A Nomogram to Individually Predict the Presence of Chemotherapy-Induced Nausea and Vomiting (cinv) in Cancer Patients

Zhihuang Hu,Wenhua Liang,Li Zhang,Yuxiang Ma,Likun Chen,Yan Huang,Hongyun Zhao,Ting Zhou,Wenfeng Fang
DOI: https://doi.org/10.1200/jco.2014.32.15_suppl.e20680
IF: 45.3
2014-01-01
Journal of Clinical Oncology
Abstract:e20680 Background: Chemotherapy-induced nausea and vomiting (CINV) remains one common side effect of cancer treatment, whose presence depends on multiple factors. Prophylactic management of CINV is far more important than the rescue therapy after it happens. Here, we aimed to build a model-based nomogram to predict the individual probability of CINV, which is crucial to guide treatment decisions regarding the use of antiemetic. Methods: Patients from eight clinical trials of antiemetic agents in China from 1985 to 2013 were used to screen eligible subjects, who should have received highly, moderately, or lowly emetogenic chemotherapy (HEC/MEC/LEC) and be evaluable for the outcome of CINV. Complete response during overall phase (0~120 hours after chemotherapy) was used to assess the efficacy of antiemetics. Information on demographics, performance status (PS), alcohol use, motion sickness, history of pregnancy vomiting (PV), antiemetic regimen and other potential risk factors of CINV were examined. Multivariate logistic regression model were used to identify significant variables and integrate them into the nomogram. Bootstraps with 1000 resample were used for calibration and calculating the concordance index (C-index). Results: A total of 993 patients were eligible for analysis. Sex/PV group (male/ female without PV history/female with PV history, P<0.001), age (P=0.059), emetogenic level of chemotherapy (HEC/MEC/LEC, P<0.001), and antiemetic regimen (Metoclopramide/5HT3RA/ 5HT3RA+ DXM/5HT3RA+ DXM+ NK-1RA, P<0.001) were independent factors in multivariate model and were built into the nomogram. The C-index of the nomogram was 0.71 (0.68, 0.74), which was higher than any prediction using single factor above (Sex/PV group 0.56, age 0.54, emetogenic level 0.54, antiemetic regimen 0.62). Alcohol use failed to enter the model due to much value missing (59.2%), although it was an independent factor of CINV in patients with this information (P=0.046). Conclusions: Factors of female, pregnancy vomiting history, age, emetogenic level of chemotherapy and earlier antiemetic regimen indicated high risk of CINV presence. A nomogram could optimize the prediction and patient stratification.
What problem does this paper attempt to address?