Clinical dermatology • Original article doi: 10.1111/j.1365-2230.2006.02220.

Improved protocol for treatment of pemphigus vulgaris with

protein A immunoadsorption
I. Shimanovich, S. Herzog,* E. Schmidt,* A. Opitz,† E. Klinker,† E-B. Bröcker,* M. Goebeler‡ and
D. Zillikens
Department of Dermatology, University of Lübeck, Germany; Departments of *Dermatology and †Transfusion Medicine, University of Würzburg, Germany;
and ‡Department of Dermatology, University Hospital Mannheim, University of Heidelberg, Germany

Summary Background. Pemphigus vulgaris is a life-threatening autoimmune blistering skin

disease, usually treated with high-dose corticosteroids in combination with other
immunosuppressants. However, this regimen may prove inadequate in severe cases
and can cause dangerous side-effects. We have recently reported protein A
immunoadsorption (PAIA) to be an effective adjuvant treatment for induction of
remission in severe pemphigus. However, in a significant number of cases, the disease
rapidly recurred once PAIA and immunosuppressive medication were tapered.
Aims. The aim of the present study was to develop a PAIA-based therapeutic regimen
that would result in a more prolonged remission of pemphigus.
Methods. Nine patients with pemphigus vulgaris were treated with a modified pro-
tocol characterized by a combination of PAIA with a higher initial dose of systemic
methylprednisolone (2 mg ⁄ kg). In addition, azathioprine or mycophenolate mofetil
was administered as a steroid-sparing agent.
Results. In all nine patients treated with this regimen, we observed a sharp decline of
circulating autoantibody levels and dramatic improvement of cutaneous and mucosal
lesions within 4 weeks of therapy. The patients remained free of clinical disease for up
to 26 months after PAIA treatment was discontinued.
Conclusion. The improved treatment protocol appears to combine highly effective
induction of clinical remission in severe or treatment-resistant pemphigus with a
prolonged subsequent symptom-free interval.

acetylcholine receptor, pemphaxin).1 The pathogenic

Introduction
Pemphigus is a group of severe autoimmune blistering
skin diseases characterized by circulating IgG autoan-
tibodies to a wide range of keratinocyte proteins,
including desmosomal constituents (desmogleins 1
and 3, plakoglobin) and acetylcholine receptors (a-9
Correspondence: Dr Detlef Zillikens, Department of Dermatolgy, University
of Lübeck, Ratzeburger Allee 160, 23538 Lübeck, Germany.
E-mail: detlef.zillikens@uk-sh.de
The first two authors contributed equally to this work.
Conflict of interest: none declared.
Accepted for publication 17 May 2006
relevance of these antibodies has been proven by
induction of pemphigus in mice and human neonates
passively transfused with IgG from pemphigus pa-
tients.2–5 While the exact role of the different antibody
subpopulations for induction of acantholysis in vivo
remains controversial,6,7 one possible hypothesis sug-
gests that functional disruption of both adhesion and
regulatory keratinocyte proteins is required for this
process.7 Commonly, levels of circulating autoantibod-
ies correlate with activity of the disease in patients with
pemphigus.8
Pemphigus is a potentially fatal condition, and a
significant part of the mortality is related to side-effects
of corticosteroids and other immunosuppressants,9

 2006 The Author(s)

768 Journal compilation  2006 Blackwell Publishing Ltd • Clinical and Experimental Dermatology, 31, 768–774
 Improved protocol for treatment of pemphigus vulgaris with protein A immunoadsorption • I. Shimanovich et al.
therefore there is a continued need to search for more
specific treatments for the condition. Protein A immu-
noadsorption (PAIA) is a therapeutic apheresis tech-
nique that selectively removes IgG from systemic
circulation. Recently, we have demonstrated that PAIA
is a highly effective and safe adjuvant treatment option
in severe and ⁄ or recalcitrant pemphigus.10 In this
report, we present an improved protocol used for
treatment of nine further patients with pemphigus
vulgaris.
Materials and methods
Patients
Our study included 9 patients with pemphigus vulgaris
aged between 45 and 69 years. Eight patients were
female and one was male. All patients had an active
form of the disease with multiple blisters and ⁄ or
erosions and revealed intercellular epidermal deposits
of IgG and C3 by direct immunofluorescence microscopy
of perilesional biopsies. By indirect immunofluorescence
microscopy, circulating autoantibodies against intercel-
lular substance of monkey oesophagus epithelium were
detected at titres between 320 and greater than 1280.
Patients’ disease activity was scored as shown in
Table 1. Seven patients had been refractory to various
past immunosuppressive treatments administered over
a period of 11–42 months, while in 2 patients PAIA
was used as the first-line therapy. Patients 8 and 9
represented relapsed cases previously treated according
to the PAIA protocol described by Schmidt et al.10 The
study was approved by the local ethics committee and
informed consent was obtained from all patients.
Desmoglein ELISA
Circulating antibodies to desmoglein 1 and 3 were
detected by ELISA according to the manufacturer’s
instructions (Medical and Biological Laboratories,
Table 1 Pemphigus activity score.
Score Disease activity
4 More than 10 fresh lesions within the preceding week
3 1–10 fresh lesions within the preceding week
2 No fresh lesions but further PAIA and immunosupression
required
1 No fresh lesions, PAIA discontinued but further
immunosuppression required
0 No lesions and no further PAIA or immunosuppression
required
 2006 The Author(s)
Journal compilation  2006 Blackwell Publishing Ltd • Clinical and Experimental Dermatology, 31, 768–774
Nagoya, Japan).8 The cut-off value was set at
14 U ⁄ mL for desmoglein 1-specific antibodies and at
7 U ⁄ mL for desmoglein 3-specific antibodies.
Treatment protocol
PAIA was performed as described previously.10 Briefly,
blood was drawn from an antecubital vein and plasma
was obtained using an automated plasma separation
device (Cobe Spectra, Gambro BCT, Lakewood, USA).
Plasma was adsorbed using a twin protein A column
(Immunosorba, Fresenius Hemocare, St. Wendel, Ger-
many) at a flow rate of 20–35 mL ⁄ min. In each
treatment cycle, 1.5–2.5 plasma volumes were processed.
The treatment protocol consisted of an induction and
a maintenance phase. In the induction phase, PAIA was
performed on three consecutive days (days 1–3). On day
1, oral methylprednisolone (2 mg ⁄ kg) and azathioprine
were started. The dose of azathioprine ranged from 50
to 250 mg ⁄ day, and was adjusted to the activity of
thiopurine S-methyltransferase as recommended by
Anstey et al.11 In cases with contraindications for
azathioprine, mycophenolate mofetil was given at a
dose of 2 g ⁄ day (Table 2). Starting from day 8, PAIA
was performed four times each at weekly, 2-weekly, 3-
weekly and 4-weekly intervals (a total of 19 treatments
over 41 weeks) and then discontinued (maintenance
phase). If no fresh lesions developed, oral methylpredn-
isolone was reduced by 20–25% every 2 weeks until a
dose of 40 mg ⁄ day was reached. Subsequently, meth-
ylprednisolone was tapered to 32, 24, 20, 16 and
12 mg ⁄ day with 2-week intervals between each reduc-
tion, then to 10, 8 and 6 mg ⁄ day with 3-week
intervals, and finally to 4 and 2 mg ⁄ day with 4 week
intervals. After methylprednisolone was discontinued,
the azathioprine (or mycophenolate mofetil) dose was
maintained unchanged for 1 month and then also
tapered by 20–25% every 4 weeks. If a patient devel-
oped < 10 new lesions within 1 week, immunosuppres-
sive medication was not reduced any further until
lesions had healed and no fresh blisters had appeared for
1 week. If a patient developed > 10 fresh lesions within
the previous week, a new induction phase was started.
In addition to this regimen, patient 5 received intrave-
nous IgG (2 g ⁄ kg, every 4 weeks) starting with cycle 10
of PAIA.
Results
Within 4 weeks of PAIA treatment, an excellent
clinical response was observed in all patients (Figs 1
and 2). In four patients, new lesion formation stopped
769
Improved protocol for treatment of pemphigus vulgaris with protein A immunoadsorption • I. Shimanovich et al.
Table 2 Treatment protocol.
Time PAIA
Day 1 (Week 1) Cycle 1
Day 2 Cycle 2
Day 3 Cycle 3
Day 8 (Week 2) Cycle 4
Weeks 3, 4, 5 (1-week intervals) Cycles 5–7
Weeks 7, 9, 11, 13 (2-week intervals) Cycles 8–11
Weeks 16, 19, 22, 25 (3-week intervals) Cycles 12–15
Week 29, 33, 37, 41 (4-week intervals) Cycles 16–19
*If TPMT activity is normal; †if azathioprine is contraindicated.
completely, and in five patients it was greatly
decreased (Table 3). A single PAIA cycle reduced the
autoantibody levels by an average of 75%. After
4 weeks and seven cycles of PAIA, desmoglein-specific
autoantibodies were decreased by an average factor of
33 (Fig. 3). Within 6 months of PAIA, eight of nine
patients were completely free of clinical disease. The
remaining patient developed perforating diverticulitis
complicated by postoperative wound dehiscence
during week 6 of treatment, and PAIA had to be
discontinued due to the poor general condition.
Twelve months after the first PAIA cycle, remission
was maintained in seven of the remaining eight
patients. In one patient, colon carcinoma was diagno-
sed 9 months after initiation of PAIA, and immuno-
suppressive therapy had to be interrupted for
3 months to allow surgery.
770 Journal compilation  2006 Blackwell Publishing Ltd • Clinical and Experimental Dermatology, 31, 768–774
Medication
Start methylprednisolone 2 mg ⁄ kg plus azathioprine (2.5 mg ⁄ kg)*
or mycophenolate mofetil (2 g ⁄ day)†
If no new blisters occur, cut methylprednisolone dose by 20-25% every
2 weeks, until adose of 40 mg ⁄ day is reached. Subsequently, reduce
methylprednisolone every second week to 32, 24, 20, 16, 12 mg ⁄ day,
respectively, then taper to 10, 8 and 6 mg ⁄ day every 3 weeks and finally to
4 and 2 mg ⁄ day every 4 weeks. After methylprednisolone is discontinued,
maintain the azathioprine dose for 1 month and then taper by 20-25%
every 4 weeks
Figure 1 Extensive skin lesions on the
back of patient 7 (a) before and (b) after
6 weeks (eight cycles) of PAIA.
At present, our patients have been followed up for
18–38 months (Table 3). Patient 1 remains in complete
clinical and serological remission, i.e. she is free of
clinical disease, has no detectable desmoglein-specific
autoantibdodies and does not require any maintenance
PAIA or immunosuppression. Two patients (2 and 5)
developed a clinical recurrence (oral erosions) after
PAIA was discontinued for 4 and 12 months, respect-
ively. Four further patients (3, 4, 6 and 7) have been
weaned off PAIA for 7–26 months. Patients 3, 6 and 7
do not show any pemphigus lesions but still have
circulating desmoglein-specific autoantibodies and re-
quire low-dose immunosuppressive medication for
remission maintenance. Patient 4 died of causes unre-
lated to pemphigus 18 months after discontinuation of
PAIA. At this point, she revealed the same disease
profile as patients 3, 6 and 7. In the remaining two
 2006 The Author(s)
 Improved protocol for treatment of pemphigus vulgaris with protein A immunoadsorption • I. Shimanovich et al.
(a)
(b)
Figure 2 Oral erosions of patient 6 (a) before and (b) after
4 weeks (six cycles) of PAIA.
patients (8 and 9), full disease control could not be
achieved due to the aforementioned interruptions of the
PAIA treatment protocol (acute diverticulitis and colon
carcinoma).
Discussion
Immunoapheresis has been used for the treatment of
different antibody-mediated autoimmune diseases,
including inhibitor-associated haemophilia, Goodpas-
ture’s syndrome, myasthenia gravis, and rheumatoid
arthritis. Pemphigus is a group of organ-specific auto-
immune disorders mediated by circulating IgG auto-
antibodies and represents an obvious choice for
immunoadsorption therapy.10,12–14
In 1998, Schoen et al. used sheep antihuman IgG
sepharose to induce a rapid clinical remission in a
patient with severe treatment-resistant paraneoplastic
pemphigus. This remission was anticipated to be
 2006 The Author(s)
Journal compilation  2006 Blackwell Publishing Ltd • Clinical and Experimental Dermatology, 31, 768–774
permanent, as the underlying benign myofibroblastic
tumour was successfully excised.12
In the protocol published by Lüftl et al., seven patients
with pemphigus vulgaris and two with pemphigus
foliaceus were treated with two tryptophan immunoad-
sorption cycles within 3 days. Each cycle was followed
by an intravenous prednisolone pulse (1.5 mg ⁄ kg) and,
starting from day 4, patients received 0.5 mg ⁄ kg oral
methylprednisolone in combination with an immuno-
suppressant. Each tryptophan immunoadsorption cycle
reduced desmoglein-specific autoantibody levels by
about 30%. All patients demonstrated a rapid improve-
ment of skin and mucous membrane lesions, and the
corticosteroid dosage could be quickly reduced. While a
stable disease remission was observed at a 6-month
follow-up, no long-term data on these patients were
presented.13
We previously reported four patients with pemphigus
vulgaris and one patient with pemphigus foliaceus
treated with PAIA.10 PAIA treatments were performed
according to the same schedule as in the present study;
however, in contrast to the current protocol, the
induction phase was followed by a single intravenous
cyclophosphamide ⁄ dexamethasone pulse and, starting
from day 8, oral methylprednisolone at a dose of
0.5 mg ⁄ kg was administered. All five patients demon-
strated an excellent response within 4 weeks of treat-
ment, and were completely free of clinical disease at a
6-month follow-up. At the time of publication in 2003,
one patient had achieved a complete clinical and
serological remission, i.e. he was free of clinical disease,
had no detectable desmoglein-specific autoantibodies
and did not need any further immunosuppression. The
remaining four cases were also completely free of
pemphigus lesions but still demonstrated circulating
desmoglein-specific antibodies and required immuno-
suppressive treatment. These patients have since been
followed up for 2–4.5 years and provide valuable
insights into the long-term effectiveness of PAIA.
At the present time, only one of the five patients
reported by Schmidt et al. remains in complete clinical
and serological remission. One further patient is free
of clinical disease but demonstrates circulating de-
smoglein-specific antibodies and keeps receiving low-
dose immunosuppressive medication. Three remaining
patients experienced a clinical recurrence once PAIA
was discontinued and systemic methylprednisolone
was tapered. These data suggest that while PAIA is
highly effective in inducing a rapid clinical remission
in severe pemphigus, it does not contribute to
achieving a permanent cure in the majority of
patients.
771
Improved protocol for treatment of pemphigus vulgaris with protein A immunoadsorption • I. Shimanovich et al.

Table 3 Disease activity, autoantibody levels and medications before, during and after PAIA.

Patient number

Treatment progression 1 2 3 4** 5†† 6 7 8‡‡ 9§§

Before PAIA
Disease activity score* 4 4 4 4 4 3 4 4 4
Methylprednisolone†,§ 100 150 140 120 200 200 120 140 140
Immunosuppressant‡,§ 2 (M) 50 (A) 50 (A) 200 (A) 2 (M) 250 (A) 150 (A) 2 (M) 150 (A)
Desmoglein 1 ELISA– 240 0 60 0 30 0 280 0 0
Desmoglein 3 ELISA– 1620 1020 930 1555 450 432 685 3260 8376
1 month after first PAIA
Disease activity score 3 2 2 2 2 3 3 3 3
No. of PAIA cycles 14 7 6 7 6 7 7 6 7
Methylprednisolone 80 80 80 80 200 150 104 70 70
Immunosuppressant 2 (M) 50 (A) 50 (A) 200 (A) 2 (M) 200 (A) 150 (A) 2 (M) 150 (A)
Desmoglein 1 ELISA 35 0 0 0 0 0 20 0 0
Desmoglein 3 ELISA 224 274 5 60 334 36 24 71 288
6 months after first PAIA
Disease activity score 2 2 2 1 2 2 2 – 3
No. of PAIA cycles 34 16 16 15 13 16 13 – 9
Methylprednisolone 16 10 10 10 30 12 10 – 90
Immunosuppressant 2 (M) 75 (A) 50 (A) 200 (A) 2 (M) 200 (A) 150 (A) – 175 (A)
Desmoglein 1 ELISA 0 0 0 0 0 0 1⁄ – 0
Desmoglein 3 ELISA 0 85 0 40 53 34 3 – 97
12 months after first PAIA
Disease activity score 1 1 1 1 1 1 1 – 3
No. of PAIA cycles 41 16 20 16 13 20 19 – 10
Methylprednisolone 4 8 2 4 10 4 6 – 8
Immunosuppressant 2 (M) 75 (A) 50 (A) 200 (A) 2 (M) 200 (A) 150 (A) – 175 (A)
Desmoglein 1 ELISA 0 0 0 0 0 0 38 – 0
Desmoglein 3 ELISA 0 730 84 78 104 80 32 – 267
At present
Disease activity score 0 3 1 – 1 1 1 – 1
No. of PAIA cycles 40 16 20 – 13 20 19 – 10
Months after 1st PAIA 38 33 32 – 20 19 18 – 19
Months after last PAIA 26 26 24 – 13 11 7 – 10
Methylprednisolone 0 4 0 – 10 24 6 – 8
Immunosuppressant 0 0 50 (A) – 2 (M) 200 (A) 150 (A) – 175 (A)
Desmoglein 1 ELISA 0 0 0 – 0 0 0 – 0
Desmoglein 3 ELISA 0 1128 388 – 55 272 223 – 183

*Disease activity was scored according to Table 1; †oral methylprednisolone, mg ⁄ day; ‡(M), oral mycophenolate mofetil, g ⁄ day, (A), oral
azathioprine, mg ⁄ day; §starting dose of methylprednisolone ⁄ immunosuppressant; –serum levels of circulating autoantibodies as detected
by ELISA with recombinant desmoglein 1 or 3, U ⁄ mL; **this patient died of causes unrelated to pemphigus 2 years after initiation of
therapy; ††starting from week 6, this patient received intravenous immunoglobulin (2 g ⁄ kg every 4 weeks) in addition to regular protocol
treatment; ‡‡in this patient, PAIA had to be discontinued at week 6 due to perforating diverticulitis; §§this patient received no immu-
nosuppression between months 9 and 12 due to surgery for colon carcinoma.

In the present study, we aimed at inducing a
prolonged disease improvement using a modified PAIA
protocol with an increased dose of concurrent immuno-
suppressants. In contrast to the protocol of Schmidt
et al., the initial cyclophosphamide ⁄ dexamethasone
pulse was omitted, and patients were started on
2 mg ⁄ kg methylprednisolone in combination with az-
athioprine or mycophenolate mofetil from the first day
of therapy. All patients treated according to the present
protocol showed a dramatic improvement of their
pemphigus lesions within the first 4 weeks of treatment.
The concurrent administration of PAIA and high-dose
systemic immunosuppressants makes it impossible to
assess the exact contribution of PAIA to our patients’
improvement. However, our previous data, demonstra-
ting that PAIA in conjunction with a much lower dose
of corticosteroids is also effective in inducing a rapid
remission of pemphigus,10 along with the fact that most
patients included in this study and in the study of
Schmidt et al. were resistant to past high-dose

 2006 The Author(s)

772 Journal compilation  2006 Blackwell Publishing Ltd • Clinical and Experimental Dermatology, 31, 768–774
 Improved protocol for treatment of pemphigus vulgaris with protein A immunoadsorption • I. Shimanovich et al.
Figure 3 Levels of circulating desmoglein
3 (dsg3)-specific autoantibodies in patients
6 (triangles, solid curve) and 7 (circles,
interrupted curve) as determined by
ELISA. (a) Desmoglein 3-specific antibody
levels immediately before and after each of
the first 10 PAIA cycles. (b) Long-term
development of autoimmune reactivity
under treatment protocol described in
Materials and Methods. Wks, time in
weeks.
immunosuppressive regimens, suggest that PAIA did
contribute to the prompt clinical improvement observed
in the current group of patients. In contrast to our
previous investigation,10 only two of seven patients of
the present study remaining on the treatment schedule
demonstrated a clinical recurrence during a long-term
follow-up of 18–38 months. Therefore, the modified
protocol of PAIA appears to be more effective in
maintaining long-term clinical remission in patients
with pemphigus. This longer-lasting effect of the treat-
ment regimen might be due to the higher cumulative
dose and the more prolonged administration of systemic
corticosteroids compared with our previous study.
We found a single PAIA cycle removed about 75% of
desmoglein-specific autoantibodies from patients’ circu-
lation. These results are in good agreement with
elimination rates reported by Schmidt et al. (approx.
80%),10 and confirm that PAIA is greatly superior to
nonselective tryptophan adsorbers in reducing serum
levels of desmoglein-specific autoantibodies.13 The levels
 2006 The Author(s)
Journal compilation  2006 Blackwell Publishing Ltd • Clinical and Experimental Dermatology, 31, 768–774
of anticholinergic receptor antibodies have not been
monitored in our patients due to the lack of a
commercially available assay system. However, based
on IgG-binding properties of the protein A matrix, these
antibodies are expected to decrease at a rate similar to
that of desmoglein-specific antibodies.
In general, PAIA was well tolerated, confirming that
it is a safe treatment modality. One of our patients
developed perforating diverticulitis and one patient was
diagnosed with colon carcinoma while on PAIA treat-
ment. However, it is highly unlikely that these adverse
events are related to PAIA therapy. One further patient
died of causes unrelated to pemphigus 18 months after
discontinuation of PAIA.
In conclusion, we describe an optimized protocol for
treatment of pemphigus with PAIA, combining high
effectiveness in induction of rapid clinical remission with
a lower recurrence rate during the long-term follow-up
period. Eventually, a randomized controlled trial should
confirm these findings on a larger group of patients.
773
Improved protocol for treatment of pemphigus vulgaris with protein A immunoadsorption • I. Shimanovich et al.
References
1 Grando SA. Autoimmunity to keratinocyte acetylcholine
receptors in pemphigus. Dermatology 2000; 201: 290–5.
2 Moncada B, Kettelsen S, Hernandez-Moctezuma JL,
Ramirez F. Neonatal pemphigus vulgaris. role of passively
transferred pemphigus antibodies. Br J Dermatol 1982;
106: 465–7.
3 Walker DC, Kolar KA, Hebert AA, Jordon RE. Neonatal
pemphigus foliaceus. Arch Dermatol 1995; 131: 1308–11.
4 Anhalt GJ, Labib RS, Voorhees JJ, Beals TF, Diaz LA.
Induction of pemphigus in neonatal mice by passive
transfer of IgG from patients with the disease. N Engl J Med
1982; 306: 1189–96.
5 Rock B, Martins CR, Theofilopoulos AN et al. The pathogenic
effect of IgG4 autoantibodies in endemic pemphigus folia-
ceus (fogo selvagem). N Engl J Med 1989; 320: 1463–9.
6 Stanley JR, Nishikawa T, Diaz LA, Amagai M. Pemphigus:
is there another half of the story?. J Invest Dermatol 2001;
116: 489–90.
7 Grando SA, Pittelkow MR, Shultz LD, Dmochowski M,
Nguyen VT. Pemphigus: an unfolding story. J Invest
Dermatol 2001; 117: 990–5.
8 Ishii K, Amagai M, Hall RP et al. Characterization of
autoantibodies in pemphigus using antigen-specific
774 Journal compilation  2006 Blackwell Publishing Ltd • Clinical and Experimental Dermatology, 31, 768–774
enzyme-linked immunosorbent assays with baculovirus-
expressed recombinant desmogleins. J Immunol 1997; 159:
2010–17.
9 Ahmed AR, Moy R. Death in pemphigus. J Am Acad
Dermatol 1982; 7: 221–8.
10 Schmidt E, Klinker E, Opitz A et al. Protein A immuno-
adsorption: a novel and effective adjuvant treatment of
severe pemphigus. Br J Dermatol 2003; 148: 1222–9.
11 Anstey AV, Wakelin S, Reynolds NJ. Guidelines for pre-
scribing azathioprine in dermatology. Br J Dermatol 2004;
151: 1123–32.
12 Schoen H, Foedinger D, Derfler K et al. Immunoapheresis in
paraneoplastic pemphigus. Arch Dermatol 1998; 134:
706–10.
13 Lüftl M, Stauber A, Mainka A, Klingel R, Schuler G, Hertl
M. Successful removal of pathogenic autoantibodies in
pemphigus by immunoadsorption with tryptophan-linked
polyvinylalcohol adsorber. Br J Dermatol 2003; 149: 598–
605.
14 Frost F, Messer G, Fierlbeck G, Risler T, Lytton S. Treatment
of pemphigus vulgaris with protein A immunoadsorption.
Case report of long-term history showing favorable out-
come. Ann N Y Acad Sci 2005; 1051: 591–6.
 2006 The Author(s)