Demographic and clinical characteristics of all RA patients
Among 755 enrolled RA patients, 242 who had missing laboratory data or abdominal ultrasound examination were excluded. Thus, a total of 513 RA patients were qualified for statistical analysis. As shown in Table 1, 78.4% included RA patients were female, with mean age of 51.8 ± 12.6 years old. The median disease duration was 60 months (IQR 21 to 120 months). According to CDAI, 85.4% RA patients were active (CDAI > 2.8), while 14.6% were remission (CDAI ≤ 2.8). Moreover, 27.5% RA patients haven’t received any previous glucocorticoid or DMARDs therapy for six months before enrollment (treatment naïve).
Prevalence of MAFLD and NAFLD in RA patients
The prevalence of MAFLD and NAFLD was 21.4% (110/513) and 20.5% (105/513) respectively. Specifically, 19.9% (102/513) RA patients fulfilled both criteria of NAFLD and MAFLD. Eight (1.6%) RA patients with chronic hepatitis B infection who were not classified as NAFLD were newly identified as MAFLD, while 3 (0.6%) RA patients with NAFLD were not complicated with metabolic abnormalities. There was no significant difference of MAFLD or NAFLD prevalence between male and female RA patients (MAFLD: 22.5% vs. 21.1%, p = 0.754; NAFLD: 20.7% vs. 20.4%, p = 0.941, Figure 1A). After adjustment for the age and sex composition ratio according to the data in the 2020 China Statistical Yearbook [22], the standardized prevalence of MAFLD and NAFLD in our RA cohort was 20.5% and 19.4%, respectively. The males had a higher standardized prevalence of MAFLD (25.4% vs. 15.6%, p = 0.020), and a higher tendency of standardized prevalence of NAFLD (23.4% vs. 15.4%, p = 0.050, Figure 1C) than the females.
The prevalence of MAFLD increased with age, and RA patients at 50-59 years had the highest prevalence (28.5%), followed by 60-69 years (23.3%) and 40-49 years (21.5%, Figure 1B). RA patients at 50-59 years had the highest standardized prevalence of MAFLD (27.2%), followed by 40-49 years (25.7%) and 60-69 years (21.0%, Figure 1D). The prevalence and standardized prevalence of NAFLD in RA patients among different age groups shared similar characteristics to MAFLD (Figure 1B, 1D).
The active RA patients had a higher prevalence of MAFLD and NAFLD than those in remission (MAFLD: 23.1% vs. 12.0%, p = 0.031; NAFLD: 22.1% vs. 10.7%, p = 0.023, Figure 1E). However, there was no significant difference of MAFLD or NAFLD prevalence among RA patients with different disease duration (Figure 1F).
Clinical and metabolic characteristics in RA patients with MAFLD or NAFLD
RA patients with MAFLD or NAFLD were older than those without (MAFLD: 54.4 ± 9.5 years vs. 51.1 ± 13.2 years, p = 0.027; NAFLD: 54.2 ± 9.7 years vs. 51.2 ± 13.2 years, p = 0.043, respectively, Table 1), but no significant differences of other RA disease characteristics were observed between 2 groups. As expected, RA patients with MAFLD or NAFLD had a significantly higher prevalence of metabolic abnormalities (Supplementary Table S1). In turn, RA patients with metabolic abnormalities, including T2DM, elevated TG, overweight/obesity, elevated WC, elevated HOMA-IR, and elevated BP, had a significantly higher prevalence of MAFLD and NAFLD (Supplementary Figure S1). Notably, RA patients with T2DM had the highest prevalence of MAFLD (43.0%) and NAFLD (39.2%), followed by RA patients with elevated TG (MAFLD: 42.9%; NAFLD: 39.7%), overweight/obesity (MAFLD: 41.1%; NAFLD: 38.1%).
Liver biochemistry and fibrosis indices in RA patients with MAFLD or NAFLD
Both RA patients with MAFLD and those with NAFLD had slightly elevated ALT, ULN < ALT < 2 ULN (MAFLD: 11.8% vs. 4.5%; NAFLD: 12.4% vs. 4.4%, both p < 0.05, Table 2). In addition, RA patients with MAFLD had a higher liver fibrosis score (NFS: median -1.56 vs. -2.19; Forns index: median 5.19 vs. 4.85, respectively), and a higher percentage of advanced fibrosis (NFS: 36.4% vs. 15.9%; Forns index: 33.6% vs. 22.6%). Similar findings were found for RA patients with NAFLD (liver fibrosis score: NFS: median -1.62 vs. -2.17 and Forns index: median 5.02 vs. 4.86, respectively; advanced fibrosis: only in NFS: 36.2% vs. 16.2%, all p < 0.05, Table 2).
Characteristics of RA patients according to the MAFLD subtypes
According to the criteria of MAFLD, RA patients with MAFLD were divided into 3 subgroups: overweight/obesity (n = 48, 43.6%), T2DM (n = 34, 30.9%) and lean/ normal weight (n = 28, 25.5%). Among the three subtypes of MAFLD, RA patients with T2DM MAFLD were older than those with overweight/obesity MAFLD (58.9 ± 7.6 years vs. 51.9 ± 9.9 years, p = 0.003), but there were no significant differences in other RA disease characteristics (Supplementary Table S2). There were no significant differences in the prevalence of metabolic abnormalities except for the highest prevalence of elevated WC in RA patients with overweight/obesity MAFLD and highest prevalence of prediabetes in RA patients with lean/normal weight MAFLD (Supplementary Figure S2).
The comparisons of liver biochemistry and fibrosis indices among RA patients with three subtypes of MAFLD were shown in Supplementary Table S3. There were no significant differences in liver biochemistry among RA patients with three subtypes of MAFLD. While RA patients with T2DM MAFLD showed the highest NFS (median -0.64 vs. -2.36 vs. -1.84, p < 0.001) and Forns index (median 5.62 vs. 4.61 vs. 4.84, p = 0.008), and the highest proportion of advanced fibrosis defined by NFS (58.8% vs. 21.4% vs. 29.2%, p = 0.004) compared with those lean/normal weight and overweight/obesity MAFLD subgroups. Furthermore, RA patients with T2DM MAFLD had a significantly higher FIB-4 level (median 0.95 vs. 0.72, p = 0.036) and advanced fibrosis ratio defined by FIB-4 (29.4% vs. 6.3%, p = 0.011) and by Forns index (52.9% vs. 22.9%, p = 0.015) compared with those with overweight/obesity MAFLD.
CVD events and 10-year CVD risk in RA patients with MAFLD or NAFLD
In this RA cohort, 56 (10.9%) patients concomitated with CVD events, including ischemic stroke (4.5%), heart failure (3.7%), myocardial infarction (2.3%), angina pectoris (1.4%) and peripheral arterial disease (0.6%). Compared with those without, RA patients with MAFLD or NAFLD had a higher incidence of CVD events (MAFLD: 17.3% vs. 9.2%; NAFLD: 18.1% vs. 9.1%), especially higher rates of myocardial infarction (MAFLD: 5.5% vs. 1.5%; NAFLD: 5.7% vs. 1.5%) and angina pectoris (MAFLD: 3.6% vs. 0.7%; NAFLD: 3.8% vs. 0.7%, all p < 0.05, Table 3). RA patients with MAFLD or NAFLD also had significantly higher 10-year CVD risk scores including China-PAR, SCORE and FRS (Table 3). The median China-PAR were significantly higher in RA patients with MAFLD or NAFLD (MAFLD: 3.7% vs. 2.1%; NAFLD: 3.7% vs. 2.1%, respectively, all p < 0.001).
A total of 166 (32.4%) RA patients had a high 10-year CVD risk. RA patients with MAFLD or NAFLD had higher proportion of high 10-year CVD risk than those without (MAFLD: 55.5% vs. 26.1%; NAFLD: 55.2% vs. 26.5%, respectively, all p < 0.001, Table 3).
Factors associated with CVD events and high 10-year CVD risk in RA patients
As shown in Table 4, univariate logistic regression analysis showed that MAFLD or NAFLD was associated with a higher risk of CVD events in RA patients (MAFLD: OR = 2.065, 95% CI 1.135-3.759, p = 0.018; NAFLD: OR = 2.215, 95% CI 1.215-4.040, p = 0.009, respectively). And the associations remained after adjustment for the potential confounders (MAFLD: adjusted OR [AOR] = 2.303, 95% CI 1.197-4.429, p = 0.012; NAFLD: AOR = 2.478, 95% CI 1.185-4.779, p = 0.007, respectively). Further analysis stratified by the MAFLD subtypes showed a significant association of T2DM MAFLD with risk of CVD events (OR = 3.044, 95% CI 1.286-7.204, p = 0.011).
Moreover, MAFLD or NAFLD was associated with the high 10-year CVD risk for patients with RA in the unadjusted analysis (MAFLD: OR = 3.533, 95% CI 2.283-5.469, p < 0.001; NAFLD: OR = 3.428, 95% CI 2.201-5.339, p < 0.001, respectively). After adjustment for the potential confounders, these associations remained (MAFLD: AOR = 3.184, 95% CI 1.777-5.705, p < 0.001; NAFLD: AOR = 2.870, 95% CI 1.597-5.156, p < 0.001, respectively).
To compare the performance of MAFLD and NAFLD in identifying CVD events and high 10-years CVD risk in RA patients, the NRI and IDI were performed to evaluate the improvement of diagnostic accuracy. The NRI and IDI was -0.011 (95% CI -0.025-0.003) and -0.002 (95% CI -0.007-0.002) for CVD events, and 0.012 (95% CI -0.014-0.038) and 0.005 (95% CI -0.003-0.013, all p > 0.05, Supplementary Table S4) for high 10-year CVD risk, respectively. Compared with NAFLD, the MAFLD criteria had no further improvement (either measured by NRI or IDI) in identifying CVD events and high 10-years CVD risk.