Serological and inflammatory correlates of sinonasal involvement in granulomatosis with polyangiitis: an exploratory phenotype-stratified analysis of 86 patients

Background Granulomatosis with polyangiitis (GPA) often involves the upper respiratory tract. How ANCA status and systemic inflammation relate to sinonasal involvement is unclear, particularly when the disease is limited to the ear, nose and throat (ENT). Methods We analysed 86 patients with GPA seen at the multidisciplinary clinic of Tashkent State Medical University in 2022–2025 and classified at assessment as ENT-limited ( n = 26) or generalized ( n = 60). Sinonasal involvement was scored with an exploratory sinonasal involvement index (SII), the unweighted sum of 12 binary clinical and endoscopic findings (range 0–12). Within each phenotype, Spearman correlations of the index with CRP, ANCA status (qualitative result) and five other laboratory parameters were estimated with 95% confidence intervals, using a Bonferroni threshold of p < 0.0071 for each family of seven tests. The correlations were compared between phenotypes with Fisher’s r-to-z test; sensitivity analyses used the Bonett-Wright variance for rank correlations and examined the attainable range of the ANCA coefficient. Results ANCA was positive in 76 of 86 patients (88.4%): in 69.2% with ENT-limited and 96.7% with generalized GPA (odds ratio for ANCA negativity in ENT-limited disease 12.4, 95% CI 2.2-130.4). The mean index was 7.35 in ENT-limited and 2.35 in generalized GPA, and in both groups about three-quarters of it came from potentially reversible mucosal findings. In ENT-limited GPA the index correlated with CRP ( r = 0.573, 95% CI 0.24–0.79, p = 0.002) and with ANCA positivity ( r = 0.628, 95% CI 0.32–0.82, p < 0.001); both associations passed the Bonferroni threshold, also in the sensitivity analysis. In generalized GPA no correlation reached nominal significance (CRP r = 0.100; ANCA r = 0.049). The difference between phenotypes was nominal for CRP ( p = 0.026; 0.035 in the sensitivity analysis). For ANCA it could not be interpreted, because with only two ANCA-negative patients in the generalized group the coefficient there cannot exceed about 0.3. Conclusions The index was associated with CRP and ANCA positivity in ENT-limited GPA and showed no association in generalized GPA, but a difference between the phenotypes was not demonstrated. Eight of 26 patients with established ENT-limited GPA were ANCA-negative. Because every patient had an established diagnosis, the study cannot estimate the diagnostic performance of ANCA testing or of individual sinonasal findings. A prospective cohort with patient-level scoring of activity and damage is needed to confirm these observations.