Long-term risk of mortality and lymphoproliferative disease and predictive classification of primary Sjogren's syndrome

Objective. Primary Sjogren's syndrome (SS) may lead to lymphoproliferative disease (LPD) and death in certain patients. We sought to determine the incidence and predictors of adverse long-term outcomes to achieve a rational predictive classification of the syndrome. Methods. Predictive modeling was performed in a cohort of 723 consecutive patients with primary SS (587 newly diagnosed [incident] cases and 136 prevalent cases). Results. During 4,384 person-years of followup, we recorded 39 deaths (7 due to lymphoma) and 38 diagnoses of LPD. The standardized mortality ratio was 1.15 (95% confidence interval [95% CI] 0.86-1.73) compared with the general population of Greece. In incident cases, the probability of LPD was 2.6% at 5 years and 3.9% at 10 years. Mortality rates were significantly higher in patients with low C4 levels at the first study visit (hazard ratio [HR] 4.39, 95% CI 2.18-8.83). LPD was independently predicted by the presence of parotid enlargement (HR 5.21, 95% CI 1.76-15.4), palpable purpura (HR 4.16, 95% CI 1.65-10.5), and low C4 levels (HR 2.40, 95% CI 0.99-5.83) at the first study visit. All patients who eventually developed lymphoma resulting in death during the followup period had either low C4 levels or palpable purpura at the first study visit. Training-validation split-cohort modeling confirmed the predictive importance of low C4 levels and palpable purpura, both of which were present in 20.9% of patients at their first visit. Conclusions. In patients with primary SS, 1 in 5 deaths is attributable to lymphoma. The presence of palpable purpura and low C4 levels at the first visit adequately distinguishes high-risk patients (type I primary SS) from patients with an uncomplicated disease course (type H [low-risk] primary SS).