De novo autoimmune connective tissue disease and mortality in patients treated with anti-programmed death receptor-1 and anti-programmed death-ligand 1 therapy: a population-level cohort study

Abstract

Using a population-level cohort analysis, our study demonstrates that, although rare, de novo autoimmune cutaneous connective tissue diseases (AiCTDs) in the setting of immune checkpoint inhibitors (ICIs) are not associated with a greater risk of mortality and overall approach a statistically significant decrease in mortality when compared with patients treated with ICIs who do not experience cutaneous immune-related adverse events. These findings are significant and highly relevant to dermatologists and oncologists caring for ICI recipients as it adds to the limited information on development of cutaneous AiCTD following ICI administration, for which enhanced understanding is critical to improving the care for this challenging patient population.

Dear Editor, Immune checkpoint inhibitors (ICIs), including antibodies targeting programmed death receptor-1 (PD-1) and programmed death-ligand 1 (PD-L1), have revolutionized cancer therapy in the last decade. However, ICIs are associated with considerable off-target immunological effects, with cutaneous immune-related adverse events (cirAEs) affecting up to 40% of ICI recipients.¹ CirAEs have been associated with improved therapeutic response, as seen in ICI-induced vitiligo and lichen planus.² And yet, despite shared aetiological mechanisms (i.e. immune dysregulation), the association between ICI and de novo autoimmune cutaneous connective tissue disease (AiCTD) remains poorly characterized outside of limited case reports and series.³ Moreover, the overall prognostic significance of ICI-induced AiCTDs is unknown. Thus, we aimed to investigate the rate of de novo AiCTD development after ICI therapy and its downstream impact on patient survival.

Briefly, we utilized the global TriNetX Diamond Network, which includes the deidentified electronic medical records of over 213 million patients from the USA and Europe and covers diverse geographic, age, racial, ethnic, income and insurance groups. Detailed methodology is provided in the online Supporting Information (https://figshare.com/s/699dad8a139f581710c7). ICI recipients were identified using record of treatment with either anti-PD-1 or anti-PD-L1 therapy. We included participants with the four most common indications for ICI therapy including malignancies of the lung, digestive tract, urinary tract or melanoma. Cutaneous eruptions of interest were derived as previously described.⁴De novo AiCTDs and cirAEs were defined as a new diagnosis of any of the included cutaneous eruptions within 12 months of the first instance of ICI therapy. Patients who developed at least one de novo AiCTD were matched on age, sex, race and cancer type to patients who developed de novo non-AiCTD cirAEs. A 12-month landmark analysis using a Cox proportional hazards model was performed to determine the prognostic association of AiCTDs with overall survival.

Of the 61 264 cancer patients treated with PD-1 or PD-L1 blockade, 6896 (11.3%) developed non-AiCTD cirAEs and 145 (0.24%) developed de novo AiCTDs within 12 months of ICI initiation. The most common AiCTDs were cutaneous small vessel vasculitis (26.2%, n = 38), cutaneous lupus erythematosus (21.4%, n = 31), dermatomyositis (15.9%, n = 23) and scleroderma (15.2%, n = 22). Compared with patients who developed cirAEs, the AiCTD cohort had a higher representation of female patients (60.7% vs. 41.2%, P < 0.001), Asian patients (6.9% vs. 3.64%, P = 0.04), Black patients (10.3% vs. 6.27%, P = 0.046) and patients with lung cancer (60% vs. 44.3%, P < 0.001). We found no statistically significant difference in systemic corticosteroid use between the two cohorts (P = 0.891) (Table S1 and Supplemental Methods; see Supporting Information). Notably, de novo AiCTDs were not associated with a greater risk of mortality [hazard ratio (HR) 0.729, P = 0.157] (Table 1) and overall approached a statistically significant decrease in mortality when compared with patients treated with ICIs who did not have cirAEs (HR 0.670; P = 0.086) (Table S2; see Supporting Information). Sensitivity analyses conducted using a 6-month landmark and expansion of the ICI cohort to all cancers demonstrated similar trends (Tables S3, S4; see Supporting Information).

Table 1.

Multivariate Cox proportional hazards model of overall survival following PD-1 or PD-L1 immunotherapy^a

AiCTD within 12 months of ICIb	Cohort (n)	HRc	P-value	Number needed to harm
Chilblains	3	N/Ad	–	–
Cutaneous lupus erythematosus	31	0.702	0.53	N/Ae
Cutaneous sarcoidosis	6	N/Ad	–	–
Cutaneous small vessel vasculitis	38	0.554	0.105	5
Dermatopolymyositis	23	0.559	0.248	14
Discoid lupus erythematosus	21	0.885	0.865	42
Eosinophilic fasciitis	5	N/Ad	–	–
Lichen sclerosus et atrophicus	21	0.694	0.528	38
Morphoea	14	0.458	0.297	N/Ae
Polyarteritis nodosa	2	N/Ad	–	–
Scleroderma	22	1.128	0.829	N/Ae
All AiCTD	145	0.729	0.157	27

AiCTD, autoimmune cutaneous connective tissue disorders; cirAEs, cutaneous immune-related adverse events; HR, hazard ratio; ICI, immune checkpoint inhibitor; N/A, not applicable; PD-1, programmed death receptor-1; PD-L1, ­programmed death-ligand 1. AiCTD: de novo autoimmune cutaneous connective tissue disease in the setting of ­immune checkpoint inhibitor therapy. ^aUsing a 12-month landmark time and the four most common cancer indications for immune checkpoint inhibitor therapy (malignancies of the lung, digestive tract, urinary tract; or melanoma). ^bEach row corresponds to a separate multivariate model with only the coefficient for the primary independent variable from each model presented in the table. ^cHR of the impact of de novo AiCTD on overall survival when compared with non-AiCTD cirAEs. ^dFor patient privacy, individual groups with fewer than 10 patients are presented as having 10 patients in the TriNetX cloud platform. As a result, precise HRs could not be derived for these individual categories. However, the reporting of the All AiCTD group is based on actual patient counts. ^eThe attributable risk is 0 in these cohorts.

Although limited by retrospective design, our ­population-​level study suggests that de novo AiCTDs in the setting of ICI therapy are rare and are not associated with worse outcomes when compared with other cirAEs. This is likely to be due to overall similar utilization of systemic immunosuppression between the two study groups (a known independent risk factor for mortality in patients with cancer who are treated with ICIs).⁵ This study adds to the limited information on development of cutaneous connective tissue diseases following ICI administration, for which enhanced understanding is crucial to provide improved care to this challenging patient population.

Funding sources

Y.R.S. is supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases of the National Institutes of Health under Award Number K23AR080791.

Data availability

The data underlying this article are available in the article and in its online supplementary material.

Ethics statement

Datasets generated by the TriNetX Platform are deidentified as per the deidentification standard defined in Section §164.514(a) of the Health Insurance Portability and Accountability Act Privacy Rule. As such, this study was exempt from institutional review board approval.

Supplementary Information

All supplementary materials are available on Figshare: https://figshare.com/s/699dad8a139f581710c7.