Lamford in his 1983 book gives 20.Ke2 indeed: Ne7 21.Qd3 Qd5 22.h4 Rd7 23.Rad1 Rhd8 24.Qb3 Qe4 with equality.
He does not commend 19.Bxe4 though.

Yesterday I played one of my favourited surpise weapons against a club player, the Albin-CounterGambit. I play black. The game went

1.d4 d5
2.c4 e5
3.dxe5 d4
4.Nf3 Nc6
5.Nbd2 Bg4
6.a3 Qe7
7.h3 Bh5
8.g4 Bg6
9.Bg2 O-O-O
10.b4
Up to here the game followed an earlier game (that I wasn't aware of between Kaminsky and Mosionzhik)
Now I continued 10..Nxe5 (instead of 10..h5 as in the mentioned game)
11.Qa4?! Nd3+
12.Kf1 Nxf2!!
13.Kxf2 Qe3+
14.Kf1! d3?! (14..Qc3 was better)
15.Qd1?? and after that blunder there is no need to mention the continuation. 

Anyway, after realising that 8.g4 was theoretical move I was curious to check the game and I found this comedy, which must be the World Chess Record for blunders in a official game with 2200+ opponents:
white: Kaminsky, Oleg (2333 according to FIDE)
black: Mosionzhik, Ilia (2228 according to FIDE)
year: 1971

10..h5
11.g4 Nxe5
12.Nxe5 Qxe5
13.Nf3 Qe6
14.Bf4 Bd6 (Fritz 8 theory stops here)
15.Bxd6 Rxd6
16.c5 Rd8
17.Nxd4 Qe5
18.e3 Be4 Now take a deep threat cause we start:
19.Bxe4? Qxe4
20.Qe2?? Ne7?
21.Qd3?? Qd5?
22.h4?? Rd7?
23.Rd1?? Rhd8?
24.Qb3?? Qe4? and at last
25.f3

Does anyone know the story of this game?