Risk Factors for Central Pontine and Extrapontine Myelinolysis After Liver Transplantation

A Single-Center Study

Abstract

Background 

Central pontine and extrapontine myelinolysis (CPM/EPM) are severe neurologic complications after liver transplantation.

Methods 

The present work retrospectively evaluated single-center prevalence of CPM/EPM and associated risk factors: cause of liver disease, hepatic encephalopathy, preoperative, intraoperative, and perioperative blood components use, serum levels, and variation of Na⁺, Cl⁻, and K⁺ and immunosuppression were compared between CPM/EPM patients and control group of transplanted patients without neurologic complications.

Results 

Among 997 transplants, CPM/EPM were diagnosed in 11 patients (1.1%), of whom four were CPM, one was EPM, and six were associated CPM and EPM. Control group consisted of 44 transplanted patients. Central pontine and extrapontine myelinolysis patients experienced higher intraoperative and perioperative serum Na⁺/24 hr variations compared to controls (16.69±5.17 vs. 9.8±3.4 mEq/L, P = 0.001). Maximum peak of intraoperative or perioperative serum Na⁺ was significantly higher in patients compared to controls (151.5±3.3 vs. 140.8±6.2 mEq/L, P≤0.001), but no difference in preoperative serum Na⁺ was detected. Three patients presented hypernatremia as isolated risk factor.

Conclusion 

Extrapontine myelinolysis can be found isolated or associated with CPM in up to two of three liver transplanted patients with myelinolysis. A marked variation of perioperative serum Na⁺ remains the main risk factor even in patients without preexisting hyponatremia; however, isolated hypernatremia may be solely responsible in some cases.

In Brief

The authors conduct a single-center analysis of liver allograft recipients and show that central pontine and extrapontine myelinosis remain uncommon, can be associated with each other, and are loosely associate with changes in serum sodium concentrations.

Central pontine myelinolysis (CPM) is a clinical syndrome first reported in 1959 by Adams et al.¹ and described as a demyelinating disorder affecting the central pons, typically in malnourished, alcoholic, and chronically debilitated patients. The original clinical description included rapidly evolving quadriplegia, dysarthria, dysphagia, and mutism. The definition of CPM evolved over the years, and the fact that the demyelinating lesions were not confined to the pons, but could also affect basal ganglia, thalamus, lateral geniculate body, and so on, led to the definition in 1962 of extrapontine myelinolysis (EPM). Categories at risk of developing demyelinating disorders also evolved over time, including septic, burn, diabetic, human immunodeficiency virus, and liver transplanted patients. These categories of patients were all linked by the presence of electrolyte imbalances (especially hyponatremia) and a possible too rapid correction thereof.

Patients affected by hepatic insufficiency are particularly prone to the development of myelinolysis because of several factors, such as malnutrition, alcohol abuse and frequent chronic hyponatremia. Patients with hepatic encephalopathy present marked depletion of myoinositol, which, together with other organic osmolytes, protects the brain from rapid changes in serum osmolality. Chronic liver disease also leads to a negative nitrogen balance, which reduces the availability of the pool of amino acids necessary for the synthesis of organic osmolytes.² Moreover, liver dysfunction entails a depletion of glycogen brain reserves: glial cells need glucose to activate the Na⁺-K⁺ ATP-ase, thus, in the absence of reserves, even small electrolyte imbalances can lead to energy depletion and cellular death.³

Central pontine myelinolysis was first described after liver transplantation by Starzl et al. in 1978.⁴ It is one of the most serious neurologic complications after liver transplantation and the incidence varies from 0.94% to 3% in current clinical series^5,6 (Fig. 1). Most cases of CPM/EPM (90%) occur within the first week after surgery and in another cohort more than 50% of neurologic symptoms—varying from stupor to locked-in syndrome - appear within the 7th post-operative day.⁶

FIGURE 1: MRI, fluid attenuated inversion recovery axial images of the pons (A) and the corresponding magnification (B). The images showed the symmetric signal alteration in the basis pontis, with a “trident” shape, typical of central pontine myelinolysis. MRI, magnetic resonance imaging.

Liver transplantation is burdened by frequent osmotic and electrolyte imbalances. Intraoperatively, an abundant infusion of sodium-rich fluids is often necessary to replace blood loss, but the ability of sodium renal excretion is impaired in cirrhotic patients.

Since 1978, 163 cases of CPM/EPM developed after liver transplantation in adults have been reported. However, limited data on the risk factors of this condition in liver transplantation have been evaluated.

RESULTS

Clinical Characteristics of the CPM/EPM Group

Between November 1990 and June 2011, 997 adult patients underwent liver transplantation at the Padua University Liver Transplant Center. One hundred forty (14%) patients developed postoperative neurologic complications and were evaluated by neurologists. In 11 patients (9 males, 2 females; mean±SD age, 53.6±8.8 years) CPM or EPM (four CPM, one EPM, and six mixed CPM/EPM) were diagnosed after liver transplantation; clinical signs and symptoms of CPM/EPM occurred in patients after a mean±SD interval time of 6.9±2.5 days after liver transplantation. The incidence of myelinolysis was 1.1%.

The control group included 44 liver transplanted patients (26 males, 18 females; mean±SD age, 51.8±9.1 years), without neurologic complications. Patients’ and controls’ clinical findings are summarized in Table 1 (Table S1, SDC, https://links.lww.com/TP/B83).

TABLE 1: Clinical characteristics of transplanted patients with CPM and/or EPM and transplanted controls without neurologic complications

Clinical manifestations were consistent with the classic description of the disease: drowsiness and stupor were reported in six of 11 patients (54.5%) and coma in four of 11 patients (36.3%); three patients experienced seizures (27.2%), without demonstrated underlying infection or peaks in serum levels of calcineurin inhibitor (CNI). Most patients presented severe muscle strength deficits (especially patients with EPM), which resulted in flaccid hemiplegia and quadriplegia in two patients; in more than half of patients (7/11, 63.6%), dysarthria was reported. Median intensive care unit stay for CPM/EPM patients was 18 days,

Risk Factors for CPM/EPM

Risk factors for the development of CPM/EPM were compared between patients and the control group without myelinolysis (Table 2).

TABLE 2: Evaluated risk factors for the development of CPM/EPM

Central pontine and extrapontine myelinolysis patients experienced a significantly higher perioperative Na⁺ variations in 24 hr compared to transplanted controls (mean±SD Na⁺ changes: 16.69±5.17 vs. 9.8±3.4 mEq/L, P = 0.001), but no difference in preoperative Na⁺ serum values was found between the two groups (132.4±7.6 vs. 134±5.2 mEq/L, P = 0.4). Central pontine and extrapontine myelinolysis patients reached perioperative maximum peak Na⁺ serum value significantly higher than controls before the onset of neurologic symptoms (151.5±3.3 vs. 140.8±6.2 mEq/L, P = <.001). Furthermore, the whole cohort of patients was divided according to the presence of very severe (<125 mEq/L) or severe hyponatremia (125‐130 mEq/L), there was no significant higher incidence of CPM/EPM (2/9 vs. 3/12; P = 1).

There was no difference of intraoperative serum K⁺ and Cl⁻ changes between patients and transplanted controls.

Mean±SD perioperative osmolality was higher in CPM/EPM group than in the control group (319.68±17.2 vs. 298.6±15.2 mOsm/kg, P = 0.004).

Immunosuppressive therapy with tacrolimus was introduced in 37 of 55 (67.2%) patients. Mean±SD tacrolimus levels until the first week after surgery were 3.84±0.33 μg/L and 4.28±1.72 μg/L in CPM/EPM patients and in control group, respectively (P = 0.4). Overtherapeutic levels of immunosuppressors were seen in one patient (9%) who developed CPM and was subsequently switched to tacrolimus, and in eight patients (18%) in control group (P = 0.6); in CPM/EPM group, four of six patients were switched to tacrolimus. In eight of 11 patients, the dose of CNIs was reduced with introduction of a non-CNI immunosuppressor. No patients withdrew immunosuppression.

The same risk factors were compared between CPM/EPM patients of “hypernatremic” and of “hyponatremic” origin (Table 3).

TABLE 3: Evaluated risk factors for the development of CPM/EPM in patients of “hypernatremia” and “hyponatremic” origins

The variation of the sodium serum level in 24 hr including surgery and after postoperative period was significantly lower in CPM/EPM patients of “hypernatremic” origin than in patients of “hyponatremic” origin (10.53±1.1 vs. 19±3.94 mEq/L; P = 0.006). A difference in preoperative serum potassium level did not reach significance.

Patients with CPM/EPM of “hyponatremic” origin presented greater intraoperative transfusion volume compared to patients with isolated hypernatremia (13.7±6.6 vs. 5±0.86 L, P = 0.05); perioperative osmolality was not different between the two groups (320.8±15.6 vs. 316±28.6; P = 0.7).

During the follow-up, there were no death directly related to the neurologic complication in the study cohort, with only two patients who died because of recidivism of hepatitis C virus–related cirrhosis. One and 5 years patients’ survival was 90.9% and 81.2% respectively.

DISCUSSION

The first report of CPM/EPM after liver transplantation dates back to 1978.⁴ Lampl and Yazdi collected the 77 cases of CPM reported in literature from 1986 and 2002.⁷ Overall, we found 115 cases of CPM, seven of isolated EPM and 41 of combined CPM/EPM after liver transplantation.^2-4,6,8-52 However, liver transplanted patients are the third largest group of CPM/EPM cases, after chronic alcoholics and patients affected by severe electrolyte imbalances.⁷

The incidence of myelinolysis in Padua University Liver Transplant Center from November 1990 to June 2011 was 1.1%, similar to the ones reported by the most recent clinical series, which vary from 0.94% to 3%.⁶

The present study registered a distribution of CPM and EPM differing from those found in literature, with a distinct predominance of the last forms, although often in combination with CPM: in our series, four of 11 patients were affected by isolated CPM, one of 11 by EPM and exceptionally six of 11 by association of CPM and EPM. Analyzing the total number of posttransplant cases reported in the literature, 83.1% of patients were suffering from CPM, 4.5% from EPM, and the remaining 12.3% from an association of the two.

Our data agreed with the autoptic findings of Gocht and Colmant⁵³ in nontransplanted patients (affected by cirrhosis, malnutrition, sepsis, and alcoholism), where 46.6% of 58 patients presented only a pontine damage, 31% an association of pontine and extrapontine lesions and 22.4% only an extrapontine damage.

In our series, neurologic signs typical of CPM/EPM appeared after a mean±SD interval of 6.9±2.6 days after liver transplantation, which is close to the data reported in literature.⁵⁴ Lee et al.⁶ documented the onset of neurologic symptoms within 2 weeks after surgery in 90.1% of their patients, whereas Lampl and Yazdi⁷ supported the possible onset of symptoms within 1 month from transplantation.

The purpose of this study was to analyze the risk factors for the development of CPM/EPM after liver transplantation, because, despite several reports of demyelinating complications after surgery, only three studies have been published with this aim.^6,28,31

In our series, CPM/EPM patients experienced a significantly higher intraoperative and perioperative Na⁺ variation in 24 hr, compared to transplanted controls. This finding is consistent with previous cases: the cause of CPM/EPM is specified only in about 70% of cases reported in literature, and 81.7% of them developed the disease after an excessive or too rapid correction of hyponatremia (the average correction of serum sodium concentration in patients affected by myelinolysis is 18.1 mEq/L/24 hr and the pretransplant sodium is 125.9 mEq/L).^2,3,6,9-52

Unlike Lee et al.’s analysis, our results did not show a significant difference in the preoperative sodium serum values between CPM/EPM patients and transplanted controls. In addition, as reported by Abbasoglu et al.,²⁸ similar variations of serum sodium concentration were registered in patients who did not develop demyelinating diseases. These data support two concepts: first, the importance of individual response of the brain to similar electrolyte imbalances for the development of the disease; second, the possibility that hypernatremia, arising after the correction of hyponatremia or in its absence, may play a role in the development of the demyelinating complications, at least in some cases. In fact, our patients reached a maximum Na⁺ serum value significantly higher than controls before the onset of neurologic symptoms.

Only three cases of myelinolysis caused by isolated hypernatremia were reported in the literature after liver transplantation⁴⁴: the mean variation of sodium in the first postoperative day was found to be 11.33±3.17 mEq/L, whereas the mean preoperative Na⁺ was 136,3±2,88 mEq/L.

The analysis of risk factors in our cohort was carried out by separating patients in two groups, one of “hypernatremic” origin and the other of “hyponatremic” origin. In the first one, change of sodium serum levels between surgery and after 24 hr was significantly lower than that in the second one (10.53±1.1 vs. 19±3.94 mEq/L, P = 0.006). Moreover, the group undergoing an overcorrection received a significantly greater intraoperative volume of blood components (13.75±6.06 vs. 5±0.86 mEq/L/die, P = 0.05). This rapid correction of sodium levels was assumed as a risk factor for the development of myelinolysis also in the study of Lee et al.⁶

The noteworthy presence of hypernatremic peaks in our cohort coexisted with an increased frequency of CPM/EPM. In detail, one patient with hypernatremic origin presented lesions in the limbic system network identical to those reported by Cagnin et al.⁵⁵ in a patient who died after extreme hypernatremia because of peritoneal lavage with hypertonic saline solution after hydatid cyst rupture. This condition might be sustained by a predisposing condition of the lymbic system toward glutamate excitotoxicity (Fig. 2). In the literature, there are no data linking the type of hydroelectrolyte imbalance to the areas of injury, thus this association remains speculative. However, this hypothesis represents an interesting starting point for future investigations.

FIGURE 2: MRI, fluid attenuated inversion recovery axial images of hippocampus, insula, and middle cell of lateral ventricle after EPM onset. Images show a marked atrophy of the tonsils and the hippocampal formation with ex vacuo dilatation of the frontal horns of lateral ventricles (A) of the insular cortex with expansion of tank Silviane (B) and of the anterior cingulate cortex (C). MRI, magnetic resonance imaging; EPM, extrapontine myelinolysis.

Higher perioperative osmolality in patients with CPM/EPM compared to controls was consistent with the raise in serum sodium concentration, in absence of significant postoperative glucose modifications. In a case report by Burns et al.,⁵⁶ CPM was determined by a rapid reduction of iperosmolarity in a patient with hyperosmolar hyperglycemia, without abnormal serum sodium. Rapid correction of hyponatremia and abrupt change of plasma osmolality might account for the development of CPM, especially in patients with liver disease in which abnormality of blood-brain barrier and decreased ability to generate new intracellular osmoles exist.³¹ Brain vascular endothelial cells shrinkage could determine entrance of circulating immune factors as cytokines, vasoactive amines, and other factors which can mediate oligodendroglial injury and demyelination.⁵⁷

The detailed analyses carried out by the Korean group of Lee et al.⁶ documented three other statistically significant differences between CPM/EPM patients and controls: lower pretransplant cholesterol serum levels, suggestive of severe liver impairment, higher MELD-Na⁺ scores, and variations of intraoperative potassium levels. In our cohort, neither MELD-Na⁺ nor changes of kaliemia were significantly different between the two groups.

Also, Abbasoglu et al.²⁸ reported in 1998 the absence of significant differences in pretransplant sodium and cholesterol serum levels and in the cause of liver disease between patients and transplanted controls; the same author noted a relevant rate of encephalopathy at the time of transplantation in the patients who developed myelinolysis. Our data supported this finding, but did not show a significant difference compared to controls.

Underlying liver disease was not associated with the risk of developing myelinolysis, in particular, alcoholic-related liver injury was not prevalent in the CPM/EPM patients included in the present study. This lack of association is consistent with other series of transplanted patients, whereas the association between alcoholic liver disease and pontinemyelinolysis in patients with chronic liver impairment was attributed to immunosuppression (6 to cyclosporine and 2 to tacrolimus).^13,29,58 Laureno and Karp⁵⁹ challenged such etiopathogenetic correlation, since calcineurin inhibitors can cause white matter lesions, similar to the reversible posterior leukoencephalopathy, easily confused with EPM. Central pontine and extrapontine myelinolysis patients included in the present study took cyclosporine more frequently than controls, but this difference was not significant; moreover, only one patient in the CPM/EPM group presented an overtherapeutic peak of the calcineurin inhibitor. Survival rate data are encouraging, and partially different than those reported by Yun et al.,⁴² who demonstrated high early mortality (40% within 3 months after transplantation).

Although this cohort is the largest described up to now, the article has some limitations because of the wide period of observation of the cases and the retrospective nature of the study. However, in this field, because of the rarity of the disease and the heterogeneity of the clinical approaches, prospective and multicenter studies will be unlikely performed.

In conclusion, liver transplanted patients affected by CPM may present also extrapontine lesions in a proportion of cases up to 63.6%, higher than currently reported in literature. The rapid increase of serum sodium concentration, also in the absence of preexisting hyponatremia, and osmotic imbalances are the main risk factors for myelinolysis, but also isolated hypernatremia may be responsible for the same damage. It is also important to emphasize that not only patients with pretransplant hyponatremia, but also all patients with preoperative, intraoperative, and perioperative serum osmotic alterations, have a risk to develop a demyelinating disease. It is therefore essential for anesthesiologists and physicians assisting the patient in the immediate postoperative period to avoid even small changes in serum sodium levels and any peak of hypernatremia.

MATERIALS AND METHODS

Clinical records of all adult recipients who underwent liver transplantation between November 1990 and June 2011 at the Padua University Liver Transplant Center were retrospectively analyzed. Living-donor liver transplants were not considered. All patients with development of neurologic symptoms after liver transplantation were investigated by clinical examination; patients with neurologic findings of suspicion for CPM/EPM were further investigated using magnetic resonance imaging scan within 24 to 72 hr, and diagnosis was made by the presence of typical radiologic features.⁶⁰ Patients were classified as suffering from CPM and EPM, depending on whether the brain lesions were limited to the pons and extended to the basal ganglia, thalamus, or cerebellum.

All transplanted patients affected by neurologic complications resulting from nondemyelinating disorders (hypoxic brain injury, infectious conditions after surgery) were excluded (Materials and Methods, SDC, https://links.lww.com/TP/B83).

Patients were further divided into two groups, regarding the presence of at least one hypernatremic peak (Na⁺>146 mEq/L) before the onset of neurologic symptoms or intraoperioperative change in serum sodium lower than 12 mEq/L/day (“hypernatremic”) or the intraperioperative correction of serum sodium level higher than 12 mEq/L/die (“hyponatremic”).

Statistical Analysis

Quantitative variables are presented as the mean±SD. The nonparametric Mann-Whitney U test was used for the comparison of averages between groups, whereas the Fisher exact test was used for the comparison of the frequency of qualitative variables. Correlations were performed using the Spearman test with description of correlation r. Statistical significance was defined as P less than 0.05 in all analyses. Statistical analyses were performed using SPSS software (Chicago IL) version 18.