Clinical Factors Predict Deterioration Among Patients With Cancer, COVID-19

ECOG performance status, cancer type, and type of prior therapy received can predict risk for clinical worsening or death in patients with cancer who contract the coronavirus disease 2019.

Fabrice Barlesi, MD

ECOG performance status, cancer type, and type of prior therapy received can predict risk for clinical worsening or death in patients with cancer who contract the coronavirus disease 2019 (COVID-19), according to new findings from the Gustave Roussy Institute in Paris, France.

“ECOG performance statuses of more than 1 and hematological malignancies were the [most influential] predictors of clinical worsening,” Fabrice Barlesi, MD, PhD, medical director of Gustave Roussy, said during a presentation at the virtual American Association for Cancer Research Annual Meeting. Clinical worsening was defined as the need for oxygen support of or in excess of 61 minutes, or death of any cause.1

Data from 137 patients with cancer and COVID-19 who were treated at Gustave Roussy showed that an ECOG performance status greater than 1 was a predictor of clinical worsening in patients with the virus on both univariate (HR, 4.6; 95% CI, 2.2-10.0; P <.0001) and multivariate (HR, 3.9; 95% CI, 1.8-8.7; P = .008) analysis. Among the 136 patients with cancer and COVID-19 who were evaluated for ECOG performance status, 60% had a status of 0 to 1 and 40% had a status higher than 1. Performance status was unknown for 1 patient.

Additionally, patients with hematologic malignancies also had a higher risk for deterioration than those with solid tumors on univariate analysis (HR, 2.7; 95% CI, 1.3-5.5; P = .008). The majority (84%) of the patients treated at Gustave Roussy had solid tumors compared with 16% who were diagnosed with a blood cancer.

In contrast with ECOG performance status and disease classification, smoking status (HR, 0.8; 95% CI, 0.2-2.7; P = .72) and body mass index (HR, 1.0; 95% CI, 0.5-2.2; P = .87) “had no impact on clinical worsening,” according to Barlesi, a professor of medicine at Aix-Marseille University in Marseille, France.

Class Of Prior Therapy May Correlate With Clinical Deterioration

Results from a univariate analysis of individuals who received chemotherapy for their disease within the past 3 months indicated that these patients had a higher propensity for clinical worsening than those who had not been treated with chemotherapy (HR, 2.60; 95% CI, 1.32-5.13; P = .06). This trend persisted in the multivariate analysis (HR, 2.00; 95% CI, 0.96-4.22). Although prior chemotherapy correlated with a greater chance of clinical deterioration, treatment with immunotherapy or targeted agents in the past 3 months did not.

Of note, disease status influenced chemotherapy-receiving patients’ risk for clinical deterioration. “Patients treated with cytotoxic chemotherapy are at a high risk for clinical worsening,” said Barlesi, who added that this risk specifically applied to patients with disease that is either active/metastatic or localized/in remission.

Conversely, cytotoxic chemotherapy treatment only increased the risk of death in patients with active disease. “The risk of death remains only for patients treated with chemotherapy at an advanced stage of active disease, meaning that we may continue to treat patients with [localized disease] with cytotoxic chemotherapy in the adjuvant or neoadjuvant setting,” Barlesi said. “We have to pay attention to factors [like this] when deciding how to treat and manage patients with cancer and COVID-19.”

Most patients (59%) had active/metastatic disease. The remaining 41% of patients were in remission or had localized disease.

ECOG performance status, cancer type, and prior chemotherapy were all characteristics predictive of clinical worsening in patients with cancer and COVID-19. However, the only factor predictive of death in both univariate and multivariate analyses of overall survival was ECOG performance status greater than 1 (HR, 3.4; 95% CI, 1.2-9.8).

COVID-19 Engulfs Gustave Roussy

France’s first 3 COVID-19 cases were confirmed on January 24, 2020, two of which were diagnosed in Paris.2 The COVID-19 outbreak in this region “significantly affected” the Institute, which managed 7251 patients from March 14 to April 15.

Of these 7251 patients, 3616 were hospitalized. Gustave Roussy clinicians subsequently tested 1302 hospitalized patients for COVID-19 using reverse transcription polymerase chain reaction. “Twelve percent came back positive for COVID-19,” Barlesi said. A fraction of these patients (23%) were fully asymptomatic.

After excluding 19 patients who did not have cancer, Barlesi and fellow Gustave Roussy investigators identified a total of 137 patients and evaluated the treatment characteristics that contributed to clinical worsening.

Gustave Roussy’s patient outcomes “seem to be comparable to [France’s],” according to Barlesi, who noted that, at the April 20 data cut-off, 20 of the 137 patients had died due to COVID-19, resulting in a mortality rate of 14.6%. On April 22, France’s mortality was 17.9% and the Paris region’s rate was 18.3%.

Clinical worsening was observed in 24.8% patients, 11% of whom were admitted to intensive care. At the time of the analysis, 16.1% of patients were still hospitalized while 69.3% were either discharged or receiving outpatient treatment.

Methods used to manage SARS-CoV-2 infection, which causes COVID-19, included a dual hydroxychloroquine and azithromycin regimen (HCQ/AZI); lopinavir in combination with retronavir; an interleukin-6 receptor antibody; or steroids. These therapies were administered to 40 patients, 5 patients, 10 patients, and 13 patients, respectively.

The absence of a standard therapeutic approach for the management of COVID-19 renders patient participation in clinical trials “essential,” Barlesi said. The Gustave Roussy is currently recruiting patients for 8 different studies evaluating a variety of treatment interventions for COVID-19, such as the phase II ONCOVID trial (NCT0441207) of HCQ/AZI in approximately 1000 French patients with advanced cancer, a positive SARS-CoV-2 test, and COVID-19 symptoms.


  1. Barlesi F, Foulon S, Bayle A, et al. Outcome of cancer patients infected with COVID-19, including toxicity of cancer research. Presented at: 2020 virtual annual meeting of the American Association for Cancer Research; April 27-28, 2020.
  2. Stoecklin SB, Rolland S, Silue Y, et al. First cases of coronavirus disease 2019 (COVID-19) in France: surveillance, investigations, and control measures, January 2020. 2020;25(6):2000094. doi: 10.2807/1560-7917.ES.2020.25.6.2000094

<<< 2020 AACR Annual Meeting