Reply: Quality of life in actual living liver donors versus potential living liver donors

We thank Butt et al.1 for their interest in our article on quality of life (QOL) in actual and potential living liver donors.2 In reply, we would like to address their methodological concerns.

The first concern is related to the composition of the control group and the retention rate. In our view, the exclusion of 13 potential donors due to additional exclusion criteria does not lower the retention rate because those patients did not constitute a valid control group from the beginning. They were rejected as donors either for psychosocial reasons or because their recipients did not undergo transplantation. Therefore, they were not comparable to actual donors, who were, of course, deemed psychologically eligible and whose recipients underwent transplantation. We consider the exclusion criteria to be necessary in order to ensure comparability with the group of actual donors. Individuals who do not meet these criteria are not suitable as controls.

Nevertheless, we would like to address the problem of the retention rate in more detail. As stated in our article,2 the overall response rate of 64% for the preoperative time point is lowered further because only 60% of these responders also returned their questionnaires at the postoperative time point. However, for the aforementioned reasons, it would have been more consistent if we had applied the exclusion criteria before we calculated the response rate. We would like to supplement this information. If the exclusion criteria constituting eligibility in the control group are applied beforehand, 33 of the 88 patients who did not donate have to be excluded. This means that of 55 eligible control patients at the time of donor evaluation (t₀), 33 also participated 3 months after transplantation, and this constitutes a response rate of 60% for the potential donors. This results in a dropout rate of 30% for the actual donors and a dropout rate of 40% for the potential donors.

We were able to compare the dropouts and the patients who completed their questionnaires at both time points with respect to sociodemographic and clinical variables as well as QOL at t₀. For this analysis, the exclusion criteria were applied beforehand so that only those potential donors who constituted a suitable control group were included. For the dropout analysis, we did not differentiate between actual and potential donors because this group difference is not relevant preoperatively. The analysis revealed no significant differences with respect to QOL, anxiety, depression, or abdominal symptoms at t₀. However, significant differences between responders (who returned questionnaires at both time points) and dropouts (who returned questionnaires only at t₀) were found with respect to the transplant center (42% dropout rate in Essen and 21% dropout rate in Hamburg; P = 0.03), donor group (46% dropout rate among patients donating for adult recipients and 13% dropout rate among patients donating for pediatric recipients; P < 0.001), donor-recipient relationship (17% dropout rate among parents donating for their children, 33% dropout rate among children donating for parents, 54% dropout rate among donors for siblings, 80% dropout rate for spouses, 39% dropout rate for other relatives, and 33% dropout rate for unrelated patients; P = 0.004), and postoperative complications in the recipients (19% dropout rate among donors for recipients with non life-threatening complications, 27% dropout rate among donors for recipients with potentially life-threatening complications, and 55% dropout rate among donors for recipients suffering retransplantation or death; P = 0.01). The last is to be considered the most important difference because complications in the recipients were found to have a negative impact on donors' QOL.3-5 However, in another study,6 no significant differences were observed between donors whose recipients suffered from complications and donors whose recipients did not. As for the donor group, a study comparing right lobe donors with left lobe donors did not find any QOL differences.7 In conclusion, the idea that QOL scores are overestimated cannot be excluded because donors whose recipients suffer from severe complications, donors for adult recipients, and those not donating for their children show a higher dropout rate. However, this is a problem that concerns actual donors as well as potential donors and therefore does not limit comparability of the 2 groups.

The problem of low retention rates, especially in longitudinal studies with several time points, and its implications concerning the generalizability of the results are valid concerns and should always be kept in mind. In clinical research, it is difficult to facilitate patients' continued participation. If the recipient experiences severe complications or even dies, it is understandable that donors want to drop out, and it would be unethical to try to prevent them from doing so.

The second concern is related to the current practice of using generic outcome measures or resorting to specific instruments originally intended for other populations, which we agree might not be appropriate. We also regret the lack of specific instruments addressing the concerns most germane to living donors. We therefore initiated a study aimed at developing a specific QOL instrument for living donors. It is supported by the German Research Foundation (SCHU 974/6-1, ER 333/2-2) and is still a work in progress.

With respect to the authors' comments about the clinical significance of changes in the outcome measures, we consider it important to point out that in living donors, initially no significant change should be dismissed because living donors undergo surgery as healthy, unimpaired individuals. Every change in QOL or physical well-being in this group should be taken seriously and warrants further consideration. However, in order to be able to further evaluate possible impairments, measures of clinically significant changes would be useful. As a first step, effect sizes could be calculated. Nevertheless, even small effects could be clinically meaningful and should not be disregarded. Specific instruments would have the advantage of being more sensitive to changes and would therefore be better suited for depicting changes of clinical significance.