If we invoke Descartes we should remember that he uses the Ontological Argument - or his own version of that protean argument. The original Ontological Argument as deployed by St Anselm (1077 or 1078 CE) also works from epistemology to metaphysics - from what we can conceive to what (necessarily) exists.
Anselm's exact argument - actually, pair of arguments - is a matter of scholarly dispute but it goes roughly like this :
(1) We have the concept of God, which is the concept of 'a-nature-than-which-no-greater-nature-can-be-thought'
(2) The concept of 'a-nature-than-which-no-greater-nature-can-be-
thought' is intelligible.
(3) Hence we can take the concept of 'a-nature-than-which-no-greater-nature-can-be-thought' as instantiable. (There could be such a nature whether there actually is or not.)
(4) A nature which is instantiated in reality is greater than one
which is not.
(5) So if a-nature-than-which-no-greater-nature-can-be-thought
were not instantiated in reality, then it would be possible to
think of a nature that is greater (for example, any nature that is
in fact instantiated in reality).
(6) But this would be a contradiction, since it is obviously impossible to think of a nature that is greater than a-nature-than-which-no-greater-nature-can-be-thought.
(7) Therefore, a-nature-than-which-no-greater-nature-can-be-thought
must indeed be instantiated in reality.
(8) Therefore, God is instantiated in reality, i.e. God exists.
I do not offer this, as presented, as an exercise in exact logic. (Getting Anselm into precise logical form is a huge task, no less because of textual than of formal difficulties.) But it represents a style of argument, preceding Descartes, that moves from epistemology to metaphysics : from concepts (the concept of God and of a-nature-than-which-no-greater-nature-can-be-thought) to reality (the existence of God).
Argument adapted and altered from Millican, 457-8; Nagasawa, 1029-30.
Endnote on St Augustine
One may add to the quotations cited in other answers :
For no one can "not know" that he himself is alive. If he is not alive, he cannot "not know" about it or anything else at all, because either to know or to "not know" implies a living subject. But, in such a case, by not positively affirming that they are alive, the skeptics ward off the appearance of error in themselves, yet they do make errors simply by showing themselves alive; one cannot err who is not alive. That we live is therefore not only true, but it is altogether certain as well. And there are many things that are thus true and certain concerning which, if we withhold positive assent, this ought not to be regarded as a higher wisdom but actually a sort of dementia. (Enchridon, VII.§ 20 : http://www.tertullian.org/fathers/augustine_enchiridion_02_trans.htm.
A version of the cogito appears in De Civitate Dei (On the City of God) :
'I am not at all afraid of the arguments of the Academicians, who say, ‘What if you are deceived?’ For if I am deceived, I am.' (De Civitate Dei, Book XI, 26).
[Nulla in his ueris Academicorum argumenta formido dicentium: Quid si falleris? Si enim fallor, sum : http://www.thelatinlibrary.com/augustine/civ11.shtml.]
Anselm, Proslogion, with the Replies of Gaunilo and Anselm, ISBN 10: 0872205657 / ISBN 13: 9780872205659
Published by Hackett Publishing Company, Inc., 2001.
Anselm, Proslogion, St Anselm : Basic Writings, trans. S.N. Deane, 2nd ed., La Salle : Open Court, 1962.
Peter Millican,'The One Fatal Flaw in Anselm's Argument'. Mind, 113, pp. 437-76.
Yujin Nagasawa, 'Millican on the Ontological Argument', Mind, New Series, Vol. 116, No. 464 (Oct., 2007), pp. 1027-1039.