Lifespan is a very central question in MMO development, and indeed central to any game model where there is expected to be more than a single release per platform - FarmVille-style casual social games, for instance.
If it were considered off topic, this SE would need to be renamed.
However, you aren't specifically asking for factors in calculating lifespan or anything, but rather just whether indefinite lifespans are possible for "large-scale" MMOs. A discussion of what constitutes success in MMO terms is a complex topic in and of itself, but there are a reasonable number of MMOs which have been continuously running for years, and at least one (Furcadia) which has been running for 20 years this year. Ultima Online is 19, though I'm not sure it's been running continuously.
As, until recently, the lead programmer of Furcadia, I'd be very interested in discussing some of the technical design decisions that were made to handle two decades of longevity (how to weather the thin times), which is in a way the exact opposite concern to the scalability issues that everyone normally worries about for MMOs.
Bartle's "Building Worlds" for example, stresses scalability very heavily, but carries the tacit assumption that if people start leaving, and the elder game becomes stale, then the game must die, and that assumption seems reasonably entrenched in the MMO design community.
I'm just a newbie to this community, but to me, this feels like an on-topic, critically important, and interesting question. In fact, it's exactly the kind of question I'd hope to see being covered in this community.