The rose bengal score is one of the most commonly used tests for evaluation of ocular surface epithelial damage. The test is used in most Sjögren's syndrome criteria. We examined 24 female and four male patients with primary Sjögren's syndrome (primary SS) in order to evaluate possible correlation between the various tests for keratoconjuncivitis sicca (KCS), and for possible correlations to xerostomia and p-IgG levels. Among the KCS tests a high rose bengal score appeared to be the key parameter, being correlated to low break-up time (P less than 0.01), low tear lysozyme (P less than 0.01), appearance of snake-like chromatin in conjunctival imprints (P less than 0.05), low sialometry (P greater than 0.01) and high p-IgG (P less than 0.01). We followed another group of patients with primary SS (30 females and four males) for a mean period of 53 (range 27-76) months. The patients were divided according to their initial response to systemic treatment with bromhexine. KCS parametres and p-IgG were measured repeatedly during the observation period. Patients responding to and continuously treated with bromhexine (2/3 of patients) improved significantly (P less than 0.05) in rose bengal score, but had increasing levels of p-IgG. Non-responders kept their low tear-production rate and had also increasing p-IgG levels. However, when subdivided according to p-IgG level, the group of patients with relatively low p-IgG improved in rose begal score, whereas the high p-IgG-group increased in rose bengal score. The rose bengal score appears to be a useful key parameter when evaluating disease level and progression.