Sautenet et al1 report the results of a systematic analysis of published trials in kidney transplantation over the last 42 years. They classify outcomes into outcome domains and categories. They found 105 different domains from 397 studies. The 4 leading outcome domains were graft function, acute rejection, graft loss, and mortality. More than half of the outcome domains were surrogate outcomes. Only 15% had patient reported outcomes. They conclude that there is a large variability in the structure of reporting outcomes for clinical trials in transplantation.
In 2004, Fritsche et al2 reported on the quality of trials of immunosuppression in kidney transplantation. This study included 63 publications. They classified the quality of the articles using the Jadad scale. On a scale of 1 to 5, the mean was 2.3 represented low over all quality. For example, 27% of the publications did not report whether graft loss was censored for death. Most results did not include confidence intervals. Even the largest trials of immunosuppression showed considerable deficits in the design and publication.
The Consolidated Standards of Reporting Trials (CONSORT) criteria were widely adapted by the majority of peer review medical journals in 2010.3 The CONSORT criteria addressed many of the deficiencies in reporting of clinical trials at that time. We would expect that these criteria have resulted in improvements in study design and publications by the major medical journals since their adaption in 2010. The study by Sautenet et al included trails published since 1974. Sixty-eight percent were published before 2010. The study does not address the impact of the era on the reporting of outcome variable. More recently, there has been significant changes in the importance of acute rejection versus longer term outcomes, as well as recognition of BK virus and donor specific anti HLA antibodies as important end points.
The nature of recent studies also obligates larger volumes of outcome data. For example, studies including genomic data can produce thousands of outcome variables, and by necessity, the reported results will not be comprehensive. This has resulted in a policy that requires investigators to submit raw genomic data to a central repository so it is available to other investigators. As a result of the more sophisticated study design and methods for clinical trials, the results of these studies include more outcome data. However, editorial policies limit the volume of text and data. As a result, investigators make a strategic decision to report only the data and outcomes that are necessary to support their study hypothesis and the conclusions.
Most medical journals limit the volume of text, figures, and tables but allow authors to submit supplement materials. As a result, authors will frequently submit supplemental materials which may include outcome data; however, supplemental files were not included in the analysis by Sautenet et al.
The authors point out that 52% of the reported outcomes are surrogate outcomes. The majority of these surrogate outcomes have not been validated. As a result of the advances in immunosuppression and surgical techniques in kidney transplantation, the 1-year patient and graft survivals exceed 90% and the risk of rejection is less than 20%. However, this has complicated the ability to adequately power clinical trials to demonstrate significant differences in clinical outcomes in reasonable periods of follow-up. As a result, there is increased pressure to develop validated surrogate outcomes for clinical trials. For example, genomic biomarkers are promising area for developing surrogate endpoints; however, the majority of the biomarker studies have not been adequately validated as surrogate endpoints for clinical trials.4
Sautenet et al suggest the need for more stringent criteria for reporting outcomes of clinical trials. However, because of editorial policies and limitations in space, it may not be possible to include all outcome data. Furthermore, studies often ask very specific questions that may only require variables pertinent to answering the hypothesis being asked. Instead of requiring the reporting of more outcomes, there may be other potential solutions to make outcome data more accessible to transplant community. This might include a more liberal use of supplemental files. Similar to the process for publishing genomic data, the transplant community might consider requiring authors to submit raw clinical data to a data repository.
Sautenet et al point out the limited inclusion of patient reported outcomes in clinical trials. Including patient reported quality of life measures will allow for better assessment of the value of new therapies. The FDA is now requiring the inclusion of patient reported outcomes in studies to support labeling claims for medical products. Including quality of life data are an important way to engage societal and industry stakeholders in the business of transplantation and to facilitate the communication of the human and societal impact of transplantation.
The study by Sautenet et al shines a light on the wide variability of reporting outcomes from clinical trials in kidney transplantation. Despite the many advances in the design and analysis of outcome data from clinical trials, this large variability of reporting might suggest a continued opportunity for quality improvement in the process.
1. Sautenet B, Tong A, Chapman JR, et al. Range and consistency of outcomes reported in randomized trials conducted in kidney transplant recipients: a systematic review. Transplantation
2. Fritsche L, Einecke G, Fleiner F, et al. Reports of large immunosuppression trials in kidney transplantation: room for improvement. Am J Transplant
3. Schulz KF, Altman DG, Moher D, Group C. CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials. BMJ
4. Kurian SM, Whisenant T, Mas V, et al. Biomarker guidelines for high-dimensional genomic studies in transplantation: adding method to the madness. Transplantation