We sought to describe the impact of chemotherapy-induced nausea and vomiting (CINV) in prior cycles on CINV and chemotherapy regimen modification in subsequent cycles. Eligible patients in this multinational prospective observational study were adults (a parts per thousand yen18 years old) receiving their first single-day highly or moderately emetogenic chemotherapy (HEC or MEC). Multivariate logistic regression was used to assess the impact of CINV in prior cycles on CINV in subsequent cycles. Other independent variables included in the model were the cycle number, age, sex, and emetogenicity of regimen. There were 598 evaluable patients in cycle 2 and 533 in cycle 3, half receiving HEC and half MEC. Patients who experienced complete response (no emesis or rescue antiemetics) in earlier cycles, relative to those with no complete response, had an adjusted odds ratio (OR) of 5.9 (95 % confidence interval (CI), 4.14-8.50) for experiencing complete response in subsequent cycles. Prior CINV was a significant and consistent predictor of subsequent CINV for all CINV endpoints: for emesis, OR 12.7 (95 % CI, 8.47-18.9), for clinically significant nausea, OR 7.9 (95 % CI, 5.66-10.9), and for clinically significant nausea and/or vomiting, OR 7.2 (5.17-10.1). Modifications to chemotherapy were recorded for 26-29 % of patients in cycles 2 and 3, with CINV as the major reason for the modification for 5-9 % of these patients. CINV in prior cycles was a strong and consistent predictor of CINV in subsequent cycles, while the incidence of chemotherapy regimen modification due to CINV was low in individual cycles.