Erin X. Wei

BACKGROUND: The recent public reporting of payments made to physicians by manufacturers of orthopedic devices provides an opportunity to assess the accuracy of physicians' conflict-of-interest disclosures. METHODS: We analyzed the reports of payments made to physicians by five manufacturers of total hip and knee prostheses in 2007. For each payment recipient who was an author of a presentation or served as a committee member or board member at the 2008 annual meeting of the American Academy of Orthopaedic Surgeons, the disclosure statement was reviewed to determine whether the payment had been disclosed. To ascertain the reasons for nondisclosure, a survey was administered to physicians who had received payments that were not disclosed. RESULTS: The overall rate of disclosure was 71.2% (245 of 344 payments). For payments that were directly related to the topic of the presentation at the meeting, the rate was 79.3% (165 of 208); for payments that were indirectly related, the rate was 50.0% (16 of 32); and for payments that were unrelated, the rate was 49.2% (29 of 59) (P=0.008). In the multivariate analysis, payments were also more likely to have been disclosed if they exceeded $10,000 (P<0.001), were directed toward an individual physician rather than a company or organization (P=0.04), or included an in-kind component (P=0.002). Among the 36 physicians who responded to the survey regarding reasons for nondisclosure (response rate, 39.6%), the reasons most commonly given for nondisclosure were that the payment was unrelated to the topic of presentation at the annual meeting (38.9% of respondents) and that the physician had misunderstood the disclosure requirements (13.9%); 11.1% reported that the payment had been disclosed but was mistakenly omitted from the program. CONCLUSIONS: In this study of self-reported conflict-of-interest disclosure by physicians at a large annual meeting, the rate of disclosure was 79.3% for directly related payments and 50.0% for indirectly related payments.

Importance: De novo bullous pemphigoid (BP) is a rare immune-mediated adverse event from immune checkpoint inhibitors (ICIs) that can necessitate permanent discontinuation of the anticancer therapy, but the risk factors for developing this toxic effect are unknown. Objective: To compare potential risk factors for BP in patients treated with ICIs who did and did not develop BP. Design, Setting, and Participants: This cohort and nested propensity score-matched case-control study was conducted at the Dana-Farber Cancer Institute, Brigham and Women's Hospital, and Massachusetts General Hospital. All patients at these facilities with de novo BP after ICI treatment were compared with all patients on the cancer registry who were treated with ICIs between October 1, 2014, and December 31, 2020. Patients with incomplete or blinded data regarding the ICI agent or total cycles were excluded. Exposures: In the cohort, assessed potential risk factors included age at ICI introduction, sex, ICI molecular target, and cancer type, which were then used as matching variables. In the propensity score-matched case-control analysis, risk factors assessed included sex, race and ethnicity, cancer stage, metastasis sites, idiopathic BP comorbidities, pre-ICI vaccination, radiation history, body mass index, and derived neutrophil-to-lymphocyte ratio. Main Outcomes and Measures: Diagnosis of BP at any point after ICI treatment, confirmed by direct immunofluorescence, indirect immunofluorescence, autoantibody serologies, or diagnostic consensus among study board-certified dermatologists. Odds ratios (ORs) and 95% CIs were calculated for all risk factors. In the secondary analysis, best overall responses to ICIs between cases and controls were compared by Fisher exact test. Results: Among 5636 patients treated with ICIs at Dana-Farber Cancer Institute, Brigham and Women's Hospital, and Massachusetts General Hospital during the study period, 35 (0.6%; median [IQR] age, 72.8 [13.4] years; 71.4% [25] male patients) developed BP. In a multivariate logistic regression model that assessed 2955 patients with complete data in the cancer registry, age 70 years or older (OR, 2.32; 95% CI, 1.19-4.59; P = .01), having melanoma (OR, 3.21; 95% CI, 1.51-6.58; P < .003), and having nonmelanoma skin cancer (OR, 8.32; 95% CI, 2.81-21.13; P < .001) were significantly associated with developing BP. In the nested 1:2 case-control comparison of all 35 cases to 70 propensity score-matched controls, a complete or partial response on initial restaging imaging was a risk factor for BP development (OR, 3.37; 95% CI, 1.35-9.30; P = .01). Bullous pemphigoid cases also more frequently exhibited overall tumor response to ICIs than matched controls (29 of 35 [82.9%] vs 43 of 70 [61.4%]; P = .03). Conclusions and Relevance: In this cohort study, age 70 years or older and skin cancer were associated with increased risk of developing ICI-associated BP. Given the association of BP with improved initial and best overall tumor responses, early identification and toxic effect-directed treatment should be prioritized, especially in individuals at risk for developing de novo BP.