Identifying the best donor for the best recipient: Parents may still be better served by nonoffspring donors
Year: 2018
Type: letter
Abstract: To the Editor: We read with great interest the report “Better graft outcomes from offspring donor kidneys among living donor kidney transplant recipients in the United States.”1 The conclusions of Holscher et al differ considerably from those in our recent paper “Inferior long-term allograft and patient outcomes among recipients of offspring living donor kidneys,”2 and we appreciate the opportunity to respond to their criticisms of our work. First, it is important to recognize that these 2 studies address distinct research questions, which demand different study designs. Although the primary focus of our study was the impact of pregnancy alloimmunization on outcomes in recipients of offspring donor kidneys, the report of Holscher et al asked broadly whether offspring or nonoffspring donor kidneys perform better. Our study design was appropriate for answering our specific research question. The differences between our studies can also be attributed to important methodologic distinctions. Holscher et al criticize our categorization of recipient age using well-validated locally weighted scatterplot smoothing3; however, we found minimal attenuation in the results when we performed sensitivity analyses using various cut-points for age categories or their piecewise spline approach (Table 1). Other substantive methodologic differences between our studies relate to how we approached (1) the sex of the recipient and (2) the characteristics of the donor. Given that many studies have demonstrated that men have worse transplant outcomes than women do, simply assessing for an interaction by sex is insufficient to evaluate the differences in expected vs achieved outcomes. Although Holscher et al describe using interaction terms in their analyses,1 our study specifically investigated sex-specific outcomes using stratified analyses throughout the study.2 With respect to donor characteristics, we chose widely accepted methods of adjusting for recipient and donor factors related to the recipient-donor relationship and our selected outcomes.2 In addition, we restricted our cohorts to individuals with ≥3 HLA matches (and adjusted for HLA matches), mimicking parent-offspring relationships, to identify a robust nonoffspring control group. In contrast, Holscher et al addressed donor age and HLA matches as mediators of the association between recipient-donor relationship and allograft outcomes and elected not to adjust for other important donor characteristics (eg, sex and BMI4). The justification for this decision is unclear and seems problematic given the contemporaneous nature of the recipient-donor relationship and HLA matches.5 We are not aware of any previous studies that addressed HLA as a mediator, and we invite future validation of this unconventional approach. In conclusion, we thank the editors for giving the transplant community the opportunity to consider both of these important, divergent papers. Holscher et al clearly illustrated how the choice of statistical adjustment of raw data dramatically impacts study results. We look forward to reading additional studies in this area, which use innovative analytic approaches that (1) are suited to address the research question at hand; (2) demonstrate clear superiority over previous methods; and (3) are founded on a strong biologic premise. Finally, we appreciate the importance of selecting the best donor for any potential recipient and welcome robust and rigorous investigation in this area. The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.
DOI: 10.1111/ajt.15202
Institution: University of Pennsylvania
Language: English
Subfield: Transplantation
Field: Medicine
Domain: Health Sciences
Open Access status: hybrid
Estimated APC paid: $4,300
OpenAlex ID: W2902052859