Pathogenic factors of cognitive dysfunction after liver transplantation: an observational study

Objectives Neurocognitive complications significantly reduce long-term health-related quality of life in patients undergoing liver transplantation; however, few studies have focused on their perioperative cognitive status. The authors designed a prospective observational study to determine the incidence and risk factors of posttransplant cognitive dysfunction. Methods This study included patients with end-stage liver disease who were on the liver transplantation waiting list. We performed an investigation with a neuropsychological battery before and 1 week after the successful transplant, analyzed the changes, and further explored the complicated perioperative factors that contribute to cognitive dysfunction. Results A total of 132 patients completed all the investigations. Compared with healthy controls and preoperative cognitive performance, 54 patients experienced deterioration, 50 patients remained unchanged, and 28 patients showed rapid improvement. Logistic regression analysis showed that age [odds ratio (OR) = 1.15, 95% confidence interval (CI, 1.07–1.22), P < 0.001], the model for end-stage liver disease (MELD) score [OR = 1.07, 95% CI (1.03–1.13), P = 0.038], systemic circulation pressure [OR = 0.95, 95% CI (0.91–0.99), P = 0.026] within the first 30 min after portal vein opening, and total bilirubin concentration [OR = 1.02, 95% CI (1.01–1.03), P = 0.036] on the seventh day post-transplant were closely related to the deterioration of cognitive function. Conclusion The incidences of deterioration, maintenance, and improvement in cognitive function were 40.9%, 37.9%, and 21.2%, respectively. Increasing age, higher MELD score, lower perfusion pressure in the early stage of the new liver, and higher total bilirubin concentration postoperatively may be independent pathogenic factors.


Introduction
Liver transplantation is the only curative treatment for end-stage liver disease due to hepatocellular carcinoma, decompensated cirrhosis, acute or chronic liver failure, and congenital biliary atresia. The number of transplants and survivors has increased significantly with the advances in immunosuppressive regimens and donation awareness. By 2019, US cadaver donors had graft survival rates of 92.1% at 1 year, 85.3% at 3 years, 79.3% at 5 years, and 59.4% at 10 years, which is higher in living donors, and the recipients' survival rates are comparable [1]. As the negative effects are eliminated after transplantation, their overall cognition and related quality of life 1 year and beyond will improve significantly [2,3]; however, whether they can fully return to normal remains controversial. Including but not limited to immediate memory, attention, language, and delayed memory may still be worse than those of healthy controls, even at the eighteenth month [4]. In addition, Sotil et al. found new-onset cognitive decline 1 year after surgery after ruling out the effects of hepatic encephalopathy and alcohol abuse [5]. Therefore, we hypothesize that this phenomenon could be related to some adverse perioperative neurological events, with the residual symptom lasting up to 3 months or years [6]. First, some pre-existing structural or functional brain lesions cannot be removed and are aggravated with the occurrence of adverse events. These lesions may result from recurrent hepatic encephalopathy [7,8], alcohol or drug abuse [9,10], normal ageing, repeated blood transfusions [11], chronic pain [12], anxiety and depression due to prolonged physical discomfort, and financial distress [13]. Second, even if a donor is secured, one will have to experience major physical trauma, various sedative and analgesic drugs, reperfusion syndrome after hepatic vein opening [14], and the neurotoxicity of immunosuppressive drugs [15], complex infections, and long-term rehabilitation. Compared to the recovery of physical function, the clinical outcome of neurocognitive impairment and its effect on the long-term quality of life in recipients is difficult to clarify, including compliance to immunosuppressive treatment, resuming work, and social relationships. Therefore, we designed this prospective observational study to determine the incidence and risk factors of postoperative cognitive dysfunction (POCD), which will only be identified if there is sufficient evidence that it leads to a sustained and significant deterioration in the quality of life.

Methods
The Ethical Committee of the Second Xiangya Hospital reviewed our clinical protocol, and all patients provided written informed consent. The donor organs included in our study were all from brain-dead donors, and all procedures have undergone rigorous ethical scrutiny. Eligible recipients were patients with end-stage liver disease on the liver transplantation waiting list. This protocol excluded patients with the following conditions: unwillingness to comply with our protocol; less than 6 years of education; language, hearing, and visual impairment or who for any reason cannot complete the investigation; a history of mental or neurological disorders, taking tranquillizers or antidepressants; a history of general anaesthesia in the past 3 months; overt hepatic encephalopathy (more than minimal hepatic encephalopathy and grade I hepatic encephalopathy according to the west haven criteria and the competing International Society of Hepatic Encephalopathy and Nitrogen Metabolism criteria); persistent hypotension or hypoxaemia (SaO 2 < 90%) at any stage; diagnosed complications that may affect cognitive performance within the first week after liver transplantation; a veno-venous bypass applied during the operation. To eliminate the practical effect of the battery in shortterm intervals, we selected 48 healthy volunteers as the control group after statistical calculations. There was no significant difference in the basic characteristics between the two groups, and all controls underwent neuropsychological testing at the same time interval.
According to the protocol, we collected the following variables of all enrolled patients before transplantation: demographic status, aetiology, model for end-stage liver disease (MELD) scores, history of overt hepatic encephalopathy, and comorbidities, all of which came from their electronic medical records. In the anaesthetic management of liver transplantation, we placed no restrictions on anaesthetics, vasoactive agents, surgical approach, blood transfusions, and postoperative analgesia, these selections are entirely up to the anaesthetist or surgeon depending on the actual situation. The following intraoperative information was also collected: donor age, cold ischaemia time, duration of anaesthesia and anhepatic phase, surgical procedure, blood transfusion volume, and mean arterial pressure (MAP) during the anhepatic and reperfusion phases.
Tthe patients were transferred to post-transplant ICU with endotracheal intubation after surgery. One specialized researcher visited regularly to review their basic conditions and evaluate whether they met any exclusion criteria. Postoperative delirium was determined using confusion assessment method-ICU according to the Diagnostic and Statistical Manual of Mental Disorders IV. Post-transplant information included ICU care duration, total bilirubin, albumin, procalcitonin, and C-reaction protein on the seventh day, and the occurrence of delirium. It is worth mentioning that basiliximab (20 mg) was administered 2 h preoperatively and 4 days postoperatively. Methylprednisolone 500 mg was administered during the anhepatic phase. Tacrolimus and mycophenolate mofetil were used for anti-rejection 3 days postoperatively. The sociodemographic, intraoperative, and postoperative information of the patients is listed in Tables 1-3.
We selected the Chinese version of the Montreal Cognitive Assessment (MoCA), a brief and commonly used global screening tool. It has higher sensitivity and specificity in patients with mild neurocognitive impairment and covers several domains [16,17], including visuospatial, multiple aspects of executive functions, naming, memory, attention, language, abstraction, delayed recall, and orientation. All investigators received professional training and could use the battery proficiently and normatively. After an initial review to rule out overt hepatic encephalopathy, neuropsychological tests were performed in a quiet room with only the patient and investigator present. We performed a second cognitive assessment on schedule if the patients did not meet the exclusion criteria. At this time, they were discharged from the ICU and free from exposure to systemic sedative analgesics, and the tracheal catheter was removed. We evaluated postoperative pain and sleep quality using the visual analogue scale and Pittsburgh Sleep Quality Index.

Statistical analysis
We used the same definition as the International StudyPOCD1 study to determine the presence of POCD [18,19]. To quantify the practice effect, we compared the change in performance of control subjects in three age groups (18-39 years, 40-59 years, and ≥60 years) between baseline and subsequent tests. For the patient group, we compared the first scores with the second test results, subtracted the average practice effect from these changes, and divided the result by the SD of the control group to obtain a z-score for the individual test outcomes. Large positive z-scores indicated deterioration from baseline. We classified patients as exhibiting POCD if their z-scores were ≥1.0, z-scores less than −1.0 belonged to the improvement group, and all others belonged to the maintenance group. Patients in the maintenance and improvement groups were assigned to the non-POCD group. Continuous variables were presented as mean ± SD, and categorical variables were presented

Results
From April 2019 to July 2020, 159 patients were enrolled in this study and 27 patients withdrew; the detailed reasons for withdrawing are listed in Fig. 1.
The detailed distribution of the MoCA scores and comparisons between the groups are shown in Fig. 2. Among 132 patients who completed all preoperative and postoperative cognitive function investigations, 54 (40.9%) showed deterioration, 50 (37.9%) maintained preoperative level, and 28 (21.2%) showed significant improvement. Logistic regression analysis showed that age, MELD score, MAP during the first 30 min after declamping the infrahepatic vena cava, and total bilirubin concentration on the seventh day after transplantation are closely related to the deterioration. Table 4 presents the results of the logistic regression analysis.
The area under the receiver operating characteristic curve was 0.849, and sensitivity and specificity were 64.81% and 84.62%, respectively.

Discussion
In defining and accurately assessing POCD, huge differences in methodology exist, such as the appropriate test battery, time interval, and how to do statistical analysis, etc. The most authoritative method is to calculate the learning effect in healthy controls and eliminate it in subjects, obtaining the Z value changes by more than one SD after calculation using multiple formulae [20]. The incidence reported in previous studies varied widely due to patient status, disease, and medical procedures. ISPOCD1 and ISPOCD2 showed that the incidence of POCD in elderly and middle-aged patients which undergone non-cardiac surgery was 25.8% and 19.2%, respectively [18,19]. In a review that analyzed all the relevant literature on liver transplant recipients, the incidence fluctuated from 0 to 50% and was inconclusive [21]. In this study, it was 41.4%, which is in line with previous data and our expectations.
Our research reached conclusions similar to most previous studies in that cognitive function gradually deteriorated during the ageing procedure [18,22]. Cholinergic neurons regulate normal memory function in the forebrain of the basal ganglia, but choline reserve gradually decreases with ageing [23,24]. Anaesthetics also affect the release of neurotransmitters (such as acetylcholine, dopamine, and norepinephrine) [25], which is prevalent and more severe in the elderly [26]. Expert consensus also recommends using anticholinergic drugs carefully to avoid adverse cognitive effects [27].
The MELD score is a mathematical model based on serum creatinine, bilirubin, and International Normalized Ratio. It has been used as an objective scale to evaluate the severity of end-stage liver disease and 3-month waiting-list mortality and acts as a general index for the distribution of liver donors worldwide [28]. Higher MELD scores indicate worse kidney function, longer intubation, longer ICU care duration, and more blood transfusion, all of which may contribute to impaired cognitive function [29].
In patients with end-stage liver disease, especially those with overt hepatic encephalopathy, the autoregulation of cerebral blood flow is impaired and can be quickly restored after successful liver transplantation [30,31]. As a result, intraoperative cerebral perfusion is particularly dependent on the MAP and PaCO 2 . Non-invasive transcranial Doppler ultrasound revealed that the cerebral blood flow increased to varying degrees after opening the infrahepatic inferior vena cava, and some even exceeded 100% compared with the pre-hepatic and anhepatic phases [31,32]. This may be due to the effect of acidic metabolites from the lower extremities, because neither central venous pressure, pulmonary capillary wedge pressure, nor cardiac output was greater after the anhepatic phase. The cerebral metabolic rate of oxygen in patients with encephalopathy increased significantly from the preanhepatic value compared to those who had never been encephalopathic. Therefore, factors known to decrease cerebral blood flow (e.g. hypotension or hypocapnia) and increase cerebral oxygen metabolism should be avoided. Higher systemic pressure, adequate anaesthesia depth, appropriate hypercapnia, and hypothermia may be beneficial.  Higher bilirubin concentrations after transplantation are detrimental to neurocognitive performance, which is also consistent with actual clinical practice because it has neurotoxic and encephalopathic effects [33,34]. Although none of the patients experienced acute or subacute rejection and major hepatobiliary complications, there were still some patients with higher serum bilirubin concentrations than others. One of the reasons is more bilirubin accumulation before surgery, more allogeneic blood transfusions, kidney insufficiency, or minor biliary tract congestion [35]. The second reason is poor donor or organ protection procedures [36] and higher donor age [37], which is one of the key factors affecting the graft survival rate. Whatever the reason, either of these has adverse effects on the fragile brain [38].
Our clinical study has some limitations. First, the sample size was small. A single-centre experience and certain biases could not be completely avoided, so large-scale multicentre studies are needed to confirm our findings. Second, considering the clinical feasibility and acceptability, we  used the Chinese version of the MoCA, which is different from previous studies (such as the ISPOCD1 study) that used a series of neuropsychological batteries; however, these may not be suitable for postoperative patients with extreme weakness and physical fatigue. Finally, we have not included some important factors that should be paid attention to, such as preoperative anxiety and depression state, donor warm ischaemia duration, anaesthesia depth, cranial imaging materials, etc. These factors may lead to bias and we hope to optimize our study in the future.

Conclusion
Compared with the preoperative results, most patients had deteriorated or remained unchanged within 1 week after successful liver transplantation, and only a few patients improved rapidly. Increasing age, higher MELD score, hypotension during the reperfusion phase, and higher total bilirubin concentration on the seventh day after liver transplantation may be closely related to impaired cognition.