Risk Factors Associated with Mortality among Patients with COVID-19: Analysis of a Cohort of 1213 Patients in a Tertiary Healthcare Center

The presence of cardio-metabolic and respiratory comorbidities, immunosuppression, and chronic kidney disease have been associated with an increase in mortality from COVID-19. The objective of this study is to establish the risk factors associated with 30-day mortality in a cohort of hospitalized patients with COVID-19. This paper conducts a retrospective and analytical study of patients hospitalized for COVID-19 in a tertiary care center. A Cox proportional hazard analysis was performed to estimate the association of comorbidities with 30-day mortality. A total of 1215 patients with a median age of 59 years were included. In the adjusted Cox proportional hazards regression model, hypothyroidism, D-dimer ≥ 0.8 μg/mL, LHD ≥ 430 IU/L, CRP ≥ 4.83 ng/mL, and triglycerides ≥ 214 mg/dL were associated with an increased risk of death. The presence of a history of hypothyroidism and biomarkers (D-dimer, lactic dehydrogenase, CRP, and triglycerides) were associated with an increase in mortality in the studied cohort.


Introduction
The COVID-19 pandemic, caused by the SARS-CoV-2 virus infection, began in the Wuhan region of China in late 2019 and reached Mexico in February 2020, with a frenzied increase during the following months. The SARS-CoV-2 infection was mainly characterized by acute respiratory symptoms and related systemic complications. The presence of cardio-metabolic (diabetes, hypertension, and obesity) and respiratory (asthma and chronic obstructive pulmonary disease) comorbidities, immunosuppression, and chronic kidney J. Clin. Med. 2022, 11, 2780 2 of 11 disease (CKD) were all aggravating factors in the evolution and possible triggers for increased mortality in Mexico [1][2][3][4][5].
Bello-Chavolla et al. conducted a study within a Mexican sample where they investigated the specific risk factors associated with mortality, in particular the impact of diabetes and obesity on lethality in COVID-19 patients. Their results showed a higher risk of death after 30 days in patients with the following description: over 65 years old, diabetes mellitus 2 (DM2), obesity, CKD, chronic obstructive pulmonary disease (COPD), immunosuppression, and hypertension [3].
At the time of writing this manuscript (May 6th, 2022), 516,326,823 cases have been recorded worldwide, with a total of 6,248,434 deaths. In Mexico, the figure has reached 5,739,680 cases and 324,334 deaths, representing a case fatality rate of 1.21% and 5.65%, respectively [6]. This fatality rate could be higher amongst high-risk populations with a greater number of comorbidities. Although there are numerous reports in the Mexican population and in the world on various clinical and biochemical risk factors associated with the risk of mortality in patients with SARS-CoV-2 infection, a high-risk analysis will allow us to understand this phenomenon with greater precision.
Thus, the aim of this study is to establish the risk factors (clinical characteristics, comorbidities, and other predictors) associated with a 30-day mortality in a cohort of patients hospitalized due to COVID-19 in a tertiary care center where the prevalence of comorbidities was more pronounced.

Study Design and Patient Population
An observational, retrospective, and analytical study was conducted among COVID-19 patients at the Hospital de Especialidades del Centro Médico Nacional Siglo XXI del Instituto Mexicano del Seguro Social (a tertiary center) in Mexico City, Mexico, during the period between 1 March 2020 and 30 May 2021. Patients were included using the following criteria: ≥18 years old, with a COVID-19 diagnosis confirmed by RT-PCR, of any gender, and with complete labs for the purposes of the study. The primary outcome was death from COVID-19 and its complications within 30 days of hospital admission. The death certificates of 243 patients of the studied cohort were obtained.
The present study was approved by the local ethics and research committee (Registry identifier: R-2020-3601-245) and was consistent with the ethical guidelines of the 1975 Helsinki Declaration and the Mexican General Health Law on Research for Health Studies.

Data Collection
Data were collected retrospectively through consultation of the electronic clinical records. The demographic data, comorbidities, symptomatology at admission, laboratory results, and therapies used were all collected.

COVID-19 Diagnosis
The identification of SARS-CoV-2 was obtained by real-time RT-PCR in nasopharyngeal exudate samples processed at the Central Epidemiology Laboratory of the National Medical Center "La Raza" of the Mexican Social Security Institute, following international standards for infectious substances.

Definition of Variables
Co-morbidities were defined as follows: a chronic obstructive pulmonary disease (COPD) diagnosis of a postbronchodilator FEV1/FVC ratio around <0.70 [7]; asthma according to the Global Initiative for Asthma 2020 criteria [8]; heart disease as a diagnosis of coronary artery disease, congestive heart failure, or heart rhythm problems; CKD as a glomerular filtration rate below 60 mL/min [9]; immunodeficiency disease as the diagnosis of any immunodeficiency disorder, whether primary or secondary; diabetes as a fasting plasma glucose (FPG) level of 126 mg/dL or higher, a 2 h plasma glucose level of 200 mg/dL or higher during a 75-g oral glucose tolerance test, or a hemoglobin A1c level of 6.5% or higher [10]; hypertension as systolic blood pressure ≥140 mm Hg and/or diastolic blood pressure ≥90 mm Hg [11]; and primary hypothyroidism as a TSH level > 4.2 µIU/mL and T4 level < 0.93 ng/dL, and in the case of central hypothyroidism, a T4 level < 0.8 ng/dL. Analysis for detecting serum glucose, d-dimer, lactic dehydrogenase, ferritin, creactive protein, and triglycerides were performed in the laboratory of the Hospital de Especialidades Centro Médico Nacional Siglo XXI. Samples for laboratory analysis were obtained at the time of hospital admission.

Statistical Methods
Descriptive and inferential statistics were used for data analysis, taking into account measures of central trends and dispersion. The chi-square test was used to compare frequencies and proportions. The Mann-Whitney U test or Student's t test was used to compare quantitative variables. The Shapiro-Wilk test was used to determine the normality of the distribution of the variables. A receiver operating characteristic (ROC) analysis was performed to determine the best cutoff points for the following continuous quantitative variables: age, D-dimer, LDH, ferritin, CRP, and triglycerides. To determine the best cut-off points, sensitivity, specificity, accuracy, and likelihood ratio (LR)+ and LR-were taken into account.
A Cox proportional hazard analysis was performed to estimate the magnitude of the relationship between the different COVID-19-associated comorbidities (and biochemical parameters) and mortality during the 30 days after admission. The variables included in the regression model were made taking into account biological plausibility and statistical significance. A Kaplan-Meier plot was used to estimate the probability of survival at 30 days after hospitalization, while the log-rank test was used to compare the difference in survival probability for different groups of COVID-19 patients.
A two-sided p value was used for the in-between group difference with respect to the primary outcome. A p value of p < 0.05 was considered statistically significant. The statistical software used was SPSS version 25.0 (IBM SPSS Statistics for Windows, IBM Corp, Armonk, NY, USA), Stata SE software version 16 (StataCorp, College Station, TX, USA), and GraphPad Prism version 8.0 for Windows (GraphPad Software, San Diego, CA, USA).

Results
During the study period, 2074 patients were hospitalized with suspected COVID-19, out of whom 1215 met our criteria for mortality risk analysis ( Figure 1). The number of deaths recorded was 653 (53.7%).

Baseline Characteristics
The median age for the total study population was 59 years (IQR, 47-69 years). The median age in the group of patients who died was 63 years (IQR, 52-72 years) vs. 54 years (IQR, 43-66 years), with a difference of 9 years (95% CI; 6-11 years) between the two groups. Of the total group, 764 patients (62.8%) were male, with similar proportions between non-survivors (NS) and survivors (Table 1).

Baseline Characteristics
The median age for the total study population was 59 years (IQR, 47-69 years). The median age in the group of patients who died was 63 years (IQR, 52-72 years) vs. 54 years (IQR, 43-66 years), with a difference of 9 years (95% CI; 6-11 years) between the two groups. Of the total group, 764 patients (62.8%) were male, with similar proportions between non-survivors (NS) and survivors (Table 1).   In the group of patients studied, the presence of fever and dyspnea was significantly more frequent than in the group of non-survivors (Table 1). Regarding biomarkers, statistically significant differences were observed in the concentrations of fasting plasma glucose, D-dimer, LDH, ferritin, CRP, and triglycerides when comparing both groups (see Table 1). Similarly, a higher proportion of type 2 diabetes (T2D), hypertension, liver disease, and hypothyroidism was observed in the non-survivor group, reaching statistical significance ( Table 1). The proportion of patients with ≥2 comorbidities was significantly higher in the non-survivor group than in the survivor group. (See Table 1).

ROC Analysis to Determine Cut-Off Points
The cut-off points of the different continuous quantitative variables and their respective sensitivity, specificity, accuracy, LR+, and LR−, are presented below: D-dimer ≥ 0. 8

Cox Proportional Hazards and Kaplan-Meier Analysis
In the adjusted Cox proportional hazards regression model, hypothyroidism, D-dimer ≥ 0.8 µg/mL, LDH ≥ 430 IU/L, CRP ≥ 4.83 ng/mL, and triglycerides ≥ 214 mg/dL were significantly associated with an increased risk of death (Table 2, Figure 2). For the Cox proportional hazards model, a chi2 LR of 91.74 was obtained with a p value < 0.001.  Figure 2).

Discussion
For the present study, we present the analysis of a group of 1215 adult patients hospitalized with COVID-19. The objective was to establish which risk factors were associated with a 30-day mortality. We anticipated that the contribution would be to explore the possible additive effect of an initial SARS-CoV-2 infection, plus the presence of comorbidities and biomarkers on 30-day mortality. In the Cox proportional hazard model, a history of hypothyroidism and biomarkers (D-dimer ≥ 0.8 µg/mL, lactic dehydrogenase ≥ 430 IU/L, CRP ≥ 4.83mg/dL, and triglycerides ≥ 214 mg/dL) were significantly associated with increased mortality.
As  [12] and Escobedo-de la Peña et al. also found a case fatality rate (CFR) that increased with age; more than half of the deceased subjects were ≥70 years-old [5]. It has been acknowledged that aging affects the function of the adaptive and innate immune system, and therefore increases susceptibility to infections, and suppressed Natural Killer (NK) cell cytolytic activity has been observed more in elderly patients compared to younger subjects. NK cells are a family of innate immune cells that play an essential role in antiviral immunity [13].
Hypertension, like other cardiovascular diseases, has been associated with increased mortality in patients hospitalized for COVID-19 [4,14,15]. In our study, it was the most frequent comorbidity in the NS group (34.46%). However, in the multivariate model, no significant association was found with mortality in the cohort of patients studied. Animal models suggest that RAAS blockers may increase ACE2 expression and potentially increase the risk of SARS-CoV-2 infection. On the other hand, a recent meta-analysis published by Chang Chu et al. concluded that angiotensin-converting enzyme inhibitors (ACEIs) reduce the risk of SARS-CoV-2 infection and all causes of COVID-19 mortality, including the risk of non-COVID-19 pneumonia [16]. The exact mechanism of hypertension related to the COVID-19 severity remains unclear [15].
T2D was the second most frequent comorbidity in the NS group (28.02%). However, no significant association was found with mortality. Our findings contrast with those reported in other studies such as the one by Muhammad M AbdelGhaffar et al. who reported an OR for T2D mortality of 1.58 (95% CI 1.14-2.19; p = 0.006 [17]. Lana pinto et al. performed a meta-analysis where the pooled OR was 3.53 (95% CI; 1.48-8.39) [18], similar to the Zeng-Hong Wu et al. OR of 1.75 (95% CI 1.31-2.36; p = 0.0002) [19]. The impaired immune system coupled with the metabolic imbalance observed in patients with T2D increases susceptibility to infection and perhaps a more severe disease. NK cells are innate lymphocytes that detect and destroy virus-infected cells. It has been observed that there is an increase in the number of dysfunctional NK cells in patients with T2D, probably secondary to an increase in reactive oxygen species. Hyperglycemia itself alters the quaternary structure of proteins in NK cells, inducing their apoptosis and decreasing their viral clearance capacity [20]. This possibly explains the increased risk of death for patients with T2D.
A history of primary hypothyroidism was associated with an increased risk of mortality, with an HR of 1.91 (95% CI; 1.08-3.39, p = 0.02), and as in the general population, it was more frequent in the female gender compared to males [21], both in the total group (68.8%, p = 0.002) and in the NS group (65%, p = 0.005). The relationship between hypothyroidism and mortality in COVID- 19 [23]. It has been shown that thyroid hormones can affect the production of NK cells [13]. Treatment with triiodothyroxine (T3) can induce IL-2 receptor expression on peripheral blood mononuclear cells; administration of low doses of thyroxine enhances the stimulatory effect of interferon on NK cells, such that there is evidence of an interaction between thyroid function and NK cell activity [13]. Based on the above, a history of hypothyroidism could influence the deterioration of the cellular immune response to SARS-CoV-2 infection, increasing the probability of death in hospitalized patients.
Regarding the association between triglyceride concentration and mortality, our analysis yielded an increased risk when its concentration was ≥214 mg/dL (HR of 1.38 (95% CI [1.17-1.63], p < 0.001). Wen Dai et al. had already found this association in their study of deceased patients who had higher levels than survivors (179 vs. 134 mg/dL, p < 0.001), with an OR for mortality of 2.3 (95% CI [1.4-3.7], p = 0.001) [24]. Previous studies have shown that infection and inflammation induce hypertriglyceridemia due to inhibition of serum clearance by lipoprotein lipase, a key enzyme in triglyceride catabolism. On the other hand, recent studies suggest a promoting effect of triglycerides in inflammation; high levels favor macrophage extravasation to tissues. This evidence may suggest a detrimental effect of hypertriglyceridemia on leukocyte activation in patients with COVID-19, placing them at high risk for severe disease [24].
The identification of biomarkers associated with mortality measured early in hospitalized patients with COVID-19 could be a useful diagnostic tool for making therapeutic decisions in high-risk patients. In fact, we found a significant increase in the risk of mortality in those patients with D-dimer concentrations ≥ 0. 8 [3.37-5.68], p < 0.001) were markers independently associated with a high risk of poor outcome. Likewise, other authors have also described the association of CPR, LDH, D-dimer, and ferritin biomarkers [2,[25][26][27]. Biomarkers (CPR, D-dimer, LDH, and ferritin) are quantitative measurements that reflect the pathophysiology of the disease and help the clinician recognize the severity of the condition. They help guide therapeutic decisions to improve patient prognosis [28]. CRP is an acute-phase protein synthesized in hepatocytes in response to IL-6; its concentrations are elevated in different inflammatory processes including infectious ones and are therefore useful in the diagnosis and analysis of the severity of the infectious process [28,29]. Elevated D-dimer concentrations, also associated with C-reactive protein levels, have been associated with a poor prognosis in patients with COVID-19 [17], an effect associated with hypoxia due to severe pneumonia and increased inflammatory response, conditions related to a state of hypercoagulability, resulting in disseminated intravascular coagulation and multiorgan failure [17].
Increased serum LDH levels are a marker of the presence of tissue injury, necrosis, and hypoxia and is an independent marker associated with increased mortality in patients with sepsis [30]. In a study by Yi Han et al., LDH was found to be a robust predictor for early recognition of lung injury and its severity due to COVID-19 [31]. In our study, it was also found to be an unfavorable prognostic factor for survival.
The exact role of ferritin in the pathophysiology of COVID- 19 has not yet been fully established. However, what is currently known is that, in response to tissue injury, cytokines stimulate the production of defense proteins by the liver, including C-reactive protein and ferritin. Transcription and translation of ferritin are induced by IL-1β, IL-6, and IFN-γ. Additionally, macrophages and damaged cells account for elevated ferritin values. Ferritin promotes the release of proinflammatory mediators and increases the inflammatory burden, resulting in a vicious circle. Ferritin achieves this by activation of NF-the other hand, recent stud high levels favor macroph mental effect of hypertrig placing them at high risk The identification of talized patients with COV decisions in high-risk pat tality in those patients wi 1.37], p = 0.01), LDH ≥ 430 (HR 1.40, 95% CI [1.11-1 ng/mL, there was a trend p = 0.12). In a meta-analys they evaluated laboratory pitalized for COVID-19, t OR 3.39; 95% CI [2.66-4.3 0.00001), and elevated Cwere markers independen authors have also descri omarkers [2,[25][26][27]. Bioma urements that reflect the p the severity of the condit prognosis [28]. CRP is an IL-6; its concentrations ar tious ones and are theref infectious process [28,29]. tive protein levels, have b 19 [17], an effect associate matory response, conditio nated intravascular coagu Increased serum LDH and hypoxia and is an ind with sepsis [30]. In a stud early recognition of lung i also found to be an unfav ԟ The strengths of our us to accurately estimate hospitalized patients with B, leading to a positive regulation of ferritin gene transcription [32]. In our study, its impact appears not to have been very significant on the risk of death.
The strengths of our study include having an acceptable sample size, which allows us to accurately estimate which covariates were associated with mortality in a group of hospitalized patients with COVID-19. The cohort of hospitalized patients was adequately characterized and followed-up on for the established time. The limitations of the study are its retrospective nature; the inclusion of severe cases referred to us by a third level center, which could limit the external validity of the study; the difficulty in obtaining all the death certificates, which would allow corroborating the final cause of death; and the presence of a probable differential bias, since the analysis of the patients' information was done according to the background info and not the morbidity control. Finally, it was not possible to evaluate the weight and body mass index of the patients, due to underreporting in the clinical record.

Conclusions
The data obtained in our study suggest that the presence of a history of hypothyroidism, a D-dimer ≥ 0.8 µg/mL, lactic dehydrogenase ≥ 430 IU/L, CRP ≥ 4.83 mg/dL, and triglycerides ≥ 214 mg/dL were associated with an increase in mortality in the studied cohort. A significant number of comorbidities probably influence mortality as a constant risk.