Outcome Measures in Solid Organ Donor Management Research - a Systematic Review

Background To systematically review published outcome measures across randomised controlled trials (RCTs) of donor management interventions. Methods The systematic review was conducted in accordance with recommendations by the Cochrane Handbook and PRISMA statement. We searched MEDLINE, EMBASE, CENTRAL, Web of Science as well as trial databases from 1980 to December 2019 for RCTs of donor management interventions. Twenty-two RCTs (n = 3432 donors) were included in our analysis. Fourteen RCTs (63.6%) reported a primary outcome relating to a single organ only. Eight RCTs primarily focused on aspects of donor optimisation in critical care. Thyroid hormones and methylprednisolone were the most commonly evaluated interventions (ve and four studies, respectively). Only two studies, focusing on single organs (e.g. kidney), evaluated outcomes relating to other organs. The majority of studies evaluated physiological or biomarker-related outcomes. No study evaluated recipient health-related quality of life. Only one study sought consent from potential organ recipients.


Abstract Background
To systematically review published outcome measures across randomised controlled trials (RCTs) of donor management interventions.

Methods
The systematic review was conducted in accordance with recommendations by the Cochrane Handbook and PRISMA statement. We searched MEDLINE, EMBASE, CENTRAL, Web of Science as well as trial databases from 1980 to December 2019 for RCTs of donor management interventions.

Results
Twenty-two RCTs (n = 3432 donors) were included in our analysis. Fourteen RCTs (63.6%) reported a primary outcome relating to a single organ only. Eight RCTs primarily focused on aspects of donor optimisation in critical care. Thyroid hormones and methylprednisolone were the most commonly evaluated interventions ( ve and four studies, respectively). Only two studies, focusing on single organs (e.g. kidney), evaluated outcomes relating to other organs. The majority of studies evaluated physiological or biomarker-related outcomes. No study evaluated recipient health-related quality of life.
Only one study sought consent from potential organ recipients.

Conclusions
The majority of RCTs evaluating donor management interventions only assessed single organ outcomes or effects on donor stability in critical care. There is a need for an evaluation of patient-centred recipient outcomes, and standardisation and reporting of outcome measures for future donor management RCTs.
PROSPERO Registration CRD42018109487 Background Solid organ transplantation is the preferred cost-effective treatment for end stage organ failure. However, the demand for organs for transplantation currently outweighs the available organ pool despite advances in organ preservation and recipient immunosuppression. Improved organ preservation strategies allow to bridge the time and distance between donor and recipient, and with more experience and better immunosuppression, we have seen a sharp decline in post-transplant morbidity and mortality. Nowadays, the remaining key problem in transplantation is the global persistent shortage of suitable deceased donor organs Due to the lack of available donor organs, mortality rates continue to remain high for those who remain on the transplant waiting list. The majority of deceased organs donated in the UK (60%) and across Europe (over 85%) are from donors with con rmed brain death (Donation after Brain Death, DBD), typically managed in intensive care units (ICUs) prior to organ procurement (1-3). The systemic sequelae of cerebral injury leading to brain death include hormonal, in ammatory and haemodynamic changes with signi cant cardiovascular instability and inevitably a degree of organ damage in the donor. As a result, organs from DBD donors are often declined due to a perceived or actual suboptimal quality. Conversely, the shortage of donor organs has led transplant centres to accept organs from older and higher risk donors often with pre-existing comorbidities enhancing the likelihood of injury prior to transplantation.
After con rmation of brain death and until retrieval of organs by an organ retrieval team, management in the ICU shifts towards donor optimisation. The cornerstones of donor optimisation include: correction of hypovolaemia, maintenance of organ perfusion, corticosteroid therapy, treatment of diabetes insipidus, and lung protective ventilation. The interventions used to delivery these goals have largely been extrapolated from general intensive care unit (ICU) patients or based on pathophysiological rationale (4,5). In the UK, recommendations for optimisation of the brain dead donor have been summarised in a recommended 'donor management bundle' (6). With emerging techniques and targeted interventions identi ed in pre-clinical research, it is important to study whether these interventions do indeed translate into improved organ donor stability, quality of organs at the point of being offered to transplant centres and ultimately improved recipient graft function, survival and health-related quality of life. Furthermore, any systemic treatment administered to the donor can affect all organs and might have different short-or long-term effects on individual organs. In the UK, on average 3.6 organs are successfully transplanted from a brain dead donor (7), thus for any donor management intervention to be translated into clinical practice the effects on all organs must be known to be accepted by the entire transplant community.
Over the past two decades, many randomised control trials (RCTs) of donor interventions or treatments have been conducted to address these evidence gaps. However, in the absence of a core outcome set, there is likely to be widespread variation on the selection and timing of relevant donor and recipient outcome measures, assessment of alternative solid organs and consent processes. We therefore aimed to systematically assess and compare published outcome measures across RCTs of brain-dead donor management interventions.

Protocol and registration
The protocol for this systematic review was registered with PROSPERO (CRD42018109487) This review was conducted in agreement with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting guidelines (8) (PRISMA Checklist, Additional File 1) Eligibility and study selection RCTs of intervention of adult deceased brain-dead donors were included. Studies involving treatments administered after organ procurement was initiated or applied to the retrieved organs alone were excluded. Preclinical and animal studies were also excluded. Studies published in more than one report, such as sub-studies, nested studies or follow up reports were combined. Investigators for studies with outstanding results or minimal information available were contacted, and studies only included if a response with further information was received. One French study was kindly translated by a native speaker to allow inclusion (9). No other studies published in other languages were identi ed.

Search strategy and data extraction
We searched MEDLINE, EMBASE, CENTRAL, Web of Science as well as trial databases (clinicaltrials.gov, WHO International Clinical Trials Registry Platform) from 1980 to July 2019 for RCTs of donor management interventions. An updated complete search was conducted on 28 Dec 2019 and a further focussed search of publications of the past year was undertaken in February 2021. A detailed search strategy is available in Additional le 2. Conference abstracts of major international conferences in the elds of transplantation and intensive or critical care were screened (include list and years). All languages were included. Conferences without available online abstracts were contacted. Two reviewers independently screened the titles and abstracts identi ed by the literature search. Any disagreement was resolved by consensus after discussion or by arbitration by a third author.
A data collection form was created and piloted on ten studies with three authors (KDB, MRE, AS) present, remaining data was extracted by three authors independently. Disagreement were resolved through discussion or by arbitration by a fourth author (RP). We extracted information regarding setting, donor and recipient characteristics, intervention types and outcomes. Two reviewers independently assessed the risk of bias of the included studies using the Collaboration tool for assessing risk of bias (10). Overall risk of bias for each study was then assigned low (all domains low); unclear (one or more domains unclear); high (one or more domains high). We assigned an 'unclear' rating when the study did not report a speci c domain in the published paper or protocol. We did not contact study authors for verbal clari cation. Microsoft Excel was used for data extraction and risk of bias assessment, RevMan was used to create the risk of bias plots (11).

Data analysis
For each study, all primary and secondary registered and reported outcome measures were recorded. Studies were grouped by type of systemic intervention (e.g. steroids) but also by primary target organ (e.g. studies clearly stating kidney graft survival or delayed graft function or creatinine as their primary outcome). For each group of studies using an identi ed intervention, all reported outcomes by donated organ or impact on donor stability were identi ed. For each group of studies with the same primary target organ, outcomes were identi ed and grouped into early (< 28 days after transplantation), late (> 28 days after transplantation) or biochemical-only outcome measures. Microsoft Excel was used for data collection and analysis. Abacus diagrams were created in Lucidchart (www.lucidchart.com). As the purpose of this systematic review was to compare the outcomes used after administration of systemic treatment to organ donors, no meta-analysis of the effects of treatments was undertaken. As part of an exploratory analysis, we also investigated the age, sex and ethnicity representativeness of included trials.

Results
Our search identi ed 17,877 records and we assessed 54 full-text articles for exclusion by screening of titles, duplicates and abstracts. Twenty-two RCTs were included in the nal analysis (PRISMA, Fig. 1). Twenty-one studies were published in English, and one French study was translated by a native speaking researcher (DM).
The 22 RCTs included a total of 3432 donors. Details of the included studies are shown in Table 1. Twenty-one RCTs were conducted in high income countries across Europe and North America, with one study conducted in Iran (12,13). Five RCTs evaluated systemic thyroid hormone therapy, four used systemic steroids and two studies used a combination of steroids and thyroid hormone. Further systemic treatments included albuterol, desmopressin, dopamine, protocolised uid therapy, therapeutic hypothermia, naloxone, phentolamine, simvastatin, vitamin C or a protocolised ventilation strategy and were studied by one trial each. Almost one third (7/22) of all studies focused primarily on kidney transplantation and a further of eight studies were aimed at optimising donor factors in critical care. Four studies primarily studied lung transplantation, two studied liver transplantation and only one study was primarily aimed at heart transplantation.
(1) 100 mg/kg dose of vitamin C 6 hours before procurement followed by vitamin C infusion (100 mg/kg) over the 6 hours; (2) standard care 40 (1) donor IL-6, donor TNF-alpha gene expression; (2) recipient liver function (2015) In ammatory cytokine levels given. Information about nested, follow-up and/or sub-studies referring to the same donor cohort is also provided.

Risk of bias in included studies
All twenty two trials were assessed for risk of bias, with eight at high risk of bias (Fig. 2). Only one trial was at low risk of bias (14). Approximately a quarter of all studies were at high risk of reporting and performance bias. Further details of risk of bias assessments can be found in Additional le 3.

Multitude of reported outcomes across studies
Eight RCTs (36.3%) aimed to study the effects of the intervention on outcomes directly relevant during the donor management period in ICU. Amongst this group, named outcomes included: vasopressor/inotrope requirements, echocardiography parameters, number of transplanted organs, routine biochemical (e.g. thyroid function) or in ammatory (e.g. TNF-alpha) markers, or an assessment of haemodynamic stability. The remaining fourteen studies identi ed one transplanted organ as their main target, with the kidney (n = 7) as the most studied graft, followed by lungs (n = 4), liver (n = 2) and heart (n = 1). Amongst each of these, there was variation of the exact outcomes measured. Effects of treatments on pancreas were only studied in one trial (15) Outcomes by studied intervention All included RCTs administered systemic treatment(s) to the randomised donor. Eleven studies studied the effects of steroids and/or thyroid hormone. The reported outcomes were grouped into outcomes affecting the major transplanted organs (kidney, liver, heart, lungs) or donor factors (such as factors haemodynamic stability or number of organs accepted for procurement). Figure 3 demonstrates that nearly all studies (21/22) of systemic treatments do not report outcome data across all the outcome domains. Only one study comparing ventilation strategies of the donor covered all outcome groups, albeit only reporting the recipient survival for each of the major transplanted organs (14). Overall, the kidney or donor factor outcome groups were most often included, with each contributing to seven intervention types. DGF was the most common reported renal outcome (6/7 studies), whilst donor factors included a variety of different outcomes such as number of transplanted organs, inotrope or vasopressor requirement, left ventricular ejection fraction or cardiac index.

Outcomes by studied primary organ
Many of the selected outcomes depended on the primary target, e.g. kidney or physiological measurements in ICU. Studies focusing on kidney or heart transplantation were more likely to provide a more comprehensive assessment of other organs, as displayed in Fig. 4. Early outcomes -de ned as within 30 days of transplantation -more commonly reported surrogate outcomes such as changes in laboratory markers or echocardiographic function. Trials of interventions aimed at donor stability only rarely assessed organ speci c function or outcomes -such as pre-transplant lung function, biochemical liver function assessment or mention of primary graft dysfunction for cardiac allografts. Furthermore, none of the donor stability RCTs assessed any long-term recipient or graft outcomes; therefore the long term effects of systemic treatment administered to nearly a quarter of all donors (23.3%, 800/3432) have not been collected by included trials. Organ-focused trials included long-term follow up of either one, three or ve years of graft survival and between six months and three years of recipient survival as shown in Fig. 5. Only one study reported the incidence of long term post-transplant complications relating to immunosuppression, such as post-transplant lymphoproliferative disease (15). More speci c measures of long term graft function after transplantation (such as creatinine or liver function tests) were only assessed in two studies (19,20). Organ speci c rejection at 30 days and 3 months was studied in two trials aimed at the kidney (21,22). Rejection episodes within 6 or 3 years follow up were reported in both trials aimed at the liver (20,23).  (15,17,24,25). Donor ethnicity was reported by three studies only (16,17,26), with two further studies (27,28) reporting the percentage of Afro-American donors in the two groups. The three studies which provide a breakdown of donor ethnicity describe a study population of between 60-80% white donors. Nine studies provide sex or age information for the recipients of the donated organs, whilst information on recipient ethnicity is not reported by any of the included trials.

Key ndings
The main ndings from this systematic review are: (a) there are a multitude of different early and long term outcomes studied; (b) bene cial or harmful effects of systemic treatments are not studied across all donated organs; (c) recipient-centred outcomes are mostly limited to duration of graft and/or recipient survival (d) donor and recipient age, sex and ethnicity are not consistently reported across all trials To date systematic reviews of effects of donor interventions have failed to demonstrate a bene t for any identi ed treatment (29-31). Published meta-analyses often demonstrate small number of identi ed trials and high heterogeneity. This systematic review set out to identify outcomes studied in the context of donor management research. Identi ed studies focus mostly on single organ outcomes or outcomes relating to donor stability without assessing if (or how) this translates to improved graft function or graft and recipient survival. The included studies can be grouped together either by studied intervention (e.g. steroids) or by the primary identi ed target (e.g. kidney).
Many of the selected outcomes appear to depend on the intended primary target organ of each study rather than the selected treatment. Therefore, grouping the trials by intervention alone does not provide a complete picture. The plots of outcomes by intended primary target organ or donor stability (Figs. 4 and 5) analyse the published outcomes further, comparing early outcomes (within 30 days) and long-term outcomes and duration of follow up. Studies primarily focused on kidney or heart transplantation provide more comprehensive assessment of early outcomes relating to other organs. Studies mainly focused on donor factors/stability fail to assess many early outcomes relating to transplanted organs other than liver function tests. In addition, they fail to assess whether the interventions aimed to improve donor stability translate to long-term bene ts for any of the recipients of transplanted organs. Early outcomes for kidney, liver and heart transplants mostly report surrogate outcomes such as biochemical serum levels or function recorded on echocardiogram. All four included studies of donor management in lung transplantation only report a measure of graft function before procurement but not after transplantation. Lung transplant recipient survival is only assessed in one of those studies, at limited to a follow up period of six months.
With an increased interest in donor management research re ected by the increasing numbers of clinical trials and the increasing demand for organs, there is a need for standardisation or creation of a core outcome set for donor management research as has been done in other settings ((32, 33). Any such discussion would bene t from involvement of donor families, hypothermia study to monitor that lowering the donor body temperature does not inadvertently affect thoracic organshowever this was not assessed as an outcome of the study (42). The available donor pool across all nations is still by far outnumbered by long waiting lists, therefore novel treatments should not omit to demonstrate their impact on every transplantable organ. In addition, demonstrating a bene t in more than just one organ group might give the proposed treatment more weight. To complicate matters from a methodological point of view, we noted that only 4/21 studies explicitly considered to seek consent from the recipients of the organs procured after donor intervention (24); only one of the included studies did indeed seek recipient consent (43).
Whilst outcome measures formed the mainstay of this review, two main methodological problems were identi ed during the appraisal of the included studies. Donor and recipient characteristics such as ethnicity and sex are not reported in a standardised manner across the studies. Both aspects matter beyond the concept of missing data and minutiae of trial methodology. Steroids are amongst the most studied interventions to improve donation outcomes, their role has also been previously often debated in the context of critical illness. Several studies of the use of steroids in sepsis have been undertaken(44) yet a de nite protective role to change clinical practice has not been proven beyond a reasonable doubt.
Critical care trials are inherently di cult due to a heterogenous, often multi-morbid population and it may simply be that any bene cial effect is too small to be observed under such circumstances. This translates to trials of treatment of the brain dead donor, with a variety of different underlying pathologies across the study population. The COVID-19 pandemic offered a rst opportunity to study the role of steroids for one de ned pathology within the critical care population and the results of the RECOVERY trial con rmed a bene cial role for dexamethasone for patients requiring oxygen or respiratory support (45)(46)(47). Both biological sex as well as ethnicity (amongst other factors) are suggested to play a role in contributing to the shape of the in ammatory response during severe illness (9). The contribution of such donor factors to their in ammatory response and how it could be shaped by treatments, is therefore an important factor to consider and ought to be routinely reported for the donor (and recipient) population in donor management studies. The latest NHS Blood and Transplant data shows that up to 31% of patients on the waiting list in the UK are from a black, Asian or minority ethnic background and spend a longer time on the waiting list than white patients. Understanding how to improve the quality and outcome of organs from ethnic minority donors -or for ethnic minority recipients -will have a big impact.
The strength of the review is the strict methodological process. Our review protocol was prospectively registered, we followed Cochrane Collaboration and PRISMA recommendations, performed a comprehensive search and carried out duplicate data extraction and risk of bias assessments. The limitations to the conclusions of this systematic review can be attributed to the clinical and methodological differences between the trials. We focused on donor interventions only, and therefore potentially promising treatment strategies of the graft after procurement and during preservation, that may be of a pre-transplant treatment potential, were excluded.

Conclusions
This systematic review provides a rst and an up-to-date systematic overview of all outcomes studied in RCTs of systemic treatments or interventions to the organ donor. Across the included RCTs, there was a wide range of outcomes that can be subdivided into donor stability factors, graft outcomes (early or late) and recipient outcomes. Not only did the majority of studies limit their selected outcomes mostly to their target domain, even within each domain there was a distinct absence of agreement about the exact outcome parameters or duration of follow up, e.g. a distinct lack of recipient-centred outcomes, such as health-related quality of life. The urgent need to improve organ utilisation to reduce mortality on the waiting lists for transplantation in the UK and Europe is nowadays addressed with an increased drive to develop methodologies for RCTs that evaluate targeted interventions in deceased donors after con rmation of brain death. Interventions that succeed in preventing or at least reducing injury of organs will increase the available organ pool for successful transplantation leading to prolonged survival of organs and patients. The results of this review can be used as a basis for future priority setting exercises, developing core outcome sets and to guide future donor management research. RJP is the PI for the Quality in Organ Donation programme.

Abbreviations
The remaining authors declare they have no competing interests.   Outcomes by intervention Legend: Abacus diagram overview of outcomes from donor management trials by intervention. Steroids and/or thyroid hormones were the most commonly studies interventions, with all other interventions only addressed in one trial each (shown alphabetically). For each studied intervention, the overall number of donors randomised is shown on the right-hand side, with a breakdown how many donors contributed to each outcome group shown in vertically aligned beads. Each bead demonstrates how many of the included trials reported this outcome (in green) and how many did not (in red)

Figure 4
Early outcomes (<30 days) by target organ Legend: Vertical Abacus diagram summary of reported early outcomes (within 30 days of donor intervention) by primary target. The total number of donors randomised to each primary target group is shown at the bottom, with the kidney as the most commonly studied primary target. Outcomes relating to other organs, donor speci c parameters (such as total number of transplanted organs or donor stability) or measured biomarkers are aligned horizontally. Each bead indicates how many of the studies reported each speci c outcome (in green) and how many did not (in red)

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.