W e read with interest the article by Offidani et al. 1 regarding the risk assessment of patients with hematologic malignancies who develop fever accompanied by pulmonary infiltrates. The authors focus on favorable prognostic parameters that are easily available (low c-reative protein, high albumin, and a favorable trend in the leukocyte count) and claim (rightly in our opinion) that their prognostic model is clinically useful. Patients with 2 or 3 of these factors were reported to have a survival rate of 95%, whereas only 46% of patients with none or only 1 of these factors survived. Clinical parameters such as hypotension or renal failure were not included in the multivariate analysis because they were clearly associated with an adverse prognosis and were found to occur in only a minority of patients. However, we wonder why the authors did not include a measure of patient performance status such as the Karnofsky performance status (KPS) or World Health Organization performance status because this variable is easily available at any time point and may add valuable information. We retrospectively evaluated a cohort of 120 mostly neutropenic patients with pulmonary infiltrates accompanying a febrile episode after they received chemotherapy for a hematologic malignancy (unpublished data). Approximately 29% of patients with a KPS of 100% at the time of hospital admission died during the course of pneumonia, but among patients with a KPS between 60 -90%, 59% died. Of those patients with an even lower KPS, 75% were reported to have died. Multivariate analysis demonstrated the significance of performance status as an independent prognostic factor with an odds ratio of 5.3 for a KPS Ͻ 60%, 2.5 for a KPS of 60 -90%, and 0.7 for a KPS of 100% at the time of hospital admission (P ϭ 0.033). We therefore suggest that a purely clinical parameter such as KPS also should be taken into account when estimating the risk for an individual patient with hematologic maligancies.