This was a retrospective study that used two patient level data (PLD) sources: (1) anonymous PLD from Symphony from Jan 2012 to Dec 2014 (3 years), and (2) Pinnacle registry data from American College of Cardiology (ACC) over the same time period with detailed clinical data such as blood pressure readings, peripheral edema flags, etc. In Symphony dataset there were 162 million (M) unique pts, 4.3M pts on celecoxib, 16.3M pts with osteoarthritis (OA)(15.4M only OA), and 2.3M pts with rheumatoid arthritis (RA) (1.4M only RA). In ACC dataset, there were 870,000 pts with edema flag. Results: In the Symphony database, edema rate was 1.0%, 2.6%, 10.6%, 7.1%, and 16.6% in normal, hypertensive, OA, RA, or combined OA/RA populations. However, in the ACC database, edema increased to 20% range and above presumably because of existing cardiovascular conditions. Logistic analysis showed a positive correlation of age and incidence of edema (p<0.0001). Comparing edema in pts without and with a disease condition, the incidence significantly increased for diabetes (P<0.0001), dyslipidemia (P<0.0001), and hypertension (P<0.0001), indicating the significant impact of disease conditions on incidence of edema. Increasing % of edema was observed for increment in doses and days of treatment with celecoxib plus any other drug combination, suggesting drug induced edema. Combination of celecoxib with ARB (olmesartan/IT-103) or ACE (lisinopril/IT-102) significantly reduced the incidence of edema.