Avaliação de modelos de predição de complicações clínicas, determinando a importância da variável infecção complexa, em pacientes oncológicos e neutropênicos febris de baixo risco.

Abstract

This dissertation is a prospective and observational study comparing the efficacy of risk asssessment models in reference treatment center. The meaning of complex infeccion was evaluated. This single institution clinical study was conducted in accordance with Good Clinical Practice and the Declaration of Helsinki and was approved by the local ethicscommittee approval. Patients were recruited throughout a 9 month period, between March 1, 2004 and December 30, 2004. Inclusion criteria were neutropenic febrile (absolute neutrophil count of < 500/mm3 and axillary temperature = 38 degrees C) secondary to chemotherapy and/or radiotherapy, = 18 years of age and admitted to Felicio Rocho Hospital, Belo Horizonte, Minas Gerais, Brazil. In total, 60 episodes of neutropenic fever were included (39 women and 21 men), with a median age of 55 (range: 18-82) years and 27 (45%) were 60 years or older. Thirty eight patients (63%) had hematologic malignancies, fifty percent wereprofoundly neutropenic (<100 neutrophils/mm3) at the onset of infection and the median duration of Hospital admission was thirteen days. Fourty four patients had serious concurrent comorbidities other than fever and neutropenia (13% had hypotension, defined as systolic blood pressure < 90 mm Hg, 28% had some degree of respiratory failure and 61,7% had dehydratation). There were 9 deaths, but only seven were related to febrile neutropenia. After multivariate analysis, we observed that need of intensive care was the only variable related to death secondary to neutropenic fever. Predefined conditions meeting these criteria, intensive care need, were: hypotension (as defined above) with or without hemodynamic support, respiratory, renal or heart failure, intensive care unity or sepsis. Sixteen patients had microbiological infeccion and the etiologic agent was mainly Gram-negative pathogens (Pseudomonas aeruginosa and E. coli, that were multissensible to conventional antibioticotherapy in the majority of cases). Of the 60 febrile neutropenic episodes, 14 were classified as low-risk by Talcott (group 4) and 24 by MASCC risk-index score (score of = 21). The sensitivity, specificity, negative predictive value, positive predictive value and accuracy of Talcottrisk assessment were, respectively 100%, 58.3%, 100%, 78.2%, 83,3% and 86.1%, 79.2%, 79.2%, 86.1% and 83.3% by MASCC risk-index score. None of the low-risk patients died (both models), but there were five episodes with MASCC score of = 21 that developed complications during febrile neutropenic episodes and they would take medical hospitalar care (false low-risk). When we subtract patients with complex infections from the group of patients with MASCC risk-index score of = 21 we got 17 patients that has been true low-risk (anyone require hospitalar care). The sensitivity, specificity, negative predictive value, positive predictive value and acuracia of this new model (risk-index score adjusted by complex infection, PACI model) were, respectively, 100%, 70.8%, 100%, 83.7%, 88,3%. Complex infection was considered if the patient had at least one of the following: 1) infection of major organs (lungs, liver, spleen, kidneys, colon, bones, joints, veins, heart, and meninges); 2) soft-tissue wound infection, abscess, or cellulitis that was > 5 cm in size and found during physical examination, or soft-tissue xxiv infections (of any size) with necrosis; 3) oral mucositis (or stomatitis) > grade 2; 4)sepsis. Talcott risk assessment and MASCC risk-index score had high sensitivity and specificity to predict who would develop medical complications and would take medicines in hospital. Nevertheless the new model (PACI model) had the best Kappa index. It means that the latter model was better than Talcott and MASCC to predict for medical complications in this population of febrile neutropenics. In summary, the modified MASCC risk-index score (PACI model) is better than other models and it has good potential for accurately predict and determine which patients may safely undergo outpatient therapy.