SARS-CoV-2-RNA viremia is associated to hypercytokinemia and critical illness in COVID-19

Rationale: whether systemic dissemination of SARS-CoV-2 has any impact on COVID-19 severity and also on the immunological alterations observed in this disease is largely unknown. Objectives: We determined the association of plasma SARS-CoV-2 RNA with clinical severity, laboratory findings and immunological parameters in a cohort of 250 patients with confirmed COVID-19 infection. Methods: Three groups of patients were studied: 50 outpatients, 100 hospitalised ward patients, and 100 critically ill. The association between plasma SARS-CoV-2 RNA and severity was evaluated using multivariate ordinal logistic regression analysis and Generalized Linear Model (GLM) analysis with a binomial distribution. The association between plasma SARS-CoV-2 RNA and laboratory parameters was evaluated using multivariate GLM with a gamma distribution. Measurements and Main Results: The prevalence of SARS-CoV-2-RNA viremia increased in parallel with severity of infection (22% in outpatients, 36 % in those hospitalised in wards, and 82% in those at the ICU). In hospitalised patients, the presence of SARS-CoV-2-RNA viremia was independently associated to critical illness: (adjusted OR= 8.30 [CI95%=4.21 - 16.34], p < 0.001). SARS-CoV-2-RNA viremia was an independent predictor of higher levels of ferritin, LDH and cytokines (involving CXCL10, CCL-2, IL-15, IL-10, IL-1ra and GCS-F), and lower of lymphocytes, monocytes and platelets counts Conclusions: SARS-CoV-2-RNA viremia is a robust marker of critical illness in COVID-19. Our findings support that hypercytokinemia in COVID-19 is a reactive event in response to the dissemination of viral material at the systemic level.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 31, 2020. . https://doi.org/10.1101/2020.08. 25 This article has an online data supplement, which is accessible from this issue's table of content online at www.atsjournals.org All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Introduction
With well over 17 million cases and 715,013 deaths globally, Coronavirus disease 2019 (COVID-19) has become the top economic and health priority worldwide. 1 As the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic continues to emerge in low income countries and resource-poor settings, tools for the early identification and management of severe COVID-19 cases are of paramount importance. It is unclear how many affected individuals need hospitalization. Amongst hospitalized patients, around 10-20% are admitted to the intensive care unit (ICU), 3-10% require intubation and 2-5% die. 2 SARS-CoV-2 RNA detection in nasopharyngeal swabs is the most common diagnostic test for SARS-CoV-2 infection; however viral RNA can be found in sputum, lung samples, peripheral blood, serum, stool samples, and to a limited extent urine. 3 4 5 While the lungs are most often affected, severe COVID-19 can also induce inflammatory cell infiltration, haemorrhage, degeneration or necrosis in other organs (spleen, lymph nodes, kidney, liver, central nervous system). 6 Whether systemic spreading of the virus or viral components has any role in the pathogenesis of the "sepsis-like" failure in different organs 6 or in immunological dysregulation 7 observed in severe COVID-19 is currently unknown.
In this study, by using a integrative approach, we evaluated first the association between the presence of SARS-CoV-2 RNA in peripheral blood plasma from COVID-19 patients and severity. We next studied the impact of SARS-CoV-2 RNA viremia on a number of biological parameters denoting tissue damage and immunological dysregulation in this disease.
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. This group neither required critical care not died during hospitalization (wards group, n=100). The third group corresponded to patients admitted to the ICU (n=100). Patient`s recruited by participating hospital are detailed in Table E1 in the online data supplement.

Blood samples:
Plasma from blood collected in EDTA tubes samples was obtained from the three groups of patients at a median collection day of 7, 8 and 10 respectively, and also. from 20 blood donors (10 men and 10 women). Ethical aspects: The study was approved by the Committee for Ethical Research of the coordinating institution, "Comite de Etica de la Investigacion con Medicamentos del Area de Salud de Salamanca", code PI 2020 03 452. Informed consent was obtained orally when All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 31, 2020. . https://doi.org/10.1101/2020.08.25.20154252 doi: medRxiv preprint clinically possible. In the remaining cases, the informed consent waiver was authorized by the Ethics committee. providing an odds ratio (OR). Generalized Linear Models (GLM) with binomial distribution were used when the outcome variable was dichotomous (outpatients vs. wards, and wards vs. ICU), also providing an odds ratio (OR). GLM with a gamma distribution (log-link) was used when the outcome variable was continuous, providing arithmetic mean ratios (AMR). In all cases, the analysis was performed first without adjustment and was later adjusted for the most relevant covariates in our study. P-values were corrected using the false discovery rate (FDR) with the Benjamini and Hochberg (q-values) procedure. (Table 1): Patients diagnosed with SARS-CoV-2 infection based on a positive nasopharyngeal test and requiring hospitalization (either general ward or ICU) were older (median 64 years of age for ward and 66 years for ICU) than those patients discharged to their home from the ER (median age 48 years of age). All rights reserved. No reuse allowed without permission.

Clinical characteristics of the patients
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 31, 2020. . https://doi.org/10.1101/2020.08.25.20154252 doi: medRxiv preprint There were no significant differences regarding age between ward and ICU hospitalised patients. Critically ill patients (ICU admitted) were more frequently male than those in the other groups. Comorbidities of obesity, hypertension, dyslipidemia and type II diabetes were more commonly found in patients requiring hospitalization, with no significant differences found in the comorbidities profile between critically ill and non-critically ill hospitalized patients. ICU patients showed significantly lower levels of O 2 saturation at the time of admission to the ICU compared to other patients admitted to the ER or the ward.
100% of ICU patients presented with pulmonary infiltrates of whom 93 % also had bilateral pneumonia, these findings were significantly higher than the incidence of pulmonary infiltrates and bilateral pneumonia found in the other two groups. Glucose levels were higher in the group of critically ill patients, who also showed higher values of INR, D-dimers, LDH, GPT, ferritin, C-reactive protein and lower haematocrit. ICU patients showed pronounced lymphopenia and lower monocyte counts; however, neutrophil counts were increased. ICU patients more frequently received experimental treatments during their hospitalization period, including hydroxicloroquine, corticoids, remdesivir, tocilizumab, lopinavir/ritonavir or beta-interferon. ICU patients stayed longer in the hospital, with 48 % fatalities reported in this group. The number of missing values for the variables registered in this study are reported in Table E2 in the online data supplement. compared to ward patients (36%) and outpatients (22%) (p<0.001). No statistical differences were found in plasma viral RNA between the outpatients and the patients in the All rights reserved. No reuse allowed without permission.

Prevalence of SARS-CoV-2-RNA viremia and specific SARS-Cov
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 31, 2020. . https://doi.org/10.1101/2020.08.25.20154252 doi: medRxiv preprint ward (p = 0.081). Critically ill patients also had a higher frequency of specific SARS-CoV-2 IgG responses than the other groups (70% in ICU compared to 52% and 49% in the outpatients and ward groups, p < 0.05, table 1 and figure 1). No significant differences were found between the group of outpatients and those admitted to the ward.  Table E3 in the online data supplement). When we compared outpatients with admitted ward patients, multivariate GLM analysis showed that viremia was not significantly associated with either group (Table 2). In contrast, when the group of ward patients was compared to ICU patients, multivariate GLM analysis showed that viremia was strongly associated with patient severity requiring critical care [OR = 8.3, p < 0.001, (CI 95% = 4.21-16.34)] ( Table   2). In the patients admitted to the ICU, no significant difference in the presence of SARS-CoV-2-RNA was found between survivors and non survivors: 42 out of 48 of non survivors had viremia (87.5%), while 40 out of 52 survivors (76.9%) had viremia, p = 0.169.

Impact of SARS-CoV-2-RNA viremia on laboratory and immunological parameters:
Multivariate GLM analysis showed that SARS-CoV-2-RNA viremia was an independent predictor of higher levels of ferritin, LDH and higher levels of chemokines (CXCL10, CCL-2), cytokines (IL-15, IL-10, IL-1ra) and GCS-F ( Figure 2 and Table E4 in the online data supplement). In contrast, viremia predicted lower lymphocytes, monocytes and platelets counts and lower concentration of IL-4 in plasma ( Figure 2 and Table E4 in the online data supplement). Patients requiring hospitalization at the ward showed significantly increased levels of IL-10, CXCL0, IL-1ra, IL-6 and TNF-α compared to All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Discussion
Here we report that SARS-CoV-2 viral RNA is detected in the plasma of the vast majority of those COVID-19 patients admitted to the ICU (82%). In hospitalized COVID-19 patients, presence of SARS-CoV-2-RNA viremia translates into an 8-fold increase in the risk of presenting critical illness, independently of age, sex and major comorbidities.
Importantly, these findings suggest that detection of viral RNA in plasma may serve as a simple test to identify those patients needing critical care.
Whether the finding of viral RNA is "true viremia" with the live virus found in the plasma and peripheral blood is unknown; however, the SARS-CoV-2 virus has been reported to be difficult to culture from blood. 4 Alternatively, but not mutually exclusive, the presence of viral RNA in the blood may represent a substantial spill over event from virally infected tissue. Importantly, SARS-CoV-2-RNA viremia was associated with higher levels of plasma LDH (a marker of necrosis and cellular injury) and lower O2 saturation, which supports viral involvement in the genesis of tissue damage and respiratory failure in patients with severe COVID-19.
Interestingly, SARS-CoV-2-RNA viremia maybe a contributing factor in the development of hypercytokinemia in patients with severe COVID-19, since the presence of viral RNA in plasma predicted higher levels of CXCL10, CCL2, IL-15, IL-10, IL-1ra and G-CSF.
Recognition of viral RNA by endosomal receptors such as TLR7 in human plasmacytoid dendritic cells and B cells, or TLR8 in myeloid cells, activate the intracellular signalling pathways enhancing cytokine production. 8 In fact, it has recently been demonstrated that All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 31, 2020. . https://doi.org/10.1101/2020.08.25.20154252 doi: medRxiv preprint SARS-CoV-2 genome has more single-stranded RNA fragments that could be recognized by TLR7/8 than the SARS-CoV-1 genome, which suggest the potential of SARS-CoV-2 to induce hyperactivation of innate immunity. 9 From a cohort of SARS-CoV-1 patients in 2003, we previously demonstrated that severe SARS patients had increased levels of CXCL10 and CCL2 in serum during the early onset of symptoms. 10 CXCL10 is a potent chemoattractant for activated Th1 lymphocytes and natural killer cells and is thought to play a role in the temporal development of innate and adaptive immunity. 11 Signalling via the CXCL10 cognate receptor, CXCR3, mediates immunopathology during other highly pathogenic respiratory virus infections such as CCL2 is one of the key chemokines that regulate migration and infiltration of monocytes/macrophages. 12 Interestingly, during SARS-CoV-1 infection, the presence of high levels of CXCL10 and CCL2 was coincident with the presence of lymphopenia 13 , as occurs in patients with severe COVID-19. 14 15 In SARS-CoV-1 severe disease, selfsustaining expression of proinflammatory chemokines has been suggested to represent a compensatory mechanism for an ineffective adaptive immune response to clear the virus. 10 IL-10 is a major immunomodulatory cytokine inducing immunosuppression. Zhao Y et al reported IL-10 (along with IL-1ra, another immunomodulatory cytokine capable of suppressing the IL-1 signalling pathway) to be associated with the severity of COVID-19. 16 In our study, both cytokines were elevated in the plasma of patients requiring hospitalization, with higher levels of IL-10 in those patients admitted to the ICU. Whether elevation of IL-10 and IL-1RA represents a mechanism of viral evasion or an attempt of All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 31, 2020. . https://doi.org/10.1101/2020.08.25.20154252 doi: medRxiv preprint 1 0 the immune system to control an exuberant inflammatory response remains to be elucidated. 17 18 The association between SARS-CoV-2-RNA viremia and IL-15 is also especially intriguing, since IL-15 was the cytokine better differentiating clinical ward patients from The presence of hyperferritinaemia has been highlighted in COVID-19, since it is, along with hypercytokinemia and high LDH, a marker of macrophage activation syndrome (MAS) . 28 29 Another signature of MAS is the decrease in platelet concentrations in blood 30 . Even though platelet counts did not show significant variation across the three severity groups (Table 1), those patients with SARS-CoV-2-RNA viremia showed significantly lower platelet counts than the other patients (198,000 vs 230,000 cells/mm3, p = 0.003). In All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
Severe COVID-19 is characterised by the frequent finding of lymphopenia 15 . Since in our study SARS-CoV-2-RNA viremia was associated with low lymphocyte and monocyte counts, it is intriguing to posit the question as to whether lymphopenia and the decrease in the monocyte counts are a result of direct cytopathic events or emigration due to localized chemokine and IL-15 expression, or a combination of these processes. Cytopenias in blood are also a shared feature between severe COVID-19 and MAS.
Our findings revealing the association between SARS-Cov-2 RNA viremia, hypercytokinemia, higher ferritin and LDH levels and lower lymphocyte, monocyte and platelets counts suggest that systemic dissemination of the virus or viral material could be the driver of the MAS-like syndrome observed in severe COVID-19.
In our work, most ICU patients with SARS-CoV-2-RNA viremia had already developed a specific IgG response against the virus (70.7 %), which would support the notion that continued viral replication is a persistent event in the course of the of antibody responses.

Additional information on cytotoxic T cells and the role of NK cells in controlling
COVID-19 disease will help link the cellular immune events with viral replication and control in disease progression.
A limitation of our study is that we did not evaluate viral load. In consequence we could not assess differences in the amount of viral RNA between severity groups, and also between survivors and non survivors. In addition, neither the potential infectivity of plasma nor the presence of live virus in plasma was assessed. Follow up studies should investigate the presence of viral RNA in specific blood cells as well.
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 31, 2020. . https://doi.org/10.1101/2020.08. 25.20154252 doi: medRxiv preprint In conclusion, the high prevalence of SARS-CoV-2-RNA viremia in critically ill patients suggests that these patients are unable to control SARS-Cov-2 replication in tissues or blood cells. Viremia is associated with hypercytokinemia and other signatures typically Ackowledgements: we thank SEIMC-GESIDA Foundation for the scientific sponsoring of this project. We thank also the "Centro de Hemoterapia y Hemodonación de Castilla y León, CHEMCYL", which provided the plasma simples used in the healthy control group.
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Conflicts of interests: the authors declare no conflicts of interests regarding this submission
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 31, 2020. All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 31, 2020.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 31, 2020.  Table 2: Multivariate generalized linear model with binomial distribution to assess the association between viremia and hospitalization at the wards in the comparison (outpatients vs wards) (left) and the association between viremia and hospitalization at the ICU in the comparison (wards vs ICU) (right).

Figure legends
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted August 31, 2020. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.