Usefulness of a general chemotherapy myelotoxicity score to predict febrile neutropenia in haematological cancers

Abstract number: 1733_224

Georgala A., Paesmans M., Schwarzbold A., Muanza F., Aoun M., Ferrant A., Bron D., Klastersky J., Moreau M.

Introduction: Chemotherapy-induced neutropenia is the most common adverse effect of chemotherapy and is often complicated by febrile neutropenia (FN). In many institutions, the decision to give granulocyte colony-stimulating factors as prophylaxis is mainly based on the myelosuppressive potential of the chemotherapy regimen (CR) but clear regimen-specific risks are not defined. The objective of this study is to validate a classification of aggressiveness of a CR and to evaluate its usefulness in a risk prediction model of FN in patients with haematological cancers (HC) at the beginning of a chemotherapy cycle.

Methodology: Patients aged above 16 years, diagnosed with HC of any type were prospectively enrolled and followed in the ``Institut Jules Bordet'' and the ``Université Catholique de Louvain'' between 2001 and 2005. Out of the 266 enrolled patients, 22.9%, 43.6% and 33.5% were followed respectively during one cycle, 2 to 4 cycles and more than 4 cycles, totalising 1053 cycles. Relevant patient's information were collected at the beginning of the first cycle (sex, age, diagnosis and concomitant diseases). At the beginning of each cycle, the CR score was computed, a blood examination performed and the temperature measured. In the follow-up, the number of days of FN (neutrophils <500/ml and fever) were counted. This outcome was dichotomised (no FN vs ≥1 day of FN). Generalised Estimating Equations (GEE) was used for the analysis as it takes into account the correlation structure between the outcome as well as the covariates within the same patient.

Results: In 35.3% of the cycles, patients experienced FN. In the final model, aggressive CR is the major predictor of FN (OR 4.4 [2.7–7.0]), compared to those not receiving an aggressive CR. The other independent predictors are a diagnosis of ``myeloid tumour'' (2.9 [1.7–5.1], a baseline monocyte count <150/ml (2.3 [1.4–3.8]), an involvement of bone marrow (2.1 [1.4–3.2]), the first cycle in the same treatment line (2.2 [1.5–3.2]) and a baseline hemoglobin dosage <12 gr/l (1.7 [1.0–2.7]). Using the estimates of the regression coefficients, a rule of prediction of FN was computed (sensitivity: 82.2%, specificity: 78.1%, positive predictive value: 68.5% and negative predictive value: 88.4%).

Conclusions: Further studies are needed to validate this score as well as investigating new factors in order to be able to better predict FN.