Trial design and study population
The trial was designed as a single center, randomized, double-blind, placebo-controlled phase II study at the Danish Colorectal Cancer Center South, Vejle Hospital, Denmark. Patients with potentially resectable or non-resectable mCRC eligible for first line treatment were included.
The inclusion criteria were age 18-75 years, ECOG performance status (PS) 0-1 (patients above 70 years were eligible if PS=0), adenocarcinoma in the colon or rectum, evaluable but not necessarily measurable disease according to Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1  and adequate biochemistry. Previous adjuvant chemotherapy for radical treatment of stage II or III colorectal cancer was allowed in patients proven disease-free for more than 6 months. The detailed eligibility criteria are listed in supplementary material (Online Resource 1).
The study was approved by The Regional Committee on Health Research Ethics for Southern Denmark (S-20150185) and the Danish Medicines Agency (2015122439), and was prospectively registered with Clinicaltrials.gov (Identifier NCT02705300). The investigation was conducted in accordance with the Declaration of Helsinki and adhered to the guidelines for good clinical practice (GCP).
All participating patients provided informed consent.
Study data were collected and managed using the Research Electronic Data Capture (REDCap) tool [20, 21] hosted by Open Patient data Explorative Network (OPEN), Odense University Hospital, Region of Southern Denmark. Data were monitored by the regional GCP Unit.
A randomization list (1:1) was generated using www.randomization.com and kept at the hospital pharmacy. Via the Clinical Research Unit of Department of Oncology, Vejle Hospital, each patient was assigned a unique code referring to the randomization list. The pharmacy would then allocated the patient to arm A or B. The treating staff and researchers were blinded to the allocation. In case of urgent need for unblinding, the unique codes were kept in sealed envelopes in a locked room in the Department of Oncology.
Treatment and procedures
FOLFOXIRI was given as the standard regimen developed by the GONO group  and administered biweekly for eight cycles with irinotecan 165 mg/m2 i.v for 30 minutes, leucovorin 200 mg/m2 i.v. for 120 minutes, oxaliplatin 85 mg/m2 i.v. for 30 minutes followed by a continuous infusion of 5-fluorouracil 3200 mg/m2 i.v. for 46 hours. After the eight cycles of FOLFOXIRI, treatment was continued with 5-fluoruracil and leucovorin biweekly for four cycles.
Capsules of δ-tocotrienol 300 mg x 3 daily or placebo x 3 daily were administered for a maximum of two years. The tocotrienol capsule consisted of the isomers δ-tocotrienol (90%) and γ-tocotrienol (10%) while the placebo capsule contained olive oil. δ-tocotrienol and placebo were made by the same manufacturer (American River Nutrition, Inc., Hadley, MA, USA) and had the same appearance.
Dose delays and modifications of chemotherapy were permitted based on toxicity and done according to the summaries of product characteristics.
In patients needing an intervention such as surgery or hepatic radiofrequency ablation (RFA) of primary tumor or metastases during study treatment chemotherapy was suspended according to departmental guidelines and δ-tocotrienol/placebo paused 7 days before and after the intervention.
Treatment effect was evaluated by CT scan according to the RECIST 1.1 criteria every eight weeks during chemotherapy, then every three months for a year, every six months the following two years, and yearly the fourth and fifth year after end of chemotherapy. In case of progression, patients were treated according to departmental guidelines.
Blood sampling and biobanking
Before each treatment cycle, blood was sampled for complete blood count, liver function, electrolytes and renal function.
Plasma samples from all patients were collected at baseline, before every treatment cycle and at every follow-up visit for later translational research.
The primary endpoint was ‘time to first non-planned hospitalization or death’ calculated from start of first cycle of FOLFOXIRI and seven months ahead in order to investigate if δ-tocotrienol reduced toxicity during the planned six months of chemotherapy. Secondary endpoints were number and duration of non-planned hospitalizations during chemotherapy; death during chemotherapy; RR, defined as the proportion of patients achieving partial or complete response according to RECIST 1.1; PFS; OS; radical resection rate; toxicity, and quality of life.
Registration of adverse events was performed before treatment start, before each cycle, after the last cycle of chemotherapy and at every follow-up until progression. Adverse events were classified according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.0.
Quality of life assessment
Quality of life was evaluated by the EORTC QLQ-C30 and EORTC QLQ-CR29 questionnaires before start of treatment, at every response evaluation and follow-up visit.
In a previous study with FOLFOXIRI, around half of the patients died or had grade 3-4 toxicities usually requiring hospitalization . Using phase II statistics, we wanted to test if time to first hospitalization or death could be reduced with a hazard ratio of 0.5 by adding δ-tocotrienol to FOLFOXIRI. With a power of 80% and a two-sided significance level of 5%, 70 patients should be included. An enrolment period of 36 months was planned.
Time to first hospitalization or death was calculated and reported according to the Kaplan-Meier method with indication of hazard ratio. PFS and OS were calculated from the day of enrolment as median survival and according to the Kaplan-Meier method. Comparisons were made using hazard ratio and the treatment arms compared by log rank test. Non-parametric methods were used to calculate and compare number of admissions, response and toxicity. Quality of life was compared using mixed regression models taking into account each patients level and also group difference at baseline.
Statistical calculations were performed using Stata/BE 17.0 (StataCorp LLC, TX, USA).