Pemphigus vulgaris

Meet the patient

A 52-year-old woman has had painful mouth ulcers for three months that her dentist treated as aphthae. Now flaccid, fluid-filled blisters are appearing on her chest and scalp, bursting within a day to leave raw, weeping erosions ringed by a ragged collar of skin. She is losing weight because eating hurts, and lateral pressure on apparently normal skin peels the epidermis away — Nikolsky positive.[1][9]

Two questions frame every pemphigus case and everything below: where in the epidermis is the split? — suprabasal, and that single fact drives the histology, the immunofluorescence and the whole differential — and which desmoglein is the antibody blocking? Hold those two and the disease falls into place.[1][2]

One antibody, one split, two faces — the desmoglein compensation theory

The whole clinical puzzle of pemphigus collapses onto a single mechanism: which desmoglein the antibody blocks. Desmoglein 3 sits in the deep epidermis and the mucosa; desmoglein 1 sits in the upper epidermis and the skin surface.[1][2]

Block Dsg3 alone and the upper-skin Dsg1 compensates, so the skin holds — you get mucosal-dominant disease, the patient whose mouth is destroyed but whose skin looks near-normal. Block both Dsg3 and Dsg1 and there is no compensation left: the skin gives way too, and you get mucocutaneous pemphigus. That is the entire theory, and examiners love it.[2]

The antibody is almost always IgG4 (pathogenic) with some IgG1, binding the cadherin extracellular domain of the desmoglein and pulling desmosomes apart so keratinocytes round up and detach — acantholysis. The split sits just above the basal layer, leaving basal cells anchored to the basement membrane like a row of tombstones: the histology signature that earns marks.[1][2]

Etymology for viva gold: pemphigus comes from the Greek pemphix, a blister; acantholysis from akantha, the prickle-cell layer, plus lysis, a loosening — the prickle cells let go of each other. Every word in the diagnosis is a description of the pathology.[2]

Meet the family — pemphigus is not one disease

The pemphigus family shares IgG- or IgA-driven acantholysis but splits on antigen, phenotype and prognosis, and recognising the variant changes the plan. The classic teaching line: an isolated oral eruption in an older patient with lymphadenopathy is paraneoplastic until proven otherwise, and a verrucous flexural plaque is vegetans, not Hailey-Hailey.[2][10]

The discriminator line: PV splits suprabasally and stains chicken-wire intercellular IgG; foliaceus splits in the uppermost granular layer and spares mucosa; paraneoplastic pemphigus adds plakin antibodies and a dual DIF pattern; IgA pemphigus is pustular with intercellular IgA.[1][13]

Drug triggers worth a named mention: thiol-containing drugs (D-penicillamine, captopril, tiopronin) and checkpoint inhibitors (pembrolizumab, nivolumab) — the latter increasingly examined. Drug-induced pemphigus often resembles foliaceus and may remit on withdrawal, though antibodies can linger for months.[2][8]

Why it happens — the immune circuit and the rituximab rationale

Pathogenic IgG4 binds Dsg3 (and often Dsg1), disrupts the desmosome, and rounds up the keratinocytes — acantholysis. The blister is therefore intraepidermal and suprabasal, fragile by design: there is no intact roof to keep it intact, so it ruptures within a day. This is why the clinical sign is a flaccid blister and an erosion, never a tense bulla.[1][2]

The same circuit is the reason rituximab works: deplete the CD20-positive B cells that produce the autoantibody and you switch off the disease at its source. That mechanistic logic, not empirical luck, is why rituximab moved from salvage to first-line.[5][6]

How common, who, and what killed them before rituximab

PV is rare but it clusters fiercely in Ashkenazi Jewish, Mediterranean, Middle Eastern and Indian or South Asian populations, driven by HLA class II alleles — HLA-DRB104:02 and DQB105:03 — that present Dsg3 peptides to autoreactive T cells.[2][3][8]

The mortality story is the viva gift. Before corticosteroids, around three-quarters died within a year — dehydration, sepsis, a lost skin barrier. Corticosteroids cut that to roughly a quarter; steroid-sparing immunosuppression to under a third; and the rituximab era has driven disease-specific mortality under 5 percent at five years. Infection has now replaced disease activity as the leading cause of death — pneumonia, bacterial sepsis, PJP under combination immunosuppression. That single pivot is why monitoring matters as much as the induction drug.[5][7][8]

Read the blister like the microscopist does

A flaccid blister that ruptures within a day, leaving a painful erosion with a collarette of detached epidermis, plus a positive Nikolsky sign, is pemphigus until disproven. The tense, robust bulla of pemphigoid it is not.[1][12]

• Skin: flaccid blisters on normal or erythematous skin across trunk, scalp, face and proximal limbs, rupturing into erosions with a peripheral collarette of detached epidermis. Nikolsky sign positive (lateral pressure shears the epidermis); Asboe-Hansen sign positive (pressure extends the blister).[1]
• Mucosa: involved in 50-70 percent, often the presenting feature and preceding skin by months — oral erosions (buccal, palate, gingival, tongue), then pharyngeal, oesophageal (dysphagia, odynophagia), conjunctival, genital, nasal. Examine every mucosal surface.[9]
• Severity: score with the Pemphigus Disease Area Index (PDAI) — 12 body areas and three activity types, total 0-250; active disease over 15, complete remission off therapy is 0 for two months.[1]

The classic trap: tense bullae on an urticarial base in an 80-year-old is bullous pemphigoid, not pemphigus — the split is subepidermal, Nikolsky is negative, and DIF is linear at the basement membrane. Confusing the two sends you down the wrong biopsy and the wrong first drug.[1][13]

Face-off — pemphigus vulgaris versus its closest mimics

The diagnostic triad — three tests, one disease

Diagnosis needs three investigations together; no single test is enough. Take the histology biopsy from a lesion and the DIF biopsy from perilesional skin — two separate sites, a recurring trainee error.[2][4]

1. Histology of a lesional biopsy — a blister cavity just above the basal layer (suprabasal acantholysis), basal cells still anchored to the basement membrane like tombstones, and rounded acantholytic keratinocytes floating free.[1][2]
2. Direct immunofluorescence of perilesional skin — the pathognomonic finding: intercellular IgG4 (and C3) deposition throughout the epidermis in a chicken-wire (fishnet) pattern. The single most specific test.[1]
3. Serology — indirect immunofluorescence on monkey oesophagus, plus anti-Dsg3 and anti-Dsg1 ELISA; titres track disease activity and guide retreatment, so they are a monitoring tool as well as a diagnostic one.[3]

A practical biomarker rule: anti-Dsg3 ELISA over 130 U/mL means active disease, and a rise over 20 U/mL in a previously seronegative patient predicts relapse within months. Use serial titres every 3-6 months to guide rituximab re-dosing.[5]

Management — steroids plus rituximab, and screen hepatitis B first

First-line is now systemic corticosteroids plus rituximab together. The days of steroid monotherapy pushed ever higher are over, because the RITUX 3 trial (NEJM 2021) proved rituximab superior to mycophenolate for complete remission off therapy, and the EADV 2020 S2K guideline made rituximab first-line.[4][5]

• Prednisolone 1-1.5 mg/kg/day (typical 40-80 mg) to induce remission, then a slow taper over months.[4]
• Rituximab depletes CD20-positive B cells and has transformed outcomes. Two protocols: the rheumatoid-arthritis protocol (1 g IV day 1 and day 15) or the lymphoma protocol (375 mg/m2 weekly for four weeks) — both effective.[5][6]
• Steroid-sparing adjuncts (mycophenolate, azathioprine) reduce steroid exposure but are inferior to rituximab for inducing remission.[4][5]

Refractory disease ladder: intravenous cyclophosphamide pulses, IVIG (2 g/kg/month) as a rapid bridge, immunoadsorption or plasmapheresis to strip antibody, and emerging agents (efgartigimod, an anti-FcRn that drops IgG; BTK inhibitors). For paraneoplastic pemphigus, treat the underlying tumour — the eruption often parallels it.[4][10][11]

Supportive care is not optional: topical steroids for residual erosions, pain control, a soft or liquid diet (nasogastric if oesophageal), and a true multidisciplinary team — dermatology, oral medicine, ophthalmology, gynaecology or urology, gastroenterology and ENT.[1]

The dosing reference — doses, routes, monitoring

How patients with pemphigus come to harm (the preventable list)

• Fatal hepatitis B reactivation after rituximab given without a screen — the textbook preventable death[4]
• Overwhelming infection under high-dose steroid plus rituximab with no PJP prophylaxis or vaccination[7]
• Malnutrition and dehydration from untreated oral and oesophageal erosions — feed early, nasogastric if needed[9]
• Missed paraneoplastic pemphigus labelled severe aphthous stomatitis while a lymphoma or thymoma goes untreated[10]
• Steroid monotherapy pushed ever higher when rituximab was the guideline first-line all along[5]
• Pregnancy on rituximab — B-cell depletion in the neonate; plan and switch[1]

Special populations

Pregnancy: pemphigus can flare; corticosteroids and azathioprine are relatively safe, but rituximab is avoided because it depletes neonatal B cells. Track anti-Dsg titres through pregnancy.[1]

Elderly and frail: prefer rituximab over escalating steroids to spare bone, glucose and infection risk, with vigilant infection surveillance — the group most likely to die of treatment, not disease.[7][8]

Paraneoplastic pemphigus: investigate and treat the underlying malignancy first; the skin disease often follows the tumour.[10][11]

Prognosis, surveillance and prevention

With rituximab, complete remission off therapy is now achievable in the majority, and anti-Dsg3 titres track activity and predict relapse — monitor every 3-6 months in active disease and 6-12 months in remission, and check immunoglobulins if rituximab is repeated.[5]

There is no primary prevention — it is autoimmune. Secondary prevention means early diagnosis, treatment to remission, infection prevention (vaccination, PJP prophylaxis) and sun protection, since UV can trigger flares. Family screening is not routine; penetrance is low.[1]

The mantra, and the memory device

The mantra: suprabasal split, chicken-wire IgG, steroids plus rituximab, screen hepatitis B, and watch for infection.[1][5]

Ward-round test — three stems, thirty seconds each