ORIGINAL ARTICLE
Sharon BAUM1,2#, Shani STEINBERG2#, Ido TZANANI1,2, Aviv BARZILAI1-3 and Anna LYAKHOVITSKY1,2
1Department of Dermatology, Sheba Medical Center, Tel Hashomer, Ramat Gan, 2Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv and 3Institute of Pathology, Sheba Medical Center, Tel Hashomer, Ramat Gan, Israel
Current research on the malignancy rate and spectrum of malignancies in patients with bullous pemphigoid is contradictory. The aims of this study were to determine the prevalence and spectrum of malignancy in patients with bullous pemphigoid and to compare demographic, clinical, therapeutic and outcome data between bullous pemphigoid patients with and without malignancy. This retrospective cohort study enrolled 335 patients (194 women and 141 men; mean age at diagnosis of bullous pemphigoid 77.5 ± 12 years) followed up at an Israeli tertiary centre between January 2009 and December 2019: 107 (32%) had malignancy and 228 (68%) did not. Malignancy occurred before and after bullous pemphigoid diagnosis in 82 (77%) and 25 (23%) patients, respectively. Bullous pemphigoid patients with cancer were older (p = 0.02) and had a higher mortality rate (p < 0.0001) than those without malignancy. The 2 groups did not differ in terms of sex, comorbidities, or clinical characteristics. Those who developed malignancy before bullous pemphigoid were younger than those who developed malignancy after bullous pemphigoid (mean age 69.3 vs 82.4 years, p < 0.0001). Overall malignancy rates did not differ between patients with bullous pemphigoid and the general population; therefore, comprehensive malignancy workup may be unnecessary. However, patients with bullous pemphigoid had a greater risk of melanoma (10.7% vs 4.3%, p = 0.0005); therefore, routine skin screening may be recommended.
Key words: bullous pemphigoid; bullous dermatoses; malignant melanoma; mortality rate.
The aims of this study were to determine the prevalence and spectrum of cancer in patients with bullous pemphigoid (BP), and to compare BP patients with and without cancer. The results show that there is no increased rate of overall malignancy in BP, while the frequency of melanoma increases. Patients with both BP and cancer are older, but there are no differences in severity or clinical features. These results suggest that routine skin screening in patients with BP is necessary to improve treatment and outcomes.
Citation: Acta Derm Venereol 2023; 103: adv00888. DOI: https://doi.org/10.2340/actadv.v103.3979.
Copyright: © Published by Medical Journals Sweden, on behalf of the Society for Publication of Acta Dermato-Venereologica. This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (https://creativecommons.org/licenses/by-nc/4.0/)
Accepted: Feb 13, 2023; Published: Mar 14, 2023
Corr: Sharon Baum, Department of Dermatology, Sheba Medical Center, Tel HaShomer, Ramat Gan, Israel. E-mail: sharon.baum@sheba.health.gov.il
#These authors contributed equally.
Competing interests and funding: The authors have no conflicts of interest to declare.
Bullous pemphigoid (BP) is the most prevalent autoimmune blistering disorder that primarily affects older individuals (1). Over the last 20 years, the prevalence of BP has increased by 1.9–4.3 time, as reported in data from the UK, France, Germany, and Israel. This trend has been attributed to population ageing, drug-induced cases, and the increased identification of non-bullous presentations (1). The overall prevalence of BP in the US is 12 cases per 100,000 individuals, with cases among women slightly outnumbering those among men (12.7/100,000 vs 11/100,000). The prevalence rate of BP is highest (123.6 cases/100,000 persons) among patients aged ≥ 90 years (2). The annual incidence of BP varies from 2.4 to 21.7 cases per million persons worldwide, with 2.5–42.7, 2.6–7.4, and 10 cases per million people in Europe, Asia, and the USA, respectively (3, 4). With increasing age, the incidence rate rises very rapidly, reaching 190–312 and up to 507 cases per million persons among individuals aged > 80 and ≥ 85 years, respectively, worldwide (1, 5).
BP is clinically distinguished by pruritic skin eruptions with the formation of tight bullae on erythematous skin, together with urticarial papules and plaques. Only papular urticarial eczematous lesions or excoriations are observed in the early stages or in atypical non-bullous variants. Most patients have severe pruritus. In 10–30% of cases, mucosal involvement may occur (6). The pathogenesis of BP is linked to tissue-bound and circulating autoantibodies directed against hemidesmosomal proteins (BP antigen 180, BP antigen 230, or both) involved in dermal-epidermal cohesion. Binding of these antibodies triggers an inflammatory cascade and the formation of subepidermal blisters (6). Histopathology reveals a subepidermal blister with an inflammatory infiltrate frequently harbouring eosinophils. Direct immunofluorescence (DIF) shows linear deposition of immunoglobulin G and/or C3 at the dermo-epidermal junction (7).
The diagnosis of BP is based on a combination of clinical features, histological findings, including positive DIF, and the detection of specific circulating immunoglobulin G (IgG) anti-basement membrane autoantibodies using indirect immunofluorescence investigations or enzyme-linked immunosorbent assays (ELISAs) (7).
BP has been linked with a variety of comorbidities, including neurological illnesses (multiple sclerosis, dementia, Parkinson’s disease, epilepsy, and stroke) (8); an elevated risk of thromboembolic events (9); psychiatric disorders (10); dermatoses (psoriasis and lichen planus) (11, 12); and various medications (13). The link between BP and malignancy has also been studied with conflicting results. Some studies have shown an association with malignancies (either solid tumours and/or haematological malignant disorders) (14–17), while others claim that this association is coincidental and may be explained by confounding factors, such as age (18, 19). Several reports have described unusual clinical BP variants in patients who also have malignancies (20, 21). Because the prevalence of BP varies by geographical area and ethnicity (22, 23), it is probable that comorbidities, such as malignancy, will also vary.
The objectives of this study were to compare the frequency of malignancies in patients with BP with that in the general population in Israel, and to describe their spectrum and range. In addition, the study aimed to evaluate the demographics, medical histories, clinical features, therapies employed, and mortality rates among BP patients with and without malignancy.
This cohort study evaluated the medical records of patients diagnosed with BP. The patients were diagnosed, treated, and followed up in the dermatology department and dermatology outpatient clinic at the Sheba Medical Center, the largest and most comprehensive tertiary centre in the middle east, between January 2009 and December 2019. The study was approved by the institutional review board (7172-09-SMC). The medical records of the patients diagnosed with BP were collected from the Sheba Hospital Chameleon System computerized files and reviewed. Inclusion criteria were: patients of any age with: a confirmed diagnosis of BP based on the typical clinical picture; typical histological appearance of subepidermal splitting; positive DIF results confirmed in Sheba Medical Center’s laboratory or in a certified laboratory at another medical centre; and follow-up of at least 10 months after BP diagnosis. Patients with a misdiagnosis or incomplete files were excluded.
Demographic data, comorbidities (general, cutaneous, and malignancy), clinical parameters (presence or absence of bullae, mucosal involvement, pruritus, and disease severity at onset), treatment received, and follow-up time-span were collected. Data were gathered on the type of malignancy and chronology of its diagnosis in relation to BP diagnosis, and all patients were divided into 2 groups: (i) those without comorbid malignancy, and (ii) those with malignancy. Patients with comorbid malignancy were subdivided into those who developed malignancy before and after BP diagnosis. All collected parameters were compared between the groups. The mortality rate of BP patients with malignancy was estimated and compared with that of BP patients without malignancy.
The malignancy rate and spectrum in the cohort of patients with BP were compared with those in the general population. To achieve this, data from the study group were compared with that from the general population (adjusted for age) using the Israel National Cancer Registry (INCR) database for 2015, which was the midpoint of the follow-up period of the study cohort.
Statistical analysis was performed using IBM SPSS statistics, version 25 (IBM Corp., Armonk, NY, USA). Dichotomous variables were compared using the χ2 test, while dichotomous and quantitative variables (age and follow-up duration) were compared using the student’s t-test for unpaired samples. A comparison of the invasive cancer proportions was performed using the Z test. All tests were performed at a confidence level of 95% (significance level 0.05), with a statistical power of 80%.
The study included a final total of 335 patients, after excluding 409 patients due to misdiagnosis or uncertain diagnosis, 41 due to lack of access to data, and 41 due to loss to follow-up. The patients’ follow-up duration from BP diagnosis ranged from 10 to 129 (mean 17.4 ± 25.6) months. Of the 335 included patients, 194 were women and 141 were men, with a mean age at BP diagnosis of 77.5 ± 12 years: women 78.5 ± 11 years; men 76.2 ± 13.3 years (p = 0.07). Out of all the patients, 107 (32%) had at least 1 malignancy, whereas 228 (68%) had no malignancy. Among the BP patients with malignancy, 14 (13%) had 2 distinct malignancies. The malignancy preceded BP onset in 82 (77%) of the 107 patients (the duration between cancer diagnosis and BP onset ranged between 4 months and 42 years; mean 9.8 ± 1.80 years), and BP preceded the malignancy in 25 (23%) patients (duration between BP diagnosis and malignancy ranged between 1 month and 7 years; mean 2.6 ± 0.85 years). Of the 107 patients with BP and associated malignancy, 7 received immunosuppressive treatment before the diagnosis of malignancy. Patients with malignancy were significantly older than those without (p = 0.02). There were no differences in sex, comorbidities, and clinical characteristics between the patient groups. Data on epidemiology, clinical characteristics, and treatments received are shown in Table I.
| Sub-category | All patients with BP n = 335 | BP with malignancy n = 107 | BP without malignancy n = 228 | p-value |
| Demography | ||||
| Age at BP diagnosis, years, mean (SD) | 77.5 (12.0) | 79.2 (10.1) | 76.7 (12.8) | 0.02 |
| Female sex n (%) | 194 (58) | 57 (53) | 137 (60) | 0.40 |
| Comorbidities, n (%) | ||||
| Smoking n (%) | 19 (6) | 5 (5) | 14 (6) | 0.58 |
| Cardiovascular | 149 (44) | 52 (48) | 97 (42) | 0.31 |
| Diabetes mellitus | 153 (46) | 44 (41) | 109 (48) | 0.25 |
| Hypertension | 248 (74) | 83 (77) | 165 (72) | 0.31 |
| Endocrine | 69 (21) | 23 (21) | 46 (20) | 0.78 |
| Pulmonary | 57 (17) | 18 (17) | 39 (17) | 0.34 |
| Neuro-psychiatric | 144 (43) | 39 (36) | 105 (46) | 0.09 |
| Coagulation | 2 (0.6) | 2 (2) | 0 (0) | 0.03 |
| Gastrointestinal | 49 (15) | 19 (18) | 30 (13) | 0.26 |
| Dyslipidaemia | 169 (50) | 50 (47) | 119 (52) | 0.35 |
| Dermatological comorbidities, n (%) | ||||
| Dermatoses (any) | 21 (6) | 6 (6) | 15 (7) | 0.73 |
| Papulosquamous diseases | 14 (4) | 4 (4) | 10 (4) | 0.78 |
| Autoimmune dermatosisa | 8 (2) | 3 (3) | 5 (2) | 0.73 |
| Clinical parameters, n (%) | ||||
| Pruritus | 282 (84) | 85 (79) | 197 (86) | 0.10 |
| Bullae | 308 (92) | 98 (91) | 210 (92) | 0.87 |
| Mucosal involvement | 66 (20) | 24 (22) | 42 (18) | 0.39 |
| Mucosal involvement, n (%) | ||||
| No | 269 (80) | 83 (78) | 186 (82) | 0.70 |
| Oral involvement | 64 (19) | 22 (20) | 42 (18) | 0.64 |
| Genital involvement | 13 (4) | 3 (3) | 10 (4) | 0.48 |
| Moderate-severe disease | 296 (88) | 92 (86) | 204 (89) | 0.35 |
| Treatment, n (%) | ||||
| Systemic Treatment | 296 (88) | 92 (86) | 204 (89) | 0.35 |
| Dapsone/tetracycline | 257 (77) | 78 (73) | 179 (78) | 0.25 |
| High-dose corticosteroids | 41 (12) | 13 (12) | 28 (12) | 0.97 |
| MTX/azathioprine | 33 (10) | 5 (5) | 28 (12) | 0.02 |
| Mycophenolate mofetil | 30 (9) | 10 (9) | 20 (9) | 0.86 |
| Rituximab | 6 (2) | 1 (1) | 5 (2) | 0.41 |
| DPP4i | 8 (2) | 6 (6) | 2 (1) | 0.08 |
| Follow-up duration of BP, months range, mean (SD) | 10–129, 17.4 (25.6) | 10–117, 20.9 (25.7) | 10–129, 15.8 (25.4) | 0.09 |
| Case-mortality rate, n (%) | 78 (23) | 49 (46) | 29 (13) | 0.000 |
| aAlopecia areata, vitiligo, urticaria. MTX: methotrexate; SD: standard deviation; DPP4i: Dipeptidyl peptidase 4 inhibitors. |
||||
The patients with BP and malignancy had comparable comorbidities and clinical characteristics. The case mortality rate in patients with BP linked with malignancy was significantly higher (p < 0.0001) (Table I).
Table II shows the comparison of demographics, comorbidities, clinical characteristics, and treatments received by patients who were diagnosed with malignancy before and after BP diagnosis. Sex, comorbidities, and clinical characteristics did not differ significantly between the 2 groups. While the age at which BP was diagnosed was not different, the age at which malignancy was discovered was much lower in patients who had malignancy before BP (69.3 vs 82.4 years, p < 0.0001). The case mortality rate was significantly higher among patients who had been diagnosed with malignancy prior to BP diagnosis (52% vs 24%, p = 0.01).
| Sub-category | Malignancy preceded BP, n = 82 | BP preceded malignancy, n = 25 | p-value |
| Demography | |||
| Age at BP diagnosis, years, mean (SD) | 79.1 (10.5) | 79.7 (8.6) | 0.29 |
| Age at malignancy diagnosis, years, mean (SD) | 69.34 (11.8) | 82.4 (8.8) | < 0.0001 |
| Time period between BP onset and malignancy – diagnosis range, mean (SD) | 4 months–42 years, 9.8 (1.80) years | 1 months–7 years, 2.6 (0.85) years | < 0.0001 |
| Female sex, n (%) | 46 (56) | 11 (44) | 0.28 |
| Comorbidities, n (%) | |||
| Smoking n (%) | 5 (6) | 0 (0) | 0.20 |
| Cardiovascular | 40 (49) | 12 (48) | 0.94 |
| Diabetes mellitus | 33 (40) | 11 (44) | 0.73 |
| Hypertension | 63 (77) | 20 (80) | 0.73 |
| Endocrine | 20 (24) | 3 (12) | 0.18 |
| Pulmonary | 13 (16) | 5 (20) | 0.77 |
| Neuro-psychiatric | 30 (36) | 9 (36) | 0.95 |
| Coagulation | 2 (2) | 0 (0) | 0.43 |
| Gastrointestinal | 14 (17) | 5 (20) | 0.73 |
| Dyslipidaemia | 36 (44) | 14 (56) | 0.28 |
| Dermatological comorbidities, n (%) | |||
| Dermatoses (any) | 4 (5) | 2 (8) | 0.55 |
| Papulosquamous diseases | 3 (4) | 1 (4) | 0.93 |
| Autoimmune dermatosesa | 1 (1) | 2 (8) | 0.07 |
| Clinical features, n (%) | |||
| Pruritus | 65 (79) | 20 (80) | 0.93 |
| Bullae | 75 (91) | 23 (92) | 0.93 |
| Mucosal involvement | 18 (22) | 6 (24) | 0.83 |
| Mucosal involvement, n (%) | |||
| No | 64 (78) | 19 (76) | 0.81 |
| Oral involvement | 18 (22) | 4 (16) | 0.51 |
| Genital involvement | 1 (1) | 2 (8) | 0.07 |
| Moderate-severe disease | 72 (88) | 20 (80) | 0.32 |
| Treatment, n (%) | |||
| Systemic Treatment | 72 (88) | 20 (80) | 0.32 |
| Dapsone/tetracycline | 62 (76) | 16 (64) | 0.25 |
| High-dose steroids | 9 (11) | 4 (16) | 0.50 |
| vMTX/azathioprine | 3 (4) | 2 (8) | 0.36 |
| Mycophenolate mofetil | 7 (8) | 3 (12) | 0.60 |
| Rituximab | 1 (1) | 0 (0) | 0.57 |
| Immunosuppressive Tx before malignancy | 4 (5) | 3 (12) | 0.20 |
| Follow-up duration of BP, months range, mean (SD) | 10–107, 18.1 (23.1) | 10–117, 29.8 (31.8) | 0.04 |
| Case mortality rate, n (%) | 43 (52) | 6 (24) | 0.01 |
| aAlopecia areata, vitiligo, urticaria. SD: standard deviation; BP: bullous pemphigoid; MTX: methotrexate. |
|||
Table III describes the distribution of malignancies in patients with BP compared with that in the general adult age-matched population. There were no differences in the overall malignancy rates between the patients with BP and the general population. Significant differences were observed in the rates of melanoma, lung cancer, and colon cancer.Patients with BP had a higher rate of melanoma than that in the general population (10.7% vs 4.3%, respectively, p = 0.0005). Of the 13 cases of melanoma, 8 (61.5%) and 5 (38.5%) preceded and followed BP diagnosis, respectively. In addition, compared with the general population, patients with BP had lower rates of lung (4.1% vs 10.9%, p = 0.01) and colon (2.4% vs 12.2%, p = 0.0008) cancers. No significant differences were found in the other malignancy rates. Breast cancer (14.8%), squamous cell carcinoma (SCC) (13.2%), prostate cancer (11.5%), melanoma (10.7%), and bladder cancer (7.4%) were the 5 most common cancer types in patients with BP. Breast cancer (20.4%), prostate cancer (12.9%), colon cancer (12.2%), lung cancer (10.9%), and bladder cancer (7.7%) were the 5 most common cancer types in the general population.
| Malignancy | Patients with BP n (%) | New patients in 2015 n (%) | p-value |
| Breast | 18 (14.8) | 4,044 (20.4) | 0.11 |
| Invasive SCC | 16 (13.2) | – | |
| Prostate | 14 (11.5) | 2,550 (12.9) | 0.61 |
| Melanoma | 13 (10.7) | 855 (4.3) | 0.0005 |
| Bladder | 9 (7.4) | 1,524 (7.7) | 0.86 |
| Renal | 8 (6.6) | 890 (4.5) | 0.28 |
| Lung | 5 (4.1) | 2,159 (10.9) | 0.01 |
| NHL | 5 (4.1) | 1,209 (6.1) | 0.32 |
| Thyroid | 4 (3.3) | 796 (4.0) | 0.65 |
| Stomach | 4 (3.3) | 749 (3.7) | 0.77 |
| Colon | 3 (2.4) | 2,412 (12.2) | 0.0008 |
| Brain | 3 (2.4) | 343 (1.7) | 0.54 |
| Uterus | 3 (2.4) | 727 (3.6) | 0.47 |
| Leukaemia | 2 (1.6) | 730 (3.7) | 0.21 |
| MDS | 2 (1.6) | 335 (1.6) | 1.00 |
| MM | 2 (1.6) | 395 (2.0) | 0.74 |
| Othera | 10 (8.2) | – | |
| Total | 121 (100) | 27,303 (100) | |
| aOther included larynx, malignant fibrous histiocytoma, soft tissue, ureter, penile cancer, pancreas, parotid gland, and unknown. MM: multiple myeloma; MDS: myelodysplastic syndrome; SCC: squamous cell carcinoma; NHL: non-Hodgkin lymphoma. |
|||
It should be noted that Israel’s National Cancer Registry does not collect information on squamous cell skin cancers. Among the 107 patients with malignancies, 93 had solid malignancies, 10 had haematological malignancies, and 4 had both solid and haematological malignancies. Compared with the general population, solid malignancies accounted for a larger proportion of all malignancies (solid and haematological) in patients with BP (91.1% vs 87%).
In this study patients with BP in the malignancy group were older and had higher mortality rates than the non-malignancy group. The literature describes BP as a disease that primarily affects older individuals (1). The average mean age of the patients in the current study was 77.5 ± 1.2 years, which is within the previously reported range of 76–80 years. There was a significant difference in the age at BP onset between the groups with and without associated malignancy, with patients in the malignancy group being much older than those in the non-malignancy group (79.2 ± 10.1 vs 76.7 ± 12.8 years, respectively, p = 0.02). As both BP and several common malignancies become more prevalent with age, these findings might be attributed to either the increasing frequency of malignancy or a demographic characteristic of BP associated with malignancy. Previously, contradictory results were obtained in patients with BP, with some studies indicating an increase in the prevalence of malignancy with age, and others not (14–19, 24, 25). This could be related to variation in the occurrence of a range of cancers among the different populations. Moreover, the current study found that, in BP patients with associated malignancy, the case mortality rate was nearly 4 times higher. This could be because the patients in the subgroup with concomitant malignancies were older.
Patients diagnosed with cancer before developing BP had a mortality rate more than twice as high as that of patients who were diagnosed with BP first. A possible explanation is that, in this study, the length of time spent with existing malignancy was longer in the group with malignancy before BP onset than in the group with BP prior to cancer (9.8 vs 2.6 years).
In addition, patients with BP had a higher incidence of melanoma, but no higher overall malignancy incidence. The fact that there were no differences in the overall malignancy rates between patients with BP and the general population is consistent with those of earlier studies, although inconsistent with those of others (14–19). Consistent with previous findings, the current study showed that patients with BP had a significantly higher rate of melanoma than that in the general population (10.7% vs 4.3%, respectively, p = 0.0005) (33). There are several suggested mechanisms underlying this association, including keratinocyte dysfunction and increased anti-BP230 autoantibodies, all of which have been observed in patients with melanoma, and human leukocyte antigen polymorphisms linked to both melanoma and BP (HLA-DQB1*03:01) (33–36). Melanoma has also recently been recognized as a risk factor for BP development after treatment with Immune checkpoint inhibitors (ICIs) (33). Notably, none of the patients in the current cohort received treatment with ICIs. In this study, 61.5% and 38.5% of the melanoma cases occurred before and after BP diagnosis, respectively. These findings indicate a causal associationbetween BP and melanoma (not only therapy) and suggest that routine melanoma screening should be considered in patients with BP.
This study found that patients with BP have a lower prevalence of lung and colon cancers than in the general population (4.1% vs 10.9% and 2.4% vs 12.2%, respectively). According to the Israel National Registry of Cancer Mortality data 2021, lung cancer is the leading cause of cancer mortality in Israel, with a low 5-year survival rate and 75% of all fatalities occurring in people aged ≥ 50 years. According to the same data, colorectal cancer is the second leading cause of cancer death in Israel, with a mortality rate close to the global mean rate, and older individuals accounting for most cancer deaths. Colorectal cancer mortality rates have also been reported to increase after the age of 55 years, peaking in patients aged ≥ 75 years. It is reasonable to assume that the prevalence of lung and colorectal cancers in the current study was understated; some of the patients with cancer may not have been referred to a dermatologist, but instead were treated by an oncologist. This is also true for other types of malignancy for which no association was found, as further discussed in the study limitations section.
In addition, the current study found that a vast majority of the cancers appeared before the diagnosis of BP. Researchers have suggested a link between BP and cancer and several pathogenic mechanisms have been proposed, most of which are reliant on the sequence in which the disorders appear; In cases where malignancy preceded the onset of BP, a hypothesis of paraneoplastic autoimmunity emerging from the disruption of central and peripheral immunity, as well as self-antigen modification, has been proposed (26). Consequently, antibodies targeting tumour-specific antigens may cross-react with basement membrane antigens, such as BP antigens, causing skin blistering (26). One putative antigen is laminin-332, a protein involved in keratinocyte-basal cell membrane adhesion, which is produced by a variety of solid malignancies (breast, pancreatic, colon, and lung cancers). Putative pathogenetic processes include tumour-produced anti-BMZ cross-reactive antigens and the creation of molecules that disrupt the basal membrane, resulting in anti-BMZ antibody production (27).
Anticancer treatments, radiotherapy, and surgery, have also been described as triggers for bullous dermatoses (28–30). When cancer develops following the onset of an autoimmune bullous disorder, it has been hypothesized that the constant activation of the immune system, the creation of proinflammatory cytokines, tissue damage, and extended inflammation all contribute to carcinogenesis (31). Chronic immunosuppressive therapy has also been linked to an increased risk of cancer in patients with BP (32). In the current study, malignancy preceded BP manifestation in most patients (77%). Thus, BP as a result of an immune response to tumour cells or anticancer therapy may be a more common scenario than BP as a primary cause of carcinogenesis. Among the anticancer drugs, ICIs are specifically linked to BP (28), as pointed out previously. The pathomechanism is not entirely known; however, M2 macrophages may play a role (29). The time between drug exposure and BP onset varies, ranging from early onset following treatment initiation to late onset after treatment discontinuation (28). Radiotherapy may cause BP by altering the antigenic properties of the basement membrane and by inducing autoantibody production (30). Notably, none of the patients in the current cohort had had any radiotherapy before BP onset.
The rate of immunosuppressive therapy differed significantly between the groups, with the malignancy group receiving immunosuppressive medication (methotrexate and/or azathioprine) at a lower rate (p = 0.02). This could be related to their poorer overall health, desire to avoid tumour aggravation, or concurrent anticancer treatment that prevented additional immunosuppression. It is also possible that BP associated with cancer may resolve following anticancer treatment, precluding the need for these immunosuppressive medications (37). There were no differences in clinical characteristics, symptoms, or illness severity between patients with and without cancer in the current study cohort. This is in contrast to several previous reports of unusual clinical BP variants (atypical presentation and refractory disease) in patients with associated malignancies (20, 21, 38–40).
This study has several limitations. There was no information regarding the patients’ anti-BP180 and anti-BP230 antibody titres, since indirect immunofluorescence (IIF) and enzyme-linked immunoassay (ELISA) are not part of the routine BP workup in our centre. Furthermore, because the data were obtained from a single tertiary centre, the sample size was limited, and there is a possibility of selection bias. While dermatologists refer most patients with severe BP to tertiary centres where they are followed closely, it is possible that less severe cases are treated in the community or by oncologists without being referred to a dermatologist, and therefore without a definitive diagnosis of BP. This can result in an underestimation of the association between BP and malignancy, especially in non-skin cancers; while most patients with skin cancer most probably see a dermatologist, patients with non-skin cancer may not be evaluated by a dermatologist, leading to an underestimation of the association between their non-skin cancer and BP. The strengths of the current study include its well-defined cohort, long-term follow-up, and reliance on data files containing complete data on clinical features and treatments provided by expert dermatologists. Several population-based studies lack complete clinical information and are prone to classification bias.
A comprehensive workup for solid or haematological malignancy in patients with BP is considered unnecessary, as no higher risk of general malignancy was found in patients with BP comperd with the general population. The sample of patients with BP had a higher prevalence of melanoma, indicating that routine skin screening may be recommended. Furthermore, large sample, multicentre prospective studies are warranted.
The study was approved by the institutional review board of Sheba Medical Center (7172-09-SMC).