Immune checkpoint inhibitor therapy may be less effective if the patient is receiving corticosteroids at the time that treatment is initiated, results of a retrospective, two-center analysis suggest.
Corticosteroid use at the time of starting a PD-1 or PD-L1 inhibitor was associated with significantly reduced overall survival and progression-free survival in patients with non–small-cell lung cancer, the authors of the analysis reported.
“Prudent use of corticosteroids at the time of initiating PD-1/PD-L1 blockade is warranted,” said Matthew D. Hellmann, MD, of Memorial Sloan Kettering Cancer Center, New York, and his coauthors. The report is in the Journal of Clinical Oncology.
The retrospective analysis from Dr. Hellmann and his colleagues comprised 640 patients treated with single-agent atezolizumab, durvalumab, nivolumab, or pembrolizumab at Memorial Sloan Kettering Cancer Center or Gustave Roussy Cancer Center, Villejuif, France, between April 2011 and September 2017.
Ninety of those patients (14%) were receiving at least 10 mg of prednisone equivalent at the time the immune checkpoint inhibitor was started, a review of patient records revealed. About one-third were receiving corticosteroids for dyspnea or other respiratory symptoms. Another 21% had fatigue and 19% had brain metastases prompting corticosteroid treatment.
Corticosteroid use at the time of PD-1/PD-L1 blockade was associated with decreased progression-free survival (hazard ratio, 1.31; P = .03) and overall survival (HR, 1.66; P less than .001), as well as decreased overall response rate in a multivariable analysis adjusted for performance status, history of smoking, and history of brain metastases, the investigators reported.
Reduced efficacy also was seen in analyses that looked at each center separately. For the Memorial Sloan Kettering cohort (455 patients), baseline corticosteroid use was associated with significantly decreased overall response rate, shorter progression-free survival, and shorter overall survival.
For the smaller French cohort, (185 patients), there was a statistically nonsignificant decrease in overall response rate, along with significant reductions in progression-free survival and overall survival.
Based on these data, it may be prudent to try other pharmacologic or nonpharmacologic strategies to manage cancer symptoms if treatment with a PD-1/PD-L1 blocker is planned, Dr. Hellmann and his coauthors said.
“These strategies could enable patients to be tapered off corticosteroids before the start of PD-1/PD-L1 blockade to potentially achieve maximum benefit from these agents,” they wrote.
However, medically necessary corticosteroids should not be avoided, such as those given for management of brain metastases, they added.
Interestingly, other recent studies have suggested that patients already on PD-1/PD-L1 inhibitors do not seem to have decreased efficacy when corticosteroids are prescribed to manage emergent immune-related adverse events.
That’s “fortunate,” given that corticosteroids are a mainstay for the management of these characteristic adverse events in patients receiving immune checkpoint inhibitors, they said.
Dr. Hellmann reported research funding from Bristol-Myers Squibb, and a consulting or advisory role with AstraZeneca, Bristol-Myers Squibb, Genentech, Janssen Pharmaceuticals, MedImmune, Merck, Mirati Therapeutics, Shattuck Labs, and Syndax. Coauthors reported disclosures related to Amgen, Boehringer Ingelheim, Pfizer, and Roche, among others.
SOURCE: Arbour KC et al. J Clin Oncol. 2018 Aug 20. doi: 10.1200/JCO.2018.79.0006.