Clinical and Laboratory Factors in Predicting Mortality Among COVID-19 RT-PCR Positive Patients: A Retrospective Observational Study From a Tertiary Care Center

Background: In coronavirus disease 2019 (COVID-19) patients, risk stratification based on clinical presentation, co-morbid illness, and combined laboratory parameters is essential to provide an adequate, timely intervention based on an individual’s conditions to prevent mortality among cases. Methods: A retrospective observational study was carried out from June to October 2020, including all reverse transcription-polymerase chain reaction (RT-PCR) positive COVID-19 non-survivors and control group survivors randomly selected after age and sex matching. Clinical and demographic information was collected from the medical records. Categorical variables were expressed by frequency and percentage. To explore the risk factors associated with mortality, univariable and multivariable logistic regression models were used. Results and discussions: All non-survivors (n = 100) and 100 survivors (out of 1,018) were analyzed. Male gender (67.4%) was the independent risk factor for COVID-19 infection. Advanced age group, diabetes, cardiovascular, neurological, and hypertensive co-morbidities were statistically associated with mortality. Cardiac arrest and acute kidney injury (AKI) were the most common complications. Mortality is significantly associated with lymphopenia and raised lactate dehydrogenase (LDH), as shown by higher odds. In addition, raised neutrophils, monocytes, aspartate aminotransferase (AST), serum creatinine, interleukin 6 (IL-6), and C-reactive protein (CRP) are also significantly associated with mortality. The most common causes of death were respiratory failure (84%) and acute respiratory distress syndrome (77%). Of the non-survivors, 92% received corticosteroids, 63% were on high-flow nasal cannula oxygen therapy, 29% were mechanically ventilated, and 29% received tocilizumab. Conclusion: Serial monitoring of neutrophils, lymphocytes, D-dimer, procalcitonin, AST, LDH, CRP, IL-6, serum creatinine, and albumin might provide a reliable and convenient method for classifying and predicting the severity and outcomes of patients with COVID-19.


Introduction
The coronavirus disease 2019 (COVID-19), which is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), originated in December 2019 in Wuhan, China. It has ever since rapidly spread worldwide causing morbidity and mortality in its way. In March 2020, the World Health Organization (WHO) declared the COVID-19 outbreak as a pandemic [1]. Though COVID-19 primarily affects the respiratory system, the clinical presentation of COVID-19 shows significant heterogeneity, ranging from asymptomatic, mild, moderate, and severe disease with multi-organ dysfunction leading to death. Therefore, risk stratification is essential to formulate better treatment plans, either to admit in hospital or community isolation or home quarantine. The most severe cases of COVID-19 are often due to respiratory failure, which often requires mechanical ventilation. This in turn leads to a higher mortality rate. In a recent study, it was seen that the mortality was 40.8% among patients receiving invasive mechanical ventilation, 39% among patients receiving extracorporeal membrane oxygenation (ECMO), and 71.6% for patients on invasive mechanical ventilation, vasoactive drugs, and new renal replacement therapy [2]. The fatality rate was very high among the severe or critically ill patients (49%) when compared to the overall (2.3%) case fatality as per the epidemiological data from the Centers for Disease Control and Prevention [3]. Some studies have reported that among COVID-19 patients, the older age group with comorbidities has been associated with a poor prognosis [4].
The risk/prognosis/mortality predictor based on clinical presentation, co-morbid illness, combined laboratory parameters (biochemical, hematological, and inflammatory markers), and imaging will play a significant role in understanding the disease and providing an adequate, timely intervention based on the individual's conditions [5]. Due to the heterogeneous nature of COVID-19, our study was planned to analyze clinical presentation with comorbidities and laboratory factors in predicting mortality among COVID-19 reverse transcription-polymerase chain reaction (RT-PCR) positive patients.

Materials And Methods
This retrospective observational study was carried out in our tertiary care center after obtaining institutional ethical committee clearance. The study included adult patients (>18 years) admitted into the COVID-19 isolation unit of the hospital with positive RT-PCR who died due to COVID-19 between June and October 2020. During the same period, patients admitted with COVID-19 with positive RT-PCR who were discharged/recovered (survivors) were randomly selected and included as the control group after age and sex matching.
Information regarding demographic details, clinical presentations, co-morbid conditions, and biochemical (renal, liver function tests, lactate dehydrogenase, D-dimer, interleukin 6 [IL-6], serum ferritin, fibrinogen), hematological, including coagulation profile, and inflammatory markers (C-reactive protein [CRP], procalcitonin) along with treatment details and duration of stay in the hospital for both groups were collected from the medical records using a standard data collection form, which was verified by other researchers for any difference.

Definitions
Confirmed cases of COVID-19, acute respiratory distress syndrome (ARDS), sepsis, and septic shock are defined by the Revised Guidelines on Clinical Management of COVID-19, Government of India [6]. Acute kidney injury was diagnosed according to the Kidney Disease Improving Global Outcomes (KDIGO) clinical practice guidelines [7].

Statistical analysis
Our study is an age and sex-matched case-control study with 100 cases and 100 controls. Data analyses were done using STATA software, version 16 (StataCorp LLC, College Station, Texas). The descriptive results for the categorical variables were displayed by frequency and percentage. For continuous variables, median and ranges were used. To explore the risk factors associated with in-hospital death, univariable and multivariable logistic regression models were used.
As this was a matched case-control study using matching individuals, conditional logistic regression was used to estimate the association between the predictors of interest with survivors and non-survivors of COVID-19 within each matched set of cases and controls, and the level of significance was set at a twotailed p-value of less than 0.05.

Results
A total of 1,118 patients were admitted into the COVID-19 isolation unit of our hospital during the study period. Statistically, a significant difference was found in the age group of patients (p < 0.0001), and no difference was found in gender analysis (p = 0.0686) among survivors and non-survivors (

TABLE 3: Analysis of comorbidities among non-survivors and survivors.
Analysis of complications among non-survivors and survivors showed cardiac arrest (p = 0.0022) and acute kidney injury (p < 0.0001) were more with the non-survivors ( Figure 1).

FIGURE 1: Complications among non-survivors and survivors.
Analysis of complete blood count and coagulation profile among non-survivors and survivors is given in Table 4. There was a statistically significant difference in the total count, polymorphs, lymphocytes, and monocytes among the cases and controls. Also, absolute neutrophil and absolute lymphocyte count showed a statistically significant difference between non-survivors and survivors. No statistical difference was found in the analysis of the coagulation profile of prothrombin time (PT) and activated partial thromboplastin time (APTT).

Variables
Non-survivors (n = 100) Survivors (n = 100) p-value  Analysis of enzymes, interleukins, sepsis markers, and other coagulation profiles among cases and controls is given in Table 5. There was a statistically significant difference in the blood urea, lactate dehydrogenase (LDH), aspartate aminotransferase (AST), IL-6, and quantitative CRP among the cases and controls. Other parameters did not show any statistical difference.

Variables
Non-survivors (n = 100) Survivors (n = 100) p-value   In univariable analysis, the odds of in-hospital mortality were higher in patients with diabetes, hypertension, renal disorder, and coronary heart disease. Neutrophilia, lymphopenia, monocytosis, elevated AST, LDH, and serum ferritin were also associated with increased mortality. In multivariate analysis, lymphopenia and elevated LDH were associated with death (  Among non-survivors, biochemical and hematological markers taken at the time of admission and 24 hours before death showed progressive high total WBC count, neutrophil count, lymphocytopenia, low serum albumin, elevated ALT, blood urea, raised LDH, D-dimer, ferritin, and procalcitonin levels and are significantly associated with mortality ( Table 7).  The most common causes of death were respiratory failure (84%) and ARDS (77%), followed by cardiac arrest (10%) and septic shock (4%). Analysis of treatment given among non-survivors and survivors showed 92% of non-survivors got corticosteroids, 63% needed high-flow nasal cannula oxygen therapy, 29% had invasive mechanical ventilation, and 29% received tocilizumab with a statistically significant difference of p < 0.0001 ( Figure 2).

Non-survivors mean value at admission
ischemia, and weakness. Among the seven patients, four of them presented with acute infarct, two of them with subarachnoid hemorrhage, and one patient had bilateral cerebellar infarct with hemorrhage. All had elevated D-dimer concentrations (>500 ng/ml). Thus, neurological manifestations of COVID-19 infection are not uncommon. Severe neurological complications are either because of direct viral invasion, immunological reaction, or hypoxic metabolic changes as evidenced by Garg et al. [14]. Coagulopathies enhance the risk of cerebral arterial and venous thrombosis in COVID-19. Li et al., in a retrospective study, noted that out of 221 patients, 11 had an acute ischemic stroke [15].
Although acute pancreatitis is a relatively common disease, its occurrence in patients with COVID-19 seems to be rare, and many questions remain unanswered [16]. As of date, there is no evidence for an association between COVID-19 and acute pancreatitis and it is unclear if pancreatitis might be caused by direct viral damage to pancreatic cells or endothelium, or thrombosis and ischemic pancreatitis. In our study, there was one survivor with chronic pancreatitis who had presented with acute illness.
The mortality and severity of complications in COVID-19 patients were significantly associated with liver dysfunction as evidenced by Wu et al. [17]. In our study, the degree of transaminitis (elevated AST) was substantially more in the mortality group compared to the survivor group (p < 0.001).
As in other studies, there was a marked rise in innate immune response, decreased adaptive immune response, an increase of markers of tissue damage, inflammation, and organ failure.
The current study identified many laboratory markers as risk factors of death in adults. Especially, lymphopenia and raised LDH were associated with higher odds of in-hospital death. In addition, elevated levels of neutrophils, monocytes, AST, and serum creatinine were also significantly associated with increased mortality.
Four patients succumbed to sepsis. Procalcitonin level was below 0.5 ng/ml in 58.2% patients, 0.5-2 ng/ml in 8% patients, 2-10 ng/ml in 3% patients, and more than 10 ng/ml in one deceased patient. Bacterial pathogens were not detected in any of these individuals. These factors conveniently favor that SARS-CoV-2 directly leads to sepsis. This needs further research.
Of the non-survivors, 92% received a combination of low molecular weight heparin, azithromycin, favipiravir, and steroids. Of the non-survivors, 63% were managed with noninvasive ventilation, and 29% of them were mechanically intubated and received tocilizumab. One deceased patient received ECMO therapy. Among survivors, all of them received a combination of azithromycin, favipiravir, and steroids, and 16% of them were managed with noninvasive ventilation.
Our study has several limitations. Due to the retrospective study design, not all laboratory markers were available for all the patients. So the role of important markers like CRP, ferritin, fibrinogen, D-dimer, PT, APTT, serum amylase, and troponin I in contributing to mortality could be underestimated. Although none of the drugs are routinely recommended for COVID-19 pneumonia, a combination of antibiotics, antiviral, and steroids was given to all critically ill patients in the current study. So poor adherence to standard supportive care and inadvertent use of a high dose of steroids also could have contributed to mortality. During the peak of the pandemic, many patients were admitted late in their course of illness, which would have altered the outcome.

Conclusions
In conclusion, serial monitoring of hematological and coagulation parameters, such as neutrophils, lymphocytes, and D-dimer, and inflammatory and tissue damage markers such as procalcitonin, AST, LDH, serum creatinine, and albumin might provide a reliable and convenient method for classifying and predicting the severity and outcomes of patients with COVID-19.

Additional Information
Disclosures