ABSTRACT
We sought to determine parameters of the acute phase response, a feature of innate immunity activated by infectious noxae and cancer, deranged by Covid-19 and establish oncological indices' prognostic potential for patients with concomitant cancer and Covid-19. Between 27/02 and 23/06/2020, OnCovid retrospectively accrued 1,318 consecutive referrals of patients with cancer and Covid-19 aged 18 from the U.K., Spain, Italy, Belgium, and Germany. Patients with myeloma, leukemia, or insufficient data were excluded. The neutrophil-lymphocyte ratio (NLR), platelet-lymphocyte ratio (PLR), prognostic nutritional index (PNI), modified Glasgow prognostic score (mGPS), and prognostic index (PI) were evaluated for their prognostic potential, with the NLR, PLR, and PNI risk stratifications dichotomized around median values and the pre-established risk categorizations from literature utilized for the mGPS and PI. 1,071 eligible patients were randomly assorted into a training set (TS, n=529) and validation set (VS, n=542) matched for age (67.9±13.3 TS, 68.5±13.5 VS), presence of 1 comorbidity (52.1% TS, 49.8% VS), development of 1 Covid-19 complication (27% TS, 25.9% VS), and active malignancy at Covid-19 diagnosis (66.7% TS, 61.6% VS). Among all 1,071 patients, deceased patients tended to categorize into poor risk groups for the NLR, PNI, mGPS, and PI (P<0.0001) with a return to pre-Covid-19 diagnosis NLR, PNI, and mGPS categorizations following recovery (P<0.01). In the TS, higher mortality rates were associated with NLR>6 (44.6% vs 28%, P<0.0001), PNI<40 (46.6% vs 20.9%, P<0.0001), mGPS (50.6% for mGPS2 vs 30.4% and 11.4% for mGPS1 and 0, P<0.0001), and PI (50% for PI2 vs 40% for PI1 and 9.1% for PI0, P<0.0001). Findings were confirmed in the VS (P<0.001 for all comparisons). Patients in poor risk categories had shorter median overall survival [OS], (NLR>6 30 days 95%CI 1-63, PNI<40 23 days 95%CI 10-35, mGPS2 20 days 95%CI 8-32, PI2 23 days 95%CI 1-56) compared to patients in good risk categories, for whom median OS was not reached (P<0.001 for all comparisons). The PLR was not associated with survival. Analyses of survival in the VS confirmed the NLR (P<0.0001), PNI (P<0.0001), PI (P<0.01), and mGPS (P<0.001) as predictors of survival. In a multivariable Cox regression model including all inflammatory indices and pre-established prognostic factors for severe Covid-19 including sex, age, comorbid burden, malignancy status, and receipt of anti-cancer therapy at Covid-19 diagnosis, the PNI was the only factor to emerge with a significant hazard ratio [HR] in both TS and VS analysis (TS HR 1.97, 95%CI 1.19-3.26, P=0.008;VS HR 2.48, 95%CI 1.47- 4.20, P=0.001). We conclude that systemic inflammation drives mortality from Covid-19 through hypoalbuminemia and lymphocytopenia as measured by the PNI and propose the PNI as the OnCovid Inflammatory Score (OIS) in this context.
ABSTRACT
Background: There is uncertainty as to the contribution of cancer patients' features on severity and mortality from Covid-19 and little guidance as to the role of anti-cancer and anti-Covid-19 therapy in this population. Method(s): OnCovid is a retrospective observational study conducted across 19 European centers that recruited cancer patients aged >18 and diagnosed with Covid-19 between 26/02 and 01/04/2020. Uni- and multivariable regression models were used to evaluate predictors of Covid-19 severity and mortality. Result(s): We identified 890 patients from UK (n=218, 24%), Italy (n=343, 37%), Spain (n=323, 36%) and Germany (n=6, 1%). Most patients were male (n=503, 56%) had a diagnosis of solid malignancy (n=753, 84%) and 556 (62%) had active disease. Mean (+/-SD) patient age was 68+/-13 years, and 670 (75%) had >1 co-morbidity, most commonly hypertension (n=386, 43%). Commonest presenting symptoms were fever (n=569, 63%) and cough (n=448, 50%), beginning 6.3 (+/-9.5 SD) days before diagnosis. Most patients (n=565, 63%) had >1 complication from Covid-19, including respiratory failure (n=527, 59%) and acute respiratory distress syndrome (n=127, 22%). In total, 110 patients (14%) were escalated to high-dependency or intensive care. At time of analysis, 299 patients had died (33%). Multi-variate logistic regression identified male gender, age>65 (p<0.0001) presence of >2 comorbidities (p=0.001) active malignancy (p=0.07) as predictors of complicated Covid-19. Mortality was associated with active malignancy (p<0.0001), age>65 and co-morbid burden (p=0.002). Provision of chemotherapy, targeted therapy or immunotherapy was not associated with higher mortality. Exposure to anti-malarials alone (chloroquine/hydroxychloroquine, n=182, p<0.001) or in combination with anti-virals (n=195, p<0.001) or tocilizumab (n=51, p=0.004) was associated with improved mortality compared to patients who did not receive any of these therapies (n=446) independent of patients' gender, age, tumour stage and severity of Covid-19. Conclusion(s): This study highlights the clinical utility of demographic factors for individualized risk-stratification of patients and supports further research into emerging anti Covid-19 therapeutics in SARS-Cov-2 infected cancer patients. Clinical trial identification: NCT04393974. Legal entity responsible for the study: Imperial College London. Funding(s): Has not received any funding. Disclosure: D.J. Pinato: Speaker Bureau/Expert testimony: ViiV Healthcare;Advisory/Consultancy, Travel/Accommodation/Expenses: Bayer;Advisory/Consultancy, Speaker Bureau/Expert testimony, Research grant/Funding (institution), Travel/Accommodation/Expenses: BMS;Honoraria (self), Advisory/Consultancy: MiNa Therapeutics;Advisory/Consultancy: Eisai;Advisory/Consultancy, Speaker Bureau/Expert testimony, Travel/Accommodation/Expenses: Roche;Advisory/Consultancy: AstraZeneca;Research grant/Funding (institution): MSD. All other authors have declared no conflicts of interest.Copyright © 2020 European Society for Medical Oncology