Risk Factors associated with COVID-19 Patients in India: A Single Center Retrospective Cohort Study

Abstract

Background and objectives: The coronavirus disease 2019 (COVID-19) outbreak has caused a worldwide pandemic, resulting in >3.8 million deaths. Our aim is to identify the risk factors associated with in-hospital mortality using survival analysis considering the characteristics and outcomes of COVID-19 patients admitted to a dedicated tertiary-care hospital in Mumbai, India.

Materials and methods: In a retrospective cohort study, 565 patients admitted from 28th March 2020 to 30th June 2020 were enrolled, and a follow-up was conducted till August 2020. To investigate the impact of COVID-19, survival analysis was performed using the Kaplan–Meier method. Potential risk factors associated with mortality were analyzed using logistic regression models for multivariate analysis and the Cox proportional hazards model for estimating hazard ratios (HRs).

Results: From the 565 positive COVID-19 cases, 49 patients died (8.7%) and 516 (91.3%) were discharged. Overall, 119 patients (20%) required intensive care unit (ICU) admission, of which 70 (58%) patients survived. The Kaplan–Meier survival curve showed a significant association of COVID-19 infection with age (≥60; p = 0.008), hypertension (p = 0.03), dialysis (p = 0.0001), lung commodities (p = 0.01), breathlessness (p = 0.0001), severe disease upon high-resolution computed tomography (HRCT) analysis (p = 0.0001), ICU admission (p = 0.0001), and low lymphocyte count at admission (p = 0.0001). Additionally, patients receiving tocilizumab (p = 0.0001) and deprived of hydroxychloroquine (HCQ) + azithromycin (azee) (p = 0.0001) were estimated at a high risk of mortality.

Interpretation and conclusion: Coronavirus disease 2019 (COVID-19) increased the risk of mortality in patients with increased age, comorbidities, and severe symptoms upon treatment with an immunosuppressant (tocilizumab). However, patients treated with HCQ + azee showed favorable results due to their antiviral effects in vitro.

Introduction

The World Health Organisation (WHO) declared COVID-19 as a global pandemic on 11th March 2020. The disease is caused by severe acute respiratory syndrome coronavirus 2 (SARS-Cov-2). China was the first country to be infected with the virus, and periodic testing and proper quarantine measures were imposed to contain the virus. Being highly contagious in <7 months, the virus spread to 214 other countries, causing the pandemic.1,2

The clinical presentation of COVID-19 is highly heterogeneous, ranging from asymptomatic to severe pneumonia with respiratory failure that could lead to acute respiratory distress syndrome or death. According to the WHO, 85% of infections cause a mild or moderate illness. In the remaining 10–15%, patients develop severe symptoms, which require hospitalization, and only about 5% of cases require intensive care. Recovery is dependent on the severity of the infection. For instance, patients with mild symptoms recover in 2 weeks, whereas those with severe symptoms recover in

3–6 weeks.3 This time to the events, here recovery is considered more critical as it can help establish the probability of the outcome and facilitate modification in treatment, if required, to achieve a more favorable result during a considerable duration of hospital stay. This time to recovery is usually estimated using a Kaplan–Meier survival curve, where the Y-axis indicates the survival and the X-axis denotes the time trend.4 Moreover, it can help to determine the duration of the hospital stay, availability of hospital beds in critical care, and recovery time of COVID-19 patients.5

Although a deeper understanding of the nature of the disease is acquired through literature, no definitive treatment has yet been proven to be effective in preventing the disease progression and the death of critical patients. This study, thus, aims to identify the risk factors for early warning and intervention using survival analysis by retrospectively analyzing the characteristics and outcomes of patients associated with mortality among COVID-19 patients admitted to Seven Hills Hospital Reliance facility, a dedicated COVID-19 care hospital in Mumbai, Maharashtra, India.

Material and Methods

The study was conducted at Seven Hills Hospital Reliance facility, Mumbai, Maharashtra, India, under the approval of the ethics committee at Sir HN Reliance Foundation Hospital (RFH) and Research Centre.

Study Design

It is a retrospective, observational analytical study without any control group for COVID19-positive cases. COVID-19 diagnosis was performed as per the WHO interim guidelines.6 Since the study did not involve any intervention in patient care, the requirement for written informed consent was waived by the Institutional Ethics Committee (IEC).

Study Population

The study population included all adult patients aged 18 years and above with reverse transcriptase quantitative polymerase chain reaction (RT-qPCR)—confirmed SARS-Cov-2 infection admitted to the hospital within the study period of March–June 2020. The diagnosis was confirmed by a positive result of RT-qPCR assay from nasal and pharyngeal swab specimens. Only those patients who were admitted for >24 hours were included in this study. Patients transferred to other facilities for further care were excluded.

Design and Data Extraction

Data including epidemiological history, demographics, clinical symptoms and signs, comorbidities, radiological assessments, laborator y f indings upon admission, treatments, and clinical outcome were extracted from electronic medical records. The primary endpoint of the outcome was time to death.

Study Variables

Data were extracted from the electronic medical record system. Baseline demographic data, arterial oxygen pressure (PaO2)/fraction of inspired oxygen (FiO2) ratio, sequential organ failure assessment (SOFA) score comorbidities, medication history, category of COVID-19 disease—mild, moderate, and severe decided by the task force committee of the hospital (Table  1), and ventilation characteristics (invasive and noninvasive ventilation) of COVID-19 patients were recorded. Clinical characteristics such as time from the onset of symptoms to hospitalization, symptoms on presentation, radiological investigation using HRCT and chest X-ray findings were also recorded in addition to abnormal laboratory parameters (low lymphocyte count) and inflammatory marker levels [serum C-reactive protein (CRP; <0.5 mg/dL), D-dimer (0–250 ng/mL), ferritin (30–400 ng/mL), lactate dehydrogenase (LDH; ≤250 U/L), and interleukin (IL- 6; 0–7 pg/mL)]. Treatment therapies as per Sir HN RFH COVID-19 treatment protocol and overall complications during hospitalization were also noted. Clinical outcomes, including in-hospital mortality (defined as the percentage of patients with COVID-19 who died in the hospital), ICU admission, use of noninvasive or mechanical ventilation, total hospital length of stay, and ICU length of stay, were all considered while performing data analysis.

Statistical Analysis

Data was entered in MS Excel (Microsoft©, USA) and converted to Stata version 15.1 ( ©Stata Corp, College Station, Texas, United States of America). The mean and standard deviation (SD) or median and interquartile range (IQR) were estimated for all continuous variables. The proportions were identified for categorical variables.

Mean values across groups were compared using the Independent Samples t-test, and median values were compared using the Mann–Whitney U test. The proportions were compared using the Chi-squared test or Fisher’s exact test for low expected cell counts. We then estimated the mortality rate among these patients (per 100 person-days), and this data was then used to estimate the survival probability of COVID-19 patients on days 3, 7, 14, 21, and 28 via the Kaplan–Meier method. Both these parameters were estimated for the whole study population using clinical and demographic characteristics as well. The difference between the two survival curves was determined using the log-rank test. We then created a multiple regression model using Cox proportional hazards regression modelling to estimate the risk factors associated with COVID-19 mortality. We also included demographic and clinical variables in the model. p-values of <0.05 were considered statistically significant.

Results

The mean (SD) age of the patients (N = 565) was estimated to be 51.6 (16.4) years. Of these, 368 (65.3%) were men and 196 (34.8%) were women.

Out of 565 patients, 516 (91.3%) patients were discharged and 49 (8.7%) patients died. Overall, 119 (20%) patients required ICU admission, of which 70 (58%) patients survived. Patients required noninvasive or mechanical ventilation for an average duration of 6 or 8 days, respectively. The average total hospital length of stay was 11 days, whereas the ICU length of stay was 9 days. On admission to the ICU, patients with a PaO2:FiO2 ratio of <200 and a SOFA score of >6 were at a high risk of mortality (41%).

The general mortality rate was estimated to be 0.81/100 person-days [95% confidence interval (CI): 0.61, 1.07/100 person-days], 0.96/100 person-days for men, and 0.50/100 person-days for women. Upon comparing the characteristics of survived and deceased, the deceased was older than or equal to 60 years (17.4%; p < 0.001) and were mostly men [10.6 vs 5.1% (men vs women); p = 0.028]. The most prevalent comorbidities were determined to be hypertension, diabetes mellitus, heart disease, thyroid and neurological diseases, chronic kidney disease, and lung disease (Table 2).

Clinical Characteristics

Symptoms

In total, 565 patients presented COVID-19 symptoms, which included constitutional symptoms (fever, rhinorrhea, sore throat, myalgia, fatigue, and anosmia), respiratory and cardiac symptoms (breathlessness, cough, and chest pain), abdominal symptoms (diarrhea, vomiting, and abdominal pain), and neurological symptoms (headache, altered sensorium, hemiplegia, convulsions, and coma) (Table  3). The median duration from the onset of symptoms to hospital admission was 5 (3/7) days. The duration from the onset of symptoms to death and that from hospital admission to death were 15 (11/25) and 10 (5/18) days, respectively.

Complications

The most common complication associated with COVID-19 was found to be sepsis [30 (5.3%)], followed by acute kidney injury [28 (5%)]. Other complications included thrombosis [19 (3.4%)], followed by cardiac [17 (3%)] and respiratory complications such as acute lung injury, pulmonary hemorrhage, and barotrauma [16 (3%)].

Clinical Outcomes

Association of Comorbidities and Severity with Mortality

Mortality among all COVID-19 patients was significantly associated with increased age (p = 0.001), hypertension (14.3 vs 5.4%; <0.001), lung disease (p = 0.03), and in patients with chronic kidney disease and are on dialysis (46.7 vs 5.4%; p < 0.001). Furthermore, a significant association of mortality was observed with male gender (p = 0.028) and diabetes mellitus (p = 0.02) (Table 2).

Association of Symptoms and Disease Severity with Mortality

Based on the severity of the disease, mortality was found to be significant in patients with severe COVID-19 symptoms (59 vs  Based on the severity of the disease, mortality was found to be significant in patients with severe COVID-19 symptoms (59 vs 0%; p < 0.001). Other symptoms, such as breathlessness (p < 0.001) and seizure (p = 0.048), were found to be significantly associated with mortality (Table  3). Patients requiring ICU admission (41.1 vs 0%; p < 0.001) were found to be at an increased risk of mortality.

Association of Laboratory and Radiological Investigations with Mortality

Out of 565 patients, 205 (37.1%) patients had low lymphocyte on admission, of which 38 (18.5%) patients died. Low lymphocyte count (p = 0.001) and severe disease on HRCT analysis (p = 0.001) were significantly associated with death. Among 49 patients who died in the hospital, HRCT was performed on 35 patients and severe disease was identified in them. In addition, high levels of CRP, LDH, ferritin, and IL6 were detected in COVID-19 patients; however, these parameters were not found to be significantly associated with mortality (Table 4).

The Kaplan–Meier survival curve showed a significant association of COVID-19 infection with age (≥60; p = 0.008) (Fig. 1), hypertension (p = 0.03) (Fig. 2), dialysis (p = 0.0001) (Fig. 3), lung comorbidities (p = 0.01) (Fig. 4), breathlessness (p = 0.0001) (Fig. 5), severe disease upon HRCT analysis (p = 0.0001) (Fig. 6), ICU admission (p = 0.0001) (Fig. 7), and lymphocyte count at admission (p = 0.0001) (Fig. 8). Cox proportional HRs demonstrated that a significant increase in mortality risk was associated with hypertension (HR 2.56; 95% CI 1.04–6.29), whereas patients with normal lymphocyte count were at low risk of mortality with HR 0.28 (95% CI 0.13–0.62, p = 0.001).

Association of Therapies with Mortality

Based on the Indian Council of Medical Research guidelines and RFH COVID-19 treatment protocol, all patients were treated with HCQ and azee on admission. However, as 52 patients were contraindicated for HCQ, 513 patients were treated with HCQ and azee, of which 478 (93.2%) patients survived. Of these 513 patients, a 100% survival rate was observed in asymptomatic (N = 24), mild (N = 358), and moderately symptomatic (N = 77) patients. However, out of 65 patients with severe COVID-19 infection, only 29 (45%) patients survived. Of the 36 patients who did not survive, the average time from the onset of illness to ICU admission was 8 days. Of these 36 patients, 14 patients were transferred from outside the hospital, patients with severe COVID-19 infection were direct admissions, and 15 patients were transferred from wards.

Patients with cytokine storm who were treated with tocilizumab were observed to have significantly higher mortality rates (47.7 vs 3.6%; p < 0.001), whereas treatment with HCQ and azee showed a low risk of mortality (6.8% vs 26.9%; p < 0.001) (Table 4).

The Kaplan–Meier survival curve results showed similar findings wherein mortality due to COVID-19 was significantly associated with patients not receiving HCQ + azee (p = 0.0001) (Fig. 9) and in patients receiving tocilizumab (p = 0.0001) (Fig. 10). Treatment with HCQ + azee was identified as a predictor of good prognosis because patients treated with this combination of drugs were at low risk of mortality (Cox proportional HR = 0.27, 95% CI 0.09–0.83, p = 0.02).

Patients with moderate to severe symptoms were treated in an awake prone position along  with the medications. Of the 139 patients, 123 (88.5%) patients survived, and 16 (11.5%) patients died. Out of 565 patients, 56 patients who received convalescent plasma therapy, 21 (37.5%) patients died. Steroids in the form of methylprednisolone or dexamethasone were used to treat 77 patients with moderate to severe disease. Of these, 53 (69%) patients survived, and 24 (31.1%) patients died.

Survival analysis results showed an overall survival rate of 0.98, 0.96, 0.91, 0.83, and 0.71 on days 3, 7, 14, 21, and 28, respectively. Based on the survival model, patients above 60 years of age showed a decline in survival (0.96–0.59). Similarly, patients on dialysis also showed a decline in survival (1.00–0.35) during the 28-day study period (Table  5). Additionally, a decline in survival probability was observed for patients who were brought into the hospital from other facilities and directly transferred to the ICU (0.81 on day 3 to 0.42 on day 28).

Discussion

The present study was an attempt to identify the determinants of mortality in COVID-19 patients who were admitted to a tertiary care hospital using survival analysis. Our initial preliminary analysis results showed that the deceased were older male patients with diabetes and hypertension as comorbidities and on dialysis. Survival analysis is the data analysis of the outcome variable till the event occurs, and this event could be recovery, relapse, or death, with the time to the event being called survival time. The inclusion of censored data facilitating covariates makes this analysis more reliable.5

The Kaplan–Meier plots demonstrated a decrease in overall survival from 0.98 to 0.71 from day 3 to 28, and the risk of mortality was associated with increased age (above 60 years), hypertension, lung involvement, dialysis, breathlessness and severe disease on HRCT analysis, and low lymphocyte count on admission.

The median duration from the onset of symptoms to hospital admission was 5 (3/7) days. The duration from the onset of symptoms to death and the period from admission to death were 15 (11/25) and 10 (5/18) days, respectively. Far from our findings, Koya et al. described the factors associated with COVID-19 deaths in the state of Tamil Nadu, and the study reported the shortest time interval of 4.6 (IQR 7.1) days from the onset of symptoms to death for the patients admitted from March to June and first 10 days of August 2020.7 Contrastingly, Mishra et  al., reported the hospital stay of patients with a definitive outcome in 17 days, which was similar to our findings, in the state of Karnataka, India, for a cohort of patients from March to April 2020 wherein the follow-up of patients was performed after 14 days of discharge.8

Gender hegemony is a probable cause of COVID-19 infection among men as men tend to go outside of the house more frequently than their female counterparts due to work or migration, making them more susceptible to the infection.9–11 This observation was also made in our study, wherein the male population was found to be more susceptible than the female population.

While observing the age variability in our study, we concluded that an enhanced risk of death among COVID-19 patients was observed with increasing age. Our study corroborated with Mishra et  al., wherein the deceased were older at admission than the rest of the population and had severe disease; however, no difference was observed in the mortality rates between the male and female population in the study carried out by Mishra et al.8 This may be due to the lack of physical activity in the elderly, which leads to a reduction in their functional capacities and an increase in lifestyle disorders such as obesity, diabetes, and enhanced risk of heart diseases.12–15 Furthermore, these comorbidities increase the susceptibility of patients to increased severity, resulting from the exacerbation of preexisting infections.

Contrastingly, Koya et  al. demonstrated decreased survival from the time of the onset of symptoms among young adults without any comorbidities.7 However, the authors suggested that further studies must be carried out to investigate the involvement of any other factors in these patients. In addition, hypothyroidism was reported as COVID-19- associated morbidity. Shang et al. reported that diabetic patients requiring insulin had a shorter survival time than nondiabetic patients when infected with COVID-19, thereby contributing considerably to the mortality rate, which was consistent with the findings of our study.16 The survival time of diabetic patients was reduced when compared with that of nondiabetic patients (0.97–0.66), suggesting the association of diabetes with COVID-19-related mortality. Guzik et al. concluded that preexisting heart conditions were also associated with an increased risk of mortality.17 In the present study population, decreased survival was found to be associated with hypertension (0.97–0.70), kidney complications (1.00–0.35), and lung involvement (0.96–0.62). Similar to our findings, worldwide literature had adjusted estimates from cohort analyzes wherein increased age, male population, hypertension, diabetes, and chronic obstructive pulmonary disease or major cardiovascular diseases were found to be associated with enhanced risk of mortality or the severity of the disease among COVID-19 patients.18–21

The initial strategy is to promote control measures to reduce transmission and enhance the use of non-pharmacological interventions, including quarantine and self-isolation. This is done to reduce hospital admission and optimize hospital capacity.22,23 

The pooled estimates showed that HCQ treatment did not significantly affect survival at 14 and 28 days in COVID-19 patients when compared with the control population [relative risk (RR): 1.003, 95% CI: 0.983–1.022]. HCQ treatment also resulted in the alleviation of symptoms at day 10 (RR: 1.044, 95% CI: 0.911–1.196). However, COVID-19 patients with comorbidities were successfully treated with HCQ (RR: 1.058, 95% CI: 1.035–1.082), resulting in negative reverse transcription polymerase chain reaction (RT-PCR) results from positive RT-PCR results on day 6 (RR: 1.123, 95% CI: 1.041–1.212). Patients receiving HCQ treatment were at a higher risk of developing cardiac side effects (RR: 2.012, 95% CI: 1.428–2.833) and gastrointestinal side effects (RR: 1.318, 95% CI: 0.730–2.380).24 Severe infections are treated with more invasive pharmacotherapeutic agents such as remdesivir, biologicals, convalescent plasma, and anticoagulant therapy, whereas inflammation caused due to infection is treated with steroids.25–27 

In the present study, 47.7% of patients receiving tocilizumab did not survive, as compared to 3.6% who did not receive this drug. Salama et  al. demonstrated the effectiveness of this drug in hospitalized patients who are not on mechanical ventilation, wherein this drug, in fact, reduced the likelihood of disease progression to the composite outcome of mechanical ventilation or death. However, the authors did not observe the benefit of survival in their study.28

In the present study, HCQ + azee showed favorable outcomes in patients with mild and moderate infections. In severe cases, the effectiveness of HCQ + azee could not be determined as the stage of disease progression and previous treatments may alter its effectiveness.

The major limitation of our study is that the data set collected is limited to one center alone. In addition, the retrospective nature of our study was an important limitation. Furthermore, limited information regarding COVID-19 clinical variables and definitive treatment was available in the literature at the time of the study. Therefore, the results may not be generalizable for the rest of the country or other population groups and may not be representative of the whole COVID-19 population of our country. Our study is limited by some sources of error, which is common to all retrospective analyzes. Thus, prospective studies need to be performed in future to properly assess the risk factors for in-hospital death from COVID-19.

Conclusion

Thus, the present study on survival analysis revealed a decrease in survival probabilities of patients with increasing age, male gender, hypertension, diabetes, dialysis, and lung involvement. Furthermore, patients presenting with low lymphocyte count and breathlessness demonstrate the severity of the disease or the dysfunction of other organs and are considered at equal risk. Population variation in terms of treatment needs to be considered, wherein treatment with HCQ + azee gave positive results, thereby pointing toward pharmacogenomic influence. Lastly, the exact stage at which immunomodulators such as tocilizumab should be administered in patients must be studied in detail via conducting hardcore analysis.

Ethics

• Ethics Committee approval: Institutional Review Board approved the study with the IEC protocol number HNH/IEC/2021/ OCS/CCM/55.
• Informed consent: The requirement for written informed consent was waived by the IEC.