%20(1).webp)
A 34-year-old woman is referred to you with a nine-year EUPD diagnosis, two completed courses of DBT, self-harm that stopped four years ago, and a covering letter saying the diagnosis "has never quite fitted". Her mother says she was away with the fairies from reception onwards. Her partner says the rows blow up in forty minutes and are over in one. You have ninety minutes.
This is the differential most often got wrong in both directions, and the one where getting it wrong costs the most. What makes it hard is not that the two conditions are subtle. It is that the features they share surface first, and the features that separate them only appear when you interrogate time and this also a crucial part of ADHA accessor's traning certificate.
One failure mode is sequential. A woman presents in her twenties with self-harm, relationship chaos and volatility. Those map onto EUPD criteria easily, the assessment stops there, and nobody takes a developmental history because the presenting problem has already been explained.
The reverse has become more common since 2020. Someone with a genuine personality disorder presents with inattention, screens positive on a self-report measure, and is titrated onto a stimulant that does nothing for the interpersonal instability actually disabling them. A positive screen is a prompt to assess, not a finding.
Everything useful here is a variant of one question: what is the shape of this over time? DSM-5-TR describes the affective instability of borderline personality disorder as episodes of intense dysphoria, irritability or anxiety usually lasting a few hours and only rarely more than a few days. That is a temporal criterion, and it gives you something to ask about.
So ask about the last five episodes specifically rather than in general. For each one, work through four things:
The ADHD pattern that emerges is usually a short fuse with a short burn: anger arriving within seconds of a frustration, disproportionate in intensity but proportionate in direction, and genuinely over within the hour, often to the patient's own embarrassment. The EUPD pattern is more often relationally triggered, persists past the trigger, and resolves through reconnection rather than through time.
Ask what the person's working life looks like in a month with no relationship trouble. In ADHD the impairment is still there, the missed deadlines, the unopened post, the abandoned projects. In EUPD without co-occurring ADHD, that month can look reasonably functional.
This tests the pervasiveness requirement in NICE guideline NG87, which asks for symptoms across two or more important settings. It is also the question partners and parents answer better than patients do, which is why an informant matters more here than almost anywhere else in adult psychiatry.
Four things commonly used to arbitrate this differential do not hold up, and knowing that is more useful than any amount of additional nuance on the ones that do.
There is also no published consensus standard to lean on. The 2024 Delphi study by Sarr and colleagues in the British Journal of Psychology generated 275 consensus statements on differentiating EUPD from autism, attachment disorders and complex PTSD. ADHD was not one of them.
DSM-5-TR requires several ADHD symptoms before age 12, and NG87 requires onset in childhood with persistence through life. That is the firmest ground in this differential and the ground most adult assessments stand on least securely.
You are asking a 40-year-old woman and, if you are lucky, a 70-year-old parent to recall behaviour from the 1990s. School reports for girls with inattentive presentations are famously unhelpful. Three things improve the yield:
This is the most likely answer in a tertiary ADHD service and the one least often written down. Ditrich and colleagues report ADHD prevalence of 30 to 60% in borderline personality disorder populations, and found the comorbid group the most impaired of any studied on both impulsivity and emotion regulation.
A conclusion of both is not a diagnostic compromise. It describes a group with a worse prognosis and a clearer treatment need. Sequencing then becomes the live question, and the evidence on treating the comorbid presentation is thin enough that it remains a matter of clinical judgement. Say so in the letter to the GP who has to sign the prescription.
DSM-5-TR Criterion E requires that ADHD symptoms are not better explained by another mental disorder, and it names personality disorder among the examples. The word doing the work is explained, not accompanied. A patient can have both; what they cannot have is ADHD entirely accounted for by EUPD.
No. Simonsen and colleagues conclude that in-depth clinical evaluation of both syndromes is required. NG87 1.3.2 makes the same point for ADHD generally: a diagnosis should not rest solely on rating scale or observational data.
Yes, and it is common. Neither diagnosis excludes the other, and Criterion E requires only that ADHD is not better explained by the other condition.
Not at present. Clinicians are working from review literature and diagnostic criteria rather than a formal consensus exercise, and that is worth stating in a medico-legal report.
A report that says plainly which discriminators the literature supports and which it does not is harder to challenge than one asserting certainty across the board. It is also more useful to the next clinician who picks up the file.
Global ADHD Network's ADHD in women webinar covers the developmental history problem in the female adult presentation in more depth. Our team is at info@globaladhdnetwork.com.
