Erythroderma at 90% BSA: admit, investigate, and decide about the drug

Decision: admit. Erythroderma is a systemic illness, and the numbers here already show decompensation, not just skin disease: pulse 104 with BP 98/60 against a usual 145/85, 4 kg of weight loss, and albumin 25 g/L. High-output cardiac failure, hypothermia, dehydration and sepsis are the ways these patients die, and he lives alone with his nearest relative two hours away.

Admission triggers present: haemodynamic compromise in an elderly man with a 20-year history of psoriasis (a group prone to cardiac decompensation from the vasodilated skin bed), hypoalbuminaemia/oedema risk, temperature instability risk despite a normal reading today, and no reliable home supervision.

First 24 hours:

• FBC with differential (eosinophilia supports drug reaction; look for atypical lymphocytes/Sézary cells)
• U&E, creatinine, LFTs, albumin, glucose, CRP, and a clotting screen if any bleeding from excoriations - LFT/renal derangement plus eosinophilia would make DRESS a live diagnosis
• Two sets of blood cultures and a skin swab (including MRSA) if febrile or crusted - infection is a common terminal event, and erythroderma also causes a staphylococcal colonisation load
• Two 4 mm punch biopsies: one in formalin for H&E, one for direct immunofluorescence if there is any blistering or mucosal involvement (to exclude pemphigus/paraneoplastic pemphigus)
• HIV and syphilis serology; blood film and, if no cause emerges, peripheral blood flow cytometry for a clonal T-cell population (Sézary)
• Chest X-ray and ECG (cardiac strain, coexisting malignancy, fluid status)
• Strict fluid balance, daily weight, four-times-daily temperature, and pressure-area checks

What I would not do reflexively is order a full malignancy screen on day one; the history and timeline matter more than a scattergun tumour hunt.

Stop allopurinol today and document it as a suspected drug reaction. It is the only new drug, started five weeks ago, which sits squarely in the latency window for both a simple morbilliform/exanthematous eruption and DRESS (typically 2–8 weeks). Gout is not an emergency; withholding urate-lowering therapy for weeks costs nothing, and restarting would be dangerous.

Before leaving the room, look for DRESS: facial oedema, lymphadenopathy, mucosal involvement, tender skin or a liver edge - none may be mentioned in the referral but all change management (a DRESS diagnosis justifies a systemic steroid taper and close hepatic/renal monitoring).

Antihypertensives - review, do not blindly stop. He is already hypotensive on ramipril and atenolol; sudden withdrawal of atenolol risks rebound tachycardia, while continuing both may worsen the low BP. My approach: continue for now with close monitoring and involve the medical team the same day to rationalise the regimen, rather than making a solo decision in the dermatology clinic.

Continue the atorvastatin - no temporal link and no competing risk.

Do not start empiric oral prednisolone unless DRESS or a clear drug hypersensitivity emerges, because he has psoriasis and systemic corticosteroids can precipitate rebound erythroderma or generalised pustular psoriasis. That is the judgement call: the biopsy will be non-specific, so the decision rests on the drug timeline, the eosinophil count, and whether systemic features of DRESS are present.

I neither agree nor refuse automatically - it depends on whether he has DRESS, not on the histology. Spongiotic dermatitis with eosinophils is exactly what a drug eruption looks like, but erythrodermic psoriasis can also show spongiosis and eosinophils, and psoriasis itself can show eosinophils after topical treatment. The histology cannot arbitrate.

Arguments for systemic steroids: eosinophils 0.9, a new drug in the right window, and a biopsy consistent with a drug eruption. If he has DRESS, delayed steroid treatment is associated with worse outcomes.

Arguments against: an untreated or misdiagnosed psoriatic erythroderma on prednisolone can rebound into pustular psoriasis, and steroids add infection and fluid-retention risk in a hypoalbuminaemic, borderline-hypotensive 74-year-old.

Practical plan:

1. Examine again for DRESS features and check LFTs, creatinine and eosinophil trend; if DRESS is present, treat with prednisolone ~0.5–1 mg/kg with a slow taper, plus gastric and bone protection, and stop allopurinol permanently.
2. If there is no DRESS, manage as erythroderma: liberal emollients (500 g/week), a low-potency topical corticosteroid to the trunk and limbs rather than nothing (to control inflammation and itch without large systemic absorption), wet-wrap or occlusion only on limbs, antipruritics, and treatment of any staphylococcal infection.
3. Hold off methotrexate, acitretin or ciclosporin until the acute phase settles and the diagnosis is clarified - starting an immunosuppressant during a possible drug reaction obscures the picture.
4. Reassess at 48–72 hours: falling eosinophils and fading erythema support a drug cause; persistence or worsening with a rising CRP demands a search for sepsis and reconsideration of Sézary syndrome or a psoriasis flare.
5. Involve cardiology/medicine early about the BP; fluid overload in an elderly erythrodermic is easier to cause than to treat.