Discussion

Our comparative analysis of two large population-based cohort data sets demonstrates that black kidney transplant patients have different baseline characteristics and post-transplant outcomes in England compared with New York State. Most importantly, black patients in England have lower mortality than their New York State counterparts, where black ethnicity in the latter cohort was found to be independently associated with increased 30-day and 1-year mortality. Our analysis, to our knowledge, is the first comparative study comparing black kidney allograft outcomes in contemporaneous population cohorts. Our study highlights significant differences between black kidney transplant patients in England versus New York State and suggests caution in translating outcomes for black transplant patients between different countries.

The strengths and weaknesses of this study must be appreciated for the correct interpretation of our results. There are likely to be numerous confounders that have an impact on black mortality post kidney transplantation that we were unable to factor in (eg, smoking, lifestyle, sociocultural factors and dialysis vintage). It should be acknowledged that New York State data may not be completely representative of other states in the USA, and we are crudely inferring that data from New York State is broadly representative of the USA. Missing data (and misclassification bias) also have an implication on the analyses performed, and this limitation is inherent with all epidemiological studies of this type. Unfortunately, while the majority of data obtainable was matched between the two cohorts, we were unable to extract any information regarding socioeconomic deprivation status or cause of death from the SPARCS data set. Our utilisation of the composite for kidney transplant rejection/failure from administration records is an inferior surrogate compared with registry-derived data regarding kidney allograft rejection and graft failure. The ability to extract more data, and to link data sets to create more comprehensive databases, allows the minimisation of confounding and provides a more robust platform from which to conduct meaningful epidemiological analysis. Further work in this area is clearly warranted and should be undertaken using more robust renal data including registry data from both the USA (eg, USRDS and Scientific Registry of Transplant Recipients) and the UK (eg, UK Transplant Registry and Renal Registry). Finally, mortality data are readily available and were the sole analytical endpoint of this analysis. However, absence of recorded death may not necessarily translate to assumption that a recipient remains alive (eg, lost to follow-up due to emigration).

Current literature suggests that black kidney transplant patients in the USA have inferior kidney allograft survival compared with white recipients,4 and risk for death is magnified among black American versus Canadian kidney transplant patients.7 Reassuringly, recently published work demonstrates significant improvement in post kidney transplant outcomes in the USA over the last two decades, more so for black versus white patients, leading to reduced disparity.5 This may be related to advances in immunosuppression and post-transplant management, possibly benefiting black kidney transplant patients more significantly who perhaps were more disproportionately overburdened by immunological barriers in the past. While some disparity remained in their analysis in relation to significantly raised adjusted HRs for black patients and risk for graft loss in the contemporary setting, the vastly improved outcomes were encouraging for transplant clinicians to actively promote kidney transplantation among black patients. Our analysis suggested worse kidney allograft outcomes since 2007 for all black kidney transplant recipients, which may be due to more high-risk kidney transplantation of medically complex or immunological difficult patients, but this requires further analysis with more transplant-specific data as it may simply reflect better administrative coding of kidney rejection/failure since 2007.

By contrast, no mortality difference between black and white patients has been observed in the few limited publications within European transplant centres.8–10 Our data therefore support the inference from existing literature about disparate outcomes for black patients in the USA versus elsewhere and raises two important questions: (1) Why outcomes for black transplant patients in the USA remain inferior to white patients, and (2) Why black transplant patients in the USA have inferior outcomes in comparison with black patients outside of the USA. It could be argued that the Black population in England differs from the USA, but this is an insufficient explanation as black people in Canada and the USA share geographical origins but still have dissimilar mortality outcomes.7 It is important to distinguish the significant differences in baseline demographics between our two black cohorts. For example, medical comorbidities such as history of diabetes and atherosclerotic diseases were less common among black transplant recipients in England compared with New York State, suggesting a higher level of cardiovascular burden at baseline for the latter. This could explain the significantly increased risk for cardiac events within the first year after kidney transplantation for black kidney transplant patients in New York State versus England. Due to the absence of cause of death from the SPARCS data, we were unable to ascertain the nature of deaths among black patients in New York State, but an increase in cardiac deaths would be a reasonable assumption. However, baseline demographics alone cannot account for the significant difference in mortality.

Perhaps the most fundamental difference between countries is the infrastructure delivering and providing healthcare. Black kidney transplant patients in England have universal health coverage as part of the taxpayer-funded NHS. This ensures complete financial access to immunosuppressive medications and other aspects of care including allied medications (eg, antihypertensive agents, glucose-lowering agents and lipid-lowering drugs) and long-term medical follow-up to support post-transplant care. Compared with the USA, countries such as England, Canada and France that demonstrate similar mortality outcomes after transplantation for black versus white patients share elements of universal health coverage,6 8–10 and this is an obvious factor to explain disparate mortality outcomes. This is also supported by our analysis also showing improved graft survival for white patients in England versus New York States, suggesting our analysis reflects better outcomes overall in England rather than any ethnicity effect. However, universal health coverage alone is unlikely to be the sole explanation for differing outcomes. Prospective clinical trials in the USA, which may attenuate the risk of any lack of access, still demonstrate poorer outcomes for black patients in subgroup analyses,15 suggesting health coverage alone does not fully explain poor outcomes. This is supported by evidence within the USA shown by Chakkera et al in their exploration of racial disparities using kidney transplant cohorts, which included patients receiving healthcare within versus outside the Department of Veterans Affairs (VA).16 The VA is the largest integrated managed healthcare system in the USA and provides comprehensive medical care to eligible veterans including coverage of prescription drugs. Perhaps surprisingly, the association of black ethnicity with poorer outcomes was consistent in non-VA versus VA users, with the latter associated with increased risk of graft loss (adjusted graft failure 1.31, 95% CI 1.10 to 1.57) and death (adjusted mortality 1.10, 95% CI 0.90 to 1.34). While noting adjusted mortality for VA users was not significantly raised in the analysis (likely due to a smaller cohort of 1646 VA users versus 77 715 non-VA users), the adjusted risks for graft failure and death were almost identical between VA and non-VA users (1.31 and 1.31, respectively, for graft failure and 1.10 and 1.11, respectively, for mortality).

This raises the question as to what underlies poor outcomes among black kidney transplant patients in the USA while acknowledging disparate outcomes appears to be narrowing between black and white patients’ post kidney transplantation.5 Factors such as non-adherence, which could be a surrogate for non-affordability of drugs in the USA, is a phenomenon that is common in transplantation cohorts everywhere (including countries with universal health coverage).17 There could be selection bias from poor access to transplantation for black people, which has been acknowledged in transplant centres across the world and has been comprehensively reviewed elsewhere.18 However, from the most up-to-date census reports, individuals of black ethnicity make up 3.5% of the English population19 and 17.6% of the New York State population20 and black patients formed 6.5% and 23.0% of our transplant cohorts, respectively. While this suggests black patients are over-represented as kidney allograft recipients in both England and New York State, respectively, it must be remembered that ESKD is more prevalent in black communities, and therefore they are more likely to need renal replacement therapy. In should be noted that disparate post-transplant outcomes between Europe and the USA are not unique to kidney allograft recipients of black ethnicity alone. Gondos et al, analysing data from the United Network for Organ Sharing and the Collaborative Transplant Study, compared kidney transplant outcomes in the USA versus Europe, respectively.21 Adjusted HRs for graft loss were higher in the USA compared with Europe and identifying the contributing factors to this disparity should be considered a high priority to improve the delivery of post-transplant care. Socioeconomic status, education and poverty are all likely to be important, and black transplant patients in the USA may be an inherent risk for higher mortality or have weaker social support networks or negative health behaviour that drives risk for adverse outcomes (such as smoking, alcohol consumption, sedentary lifestyle and poor diet).22–24 It is clear that our understanding of the social epidemiological impact on health inequalities is limited,25 but deepening our understanding of these issues may further reduce the disparity in mortality outcomes seen in the USA for black kidney allograft recipients.

Our analysis should be interpreted in the context of the well-recognised observation of reduced risk for death for black versus white ESKD patients after commencement of dialysis, and this phenomenon has been noted in both the USA2 and the UK.3 However, inferring that all black patients on dialysis have reduced risk for death on dialysis may be incorrect when interpreting large population cohorts in the context of competing factors. For example, Kucirka  et al conducted an observational cohort study of 1 330 007 incident end-stage renal disease patients in the USRDS and showed black patients have lower mortality on dialysis than white patients (adjusted HR 0.84; 95% CI 0.83 to 0.84; p<0.001).26 However, after stratification by age and treating kidney transplantation as a competing risk, black patients below the age of 50 had significantly higher mortality than their white counterparts on dialysis. Teasing out the differences that help explain our disparate mortality outcomes between black patients in England versus New York State will likely prove more challenging than just a simple explanation. Caution must also be exercised in the interpretation of epidemiological data to ensure false inferences are not made from statistical analyses. For example, Yeates et al found black kidney transplant patients in Canada had significantly lower post-transplantation mortality compared with white patients (HR 0.49; 95% CI 0.28 to 0.88; p=0.02).6 This finding is surprising, and the only documented example of this observation, which requires corroboration before any legitimate interpretation, can be made. However, there are recognised steps that can be taken to reduce the risk of post-transplant mortality for black patients including greater use of living-kidney donors, which has been shown to improve outcomes and to reduce racial disparity of post-transplant outcomes.27

To conclude, this is the first comparative study examining mortality for black kidney transplant patients between two different countries and found black kidney transplant patients in New York State have a greater hazard for death and kidney transplant failure/rejection compared with their black counterparts in England. Our results suggest that further work is essential to investigate poor outcomes for black patients after kidney transplantation in the USA to aid our inadequate understanding of their social, environmental and/or biological pressures. While macro-level factors like universal health coverage may be an important factor contributing to post-transplantation mortality and allograft loss for black patients, our study suggests additional micro-level factors are likely to be concomitantly important and are in need of deeper evaluation.