Annals of Intensive Care (Oct 2023)

Prognosis of critically ill immunocompromised patients with virus-detected acute respiratory failure

  • Guillaume Dumas,
  • Maxime Bertrand,
  • Virginie Lemiale,
  • Emmanuel Canet,
  • François Barbier,
  • Achille Kouatchet,
  • Alexandre Demoule,
  • Kada Klouche,
  • Anne-Sophie Moreau,
  • Laurent Argaud,
  • Florent Wallet,
  • Jean-Herlé Raphalen,
  • Djamel Mokart,
  • Fabrice Bruneel,
  • Frédéric Pène,
  • Elie Azoulay

DOI
https://doi.org/10.1186/s13613-023-01196-9
Journal volume & issue
Vol. 13, no. 1
pp. 1 – 11

Abstract

Read online

Abstract Background Acute respiratory failure (ARF) is the leading cause of ICU admission. Viruses are increasingly recognized as a cause of pneumonia in immunocompromised patients, but epidemiologic data are scarce. We used the Groupe de Recherche en Réanimation Respiratoire en Onco-Hématologie’s database (2003–2017, 72 intensive care units) to describe the spectrum of critically ill immunocompromised patients with virus-detected ARF and to report their outcomes. Then, patients with virus-detected ARF were matched based on clinical characteristics and severity (1:3 ratio) with patients with ARF from other origins. Results Of the 4038 immunocompromised patients in the whole cohort, 370 (9.2%) had a diagnosis of virus-detected ARF and were included in the study. Influenza was the most common virus (59%), followed by respiratory syncytial virus (14%), with significant seasonal variation. An associated bacterial infection was identified in 79 patients (21%) and an invasive pulmonary aspergillosis in 23 patients (6%). The crude in-hospital mortality rate was 37.8%. Factors associated with mortality were: neutropenia (OR = 1.74, 95% confidence interval, CI [1.05–2.89]), poor performance status (OR = 1.84, CI [1.12–3.03]), and the need for invasive mechanical ventilation on the day of admission (OR = 1.97, CI [1.14–3.40]). The type of virus was not associated with mortality. After matching, patients with virus-detected ARF had lower mortality (OR = 0.77, CI [0.60–0.98]) than patients with ARF from other causes. This result was mostly driven by influenza-like viruses, namely, respiratory syncytial virus, parainfluenza virus, and human metapneumovirus (OR = 0.54, CI [0.33–0.88]). Conclusions In immunocompromised patients with virus-detected ARF, mortality is high, whatever the species, mainly influenced by clinical severity and poor general status. However, compared to non-viral ARF, in-hospital mortality was lower, especially for patients with detected viruses other than influenza.

Keywords