Characteristics of viral pneumonia in non-HIV immunocompromised and immunocompetent patients: a retrospective cohort study

Background Concerning viral pneumonia, few large-scale comparative studies have been published describing non-HIV immunocompromised and immunocompetent patients, but the epidemiological characteristics of different viruses or underlying diseases in immunocompromised hosts are lacking. Methods We retrospectively recruited patients hospitalised with viral pneumonia from six academic hospitals in China between August 2016 and December 2019. We measured the prevalence of comorbidities, coinfections, nosocomial infections, and in-hospital mortalities. Results Of the 806 patients, 370 were immunocompromised and 436 were immunocompetent. The disease severity and in-hospital mortality of immunocompromised patients were higher than those of immunocompetent patients. During the influenza season, an increased number of cases of influenza virus (IFV) infection were found in the immunocompromised group, followed by cases of cytomegalovirus (CMV) and respiratory syncytial virus (RSV) infection. During the non-influenza season, CMV was the main virus detected in the immunocompromised group, while RSV, adenovirus (AdV), parainfluenza virus (PIV), and rhinovirus (HRV) were the main viruses detected in the immunocompetent group. Pneumonia caused by Pneumocystis jirovecii (22.4%), Aspergillus spp. (14.1%), and bacteria (13.8%) were the most frequently observed coinfections in immunocompromised patients but not in immunocompetent patients (Aspergillus spp. [10.8%], bacteria [7.1%], and Mycoplasma spp. [5.3%]). CMV infection and infection with two-or-more viruses were associated with a higher in-hospital mortality rate than non-IFV infection. However, patients with IFV and non-IFV infection in immunocompromised patients had similar disease severity and prognosis. Conclusions Immunocompromised patients have a high frequency of coinfections, and a higher mortality rate was observed among those infected with CMV and two-or-more viruses. In addition, patients with IFV and non-IFV infection in immunocompromised patients had similar same disease severity and prognosis. The type of viral infection varied with seasons. Supplementary Information The online version contains supplementary material available at 10.1186/s12879-021-06437-5.

Conclusions: Immunocompromised patients have a high frequency of coinfections, and a higher mortality rate was observed among those infected with CMV and two-or-more viruses. In addition, patients with IFV and non-IFV infection in immunocompromised patients had similar same disease severity and prognosis. The type of viral infection varied with seasons.
Keywords: Viral pneumonia, Immunocompromised, Immunocompetent, Prognosis Background Among transplant recipients and patients with haematological malignancy, viral pneumonia often leads to severe respiratory disease and death [1]. Viral lower respiratory tract infections in immunocompromised patients have generally been ascribed to herpes virus (HSV) and cytomegalovirus (CMV) [2]. In recent years, influenza virus (IFV), parainfluenza virus (PIV), respiratory syncytial virus (RSV), and rhinovirus (HRV) have also been recognised as causes of serious infections, especially in patients undergoing treatment for haematologic malignancies and haematopoietic stem cell transplantation. These patients have a higher tendency to develop severe pneumonia, and a mortality rate as high as 25-70% has been reported [3][4][5][6][7]. These patients might experience prolonged viral shedding that potentially result in a longer duration of infection, a higher nosocomial transmission rate, and a higher mortality rate than those of immunocompetent hosts [8,9]. Concerning non-HIV immunocompromised patients with viral pneumonia, few large-scale epidemiological studies and comparative studies have investigated different viruses or underlying diseases; however, investigations on the epidemiological and etiologic characteristics are lacking.
The objective of this study was to examine the epidemiological and etiologic characteristics and to identify the most common types of viruses that cause viral pneumonia in non-HIV immunocompromised and immunocompetent patients.

Study design and participants
We retrospectively recruited patients with communityacquired pneumonia (CAP) who were hospitalised between August 2016 and December 2019 at one of the six secondary and tertiary academic hospitals in China. The diagnosis of CAP was based on the American Thoracic Society and Infectious Disease Society of America (ATS/ IDSA) guidelines [10]. Immunocompromised patients were selected if they met any of the following inclusion criteria: (1) solid-organ, stem cell, or bone marrow transplant recipients; (2) undergoing chemotherapy for any haematological disease (including acute lymphocytic leukaemia, acute myeloid leukaemia, chronic lymphocytic leukaemia, myeloma, or lymphoma) or the presence of a solid tumour within 6 months of admission or neutropenia (neutrophil count < 500 cells/mm 3 ); (3) chest radiation therapy within 3 months of admission; (4) an autoimmune disease (including but not limited to systemic lupus erythematosus, rheumatoid arthritis, polymyalgia rheumatica, and interstitial lung disease) and receiving immunosuppressive therapy (including chronic glucocorticoid treatment: oral prednisone > 10 mg/d or the equivalent for ≥3 weeks) or methotrexate > 12.5 mg/week, cyclosporine, azathioprine, or biological modifiers such as etanercept or infliximab within 3 months of admission; and (5) history of splenectomy or cirrhosis [1,11,12]. Patients were excluded if they (1) were aged < 14 years, (2) experienced pneumonia onset ≥48 h after admission, or (3) tested positive for human immunodeficiency virus.

Study quality control
Key investigators, including clinicians, statisticians, microbiologists, and radiologists, worked together to draft the protocol and created a single formatted case report form (CRF) that was used by all centres. Before the initiation of the study, all investigators from the six centres received training related to the study protocol, including the screening process, definitions of underlying diseases, and the format of the CRF. After data were collected, CRFs were reviewed by a trained researcher to ensure completeness and data quality.

Data collection
The data were collected and included information on patient and disease characteristics, initial oxygenation strategy, laboratory and microbiological data (blood, nasopharyngeal swabs, sputum, and/or bronchoalveolar lavage samples; bacterial or fungal cultures; viral nucleic acid detection; and antibiotic susceptibility patterns), associated organ dysfunction, and patient outcomes at hospital discharge.

Microbiological methods
Microbiological samplings were performed, bronchoalveolar lavage (BAL) or sputum samples were obtained by the treating physicians, and microorganisms were identified and tested for drug susceptibilities. Bronchoscopic examinations were performed according to general guidelines. Lidocaine spray was applied to the upper airway and carina as a local anaesthetic, and airways were thoroughly examined. BAL was performed by administering 60-120 mL of sterile saline solution 2-4 times into the distal bronchial tree, either at the affected lobe or in the middle lung lobe with more radiographic abnormalities. BAL specimens were aliquoted and immediately transported to laboratories. Sputum, BAL samples, or nasopharyngeal swabs were used for atypical pathogen and viral polymerase chain reaction (PCR) amplification tests.

Pathogen-specific diagnostic criteria
To diagnose pneumonia caused by Aspergillus, one or more of the following criteria were required: (1) histopathologic or direct microscopic evidence of dichotomous septate hyphae with a positive culture for Aspergillus from tissue sample, (2) a positive Aspergillus culture from BAL fluid sample, (3) a galactomannan optical index in BAL fluid ≥1, (4) a galactomannan optical index in serum ≥0.5; (5) Aspergillus species identified on culture and microscopically [13,14].
The diagnosis of Pneumocystis jirovecii pneumonia (PCP) required one of the following: (1) high-resolution computed tomography imaging showing diffuse ground glass opacity with patchy distribution; (2) microscopic examination of the respiratory sample revealing the presence of Pneumocystis cystic or trophic forms; or (3) a positive PCR test result for Pneumocystis [15].
Coinfection was considered if bacteria or fungi were isolated from lower respiratory tract specimens (qualified sputum, endotracheal aspirate, and BAL) within 48 h of hospitalisation. A nosocomial infection was diagnosed when patients showed clinical signs or symptoms of pneumonia or bacteraemia and had a positive culture of a new pathogen obtained from lower respiratory tract specimens and/or blood samples taken ≥48 h after admission.

Statistical analysis
The demographics, clinical characteristics, and pathogen testing results are expressed as the mean (± standard deviation), median (interquartile range), or number (percentage). Group comparisons were conducted using Student's t-test or the Wilcoxon rank-sum test for continuous variables with and without normal distributions, respectively. Categorical variables of the two groups were compared using the χ 2 test.
Statistical analyses were performed using SPSS version 19.0 (SPSS, Inc., Chicago, Illinois). All tests were twosided, and P-values < 0.05 were considered statistically significant.

Patient and public involvement
No patient or the public were involved in the development of the research question, study design, recruitment, and the conduct of the study.

Results
A total of 860 adult patients with positive respiratory viral nucleic acid test results were selected. After excluding patients with upper respiratory tract infections (n = 24) and those who failed to meet the diagnostic criteria for pneumonia (n = 30), 806 patients with viral pneumonia were included in the final analysis. These included 370 immunocompromised and 436 immunocompetent patients. Approximately 34.3% (127/370) of the immunocompromised patients were women with a median age of 60 years. The main presenting symptoms were fever (74.6%), cough (92.4%), and dyspnoea (66.2%). The most common underlying immune-related diseases were connective tissue disease (36.2%), interstitial lung disease (44.6%), solid-organ transplantation (16.2%), and nephrotic syndrome or chronic glomerulonephritis (12.4%). Ddimer levels, pneumonia severity index (PSI) scores, rates of non-invasive mechanical ventilation, septic shock, and in-hospital mortality were higher in the immunocompromised group than in the immunocompetent group (P < 0.05) ( Table 1).
During the influenza season (November, December, January, and February), an increase in the number of IFV infection cases (22.4%) was found in the immunocompromised group, followed by CMV (15.4%) and RSV (13.0%) infection cases. In the immunocompetent group, IFV (43.5%) was most frequently detected, followed by RSV (14.9%). During the noninfluenza season, CMV (42.7%) was the main virus detected in the immunocompromised group. However, in the immunocompetent group, there was no dominant virus; the order of detection was as follows: IFV (9.4%), PIV (7.6%), AdV (7.3%) HRV (7.1%), and RSV (5.7%) ( Table 2 and Figs. 1 and 2). Regarding coinfections in immunocompromised patients, PCP (22.4%), Aspergillus (14.1%) and bacteria (13.8%) were most frequent, with Klebsiella pneumoniae (4.1%), Pseudomonas aeruginosa (3.0%), and Among the secondary nosocomial bacterial infections, Acinetobacter baumannii, P. aeruginosa, and K. pneumoniae were most commonly detected as causative agents ( Table 2). The CMV infection group had more patients with nephrotic syndrome and high rates of PCP infection and ground glass shadows on computed tomography (CT) (P < 0.05). In the non-IFV group, there were fewer patients who required non-invasive ventilator use and intensive care unit treatment than other groups. Further, the non-IFV group was associated with a lower in-hospital mortality rate than CMV and two-or-more viruses' groups. However, patients with IFV and non-IFV infection in immunocompromised patients had similar disease severity and prognosis (Table 3). Patients with nephrotic syndrome and chronic glomerulonephritis had the highest rate of CMV infection (89.1%), organ transplant patients had the highest rate of RSV infection (35.0%), patients with haematopoiesis diseases had the highest rates of AdV (22.7%) and HRV (18.2%) infections, and malignant solid patients with radiotherapy and chemotherapy had the highest rate of PIV infection (23.5%). Patients with nephrotic syndrome and chronic glomerulonephritis had a low oxygenation index and lymphocyte count, high rate of CMV and PCP infection, were more likely to require additional non-invasive ventilator use and intensive care unit treatment, and had a high in-hospital mortality rate. The in-hospital mortality rate of patients with connective tissue disease was the second highest (30%), while that of solidorgan transplantation patients was the lowest (10%) ( Table 4). Viral shedding was significantly longer in immunocompromised hosts than in immunocompetent hosts (Table 5).

Discussion
This study was a large-scale, multicentre, retrospective study of the aetiology of and clinical risk factors for CAP in immunocompromised patients. The main findings were as follows: (1) The disease severity and in-hospital mortality rate of immunocompromised patients were higher than those of immunocompetent patients; (2) during the influenza and non-influenza seasons, the distribution of viruses in the immunocompromised group differed; (3) among the coinfections of immunocompromised patients, PCP was the main pathogen, followed by Aspergillus and bacteria, and in the immunocompetent group, Aspergillus was the most common pathogen, followed by bacteria and Mycoplasma; (4) the in-hospital mortality rate of the non-IFV infection group was lower than those of the CMV group and the two-or-more viruses group, but had similar prognosis with IFV group; (5) the type of virus infection varied according to the underlying diseases detected; (6) viral shedding was significantly longer in immunocompromised hosts than in immunocompetent hosts.  In recent years, several studies have focused on respiratory virus infection in patients after haematopoietic cell transplantation (HCT) [16][17][18][19][20][21]. Sachiko studied HRV in the lower respiratory tract of patients with HCT and found that 55% of patients had coinfections and that the 90-day mortality rate was 41% [16], which was similar to that of lower respiratory tract infections caused by RSV, PIV, or IFV [17][18][19]. Among the immunocompromised patients with IFV pneumonia, approximately 60% had an associated infection with at least one other organism, and the mortality rate among these patients was 15-30% [20]. The mortality rate among haematologic malignancy patients with RSV is approximately 18%,  Fig. 1 Virus detection of immunocompromised hosts in different months and in HCT recipients who developed RSV lower respiratory tract infections, it can be as high as 83% [21]. Similarly, our study showed that the disease severity and in-hospital mortality (26.5% vs 18.8%) of immunocompromised patients were higher than those of immunocompetent patients. CMV, especially with PCP coinfection, has a high mortality rate in immunocompromised patients [22,23]. However, at present, there are few comparative studies examining CMV and other respiratory viruses. Our findings indicated that during the influenza season, IFV, CMV, and RSV were the main viruses detected in immunocompromised hosts, while during the non-influenza season, we need to pay attention to CMV, IFV, PIV, AdV, HRV and RSV as these were more readily detected. Non-CMV viral infections may also exist with a PCP coinfection, albeit less frequent. Comparably we found no difference in the rate of virus-Aspergillus coinfections irrespective of the type of viral infection [13,24].
The disease severity in, complications in, and outcomes of immunocompetent patients with CAP were similar between IFV-and non-IFV-related respiratory diseases [25][26][27]. We found that the in-hospital mortality rate was significantly higher in immunocompromised patients with CMV or two-or-more viral infections than the non-IFV infections. This suggests that when a viral infection is suspected in an immunocompromised patient, healthcare providers should also determine the presence of CMV and other viral aetiologies, as early diagnosis and treatment are essential in improving the outcomes. In addition, the highest mortality rate was observed among patients with nephrotic syndrome or chronic glomerulonephritis, for which there was a higher rate of CMV and PCP infection. This indicates that routinely screening for PCP and CMV infections should be considered for this group of patients. Moreover, the higher incidence of CMV and PCP and mortality rates associated with nephrotic syndrome patients may be related to the lack of routine prevention of infection when using immunosuppressants or glucocorticoids.
It has been suggested that viral respiratory infections in immunocompromised patients involve persistent viral shedding, rendering these patients contagious for prolonged periods [28][29][30]. Memoli et al. reported that the viral shedding period of immunocompromised patients was longer than that of immunocompetent patients with IFV pneumonia (19.04 vs. 6.38 days, respectively; P < 0.05) [29]. Virus detection for ≥30 days was reported in 29% of infected patients with haematological disorders [28]. In this study, we demonstrated that both influenza A virus subtype H1N1 and RSV infections had a longer viral shedding period in immunocompromised hosts, which made it necessary to extend the duration of antiviral therapy.
There were some limitations to this study. First, it had a retrospective design and might not have included all patients. Second, as it was a multicentre research, not every patient with pneumonia underwent a full array of pathogen testing. Therefore, pathogen identification and diagnosis could have been incomplete. Third, many patients had been previously administered antibiotics. Despite these limitations, our results were consistent with the literature and provide a detail insight into the clinical and pathogenic characteristics and outcomes of different viral infections in immunocompromised hosts.  The in-hospital mortality between non-IFV and IFV patients was not statistically different (P = 0.324), but the non-IFV group was associated with a lower in -hospital rate than that of CMV group and two or more viruses' group (P<0.05)