Introduction: Viral infection among solid organ transplant (SOT) and hematopoietic stem cell transplantation (HSCT) patients has been described in previous studies, with respiratory viral infection (RVI) among this immunocompromised population being associated with increased mortality and morbidity.1-3 Immunocompromised patients with RVI demonstrated prolonged viral shedding and were at higher risk of progression to lower respiratory tract infection (LRTI), which can lead to complications including graft dysfunction, graft loss, and bronchiolitis obliterans among lung transplant recipients.1 Early treatment is crucial in improving recovery from RVI. Therefore, a better understanding of diagnostic methods of viral detection may lend insight to clinical management.1
The Bauer study was a prospective multinational observational study of 1611 immunocompromised patients with acute hypoxemic respiratory failure (ARF) admitted to the ICU.4 A secondary analysis of the Bauer study examined the differential results for patients who had undergone NIT versus NIT-FOB.5 We conducted further investigation of the results from the secondary analysis of the Bauer study. Our objective was to determine whether techniques used to diagnose patients with non-CMV viral illness resulted in differential patterns of non-CMV viral infection diagnosis. This is the first such report, to the authors’ knowledge, to perform this sub analysis of the Bauer study data. 
Materials and Methods:
Data Collection : The data obtained from the secondary analysis of the Bauer study was narrowed down through the following process:
1)  Examined Table 6S, which lists the primary and secondary diagnoses for patients who underwent NIT versus NIT-FOB5
2)  Resolved the data from Table 6S to only include data from patients with primary diagnoses of non-CMV viral infection. (One subject with Herpes Simplex Virus was excluded since no serotype was reported.) 
3)  Further resolved data to analyze a cohort of non-CMV and non-influenza patients
Statistical Analysis: The statistical analysis included Fisher’s Exact Test, two-sample proportion hypothesis test, as well as descriptive analysis of the data. A p-value of <0.05 was considered significant. The data comparison included the following comparisons, with a description of yield as percentage and number:
1)  NIT and NIT-FOB cohorts, non-CMV viral infection patients
2)  NIT and NIT-FOB cohorts, non-CMV viral infection and non-influenza patients
Results:
TABLE 1 - Description of primary diagnoses of immunocompromised patients with ARF using NIT-FOB versus NIT. Data was pooled together (a specific virus vs. all the rest of patients in the group) when constructing contingency tables for Fisher’s exact test.
A two-sample proportion hypothesis test confirms that diagnosis of non-CMV viral infection was more frequent among patients in the NIT-FOB group compared to in the NIT group (p < 0.00001 with Yates’s correction for continuity).
The NIT-FOB group also showed a higher frequency of rhinovirus diagnosis (Fischer’s p = 0.0108) and respiratory syncytial virus (Fischer’s p = 0.0377). Respiratory syncytial virus (RSV) was notably the most commonly diagnosed non-CMV, non-influenza viral infection in both the NIT-FOB and NIT groups.
Finally, the NIT-FOB group had a higher presentation influenza (Fischer’s p = 0.0001)
Discussion:  Our analysis of non-CMV viral infection between two patient cohorts revealed that compared to the NIT cohort, the NIT-FOB cohort had a higher presentation of RSV, Rhinovirus, and Influenza. 
The exact mechanism leading to this observed difference in non-CMV viral LRTI expression may be multifactorial. Although both  the cohorts we analyzed demonstrated severe respiratory compromise, the use of flexible bronchoscopy to establish the diagnosis of LRTI in the ICU may reflect the severity of illness, the necessity of ruling out luminal obstruction, or the placement of an endotracheal tube (which facilitates the use of FOB in this population), among other indicators. Thus, while the higher presentation of RSV, Rhinovirus, and Influenza in the NIT-FOB cohort may be due to FOB increasing exposure to viral agents as it accesses the distal airway, patients in the NIT-FOB cohort may simply have been more immunocompromised and prone to infection by these viral agents. 
The secondary analysis of the Bauer study found that FOB does improve diagnostic success for immunocompromised patients with ARF, with bronchoscopy yielding a diagnosis in approximately half the cases it was carried out.5 However, this study also found that FOB was associated with increased ICU and hospital mortality, even after measures were taken to adjust for disease severity and other potential confounding variables.5This increased mortality is somewhat paradoxical, as in general, successful identification of the cause of respiratory failure has been shown to be associated with better outcome.4 This information highlights the necessity of a focused effort on improving the resolution of non-invasive diagnostic tests. 
If the differential presentation of rhinovirus, RSV and influenza between the NIT and NIT-FOB groups is not due to discrepancies in disease burden, improving non-invasive diagnostic techniques for these three viral infections may be crucial to lowering hospital mortality for patients with hypoxemic respiratory failure. More precise noninvasive diagnostic methods are needed, and this study suggests that rhinovirus, RSV and influenza detection be prioritized in the improvement of noninvasive diagnostic techniques for viral infection. 
Furthermore, our study shows that no significant differences in diagnosis of other non-CMV viruses was observed. Thus non-invasive methods may be preferable to either rule out or establish a diagnosis of other non-CMV viral infections.