Exploring early core symptoms in post-hepatectomy patients: a longitudinal study

Haoran Zhang, Shaoxue Li, Mengmeng Yuan, Jingya Liu, Hua Fan, Shuwen LiResearch Square 2023 · DOI

TL;DR

This longitudinal study uses network analysis to identify "core" symptoms—those most strongly connected to others—in patients recovering from liver cancer surgery. By tracking 249 patients, researchers found that while overall symptom severity decreases before discharge, the core drivers of distress shift from localized surgical trauma (like dry mouth and nausea) to systemic recovery challenges (like shortness of breath and disturbed sleep).

What Problem Does This Paper Address?

Primary liver cancer (PLC) is a global health burden, particularly in China, where chronic Hepatitis B is prevalent. While hepatectomy (surgical removal of part of the liver) is a vital radical treatment, the postoperative period is fraught with a complex "symptom cluster."

Traditionally, clinicians focus on the most severe symptoms reported by patients. However, the authors argue that severity doesn't always equal importance. Some symptoms act as "hubs" or "core symptoms" that trigger or exacerbate others. If a clinician can identify and treat these central nodes, they might achieve a "domino effect" that improves the patient's overall recovery more efficiently than treating symptoms in isolation.

Methods

The researchers conducted a longitudinal study following 249 PLC patients at two hospitals in Eastern China. They assessed symptoms at two critical time points:

  • T1: 1–2 days after surgery (peak acute recovery).
  • T2: 1–2 days before discharge (transition to home care).

Tools used: * MD Anderson Symptom Inventory (MDASI): A standard 13-item tool for cancer symptoms. * TSM-PLC: A specific 6-item module for liver cancer (covering abdominal distension, jaundice, etc.).

The "Secret Sauce": Network Analysis Instead of just averaging scores, the team used Network Analysis (specifically the EBIC glasso algorithm). In this model, symptoms are "nodes" and the correlations between them are "edges." They calculated Centrality Indices (Strength and Expected Influence) to determine which symptoms were the most influential "hubs" in the network.

Datasets

The study utilized a primary clinical dataset of 249 participants (88.6% of those eligible) recruited between March 2021 and September 2022.

  • Demographics: Mostly male (73.3%), aged 45–70, with a high prevalence of Hepatitis B (67.5%).
  • Instruments: The MDASI and TSM-PLC provided the 19 symptom variables used to construct the networks.

Key Insights and Ideas

  • Severity $\neq$ Centrality: A major "aha moment" was confirming that the most severe symptom isn't always the most central. For example, while "dry mouth" was both severe and central at T1, "shortness of breath" became a core symptom at T2 despite having a relatively low severity score.
  • The Shift from Local to Systemic: At T1, core symptoms (dry mouth, nausea, abdominal distension) were largely linked to the physical trauma of surgery and anesthesia. By T2, the core symptoms (disturbed sleep, shortness of breath) reflected systemic dysfunction and the challenges of organismal recovery.
  • Network Stability: Interestingly, the overall structure of the symptom network (how tightly symptoms are interconnected) remained statistically similar between T1 and T2, even though the specific "core" symptoms within that structure changed.

Key Results

  • T1 Core Symptoms: Dry mouth (highest centrality), pain, abdominal distension, lack of appetite, and nausea.
  • T2 Core Symptoms: Lack of appetite (ascended to top priority), pain, shortness of breath, and disturbed sleep.
  • Symptom Reduction: Overall severity scores dropped significantly across the board from T1 to T2. For instance, dry mouth dropped from a mean of 6.25 to 1.40.
  • Centrality Stability: The "Bridge" between symptoms was stable ($p = 0.14$ for network invariance), suggesting that while individual symptoms fade, the way they influence each other remains consistent during the hospital stay.

Limitations and Future Work

  • Correlation vs. Causation: Network analysis shows associations, not necessarily that one symptom causes another.
  • Sample Size: While 249 is a solid clinical cohort, network models often benefit from even larger datasets to ensure the stability of "edges."
  • Psychological Gap: The study focused heavily on physical symptoms; the authors note that including more psychological distress measures might provide a fuller picture of the patient experience.
  • Next Steps: The authors suggest that future research should move from observation to intervention—testing whether targeted treatment of these "core" symptoms actually leads to faster discharge or better long-term quality of life.

Summary updated

Abstract

<title>Abstract</title> <bold>Background: </bold>Patients with primary liver cancer (PLC) experience a range of symptoms in the early postoperative period. Symptoms include cancer-related symptoms and adverse effects of treatment. Exploring the core symptoms and their dynamics in the early post-hepatectomy patientsmay help provide better symptom management programs. <bold>Objective: </bold>The purpose of this study was to identify the core symptoms in early post-hepatectomy patients and to explore the trajectory of their dynamics. <bold>Methods: </bold>249 participants with PLC underwent curative surgical procedures between March 2021 and September 2022 at the Hepatobiliary Surgery Departments of two Eastern Chinese hospitals. A comprehensive symptom assessment was administered to the patients 1-2 days after surgery (T1) and 1-2 days before discharge (T2). Network analysis was used to identify core symptoms in early post-hepatectomy patients based on symptom severity. <bold>Results: </bold>The network analysis identified the core symptoms of dry mouth, pain, abdominal distension, lack of appetite, and nausea at T1. In contrast, at T2, lack of appetite, pain, shortness of breath, and disturbed sleep were revealed. At the two early time points, while the results of the network structure test showed no statistically significant difference in the overall symptom network structure, there was a significant difference in the results for the core symptoms. <bold>Conclusion: </bold>Our study demonstrates the need to identify core symptoms in the early in early post-hepatectomy to improve symptom management in PLC patients. It is crucial to apply network analysis as an essential component of cancer care.
BibTeX
@inproceedings{zhang2023exploring,
  author = {Haoran Zhang and Shaoxue Li and Mengmeng Yuan and Jingya Liu and Hua Fan and Shuwen Li},
  title = {{Exploring early core symptoms in post-hepatectomy patients: a longitudinal study}},
  booktitle = {Research Square},
  year = {2023},
  doi = {10.21203/rs.3.rs-3672143/v1},
}
Figure