Clinical characteristics of patients among non-cancer and cancer patients
We have obtained and analyzed 64 RT-PCR confirmed 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 (Table 1). The source of infection mostly occurred from hospital 41 (34.1%) out of 120 and 29 (45.31%) out of 64 for non-cancer and cancer cases respectively. When assessing the patient’s comorbid factors, diabetes ranks the top (31/120[25.83%] vs 13/64 [20.31%) for non-cancer cases. However, hypertensive condition was deadliest in cancer cases (19/64 [29.69%] vs 21/120 [17.5%]) than non-cancer cases. Typical Covid-19 related signs and symptoms fever (65/120 [54.17%] vs 29/64 [45.31%]) and cough (43/120 [35.83%] vs 26/64 40.63%]) was the highest in both cancer and non-cancer cases. Table 1 summarizes the patients clinical characteristics.
Table 1
Characteristics of Covid-19 positive non-cancer and cancer patient
Characteristics
|
Covid-19 patients without cancer (n = 120)
|
Covid-19 patient with cancer (n = 64)
|
p-values
|
Age (years, median/IQR
|
53.0/9.0
|
55/10.5
|
< 0.0001
|
Sex
|
Male
|
75/62.5%
|
48/75.0%
|
0.63
|
Female
|
45/37.5%
|
16/25%
|
0.59
|
Source of infection
|
Hospital
|
41/34.17%
|
29/45.31%
|
0.16
|
Family
|
14/11.67%
|
15/23.44%
|
0.77
|
Unknown
|
65/54.17%
|
20/31.25%
|
0.15
|
Underlying medical conditions
|
Chronic renal disease
|
9/7.5%
|
¾.69%
|
0.69
|
Cardiac disease
|
11/9.17%
|
15/23.44%
|
0.003
|
Diabetes
|
31/25.83%
|
13/20.31%
|
1.0
|
Hypertension
|
21/17.5%
|
19/29.69%
|
0.05
|
Asthma
|
5/4.17%
|
1/1.56%
|
1.0
|
Pulmonary disease
|
4/3.33%
|
1/1.56%
|
1.0
|
No known disorders
|
39/32.5%
|
12/18.75%
|
0.07
|
Hospitalization status
|
Discharged
|
109/90.33%
|
45/70.31%
|
0.56
|
Deceased
|
9/7.5%
|
19/29.69%
|
0.04
|
Duration in hospital stay (days)
|
13.6
|
23.8
|
0.0001
|
Typical Covid-19 signs and symptoms
|
Fever
|
65/54.17%
|
29/45.31%
|
0.03
|
Cough
|
43/35.83%
|
26/40.63%
|
0.02
|
Nausea
|
3/2.5%
|
6/9.34%
|
0.07
|
Fatigue
|
2/1.67%
|
¾.69%
|
0.08
|
Sore throat
|
7/5.83%
|
2/3.13%
|
1.0
|
Treatments during hospitalization
|
Antiviral treatment
|
49/40.83%
|
19/29.61%
|
0.39
|
Oxygen therapy
|
19/15.83%
|
23/35.94%
|
0.0001
|
Antibiotic treatment
|
46/38.33%
|
25/39.06%
|
0.04
|
Invasive mechanical ventilation
|
6/5.0%
|
13/20.31%
|
0.04
|
Cancer type
|
Breast cancer
|
N/A
|
16/25.0%
|
N/A
|
Ovary
|
N/A
|
5/7.81%
|
N/A
|
Lung/NSCLC
|
N/A
|
9/14.06%
|
N/A
|
Hematologic cancer
|
N/A
|
3/4.69%
|
N/A
|
Bladder
|
N/A
|
6/9.38%
|
N/A
|
Colon
|
N/A
|
7/10.94%
|
N/A
|
HNSCC
|
N/A
|
6/9.38%
|
N/A
|
Rectal
|
N/A
|
5/7.81%
|
N/A
|
Pancreas
|
N/A
|
3/4.69%
|
N/A
|
Esophageal
|
N/A
|
4/6.25%
|
N/A
|
Overall, intensive care unit (ICU) admission, disease severity and invasive ventilation support requirements was urgently needed for cancer patients compared to non-cancer patients (Fig. 1B) and the death was principally attributed to Covid-19 positive cancer patients. In the cancer cohort, the death rate was strikingly higher (29.69%) compared to non-cancer cases (7.5%], Fig. 1B, p < 0.05) given the fact of similar age distribution between cancer and non-cancer patients. The requirement of ICU admission was greater (20.31%, p = 0.01) in cancer patients than non-cancer cohort, higher rate of having severe condition (37.5%; p = 0.05), and needing and utilizing invasive ventilation support (IVS; 12.5%; p = 0.01; Fig. 1B).
As of August 15, 2020, 28 (10.32%) patients had died in the non-cancer (9/120; 7.5%) and cancer (19/64; 29.69%) cohort, all within 15 days with reports stating that the death was substantially attributed to Covid-19 in all patients. The patients who died from Covid-19 related complications in the cancer and 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 stratified 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 identified 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. Covid-19 positive non-cancer patients with cardiac disease (odd ratio: 0.33, 95% CI: 0.09–1.27), diabetes (odd ratio: 0.62, 95% CI 0.16–2.43), and hypertension (odd ratio: 0.82, 95% CI 0.28–2.42) remained the deadliest comorbidity factors among all non-cancer patients (Supplementary Fig. S1E). According to multivariate regression analysis, Covid-19 positive cancer patients had overall higher OR (OR, 2.59; 95% CI 1.47–5.26; p < 0.001; Supplementary Table S1a) for death, ICU admission (OR 4.22; 95% CI 3.17–6.26, p < 0.001; Supplementary Table S1b) and severe symptoms (OR 2.26; 95% CI 1.77–4.61, p < 0.01; Supplementary Table S1c) when considering with other underlying conditions.
Incubation and serial interval period analysis for non-cancer and cancer patients
Types of malignancies in the Covid-19 patients may greatly influence the patient’s outcomes. To expand our understanding of patient’s malignancies and symptom onset between cancer and non-cancer patients, we have investigated the incubation period and serial interval time. Our study composed of several cancer types where breast cancer was overrepresented (16/64, 25.0%), followed by lung (9/64,14.06%), colon (7/64, 10.94%) and head and neck cancer (6/64, 9.38%) (Fig. 2A). The overall death rate in cancer patients was 29.69%. The death rate among lung cancer patients was higher (7/64, 10.93%), while hematologic and ovarian cancer patients had the second highest death rate (3/64, 4.69%; 3/64, 4.69%). Breast, colon, and gastrointestinal cancer each had 2 deaths out of 64 (2.94%) (Supplementary table S2). The patients who had a high death rate also accompanied with severe and critical symptoms, longer ICU supports and invasive mechanical ventilation requirements (Supplementary Table S2).
Many recently published manuscripts reported that the mean incubation period for Covid-19 positive cases was between 5.6 to 6.7 days and the pooled parameter estimates the median period of 5.1 and 11.7 days 14. In the non-cancer cohort, we find that the median incubation period is 5.41 days with the gamma distribution: shape 4.84 (95% CI: 3.41–5.93) and scale 1.2 (95% CI 0.81–1.34). On the other hand, the mean incubation period is 5.81 (95% CI: 5.17–6.50). In the cancer cohort, the median incubation period is 4.10 days with the gamma distribution: shape is 1.89 (95% CI: 1.32–2.42) and scale 2.61 (95% CI: 1.47–3.79). The calculated mean is 4.99 (95% CI: 3.60–6.71) days. Summary of the results is presented in supplementary table s3 and fitted Weibull and log normal distributions is in supplementary table s4. Although our results are consistent with recently published data for non-cancer group, while the cancer patient’s incubation time was shorter than the non-cancer counterparts. These results suggest that the estimated mean incubation period for cancer patients is likely shorter than the non-cancer Covid-19 positive patients. Supplementary table s5 shows the recently published estimated mean incubation period and serial interval for Covid-19 positive non-cancer patients. A heatmap presentation of disease progression in Figs. 2B and 2C for 120 non-cancer and 64 cancer cases shows that Covid-19 related symptom was 3.48+/-2.17 (mean+/-sd) days after first possible exposure to virus and most cases were confirmed 2.43+/-1.47 (mean+/-sd) days after the symptom onset. The mean length of hospital stay for non-cancer cases were 13.6+/-7.27 (mean+/-sd), and 16.11+/-8.69 (mean+/-sd) before the individuals were either recovered, discharged or deceased (Figs. 2B, 2C). The mean serial interval of the fitted normal distribution is 5.09 (95% CI: 4.87–6.07) days for non-cancer patients, while it was 3.69 (95% CI:3.02–4.39) days for cancer cases (Fig. 3A, 3B, 3C; supplementary tables s3; s4, and s6; 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 confirmed 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 specific gene target quantified using two separate assay which specifically target SARS-CoV-2. We first 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 significant 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 transformed the Ct value into quantitative assessment of viral load. For this, Ct values were divided into three quartiles Q1, Q2 and Q3. Q1 represents the high viral load (Ct values less than 21.5; and viral load 5.30.1-5.778 log10 RNA copies/mL), Q2 medium viral load (Ct value between 21.5–27.3; 3.699-5.30 log10 RNA copies/mL) and low viral load (Ct values over 27.3; 2.67–3.69 log10 RNA copies/mL). The median Ct values for the SARS-CoV-2-specific gene target for cancer (MOLgen detection assay) was 21.5 (IQR-17.3-29.57) compared with a median Ct values of over 23.7 (IQR = 22.24–32.4) in non-cancer patients (Fig. 4A). Patients age has shown the significant positive correlation with high viral load (Fig. 4B, p < 0.0001; R2 = 0.50)), but there was no significant difference in viral load between male and female (Fig. 4C, p = 0.32).
Using this cut-off, compared with Covid-19 patients without cancer and low viral load, cancer patients and high viral load had higher observed severe patient’s condition (OR,1.42; 95% CI 1.17–3.003; p < 0.0001), higher death (OR, 1.62; 95% CI 1.27–3.25; p = 0.00016); ICU admission (OR 1.25; 95% CI 1.07–2.81, p = 0.00029) and higher chances of requiring IVS (OR, 1.10; 95% CI 0.91–2.38; p = 0.098; Fig. 4D; supplementary table s7). We then analyzed the possible occurrence of severity, death, ICU admission and IVS requirement in cancer and non-cancer patients’ group by cumulative survival analysis. We found that the severe condition occurred earlier in cancer patients where SARS-CoV-2 viral load is higher than low SARS-Cov-2 viral load in non-cancer patients (Figs. 4E; 4F). These results are consistent with results obtained for multi-variate regression results after adjusting for patients underlying medical conditions (supplementary table s1a; s1b and s1c).
High viral load is associated with severe condition and high mortality in certain type of cancer patients
Next, we sought to compare SARS-CoV-2 viral load and patient’s severity and mortality in several cancer type and non-cancer cohort. The mortality rate was 41.38% for high viral load; 23.81% in medium viral load and 14.29% in low viral load cancer patient, while the mortality was lower in non-cancer patients (22.22% for high viral load, 5.13% in medium viral load and 1.85% in low viral load patients; Table 2). We then compared viral load analysis between the different type of cancers and non-cancer patients using similar cutoff criteria above. As shown in Figs. 5A; 5B, lung, colon, hematologic, esophageal and breast cancer presented high SARS-CoV-2 viral load with higher risk of severe patients’ conditions, have relatively high death rate, ICU admission and comparatively higher chance of requiring IVS in the shortest possible time.
Table 2
Mortality rate of cancer and non-cancer patients based on SARS-CoV-2 viral load
Non-cancer patients COVID-19 positive
|
Viral load
|
% Mortality
|
Low (n = 54)
|
(n = 1; 1.85%)
|
Medium (n = 39)
|
(n = 2; 5.13%)
|
High (n = 27)
|
(n = 6; 22.22%)
|
Cancer patients with COVID-19 positive
|
Low (n = 14)
|
(n = 2; 14.29%)
|
Medium (n = 21)
|
(n = 5; 23.81%)
|
High (n = 29)
|
(n = 12; 41.38%)
|
Among the 64 Covid-19 positive cancer patients, patients have received different therapeutic regimens, such as, chemotherapy, surgery or radiotherapy treatments. Thirty-three (33/64; 5156%) had chemotherapy, 5 (5/64; 7.81%) had radiotherapy, 10 (10/64; 15.63%) had surgery and 16 (16/64; 25.0%) had surgery before these patients were tested positive for Covid-19 (supplementary table s8). Patients who received chemotherapy tended to show high rate of developing severe condition in patients with high SARS-CoV-2 viral load compared to low viral load (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.