Twelve years ago we reported that lymphocytic alveolitis [or bronchoalveolar lavage (BAL) lymphocytosis] correlates with clinical pulmonary involvement in primary Sjogren's syndrome (pSS). Our thesis was based on subtle clinical and functional evidence of interstitial lung disease (ILD) in pSS patients with "high lymphocytic alveolitis" (>15% lymphocytes in BAL). This report is a follow-up study of these patients. Basic clinical and functional re-evaluation of the 22 patients with pSS, studied in 1991, emphasized the differences between those with alveolitis and those without alveolitis. There was no significant functional decline. There were, however, two statistically significant differences between the two groups: (1) only patients with BAL lymphocytosis had to be treated with steroids (5/12 vs. 0/10, P < 0.05) and (2) only patients with BAL lymphocytosis had died in the mean time (6/12 vs. 0/10, P < 0.01). The causes of death were various. On only two occasions were they related to respiratory infections while there were no deaths from respiratory failure secondary to ILD. BAL lymphocytosis appears to be a surprisingly serious index of dismal prognosis in patients with pSS. We offer no unifying pathophysiologic mechanism for it and, therefore, all we propose is that BAL is performed early, in as many patients with pSS as possible. These patients should then be followed up systematically, in order to evaluate if BAL lymphocytosis has any pathophysiologic importance in the development of clinically serious pSS, which is serious enough to lead to death.