Viral Respiratory Infections After Allogeneic Hemaotpoietic Stem Cell Transplantation: Outcome and Prognosis

Influenza (FLU, parainfluenza (PIF), respiratory syncitial (RSV) virus, Adenovirus (ADV) and less frequently Epstein Barr virus (EBV), Herpes Simplex (HSV) and Cytomegalovirus (CMV) are responsible for VRI leading to a mortality rate of 5 to 30% after HSCT. New VRI have been recently described with paradoxical conclusions concerning their pathogenicity after HSCT: rhinovirus (RV), metapneumovirus (MPV), coronarovirus (CV), Human Herpes Virus 6 (HHV6) and poliomavirus KI (KI).

This study aims to describe VRI infections, outcomes and risk factors for 100-day mortality in patients who received a first HSCT in our center from 01/2007 to 10/2010. Diagnosis of VRI was performed on nasopharyngeal secretion (by nasopharyngeal aspiration (NPA)) or broncho-alveolar lavage (BAL). Viruses were detected by a multiplex PCR assay (RespiFinder19) that detects FLU, PIF, RSV, ADV, RV, MPV, CV (OC43, 229E, NL63), (and C. pneumoniae, M. pneumoniae, L. pneumophila, B. pertussis) prospectively from 09/2009 and retrospectively before this date. Real time quantitative PCR for ADV, HSV, CMV, HHV6, KI and EBV were also performed from 01/2007. Other bacterial and fungal infections were documented by usual method (sputum, BAL, aspergillus galactomannan antigenemia, radiological findings, L. pneumophila antigen and S. pneumonia urinary antigens). Factors associated with 100-day mortality were estimated using Cox proportional cause-specific Hazard models.

127 patients with at least one episode of VRI were studied. Median age was 41 years (range: 8 to 66) with 38% women. The majority of patients was transplanted for hematological malignancy (87%), after a reduced intensity conditioning regimen (66%) with peripheral blood stem cells (63%) [Bone marrow 22%, cord blood 15%]. VRI was essentially diagnosed by NPA (90%) and was associated with a documented bacterial or fungal co-infection in 12% of patients. At VRI diagnosis, 143 viruses were identified in the 127 patients (concomitant viruses in some patients). The distribution of virus was: RV (32%), CV (17%), PIF (17%), ADV (11%), RSV (9%), MPV (7%), HHV6 (5%), FLU (4%), CMV (3%), PM (3%), EBV (2%), HSV (1%). After excluding patients with several virus or another co-pathogen, distribution of virus was RV (28%), CV (12%), PIF (10%), ADV (7%), RSV (6%), MPV (5%), HHV6 (2%), FLU (2%), CMV (1%), PM (2%), EBV (1%), HSV (1%). Thirty-two patients had nosocomial VRI and 36 were hospitalized during VRI evolution. At diagnosis, lower respiratory tract was involved in 52% of patients and 14% required oxygenotherapy. Four patients were transferred into intensive care unit. Seven patients died within 30 days and 14 within 3 months leading to a 3-month survival at 89%. None of them had an isolated VRI as sole cause of death: VRI was associated with active GVHD in 3 patients, active GVHD and co-infection in 3 patients, progressive malignancy in 3 patients, engraftment failure and co-infection in 2 patients, acute non-respiratory organ failure in 3 patients. Virus distribution in the 14 patients who died within 3 months was: CV (n=3), HHV6 (n=3), PIF (n=2), RV, RSV, MPV, FLU, KI, ADV (1 patient for each). Risk factors in univariate analysis for death at 3 months identified ongoing steroid therapy ≥ 1 mg/kg (HR: 4.65, P=0.004), lymphocytes count < 0.5 G/L (HR: 12.4, P=0.04), lower tract involvement (HR: 3.69, P=0.045), presence of a co-infection (HR: 3.38, P=0.043) and oxygenotherapy (HR: 4.57, P=0.008). The delay between transplantation and VRI has no evident effect on outcome. Multivariate analysis showed that lymphocyte count and steroid dose at VRI onset were associated with the 3-month survival. More precisely, patients with lymphocyte count > 0.5 G/L had the best prognosis, while those with lymphopenia had a proportionally increasing risk of death according to the steroid dose (P=0.0002), with HR at 5.75 (95%CI 2.31 to 14.3) if they received less than 1 mg/kg and 33.0 (95%CI 5.32 to 204.9) if they received 1 mg/kg or more. Importantly, the prognostic value of these factors remained true during follow-up (Figure 1).

View largeDownload slide

View largeDownload slide

Close modal

RV, MPV, CV (OC43, 229E, NL63) are frequently isolated in cases of respiratory symptoms. Early mortality attributable to all VRI is relatively low and occurred only in patients with comorbidity. Prognosis of VRI may be related to the underlying immune defect induced by GVHD and its treatment, especially on lymphocyte counts.

Peffault de Latour:Alexion: Consultancy, Research Funding.