I would have had the same options as I had later - a4, e3 and e4. 4....e6 is more widely played, with very similar results, so I am not sure why Kenneth thinks dxc4 is slightly better.

What did you have prepared for 4..dxc4? which is slightly better than black's move here? Charles, I say slightly better here for two reasons: one, as black I prefer to destabilize white's pawn structure a bit, and two, dxc4 scores slightly better in GK'.s database in several similar variations. I'll add further that since this variation is widely played it is pretty much a main line of the QGD.

breaking the pin with Be7 is on the k-side and a very reasonable move here

Kenneth is right. However 5...h6 does avoid 5....Be7 6. e3 0-0 7. Qc2 h6 8. h4!? 5..Nb-d7 and going into the Cambridge Springs variation is another possibility.

Bh4 is different, but Bxf6 also good.

I quite like 7.e4, here, threatening e5, and striking in the centre before Black makes his claim on the Q-side.

I agree with archduke 7.e4 makes a serious threat. I well add that a4 does NOT stop b5. Example 7.a4 b5 8.axb5 cxb5 9.Nxb5 Qb6 10.Na3 Qxb2

But in the line given by Kenneth, white does have 11. Nxc4 winning the pawn back. This happened in Klecker Tadeas (2075) vs. Sveshnikov Evgeny (2506) which ended in a draw. I must admit e4 does score better and is more popular.

8...b5 is more obvious and seems more apt, as well.

...b5 can still be played here and it is still effective 8..Qa5 is part of what makes the earlier ..h6 not the best choice.

David Jameson commented by email "Glad to know I am not the only player to have thought Qa5 seemed ok, only to quickly realise the error."