Risk Factors of Secondary Infections in Severe and Critical Patients Hospitalized with COVID-19: A Case-Control Study

Jie Li Department of Pharmacy,Union Hospital,Tongji Medical College,Huazhong University of Science and Technology Jingchao Cao Department of Pharmacy,Union Hospital,Tongji Medical College,Huazhong University of Science and Technology Peishan Cai Wuhan Union Hospital Baoxia Shi Tongji Medical College A liated Wuhan Puai Hospital Jie Cao Wuhan Union Hospital Yu Zhang Department of Pharmacy,Union Hospital,Tongji Medical College,Huazhong University of Science and Technology Junwei Wang (  wjwhzkj@163.com ) Wuhan Union Hospital

Inclusion criteria for case group: inpatients with COVID-19 who were conformed as secondary bacterial and fungal pneumonia or bloodstream infections 48 hours after admission. Inclusion criteria of control group: inpatients with COVID-19 who had no bacterial or fungal infections from admission to discharge.
Case group included 56 inpatients with secondary infections (pneumonia and bloodstream infections caused by bacteria or fungal); control group included 182 inpatients (no bacterial or fungal infections).

Data collection
The demographic, laboratory ndings and treatments of cases and controls were extracted from electronic medical records. Data of case group were recorded from hospital admission to secondary infections. Data of control group were recorded from hospital admission to discharge or death. To ensure the accuracy of data, two independent researchers were arranged to review and check the data form.

De nitions
Disease severity status on admission was de ned according to the diagnostic and treatment guideline for COVID-19 (version 7) [5]. Comorbidities was identi ed according to the International Classi cation of Diseases and Injuries-10 diagnostic codes. In our study secondary infections referred to secondary pneumonia and bloodstream infections caused by bacteria or fungal. Secondary infections were con rmed when inpatients showed clinical symptoms or signs of pneumonia or bloodstream infections, as well as positive culture of a new pathogen obtained from lower respiratory tract specimens or blood samples 48 h after admission [10,11]. Ventilator-associated pneumonia was diagnosed according to the guidelines for treatment of hospital-acquired and ventilator-associated pneumonia [12]. Anemia was de ned as blood hemoglobin of less than 90 g/L. Hypoproteinemia was de ned as blood albumin of less than 25 g/L. Electrolyte disturbance was con rmed when at least one of sodium, potassium, calcium and magnesium of human plasma was lower or higher than the normal range.

Statistical analysis
Continuous variables were presented as median (IQR); categorical variables were present as frequencies. Differences between cases and controls were compared with Mann-Whitney U test, χ² test, or Fisher's exact test. Subgroup analysis was presented as percentage. Multivariable binary logistic regression models were used to explore the risk factors associated with secondary infections.
Severity of illness on admission, ICU admission, mechanical ventilator, central venous catheterization, gastric catheterization, catheter catheterization, and sedatives were common in the cases, while almost none of these factors was observed in the controls. Therefore, regression analysis was not performed on these factors.
Considering the total number of case patients (n = 56) in our study, we chose six variables based on previous ndings of pneumonia, bloodstream infections and clinical importance. 7,[13][14][15][16] The six variables included sex, age 65 or older, comorbidity (heart diseases), laboratory ndings (hypoproteinemia) on admission and drug therapy (corticosteroids and PPI) during hospitalization. P value less than 0.05 was regarded as statistically signi cant. All statistical analyses were performed by IBM SPSS Statistics 26.0.

Demographic
A total of 238 inpatients with COVID-19 were enrolled, the median age of 238 inpatients was 59.0 years old (IQR 50.0-68.0), ranging from 21 to 91-year-old. The median age of case group was higher than that of the control group (IQR 66.0 vs 57.0).
The proportion of males in the cases was signi cantly higher than that of the control group (67.9% vs 41.8%). Similarly, the proportion of patients (age, ≥ 65 years) in the cases was nearly one times that of the control group (55.4% vs 29.1%) ( Table 1). 62.2% of patients suffered from at least one comorbidity, and this proportion was as high as 91.1% in the cases, which was signi cantly higher than 53.3% in the controls.

Laboratory ndings and treatment
Over half of the cases appeared electrolyte disturbance, while the proportion was 34.6% in controls. Odds of hypoproteinemia (albumin of less than 25 g/L) was higher in case group. 75.0% of patients used PPI in cases, while the proportion was only 29.1% in controls. 69.6% of cases used corticosteroids, which was only 8.8% in the controls ( Table 1).
Severity of illness on admission, ICU admission, ventilator, central venous catheterization, gastric catheterization, catheter catheterization, and sedatives were common in the cases, however almost none of them was observed in the controls (Table 1).

Subgroup analysis
Among 56 patients of the case group, the number of severe and critical patients were 20 (35.7%) and 36 (64.3%), respectively. Severity of illness on admission, ICU admission, ventilator, central venous catheterization, gastric catheterization, catheter catheterization, and sedatives were more common in critical patients (Table 2).
Of the 56 cases with secondary infections, 50 had pneumonia and 6 had bloodstream infections, respectively. 11 of pneumonia patients subsequently developed bloodstream infections; half of 6 bloodstream infections patients occurred pneumonia. 4 of the 14 inpatients with both pneumonia and bloodstream infections were infected by the same pathogens (Carbapenem-Resistant Klebsiella pneumoniae) (

Discussion
In this case-control study, we identi ed several risk factors of secondary infections in severe and critical patients hospitalized with COVID-19. Signi cant risk factors were male, age 65 or older, heart diseases, hypoproteinemia, corticosteroids and PPI.
As previously covered risk factors for hospital acquired pneumonia (HAP), male, age 65 or older were risk factors of secondary infection [13,17]. Recent studies related to COVID-19 reported that male was a risk factor associated with disease severity status, and age 65 or older was a risk factor related to death [7,15,18,19].
In multivariate regression model, heart disease was the only underlying comorbidity associated with secondary infections. It had been uncovered that, acute cardiac events and poor prognosis appeared on patients with coronary heart disease were related to in uenza and lower respiratory tract virus infection [20]. Previous report indicated that MERS-CoV would damage the heart muscle, when COVID-19 appeared, some researchers also con rmed its heart damage effect [21]. So what was the relationships between heart diseases and secondary infections in patients with COVID-19? We speculated that patients with heart disease once infected with COVID-19 were more likely to develop into severe illness, thus exposing inpatients to invasive devices such as mechanical ventilation and central venous catheterization.
Diabetes was generally considered as a risk factor for infections [22], but in our study, there was no statistically signi cant difference between cases and controls. Previous risk factors of pneumonia reported chronic obstructive pulmonary disease (COPD) to be one risk factor associated with secondary infection [23].We also found the proportion of COPD in the cases was more than three times that in the controls.
Unfortunately, we did not record enough patients with COPD in the two groups. Despite the incidence of anemia between cases and controls were notable difference, there were only 3 patients in each group. Therefore anemia couldn't be included in univariable regression.
In a sepsis study, albumin could be used as a predictor of disease severity [24]. Hypoproteinemia was a risk factor of carbapenem-resistant Klebsiella pneumoniae bacteremia in non-transplant patients [14]. In our study, we found that hypoproteinemia was also a key risk factor of secondary infections. Report showed COVID-19 attacked the body's immune and disrupted immune response [6]. Besides, hypoproteinemia would weaken immunity to be susceptible to infections, and systemic edema, ascites and pleural effusion caused by hypoproteinemia might cause infections.
Among factors identi ed by multivariate regression, corticosteroids' risk ratio was notable. Corticosteroids could suppress the immune system if taken for long time or large doses. A research of BSI in ICU reported immunosuppressants were associated with BSI [9]. One possible cause of hospital-acquired pneumonia (HAP) was that diseases on admission destroyed patients' immune system, thus making patients be susceptible to infections [25]. In addition, COVID-19 attacked the human immune system and made corticosteroids' in uences more prominent [6]. Therefore, the use of corticosteroids to treat in ammation was a double-edged sword. It was necessary to comprehensively evaluate the patient's condition before rationally using corticosteroids in the short term.
Taking proton pump inhibitors (PPI) increased the risk of secondary infections in patients hospitalized with COVID-19. Herzig SJ et al. reported acid-suppressive medication use was associated with 30% increased odds of hospital-acquired pneumonia. Statistically signi cant risk was demonstrated only for use of PPI [16].
Undoubtedly, risk factors associated with secondary infections were severity of illness on admission, ICU admission, ventilator, central venous catheterization. Disease severity status on admission was a leading cause for secondary infections. Critical illness of patients usually had decreased level of consciousness and needed to be admitted to the ICU. Once it happened, invasive devices such as ventilator and central venous catheterization were required. These were the major risk factors of secondary infections con rmed in many previous studies [26][27][28][29][30].
Unfortunately, of the same pathogens from sputum and blood culture 4 inpatients all died. We speculated prolonged pneumonia or bloodstream infections could cause another infection.
In recent years, scholars reported gastric catheterization and urinary catheterization were risk factors of nosocomial infections [14, 19,31].
Intravenous sedatives were typically used to enhance inpatients comfort and patient-ventilator synchrony. A systematic review about the relationship between sedatives and healthcare-associated infection reported that, the three most common sedatives (benzodiazepines, propofol, and dexmedetomidine) for mechanically ventilation patients had different pharmacologic and immunomodulatory effects, which might impact infection risk. 32 Further observations were demanded to con rm the relationship between gastric catheterization, urinary catheterization, and sedatives with secondary infections.
However, our study has several limitations. Firstly, control group could not include enough critical inpatients due to the characteristics of pneumonia caused by COVID-19. Almost none of these phenomena were observed in the controls (severity of illness on admission, ICU admission, ventilator, central venous catheterization, gastric catheterization, urinary catheterization, and sedatives). More prospective studies were needed to quantify the odds of secondary infections increased by these factors. Secondly, due to the lack of certain records on admission, some factors that might be related to secondary infections could not be explored, e.g. smoking history and BMI (body mass index). Finally, limited by the number of inpatients enrolled in case group, some variables could not be included in the multivariate regression model simultaneously.
To the best of our knowledge, this is the rst report of risk factors for secondary infections in severe and critical patients hospitalized with COVID-19. Severe and critical inpatients with male, age (≥ 65 years), heart diseases, hypoproteinemia, treated with corticosteroids, and PPI need to be observed carefully and to be intervened with drugs to prevent the occurrence of secondary infections as early as possible. In particular, inpatients needing ICU admission and invasive devices also need to be given optimal cares and to be minimized the duration.