Comprehensive interoperability between eHealth/pHealth systems requires properly represented shared knowledge. Formal ontologies allow specifying the semantics of health knowledge representation in a well-defined and unambiguous manner. The objective of this paper is to formally analyze – from a system-theoretical architectural perspective – existing clinical ontologies. The paper defines important ontology requirements for semantically interoperable pHealth/eHealth systems. Then, based on those requirements, 17 criteria are defined and used for analyzing 129 clinical ontologies. Statistical results confirm that most ontologies do not meet the defined criteria. OBO foundry defines a good approach to meet all defined criteria, but it does not cover yet the clinical domain as a whole. SNOMED CT was found the more comprehensive one, despite several restrictions.