SARS-CoV-2 viral load predicts the severity and mortality in patients with cancer


 Emerging evidence suggests that patients with cancer are at increased risk of detrimental Covid-19 outcome. The relationship between SARS-CoV-2 viral load and risk factors and outcome of SARS-CoV-2 positive cancer patients remains largely unexplored. We assessed the outcomes of Covid-19 infection in 64 cancer patients and 120 non-cancer and measured SARS-CoV-2 viral load from nasopharyngeal swab samples using cycle threshold (Ct) values who were admitted to two geographically distinct hospitals. We also assessed the incubation period and serial interval time differences between the non-cancer and cancer groups. Our results indicated that the overall mortality rate was higher among cancer patients with a high SARS-CoV-2 viral load. Covid-19 positive cancer patients with higher viral load are more prone to severe outcomes compared to non-cancer and low viral load patients. In addition, patients with lung and hematologic cancer have higher tendencies of severe events in proportion to high viral load. Higher attributable mortality and severity were directly proportional to high viral load particularly patients who are receiving anticancer treatment. Importantly, we found that the incubation period and serial interval time is fairly shorter in cancer patients compared with non-cancer cases. Our report suggests that high SARS-CoV-2 viral loads may play significant role in the overall mortality and severity of Covid-19 positive cancer patients and warranted further study to explore the disease pathogenesis and their use as prognostic tools.


Introduction
Individuals who are infected with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) are phenotypically diverse depending on the history of active malignancies. The Covid-19 symptoms ranging from mild to severe and may have an extreme outcome. Patients with cancer and Covid-19 have been identi ed as increased risk of mortality and morbidity 1 .
Recent reports suggest that age, male sex, smoking status and comorbidities (hypertension, cardiovascular disease and diabetes) have been widely identi ed as risk factors and severely impact on patients mortality in patients with cancer 2,3 . Among other comorbid factors, cancer was identi ed as a susceptible group and have been identi ed as an increased risk of infection with Covid-19 and increased risk of death and severe outcomes. These severe phenotypes require special attention with higher intensive care, rapidly deteriorating patient conditions and increased risk of death 4-6 with a clear contrast between the patients with and without cancer. Furthermore, the clinical phenotypes of patients with cancer and the effects of anticancer treatment greatly in uence the outcome of the patient's severity and survival.
It was reported that high SARS-CoV-2 viral load was independently associated with in-hospital mortality among Covid-19 positive population 7 . In addition, high SARS-CoV-2 viral load in patients with cancer have been reported recently 8,9 Although this report highlighted the relationship between viral loads and mortality of cancer and non-cancer patients, the impact of viral loads on outcomes of patients with cancer with a speci c cancer type and anti-cancer treatment require additional study to determine whether intensity of SARS-CoV-2 viral load may precisely predict the outcome of cancer patients.
We aimed to illustrate the clinical characteristics and outcome of patients with and without cancer and presented evidence the detrimental effects of SARS-CoV-2 viral loads among patients with and without cancer. Here we show that SARS-CoV-2 high viral loads adversely affect the certain type of cancer patients particularly who are under chemotherapeutic treatment which may provide predictive tool for cancer patients mortality.

Study design and patients
In this study, we collected data from a primary (Bangladesh) and a tertiary (Riyadh) health care centers among patients with Covid-19 who have been diagnosed with cancer recently or in the past. All cancer patients enrolled in this study were con rmed with Covid-19 who were admitted to the hospitals from June 30, 2020 to August 7, 2020. A control cohort without cancer with con rmed Covid-19 infection who were admitted to same hospitals during the same time period were collected. The cancer cohort comprised 64 cases, while the non-cancer cohort comprised 120 cases. We excluded patients who displayed radiological or clinical diagnosis of Covid-19, but without a positive RT-PCR results. For cancer cohort, patients with active cancer were de ned as those undergoing anticancer treatment with curative, radical, adjuvant or neoadjuvant therapy or treated in the last 12 months with radiotherapy, surgery, chemotherapy. Both datasets contained mainly information on exposure time, time of symptoms onset, as such all cancer and non-cancer cases, by de nition are symptomatic. Four clinical outcomes were monitored up to August 30, 2020. We have classi ed disease severity in 4 main categories. (a) "mild cases" if the patients do not show any serious symptoms described in moderate, severe and deceased categories but positive nasopharyngeal swab RT-PCR test and did not require serious medical intervention; (b) "moderate cases" if the patients show fever, some respiratory symptoms and any evidence of pneumonia by radiography; (c) "severe cases" de ned as cases experiencing any of the following symptoms: breathing rate 30 or higher breaths/minute, the oxygen saturation level below 93%, one or multiple organ failures, requiring intensive care unit (ICU) and invasive mechanical ventilation support (IVS), a ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen (PaO2:FiO2) of less than 300 mm Hg, or in ltrates in more than 50% of the lung eld within 24 to 48 hours; and (d) "deceased cases" patients admitted to the hospitals with Covid-19 related symptoms and died during their hospital stay. This study was considered exempt from the requirements of institutional review board (IRB) approval and was approved by the central ethics committee of Bangladesh Medical Research Council (study #2021-2023/62 (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20) and King Faisal Specialist Hospital and Research Centre (RAC # 2200031).
SARS-CoV-2 Viral RNA load measurements Nasopharyngeal swab samples were collected from cancer and non-cancer patients and samples were stored in room temperature and processed within reasonable time of collection. Nucleic acid was extracted from samples using MOLgen-SARS-CoV-2 and Sansure Biotech SARS-CoV-2 assay protocol according to the manufacturer's instructions. Extracted RNA was eluted using elusion buffer. Levels of SARS-CoV-2 viral load was determined using US CDC real-time RT-PCR primer and probe set for 2019-nCoV N1 and N2. Ct values obtained from the N1 primers set were converted to RNA copies/mL using N1 quantitative PCR (qPCR) in 10-fold dilution standard curve as described previously 10,11 . Values were log 10 transformed. We converted Ct values into quantitative assessments of viral load-high, Ct value < 21, medium, Ct value 21-26, and low, Ct value >26. These cutoff values (high, medium and low) were used to determine the risk of severity and death of all patients used in this study.

Data collection and procedures
For this study, we collected data from patients by our data collection team with the assistance participating hospitals registry personnel. Patients demographics, cancer details, cancer treatment information, age, sex, number of comorbidities which require active treatment, surgical history and diagnosis of Covid-19, cancer status, treatment during hospitalization were collected.

Outcomes
We considered and analyzed four clinical outcome/endpoints: severe illness requiring admission to the hospital, death, admission to ICU, needing invasive mechanical ventilation support (IVS).

Incubation period analysis
We recorded daily mortality and hospitalization incidence and plotted with cumulative numbers of con rmed and discharged cases based on symptom onset date for non-cancer and cancer cohorts. We excluded cases that do not have de nitive symptom onset date. We generated a source of disease plot based on the best information available during the study period. For incubation period analysis, we used direct estimation methods 12 based on the earliest and latest possible exposure times and reported symptom onset times. The parametric distributions (gamma distribution, Weibull distribution, and lognormal distribution) of incubation period was estimated using interval censoring using maximum likelihood of a time falling in a de ned window. We used interval censoring because it is not possible to know the exact time of exposure.

Serial interval analysis
We estimated serial intervals in a direct method as the difference between the symptom onset dates. We excluded asymptomatic cases as well those cases that do not have exact symptom onset dates. We tted the normal, log normal, gamma and Weibull distributions using R packages " tdstrplus" by maximum-likelihood of a time falling in a de ned window. We computed each t distribution using Akaike information criteria (AIC) scores and calculated con dence intervals for parameters from 1000 bootstrap replicates 13 .

Statistical analysis and data visualization
All statistical analyses were performed using R (R Development Core Team, 2013). We assessed and reported clinical outcomes of Covid-19 positive non-cancer and cancer patients whether patients died or discharged and the effects of anti-cancer therapy on patients underlying conditions. We calculated the percentages of patients in each category for the categorical data. The Wilcoxon rank sum test was used for continuous data, and two-sided Fisher's exact test was used to compare categorical data for different categories of data without multi-test. Multivariate logistic regression analysis was used to estimate the odd ratio (OR) and 95% con dence interval (CI) of each factor of interest with confounders/outcomes. The OR was adjusted to chronic renal disease, cardiac disease, diabetes, hypertension, asthma and pulmonary disease. A two-sided P values < 0.05 was used to indicate the statistical signi cance. We constructed a multivariate logistic analysis model to identify variables that were independently associated with high and low viral load. Correlation analysis was performed using Spearman-rank based testing.

Clinical characteristics of patients among non-cancer and cancer patients
We have obtained and analyzed 64 RT-PCR con rmed COVID-19 positive cancer patients from June 30, 2020, to August 7, 2020 from a primary care hospital in Bangladesh and a tertiary care hospital King Faisal Specialist Hospital and Research Centre, Riyadh. In addition, we have collected 120 COVID-19 positive patients without the history of cancer from the same hospitals. The latter group was used as a control group to compare the parameters between cancer and non-cancer patients. The median age of cancer patients was 55 (inter quartile range [IQR)-10.5], and 53 (IQR 9.0) for non-cancer patients (p < 0.0001; Table 1). In our patient population age distribution range for cancer and non-cancer were close. Figure 1A shows the age distribution of patients in non-cancer and cancer patients. In total, the two cohort comprises with 123 males and 61 females' cases and male female ratio is 2:1 (  non-cancer patients were mostly older age group. The age distribution of death rate in cancer patients was between 40-59 years, while for non-cancer patients it was between 40-69 years of age ( Supplementary Fig.S1A). We also strati ed the severe condition of all cancer and non-cancer patients by age group. In general, the requirement of ICU support for cancer patient was higher in the older age group (40-69 years), but in the non-cancer group only patients age between 60-79 needed for ICU support requirement ( Supplementary Fig. S1B). The symptoms of severity in cancer and non-cancer patients increased with the increase of age ( Supplementary Fig. S1C), and ventilation support was proportionally higher in cancer patients age group between 40-69 years ( Supplementary Fig. S1D). In contrast, no invasive ventilation support was needed for non-cancer patients at any given age group ( Supplementary  Fig. S1D).
The overall pooled odd ratio (OR) of all identi ed comorbidity for non-cancer patient was 0.45 (95% CI: 0.240.86, Supplementary Fig. 1E). A forest plot of the potential underlying conditions is shown in Supplementary Figure S1E.  Supplementary Fig. S2).

SARS-CoV-2 viral load is higher in cancer patients than non-cancer patients
We report SARS-CoV-2 viral load analysis from two commercial viral detection assay kits (details in Materials and methods section). We analyzed the impact of SARS-CoV-2 viral load for all 64 cancer patients as well as 120 non-cancer patients with con rmed Covid-19 diagnosis. All patients were tested using nasopharyngeal swab. To investigate the impact of SARS-CoV-2 viral load, we have obtained the Ct values for SARS-CoV-2 speci c gene target quanti ed using two separate assay which speci cally target SARS-CoV-2. We rst compared the results with that of Roche cobas SARS-CoV-2 ORF-1ab and E gene. Results show that MOLgen-SARS-CoV-2 and Sansure Biotech SARS-CoV-2 N gene is highly correlated with that of the Roche cobas ORF-1ab and E genes and no signi cant Ct value variations were found among the two detection kits ( Supplementary Fig. S3). The only differences between the two assay kits was data generated from MOLgen SARS-CoV-2 detection was 2 cycle higher than the Ct values obtained from Sansure SARS-CoV-2 detection kit for N2 gene target for the cobas Ct values ORF-1ab target.
We    Fig. 5C; 5D). However, radiotherapy and surgery had medium to no effects in developing severs condition with lower low viral load. When mortality/death was assessed, chemotherapy, surgery and radiotherapy treated patients having high viral load and lead to higher rate of death, higher chances of ICU admission and higher use of IVS ( Fig. 5C; 5D). Additionally, the relationship between patients who received chemotherapy and underlying conditions and disease severity as well as disease outcome for Covid-19 positive cases shown in Fig. 5E and 5F.

Discussion
Our study demonstrated that the differences in SARS-CoV-2 viral load in Covid-19 positive cancer and non-cancer patients may play a role in the prediction of mortality and the extent of disease severity. SARS-CoV-2 viral load is signi cantly higher in cancer patients with increasing disease severity and mortality compared to non-cancer patients. Several phenomenal observations have emerged from our study. First, cancer patients with active Covid-19 infection showed shorter incubation period and serial interval time when compared with non-cancer patients. Secondly, patients with cancer infected with Covid-19 and high viral load is more likely to experience severe and deleterious outcomes compared to patients with non-cancer and low viral load. Thirdly, lung and breast cancer patients with high viral load demonstrated higher gravity of severe events, i.e death, ICU support, invasive ventilation requirement compared with non-cancer and low viral load individuals. Lastly, cancer patients who are under active anti-cancer treatment or have been previously treated with anticancer agents, particularly chemotherapeutic treatment showed higher death rate and higher chances of experiencing critical symptoms due to high SARS-CoV-2 viral load. These results may be useful for the consideration of prognostic tools to monitor and stratify the patients for delivering relevant treatment.
Since the emergence of Covid-19, the virus has rapidly spread all over the world and many countries are grappling for the search of epidemiological characteristics and control of the transmission of the virus, including its assessment of overall outcome and its impact in the society. An astonishing feature of Covid-19 pandemic is that elderly male reported severe disease and higher mortality than females [15][16][17][18][19] . Moreover, age, smoking status, and comorbidities, such as, hypertension and cardiovascular disease are the risk factors for severe disease and mortality among patients with cancer and non-cancer Covid-19 patients 8,20 . Since the start of the global pandemic of Covid-19, patients with cancer in most countries became the center point of concern due the highly vulnerability, increased risk of contracting Covid-19, severe outcomes and comparatively higher rate of requiring intensive care and increased risk of death 6,21 . Given the similar epidemiological features between cancer and non-cancer Covid-19 patients, it is highly likely that cancer patients endow some additional features different from non-cancer individuals.
There are considerably wide gaps remains in our understanding of SARS-CoV-2 pathogenesis, including whether levels of viral load and disease severity in patients with cancer. To ascertain these additional attributes, this study was conducted to nd out the possible effects of SARS-CoV-2 viral load in cancer patients and possible outcomes. Our study evaluating the patient's severe condition and death not only con rm the association of higher-level viral load with these factors, but these factors have an additive effect which may contribute to the increased risk of mortality and severe condition in patients with multiple risk factors.
Perhaps the most important observation in our study is that, Covid-19 patients with cancer and high viral load had signi cantly more severe outcomes, higher rates and longer stay in the hospitals and ICU and invasive ventilation supports compared to low viral load and non-cancer patients, con rming fraction of Finally, we analyzed the serial interval and incubation period differences between non-cancer and cancer cohorts, which are key parameters for transmission modeling and assists governments and policy makers to respond to the pandemic. These two parameters in uence the disease incidence and prevalence of transmission 12 . In the non-cancer and cancer cohort, we estimated the serial interval, and obtained a shorter serial interval (3.9 days) for cancer patients' cohort, while the non-cancer patients serial interval was 5.09 days 12 . Furthermore, the estimated serial interval is shorter than the incubation period in cancer cohorts suggesting that the pre-symptomatic transmission for cancer patients. The striking observation was that the source of Covid-19 infection in most cancer patients was from hospital facilities probably due to the frequent treatment visits during the pandemic. This study thus suggests that the hospital related infection could be avoided by delaying or cancelling cancer treatment if possible.
As usual our study carries several limitations. Although our study cohort for cancer is comparatively smaller than recently published studies and only symptomatic cases were included who seeking help from health care providers, but larger cohort as well as asymptomatic cases, and hidden cases who did not report for assistance could further extends our understandings of actual characteristics of cancer patients' treatment, and Covid-19 infection. Furthermore, due to the government pandemic restrictions it was not possible to collect all the available patient's data and records from the health care providers which potentially impeded our analysis. With some larger number of patient's analysis would be able pinpoint more unanswered questions. Despite all these limitations, our study is unique when compared with the larger studies, representative of small numbers of cancer patients among hundreds of unidenti ed cases. Future work with diverse cancer type and larger patients' cohort and long-term followups will de ne the speci c risk of Covid-19 on outcomes in much greater gravity in patients with cancer.

Declarations
Ethics declaration

Con ict of interest
The authors declared no con ict of interest