Correlation Does Not Mean Causation; Except When it Does
It's important to remember that epidemiological axioms -- "Correlation does not equal causation" -- do not always apply to individual patient scenarios.
I saw a 40 year-old woman with a recent diagnosis of interstitial cystitis who told me of some weird symptoms she'd had a few days prior. She was shopping and suddenly found she couldn't read. Being a "person" first and a "patient" a distant second, she ignored it and continued shopping. On the drive home her left arm and hand went numb. Then a few hours later she began with a pounding left-sided headache accompanied by nausea and photophobia. Her mother said it sounded like a migraine, and suggested she take Excedrin, which she did. It took awhile, but the headache went away. She had never had a headache nor preceding symptoms like this before, and there was no family history of migraine.
Her neuro exam was completely normal but her age and the absence of any history or family history of migraine alarmed me enough to get an MRI of her brain. It was normal. When I called to tell her that, she relayed that she'd just had another similar episode. She also mentioned that the urologist had given her something called Urelle for her interstitial cystitis, telling her to take it on an as-needed basis. She had also been given a special diet for her bladder that was actually working quite well, so she hadn't taken the medication much. However she now realized that the two times she'd taken it were on the days she'd had the neuro symptoms and the headaches.
It turns out that Urelle contains several things, including hyoscyamine: a smooth muscle relaxant. It's used for bladder spasms because the bladder contains smooth muscle. Another function of smooth muscle in the body, though, is in the walls of blood vessels. Relax it there and blood vessels dilate. One of the suggested mechanisms of migraine is dilation of blood vessels in the brain. I may be just a little old family doc, but the causal connections seemed damn plausible to me. I suggested she stop taking the Urelle, which she was happy to do. She knows to call me if she has another episode even without it, but it hasn't happened yet.
Moral of the story: to explain new symptoms, think of new medications. Although the plural of "anecdote" is not "data", correlation very often equals causation in patient care. (Just not in epidemiology.)