I can't really say what the significance of the result is that you show without some more details on what exactly is being shown in the picture and where it is from.

Here though is a counter-example from the well-cited study of [Fich, Blitz & Stark (1989)][1] for our own Galaxy, which appears to show excellent agreement where the two techniques overlap between 5 and 9 kpc. In this picture the CO measurements are the solid bars and the 21cm measurements are the inner ones with dashed points.

However, in distant galaxies spatial resolution can play an important role. The spatial resolution of the 21cm observations will be much lower than the CO observations. This will "smear" the observed rotation curve so that you get contributions from gas at a range of radii. This will become very important in the inner regions where the velocity changes rapidly with radius, The exact effect will depend on the intrinsic profile.

I don't really see how this explains the differences in the curves you show and in general I believe the agreements are much better than this, though it is normally the case that the CO measurements are trusted more in the inner regions. Is there anything peculiar about this galaxy? A molecular ring or a bar or some such feature?

![Rotation curve from Fich et al. (1989)][2]


  [1]: http://adsabs.harvard.edu/abs/1989ApJ...342..272F
  [2]: https://i.sstatic.net/QIfw5.png