Recent studies have shown elevated inorganic pyrophosphate (PPi) levels in most knee joint fluid supernates from patients with pseudogout (PG) or osteoarthritis (OA) and more modestly elevated levels in some supernates from patients with gout or rheumatoid arthritis (RA) relative to PPi levels found in the venous blood plasma of normal or arthritic subjects. We measured the intraarticular PPi pool and its rate of turnover to better understand the significance of the joint fluid-plasma PPi gradient. Preliminary studies in rabbits showed that (32-P)PPi passed from joint space to blood and vice versa without detectable hydrolysis. Incubation of natural or synthetic calcium pyrophosphate dihydrate (CPPD) microcrystals with synovial fluid in vitro in the presence of (32P)PPi tracer showed no change in PPi specific activity in the supernate over a 19-h period so that exchange of PPi in solution with that in CPPD microcrystals could be ignored. Clearance rates of (32P)PPi and of (33P)Pi, as determined by serially sampling the catheterized knee joints of volunteers with various types of arthritis over a 3-h period, were nearly identical. The (32P)PPi/(32P)Pi was determined in each sample. A mixture of a large excess of cold PPi did not influence the clearance rate of either nuclide. The quantity of PPi turned over per hous was calculated from the pool size as determined by isotope dilution and the turnover rate. The residual joint fluid nuclide was shown to be (32P)PPi. The PPi pool was generally smaller and the rate of turnover was greater in clinically inflamed joints. The mean plus or minus SEM pool size (mu-moles) and turnover rate (percent/hour) in PG knees was 0.23 plus or minus 0.07 and 117 plus or minus 11.9, hydrolysis rate (%/h) to Pi was 27.7 plus or minus 13.2; in OA knees: 0.45 plus or minus 0.26 and 72 plus or minus 9.2, hydrolysis 6.9 plus or minus 0.9; in gouty knees: 0.8 plus or minus 0.41 and 50 plus or minus 11.6, hydrolysis 9.8 plus or minus 2.8; and in RA knees: 0.14 plus or minus 0.14 and 114 plus or minus 35.8, hydrolysis 236 plus or minus 116. PPi turnover (mumoles/hour) correlated with the degree of OA change present in the joint as graded by radiologic criteria irrespective of the clinical diagnosis. Mean PPi turnover in joints with advanced OA was greater than in those with mild or moderate changes (P smaller than 0.001), but the mild and moderate groups showed no significant difference. We conclude that synovial PPi turnover and elevated PPi fluid concentrations are not specific for PG patients, and that these factors alone cannot be the only determinants of CPPD crystal deposition.