British Association of Dermatologists' Guidelines for the Management of Pemphigus Vulgaris 2017

K.E. Harman; D. Brown; L.S. Exton; R.W. Groves; P.J. Hampton; M.F. Mohd Mustapa; J.F. Setterfield; P.D. Yesudian


The British Journal of Dermatology. 2017;177(5):1170-1201. 

In This Article

Adjuvant Drugs

Azathioprine (Strength of recommendation B, Level of evidence 1+)

Introduction. Azathioprine is a commonly prescribed adjuvant drug in PV and was first used successfully in 1969 by Krakowski et al.[88] Numerous small case series have reported a corticosteroid-sparing effect.[89–92] The complete remission rates of 28–45% exceed those seen in historical controls treated with corticosteroids alone.[9,59,91] Mortality rates of 1·4–7% are lower than those seen in historical controls treated with corticosteroids alone.[9,10,59,91]

Azathioprine as a single agent. In three cases, azathioprine was used successfully as a monotherapy to induce and maintain clinical remission with a fall in antibody titre.[90,93] However, there is a latent period of at least 6 weeks before the effects of azathioprine are seen,[89–91,93] and its use as a monotherapy to induce remission should therefore be reserved for mild cases only, if a delay in achieving disease control can be tolerated.

Comparison with oral corticosteroids. The role of azathioprine as a corticosteroid-sparing drug has been demonstrated in a number of small studies. Chaidemenos et al. compared high-dose prednisolone (1·5 mg kg−1 per day; n = 17) with low-dose prednisolone (40 mg on alternate days) plus azathioprine 100 mg per day (n = 19) in a retrospective comparison.[60] Both regimens were effective. Analysis of the 30 responders showed that the high-dose prednisolone group achieved a faster remission with greater side-effects. The combination group had a significantly lower total prednisolone usage but a longer time until complete or partial remission. This was not an intention-to-treat analysis and the study was not powered adequately. In a retrospective study, described in section 14.4.2, the times to remission and complete remission off-treatment showed no significant differences when adding azathioprine (100 mg per day) to prednisolone (100 mg per day starting dose).[94]

In a large, unblinded RCT described in section 18·1, patients randomized to the prednisolone plus azathioprine arm (n = 30) had required lower cumulative corticosteroid doses at 1 year than those treated with prednisolone alone (n = 30), although efficacy was similar in these two arms.[53] In a subsequent study, the same authors performed a double-blind RCT comparing prednisolone (initial dose 2 mg kg−1 per day; n = 28) plus placebo with prednisolone plus azathioprine (2·5 mg kg−1 per day; n = 28) over 12 months.[95] Disease severity was measured using the Pemphigus Vulgaris Disease Activity Index (PVDAI) and included an intention-to-treat analysis. No significant differences were seen in the mean PVDAI scores or the corticosteroid doses between the two groups over the 12 months. However, subgroup analyses revealed differences in the two arms towards the end of the trial: in the final 3 months there were significant differences in the PVDAI and mean daily and cumulative prednisolone doses, favouring prednisolone plus azathioprine. The mean PVDAI of the prednisolone-only group was 2·41 and for prednisolone plus azathioprine it was 0·47 (P = 0·045, intention to treat).

Comparison with other adjuvant drugs. Trials comparing azathioprine with mycophenolate mofetil and cyclophosphamide are described in section 18. In summary, two trials have compared azathioprine with mycophenolate mofetil[53,71] and there is evidence to suggest that azathioprine has a superior corticosteroid-sparing effect.[52] There is also some evidence that azathioprine may be less effective at achieving disease control.[52,71] One retrospective study suggested that azathioprine might be less effective than oral cyclophosphamide.[94] Three trials have compared azathioprine with pulsed cyclophosphamide regimens: one RCT showed no significant differences;[70] a nonrandomized trial favoured pulsed cyclophosphamide, which showed a lower cumulative corticosteroid dose although efficacy was similar;[72] and a single-centre RCT showed lower cumulative corticosteroid doses in the azathioprine arm compared with the pulsed cyclophosphamide arm, which did not reach significance in the authors' analysis[53] but was considered significant in an independent Cochrane review.[52]


Azathioprine is a well-established choice of adjuvant drug for the management of pemphigus. A reasonable duration of treatment is needed to test efficacy, and treatment failure should only be determined after at least 3 months at a dose of 2·5 mg kg−1 in patients with normal thiopurine methyltransferase levels.[47] Although there remains a lack of high-quality prospective randomized trials there is some evidence to suggest that the coadministration of azathioprine reduces the cumulative corticosteroid dose and has a superior corticosteroid-sparing effect compared with mycophenolate mofetil.

Mycophenolate Mofetil (Strength of Recommendation B, Level of Evidence 1+)

Introduction. Mycophenolate mofetil is often used as a first-line adjuvant to corticosteroid agents. Total daily doses of 2–3 g are given typically in two divided doses with prednisolone, thus 1–1·5 g twice daily. In patients who experience gastrointestinal side-effects, mycophenolic acid can be given as an alternative, with the approximate equivalent dose being 720–1080 mg twice daily.

Several small, unblinded trials have suggested that mycophenolate mofetil is beneficial in pemphigus treatment. In a series of 12 patients who had relapsed on corticosteroids plus azathioprine, 11 improved on mycophenolate mofetil (2 g per day) and prednisolone (2 mg kg−1), allowing a reduction in the prednisolone dose to 5 mg per day or less during the follow-up of 1 year. The patients responded rapidly, with a fall in IIF titres, and were free of lesions within 8 weeks of initiating mycophenolate mofetil.[96] However, based on nine patients, Nousari et al. commented that higher doses of mycophenolate mofetil (2·5–3 g per day) were often required to induce remission in PV, and at least 8 weeks of treatment was necessary before clinical and immunological improvement was observed.[97] There have been more than 30 case series since these. Examples of the number of patients achieving disease control in previously refractory patients with PV then treated with mycophenolate mofetil as a corticosteroid-sparing agent include 71% (22 of 31),[98] 73% (eight of 11)[99] and 78% (14 of 18).[100] Few adverse effects were reported.

Comparison with oral corticosteroids. In the study described in section 18·1, 30 patients with PV were given prednisolone alone (initial dose 2 mg kg−1) and in another 30 it was combined with mycophenolate mofetil (1 g twice daily) in this single-centre unblinded RCT.[53] There were no significant differences in efficacy in these two arms. The cumulative corticosteroid dose in the mycophenolate mofetil arm was lower but did not reach statistical significance.[52]

In an unblinded RCT,[61] 47 patients (36 PV, 11 PF) were allocated randomly to receive methylprednisolone alone or methylprednisolone (initially 1 mg kg−1 prednisolone equivalent) and mycophenolate mofetil (1·5 g twice daily). Disease activity was scored according to the number of lesions present. The authors reported no difference in the time to achieve disease control, induction of partial and complete remissions on or off minimal therapy, or the total amount of corticosteroids administered. There was no difference in the frequency of relapses or the development of side-effects and complications.[61]

There has been one double-blinded, placebo-controlled RCT.[54] In this multicentre study, 94 of 96 randomized patients were treated and 75 completed the study. Patients were allocated to either prednisone 1–2 mg per day (initial dose) plus placebo (n = 37), prednisone plus mycophenolate mofetil 2 g per day (n = 22) or prednisone plus mycophenolate mofetil 3 g per day (n = 37). The primary outcome measure was the proportion of patients in each arm responding to treatment as determined by an absence of new or persistent lesions and a prednisone dose of ≤ 10 mg daily dose from weeks 48–52. While the authors found no significant difference in the primary end points, the time to initial response was faster and the time to a sustained response was 12 weeks shorter in both mycophenolate mofetil-treated arms. In addition, the cumulative corticosteroid dose taken over weeks 12–52 of the study was significantly lower in the combined mycophenolate mofetil arm compared with the placebo arm (P = 0·028). Efficacy was similar in both mycophenolate mofetil arms, but infectious adverse events were higher in those taking 3 g daily. In both these arms infections were more common than in the placebo arm.[54]

Comparison with other adjuvant drugs. Studies comparing mycophenolate mofetil with azathioprine and cyclophosphamide are described in section 18. In summary, there is evidence that mycophenolate mofetil has an inferior corticosteroid-sparing effect compared with azathioprine[52,53,71] but may be more effective at achieving disease control.[52,71] Adverse events were not significantly different in these two studies but one did show fewer grade 3 and 4 adverse events with mycophenolate mofetil.[71] Mycophenolate mofetil has an inferior corticosteroid-sparing effect compared with pulsed cyclophosphamide.[52]


On the basis of current knowledge, there is evidence that mycophenolate mofetil has a corticosteroid-sparing effect. It could be considered as an alternative to azathioprine in patients unresponsive to treatment or where comorbidities or baseline investigations preclude azathioprine. It has a more favourable side-effect profile than azathioprine and is well tolerated. Treatment failure has been defined by international consensus as failure to respond to 3 g daily for 3 months.[47]

Rituximab (Strength of Recommendation B, Level of Evidence 1+)

Introduction. Rituximab is a chimeric murine–human monoclonal antibody of the IgG1 subclass, directed against the B-lymphocyte-specific antigen CD20,[101] expressed by early B cells in the bone marrow, autoantigen-specific B cells, memory B cells and mature B cells. Following treatment with rituximab there is rapid and sustained depletion of circulating and tissue-based B cells that is maintained for at least 6–12 months. Recent data suggest that rituximab may also affect T-cell function and modulate autoreactive T cells and production of T-cell cytokines.[102] Further details are provided in Appendix S2 (see Supporting Information).

Efficacy. In an unblinded RCT, 90 newly diagnosed and treatment-naive patients with moderate and severe PV (n = 74) and PF (n = 16) were treated with rituximab (1 g on days 0 and 14 and 0·5 g at 12 and 18 months) in combination with short-term prednisolone (0·5–1 mg kg−1 for 3–6 months) compared with prednisolone alone (1–1·5 mg kg−1 for 12–18 months).[45] There was a significant difference in primary outcome: 89% of patients in the rituximab arm were in complete remission off all treatment at 2 years compared with 34% of patients treated with prednisolone alone (P < 0·001). The rates were 89% vs. 28% in those with PV (P < 0·001). There were fewer severe adverse events in the rituximab-treated patients, which probably reflects the fact that prednisolone doses were higher and more prolonged in those treated with prednisolone alone. The lack of blinding is a flaw of this trial, and in particular the risk of withdrawal bias as dropout rates were higher in the prednisolone-only arm. However, reanalysis assuming that all withdrawals in the prednisolone-only arm went on to achieve remission off treatment still leads to a highly significant result. Nevertheless, the guideline group felt it appropriate to downgrade the recommendation rating to B based on this single unblinded RCT.

Previous studies of rituximab have considered its use in patients resistant to other therapies: multiple case series (reviewed in Ahmed and Shetty[103] and Wang et al.)[104] suggest that it is of utility in the treatment of PV, PF and paraneoplastic pemphigus, with rates of remission in refractory disease of up to 86% following a single cycle of treatment.[105] In a meta-analysis of 578 patients with pemphigus (496 PV), remission was achieved in 76% of patients following a single cycle of rituximab, and 39% were able to come off adjuvant treatments.[104] In this study the mean times to disease control and remission were 1·1 and 5·8 months, respectively. Relapse occurred in 40% of patients after an average duration of 14·5 months. Similar data are reported in an analysis of 451 patients with PV from case series: remission was achieved in 74–87% after a single cycle (16–58% remained on other therapies, 27–58% off adjuvant treatments).[103] They reported clinical responses within 6 weeks and a relapse rate of up to 65% occurring 13–17 months after rituximab.

In a single case, rituximab was used as a sole agent, and complete healing had been achieved 6 weeks after starting treatment.[106]

Dose. Initial studies employed a dosing regimen derived from the treatment of patients with lymphoma, using four, weekly infusions of 375 mg m−2.[105] A comparison of repeated weekly treatments of 375 mg m−2 suggested that patients with pemphigus who received three or more infusions demonstrated more rapid complete remission of disease compared with those who received only one or two infusions (149 vs. 443 days) and lower levels of relapse (0% vs. 67%).[107]

More recently, an alternative regimen has been introduced, based on that employed in the treatment of rheumatoid arthritis (RA). A regimen of two infusions of rituximab 1 g, 2 weeks apart, has now been shown to be effective in retrospective[108,109] and prospective[45,110] studies. Modified protocols have been used, but data suggest that the 'low-dose' RA protocol (2 × 0·5-g infusions) has a lower response rate and shorter time to relapse than the standard RA protocol.[111]

Comparisons of the standard RA and lymphoma protocols have failed to show consistent superiority of either one. Two studies reported no significant differences,[104,111] although there was a trend to better outcomes in the lymphoma-protocol-treated patients in the study by Wang et al.[104] In their analysis, Ahmed and Shetty showed significantly better clinical responses in the RA-protocol-treated patients but with higher relapse rates (nonsignificant).[103] Regarding cost, the RA protocol is less expensive, in terms of both drug cost and the associated expense of requiring two rather than four intravenous infusions. Lower-dose treatment (two infusions of 500mg each, two weeks apart) has been studied and reported to be effective,[112,113] although this approach may be associated with poorer response and increased rates of relapse.[111,114] Lower-dose rituximab (500 mg) has been used to control relapse following successful treatment with a standard 2 × 1 g induction regimen.[115]

In a single report of resistant oral pemphigus rituximab was used intralesionally.[116]

Combination with other therapies. In general, rituximab has been used as part of combination therapy including systemic corticosteroids together with cytotoxic immunosuppression, or as an adjunct to treatment with intravenous immunoglobulin (IVIg)[117] or immunoabsorption.[118–120] Of 372 patients with PV who received rituximab reported in the study by Ahmed and Shetty, 79–97% were treated concomitantly with adjuvant corticosteroids and/or immunosuppressants (59–69% with both).[103] Two studies have employed rituximab together with prednisolone alone, and in one study of 42 patients to good effect.[121] The other, an RCT with 46 newly diagnosed patients (38 PV), resulted in complete remission off treatment in 89% of patients.[45] At present, there are insufficient comparative data to indicate which of these approaches is preferable, from the perspective of either efficacy or adverse effects.

Adjuvant systemic immunosuppressive drugs can be continued with concomitant use of rituximab, but dose reduction should be considered to decrease the risk of infections and other adverse effects related to immunosuppression.

Novel anti-CD20 agents. A number of novel anti-B-cell antibodies are currently in development,[122] although to date only one has been reported to have been used in a patient with pemphigus. Veltuzumab (a humanized anti-CD20 antibody) was used as two subcutaneous injections 2 weeks apart in a patient with severe pemphigus that was refractory to conventional immunosuppression and several cycles of rituximab.[123] Complete remission resulted and was sustained for 2 years, at which point the patient was re-treated, again with induction of remission. While rituximab resistance is rare, such novel agents will undoubtedly be of benefit in some patients and may also be more convenient as a result of the subcutaneous route of administration.


The superior efficacy of rituximab and short-term corticosteroids compared with corticosteroids alone has been demonstrated in a single unblinded RCT of newly diagnosed patients with pemphigus. There is also evidence that rituximab is effective in treatment-resistant disease, and in most of these cases it has been given in combination with standard immunosuppression. Rituximab is effective for all forms of pemphigus. The 2 × 1 g RA dosing protocol is preferred due to cost considerations, with similar efficacy to the lymphoma protocol.


Introduction. Cyclophosphamide treatment regimens in PV vary from daily oral administration to fortnightly or monthly pulses, or a combination of these.[124] Large, comparative trials examining differing doses and regimens are lacking in PV. However, it is interesting to note that studies analysing pulsed intravenous and daily oral cyclophosphamide therapies in the treatment of vasculitis suggest equal efficacy, but with a lower cumulative dose and rate of complications for pulsed treatment,[125,126] although the risk of relapse may be higher.[126,127] Guidelines produced for the treatment of antineutrophil cytoplasmic autoantibody (ANCA)-associated vasculitis also recommend discontinuation of cyclophosphamide, both oral and intravenous, after 3–6 months, with transfer to an alternative maintenance therapy, azathioprine or methotrexate, because of the risk of haemorrhagic cystitis, cancer and infertility associated with prolonged exposure to cyclophosphamide. The recommended total duration of cyclophosphamide treatment in this context is up to a maximum of 6 months.[128]

Oral cyclophosphamide (Strength of recommendation D, Level of evidence 3). Early studies reported the corticosteroid-sparing effects of cyclophosphamide at doses of 50–200 mg per day in case series of up to six patients.[129–133] Prolonged remission with cessation of all therapy was observed in some cases.[130] In a retrospective case series including 20 patients with PV who had failed or were intolerant to azathioprine or mycophenolate mofetil, or had severe PV, cyclophosphamide 2–2·5 mg kg−1 with prednisolone, initially at 1 mg kg−1 per day, led to remission on minimal prednisolone doses (< 12·5 mg per day) in 85% of patients.[134] A larger retrospective study included 51 patients treated with cyclophosphamide 100 mg daily (1·1–1·5 mg kg−1) and prednisolone 100 mg daily (1·1–1·5 mg kg−1) compared with prednisolone alone (n = 20) or combined with azathioprine (n = 16) or ciclosporin (n = 14). The time to clinical and immunological remission was significantly shorter in the cyclophosphamide arm, with lower cumulative corticosteroid doses, suggesting that cyclophosphamide is more effective than prednisolone alone and is superior to azathioprine and ciclosporin.[94] However, in an earlier study superiority was not demonstrated: the efficacy of prednisolone (40 mg per day) alone was compared with prednisolone/cyclophosphamide (100 mg) and prednisolone/ciclosporin (5 mg kg−1) in 28 patients with oral pemphigus.[27] There was no significant difference in the duration to achieve remission or in relapse rates between the three groups, but cyclophosphamide and ciclosporin were given for a brief period of only 2–3 months.[27] Treatment failure for oral cyclophosphamide has been defined by international consensus as failure to achieve disease control after 3 months of treatment at 2 mg kg−1 per day.[47]


Oral cyclophosphamide 1–2 mg kg−1 could be considered as an alternative to azathioprine or mycophenolate mofetil. Due to concerns about its toxicity, it is best reserved for patients with recalcitrant or severe PV.

Intravenous cyclophosphamide (Strength of recommendation B, Level of evidence 2+). Pulsed intravenous cyclophosphamide with dexamethasone or methylprednisolone: This refers to the intermittent administration of high doses of intravenous corticosteroids and cyclophosphamide, usually three daily doses of dexamethasone (100 mg) or methylprednisolone (500–1000 mg) and a single dose of cyclophosphamide (500 mg) given monthly. Doses and frequency are arbitrary.

Dexamethasone–cyclophosphamide pulse (DCP) therapy for PV, first described in 1984 by Pasricha and Ramji,[135] is widely used in India for all types of pemphigus. The originally described regimen comprises four phases. Phase 1 consists of monthly intravenous dexamethasone 100 mg on three consecutive days with 500 mg intravenous cyclophosphamide on day 2. Low-dose daily oral cyclophosphamide (50 mg) is administered between pulses. Pulsing is continued until clinical remission and followed by a consolidation phase of a further six DCP courses (phase 2). Oral cyclophosphamide is then continued alone (phase 3) and if there are no relapses after 1 year, all treatment is withdrawn (phase 4). Minor modifications have been made to the regimen, including extending phase 2 to 9 months, reducing phase 3 to 9 months and the addition of daily oral corticosteroids if needed during phase 1.[136] Using the original regimen, 81% (346 of 425) of patients with pemphigus were in remission and had been off all therapy for at least 2 years, with 74% (313 of 425) for more than 5 years.[136] Four percent of patients died during treatment. Using the modified regimen, 86% (106 of 123) of patients with pemphigus had completed treatment and had been off therapy for at least 2 years, with 50% (62 of 123) for more than 5 years. The mortality rate was 2%.

Many other retrospective case series describing the encouraging results of this treatment approach have been published both from Indian centres[137–143] and from other countries around the world including Iran, South Africa, the U.K. and Serbia.[144–147] In one study, 100% of 32 patients with PV completed the regimen and were off treatment, in remission.[138]

Advocates of the DCP regimen claim relative freedom from corticosteroid side-effects, but 20–85% of menstruating women developed amenorrhoea;[145,147–149] azoospermia in men was also noted. Haemorrhagic cystitis occurred in 0·6%[149] and pituitary–adrenal suppression in 55% (17 of 31) of patients.[150]

Dexamethasone–cyclophosphamide pulse regimen compared with oral corticosteroids: DCP therapy has not been tested rigorously against other treatment protocols in controlled trials, but one study has compared the 6% mortality achieved in 50 patients (45 PV) on DCP therapy with an estimated 25–30% mortality in historical cohorts on conventional corticosteroid therapy at the same institute.[137] More recent studies also indicate an advantage of combining pulsed cyclophosphamide with conventional corticosteroids compared with corticosteroids alone; neither used DCP therapy. In a controlled, open-label study, described in section 18·1, the addition of intravenous pulsed cyclophosphamide 1 g monthly for 6 months, then every 2 months, to conventional oral prednisolone resulted in significantly lower cumulative corticosteroid dose at 1 year.[53] Similarly, in a randomized, prospective unblinded trial, 60 patients with PV were randomized to receive prednisolone 1 mg kg−1 per day with or without monthly intravenous cyclophosphamide 15 mg kg−1 for 1 year. There were no significant differences in the two treatment arms, but many outcomes tended to be better in the arm that included pulsed cyclophosphamide, with reduced relapse rates and cumulative corticosteroid doses.[58]

Dexamethasone–cyclophosphamide pulse regimen compared with alternative pulsing protocols: One study has compared DCP therapy with an alternative pulsing protocol: in a prospective, randomized open-label trial, 28 patients with PV received either DCP therapy or conventional oral prednisolone 1·5 mg kg−1 plus monthly cyclophosphamide pulses 15 mg kg−1. Most efficacy parameters were similar, although the time to achieve remission was significantly shorter in the oral prednisolone plus cyclophosphamide 15 mg kg−1 arm.[79] However, the period of study was 1 year only.

Comparison of pulsed cyclophosphamide with other adjuvant drugs: Modified DCP regimens used in several trials have failed to demonstrate consistent superiority over other corticosteroid/adjuvant PV treatment regimens. Studies comparing pulsed cyclophosphamide with azathioprine and mycophenolate mofetil are summarized in section 18. In summary, three trials have compared azathioprine with pulsed cyclophosphamide regimens: one RCT showed no significant differences;[70] a nonrandomized trial favoured pulsed cyclophosphamide, which showed a lower cumulative corticosteroid dose although efficacy was similar;[72] and a single-centre RCT showed lower cumulative corticosteroid doses in the azathioprine arm, which did not reach significance in the authors' analysis but was considered significant in an independent Cochrane review.[52] Pulsed cyclophosphamide has a superior corticosteroid-sparing effect compared with mycophenolate mofetil.[52,53]

Dose: The dose of intravenous cyclophosphamide most commonly reported for the treatment of PV is a fixed dose of 500 mg monthly, but this is arbitrary and is often combined with 50 mg per day oral cyclophosphamide. Three studies have given intravenous cyclophosphamide 15 mg kg−1 monthly combined with conventional oral corticosteroids and without daily oral cyclophosphamide.[58,79,81] In another study, a fixed dose of intravenous cyclophosphamide 1 g monthly was given, without daily oral cyclophosphamide, and combined with conventional oral corticosteroids.[53] It is common practice to combine intravenous cyclophosphamide with mesna to reduce the risk of haemorrhagic cystitis.[124]

The dose of 15 mg kg−1 for an intravenous cyclophosphamide pulse is commonly used in the treatment of other severe autoimmune diseases. For example, in remission induction of ANCA-associated systemic vasculitis, pulsed intravenous cyclophosphamide 15 mg kg−1 (maximum dose 1500 mg) is given initially every 2 weeks, reducing to every 3 weeks and continued for a maximum of 6 months.[128]


There is some evidence that pulsed cyclophosphamide therapy may reduce cumulative corticosteroid dose. There is no consistent evidence that it is more effective than other adjuvant drugs, so in view of concerns about long-term toxicity and the practical disadvantages of administering regular intravenous treatment, it is best reserved for severe or recalcitrant cases of PV.

Intravenous Immunoglobulin (Strength of Recommendation B, Level of Evidence 2++)

Many reports have suggested the utility of IVIg in patients with PV,[151–154] and a recent double-blind, placebo-controlled study in 61 patients has confirmed this in a robust way.[155] Patients with PV treated with a single cycle of IVIg (either 1 g kg−1 or 2 g kg−1 divided over 5 days) did significantly better, measured according to the need for additional treatment, than those treated with placebo, and a dose–response effect was demonstrated. Clinical improvement was measured objectively and seen by day 8 in the higher-dose treatment arm. In addition, a significant fall in desmoglein antibody titres was demonstrated in both treatment groups with no fall in the placebo group. A placebo-controlled crossover trial of IVIg in a single patient also confirmed its efficacy, with significantly improved disease activity scores and lower indirect immunofluorescence titres and desmoglein 1 and 3 antibody levels.[156]

In pemphigus, IVIg is generally used at high dose, typically 2 g kg−1 in divided doses over several days, together with corticosteroids with or without cytotoxic immunosuppressive agents such as azathioprine or mycophenolate mofetil. Treatment is given at monthly intervals and may need to be prolonged for continued effect. Thus, multiple treatments will be needed if used to maintain remission. IVIg seems to act by increasing catabolism of pathogenic antibodies.[157,158] It is generally well tolerated[154] and has the attraction over other adjuvant therapies that it does not increase the risk of infection. IVIg has been used to treat pemphigus in pregnancy[159] and in children.[160]

While uncommon, adverse effects of IVIg do occur, including headache, aseptic meningitis and anaphylaxis, which is a particular risk in patients who are IgA deficient.


IVIg could be considered as maintenance treatment in patients with refractory disease unresponsive to other adjuvant drugs. In view of reports of a rapid action in some cases, it may also be used to help induce remission in patients with severe PV while slower-acting drugs take effect. IVIg should be considered as part of the acute management of severe or widespread pemphigus and in patients who are at particularly high risk of infection.

Methotrexate (Strength of Recommendation D, Level of Evidence 3)

Methotrexate has been used as an immunomodulatory and corticosteroid-sparing agent in a variety of skin diseases. Studies from the late 1960s and early 1970s[14,161–163] attributed high morbidity and mortality to methotrexate and hence it fell out of favour for its adjuvant use in PV. Three of four patients cited in one report died, but high doses of methotrexate had been used (125–420 mg per week) in combination with prednisolone 40–240 mg per day.[14] There have been no controlled trials evaluating the role of methotrexate in the treatment of PV.[162–166]

A retrospective review of 116 patients with PV revealed clinical improvement in 83% (96 of 116) when methotrexate was used in doses of 10–50 mg per week, in combination with corticosteroids. Thirteen patients did not improve, two had it discontinued for unknown reasons, and five died from causes unrelated to methotrexate therapy. Of the responders, 14 patients were clear at a mean of 2·6 years (range 3 months to 18 years) after discontinuation of all systemic therapy.[167]

Two retrospective studies have shown a corticosteroid-sparing effect with the use of methotrexate in PV. In a 25-year survey of 53 patients treated with methotrexate and systemic corticosteroids, there was a 50% reduction in the dose of corticosteroids,[168] and in the second study, prednisolone (mean dose prior to treatment with methotrexate 20 mg per day, range 3–40) was discontinued in six of nine patients.[169]

In 2012, a retrospective review of methotrexate use in PV reported its effectiveness in moderate-to-severe cases as an adjuvant to systemic corticosteroids. A predetermined severity score was used by the authors, which included the number of erosions, percentage of body surface involved and the dose of prednisolone used. In total 30 patients were identified and used methotrexate 15 mg per week. Of the 25 patients described as having severe or moderate disease in the study, 84% (21 of 25) improved their severity score within 6 months (P < 0·001). Only 13% (four of 30) experienced side-effects. The dose of prednisolone was reduced (range 2·5–85 mg) in 23 patients (77%), and in 21 patients (70%) the decrease was 50% or more.[170]

A retrospective review by Tran et al.[171] on the adjunctive use of methotrexate in patients with PV has demonstrated its effectiveness as a corticosteroid-sparing agent; 23 patients with PV were treated with methotrexate, of whom 21 (91%) experienced improvement (as measured by reduction in the prednisolone dose). Sixteen patients (70%) were eventually weaned off prednisolone completely. The mean dose of methotrexate used in this study was 18·9 mg per week (range 15–25).


Given the limitations of the data available, it would be difficult to recommend methotrexate as a first-line agent in the treatment of PV.[172] Methotrexate could be considered as an adjuvant drug if more established drugs cannot be used or have failed. International consensus has defined treatment failure as persistent disease despite methotrexate 20 mg per week for at least 12 weeks.[47]

Dapsone (Level of Evidence 1−)

Dapsone has been reported to be beneficial as an adjuvant drug in several case reports of PV.[173–177] However, in three of these cases, it was started either with or shortly after prednisolone, and in two cases it was started after long-standing prednisolone was increased to high doses. Therefore, it is difficult to be certain whether dapsone had a significant role if any.

In a case series, five of nine patients with PV in the maintenance phase of treatment and who had been unable to reduce their prednisone dose below 15 mg per day experienced a mean ± SEM drop of 67 ± 7·1% in prednisone dose after 4 months of maximal dapsone treatment and an 84% ± 3·5% drop in prednisone dose after 8 months of dapsone treatment.[178]

There has been one double-blind, placebo-controlled RCT undertaken to look for a potential corticosteroid-sparing effect of dapsone. Nineteen patients with PV on maintenance treatment with corticosteroids and/or immunosuppression were randomized to additional dapsone (n = 9) vs. placebo (n = 10). The primary outcome measure was reduction of prednisolone to 7·5 mg daily for at least 30 days within 1 year of achieving the maximum dapsone dose (150–200 mg per day). The results were based on an intention-to-treat analysis and did not show a statistical difference: 56% (five of nine) of the dapsone group were treated successfully, three failed treatment and one left the study. Among the placebo group 30% (three of 10) were treated successfully, 57% (four of seven) of those who failed treatment were treated with dapsone and in 75% (three of four) of those it was successful. Among those who completed the dapsone trial, 73% (eight of 11) vs. 30% (three of 10) on placebo showed a corticosteroid-sparing effect with dapsone. However, the study numbers are very small and at best may show only a slight trend for a corticosteroid-sparing effect with dapsone.[77]


There is weak evidence to suggest that dapsone may have a corticosteroid-sparing effect. Larger placebo-controlled RCTs are needed.

Tetracyclines/Nicotinamide (Level of Evidence 3)

Tetracyclines have been used in the treatment of PV, with or without nicotinamide, in varying combinations. Sixteen patients were given nicotinamide 1·5 g and tetracycline 2 g daily. In 12, no systemic corticosteroids were given, and of these, three cleared and three improved.[179,180] Of the four patients given additional prednisolone, there was clearance in one, partial improvement in two and no response in another.[179]

Thirteen hospitalized patients with PV were given tetracycline 2 g daily for a month followed by 1 g per day for the next 4 weeks in combination with oral prednisolone. They had a faster response rate and required lower doses of prednisolone compared with seven historical corticosteroid-treated controls.[181]

Two studies using minocycline 50–200 mg per day as an adjuvant drug reported improvement and a corticosteroid-sparing effect in 54% of patients (seven of 13).[182,183]


Tetracyclines, with or without nicotinamide, are not widely used for the treatment of PV, and evidence of their corticosteroid-sparing role is weak, but they could be considered as adjuvant treatment, perhaps in milder cases of PV.

Sulfasalazine and Pentoxifylline (Level of Evidence 2−)

A double-blind, placebo-controlled clinical trial in 64 patients with PV was carried out to ascertain the value of sulfasalazine and pentoxifylline as an adjuvant therapy for PV. Patients were not randomized. The drugs were chosen as low-cost, antitumour necrosis factor (anti-TNF) agents. All patients received standard pulsed therapy with intravenous corticosteroid (500 mg on five consecutive days) and pulsed cyclophosphamide (on day 1) in a 2- to 4-weekly cycle with oral cyclophosphamide (100 mg per day) and oral corticosteroid (60 mg per twice weekly) between the cycles. In addition, group 1 (n = 42) were treated with oral sulfasalazine 500 mg twice daily and pentoxifylline 400 mg twice daily for 8 weeks, while group 2 (n = 22) received a placebo. The serum level of TNF-α was higher statistically in both groups of patients than in the healthy individuals. There was a statistically significant decrease in the serum levels of TNF-α in patients in group 1 compared with those in group 2 at 6 and 8 weeks. There was also a rapid clinical improvement in patients in group 1 compared with those in group 2.[184]


There is some evidence to support the use of pentoxifylline and sulfasalazine as adjuvant therapy in the treatment of PV, but further studies are required.

Chlorambucil (Level of Evidence 3)

Like cyclophosphamide, chlorambucil is a nitrogen mustard alkylating agent. Since the last guidelines were compiled, there have been no new reports of the use of chlorambucil in PV. The biggest series, published in 2000, involved seven patients with PV who had failed to respond to other corticosteroids and immunosuppressants. They were given oral chlorambucil 4 mg per day, titrated upwards according to clinical response. There was improvement or remission in five patients and a corticosteroid-sparing effect was noted. A fall in IIF titres was reported in three of four cases.[185] The lack of bladder toxicity with chlorambucil is an advantage compared with cyclophosphamide.


Chlorambucil could be considered as an adjuvant drug if more established options cannot be used, but there are limited data to support its use.

Gold (Strength of Recommendation D, Level of Evidence 3)

Gold is a historical treatment, rarely used now in the treatment of PV. Most studies have used intramuscular gold, given as sodium aurothiomalate, initially at a dose of 50 mg per week as an intramuscular injection if test doses were tolerated. It was used successfully as monotherapy in five patients, with an associated fall in IIF titres.[186,187] However, it has been used more commonly as an adjuvant drug, and corticosteroid-sparing effects are reported.

Penneys et al. reported a series of patients receiving gold for up to 4 years, with 14 of 15 patients responding. Eight achieved remission off-treatment after a mean of 21 months, seven achieved remission on treatment and one stopped due to side-effects.[188] In a retrospective review of 26 patients treated with gold over 10 years, a response was seen in 62% and complete remission off-treatment had occurred in four patients. Toxicity was seen in 42%. The average dose of prednisolone was reduced from 55 mg per day before gold to 9 mg per day at the end of the study.[189] A more recent study used gold as an adjuvant therapy in 13 patients, with prednisolone doses ranging from 7·5 to 100 mg. The addition of 50 mg intramuscular gold was felt to be beneficial as seven patients went into complete remission and four were able to reduce prednisolone doses.[190]

Significantly, there are also case reports implicating gold as a trigger for pemphigus. Gold compounds contain a thiol group, which has previously been implicated in drug-induced pemphigus. Lo Schiavo et al. reported a convincing case of gold-induced pemphigus in a patient with rheumatoid arthritis, with complete resolution on withdrawal of gold.[191]


Gold is now a historical treatment in most developed healthcare systems. It could be considered as an alternative to more established adjuvant drugs if they cannot be used or are unavailable. However, the lack of randomized trial data makes the magnitude of an effect uncertain and there is a risk of gold acting as a disease trigger.

Ciclosporin (Level of Evidence 1−)

There are a number of case reports suggesting that ciclosporin is a useful adjuvant with corticosteroid-sparing effects in PV.[64,192–194] However, a small prospective single-centre RCT of 33 patients comparing oral methylprednisolone 1 mg kg−1 alone vs. methylprednisolone with ciclosporin 5 mg kg−1 found no statistically significant difference in outcome measures such as time to healing, complete remission rate and cumulative corticosteroid dose. More side-effects were encountered in the ciclosporin group during a mean follow-up period of 5 years.[195] There were no deaths, and 10 patients (five from each group) were in complete remission, off all therapy, while the others were taking an average of prednisone 2·5 mg per day.[195] Olszewska et al. reported a retrospective series of 101 patients with PV treated with prednisolone alone (n = 20) or in combination with adjuvant immunosuppressants, including azathioprine (n = 16), ciclosporin (n = 14) and oral cyclophosphamide (n = 51).[94] Cyclophosphamide plus prednisolone was significantly better at inducing remission than prednisolone alone. Ciclosporin did not add any significant benefit. The proportion remaining relapse free 5 years after discontinuation of treatment was lowest in the ciclosporin group, at 43%, and highest in the cyclophosphamide group at 69%.[94]


On the basis of current evidence, ciclosporin cannot be recommended as an adjuvant drug in PV.