Corticosteroid Use at Start of PD-1/PD-L1 Inhibitor Therapy Affects Outcomes in NSCLC

Get Permission

Kathryn C. Arbour, MD

Kathryn C. Arbour, MD

IN A STUDY reported in the Journal of Clinical Oncology, Kathryn C. Arbour, MD, of Memorial Sloan Kettering Cancer Center, and colleagues found that baseline treatment with corticosteroids was associated with poorer efficacy of programmed cell death protein 1/programmed cell death ligand 1 (PD-1/PD-L1) inhibitors in patients with advanced non–small cell lung cancer (NSCLC).1 As stated by the investigators, “On-treatment corticosteroids for treatment of immune-related adverse events do not seem to affect efficacy [of these agents], but the potential impact of baseline corticosteroids at the time of treatment initiation is unknown.”

Study Details

CLINICAL TRIALS of PD-1/PD-L1 inhibitors generally have excluded patients receiving baseline corticosteroids. Thus, the study examined the effect of baseline corticosteroids in PD-1/ PD-L1 inhibitor–naive patients starting single-agent therapy in the “real-world” clinical setting. The study included patients with advanced NSCLC who were treated with single-agent PD-1/PD-L1 inhibitors (pembrolizumab [Keytruda], nivolumab [Opdivo], atezolizumab [Tecentriq], or durvalumab [Imfinzi]) between April 2011 and September 2017 at Memorial Sloan Kettering Cancer (MSK; n = 455) and Gustave Roussy Cancer Center (GRCC; n = 185). The efficacy of treatment was assessed by local specialized radiologists using the Response Criteria in Solid Tumors v1.1, with all patients being followed until death or data lock in March 2017 for MSK and December 2017 for GRCC.

Use of Corticosteroids at the Start of Treatment

OVERALL, 90 of 640 patients (14%) treated with single-agent PD-1/PD-L1 blockade received corticosteroids (equivalent of ≥ 10 mg of prednisone) daily at the start of treatment, including 53 (12%) in the MSK cohort and 37 (20%) in the GRCC cohort. A total of 17 patients (3%) received the equivalent of < 10 mg of prednisone at the initiation of immune checkpoint inhibitor treatment and were included in the no-corticosteroid group. Indications for steroid treatment included dyspnea (33%), fatigue (21%), and brain metastases (19%). Clinicopathologic characteristics were generally similar in the corticosteroid and no-corticosteroid groups except for a higher proportion of patients in the corticosteroid group with an Eastern Cooperative Oncology Group (ECOG) performance status ≥ 2 in the MSK cohort (30% vs 8%, P < .01) and GRCC cohort (43% vs 17%, P < .01) and a higher proportion having a history of brain metastases in the MSK cohort (42% vs 23%, P < .01) and GRCC cohort (51% vs 13%, P < .01).

Association With Outcomes

IN THE MSK cohort, use of corticosteroids at ≥ 10 mg at the start of checkpoint inhibitor treatment was associated with a decreased overall response rate (6% vs 19%, P = .02), shorter median progression-free survival (1.9 vs 2.6 months, hazard ratio [HR] = 1.7, P = .001), and shorter median overall survival (5.4 months vs 12.1 months, HR = 2.1, P < .001). In the GRCC cohort, such use was associated with a nonsignificantly lower overall response rate (8% vs 18%, P = .2), reduced median progression-free survival (1.7 vs 1.8 months, HR = 1.5, P < .001), and reduced median overall survival (3.3 vs 9.4 months, HR = 2.0, P < .001). In multivariate analysis of the pooled population adjusting for smoking history, performance status, and history of brain metastases, baseline corticosteroid use remained significantly associated with decreased progression-free survival (HR = 1.31, P = .03) and overall survival (HR = 1.66, P < .001), with a borderline-significant effect on response rate (odds ratio [OR] = 0.42, P = .053).

In subgroup analyses of the pooled cohorts, hazard ratios for progression-free survival favored no/low corticosteroids vs the equivalent of ≥ 10 mg of prednisone for history and no history of brain metastases, nonsquamous and squamous histology, age < 65 and ≥ 65 years, men and women, ECOG performance status 0 to 1 and ≥ 2, and ever-smokers (but not for never-smokers). Hazard ratios for overall survival favored no/low corticosteroids for all of these subgroups.

“Baseline corticosteroid use of ≥ 10 mg of prednisone equivalent was associated with poorer outcome in patients with non–small cell lung cancer who were treated with [PD-1/PD-L1] blockade.”
— Kathryn C. Arbour, MD

Tweet this quote

Analysis in the pooled cohorts showed a similar reduction in efficacy with use of > 20 mg or 10 mg to 19 mg of corticosteroids compared with use of < 10 mg. Data on the timing of use, available only in the MSK cohort, showed that patients who received and discontinued corticosteroids at days 1 to 30 before initiation of immune checkpoint inhibitor treatment (66 patients) had progression-free and overall survival intermediate between those who received corticosteroids on the day that immune checkpoint inhibitor started (53 patients) and those who received no corticosteroids within 30 days of the start of treatment.

Three patients in each cohort who were receiving corticosteroids at the start of treatment exhibited a partial response to PD-1/ PD-L1 therapy. These 6 patients were receiving 10 mg to 20 mg of corticosteroids for palliative purposes, including fatigue, respiratory symptoms, and pain, and all had an ECOG performance status of 1. Response persisted for more than 15 months in 4 of the 6 patients.

The investigators concluded: “Baseline corticosteroid use of ≥ 10 mg of prednisone equivalent was associated with poorer outcome in patients with non–small-cell lung cancer who were treated with [PD-1/PD-L1] blockade. Prudent use of corticosteroids at the time of initiating [PD-1/PD-L1] blockade is recommended.” ■

DISCLOSURE: For full disclosures of the study authors, visit


1. Arbour KC, Mezquita L, Long N, et al: Impact of baseline steroids on efficacy of programmed cell death-1 and programmed death-ligand 1 blockade in patients with non–small-cell lung cancer. J Clin Oncol 36:2872-2878, 2018.

Related Articles

Corticosteroids and Immunotherapy: Guilty or Guilt By Association?

Thomas E. Stinchcombe, MD

Thomas E. Stinchcombe, MD

FIRST-LINE TREATMENT options for advanced non–small cell lung cancer (NSCLC) include single- agent immunotherapy for patients with a programmed cell death ligand 1 (PD-L1) tumor proportion score of at least 50% or in combination with chemotherapy in unselected...