Some people make these balloons using condom tips which they tie on with thread. Others buy small balloons from Harvard Apparatus.
Both of these are latex solutions which are not ideal because latex has a high compliance (elasticity) and therefore causes a damping of the pressure waveform. Whilst the Harvard balloons and the condom tips will easily and accurately record systolic and diastolic values, they are useless if you are interested in recording the dp/dt of the rise in pressure which many people want.
Therefore it is better to form a balloon from thin sandwich wrapping such as "Clingwrap" or similar. I think that there is a well known brand called Seranne wrap in North America but I am unsure about the spelling. This material is thin, strong and has a very low compliance which makes it ideal. Basically, the user ties it over the end of a cannua leaving some bunched at the tip so that when it is dropped into the heart it can be inflated.
Any balloon should be filled with water as hydrolic transmission of the pressure waveform is more true than gaseous transmission because gases compress.
This is what the guys at St. Thomas| Hospital do. This method is referred to in the H.E.L.P. section of the International Society for GHeart Research web site: http://www.ishrworld.org/ under the title of "The Isolated Blood And Perfusion fluid Perfused Heart"
Further information and instructions can be seen here.