1458P - Development of a prognostic system to predict the response to treatment of neutropenic fever

Date 09 October 2016
Event ESMO 2016 Congress
Session Poster display
Topics Supportive Measures
Presenter hamdy Zawam
Citation Annals of Oncology (2016) 27 (6): 497-521. 10.1093/annonc/mdw390
Authors H. Zawam1, A. Selim2, R. Salama3, N. Hanna2, W. Edesa3
  • 1Clinical Oncology Department, Cairo University, 12345 - Cairo/EG
  • 2Clinical Oncology Department, Cairo University, Cairo/EG
  • 3Clinical Oncology Department, Cairo University, 1 - Cairo/EG



Febrile neutropenia (FN) remains one of the most commonly encountered oncologic emergencies in patients (pts) with hematological malignancies. Development of a prognostic system will help to improve the outcome in those pts.


This is a prospective observational study including 142 pts with haematological malignancies who presented to Kasr Al Ainy Center of Clinical Oncology during the period 1st of June 2014 to October 2015. This group of pts suffered from 270 episodes. According to the MASCC score, high risk pts were treated inpatients. All admitted pts were subjected to blood, sputum, stool and urine cultures withdrawal and Galactomann test. PCR (polymerase chain reaction) of sepsis and BAL (broncho-alveloar lavage ) were done in certain cases. Empirical antibiotics were started immediately, antifungal and antiviral tretements were recieved according to the guidelines.


The different diagnostic modalities were analysed in addition to the results of treatment by different classes of antibiotics. The most frequent diagnosis in our study were AML (55 pts), followed by ALL (27pts), NHL (35 pts).The disease status was found to be highly significant and affects the control of neutropenic fever episode. The more the patient developed neutropenic episodes the more the risk of mortality. In our study, the MASCC score was highly significant. 62% of the identified pathogens were gram positive detected by blood culture, while gram negative bacteria were the commonest pathogens identified by other diagnostic modalities.

Multivariate analysis to determine pretreatment variables of independent prognostic value

Variable odds ratio 95% CI P-value Weighted partial score
Previous FN episode 1.47 0.307 7.023 0.63 -
Disease burden 3.677 1.2 11.265 0.023 2
Hypotension 5.609 1.711 18.392 0.004 3
Prevoius fungal infection 1.905 0.407 8.91 0.413 -
Uncontrolled disease 4.222 1.019 17.493 0.047 2.5
The sum of the weighted partial scores of the three significant variables resulted in a prognostic score ranging from 0 (best prognosis) to 7.5 (worst prognosis). Cutoff value of 4.5 was determined and it divides patients into two groups, ≤4.5 vs. >4.5


Many risk factors affect the outcome of FN and should be taken into consideration for every pt.

Clinical trial identification

Legal entity responsible for the study

Hamdy Zawam




All authors have declared no conflicts of interest.