• comparison of the evolution between day 1 and day 14 of the mean echocardiographic score of pericardial effusion between treated and untreated (placebo) groups
  

• Number of tamponades [ Designated as safety issue: Yes ]
  
• Number of patients in whom the individual echographic grade is decreasing of at least one point
  
• Number of pericardiotomy
  
• Creatinemia
  
• Haemoglobinemia
  
• PE evolution in patients having an inflammatory syndrome (C reactive Protein >30)
  
• PE evolution in patients receiving a vitamin K antagonist
  

Following cardiac surgery, the incidence of Pericardial effusion (PE) is high (50-85%) . The risk of tamponade is well acknowledged : about 2%. We published in 2004 in CHEST a study which allows us to know the natural history of post-operative PE and to validate, for the first time the use of an echocardiographic classification for predicting the occurrence of a tamponade. NSAID are widely used in this setting, but no study has ever been conducted trying to assess their efficiency. The aim of the study is therefore obvious : must we use NSAID in order to prevent post operative cardiac tamponades ?.

In order to answer this question, we are going to conduct a double-blind randomized study comparing diclofenac to a placebo.

Every patient hospitalized in a post operative cardiac rehabilitation center less than 30 days after cardiac surgery and presenting at the first TTE (Trans Thoracic cardiac Echography) a PE of severity > 2 (that is to say about 10 % of the totality of the patients having undergone a cardiac operation) will be included.: after randomisation, patients will receive a placebo or diclofenac (50 mg ) bid, in a double blind way, during 14 days.

Trans thoracic cardiac echography, creatininemia, haemoglobinemia, International Normalized Ratio (for patients receiving a vitamin K antagonist) will be performed once a week during 2 weeks.

Clinical assessment will be done every day (there will be no outpatient

Primary end point : evolution of the mean echocardiographic score in each group

-Secondary end-points :

Number of tamponades Number of patients in whom the individual echographic grade is decreasing of at least one point Number of pericardiotomy Creatininemia Haemoglobinemia PE evolution in patients having an inflammatory syndrome (C reactive Protein >30) PE evolution in patients receiving a vitamin K antagonist

86 patients per group are necessary; therefore we will include a total of 200 patients