Laboratory features of severe vs. non-severe COVID-19 patients in Asian populations: a systematic review and meta-analysis

Background More severe cases of COVID- 19 are more likely to be hospitalized and around one-fifth, needing ICU admission. Understanding the common laboratory features of COVID-19 in more severe cases versus non-severe patients could be quite useful for clinicians and might help to predict the model of disease progression. This systematic review and meta-analysis aimed to compare the laboratory test findings in severe vs. non-severe confirmed infected cases of COVID-19. Methods Electronic databases were systematically searched in PubMed, EMBASE, Scopus, Web of Science, and Google Scholar from the beginning of 2019 to 3rd of March 2020. Heterogeneity across included studies was determined using Cochrane’s Q test and the I2 statistic. We used the fixed or random-effect models to pool the weighted mean differences (WMDs) or standardized mean differences and 95% confidence intervals (CIs). Findings Out of a total of 3009 citations, 17 articles (22 studies, 21 from China and one study from Singapore) with 3396 ranging from 12 to1099 patients were included. Our meta-analyses showed a significant decrease in lymphocyte, monocyte, and eosinophil, hemoglobin, platelet, albumin, serum sodium, lymphocyte to C-reactive protein ratio (LCR), leukocyte to C-reactive protein ratio (LeCR), leukocyte to IL-6 ratio (LeIR), and an increase in the neutrophil, alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bilirubin, blood urea nitrogen (BUN), creatinine (Cr), erythrocyte Sedimentation Rate (ESR), C-reactive protein (CRP), Procalcitonin (PCT), lactate dehydrogenase (LDH), fibrinogen, prothrombin time (PT), D-dimer, glucose level, and neutrophil to lymphocyte ratio (NLR) in the severe group compared with the non-severe group. No significant changes in white blood cells (WBC), Creatine Kinase (CK), troponin I, myoglobin, IL-6 and K between the two groups were observed. Interpretation This meta-analysis provides evidence for the differentiation of severe cases of COVID-19 based on laboratory test results at the time of ICU admission. Future well-methodologically designed studies from other populations are strongly recommended.

been more than 4000,000 confirmed reported COVID-19 cases affected by and more than 280,000 reported deaths until May 13th [1]. While there are parallels between COVID-19 and the severe acute respiratory syndrome (SARS), variations in the clinical characteristics of the diseases caused by the two viruses have been noted [2]. Urgent identification of clinical laboratory predictors of disease progression toward severe/critical form is an urgent necessity for clinicians to be able to stratify risks, distinguish and differentiate severe patients from the mild/moderate form of COVID-19.
Based on the clinical symptoms and laboratory test results, patients are categorized as mild, moderate, severe, and critical types [2,3]. Mild/moderate cases include most of the affected patients (81%). Although severe and critical ones comprise only 14% and 5% of infected cases, respectively [4] they mainly need hospitalization. Almost 20% of hospitalized patients need intensive care unit (ICU) [5]. As such, the mortality rate of ICU admitted COVID-19 patients is reported quite high, nearly 61.5% die due to many different reasons [6].
Apart from the clinical symptoms and pulmonary computed tomography (CT) findings, a large number of COVID-19 confirmed patients showed laboratory fluctuations including complete blood count (CBC) variables, cardiac and coagulation parameters, renal and liver function tests, and inflammation-related factors [7,8].
Recently, combinations of some laboratory tests have been used in some settings to show the hyperinflammation state and prognosis. These combinations include neutrophil to lymphocyte ratio (NLR) and lymphocyte to C-reactive protein ratio (LCR) [9][10][11].
Among the CBC parameters of COVID-19 confirmed cases, decreased lymphocytes and normal or increased monocytes have pointed out previously [7]. Combined evidence so far appears inconsistent [6,7,[12][13][14][15]. In the previous meta-analysis, the most prevalent laboratory features that were decreased in all confirmed cases consist of albumin, high CRP, high lactate dehydrogenase (LDH), lymphopenia, and high ESR, while increase in bilirubin, cytokines, and leukocytes was less frequent [5]. Since the pandemic outbreak of COVID-19, a vast number of studies investigated the laboratory changes in confirmed COVID-19 patients, leaving the association between the routine laboratory features and the disease severity as less attention subject. Therefore, in this study, we conducted a systematic review and meta-analysis to quantify the results of previously published studies, comparing the CBC indices cardiac and coagulation parameters, electrolyte factors, renal and liver function tests, inflammation-related factors, and some new combined inflammatory laboratory tests in severe/critical versus non-severe confirmed infected cases of COVID-19.

Methods
The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines were used for performing and reporting our systematic review and meta-analysis.

Search strategy
Electronic databases were systematically searched in PubMed, EMBASE, Scopus, Web of Science, and Google Scholar from the beginning of 2019 to 3rd of March 2020. The reference lists of relevant studies and previous reviews manually were checked to retrieve more studies. Search terms included "2019 novel coronavirus infection" OR "COVID-19" OR "COVID19" OR "coronavirus disease 2019" OR "coronavirus disease-19" OR "2019-nCoV disease" OR "2019 novel coronavirus disease" OR "2019-nCoV infection" OR "2019-nCoV" OR "2019 novel coronavirus" OR "2019 coronavirus" OR "novel coronavirus" OR (2019 AND coronavirus). Additional file 1: Appendix 1 provides of search strategy from Scopus database.

Inclusion and exclusion criteria
Two independent researches (M.A-K and N.Z) assessed all retrieved reports using our inclusion and exclusion criteria. Differences were resolved through consensus or discussion with a third author (R.T). Studies that met the following criteria were included in our study: original studies with cross-sectional, case-control, and cohort design in English language; studies that investigated laboratory features [includes CBC (neutrophil, lymphocyte, monocyte, eosinophil, hemoglobin, and platelet), liver and kidney functions [alanine aminotransferase (ALT), aspartate aminotransferase (AST), albumin, total bilirubin (TBIL), blood urea nitrogen (BUN), and creatinine (Cr)], myocardial enzymes [creatine kinase (CK), troponin I, and myoglobin], inflammatory factors [ESR, CRP, LDH, procalcitonin (PCT), and IL-6] serum electrolytes (sodium, potassium), coagulation functions [fibrinogen, prothrombin time (PT), and D-dimer], and glucose level; we also combined inflammatory markers including NLR, LCR, leukocyte to C-reactive protein ratio (LeCR), and leukocyte to IL-6 ratio (LeIR). For a variable to be included in analysis, at least three studies that reported or appropriated to calculate the mean changes (standard deviation (SD)) of intended laboratory features in severe vs. non-severe COVID-19 were needed. Severe cases were defined based on severe cases of America thoracic society [16] or interim guidance of World Health Organization definition [17] or severe/critical case based on China's National Health Commission definition [18] or Acute Respiratory Distress Syndrome (ARDS) based on Berlin definition [19] or admission to ICU. Otherwise, patients were defined as non-severe. We excluded studies

Data extraction
Two researchers (SMAK and NZ) were independently extracted the following data from the included studies: first author's name, year of publication, type of publication, country, patient characteristics, total sample size, number of patients in severe and non-severe groups, mean (SD) of laboratory parameters in severe and nonsevere groups. Differences between the two researchers were resolved by consensus or discussion with a third author (PN-S).

Quality assessment
The quality of the included articles was critically assessed using the Newcastle-Ottawa Scale (NOS). This tool was conducted according to three aspects including selection, comparability, and exposure/outcome. A study with an NOS score of ≥ 7 was considered as good quality. Table 1 shows the results of the quality assessment of included studies based on NOS.

Statistical analysis
All statistical analyses were performed using STATA version 12.0 (Stata Corp., College Station, TX). Heterogeneity among included studies was assessed using Cochrane's Q test or the I 2 statistic. I 2 above 70% and Cochrane's Q test with P < 0.05 were considered as the existence of significant heterogeneity.
Laboratory factors were estimated as the mean (SD) difference with 95% confidence intervals (CIs) between severe and non-severe groups. Weighted mean difference (WMD) with the random-effects model (DerSimonian-Laird method) was used to pool the mean differences of each laboratory factors in studies with the same clinical units and measures; otherwise, the standardized mean difference (SMD) was applied. We used a series of sensitivity analysis to assess the robustness of our findings; applying the leave-one-out method to test the impact of each included study on the pooled WMDs or SMDs. The potential evidence of publication bias was assessed using the Egger regression and Begg's rank correlation tests.

Results
A total of 3009 citations were identified through electronic database searches. Of these, 1021 were duplicate reports. After screening titles and abstracts, 1694 articles were excluded and 294 full-text articles were retrieved for more assessment. Finally, 17 articles (or 22 studies) were found to be eligible for this meta-analysis. We summarized the process of study identification and selection in Fig. 1.

Main outcomes
The mean difference forest plots of laboratory features in severe vs. non-severe hospitalized patients with COVID-19 are shown in Additional file 2: Appendix 2: 2a-h.

CBC with differential counts of WBC
Using random-effects model, our meta-analyses showed a significant decrease in the WMD of lymphocyte [WMD = − 0.43 × 10 9 per L; 95% CI − 0.56,   Fig: A-F).

Myocardial enzymes and myoglobin
The pooled findings showed no significant differences between the two groups of COVID-19 patients    Fig: A-D). Table 2 shows a summary of laboratory features in severe vs. non-severe hospitalized patients with COVID-19.

Sensitivity analysis
We evaluated the effect of each study on the strength of the pooled WMDs or SMDs by excluding each study from the meta-analysis. We found no significant differences between the pre-and post-sensitivity pooled effect sizes for lymphocyte, hemoglobin, platelet, neutro-  [7], the study on LeIR (SMD = − 0.52, 95% CI − 1.35, 0.31), we found significant differences between pre-and post-sensitivity pooled effect sizes.

Discussion
To the best of authors' knowledge, this is the first metaanalysis that elaborated on the differences between laboratory tests results of severe and non-severe confirmed cases of COVID-19. The results showed the significant decreased levels of lymphocyte, monocyte, eosinophil, hemoglobin, and platelet, while elevated neutrophil counts among the CBC indices in severe vs. non-severe patients. ALT, AST, TBIL, BUN, and Cr levels showed the significant increase, following a decrease in albumin level as the main liver and kidney outcomes in severe patients compared to non-severe ones. Inflammatory/ infection markers (ESR, CRP, LDH, and PCT, but not IL-6), coagulation function tests (fibrinogen, PT, and D-dimer), and glucose were positively associated with the COVID-19 severity. However, serum sodium decreased in severe patients in comparison with non-severe group. Besides, potassium, and other cardiac-related factors were not associated with COVID-19 severity. It is worth to mention that combined ratios significantly decreased (LCR, LeCR, and LeIR6) and increased (NLR) in severe/ critical compared to non-severe COVID-19 patients. The laboratory features presented by this meta-analysis could, to a large extent, be attributed to low oxygen saturation, respiratory failure, septic shock, and/or multiple organ dysfunction or failure compatible with the ARDS course of severe and critical COVID-19 types [2,3].
COVID-19 is a systemic disorder affecting multiple organs. Abnormal kidney and liver function test results and elevated serum glucose might happen in more severe cases due to the following explanations. These include, but not limited to hypoxia, hypoperfusion as well as thrombosis caused by ARDS, shock and disseminated intravascular coagulation respectively. Aggravation of underlying disease in severe cases that are older and suffer from concomitant comorbidities also is another possibility for this feature. Drug-induced damages [23] to the liver are also one potential factor. Secondary hemophagocytic lymphohistiocytosis (HLH) characterized by hypercytokinemia with multiorgan failure [28] which occurred after viral infection in adults [29,30] is another possible explanation for the stated laboratory features. Although one no peerreviewed study by Zheng et al. [15] has demonstrated that organ damage in COVID-19 is mostly endorsed as organ damage caused by the virus itself rather than HLH. This claim needs more investigation and autopsy studies.
Myocardial damage in COVID-19, especially in severe cases, has been demonstrated previously [22,24]; however, our analysis did not find significant higher level of myocardial enzymes in severe compared to non-severe cases. We believe more studies on the pathophysiology of such injury may help this issue to be clearer, one possibility could be interstitial infiltration of mononuclear cells instead of direct damage of myocardium in severe cases [23] which it is not in line with the finding of Zheng et al. [15].
The present findings showed a decreased in lymphocyte, eosinophil and monocyte counts but slightly increased of neutrophil in severe cases. Lymphopenia as a dysregulation in the immune response is mapped through more decrease in T cell and especially T helper cells in severe cases of COVID-19 [31]. The mechanism of this lymphopenia seems to be both due to the cytotoxic action of virus [6] and the collective characteristics of severe patients which are more likely to be older and have underlying diseases [24,26,32] which make them more susceptible to endothelial dysfunction and its correlated lymphopenia [33]. Findings of current study on decreased eosinophil and monocytes but increased neutrophils in severe cases need more investigations through future studies. These results highlight the potential of simple available tests besides the lymphocytes count for early screening in severe and critically ill COVID-19 cases.
The NLR was a commonly used index for determining the bacterial infections severity and for the prognosis of pneumonia and tumor patients [34,35]. The current study suggests the increased NLR and decreased LCR of severe COVID-19 patients in which been recently reported [36] but to the best of our knowledge, our metaanalysis for the first time showed decrease in LeCR, and LeIR in severe COVID-19 patients compared to nonsevere, indicating the poor prognosis of this pneumonia.
Lower hemoglobin level in more severe cases could be due to underlying medical conditions, malnutrition or coagulation abnormality. Coagulation abnormality is also contained in low platelet count, increased fibrinogen, D-dimer level and prolonged PT in severe patients.
An increase in PCT and neutrophil may be correlated with the concomitant bacterial infection in severe cases. Therefore, high PCT and neutrophil in two laboratory tests might be beneficial for the prognosis of severe cases. Despite that our meta-analysis found higher levels of ESR, CRP and LDH as inflammatory markers in severe cases of COVID-19, the results did not show increased IL-6 level in severe cases. This finding was mainly affected by the results of the study by Liu et al. [37] and should be confirmed with more investigations on IL-6 as well as other cytokines and chemokines.
Our findings provide evidence for the differentiation of severe cases of COVID-19 based on laboratory test results collected at the admission time. Our results also may provide the potential for efficient resource allocation in the era of scarcity of available resources including ICU.
However, inadequate evidence from outside China [27] may limit the generalizability of our results especially in terms of availability of the costly laboratory tests in low resource settings. Accordingly, as more data become available from outside of China in the ongoing weeks, it is highly recommended that new data are interpreted and compared with the present findings as the virus may differently affect based on both genetic and/or interplay environmental factors diversity. Moreover, heterogeneity of included studies as another limitation implies the need for further studies.

Conclusions
In summary, our meta-analysis provides evidence for the differentiation of severe and non-severe cases of COVID-19 based on the laboratory test results on the time of admission. The results of CBC test, liver and kidney function tests, inflammatory/infection markers, serum electrolytes and glucose were significantly different between severe and non-severe cases of COVID-19. However, to confirm our results, we think further studies, particularly from other populations, are needed.