Is it possible to withdraw authorship almost 2 years after the paper has been published in a journal? Reasons could be anything ranging from disagreement on data interpretation, new data contradictory to the original one, etc.
I wouldn't retract authorship, science is all about revising research based on new data or a better explanation. I'd suggest publishing a new paper, and explaining what you disagree with or have found better interpretations of from the old paper. A withdrawal of authorship (even if possible) would be interpreted as some form of manipulation of data in the prior paper.
Is it possible to withdraw authorship almost 2 years after the paper has been published in a journal?
No, as a general rule you can't. You can publish an erratum if there's a clear-cut mistake in a paper that needs to be corrected or noted, and you can retract the paper if the mistake is so significant that nothing of any value can be salvaged. Retraction is a big deal, which you should resort to only if you have to. In practice, papers are retracted either for fraud or for shockingly big mistakes.
Other than that, you can't make any changes to published papers. In particular, you can't withdraw as an author just because you changed your mind or regret having been an author.
Reasons could be anything ranging from disagreement on data interpretation, new data contradictory to the original one, etc.
Reasonable disagreement on data interpretation would not be viewed as a compelling reason. If you disagreed at the time, then it should have been handled then; if you previously agreed but now disagree, then the paper correctly reflects your prior agreement.
New data contradictory to the old could be grounds for an erratum or retraction if it indicates fraud or serious methodological error in the original paper. On the other hand, the fact that the conclusions should be updated in light of new data is not in itself an error in the original paper.