Hypertriglyceridemia causes ~9% of pancreatitis, the third most common cause after alcohol and gallstones. It is a risk factor for severe pancreatitis, making it more frequent among ICU patients with pancreatitis. I see this a few times each year. Nonetheless, it has low penetration into educational curricula or our collective awareness.
The treatment of hypertriglyceridemic pancreatitis is an ongoing source of confusion. This post will attempt to sort this out, given existing evidence.
Definition & diagnosis
Defining hypertriglyceridemic pancreatitis
No evidence-based definition of hypertriglyceridemic pancreatitis exists (Carr 2016). The general consensus in the literature seems to be that in a patient with pancreatitis, a triglyceride level >1,000 mg/dL makes hypertriglyceridemia the likely cause. However, some patients have chronically elevated triglyceride levels in the thousands, without pancreatitis. Therefore, a triglyceride level >1,000 mg/dL doesn't prove the diagnosis of hypertriglyceridemic pancreatitis.
Ultimately, the diagnosis of hypertriglyceridemic pancreatitis requires judgement, based on the following considerations:
- Severely elevated triglyceride level (at least >500 mg/dL, generally >1,000 mg/dL)
- No alternative cause of pancreatitis based on history, imaging, and laboratory studies.
- The presence of diabetes, pregnancy, obesity, or chronic hyperlipidemia supports this diagnosis.
Diagnosing the underlying cause of hypertriglyceridemia
The following are common causes of hypertriglyceridemia, which may lead to hypertriglyceridemic pancreatitis:
- Hereditary hypertriglyceridemia
- Diabetes (type I or II)
- Medications including propofol, valproic acid, protease inhibitors, thiazides, olanzapine, mirtazapine, estrogen-containing oral contraceptives, tamoxifen, clomiphene, and isotretinoin (Scherer 2014)
Identifying the cause of hypertriglyceridemia generally isn't critical for management. However, any potentially causative drugs should be discontinued.
Other diagnostic issues in these patients
Severe hypertriglyceridemia may cause the serum to appear milky (latescent). Although this may be a useful diagnostic sign, it can also interfere with certain laboratory tests. In particular, measurements of amylase and sodium may be artificially low (pseudohyponatremia; Melnick 2016). In extreme cases, the lab may be unable to run multiple tests.
General treatment of hypertriglyceridemic pancreatitis
General therapies for pancreatitis
The basics of treating hypertriglyceridemic pancreatitis are the same as for treating any patient with pancreatitis. Resuscitation should be performed using a judicious volume of fluid and vasopressors, similar to sepsis resuscitation (rather than blindly drowning patients with massive fluid resuscitation). Intubated patients benefit from early enteral nutrition. Overall, the treatment of pancreatitis largely centers on high-quality supportive care, similar to any other critically ill patient.
Oral anti-lipemic therapies
Patients generally receive lipid lowering medications (e.g. gemfibrozil 600 mg BID). Dietary fat restriction is also sensible.
Treatments directed at lowering triglyceride levels
When is specific therapy needed to lower triglyceride levels?
Triglyceride levels fall naturally over time, without any specific intervention (figure below). This may reflect reduced fat intake and fluid resuscitation.
When are additional therapies beneficial to accelerate the fall in triglyceride levels? This question hasn't been investigated rigorously. Given that pancreatitis may be very difficult to treat once it gets out of hand, it seems rational to attempt to lower triglyceride levels in efforts to limit disease severity.
Overview of therapies to lower triglyceride levels
The fall in triglyceride levels may be accelerated using plasmapheresis or insulin.
Plasmapheresis (a.k.a. plasma exchange): This involves placement of a hemodialysis catheter to allow for rapid exchange of blood. Blood is fractionated into cells and plasma, and plasma is replaced with either albumin or fresh frozen plasma. Plasmapheresis may be performed in various ways:
- Centrifugal plasmapheresis: This is more common in the United States. Blood is separated into plasma and cells via continuous centrifugation. Large volumes of plasma may be replaced over a few hours. This procedure may be repeated daily as needed.
- Plasma filtration: This seems to be more common in Asia. Plasma is separated from blood via a filter, using a setup which is similar to continuous renal replacement therapy.
Insulin will reduce triglyceride levels, by reducing the synthesis of triglycerides and also by accelerating their metabolism. To achieve rapid lowering of triglycerides, insulin may be infused at a fixed rate (e.g. ~0.25 units/kg/hr). An infusion of dextrose is often required to avoid hypoglycemia.
RCT of plasmapheresis vs. insulin: He et al 2016
Numerous case reports and case series showcase the use of a single technique (e.g. plasmapheresis). Such uncontrolled studies provide little useful information. Only one RCT has been performed to compare plasmapheresis vs insulin.
This was a prospective RCT involving 66 patients with hypertriglyceridemic pancreatitis who were randomized to receive plasmapheresis (using plasma filtration) versus insulin. Insulin was titrated with a goal of treating hyperglycemia (rather than specifically titrating insulin to lower triglyceride levels). Plasmapheresis cleared triglycerides faster:
How is it possible that plasmapheresis could drop triglyceride levels faster, but lead to worse clinical outcomes? The answer may be that triglyceride level is a flawed surrogate outcome. Triglycerides are actually rather inert. Before causing harm, triglycerides are metabolized into free fatty acids:
Triglycerides –> Free fatty acids –> Pancreatitis
Free fatty acids are probably the most important substance to measure, but we cannot measure them in routine clinical practice. So, instead, we settle on measuring triglyceride level.
Insulin works via a variety of pathways to reduce levels of both triglycerides and free fatty acids. Insulin may rapidly reduce free fatty acid levels, via mechanisms that are independent of triglyceride levels (Chow 2011, Afari 2015). Thus, it is possible that insulin could rapidly reduce free fatty acid levels, before the total level of triglycerides was reduced. This could explain how insulin is more effective than plasmapheresis clinically, despite causing a slower clearance of triglycerides.
Does plasmapheresis add to insulin infusion?
Miyamoto 2017 performed a post-hoc analysis of a registry including 1,159 pancreatitis patients at 44 hospitals in Japan. 30 patients with hypertriglyceridemic pancreatitis were studied, of whom 10 were treated with plasmapheresis. Characteristics of patients in both groups were similar, including that most patients received insulin (80% overall). Plasmapheresis had no detectible effect on the triglyceride level:
This study is limited by a lack of details regarding the precise regimen of plasmapheresis and insulin utilized in these patients. However, it does suggest that in a clinical context where most patients are receiving insulin, plasmapheresis may provide little added value.
Conclusions on insulin vs. plasmapheresis
Plasmapheresis has a number of drawbacks compared to insulin infusion:
- More invasive (requiring hemodialysis catheter).
- More expensive (in He 2016, plasmapheresis doubled the total hospitalization cost).
- Requires anticoagulation, which causes complications (citrate may cause hypocalcemia, whereas heparin may cause bleeding and possibly increased mortality; Gubensek 2014).
- If fresh frozen plasma is used for replacement, this may cause allergic or infectious complications.
- Extracorporeal blood circuits can cause leukocyte activation, increasing inflammation.
- Plasmapheresis requires buy-in from hematology consultants and the assistance of specialized nurses (who may be unavailable during off hours). Most hospitals lack plasmapheresis capabilities. These factors may lead to considerable delays before therapy can be initiated.
Available evidence suggests that insulin is equivalent or superior to plasmapheresis. This data is admittedly quite limited. However, when combined with the drawbacks of plasmapheresis, this suggests that insulin may be a first-line therapy for hypertriglyceridemic pancreatitis.
Plasmapheresis used to be my first-line approach to hypertriglyceridemic pancreatitis, based on the recommendation of several authors (UpToDate still recommends it as “potentially lifesaving”)(1). Due to logistic headaches and emerging evidence, I've transitioned to insulin infusions instead. Consistent with published literature, both approaches seem to yield excellent outcomes. However, insulin infusion may be safer and easier.
Nuts & bolts of using an insulin infusion
Running the insulin infusion isn't difficult, with many similarities to the management of diabetic ketoacidosis. The insulin infusion should generally be infused between 0.1-0.3 units/kg/hr depending on the patient's degree of insulin resistance. Higher insulin infusion rates will achive more rapid decline in triglyceride levels. An infusion of dextrose (usually 10% dextrose in water) is often needed to avoid hypoglycemia, with titration of the dextrose infusion based on the serum glucose to target a glucose of ~150-200 mg/dL. This is fairly easy to do within an ICU, with attention to avoid the following complications:
- Hypokalemia, hypophosphatemia (cycle electrolytes and replete as needed).
- Hypervolemia (follow volume status and consider furosemide if the patient is retaining lots of volume from the dextrose infusion)
It's unclear exactly when to stop the insulin infusion. It might be ideal to continue the infusion until the triglyceride level is <500 mg/dL. However, this can take a while. For patients with mild pancreatitis who are already clinically improving, targeting a triglyceride level <1,000 mg/dL may be reasonable.
Many patients with hypertriglyceridemic pancreatitis have underlying diabetes. These patients may be bridged from an insulin infusion to subcutaneous long-acting insulin. The dose of subcutaneous insulin varies, with some type II diabetic patients requiring surprisingly high doses of long-acting insulin.
Heparin to reduce triglyceride levels?
Triglycerides –> Free fatty acids –> Pancreatitis
Heparin stimulates the metabolism of triglycerides into free fatty acids. This may seem like a good idea, because it will reduce triglyceride levels. However, increasing free fatty acid levels is potentially dangerous as this may directly injure the pancreas. In one large series of patients with hypertriglyceridemic pancreatitis undergoing plasmapheresis, anticoagulation with heparin was associated with a 10-fold higher mortality compared to patients treated with citrate (11% vs. 1%; Gubensek 2014).
Pending further study, heparin shouldn't be used intentionally to lower triglyceride levels. Alternative medications for DVT prophylaxis may be considered (e.g. fondaparinux).
- Hypertriglyceridemic pancreatitis is common among critically ill pancreatitis patients, but receives little attention.
- Measuring a triglyceride level should be considered in all patients with pancreatitis.
- The diagnosis of hypertriglyceridemic pancreatitis is based upon an elevated triglyceride level (typically >1,000 mg/dL) and exclusion of other likely causes.
- In addition to general therapies for pancreatitis, patients with hypertriglyceridemic pancreatitis may also benefit from treatments targeted at reducing triglyceride levels (either plasmapheresis or insulin infusion).
- Some recent evidence suggests that insulin infusion may be more effective than plasmapheresis clinically. Insulin infusion is also safer and easier to initiate immediately.
- The myth of large-volume resuscitation in pancreatitis
- Abdominal compartment syndrome (this is what happens if you ignore #1)
- New guidelines simplify ICU nutrition
- Gelrud A and Whitcomb DC. Hypertriglyceridemia-induced acute pancreatitis. UpToDate, accessed 3/6/17.
Image credits: Bomb
Latest posts by Josh Farkas (see all)
- PulmCrit- Toxicology dogmalysis: the osmolal gap - August 14, 2017
- PulmCrit – Central venous oxygen saturation: signal or noise? - July 17, 2017
- PulmCrit- The gag reflex shouldn't be tested in living patients - July 4, 2017