General prognostic models may neglect vulnerable subgroups in ANCA-associated vasculitis

Abstract

Background: ANCA-associated vasculitis is an organ and life-threatening disease with the highest incidence in elderly patients. However, few studies have focussed on characteristics and treatment outcomes in a direct comparison of elderly and younger patients.

Methods: In a retrospective, single-centre, renal biopsy-cohort, patients were dichotomized by age ≥ 65 years to analyse baseline clinical, histological, laboratory and immunological characteristics and outcome differences in elderly and younger patients as regard to mortality, renal recovery from dialysis and eGFR after two years.

Results: In the biopsy registry, n = 774 patients were identified, of whom 268 were ≥ 65 years old. Among them, ANCA-associated vasculitis was the most prevalent kidney disease (n = 54 ≈ 20%). After a follow-up of 2 years, overall mortality was 13.4%, with 19% and 4% in patients ≥ and < 65 years of age, respectively. While 41% of elderly and 25% of younger patients were dialysis-dependent at the time of biopsy, renal recovery was achieved in 41% and 57% of patients, respectively. The accuracy of prediction differed significantly between the whole cohort and elderly patients as regard to mortality (sensitivity 46% vs. 90%, respectively) and between younger and elderly patients as regard to eGFR (r² = 0.7 vs. 0.46, respectively). Age-group-wise analysis revealed patients above 80 years of age to have particularly dismal renal outcome and survival.

Conclusion: In our cohort, ANCA-associated vasculitis is the single most frequent histopathological diagnosis among the elderly patients in our cohort. Elderly and younger patients have comparable chances of recovering from dialysis-dependent renal failure, with comparable residual independent kidney function after two years. This study suggests (1) relevant predictors differ between age groups and hence (2) models involving all patients with ANCA-associated vasculitis neglect important features of vulnerable subgroups, i.e., patients above 80 years old.

Keywords: ANCA-associated vasculitis; Elderly; Mortality; Outcome; Renal recovery; Renal replacement therapy; Survival.

Fig. 1

a Distribution of incidental biopsy-proven renal diseases. The upper figure depicts the distribution of renal pathology findings in the entire cohort. The lower figure shows the respective distribution in patients aged 65 years or older. Miscellaneous comprises, among others, glomerulonephritis with detection of antibodies against the glomerular basement membrane, endocapillary (proliferative) glomerulonephritis, fibrillary glomerulonephritis, mesangiocapillary glomerulonephritis, Alport‘s disease nephropathy, thrombotic micoangiopathies (including atypical hemolytic uremic syndrome, malignant hypertension), kidney infarction, cholesterol embolism, renal crisis in sclerodermia, Fabry‘s disease, sarcoidosis, renal lymphoma, nephronophthisis, nephrocalcinosis, Bunya-Viridae nephritis and others. b Incidence as a fraction of whole cohort of biopsy-proven kidney disease in respective age groups. (blue line depicts IgAN, red line indicates nephrosclerosis, green line depicts ANCA-ass. vasculitis)

Fig. 2

Flow-diagram of screening cohort and AAV patients’ clinical courses; highlighting end stage kidney disease and death. Percentage rates relate to the absolute number of cases in the respective level above. KRT kidney replacement therapy; CKD G5D end stage renal disease

Fig. 3

a Absolute cases of AAV (blue line) and two-year mortality (red line) in the respective age group. Kaplan–Meier plot of survival for the respective age groups (b) and the KRT and non-KRT group (c), respectively. d Proportion of patients with initial KRT dependency (blue line) and respective recovery rate (red line). e Descriptive outcome of patients with initial KRT-dependency. f Residual renal function in patients that recovered from initial KRT-dependency in the respective age groups

Fig. 4

Truncated Violin plot of predicted and observed events (death within 2 years follow-up) in the whole cohort (a) and elderly patients (b); cut off (differentiation) probability 0.5. Independent variables: age, initial KRT-dependency. The one additional (younger) deceased patient alters the model, so that probability for every data point is not equal in both models

Fig. 5

a Kaplan-Maier curve demonstrating renal survival of patients and number of patients at risk in their respective groups. Dashed lines indicate three and 24 months. Three months depict the earliest point in time to assign a patient to end stage renal disease. Grade of chronic kidney disease in the respective risk group (b)

Fig. 6

Multiple linear regression for eGFR after two years in the respective cohort. The coefficient of determination (r²) depicts a measure of association, with higher values indicating tighter association (in all data sets p < 0.0001). Independent variables: younger patients (serum albumin, age, initial eGFR, percentage of regular glomeruli) whole group (age, initial eGFR) and elderly patients (age, initial eGFR, haemoglobin level)