I would imagine this has been checked already but have you scoped the coolant temp signal to the ecu throughout the warm up period? I had a Renault do something similar where the temp sensor signal would give a false value not long after first starting up but still plausible as far as the ecu was concerned so didn't throw up any fault code. Other than that I have had other variable valve timing problems (Rover VVC! ) causing engine vac problems and obviously upsetting fuelling due to manifold vacuum change/compression etc so the hydraulic lifter issue could well be the problem. Any chance of getting 4 gas readings before during and after the problem occurs?

Liam