It's well known that some special classes of formulas have got specific model-theoretic properties in classical logic. For instance, universal, existential, positive, and Horn formulas are stable on subsystems, extensions, homomorphic images, and products, respectively. Moreover, each formula stable on subsystems is equivalent to a universal formula, and so on.
My question is about analogues of such statements for intuitionistic logic. Could you please tell me where to find an overview of these results?