Co-infection in patients with hypoxemic pneumonia due to COVID-19 in Reunion Island

Abstract This study aimed to evaluate the incidence of co-infection with different types of pathogens in patients with hypoxemic pneumonia due to coronavirus disease 2019 (COVID-19) in Reunion Island. This observational study using a prospectively collected database of hypoxemic pneumonia due to COVID-19 cases was conducted at Félix Guyon University Hospital in Reunion Island, France. Between 18 March 2020 and 15 April 2020, 156 patients were admitted to our hospital for COVID-19. A total of 36 patients had hypoxemic pneumonia (23.1%) due to COVID-19. Thirty of these cases (83.3%) were imported by travelers returning mainly from metropolitan France and Spain. Patients were screened for co-infection with other pathogens at admission: 31 (86.1%) by multiplex polymerase chain reaction (PCR) and 16 (44.4%) by cytobacteriological examination of sputum culture. Five patients (13.9%) were found to have co-infection: 1 with influenza virus A H1N1 (pdm09) associated with Branhamella catarrhalis, 1 with Streptococcus pneumoniae associated with Haemophilus influenzae, 1 with Human Coronavirus 229E, 1 with Rhinovirus, and 1 with methicillin-susceptible Staphylococcus aureus. Patients with co-infection had higher D-dimer levels than those without co-infection (1.36 [1.34–2.36] μg/mL vs 0.63 [0.51–1.12] μg/mL, P = .05). The incidence of co-infection in our cohort was higher than expected (13.9%). Three co-infections (with influenza virus A(H1N1) pdm09, Streptococcus pneumoniae, and Staphylococcus aureus) required specific treatment. Patients with hypoxemic pneumonia due to COVID-19 should be screened for co-infection using respiratory cultures or multiplex PCR. Whilst our study has a number of limitations, the results from our study suggest that in the absence of screening, patients should be commenced on treatment for co-infection in the presence of an elevated D-dimer.


Introduction
Coronavirus disease 2019  first emerged in China in December 2019. [1] Since then, the disease has spread rapidly across all continents. The COVID-19 pandemic is overwhelmingly associated with international travel. [2][3][4][5] As of 29 September 2020, there were nearly 34 million cases of COVID-19 infection worldwide, with more than 1 million deaths. [6] Reunion Island, a French overseas department located in the Indian Ocean region, is connected to Paris, South Africa, India, China, and Thailand through several daily flights. On 29 September 2020, 3882 cases of COVID-19 had been reported on Reunion Island (first case detected on 11 March 2011). During the first epidemic period (March to July 2020), 667 cases of COVID-19 had been reported. The vast majority of these cases were imported from Europe (France, Spain, and Italy) and Comoros Archipelago. [7] To date, studies that explored cases of co-infection between COVID-19 and other microorganisms were mainly carried out in Asia. [8][9][10][11][12][13] The aim of our study was to evaluate the incidence of co-infection with different types of pathogens in patients with hypoxemic pneumonia due to COVID-19 in Reunion Island during the first epidemic period.

Methods
The present observational study was approved by the Ethics Committee of the French Society of Pulmonary Medicine and was declared to the Commission nationale de l'informatique et des libertés (French Data Protection N°2206739). The need for informed consent was waived, as the study was non-interventional and followed our usual protocol. However, all patients or their legally authorised representative were verbally informed about the process of data collection and could refuse to participate in the study.
This study complies with the Strengthening the Reporting of Observational studies in Epidemiology recommendations statement. [14] 2.1. Selection of the study sample This retrospective observational cohort study using a prospectively collected database was conducted between 18 March 2020 and 15 April 2020 at Félix Guyon University Hospital, the only hospital authorized to manage patients with COVID-19 in Reunion Island, France. All patients diagnosed with COVID-19 in another hospital of the Island were systematically transferred to our hospital.
During this study, hospitalization policy has changed according to the guidelines from the French Ministry of Health, depending on the stage of the outbreak. From 11 to March 24, 2020, all diagnosed cases were systematically hospitalized for 14 days. As of March 25, 2020 (start of the spread of COVID-19 in Reunion Island), only cases with severity's signs were hospitalized.
All consecutive patients with COVID-19 confirmed by polymerase chain reaction (PCR) and presenting with hypoxemic pneumonia were evaluated.
All patients with hypoxemic pneumonia due to COVID-19 were treated with a 3rd generation cephalosporin not active against Pseudomonas aeruginosa.
In the absence of contraindications, patients with symptoms for less than 10 days were treated with oral hydroxychloroquine for a period of 10 days, in association with azithromycin for a period of 5 days. [17,18]

Microbiological investigations
Samples used (nasopharyngeal swab in non-intubated and tracheal aspirate in intubated patients) for COVID- 19  Pneumococcal and Legionella urinary antigen tests, cytobacteriological examination of sputum cultures, and serology of atypical respiratory pathogens were performed at the physician's discretion (DiaSorin-LIAISON XL (CLIA) for Mycoplasma pneumoniae and NovaLISA/NovaTec technique (Biomnis, EIA) for Chlamydia pneumoniae).

Data collection and study outcome
Patient comorbidities at hospital admission were recorded. Clinical and biological data were collected at the time of diagnosis of hypoxemic pneumonia due to COVID-19.
The primary outcome was to identify the microorganisms responsible for co-infection in patients with hypoxemic pneumonia due to COVID-19.
The secondary outcome was to evaluate the clinical, biological, and computed tomography scan characteristics of patients with COVID-19 pneumonia complicated by co-infection.

Statistical analysis
Results were expressed as total numbers (percentages) for categorical variables and as medians [25th-75th percentiles] for continuous variables. Categorical variables were compared using the Chi-square test or the Fisher exact test, as appropriate. A P-value < .05 was considered significant. Analyses were performed using SAS statistical software (8.2, Cary, NC, USA).

Population
In summary, 156 patients were admitted to our hospital for COVID-19 over the study period. Among the 156 patients median age was 50 (33-62) years old, 83 were male gender (53%), 36 had hypertension (23%), 47 had body mass index > 25 kg/m 2 (30%) and 22 had diabetes mellitus (14%). A total of 36 patients were found to have hypoxemic pneumonia (23.1%). Characteristics of patient with hypoxemic pneumonia are described in Tables 1 and 2 Cytobacteriological examination of the sputum sample was performed in 16 patients. One patient tested positive for methicillin-susceptible Staphylococcus aureus, and another tested positive for Branhamella catarrhalis (this patient had a coinfection with the influenza virus A H1N1).
Lastly, 15 pneumococcal and legionella urinary antigen tests, 24 serologies for Mycoplasma pneumoniae and Chlamydia pneumoniae were performed. All these tests gave negative results.
Four patients (10.5%) (all of whom were hospitalized in intensive care unit) developed a nosocomial infection: 2 developed central venous catheter-related infection due to coagulase-negative Staphylococci (1 of which was associated with bacteremia), 2 developed bacteremia secondary to ventilator-associated pneumonia due to Pseudomonas aeruginosa in the first case and to Burkholderia cepacia in the second case. Median in-hospital length of stay was not significantly different between the 2 groups of patients: 26 [19][20][21][22][23][24][25][26][27][28][29][30][31] days in the group with nosocomial infection and 12 [10][11][12][13][14][15] days in the group without nosocomial infection (P = .09). Ten of the 36 patients with hypoxemic pneumonia were admitted to ICU (27.8%). Of these, 3 patients received high-flow nasal cannula oxygen therapy and 2 received invasive mechanical ventilation. Among the 36 patients, there were no deaths at follow-up. At the time of writing, all hospitalized patients were discharged from hospital.

Discussion
The incidence of co-infection in our patients with hypoxemic pneumonia in Reunion Island was higher than expected (13.9%). In the published literature, the incidence of co-infection in patients with COVID-19 pneumonia is highly variable, ranging from 0% to 20%. [8][9][10][11][12][13] To date, studies that explored cases of coinfection between COVID-19 and other microorganisms were mainly carried out in Asia. [8][9][10][11][12][13] It appears that the incidence of co-infection was higher in studies that evaluated intensive care patients. [9,10] The different microorganisms isolated were very different according to the studies with a high proportion of Mycoplasma pneumoniae in some studies and especially in China. [9][10][11][12][13] The Influenza virus was also one of the most common pathogens responsible for coinfection. [9,11] It should be noted that these microorganisms are treatable and known to be true pathogens of the upper airways.
In our study, 2 different bacteria (1 strain of Staphylococcus aureus and 1 strain of Branhamella catarrhalis) were isolated in Table 1 Demographic characteristics and comorbidities at hospital admission of the 36 patients with hypoxemic pneumonia due to COVID-19.  (20) .55

Co-infection
Results are expressed at n (%) or median [25th-75th] as appropriate. NYHA = New York Heart Association. * included any immunosuppressive diseases, such as congenital or acquired immunodeficiency, hematologic diseases, treatment with immunosuppressive drugs within the previous 30 days, or corticosteroids in daily doses of at least 10 mg/day of a prednisone equivalent for more than 2 weeks,. † angiotensin-converting enzyme or angiotensin II receptor blockers. www.md-journal.com Table 3 Chest computed tomography scan characteristics of 34 patients with hypoxemic pneumonia due to COVID-19.  sputum samples taken from 2 patients. This is unsurprising, as viruses, and especially the influenza virus, are known to be frequently associated with bacteria such as Staphylococcus aureus. [16] The only biomarker to be associated with co-infection in our cohort was D-dimer levels. Several studies have found D-dimer levels to be a prognostic marker of severity in COVID-19 infection. [17] In view of this, patients with severe COVID-19 infection and high D-dimer levels should be treated without delay, ideally with an antibiotic not active against Pseudomonas aeruginosa (e.g., 3rd generation cephalosporin or amoxicillin/ clavulanate). Alternatively, oseltamivir treatment may be administered to patients at risk of developing co-infection with the Influenza virus.

Co-infection
Azithromycin may be an interesting molecule for the treatment of COVID-19 patients with co-infection. Indeed, azithromycin has a broad antibacterial spectrum, which is active against atypical germs like Mycoplasma pneumoniae. [18] Moreover, it has well-known anti-inflammatory activity and may be effective against COVID-19 even in the absence of pyogenic infection. [18] It should be noted, however, that the hydroxychloroquine/ azithromycin treatment was not associated with better outcome in some studies. [19,20] Our study has many limitations that must be acknowledged. First, our cohort was relatively small. Another limitation of our study is that our microbiological results are difficult to extrapolate because of the variable distribution of infective microorganisms across geographical areas. While the co-infections analyzed in our study were diagnosed in the southern hemisphere, the vast majority of our patients had recently traveled to the northern hemisphere, making it almost impossible to trace the source of their coinfection. It is indeed difficult to determine whether a patient coinfected with Influenza caught the virus in Europe, where Influenza season in nearing its end, or in Reunion Island, where the virus is present all year round with a peak during the southern winter (May to October). [21,22] Only data concerning patients with hypoxemic pneumonia were collected. However, it is not recommended to initiate an anti-infective treatment only in a patient without severe pneumonia due to COVID-19 (i.e. hypoxemic pneumonia). [23] It is difficult to ascertain whether the infectious agents are actually implicated in disease, rather than occupying an ecological niche in the nasopharynx or the lung without causing disease. Nevertheless, in clinical practice, it is difficult not to treat the presence of some microorganisms as Staphylococcus spp or Influenza in a patient with severe pneumonia. Lastly, few of our patients underwent all the microbiological diagnostic tests (sputum cultures, serologies, antigen tests). Cytobacteriological examination would have likely been more sensitive had we used deep respiratory samples instead of sputum samples. However, we preferred not to subject our patients to invasive procedures when not entirely necessary, especially since the vast majority of them did not require invasive mechanical ventilation.
Likewise, it may be that multiplex PCR would have performed better on deep respiratory samples. Yet, studies comparing the diagnostic performance of PCR on nasopharyngeal vs. deep respiratory samples have found no major difference between the 2 techniques, except in the case of some microorganisms such as Legionella pneumophilia. [24]

Conclusion
The incidence of co-infection in our patients with hypoxemia pneumonia due to COVID-19 was higher than expected (13.9%).
COVID-19 patients who present with hypoxemic pneumonia should be be screened for co-infection using respiratory cultures or multiplex PCR in order to isolate treatable pathogens like influenza virus or pyogenic germs. Whilst our study has a number of limitations, the results from our study suggest that in the absence of screening, patients should be commenced on treatment for co-infection in the presence of an elevated D-dimer. Larger studies should be conducted to determine the clinical, biological, and radiological characteristics that indicate the presence of coinfection in patients with COVID-19.