Introduction

Pneumocystis jirovecii (P. jirovecii) is an opportunistic fungus responsible for Pneumocystis pneumonia (PCP) in deeply immunocompromised patients. Spontaneous evolution of PCP is fatal and this invasive fungal disease was a major cause of morbidity and mortality among HIV-infected people during the 80 and 90 s [1]. Nowadays, patients with hematological malignancy (HM), solid-organ tumor, solid-organ transplantation, primary immune deficiency and/or receiving long-term (> 3 months) or high-dose (> 0.5 mg/kg/day) corticosteroids or other immunosuppressive drugs present an increased risk of develo** PCP [2]. Thus, the at-risk population is now clearly recognized and PCP incidence in immunocompromised HIV-uninfected patients with no prophylaxis is estimated between 5 and 15% [3]. Pneumocystis detection in individuals with mild immunodeficiency or impaired respiratory function may also reflect pulmonary colonization. Pneumocystis colonization has been described in patients with chronic lung diseases such as chronic obstructive pulmonary disease but also occurs in immunosuppressed population [4]. Meanwhile, Herpesviridae/Herpesviruses (HVs), such as Cytomegalovirus (CMV), Epstein-Barr virus (EBV), Herpes simplex viruses type 1 and type 2 (HSV-1 and  -2) or Varicella zoster virus (VZV) can also be responsible for severe pneumonia, especially in critically ill patients [5]. Moreover, rising numbers of critically ill patients are henceforth immunocompromised and account for approximately 30% of all ICU admissions [6]. Pulmonary co-infections with P. jirovecii and CMV have widely been described [7,8,9,10,11,12,12, 17]. Neutropenia induced by antiviral treatment (ganciclovir) would especially increase the risk of develo** co-infections [17].

There are conflicting reports regarding morbidity and mortality due to CMV co-infection during PCP. Some authors reported no significant difference of morbidity and mortality rates between PCP patients develo** a simultaneous CMV pneumonia and those with no viral pneumonia [9, 20,21,22,23]. However, the presence of CMV in BALF samples from HIV-infected patients with PCP has been correlated with a poor prognosis [10, 24] and numerous studies described a relationship between CMV infection and increased mortality rate in critically ill and SOT patients who developed PCP [11, 18, 25,26,27,28]. In the present study, we assessed the relationship between different HVs and the mortality rate. Although the 30-day mortality did not vary significantly according to the result of HV detection, our data indicated that the combination of [EBV + CMV] was significantly associated with an increased 30-day mortality rate in PCP patients compared to those without HV, with EBV, [EBV + CMV] and [EBV + HSV-1] (p = 0.0017). This increased mortality appears to be independent from the deep of the underlying immunodeficiency and rather linked to the specific [EBV + CMV] co-infection. Moreover, the absence of antiviral treatment was significantly associated with an increased day-30 mortality rate in PCP patients with at least one Herpesviridae detected in BALF. These results suggest that HV infections are important to test and treat in patients with PCP.

The diagnosis of pulmonary mixed infections remains challenging but is crucial because clinical manifestations are severe and associate with poor prognosis [14]. Pulmonary infections associating P. jirovecii and more than one viruses has already been described [13, 19, 27, 29]. Maartens et al. conducted a prospective study including 284 HIV-infected patients with pulmonary symptoms [30]. Simultaneous testing for community-acquired bacteria and viruses, mycobacteria, CMV and P. jirovecii was performed using PCR assays in induced sputa from all patients. Various respiratory viruses were detected in 203/284 patients (71.5%) with the highest prevalence in PCP patients (22/26, 85%). CMV was detected in 5/26 patients with PCP (19.2%) and multiple co-infections with respiratory viruses such as Human metapneumovirus A/B, Enterovirus, Influenza A or Parainfluenza virus were identified in PCP patients. In our study, Pneumocystis and HVs detections were performed in either a single BALF sample or two BALF samples collected less than two days apart. Thus, we demonstrated the simultaneous presence of HVs and P. jirovecii in 70.1% of PCP patients and in 58.3% of colonized patients. Moreover, at least two HVs were detected in one-third of BALF samples, regardless of the Pneumocystis status. In this context, syndromic molecular testing targeting the main pulmonary microorganisms represents an interesting method to reduce time and cost related to pre-analytical process. The usefulness of syndromic approach has already been assessed with a multiplex PCR TaqMan® array card (ThermoFisher) which simultaneously detects 24 viruses (including CMV and HSV 1 + 2), eight bacteria and two fungi (including P. jirovecii) [31] and with the “FTD respiratory pathogens 33” kit (Fast-Track Diagnostics, Esch-sur-Alzet, Luxembourg) targeting 33 microorganisms (including P. jirovecii and CMV) [30]. Recent studies proposed next-generation sequencing methods for a large-scale approach [13, 27, 29, 32]. These techniques allow thus prompt management of polymicrobial infections frequently associated with a poor prognosis.

Conclusion

Our study describes the detection of HVs in the lungs from Pneumocystis-infected patients with either PCP or colonization over a 7-year period. Distribution of each HV, alone or in combination was similar whatever the clinical form of Pneumocystis infection and EBV was the most frequent virus. Multiple viral infections were also observed in one-third of patient from each group. In addition, our results suggest that an increased 30-day mortality rate can be observed in PCP patients, especially in the case of [EBV + CMV] co-infection and in the absence of antiviral treatment.