Letter reply

We appreciate the interest of Dr. Akbal and colleagues in our work. Below, we address their concerns.

In our study, we wished to know the impact of our intervention on the quality of life of patients with dysplastic Barrett’s esophagus. Akbal and colleagues suggest that “QoL should be compared with healthy individuals rather than a sham group.” Such a comparison is not relevant to our question, since one would not consider performing RFA on healthy individuals. It is quite possible, even likely, that healthy controls would report better QoL than our patients before or after treatment. Such a comparison tells us nothing about the impact of our treatment on the patients in whom we would plausibly use RFA. Moreover, the strength of a randomized controlled trial is that it minimizes potential bias by the randomized nature of the allocation of participants to the treatment group or control group. It would be a squandered opportunity not to take advantage of this strong study design to assess changes in QoL associated with the intervention in a relevant patient population.

Akbal et al. misunderstand our study design when they suggest that we “compare QoL among both sham and RFA groups before and after interventions.” This is exactly what we did do. As we note in the manuscript, there were no significant differences in QoL between the groups at baseline. It was only after the intervention that the groups had the dramatic differences we document.

Akbal and colleagues are correct when they suggest that numerous variables may impact QoL, specifically “gender and marital status.” However, the process of randomization should produce comparable groups with regards to known and unknown potential confounders. As documented in the manuscript, randomization was successful with respect to the measured potential confounders in this trial.

As Akbal and colleagues point out, the treatment group in our study had greater exposure to endoscopy than the sham group. As they note, greater exposure to endoscopy has been associated with reduced QoL. Therefore, because our treatment group had greater exposure to endoscopy compared with the control group, any bias generated due to the extra endoscopies would be expected to bias QoL measures against, not in favor of, the intervention. If this is the case, our study may actually underestimate the beneficial effect of the intervention on QoL.

We disagree with Akbal et al.’s contention that the “sham group had a negative doctor–patient relationship.” We performed this study in clinical equipoise – there was uncertainty as to the value of the intervention compared with standard surveillance endoscopy. While we are pleased to have demonstrated that this intervention benefits patients at a fundamental level, if we had been certain of these results beforehand, we would have not undertaken the study. The level of care, and the positivity of interactions between the caregivers and the patients were the same in both arms of this study.

While we look forward to future data from other groups describing the impact of ablative therapy on QoL in dysplastic Barrett’s esophagus, we are comfortable with our interpretation of our data, and feel that we have provided a conservative and accurate estimate of the positive impact of RFA on the QoL of the patients in the AIM Dysplasia trial.