BACKGROUND: Primary angle closure glaucoma (PACG) accounts for nearly 50% of global glaucoma blindness. There are currently no public health strategies to deal with this problem. Screening and prophylactic treatment of primary angle closure suspects (PACS) with laser peripheral iridotomy (LPI) may form a feasible population-level intervention. However, more information about the natural history of PACS is required before such an approach could be considered. METHODS: Six hundred and forty-four participants aged 50 years with a central anterior chamber depth (cACD) of <2.53 mm underwent a full slit-lamp examination in 1999. Of these, 160 participants diagnosed as having occludable angles by gonioscopy (ISGEO classification) were excluded from all further analysis, leaving 484 for follow-up. Six years later, 95/484 (19.6%) had died. A total of 201 of 389 participants traced (51.7%) were re-examined. The potential risk factors for the development of an occludable angle were assessed using the chi squared test, t test and the Wilcoxon rank sum test. RESULTS: At follow-up, 41 participants (20.4%, 95% CI: 14.8 to 25.7) were diagnosed as having incident PACS. Narrower angles, identified by grading of limbal chamber depth and gonioscopy at baseline, were strongly associated with incident occludable angles (p = 0.01 and p<0.01, respectively). There was weak evidence of an association with change in cACD (p = 0.05), and no evidence of an association with age, gender, and baseline cACD for the development of occludable angles. CONCLUSIONS: Narrower angles as determined by limbal chamber depth grading and gonioscopy at baseline were the main risk factors identified for the development of occludable angles.