Nephrol Dial Transplant. 2017 May 1;32(5):890-900. doi: 10.1093/ndt/gfx036.
Barriers to living donor kidney transplantation in the United Kingdom: a national observationalstudy.
Wu DA, Robb ML, Watson CJE, Forsythe JLR, Tomson CRV, Cairns J, Roderick P, Johnson RJ, Ravanan R, Fogarty D, Bradley C, Gibbons A, Metcalfe W, Draper H, Bradley AJ, Oniscu GC.
PMID: 28379431 Full free text available at NDT
Check out the #NephJC summary of the Purnell et al study in JAMA as well as the brilliant visual abstract. One could summarise the study as showing that not only have living donation rates not improved in the US in the last 2 decades, but the disparity of lower rates of living donation amongst black/Hispanic/Asian transplant recipients is getting worse. Is this purely an American problem, related to its history of slavery, and institutional racism? Has the Civil Rights movement, affirmative action and all else that has happened come to naught? That might be the easy conclusion, but the wrong one to draw. In the Netherlands 44% of patients on the transplant waitlisted are of a non-European background (non-Europeans in Netherlands include individuals from Indonesia, Morocco, Turkey, Surinam, Antilles, and other countries), but make up only 15% of living donor kidney transplants. In a follow up study of 1338 transplants done at Erasmus (ie Rotterdam, Netherlands), despite adjusting for socioeconomic factors (which explained some of the disparity), non-Europeans were less likely to receive a living donor transplant. Most of these studies are retrospective, but the ‘Access to transplantation and transplant outcome measures’ (ATTOM) research programme is a well conducted prospective study that can help unpack the issue of transplant disparity more fully.
We can look closely at their 2017 study which did precisely that.
What did they do?
The ATTOM study (check out detailed methodology here) is a prospective study of transplant recipients (as opposed to waitlisted patients, as Purnell et al did) from 23 UK transplant centres, recruited over a 12 month period (2011-3). Of the 3005 patients who were transplanted during that period, 2055 consented and/or had data to be eligible in this study. Ethnicity in this study was coded as White, Black, Asian (which here includes mainly people from the Indian subcontinent, and not of Chinese origin), or other (including patients of Chinese and mixed origin). Apart from the usual demographic and comorbid condition details, they also collected educational attainment and employment status. The primary analysis was a logistic regression to estimate the odds of receiving a deceased versus living kidney transplant. Some additional sensitivity analyses were also conducted.
What did they find?
About 60% of all transplants were deceased donor and 40% living donors. The table 1 has all the details of the differences between those who had a live versus a deceased donor transplant. For ethnicity, white recipients were more likely to have a living donor transplant, while Asian, black and others were more likely to have had a deceased donor transplant. The Geographic variation is also interesting - more living donors in Northern Ireland (perhaps related to this?), and less so in Scotland.
Table 1 from Wu et al, NDT 2017
In the adjusted analysis (table 4 and figure 2), black and Asian ethnicity were still associated with lower odds of receiving a living donor transplant, with white patients being referent group. Other factors that increased odds of getting a living donor transplant:
- Car ownership
- Home ownership (car and home were significant individually, even after adjustment)
- Younger age
- Being married
- More educated
- Living in Northern Ireland
Figure 2 from Wu et al, NDT 2017
There are some interesting donor details to parse presented in table 2 and 3. Compared to deceased donors, living donors were more likely to be younger and women - but also more likely to be white, suggesting a paucity of non-whites overall in the donor pool (deceased and living). Spouses and siblings made up roughly half the living donors, though pooled/altruistic donors (ie non-directed, presumably), notably, made up a not-insignificant 11.5%. Additional sensitivity analysis restricted only to white recipients showed that the relationship of the socioeconomic factors reported above also help true in this analysis. Lastly, an analysis of living donor recipients showed that Asian ethnicity (but not black, compared with white as referent) was also associated with a lower likelihood of pre-emptive transplantation.
How does this compare with the JAMA study?
There are many differences between the ATTOM study and the one JAMA published recently.
- This is a prospective study, consisting only of transplant recipients and comparing who received living versus deceased donor transplants. Purnell et al have a study of of waitlisted candidates, and showing trends over the last two decades of the odds of receiving a live donor transplant within 2 years of being listed.
- The ATTOM study was done in UK - which has universal healthcare, not the case in the US - though interestingly in the Purnell study, the rates of living donation were higher in those with private insurance compared to Medicare/Medicaid/Uninsured (the absolute rates were higher in whites AND other ethnicities, so the relative numbers, the Hazard ratios, remained similar).
- The ATTOM study adjusted for a whole lot of socio-economic factors, and showed that the differences for ethnicities persisted after adjustment. Purnell et al adjusted for biological factors only in the main analysis, and not socio-economic factors. That analysis, done for one factor at a time, is provided in the supplementary appendix. Why do it this way? Indeed, many social epidemiologists do suggest that adjusting away socio-economic factors is not appropriate, when trying to understand the relationship between race/ethnicity and outcomes. A textbook definition of confounding is a risk factor for the outcome (in this case living donation), independent of the putative risk factor (ie race/ethnicity). The confounding factor must be associated with the putative risk factor (ie low socioeconomic conditions and race/ethnicity link), but most importantly, it should not be in the causal pathway between exposure and disease. We really cannot say that about the relationship between race/ethnicity and living donation, can we? If the effect of race/ethnicity is mediated via lower socio-economic status, should we ‘adjust’ for it? Read this excellent review for a longer discussion of the topic.
Overall, the issue with differences between white and other race/ethnicities seems to spillover beyond national boundaries, and considerations of universal healthcare. Increasing donation rates will require more focused interventions, targeting these groups with effective interventions.