Socioeconomic, demographic and policy comparisons of living and deceased kidney transplantation rates across 53 countries

Kidney transplantation (KT) has become the preferred modality for the treatment of end-stage kidney disease. It offers improved survival, decreased morbidity, fewer cardiovascular events, lower healthcare costs and significantly improved quality of life when compared with dialysis. Systematic reviews demonstrate that this applies to both living and deceased donor kidney transplants and all forms of dialysis internationally.1

We estimate from our data collection (see Methods) and publications by Schieppati and Remuzzi (2005)2 that more than 1.7 million people are on dialysis worldwide and of these, about 270 000 are on kidney transplant waiting lists. Of the 104 000 solid organ transplants performed in 2010, about 71 000 people received a kidney transplant (33 000 from live donors and 38 000 from deceased donors3), representing only 4% of the world's dialysis population. Despite efforts to improve the number of kidneys available for transplantation, dialysis growth has been estimated at approximately 8% per annum whereas kidney transplantation rates (KTR) have grown at only 4%, meaning that the gulf between the numbers of people who require dialysis and those who will receive a transplanted kidney is increasing over time.3-5 While not all dialysis patients will be suitable for transplantation, the shortfall between the number of people on dialysis and those able to receive a kidney transplant, an issue noted previously by those concerned about both the care of patients with end-stage renal failure (ESRF) and the benefits and ethics of transplantation,6 remain highly problematic, particularly in light of increases in the survival advantage associated with transplantation over the past two decades.1

Living donor (LD) and deceased donor (DD) (comprised of both donation after brain death (DBD) and donation after cardiocirculatory death (DCD) donors) KTR vary greatly across religions, countries and geographic regions. Although many possible explanations for this wide variation have been proposed,6 the major determinants of kidney transplantation rates remain unclear. Previous studies of KTR have focused on the effects that specific, and largely singular, factors play in living or deceased transplant rates. These have included the effects on transplantation rates of religion,7, 8 probabilities of need,9 demographic factors and attitudes,10 marital status,11 gender,12 geographic variation,13 explicit and presumed consent policies14-17 and socioeconomic status.13 While these studies have provided valuable insight into the potential for single locus variables to influence rates of renal transplantation, few have focused upon more than one country, and, while one13 has focused on several of the variables included in our study, none has attempted to associate the relationships of so many different factors for both living and deceased KTR during a specific, homogeneous time frame, across such a large number of countries. Our univariate and multivariate analysis of 15 socioeconomic, policy and demographic factors across 53 different countries provides, to date, the largest and most comprehensive analysis of factors that may potentially influence KTR.

Methods

Data for both LD and DD KTR (in donors per million population) were drawn from published material for 53 countries by accessing publicly available registries listed in Table 1a. The fifteen socioeconomic, demographic and policy factors listed in Table 2 were then collected for each of these countries from the sources listed in Table 1b. These 15 factors chosen were selected because; relevant data was available across all (or most) of the countries studied; these factors had previously been shown to be related to organ donation and/or transplantation rates in more limited studies; and/or because it seemed likely that these factors may influence rates of organ donation and/or kidney transplantation.

This information was then cross-checked for accuracy by comparing data integrity across multiple sources (when multiple sources of the same data points were available). In cases where there was inconsistency between sources we requested verification of the correct rate from the national and international nephrology organizations listed in Table 1a. If verification was unavailable, we used the donor rate from the source that had been updated most recently.

In order to form a logical and equitable distribution of the countries into geographical areas, the 53 countries studied were then grouped into nine regions. The regions were based largely on the six world regions defined by the WHO with a further three defined in order to detect potential differences where cultural aggregations were known to exist (Europe East, Europe North, Europe South, Asia North and Asia South). Mexico and Cuba were included in the America South region because of common language and culture. The nine regions were defined as follows: America North (Canada, USA); America South (Argentina, Brazil, Chile, Colombia, Cuba, Mexico, Uruguay); Asia North (Hong Kong, Japan, South Korea, Taiwan); Asia South (India, Malaysia, Thailand, Singapore); Europe East (Bulgaria, Croatia, Cyprus, Czech Republic, Estonia, Greece, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia, Slovenia); Europe North (Austria, Belgium, Denmark, Finland, Germany, Ireland, Luxembourg, Netherlands, Norway, Sweden, Switzerland, UK); Europe South (France, Italy, Portugal, Spain); Middle East (Iran, Israel, Jordan, Saudi Arabia, Turkey); and Oceania (Australia, New Zealand).

The existence of a nationally managed organ donation and transplant programme was defined as a national body (governmental or otherwise) tasked with management and regulation of organ donation and transplant policies. We defined presumed consent countries (also known as ‘opt-out’), as those with legislation in place that defines a deceased organ donor as someone who has not expressed written opposition to becoming a deceased organ donor. Explicit consent countries (also known as ‘informed consent’ or ‘opt-in’) are those without presumed consent legislation. It is noteworthy that based upon the information received from their respective national transplantation societies, we have defined both Norway and Israel as explicit consent countries, which conflicts with data from previous studies.15, 16

Predominant religion of each country was determined using the US CIA Factbook (Table 1b) and then grouped into the following categories: Catholic; Protestant; Orthodox (Christian); Jewish; Muslim and Eastern. In order to facilitate statistical analysis of the religions in Asia, we grouped ones unique to particular countries (i.e. Shintoism (Japan), Confucianism (Hong Kong) and Hinduism (India)), together with Buddhism, into ‘Eastern’ faith traditions. Where there was no majority religion (>50%), the predominant religion was selected.

Univariate and multivariate linear regression models were fitted separately for DD KTR and LD KTR. For each model, P-values were calculated from the corresponding t- and F-tests and r² also estimated. For the multivariate model, all variables were initially included except for organ transplant waiting list, dialysis population and government expenditure on education (because of incomplete data available for these factors). Variables that were highly correlated were also excluded from the multivariate model if the Variance Inflation Factor (VIF) was greater than 10.18 The remaining variables in the model were removed one at time, starting with the least significant, until only significant variables remained. Separate multivariate models were also built which included transplant waiting list, dialysis population, education expenditure (both per pupil and as a percentage of total government expenditure), using the same process as described above. All analyses were conducted in Prism™ (GraphPad Software Inc., La Jolla, CA, USA) and Stata^™ 12.1 (StataCorp LP, College Station, TX, USA). Results were considered statistically significant if P < 0.05.

Results

Deceased donation kidney transplant rate – univariate analysis

Multiple factors appear to be associated with the DD KTR. Higher rates were observed in countries with presumed consent legislation (Fig. 1a), existence of a nationally managed organ donation programme (Fig. 1c) predominantly Catholic faith traditions (Fig. 2a), the European geographical region (particularly Europe South – Fig. 2c), higher GDP, education expenditure both per pupil and as a percent of national budget, national health expenditure per capita, and an older population (both mean age and percent population >80 years old). Of all the factors analysed, the most highly significant associations with DD KTR were presumed consent legislation (Fig. 1a), existence of a nationally managed organ donation programme (Fig. 1c), population age, geographic region, and UNHDI (United Nations Human Development Index) (all P < 0.0001) (Table 2).

Table 2. Results from the univariate analysis

		DD KTR				LD KTR
Factor	n	r2	Mean change	(95% CI)	P	R2	Mean change	95% CI	P
Presumed consent legislation	53	0.23	12.53	(6.09; 19.06)	<0.0001	0.22	−8.7	(−12.96; −3.68)	0.003
% Population >80 years old	53	0.40	5.49	(3.57; 7.49)	<0.0001	0.03	−1.08	(−2.64; 0.74)	0.22
Median population age	53	0.26	1.16	(0.61; 1.72)	<0.0001	0.03	−0.29	(−0.74; 0.17)	0.21
Religion	53	0.47			<0.001	0.24			0.02
Catholic	21		Reference				Reference
Protestant	12		1.44	(−5.81; 8.68)			9.38	(3.22; 15.54)
Orthodox	7		−8.93	(−18.97; 1.11)			8.14	(−0.39; 16.67)
Jewish	1		−13.32	(−34.11; 7.46)			3.70	(−13.96; 21.37)
Muslim	5		−21.42	(−31.46; −11.38)			12.02	(3.49; 20.56)
Eastern	7		−20.40	(−28.75; −12.05)			3.21	(−4.25; 10.68)
Kidney transplant waiting list (pmp)	33†	0.05	0.05	(−0.03; 0.13)	0.22	0.10	0.06	(−0.005; 0.13)	0.69
Adult diabetes rate (pmp)	53	0.06	0.18	(0.02; 0.38)	0.07	0.004	0.03	(−0.11; 0.18)	0.65
Dialysis rate (pmp)	33†	0.05	0.01	(−0.007; 0.03)	0.23	0.002	0.001	(−0.009; 0.01)	0.83
Nationally managed deceased donor programme	53	0.30	19.25	(11.02; 27.48)	<0.0001	0.01	2.21	(−5.06; 9.49)	0.55
Geographic region	53	0.62			<0.0001	0.21			0.18
America North	2		Reference				Reference
America South	7		−13.53	(−27.86; 0.79)			−12.08	(−26.78; 2.62)
Asia North	4		−23.87	(−38.82; −8.92)			−11.08	(−26.96; 4.80)
Asia South	4		−25.52	(−40.99; −10.04)			−10.40	(−26.29; 5.48)
Europe East	13		−6.07	(−19.64; 7.50)			−11.29	(−25.22; 2.64)
Europe North	12		−1.75	(−15.40; 11.89)			−6.60	(−20.60; 7.40)
Europe South	4		10.75	(−4.73; 26.22)			−13.73	(−29.61; 2.16)
Middle East	5		−22.99	(−37.94; −8.04)			0.20	(−15.15; 15.54)
Oceania	2		−13.33	(−31.19; 4.54)			−4.14	(−22.47; 14.20)
Health expenditure (Log(US $ per capita))	51†	0.32	7.46	(4.35; 10.57)	0.001	0.005	0.64	(−2.04; 3.33)	0.63
Health expenditure (% Gov't Budget)	51†	0.12	1.23	(0.29; 2.17)	0.12	0.002	−0.12	(−0.83; 0.60)	0.74
Education expenditure (Log(US $ per Pupil))	41†	0.17	6.94	(1.97; 11.91)	0.007	0.03	−2.35	(−6.43; 1.72)	0.25
Education expenditure (% Gov't Budget)	48†	0.13	−1.01	(−1.77; −0.24)	0.01	0.003	−0.10	(−0.66; 0.45)	0.7
UN HDI	52‡	0.30	−0.26	(−0.37; −0.15)	<0.0001	0.003	−0.02	(−0.11; 0.08)	0.58
GDP (log(US $ per capita)	53	0.13	0.22	(0.06; 0.38)	0.008	0.04	0.09	(−0.03; 0.21)	0.13

• †<53 countries analysed due to incomplete data available. ‡Taiwan is not included in UN HDI rankings. n = number of countries; r² = Correlation Coefficient; P-values calculated from t- or F-tests. DD, deceased donor; KTR, kidney transplantation rates; LD, living donor; UN HDI, United Nations Human Development Index.

Discussion

This study demonstrates that clear differences exist between factors that influence living (LD) and deceased donor (DD) kidney transplantation rates (KTR). Presumed consent, GDP and religion were the only factors that influenced rates of LD KTR – with presumed consent demonstrating an unexpectedly high negative association with LD KTR in the univariate model. In contrast to LD, increased DD KTR was positively associated with presumed consent as well as with a number of social, medical, economic and demographic factors. The fact that the greatest increases observed in all of the factors in our univariate analysis were existence of a nationally managed donation and transplantation programme (which increased DD KTR by 19.25 transplants pmp (95% CI: 11.02, 27.48; P < 0.0001)) and presumed consent legislation (which increased DD KTR by 12.53 transplants pmp (95% CI: 6.09, 19.06; P < 0.0001)) demonstrates the important roles that nationally managed legislative and political reform can play in increasing DD KTR. While the multivariate analysis showed a reduced rate of increase (relative to the univariate analysis), it nevertheless increased at a rate of 11.91 transplants pmp in the multivariate DD KTR model. Interestingly, presumed consent decreased LD KTR by 8.7 transplants pmp (95% CI: −13.32, −4.01; P = 0.003) in the univariate model. The antipodal effect that presumed consent played on LD and DD KTR is both significant and surprising and is the subject of ongoing research.

Because LD kidney transplants generally originate from a donation by the recipient's relative, we were surprised to find that need (as expressed by countries with higher rates of dialysis and larger kidney transplant wait-list) was not associated with increased LD KTR. Although because of limitations in the data sets available to us we were unable to exclude age bias (i.e. older potential recipients being excluded from waiting lists), the fact that not all living kidney recipients may have been listed on kidney transplant waiting lists, and certain qualitative factors (such as lack of awareness and fear and mistrust of the transplant system) as potential reasons for this result, our finding that need does not appear to influence LD KTR remains unexplained.

Likewise, factors that may be associated with higher standards of medical practice and optimal donor care such as GDP, health care expenditure and education expenditure were also not associated with increased LD KTR. One reason for this may relate to the fact that living related organ donation is, simultaneously, both an intimate and interpersonal exchange and, in most developed countries, a highly regulated clinical ‘exchange’. It is possible, therefore, that this regulation may interfere with, or even preclude, associations with broader socioeconomic variables.

In contrast to LD KTR, a series of factors that are important to the construction, maintenance and functioning of an effective deceased organ donation and transplant programme were associated with higher DD KTR. These include: a nationally managed organ donation programme; presumed consent; the UNHDI; health expenditure per capita and GDP. Likewise, factors that may be associated with higher population awareness of need and with support for organ donation and transplantation were also positively associated with increased DD KTR, including age (both median population age and percentage of population aged >80 years) and education expenditure (both per student and percent of overall government expenditure).

Similar to LD KTR Rates, our univariate analysis demonstrated that factors that one might anticipate may be associated with need and increased demand for deceased kidney transplants, such as rates of adult diabetes, the number of people on dialysis and the number of people on transplant waiting lists, were not associated with DD KTR.

While both geographic region and religion were also strongly associated, both positively and negatively, with increased DD KTR, with both European and predominantly Christian countries having higher rates of DD KTR, it is not clear that the association between increased DD KTR rates observed in Christian countries were a consequence solely of the influence of Christian values or beliefs about organ donation. Virtually all Abrahamic and Eastern faith traditions valorize giving, altruism, compassion and justice and these values play a strong and defining role in the rituals surrounding death, dying and burial as well as the definition of the cultural meaning and value of the body's post mortem integrity. This makes broad faith-based interpretation of the wide variance observed in LD & DD KTR between faith traditions difficult. Therefore, while it is tempting to suggest that Christian values (particularly those of the Roman Catholic church) may influence KTR,19 because of similar support of organ donation and transplantation by other faith traditions,20, 21 such a concise explanation may not be appropriate. An explanation such as this also fails to account for why LD KTR in Catholic countries was among the lowest in our cohort (Fig. 2b). The link between religion and organ donation, therefore, seems both complex and paradoxical.

A number of limitations to this study may affect our conclusions. We did not examine the impact of unique, and in some cases extraordinary, policies operating in countries, such as the effect of remuneration to donors and donor families (as is the case for Saudi Arabia and Iran) as these policies apply in very few countries (making meaningful international comparison impossible), and may be inconsistent with international declarations. The effect that remunerative policies may have on KTR, although largely inconsistent with international declarations, is perhaps, worthy of further examination in future studies. Additionally, because we restricted our analysis to 2008/2009 kidney transplant rate data, it is possible that idiosyncratic or inconsistent results that may have occurred during this specific period (such as short-term effects caused to transplantation rates by the publication of the Declaration of Istanbul) and could be very different from data analysed over longer time-frames. In addition, while we managed to generate an extraordinarily complete data set for the countries under study, this required collection of data from multiple sources. Given that processes for collecting donor, transplant and socioeconomic metrics differ between both the organizations collecting them and the countries from which they were taken, it is difficult to be absolutely certain as to the veracity of all the data collected. In spite of these limitations, our results provide a solid, empirical foundation for further debate regarding policy development in organ donation and transplantation. Also, while we fully recognize that the religious pluralism that characterize most countries' religious identities make categorization with one dominant faith tradition difficult, all of the countries in this study did have a clearly dominant religion that was appropriate for statistical analysis. Finally, while we would like to have included other factors in our analysis, including; the number of ICU beds per population; the use of extended criteria donors (ECD); the existence of policies for paired kidney exchange; the existence of a nationally funded universal health system, and the specific levels of financial remuneration received nationally by medical institutions for kidneys retrieved (both LD and DD), this information proved impossible to collect across such a large number of countries.

There are several implications that can be drawn from our analysis. First, efforts to increase KTR should regard LD and DD as completely separate and distinct entities with factors influencing them unique to each. Second, the absence of identifiable factors linked to LD KTR suggests that it may be more influenced by interpersonal factors than by common, socioeconomic variables that are globally monitored. As noted in previous analyses, living kidney donation programmes frequently play the central role in transplantation programmes in many countries.22 Because the factors associated with increased LD KTR appear complex and, not readily amenable to simple policy changes, more study is needed to better understand how donor and recipient outcomes in LD KT can be improved. In contrast, rates of transplantation of kidneys from deceased donors seem to be strongly correlated to many socioeconomic factors that may be amenable to government programmes and legislative reform. The existence of a national programme and presumed consent legislation stand out as very strongly associated with increased DD KTR.

Finally, while the highly significant association between increased DD KTR and presumed consent legislation suggests that restricting decisions on consent to donate to the pre-mortem donor may be helpful in raising KT rates, we believe that this conclusion is unwarranted as, with the exception of Singapore, virtually all presumed consent countries require familial consent for donation to proceed. Instead, we suggest the adoption of presumed consent legislation represents a predisposition on the part of a country's citizens to regard organ donation as an accepted and natural part of death and dying. Hence we believe that presumed consent should not be viewed as a binary phenomenon, singularly linked to consent,23 but as an indicator that represents the confluence of a series of factors and attitudes that permit the optimization of the processes used to identify potential donors very early in critical care. We believe these factors and cultural attitudes may help to normalize and integrate the concept of brain death with the application of technology at the end of life and the process of becoming an organ donor, into the rituals and social practices surrounding death and dying – rather than keeping these as separate and distinct processes, as is typically the case in many explicit consent countries.