To the editor,

The role of respiratory viruses, including influenza and non-influenza viruses, in severe lower respiratory tract infections (LRTI) is of increasing concern. However, the approaches for their diagnosis and management may vary among intensivists given the lack of guidance. We consequently aimed at describing the practice patterns of intensivists worldwide using a 39-question self-administered online questionnaire (Additional file 1: Appendix S1). After approval from Research Committee members of the European Society of Intensive Care Medicine (ESICM), the questionnaire was openly accessible on its website from January 24 to June 4, 2019. There were 229 respondents from 53 countries, mainly in Europe (78%), who worked predominantly in mixed intensive care units (74%) (Additional file 1: Table S1).

In case of an acute respiratory failure suggestive of community-acquired LRTI, 83% of the respondents considered searching for respiratory viruses. This attitude was encouraged in case of pre-existing respiratory condition, immunocompromised status or presence of coryza/rhinorrhea. Interestingly, more than three-quarters of respondents confirmed their attitude even in the absence of influenza-like illness or recent close contact with somebody presenting influenza-like illness.

Nucleic acid amplification tests, such as multiplex polymerase chain reaction (mPCR), were available routinely (95%), and largely used as the first-line test to search for respiratory viruses. The test was available only during the opening hours in half of the cases (56%), with a turnaround time below 2 h in only 11% of cases. Large panels (> 5 viruses) and those including atypical bacteria were available for 70% and 47% of the respondents, respectively. Most of them (92%) had some knowledge of the panel composition, but 53% ignored the mPCR cost. The nasopharyngeal swab was the preferred type of proximal sample (66%). In intubated patients, a proximal sample (27%), a distal sample (45%) or both (29%) were collected. Interestingly, 62% of the respondents considered repeating the mPCR in case of negative results despite a high clinical suspicion of viral LRTI (Table 1 and Additional file 1: Table S2).

Table 1 Main results of the survey
Full size table

There were wide variations regarding the putative causal role of non-influenza respiratory viruses in severe community-acquired pneumonia (Additional file 1: Figure S1A), but most respondents (96%) considered bacteria–virus coinfection as a risk factor of severity, in line with previous reports [1, 2]. Only 29% of the respondents systematically applied measures to prevent droplet transmission regardless of the season, whereas 49% did so only during the epidemic Flu season. In case of viral documentation, the application (or continuation) of confinement measures depended on the viral species documented, in parallel with its putative causal role (Additional file 1: Figure S1B). Strikingly, conditions of interruption of confinement measures varied widely among respondents. A majority (65%) interrupted them after a specified period (ranging from 5 to more than 10 days), whereas 24% did so after the complete resolution of both the fever and respiratory symptoms and 11% after a negative result of an additional respiratory mPCR (Additional file 1: Table S3).

In case of severe community-acquired pneumonia with documentation of a non-influenza respiratory virus (i.e., Respiratory Syncytial Virus), 77% of the respondents did not consider to prescribe antiviral treatment (i.e., ribavirin) if the patient was non-immunocompromised. Finally, in case of no bacterial documentation despite usual microbiological investigations, regardless of any additional information relating to blood tests, only 30% of respondents declared that the viral documentation encouraged them (certainly of likely) to stop antibiotics early (Additional file 1: Table S4).

This survey has several limitations. First, respondents were mainly from Europe, so caution should be exercised on generalizing our results. Second, the survey response rate was unknown, because the questionnaire was openly accessible on the ESICM website. Therefore, how our cohort of respondents was representative of intensivists worldwide is questionable. In particular, we may suspect that physicians interested in respiratory viruses were more prone to participate. This may have led to an overestimation of mPCR and confinement measures prescription rates. Third, some interesting points were not addressed in our questionnaire, in that we decided to limit to 40 items. For example, it would have been interesting to collect additional data regarding the characteristics of the respondents (gender, years of experience in critical care, proportion of immunocompromised patients in their ICU population). The influence of biomarkers such as procalcitonin on antimicrobial treatments could also have been addressed.

In this survey, we observed that a majority of intensivists (i) searched for respiratory viruses in case of severe community-acquired LRTI in adults, whatever the season, the medical history and clinical presentation; (ii) had access to large-panel respiratory mPCR; (iii) used them as first-line diagnostic test in routine practice; (iv) had some knowledge about the panel of the mPCR that they use, but markedly less about the cost. However, we observed strong heterogeneity regarding how intensivists took into account mPCR results for infection control (confinement measures). Moreover, the impact of the mPCR results on patient care (antiviral treatment and antibiotics management) looked limited. These findings highlight the need for further studies in the field of respiratory viruses, aiming to better integrate nucleic acid amplification tests and their results in the management of severe LRTI in adults.