Results
Baseline Clinical Characteristics of Patients With AAV
The median age of the study subjects was 60.5 years, and 35.0% of them were men. The mean value of BMI was 22.2 kg/m2, and the median BVAS and FFS were 12.0 and 1.0, respectively. MPA was the most common (55.1%) disease subgroups among AAV. Regarding organ involvement, chest (61.7%) and renal (59.3%) involvements were the most frequent clinical presentations, followed by those of ear, nose, and throat and general manifestations. RF positivity was found in 109 (50.9%) patients, while 174 (81.3%) patients were ANCA-positive. During the median observation period of 33.8 months, 23 (10.7%), 39 (18.2%) and 70 (32.7%) patients experienced all-cause mortality, ESKD and relapse, respectively (Table 1). On dividing the patients into the RF <12 IU/ml, 12 IU/ml ≤ RF <120 IU/ml, and 120 IU/ml ≤ RF groups, differences were noted concerning the age, diagnosis and pattern of organ involvement. Furthermore, there were differences in the laboratory data of WBC, neutrophil and platelet counts, as well as the ESR, CRP, creatinine and albumin levels (Supplementary Table S1, available at Rheumatology online).
Difference in Organ Involvement Patterns According to RF and ANCA Status
Based on RF and ANCA positivity, patients were divided into four groups: (i) RF (+)/ANCA (+) (n = 94); (ii) RF (−)/ANCA (+) (n = 80); (iii) RF (+)/ANCA (−) (n = 15); and (iv) RF (−)/ANCA (−) (n = 25). General manifestations were most common in the RF (+)/ANCA (+) group (58.5%), whereas cutaneous manifestation was the most common in RF (+)/ANCA (−) group (60.0%). In contrast, chest and renal manifestations were shown to be comparable in the RF (+)/ANCA (+) and RF (−)/ANCA (+) groups (P = 0.357 and P = 0.409, respectively) and were more frequent in these groups than in the other groups. Nervous system involvement was most frequently found in the RF (+)/ANCA (−) group (Table 2 and Supplementary Figure S2, available at Rheumatology online). When the clinical characteristics of RF (+)/ANCA (+) and RF (−)/ANCA (+) groups were compared, the RF (+)/ANCA (+) patients were older, and this group had a lower proportion of those with c-ANCA (or PR3-ANCA) positivity. Considering the laboratory test results, patients of the RF (+)/ANCA (+) group presented with higher WBC, neutrophil and platelet counts, as well as ESR and CRP levels, whereas the levels of creatinine and albumin were lower than those of the RF (−)/ANCA (+) patients. Haematuria was observed more frequently in the RF (+)/ANCA (+) group (P = 0.032) (Table 3). Consistently, patients with renal involvement in the RF (+)/ANCA (+) group were older and had higher laboratory marker levels indicating systemic inflammation and lower creatinine and albumin than those in the RF (−)/ANCA (+) group. The proportion of those with haematuria was also lower in the RF (+)/ANCA (+) group (Supplementary Table S2, available at Rheumatology online).
Comparison of Patient Outcomes and Medications According to RF and ANCA Status
For patient outcomes, there was a significant difference in progression to ESKD between the groups; the RF (−)/ANCA (+) patients experienced the highest rates (28.8%) of ESKD occurrence (Supplementary Table S3, available at Rheumatology online). In addition, the Kaplan–Meier analysis revealed that there was a significant difference in the ESKD- and relapse-free survival rates between the groups (P = 0.008 and P = 0.036, respectively), although there was no difference in the overall survival rate (P = 0.422) (Figure 1A–C). Moreover, the RF (+)/ANCA (+) group exhibited a significantly higher ESKD-free survival rate compared with the RF (−)/ANCA (+) group (P = 0.013) (Supplementary Figure S3A, available at Rheumatology online). Similarly, in a subgroup analysis of patients with renal involvement in the RF (+)/ANCA (+) and RF (−)/ANCA (+) groups, the ESKD-free survival was higher in the RF (+)/ANCA (+) group (P = 0.009) (Supplementary Figure S3B, available at Rheumatology online). Regarding medications that were administered during the follow-up period, a difference was noted in the usage of cyclophosphamide, mycophenolate mofetil, azathioprine and tacrolimus (Supplementary Table S3, available at Rheumatology online).
Figure 1.
Comparison of patient outcomes according to RF and ANCA status
Patients were divided into four different groups: RF (+)/ANCA (+), RF (−)/ANCA (+), RF (+)/ANCA (−) and RF (−)/ANCA (−), and the patient outcomes of: (A) overall, (B) ESKD-free and (C) relapse-free survival rates were compared between the groups. ESKD: end-stage kidney disease.
Further, a subgroup analysis of patients with renal manifestation (n = 127) was performed to elucidate the risk factors for ESKD in our patients. In a univariate analysis, the rise in FFS [hazard ratio (HR) 1.588, 95% CI: 1.115, 2.263, P = 0.010] and creatinine level (HR 1.793, 95% CI: 1.560, 2.061, P < 0.001) was associated with increased risk of ESKD, while RF positivity (HR 0.458, 95% CI: 0.226, 0.927, P = 0.030) and increased haemoglobin level (HR 0.783, 95% CI: 0.649, 0.944, P = 0.011) was related to decreased risk of ESKD. However, in a multivariate adjusted model, only increased creatinine levels were predictive of ESKD (HR 1.762, 95% CI: 1.519, 2.044, P < 0.001) (Table 4). On analysing factors associated with ESKD in patients with renal manifestation and FFS ≥2 (n = 72), RF positivity seemingly had a protective effect (HR 0.372, 95% CI: 0.157, 0.878, P = 0.024). Nonetheless, in a multivariate analysis, this tendency was revealed to be insignificant (Supplementary Table S4, available at Rheumatology online).
Patient Characteristics and Outcomes in RF (+)/ANCA (+) and RF (−)/ANCA (+) Groups After PSM
Considering that there are substantial differences in baseline patient characteristics between the RF (+)/ANCA (+) and RF (−)/ANCA (+) groups, a PSM analysis was carried out. However, even after PSM was performed, it was found that WBC, neutrophil and platelet counts, as well as ESR and CRP levels were higher and albumin levels were lower in the RF (+)/ANCA (+) group (Table 5).
On comparing the difference in patient outcomes after PSM, it was found that the overall, ESKD-free and relapse-free survival rates were comparable after PSM (P = 0.528, P = 0.807 and P = 0.826, respectively) (Supplementary Figure S4A–C, available at Rheumatology online). Identification of factors associated with ESKD in these patients also showed that creatinine level (HR 1.786, 95% CI: 1.323, 2.412, P < 0.001) was the sole predictor, which is consistent with the results without performing PSM (Supplementary Table S5, available at Rheumatology online).
Rheumatology. 2022;61(4):1366-1375. © 2022 Oxford University Press
