Reply to E. Javor et al

In their recent letter regarding our paper on supplement use during chemotherapy, in which we reported that use of B₁₂ supplements before and during treatment, and of iron only during treatment, was associated with poorer disease-free and overall survival, Javor et al¹ note that iron and vitamin B₁₂ are commonly given to patients with cancer, with a goal to prevent anemia and increase performance status. They query if use of these supplements could differentially affect patient outcomes depending on the assigned treatment schedule (usual dosing or dose dense) or if they are affected by hemoglobin status, noting that adjustment for anemia or treatment arm may be insufficient to address this possibility. Thus, the authors request that we stratify survival analysis by treatment arms and presence of baseline anemia, testing for possible interactions.

We appreciate this comment and suggestion. It is often difficult in an observational study to determine true relationships between exposures and outcomes, and the approaches suggested are reasonable. However, we are somewhat constrained by the limitations of the data. Because this was a multisite cooperative-group trial and only limited clinical data were required, we do not have information on baseline hemoglobin levels. However, it should be noted that the protocol for the trial states that no dose modifications for grades 2 to 4 anemia are allowed, and transfusions are acceptable, with early use of darbepoetitin recommended.

We did attempt to stratify by treatment arms to assess relationships between supplements and outcomes, but the numbers were too small. For example, for vitamin B₁₂, 80% of the patients were in the No, No (NN) group (ie, receiving supplements neither before nor during treatment) and 86% of patients were in the NN group for iron. Those analyses might have been possible if there were only two treatment arms, one of which was dose dense, but the treatment arms were six distinct groups. Thus, the data were too sparse across the other use categorizations (Yes, No; No, Yes; Yes, Yes) to include an additional main effect for treatment and a treatment-by-interaction model term, relative to the model fitting process converging. However, in an ad hoc analysis to respond to this question, we considered the treatment arms in the context of a 3 × 2 factorial design relative to the treatment sequence. Examining the interaction terms with supplement use in this fashion led to a stable model fit in the age-only adjusted model but yielded no significant interaction terms.

We thank Javor et al¹ for their thoughtful read of our paper and appreciate their comments. A larger study, or perhaps randomized trials of supplementation with iron or vitamin B₁₂ during chemotherapy, might be able to more fully address this question.

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Reply to E. Javor et al

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).

No potential conflicts of interest were reported.