The only way for a diode as drawn in your schematic I can imagine would be a second (separate) n-well guardRing with n+ diff contacts surrounding the VPNP's n-well - and, possibly, also the VPNP's collector p+ substrate contacts. In this case its cathode pin (n+ diff) should be connected to the highest positive potential (probably VDD). But of course its anode connection shouldn't point to the base but to the collector of the VPNP = substrate connection.
This construction also would implicate a parasitic lateral NPN, which is totally off, however. Would be nice to see the layout.