Aeration of Prussak's space is independent of the supradiaphragmatic epitympanic compartments.

HYPOTHESIS: This study's aim was to find out how well various microdissection approaches reveal the basic anatomy of the epitympanum, especially the pathways of aeration to Prussak's space, without the help of serial sections, which many find difficult to interpret. BACKGROUND: The basic studies where made during the latter half of the 19th and the first half of the 20th century. Conflicting concepts have later been published, and doubtful information has been included even in textbooks. METHODS: We have studied 145 temporal bones via microdissection to record the state of the soft tissue structures of the epitympanum, particularly upon Prussak's space with its boundaries. A normal surgical otomicroscope was used in the evaluation, and the findings were recorded via black and white and/or color photography; for recent cases, a digital video camera was used. RESULTS: The epitympanic diaphragm separates the large upper floor compartments from the small, laterally placed lower floor unit. The latter consists of Prussak's space and the posterior pouch, at times also of the lower lateral attic. The tympanic isthmus connects the upper unit to the medial tympanum. Defects in the diaphragm create additional airways to the upper unit, in 29% via the tensor fold and in 19% via the lateral incudomalleal fold. In only 7% was there a small opening in the roof of Prussak's space connecting it to the upper unit. Effective aeration of Prussak's space was independent of the upper floor compartments. CONCLUSION: Microdissection is a reliable and sufficient method for teaching epitympanic anatomy. All important structures can be identified and defects in the epitympanic diaphragm verified. Data obtained via serial sections are invaluable in research but not essential in the training of ear surgeons. The two-floor structure of the epitympanum with an independent aeration of the two units should be the starting point for all anatomy teaching.