In an ongoing attempt to investigate qualitative salivary parameters in diseases affecting salivary glands, patients with rheumatoid arthritis (RA) were examined. Patients were selected from the Oslo RA register for the present study if they fulfilled the following criteria: age 52-74 years, disease duration 10-20 years, and disability score as assessed by the Modified Health Assessment Questionnaire < or = 2.5. From these 105 patients, two subgroups of patients were selected, one group with pronounced sicca symptoms from eyes and mouth, and one group without such symptoms. Sicca symptoms were assessed using a postal questionnaire with the questions on dry mouth and dry eyes of the European classification criteria for Sjögren's syndrome. Patients were excluded from further examinations if they used medication that could cause dryness in eyes or mouth. Thus, nine patients remained in the sicca group (having four or more sicca symptoms), and ten matched RA patients were selected for the nonsicca group. A healthy sex- and age-matched control group (n = 10) was also examined. In a preliminary report we have shown that differences in flow rates between sicca and nonsicca RA patients were limited to lower values of unstimulated whole saliva. To further evaluate salivary changes in RA, a disease frequently associated with secondary Sjögren's syndrome, we have studied qualitative salivary parameters in these patients,' including secretory rates of proline-rich proteins (PRPs), statherins, and histatins. In the present report, phenotypes of PRPs, the ratio of PRPs derived from the two loci (PRH1 and PRH2), and PRP concentration and output in parotid and submandibular saliva derived from the two loci are presented. Parotid (PS) and submandibular saliva (SS) were collected from all individuals using 2% citric acid as a saliva stimulus. PRPs in PS and SS were identified using a SMART microchromatographic system with a Mono Q column and a Tris-HCl/NaCl gradient (method adapted from ref. 5). For PRPs, the primary polypeptide products are coded for on two loci (PRH1 and PRH2), which have three and two commonly occurring gene variants, respectively. On PRH1, the proteins PIF-s, Db-s, and Pa are coded for, whereas PRP-1 and PRP-2 are coded for on the PRH2 locus. As each protein variant has a postranscriptional cleavage product, individuals will exhibit four, six, or eight PRPs in their saliva, depending on whether they are homozygous at both, one, or neither of the two loci. Accordingly, 18 possible phenotypes may exist, but as few as three phenotypes were found in 79% of the 127 healthy individuals examined by Hay et al. The SMART system allows the determination of the different acidic PRPs present in saliva. Concentrations of the various phenotypes were calculated by peak integration versus pure PRP standards. Total PRP concentration derived from each locus was calculated as the sum of the concentrations of PRP variants from that locus.