Key PointsQuestion
Do women derive less advantage from immune checkpoint inhibitors, compared with standard systemic therapy, in the treatment of advanced solid-organ malignant neoplasm?
Findings
In this systematic review and meta-analysis of 23 randomized clinical trials of immunotherapy for advances solid-organ cancers including 9322 men and 4399 women, overall survival from immunotherapy was found in both men and women, with no statistically significant differences between the sexes.
Meaning
The response to immune checkpoint inhibitors does not appear to differ on the basis of patient sex.
Importance
Sex-associated differences in immune response are known, but a meta-analysis suggested men, compared with women, derive greater value from immunotherapy for advanced solid-organ malignant neoplasms. However, methodologic concerns and subsequent trials have placed these results in doubt.
Objective
To perform an updated, comprehensive meta-analysis that assesses the efficacy of immunotherapy in advanced cancers according to patient sex.
Design, Setting, and Participants
A systematic review of studies (n = 23) indexed in MEDLINE (PubMed), Embase, and Scopus from inception of these databases to October 2, 2018, was conducted. Randomized clinical trials that compared immunotherapy with standard of care in the treatment of advanced solid-organ malignant neoplasms were included if overall survival was reported as an outcome and if data stratified by patient sex were available. Observational studies, editorials, commentaries, review articles, non–peer-reviewed publications, studies that compared various immunotherapy regimens, studies that reported other measures of oncologic response, and studies that reported subgroup analyses for 1 sex only were excluded.
Main Outcomes and Measures
Overall survival, with a test for heterogeneity between women and men, to assess the null hypothesis that no difference in the survival advantage of immunotherapy exists by patient sex.
Results
This meta-analysis included 23 randomized clinical trials that reported on 9322 men (67.9%) and 4399 women (32.1%); the age of most patients was in the 70s. An overall survival benefit of immunotherapy was found for both men (hazard ratio [HR], 0.75; 95% CI, 0.69-0.81; P < .001) and women (HR, 0.77; 95% CI, 0.67-0.88; P = .002). Random-effects meta-analysis of study-level differences in response to immunotherapy demonstrated no statistically significant difference between the sexes (I2 = 38%; P = .60). Subgroup analyses according to disease site, line of therapy, class of immunotherapy, study methodology, and representation of women recapitulated these findings.
Conclusions and Relevance
Stratified analyses demonstrated no statistically significant association of patient sex with the efficacy of immunotherapy in the treatment of advanced cancers using overall survival as the outcome.
Women and men differ in their immunologic response to both foreign and self-antigens, with women typically having stronger innate and adaptive immune responses.1 Compared with men, women not only experience a higher prevalence of systemic autoimmune disease1,2 but also have a greater response to vaccination and a lower severity and prevalence of many infectious conditions.1,3,4 In oncology, differences in immune response have been postulated to underlie observed differences in prevalence and mortality from many cancers.5,6
Immune checkpoint inhibitors targeting cytotoxic T-lymphocyte antigen-4 (CTLA-4) and programmed cell death 1 (PD-1) have demonstrated higher efficacy than standard of care (SOC) chemotherapeutic approaches in several malignant neoplasms. Sex hormone modulation of the PD-1/programmed cell death 1 ligand 1 (PD-L1) pathway has been demonstrated in animal models.7,8 Thus, it has been postulated that the advantages of immunotherapy may vary according to patient sex.9 Recently, Conforti et al10 found in a meta-analysis of randomized clinical trials that men derived greater value from immune checkpoint inhibitors compared with women (hazard ratio [HR], 0.72 [95% CI, 0.65-0.79] vs 0.86 [95% CI, 0.79-0.93]; P = .002). However, another recent analysis has presented conflicting data: No difference in advantages between nivolumab and everolimus was seen among men and women with metastatic renal cell carcinoma.11
The Conforti et al10 meta-analysis demonstrated a difference between patient sex, which presents a number of limitations that preclude strong conclusions from being drawn from the data set. First, the meta-analysis included a limited subset of approved immunotherapeutic agents. Second, several comprehensive and updated studies that met the inclusion criteria, including those with a more robust representation of female patients, have been published since the Conforti et al10 literature review.
To address these concerns, we performed a systematic review and meta-analysis that examine the association of patient sex with the advantages of immunotherapy in patients with advanced cancer. We used a more contemporary and comprehensive literature search strategy.
This systematic review and meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines.12 The study protocol was registered with PROSPERO.
We included randomized clinical trials. Observational studies (whether cohort or case-control in design), editorials, commentaries, and review articles were excluded. Publications that were not subject to peer review (ie, reports of data from the National Vital Statistics System and dissertations or theses) were also excluded. To prevent the duplication of patients used in our analyses, we selected 1 study (when more than 1 was published about the same patient cohort), on the basis of contemporary timing, cohort size, and granularity of data reported.
We included studies that compared immunotherapy for metastatic cancers with other systemic treatment regimens, including chemotherapy-based regimens and those that used other targeted therapies. Analyses that examined immunotherapy-chemotherapy combinations compared with chemotherapy alone were also included. However, studies that compared various immunotherapy regimens were excluded.
Outcome and Exposure of Interest
The outcome of interest was overall survival (OS) and whether OS was reported as the primary or secondary outcome of the original study. Studies that reported other measures of oncologic response, including progression-free survival and objective response rate (without OS data) were excluded as these may not be comparable across histologic subtypes.
We sought to examine whether patient sex modified the association between immunotherapy (compared to chemotherapy) and OS. Studies that did not report analyses stratified by sex in the original trials were excluded. Furthermore, to exclude ecologic bias, we excluded studies that reported subgroup analyses for 1 sex only.
Search Strategy and Review Method
To perform this present analysis, we updated a previous relevant systematic review by Conforti et al10 that used MEDLINE (PubMed), Embase, and Scopus from inception of these databases to November 30, 2017, to identify phase 2 or 3 randomized clinical trials for the agents ipilimumab, tremelimumab, nivolumab, and pembrolizumab. In this update, we expanded the literature search for previously included agents from November 30, 2017, to October 2, 2018. We expanded the search criteria to include atezolizumab, durvalumab, and avelumab and searched the relevant databases from inception to October 2, 2018. References from review articles, editorials, and included studies were reviewed and cross-referenced to ensure completeness. No limitations were placed regarding publication language or publication year. After the literature search, we excluded all duplicates. References from review articles, commentaries, editorials, included studies, and conference publications of relevant medical societies were reviewed and cross-referenced to ensure completeness.
We (M.B. and Z.K.) performed study selection independently, and we resolved disagreements by consensus with the primary author (C.J.D.W.). Titles and abstracts were used to screen for initial study inclusion. Full-text reviews were performed if the abstracts were insufficient for determining if the studies met the inclusion or exclusion criteria. We (C.J.D.W., M.B., and Z.K.) developed a data extraction form by consensus. One of us (M.B.) performed all of the data extraction, and two of us (C.J.D.W. and Z.K.) conducted independent verification.
Study characteristics, including first author, year of publication, study design, phase, type of therapy (anti–PD-L1 or anti–CTLA-4), line of therapy, underlying malignant neoplasm, and baseline demographic characteristics were extracted. In addition, outcome information, including the HR (with 95% CI) for death, stratified by patient sex, was abstracted. A risk-of-bias assessment was conducted using the Cochrane Collaboration tool for assessing risk of bias.13
Assessment of Heterogeneity and Statistical Analysis
We identified heterogeneity using the Q test. Heterogeneity was estimated using the DerSimonian-Laird method and was quantified using I2 values.14
Meta-analysis was performed using Review Manager, version 5.3 (Nordic Cochrane Centre). We used the inverse variance technique for meta-analysis of HRs. Because of the clinical heterogeneity inherent in the data, we used random-effects models for all meta-analyses. To assess the differences between the sexes in each study while accounting for study-level associations, we made calculations using log HR and then assessed whether the variations differed from the null using the χ2 test.14 All reported P values are 2-sided, and P = .05 was used to indicate statistical significance.
We performed a number of prespecified subgroup analyses to assess the potential association of oncologic and methodologic factors in effect modification of patient sex with immunotherapy efficacy. We considered subgroups, including disease site (melanoma, non–small cell lung cancer [NSCLC], and other tumor sites), line of therapy (first line and subsequent), class of immunotherapy (anti–CTLA-4, anti–PD-1, and anti–PD-L1), and study methodology (various immunotherapies vs chemotherapy alone, immunotherapy-chemotherapy combination vs chemotherapy alone).
To assess the degree to which the relative underrepresentation of women in these trials may contribute to the previously observed differences in outcome between women and men, we performed a stratified analysis according to the proportion of women in each study. We categorized the included studies according to whether women represented less than 20%, 20% to 30%, 30% to 40%, or 40% or more of the study cohort.
Literature Search Results
The literature search identified 57 unique references. After a full-text review of 17 studies, we identified 7 relevant clinical trials for inclusion, in addition to 16 trials included from the Conforti et al10 meta-analysis. Thus, a total of 23 trials15-37 were included in the present meta-analysis (Figure 1). All trials included a subgroup analysis, stratified by sex, that compared the intervention group with the control group, with an HR for OS.
Characteristics of Identified Trials
Table 1 lists the main characteristics of the 23 trials. In total, 13 721 patients were included, of which 9322 (67.9%) were men and 4399 (32.1%) were women; the age of most patients was in the 70s. All studies enrolled patients within the past decade, and most trials were published in the past 3 years. Only 2 studies18,29 (9%) evaluated OS as a secondary end point (the primary end point was progression-free survival), both of which allowed crossover to immunotherapy at the time of disease progression. There were 11 trials (48%) for patients with NSCLC,16-18,20-22,29,31,34-36 4 (17%) for melanoma,24,30,32,33 2 (9%) for clear cell renal cell carcinoma,26,27 2 (9%) for SCLC,28,37 and 1 (4%) each for urothelial carcinoma,15 head and neck squamous carcinoma,19 mesothelioma,25 and gastric or gastroesophageal carcinoma.23 Most trials evaluated immunotherapy after previous systemic therapy failure; however, 11 trials (48%) assessed the efficacy for OS in the first-line setting.18,20,21,27-30,32,33,35,37 Most trials used a PD-L1 or PD-1 inhibitor as the immunotherapy agent, whereas 6 trials (26%) used a CTLA-4 inhibitor.21,25,27,28,30,32 Most study designs included immunotherapy vs SOC, but 6 trials20,21,28,32,35,37 (26%) were designed as immunotherapy and SOC vs SOC alone.
Several trials had unique designs that may warrant further explanation. The KEYNOTE 010 trial22 was unique in that it tested 2 doses of pembrolizumab (2 mg/kg and 10 mg/kg) vs docetaxel among patients with NSCLC, with an overall pooled HR for OS of 0.67 (95% CI, 0.56-0.80). In the CheckMate 214 study, Motzer et al27 randomized 1096 patients with advanced clear cell renal cell carcinoma to receive both an anti–PD-1 (nivolumab) and an anti–CTLA-4 (ipilimumab) agent vs sunitinib; however, Motzer et al27 performed OS analysis on only a sex subgroup among 847 patients with intermediate- or poor-risk disease, demonstrating a survival advantage for both sexes (men HR, 0.71 [95% CI, 0.55-0.92]; women HR, 0.52 [95% CI, 0.34-0.78]).
The median age of patients included was typically in the 70s; however, in 2 trials, the median age was in the 60s.30,32 Most studies tended to have short follow-up, although 3 trials (13%) had a median follow-up of 24 months or more.24,27,34 Overall, all but 7 studies18,21,24,25,28,30,36 (30%) showed an OS advantage for patients who received immunotherapy compared with the control group. In subgroup analyses, 14 studies15-17,19,20,22,23,26,27,29,31-33,35 (61%) demonstrated a survival advantage from immunotherapy among men and 7 studies20,22,27,31,33-35 (30%) showed this advantage among women.
Risk of bias of the included trials is shown in the eTable in the Supplement. All trials included random-sequence generation and were at low risk for selection bias. There was intermittent reporting of allocation concealment; several studies were at risk for selection bias because of this criterion. Generally, all studies were at low risk for attrition and reporting bias. Several studies were unblinded and were thus at risk for performance and detection bias; however, for the outcome of OS, such a lack of blinding is likely inconsequential as blinding is unlikely to affect the outcome.
Meta-analysis of the available literature demonstrated a statistically significant advantage in OS for patients who received immunotherapy compared with other systemic therapies (HR, 0.75, 95% CI, 0.70-0.81; P < .001; I2 = 61%). Compared with SOC systemic therapy, an OS advantage of immunotherapy was observed for both men (HR, 0.75; 95% CI, 0.69-0.81; P < .001) and women (HR, 0.77; 95% CI, 0.67-0.88; P = .002); however, we found no statistically significant difference in OS advantage between the sexes (P = .60; I2 = 38%) (Table 2, Figure 2). Statistically significant heterogeneity was demonstrated among both men (tau2 = 0.02; χ2 = 51.67; P = .003; I2 = 57%) and women (tau2 = 0.07; χ2 = 62.29; P < .001; I2 = 65%).
We performed a number of subgroup analyses according to disease site, line of therapy, class of immunotherapy, and study methodology. No statistically significant differences in the efficacy of immunotherapy were found between men and women in any of these analyses (Table 2). Finally, we examined for the effect of the prevalence of women in the study cohort. Again, no statistically significant differences were demonstrated among these subgroups (Table 2).
Contrary to the published meta-analysis by Conforti et al,10 which suggested a greater immunotherapy advantage compared with SOC systemic therapy for men than women, the present analysis found no difference in OS from immune checkpoint inhibitors when comparing the efficacy of these treatments between the sexes. Furthermore, when assessing several subgroup analyses, including disease site, line of therapy, class of immunotherapy, and study methodology, we could not demonstrate any significant sex-associated differences in efficacy.
These conflicting results may be explained in a number of ways. First, we excluded 3 trials, which were included in Conforti et al,10 that compared various immunotherapy regimes.38-40 By including only those trials that compared an immunotherapy group with a nonimmunotherapy control, we were able to specifically assess the association of sex with response to immunotherapy. Second, we expanded the search criteria to include immunotherapy agents that were not considered in Conforti et al.10 The resulting search included a trial of atezolizumab in patients with NSCLC.31 This trial demonstrated a greater net value of immunotherapy for women (HR, 0.64; 95% CI, 0.49-0.85) than for men (HR, 0.79; 95% CI, 0.64-0.97). Because this trial was large (n = 850 patients), it contributed considerably to the pooled HR effect.
Third, we updated the search previously performed and identified 7 recent large trials that have been published since the end date for inclusion in the Conforti et al10 meta-analysis. Gandhi et al20 in KEYNOTE 189 tested pembrolizumab plus platinum chemotherapy (n = 410) vs placebo plus platinum chemotherapy (n = 206) in the first-line setting among patients with NSCLC. KEYNOTE 189 included 363 men and 253 women and noted a strong OS advantage from immunotherapy among women (HR, 0.29; 95% CI, 0.19-0.44) compared with men (HR, 0.70; 95% CI, 0.50-0.99). Motzer et al27 in CheckMate 214 tested nivolumab plus ipilimumab (n = 425) and the tyrosine-kinase inhibitor sunitinib (n = 422) in the first-line setting among intermediate- and poor-risk patients with clear cell renal cell carcinoma. CheckMate 214 had 615 men and 232 women and also found a strong OS advantage from immunotherapy among women (HR, 0.52; 95% CI, 0.34-0.78) compared with men (HR, 0.71; 95% CI, 0.55-0.92). KEYNOTE 407 tested first-line pembrolizumab vs saline placebo (plus carboplatin and either paclitaxel or nanoparticle albumin-bound paclitaxel) among patients with NSCLC.35 Although women made up only 18.6% of participants in KEYNOTE 407, they had a remarkable immunotherapy treatment advantage (HR, 0.42; 95% CI, 0.22-0.81) compared with men (HR, 0.69; 95% CI, 0.51-0.94). Taken together, the present meta-analysis provides a more specific assessment of the research question while including a greater number of immunotherapy agents and an updated search.
Small samples sizes may result in an elevated false discovery rate41 or even false-positive results.42 Meta-analyses of such trials may propagate such findings by enhancing the statistical power of these small subgroup analyses. To explore the final hypothesis that the representation of women in a study may mediate observed differences in immunotherapy efficacy between men and women, we performed a stratified analysis. We found no statistically significant difference in outcomes between men and women regardless of the proportion of women in the study cohort. The effect estimates favored men in studies with study cohorts composed of less than 20% women, but the results were very comparable in the other subgroups. Trials with an underrepresentation of women may present spurious results for sex-specific subgroup analyses, as evidenced by the wide CIs when less than 20% of the cohort represented is women. Six20,27,31,34,36,37 of the 7 trials included in this meta-analysis but not included in the Conforti et al10 study had more than 27% women representation, with 2 trials20,31 having more than 38% women inclusion.
Strengths and Limitations
The strengths of this meta-analysis include the strict methodologic inclusion criteria that required the comparison between an immunotherapy group and a nonimmunotherapy control group; the broad inclusion of all approved immunotherapy agents; and the rigorous, up-to-date search strategy. As a result, this analysis provides a comprehensive assessment of the association of patient sex with response to immunotherapy compared with nonimmunotherapy, including data on more than 13 000 patients. Furthermore, we undertook several subgroup analyses in an attempt to ascertain any differences in immunotherapy efficacy between the sexes.
This analysis has several limitations. First, it relies on published clinical trial subgroup HRs and not on individual patient-level data. Second, residual confounding is possible in that differences other than sex contribute to immunotherapy response and OS. Third, differences in outcomes between men and women may be ascribed to other factors (including differences in lifestyle, comorbidities, incidence of autoimmune diseases, and other factors) that are unaccounted for in clinical trials. Fourth, as in all clinical trials, the included studies are at risk for having nongeneralizable results (the so-called efficacy-effectiveness gap) because of referral bias and strict inclusion criteria, among other factors, that result in the underrepresentation of uninsured, low-income, and minority populations. Meta-analysis of these trials, such as the one we performed, is subject to the same limitations. Finally, the trials that we excluded because of a lack of published sex-subgroup analyses may demonstrate sex differences if analyzed in this fashion.
In this contemporary meta-analysis of all available immunotherapy clinical trials across all disease sites, we found no difference in immunotherapy efficacy or OS between women and men. Contrary to findings of a previous analysis, we found no evidence that sex should be considered when deciding whether to offer immunotherapy to patients with advanced cancers.
Accepted for Publication: October 10, 2018.
Corresponding Author: Christopher J. D. Wallis, MD, PhD, Division of Urology, Department of Surgery, University of Toronto, 149 College St, Room 503G, Toronto, ON M5T 1P5, Canada (wallis.cjd@gmail.com).
Published Online: January 3, 2019. doi:10.1001/jamaoncol.2018.5904
Author Contributions: Drs Wallis and Klaassen had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Wallis, Hamid, Pal, Klaassen.
Acquisition, analysis, or interpretation of data: Wallis, Butaney, Satkunasivam, Freedland, Patel, Hamid, Klaassen.
Drafting of the manuscript: Wallis, Patel, Hamid, Klaassen.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Wallis, Satkunasivam.
Administrative, technical, or material support: Wallis, Butaney, Patel, Pal, Klaassen.
Supervision: Patel, Pal, Klaassen.
Conflict of Interest Disclosures: Dr Freedland reported receiving grants from Merck outside of the submitted work. Dr Patel reported receiving scientific advisory income from AstraZeneca, BMS, Illumina, Tempus, and Novartis. Dr Patel's university receives research funding from Bristol-Myers Squibb, Eli Lilly, Fate, Incyte, AstraZeneca/MedImmune, Merck, Pfizer, Roche/Genentech, Xcovery, Fate Therapeutics, Genocea, and Iovance. Dr Pal reported receiving personal fees from Genentech, Pfizer, and BMS outside of the submitted work. No other disclosures were reported.
8.Polanczyk
MJ, Hopke
C, Vandenbark
AA, Offner
H. Estrogen-mediated immunomodulation involves reduced activation of effector T cells, potentiation of Treg cells, and enhanced expression of the PD-1 costimulatory pathway.
J Neurosci Res. 2006;84(2):370-378. doi:
10.1002/jnr.20881PubMedGoogle ScholarCrossref 11.Graham
J, Abdel-Rahman
O, Choueiri
TK, Heng
DYC; International mRCC Database Consortium. Re: Fabio Conforti, Laura Pala, Vincenzo Bagnardi, et al. Cancer immunotherapy efficacy and patients’ sex: a systematic review and meta-analysis.
Lancet Oncol 2018;19:737-46.
Eur Urol. 2018;S0302-2838(18)30482-2.
doi:
10.1016/j.eururo.2018.07.004PubMedGoogle Scholar 13.Higgins
JP, Altman
DG, Gøtzsche
PC,
et al; Cochrane Bias Methods Group; Cochrane Statistical Methods Group. The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials.
BMJ. 2011;343:d5928. doi:
10.1136/bmj.d5928PubMedGoogle ScholarCrossref 23.Kang
YK, Boku
N, Satoh
T,
et al. Nivolumab in patients with advanced gastric or gastro-oesophageal junction cancer refractory to, or intolerant of, at least two previous chemotherapy regimens (ONO-4538-12, ATTRACTION-2): a randomised, double-blind, placebo-controlled, phase 3 trial.
Lancet. 2017;390(10111):2461-2471. doi:
10.1016/S0140-6736(17)31827-5PubMedGoogle ScholarCrossref 24.Larkin
J, Minor
D, D’Angelo
S,
et al. Overall survival in patients with advanced melanoma who received nivolumab versus investigator’s choice chemotherapy in CheckMate 037: a randomized, controlled, open-label phase III trial.
J Clin Oncol. 2018;36(4):383-390. doi:
10.1200/JCO.2016.71.8023PubMedGoogle ScholarCrossref 25.Maio
M, Scherpereel
A, Calabrò
L,
et al. Tremelimumab as second-line or third-line treatment in relapsed malignant mesothelioma (DETERMINE): a multicentre, international, randomised, double-blind, placebo-controlled phase 2b trial.
Lancet Oncol. 2017;18(9):1261-1273. doi:
10.1016/S1470-2045(17)30446-1PubMedGoogle ScholarCrossref 28.Reck
M, Luft
A, Szczesna
A,
et al. Phase III randomized trial of ipilimumab plus etoposide and platinum versus placebo plus etoposide and platinum in extensive-stage small-cell lung cancer.
J Clin Oncol. 2016;34(31):3740-3748. doi:
10.1200/JCO.2016.67.6601PubMedGoogle ScholarCrossref 31.Rittmeyer
A, Barlesi
F, Waterkamp
D,
et al; OAK Study Group. Atezolizumab versus docetaxel in patients with previously treated non-small-cell lung cancer (OAK): a phase 3, open-label, multicentre randomised controlled trial.
Lancet. 2017;389(10066):255-265. doi:
10.1016/S0140-6736(16)32517-XPubMedGoogle ScholarCrossref 34.Antonia
SJ, Villegas
A, Daniel
D,
et al; PACIFIC Investigators. Overall survival with durvalumab after chemoradiotherapy in Stage III NSCLC [published online September 25, 2018].
N Engl J Med. 2018. doi:
10.1056/NEJMoa1809697PubMedGoogle Scholar 36.Barlesi
F, Vansteenkiste
J, Spigel
D,
et al. Avelumab versus docetaxel in patients with platinum-treated advanced non-small-cell lung cancer (JAVELIN Lung 200): an open-label, randomised, phase 3 study [published online September 21, 2018].
Lancet Oncol. 2018;S1470-2045(18)30673-9. doi:
10.1016/S1470-2045(18)30673-9PubMedGoogle Scholar 37.Horn
L, Mansfield
AS, Szczęsna
A,
et al; IMpower133 Study Group. First-line atezolizumab plus chemotherapy in extensive-stage small-cell lung cancer [published online September 25, 2018].
N Engl J Med. 2018. doi:
10.1056/NEJMoa1809064PubMedGoogle Scholar 39.Hodi
FS, Chesney
J, Pavlick
AC,
et al. Combined nivolumab and ipilimumab versus ipilimumab alone in patients with advanced melanoma: 2-year overall survival outcomes in a multicentre, randomised, controlled, phase 2 trial.
Lancet Oncol. 2016;17(11):1558-1568. doi:
10.1016/S1470-2045(16)30366-7PubMedGoogle ScholarCrossref