Association Between Medication Use and Bullous Pemphigoid: A Systematic Review and Meta-analysis

Sian-De Liu; Wei-Ti Chen; Ching-Chi Chi

doi:10.1001/jamadermatol.2020.1587

. 2020 Jun 17;156(8):1–11. doi: 10.1001/jamadermatol.2020.1587

Association Between Medication Use and Bullous Pemphigoid

A Systematic Review and Meta-analysis

Sian-De Liu ¹, Wei-Ti Chen ^2,^3,^4,⁵, Ching-Chi Chi ^2,^4,^5,^✉

PMCID: PMC7301306 PMID: 32584924

This systematic review and meta-analysis assesses the association of the use of certain classes of medications among patients and the development of bullous pemphigoid.

Key Points

Question

Is there an association between use of medications and the development of bullous pemphigoid?

Findings

In this systematic review and meta-analysis of 13 case-control studies, 1 cohort study, and 1 randomized clinical trial with 285 884 participants, there was a significant association of the development of bullous pemphigoid with the prescribed use of aldosterone antagonists, dipeptidyl peptidase 4 inhibitors, anticholinergics, and dopaminergic medications.

Meaning

The findings of this systematic review and meta-analysis suggest that medications should be prescribed judiciously, particularly in high-risk patients who are elderly and have disabling neurologic disorders.

Abstract

Importance

The association between the use of medications and the development of bullous pemphigoid (BP) is unclear.

Objective

To assess the associations between previous exposure to certain medications and BP.

Data Sources

For this systematic review and meta-analysis, PubMed, the Cochrane Central Register of Controlled Trials, and Embase were searched for relevant studies from inception to February 20, 2020.

Study Selection

Case-control or cohort studies and randomized clinical trials that examined the odds or risk of BP in patients with previous medication use were included. No geographic or language limitations were imposed.

Data Extraction and Synthesis

The Meta-analysis of Observational Studies in Epidemiology (MOOSE) guideline was followed. The Newcastle-Ottawa Scale was used to evaluate the risk of bias of included observational studies; Cochrane Collaboration’s tool was used for randomized clinical trials. Aggregate data were used to conduct a random-effects model meta-analysis if the included studies were sufficiently homogenous. Subgroup analyses were performed for use of various medications of the same category.

Main Outcomes and Measures

Odds ratio (OR), hazard ratio, and risk ratio of bullous pemphigoid in association with medication use.

Results

This meta-analysis included 13 case-control studies, 1 cohort study, and 1 randomized clinical trial with a total of 285 884 participants. The meta-analysis of case-control studies showed a significant association of BP with previous use of aldosterone antagonists (pooled OR, 1.75; 95% CI, 1.28-2.40), dipeptidyl peptidase 4 inhibitors (pooled OR, 1.92; 95% CI, 1.55-2.38), anticholinergics (pooled OR, 3.12; 95% CI, 1.54-6.33), and dopaminergic medications (pooled OR, 2.03; 95% CI, 1.34-3.05). One cohort study found an increased risk of BP among patients receiving dipeptidyl peptidase 4 inhibitors (hazard ratio, 2.38; 95% CI, 1.16-4.88; P = .02). One trial found a higher occurrence of BP in patients with diabetes receiving linagliptin (0.2% in diabetes group vs 0% in the placebo group).

Conclusions and Relevance

The findings of this systematic review and meta-analysis suggest that aldosterone antagonists, dipeptidyl peptidase 4 inhibitors, anticholinergics, and dopaminergic medications are associated with BP. These medications should be judiciously prescribed, particularly in high-risk patients who are elderly and have disabling neurologic disorders.

Introduction

Bullous pemphigoid (BP) is the most common autoimmune blistering disease, characterized by subepidermal separation and inflammation with abundant eosinophils leading to tense bullae.^1,2 Autoantibodies targeting 2 main structural proteins of the dermal-epidermal junction, BP antigen 1 (BPAG1 or BP230 antigen) and BPAG2 (or termed BP180 antigen), are involved in the pathogenesis of BP.^3,4 Bullous pemphigoid has been associated with neuropsychiatric diseases including cerebrovascular disease and dementia.^5,6 The annual incidence of BP was estimated to range from 2.4 to 21.7 per million population in different populations worldwide.^{7,8,9,10,11,12,13,14} Various studies have shown an increase in the incidence of BP of 1.9- to 4.3-fold compared with that previously reported over the past 2 decades.^{11,12,15,16,17} The possible explanations for this surge are increasing life expectancy of the populations, increasing use of certain medications, increasing awareness of the atypical variants of BP, and better diagnostic methods.^18,19,20

Many studies have suggested potential associations of medication use with BP.^21,22,23 Certain medications could induce antibody production by acting as haptens to bind to proteins in the lamina lucida, and some may unmask hidden antigens or stimulate an autoimmune response.²⁴ However, the levels of evidence for most suspected medications were low because studies have lacked control groups.^{25,26,27,28,29,30,31,32} Two case-control studies examined use of diuretics but reported conflicting data on the association between loop diuretics and BP.^21,33 To our knowledge, there have been no comprehensive studies to clarify the association between use of various medications and BP. To close this gap in knowledge, we conducted a meta-analysis to systematically acquire and evaluate the evidence on the association between medication use and BP.

Methods

Eligibility Criteria and Evidence Search

In accordance with the Meta-analysis of Observational Studies in Epidemiology (MOOSE) reporting guideline,³⁴ we conducted a aystematic review and meta-analysis on the association between use of various medications and BP. We searched the PubMed, Cochrane Central Register of Controlled Trials, and Embase databases from their inception to February 20, 2020. We also checked the bibliographies of relevant reviews for potentially eligible studies. No limitations on publication status, language, and geographic regions were imposed. The search strategy is listed in eTable 1 in the Supplement. This meta-analysis has been registered with PROSPERO (CRD42018115041).³⁵

Selection of Studies

Case-control studies, cohort studies, and randomized clinical trials (RCTs) were eligible. For eligible case-control studies, the case group was composed of patients with BP, the control group was composed of individuals without BP, and the outcome was the odds ratio (OR) of use of medications in association with BP. For eligible cohort studies or RCTs, the exposed or experimental group was composed of participants who received the study medications, the unexposed group was composed of people without use of the study medications, and the outcome was the hazard ratio (HR) or risk ratio (RR) of BP. We excluded studies on chemotherapeutic agents and checkpoint inhibitors for patients with cancer because the immune system in these participants was complex and data pooling of persons without cancer would be inappropriate. Two of us (S.-D. L., W.-T.C.) independently selected relevant studies by scanning the titles and abstracts of search results. The full text of potential studies was obtained and examined for eligibility. Disagreement was resolved by discussion with a another one of us (C.-C.C.).

Data Extraction and Risk of Bias Assessment

Data including author name, publication year, country, setting, data source, medication name, and proportion of participants with BP were extracted. We also extracted age, percentage of female participants, diagnostic methods for BP, and the effect estimates between medications and BP including OR, HR, or RR with corresponding 95% CI. When studies involved the same data source population, we included the one with the longest period of study. One of us (S.-D. L.) was responsible for data extraction, and another (W.-T. C.) confirmed the accuracy of data.

The Newcastle-Ottawa Scale was used in the risk of bias assessment of included observational studies.³⁶ For included case-control studies, we evaluated the following 8 domains: (1) adequacy of case definition, (2) representativeness of cases, (3) selection of controls, (4) definition of controls, (5) comparability of cases and controls, (6) ascertainment of exposure, (7) same method of ascertainment for cases and controls, and (8) nonresponse rate.³⁶ For included cohort studies, we evaluated the following 8 domains: (1) representativeness of the exposed cohort, (2) selection of the nonexposed cohort, (3) ascertainment of exposure, (4) outcome of interest not present at the start of study, (5) comparability of cohort, (6) assessment of outcome, (7) follow-up duration, and (8) adequacy of follow-up of cohorts.³⁶ For included RCTs, we used the Cochrane Collaboration’s tool to assess their risk of bias, which contains 6 domains including (1) randomization process, (2) deviations from intended interventions, (3) missing outcome data, (4) measurement of the outcome, (5) selection of the reported result, and (6) overall bias.³⁷

Statistical Analysis

Effect estimates were expressed as a pooled OR with 95% CI for case-control studies, a pooled HR with 95% CI for cohort studies, and a pooled RR with 95% CI for RCTs. Random-effects model meta-analyses were conducted because we assumed considerable clinical heterogeneity across included studies³⁸ (eg, different compounds in the same pharmacologic category and various regimens, populations, and settings). The I² statistic was calculated to assess the statistical heterogeneity between studies, with an I² of more than 50% indicating substantial heterogeneity.³⁹ Because neuropsychiatric diseases per se have been associated with BP,⁵ we performed a sensitivity analysis to examine the association of BP with use of psycholeptics and anti-Parkinson medications after excluding studies that did not adjust for the presence of neuropsychiatric diseases. The Review Manager, version 5.3 (The Nordic Cochrane Centre, The Cochrane Collaboration) was used for conducting meta-analyses. A 2-sided P < .05 defined statistical significance.

Results

Study Selection

As shown in Figure 1, our search identified 3230 citations. After scanning the titles and abstracts, the full text of 148 articles were examined. Eventually we included 13 case-control studies,^{15,21,22,33,40,41,42,43,44,45,46,47,48} 1 cohort study,⁴⁹ and 1 RCT⁵⁰ with a total of 285 884 participants.

Characteristics of Included Studies

The Table summarizes characteristics of included studies. Included studies were from France (n = 4),^33,40,41,43 Finland (n = 3),^46,47,48 Korea (n = 1),¹⁵ Singapore (n = 1),⁴⁵ Switzerland (n = 1),⁴⁴ Taiwan (n = 1),⁴⁹ and the UK (n = 1).²¹ A multicenter RCT was conducted across 27 countries.⁵⁰ The sample size ranged from 36 to 227 187 participants with a mean (SD) range of age from 65.6 (9.1) years to 82.4 (8.7) years. The numbers of cases and controls as well as the ORs (95% CI) for each class of medications are shown in eTables 2 to 13 in the Supplement.

Table. Characteristics of Included Studies.

Source	Study design, participants, No.	Female, % (age, mean [SD])	BP definition	Data sourced or medication studied	Control group	Multivariate adjustment factors	Newcastle-Ottawa Scale score^a
Source	Study design, participants, No.	Female, % (age, mean [SD])	BP definition	Data sourced or medication studied	Control group	Multivariate adjustment factors	Selection	Comparability	Exposure	Outcome
Bastuji-Garin et al,⁴⁰ 1996 (France)	Case-control, 332	Cases: 50.0 (79.2 [10.1]); controls: 50.0 (77.0 [10.9])	Clinical and immunopathologic features	Multicenter questionnaire	Patients with skin tumors matched for sex and age within 5 y at dermatology clinic	Number of medications, diuretics, and neuroleptics	3	2	1	NA
Bastuji-Garin et al,³³ 2011 (France)	Case-control, 546	Cases: 64.3 (82.4 [8.7]); controls: 61.5 (82.2 [8.6])	Clinical and immunopathologic features	Multicenter questionnaire	Patients with skin tumors or acute conditions matched for sex, age within 5 y, center, and place of residence	Age, sex, and place of residence	3	2	1	NA
Nozu and Mita,⁴² 2010 (Japan)	Case-control, 36	Cases: 80.0 (87.8 [3.2]); controls: 38.7 (83.3 [1.7])	Clinical and immunopathologic features	Medical record review from a hospital	Disabled inpatients owing to cerebral stroke	NA	1	0	2	NA
Lloyd-Lavery et al,²¹ 2013 (UK)	Case-control, 220	Cases: 59.3 (81.5 [9.7]); controls: 49.3 (78.3 [8.5])	Clinical and immunopathologic features	Medical record review from hospitals	Patients with skin tumors matched for sex and age within 5 y at dermatology clinic	Sex, age, cerebrovascular disease, and dementia as well as patient comorbidities for drugs that were significantly associated with BP in the univariate analysis	3	2	2	NA
Schaffer et al,⁴⁴ 2017 (Switzerland)	Case-control, 193	Cases: 52.2 (77.6 [not reported]); controls: 33.5 (76.5 [not reported])	Clinical and immunopathologic features	Hospital database	Patients with diabetes matched for age at endocrinology clinic	Sex, age, and calendar year	2	1	2	NA
Tan et al,⁴⁵ 2017 (Singapore)	Case-control, 420	Cases: 50.0 (78.0 [11.2]); controls: (76.0 [12.6])	Established criteria with corresponding histologic and immunologic investigations	Hospital record review	Matched for sex and age within 5 y at dermatology clinic	NA	3	1	2	NA
Benzaquen et al,⁴¹ 2018 (France)	Case-control, 183	Cases: 50.8 (78.7 [7.0]); controls: 50.8 (79.3 [7.0])	Clinical and immunopathologic features	Medical record review from 3 hospitals	Patients matched for sex and age within 5 y at endocrinology clinic	Neurologic, metabolic or endocrine, and other dermatologic comorbidities	3	2	2	NA
Kridin and Bergman,²² 2018 (Israel)	Case-control, 410	Cases: 53.7 (79.1 [9.1]); controls: 53.7 (79.0 [9.0])	Clinical and immunopathologic features	Hospital medical record review	Patients with diabetes and with hernia or skin lesion matched for sex, age, year of diagnosis, and ethnicity	Neurologic diseases and metformin intake	4	2	2	NA
Lee et al,¹⁵ 2019 (Korea)	Case-control, 1340	Cases: 49.0 (75.3 [10.0]); controls: 49.0 (75.3 [10.0])	ICD-10 code L12.0	National insurance database	Patients with diabetes matched for sex, age, and year of diagnosis	Neurologic, malignant, and psychological comorbid disorders and receipt of spironolactone and psycholeptics	3	2	2	NA
Rosenstock et al,⁵⁰ 2019 (27 countries)	Randomized clinical trial, 6979	Treatment: 38.5 (66.1 [9.1]); controls: 35.7 (65.6 [9.1])	According to clinical investigations and diagnosis in the centers	Linagliptin, 5 mg/d	Placebo once daily added to usual care	NA	NA	NA	NA	NA
Varpuluoma et al,^46,47,48 2018, 2019 (Finland)	Case-control, 16 338	Cases: 59.7 (76.6 [not reported]); controls: 60.0 (76.7 [not reported])	ICD-9 codes 694.5A and 694.5B; ICD-10 code L12.0	National insurance database	Patients with basal cell carcinoma matched for sex, age, and year of diagnosis within 2 y	Diabetes, Alzheimer disease, vascular dementia, other or unspecified dementia, Parkinson disease, multiple sclerosis, subarachnoid hemorrhage, intracerebral hemorrhage, cerebral infarction, and epilepsy	2	2	2	NA
Hung et al,⁴⁹ 2020 (Taiwan)	Cohort, 31 700	Exposure: 48.2 (66.0 [11.9]); controls: 78.8 (66.0 [11.8])	ICD-9-CM diagnosis code: 694.5	National insurance database	Matched for sex, age, and index year	Age, sex, comorbidity (Charlson Comorbidity Index), season, location, urbanization level, and level of care	3	2	NA	3
Plaquevent et al,⁴³ 2019 (France)	Case-control, 227 187	Cases: 46.3 (77.9 [9.3]); controls: NA	Clinical and immunopathologic features	Cases: medical record review from 21 departments; references: national insurance database	Indirect age standardization on the randomly sampled general population with stratification on sex and age	NA	3	1	1	NA

Open in a new tab

Abbreviations: BP, bullous pemphigoid; ICD-9, International Classification of Diseases, Ninth Revision; ICD-9-CM, International Classification of Diseases, Ninth Revision, Clinical Modification; ICD-10, International Statistical Classification of Diseases and Related Health Problems, Tenth Revision; NA, not available.

^{^a}

Newcastle-Ottawa Scale (total points) with selection, comparability, and exposure or outcome domains, in which the total score ranges from 0 (lowest quality) to 9 points (highest quality).

Risk of Bias in Included Studies

The risk of bias for included studies is presented in the Table. Most studies did not mention the history of outcome for the control group; thus, they were judged as having an unclear risk of bias.^{21,33,40,42,43,44,45,46,47,48} Five studies defined cases with only the International Classification of Diseases diagnosis codes in the database and thus were rated with unclear risk of bias.^{15,44,46,47,48} For comparability, 3 studies were rated with an unclear risk of bias because there was no adjustment for confounders except sex and age.^43,44,45 Based on exposure to medications, all forms of medical records could not be viewed as medication consumption in reality; thus, most case-control studies were rated as an unclear risk of bias in the exposure domain.^{15,21,22,33,40,41,43,44,45,46,47,48} An included RCT demonstrated low bias risk in each domain of risk of bias.⁵⁰

Association of Medication Use With BP

Diuretics

The data for use of diuretics are shown in Figure 2. Five case-control studies with 2474 participants examined the association between use of aldosterone antagonists and BP.^{15,21,33,40,42} Use of aldosterone antagonists was significantly associated with BP (pooled OR, 1.75; 95% CI, 1.28-2.40; I² = 4%). Five case-control studies with 2638 participants provided data on the association of use of thiazides or loop diuretics with BP.^{15,21,33,40,45} We found no significant associations of BP with thiazides (pooled OR, 1.16; 95% CI, 0.93-1.44; I² = 0%) and loop diuretics (pooled OR, 1.29; 95% CI, 0.88-1.90; I² = 55%).

Antidiabetic Medications

The data for use of antidiabetic medications are shown in Figure 3. Six case-control studies with 245 651 participants examined the association between use of dipeptidyl peptidase 4 (DPP-4) inhibitors and BP.^{15,22,41,43,44,47} Use of DPP-4 inhibitors was significantly associated with BP (pooled OR, 1.92; 95% CI, 1.55-2.38; I² = 46%). The sensitivity analysis excluding studies that lacked adjustment for neurologic diseases^43,44 showed a significant association of BP with use of DPP-4 inhibitors (pooled OR, 2.14; 95% CI, 1.55-2.96; I² = 58%) (eFigure 10 in the Supplement).

Three case-control studies with 17 168 participants examined the association between use of metformin and BP.^22,45,46 We found no significant associations of BP with metformin (pooled OR, 1.00; 95% CI, 0.81-1.23; I² = 8%). The sensitivity analysis excluding the study by Tan et al,⁴⁵ which lacked adjustment for neurologic diseases, also showed no significant association (pooled OR, 1.02; 95% CI, 0.87-1.21; I² = 0%) (eFigure 10 in the Supplement).

Two case-control studies with 16 758 participants found no significant association between use of sulfonylureas and BP (pooled OR, 0.98; 95% CI, 0.78-1.22; I² = 0%).^45,47 The sensitivity analysis excluding the study by Tan et al,⁴⁵ which lacked adjustment for neurologic diseases, also revealed no significant association (OR, 0.99; 95% CI, 0.79-1.24) (eFigure 10 in the Supplement). A case-control study with 16 338 participants provided data on the association of use of thiazolidinediones and use of glucagon-like peptide 1 analogues with BP.⁴⁷ No significant associations of BP with thiazolidinediones (OR, 1.16; 95% CI, 0.63-2.14) and glucagon-like peptide 1 analogues (OR, 0.95; 95% CI, 0.11-8.52) were found.

A cohort study with 31 700 participants examined the association between use of DPP-4 inhibitors and the risk of BP.⁴⁹ Treatment with DPP-4 inhibitors was significantly associated with the risk of BP (adjusted HR, 2.38; 95% CI, 1.16-4.88; P = .02) among patients with diabetes. One RCT on 6991 adult patients with diabetes found a BP occurrence rate of 0.2% (7 of 3494 patients) in the linagliptin group but of 0% (0 of 3485 patients) in the placebo group.⁵⁰

Psycholeptics

As shown in eFigure 1 in the Supplement, 5 case-control studies with 1317 participants examined the association between BP and use of psycholeptics (Anatomical Therapeutic Chemical Classification System code N05).^{21,33,40,41,42} Use of antipsychotics (N05A) was significantly associated with BP (pooled OR, 1.63; 95% CI, 1.15-2.32; I² = 0%). The sensitivity analysis excluding 3 studies that did not adjust for neuropsychiatric diseases^33,40,42 showed no significant association of BP with use of antipsychotics (pooled OR, 1.54; 95% CI, 0.68-3.50; I² = 26%) (eFigure 11 in the Supplement).

Three case-control studies with 1098 participants examined the association between use of anxiolytics (N05B) and BP.^21,33,40 Use of anxiolytics (N05B) was not significantly associated with BP (pooled OR, 1.16; 95% CI, 0.65-2.06; I² = 57%). The sensitivity analysis excluding 2 studies that did not adjust for neuropsychiatric diseases^33,40 found no significant association of BP with use of anxiolytics (OR, 2.80; 95% CI, 0.60-13.07) (eFigure 11 in the Supplement).

A case-control study with 546 participants provided data on the association between use of sedatives (N05C) and BP but did not adjust for neuropsychiatric diseases in statistical analysis.³³ Use of sedatives (N05C) was not significantly associated with BP (OR, 1.16; 95% CI, 0.70-1.92).

Anti-Parkinson Medications

The data for use of anti-Parkinson medications are shown in Figure 4. Two case-control studies with 16 884 participants examined the association between use of anticholinergics and BP.^33,48 Use of anticholinergics was associated with BP (pooled OR, 3.12; 95% CI, 1.54-6.33; I² = 0%). The sensitivity analysis after excluding the study by Bastuji-Garin et al,³³ which lacked adjustment for neurologic diseases, revealed a significant association of BP with use of anticholinergics (OR, 2.94; 95% CI, 1.42-6.09) (eFigure 12 in the Supplement)

Three case-control studies with 17 304 participants examined the association between use of dopaminergic medications and BP.^33,45,48 Use of dopaminergic medications was associated with BP (pooled OR, 2.03; 95% CI, 1.34-3.05; I² = 0%). The sensitivity analysis after excluding 2 studies that did not adjust for neurologic diseases showed a significant association of BP with use of dopaminergic medications (OR, 1.85; 95% CI, 1.10-3.11) (eFigure 12 in the Supplement)

Analgesics

The data for use of analgesics are shown in eFigure 2 in the Supplement. A case-control study with 546 participants provided data on the association between use of opioids and salicylates with BP.³³ This study reported negative associations of the use of opioids (OR, 0.50; 95% CI, 0.29-0.86) and salicylates (OR, 0.55; 95% CI, 0.31-0.98) with BP.

Antihypertensive Medications

The data for use of antihypertensive medications are shown in eFigure 3 in the Supplement. Five case-control studies with 2858 participants examined the association between use of angiotensin-converting enzyme inhibitors and BP.^{15,21,33,40,45} Use of angiotensin-converting enzyme inhibitors was not significantly associated with BP (pooled OR, 1.14; 95% CI, 0.96-1.35; I² = 0%). Three case-control studies with 2306 participants examined the association between use of angiotensin II receptor blockers and BP.^15,33,45 Use of angiotensin II receptor blockers was not significantly associated with BP (pooled OR, 1.22; 95% CI, 0.99-1.49; I² = 0%). A case-control study with 420 participants provided data on the association between use of α-blockers and BP.⁴⁵ Use of α-blockers was not significantly associated with BP (OR, 0.65; 95% CI, 0.22-1.98). Five case-control studies with 2858 participants examined the association between use of β-blockers and BP.^{15,21,33,40,45} Use of β-blockers was not significantly associated with BP (pooled OR, 1.07; 95% CI, 0.77-1.48; I² = 43%). Five case-control studies with 2858 participants examined the association between use of calcium channel blockers and BP.^{15,21,33,40,45} Use of calcium channel blockers was not significantly associated with BP (pooled OR, 1.09; 95% CI, 0.86-1.39; I² = 21%). Two case-control studies with 878 participants provided data on the associations between use of central antihypertensive medications and vasodilators with BP.^33,40 No significant associations of BP with central antihypertensive medications (pooled OR, 1.45; 95% CI, 0.83-2.53) and vasodilators (pooled OR, 0.82; 95% CI, 0.55-1.23) were found.

Antithrombotics

The data for antithrombotic use are shown in eFigure 4 in the Supplement. Three case-control studies with 1186 participants examined the association between use of aspirin and BP.^21,33,45 Use of aspirin was not significantly associated with BP but with substantial heterogeneity (pooled OR, 1.04; 95% CI, 0.46-2.35; I² = 72%). A case-control study²¹ with 220 participants provided data on the association of warfarin use and clopidogrel use with BP. No significant associations of BP with warfarin (OR, 0.77; 95% CI, 0.25-2.32) and clopidogrel (OR, 11.28; 95% CI, 0.58-221.06) were found. Two case-control studies with 640 participants examined the association between use of dipyridamole and BP.^21,45 Use of dipyridamole was not significantly associated with BP (pooled OR, 5.58; 95% CI, 0.73-42.56; I² = 0%).

Lipid-Lowering Agents

The data for use of lipid-lowering agents are shown in eFigure 5 in the Supplement. Five case-control studies with 2709 participants examined the association between use of statins and BP.^{15,21,33,41,45} Use of statins was not significantly associated with BP (pooled OR, 0.94; 95% CI, 0.77-1.15; I² = 7%). Two case-control studies with 966 participants examined the association between use of fibrates and BP.^33,45 Use of fibrates was not significantly associated with BP (pooled OR, 0.50; 95% CI, 0.22-1.18; I² = 0%).

Antidepressants

The data for use of antidepressants are shown in eFigure 6 in the Supplement. Two case-control studies with 640 participants examined the association between use of tricyclic antidepressants and BP and between use of selective serotonin reuptake inhibitors and BP.^21,45 No significant associations of BP with tricyclic antidepressants (pooled OR, 1.24; 95% CI, 0.48-3.18) and selective serotonin reuptake inhibitors (pooled OR, 3.56; 95% CI, 0.83-15.20) were found.

A case-control study²¹ with 220 participants provided data on the association between use of tetracyclic antidepressants and BP and on the association between use of serotonin norepinephrine reuptake inhibitors and BP. Use of tetracyclic antidepressants was not associated with BP (OR, 4.81; 95% CI, 0.49-46.98) nor was use of serotonin norepinephrine reuptake inhibitors (OR, 6.49; 95% CI, 0.71-59.05).

Nonsteroidal Anti-inflammatory Drugs

The data for use of nonsteroidal anti-inflammatory drugs (NSAIDs) are shown in eFigure 7 in the Supplement. Four case-control studies with 2401 participants examined the association between use of NSAIDs and BP.^15,33,40,41 Use of NSAIDs was not significantly associated with BP (pooled OR, 1.08; 95% CI, 0.65-1.79; I² = 35%).

Antibiotics

The data for use of antibiotics are shown in eFigure 8 in the Supplement. Three case-control studies with 1098 participants examined the association between use of antibiotics and BP.^21,33,40 Use of antibiotics was not significantly associated with BP (pooled OR, 1.71; 95% CI, 0.99-2.95; I² = 4%).

Gastrointestinal Tract Medications

The data for use of gastrointestinal tract medications are shown in eFigure 9 in the Supplement. Three case-control studies with 823 participants examined the association between use of proton pump inhibitors and BP.^21,41,45 Use of proton pump inhibitors was not significantly associated with BP (pooled OR, 1.01; 95% CI, 0.66-1.55; I² = 0%). Two case-control studies of 640 participants found no association between use of histamine-2 receptor antagonists and BP (OR, 1.52; 95% CI, 0.51-4.55).^21,45

Discussion

To our knowledge, this meta-analysis is the first to comprehensively examine the associations between use of various medications and BP. The evidence from case-control studies showed significant associations of BP with use of aldosterone antagonists, DPP-4 inhibitors, anticholinergics, and dopaminergic medications. The sensitivity analysis excluding studies that did not adjust for neuropsychiatric diseases showed similar results. In addition, the evidence from 1 cohort study revealed a 2.38-fold HR for BP among patients receiving DPP-4 inhibitors compared with controls.⁴⁹ One RCT found similar results for linagliptin.⁵⁰

A number of pathogenetic hypotheses^{51,52,53,54,55,56,57} have been proposed to elucidate the mechanisms of medication-induced BP. Many studies^24,53,54,55 found that medications could disrupt immune balance and consequently induce antibasement membrane antibody formation, such as anti-BP180 and anti-BP230 autoantibodies. First, cross-reactivity to the medication with the same or similar chemical structure occurred in a susceptible person who had already been sensitized to another medication with a certain chemical structure. Therefore, previous medication exposure may be associated with the increasing prevalence of BP, especially among elderly people, who are more likely to receive multiple medications.^51,52 Second, molecular mimicry is also involved in immune pathogenesis of medication-induced BP. By means of binding to micro RNA and other transcriptional and translational regulators, medications may be misidentified as non–self-antigens by the immune system, leading to activation of CD4+ T cells and initiation of the autoimmune cascade.^53,54 Third, medications may also cause the inactivation of endogenous regulatory processes, leading to an immune dysregulation. For example, a reduction in number, function, or skin homing of Foxp3-expressing CD4+ CD25+ T-regulatory cells was found in patients with BP, which could cause uncontrolled B-cell clones to release autoantibodies against the basement membrane.^24,55 Fourth, specific medications (eg, furosemide) could potentially change antigenic properties in the basement membrane by binding to protein molecules and cause the exposure of a hidden antigenic site in the basement membrane.^30,51 Conversely, medications containing sulfur in the molecule and thiol formation after xenobiotic metabolism may have the ability to directly disrupt the dermo-epidermal junction by interacting with the sulfhydryl groups in the desmosome.^56,57

A study by Bastuji-Garin et al³³ reported negative associations of use of opioids and salicylates with BP. The reasons for the negative associations are unclear, and further studies are warranted for confirmation.

Limitations

This study has some limitations. First, there was only 1 RCT that examined the association between BP and previous medication exposure. However, it is unlikely that there would be plentiful RCTs that examine the association between use of various medications and BP. Second, a limitation may be inherent to the use of International Classification of Diseases diagnosis codes in identifying BP because validation in the diagnosis of BP beyond medical records was not practically achievable in a study by Varpuluoma et al.^46,47,48 However, after exclusion of this study, the associations between DPP-4 inhibitor use and BP remained significant. Third, it may be arguable that patients with more severe neurologic disease are more likely to take medications and thus are more likely to develop BP. However, there is a lack of studies examining the association of BP with various severity of Parkinson disease.

Conclusions

In this study, an association of BP with use of aldosterone antagonists, DPP-4 inhibitors, anticholinergics, and dopaminergic medications was found. These findings suggest that caution should be exercised about the use of these medications, particularly in high-risk patients who are elderly and have disabling neurologic disorders.

Supplement.

eTable 1. PubMed, Cochrane Central Register of Controlled Trials, and Embase Databases Search Strategy

eTable 2. Characteristics of Included Studies for the Class of Diuretics

eTable 3. Characteristics of Included Studies for the Class of Antidiabetic Drugs

eTable 4. Characteristics of Included Studies for the Class of Psycholeptics

eTable 5. Characteristics of Included Studies for the Class of Anti-Parkinson Drugs

eTable 6. Characteristics of Included Studies for the Class of Analgesics

eTable 7. Characteristics of Included Studies for the Class of Antihypertensives

eTable 8. Characteristics of Included Studies for the Class of Antithrombotics

eTable 9. Characteristics of Included Studies for the Lipid-Lowering Agents

eTable 10. Characteristics of Included Studies for the Class of Antidepressants

eTable 11. Characteristics of Included Studies for the Class of Nonsteroidal Anti-inflammatory Drugs

eTable 12. Characteristics of Included Studies for the Class of Antibiotics

eTable 13. Characteristics of Included Studies for the Class of Gastrointestinal Tract Medications

eFigure 1. Association Between Psycholeptics and Bullous Pemphigoid

eFigure 2. Association Between Analgesics and Bullous Pemphigoid

eFigure 3. Association Between Antihypertensive Drugs and Bullous Pemphigoid

eFigure 4. Association Between Antithrombotics and Bullous Pemphigoid

eFigure 5. Association Between Lipid-Lowering Agents and Bullous Pemphigoid

eFigure 6. Association Between Antidepressants and Bullous Pemphigoid

eFigure 7. Association Between Nonsteroidal Anti-inflammatory Drugs and Bullous Pemphigoid

eFigure 8. Association Between Antibiotics and Bullous Pemphigoid

eFigure 9. Association Between Gastrointestinal Tract Drugs and Bullous Pemphigoid

eFigure 10. Sensitivity Analysis on Association Between Antidiabetic Drugs and Bullous Pemphigoid

eFigure 11. Sensitivity Analysis on Association Between Psycholeptics and Bullous Pemphigoid

eFigure 12. Sensitivity Analysis on Association Between Anti-Parkinson Drugs and Bullous Pemphigoid

Click here for additional data file.^{(2.3MB, pdf)}

References

1.Schmidt E, Zillikens D. Pemphigoid diseases. Lancet. 2013;381(9863):320-332. doi: 10.1016/S0140-6736(12)61140-4 [DOI] [PubMed] [Google Scholar]
2.Bağcı IS, Horváth ON, Ruzicka T, Sárdy M. Bullous pemphigoid. Autoimmun Rev. 2017;16(5):445-455. doi: 10.1016/j.autrev.2017.03.010 [DOI] [PubMed] [Google Scholar]
3.Hammers CM, Stanley JR. Mechanisms of disease: pemphigus and bullous pemphigoid. Annu Rev Pathol. 2016;11:175-197. doi: 10.1146/annurev-pathol-012615-044313 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.van Beek N, Zillikens D, Schmidt E. Diagnosis of autoimmune bullous diseases. J Dtsch Dermatol Ges. 2018;16(9):1077-1091. [DOI] [PubMed] [Google Scholar]
5.Kridin K, Zelber-Sagi S, Kridin M, Cohen AD. Bullous pemphigoid and neuropsychiatric medications: an influence of drugs or of underlying conditions? J Am Acad Dermatol. Published online April 17, 2019. doi: 10.1016/j.jaad.2019.03.091 [DOI] [PubMed] [Google Scholar]
6.Taghipour K, Chi C-C, Vincent A, Groves RW, Venning V, Wojnarowska F. The association of bullous pemphigoid with cerebrovascular disease and dementia: a case-control study. Arch Dermatol. 2010;146(11):1251-1254. doi: 10.1001/archdermatol.2010.322 [DOI] [PubMed] [Google Scholar]
7.Bernard P, Vaillant L, Labeille B, et al. ; Bullous Diseases French Study Group . Incidence and distribution of subepidermal autoimmune bullous skin diseases in three French regions. Arch Dermatol. 1995;131(1):48-52. doi: 10.1001/archderm.1995.01690130050009 [DOI] [PubMed] [Google Scholar]
8.Marazza G, Pham HC, Schärer L, et al. ; Autoimmune Bullous Disease Swiss Study Group . Incidence of bullous pemphigoid and pemphigus in Switzerland: a 2-year prospective study. Br J Dermatol. 2009;161(4):861-868. doi: 10.1111/j.1365-2133.2009.09300.x [DOI] [PubMed] [Google Scholar]
9.Baican A, Baican C, Chiriac G, et al. . Pemphigus vulgaris is the most common autoimmune bullous disease in Northwestern Romania. Int J Dermatol. 2010;49(7):768-774. doi: 10.1111/j.1365-4632.2009.04345.x [DOI] [PubMed] [Google Scholar]
10.Nanda A, Dvorak R, Al-Saeed K, Al-Sabah H, Alsaleh QA. Spectrum of autoimmune bullous diseases in Kuwait. Int J Dermatol. 2004;43(12):876-881. doi: 10.1111/j.1365-4632.2004.02292.x [DOI] [PubMed] [Google Scholar]
11.Bertram F, Bröcker E-B, Zillikens D, Schmidt E. Prospective analysis of the incidence of autoimmune bullous disorders in Lower Franconia, Germany. J Dtsch Dermatol Ges. 2009;7(5):434-440. doi: 10.1111/j.1610-0387.2008.06976.x [DOI] [PubMed] [Google Scholar]
12.Joly P, Baricault S, Sparsa A, et al. . Incidence and mortality of bullous pemphigoid in France. J Invest Dermatol. 2012;132(8):1998-2004. doi: 10.1038/jid.2012.35 [DOI] [PubMed] [Google Scholar]
13.Zillikens D, Wever S, Roth A, Weidenthaler-Barth B, Hashimoto T, Bröcker EB. Incidence of autoimmune subepidermal blistering dermatoses in a region of central Germany. Arch Dermatol. 1995;131(8):957-958. doi: 10.1001/archderm.131.8.957 [DOI] [PubMed] [Google Scholar]
14.Gudi VS, White MI, Cruickshank N, et al. . Annual incidence and mortality of bullous pemphigoid in the Grampian Region of North-east Scotland. Br J Dermatol. 2005;153(2):424-427. doi: 10.1111/j.1365-2133.2005.06662.x [DOI] [PubMed] [Google Scholar]
15.Lee SG, Lee HJ, Yoon MS, Kim DH. Association of dipeptidyl peptidase 4 inhibitor use with risk of bullous pemphigoid in patients with diabetes. JAMA Dermatol. 2019;155(2):172-177. doi: 10.1001/jamadermatol.2018.4556 [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Langan SM, Smeeth L, Hubbard R, Fleming KM, Smith CJP, West J. Bullous pemphigoid and pemphigus vulgaris—incidence and mortality in the UK: population based cohort study. BMJ. 2008;337(7662):a180. doi: 10.1136/bmj.a180 [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Kridin K, Ludwig RJ. The growing incidence of bullous pemphigoid: overview and potential explanations. Front Med (Lausanne). 2018;5:220. doi: 10.3389/fmed.2018.00220 [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Brick KE, Weaver CH, Savica R, et al. . A population-based study of the association between bullous pemphigoid and neurologic disorders. J Am Acad Dermatol. 2014;71(6):1191-1197. doi: 10.1016/j.jaad.2014.07.052 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Stavropoulos PG, Soura E, Antoniou C. Drug-induced pemphigoid: a review of the literature. J Eur Acad Dermatol Venereol. 2014;28(9):1133-1140. doi: 10.1111/jdv.12366 [DOI] [PubMed] [Google Scholar]
20.Chen YJ, Wu CY, Lin MW, et al. . Comorbidity profiles among patients with bullous pemphigoid: a nationwide population-based study. Br J Dermatol. 2011;165(3):593-599. doi: 10.1111/j.1365-2133.2011.10386.x [DOI] [PubMed] [Google Scholar]
21.Lloyd-Lavery A, Chi C-C, Wojnarowska F, Taghipour K. The associations between bullous pemphigoid and drug use: a UK case-control study. JAMA Dermatol. 2013;149(1):58-62. doi: 10.1001/2013.jamadermatol.376 [DOI] [PubMed] [Google Scholar]
22.Kridin K, Bergman R. Association of bullous pemphigoid with dipeptidyl-peptidase 4 inhibitors in patients with diabetes: estimating the risk of the new agents and characterizing the patients. JAMA Dermatol. 2018;154(10):1152-1158. doi: 10.1001/jamadermatol.2018.2352 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Guo J-Y, Chen H-H, Yang Y-C, Wu P-Y, Chang M-P, Chen C-C. The association of dipeptidyl peptidase IV inhibitors and other risk factors with bullous pemphigoid in patients with type 2 diabetes mellitus: a retrospective cohort study. J Diabetes Complications. 2020;34(3):107515. doi: 10.1016/j.jdiacomp.2019.107515 [DOI] [PubMed] [Google Scholar]
24.Ruocco V, Sacerdoti G. Pemphigus and bullous pemphigoid due to drugs. Int J Dermatol. 1991;30(5):307-312. doi: 10.1111/j.1365-4362.1991.tb03867.x [DOI] [PubMed] [Google Scholar]
25.Fellner MJ, Katz JM. Occurrence of bullous pemphigoid after furosemide therapy. Arch Dermatol. 1976;112(1):75-77. doi: 10.1001/archderm.1976.01630250039012 [DOI] [PubMed] [Google Scholar]
26.Ameen M, Harman KE, Black MM. Pemphigoid nodularis associated with nifedipine. Br J Dermatol. 2000;142(3):575-577. doi: 10.1046/j.1365-2133.2000.03389.x [DOI] [PubMed] [Google Scholar]
27.Czechowicz RT, Reid CM, Warren LJ, Weightman W, Whitehead FJ. Bullous pemphigoid induced by cephalexin. Australas J Dermatol. 2001;42(2):132-135. doi: 10.1046/j.1440-0960.2001.00497.x [DOI] [PubMed] [Google Scholar]
28.Esposito I, Moretta G, Peris K, De Simone C. Linagliptin-induced bullous pemphigoid. Int J Dermatol. 2017;56(12):1467-1469. doi: 10.1111/ijd.13705 [DOI] [PubMed] [Google Scholar]
29.Durdu M, Baba M, Seçkin D. A case of bullous pemphigoid induced by aspirin. J Am Acad Dermatol. 2011;65(2):443-444. doi: 10.1016/j.jaad.2010.02.032 [DOI] [PubMed] [Google Scholar]
30.Lee JJ, Downham TF II. Furosemide-induced bullous pemphigoid: case report and review of literature. J Drugs Dermatol. 2006;5(6):562-564. [PubMed] [Google Scholar]
31.Kanahara SM, Agrawal A. Drug-induced bullous pemphigoid. J Gen Intern Med. 2016;31(11):1393-1394. doi: 10.1007/s11606-016-3679-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Murad AA, Connolly M, Tobin A-M. Rosuvastatin-induced pemphigoid. BMJ Case Rep. 2012:bcr1120115180. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Bastuji-Garin S, Joly P, Lemordant P, Sparsa A. Risk factors for bullous pemphigoid in the elderly: a prospective case-control study. J Invest Dermatol. 2011;131(3):637-643. doi: 10.1038/jid.2010.301 [DOI] [PubMed] [Google Scholar]
34.Stroup DF, Berlin JA, Morton SC, et al. ; Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. Meta-analysis of observational studies in epidemiology: a proposal for reporting. JAMA. 2000;283(15):2008-2012. doi: 10.1001/jama.283.15.2008 [DOI] [PubMed] [Google Scholar]
35.Liu S-D, Chen W-T, Chi C-C. Association between drug use and bullous pemphigoid. Published 2018. Accessed January 6, 2020. https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42018115041
36.Wells GA, Shea B, O’Connell D, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analysis. Published 2019. Accessed February 1, 2020. http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp
37.Sterne JAC, Savović J, Page MJ, et al. . RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019;366:l4898. doi: 10.1136/bmj.l4898 [DOI] [PubMed] [Google Scholar]
38.Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. A basic introduction to fixed-effect and random-effects models for meta-analysis. Res Synth Methods. 2010;1(2):97-111. doi: 10.1002/jrsm.12 [DOI] [PubMed] [Google Scholar]
39.Higgins J, Thomas J, Chandler J, eds. Cochrane handbook for systematic reviews of interventions, version 6.0. Updated July 2019.Cochrane, 2019. Accessed May 10, 2020. http://www.training.cochrane.org/handbook
40.Bastuji-Garin S, Joly P, Picard-Dahan C, et al. . Drugs associated with bullous pemphigoid. a case-control study. Arch Dermatol. 1996;132(3):272-276. doi: 10.1001/archderm.1996.03890270044006 [DOI] [PubMed] [Google Scholar]
41.Benzaquen M, Borradori L, Berbis P, et al. . Dipeptidyl peptidase IV inhibitors, a risk factor for bullous pemphigoid: retrospective multicenter case-control study from France and Switzerland. J Am Acad Dermatol. 2018;78(6):1090-1096. doi: 10.1016/j.jaad.2017.12.038 [DOI] [PubMed] [Google Scholar]
42.Nozu T, Mita H. Bullous pemphigoid and percutaneous endoscopic gastrostomy. Intern Med. 2010;49(11):971-975. doi: 10.2169/internalmedicine.49.3346 [DOI] [PubMed] [Google Scholar]
43.Plaquevent M, Tétart F, Fardet L, et al. ; French Investigators for Skin Adverse Reaction to Drugs; French Study Group on Autoimmune Bullous Skin Diseases . Higher frequency of dipeptidyl peptidase-4 inhibitor intake in bullous pemphigoid patients than in the French general population. J Invest Dermatol. 2019;139(4):835-841. doi: 10.1016/j.jid.2018.10.045 [DOI] [PubMed] [Google Scholar]
44.Schaffer C, Buclin T, Jornayvaz FR, et al. . Use of dipeptidyl-peptidase IV inhibitors and bullous pemphigoid. Dermatology. 2017;233(5):401-403. doi: 10.1159/000480498 [DOI] [PubMed] [Google Scholar]
45.Tan CWX, Pang Y, Sim B, Thirumoorthy T, Pang SM, Lee HY. The association between drugs and bullous pemphigoid. Br J Dermatol. 2017;176(2):549-551. doi: 10.1111/bjd.15195 [DOI] [PubMed] [Google Scholar]
46.Varpuluoma O, Försti A-K, Jokelainen J, et al. . Vildagliptin significantly increases the risk of bullous pemphigoid: a Finnish nationwide registry study. J Invest Dermatol. 2018;138(7):1659-1661. doi: 10.1016/j.jid.2018.01.027 [DOI] [PubMed] [Google Scholar]
47.Varpuluoma O, Försti A-K, Jokelainen J, et al. . Oral diabetes medications other than dipeptidyl peptidase 4 inhibitors are not associated with bullous pemphigoid: a Finnish nationwide case-control study. J Am Acad Dermatol. 2018;79(6):1034-1038.e5. doi: 10.1016/j.jaad.2018.05.030 [DOI] [PubMed] [Google Scholar]
48.Varpuluoma O, Jokelainen J, Försti A-K, et al. . Drugs used for neurologic and psychiatric conditions increase the risk for bullous pemphigoid: a case-control study. J Am Acad Dermatol. 2019;81(1):250-253. doi: 10.1016/j.jaad.2019.02.017 [DOI] [PubMed] [Google Scholar]
49.Hung C-T, Liu J-S, Cheng C-Y, et al. . Increased risk of bullous pemphigoid in dipeptidyl peptidase 4 inhibitors: a nationwide, population-based, cohort study in Taiwan. J Dermatol. 2020;47(3):245-250. doi: 10.1111/1346-8138.15195 [DOI] [PubMed] [Google Scholar]
50.Rosenstock J, Perkovic V, Johansen OE, et al. ; CARMELINA Investigators . Effect of linagliptin vs placebo on major cardiovascular events in adults with type 2 diabetes and high cardiovascular and renal risk: The CARMELINA randomized clinical trial. JAMA. 2019;321(1):69-79. doi: 10.1001/jama.2018.18269 [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Patsatsi A, Vyzantiadis T-A, Chrysomallis F, Devliotou-Panagiotidou D, Sotiriadis D. Medication history of a series of patients with bullous pemphigoid from northern Greece—observations and discussion. Int J Dermatol. 2009;48(2):132-135. doi: 10.1111/j.1365-4632.2009.03839.x [DOI] [PubMed] [Google Scholar]
52.Park BK, Sanderson J, Naisbitt D. Drugs as haptens, antigens, and immunogens In: Pichler WJ, ed. Drug Hypersensitivity. Karger Publishers; 2007:55-65. doi: 10.1159/000104189 [DOI] [Google Scholar]
53.Baum H, Butler P, Davies H, Sternberg MJ, Burroughs AK. Autoimmune disease and molecular mimicry: an hypothesis. Trends Biochem Sci. 1993;18(4):140-144. doi: 10.1016/0968-0004(93)90022-F [DOI] [PubMed] [Google Scholar]
54.Bowman C, Delrieu O. Immunogenetics of drug-induced skin blistering disorders. part I: perspective. Pharmacogenomics. 2009;10(4):601-621. doi: 10.2217/pgs.09.11 [DOI] [PubMed] [Google Scholar]
55.Ruocco V. Blistering eruptions caused by drugs: commentary. Clin Dermatol. 1993;11(4):433-435. doi: 10.1016/0738-081X(93)90148-6 8124640 [DOI] [Google Scholar]
56.Brenner S, Bialy-Golan A, Ruocco V. Drug-induced pemphigus. Clin Dermatol. 1998;16(3):393-397. doi: 10.1016/S0738-081X(98)00010-8 [DOI] [PubMed] [Google Scholar]
57.Breathnach SM, Hintner H. Adverse Drug Reactions and the Skin. Blackwell; 1992. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials