Article Text
PDF
Abstract
Objectives To determine association of biomarkers—high-sensitivity C reactive protein (hsCRP), D-dimer, interleukin-6 (IL-6), lactic dehydrogenase (LDH), ferritin and neutrophil–lymphocyte ratio (NLR)—at hospitalisation with outcomes in COVID-19.
Design and Setting Tertiary-care hospital based prospective registry.
Participants Successive virologically confirmed patients with COVID-19 hospitalised from April 2020 to July 2021 were prospectively recruited. Details of clinical presentation, investigations, management and outcomes were obtained.
Primary and secondary outcome measures All biomarkers were divided into tertiles to determine associations with clinical features and outcomes. Primary outcome was all-cause deaths and secondary outcome was oxygen requirement, non-invasive and invasive ventilation, dialysis, duration of stay in ICU and hospital. Numerical data are presented in median and interquartile range (IQR 25–75). Univariate and multivariate (age, sex, risk factors, comorbidities, treatments) ORs and 95% CIs were calculated.
Results 3036 virologically confirmed patients with COVID-19 were detected and 1251 hospitalised. Men were 70.0%, aged >60 years 44.8%, hypertension 44.1%, diabetes 39.6% and cardiovascular disease 18.9%. Median symptom duration was 5 days (IQR 4–7) and oxygen saturation 95% (90%–97%). Total white cell count was 6.9×109/L (5.0–9.8), neutrophils 79.2% (68.1%–88.2%), lymphocytes 15.8% (8.7%–25.5%) and creatinine 0.93 mg/dL (0.78–1.22). Median (IQR) for biomarkers were hsCRP 6.9 mg/dL (2.2–18.9), D-dimer 464 ng/dL (201–982), IL-6 20.1 ng/dL (6.5–60.4), LDH 284 mg/dL (220–396) and ferritin 351 mg/dL (159–676). Oxygen support at admission was in 38.6%, subsequent non-invasive or invasive ventilatory support in 11.0% and 11.6%, and haemodialysis in 38 (3.1%). 173 (13.9%) patients died and 15 (1.2%) transferred to hospice care. For each biomarker, compared with the first, those in the second and third tertiles had more clinical and laboratory abnormalities, and oxygen, ventilatory and dialysis support. Multivariate-adjusted ORs (95% CI) for deaths in second and third versus first tertiles, respectively, were hsCRP 2.24 (1.11 to 4.50) and 12.56 (6.76 to 23.35); D-dimer 3.44 (1.59 to 7.44) and 14.42 (7.09 to 29.30); IL-6 2.56 (1.13 to 5.10) and 10.85 (5.82 to 20.22); ferritin 2.88 (1.49 to 5.58) and 8.19 (4.41 to 15.20); LDH 1.75 (0.81 to 3.75) and 9.29 (4.75 to 18.14); and NLR 3.47 (1.68 to 7.14) and 17.71 (9.12 to 34.39) (p<0.001).
Conclusion High levels of biomarkers—hsCRP, D-dimer, IL-6, LDH, ferritin and NLR—in COVID-19 are associated with more severe illness and higher in-hospital mortality. NLR, a widely available investigation, provides information similar to more expensive biomarkers.
- COVID-19
- Molecular diagnostics
- Adult intensive & critical care
- CHEMICAL PATHOLOGY
- Epidemiology
Data availability statement
All data relevant to the study are included in the article or uploaded as supplemental information. All the data relevant to the study are in the manuscript.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
STRENGTHS AND LIMITATIONS OF THIS STUDY
Prognostic evaluation of multiple biomarkers at admission in COVID-19.
Comprehensive clinical details available for all the patients.
We did not evaluate associations with biochemical markers of cardiac, renal, hepatic and gastrointestinal dysfunction.
Introduction
The pandemic of SARS-CoV-2 infection and COVID-19 continues unabated in many regions of the world.1 The current wave of the epidemic has been triggered by the omicron variants of COVID-19 although previous variants (ancestral, alpha, beta, gamma and delta) are also present in certain parts of the world.2 3 Studies have reported that adverse outcomes in the omicron wave are not as severe as the delta variant; a significant number of patients are being hospitalised and high transmission rate of these variants has resulted in more hospital admissions as compared with the previous waves in some countries.3–6 A number of prognostic markers have been identified to indicate severity of COVID-19.7 8 These include clinical markers such as tachypnoea, tachycardia and hypoxia, radiological abnormalities, and abnormalities of hepatic, renal and cardiac functions. Novel biomarkers that have been identified as important include interleukins (ILs) 6, 4 and 10, procalcitonin, C reactive protein (CRP), serum amyloid A, neutrophil, lymphocyte, monocyte and platelet counts, creatinine, alanine aminotransferase (ALT), aspartate aminotransferase (AST), lactic dehydrogenase (LDH), creatine kinase (CK), CK-MB isoenzyme, activated partial thromboplastin time (aPTT) and prothrombin time.7 Thrombotic biomarkers are also important and include markers of platelet activation, platelet aggregation, endothelial cell activation or injury, coagulation and fibrinolysis, and fibrinogen, D-dimer and aPTT.8 A systematic review and meta-analysis reported significant association of lymphopenia, thrombocytopenia and elevated levels of CRP, procalcitonin, LDH and D-dimer with adverse outcomes in COVID-19.9
Limited studies have evaluated the prognostic importance of novel biomarkers in developing countries where the burden of COVID-19 is the highest.10 Only small studies are available from developing countries including India.11–17 Measurement of laboratory biomarkers is expensive and not routinely available in most developing countries. We initiated a prospective COVID-19 registry at our hospital to evaluate disease pattern and outcomes.18 The present study aims to evaluate association of select novel biomarkers—high-sensitivity CRP (hsCRP), D-dimer, IL-6, LDH, ferritin and neutrophil–lymphocyte ratio (NLR) at hospital admission with clinical presentation, investigations and outcomes in COVID-19. We also evaluated whether NLR, a low-cost and widely available biomarker, is as important as others for prognostication.
Methods
We initiated a registry of all patients with virologically confirmed COVID-19 admitted to our hospital since April 2020.18
Setting: this is a 220-bed tertiary care hospital with major focus on critical care and cardiovascular sciences. It was designated an advanced COVID-19 care hospital by the Government of Rajasthan and more than 20% beds in general wards and intensive care units (ICUs) were initially reserved for patients with COVID-19. This proportion was later increased to 50% and 75% as the number of critically ill patients in the region increased. The hospital provided subsidised treatment to all the admitted patients according to the state government regulations. We developed a protocol for admission so that only those patients fulfilling definite clinical criteria were hospitalised.18 The case report form was updated and modified from that available at the regional COVID-19 care hospital at Health Sciences University.19
Patients: all patients presenting to medical and emergency departments with symptoms suggestive of upper respiratory infections were screened with reverse transcriptase-PCR (RT-PCR). Details of the sample collection and testing protocol have been reported.20 We obtained details of case history; vital monitoring for all patients was recorded and a flow chart of all in-hospital investigations maintained. Haematological investigations included complete blood counts performed using XS-1000i, a six-part fully automated analyser from Sysmex, USA. Biochemical investigations focused on measurement of blood glucose, renal function tests (urea, creatinine, electrolytes, uric acid) and liver function tests (bilirubin, AST, ALT, alkaline phosphatase, proteins, albumin, gamma glutaryl transferase) that were measured using Roche Cobas 6000, a fully automated molecular biochemistry and immunoassay analyser. COVID-19-specific biomarkers measured at admission were hsCRP, IL-6, LDH and ferritin using Roche Cobas 6000. D-dimer was measured using a fully automated coagulation analyser, ECL-760 ERBA and ACL Pro, from Instrumentation Laboratories, USA.
All the hospitalised patients received clinical management according to national and international protocols.18 Essentially, hydration and oxygenation were maintained using oral or intravenous fluids, and nasal cannula-based oxygen supplementation was provided as needed. Steroids were used only in patients needing non-invasive or invasive ventilatory support. Remdesivir was used according to international guidelines.21 22 Anti-IL drugs (tocilizumab, bevacizumab) were used infrequently, and non-evidence-based therapies such as oral hydroxychloroquine, ivermectin, anti-viral drugs (eg, favipiravir or ritonavir) or plasma therapy were not recommended for hospitalised patients.
Statistical analyses
All the data were computerised and entered in MS Excel worksheets. We focused on clinical history at presentation, admission haematological and biochemical investigations, radiological imaging-CT scan, medical therapies, oxygenation, ventilation and in-hospital outcomes. Long-term follow-up data are not yet available. Descriptive analyses have been performed using SPSS package (V.22). The categorical variables have been reported as numbers and per cent, while continuous variables are reported as medians and 25th–75th percentile IQR. As most of the biochemical variables had a skewed distribution, we used medians and IQR for descriptive statistics. Intergroup comparisons have been performed using χ2 test for categorical variables and Kruskal-Wallis test for non-parametric continuous variables. The biochemical variables (hsCRP, IL-6, LDH, D-dimer, ferritin) and haematological variables (NLR) were divided into tertiles, and clinical and other details tabulated accordingly. To identify correlation among various biomarkers, we performed both parametric (Pearson’s r) and non-parametric (Spearman’s r) analyses. Receiver operating characteristics (ROC) curves and area under the curve (AUC) were created for each of the biomarker for identification of sensitivity and specificity. To determine ORs and 95% CIs for deaths in second and third tertiles of various biomarkers versus first tertile, we initially performed a univariate logistic regression. Multivariate logistic regression was performed using variables likely to confound the outcomes such as age, sex, risk factors, comorbidities and medical treatments (steroids, remdesivir, tocilizumab, anticoagulants). We did not adjust for disease severity and mortality outcomes, as these were the primary focus in the present study. P values of <0.05 are considered significant.
Patient and public involvement
Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Results
In the successive waves of COVID-19 epidemic from April 2020 to July 2021, we evaluated 16 146 suspected patients with nasopharyngeal samples for SARS-CoV-2 antigen RT-PCR test. Of these, 3036 (18.8%) tested positive for the virus and 1251 (41.2%) were hospitalised. Data of these 1251 successively admitted patients have been obtained and analysed. Men were 876 (70.0%), aged >60 years were 560 (44.8%) and prevalence of important comorbidities were in: hypertension 551 (44.1%), diabetes 495 (39.6%), cardiovascular disease 241 (18.9%), and asthma or chronic obstructive pulmonary disease 72 (2.7%). Median symptom duration was 5 days (IQR 4–7) and oxygen saturation (SpO2) at admission was 95% (90%–97%). Total white cell count was 6.9×109/L (5.0–9.8), neutrophils 79.2% (68.1–88.2), lymphocytes 15.8% (8.7%–25.5%) and platelets 2.2×109/L (1.7–2.8). At admission, median creatinine level was 0.93 mg/dL (0.78–1.22), serum creatinine >1.2 mg/dL (laboratory upper limits of normal) was in 312 (25.4%) and >2.0 mg/dL in 104 (8.5%). Levels of various biomarkers were hsCRP 6.9 mg/dL (2.2–18.9), D-dimer 464 ng/dL (201–982), IL-6 20.1 ng/dL (6.5–60.4), LDH 284 mg/dL (220–396) and ferritin 351 mg/dL (159–676). Other details including liver functions tests and radiography have been reported earlier.18 Oxygen support at admission was in 480 (38.6%), and during hospitalisation, nasal cannula-based oxygen support was provided in 357 (28.7%), non-rebreather masks in 139 (11.2%), non-invasive ventilatory support in 137 (11.0%) and invasive ventilation in 144 (11.6%). Prone positioning (proning) was performed in 676 (54.3%) patients. All the patients received antibiotics, and clinically and microbiologically confirmed secondary infection developed in 133 (10.7%) patients. Although acute kidney injury was observed in 104 (8.5%), incidence of severely deranged renal function requiring haemodialysis was in 38 (3.1%). Of the total hospitalised patients, 1055 (84.9%) were discharged to home-based care, 15 (1.2%) transferred to hospice care and 173 (13.9%) died.
The patients have been divided into tertiles of various biomarkers—hsCRP (table 1, n=1130), D-dimer (table 2, n=961), IL-6 (table 3, n=1038), LDH (table 4, n=721), ferritin (table 5, n=998) and NLR (table 6, n=1214). Details of demography, risk factors, comorbid conditions, clinical features and investigations, medical management and outcomes for each of the biomarkers are provided in tables 1–6.
High-sensitivity C reactive protein tertiles, clinical features and outcomes
D-dimer tertiles, clinical features and outcomes
Interleukin-6 tertiles, clinical features and outcomes
Lactic dehydrogenase tertiles, clinical features and outcomes
Ferritin tertiles, clinical features and outcomes
Neutrophil:lymphocyte ratio tertiles, clinical features and outcomes
For each biomarker, those in the highest tertile have the most adverse clinical characteristics and significantly greater levels of clinical abnormalities—fever, shortness of breath, hypoxia, leucocytosis, lymphopenia—and higher levels of other biomarkers. Use of steroids, anticoagulants, remdesivir and anti-inflammatory drugs as well as proning, oxygenation and non-invasive and invasive ventilatory support is also greater in those in the second and third tertiles (tables 1–6). Evidence of bacteriologically confirmed secondary infection requiring escalation of antibiotics was observed in the highest tertile groups for all the biomarkers. Acute kidney injury advance disease requiring haemodialysis was low (3.1%) and was significantly more in the highest tertile groups only for D-dimer, IL-6 and ferritin biomarkers.
There is a strong correlation among various biomarkers (table 7). Results of parametric (Pearson’s) and non-parametric (Spearman’s) correlation among various biomarkers show significant intercorrelation among all the biomarkers (p<0.001). Non-parametric analysis shows greater correlation (r) of hsCRP with IL-6 (0.42), LDH (0.54), ferritin (0.42) and NLR (0.48), and NLR with hsCRP (0.48), LDH (0.48) and ferritin (0.38) (table 7). ROC curves are shown in figure 1. The AUC statistics were significant for all the biomarkers with AUC of >0.70, which is considered as of fair instrumental sensitivity for true positive outcomes. The highest AUC ±1 SD was observed for hsCRP (0.793±0.032) followed by NLR (0.778±0.031), LDH (0.773±0.032), D-dimer (0.752±0.034), ferritin (0.735±0.032) and IL-6 ((0.072±0.034).
Receiver operating characteristics curve shows that area under the curve for each parameter was more than 0.7 (range 0.720–0.793, p<0.001). hsCRP, high-sensitivity C reactive protein; IL-6, interleukin-6; LDH, lactic dehydrogenase.
Correlation matrix (parametric Pearson’s r and non-parametric Spearman’s r) among the biomarkers evaluated
Incidence of deaths in various biomarker groups is shown in figure 2. For all the biomarkers, incidence of death is the highest in the third tertile. Univariate, age and sex, multivariate (age, sex, risk factors, comorbidities) and multivariate and treatment-adjusted ORs (95% CI) for deaths in second and third versus first tertiles are in table 8. As compared with the lowest tertile (reference group, T1), deaths were significantly greater in second and third tertiles for all the biomarkers (figure 3). Multivariate-adjusted ORs (95% CI) in second and third tertiles, compared with the first, respectively, were for hsCRP 2.24 (1.11 to 4.50) and 12.56 (6.76 to 23.35); D-dimer 3.44 (1.59 to 7.44) and 14.42 (7.09 to 29.30); IL-6 2.56 (1.13 to 5.10) and 10.85 (5.82 to 20.22); ferritin 2.88 (1.49 to 5.58) and 8.19 (4.41 to 15.20); LDH 1.75 (0.81 to 3.75) and 9.29 (4.75 to 18.14); and NLR 3.47 (1.68 to 7.14) and 17.71 (9.12 to 34.39) (table 8). We also calculated multivariate ORs (95% CI) for comparison of second and third tertiles (table 8). Compared with the second tertile, the ORs were significantly greater in the third tertile for all the biomarkers—hsCRP 5.60 (3.57 to 8.78), D-dimer 4.20 (2.64 to 6.68), IL-6 4.23 (2.72 to 6.58), ferritin 2.84 (1.86 to 4.33), LDH 5.32 (3.12 to 9.05) as well as NLR 5.14 (3.40 to 7.78).
Univariate, age and sex, multivariate (age, sex, risk factors, comorbidities) and multivariate (plus treatments)-adjusted OR for association of various biomarkers with in-hospital deaths among second and third tertiles compared with the first
In-hospital deaths (%) in various biomarker tertiles. hsCRP, high-sensitivity C reactive protein.
Multivariate (age, sex, risk factors, comorbidity, treatments)-adjusted ORs and 95% CIs for association of various biomarkers with in-hospital mortality. T1 (reference), T2 and T3 are the respective tertiles. hsCRP, high-sensitivity C reactive protein.
Discussion
This study shows that among consecutive patients hospitalised with COVID-19, increasing levels of biomarkers: hsCRP, D-dimer, IL-6, LDH, ferritin and high NLR are associated with greater illness severity, oxygenation, non-invasive and invasive ventilation and exponentially higher in-hospital mortality. We also show that NLR, a simple, widely available and inexpensive investigation, provides prognostic information that is similar to the more expensive biomarkers.
A meta-analysis that evaluated influence of biomarkers on poor outcomes (measured as SpO2 <90%, invasive mechanical ventilation, severe disease, ICU admission and mortality) among hospitalised patients with COVID-19 from December 2019 to August 2020 included 32 studies with 10 491 patients.9 Biomarkers included were lymphocyte, platelets, D-dimer, LDH, CRP, AST, ALT, creatinine, procalcitonin and CK. This study reported a significant association between lymphopenia, thrombocytopenia and elevated levels of CRP, procalcitonin, LDH, D-dimer with COVID-19 severity with ORs varying from 3.33 (2.51 to 4.41) for lymphopenia, 4.37 (3.37 to 5.68) for elevated CRP and 5.48 (3.89 to 7.71) for LDH. We used mortality as the endpoint and the results of our study are not directly comparable; however, the ORs are similar when we compared tertile 2 with tertile 3 in our study (table 8). Another meta-analysis reviewed impact of biomarkers of endothelial dysfunction (von Willebrand factor, tissue type plasminogen activator, soluble thrombomodulin, plasminogen activator inhibitor-1) and reported significant association with poor outcomes among 1187 patients from 17 studies.23 We did not evaluate these biomarkers and cannot directly compare our results; however, significance of D-dimer, another marker of coagulation and vascular dysfunction, in the present study suggests importance of coagulation markers in COVID-19. A more recent review reported that biomarkers useful for risk prediction in COVID-19 include several proinflammatory cytokines, neuron-specific enolase, LDH, AST, neutrophil count, NLR, troponins, CK-MB, myoglobin, D-dimer, brain natriuretic peptide and N-terminal pro-hormone. Some of these biomarkers can be readily used to predict disease severity, hospitalisation, ICU admission and mortality, while markers of metabolomic and proteomic analysis have not yet been translated to clinical practice.24 The International COVID-19 Thrombosis Biomarkers Colloquium and international agencies recommend routine use of some of these biomarkers for prognostication.8 22 Our study shows a strong correlation among various biomarkers (table 7) and suggests that clinical use of any one or two of these biomarkers may be used for prognostication.
Our study also shows that a widely available biomarker—NLR—is as predictive as other biomarkers. This is similar to previous studies,9 24 although the magnitude of risk is much higher as compared with them (figure 2). The relative risk is similar to previous studies on comparison of second and third tertiles (table 8). Regolo et al25 compared NLR with CRP in 411 elderly patients with COVID-19; 33% of patients died during hospitalisation. When divided into tertiles according to NLR values, it was observed that NLR was a better predictor of mortality than CRP with largest AUC (0.772) and high specificity (71.9%) and sensitivity (72.9%). Other studies, considerably smaller than ours, have also identified the role of NLR in identification of severe disease with similar AUC on ROC analyses (figure 1).26 27 Reviews have concluded that this is an important marker,7 8 23 28 and the present study, which is larger than most of the previous studies, confirms this observation. This finding is important in the context of India and low and lower middle-income countries as the cost of this test is minuscule compared with the biomarkers evaluated in the present study.
Our study has limitations and strengths, apart from those mentioned above. This is one of the larger studies that has evaluated multiple biomarkers in mild, moderate and severe COVID-19 cases and shows an exponential increase in deaths with rising levels of hsCRP, D-dimer, IL-6, LDH, ferritin and NLR, and the strength of association is maintained after multivariate adjustment (figure 3). This is also one of the larger studies from India—a country with one of the highest deaths from COVID-19.10 29 30 We did not assess well-known markers of COVID-19 severity such as serum creatinine and other markers of renal dysfunction, liver enzymes (transaminases, etc), markers of cardiac damage (troponins, myocardial-bound CK) or pulmonary involvement in the present study, and this is a study limitation. Other limitations include lack of assessment of novel biomarkers of coagulation and endothelial dysfunction.8 9 23 24 We also do not have long-term data in these patients and cannot comment on importance of biomarkers in incidence of post-acute syndrome of COVID-19. Only few studies have addressed this question.31 32
In conclusion, the present study in hospitalised patients with COVID-19 shows that higher levels of multiple biomarkers—hsCRP, D-dimer, IL-6, LDH, ferritin and NLR—at admission are associated with greater illness severity and significantly higher in-hospital mortality. The study also shows that NLR, a universally available investigation, provides prognostic information similar to the less available biomarkers. Our study is important in the context of occurrence of multiple waves of COVID-19 in many developed and developing countries.1–3 33 34 We demonstrate that a simple investigation—NLR—can provide important prognostic information and can inexpensively triage patients presenting to primary care into low risk—who can be advised home-based care—and the intermediate and high-risk groups to more intensive care settings.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplemental information. All the data relevant to the study are in the manuscript.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and was approved by the ethics committee of Eternal Heart Care Centre and Research Institute, Jaipur, India (Government of India registration, CDSCO No. ECR/615/Inst/RJ/2014/RR-20) before initiation of the study. Informed consent from all the patients or next of kin was obtained for anonymised data publication.
References
Footnotes
Twitter @rajeevgg
Contributors RSK and RG initiated the project. KM, HCA, JBG, YS and AM collected the data. RG and KS performed all statistical analyses. RG had main responsibility of writing the article. SSingh and SSharma performed the biochemical and pathological investigations. RSK, RG and KS contributed to the structure and content of the manuscript, and all authors have read and approved the final draft. RSK acts as the guarantor.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.