Baseline patient characteristics of different subgroup analyses

To investigate different disease- and treatment-related outcomes in elderly AAV patients, the patient population was divided into AAV patients > 65 years and ≤ 65 years and followed-up for 5 years after disease onset (Fig. 1). Detailed patient characteristics of both subgroups are given in Table 1. By definition, elderly AAV patients were significantly older with a median (IQR) age of 71 (69–74) compared to 57 (37–60) (Table 1; P < 0.001). The disease activity measured by the DEI and the BVAS score as well as the organ systems involved were comparable between both groups. Patients > 65 years tended to have more comorbidities at disease onset such as condition after myocardial infarction and diabetes type 2 with 10% and 17% versus 3% and 7% in patients ≤ 65 years, respectively (Table 1; P = 0.08 and P = 0.09). Kidney function in elderly patients was significantly worse with a median (IQR) estimated glomerular filtration rate (eGFR) of 18 (12–49) compared to 30 (23–63) at disease onset in younger patients (Table 1; P = 0.03).

Table 1 Patient characteristics, outcomes, and complications of AAV patients \(\le\) 65 years and > 65 years

The population of elderly AAV patients was further subdivided into patients with and without DAH. Baseline characteristics including age, gender distribution, and comorbidities at disease onset were comparable between both subgroups (Table 2). The disease activity tended to be higher in patients with DAH as compared to patients without DAH with a median (IQR) DEI of 7 (5–8) and 5 (3–6) (Table 2; P = 0.07) and a median (IQR) BVAS score of 21 (15–35) and 16 (11–28) (Table 2; P = 0.09). Furthermore, significant differences in kidney function or the incidence of end-stage kidney disease between elderly AAV patients with and without DAH at disease onset.

Table 2 Patient characteristics, outcomes, and complications of AAV patients > 65 years with (+) and without (−) DAHDisease- and treatment-related outcomes and complications of elderly patients with ANCA-associated vasculitis

The majority of patients in both groups received intravenous CYC as induction therapy combined with a steroid pulse (Supplemental Table S1. Only 3 (4%) patients ≤ 65 and 4 (6%) patients > 65 years had RTX for induction therapy, respectively. The cumulative CYC dose per kilogram body weight was significantly lower in patients > 65 years with a median (IQR) of 34 mg (27–68) compared to 43 mg (28–86) (Supplemental Table S1; P = 0.03). For maintenance therapy, immunosuppressive drugs combined with steroids were not different between both groups (Supplemental Table S1). At disease onset, the administered oral steroid dose was significantly lower in elderly patients with a median (IQR) of 48 mg (35–61) than in patients ≤ 65 years with 60 mg (44–70) (Supplemental Table S1; P < 0.001). However, 3, 6, and 12 months after disease onset, the steroid dose no longer differed between the two cohorts.

Disease-related outcomes such as relapse rate, time to relapse, refractory disease, and disease activity measured by the DEI and the BVAS score 3, 6, and 12 months after disease onset were all comparable between patients > 65 and ≤ 65 years (Table 1). The relapse-free survival as a primary outcome between both groups is shown in Fig. 2A (log-rank P = 0.49). The overall survival of patients > 65 years was significantly lower with 17 (24%) compared to 6 (9%) deaths during the 5 years follow-up period (Fig. 2B; log-rank P = 0.01). Death by infectious complications occurred more frequently in the elderly cohort with 10 (14%) patients compared to 2 (3%) in the ≤ 65 years group (Fig. 2C; log-rank P = 0.02). In addition, the incidence of pneumonia was considerably higher in AAV patients > 65 years with 26 (37%) affected patients than in patients ≤ 65 years with only 9 (13%) patients (Fig. 2D; log-rank P = 0.001). Notably, the majority of pneumonias in both subgroups occurred within the first 12 months after disease onset. Opportunistic pathogens (Aspergillus species, Pneumocystis jirovecii, or the cytomegalovirus) were significantly more frequently the cause of pneumonia in the > 65 years cohort with 9 (13%) versus 2 (3%) affected patients (Table 1; P = 0.03). Three out of 35 patients (9%) with pneumonia had chronic lung disease before the pneumonia occurred. The risk of septic disease courses was also higher in elderly patients (Table 1; P = 0.04). Beside infectious complications, the incidences of treatment-related side effects such as steroid-induced diabetes (P = 0.04), osteoporosis (P = 0.008), new-onset arterial hypertension (P = 0.07), and leukopenia (P = 0.06) were increased in elderly AAV patients.

Fig. 2

Primary outcomes in AAV patients \(\le\) 65 years and > 65 years old. Primary outcomes—relapse-free survival (A), patient survival (B), death by infectious complications (C), and incidence of pneumonia (D)—in AAV patients \(\le\) 65 years and > 65 years old during a follow-up period of 5 years. To estimate the univariate probability, Kaplan–Meier estimators and the log-rank test were utilized, respectively

Impact of diffuse alveolar hemorrhage in elderly patients with ANCA-associated vasculitis

To examine the impact of DAH in elderly AAV patients, we further subdivided patients > 65 years into individuals with (N = 18) and without (N = 52) DAH. The CYC or RTX induction dose was not different between both subgroups. However, AAV patients with DAH received significantly more frequently steroid pulse doses (18 (100%) versus 40 (77%), P = 0.02; Supplemental Table S2) as well as plasma exchange therapy (8 (44%) versus 5 (10%), P = 0.001; Supplemental Table S2) for induction therapy. In addition, with a median (IQR) of 60 mg (49–77) and 20 mg (15–34) the oral steroid dose at disease onset as well as 3 months after induction therapy was significantly higher in AAV patients with DAH as compared to patients without DAH with a median (IQR) of 40 mg (28–58) and 12 mg (8–20), respectively (Supplemental Table S2). Six and 12 months after disease onset, the oral steroid dose was comparable between both groups.

Although the initial DEI and the BVAS score tended to be higher in elderly AAV patients with DAH, disease-related outcomes including the primary outcome relapse-free survival were not significantly different compared to patients without DAH (Table 2 and Fig. 3A). Disease activity as well as renal outcome parameters such as ESKD were comparable during follow-up (Table 2). In contrast, overall patient survival was significantly lower in elderly patients with DAH with 8 (44%) compared to 9 (17%) deaths during the 5 years follow-up (P = 0.02; Fig. 3B). Notably, 6 out of 8 fatal outcomes in elderly DAH patients were associated with severe infectious complications, especially pneumonia during the first 6–12 months. Both death by infectious complications [6 (33%) versus 4 (8%), P = 0.004] as well as the incidence of pneumonia [11 (61%) versus 15 (29%), P = 0.01] were significantly increased in patients with DAH (Fig. 3C + D). Sepsis [6 (33%) versus 4 (8%), P = 0.004] and pneumonia with opportunistic pathogens [5 (28%) versus 4 (8%), P = 0.03] occurred also more frequently in elderly AAV patients with DAH (Table 2). In addition, with a median (IQR) of 2 (2–4) in elderly patients with and 1 (0–2) in patients without DAH, the VDI 1 year after disease onset was tremendously higher (P < 0.001; Table 2). However, the VDI scores irreversible physical damage \(\ge\) 3 months after disease onset and is either affected by AAV disease activity or by treatment-related side effects and by implication there is no attribution of cause.

Fig. 3

Primary outcomes in elderly AAV patients with (+) and without (−) DAH. Primary outcomes—relapse-free survival (A), patient survival (B), death by infectious complications (C), and incidence of pneumonia (D)—in “elderly” AAV patients > 65 years old with (+) and without (−) diffuse alveolar hemorrhage (DAH) during a follow-up period of 5 years. To estimate the univariate probability, Kaplan–Meier estimators and the log-rank test were utilized, respectively

We also performed a detailed characterization of AAV patients ≤ 65 with (N = 16) and without (N = 53) DAH (Supplemental Tables S3 + S4). Except for an increased DEI and BVAS score at disease onset in DAH patients, baseline characteristics and disease-related outcomes were not significantly different between both subgroups (Supplemental Table S3). Comparably to elderly AAV patients with DAH, patients ≤ 65 with DAH more often received plasma exchange and had a significantly higher oral steroid dose at disease onset (Supplemental Table S4). However, whereas younger AAV patients with DAH also had more infectious episodes per patient, severe infectious complications such as pneumonia and sepsis were similarly common and death by infectious complications was consequently not significantly increased as compared to younger AAV patients without DAH (Supplemental Table S3). Overall death during follow-up tended to be higher without being statistically significant (Supplemental Table S3; P = 0.10).

Predictors of death by infectious complications in all patients (N = 139) with ANCA-associated vasculitis

To investigate predictors for death by infectious complications, we performed a multivariable cox regression analysis including all 139 AAV patients enrolled in our study. We controlled for the confounders age, DEI and BVAS score at disease onset, DAH, CYC induction dose, plasma exchange therapy, and the glucocorticoid dose after 3 months. As age was included as a confounding factor in our analysis, we did not subdivide patients into > 65 and ≤ 65 years. Only higher age at disease onset [OR 1.62, CI (1.45–2.12), P = 0.01] and higher GC dose after 3 months [OR 1.23, CI (1.05–1.65), P = 0.04] independently predicted death by infectious complications (Table 3).

Table 3 Predictors of death by infectious complications in all AAV patients (N = 139)