro is clinically significant. 
      • The median time to initiation of SQ insulin administration and duration of hospitalization, which are probably better outcome markers than resolution of hyperglycemia, did not differ between the two groups
      • CRI protocol: 2.2U/kg of regular or lispro insulin added to 240mL 0.9% NaCl
Blood glucose concentration (mg/dL) Solution to fulfill fluid requirements Rate of insulin solution (mL/h) Dose of administered insulin (U/kg/h)
>300 0.9%NaCl or Ringer’s 10 0.09
200-300 + 2.5% dextrose 7 0.063
150-199 + 2.5% dextrose 5 0.063
100-149 + 5% dextrose 5 0.045
<100 + 5% dextrose Stop insulin Dextrose only
  1. Diabetic ketoacidosis and hyperosmolar hyperglycemic state in cats. Rand. VCNA-SAP 2013. (PMID: 23522177)
    • Review paper, describes the 4 feline protocols that were published to date (2013) and are all detailed above. Presents a clearer protocol for IM & SC glargine protocol:
      • Initial dose of 2U glargine per cat subcutaneously on initiation of fluid resuscitation
      • Begin 1U per cat glargine IM 1-2h later (when volume resuscitated), can delay up to 4 hours if very hypokalemic, but recommend no longer than 4 hours
      • Repeat IM glargine 4+ hours later if BG >250mg/dL (14mmol/L)
      • Continue SC glargine q12h
      • Provide IV glucose to maintain blood glucose levels ~216-255mg/dL (12-14mmol/L) in the first 24h
        • Dextrose supplementation recommendations:
        • When BG <15mmol/L (~270mg/dL) supplement IV fluids at 2.5% dextrose
        • When BG <8mmol/L (~140mg/dL) supplement IV fluids at 5% dextrose
        • Their algorithm suggests IV fluids are being supplied at 60-150cc/kg/d

Other useful information:

  • Gilor et al. Synthetic insulin analogs and their use in dogs and cats. VCNA Small Animal 2010. (PMID: 20219490)
    • Good review of insulin pharmacology in dogs and cats

 

Last Update 5NOV19 -ap