Incidence, Clinical Features, and Risk Factors of Early Neurological Complications After Orthotopic Liver Transplantation: A Single-Center Experience

Introduction

Orthotopic liver transplantation (OLT) is the best therapeutic strategy for severe acute and chronic hepatic diseases from many etiologies – infectious, autoimmune, metabolic, toxic, and even neoplastic [1–3] – allowing better survival and quality of life in patients with end-stage liver disease [4]. Despite advancements in surgical technology and medical expertise over time, OLT can still lead to serious neurological complications (NCs) immediately after surgery or after many months or even years.

NCs occur in roughly 15% to 30% of patients undergoing OLT [5–9], much more frequently than in any other solid organ transplant with comparable technical difficulty (eg, heart, lung) [10,11], and are associated with higher mortality, higher risk of re-transplantation, and increased rates of rejection and infective complications [12–14]. Neurological disorders also result in disability and impairment of quality of life, as well as huge economic costs for healthcare systems [15,16].

NCs occurring in the transplantation setting can involve both central and peripheral nervous systems and can appear at any time after surgery. Many authors have proposed a temporal classification of NC after OLT according to the time of onset from transplant: acute (within 1 month), subacute (from 1 to 6 months), and chronic (after 6 months) [17,18].

Many studies have explored risk factors for OLT-related NCs with the aim of preoperatively identifying all patients with a higher risk profile. To date, known pre-transplant risk factors are older age [19], portosystemic encephalopathy (PSE) [20], higher model for end-stage liver disease (MELD) scores [19], previous alcohol abuse [21], malnutrition, kidney failure [22], and severe ascites [23]. Moreover, intra-transplant risk factors are sodium shift [24], coagulopathy, hypercholesterolemia [25], hypomagnesemia [26], and co-administration of tacrolimus (FK) with drugs that modify the immunosuppressant drug metabolism [27,28]. All this information can be used preoperatively to approximately assess the risk profile of an OLT candidate without an already standardized operative risk calculator. Better patient stratification before surgery may allow the undertaking of preventive measures to make an early diagnosis soon after the appearance of neurological impairment and to start an adequate therapeutic strategy as soon as possible to reduce long-term sequelae.

Since NCs after OLT have a strong impact on postoperative adverse outcomes, we assessed the incidence of NCs in our cohort of liver transplant patients to characterize the etiology of neurological symptoms/signs/syndromes, and to identify independently associated risk factors for their perioperative occurrence. Then, we analyzed all data and interpreted them from a neurological perspective to determine if the standard perioperative management could be improved.

Material and Methods

We performed a single-center retrospective study to determine the incidence of early postoperative NCs in our cohort of advanced liver failure patients undergoing liver transplant (primary outcome). Additionally, we clinically characterized the range of NCs according to etiology and assessed the preoperative, intraoperative, and postoperative risk factors for their occurrence (secondary outcomes). The study was approved by the local ethics committee (IRRB/27/21) and all enrolled patients signed an informed consent form for retrospective research data collection. To achieve all the above-mentioned aims, we reviewed the electronic medical charts from all adult patients who underwent OLT from June 2014 to October 2021. Inclusion criteria were: adult (≥18 years old) end-stage acute or chronic liver disease patients who underwent cadaveric OLT in the period between June 2014 to October 2021, and whose informed consent forms were available. Exclusion criteria were the absence of informed consent, age <18 years old and the absence of cadaveric OLT. No other exclusion criteria were applied. Patients were consecutively enrolled.

The process of data collection consisted of manual review of all medical documents in the electronic chart. We collected: a. Demographic data: height, weight, body mass index, sex, and age at OLT; b. Clinically relevant pre-transplant information: smoking habits; comorbidities, including pre-existing neurological comorbidities; use of central nervous system (CNS)-active drugs; liver disease etiology; severity and complications of liver disease including portosystemic encephalopathy (PSE) with at least 1 episode of PSE grade II according to the West Haven Criteria [29]; transjugular intrahepatic portosystemic shunt (TIPS) implantation; and already hospitalized at the time of OLT (eg, for acute liver failure or acute-on-chronic liver failure). We considered blood cholesterol level as a surrogate for nutritional status. As for liver disease etiology, if alcoholic, we considered the alcohol withdrawal time (more or less than 5 years before transplant), whereas if viral, we specified whether the infection had been eradicated; c. Intraoperative relevant information: split versus whole-organ transplant; plasma laboratory tests a few hours before OLT and immediately after surgery: sodium, magnesium, potassium, creatinine, lymphocyte count, and albumin; transplant technical details: time of cold and warm ischemia, type of graft preservation liquid, use of extracorporeal bypass, percentage of histologically documented graft macro- and micro-steatosis; d. Postoperative relevant clinical information: blood levels of sodium and magnesium on postoperative day (POD) 2 and sodium level shifts from POD 2 to the values assessed before and soon after surgery; occurrence of renal impairment including the need for continuous renal replacement therapy (CRRT) (defined according to KDIGO guidelines) [30]; infections; type of immunosuppressive induction drug, if administered; type of immunosuppressive maintenance regimen; graft rejection (diagnosed by biopsy and histological examination); pluri-complicated clinical course (more than 3 major systemic complications); administration of fluconazole; length of hospital stay; length of intensive care unit (ICU) stay; ICU re-admission; need for rehabilitation after hospital discharge; in-hospital death; and occurrence of neurological complications. By major systemic complications, we intend all relevant conditions leading to therapy changes or surgical procedures.

The follow-up period for each patient was from the time of transplantation to discharge or death. Information related to the post-discharge period was not collected, so we only focused on acute NCs occurring during surgical hospitalization. NCs were classified qualitatively, approximately following the Wu S-Y et al [9] and Dhar et al [12] models as follows: (1) encephalopathy (delirium, psychosis and/or consciousness impairment without neurological focal deficit and/or without structural changes on neuroimaging); (2) seizures/status epilepticus; (3) drug-related neurotoxicity: neurological symptoms/syndromes potentially related to calcineurin inhibitors (CNI) with improvement after drug reduction/suspension; (4) ischemic and hemorrhagic acute stroke; (5) central nervous system (CNS) infection; (6) osmotic demyelination syndrome (ODS); (7) cerebral edema; (8) other minor neurological manifestations; (9) peripheral nervous system disorders.

NCs were also classified according to a severity grading of fatal, major (if they caused a severe cognitive or motor disability), or minor (if they caused just a mild and/or transitory neurological deficit). All NC diagnoses were confirmed by a neurologist with extensive expertise in SOT-related NCs, neurotoxicity by immunosuppressive drugs, and end-stage organ diseases. Neurologists were consulted upon request by anesthesiologists during the ICU stay or by the transplant surgical team in step-down units. The neurology consultant, if clinically indicated, managed the neurological diagnostic work-up, including neuroimaging investigations.

Data were collected retrospectively. Continuous and categorical variables are expressed as median with interquartile range and as frequency with percentage, respectively. To compare continuous variables, the Wilcoxon test or t test was used when appropriate. Fisher’s exact test was used to compare categorical variables among groups. To explore the differences between patients with neurological complications (NCs) and patients without neurological complications (nNCs), logistic regression models were used. Levels of significance were set with P<0.05. Statistical analyses were performed using SAS software version 9.4.

Results

PATIENTS WITH NCS:

Ninety-seven out of 376 (25.8%) patients had at least 1 neurological manifestation and 24 (6.38%) had more than 1 neurological manifestation (eg, neurotoxicity and seizures, or neurotoxicity and critical illness polyneuropathy). Among all NCs, the most frequent was delirium/metabolic encephalopathy, affecting 14.1% (53/97) of patients, followed by seizures (3.7%) and clinical manifestations of neurotoxicity (3.7%), then ODS (2.9%) and peripheral nervous system injury (2.9%). Less frequently, we registered hemorrhagic stroke (1.3%), ischemic stroke (1.1%), cerebral edema (1.1%), CNS infections (0.3%) and other minor neurological manifestations (2.1%), such as headache, torticollis, or reactive depression. Most patients with NCs (73%) underwent instrumental diagnostic investigations: 32% had neuroimaging (brain computed tomography or magnetic resonance imaging) assessment, 14.4% had electroencephalography (EEG), and 26.8% received both exams. A structural brain abnormality was confirmed in 7.2% of patients in the whole population (27.8% in the NC group). Most patients (26 out of 40) who underwent EEG only had mild abnormalities, such as slow theta waves without epileptic significance; only 4 patients presented with major neurophysiological abnormalities such as epileptic discharges or status epilepticus. In 39 out of 97 (40%) patients, therapeutical changes were made, such as the introduction of antiseizures medication or the modifying of the immunosuppression regimen. In 20 (20.6%) patients, tacrolimus was stopped and in 4 (4.12%) patients the dosage was just reduced. This suggests that the administration of tacrolimus was withheld even when neurotoxicity was only suspected. In fact, neurotoxicity syndromes often do not have a very specific clinical picture (eg, tremors and consciousness impairment) but can lead to a severe neurological impairment if not immediately stopped. All patients with NC had an average blood concentration (measured at the time of the NC) of tacrolimus within the normal range. Most of the patients with NC (77/97) had minor sequelae or none, whereas 9 of them died because of neurological disease and 6 died because of non-neurological causes; finally, 11 patients had severe sequelae (Table 3).

PREOPERATIVE VARIABLES:

We compared the group of patients with neurological complications (NC) and the group of patients without neurological complications (nNC), looking for any demographic or clinically associated risk factors. Among preoperative variables, we found that the group of patients with NC had higher MELD scores and lower plasma concentrations of albumin (Table 1) as well as a higher probability of having a history of chronic renal impairment (OR 2.02), at least 1 episode of PSE (OR 3.81), and a higher probability of being in critical condition before transplant (ie, having acute liver failure or already being hospitalized) (Table 1). Interestingly, hepatitis C virus (HCV) cirrhosis etiology was correlated with a lower probability of developing NCs (Table 2). No preoperative neurological comorbidities (any type), including chronic cerebrovascular disease, considered separately, were found to be risk factors for developing acute NC early after OLT.

PERIOPERATIVE VARIABLES:

Patients with NC had higher plasma sodium concentrations at POD 2 (P=0.0291) and higher delta sodium between POD 2 and preoperative concentrations (P=0.0024) (Table 4). Moreover, patients with NCs had a higher probability of having postoperative acute renal impairment (P<0.00001, OR 3.74), the need for CRRT (P<0.0001, OR 7.1), postoperative infections (P<0.0001, OR 4.48), graft rejection (P=0.0006, OR 3.19) and more than 3 systemic complications (P<0.0001, OR 4.39) (Table 4).

OUTCOMES:

Compared to nNC patients, patients with NCs had longer in-hospital length of stay (P<0.0001), longer ICU stay (P<0.0001), higher probability of ICU re-admission (P<0.0001, OR 6.65), higher probability of in-hospital death (P<0.0001, OR 7.67), and higher probability of needing rehabilitation after hospital discharge (P<0.0001, OR 8.7) (Table 5).

Discussion

In our cohort of OLT recipients, we registered acute NCs soon after transplantation in almost 26%, a frequency exactly reproducing the findings reported in the literature [6–9,33]. The considerable number of studied phenomena in our cohort may even be underestimated due to the narrow observation time window, or it is possible that not all neurological findings were documented in the patient charts. The higher frequency of NCs seen in the liver transplantation setting, more than any other technically difficult solid organ transplant (eg, heart or lungs), has been commented on in the literature as being hypothetically due to brain damage caused by liver dysfunction. Brain injury related to liver function impairment could act as a predisposing factor for acute perioperative neurological diseases [16,31]. Consistently, our data strongly confirm that the main risk factor for post-transplantation NCs is the severity of the baseline liver disease with its complications, such as PSE, but renal impairment can also induce neurological dysfunction. Extra-cranial peripheral organ dysfunctions with a well-documented associated neurological impairment [32] can act as a predisposing factor for further perioperative NCs, but primary neurological diseases, including common cerebrovascular diseases, as comorbidities are not risk factors for postoperative NCs. We believe this result is important for neurologists consulted for their opinion on liver transplantation eligibility for end-stage liver disease patients with neurological comorbidities. Whether neurological comorbidities do not influence the neurological outcomes after liver transplantation needs to be confirmed in larger multicenter prospective studies, and a more detailed clarification of the type of neurological comorbidities (eg, neurodegenerative vs non-neurodegenerative disorders) is required. Although not standardized among centers, neurological consultation during the work-up for OLT eligibility is mostly considered for liver disease patients with a known neurological comorbidity or older cirrhotic patients; however, we believe that future efforts should focus on establishing a transplantation risk score calculator, including all statistically confirmed risk factors. Our study, consistent with the literature, acknowledges the severity of liver disease and other extra-cranial/peripheral organ dysfunctions potentially affecting the brain, such as renal impairment, as risk factors for early NCs after OLT. This strategy would allow for better preoperative stratification of liver transplant candidates, considering their risk of negative neurological outcomes, and for close observation of high-risk profile patients for postoperative neurological impairment to allow early recognition with fewer high-impact sequelae. In extreme cases, a cirrhotic patient’s risk profile could argue against transplantation if the probability of highly disabling NCs is prohibitive. We are aware that this objective requires a prospective study in a larger population to carefully establish the role of neurological comorbidities, including the more frequent cerebral small-vessel disease or neurodegenerative diseases.

Finally, when NCs occur after a liver transplant, there are many adverse postoperative-associated outcomes: renal impairment, infections, rejection, and pluri-complicated courses. NCs also impact the length of hospital and ICU stays, ICU re-admission, in-hospital death, and patient discharge to rehabilitation facilities. All these adverse postoperative outcomes indicate the enormous size of the problem in terms of healthcare costs and patients’ and families’ quality of life.

A strength of our study is that all neurological diagnoses were confirmed by a neurology consultant well-integrated in the multidisciplinary transplantation team, and the main limitation of this investigation is its retrospective and single-center design.

Conclusions

NCs after OLT are common and are often severe, with a strong impact on survival, performance status, and long-term quality of life. NCs can nullify the health gain derived from the organ transplantation and, unfortunately, lead to a loss of functional autonomy. Our study, despite the limitation connected to its retrospective nature, shows the importance of identifying risk factors to focus on prevention of NCs, as treatment of brain injuries is often ineffective. Our results suggest the importance of extra-cranial conditions, such as the severity of pre-transplant liver disease or renal impairment, instead of neurological comorbidities, as potential risk factors for developing NCs after OLT.