1 Hyonam Kidney Laboratory, Soon Chun Hyang University, Seoul, Korea and 2 Divisions of Baxter Novum and Renal Medicine, Department of Clinical Science, Karolinska University Hospital Huddinge, Stockholm, Sweden
Correspondence and offprint requests to: Hi Bahl Lee, MD, PhD, Professor of Medicine/Nephrology, Hyonam Kidney Laboratory, Soon Chun Hyang University, 657 Hannam-dong, Yongsan-ku, Seoul 140-743, Korea. Email: hblee{at}hkl.ac.kr
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Methods. The influence of centre and patient characteristics on patient survival was investigated in 132 Korean and 106 Swedish incident PD patients, who underwent initial biochemical measurements and assessment of adequacy of dialysis, nutritional status, RRF and peritoneal transport characteristics.
Results. At the start of PD, Korean patients had a higher prevalence of diabetes, peritoneal Kt/Vurea, peritoneal creatinine clearance and peritoneal fluid removal, and lower body mass index, RRF and dialysate to plasma creatinine concentration ratio (D/P Cr) compared with Swedish patients. Significantly more patients from Korea were placed on temporary haemodialysis before PD (100 out of 132) when compared with Swedish patients (21 out of 106). During the follow-up, there was a significantly higher rate of transfer to other units in Korea and a significantly higher rate of kidney transplantation in Sweden. On KaplanMeier analysis, overall patient survival did not differ and relative risk for death was also not different between the two centres even after adjustment for age, diabetes, cardiovascular disease, RRF and D/P Cr. On Cox proportional hazards multivariate analysis, age, diabetes, RRF and D/P Cr were found to be independent predictors of mortality in the combined cohort of patients. While age, diabetes and D/P Cr were independent predictors of mortality in Korean patients, age and RRF independently predicted mortality in Swedish patients.
Conclusion. Although there were significant differences in centre and patient characteristics, we were unable to confirm a survival advantage for Korean over Swedish PD patients. The results of this study suggest that the reported difference in survival between Asian and Caucasian dialysis patients may have been due, in part, to differences in centre and patient characteristics rather than to race as such. The genetic influence on patient characteristics remains, however, to be elucidated.
Keywords: nutrition; patient survival; peritoneal dialysis; peritoneal transport rate; race; residual renal function; risk factor for mortality; technique survival
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
On the other hand, several reports in the literature suggest that racial difference may influence patient survival in dialysis populations [14]. A study from the USA [2] found that Caucasian-American dialysis patients had a higher mortality risk than Asian-American patients, even after adjusting for demographics, diabetes, co-morbidity and nutritional factors. Similarly, data from the Canadian Dialysis Registry [3] showed that the risk of death in Caucasian patients was significantly increased when compared with Asian and African-American patients after adjusting for co-morbidity. Finally, several reports from East Asian countries [57] consistently demonstrated a higher patient survival rate in East Asians on PD than in Caucasian patients in Western countries.
However, centre and patient characteristics may differ between study populations, and these differences may affect patient survival. It is important to evaluate systematically the difference in all predictors of death before the difference in survival is attributed to racial difference. We therefore compared the centre and patient characteristics between incident PD patients from one Korean and one Swedish centre and examined whether these differences influence patient survival in patients of different racial background.
![]() |
Patients and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Comparisons of centre and patient characteristics at the start of PD between the two centres are shown in Tables 1 and 2, respectively.
|
|
Nutritional assessment
The nutritional status of patients was assessed by subjective global assessment (SGA; Korean patients only), biochemical measurements, body mass index, calculation of normalized protein equivalent of total nitrogen appearance (nPNA) and urea kinetic studies.
SGA. We used a seven-point Likart-type scale of four items: weight loss, anorexia, subcutaneous fat and muscle mass. Each item was given scores to produce a global assessment. Scores of 12 represented severe malnutrition; 35, moderate to mild malnutrition; and 67, normal nutrition.
Body mass index. The body mass index was calculated as weight (in kg)/[height in m)]2.
Estimated protein intake. Dietary protein intake was estimated from the protein equivalent of total nitrogen appearance (PNA) using the equation PNA = 15.1 + 6.95 urea nitrogen appearance (UNA) (g/24 h) + protein loss (g/24 h). UNA and protein losses were determined from the measured urea and protein excretion in dialysate and urine. PNA was normalized to desirable body weight (nPNA) obtained from the Metropolitan height and weight table [8].
Adequacy of dialysis
Weekly total Kt/Vurea and weekly total creatinine clearance (CCr) were calculated from a 24 h collection of dialysate and urine. The distribution volume of urea (V), which is generally assumed to be equal to total body water, was calculated from the Watson equation.
Peritoneal, renal and total fluid removal in ml/day were also obtained as an adequacy index.
Residual renal function
RRF was estimated by calculating the mean of renal clearances of urea and creatinine from a 24 h urine collection.
Peritoneal equilibration test (PET)
The PET was performed with a standard 4 h dwell period using a 4.25% (Korean patients) or 2.27% (Swedish patients) glucose concentration for a 2 l volume exchange.
Biochemical analyses
A fasting venous blood sample was taken before the morning exchange. Blood chemistries were analysed by standard techniques. Concentrations of creatinine in dialysate and blood samples were measured by the Jaffe method and were corrected for glucose interference. Serum albumin was determined by the bromcresol purple method in the Swedish patients and by the bromcresol green method in the Korean patients. The serum albumin was corrected by the formula suggested by Joseph et al. [9] since the bromcresol purple method is reported to yield a lower value of serum albumin concentration when compared with the bromcresol green method [9].
Statistical analysis
Student's t-test or KruskalWallis test was used to compare the difference in clinical characteristics between different subgroups. 2 test or Fisher's exact test was used to compare the nominal variables between different subgroups. Actuarial survival rates were determined by the KaplanMeier method. A log-rank test was used to compare the patient and technique survival between subgroups. The Cox proportional hazards model was used to identify the factors predicting patient mortality. The Cox model for multivariate analysis was constructed by those factors significant at univariate analysis. Data are presented as mean±SD. A difference was considered significant when the P-value was <0.05.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Patient characteristics at the start of PD
Patient characteristics in Korean and Swedish PD centres are shown in Table 2. At the start of PD, Korean PD patients had a higher prevalence of diabetes, higher peritoneal Kt/Vurea, peritoneal CCr and peritoneal fluid removal, but lower body mass index, dialysate to plasma creatinine concentration ratio (D/P Cr), RRF, renal Kt/Vurea, total Kt/Vurea, renal CCr, total CCr and renal fluid removal compared with Swedish PD patients. There was no difference in age, gender, prevalence of CVD, corrected serum albumin, nPNA or total fluid removal.
Final status of study patients
The final status of Korean and Swedish patients is shown in Table 3. At the end of follow-up, 24 Korean (18.2%) and 22 Swedish (20.8%) PD patients had died and 39 Korean (29.6%) and 31 (29.2%) Swedish patients had transferred to HD. Two Korean (1.5%) and 25 Swedish (23.6%) patients underwent kidney transplantation. Thirteen Korean (9.8%) and no Swedish patients were transferred to other units.
|
Comparison of National Registry data from Korea and Sweden
In order to examine if the two centres represent the national dialysis practice in the respective countries, the National Registry data [10,11] were compared, as shown in Table 4. Significantly more patients received kidney transplantation in Sweden, whereas more patients were on dialysis (both HD and PD) in Korea. The incidence and prevalence of diabetes were significantly higher among Korean patients, while the prevalence of CVD and CVD death was higher in Sweden. There was no difference in prevalence and incidence of end-stage renal disease (ESRD) between the two countries.
|
|
|
|
Technique survival rate was not different between the two centres (P = 0.65). The 1-, 2- and 3-year technique survival rate were 86.0, 73.6 and 60.5% for the Korean, and 89.8, 65.9 and 51.9% for the Swedish patients.
Patient characteristics in surviving and the deceased patients
Patients who died during the follow-up period were older, had a higher prevalence of diabetes or CVD and lower corrected serum albumin concentration, RRF, renal CCr, total CCr and total fluid removal compared with those who survived (Table 6).
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
PD patient survival is influenced by many factors. Better preservation of RRF has been reported to be associated with better clinical outcome. In the present study, we found that RRF had an impact on mortality in the Swedish patients, but not in the Korean patients. This may be explained by the fact that, in Korean patients, the low RRF at the start of PD was due mostly to temporary HD for 68 weeks prior to PD in most (100 out of 132) of the patients but not due to delayed start of dialysis itself.
Since the ADEMEX study [13], the role of peritoneal solute clearance in patient survival has been questioned. In anuric patients from Hong Kong [6], however, higher peritoneal clearance of small solutes was associated with better patient survival. Although renal clearance and peritoneal clearance may not be equivalent, peritoneal solute clearance may, therefore, make a contribution to patient survival. Indeed, in the present study, despite low RRF, Korean patients achieved an average total Kt/Vurea of 2.17 and equal survival probability compared with Swedish patients.
On the other hand, inadequate fluid removal has been shown to be a potential risk factor for mortality in PD patients [14]. Indeed, in the present study, patients who died during the follow-up had lower total fluid removal than those who remained on PD. There was no difference in total fluid removal or patient survival rate between Korean and Swedish patients. The findings that Korean patients had lower renal but higher peritoneal fluid removal and achieved equal total fluid removal suggest that high peritoneal fluid removal may compensate for low renal fluid removal and improve patient survival. This observation is in agreement with data from Europe [15] showing that in anuric APD patients, low peritoneal fluid removal was associated with poor patient survival.
In the present study, a high peritoneal transport rate was an independent predictor of mortality in Korean patients but not in Swedish patients. Although the reasons for this discrepancy are not clear, it is possible that already high D/P Cr in Swedish patients at the start of PD may have reduced the impact of a high peritoneal transport rate on survival, or higher RRF in the Swedish patients at the start of PD may have compensated for the higher D/P Cr. Indeed, D/P Cr lost its predictive power for mortality in Swedish patients when the multivariate analysis included RRF. Although PET tests were carried out using a 4.25% solution in Korean patients and a 2.27% solution in Swedish patients, D/P Cr should not be significantly different between the two techniques [16,17]. Heimbürger et al. [18], however, observed significant differences in D/P Cr using different solutions.
The incidence and prevalence of diabetes in the Korean dialysis population have steadily increased over the past decade and, in this study, diabetes was more common in Korean than in Swedish patients. Although there were more diabetics in the Korean centre, there was no difference in overall prevalence of CVD between Korean and Swedish patients. In this study, a quantitative measure for the severity of CVD was not available to explain the higher CVD mortality among Swedish patients. In 2002, national average CVD mortality in PD patients was 38.9% in Korea and 48% in Sweden [10,11].
Interestingly, there was also no difference in technique survival between the Korean and Swedish patients in the present study. This is in contrast to a review which found that technique survival is substantially better in Asian than in Western countries [4].
There are several obvious limitations in this study. Since only one centre each with a limited number of patients and a short duration of follow-up from the two countries was included, it may be difficult to extrapolate the results to all Korean and Swedish PD patients or to Asian and Caucasian patients. Although the two centres may not represent the practice patterns in the respective countries in all aspects, they do not deviate substantially from the respective National Registry data. Registry data from both countries [10,11] showed that the prevalence of diabetes at the start of dialysis was 40.7 and 35%, and the cardiac cause of death in PD patients was 38.9% (2003) and 48% (19982001) for Korean and Swedish patients, respectively.
There were several important differences in the PD practice between the two centres. While most of the Korean patients were treated with HD for 68 weeks before PD was started to allow time for wound healing after subcutaneous implantation of the PD catheter, most of the Swedish patients were on conservative therapy before PD. As a result, the RRF was significantly lower in Korean than in Swedish patients at the start of PD. The Korean patients, however, received a significantly higher dose of PD to achieve adequate total solute removal and equal total fluid removal to Swedish patients.
Another major difference was that the transplantation rate was much higher in Swedish patients than in Korean patients. In general, a healthier and younger patient is more likely to receive a renal transplant, and, by censoring transplanted patients, the remaining patients should therefore have an increased proportion of less healthy patients remaining at risk. This would negatively impact the overall survival in Swedish patients. On the other hand, those who were transplanted would have stayed on PD for a shorter duration and their survival would appear better when censored at transplant. Total duration of PD was not different between Korean and Swedish patients.
There were several important differences in patient characteristics between the two centres. There were significantly more diabetics and a lower prevalence of high D/P Cr among Korean than Swedish patients.
In summary, there were significant differences in centre and patient characteristics between the two centres and yet no significant difference was observed in overall PD patient survival or relative risk of death after adjusting for independent predictors of death. Higher prevalence of diabetes, lower RRF, lower renal and total Kt/Vurea and creatinine clearance and lower renal fluid removal in Korean patients than in Swedish patients may be disadvantageous for Korean patients. On the other hand, possibly fewer or less severe CVD, relatively lower D/P creatinine, higher peritoneal Kt/Vurea and creatinine clearance, and higher peritoneal fluid removal may have been advantageous to Korean patients. The prevalence of transplantation may have affected PD patient survival in both ways. There were no significant differences in age, gender, prevalence of CVD, corrected serum albumin or nPNA between the two centres at the start of PD. The advantages may have compensated for disadvantages in Korean patients to achieve equal survival to Swedish patients. In future studies comparing patient survival between centres in different countries and/or races, all those centre and patient factors known to influence survival should be systematically studied and compared.
We conclude that the reported difference in patient survival rate between Asians and Caucasians may have been due, in part, to a difference in centre and patient characteristics rather than to race itself. The genetic influence on patient characteristics remains, however, to be elucidated.
Conflict of interest statement. None declared.
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|