Parameters related to death at one year in colorectal cancer patients

Quintana, Jose M

Parameters related to death at one year in colorectal cancer patients

Sunday, 17 August 2014

Exhibit hall (Dena'ina Center)

Jose M Quintana, PhD , Hospital Galdakao-Usansolo / Red de Investigación en Servicios Sanitarios y Enfermedades Crónicas (REDISSEC), Galdakao, Spain

Urko Aguirre, MS , Hospital Galdakao-Usansolo / Red de Investigación en Servicios Sanitarios y Enfermedades Crónicas (REDISSEC), Galdakao, Spain

Nerea M González, PhD , Hospital Galdakao-Usansolo / Red de Investigación en Servicios Sanitarios y Enfermedades Crónicas (REDISSEC), Galdakao, Spain

Santiago Lazaro, MD , Hospital Galdakao-Usansolo, Galdakao, Spain

Cristina Sarasqueta, MD , Hospital Universitario Donostia / Red de Investigación en Servicios Sanitarios y Enfermedades Crónicas (REDISSEC), Donostia, Spain

Antonio Escobar, PhD , Hospital Universitario Basurto / Red de Investigación en Servicios Sanitarios y Enfermedades Crónicas (REDISSEC), Bilbao, Spain

Marisa Baré, PhD , Parc Tauli Sabadell-University Hospital/ Red de Investigación en Servicios Sanitarios y Enfermedades Crónicas (REDISSEC), Sabadell, Spain

Nerea Fernández de Larrea, MPH , Departamento de Salud de Madrid / Red de Investigación en Servicios Sanitarios y Enfermedades Crónicas (REDISSEC), Madrid, Spain

Maximino Redondo, PhD , Hospital Costa del Sol / Red de Investigación en Servicios Sanitarios y Enfermedades Crónicas (REDISSEC), Marbella, Spain

INTRODUCTION: There is controversial information on the variables that may be related to death at one year after surgical removal in patients with colon cancer. The goal of the study was to develop predictive models, with variables available at the time patients underwent the surgical removal, and to relate them to death at one year.

METHODS: Prospective cohort study of patients diagnosed of colon cancer that underwent curative surgery in any of the 16 participant hospitals. Patients were followed from the moment they got in touch with the index hospital through diagnostic, pre-intervention, at admission, intervention and in the follow-up at one year after discharge. Patient sociodemographic (age, gender), clinical (Charlson comorbidity index, TNM stage, location of tumour, complications at admission) variables were retrieved from the patient medical record. For the statistical analysis we developed a multivariate linear model where all the statistically significant variables from the univariate analysis were included as well as age, the location of the tumour and the tumour stage (based on the TNM classification of malignant tumours).

RESULTS: A total of 997 patients were included in the analysis. Variables related to mortality at one year were presence of comorbidities, as measured by the Charlson Comorbidity Index (OR: 1.34), major complications at admission index (OR:3.03), length of stay at admission index (more than 9 days, OR:1.99), location of the tumour(p<0.05) and the tumour stage (based on the TNM classification(p<0.01)). AUC for the whole model was 0.79

CONCLUSIONS: Comorbidities, severity of the disease, length of stay and complications at the time the patient was first admitted to the hospital for the removal of the tumour already predict death at one year. These results may help to identify patients to have a closer follow-up as to prevent death, specially in those having complications at admission index.