Transient extreme insulin resistance in a patient requiring extracorporeal membrane oxygenation for cardiogenic shock: A case report

Abstract

BACKGROUND

Acute and extreme insulin resistance with persistent hyperglycemia requiring excessively high doses of insulin before rapidly resolving is rare and has been referred to as transient and extreme insulin resistance (TEIR). The underlying pathophysiology and optimal management of TEIR are poorly understood, and previous reports of TEIR in the literature are sparse. This report is the first description of TEIR in a patient requiring mechanical circulatory support (MCS).

CASE SUMMARY

A 62-year-old male developed cardiogenic shock and was placed on veno-arterial extracorporeal membrane oxygenation following percutaneous coronary intervention and successful revascularization. Over the next 24 hours, glucose levels rose and remained markedly elevated despite increasing insulin infusion rates and repeated boluses. The insulin infusion rate peaked at 450 units/hour, and the patient received 4300 units (33 units/kg) of insulin over the 24-hour period of peak insulin resistance. Insulin resistance resolved rapidly, necessitating an abrupt decrease in the insulin infusion rate and development of rebound hypoglycemia.

CONCLUSION

Onset of TEIR did not seem to correlate with end-organ hypoperfusion or vasoactive drug dosing.

Key Words: Veno-arterial extracorporeal membrane oxygenation; Insulin resistance; shock; Mechanical circulatory support; Myocardial infarction; Diabetes; Case report

Core Tip: This paper presents the first report of transient and extreme insulin resistance (TEIR) in a patient receiving mechanical circulatory support (MCS). TEIR is an exceedingly rare condition that occurs in critically ill patients and carries with it a high mortality rate and uncertain pathophysiology. Due to the implementation of MCS and vasoactive medication, this case is uniquely positioned to provide an observation on some of the previously proposed mechanisms of TEIR (tissue hypoperfusion and vasoactive drugs).

INTRODUCTION

Insulin resistance is common in critically ill patients; however, acute and extreme insulin resistance with persistent hyperglycemia requiring excessively high doses of insulin before rapidly resolving is rare[1,2]. This clinical scenario has been described as transient and extreme insulin resistance (TEIR), with extreme insulin resistance being defined as a patient requiring more than 3 units/kg/day of insulin[3]. The underlying pathophysiology and optimal management of TEIR are poorly understood, and previous reports of TEIR in the literature are sparse[4,5]. This report is the first description of TEIR in a patient requiring veno-arterial extracorporeal membrane oxygenation (VA-ECMO) for post-myocardial infarction cardiogenic shock.

CASE PRESENTATION

Chief complaints

The patient presented with chest pain.

History of present illness

The patient presented to an emergency department with an inferior-posterior ST-elevation myocardial infarction. He was initially treated with thrombolytic therapy and subsequently transferred to a tertiary care center for rescue percutaneous coronary intervention. Shortly after arrival he deteriorated, developing a ventricular fibrillatory arrest lasting 20 minutes before a return of spontaneous circulation with cardiopulmonary resuscitation. He was brought to the cardiac catheterization lab where his left ventricular ejection fraction was found to be 30%, and an emergency rescue percutaneous coronary intervention of the culprit right coronary artery occlusion was performed successfully.

Following revascularization, he was transferred to the coronary care unit (CCU) where he developed progressive cardiogenic shock with tachycardia, hypoxemia, pulmonary edema, and escalating vasopressor requirements (norepinephrine 0.6 mcg/kg/min). He was profoundly acidotic (pH < 7.0), with rising lactate levels (10.8 mmol/L), and anuria. The decision was made to initiate VA-ECMO support. A 25 French multistage venous cannula was placed in the left common femoral vein, a 19 French arterial cannula was placed in the left common femoral artery, and an 8 French distal perfusion catheter was placed in the left superficial femoral artery in the CCU. The patient was then transferred to the intensive care unit (ICU) for closer monitoring. They were stabilized on 3.5 Liters per minute of ECMO flow, which achieved a central venous oxygen saturation > 70%. A pulse pressure > 20 mmHg was maintained following initiation of an epinephrine infusion and correction of the acid-base status. A trans-esophageal echocardiogram was performed, and a pulmonary arterial catheter was inserted to rule out the need for a left ventricular vent. Targeted temperature management (TTM) was initiated, and the patient was cooled to 34 °C. Over the next 24 hours the patient demonstrated signs of improved end organ perfusion, and the lactate concentration (based on arterial blood gas samples) decreased from 10.8 to 3.9 mmol/L. The patient developed an acute kidney injury with a creatinine concentration peaking at over 400 mmol/L.

History of past illness

The patient was a 62-year-old male, weighing 117 kg, with a history of ischemic heart disease (coronary artery bypass grafting 8 years prior), type 2 diabetes mellitus (diagnosed 8 years prior, A1C near time of admission was 7.1%). To manage his diabetes, he was prescribed Empagliflozin, Gliclazide, and Janumet; however, medication compliance was questionable. He was not taking insulin at home.

Personal and family history

His medical history also included hyperlipidemia, obstructive sleep apnea, nasal papilloma, and chronic tobacco and marijuana use. Family history was unknown.

Physical examination

Upon admission to the ICU the patient presented with stable vitals on ECMO support (Temperature 36.0 °C, heart rate 93, sinus rhythm, blood pressure 110/65, mean arterial pressure 65, SpO₂ 100%). Pedal pulses were absent to feet bilaterally.

Laboratory examinations

On arrival to the ICU initial laboratory tests revealed potassium 3.3 mmol/L, sodium 144 mmol/L, chloride 101 mmol/L, creatinine 297 μmol/L, urea 11.7 mmol/L, creatine kinase 2892 U/L, and troponin 11800 ng/L. Hemoglobin was 154 g/L and platelets 226 × 10⁹/L. Arterial blood gases revealed pH 7.24, pCO₂ 38 mmHg, pO₂ 249 mmHg, bicarbonate 16 mmol/L, and lactate 10.8 mmol/L. Further workup revealed international normalized ratio 1.2, and patrial thromboplastin time 59.6 seconds. Initial glucose on presentation to ICU was 18.2 mmol/L. Potassium levels dropped the first night in the ICU to 2.9 mmol/L, which required six doses of 20 mmol of potassium chloride intravenously over the course of the night to target a minimum value of 3.5 mmol/L.

On post-admission day 2: Laboratory examinations revealed potassium 3.5 mmol/L, sodium 139 mmol/L, chloride 100 mmol/L, creatinine 351 μmol/L, and urea 13.8 mmol/L. Arterial blood gases showed pH 7.4, pCO₂ 33 mmHg, pO₂ 86 mmHg, bicarbonate 20 mmol/L, lactate 5.6 mmol/L.

On post-admission day 3: Laboratory examinations revealed potassium increased to 6.3 mmol/L and creatinine increased to 420 μmol/L.

On post-admission day 4: Laboratory examinations revealed potassium 5.4 mmol/L, urea 12.1 mmol/L and creatinine 348 μmol/L. Potassium values normalized by post-admission day 6 (3.8 mmol/L).

Imaging examinations

Percutaneous coronary angiography revealed a totally occluded right coronary artery and an ejection fraction of 30%. Chest X-ray on admission to the ICU showed central lines in good positions and clear lungs.

FINAL DIAGNOSIS

Inferior STEMI requiring rescue percutaneous coronary intervention of right coronary artery, cardiogenic shock requiring ECMO, acute kidney failure requiring temporary dialysis.

TREATMENT

Initial glucose on presentation to hospital was 26 mmol/L, it had declined to 18.5 mmol/L at the time of VA-ECMO cannulation at which point an intravenous Humulin R infusion was started at a rate of 10 units/hour. Over the 24 hours following VA-ECMO cannulation, despite escalating insulin infusion rates and repeated boluses (totaling 215 units), the patient's glucose levels remained markedly elevated, peaking at 27 mmol/L (Figure 1). During this period the patient’s epinephrine infusion ranged from 0.06-0.1 mcg/kg/min, and norepinephrine infusion was weaned down from 0.6 μg/kg/min to 0.06 μg/kg/min (Figure 2). Dopamine infusion (at 3 μg/kg/min) was started 22 hours after initiation of ECMO to assist in weaning off epinephrine infusion. Regular electrolyte monitoring was performed, and potassium, calcium, and magnesium were replaced as required. Potassium values dropped during the high rates of insulin infusion and required repeated replacement. Glucose concentration was checked every 30-60 minutes, and insulin requirements reached a maximum of 450 units/hour roughly 24 hours after VA-ECMO cannulation (Figure 1). At this point glucose levels began declining and the frequency of glucose measurement was intensified. Over the next four hours a rapid de-escalation of the insulin infusion dose (from 450 units/hour to 350 units/hour to 250 units/hour to 100 units/hour to 50 units/hour) was required with infusion rates decreasing from 450 units/hour to 50 units/hour. By 36-hours post-cannulation, a normal blood glucose (6.9 mmol/L) was achieved on 5 units/hour of insulin, which was weaned shortly afterwards. The patient did experience some rebound hypoglycemia, lowest reading 3.3 mmol/L.

Figure 1 Blood glucose levels (maroon, left Y-axis), serum lactate levels (green, left Y-axis), insulin bolus (orange, left Y-axis) compared with insulin infusion rates (blue, right Y-axis) since start of veno-arterial extracorporeal membrane oxygenation.

Figure 2 Infusion rates of vasoactive medication (left Y-axis), vs infusion rates of insulin (right Y-axis) since start of veno-arterial extracorporeal membrane oxygenation.

OUTCOME AND FOLLOW-UP

Two days after ECMO initiation, an improvement in cardiac function was observed. His epinephrine infusion was weaned and repeat echocardiography demonstrated a left ventricular ejection fraction of 45% with moderate right ventricular dysfunction. He subsequently underwent successful ECMO decannulation, following which he developed hyperkalemia. The hyperkalemia was managed with a dose of sodium polystyrene sulfonate (30 g) and hemodialysis, which was performed the same day as ECMO decannulation and the day after. The patient continued to recover and was successfully discharged from hospital 11 days after initial hospital admission.

DISCUSSION

This case illustrates an episode of TEIR following myocardial infarction and cardiogenic shock requiring VA-ECMO cannulation, with a peak insulin infusion rate of 450 units/hour. The patient received 4300 units (33 units/kg) of insulin over the 24-hour period of peak insulin resistance. Subsequently, there was a rapid reversal of insulin resistance that coincided with a mean lactate (during the 12-hour peak of insulin infusions) of 4.85 mmol/L and moderate dose of catecholamine infusion support. The patient then recovered sufficiently from myocardial dysfunction to facilitate ECMO decannulation; however, he subsequently required transient hemodialysis for acute hyperkalemia possibly related to reperfusion of subclinical ischemia in the cannulated leg. The patient was eventually discharged neurologically intact. To the best of our knowledge, this is the first case report documenting TEIR in a patient requiring VA-ECMO for post-myocardial infarction cardiogenic shock.

Previous case reports of patients presenting with TEIR in the context of adverse cardiac events are summarized in Table 1. Myocardial infarction and subsequent cardiac arrest were the most common presentations in this group of patients. In our case and two of the previously reported cases, patients required dialysis for their acute kidney injury. As in our case, two of the case reports also employed TTM. The duration of insulin infusion varied considerably among reported cases, ranging from 25 hours up to 26 days. Interestingly, all but one of the case reports observed a resolution of the extreme insulin resistance within 24 hours of the peak insulin infusion rate (Table 1). We observed that the insulin infusion was able to be weaned from a peak of 450 units/hour to 0 units/hour within 12 hours (Figure 1).

Table 1 Case reports of transient extreme insulin resistance in patients with adverse cardiac events.

Ref.	Myocardial infarction	Concurrent conditions	Therapeutic hypothermia	Peak insulin infusion rate	Peak insulin daily dose	Peak to resolution time	Duration of insulin infusion	Rebound hypoglycemia	Vasopressors	Peak lactate	Mechanical circulatory support	Outcome
Current paper	STEMI	Cardiac arrest (20 minutes), metabolic acidosis, acute kidney injury (dialysis)	Yes (34 ℃)	450 units/hour	4300 units	13 hours	38 hours	No	Epinephrine Norepinephrine dopamine	10.8 mmol/L	Veno-arterial extracorporeal membrane oxygenation	Survived
Wei et al[5]	NSTEMI	Cardiac arrest (6 minutes), sepsis, mixed acidosis, pneumonia, acute respiratory failure, AKI (dialysis)	Yes (no temperature reported)	960 units/hour	18 224 units	8 hours	11 days	Yes	Norepinephrine	19 mmol/L	No	Deceased (recurrent sepsis)
July et al[14]	STEMI	Cardiac arrest (unknown duration). acute kidney injury (renal replacement therapy), metabolic acidosis	Yes (32 ℃)	280 units/hour	Not reported	24 hours	5 days	No	No	1.9 mmol/L	No	Deceased (anoxic brain damage)
Oo et al[6]	NSTEMI	Diabetic ketoacidosis, acute on chronic kidney injury	No	120 units/hour	2400 units	> 5 days	26 days	No	No	Not reported	No	Survived
Surani et al[16]	None	Cardiac arrest (unknown duration)	No	832 units/hour	> 15 000 units	6 hours	25 hours	No	Norepinephrine	Not reported	No	Not reported
Yokoyama et al[9]	NSTEMI	Diabetic ketoacidosis, rhabdomyolysis, acute kidney injury, pneumonia, disseminated intravascular coagulation.	No (however, body temperature was 35 ℃ on arrival)	10 000 units/hour	91 580 units (over 25 hours)	6 hours	25 hours	Yes	Dopamine	Not reported	No	Survived

AKI: Acute kidney injury.

ECMO and TEIR

The role that myocardial dysfunction plays in TEIR remains poorly understood. Oo et al[6] reported a case of TEIR in a patient who suffered a myocardial infarction; however, in their case the insulin resistance improved following coronary revascularization (coronary artery bypass grafting). In the present report, insulin resistance persisted and worsened following revascularization of the culprit coronary artery. The interpretation of this is complicated and further confounded by the cardiogenic shock and initiation of VA-ECMO shortly following revascularization.

The potential role of VA-ECMO in the pathophysiology of TEIR has not been discussed. One of the proposed mechanisms of TEIR is rooted in an overactivation of the inflammatory system[5]. The exposure of circulating blood to the ECMO circuit causes a response similar to that of the Systemic Inflammatory Response Syndrome[7]. This results in a cascade of proinflammatory, coagulation, and immunological changes, which may play an important role in TEIR. Two key cytokines that have been demonstrated to play a role in insulin resistance are interleukin-6 (IL-6) and tumour necrosis factor-alpha (TNF-α), both of which become elevated during ECMO[7,8]. The possibility that VA-ECMO may have contributed to the pathophysiology of TEIR in this case warrants consideration. However, if increased IL-6 and TNF-α levels induced by ECMO initiation was the principal driver of the insulin resistance, then it would be expected to see the insulin resistance continue until after ECMO decannulation. We observed that TEIR was resolved while the patient was still supported with VA-ECMO, suggesting that mechanical circulatory support was unlikely to be the primary cause of insulin resistance in this case.

This case also provides a unique perspective on the role of hemodynamics and end organ perfusion on the pathophysiology of TEIR. It has been proposed that tissue hypoperfusion in shock states may contribute to insulin resistance, by decreasing the delivery of insulin to the peripheral tissues[9]. In our case, we were able to restore end organ oxygen delivery following VA-ECMO cannulation as evidenced by a normal central venous oxygen saturation; however, TEIR developed over the next 24 hours and did not correlate with the severity of end organ hypoperfusion. Lactate, which is a product of anaerobic glycolysis, increases during episodes of tissue hypoxia[10]. Interestingly, the peak lactate concentration was observed 24 hours prior to the onset of TEIR, and the mean lactate level during the 12 hours period of peak insulin requirement was 4.8 (Figure 1). This suggests that inadequate tissue perfusion may not have been the primary driver of insulin resistance in this case.

Vasopressors and TEIR

It has been demonstrated previously that exogenous norepinephrine decreases insulin sensitivity, likely via action on the liver and skeletal muscle[11]. Similarly, epinephrine has been shown to block the inhibitory effect of insulin on hepatic glucose production and impair tissue sensitivity to insulin through β-adrenergic mechanisms[12]. We observed that the TEIR subsided before the intravenous norepinephrine and epinephrine doses were weaned (Figure 2), suggesting that exogenous norepinephrine and epinephrine were not likely the principal cause of insulin resistance in this case.

Autoimmunity and TEIR

Another possibility to be reviewed is an antibody to the insulin receptor or an insulin autoantibody. Insulin autoantibodies, when present, can produce a similar phenomenon that was seen in this case, resulting in extreme insulin resistance, and rapid resolution of symptoms[13]. The rapid resolution is within the context of being treated with an immunosuppressive therapy such as mycophenolate mofetil[13]. While insulin auto-antibodies were not screened for during this patient’s admission, the acute timeline and rapid resolution without the use of an immunosuppressive agent makes this diagnosis unlikely.

TTM and TEIR

TTM was utilized as a part of the post-resuscitation care to minimize the risk of neurologic injury in this case. Severe insulin resistance during TTM has been previously reported[14-16]. July et al[14] documented a case of TEIR in a patient receiving TTM, and described a clear correlation between onset of TEIR and initiation of therapeutic hypothermia. Interestingly, they also observed a resolution of TEIR once the patient was rewarmed from approximately 32 to approximately 34 °C[14]. Patient’s temperature and insulin infusion rate in our case are displayed in Figure 3. The insulin infusion rate rose rapidly 6 hours after the patient was cooled to 34 °C. This trend continued until 3 hours after the core body temperature normalized (36.1 °C), at which point the insulin requirements declined quickly. However, any correlation between therapeutic hypothermia and insulin resistance in this case must be interpreted with caution due to the presence of multiple confounding factors.

Figure 3 Insulin infusion rates (right Y-axis), compared with core body temperature (left Y-axis), over time since initiation of veno-arterial extracorporeal membrane oxygenation.

CONCLUSION

Transient extreme insulin resistance may develop in a variety of clinical scenarios and necessitates close monitoring and high doses of insulin to manage hyperglycemia. Once insulin resistance begins to resolve, monitoring must be intensified, and rapid de-escalation of infusion rates are required. In this case, the onset of TEIR did not seem to correlate with the timing of end-organ hypoperfusion or vasoactive drug dosing. While there is a possible correlation between therapeutic hypothermia and the severity of insulin resistance in this case, multiple confounding factors were present. More research is needed to better characterize the pathophysiology of TEIR and identify optimal management strategies.