New Psychosis on High-Dose Steroids: Is This Neuropsychiatric Lupus, or the Steroids Themselves
She's psychotic, seizing, and already on sixty milligrams of prednisone — which means the same drug meant to be treating her disease might also be causing part of what's happening to her. The disagreement is what to do before anyone actually knows which is true.
A.F., a 24-year-old graduate student in her final year of a doctoral program, was admitted three days ago with new-onset auditory hallucinations and a single generalized tonic-clonic seizure, on a background of SLE diagnosed eighteen months ago with prior mucocutaneous and articular involvement only. She had been started on prednisone 60mg daily nine days before admission for a presumed disease flare, and her psychiatric symptoms began on day six — inside the first two weeks, when most corticosteroid-induced psychiatric reactions declare themselves. Her dose places her in the middle band of the Boston Collaborative Drug Surveillance Program's dose-incidence data, where 41 to 80mg daily carried a 4.6 percent rate of psychiatric disturbance against 1.3 percent at 40mg or less and 18.4 percent above 80 — elevated enough to take seriously, and well short of the near-one-in-five risk the phrase “high-dose steroids” tends to summon. Lumbar puncture showed no pleocytosis and negative infectious studies; brain MRI was nonspecific, showing no discrete lesion. Her anti-dsDNA is markedly elevated and complement is low, both consistent with active systemic disease, though neither is a specific marker for neuropsychiatric involvement itself. Her mother, at the bedside since admission, describes her as having been “not herself” for closer to two weeks — quieter, sleeping poorly, withdrawn from her usual calls home — which predates the seizure and complicates a clean read of exactly when today's presentation actually began, since a slower prodrome sits less comfortably with a purely medication-timed explanation than the seizure's own sudden onset does.
The diagnostic problem is not academic: the two leading explanations point toward opposite treatments. If this is inflammatory neuropsychiatric lupus, guideline-recommended management is high-dose glucocorticoids plus cyclophosphamide or a biologic — more immunosuppression, on top of what she's already receiving. If this is corticosteroid-induced psychosis, the correct response is a cautious steroid taper and an antipsychotic, and adding cyclophosphamide would mean subjecting a 24-year-old to a gonadotoxic drug for a diagnosis she may not actually have. Her active serology argues for real systemic disease activity somewhere; it does not, on its own, prove the brain is where that activity is happening this week, and no single test in front of the group today can settle which explanation is correct.
Inpatient consult, day three of new psychiatric symptoms
Her anti-dsDNA is markedly elevated and her complement is low — real, active systemic disease, in a patient who now has new seizures and hallucinations. That combination meets a reasonable threshold for presumed neuropsychiatric lupus, and delaying real immunosuppression while we wait for more certainty risks the kind of neurologic injury that doesn't reverse. I'd start cyclophosphamide alongside the steroids she's already on.
I want to name the timeline plainly: her symptoms started on day six of a new 60-milligram prednisone course, inside the first two weeks, which is where most of these declare themselves. The Boston Collaborative data puts her dose band at 4.6 percent — real, and worth saying out loud, well below the 18 percent above 80 milligrams that people tend to picture when they hear “high-dose steroids.” That is not an incidental detail either way. Her serology tells us her disease is active somewhere — it doesn't tell us her brain is where that activity is happening this week.
I'd try a cautious steroid taper with an antipsychotic on board before committing her to cyclophosphamide, which is gonadotoxic in a 24-year-old who hasn't even had the fertility-preservation conversation yet.
You're both reasoning from real signals that point in opposite directions, and I don't think either of you is wrong about the evidence you're citing — I think the honest answer is that neither test in front of us actually settles which explanation is correct this week.
I'd rather not force a binary choice between full cyclophosphamide induction and a steroid taper alone. Rituximab treats presumed autoimmune CNS activity without cyclophosphamide's gonadotoxicity, and buys us time before committing to cyclophosphamide. I want to be honest about the cost, because it isn't nothing: if we taper the steroid, start rituximab and start an antipsychotic inside the same twenty-four hours, then whatever happens next, none of us will know which of the three did it. We are buying safety by spending the cleanest diagnostic test available to us. I still think that's the right trade in a patient who has already seized once — but nobody should read her improving as proof the psychiatrist was right.
Agreed: start rituximab induction, taper prednisone cautiously rather than continuing at 60mg, and add low-dose risperidone for symptomatic control. Cyclophosphamide is held in explicit reserve, not started today.
that is suggestive of the psychiatrist's read but cannot confirm it, since rituximab and risperidone were started in the same window and either could account for the improvement. The group accepted losing that clarity as the price of not leaving a seizing patient untreated, and named the loss rather than planning to read the outcome as an answer.
the rheumatologist's original concern was correct, and cyclophosphamide moves from held-in-reserve to actively indicated, with fertility preservation discussed before it starts.
Genuinely unresolved, and named as such rather than smoothed over: whether presumed NPSLE ever met threshold for treatment today at all, or whether the group treated diagnostic uncertainty as license to add a biologic that may turn out to have been unnecessary. No voice claimed the other's read was wrong — only that the evidence in front of them this week couldn't decide it.