Divalproex or Fluoxetine First-Line for Intermittent Explosive Disorder
A married electrician's explosive rage at home has real evidence behind two very different drugs — and the trial that tested the mood stabilizer's specific value for his exact diagnosis is harder to read than either voice at the table first assumes.
M.K., a 41-year-old man, has been married to his wife Renee for twelve years, and she came to this visit with him, which he says he wouldn't have agreed to a year ago. Last week he punched their front door hard enough to crack the frame and screamed at her over a misplaced set of keys — the kind of trigger, he says, that he knows doesn't warrant the reaction even while it's happening.
He's a self-employed electrician who has never raised his voice at a client or a coworker in fifteen years of work; the pattern is almost entirely confined to home, roughly weekly for the past year, and Renee is the one who finally insisted on this appointment.
A structured screening today did not find a personality disorder underneath the pattern — his difficulty is specifically episodic and reactive, not a pervasive, longstanding way of relating to people, which sharpens rather than simplifies today's question. Two years ago he tried sertraline for situational anxiety around a business slowdown; it helped the anxiety but he stopped after three months because of sexual side effects, not because it failed — he was never on it specifically for outbursts like these. His brother was diagnosed with bipolar I disorder in his thirties, a family history that sits in the room without being decisive on its own.
The real complication isn't any single fact in his history — it's what the actual trial evidence for a mood stabilizer in a patient who looks like him turns out to say once you read past the headline. Hollander and colleagues (2003, Neuropsychopharmacology) randomized 246 patients across Cluster B personality disorder, intermittent explosive disorder, and PTSD to divalproex or placebo: the combined intent-to-treat analysis found no overall treatment effect, and the positive result the trial is usually cited for was specific to the Cluster B personality disorder subgroup — the one group M.K.'s own screening rules him out of. Whether that positive finding, or the trial's null combined result, is the more relevant piece of evidence for him depends entirely on which subgroup a reader assumes describes his own diagnosis.
Initial visit, with his wife present
I'd start with fluoxetine, and the reason is specifically about which trial actually describes him. Coccaro, Lee, and Kavoussi tested fluoxetine directly in a hundred adults with IED and found a real, sustained reduction in aggression from week two. Hollander's divalproex trial did randomize 116 IED patients alongside Cluster B personality disorder and PTSD patients, but the positive finding that trial is usually cited for belongs specifically to the Cluster B subgroup — and M.K.'s own screening today rules a personality disorder out.
I'd weight his history differently. He already tried an SSRI — sertraline, two years ago — and stopped it for sexual side effects he found genuinely intolerable, not because it didn't work. That's real information about how he tolerates the class, even though it wasn't prescribed for this exact problem. And his brother's bipolar I diagnosis isn't nothing — a mood-stabilizing mechanism may fit whatever's underlying this better than another serotonergic drug does.
You're reading the Cluster B subgroup finding as disqualifying, but divalproex's trial still included a real subgroup of 116 IED patients — that's not zero evidence in his diagnosis, even if the headline result belongs to a different group.
I think you're both arguing an efficacy question that may be closer to a genuine tie than either of you wants to grant — and if that's true, the practical difference between the two drugs matters more than either of your positions is currently weighting it.
Divalproex needs baseline and periodic liver function tests and a blood count, carries real hepatotoxicity and platelet risk, and needs level monitoring for the rest of however long he's on it. Fluoxetine needs almost none of that. Where the efficacy case for his exact diagnosis is genuinely uncertain on both sides, the simpler option isn't a fallback for indecision, it's a legitimate way to break a tie neither of you has actually resolved with the trial data alone.
Agreed: fluoxetine starts today, with baseline liver function tests and a blood count already drawn so a later switch to divalproex would not be delayed if needed, and a couples-focused anger-management referral made alongside the medication.
Not agreed: how much weight the family history and the prior SSRI intolerance should carry going forward if fluoxetine only partially helps. The second psychiatrist wants those two factors to lower the bar for switching to divalproex sooner rather than trying a second SSRI first; the first psychiatrist would want to establish that fluoxetine specifically, not the SSRI class in general, has been given a fair trial before conceding the point.