I'm confused by the CPT theorem. It states (more or less) that a Lorentz invariant quantum field theory needs to be CPT invariant. But what does it actually mean for a QFT to be CPT invariant? It surely means that it's Lagrangian is. What about the Hamiltonian? Does the invariance get inherited by it as well? What about other observables that I can measure in an experiment? Are all of those also invariant under CPT (even though they might not be Lorentz invariant)?
Related to this: If I have a Lagrangian (density) that transforms in a specific way under C, P, and T, and I derive a low energy Hamiltonian from it, does this one necessarily inherit the same transformation properties? And if yes, is the reverse true? If a Hamiltonian (of an isolated system) is odd under C, P, or T, will this necessarily correspond to violations of the same symmetries at high energies?