Effect of insulin on hyperkalemia during anhepatic stage of liver transplantation

Abstract

AIM: To investigate the effectiveness of insulin on decreasing serum potassium concentration during anhepatic stage of orthotopic liver transplantation.

METHODS: Sixteen patients with serum potassium concentrations greater than 4.0 mmol/L at the onset of anhepatic stage were randomized into two groups. The patients in control group (n = 8) received no treatment, while those in treatment group (n = 8) received an intravenous bolus injection of regular insulin (20 U) 10 min into the anhepatic stage, followed by a glucose infusion (500 mL 50 g/L dextrose) over 15 min.

RESULTS: In control group, potassium concentration underwent no changes whereas in treatment group, it decreased from 4.8 ± 0.48 mmol/L to 4.19 ± 0.55 mmol/L (mean ± SD) within 15 min and to 3.62 ± 0.45 mmol/L 60 min after the therapy. The potassium concentration was lower in treatment group than in control group within 30 min of treatment (3.94 ± 0.57 vs 4.47 ± 0.42 mmol/L, respectively; P < 0.05), and increased similarly 30 s after graft reperfusion in both groups of patients, but remained lower in treatment group (5.81 ± 1.78 vs 7.44 ± 1.75 mmol/L, respectively; P < 0.05). The potassium concentration returned to pre-reperfusion levels within 5 min after graft reperfusion.

CONCLUSION: In patients undergoing orthotopic liver transplantation, the administration of insulin rapidly decreases serum potassium concentration even in the absence of the liver, suggesting an important contribution by extrahepatic tissues in insulin-stimulated uptake of potassium.

Key Words: $[Keywords]

INTRODUCTION

Severe hyperkalemia is a serious complication in orthotopic liver transplantation (OLT) and usually results from large potassium (K⁺) load following massive blood transfusion or graft reperfusion, especially in patients with renal insufficiency[1,2]. The very short-lived but severe hyperkalemia of greater than 7-8 mmol/L caused by washout of K⁺ from the graft may be a factor in the circulatory impairment associated with reperfusion. It was reported that 18 cases of cardiac arrest during OLT in the First People’s Hospital of Tianjing, China, happened at the minute of reperfusion[3]. Therefore, successful treatment of hyperkalemia, particularly during the anhepatic stage, can be of vital importance during OLT. Management strategies for decreasing serum K⁺ concentration include administration of washed packed red cells, hemodialysis, autotransfusion, diuretics, as well as other measures to promote K⁺ redistribution by, for instance, alkalinization and application of β-adrenergic agonist or insulin. However, some of these techniques are cumbersome or have not been well studied in the setting of OLT.

Insulin decreases serum K⁺ concentration in healthy persons and in patients with renal failure[4]. The effect of insulin on serum K⁺ in patients with liver disease is unknown; in normal patients, the liver accounts for about 70% of the total K⁺ removed from the circulation during the first hour after insulin administration[5]. Other metabolic actions of insulin can be abnormal in the setting of hepatic dysfunction: Patients with liver disease are frequently resistant toward the glucose-decreasing effects of insulin[6-9]. Furthermore, during OLT, the anhepatic stage excludes any hepatic function. Thus, the purpose of this study was to determine the effectiveness of insulin on decreasing serum K⁺ concentration during anhepatic stage of OLT.

MATERIALS AND METHODS

Sixteen adult patients undergoing OLT with serum K⁺ concentrations greater than 4.0 mmol/L at the onset of anhepatic stage were studied. Patients with diabetes mellitus were excluded because of abnormal glucose metabolism, along with those having renal insufficiency due to potential decrease in renal K⁺ excretion. Anesthesia was induced and maintained with isoflurane, fentanyl, and pancuronium. Intraoperative management followed the standard care procedures used during OLT at our hospital including the use of pulmonary arterial and femoral arterial catheters with frequent determination of hemodynamic variables. Normocarbia was maintained (PaCO₂ of 35-40 mmHg, 1 mmHg = 0.133 kPa) and pH kept within the normal range (7.3-7.45). Venovenous by-pass was not performed during the anhepatic stage in all cases.

The patients were randomized into control group (n = 8, without insulin therapy) and treatment group (n = 8) receiving an intravenous bolus injection of regular insulin (20 U) 10 min after the start of anhepatic stage, and immediately followed by an intravenous glucose infusion (500 mL 50 g/L dextrose) over 15 min. No other techniques were used to deliberately decrease K⁺ concentration in any of the cases. Serum K⁺ and glucose concentrations were determined at the time points of 5 min before, 5, 10, 15, 30, 45 and 60 min after the administration of insulin and glucose (IG-5′, IG+5′, IG+10′, IG+15′, IG+30′, IG+45′, and IG+60′, respectively), and at 30 s, 5 and 30 min after graft reperfusion (III+30″, III+5′, and III+30′, respectively). Blood urea nitrogen and creatinine concentrations were determined at the beginning of the procedure. Urine output and urine K⁺ concentration were measured during the anhepatic stage.

Statistical analysis

The results presented as mean ± SD were analyzed using one-factor analysis of variance and analysis of variance for repeated measures. A P value less than 0.05 was considered statistically significant.

RESULTS

The two patient groups matched for age, weight, and height had similar baseline blood urea nitrogen and creatinine without obvious differences in urine output, urine K⁺ concentration, renal K⁺ excretion during the anhepatic stage or transfusion requirements.

Immediately before the administration of insulin and glucose (IG-5′), serum K⁺ concentration was slightly decreased in control group than in treatment group (P > 0.05), and the concentration was ≥ 5.0 mmol/L in one patient in control group and in 3 in treatment group. During anhepatic stage, K⁺ concentration exhibited no obvious changes in control group, but increased 30 s after graft reperfusion to 7.44 ± 1.75 mmol/L followed by reduction to pre-reperfusion levels at 5 and 30 min after reperfusion. In treatment group, K⁺ concentration decreased within 15 min after insulin administration, and remained lower than that in control group within 30 min after treatment. Immediately after graft reperfusion, K⁺ concentration increased in treatment group to the level of control group, but as K⁺ concentration had been kept below the pre-reperfusion level, the peak K⁺ concentration in treatment group (5.81 ± 1.78 mmol/L) was lower than that the control group (P < 0.05).

As shown in Figure 1, hypokalemia (K⁺ ≤ 3.0 mmol/L) was not seen in control group. In treatment group, however, hypokalemia occurred in one case during anhepatic stage (serum K⁺ concentration of 2.8 mmol/L) and in 3 till 5 and 30 min after graft reperfusion, but cardiac arrhythmia was not incurred.

Figure 1  Serum potassium concentration during liver transplantation in the control group and treatment group. ^aP < 0.05 compared with IG-5′; ^cP < 0.05 compared with control group. IG represents insulin and glucose treatment and III stands for neohepatic stage (n = 8).

Measurements of glucose levels in these patients are presented in Table 1. Hypoglycemia was observed in none of the patients. Two patients in control group had lower-than-normal serum glucose during the anhepatic stage. Two patients in control group and 7 in treatment group developed hyperglycemia (serum glucose ≥ 250 mg/L) before or after graft reperfusion, but the hyperglycemia was usually mild. The pH values were similar in the two groups.

Table 1 General clinical data of two groups¹ of patients (n = 8).

	Control group	Treatment group
Age (yr)	40.2 ± 10.5	42.5 ± 9.8
Weight (kg)	54.3 ± 11.6	52.6 ± 10.9
Height (cm)	162 ± 9.5	160 ± 8.7
Baseline BUN (mg/L)	26 ± 20	25 ± 17
Baseline creatinine (mg/L)	1.6 ± 1.8	1.4 ± 1.7
Urine output (mL)	150 ± 160	202 ± 161
Urine K⁺ (mmol/L)	40.5 ± 9.4	39.7 ± 11.8
Renal K⁺ excretion (mmol)	5.5 ± 4.1	7.8 ± 4.6
Packed RBC (mL)	855 ± 410	910 ± 480
Fresh frozen plasma (mL)	420 ± 245	460 ± 255

¹The two patient groups are matched for the indexes listed in the table (P > 0.05).

Two patients in each group required low-dose bolus injections of epinephrine (≤ 50 μg) upon graft reperfusion to correct transient hypotension (systolic blood pressure ≤ 9.3 kPa).

DISCUSSION

Hyperkalemia following massive transfusion, renal insufficiency, and a sudden release of K⁺ from the graft on reperfusion remains the major cause for morbidity and mortality during OLT. A short-lived, severe post-reperfusion hyperkalemia may contribute to hypotension and bradycardia, known as the “postreperfusion syndrome”[10-13]. A severe form of postreperfusion syndrome results in cardiac arrest, caused by a brief episode of hyperkalemic cardioplegia in addition to other metabolic and hemodynamic changes[3]. Because of the severe potential complications of hyperkalemia during OLT, it is of utmost importance to decrease the baseline K⁺ concentration effectively and quickly before graft reperfusion.

In healthy persons, total body K⁺ content depends on the balance between its intake and renal excretion. The plasma K⁺ concentration is a function of total body K⁺ content and its distribution between the extracellular and intracellular compartments. Normally about 2% of the total body K⁺ is in the extracellular fluid compartment. Intracellular K⁺ concentration depends mainly on the activity of sodium-potassium ATPase and is influenced by the acid-base status and by hormones such as epinephrine, aldosterone, and insulin[14-17]. After an acute K⁺ load, homeostasis is maintained by an insulin-dependent translocation of K⁺ into the intracellular space, with liver and muscles representing the major buffering systems.

Frequently, K⁺ homeostasis is abnormal in patients with liver cirrhosis. Decreased total body K⁺ content results in chronic hypokalemia and is probably due to hyperaldosteronism, diuretic therapy, and gastrointestinal losses. Furthermore, tolerance to exogenous K⁺ loading is reduced due to a decreased K⁺ uptake by the cirrhotic liver, possibly because of reduced sensitivity to insulin, and also attributable to reduced muscle mass.

Therapy for hyperkalemia is directed both at decreasing the harmful effects of hyperkalemia on the heart and at lowering the K⁺ concentration. Calcium chloride, epinephrine, and magnesium sulfate may decrease the effects of hyperkalemia on myocardial cells. The K⁺ load may be lessened by the use of fresh blood products or by washing of packed red cells before their administration, because K⁺ concentration in 35-day-old packed red cells (with ± 70 mL of plasma) may reach 76 mmol/L. Potassium can be removed by the use of hemodialysis, exchange transfusion, and diuretics. Hemodialysis is the most efficient way to rapidly decrease K⁺ concentration. Exchange autotransfusion has also been used effectively to decrease the K⁺ concentration during OLT. Dialyzing the effluent blood from the liver upon graft reperfusion before it enters the systemic circulation reduces the degree of reperfusion hyperkalemia and incidence of cardiac arrest in an animal model. However, all the techniques described above are fairly complicated and impractical in the clinical setting. In addition, diuretics are relatively ineffective and of slow action. Finally, K⁺ concentration can be decreased by promoting redistribution of K⁺ using alkalinization, β₂-adrenergic agonists, and insulin.

Alkalosis decreases plasma K⁺ concentration by shifting K⁺ to intracellular compartment in addition to increasing distal tubular K⁺ excretion by the kidney. However, intravenous sodium bicarbonate seems to be ineffective in rapidly correcting hyperkalemia in patients with renal insufficiency. Moreover, its hypertonicity may raise K⁺ concentration as a result of cellular dehydration, which leads to increased intracellular K⁺ concentration followed by passive K⁺ diffusion out of cells.

Epinephrine directly stimulates K⁺ uptake by skeletal and cardiac muscle cells and hepatocytes by a β₂-adrenergic receptor-mediated increase in cAMP, resulting in activation of sodium-potassium ATPase. However, epinephrine only produces a small decrease in K⁺ concentrations in merely half of the patients with renal failure, and tachycardia is a common side effect of epinephrine.

Insulin quickly and effectively decreases serum K⁺ concentrations in healthy volunteers. In patients with renal failure, insulin takes effect within 10 min and is much more effective than epinephrine. The effect of insulin on hyperkalemia is dose-dependent, even in patients with diabetes mellitus. Insulin-induced K⁺ uptake is mediated by direct stimulation of the activity of sodium-potassium ATPase and is independent of insulin-stimulated glucose uptake. Both insulin and β₂-adrenergic agonists stimulate sodium-potassium ATPase, but their effects are additive, indicating that their intracellular signals are different. Under normal conditions, the cellular uptake of K⁺ by splanchnic tissues during the first 60 min after insulin administration accounts for approximately 70% of the total decrease in extracellular K⁺ content; thereafter K⁺ is taken up predominantly by skeletal and cardiac muscle and adipose tissues. Obviously, glucose should be administered along with insulin to avoid hypoglycemia. As hypertonicity may arise from the injection of 500 g/L glucose, 50 g/L glucose administered through continuous infusion can be a better alternative.

The results of this study demonstrated that insulin effectively decreased serum K⁺ concentration within 15 min, and the effects lasted for at least 60 min even in the absence of the liver and in patients with abnormal K⁺ homeostasis or insulin resistance. The reduction in K⁺ concentration could not be explained by increased renal K⁺ excretion, which scarcely differed between the groups. Many tissues, including the skeletal and cardiac muscles and adipose tissues, probably contributed to the decrease in serum K⁺ concentration. This study did not include patients with renal insufficiency or with insulin-dependent diabetes, though insulin is also effective in decreasing K⁺ concentration in these patients. Although K⁺ concentration 30 s after graft reperfusion was greater in control group, the incidence of hemodynamic instability (defined as the need for epinephrine to maintain arterial blood pressure) did not differ between the groups. The number of patients in our study might have been too small to reveal such a difference. Another possibility is that the degree of potassium increase is more important than the potassium level itself, and gradual decrease in K⁺ concentration by insulin followed by a rapid increase in K⁺ concentration on reperfusion might lead to more cardiac irritability. Finally, arrhythmias and hemodynamic disturbances at the time of reperfusion were due to several factors, and not to potassium shifts alone.

Three patients in treatment group had hyperglycemia. This could probably be avoided by administering a smaller dose of glucose or by decreasing the infusion rate of 5% dextrose. Hypokalemia (≤ 3.0 mmol/L) was also observed in 3 patients of treatment group during the neohepatic stage, but cardiac arrhythmias did not occur.

In conclusion, the administration of insulin rapidly decreases serum K⁺ level during the anhepatic stage of OLT, suggesting an important contribution by extrahepatic tissues in insulin-stimulated uptake of K⁺.