Borderline Personality Disorder: When, If Ever, to Start Medication
A 24-year-old presents after her third self-harm crisis in eighteen months, referred for outpatient follow-up with an 11-week wait before psychotherapy starts. The disagreement isn't which drug -- it's whether a national guideline's case against medicating this diagnosis still holds during the gap before therapy is even available.
R.D., a 24-year-old veterinary technician, has worked the overnight shift at a busy emergency animal hospital for three years — the job she describes as the one place she reliably feels competent, steady hands with a frightened animal, calm under pressure, the version of herself she likes best. The version that showed up in the emergency department two nights ago was different: a coworker found fresh superficial cuts on her forearm the morning after their shift and drove her straight to the ED rather than to work. It was R.D.'s third self-harm-related ED visit in eighteen months, but the first serious enough to prompt overnight admission, and the first time anyone connected the pattern to something with a name rather than treating each visit as its own isolated crisis.
Her outpatient therapist diagnosed borderline personality disorder roughly two years ago, after R.D. had carried a diagnosis of treatment-resistant major depression for nearly five years — two SSRI trials, one SNRI trial, and a partial course of ECT a prior psychiatrist recommended and she never completed. None of it addressed what she actually experiences: not a sustained low mood but sharp, hours-long crashes, almost always triggered by a specific rupture in a relationship, followed by a return to baseline that looks, from the outside, like nothing happened. This episode followed the end of a fourteen-month relationship three nights earlier; the argument that ended it is, by her own account, one she can no longer clearly reconstruct. She has no other psychiatric or medical history, takes no current medications, and has never been hospitalized before this admission.
That pattern — repeated, sharply time-limited crises clustered around interpersonal rupture, embedded in an otherwise stable, high-functioning life — is closer to the clinical picture borderline personality disorder actually describes than the chronic, pervasive low mood her earlier chart implied, and it changes what a medication is being asked to do. A trial aimed at sustained mood elevation, the model her prior SSRIs were built around, was never well matched to symptoms that resolve within days regardless of what's prescribed. The question her outpatient team now has to answer isn't only which drug, but whether a drug has a defined job here at all. The two most-cited practice documents in psychiatry have converged on that point rather than split over it: NICE said in 2009 that drug treatment should not be used for this disorder or its associated symptoms, and APA's 2024 revision — which retired the symptom-domain algorithm the 2001 version was built around — found no pharmacotherapy effective for the core symptoms. Both point the same direction: to the psychotherapy. Which, for R.D., is eleven weeks away.
Outpatient follow-up, three days post-discharge
She has an intake slot with the DBT program in eleven weeks, and three ED visits in eighteen months tells me that gap is not a neutral waiting period — it's the highest-risk window she's going to face. I'd start a low-dose SSRI today, specifically framed to her as targeting the affective-instability piece, not as treatment for “BPD” as a diagnosis. I want to be careful about how I'm grounding that, though, because the guideline picture has moved: the symptom-domain algorithm most of us trained on comes from the APA's 2001 practice guideline, and APA replaced it in 2024. The updated guideline is blunt that no pharmacotherapy has established efficacy for the core symptoms of the disorder. What it does still allow — as a suggestion rather than a recommendation — is psychotropic treatment that's time-limited, aimed at a specific measurable target symptom, and adjunctive to psychotherapy. My argument is that an eleven-week wait for the psychotherapy is exactly the situation that provision was written for, not an exception to it.
I don't doubt the eleven-week wait is a real risk window — that part isn't in dispute. But “targeting a symptom domain, not the disorder” is doing more work than it can bear here.
You're right that the 2024 guideline keeps a narrow adjunctive provision — but read what the same document concluded to get there. It found no evidence of efficacy for the core symptoms, and it wrote its pharmacotherapy recommendations specifically to limit polypharmacy and prolonged medication treatment. That's not a green light with conditions attached; it's a warning with a narrow exception. NICE's 2009 guideline goes further still, stating that drug treatment should not be used specifically for the disorder or for the individual symptoms and behaviors associated with it. And the polypharmacy concern isn't a hypothetical slope: Paton and colleagues' 2015 audit of UK mental health services found 92 percent of patients carrying this diagnosis were prescribed a psychotropic, and roughly two-thirds were on two or more — most combinations never tested together for this indication. She has no medication now. What gets started today is what somebody inherits in five years.
I think you're both right about different questions. NICE's guideline doesn't only say no drugs, ever — the same document that rules out treating the disorder explicitly allows cautious short-term use of sedative medication as part of a crisis plan, agreed with the patient and lasting no longer than a week. It even names the shape of the drug it has in mind: something with a low side-effect profile, low addictive potential, minimal misuse potential, and relative safety in overdose — a sedative antihistamine is the example it gives. That is a different prescription from the one you two have been arguing about. What she needs in the next seventy-two hours and what she'd be taking in month four of an eleven-week wait aren't the same prescription, even if they'd start from the same bottle. I'd rather write something narrow, dated, and reviewed at her one-week follow-up than either leave her with nothing or start something that quietly becomes permanent because nobody put a stop date on it.
Agreed within the visit: hydroxyzine 25 mg at bedtime, explicitly time-limited and reviewed at a one-week follow-up rather than left open-ended. R.D.'s DBT referral was escalated to the program's cancellation list given the acute presentation, though the eleven-week baseline estimate stands unless an earlier slot opens.
Not agreed, and left as an explicit open question for the one-week visit rather than resolved today: whether the affective-instability-targeted SSRI should start if the DBT wait is still running at that point.
Hydroxyzine tapers off at the review visit as planned; no SSRI trial starts unless a new crisis emerges independently of the access question.
The affective-instability-targeted SSRI becomes the next real decision — not deferred indefinitely, but revisited with the actual waiting time in hand rather than an assumption made today.
Both physicians agreed this second decision, whenever it comes, deserves its own full discussion rather than being made today under crisis conditions — the same objection each had to a symptom-targeted prescription being written at this visit.