Risk of Incident Venous Thromboembolism Among Patients With Bullous Pemphigoid or Pemphigus Vulgaris: A Nationwide Cohort Study With Meta‐Analysis

Abstract

Background

Bullous pemphigoid (BP) and pemphigus vulgaris (PV) share similar pathophysiology with venous thromboembolism (VTE) involving platelet activation, immune dysregulation, and systemic inflammation. Nevertheless, their associations have not been well established.

Methods and Results

To examine the risk of incident VTE among patients with BP or PV, we performed a nationwide cohort study using Taiwan's National Health Insurance Research Database and enrolled 12 162 adults with BP or PV and 12 162 controls. A Cox regression model considering stabilized inverse probability weighting was used to calculate the hazard ratios (HRs) for incident VTE associated with BP or PV. To consolidate the findings, a meta‐analysis that incorporated results from the present cohort study with previous literature was also conducted. Compared with controls, patients with BP or PV had an increased risk for incident VTE (HR, 1.87 [95% CI, 1.55–2.26]; P<0.001). The incidence of VTE was 6.47 and 2.20 per 1000 person‐years in the BP and PV cohorts, respectively. The risk for incident VTE significantly increased among patients with BP (HR, 1.85 [95% CI, 1.52–2.24]; P<0.001) and PV (HR, 1.99 [95% CI, 1.02–3.91]; P=0.04). In the meta‐analysis of 8 studies including ours, BP and PV were associated with an increased risk for incident VTE (pooled relative risk, 2.17 [95% CI, 1.82–2.62]; P<0.001).

Conclusions

BP and PV are associated with an increased risk for VTE. Preventive approaches and cardiovascular evaluation should be considered particularly for patients with BP or PV with concomitant risk factors such as hospitalization or immobilization.

Nonstandard Abbreviations and Acronyms

BP

bullous pemphigoid

IPW

inverse probability weighting

NHIRD

National Health Insurance Research Database

PV

pemphigus vulgaris

Clinical Perspective.

What Is New?

• This nationwide cohort study including 24 324 subjects demonstrated that patients with bullous pemphigoid or pemphigus vulgaris had a 1.76‐fold risk for incident venous thromboembolism.

• The association remained significant in age‐ and sex‐stratified analyses.

• The increased risk of venous thromboembolismin associated with bullous pemphigoid or pemphigus vulgaris was confirmed by a meta‐analysis that incorporated data from this cohort with previous literature.

What Are the Clinical Implications?

• Evidence from our study suggests an increased risk of incident venous thromboembolism among patients with bullous pemphigoid or pemphigus vulgaris.

• Imperative vascular examination and prompt anticoagulation may benefit such patients presenting symptoms suggestive of venous thromboembolism.

Bullous pemphigoid (BP) and pemphigus vulgaris (PV) are the 2 most common autoimmune bullous dermatoses predominantly affecting the elderly. ¹ , ² , ³ The pathophysiology of BP and PV is mediated by autoantibodies against structural proteins of cell junctions, causing blisters and erosions on patients' skin or mucous membranes. ⁴ , ⁵ , ⁶ , ⁷ An elevated risk of cardiovascular comorbidities and related mortality has been reported in patients with BP or PV. ⁸ , ⁹ , ¹⁰

Venous thromboembolism (VTE), comprising deep vein thrombosis (DVT) and pulmonary embolism (PE), is a global health burden associated with serious morbidity and mortality. ¹¹ , ¹² Discoveries of immune alterations and vascular inflammation involved in the thrombogenic mechanism have advanced our knowledge of VTE pathogenesis. ¹³ , ¹⁴ , ¹⁵ Several chronic inflammatory and autoimmune disorders have been reported as risk factors for VTE. ¹⁶ , ¹⁷ , ¹⁸ BP and PV share similar mechanisms with VTE involving immune dysregulation, activation of complement cascade, and systemic inflammation. ¹⁹ , ²⁰ , ²¹ Several case reports have presented the coexistence of BP or PV with VTE. ²² , ²³ , ²⁴ Previous epidemiological studies have also explored the association of BP or PV with VTE, but a sustained debate remains. Some studies demonstrated an increased risk of VTE in patients with BP or PV, ²⁵ , ²⁶ whereas others showed contradictory results. ²⁷ , ²⁸ Moreover, most existing studies were from Western countries, thereby limiting their generalizability to other countries or races. Therefore, data from Asian populations are required and a comprehensive meta‐analysis is warranted to synthesize the overall effect estimates incorporating different populations.

Herein, we first conducted a nationwide retrospective cohort study to investigate whether patients with BP or PV had an increased risk  developing VTE. To consolidate our findings, a systematic review and meta‐analysis encompassing the present cohort study and previous studies was subsequently performed.

Methods

The National Health Insurance Research Database (NHIRD) is managed and supervised by the Health and Welfare Data Science Center, Ministry of Health and Welfare in Taiwan. This data set can be used only in the Health and Welfare Data Science Center. Researchers interested in analyzing this data set have to formally request access through the Ministry of Health and Welfare in Taiwan (website: https://dep.mohw.gov.tw/DOS/cp‐2516‐59203‐113.html). The requirement for informed consent was waived due to the anonymized data encrypted by the Ministry of Health and Welfare.

Data Source

We performed a nationwide retrospective cohort study using the Taiwan's NHIRD, which contains health care data of ≈23.6 million individuals, representing the entire Taiwanese population. ²⁹ , ³⁰ The NHIRD has been widely used in published epidemiological studies. ³⁰ , ³¹ , ³² Diagnostic and procedure codes were derived from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD‐9‐CM, before 2016) and Tenth Revision (ICD‐10‐CM, after 2016) codes. A data subset of NHIRD, the Catastrophic Illness Dataset, was applied to ascertain the diagnosis of PV. A brief introduction to the Catastrophic Illness Dataset and the reason for using this data set are provided in Data S1. The Research Ethics Committee of Hualien Tzu Chi Hospital approved this study (REC No: IRB107‐152‐C).

Study Population

The study population comprised all patients aged ≥20 years who received a new diagnosis of BP or PV from dermatologists between January 1, 2001 and December 31, 2017. A diagnosis of BP was defined as a diagnosis confirmed at least 3 times in an outpatient department within 1 year or a discharge diagnosis of ICD‐9‐CM code 694.5 and ICD‐10‐CM code L12. These diagnostic codes have been validated with a high positive predictive value of 98%. ³³ , ³⁴ Patients with a diagnosis of PV were identified through the catastrophic illness registry using ICD‐9‐CM code 694.4 and ICD‐10‐CM code L10.0‐L10.9. Patients who received a diagnosis of BP or PV before 2001 were excluded to ensure that newly diagnosed patients were recruited.

The exposed cohort comprised all patients with BP or PV, whereas the unexposed cohort comprised patients without BP or PV. Among the unexposed cohort, we excluded patients with diagnoses of other bullous disorders (ICD‐9 codes: 694.0–694.3, 694.6, 694.8, 694.9; ICD‐10 codes: L11, L13, L14). Each patient in the exposed cohort was individually matched with 1 control subject in the unexposed cohort by age and sex. The index date for patients with BP was the date when the first BP diagnosis was made, whereas the index date for patients with PV was the date of the application for the catastrophic illness certificate. The index date for patients in the unexposed cohort was assigned as the same date as the matched patient. Study subjects with a history of VTE before the index date were excluded.

Outcomes

The primary outcome was any incident event of VTE, including DVT (ICD‐9‐CM codes: 453.2, 453.4, 453.5, 453.7, 453.8, 453.9; ICD‐10‐CM codes: I82.2, I82.4, I82.5, I82.60, I82.62, I82.89, I82.9) and PE (ICD‐9‐CM code: 415.1; ICD‐10‐CM code: I26). The diagnostic accuracy of the aforementioned codes has been validated, with high positive predictive value found (94%). ³⁵ An outcome event was defined as a diagnosis made either in the outpatient or inpatient department after the index date. We also conducted individual outcome analyses considering any incident event of DVT and PE. All enrollees were followed from the index date until the occurrence of VTE, death, or December 31, 2018 (the last date of our database), whichever came first. In the individual outcome analysis, patients were followed up until that outcome was observed; that is, patients were not censored if the other outcome occurred earlier.

Covariates and Confounders

We obtained baseline data of subjects from reimbursement claims of NHIRD. Preexisting comorbidities and baseline medication use (Tables S1 and S2), considered potential confounders, were chosen based on previous studies. ³⁶ Preexisting comorbidities were defined as diagnoses on discharge or diagnoses confirmed at least twice in the outpatient services within the year before the index date, using the ICD‐9‐CM, ICD‐10‐CM, and procedure codes. We quantified enrollees' overall systemic health status using the Charlson Comorbidity Index according to preexisting comorbidities. ³⁷ Medication use at baseline was defined as a drug prescribed for at least 30 days within the year before the index date. Monthly income levels were evaluated based on income‐related National Health Insurance premiums.

Subgroup and Stratified Analyses

The exposed cohort was divided into 2 subgroups: a BP cohort comprising only patients with BP and a PV cohort comprising only patients with PV. Additionally, age‐ and sex‐stratified subanalyses were performed. To investigate the effect of cardiovascular and pulmonary comorbidities on VTE risks associated with the exposed and unexposed cohorts, ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ we conducted stratified analyses for patients with and without comorbidities such as hypertension, atrial fibrillation, stroke, heart failure, coronary artery disease, chronic obstructive pulmonary disease, hyperlipidemia, diabetes, and varicose veins. For each comorbidity, we compared the risk of VTE between the exposed and unexposed cohorts in 2 strata: patients with and without that specific comorbidity.

Statistical Analysis

We applied a propensity score‐based approach, stabilized inverse probability weighting (IPW), to mitigate potential selection bias and to balance systematic differences in baseline characteristics between the study cohorts. ⁴² Standardized mean difference was used to examine the difference in baseline characteristics between the exposed and unexposed cohorts, with a value of <0.1 indicating a negligible difference. A univariable Cox regression model considering IPW was applied to estimate cumulative incidences and hazard ratios (HRs) for each outcome. The cumulative incidence curves were illustrated using Kaplan–Meier methods. If any covariate was imbalanced between the study cohorts after stabilized IPW, we added that covariate into the regression model for additional adjustment to test the robustness of our primary analyses. A 2‐sided probability value of <0.05 was considered significant. All statistical analyses in the cohort study were performed using SAS software (version 9.4; SAS Institute, Cary, NC).

Systematic Review and Meta‐Analysis

We conducted a systematic review and meta‐analysis to examine the risk of incident VTE among patients with BP or PV and incorporated data from the present study with previous literature. The MEDLINE, Embase, Cochrane Library, and Web of Science databases were searched for relevant publications from their individual inception to October 20, 2022. We included observational studies that reported risk estimates of VTE in patients with BP or PV compared with controls without BP/PV, including cohort, case–control, or cross‐sectional studies. No restrictions on language or geographic locations were imposed. The Newcastle‐Ottawa Scale was applied to assess the risk of bias of included studies. ⁴³

The Review Manager version 5.4.1 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2020) was used to conduct random‐effects meta‐analyses. Data were presented as pooled relative risks (RRs) with 95% CIs. ⁴⁴ A P value of <0.05 was deemed significant. We quantified between‐study heterogeneity by I² statistics, with an I² of >50% indicating at least moderate heterogeneity. ⁴⁵ Predefined subgroup meta‐analyses based on BP or PV, study design, and geographic locations were performed to determine if certain study‐level factors may affect the pooled results. The detailed methodology regarding meta‐analysis was provided in Data S2 and Table S3.

Results

Results of the Cohort Study

We originally recruited 24 324 study subjects (12 162 adults with BP or PV and 12 162 controls) from the NHIRD. After applying the stabilized IPW, the exposed cohort comprised 12 021 patients with BP or PV, whereas the unexposed cohort comprised 13 043 control subjects. All the baseline characteristics, except for stroke, dementia, and antipsychotics use, were balanced between the study cohorts after stabilized IPW (Table 1). The baseline characteristics of study cohorts before stabilized IPW are provided in Table S4. The overall mean length of follow‐up was 4.5 years.

Table 1.

Baseline Characteristics of the Study Population After Stabilized Inverse Probability Weighting

Characteristics*	With BP or PV (n=12 021)	Without BP or PV (n=13 043)	SMD†
Mean age (SD), y	74.4 (15.62)	75.5 (15.93)	0.0672
Sex, %
Male	52.57	52.11	0.0092
Female	47.43	47.89	0.0092
Income level (New Taiwan Dollar), %
Financially dependent	32.95	34.43	0.0313
15 840–24 999	41.42	40.73	0.0140
25 000–39 999	9.87	9.24	0.0214
≥40 000	15.76	15.60	0.0044
Mean Charlson Comorbidity Index (SD)	1.41 (1.53)	1.57 (1.81)	0.0955
Comorbidities, %
Hypertension	51.52	53.17	0.0330
Atrial fibrillation	3.78	4.99	0.0591
Stroke	26.98	31.58	0.1012
Heart failure	7.09	8.09	0.0378
Coronary artery disease	14.82	16.11	0.0357
Chronic obstructive pulmonary disease	16.19	19.24	0.0799
Chronic kidney disease	6.13	7.25	0.0448
Cirrhosis	1.16	1.24	0.0073
Hyperlipidemia	15.51	15.75	0.0066
Fracture of lower limbs	3.18	3.74	0.0306
Gout	5.86	5.67	0.0082
Malignancy	6.81	6.59	0.0088
Pregnancy	0.18	0.15	0.0074
Thyroid dysfunction	1.33	1.58	0.0209
Dementia	14.77	20.39	0.1480
Epilepsy	2.46	4.23	0.0986
Schizophrenia	0.59	0.52	0.0094
Anxiety	8.20	9.27	0.0379
Depression	6.85	8.33	0.0559
Bipolar disorder	0.64	0.59	0.0064
Autoimmune disease	2.73	2.83	0.0061
Diabetes	25.62	28.36	0.0618
Varicose veins	0.26	0.69	0.0626
Baseline medication use, %
Hormone therapies	1.13	1.09	0.0038
Statins	14.19	14.05	0.0040
Antiplatelets	31.38	32.25	0.0187
Warfarin	1.31	1.52	0.0178
Dabigatran	0.28	0.43	0.0252
Rivaroxaban	0.36	0.59	0.0335
Apixaban	0.14	0.13	0.0027
Edoxaban	0.02	0.01	0.0082
Antipsychotics	11.12	14.87	0.1117
Dipeptidyl peptidase‐4 inhibitors	6.13	7.63	0.0593

BP indicates bullous pemphigoid; PV, pemphigus vulgaris; and SMD, standardized mean difference.

^*All covariates listed were used to calculate the propensity score for inverse probability weighting.

^† A standardized mean difference of <0.1 indicates a negligible difference.

Overall, patients with BP or PV were at an increased risk for incident VTE compared with controls (HR, 1.87 [95% CI, 1.55–2.26]; P<0.001; Table 2). When divided by subgroups, 240 patients in the BP cohort and 24 patients in the PV cohort developed VTE. The incidence of VTE was 6.47 and 2.20 per 1000 person‐years in the BP and PV cohorts, respectively. The risk for incident VTE was significantly increased among patients with BP (HR, 1.85 [95% CI, 1.52–2.24]; P<0.001) and PV (HR, 1.99 [95% CI, 1.02–3.91]; P=0.04). The cumulative incidence curves are presented in Figure 1. Regarding the individual outcome analyses (Table 2), patients with BP or PV were at a significantly higher risk for incident DVT (HR, 2.09 [95% CI, 1.68–2.61]; P<0.001) but not for incident PE (HR, 1.39 [95% CI, 0.99–1.97]; P=0.061). The results of sensitivity analyses by additional adjustment for stroke, dementia, and antipsychotic use were similar to our primary analysis (Table S5).

Table 2.

Risk of Venous Thromboembolism Among Patients With Bullous Pemphigoid or Pemphigus Vulgaris

Outcomes/comparison	Exposed cohort			Unexposed cohort
	n	Events	IR*	n	Events	IR*	HR† (95% CI)	P value
Venous thromboembolism
Overall cohort vs unexposed cohort	12 021	264	5.42	13 043	183	2.85	1.87 (1.55–2.26)	<0.001
BP cohort vs unexposed cohort	10 572	240	6.47	11 569	178	3.41	1.85 (1.52–2.24)	<0.001
PV cohort vs unexposed cohort	1439	24	2.20	1438	13	1.10	1.99 (1.02–3.91)	0.04
Deep vein thrombosis
Overall cohort vs unexposed cohort	12 021	209	4.28	13 043	129	2.01	2.09 (1.68–2.61)	<0.001
BP cohort vs unexposed cohort	10 572	187	5.03	11 569	124	2.37	2.08 (1.66–2.61)	<0.001
PV cohort vs unexposed cohort	1439	22	1.98	1438	13	1.10	1.80 (0.91–3.57)	0.093
Pulmonary embolism
Overall cohort vs unexposed cohort	12 021	68	1.37	13 043	62	0.96	1.39 (0.99–1.97)	0.061
BP cohort vs unexposed cohort	10 572	63	1.68	11 569	63	1.19	1.36 (0.96–1.93)	0.083
PV cohort vs unexposed cohort	1439	4	0.34	1438	0	0	NA

BP indicates bullous pemphigoid; PV, pemphigus vulgaris; HR, hazard ratio; IR, incidence rate; and NA, not applicable.

^*Incidence rate per 1000 person‐years.

^† The HRs were calculated using a univariable Cox regression model with inverse probability weighting and using the corresponding unexposed cohort (patients without BP or PV) as the reference.

Figure 1. Cumulative incidence curves for venous thromboembolism in the overall, bullous pemphigoid, and pemphigus vulgaris cohorts.

BP indicates bullous pemphigoid; and PV, pemphigus vulgaris.

In the age‐ and sex‐stratified analyses, the increased risks for incident VTE associated with BP or PV remained significant across all age and sex subgroups (Table 3; see Table S6). In the analyses stratified by various cardiovascular and pulmonary comorbidities, a similar trend of higher VTE risk in patients with BP or PV was still observed in each comorbidity stratum, though some did not reach statistical significance; this could be due to the reduced sample size/event number after dividing patients into different comorbidity subgroups, especially those with certain comorbidities (eg, atrial fibrillation, heart failure, varicose veins) (Table S6).

Table 3.

Risk of Venous Thromboembolism in Patients With Bullous Pemphigoid or Pemphigus Vulgaris Compared With Controls, Stratified by Age and Sex

Outcomes	Venous thromboembolism			Deep vein thrombosis			Pulmonary embolism
	HR*	95% CI	P value	HR*	95% CI	P value	HR*	95% CI	P value
Age
<65 y	1.95	1.21–3.14	0.006	2.05	1.19–3.52	0.010	1.62	0.59–4.44	0.345
≥65 y	1.99	1.62–2.45	<0.001	2.30	1.81–2.93	<0.001	1.42	0.99–2.06	0.060
Sex
Male	2.05	1.57–2.67	<0.001	2.46	1.82–3.34	<0.001	1.09	0.64–1.85	0.760
Female	1.66	1.27–2.17	<0.001	1.73	1.25–2.38	0.001	1.58	1.01–2.47	0.046

HR indicates hazard ratio.

^*The HRs were calculated using a univariable Cox regression model with inverse probability weighting and using the unexposed cohort (patients without bullous pemphigoid or pemphigus vulgaris) as the reference.

Results of the Meta‐Analysis

After including the present cohort study, a total of 8 studies ²⁵ , ²⁶ , ²⁷ , ²⁸ , ⁴⁶ , ⁴⁷ , ⁴⁸ with 61 600 subjects were eligible for a meta‐analysis. As illustrated in Figure 2, ²⁵ , ²⁶ , ²⁷ , ²⁸ , ⁴⁶ , ⁴⁷ , ⁴⁸ our meta‐analysis indicated that patients with BP or PV were at an increased risk for incident VTE (pooled RR, 2.17 [95% CI, 1.80–2.62]; I²=72%), DVT (pooled RR, 1.97 [95% CI, 1.74–2.23]; I²=0%), and PE (pooled RR, 1.69 [95% CI, 1.31–2.17]; I²=26%) when compared with controls. The Preferred Reporting Items for Systematic Reviews and Meta‐Analyses flow diagram, summary of risk of bias assessment, and characteristics of included studies are respectively provided in Figures S1 and S2 and Table S7. Subgroup meta‐analyses generated similar results and revealed that study design might be a potential effect modifier to influence the pooled RR (see Figures S3 through S5).

Figure 2. Forest plots of the meta‐analysis.

The meta‐analysis illustrated a significant association of bullous pemphigoid or pemphigus vulgaris with venous thromboembolism, deep vein thrombosis, and pulmonary embolism. IV indicates inverse variance; and VTE, venous thromboembolism.

Discussion

To the best of our knowledge, this is the largest cohort and most comprehensive study examining current evidence regarding the risk of incident VTE among patients with BP or PV. Our nationwide cohort study illustrated a 1.76‐fold HR of developing VTE among patients with BP or PV when compared with controls. The increased risk for incident VTE associated with BP or PV remained significant on stratification by age and sex. The pooled results from meta‐analysis further confirmed the significant association of BP or PV with incident VTE.

Although the exact mechanism was not fully elucidated, previous studies have provided mechanistic clues to support our findings. First, activation of coagulation cascades has been observed in patients with BP and PV. ⁴⁹ Serum markers for both thrombin formation (plasma prothrombin fragments F1+2) and fibrin degradation (D‐dimer) were significantly higher among patients with BP than controls. ²⁰ , ⁵⁰ Skin tissue factor, an initiator of coagulation, was also expressed in patients with BP. ⁵⁰ Second, multiple proinflammatory cytokines were found to be increased in patients with BP and PV, suggesting that BP and PV are autoimmune diseases with systemic inflammation. ⁵¹ , ⁵² , ⁵³ It is believed that type 2 inflammation triggers the formation of autoantibodies in BP. ⁵⁴ The levels of interleukin (IL)‐4, IL‐5, and IL‐13 were reportedly elevated in the serum, blister fluid, and skin biopsy specimens among patients with BP. ⁵⁵ These cytokines may also involve the pathogenesis of VTE. ⁵⁶ Serum levels of soluble E‐selectin and vascular endothelial growth factor in BP were associated with those of circulating autoantibodies, which supported the interaction between activated endothelial inflammatory and immune reactions. ⁵⁷ , ⁵⁸ Eosinophils in BP and PV are also believed to involve in coagulation activation at skin levels. ⁵⁹ Third, antiphospholipid antibodies have been detected in patients with BP and PV. ⁶⁰ , ⁶¹ Antiphospholipid antibodies comprised a heterogeneous group of antibodies that could produce a prothrombotic state and are related to thrombotic events. ⁶¹ Future research is warranted to clarify the underlying mechanisms between inflammation, immune dysregulation, and coagulation.

A cross‐sectional study reported associations of BP with VTE (odds ratio [OR], 1.64 [95% CI, 1.47–1.83]) and PV with VTE (OR, 1.96 [95% CI, 1.68–2.28]). ²⁵ However, the authors could not elucidate the temporal relationship. Two case–control studies reported that BP might not be a risk factor of VTE. ²⁷ , ²⁸ Several cohort studies have also investigated the risk of VTE among patients with BP or PV. ²⁶ , ⁴⁶ , ⁴⁷ , ⁴⁸ However, these studies mainly focused on people in Western countries, with a resultant lack of generalizability. Our nationwide cohort analysis used data from an Asian population to complement this topic with the largest sample size to date. Recently, a retrospective cohort study suggested BP as a risk factor for VTE (HR, 2.02 [95% CI, 1.01–4.06]). ⁶² In comparison to that study, we additionally investigated the VTE risk among patients with PV. Moreover, we conducted a systematic review and meta‐analysis to compile all the available data. The findings from our study provided comprehensive real‐world evidence regarding the risk of VTE among patients with BP and PV and also extended the generalizability based on previous literature.

A previous meta‐analysis from 2018 has evaluated the association of BP with VTE. ⁶³ Nevertheless, this analysis enrolled only 4 studies from Europe. Because recent studies have provided additional new data, our meta‐analysis enrolled more studies across different populations (one from the United States, one from Israel, and our nationwide cohort study from Taiwan) to provide better generalizability.

The substantial heterogeneity in our meta‐analysis may be partially explained by the study design of included studies. The pooled RR from cohort studies was higher than that from case–control and cross‐sectional studies. Although the heterogeneity was still high in the subanalysis regarding cohort studies, the conclusion from these studies was consistent. Further investigations with individual patient data may be instrumental in providing in‐depth analyses.

The risk for PE among patients with BP or PV was not significant in our cohort study, which might be attributed to the low incidence of PE in the exposed cohort. However, when the estimates were pooled together in the meta‐analysis (Figure 2), a significant association of BP or PV with PE was indicated. This finding may emphasize the value of meta‐analysis to provide sufficient statistical power.

The main strength of this study is that we not only conducted a population‐based cohort study but also incorporated it into the existing literature to strengthen the evidence. However, the results need to be interpreted in context with its limitations. First, NHIRD used in our cohort analysis lacks data on body mass index, smoking history, physical activity, lifestyle, and laboratory data. Given the application of stabilized IPW to minimize potential confounding effects, residual confounders might have still existed. Second, the diagnoses of BP, PV, and VTE were mainly based on ICD codes. Although the codes have been reportedly validated with high positive predictive values, misclassification bias was inevitable. Third, the present study did not include the increasingly prescribed immune checkpoint inhibitors, which might be considered as potential confounders, ⁶⁴ because these medications were not reimbursed by Taiwan's National Health Insurance program until 2019. Fourth, a novel treatment for pemphigoid diseases, rituximab, was not studied because the National Health Insurance program did not cover this drug during the study period. As case reports have discussed the potential effect of rituximab on VTE, ¹⁰ , ⁶⁵ further studies are needed to focus on the impact of rituximab on VTE risk among patients with BP and PV.

Conclusions

In conclusion, BP and PV are associated with increased risk for VTE based on real‐world evidence from this cohort study and meta‐analysis. Preventive approaches and cardiovascular evaluation should be considered particularly for patients with BP or PV and concomitant risk factors, such as hospitalization or immobilization. Prompt anticoagulation and multidisciplinary care could be considered in patients with BP or PV presenting symptoms suggestive of VTE, such as unexplained dyspnea, chest pain, or leg swelling. Based on its significant morbidity, the risk of VTE among patients with BP and PV should deserve more public health concern.

Sources of Funding

This study received a grant from the Hualien Tzu Chi Hospital (TCRD108‐21). The funders had no role in the study design, data analysis, data interpretation, drafting of the article, decision to submit for publication, or approval of the article for publication.

Disclosures

None.

Supporting information

Data S1–S2

Tables S1–S7

Figures S1–S5

References ⁶⁶ , ⁶⁷ , ⁶⁸ , ⁶⁹ , ⁷⁰ , ⁷¹ , ⁷² , ⁷³