Analysis of Intravenous Insulin Dosing Requirements for Treatment of Severe Hypertriglyceridemia

Abstract

Purpose: Insulin infusion therapy is commonly utilized for treatment of severe hypertriglyceridemia, however, data supporting a standardized dosing approach is lacking. This study aimed to determine the average initial insulin dose utilized for treatment of hypertriglyceridemia and the associated reduction in serum triglycerides. Methods: This single-center, retrospective, descriptive analysis conducted at an academic medical center included adult hospitalized patients with serum triglyceride levels greater than 1000 mg/dL receiving treatment with an intravenous insulin infusion between November 2017 and August 2020. Data was extracted from the electronic medical record. The primary outcome was the mean weight-based intravenous insulin dose resulting in resolution of hypertriglyceridemia. Secondary outcomes included time to resolution of hypertriglyceridemia, adverse events associated with insulin treatment, incidence of rebound hypertriglyceridemia, and use of additional lipid-lowering therapies. Results: Data from 32 hospital encounters was analyzed. The mean initial triglyceride level was 3229 mg/dL. The average insulin doses observed on days 1 and 2 of therapy were 0.07 and 0.05 units/kg/hour, respectively. The mean percent triglyceride reduction at 48 hours was 40%. Mean time to resolution of hypertriglyceridemia, discontinuation of insulin infusion, or discharge was 5.7 days. Hypoglycemia and hypokalemia were observed in 9% and 29% of patients, respectively. Conclusion: The results of this study provide new guidance for insulin dosing for hypertriglyceridemia. Serum potassium levels and blood glucose should be monitored closely during therapy.

Background

Severe hypertriglyceridemia (HTG), defined as a triglyceride (TG) level ≥1000 mg/dL, is associated with an increased risk of pancreatitis and cardiovascular disease, and higher rates of pancreatitis-related complications when HTG is the inciting cause.^1,2 Insulin is considered a first line treatment option for management of severe HTG, although the optimal dose and route of administration has yet to be elucidated. Plasmapheresis may also be considered for acute management of HTG, however it is unclear whether this treatment reduces the time to resolution of severe HTG when combined with insulin treatment. ¹ Additionally, the high cost, need for specially trained personnel, and invasive nature of plasmapheresis limits its utility. Current guidelines do not recommend the use of plasmapheresis in the setting of very severe HTG with pancreatitis. Furthermore, treatment options such as fibrates, omega-3 fatty acids, niacin, and HMG-CoA reductase inhibitors are not ideal for use as monotherapy in acute management due to their delayed onset of effect and are most effectively used as adjunct therapy.^{2 -4}

Insulin reduces TG levels through stimulation of lipoprotein lipase and subsequent degradation of chylomicrons and has been shown to be an effective treatment for HTG when administered by either the subcutaneous or intravenous (IV) routes. However, more rapid reductions in TG levels are achieved with IV insulin administration. ⁴ Case reports have provided evidence of the utility of IV insulin for a more rapid resolution of severe hypertriglyceridemia, however, the optimal dose has yet to be defined.^{5 -17} Without specific guidelines supporting a standardized approach to IV insulin dosing for HTG treatment, literature evaluating this population has suggested a wide range of starting doses from 0.1 to 0.4 units/kg/hour. ³ Given the risk of significant adverse effects associated with delayed resolution of HTG and the known safety risks associated with IV insulin administration, the authors of the present study proposed that elucidating a standardized insulin initiation dose would serve to provide guidance to the medical community on IV insulin dosing for the rapid and safe treatment of this disease. Therefore, the purpose of this study was to determine the average dose of IV insulin utilized and the associated reduction in serum TG concentration.

Methods

This retrospective, observational, descriptive analysis was approved by the University of Tennessee College of Medicine at Chattanooga Institutional Review Board, reference number 20-063. It was conducted at a single urban academic medical center with approximately 500 beds. Inclusion criteria consisted of patients greater than 18 years of age admitted to the study institution with a diagnosis of HTG, defined as a TG > 1000 mg/dL, and treated with an IV insulin infusion from November 1, 2017 to July 1, 2020. Pregnant patients were excluded from the study.

The primary endpoint of the study was the average weight-based IV insulin dose administered per day to achieve resolution of HTG. Resolution of HTG was defined as TG < 500 mg/dL or hospital discharge, whichever occurred sooner. If applicable, insulin dosing and TG levels were excluded from the primary outcome analysis for the day(s) patients received plasmapheresis. Secondary endpoints included time to resolution of HTG, incidence of hypokalemia and hypoglycemia, incidence of rebound HTG within 72 hours of insulin infusion discontinuation, and rate of concomitant use of additional lipid-lowering therapies. Hypoglycemia was defined as a blood glucose less than 70 mg/dL (either by point of care or venipuncture test modality) and hypokalemia was defined as serum potassium less than 3.4 mg/dL. Rebound HTG was defined as an increase in serum triglycerides of 25% from the nadir.

Data was collected via Research Electronic Data Capture (REDCap) database through the University of Tennessee Health Science Center. Statistical analysis of the data was preformed using Microsoft Excel and is reported as descriptive data. ¹⁸ Pearson’s correlation coefficient was utilized to compare average IV insulin dose and percent triglyceride lowering effect.

Results

We identified 25 unique patients across 32 separate hospital encounters diagnosed with severe HTG who met inclusion criteria. Patient demographics are summarized in Table 1.

Table 1.

Demographics (n = 32 Unless Otherwise Noted by ^*).

Demographics
Age*, years mean, (SD)	43.3 ± 8.6
Male* (n, %)	19 (76)
Weight, kg mean, (SD)	97.8 ± 24.7
Initial TG level, mg/dL mean (SD)	3229 ± 2018.6
ICU level of care (n, %)	19 (59.4)
Length of stay, days (median, IQR)	6.5 (5,9.25)
Comorbidities*	n (%)
Pancreatitis	14 (56)
Diabetes mellitus, Type 1	1 (4)
Diabetes mellitus, Type 2	19 (76)
Familial hypertriglyceridemia	2 (8)
Hyperlipidemia	12 (48)
Chronic kidney disease	1 (4)
End stage renal disease	0
Active problems on admission	n (%)
Diabetic ketoacidosis	10 (31.3)
Pancreatitis	32 (100)
Acute kidney injury	5 (15.6)
Etiology of hypertriglyceridemia	n (%)
Familial	6 (18.7)
Alcohol	1(3.1)
Unspecified	25 (78.1)

^*n = 25.

Renal insufficiency was present in a total of 6 encounters. In 5 (15.6%) of these cases, acute kidney injury was present on admission. Chronic kidney disease without acute kidney injury was present in one case. The mean patient weight was 97.8 ± 24.7 kg. The average initial serum triglyceride level was 3229 mg/dL. Acute pancreatitis was present in all encounters and 59.4% of encounters required intensive care unit (ICU) admission. The primary outcome was analyzed for days 1 and 2 of treatment. Twenty-six of 32 total hospital admissions were included in the day 1 analysis. Four admissions were excluded due to missing serum triglyceride levels and 2 admissions were excluded due to receiving plasmapheresis. On day 1 of treatment, the average insulin dose was 0.07 ± 0.03 units/kg/hour and resulted in an average triglyceride reduction of 43.93 ± 25.03%. The average insulin dose and associated triglyceride reduction on day 1 are depicted in Figure 1. There was no correlation between average insulin dose and percent triglyceride reduction (R ²  = .0242). Twenty-four encounters were included in the day 2 analysis. Four patients were excluded due to missing serum TG levels and 4 patients were excluded due to receiving plasmapheresis. On day 2, the average insulin dose was 0.05 ± 0.03 units/kg/hour with an average TG reduction of 29.37 ± 18.27%. In the total cohort, patients received insulin infusion therapy for a mean duration of 3.75 ± 3.1 days and the average time to resolution of HTG was 5.22 ± 2.97 days). Rebound elevation in serum TG levels after discontinuation of insulin therapy was observed in 3 encounters, none of which required re-initiation of IV insulin therapy. In these cases, the mean rebound increase in serum triglycerides was 31.6% (range 26.1%-35.7%).

Figure 1.

Average insulin dose on treatment day 1 and associated triglyceride reduction.

The average IV insulin dose, incidence of hypokalemia, and incidence of hypoglycemia for each day of therapy is displayed in Figure 2 and includes all admissions, regardless of plasmapheresis status. Hypoglycemia was observed in 9.4% of encounters on therapy day 1 but was not documented during any encounters after day 1 of insulin therapy. Among the entire study sample, dextrose-containing maintenance IV fluids were administered with insulin in 26 (81%) encounters. Hypokalemia was observed on days 1 to 5 of insulin therapy and ranged from 8% to 26%.

Figure 2.

Average daily IV insulin dose and associated incidence of adverse effects per day of therapy.

Rates of administration of additional HTG treatment modalities was also assessed, including oral pharmacologic therapies and plasmapheresis. Use of oral lipid-lowering therapies prior to admission was reported in 31% of all patients and such therapies were either continued or initiated during the hospitalization in all but one encounter. Table 2 provides details regarding specific pharmacologic agents used during admission. Heparin was administered during 13 (40.6%) admissions with 2 of these receiving IV therapeutic dosing and the remainder of admissions receiving heparin for deep vein thrombosis prophylaxis.

Table 2.

Concomitant Oral Lipid Lowering Therapies During Hospitalization.

Medication class	n (%)
HMG-CoA reductase inhibitor	19 (59.4)
Fibrate	26 (81.3)
Omega-3 fatty acids	9 (28.1)
Niacin	3 (9.4)

Plasmapheresis therapy was provided during 8 separate admissions (25%). ICU and hospital length of stay was compared between patients who did and did not receive plasmapheresis. Of these, one admission was excluded from analysis due to an extensive ICU and hospital length of stay, 93 and 94 days, respectively. A shorter length of ICU and hospital length of stay were observed in the cohort that did not receive plasmapheresis. The results are summarized in Table 3.

Table 3.

ICU and Hospital Length of Stay in Patients Receiving Plasmapheresis Versus No Plasmapheresis.

	Plasmapheresis a	No plasmapheresis b
ICU length of stay, days (mean, SD)	10.75 ± 5.56	6 ± 5.55
Hospital length of stay, days (mean, SD)	9.57 ± 6.43	8.04 ± 6.99

^aPlasmapheresis was performed during 8 admissions with 5 of these patients requiring an ICU admission. One of these admissions was excluded from analysis due to an extended ICU and hospital length of stay.

^bIn 24 admissions plasmapheresis was not utilized, 14 of these admissions required ICU admission.

Discussion

Our study results confirm the previously documented efficacy of IV insulin therapy in rapid resolution of severe HTG and addresses a gap in current literature on selection of an initial insulin infusion dose. Based on these summarized results, the authors suggest that using an initial IV insulin dose of 0.07 units/kg/hour in combination with dextrose-containing maintenance fluids is a reasonable strategy for initial treatment of severe HTG. The percent reduction in serum triglycerides at 24 hours observed in our study mirrors a previous report, by Thuzar et al who observed a 40% reduction in serum TG levels after 24 hours of insulin therapy. ¹⁷ Although 29% of encounters in the present study received treatment with concomitant oral lipid lowering therapies, this would not be expected to have a clinically meaningful impact on the IV insulin dose requirement due to their delayed effect on the reduction of serum triglycerides. This dosing strategy resulted in a rapid TG lowering effect, with a low incidence of hypoglycemia, hypokalemia that occurred in approximately 25% of patients. Therefore, serum potassium should be closely monitored and aggressively replaced. Dextrose containing maintenance fluids were concomitantly administered with IV insulin in the majority of encounters. When use of dextrose-containing fluids occurred, dextrose 5% was initiated in all cases except one, where dextrose 10% was utilized. However, in 10 of these encounters, we observed that the dextrose 5% was transitioned to dextrose 10%. Further examination revealed only one episode of a documented hypoglycemic event, suggesting that the change from 5% dextrose to 10% dextrose was unrelated to hypoglycemia. It should be noted that the incidence of renal impairment in this study population was low, and that such patients may require a lower initial insulin infusion rate to minimize risk of hypoglycemia.

Adverse effects including rates of hypoglycemia and hypokalemia have not been consistently reported in the literature, thus it is difficult to ascertain relative safety of the proposed standard insulin dosing regimen compared to historic case reports. It is likely that initiation of IV insulin with concomitant dextrose-containing IV fluids would be an ideal strategy for prevention of hypoglycemia, particularly for patients on bowel rest who are not eating, regardless of presence of diabetes except in cases in which patients present with hyperglycemia.

Notable limitations of this study include the retrospective nature of data collection, small sample size, and lack of ability to ascertain information on previous insulin dosing and serum TG levels from outside healthcare facilities for patients who received hospital care prior to transfer to the study center or admitted from an outside healthcare facility. Additionally, the first serum TG level used in this study as the baseline record was the first level obtained at the study institution. It is possible that patients transferring from outside healthcare facilities may have had higher baseline TG levels. The impact of heparin dose on TG level reduction was not assessed, however, of 8 encounters in which heparin was prescribed only 2 patients received therapy via the intravenous route. Information regarding HmG-CoA reductase inhibitor dose for individual patients was not collected, as this was not expected to have a clinically meaningful influence on time to resolution of severe hypertriglyceridemia. Finally, specific data regarding rate of dextrose-containing fluid administration is not reported. Selection of maintenance fluid infusion rate was left to the discretion of the ordering provider.

Conclusion

The results of this analysis provide new information that may guide selection of a specific weight-based insulin dose that results in rapid reduction in serum TG levels, and characterizes rates of hypoglycemia and hypokalemia associated with the observed average insulin dose of 0.07 units/kg/hour during the first 24 hours of therapy. This will serve to aid other healthcare facilities in development of standardized treatment protocols for severe HTG. Further studies are warranted to determine whether plasmapheresis in addition to insulin therapy confers additive clinical benefit.