Predicting Disease Severity and Outcome in COVID-19 Patients: A Review of Multiple Biomarkers Open Access

Databases including PubMed, medRxiv, and bioRxiv were searched for peer-reviewed and non–peer-reviewed papers. Search terminology included COVID-19 search terms (COVID-19, SARS-CoV-2, 2019-ncov) and laboratory parameters: white blood cell count (WBC), lymphocyte count, platelet/thrombocyte count, coagulation, prothrombin time (PT), D-dimer, fibrinogen, LDH, ferritin, CRP, SAA, procalcitonin, and interleukin 6 (IL-6).

A total of 72 papers, including 54 peer-reviewed papers and 18 non–peer-reviewed preprints, with data including hematologic, coagulation, biochemical, or inflammatory parameters and their role and predictive value in mild, severe, and critically ill patients were reviewed. Within the total reviewed papers, 43 cohort studies and 11 systematic reviews (including meta-analysis) were extracted. Where applicable, statistical significance was represented in the literature by minimum P values of P ≤ .05. Relevant data were collected by 2 independent reviewers.

Most of the studies analyzed were performed with COVID-19–affected patients in China, whereas a minority were from Italy, the United Kingdom, Iran, Brazil, and the United States. In some studies, the clinical staging of patients was generally categorized into 4 groups based on treatment guidelines issued by the Chinese National Health Committee.²¹  Patients with uncomplicated illness demonstrated mild clinical symptoms without radiologic findings of pneumonia, whereas those with mild disease had fever, respiratory symptoms, and findings consistent with pneumonia on chest imaging studies. Severe disease was defined clinically as respiratory distress with respiratory rate ≥ 30/min, pulse oximeter oxygen saturation ≤93% at rest, and oxygenation index (artery partial pressure of oxygen/inspired oxygen fraction, PaO2/FiO2) ≤300 mm Hg. Critically ill disease was defined as respiratory failure requiring mechanical ventilation, shock, or other organ failure requiring intensive care unit (ICU) monitoring.

Zhang et al⁶  analyzed the clinical characteristics of 140 patients with an overall median age of 57 years. No gender predominance was observed between severe and nonsevere illness. Initial laboratory results on admission revealed normal WBC with decreased lymphocyte and eosinophil counts in most patients. Serial follow-up testing (starting at >3 days) demonstrated further decreases in lymphocyte and eosinophil counts, and these changes were positively correlated with disease severity. Significantly higher levels of D-dimer (>0.4 μg/mL; to convert to nmol/L, multiply by 5.476), CRP (4.76 mg/dL), and procalcitonin (0.01 μg/dL) were associated with a severe disease course. Older age, multiple comorbidities, and prominent laboratory abnormalities were also associated with severe disease course. Allergic diseases, asthma, and chronic obstructive pulmonary disease were not risk factors for SARS-CoV-2 infection in this study.

Other studies have demonstrated an association of severity, disease progression, and adverse outcome with older individuals (>60 years) with comorbidities who were predominantly males.^22–25  Hypertension,^5,22,23,26  diabetes mellitus,^5,22,23,26  cardiovascular disease,^5,22,23,26  and high body mass index²⁴  were among the most commonly reported risk factors. Inconsistent results were reported for chronic obstructive pulmonary disease and asthma as risk factors.^6,24,27  Two studies analyzing the correlation between ABO blood group and clinical outcome suggested that blood group A individuals as well as those with rhesus antigen positivity were more susceptible to SARS-CoV-2 infection compared with blood groups O or AB.^28,29  However, blood groups are not among the risk factors for either disease severity or complications. The leading cause of death in SARS-CoV-2 infection was acute respiratory distress and respiratory failure, but systemic involvement with end organ damage, including sepsis, thrombotic or hemorrhagic events, cardiac failure, liver failure, and renal failure, also contributed to death.^5,22,23,30 

A few studies examined the presence of changes in blood tests found concurrently with initial positive SARS-CoV-2 RT-PCR results.^25,26  Most molecular diagnostic tests have used the RT-PCR targeting different SARS-CoV-2 genomic regions, including the ORF1b or ORF8 regions, and the nucleocapsid (N), spike (S) protein, RNA-dependent RNA polymerase (RdRP), or envelope (E) genes.³¹ 

An early study was performed on patients admitted to the emergency department in Italy.²⁵  Retrospective analysis was used to compare the results of routine blood tests in COVID-19–positive (n = 105) and COVID-19–negative (n = 102) patients where infection status was confirmed by molecular testing. On admission, samples from the RT-PCR positive group, who were predominantly male, showed significant differences (P < .05) in levels for WBC count (6.47 ± 2.61 × 10³/μL), CRP (8.71 ± 8.12 mg/dL), aspartate aminotransferase (AST; 56.2 ± 40.8 U/L; to convert to μkat/L, multiply by 0.0167), alanine aminotransferase (ALT; 47.9 ± 40.9 U/L; to convert to μkat/L, multiply by 0.0167), and LDH (388.0 ± 154.5 U/L). No association was found between positive status and platelet count. The study demonstrated that routine blood tests had detection rates similar to those of RT-PCR. A positive predictive value and negative predictive value of 83.3% and 90.6%, respectively, were reported.

A similar study evaluated the accuracy of laboratory parameters in predicting cases with positive RT-PCR for COVID-19.²⁶  Based on the result of RT-PCR testing, patients were classified into positive (n = 70) and negative (n = 130) groups. Mean patient age was 41.3 ± 14.6 years. No sex predominance was observed. Significantly lower WBC count (4.04 ± 1 × 10³/μL), increased neutrophil count, and hypoalbuminemia (2.9 ± 0.8 g/dL; to convert to g/L, multiply by 10), as well as elevated levels of CRP, LDH (465.2 ± 100.2 U/L), AST (32.1 ± 8.01 U/L), and ALT (37.8 ± 7.9 U/L), were observed in the RT-PCR positive group compared with the negative group. Overall, neutrophil counts, CRP, LDH, and ALT had a good predictive value (area under the curve [AUC] >0.8) and were proposed as useful markers to predict the result of molecular testing.

Several clinical pathology studies have reported changes in hematologic parameters (Table 1). A study of the viral genomes has suggested 2 major viral genetic variations of SARS-CoV-2, Clade I and Clade II.³²  Both strains showed similar virulence and clinical outcomes.³²  Decreased lymphocyte count on admission, especially CD4⁺ and CD8⁺ T lymphocytes, was the most important predictor of disease progression and outcome, along with age.³² 

A large multicentric, retrospective study including 548 confirmed COVID-19 patients demonstrated significantly different hematologic and immunologic parameters on admission and at end point between survivors and nonsurvivors.³³  The overall median age was 56 years, with significantly older age for nonsurvivors. On admission, marked lymphopenia, decreased eosinophil count, increased neutrophil count, and thrombocytopenia were predominantly seen in severe and critical cases, and in nonsurvivors. High neutrophil to lymphocyte ratio (NLR) as inflammatory biomarker and indicator of systemic inflammation was also observed among those groups. Low CD3, CD4, and CD8 counts were seen in severe cases. During hospitalization, the longitudinal variation in levels of lymphocytes, eosinophils, and platelets showed an increasing trend in survivors and a downward trend in nonsurvivors. Additionally, nonsurvivors showed an upward trend of neutrophil counts.

The dynamics of hematologic parameters and coagulation factors were compared between mild and severe groups in 1 retrospective study.³⁴  The mean age of the cohorts (n = 75) was 46.6 ± 14 years. On admission, significantly severe lymphopenia in addition to higher NLR, D-dimer, and fibrinogen levels was observed in the severe group compared with the mild group (P < .05). Serial measurements of NLR (on days 1, 4, and 14) and D-dimer levels (on days 1, 7, and 14) differed significantly between the 2 groups. Another study suggested that elevated NLR was significantly associated with disease severity and was an independent biomarker for poor clinical outcomes.³⁵ 

In summary, leukocytosis characterized by neutrophilia; lymphopenia with low CD3, CD4, and CD8 subset counts and/or percentages; and increased NLR were associated with severe disease in several studies (Table 1). Lymphopenia and clinically severe disease were the main risk factors contributing to longer hospitalization.³⁶  Thrombocytopenia was more likely to occur in severe or critically ill and fatal cases (Table 1). Other laboratory findings, including decreased WBC count,^25,26  decreased neutrophil count,²⁵  decreased eosinophil count,^{3,6,25,33,37,38}  and variable monocyte count,^25,36,39  have been described among COVID-19 patients.

The major coagulation indices, including PT, INR, D-dimer, or fibrinogen/fibrin degradation products (FDPs) in SARS-CoV-2–infected patients were analyzed among the different disease severity groups, and results were compared to healthy controls in several studies (Table 1). Coagulation abnormalities were examined in a group of 22 COVID-19 patients with acute respiratory failure requiring ICU admission.⁴⁰  Patients were predominantly male and 67 ± 8 years of age. Significantly higher plasma fibrinogen (517 ± 148 mg/dL; to convert to g/L, multiply by 0.01) and D-dimer (5.3 ± 2.1 μg/mL) levels were observed compared with healthy controls (n = 44).

Similar results were reported in another retrospective study of 94 patients with confirmed SARS-CoV-2 infection.⁴¹  Levels of D-dimer (10.36 ± 25.31 μg/mL), fibrinogen (502 ± 153 mg/dL), and FDP (33.83 ± 82.28 μg/mL; to convert to mg/L, multiply by 1) were significantly higher in all SARS-CoV-2 cases than in healthy controls (n = 40). Moreover, D-dimer (19.1 ± 35.5 μg/mL versus 2.1 ± 2.9 μg/mL) and FDP (60 ± 10.9 μg/mL versus 7.9 ± 11.4 μg/mL) values were higher in severe cases compared with mild cases.

Few selected retrospective studies compared coagulation function in patients with mild, severe, and critical disease. One large retrospective study (n = 380) demonstrated that PT, D-dimer, and FDP were significantly higher in critically ill patients compared with those with mild and severe disease, and their levels correlated positively with disease severity.³⁰  Sepsis-induced coagulopathy and disseminated intravascular coagulation (DIC) scores were increased over time during disease progression.³⁰  Another study suggested that coagulation parameters have good predictive value and ability to discriminate between mild, severe, and critical disease states (AUC >0.8).⁴² 

In summary, significant coagulation abnormalities were observed in patients with SARS-CoV-2 compared with healthy controls, and they correlated positively with disease severity. Monitoring coagulation parameters may prove helpful for the early identification of severe cases.

One of the earliest host responses to viral or bacterial infection is the activation of acute-phase reactants, including CRP, ferritin, SAA, albumin/prealbumin, procalcitonin, erythrocyte sedimentation rate (ESR), and proinflammatory cytokines, among others. CRP, a member of the short pentraxins group, is a 25-kDa protein synthesized in the liver and known to interact with complement factor c1q in the activation of the complement cascade.⁴³  Cytokines are mainly secreted from hematopoietic cells, such as lymphocytes and macrophages, and may play an important role in the clinical course of disease and outcome. During the acute phase of illness, a dysregulation of inflammatory response may occur with massive release of cytokines (cytokine release syndrome or cytokine storm), causing damage of single or multiple organs.

The most commonly measured cytokines in patients with SARS-CoV-2 infection were IL-2/2R, IL-6, IL-10, tumor necrosis factor-α (TNF-α), and interferon-γ (Table 2). Several studies reported that IL-6 was the most sensitive cytokine and may be used as a biomarker for evaluating prognosis.^3,22,33,44  An elevated level of IL-6 (>10 pg/mL) was proposed as the significant driving force for the cytokine storm and as contributing to the multiple-organ failure seen in advanced cases.⁴⁵  An extremely high IL-6 level (up to 100 pg/mL) was closely correlated with the incidence of detectable serum viral particles by RT-PCR.⁴⁴  Other inflammatory markers, including interferon-γ inducible protein (IP-10) and human monocyte chemotactic protein (MCP-1/MCP-3), were also significantly increased in severe cases.^5,45  A study examining cerebrospinal fluid cytokine levels in a case series of 3 patients with respiratory and neurologic complications has reported markedly increased levels of IL-6, IL-8, and IL-10.⁴⁶ 

A comparison study between survivors and nonsurvivors showed an upward trend of acute-phase proteins, including CRP, ferritin, SAA, procalcitonin, and cytokine IL-6, in nonsurvivors, and a stable or downward trend in survivors.³³  Within the survivor group, significant differences in CRP,³  ferritin,³⁷  SAA,^22,47–49  procalcitonin,⁴⁷  and IL-6³⁷  were reported between mild, severe, and critically ill patients. There was also a downward trend of prealbumin and albumin as the disease progressed from mild to severe or critical condition.⁴⁷  Increased ESR levels have also been reported.^50,51 

The inflammatory markers among mild, severe, and critically ill patient groups were analyzed to identify their correlation with disease progression.³⁷  Significant differences in serum CRP, ferritin, and procalcitonin were reported between the groups. Interleukin-2 receptor, IL-6, IL-10, and TNF-α levels were significantly lower in mild compared with severe and critically ill groups (P < .001). The study concluded that CRP >3.07 mg/dL, ferritin >2252 ng/mL (to convert to μg/L, multiply by 1), IL-2R >0.7935 U/L, and IL-6 <100 pg/mL were associated with disease progression.

Several studies reported increased levels of SAA among patients with mild and severe disease.^22,47–49  Gradual increase of SAA was seen as the disease progressed from mild to severe.⁴⁹  A markedly increased SAA level (>300 μg/mL; to convert to mg/L, multiply by 1) was observed in severe and critically ill patients and positively correlated with disease severity.^22,47  Serial measurements of SAA may aid in monitoring the extent of pneumonia, and the levels correlated well with the dynamic changes seen on serial computerized tomography scans.^49,52 

One study of 75 patients in Italy measured the level of Presepsin (PSP) in addition to routine laboratory tests.⁵³  Presepsin, or soluble cluster of differentiation CD14-subtype (sCD14-ST), is a regulatory factor that modulates immune responses by interacting with T and B cells and has been demonstrated to be a better marker for early diagnosis of sepsis compared with other sepsis markers in non-COVID19 patients.⁵⁴  Prespesin was significantly higher in ICU patients and fatal cases than in non-ICU patients. High PSP level (>25 ng/dL) correlated with longer ICU stay, but the level correlated poorly with CRP, LDH, and procalcitonin levels measured on the same day. The study suggested that PSP may be used for risk stratification of SARS-CoV-2 patients and early identification of disease severity and longer hospitalization times.

In summary, commonly used inflammatory biomarkers, such as CRP, ferritin, procalcitonin, SAA, and IL-6, and, less commonly used, PSP, were significantly increased in patients with COVID-19. The level of these biomarkers correlated with disease severity. Initial and serial measurement of these markers along with other parameters might aid in risk stratification and follow-up assessment of disease progression and improvement. At present, assays for SAA and PSP are not available in the United States, but their implementation should be considered.

Significant elevations of LDH, creatine kinase (CK), liver enzymes (AST and ALT), total bilirubin, blood urea nitrogen (BUN), and creatinine were commonly reported in severe and critically ill patients (Table 2).

The liver biochemistry and its association with other biomarkers were analyzed in COVID-19 patients among different disease severities (n = 60).⁵⁵  On admission, most nonintubated patients (69%) demonstrated elevated AST (median, 46 U/L) and ALT (median, 30 U/L). The dynamics of AST levels correlated with the level of LDH, CK, and ferritin, but not with CRP. In critically ill and intubated patients, a marked AST predominant increase was observed (69 and 364 U/L, respectively). No clear association was found between preexisting liver disease and clinical outcome of COVID-19 patients. Another study of the liver enzymes concluded that higher levels of bilirubin, AST, and γ-glutamyl transferase were predominantly observed in fatal cases.⁵⁶ 

The clinical features and outcome of patients with high brain natriuretic peptide (BNP) levels (>100 pg/mL; to convert to ng/L, multiply by 1 × 10³) were analyzed and the findings compared with a group of patients with normal BNP levels.⁵⁷  The high BNP group demonstrated higher CRP, AST, and c-Tn-I and were more likely to develop pneumonia requiring ICU admission.

Gong et al⁵⁸  reported that on admission, higher levels of serum LDH, CRP, direct bilirubin, and BUN, and lower levels of albumin were found to be associated with higher odds for severe disease and disease progression. Lastly, COVID-19 patients mounted with a marked acute cortisol stress response, and the level was significantly higher compared with non-COVID-19 individuals, 22.44 versus 18.81 μg/dL (to convert to nmol/L, multiply by 27.588), respectively.⁵⁹ 

Several studies compared the hematologic, biochemical, inflammatory, and coagulation parameters in mild, severe, and critical disease, and reported comparable findings.^8,60  However, in some of the studies statistical analysis was limited because of an unavailability of serial test specimens in many patients, resulting in smaller total sample size. The median age in mild, severe, and critical groups ranged from 50 to 51.4 years,^8,60  63.9 to 64 years,^30,60  and 66 years,⁷  respectively.

Chen et al⁶⁰  reported that CRP, ferritin, LDH, and ALT were significantly higher in severe cases compared with mild cases. As the disease progressed from mild to severe or critical, a downward trend for lymphocytes, prealbumin, and albumin was observed, whereas the opposite was seen for WBC count, neutrophil count, CRP, and LDH.^8,47  Other significant findings, including increased SAA, NLR, PT, D-dimer, FDP, and inflammatory cytokines IL-2R, TNF-α, and IL-10, have been reported.^30,47,60 

Association of serum viral load, disease severity, and outcome was analyzed in several studies.^44,51,61,62  In critically ill patients, serum viral load was correlated with extremely high levels of IL-6 (up to 100 pg/mL) and higher mortality.⁴⁴  Prolonged viral shedding (up to 57 days), as demonstrated by serial RT-PCR assay and anti–SARS-CoV-2 immunoglobulin M (IgM), and higher levels of ESR, CRP, ferritin, and IL-4, was observed among patients in the poor recovery group.⁵¹  The risk of intubation and in-hospital mortality increased significantly with high viral load on admission compared with low or mild viral load.⁶¹  Other reported laboratory characteristics of critically ill patients requiring ICU admission were progressive decrease in lymphocyte count with worsening lymphopenia, decreased eosinophil count, thrombocytopenia, elevated CRP, elevated procalcitonin, increased IL-6 and IL-10, increased liver enzymes, increased total bilirubin, decreased renal function, hypercoagulable state higher, and higher incidence of complications.^3,7,44  Contrary to previously described findings, another study found no significant association between viral load, admission to ICU, and outcome.⁶²  Higher viral load was associated with shorter duration of symptoms and shorter hospital stay.⁶² 

In summary, progressive worsening of laboratory parameters—including decreased lymphocyte count, and increased NLR, CRP, ferritin, IL-6, IL-10,—certain coagulation parameters, and serum viral load were observed as the disease progressed. However, lymphocytopenia and age appeared to be the most important determinant of disease severity.³² 

Retrospective studies analyzing the nonsurvivor/fatal cases demonstrated a median patient age from 71 to 72.5 years.^22,23  An unfavorable prognosis was observed predominantly in male patients and those with underlying medical conditions, including diabetes mellitus, hypertension, and cardiac disease.²² 

On admission, most of the patients demonstrated lymphopenia, neutrophilia, thrombocytopenia, prolonged PT, elevated D-dimer, high lactate level, lower oxygen saturation, high NLR (>5.0), and systemic immune-inflammation index (>500).^22,23  All patients had high initial CRP and IL-6 (>10 pg/mL) levels, and a large subset of patients had elevated LDH, D-dimer, procalcitonin, and c-Tn-I levels. These groups were reportedly more likely to develop sepsis, have rapid disease progression, and experience early death (within 3 days of admission).²³  Laboratory results obtained 24 hours prior to death showed elevated serum AST, ALT, CK-MB, myoglobin, BUN, and creatinine levels.²²  In this study, the leading cause of death was respiratory failure followed by sepsis and multiple-organ failure.²² 

Furthermore, a study by McRae et al⁶³  analyzed the COVID-19 Severity Score and clinical decision support tools to predict mortality. To predict mortality, it combined measurements of multiple biomarkers: CRP, N-terminus pro-B type natriuretic peptide, myoglobin, D-dimer, procalcitonin, CK, and c-Tn-I, and risk factors in a statistical learning algorithm. The analysis found that COVID-19 Severity Scores were significantly higher for the group that died versus the group that was discharged, with median (interquartile range) scores of 59 (40–83) and 9 (6–17), respectively. These encouraging results have the potential to empower health care sustenance to save lives by prioritizing critical care for patients at risk of adverse outcomes.

Several systematic reviews with or without meta-analysis of hematologic, biochemical, and inflammatory marker abnormalities associated with disease severity were also reviewed, and the results are summarized in Table 3. Lymphopenia,^24,64–69  increased neutrophil count,^67,69  leukocytosis,²⁴  and thrombocytopenia^64,66,69  were the most common significant hematologic findings. One study demonstrated normal WBC count accompanied by lymphopenia.⁷⁰  Prolonged PT^41,71  or elevated D-dimer/FDP^67–69,71  represented derangements in the coagulation profile. Elevated inflammatory markers—CRP,^{24,65,66,68,69,72,73}  ferritin,^64,72,73  procalcitonin,^24,66  SAA,^69,72  and IL-6^64,67,69,72 —and other biomarkers—LDH,^24,66–69  CK,²⁴  CK-MB,⁶⁷  AST/ALT,^24,41,66,67  bilirubin,^24,74  BUN,^67,69  and creatinine^67,69 —were also among significant findings. Patients with severe and fatal disease had significantly higher WBC count and lower lymphocyte and platelet counts compared with nonsevere disease or survivors.⁶⁴  Thrombocytopenia and elevated levels of procalcitonin and c-Tn-I were associated with severe disease.⁶⁶  In severe and fatal disease, biomarkers of inflammation, cardiac and muscle injury, renal and liver function, and coagulation parameters were significantly elevated.⁶⁴  Additionally, IL-6, IL-10, and serum ferritin were strong discriminators for severe disease.⁶⁴ 

A great proportion of COVID-19 studies concluded risk factors contributing to severe disease and adverse outcomes include advanced age (>60 years) and male sex, with comorbidities, such as hypertension, diabetes mellitus, and cardiovascular disease. The most common initial clinical symptoms were fever, cough, dyspnea, and fatigue.^4–8  However, it is unclear whether the symptoms will become more insidious as the pandemic progresses and will gradually evolve into a virus similar to influenza or remain latent in humans for a long period of time, as suggested by Chen et al.⁴⁸ 

Wide ranges of laboratory abnormalities were reported with different disease severity, but marked changes were more commonly seen in samples from severe and critically ill patients.²⁴  Hematologic parameters, including lymphopenia, leukocytosis with increased neutrophil count, increased NLR, and thrombocytopenia, were the most common findings observed and positively correlated with disease severity.^22,30,47  A strikingly decreased lymphocyte count was associated with severe disease and higher complication rate. The decreases in both CD4 and CD8 T lymphocytes are best explained by the roles these T-lymphocyte subsets play in eliminating virus-infected cells, and this is consistent with low lymphocyte counts being associated with poor case outcomes.^32,64 

An upward trend of CRP, ferritin, SAA, procalcitonin, and the most prominent cytokine, IL-6, and a downward trend of albumin and/or prealbumin were frequently observed during progression from mild to severe/critical condition, and in nonsurvivors. Serial measurements of these markers can be used to predict disease course, severity, and mortality.^22,37,47  It has been postulated that SARS-CoV-2 may target alveolar macrophages via the angiotensin converting enzyme 2 (ACE2) receptor, leading to an increase in cytokine secretion, including IL-6 and TNF-α, which subsequently induces the elevation of various APPs, such as CRP, SAA, and complement factor, which are significantly upregulated in the severely ill group. Changes in coagulation parameters, including prolonged PT, elevated D-dimer, and elevated fibrinogen or FDP, were common findings in severe disease and nonsurvivors.^4,30  Prolonged PT and higher serum D-dimer levels were postulated to demonstrate a hypercoagulable state rather than consumptive coagulopathy. It was proposed that hyperfibrinogenemia leads to fibrin polymerization, thrombus formation, and eventually complications or adverse outcome.⁴⁰  Other biomarkers, such as LDH, CK, BNP, AST, and ALT, have been associated in several studies with severe and critically ill disease, and their levels likely indicate adverse outcome.^24,55,56  AST-dominant elevations were common in COVID-19 patients and appeared to reflect true hepatic injury.⁵⁵  Additionally, AST level correlated with markers for muscle injury, including LDH and CK.⁵⁵ 

To date, molecular testing identifying viral particles on nasopharyngeal specimens by RT-PCR remains the gold standard in the diagnosis of SARS-CoV-2 infection. Concurrent antibody testing can aid in increasing detection sensitivity.⁷⁵  Other specimen sources, such as self-collected saliva, which is less painful and does not require trained personnel for collection, should be considered as preferable alternatives for SARS-CoV-2 screening of health care workers and asymptomatic cases.^11,76  The sensitivity of SARS-CoV-2 detection from saliva samples was demonstrated to be comparable to that of nasopharyngeal swabs in early hospitalization.⁷⁶  The results were more consistent during extended hospitalization and recovery, most likely because of less temporal SARS-CoV-2 variability.⁷⁶  Routine blood tests may also be used as early predictors of the molecular testing result. They can serve as an alternative for identifying SARS-CoV-2 infection in countries with heavy outbreaks and a shortage of RT-PCR reagents or specialized labs, because they have been shown to have detection rates comparable with those of molecular tests.²⁵  Moreover, early recognition of severe disease or disease that is likely to progress is absolutely essential for timely triage of patients. Measurement of soluble PSP or sCD14-ST in peripheral blood may be used for early diagnosis of sepsis and for risk stratification.⁵³  The use of proteomic approaches and nontraditional samples, like cerebrospinal fluid, may identify additional biomarkers that may be helpful in the pathophysiology and prognosis of COVID-19 disease.^20,46 

Interpretation of the results of several studies presented in this review is limited because of the predominantly retrospective study designs, small sample sizes, multiple sampling biases (ie, most were single-center studies with cohorts from east Asian ethnic groups), lack of uniformity of disease severity definition based on variable RT-PCR methods, and lack of exact timeline of laboratory sample collection, as well as lack of serial sample measurements. This information is important because a defined timeline of collection and serial sampling performance may aid in clinical decision-making during the acute phase of disease. In addition, a great proportion of studies were cut short and failed to report final outcomes because of the need for prompt data publishing during the current pandemic. Overall, the reader should be cautioned when viewing medRxiv papers presented prior to peer review. It is recommended that updated peer-reviewed published versions of these original studies be evaluated when/if available.

The dynamic changes in biomarker levels may assist in predicting disease course, prognosis, and outcome. Indicators of systemic inflammation, such as NLR and systemic immune-inflammation index or coagulopathy screening using a DIC scoring system, could be appropriately used to predict disease severity, possible complications, and outcome. Finally, COVID-19 Severity Score and clinical decision support tools can additionally be used employing combined measurements of multiple biomarkers to predict mortality. In summary, WBC, lymphocyte, and platelet counts, CRP, ferritin, and IL-6 may be potential prognosticators of progression to critical illness. Therefore, prospective studies with larger cohorts, clearly defined disease severity, and serial measurements with defined sampling collection timelines are imperative to further confirming the correlation and significance of the current findings.