Letter to the Editor: Hemodynamic role in hepatic encephalopathy pathogenesis following transjugular intrahepatic portosystemic shunt placement

Abstract

Hepatic encephalopathy (HE) is a common complication following transjugular intrahepatic portosystemic shunt (TIPS) placement and is thought to result from reduced hepatic detoxification of neurotoxins. However, Zhuang et al in the World Journal of Radiology demonstrate that this understanding may be incomplete through exploration of the pathophysiology from a hemodynamic perspective. They found significant changes in cerebral blood flow (CBF) and spontaneous brain activity following TIPS. Initially, CBF increased in specific areas within the first month, followed by a return to baseline at three months. These CBF changes coincided with increased spontaneous brain activity that persisted in some areas, suggesting compensation for functional loss in other regions. Together with the fact that serum ammonia remained unchanged in this cohort and prior work showing that, after three months, HE risk returns to pre-TIPS levels, these results suggest that HE may develop in individuals susceptible to these hemodynamic and adaptive changes, warranting further research.

Key Words: Arterial spin labeling; Cerebral blood flow; Hemodynamics; Hepatic encephalopathy; Digit symbol substitution test; Resting-state functional magnetic resonance imaging; Transjugular intrahepatic portosystemic shunt

Core Tip: The pathogenesis of hepatic encephalopathy (HE) following transjugular intrahepatic portosystemic shunt (TIPS) placement remains incompletely understood. While hyperammonemia is classically implicated as the driver of HE development, Zhuang et al also suggest that postoperative hemodynamic changes play a key role. Using three-dimensional arterial spin labeling and resting-state functional magnetic resonance imaging, the authors demonstrate brain-region-specific alterations in cerebral blood flow and resting spontaneous brain activity, as measured by amplitude of low-frequency fluctuation, following TIPS placement. Their findings challenge the conventional understanding of post-TIPS placement HE and highlight the diagnostic potential of resting-state functional magnetic resonance imaging for characterizing HE during the early postoperative period.

TO THE EDITOR

It is with great interest that we read the high-quality article by Zhuang et al[1] entitled “Role of cerebral blood flow changes in post-transjugular intrahepatic portosystemic shunts hepatic encephalopathy”, recently accepted for publication in the World Journal of Radiology. In this prospective study of 18 patients with liver cirrhosis and portal hypertension undergoing non-emergent transjugular intrahepatic portosystemic shunting (TIPS), alongside 18 sex- and age-matched healthy controls, the authors present a fascinating analysis of the hemodynamic effects of TIPS on brain function by region. Their use of three-dimensional arterial spin labeling (3D-ASL) and resting-state functional magnetic resonance imaging (rs-fMRI) represents a promising approach to assess changes in cerebral blood flow (CBF) and amplitude of low-frequency fluctuation (ALFF) underlying the pathogenesis of post-TIPS hepatic encephalopathy (HE).

While previous studies have elucidated a pathophysiological mechanism involving the systemic delivery of neurotoxins such as ammonia and vasodilators, Zhuang et al[1] propose that these factors alone cannot fully explain the development of post-TIPS HE. Instead, they argue that the synergistic effects of neurotoxin overload and postoperative hemodynamic changes, including brain-region-specific CBF changes, constitute a more viable mechanism, especially during the first 3 months postoperatively. Their findings challenge the conventional understanding of post-TIPS HE pathogenesis and highlight rs-fMRI as a potentially diagnostically sensitive tool for HE.

The primary aim of this article is to outline the key contributions of the study and contextualize them within current radiological and clinical practice, particularly as they relate to TIPS associated HE. We commend the authors for their novel approach and seek to both expand their findings and provide a critical perspective.

BLOOD FLOW AND SPONTANEOUS ACTIVITY SUPPORT A ROLE OF HEMODYNAMICS IN HE AFTER TIPS

In the study by Zhuang et al[1], patients experienced significant brain-region-specific alterations in CBF and spontaneous brain activity after TIPS placement, compared with the pre-TIPS baseline. Specifically, the authors note a transient increase in CBF in the left fusiform gyrus (L-FG) at one month postoperatively, an area that contains the Visual Word Form Area and is critical for high spatial frequency detail discrimination in visual stimuli used in reading, being involved in the indirect route within the dual-route model of language processing[2-6]. It has been shown that damage in this region produces pure alexia, the isolated inability to read, and impairments in object naming[2].

At the same time point at which CBF was increased in the L-FG, there was increased ALFF in the left superior temporal gyrus (L-STG). The L-STG is also involved in indirect route as well as phonological analysis involved in understanding both spoken and written language[7]. These functions of the L-STG are supported by functional neuroimaging studies showing increased activity in word perception and naming compared to object perception and naming[3]. As the authors note, an increase in ALFF from baseline may be interpreted as heightened intrinsic neural activity, which, in pathological contexts, potentially reflects the compensatory re-localization of function lost in other brain areas[1,8]. Additionally, at three months post-operatively, CBF across all regions and ALFF in the L-STG return to pre-TIPS placement levels, suggesting this functional alteration is transient and supporting an association between the L-FG and L-STG. Taken together, the significant overlap in the roles of the L-FG and L-STG, along with increased CBF in one being associated with increased ALFF in the other at the same time point, supports the authors’ proposed mechanism for HE etiology. Namely, increased perfusion to the L-FG delivers neurotoxins shunted away from hepatic clearance by the TIPS ultimately to the parenchyma of that region, potentially causing damage. While this is a compelling mechanism, there is insufficient evidence to claim with a high degree of certainty that it is accurate; thus, the functional neuroanatomy correlates we present should be viewed as circumstantial evidence in support of the proposed mechanism.

The authors also have findings that do not neatly fit the proposed model. For instance, the left orbitofrontal superior gyrus and left precuneus showed increased ALFF at three months compared to pre-TIPS placement without a corresponding increase in CBF in regions that were closely related over the same period[1]. Furthermore, the study by Zhuang et al[1] is limited by a small sample size of 18, with only eight of those subjects completing the three-month follow-up, an unbalanced male-to-female ratio (5:1), and its single-center design. Despite these limitations, the article presents compelling evidence for a new understanding of the etiology and pathophysiology of HE following TIPS placement, which merits further investigation.

STUDY DESIGN STRENGTHS

Zhuang et al[1] made several key methodological choices that strengthen their argument. Both rs-fMRI and 3D-ASL have previously been used separately to evaluate brain activity after TIPS. To our knowledge, however, this is the first study to combine both rs-fMRI and 3D-ASL into the same study to analyze HE post-TIPS placement[9,10]. Of note, while HE is a clinical diagnosis primarily diagnosed via the West Haven Criteria[11], the authors did incorporate Child-Pugh score, various cognitive tests, and the aforementioned neural/perfusion tests into their investigation. Additionally, prior studies have used alternative perfusion measurement modalities, namely [¹⁵O]-water positron emission tomography scans, which lack the same functional/spatial resolution as the modalities used in this study. PET remains the reference standard for quantitative CBF measurement, but previous studies have established the efficacy and strong correlation between PET and 3D-ASL in the assessment of CBF[12]. The authors were able to effectively leverage their multimodal study design to show through complementary methods that hyperammonemia alone may not fully explain HE development. This mechanism is further supported by previous research that has shown similar transient cerebral hyperperfusion following TIPS placement[9]. In sum, the claim by Zhuang et al[1] that increased perfusion post-TIPS placement alongside hyperammonemia drives HE after TIPS is novel and supported in its individual components by both previous research and their own multimodal findings.

In addition to the strengths of using a multimodal study design, the time window for brain analysis was also ideal, as HE has the highest risk of developing within three months post-TIPS placement[13]. While some studies have investigated CBF for up to three months after TIPS, several studies on CBF pre- and post-TIPS placement only measured CBF up until one month postoperatively, missing the latter two months of the higher risk window for HE development following TIPS placement[14]. Therefore, conducting brain imaging at one month and three months captured the primary risk window and was an informed decision. Overall, the argument that altered cerebral hemodynamics and increased delivery of neurotoxins to the brain synergistically predispose the brain to HE is strongly argued by the authors.

ANALYSIS OF STATISTICAL METHODOLOGY AND SAMPLE SIZE

While Zhuang et al’s article[1] provides a compelling insight into the etiology and pathophysiology of post-TIPS HE, there are a few but noticeable methodological considerations that limit the generalizability of this study and warrant attention. From a statistical point of view, while a Gaussian random field correction was applied appropriately to deflate the chance of type I error at the cluster level to control for comparing multiple voxel locations (spatial multiplicity), the study authors did not control for the family-wise error rate (FWER) caused by repeated-measure comparisons across multiple timepoints (temporal multiplicity). FWER is the cumulative probability of type I error after multiple pairwise tests, which rises as the number of pairwise tests increases due to spatial or temporal multiplicity. In this study, the consequence of a type I error with respect to temporal comparisons would be, for example, stating that there is a significant change in CBF between two time points when there is no difference. An argument could be made that adjusting for the FWER, such as with a Dunn or Holm correction, may obscure a detectable effect by inflating the type II error; however, it is proper statistical practice to follow pairwise testing with post hoc corrections.

The study initially included 18 patients, the sample size decreased to 12 at 1 month and 8 at 3 months. With such severe attrition and even more limited sample size, the reliability and reproducibility of the voxel-wise testing, longitudinal correlation/inference, and correlation analyses are called into question. Due to these limitations, future research using larger sample sizes is needed to support the validity of the authors' findings. Finally, it is important to note that only 3 of the 18 patients were diagnosed with HE by the end of the study, meaning that the majority of patients did not develop the outcome the study sought to explain. What Zhuang et al[1] did manage to show, however, were significant changes in brain perfusion and similar findings in the entire patient cohort. Still, the authors did not compare brain changes in HE patients against those who did not develop HE. Without a concrete comparison, which the authors understandably could not reasonably perform with such a limited sample size, such claims regarding the synergy between increased CBF and hyperammonemia being the main driver behind specifically HE would be overstating the study findings. In our view, the authors did provide a reasonable mechanism behind the development of HE, but they do not have sufficient statistical evidence to support this finding with this study alone. Then, the study distills to several noteworthy and interesting cerebral perfusion findings in TIPS patients with questionable generalizability to HE patients specifically.

CONCLUSION

Zhuang et al[1] present evidence that provides immense value in understanding the pathophysiology contributing to post-TIPS placement HE. Of primary significance, they illustrate regional alterations in CBF and ALFF utilizing 3D-ASL and rs-fMRI, independently of hyperammonemia, and correlate these changes with neuropsychological test performance and liver function parameters. These early results align with previous studies highlighting cerebral hyperperfusion following TIPS placement. Despite these contributions by Zhuang et al[1], limitations exist in the article. Of note is the attrition, the low number of three patients who developed HE in the study, and the lack of adjusting for temporal multiplicity caused FWER, such as with Dunn or Holm correction. Future studies with larger sample sizes, adjusted statistical methods, and prolonged follow-up periods may improve the generalizability of these results and provide further insight into the multifactorial mechanism underlying post-TIPS HE. We appreciate the information brought forth by Zhuang et al[1] in advancing the understanding of HE in the postoperative period and the multidisciplinary dialogue they invite.