Background
Antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV) is a group systemic vasculitides, which can affect small vessels ranging from capillaries to intraparenchymal arterioles and venules of almost all the organs [
1]. AAV consists of three variants, such as microscopic polyangiitis (MPA), granulomatosis with polyangiitis (GPA) and eosinophilic granulomatosis with polyangiitis (EGPA) [
1,
2]. MPA and GPA mainly present similar pulmonary and renal symptoms despite different genetic and antigenic backgrounds, whereas EGPA often exhibit both allergic and necrotising vasculitic features [
1‐
3].
Since AAV can bring out an amount of inflammatory burden and damage in major organs including lungs, kidneys and heart, AAV can be occasionally fatal. So far, there have been diverse reports regarding all-cause mortality in patients with AAV from different ethnic and geographic backgrounds: one year-cumulative patient survival rates were from 82 to 95% in Western countries [
4], and that was reported to be 79.1% in Japan [
5]. In Korean patients with AAV, 10-year cumulative patient survival rate was estimated up to 92.8% [
6].
Renal and pulmonary involvements and five factor score (FFS (2009)) ≥ 2 at diagnosis of AAV have been reported as AAV-related risk factors for all-cause mortality [
6,
7], together with conventional risk factors for all-cause mortality among the general population including age, male gender, diabetes mellitus (DM) and hypertension (HTN) [
8]. Although infectious or immunosuppressive drug-related causes are major aetiologies of death in AAV patients, identifying predictors of all-cause mortality at diagnosis in immunosuppressive drug-naïve patients may have clinical implications in the real settings.
The fibrosis-4 index (FIB-4), which is calculated based on age, aspartate aminotransferase (AST), alanine aminotransferase (ALT) and platelet count, was first proposed to assess liver fibrosis in hepatitis C virus (HCV)-monoinfected patients [
9]. The critical cut-off of FIB-4 for predicting significant liver fibrosis (S2 or greater) is currently set as 1.45 [
10]. Recently, it has been reported that the baseline FIB-4 is significantly associated with poor outcomes in patients with liver diseases, such as hepatocellular carcinoma development or all-cause mortality [
11,
12]. FIB-4 was also reported to be an independent risk factor of chronic kidney disease in patients with non-alcoholic fatty liver disease [
13]. The association of FIB-4 with all-cause mortality was also demonstrated in patients with heart failure, which is a non-liver disease [
14]. With this report, we searched previous reports on the association between the baseline FIB-4 and all-cause mortality in patients with AAV. However, to our best knowledge, there was no report regarding the association of FIB-4 at the time of diagnosis of AAV with all-cause mortality during follow-up to date. In our retrospective AAV-cohort, all-cause mortality was observed only in patients with MPA and GPA, but not EGPA [
6]. Hence, in this study, we investigated whether FIB-4 at diagnosis could be associated with all-cause mortality in 132 immunosuppressive drug-naïve patients with MPA and GPA.
Discussion
In this study, we compared the predictive potential of FIB-4 at diagnosis ≥1.45 with those of conventional and AAV-related risk factors for all-cause mortality in 132 immunosuppressive drug-naïve patients with MPA and GPA. In the multivariable Cox Hazard model analysis, old age (≥ 65 years), DM, HTN, FFS (2009) ≥ 2 and FIB-4 ≥ 1.45 at diagnosis were included. Interestingly, we found that only FIB-4 at diagnosis ≥1.45 was associated with all-cause mortality during follow-up in patients MPA and GPA. We considered two reasons why FIB-4 ≥ 1.45 at diagnosis exhibited the relatively high predictive power for all-cause mortality. First, inflammation may accelerate systemic fibrotic change through various inflammatory signals. Thus, the extent of liver fibrosis may indirectly reflect the accumulated amount of inflammatory burden in non-liver diseases such as heart failure [
14]. We assume that FIB-4 at diagnosis could predict all-cause mortality in MPA and GPA patients in a similar manner. Next, we assume that FIB-4 might be influenced by both conventional and AAV-related risk factors for all-cause mortality prior to or at the time of diagnosis. To prove this assumption, we conducted the univariable and multivariable logistic regression analysis based on FIB-4 at diagnosis ≥1.45 using those risk factors at diagnosis. In the univariable analysis, age at diagnosis ≥65 years (OR 3.812) and DM at diagnosis (OR 5.200) were significantly associated with FIB-4 at diagnosis ≥1.45. BVAS at diagnosis ≥16 and FFS (2009) at diagnosis ≥2 exhibited a tendency to be associated with FIB-4, so they were also included in the multivariable analysis. In the multivariable analysis, only age at diagnosis ≥65 years (OR 3.088, 95% CI 1.157, 8.239) and DM at diagnosis (OR 4.556, 95% CI 1.678, 12.368) were significantly associated with FIB-4 at diagnosis ≥1.45 (Additional file
1: Table S1).
Since, age is one of variables comprising an equation of FIB-4, it is naturally accepted that FIB-4 may be directly correlated with age [
9]. Whereas, pre-existing DM could induce non-alcoholic fatty liver disease, one of the systemic complication of DM, which can lead to an increase in FIB-4 within 3 years from baseline [
17,
18]. The link between male gender and FIB-4 is controversial: according to a previous review regarding an effect of gender on the outcome of liver diseases, differences in incidence and severity of liver diseases between men and women may vary based on a disease-type [
19]. The relation between pre-existing systemic HTN and FIB-4 is also controversial: non-alcoholic fatty liver disease was identified as an independent risk factor for the development of systemic HTN [
20]. Therefore, we conclude that FIB-4 at diagnosis itself may be an independent predictor of all-cause mortality in MPA and GPA patients and at the same time, it may reflect the effect of old age and DM on subclinical liver fibrosis, leading to an increase in FIB-4 at diagnosis.
Despite no significant association in the multivariable logistic regression analysis, BVAS and FFS (2009) at diagnosis exhibited a tendency to be associated with FIB-4 at diagnosis. In cases of severe MPA and GPA, as a counterpart of TH1 and TH17 cells, which are important participants in the pathogenesis of AAV, Treg cells may enhance the production of transforming growth factors (TGF)-β, which can initiate and accelerate liver fibrosis [
21]. Moreover, the higher extent of inflammatory burden may provoke the augmented production of reactive oxygen species, which can subsequently promote cell differentiation of fibroblasts to myoblasts in liver [
22]. Since BVAS and FFS (2009) were calculated at diagnosis of MPA and GPA, together with FIB-4 at diagnosis, the effect of BVAS and FFS (2009) at diagnosis on FIB-4 at diagnosis might be negligible. However, considering an asymptomatic latency prior to diagnosis of MPA and GPA, BVAS and FFS (2009) at diagnosis might theoretically influence FIB-4 at diagnosis. Therefore, we also conclude that FIB-4 at diagnosis may reflect the subtle effect of the inflammatory burden of AAV on subclinical liver fibrosis, leading to an increase in FIB-4 at diagnosis.
In addition to conventional and AAV-related risk factors, we investigated the predictive potential of comorbidities at diagnosis for all-cause mortality. Interstitial lung disease at diagnosis (HR 6.981, 95% CI 1.535, 31.756) was significantly associated with all-cause mortality in MPA and GPA patients. However, chronic kidney disease over stage 3 (HR 2.615, 95% CI 0.883, 7.742), ischaemic heart disease (HR 3.493, 95% 0.703, 17.361) and cerebrovascular disease (HR 1.213, 95% CI 0.324, 4.546) at diagnosis exhibited no significant association with all-cause mortality. We added interstitial lung disease at diagnosis to the multivariable Cox Hazards model analysis together with 5 variables described in Table
3. Nevertheless, only FIB-4 at diagnosis ≥1.45 was significantly associated with all-cause mortality. Because interstitial lung disease is not an established conventional or AAV-related risk factor for all-cause mortality, we did not include interstitial lung disease in the multivariable Cox Hazards model in this study.
In this study, we first demonstrated that FIB-4 at diagnosis ≥1.45 is an independent predictor of all-cause mortality during follow-up in patients with MPA and GPA, and furthermore its predictive potential is higher than those of conventional and AAV-related risk factors for all-cause mortality. However, our study also has several issues. First, despite the significant association between FIB-4 at diagnosis and all-cause mortality, we could not clarify the direct mechanism of FIB-4 at diagnosis to estimate all-cause mortality. Second, we could not provide the concrete data on liver fibrosis by liver histology or transient elastography. Third, our study was designed as a retrospective study, we could not strictly control the confounding factors. Particularly, we could not confirm the repeated results of ANCAs in patients without ANCA. Fourth, this study was conducted in a single centre, the number of deceased patients was too small to augment the statistical power. Future prospective and multi-centric studies with a larger number of patients will validate the clinical significance of FIB-4 at diagnosis in predicting all-cause mortality during follow-up of MPA and GPA in real-world clinical practice.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.