br LIMITATIONS br Some limitations accompanied our study

LIMITATIONS
Some limitations accompanied our study. Current cancer immunotherapy is still in a state of flux, with rapid changes in the number of agents available, as well as in the approved indications for the existing agents. As the demographics of the patients receiving immune checkpoint therapy change over time, the relative percentages of immune-related adverse effects that require ED visits will also change, which could affect the generalizability of our results. A limitation of our methodology was that the physicians performing the data abstraction were not blinded to the clinical outcomes or the previous abstraction results when performing data confirmation. Given the retrospective nature of our study, using electronic medical records might have resulted in erroneous data, yet the strong medical record system in our institution and the large number of patients included helped in overcoming this limitation. Also, using the last confirmed contact with the health care system when a patient death could not be verified might have resulted in underestimation of the true survival rate. Another limitation of our study is that it NPS-2143 involved only one institution, and the sample size is limited. Our findings about the association of colitis with good prognosis and the association of pneumonitis with poor prognosis are interesting but will need to be further investigated and confirmed in different and larger cohorts.
DISCUSSION
Of all the immune-related adverse effects investigated in our study, diarrhea was the most common one associated
El Majzoub et al Adverse Effects of Immune Checkpoint Therapy in Cancer Patients
Table 4. Univariable and multivariable Cox regression analysis for overall survival for each immune-related adverse effect.
Multivariable,
*HR adjusted for age, sex, race, Charlson comorbidity index, cancer type, preexisting autoimmune disease, and presence of allergy.
†HR adjusted for age, sex, race, Charlson comorbidity index, cancer type, and presence of allergy.
with ED visits, ranging from 6.4% to 18.4% among different treatment groups. Although hypophysitis is a common immune-related adverse effect in ED visits, its relative prevalence was different among the 4 medication groups in our study (Table 3). ED visits related to hypophysitis appeared to be associated more with combination therapy and ipilimumab than with pembrolizumab or nivolumab. Thyroiditis also was a common immune-related adverse effect among the ED visits, and its relative prevalence was also different among the 4 treatment groups, in a pattern similar to bronchi for hypophysitis. These patterns reflected the previous finding that patients receiving combination therapy are at increased risk for thyroid dysfunction and hypophysitis.44 Pancreatitis was another immune-related adverse effect for which the relative prevalence differed among the 4 treatment groups (Table 3). ED visits related to pancreatitis appeared to be associated more with combination therapy and nivolumab than with pembrolizumab or ipilimumab.
Although myocarditis is a potentially fatal complication of immune checkpoint therapy,27,44 only one case of
(nonfatal) myocarditis was observed, in a patient who visited the ED after receiving more than one immune checkpoint therapy medication. Other identified immune-related adverse effects, such as colitis, pneumonitis, dermatitis, adrenalitis, nephritis, vasculitis, and hematologic and eye adverse effects, were less common in our study and did not show significant differences in proportions among the 4 treatment groups. Overall, the prevalence of the immune-related adverse effects among immune checkpoint therapy–treated cancer patients who visited the ED generally reflected the incidence of immune-related adverse effects in the literature (Table E1, available online at http://www.annemergmed.com).