Power of 12
Some background first. A decade ago a lot of the definitions used to define metropolitan statistical areas changed. MSAs are defined primarily by commuting flows. If the population of an areas is large enough, and pairs, or groups, of counties have sufficiently high flows of commuters between them, they are considered to be integrated in one of various ways depending on how big the area is, and how big the commuting flows are.
There used to be metropolitan statistical areas, typically collections of counties, or in New England collections of municipalities. If two or more metropolitan statistical areas were near each other, the collection of MSAs were called Combined Metropolitan Statistical areas (CMSAs).
Then those definitions changed changed. First a new set of areas were created called micropolitan statistical areas were created. Places not big enough to really be considered metropolitan statistical areas, but still bigger than a lot of other smaller areas. CMSAs as previously defined went away, so if you catch anyone talking about the Pittsburgh CMSA, they are a couple decades out of touch. Now, really big metropolitan statistical areas can have subdivisions called metropolitan divisions, a bit analogous to what long ago were called Primary Metropolitan Statistical Areas (PMSA), also a term no longer in use.
The collection of metropolitan AND micropolitan statistical areas are types of the more generic core based statistical areas (CBSAs). Now if adjacent CBSAs (metro or micro areas) are sufficiently integrated by commuting patterns they form a separate Combined Statistical Areas (CSA). Notice how you really ought not use the acronym MSA any longer, since it does not distinguish between metros and micros.
In 2003, the Pittsburgh MSA was defined to be 7 counties (Allegheny, Armstrong, Beaver, Butler, Fayette, Washington and Westmoreland). Armstrong was new to the MSA just that year. There were 2 nearby micropolitan statistical areas: Indiana County formed one, and Lawrence County (the New Castle micropoltian area) formed the other. Indiana was kind of an independent CBSA without sufficient commuting into the Pittsburgh MSA to be linked to it within the CSA definition at the time. However, the New Castle micro (otherwise known as Lawrence County) did have sufficient commuting into Pittsburgh to be linked with it, so there was an 8 county CSA called the Pittsburgh-New Castle Combined Statistical Area.
Who is still reading?
So last year, and I am embarrassed to say I did not catch this at the time, but the occasional redefinition of CBSAs had some major changes to the definitions of CBSAs regionally as of February 2013. Two big changes. 1) Indiana County is now integrated into the Pittsburgh CSA. Still not integrated into the MSA, but the CSA. A bigger change is that the Weirton-Steubenville Metropolitan Statistical Area became more formally linked to Pittsburgh. It remains a separate metro area, but the two counties ( Brook and Hancock counties in WV and Jefferson County in OH) are also linked to Pittsburgh in a much bigger CSA. So now there exists a 12 county Pittsburgh-New Castle-Weirton Combined Statistical Area.
So with the most recent revisions, one definition of Pittsburg stretches into both Ohio and West Virginia for the first time. Only 20 more counties to go.
I should have caught this sooner since a couple years ago I wonked out on what anticipated changes that were coming down the road. But while I am slow in noticing... Wikipedia gnomes know all. They even have a map of the 'new' Pittsburgh CSA: