Pretreatment Platelet Count Predicts the Outcome of Invasive Aspergillosis among Myeloma Patients Undergoing Myelosuppressive Chemotherapy

Invasive pulmonary aspergillosis (IPA) is a major cause of mortality in cancer patients with chemotherapy-induced neutropenia. While duration of neutropenia is a significant risk factor for and an outcome predictor of IPA, many patients with IPA recover from neutropenia and yet fail to respond to therapy. In addition, the duration of neutropenia (absolute neutrophil count; ANC <1000/μL) cannot be predicted at time of initiation of chemotherapy. Because platelet count serves as a surrogate marker for bone marrow reserve and because platelets are potent effector cells involved both in innate and adaptive immune response against infections including aspergillosis, we hypothesized that pretreatment platelet count may predict the outcome of IPA. Pretreatment platelet count was defined as the platelet value obtained immediately prior to commencing the myelosuppressive chemotherapy associated with development of IPA.

Between November 2003 and February 2008, 78 consecutive myeloma patients developed IPA following chemotherapy-induced neutropenia. This included two populations: recipients of autologous stem cell transplantation (ASCT) (32 patients) and patients treated with non-myeloablative chemotherapy (NMCT) (46 patients). The diagnosis of IPA was based on conventional diagnostic criteria plus ≥ 2 consecutive positive assays for the FDA-approved serum galactomannan index (GMI) (optical density ≥0.5). Following the diagnosis of IPA, GMI testing was repeated to monitor the rapidity of IPA response. Response of IPA was defined as normalization of GMI with survival and was considered rapid if it occurred within 8 days after ANC ≥ 1000/μL. Overall, 664 GMI tests were positive (median of 5 per patient).

Univariate and multivariate logistic regression analysis were applied to identify the pretreatment variables that predict IPA outcome. These variables included age, gender, body measurements, myeloma remission status, ANC, platelet and lymphocyte counts, renal and liver function tests, serum albumin and the number of CD34 + cells infused (for 32 ASCT recipients). All patients developed neutropenia which was severe (ANC <100/μL) in 53. Response was rapid (RR) among 36 patients (median 2 days after ANC ≥ 1000/μL, range −6 to +8 days), slow (SR) among 25 patients (median 25 days after ANC ≥ 1000/μL, range +12 to + 64 days) while 17 were nonresponders (NR) and died with IPA. The incidence of severe neutropenia did not significantly differ among these 3 groups. However, the duration of severe neutropenia was significantly shorter among patients with higher pretreatment platelet count and among the various response groups (6.72 ± 4.68, 8.55 ± 5.51 and 13.79 ± 11.47 days for RR, SR and NR respectively (P = 0.0019). By multivariate analysis, prognostic factors for poor response (SR and NR) among the 78 patients were lower pretreatment platelet count (platelet unit=25.000 platelets/μL) (P=0.0095; Odds Ratio (OR) 1.284; 95% confidence interval (CI) (1.063–1.550) and receipt of NMCT (P=0.0166; OR 3.686; 95% CI; 1.267–10.722). Lower pretreatment platelet count was the only significant predictor of poor response when the ASCT and NMCT patient populations were separately analyzed (P=0.0290; OR 1.356; 95% CI-1.032–1.783) and (P=0.0377; OR 1.295; 95% CI −1.015–1.653), respectively.

We conclude that pretreatment platelet count is a strong predictor of the quality and rapidity of response of IPA among neutropenic patients with myeloma. These novel findings, which are available prior to commencing myelosuppressive chemotherapy, allow us to predict, for the first time, the likelihood and rapidity of response of IPA and may be used to individualize antifungal prophylaxis and therapy.