INTRODUCTION
End-stage kidney disease (ESKD) impacts the lives of thousands of individuals in South Africa and globally. Due to the high cost of renal replacement therapy (RRT), only a fraction of the patients is afforded the benefit of RRT in developing countries. The 2012 South African Renal Registry annual report showed that in the public sector, 8559 ESKD patients (3182 in the public sector and 5377 in the private sector) were on RRT for a population of 43.6 million.(1) Of these 1607 (18.8%) patients had had kidney transplants.(1) The most recent report in 2016 shows that 10,257 patients were on RRT in South Africa for a population of 55.9 million, and 3175 of these were in the public sector, with only 1425 (13.9%) having kidney transplants.(2) The majority of transplants (893/1425) were in the public sector.(2) With an increasing ESKD population and declining transplant numbers, it is vital to conserve and utilise the available resources cost-effectively in order to address the increasing demand for RRT.
Renal transplantation is the most effective therapy for ESKD.(3) This is achieved by a multidisciplinary team of nephrologists, transplant surgeons, pathologists, nurse coordinators as well as consultation services from urology, vascular surgery, cardiology, infectious diseases and psychiatry. Immunosuppressive therapy is essential, even in patients with ideal tissue matches, in order to prevent allograft rejection.(4) Successful transplantation is associated with decreased mortality and an improved quality of life. It frees a patient from chronic dialysis, and fluid and dietary restrictions. Anaemia and infertility are also corrected. The donor can be a living related, unrelated living or a cadaveric donor.(3,5)
In the late 1950s, 6-mercaptopurine and azathioprine (AZA) were developed.(6) AZA is one of the oldest agents still in use to date. Prior to the availability of AZA and corticosteroids, renal transplantation was a challenge except in identical twins. This regimen changed in the 1980s after trials showed improved allograft survival with cyclosporine therapy (CYA).(7) The subsequent introduction of mycophenolate mofetil (MMF), tacrolimus and sirolimus has permitted the use of combination immunosuppressive therapy for successful cadaveric transplantation.(8) These breakthrough discoveries, as well as the advancement in organ preservation and tissue typing techniques, have paved the way for modern transplantation. However, the comparative performance of MMF and AZA has not been well studied in South Africa. This study, therefore, set out to compare graft and patient survival in patients treated with MMF vs AZA after cadaveric renal transplantation.
METHODS
A retrospective cohort study was conducted. The data source was clinical records of renal transplant recipients at the Renal Transplant Unit at Charlotte Maxeke Johannesburg Academic Hospital from 1985 to 2013, who were treated with one of the two immunosuppressive agents required for analysis in the study and followed up for a minimum of 5 years. Data collected included patient characteristics, immunosuppressive treatment as well as graft and patient survival. The study was approved by the University of the Witwatersrand Human Research Ethics Committee.
Study setting and population
In total, 208 renal transplant patients were studied. Included were patients with a first cadaveric renal transplant and on different regimens, including: (CYA, AZA, prednisone); (CYA, MMF, prednisone); (CYA, AZA, prednisone up to time of change in regimen) and induction therapy with interleukin-2 receptor antagonist (basiliximab/dacluzimab). Excluded from the study were patients on tacrolimus, patients with graft failure <3 months after transplantation, living donor transplants and repeat renal transplants.
Data collection, management and analysis
Data was obtained by consecutively reviewing clinical records of renal transplant patients. Acute allograft rejection was diagnosed by a rise of 25% in the baseline serum creatinine and histologically via renal biopsy. Treatment outcomes were changed in serum creatinine levels, number of acute rejections, graft survival and patient survival. Predictor variables studied included socio-demographic characteristics (age at transplant, gender and race); risk factors for ESKD (hypertension, diabetes, glomerulonephritis, others); serum creatinine levels measured at intervals (at 3, 6 months then yearly) for a period of 60 months post transplantation; number of acute infections (classified broadly as the presence or the absence of infections due to limitation of the data availability) and adverse effects (e.g. nausea, vomiting, anaemia, leucopenia, diarrhoea).
Data was entered into Microsoft Office Excel Spreadsheet and transferred to STATA 14 (StataCorp, College Station, Texas) for statistical analysis. For continuous data, the Student t-test was used to compare normally distributed variables, and the Mann–Whitney test was used to compare non-normally distributed data. For variables expressed as frequencies and percentages, the chi-square and Fischer's exact tests were used. Kaplan–Meier curves were used to assess survival for each group. Cox regression model was used to determine the risk factors associated with the acute rejection in each group of patients on different drugs. The goodness of fit of the model for the data was assessed using the Cox Snell residual and Nelson–Aalen cumulative Hazard test.
RESULTS
Study participant characteristics
A total of 208 patient records, which met the inclusion criteria, were reviewed for the study. Of these, 101 patients were treated with AZA and 107 patients treated with MMF. Mean age was 38.1 ± 11.0 and 39.7 ± 10.7 years in the AZA and MMF groups, respectively. There were more males than females in both study groups, 64 (63.4%) in the AZA and 81 (75.5%) in the MMF group. Hypertension was the most common risk factor for ESKD in both groups, with 54.5% in the AZA and 36.5% in the MMF groups, respectively. Infections developed in 42.6% of the patients on AZA compared to 69.2% on MMF (p = 0.01). A total of 16 patients developed acute allograft rejection during the 5-year follow-up period, with 6 (6%) occurring in the AZA group and 10 (10%) in the MMF group (p = 0.3). The median time to rejection was significantly longer in the MMF group at 52 weeks compared to 12 weeks in the AZA group (p = 0.03). The mortality rate was 1% in the MMF and 16.8% in the AZA group, respectively (p < 0.001). The other characteristics of the study participants are shown in Table 1.
Characteristics of the study population.
| Variable | Azathioprine, n (%) | Mycophenolate mofetil, n (%) | p-Valued |
|---|---|---|---|
| Age (years) mean (SD)a | 38.1 (11.0) | 39.7 (10.7) | 0.29 |
| Gender | |||
| Male | 64 (63.4%) | 81 (75.7%) | 0.05 |
| Female | 37 (36.6%) | 26 (24.3%) | 0.05 |
| Ethnicity | |||
| Black | 31 (30.7%) | 83 (77.5%) | <0.01 |
| White | 64 (63.4%) | 19 (17.8%) | <0.01 |
| Indian | 6 (5.9%) | 5 (4.7%) | 0.70 |
| Hypertension | 46 (45.5%) | 68 (63.5%) | 0.01 |
| Diabetes | 2 (2.0%) | 1 (0.9%) | 0.61 |
| Both diabetes and hypertension | 0 (0.0%) | 6 (5.6%) | 0.03 |
| Glomerulonephritis | 22 (21.8%) | 3 (2.8%) | <0.01 |
| Unknown cause of CKD | 11 (10.9%) | 23 (21.5%) | 0.04 |
| Other risk factorb | 19 (18.8%) | 6 (5.6%) | 0.05 |
| Graft failure/rejection | 6 (6.0%) | 10 (10.0%) | 0.30 |
| Time to rejection (in weeks), median (IQR)c | 12 (12–16) | 52 (44–52) | 0.03 |
| Infections | 43 (42.6%) | 74 (69.2%) | 0.01 |
| Mortality | 17 (16.8%) | 1 (1.0%) | <0.01 |
SD = standard deviation.
Other risk factors include pyelonephritis, cortical necrosis, analgesic nephropathy, posterior urethral valves, post-obstetric obstruction, obstructed ureter, medullary cystic disease, reflux nephropathy, poisoning.
IQR = inter-quartile range.
p–Value determined by the t-test for age, Mann–Whitney test for time to rejection and Fishers exact (where frequencies were less than or equal to 5) or chi-square test for all other variables.
Graft survival of patients on azathioprine and mycophenolate mofetil
Graft survival, defined as the maintenance of stable creatinine levels, was similar between the two groups over the long period of follow-up (60 months) (Figure 1). Over the course of the follow-up period, Kaplan–Meier curves of graft survival for transplant recipients receiving either AZA or MMF did not diverge significantly from each other, suggesting that the graft survival was not different between the two groups (p = 0.37).
Renal function
Over a period of 5 years of observation after renal transplantation, there was a steady improvement in renal function in both groups. However, the MMF cohort generally had lower serum creatinine levels at all-time intervals (p < 0.05) as shown in Table 2.
Distribution of serum creatinine levels for azathioprine and mycophenolate mofetil in the study population over a 5-year period of observation.
| Variable | Azathioprine (μmol/l) | Mycophenolate mofetil (μmol/l) | Total (μmol/l) | p-Value* |
|---|---|---|---|---|
| Serum creatinine at 3 months median (IQR)a | 142 (120–198) | 124 (99–153) | 134 (108–166) | <0.01 |
| Serum creatinine at 6 months median (IQR) | 132 (115–174) | 120 (91–144) | 126 (104–159) | <0.01 |
| Serum creatinine at 12 months median (IQR) | 131 (113–172) | 113 (92–146) | 122 (98–154) | 0.01 |
| Serum creatinine at 24 months median (IQR) | 133 (109–165) | 113 (96–141) | 124 (102–148) | 0.04 |
| Serum creatinine at 36 months median (IQR) | 129 (110.5–154) | 121 (93.5–143.5) | 126 (99.5–148.5) | 0.01 |
| Serum creatinine at 48 months median (IQR) | 128 (111–155) | 121 (94–142) | 125.5 (105–149) | 0.05 |
| Serum creatinine at 60 months median (IQR) | 129 (111–151) | 115 (93.5–153.5) | 126 (106–152) | 0.15 |
IQR = interquartile range.
p-Value determined by the Mann–Whitney test.
Adverse effects of azathioprine and mycophenolate mofetil
With regards to adverse effects, significantly more patients (16.7%) on MMF experienced nausea compared to 5% in the AZA group. Similarly, more patients (11.8%) treated with MMF had vomiting compared to 4% in the AZA group. A highly significant number (41%) of the patients on AZA had anaemia as compared to none treated with MMF. Neutropaenia developed in 49% of patients treated with MMF compared to 0% on AZA (p < 0.01). Abdominal pain and diarrhoea were reported in 39% and 29.7%, respectively, in patients treated with MMF. In general, AZA-treated patients had more haematological side effects (except for neutropaenia), while MMF-treated patients had more gastrointestinal side effects (Table 3).
Adverse effects of azathioprine and mycophenolate mofetil in the study population.
| Variable | Azathioprine (N = 101), n (%) | Mycophenolate mofetil (N = 107), n (%) | Total (N = 208), n (%) | p–Value* |
|---|---|---|---|---|
| Nausea | 5 (5.0) | 17 (16.7) | 22 (10.8) | 0.01 |
| Vomiting | 4 (4.0) | 12 (11.8) | 16 (7.9) | 0.04 |
| Anaemia | 41 (40.6) | 0 (0.0) | 41 (19.7) | <0.01 |
| Neutropaenia | 0 (0.0) | 49 (49.0) | 49 (24.4) | <0.01 |
| Leucopaenia | 10 (9.9) | 0 (0.0) | 10 (4.8) | <0.01 |
| Thrombocytopaenia | 7 (6.9) | 0 (0.0) | 7 (3.4) | 0.01 |
| Hepatotoxicity | 8 (7.9) | 0 (0.0) | 8 (3.9) | <0.01 |
| Diarrhoea | 0 (0.0) | 30 (29.7) | 30 (14.9) | <0.01 |
| Abdominal pain | 0 (0.0) | 39 (39.0) | 39 (19.4) | <0.01 |
p-Value determined using the Fishers exact (where frequencies were less than or equal to 5) or chi-square tests as appropriate.
Predictors of allograft rejection for patients on mycophenolate mofetil
Unadjusted and adjusted Cox hazard ratios (HR) of allograft rejection for MMF were computed for selected variables, and the final model was tested for goodness of fit. Serum creatinine at 3 months was found to be a strong predictor of graft function in patients on MMF (p < 0.001) compared with other time points over 60 months. After adjusting for confounders, serum creatinine at 3 months remained a strong predictor of outcome. Patients experiencing abdominal pain while on treatment with MMF were 62% less likely to develop graft failure than patients on treatment with AZA [unadjusted HR = 0.38 (0.14–1.05)]. This finding, however, is not statistically significant (p = 0.06). After adjusting for the other confounders, the association with abdominal pain was marginally significant [adjusted HR 0.18 (0.01–1.02); p = 0.05]. The other findings are detailed in Table 4.
Unadjusted and adjusted hazard ratios (HR) of allograft loss across selected variables for mycophenolate mofetil.
| Variable | Unadjusted HR (95% CI) | p-Value | Adjusted HR (95% CI) | p-Value* |
|---|---|---|---|---|
| Drug | ||||
| Azathioprine | 1.00 (Reference) | – | 1.00 (Reference) | – |
| Mycophenolate mofetil | 1.58 (0.57–4.33) | 0.38 | 1.16 (0.15–8.77) | 0.88 |
| Gender | ||||
| Male | 1.00 (Reference) | – | – | – |
| Female | 0.32 (0.07–1.40) | 0.13 | – | – |
| Ethnicity | ||||
| Black | 1.00 (Reference) | – | – | – |
| Other races | 0.55 (0.19–1.58) | 0.27 | – | – |
| Risk factor | ||||
| All other risk factors | 1.00 (Reference) | – | – | – |
| Hypertension | 1.57 (0.55–4.52) | 0.40 | – | – |
| Serum creatinine at 3 months | 1.00 (Reference) | 1.00 (Reference) | ||
| Serum creatinine at 12 months | 0.98 (0.96–1.00) | 0.08 | 0.97 (0.92–1.02) | 0.26 |
| Serum creatinine at 24 months | 0.99 (0.97–1.01) | 0.16 | 1.03 (0.97–1.03) | 0.35 |
| Serum creatinine at 36 months | 0.98 (0.96–1.00) | 0.081 | 0.97 (0.91–1.03) | 0.34 |
| Nausea | ||||
| Azathioprine | 1.00 (Reference) | – | – | – |
| Mycophenolate mofetil | 0.35 (0.11–1.07) | 0.07 | – | – |
| Vomiting | ||||
| Azathioprine | 1.00 (Reference) | – | – | – |
| Mycophenolate mofetil | 0.35 (0.10–1.24) | 0.10 | – | – |
| Anaemia | ||||
| Azathioprine | 1.00 (Reference) | – | – | – |
| Mycophenolate mofetil | 1.73 (0.39–7.59) | 0.47 | – | – |
| Leucopaenia | ||||
| Azathioprine | 1.00 (Reference) | – | – | – |
| Mycophenolate mofetil | 0.73 (0.10–5.54) | 0.76 | – | – |
| Abdominal pain | ||||
| Azathioprine | 1.00 (Reference) | – | 1.00 (Reference) | – |
| Mycophenolate mofetil | 0.38 (0.14–1.05) | 0.06 | 0.18 (0.01–1.02) | 0.05 |
Cox regression model.
DISCUSSION
In this study, the mean age at time of renal transplantation was similar in both groups with more male patients being transplanted than females. This gender disparity is consistent with studies performed in the United States that have shown that more male than female patients have access to haemodialysis and kidney transplantation.(9) Hypertension was the most common cause of ESKD in the study population, similar to a study conducted in the United States.(10) This study illustrated a steady improvement in serum creatinine after renal transplantation over the study period of 60 months. The serum creatinine at 3 months was found to be a strong predictor of allograft function, the primary outcome of this analysis, compared with the other time points over 60 months. Previous studies have reported serum creatinine levels at 6 and 12 months as significant predictors of long-term allograft survival.(11,12) Both drugs were reasonably tolerated, with nausea and vomiting as the common adverse effects. The prevalence of gastrointestinal side effects was higher in the MMF group compared to the AZA group. This finding is consistent with previous studies which have also shown the same side effect.(13,14) The presence of abdominal pain was found to be protective of allograft function. This association was, however, marginally significant (p = 0.06) when adjusted for confounding factors. With regard to haematological side effects, anaemia was commoner in the AZA group compared to the MMF group. This finding is also consistent with previous studies.(15)
There is an association between immunosuppressive drugs and infection. However, the specific contribution of different drugs to infections has been poorly analysed.(16) In this study, we collected data on the type of infections. However, the data was not sufficient for statistical analysis. With this limitation, the only possibility was to dichotomise the data to the presence or the absence of infections as shown in Table 1, where MMF-treated patients had more infections.
The Kaplan–Meier survival curves showed similar graft survival for both drugs (p < 0.37). While serum creatinine levels were lower in the MMF group, this study did not show that MMF was better than AZA in preventing allograft loss over the study period. Findings of this study are consistent with previous studies that have shown that treatment benefits and risks are similar in patients treated with AZA and MMF and that MMF therapy is not superior in preventing long-term graft loss when compared to AZA.(17,18) The effectiveness of both drugs was demonstrated in this study. Patient survival on the other hand was superior in the MMF group. However, patient characteristics may have contributed to the improved survival with MMF. There were significant differences in ethnicity between the two groups. The AZA patients were transplanted mainly in the 1990s, where white patients, a group with lower immunological risk, formed the majority of transplant recipients in South Africa. This changed in the post-apartheid period when MMF therapy was commenced and black patients constituted the majority. The latter group is known to have higher immunological risk.(19) Also, the AZA group received higher doses of steroids, which may have contributed to more complications and poorer outcomes.
Limitations: Due to the retrospective nature of the study, there were limitations related to missing data, such as details of infections. This study was also limited by the fact that we used a convenience sample since the sampling frame is not known. This often suffers from bias since the sample is not chosen at random. Despite this limitation, this is the first study in South Africa to compare the cadaveric renal transplant recipients treated with either AZA or MMF.
In conclusion, this study has shown that patient survival was superior in the long term in cadaveric renal transplant patients receiving MMF therapy. While both drugs are effective in the maintenance of allograft function in renal transplant patients in the long term, serum creatinine levels were lower in the MMF group.