Time of diagnosis has a major effect on how your disease is modeled (the term I would use instead of "diagnosed"). Technically speaking, if you've ever had a manic episode, you're considered to have Bipolar I, the most severe version. If you're 28 and had one at 17 and never again... you probably don't have Bipolar I. That could be cyclothymia aggravated by a hormonal or sleep disturbance. It could be a one-off. Perhaps out of a fear of getting it wrong, the medical assumption is that a person who's had one episode is prone to having more. That might be a good policy for doctors (i.e. don't prescribe certain medications to people with histories of depression, or (hypo|)mania) but it's not a great model of how mental illness actually works. Some cases get better with age, and some get worse, and we really don't know why.
As addiction is a mental illness with behavioral complications, I'm not surprised. An episode of addiction is something one has to be aware of for one's life, but the treatment (don't use the drug) is something that, evidently, many people follow perfectly.