Sex differences in clinical characteristics and risk factors for mortality among severe patients with COVID-19: a retrospective study

Demographics and characteristics

1530 patients with confirmed SARS-CoV-2 detection were admitted to Eastern Renmin Hospital of Wuhan University as of Apr 10, 2020. After excluding 776 patients that were still hospitalized or not diagnosed as severe illness as of Apr 10, 2020, as well as 103 inpatients without available key information in their medical records, we included 651 severe patients in the final analysis. Among these patients, the median age was 62 years (IQR, 49-71), ranging from 18 years to 98 years, 287 (44.1%) patients were elderly, and 332 (51.0%) patients were male (Table 1). All these patients were Wuhan citizen, of whom 5.2% patients were current smokers, and 35.8% patient had a clear contact history with confirmed COVID-19 patients. They all presented bilateral involvements of typical chest computerized tomography manifestations of COVID-19 pneumonia on admission. 274 of 651 (42.1%) patients had comorbidities. The most common comorbidity was hypertension (183 [28.1%]), followed by diabetes (116 [17.8%]), cardiovascular disease (83 [12.7%]), chronic lung disease (31 [4.8%]), cerebrovascular disease (28 [4.3%]), chronic kidney disease (20 [3.1%]), chronic liver disease (20 [3.1%]), and malignancy (16 [2.5%]). The most common symptoms on hospital admission were fever (540 [82.9%]), dry cough (338 [51.9], fatigue (217 [33.3%]), dyspnea (207 [31.8%]), expectoration (133 [20.4%]), and anorexia (97 [14.9%]). Dizziness (25 [3.8]), headache (23 [3.5%]), nausea (18 [2.8%]), vomiting (18 [2.8%]), abdominal pain (6 [0.9%]), and hemoptysis (4 [0.6%]) were rare (Table 1). Compared to the female patients, male patients were more likely to report clear exposure history (P =0.003) and have fever (P =0.0046).

Laboratory findings

Laboratory values on admission were summarized in Table 2. The most common hematologic abnormalities were lymphocytopenia (327 [50.2%]), neutrophilia (153 [23.5%]), mononucleosis (144 [22.1%]), and thrombocytopenia (63 [9.7%]). Most patients presented decreased T-lymphocyte, including CD3 (226 of 335 [67.5%]), CD4 (215 of 335 [64.2%]), CD8 (204 of 335 [60.9%]), CD16+56 (121 of 335 [36.1%]), and CD19 (95 of 335 [28.4%]). Most patients had elevated cardiac indices, including lactate dehydrogenase (269 of 427 [63.0%]), N-terminal pro-B-type natriuretic (102 of 324 [31.5%]), myoglobin (83 of 355 [23.4%]), Hs-TnI (95 of 432 [22.0%]), creatine kinase-MB (78 of 466 [16.7%]), and creatine kinase (49 of 432 [11.3%]). Coagulation abnormalities were also obvious: 340 of 544 (62.5%) patients had raised D-dimer, 111 of 431 (25.8%) had prolonged activated partial thromboplastin time, and 83 of 434 (19.1%) had prolonged prothrombin time. More than half of patients had COVID-19 related inflammation indicated by elevated C-reactive protein (277 of 430 [64.4%]) and procalcitonin (164 of 398 [41.2%]). Some patients demonstrated liver injury with elevated aspartate aminotransferase (191 of 607 [31.5%]), alanine aminotransferase (153 of 607 [25.2%]), and total bilirubin (35 of 450 [7.8%]). Kidney injury indicated by elevated blood urea nitrogen (115 of 515 [22.3%]) and creatinine (161 of 513 [31.4%]) was also considerable.

Compared to female patients, male patients showed significantly higher neutrophil count (P = 0.026), lower counts of lymphocyte (P < 0.001) and platelet (P = 0.001) (Table 2). Male patients also presented significantly lower counts of CD3 (P = 0.009), CD4 (P = 0.003), CD8 (P = 0.01) and CD19 (P = 0.013) T-cells as compared with female patients. Additionally, male patients had higher levels of alanine aminotransferase (P < 0.001) and aspartate aminotransferase (P < 0.001), and longer prothrombin time (P = 0.01) and activated partial thromboplastin time (P < 0.001). Total bilirubin, urea nitrogen, creatinine, lactate dehydrogenase, creatine kinase, myoglobin, C-reactive protein, and procalcitonin were also significantly higher in male patients than female patients (all P < 0.05) (Table 2).

Treatments, complications and clinical outcomes

Among the 651 patients, 95.6% received antiviral medications (eg, lopinavir, ritonavir), 78.3% patients received antibiotics, 34.1% received glucocorticoid, and only 2.0% received continuous renal replacement therapy (Table 3). 336 of 651 patients (51.6%) patients received high-flow nasal cannula oxygen therapy, 62 (9.5%) patients received noninvasive mechanical ventilation, and 45 (6.9%) patients required invasive mechanical ventilation, of whom 3 received extracorporeal membrane oxygenation as rescue therapy. ARDS (266 [40.8%]) was the most common complication, followed by secondary infection (103 [15.8%]), acute cardiac injury (93 [14.3%]), coagulopathy (90 [13.8%]), septic shock (63 [9.7%]), acute kidney injury (58 [8.9%]) and arrhythmia (18 [2.8%]). The media time from first symptoms to hospital admission was 11 days (IQR, 6-17), and 11.5 days (IQR, 4.8-18.5) to ARDS. The median length of hospital stay was 20 days (IQR, 9-30).

Male patients were more likely to receive glucocorticoid therapy (133 [40.1%] vs 89 [27.9%]) as compared with female patients. The frequencies of ARDS (160 [48.2] vs 106 [33.2]), secondary infection (72 [21.7%] vs 31 [9.7%]), acute cardiac injury (60 [18.1%] vs 33 [10.3]), coagulopathy (58 [17.5%] vs 32 [10.0%]), acute kidney injury (38 [11.2%] vs 20 [6.2%]) and arrhythmia (14 [4.2%] vs 4 [1.3%]) were significantly higher than female patients (all P < 0.05). Of note, the mortality in male patients was much higher than female patients (P < 0.05).

Risk factors for in-hospital death

As shown in Table 4, univariable logistic regression analysis revealed that sex, old age, leucocytosis, neutrophilia, lymphocytopenia, thrombocytopenia, CD3, CD4 and CD8 T-cells, alanine aminotransferase, aspartate aminotransferase, total bilirubin, blood urea nitrogen, creatinine, lactate dehydrogenase, creatine kinase, myoglobin, Hs-TnI, C-reactive protein, prothrombin time, activated partial thromboplastin time, and procalcitonin were associated with in-hospital death in severe patients with COVID-19. Multivariable logistic regression model further showed that neutrophilia (OR 6.845, 95% CI 1.227-38.192, p = 0.028), thrombocytopenia (OR 19.488, 95%CI 3.030-25.335, p = 0.002), Hs-TnI greater than 0.04 pg/mL (OR 6.058, 95%CI 1.545-23.755, p = 0.010), and procalcitonin greater than 0.1 ng/mL (OR 6.350, 95%CI 1.396-28.882, p = 0.017) on admission were associated with the incidence of in-hospital death (Table 4).

The Kaplan-Meier survival curve showed that the cumulative survival rate during hospitalization was significantly lower in patients with neutrophilia, thrombocytopenia, Hs-TnI greater than 0.04 pg/mL or procalcitonin greater than 0.1 ng/mL as compared with the corresponding normal indices both in male and female patients (Figure 1). Compared to female patients, the cumulative survival rate of male patients with either of these risk factors was relatively lower. Male patients were less tolerant to neutrophilia, thrombocytopenia, Hs-TnI greater than 0.04 pg/mL and procalcitonin more than 0.1 ng/mL relative to female patients.

Figure 1. Kaplan-Meier survival curves for mortality between male and female patients. (A) Neutrophil count. (B) Platelet count. (C) Hs-TnI ≤ 0.04 pg/mL. (D) procalcitonin ≤ 0.1 ng/mL.

Study design and participants

This retrospective cohort study was conducted at Eastern Renmin Hospital of Wuhan University, Wuhan, China, which was an assigned hospital to treat COVID-19 by the government. All patients who were diagnosed with COVID-19 according to WHO interim guidance [24] were screened from Jan 25, 2020 to Apr 10, 2020. We included all severe patients with COVID-19 in this study. Severe patients were those who meet any of the following criteria according to the Chinese management guideline for COVID-19 (version 7.0) [25]: 1) shortness of breath, respiratory frequency ≥ 30/min; 2) blood oxygen saturation ≤ 93%; 3) the ratio of arterial partial pressure of oxygen (PaO2) to fraction of inspired oxygen (FiO2) ≤ 300 mmHg; 4) chest CT imaging showing lung infiltrates > 50% within 24 to 48 hours.

This study was reviewed and approved by the Ethics Commission of Renmin Hospital of Wuhan University (approval number WDRY2020-K118). Written informed consent was waived by the designated hospital due to the emerging infectious disease.

Data collection

The demographic characteristics (age, sex, exposure history), underlying comorbidities (eg, hypertension, diabetes, cardiovascular disease), clinical symptoms (eg, fever, dry cough, dyspnea), laboratory findings (eg, hematologic, biochemical, coagulation function), treatments (eg, antiviral, antibiotic, glucocorticoid therapy), complications (eg, acute respiratory distress syndrome [ARDS], secondary infection, acute cardiac injury), and outcomes (discharge and mortality) were collected from electronic medical records by three investigators. Another two investigators independently reviewed all the data to verify data accuracy.

Outcomes

The primary outcomes were in-hospital mortality among male and female patients with COVID-19. The secondary outcomes included incidence of ARDS, acute cardiac injury, secondary infection, coagulopathy and acute kidney injury.

ARDS was diagnosed as acute-onset hypoxemia with bilateral pulmonary opacities on chest imaging which was not fully explained by other disease according to the Berlin Definition [26]. Secondary infection was defined if patients showed clinical symptoms or signs of bacteremia and a positive culture of a new pathogen was obtained from sputum or blood samples after admission [25]. Acute cardiac injury was defined when the blood concentration of cardiac biomarkers (eg, hypersensitive troponin I, Hs-TnI) was above the 99^th percentile reference limit or new abnormalities presented in electrocardiography and echocardiography [5]. Coagulopathy was defined as a 3-second extension of prothrombin time or a 5-second extension of activated partial thromboplastin time [12]. Sepsis was diagnosed according to the 2016 Third International Consensus Definition for Sepsis and Septic Shock [27]. Acute kidney injury was defined according to the Kidney Disease Improving Global Outcomes (KDIGO) criteria [28].

Statistical analysis

Continuous variables were expressed as median (interquartile range, IQR), and compared by Mann-Whitney U test. Categorical variables were expressed as number (%), and compared by χ² test or Fisher’s exact test between male and female patients. To explore the risk factors associated with in-hospital mortality, univariable and multivariable binary logistic regression models were used. The Kaplan-Meier curve was performed with log-rank test for in-hospital mortality among female and male patients. A two-sided p value less than 0.05 was considered statistically significant. Data analysis was conducted with SPSS software (version 22.0). Statistical charts were generated by Graphpad Prism 8.0.