A Systematic Review Assessing the Underrepresentation of Cancer Patients in COVID-19 Trials

The new severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and the resulting illness, coronavirus disease 2019 (COVID-19), have emerged as a global pandemic that threatens thousands around the world [1, 2]. Several clinical features are linked to mortality due to COVID-19 patients, including old age and the presence of comorbid conditions, such as cardiovascular disease, diabetes, chronic respiratory disease, hypertension, and cancer [3, 4].

In a multicenter study in China including 105 hospitalized COVID-19 patients with cancer and 536 age-matched COVID-19 noncancer patients, Dai et al. [4] observed a high mortality rate with the need of ICU and mechanical ventilation in those with cancer patients compared to the matched sample of COVID-19 patients without cancer [5]. In a large cohort study, data were collected on adult patients with active or previous malignancy, with confirmed COVID-19 infection. Over 900 patients with COVID-19 and cancer were included. Patients with cancer appeared to be at increased risk of mortality and severe illness due to COVID-19, regardless of whether they have active cancer, are on anticancer treatment, or both. Several cancer-specific factors were associated with increased 30-day all-cause mortality as cancer status (present, stable, or responding disease vs. remission or no evidence of disease OR 1.79 [95% CI: 1.09–2.95]) and present, progressive disease versus remission or no evidence of disease OR 5.20 [95% CI: 2.77–9.77]) [6]. In a cohort of 4,184 patients with COVID-19, including 233 with active cancer, the authors found that patients with an active cancer diagnosis were more likely to die from COVID-19. Those with hematological malignancies were at the highest risk of death. Patients receiving cancer-directed therapy within 3 months before hospitalization had no overall increased risk of death [7]. All these data are derived from observational cohort studies.

Cancer patients are probably at a higher risk of developing severe COVID-19 because the malignancy and chemotherapy treatment may negatively affect the immune system. Moreover, their immunocompromised condition may increase the risk of infection [8]. Many medications are being actively investigated as COVID-19 therapy or prophylaxis, with over 7,000 studies currently registered at the clinicaltrials.gov registry.

Since patients with cancer seem to be at increased risk of mortality and adverse outcomes according to the data from observational studies, we aimed to examine the representation of cancer patients (hematological malignancies and solid tumors) in COVID-19 clinical therapeutic and prophylactic interventional trials. Our goals were to evaluate if patients with cancer were explicitly or indirectly excluded from the clinical therapeutic research, whether their proportion and outcomes are reported, and to assess their outcomes.

This review was conducted and reported in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [9]. In this review, all registered randomized clinical trials (RCTs) published between December 2019 (the beginning of the pandemic) and August 2021 were included. Only phase III trials were included. Participants in the trials were adult patients (age ≥18 years) with COVID-19. Clinical therapeutic trials comparing pharmacological agents to any another intervention or no intervention (control) in patients with COVID-19 were included. COVID-19 diagnosis was considered as defined in the individual trials. Since many medications have been studied, and many trials were published during this time with conflicting results regarding efficacy, we chose to include only acceptable interventions according to recent guidelines [10]. We have included only trials evaluating medications recommended by National Institutes of Health (NIH) guidelines. In August 2021, the recommended drugs were steroids (preferably dexamethasone, but hydrocortisone or methylprednisolone as well), the anti-IL-6 receptor antibody tocilizumab, the antiviral agent remdesivir, and the monoclonal antibody casirivimab-imdevimab REGN-COV2 [10]. Search terms used for PubMed search were “COVID19 OR COVID-19 OR SARS-CoV-2 OR 2019-nCoV OR SARS2,” combined with the Cochrane filter for RCTs [11].

Two reviewers independently extracted data from the included trials and evaluated the quality of the methodologies (S.B., D.F.). If the two reviewers were not in agreement, a third evaluator (A.G.) extracted the data and the results were obtained by consensus. We assessed all possible sources of bias that are relevant, including allocation concealment, generation of the allocation sequence, blinding, incomplete outcome data reporting, and selective outcome reporting. We rated each domain as low risk of bias, unclear risk (lack of information or uncertainty over the potential for bias), or high risk of bias according to criteria specified in the Cochrane Handbook, version 5.1.0 [11].

Two reviewers independently extracted the following data from full-text RCTs: publication status and site, setting, inclusion, and exclusion criteria according to the cancer status with emphasis on criteria that may eventually lead to exclusion of cancer patients like exclusion according to laboratory tests (e.g., neutropenia or thrombocytopenia) or exclusion of certain medications (such as immunosuppressive therapy or chemotherapy). We extracted data regarding the primary endpoint of each trial. We aimed to extract data regarding outcomes (death, ICU admission, mechanical ventilation, etc.) for the subgroup of cancer patients.

The search yielded 541 potentially relevant RCTs, 22 of which were considered suitable [12-33]. A study flowchart according to PRISMA, showing the flow of trials included in the meta-analysis and reasons for exclusion, is shown in Figure 1. The study flowchart according to PRISMA, showing the flow of trials included in the systematic review and reasons for exclusion, is shown in Table 1, which presents the characteristics of the included trials.

Studies were qualitatively compared and summarized. Ninety-one percent (20/22) of trials included patients with severe COVID-19 disease. Fifteen trials included only severe patients, four trials included moderate-to-severe patients, and one trial included patient with any severity. The included trials evaluated the following interventions: nine trials randomized patients to steroid versus standard care (of them 4 trials dexamethasone, 2 trial hydrocortisone, 3 trials methylprednisolone), nine trials randomized patients to tocilizumab versus standard care, three trials randomized patients to remdesivir versus standard care, and one trial randomized REGN-COV2 versus standard care.

Regarding sequence generation, all the trials except one were considered low risk for bias. Regarding allocation concealment, results were variable. Nineteen were considered low risk for bias. The results of allocation concealment are depicted in Table 1. Forty-one percent of trials were double blind. All trials except one were considered as low risk for incomplete outcome reporting.

All trials included patients with solid cancer and hematological malignancies in the formal reported inclusion criteria. However, only two trials reported an accurate number of cancer patients [27, 29]. Hermine et al. [26] in their CORIMUNO-19 trial assessing tocilizumab for patients with moderate or severe pneumonia requiring oxygen included 9 patients with active solid malignancy of 131 total: 4/61 (7%) in the tocilizumab group versus 5/67 (8%) in the standard care group [27]. Veiga et al. [28] in their trial assessing tocilizumab for patients with severe or critical COVID-19 in nine hospitals in Brazil included 9/129 patients with active solid malignancy: 4/65 (6%) in the tocilizumab group versus 5/64 (8%) in the standard care group. In addition, there was 1 patient with a hematological malignancy in the tocilizumab group [29].

Ten trials excluded neutropenic patients, seven of which excluded patients with an absolute neutrophil count (ANC) <500 k/μL and three trials excluded patients with an ANC <1,000 k/μL [19, 25, 28, 30-33]. Seven trials excluded thrombocytopenic patients (platelets <50 k/μL).

Eleven trials excluded patients that treated with any immunosuppression treatment, most of which did not specify the exact medications [12, 14, 16, 18, 19, 25, 26, 28, 31-33].

None of the two trials that included cancer patients reported separate outcomes for this population.

We performed a systematic review to assess the representation of cancer patients in RCTs of NIH-approved interventions, for treatment of patients with COVID-19. We included 22 RCTs. Although none formally excluded cancer patients, only two trials reported how many cancer patients were included. None of the trials reported outcomes for cancer patients.

We included only trials evaluating medications recommended by NIH guidelines at the time of our systematic review [10]. This was done in order to assess the most accepted interventions. We collected our data until August 2021 and the recommended medications were steroids, tocilizumab, remdesivir, and REGN-COV2, although, as expected during a new pandemic, the guidelines are constantly changing. According to the recent living guidelines of WHO, the recommended medications are dexamethasone and tocilizumab. The evidence regarding remdesivir is currently inconclusive [34]. In addition, novel antiviral drugs, Paxlovid and molnupiravir, are now recommended for patients with high risk of progression to severe disease [35].

Observational cohort studies, conducted worldwide, including hospitalized patients with COVID-19 of variable severity, have demonstrated that cancer patients have inferior outcomes due to factors including the underlying malignancy, treatment-related immunosuppression, or increased comorbidities [3-6]. A large cohort of cancer patients with COVID-19 in the USA suggests a significant risk posed to patients with cancer infected with COVID-19, with a significant increase in mortality observed. Most susceptibility appeared to be in hematological or lung malignancies [36].

Despite a higher mortality rate shown in this population, it seems that cancer patients are underrepresented in clinical trials. As we showed, only two of the included RCTs reported on the actual number of cancer patients included. Moreover, in these two trails, they comprised only around 6–8% of the total study population. None of the RCTs reported outcomes of cancer patients, not from the point of view of efficacy nor of safety. On one hand, some of the recommended medications for COVID-19 are widely used in other known indications in cancer patients. Steroids, for example, play a significant role in the treatment of cancer. Another example is tocilizumab, which is indicated for the treatment of lymphoma patients treated with chimeric antigen receptor T (CAR-T). Cells that suffered CAR-T cell-induced cytokine release syndrome [37, 38]. On the other hand, some of these medications have life-threating adverse events that might be more problematic in cancer patients compared to the general population. For example, tocilizumab increases the risk for infections, including opportunistic infections, and risk for GI perforation [39].

Our systematic review is limited by the paucity of data regarding inclusion of hematological and solid malignancy patients in the trials, and the paucity of data regarding the proportion of patients included. The scarcity of outcome data limited conduction of meta-analysis. Nevertheless, the literature search was comprehensive and included all available trials.

Despite the surging prevalence of COVID-19, it is suggested that vaccination programs worldwide may have broken the link between infection and hospitalization and death. Given that cancer or its treatment may impact immunity, characterization of immune response to COVID-19 vaccines in cancer patients represents a priority.

As we showed the lack of representation of cancer patients in interventional trials for treatment of COVID-19, they were also virtually excluded from the pivotal randomized controlled vaccine studies. Nevertheless, there are data regarding efficacy or immune response to COVID-19 vaccines in this population from retrospective and prospective cohort studies.

In large cohort, Wu et al. [39] determined the association between SARS-CoV-2 vaccination and SARS-CoV-2 infections among a population of patients with cancer. A total of 184,485 patients met eligibility criteria, and 113,796 were vaccinated. Overall vaccine effectiveness in the cohort was 58% (95% CI: 39–72%). Patients who received chemotherapy within 3 months prior to the first vaccination dose were estimated to have a reduced vaccine effectiveness of 57% (95% CI: −23% to 90%) versus 76% (95% CI: 50–91%) for those who had not received systemic therapy for at least 6 months prior [40]. In a cross-sectional Israeli study, Gurion et al. [40] assessed the ability of lymphoma patients to generate a sufficient humoral response after two injections of BNT162b2 Pfizer vaccine. Positive serological responses were observed in only 51% of the 162 patients. The rate of seropositivity increased from 3% in patients vaccinated within 45 days from the last monoclonal antibodies’ administration to 80% in patients vaccinated >1 year after this therapy. The latter percentage was equal to that of patients never exposed to monoclonal antibodies [41]. Similar results were shown in chronic lymphocytic leukemia (CLL) patients. In a prospective study conducted in the framework of the European Research Initiative on CLL (ERIC) and including 167 CLL patients, response rates to the BNT162b2 Pfizer vaccine were 55.2% in treatment-naïve patients but only 16.0% in patients undergoing active treatment [42]. These findings were also confirmed in a multicenter study that enrolled 373 CLL patients from 9 Israeli hospitals. Serological response to the vaccine was 61% in treatment-naïve patients and between 23% and 24% in those treated with Bruton kinase and BCL2 inhibitor agents. Of note, the response to vaccine dropped to 5% in patients given an anti-CD20 antibody during the year that preceded vaccination [43].

Recently, the CAPTURE (NCT03226886) prospective cohort study assessing COVID-19 immunity in 585 patients with cancer following administration of two doses of BNT162b2 or AZD1222 vaccines, administered 12 weeks apart, was published [43]. Seroconversion rates after two doses were 85% and 59% in patients with solid and hematological malignancies, respectively. Patients with hematological malignancies were more likely to have undetectable neutralizing antibody titers and a lower titer compared with solid cancer patients. Notably, vaccine-induced T-cell responses were detected in 80% of patients and were comparable between vaccines or cancer types.

Thus, the results of humoral response to COVID-19 vaccination in cancer patients and specifically in hematologic cancers are less than optimal. Thus, in addition to vaccination, better treatment strategies are warranted. In addition, cancer patients classified as clinically extremely vulnerable should still be advised to take additional precautions of their own accord such as social distancing, etc.

Our systematic review shows that cancer patients are underrepresented in COVID-19 interventional therapeutic trials, and their outcomes are underreported. The fact that the reported efficacy of vaccines is less than expected reinforces the fact that future studies should attempt to focus on this special vulnerable population of cancer patients. Proactive strategies are needed to improve early identification of COVID-19 positivity in cancer patients, measures to reduce the probability of infection are needed, and therapeutic options should be sought.

An ethics statement is not applicable because this study is based exclusively on the published literature.

The authors have no conflicts of interest to declare.

None.

Shira Buchrits and Danielle Fredman extracted the relevant data from the articles. Anat Gafter-Gvili and Shira Buchrits wrote the manuscript. Kim Ben Tikva Kagan critically reviewed the manuscript.

All data generated or analyzed during this study are included in this article. Further enquiries can be directed to the corresponding author.