Valproic acid is a remarkable molecule.

It is one of the quintessential examples of a serendipitous antiseizure medication (ASM) discovery. As an amphiphilic molecule it’s an effective solvent and happened to be used in a series of experiments as the carrier for candidate ASMs—and as the unexpectedly effective experimental control.

It is available as an injectable solution; it is available as delayed-release or enteric-coated tabs; as extended-release tablets; as extended-release sprinkles; and as an oral solution. (The precise ionic form is slightly different for each, with valproic acid in the injectable, eg, and divalproex semisodium in the tablets.)

It is effective in status epilepticus and can be infused faster than phenytoin.

As for mechanisms of action? Sodium channel blockade; regulation of GABA through GABA transaminase and TCA cycle substrates; and antagonism of Ca++ channels (including the T-type channels implicated in childhood absence epilepsy or CAE).

It works for focal epilepsy; it works for idiopathic generalized epilepsy; it works for many types of DEE. For CAE it’s basically just ETX or VPA. Hell, it works for bipolar depression. The only epilepsy it’s truly problematic for is that due to mitochondrial dysfunction.

In short—it’s just great.

Except for hepatocellular toxicity. Except for messing with the urea cycle and causing hyperammonemia. Pancreatitis, weight gain, alopecia. One of the only truly hormonal adverse effects among the ASMs: polycystic ovarian syndrome.

It is deeply problematic in pregnancy and is almost completely contraindicated in people who could get pregnant.