Prognostic significance of liver stiffness in patients with primary biliary cholangitis: validation of Baveno VII criteria

Background The role of liver stiffness measurements (LSM) in patients with primary biliary cholangitis (PBC) remains to be further elucidated. Aims To clarify the prognostic role of LSM and to validate the “novel concepts” proposed by the Baveno VII Working Group. Methods An analysis of the prognostic significance of LSM was performed involving 672 patients. Results LSM and ΔLSM/ΔT were independent risk factors for liver decompensation, liver transplantation, or liver-related death (primary outcomes, p < 0.001, both). A rule of 5 kPa for LSM (10–15–20 kPa) could be used to denote progressively higher relative risks of primary outcomes. Patients with LSM < 10 kPa have a negligible 3-year risk of primary outcomes (< 1%). Cut-off values of 10 and 15 kPa can be used to classify PBC patients into low-, medium-, and high-risk groups. A clinically significant decrease in LSM, evaluated at 6, 12, or 24 months elastography tests, was associated with a substantially reduced risk of primary outcomes (p < 0.05, all), which can be defined as a decrease in LSM of >  − 20% associated with LSM < 20 kPa or any decrease to LSM < 10 kPa. A clinically significant increase in LSM, evaluated at 6, 12, or 24 months elastography tests, was associated with a substantially raised risk of primary outcomes (p < 0.05, all), which can be defined as an increase in LSM of ≥  + 20% or any increase to LSM ≥ 15 kPa. Conclusions LSM can be used to monitor disease progression and predict long-term prognosis in patients with PBC. Supplementary Information The online version contains supplementary material available at 10.1007/s12072-023-10587-w.


Introduction
Primary biliary cholangitis (PBC) is a slowly progressive disease with a natural course of 10-15 years, but high-risk patients can progress rapidly to decompensated cirrhosis and even death [1].Early identification of such patients and timely initiation of drug therapy may improve long-term prognosis.
Liver stiffness measurements (LSM) by transient elastography technique, have been shown in several studies to be of prognostic value in patients with PBC.Corpechot et al. [11] found that PBC patients had a fivefold increased risk of liver failure, liver transplantation or death when LSM were > 9.6 kPa.In addition, worsening liver stiffness was associated with an increased risk of adverse outcomes.When the cut-off value of 2.1 kPa/year was taken, patients had an 8.4-fold higher risk of liver failure, liver transplantation or death [11].An international multicenter study of 3985 PBC patients from 12 countries found that baseline LSM at cut-off values of 8 and 15 kPa allowed for the classification of PBC patients into low-, medium-, and high-risk groups, facilitating the early implementation of individualized interventions for patients in different risk strata [12].
In 2021, the Baveno VII Working Group [13] first proposed the "Rule of 5 kPa", defined as 5 kPa for LSM  should be used to represent the relative risk of progressively increasing decompensated events and liver-related deaths regardless of the etiology of chronic liver disease (B.1); the first proposal that patients with chronic liver disease with LSM < 10 kPa have a very low risk of decompensation or liver-related death within 3 years (≤ 1%, A.1); and the first proposal for a clinically significant decrease in LSM (CSDL), defined as a reduction in liver stiffness of at least 20% and < 20 kPa, or to less than 10 kPa, was associated with a significant reduction in the risk of patient decompensation or liver-related death (C.2).These "new concepts" have yet to be validated in patients with PBC.Therefore, this study was designed to assess the predictive value of LSM for long-term prognosis in patients with PBC through a retrospective cohort analysis and to validate the "novel concepts" proposed by the Baveno VII Working Group.

Study population
Patients attending the Xijing Hospital of the Air Force Military Medical University from January 2016 to January 2022 who met the inclusion and exclusion criteria were retrospectively included in the study.In total, 672 patients with PBC and 2552 valid LSM were included.The average number of LSM per patient was 3.2 ± 2.1 (interquartile range [IQR], 1-5).The average time interval between two successive LSM was 9 ± 5 months (IQR, [6][7][8][9][10][11]. Relevant serum biochemical, imaging, and other tests were performed at each elastography test.This study design was approved by the Ethics Committee of the Xijing Hospital of the Air Force Military Medical University.

Diagnostic criteria
Referring to the diagnostic criteria established in the 2017 Guidelines [14] for the diagnosis and treatment of PBC, PBC can be diagnosed by meeting any two of the following three criteria: (1) Elevated ALP, an indicator of cholestasis, and imaging that rules out extrahepatic or intrahepatic bile duct obstruction; (2) Positive anti-mitochondrial antibody (AMA)/AMA-M2 or positive for anti-sp100 and/or anti-gp210; (3) Pathology with typical evidence of PBC.

Inclusion criteria
(1) Perform at least one elastography test; (2) Meet the diagnostic criteria and be at least 18 years of age; (3) Standardized access to UDCA treatment.

Exclusion criteria
(1) Medication irregularities or the absence of pertinent clinical information; (2) Those with failed or unreliable LSM; 1 3 (3) Those with combined autoimmune hepatitis, viral hepatitis, primary sclerosing cholangitis, alcoholic liver disease, hepatomegaly, non-alcoholic fatty liver disease, IgG4-associated cholangitis, hemochromatosis, and drug liver injury; (4) Those with a previous decompensated event, a history of malignancy, or another end-stage disease.

Clinical data collection
Basic data such as gender, age, and body mass index (BMI) were collected; laboratory indicators such as ALP, GGT, TB, alanine aminotransferase (ALT), AST, albumin (ALB), platelets (PLT), immunoglobulin M (IgM), immunoglobulin G (IgG), and AMA were collected; imaging data such as abdominal ultrasound, electronic gastroduodenoscopy, and electronic computed tomography were collected; and pathological data were also collected.

Instruments and measurement methods
Elastography test: The LSM of a candidate with PBC was measured by trained physicians using FibroTouch [15][16][17][18][19][20] and expressed in kilopascals (kPa).The median of 10 or more consecutive successfully tests was taken as the LSM value.Results with deviation values greater than 1/3 of the median data were considered invalid [21,22].Liver pathology assessment: Liver tissue specimens were obtained by ultrasound-guided percutaneous liver puncture and specimens less than 8 mm in total length were considered unacceptable.Pathological assessment was performed independently by two experienced pathologists (at least one of whom was of associate chief medical officer rank or above).Cases, in which there was a disparity in fibrosis grade between the two pathologists, were resolved by consensus agreement.The METAVIR scoring system [23] was used to assess the staging of liver fibrosis in PBC.Stages 0-4 represent normal liver, confluent zone fibrosis, confluent zone fibrosis with minimal bridging fibrosis, bridging fibrosis, and cirrhosis, respectively.

Defining the primary outcome
The primary outcome was defined as the occurrence of any of the following events: liver-related death, liver transplantation, or cirrhotic decompensation.Patients with cirrhosis and ascites, hepatic encephalopathy, or variceal bleeding were defined as having cirrhotic decompensation.For patients who have died or received a liver transplant, data were censored at the time of death or transplant.For surviving patients without a liver transplant, data were censored at the onset of cirrhosis-associated complications or at the last follow-up visit.If more than one cirrhosis-related complication occurred during the follow-up period for patients who survived but did not have a liver transplant, data were censored at the first occurrence of a cirrhosis-associated complication.In our study cohort, no patients developed hepatocellular carcinoma (HCC) during follow-up, so we did not identify HCC as a primary outcome.

Statistical analysis
R v4.1.3(Tsinghua, Beijing, China) and SPSS 26.0 (IBM, NY, USA) were employed to analyze the data.Frequencies and percentages, median and IQR and means ± standard deviations were used to describe the data depending on the type of data, and comparisons between groups were made via the Fisher exact probability test, chi-square test, Studentt test, or Mann-Whitney U test.A generalized linear regression model was used to calculate the amount of change in LSM per unit of time, defined as ΔLSM/ΔT (kPa/year).A univariate Cox regression analysis was performed to screen PBC patients for risk factors for the primary outcomes.Statistically significant variables from the screening results were included in multivariate Cox regression analysis with forward stepwise selection, and the hazard ratio (HR) was determined.LSM progression rate was defined as (LSM at a particular point in time-baseline LSM)/ baseline LSM.
Optimal cutoffs for LSM progression rates were calculated using the R language package "survival".The Kaplan-Meier (K-M) method was used to calculate survival rates without primary outcomes, and the Log-rank and Cox test was used to compare differences in primary outcomes between groups.A two-sided p < 0.05 was considered a statistically significant difference.

Case screening process and study design
Figure 1 illustrates the case screening process and study design.The number of patients who underwent elastography testing during the follow-up period was 792, and 120 patients with previous cirrhotic decompensation events were excluded.Finally, 672 patients were included to explore the prognostic value of LSM and validate the "Rule of 5 kPa" and "Relationship between LSM < 10 kPa and prognosis" proposed by Baveno VII.During the follow-up period, a total of 297 patients with compensated advanced chronic liver disease (LSM ≥ 10 kPa) [13] were included to verify "CSDL" proposed by Baveno VII, who met the criteria of "at least two elastography tests during the follow-up period and at least half a year between the first and last test".

Baseline characteristics of patients with or without the primary outcome
The mean follow-up time for the 672 included patients was 37 ± 14 months.44 patients (7%) had primary outcomes during follow-up, including 6 (14%) liver-related deaths, 2 (4%) liver transplants, 18 (41%) variceal bleeding, and 18 (41%) ascites.The mean age of the patients at enrollment was 54 ± 9 years; 14% were male; the median BMI was 22.10 kg/m 2 ; and 91% were seropositive for AMA or AMA-M2.It was noted that the patients had received UDCA treatment for 28 ± 39 months (IQR, 0-46 months) prior to enrollment, and 244 patients (36%) began UDCA therapy after enrollment.

Validation of the "Rule of 5 kPa" and "Relationship between LSM < 10 kPa and prognosis" proposed by Baveno VII
The LSM cut-off values of 10, 15, and 20 kPa were used to divide the patients into four groups and the K-M method was used to compare the primary outcome-free survival between the groups (Fig. 2).Log-rank test analysis showed that primary outcome-free survival was statistically significant among the LSM < 10 kPa vs. 10 ≤ LSM < 15 kPa vs. 15 ≤ LSM < 20 kPa groups (p < 0.05, all).Limited to sample size, primary outcome-free survival was not statistically significant between 15 ≤ LSM < 20 kPa and LSM ≥ 20 kPa groups (p = 0.213).
Considering the negligible 3-year risk of primary outcomes in patients with LSM < 10 kPa and the cut-off value nature of 10 kPa for compensated advanced chronic liver disease [13], and 15 kPa has been well established as the cut-off value for the high-risk group [12].Therefore, we took baseline LSM threshold values of 10 and 15 kPa to categorize PBC patients into low-, medium-, and high-risk groups (Supplementary Fig. 1).

Defining a clinically significant increase in LSM in median-risk patients
Median-risk (10 ≤ LSM < 15 kPa) patients were equivalent to a T-junction and were more likely to progress to the high-risk group as compared to low-risk patients or to regress to the low-risk group during treatment as compared to high-risk patients.We previously verified CSDL to assess remission, and now we try to define CSIL to evaluate progression.Since LSM ≥ 15 kPa can be used to define high-risk patients, we tried to define LSM progression to 15 kPa as the progression group.Of the 213 patients included with 10 ≤ LSM < 15 kPa, 65 progressed from LSM < 10 kPa to ≥ 10 kPa with a mean time of 14 ± 8 months (IQR, [7][8][9][10][11][12][13][14][15][16][17][18][19], and no patient with LSM < 10 progressed directly to ≥ 15 without at least one LSM monitoring with 10 ≤ LSM < 15 kPa.Evaluated at 6, 12, and 24 months, Cox analysis showed that the primary outcome-free survival was all statistically significant (p < 0.05, all; Supplementary Fig. 2).
Further, we try to define the optimal LSM progression rate (ΔLSM/baseline LSM).At 12-month elastography test, we calculated the LSM progression rate.Univariate and multivariate Cox regression analysis revealed that LSM progression rate was an independent risk factor for the occurrence of the primary outcomes (p < 0.05; Supplementary Table 2), and the optimal cut-off value was 21.58% (Supplementary Fig. 3).Therefore, we took 20% as the optimal LSM progression rate.Evaluated at 6, 12, and 24 months, Cox analysis showed that the primary outcome-free survival was all statistically significant (p < 0.05, all; Supplementary Fig. 4).

Discussion
In this study, we assessed the prognostic value of LSM in patients with PBC and validated the "novel concept" proposed by Baveno VII.It was found that: (1) LSM and ΔLSM/ΔT were independent risk factors for primary outcomes; (2) the LSM 5 kPa scale (10-15-20 kPa) represented a progressively higher risk of the primary outcomes; (3) patients with LSM < 10 kPa have a low risk of primary outcomes at 3 years (< 1%); (4) cut-off values of 10 and 15 kPa can be used to classify patients into low-, medium-, and high-risk groups; (5) a clinically significant decrease in LSM was associated with a significantly lower risk of primary outcomes; and (6) a clinically significant increase in LSM was associated with substantially raised risk of primary outcomes.
Second-line drug therapy (obeticholic acid or fibrates) is recommended for PBC patients with a poor UDCA response, regardless of disease stage [14,24].The finding that elevated LSM and ΔLSM/ΔT were independent risk factors for poor prognosis in PBC was largely consistent with the finding of Corpechot et al. [11], provides clues for early identification of patients with poor response and timely initiation of second-line drug therapy, although more clinical data are 1 3 needed to determine the optimal prognostic cut-off.At the same time, this study found that LSM and ΔLSM/ΔT were more effective in assessing the prognosis of PBC patients than the classical indicators of ALP and TB.Future attempts should be made to integrate LSM and/or ΔLSM/ΔT into biochemical response criteria, which may optimize the effectiveness of prognosis assessment in PBC patients.The heterogeneity of disease progression in patients with PBC poses a serious challenge for fine-tuned management.An international multicenter study [11] of 3985 PBC patients from 12 countries found that baseline LSM cut-off values of 8 and 15 kPa were used to classify PBC patients into low-, medium-, and high-risk groups.The present study confirmed the "Rule of 5 kPa" (10-15-20 kPa) based on LSM, which represents a progressively higher relative risk of the primary outcomes, and facilitates the early implementation of individualized interventions for patients in different risk strata.
Our study confirmed that CSDL, defined as a reduction in liver stiffness of at least 20% and < 20 kPa, or to less than 10 kPa, can be evaluated at 6-, 12-, and 24-month elastography tests, and was associated with a significant reduction in the risk of the primary outcomes, suggesting that it could be used as a substitute for the primary endpoint event in clinical trials.This is supported by the results of the recent phase III bezafibrate trial in patients with PBC who responded poorly to UDCA (BEZURSO).Patients in the placebo group had a significant increase in LSM, while patients in the bezafibrate group had a stable or even reduced LSM [25].We further defined CSIL to evaluate the progression in patients with median risk or low.Our study demonstrated that CSIL, defined as an increase in LSM of ≥ + 20% or any increase to LSM ≥ 15 kPa, was associated with a substantially raised risk of primary outcomes, which can be evaluated at 6-, 12-, and 24-month elastography tests.
According to these findings, treatment and monitoring strategies based on LSM were established.Baseline LSM values of 10 and 15 kPa divide patients into low-, median-, and high-risk groups.For patients with median to high risk, CSDL can dynamically assess the disease remission status under current treatment; for patients with median risk or who progress from low to median risk, CSIL can dynamically evaluate the disease progression.Their clinical application may optimize the clinical monitoring and medication strategy for PBC patients.
The study was a single-center retrospective cohort study with some bias in the selection of patients; the follow-up cohort included a high proportion of early-stage (F0-F2) patients with a short follow-up period and a low proportion of patients with primary outcomes; given the small number of patients with paired liver biopsies and varying time intervals, it is difficult to evaluate the relationship between progression or remission of liver stiffness and pathological results; the patients included were a mixed group of primary and treated patients and could have affected the results.Future validation should be conducted in a multicenter, prospective, large sample size, and long-term follow-up clinical cohort study.
In summary, LSM can be used to monitor disease progression and predict long-term prognosis in patients with PBC.

Fig. 2
Fig. 2 Kaplan-Meier plots of the primary outcome-free survival of patients with PBC based on the rule of 5 kPa (10-15-20 kPa)

Fig. 3 Fig. 4
Fig. 3 Kaplan-Meier plots of the primary outcome-free survival of patients with a clinically significant decrease in LSM (CSDL) and without CSDL.a Evaluated at 6-month elastography test; b Evaluated at 12 months; c Evaluated at 24 months

Table 1 Baseline
ALP alkaline phosphatase, GGT gamma-glutamyl transferase, ALT alanine aminotransferase, TB total bilirubin, ALB Albumin, AST aspartate aminotransferase, IgG immunoglobulin G, IgM immunoglobulin M, PLT platelets, ULN upper limit of normal, LLN lower limit of normal, LSM liver stiffness measurements

Table 2
ALP alkaline phosphatase, GGT gamma-glutamyl transferase, ALT alanine aminotransferase, TB total bilirubin, ALB Albumin, AST aspartate aminotransferase, IgG immunoglobulin G, IgM immunoglobulin M, PLT platelets, ULN upper limit of normal, LLN lower limit of normal, CSDL a clinically significant decrease in LSM, LSM liver stiffness measurements