Introduction

Immune checkpoint inhibitors (ICIs) are antineoplastic agents that mediate anticancer effects by upregulation of the immune response to enhance immune surveillance. ICIs may cause a variety of immune mediated adverse events (irAEs) by enhancing autoimmunity that can involve various organ systems. Hematologic irAEs have been reported in the literature including auto-immune hemolytic anemia (AIHA), hemophagocytic lymphohystiocytosis (HLH), immune thrombocytopenic purpura (ITP), and thrombotic thrombocytopenic purpura (TTP). Heparin-induced thrombocytopenia (HIT) is also an immune related process particularly relevant for cancer patients who are frequently exposed to heparin products. Given the nature of HIT as an autoimmune antibody mediated disease process, it is mechanistically plausible (although unexplored) that ICIs increase the risk of HIT. All of these events can be severe or life-threatening, and exact estimates of their incidence and natural history are limited given their rarity.

Methods

We conducted a retrospective cohort study of patients who received ICI therapy at Beth Israel Deaconess Medical Center from January 1, 2016 through June 30, 2024. Hematologic irAEs were identified utilizing ICD-9 and ICD-10 billing codes (including AIHA, HLH, TTP and ITP) and manually verified by two physicians. HIT was also included in the exploratory analysis as a secondary endpoint. The date of initiation of ICI therapy was used as index period, and patients were followed to last follow up, date of death, or date of irAE development, whichever was earliest. Outcomes were reported as incidence rates with 95% confidence intervals with death as a competing risk.

Results

We included 763 patients with a mean age of 65.7 years (IQR 59.2 – 74.5) at the time of initiating ICI therapy; 38.8% were female, and 74.7% White. The most common malignancies included lung (27.4%), melanoma (17.2%), and hepatobiliary (16.1%). Metastatic disease was present in 33.2%. Specific ICI agents included nivolumab (60.7%), atezolizumab (19.3%), and durvalumab (18.4%); and 21.4% of patients received more than one. Mortality rates at 6 and 12 months after ICI initiation were 20.1% and 32.1%, respectively.

The incidence rate for hematologic irAEs was 0.28 events per 100 person years (95% CI 0.08 – 0.71). Four patients had one of the hematologic irAEs (not including HIT) which were queried for; 3 had AIHA and one had HLH. None were identified as having TTP or ITP. Preceding ICI therapy included pembrolizumab (n=2), nivolumab (n=1) and combination nivolumab/ipilimumab (n=1). Mean time to development of irAE was 75.8 days (ranging from 29 days to 144 days). Underlying cancers included hepatobiliary, lung, breast, and melanoma.

Of the AIHA cases, 2 were direct antiglobulin test (Coombs) negative and one was direct antiglobulin test positive. All cases responded well with steroid immunosuppression although the dosage varied (0.5-1mg/kg) as well as length of taper (4 to 8 weeks). One of the patients with an irAE (after nivolumab/ipilimumab) was later rechallenged with a nivolumab and tolerated this successfully without recurrence of AIHA. Three of the 4 patients are alive at time of data censoring (with length of follow up 2.0, 6.2, and 7.8 years), and one passed away 41 days after being diagnosed with AIHA due to septic shock from pneumonia.

Two patients in the cohort developed HIT after ICI initiation in the setting of recent heparin exposure (7 and 13 days prior). One had lung cancer treated with pembrolizumab and the other had hepatobiliary cancer treated with nivolumab. In both cases the interval between initiation of ICI and diagnosis of HIT was 29 days. Timing of death was 2 weeks and 4 weeks after diagnosis of HIT.

Conclusions

Over eight years, we systematically analyzed a cohort of patients with solid malignancies receiving ICI therapy and found that hematologic irAEs are rare but potentially serious. AIHA was the most frequent hematologic irAE; antiglobulin test results were variable but steroid responsive. Although no irAE ITP or TTP cases were identified, alterations in hematologic parameters are common in malignancy, including from chemotherapy or the cancer itself, so some may have gone unrecognized. HIT was diagnosed in two cases shortly after ICI initiation. These data highlight the need for multicenter databases employing a systematic approach to capture and characterize hematologic irAEs across diverse patient populations.

This content is only available as a PDF.
Sign in via your Institution