A prospective study of predictive factors for chemotherapy-induced nausea and vomiting.

Abstract

e19720 Background: Despite effective antiemetics, patients (pts) still develop nausea and vomiting after chemotherapy. While the emetogenic potential of different chemotherapy drugs has been established, the contributing patient and disease factors are less understood, and are the subject of this prospective study. Methods: Cleveland Clinic (CC), pts with locoregionally advanced, non-metastatic cancer of the head and neck (HN), esophagus, and gastro-esophageal junction (E/GEJ) are frequently treated in the hospital with concurrent chemoradiation using 96 hour infusions of cisplatin (20 mg/m2/day) and 5FU (1000 mg/m2/day). Between 7/09 and 9/10, all pts receiving their first course of chemotherapy with this regimen were entered in this study. A standard antiemetic prophylactic and therapeutic algorithm was utilized. Demographics, disease characteristics, potential risk factors, and maximum CTCAE v3.0 grade nausea and vomiting were recorded. Logistic regression analysis was used to identify predictors for the primary endpoint of > grade 1 chemotherapy induced nausea and/or vomiting (CINV). Results: The study enrolled 120 consecutive pts; 87% were male, 90% white, 61 had HN cancer, and the median age was 57 (range 26-76) years. Alcohol use was moderate or heavy in 33% and pts lived a median 39 (range 4-3000) miles from the CC. Univariable predictors for CINV included limited (vs. moderate or heavy) alcohol use, greater distance from the CC, history of psychiatric disorder, and E/GEJ (vs. HN) cancer. Although a body mass index (BMI) <25 predicted for greater CINV than BMI >32, this was felt to reflect the chemotherapy dose reductions sometimes used in obese pts. On multivariable analysis, only limited alcohol intake (OR 2.64, 95% CI 1.12-6.22, P=0.026) and an E/GEJ primary (OR 3.53, 95% CI 1.50-8.31, P=0.004) proved predictive for CINV. Pts with HN cancer and moderate or heavy alcohol use had a 41% chance of CINV in contrast to an 86% chance of CINV in pts with E/GEJ cancer and limited alcohol. Conclusions: The importance of disease site in predicting for CINV in our series likely reflects the radiation therapy port used in this concurrent treatment. The only prospectively identified patient predictor for CINV was a limited alcohol intake.