key: cord-0788894-hwj42cq3
authors: Payán-Pernía, Salvador; Gómez Pérez, Lucía; Remacha Sevilla, Ángel F; Sierra Gil, Jordi; Novelli Canales, Silvana
title: Absolute Lymphocytes, Ferritin, C-Reactive Protein, and Lactate Dehydrogenase Predict Early Invasive Ventilation in Patients With COVID-19
date: 2020-12-18
journal: Lab Med
DOI: 10.1093/labmed/lmaa105
sha: d26b0ca2bcee9e13d61d0066ab19e17cd551f80a
doc_id: 788894
cord_uid: hwj42cq3

OBJECTIVE: Early detection of patients with COVID-19 who will need mechanical invasive ventilation (MIV) may aid in delivering proper care and optimizing the use of limited resources. METHODS: In this single-center retrospective observational study, we aimed to identify simple laboratory parameters that in combination with ferritin (a surrogate marker of severe inflammation) may help predict early (first 48 hours) MIV. A total of 160 patients with COVID-19 in whom serum ferritin, absolute lymphocyte count (ALC), platelet count, C-reactive protein (CRP), and lactate dehydrogenase (LDH) had been analyzed at admission were included. RESULTS: We found that ferritin, LDH, ALC, and CRP predicted with 88% accuracy the probability of early MIV. Results indicated that LDH showed the greater area under the curve (AUC), with a value of 89.1%. Using the AUC, we established cutoff values for clinical application. Finally, we developed a classification tree based on LDH for its clinical use. CONCLUSION: Ferritin, LDH, ALC, and CRP predict with 88% accuracy the probability of early MIV.

As of October 4, 2020, the World Health Organization reported a total of 34,804,348 confirmed COVID-19 cases of infection globally, including 1,030,738 deaths. 1 The most common clinical features at the onset of the illness caused by SARS-CoV-2 are fever, fatigue, and dry cough. Patients with severe illness may develop dyspnea and hypoxemia within 1 week after onset of the disease, which may quickly progress to acute respiratory distress syndrome (ARDS) or end-organ failure. 2 Since the outbreak in December 2019, the sudden increase in COVID-19 cases of infection is putting high pressure on healthcare services worldwide, with particular significance in intensive care units (ICU). Reported rates of ICU admission represent up to one-quarter of hospitalized patients, but rates vary among countries. 3 These differences may relate to the availability of ICU beds, variations in practice and admission criteria, and differences in predisposing factors and testing availability.

Researchers have learned that ARDS is the most common complication for ICU admission; in a series of 1300 patients admitted to the ICU in the Lombardy region of Italy, 88% required endotracheal intubation and mechanical ventilation. 3 Similarly, two-thirds of patients with COVID-19 who required critical care in the United Kingdom had mechanical ventilation within 24 hours of admission. 4 Therefore, early detection of patients who will need mechanical invasive ventilation (MIV) may aid in delivering proper care and optimizing the use of limited resources, and this is of particular interest in lower-and middle-income countries.

Several laboratory parameters have been associated with worse outcomes in patients with COVID-19: elevated liver enzymes, ferritin, IL-6, lactate dehydrogenase (LDH), C-reactive protein (CRP), D-dimer, prothrombin time, troponin, and creatine phosphokinase, along with lymphopenia and acute kidney injury. [5] [6] [7] Hyperferritinemia has been linked to macrophage activation syndrome (MAS), which is present in serious inflammatory disease 8 ; MAS is quite possibly the origin of the severest clinical manifestations of SARS-CoV-2 infection. 9 In this setting, we aimed to identify simple laboratory parameters that in combination with ferritin may help to predict early (first 48 hours) MIV by orotracheal intubation.

This retrospective observational study was performed at Hospital de la Santa Creu i Sant Pau, a first-level hospital in Barcelona, Spain. The study was conducted according to the Declaration of Helsinki and approved by the institutional ethics committee. For patient enrollment, all consecutive blood tests that included ferritin between March 15, 2020 and April 6, 2020 were reviewed. In this way, patients diagnosed with COVID-19 in whom serum ferritin, ALC, platelet count, CRP, and LDH had been analyzed at admission were selected. In all patients, these parameters were determined in the same sample or with a maximum time difference of 24 hours. We assessed ALC on a Sysmex XN-10 (Sysmex Corporation, Kobe, Japan) analyzer. Serum ferritin was measured on an Architect c16000 System (Abbott Laboratories, IL) using a 2-step chemiluminescent microparticle immunoassay. We determined CRP and LDH in serum using the Alinity c system (Abbot Laboratories, IL), the former using a particle-enhanced immunoturbidimetric assay, the latter using spectrophotometry.

One hundred sixty patients aged 23 to 75 years were recruited. Informed consent was obtained from all participants. One hundred fifty-eight patients had a laboratoryconfirmed SARS-CoV-2 infection according to World Health Organization guidance 10 : a positive result of real-time reverse-transcriptase polymerase chain reaction (RT-PCR) assay of a nasopharyngeal swab. In 2 patients, RT-PCR was negative and the diagnosis of COVID-19 was made presumptively based on a compatible clinical presentation and an exposure risk. Patients who received tocilizumab, a monoclonal antibody against IL-6, at any time during their hospital stay were excluded for the study because this agent has been associated with a decrease in CRP and ferritin levels. 11 Patients with active neoplasia were also excluded because of iron overload.

Demographic, clinical, laboratory, and outcome data were extracted from electronic medical records. An image on chest radiograph was considered typical of COVID-19 in the presence of consolidation and ground-glass opacities, with bilateral, peripheral, and lower lung zone distributions. All patients were evaluated until death or hospital discharge.

Descriptive analyses of the variables were expressed as mean (range) or the number of patients (%). We chose MIV as the dependent variable, and the independent variables were LDH, ALC, platelet count, ferritin, and CPR. To analyze the association between the independent variables and MIV, the Pearson χ 2 test was used, considering a type I error < 5%. A binomial logistic regression analysis was used for the joint evaluation of variables associated with MIV. The significant variables with P <.05 in the univariate analysis were selected for the regression analysis performed by the stepwise backward method (likelihood ratio). The variables that kept P ≤.05 after adjustments remained in the multiple regression model. A receiver operating characteristic (ROC) analysis was performed to measure the diagnostic/predictive accuracy of each significant variable.

Moreover, we developed a classification tree analysis using the chi-squared automated interaction detection (CHAID) growing method. This nonparametric analysis examines interactions among variables to create a decision tree without assuming that independent and dependent variables are linked by linear relationships. All statistical analyses were carried out using SPSS, version 21.0. Because the sample size was small (n = 160), we could not conduct an internal validation analysis. Table 1 shows the demographic and clinical characteristics of the 160 patients included in the study. Overall, 58.10% were men. The median age was 57 years (minimum: 23; Science www.labmedicine.com maximum: 75). A total of 32 patients (20%) required endotracheal intubation and mechanical ventilation, which happened in 96.9% of these patients within 48 hours after the emergency room admission.

Four independent variables were selected for logistic regression model fitting ( Table 2) : LDH, ALC, ferritin, and CRP. The platelet count was excluded (P >.05). Although each variable in the equation remained statistically significant, CRP and ferritin were close to the limit of significance (P =.046 and P =.045, respectively). Without including independent variables in the model, the probability of not being intubated was 80%. After the inclusion of the independent variables, the model's capacity to predict no intubation improved to 88%.

In the ROC analysis, LDH showed the greater AUC, with a value of 89.1%, followed by CRP (80.5%), ALC (77.6%), and ferritin (77.5%). Using the AUC, we established cutoff values for clinical application, as shown in Table 3 . We found that LDH might represent the most useful discrimination variable for clinical application: when it was <219 U/L, patients did not need MIV, with a sensitivity of 100% and a specificity of 11.7%. No patients with CRP <65.65 mg/L or an ALC >1.56 × 10 9 /L needed MIV. Ferritin levels <300 μg/L predicted no MIV with a sensitivity of 93.8%.

On the classification tree analysis (Figure 1) , LDH >598 U/L increased the likelihood of MIV, and an <424 U/L defined a category with the lowest probability of MIV.

Our work provides an approximation for the prediction of early (first 48 hours) MIV in patients with COVID-19 by means of simple laboratory tests that can be easily collected in any hospital. Although our results are concordant with those published by other authors about risk factors for severe disease and death, they refer specifically to the risk of MIV, which is of particular interest in the context of the high pressure on ICUs. Therefore, our data can help prioritize patients quickly when healthcare resources are limited.

We found that LDH is the biomarker that better can predict early MIV; with a sensitivity of 100%, patients will not be intubated if LDH on admission is <219 U/L. Based on LDH, we developed a classification tree to estimate the risk of MIV quickly. Research has shown that LDH is an intracellular enzyme found in nearly all organ systems that catalyzes the interconversion of pyruvate and lactate, with concomitant interconversion of Reduced nicotinamide adenine dinucleotide and Nicotinamide adenine dinucleotide. Abnormal values can not only result from cardiac damage or hemolysis but also from multiple organ injury and decreased oxygenation with upregulation of the glycolytic pathway. 

Classification tree. The root node (node 0) shows the binary (Yes/ No) distribution of the outcome of the dependent variable, MIV. The χ 2 test is then applied to ensure that the branch is associated with a statistically significant predictor of MIV-in this case, LDH. In severe COVID-19 infection, a deviation of the protective immune response into a dysfunctional program occurs, leading to cytokine release syndrome with severe inflammation and, eventually, multisystemic failure. A better understanding of the mechanisms lying at the root of immune response failure is needed; serum levels of inflammatory markers, such as CRP, ferritin, IL-6, and other cytokines, are increased in COVID-19. 13 In the setting of ongoing inflammation, evidence supports a role for ferritin in modulating the immune response, via its induction of anti-inflammatory cytokines and limitation of free radical damage. Alternatively, emerging work suggests a potential causative role of ferritin in the inflammatory pathology of disease. 14 Transcriptional induction of the CRP gene mainly occurs in hepatocytes in the liver in response to increased levels of inflammatory cytokines, especially IL-6. Similar to ferritin, evidence suggests that CRP is an important regulator of inflammatory processes and not just a marker. 15 Lymphopenia, the fourth marker to predict early MIV, is a common feature in patients with COVID-19 and is more pronounced in severe cases of infection. It affects mainly T cells, including CD4 Th1 and Tregs, but particularly CD8. Although circulating CD8 in patients with severe COVID-19 has exhibited phenotypes associated with abnormal functionality and exhaustion, CD4 cells have been shown to express activation markers. In addition, natural killer lymphocytes have decreased in patients with both moderate and severe cases of the disease. 9,16 Injured alveolar epithelial cells could lead to the infiltration of lymphocytes. 16

This study identifies 4 simple laboratory parameters that predict with 88% accuracy the probability of early MIV, enabling early intervention and optimization of healthcare resources. LM

COVID-19) weekly epidemiological update and weekly operational update

Clinical characteristics of 138 hospitalized patients with 2019 novel coronavirus-infected pneumonia in Wuhan, China

COVID-19 Lombardy ICU Network. Baseline characteristics and outcomes of 1591 patients infected with SARS-CoV-2 admitted to ICUs of the Lombardy region

Covid-19: most patients require mechanical ventilation in first 24 hours of critical care

Hematologic, biochemical and immune biomarker abnormalities associated with severe illness and mortality in coronavirus disease 2019 (COVID-19): a meta-analysis

Clinical predictors of mortality due to COVID-19 based on an analysis of data of 150 patients from Wuhan, China

UK. COVID-19: consider cytokine storm syndromes and immunosuppression

The hyperferritinemic syndrome: macrophage activation syndrome, Still's disease, septic shock and catastrophic antiphospholipid syndrome

Pathological inflammation in patients with COVID-19: a key role for monocytes and macrophages

Laboratory testing strategy recommendations for COVID-19: interim guidance. World Health Organization

Pilot prospective open, single-arm multicentre study on off-label use of tocilizumab in patients with severe COVID-19

Lactate dehydrogenase levels predict coronavirus disease 2019 (COVID-19) severity and mortality: a pooled analysis

Immune response, inflammation, and the clinical spectrum of COVID-19

Hyperferritinemia and inflammation

Role of C-reactive protein at sites of inflammation and infection

Longitudinal characteristics of lymphocyte responses and cytokine profiles in the peripheral blood of SARS-CoV-2 infected patients