340 Nocardiosis in Allogeneic Hematopoietic Stem Cell Transplant Recipients: A Matched Case-Control Study of Risk Factors, Clinical Features and Outcomes

Track: Contributed Abstracts
Saturday, February 16, 2013, 6:45 PM-7:45 PM
Hall 1 (Salt Palace Convention Center)
Nadia M Bambace, MD, FRCPC , Hematology/Stem Cell Transplantation, Maisonneuve Rosemont / University of Montreal, Montreal, QC, Canada
Louise Poirier, MD, FRCPC , Department of Microbiology-Infectious Diseases, Maisonneuve Rosemont/University of Montreal, Montreal, QC, Canada
Sandra Cohen, MD, FRCPC , Hematology/Stem Cell Transplantation, Maisonneuve Rosemont / University of Montreal, Montreal, QC, Canada
Thomas Kiss, MD, FRCPC , Hematology/Stem Cell Transplantation, Maisonneuve Rosemont / University of Montreal, Montreal, QC, Canada
Guy Sauvageau, MD, PhD, FRCPC , Inst De Recherches, Clinic De Montreal, Montreal, QC, Canada
Jean Roy, MD, FRCPC , Hematology/Stem Cell Transplantation, Maisonneuve Rosemont / University of Montreal, Montreal, QC, Canada
Denis-Claude Roy, MD, FRCPC , Hematology/Stem Cell Transplantation, Maisonneuve Rosemont / University of Montreal, Montreal, QC, Canada
Miguel Chagnon, M.SC., P.Stat. , Université de Montréal, Montreal, QC, Canada
Silvy Lachance, MD, FRCPC , Hematology/Stem Cell Transplantation, Maisonneuve Rosemont / University of Montreal, Montreal, QC, Canada
Nocardial infection is emerging as an important cause of morbidity and mortality among hematopoietic stem cell transplant (HSCT) recipients. Risk factors and outcomes in this population remain undefined.

We performed a matched case-control study (1:2 ratio). Cases were defined as recipients of allogeneic HSCT with a microbiological diagnosis of nocardial infection. Control subjects were matched for age, timing and transplant type. Between January 2007 and December 2011, among 440 allogeneic HSCT recipients, 11 (0.03%) were diagnosed with nocardiosis.

 Infection occurred at a median of 510 days (range 139-1042) after HSCT and was disseminated in 45% of cases. Diagnoses clustered in the fall (68%). Pulmonary involvement with nodular infiltrates occurred in 91% of cases.   Final culture results were available 55.7 days after diagnostic testing (range 14-120 days). Nocardia nova was the strain most commonly isolated (27%). Co-infection with S.aureus, Pseudomonas, CMV and Mycobacterium sp. occurred in 73% of all cases. Trimetroprim-Sulfametoxazole was not protective in 4 out of 11 patients (36%) receiving it for Pneumocystis jirovecii pneumonia prophylaxis.

 In univariate analysis, chronic GVHD (p= .011) was associated with nocardial infection. Other associated conditions included bronchiolitis obliterans syndrome (BOS) (p= .033), steroid-induced diabetes mellitus (p=< .001), and opportunistic infection within the preceding 6 months (p= < .001). Positive CMV serologic status or recent CMV infection were not significant variables. High-dose corticosteroid treatment within the preceding 6 months (p =.005), tacrolimus therapy (p = .002), antifungal prophylaxis (p = <.001), prior autologous transplant (p = .008), and rituximab treatment within 12 months (p = < .001) were specifically associated with nocardiosis. Patients with Nocardia infection had significantly higher mean tacrolimus levels (p = .043), LDH levels (p = .040), and neutrophil counts (p= .002) than controls.

 Nocardial infection is an infrequent delayed complication of allogeneic HSCT primarily affecting recipients with chronic GVHD. This has biologic correlation, since these patients often require prolonged immunosuppressive therapies, targeting both B and T cells, and may have underlying anatomic factors interfering with microbial clearance, such as bronchiectasis in BOS.  High-dose corticosteroid treatment and its consequences, notably steroid-induced DMII, may increase susceptibility to this infection. Nocardiosis should be promptly considered and carefully investigated in susceptible cGVHD patients, given its adverse impact on prognosis, as demonstrated by the  significant reduction in the overall survival of the infected cohort (72.7% vs 100%, p=.013).