Immunosuppression and herpes viral reactivation in intensive care unit patients: one size does not fit all

More than 20 years after the initial description of cytomegalovirus (CMV) pneumonia in “non-immunocompromised” ICU patients by Papazian et al. [1], the treatment of herpes viruse reactivation in ICU is still a matter of debate. Recently, Mirouse et al. [2] reported a unique cohort of varicella-zoster virus (VZV)-related community acquired pneumonia in the ICU over 20 years. This highlighted that VZV infections in ICU patients are rare, with roughly a hundred cases over 20 years, in 29 French ICUs. Conversely, addressing a wider range of herpes viruses, two major papers recently reported a high frequency of viral reactivation in critically ill patients, and both showed variable plasma levels of various herpes viruses (Table 1) [3, 4]. Interestingly, almost 30% of the patients had multiple viremia events, and this viremia usually lasted until ICU discharge. The timing of viral reactivation was also informative, with herpes simplex virus (HSV)1 and Epstein-Barr virus (EBV) being detected earlier in the ICU course than CMV or human herpesvirus (HHV)6. Several herpes viruses have been associated repeatedly with mortality and the occurrence of secondary infections. The most validated association is between CMV and mortality, either in ICU or at various times postICU [3–5]. In the MARS cohort, this association remained significant after adjusting for confounders, time-dependent bias, and competing risks [4]. Importantly, this association between CMV viremia and ICU mortality was significant (adjusted sdHR = 3.2 [1.4–7.1]) while taking into account other viral reactivation, which was never considered before [4]. Multiple reactivations might also suggest increased severity since the association of HHV6 and CMV reactivation was also associated with an increased risk of death in critically ill patients [6]. Association between EBV reactivation and mortality was also described [4, 7]. Data are less clear for other neurotropic herpes viruses such as HSV2 or VZV, which exhibit lower reactivation rates. Finally, in septic patients, CMV/EBV and HSV1 have also been associated with an increased rate of secondary fungal and bacterial infections, respectively [4]. This last result supports the hypothesis that immunosuppression might play an important role in this reactivation. Most data related to critically ill patients made a clear distinction between immunocompromised (mainly oncologic treated patients, solid or bone marrow transplants) and so-called immunocompetent patients. However, we know that 30 to 50% of these critically ill patients exhibit signs of immunosuppression [8]. Clearly, the wording immunocompetent—to identify a category of patients that do not exhibit a severe iatrogenic or congenital immunosuppression—is misleading and it is time to propose several levels of immunosuppression (in terms of depth or type of immunosuppression) to better reflect the various levels of depressed immune status in ICU patients. Herpes viremia might therefore reflect either a lack of a latent virus control or a true clinical viral infection. The ability to describe several types or levels of immunosuppression is important and supports a wider use of quantitative tools to measure viral titers in addition to host response biomarkers of the immune * Correspondence: [email protected] EA7426 “Pathophysiology of Injury-Induced immunosuppression”, Hospices Civils de Lyon—Université Claude Bernard Lyon 1—bioMérieux, Lyon, France Anesthesiology and Critical Care Medicine, Hospices Civils de Lyon—Université Claude Bernard Lyon 1, Lyon, France