I think your interpretation is correct, no conclusion can be drawn from the results of this study.
If a similar study were repeated and adequately powered, I wouldn’t consider it p-hacking. The larger sample should only decrease the uncertainty in the primary outcome. P-hacking would be like if they set out to measure like 20 different outcomes to see if some turn out to be statistically significant by pure chance.
As an aside, I think it might be hard to gather enough participants for an adequately powered study given how close the outcomes were in this study. I’m no cardiac surgeon, but I think mechanical heart valves are generally less favored compared with bioprosthetics. Furthermore, the addition of aspirin to warfarin may cut down the number of participants as well, as aspirin is not routinely recommended.
The link to the follow up study seems to be split, but I was able to pick together the link There are two major differences between the two studies: 1. The first one looked at mitral valves, and this one studied aortic valves; 2. The target INRs were different, although given the difference between aortic and mitral valves, both correspond to an INR 0.5 points lower than standard targets.
In your case, given the risks involved, a mechanical valve sounds like the best choice. I mostly see valve replacements in the elderly, so there’s my personal observational bias. I’m from the US. From what I can tell about our warfarin recommendations, On-X aortic valves do get warfarin and aspirin at lower INR targets. In contrast, there is no such recommendation for On-X valves in the mitral position (yet).