Originally published in Volume 40 Issue 3 of Artificial Organs, 04 March 2016
Since the development of silicone peritoneal catheter and plastic bag systems, peritoneal dialysis (PD) has gained acceptance as a mode of long-term renal replacement therapy (RRT) over the past 40 years. Although hemodialysis (HD) is the most common modality of RRT, PD remains a critical component of most dialysis programs around the world. Notably, the HD to PD distribution ratio is highly variable between countries. For example, around 80% of end stage renal disease (ESRD) patients in Hong Kong are treated by PD.
The “dosage” of PD is often called the “adequacy of dialysis.” Basically, the adequacy of PD is a concept similar to that of HD, and is usually quantified by the measurement of small solute clearance, expressed as Kt/V or weekly creatinine clearance. It is important to note that the term Kt/V in PD is used as an extrapolation from HD, which is usually expressed as urea clearance per session in HD by convention, but per week in PD because of the continuous nature of the former. In PD, Kt/V denotes the proportion of total body water (i.e., the distribution space of urea) of which urea is cleared per unit time. In other words, if the weekly Kt/V equals one, the amount of urea cleared per week equals the amount of urea present in onefold of the total body water.
What dosage of PD would be adequate? The original PD technique proposed by Popovich et al. 1 in 1976 was continuous ambulatory peritoneal dialysis (CAPD), which was based on kinetic modeling of peritoneal urea transport. The assumptions at that time were: for an anuric 70-kg man with a total body water of 42 L, a steady-state serum urea concentration of 21.4 mmol/L required daily dialysate drainage of 14 L. This target could be achieved by doing 2 L PD cycles five times daily, with an additional 2 L ultrafiltration. The resultant weekly urea clearances would be 84 L, and the corresponding weekly Kt/V would, therefore, be 2.0. By similar assumptions, Teehan et al. 2 calculated that an anuric 70-kg patient who has a dietary protein intake 1.2 g/kg/day would require 13.5 L of dialysis drainage each day to maintain a steady-state serum urea concentration of 25 mmol/L. This corresponds to a weekly Kt/V of 2.25. It is important to note that both models predict that five 2 L exchanges, rather than four exchanges in the conventional regimen, are required for adequate removal of uremic toxins.
It soon became obvious that a theoretical target of weekly Kt/V of 2–2.25 is very much lower than the Kt/V of 1.0 per session for thrice weekly HD (i.e., total Kt/V of 3.0 per week). To explain the discrepancy, the peak urea concentration hypothesis was proposed 3. According to this hypothesis, the peak serum urea level, rather than the average level, is the major cause of uremic syndrome. Continuous therapy, such as CAPD, therefore, requires a much lower total Kt/V than that of HD to maintain an acceptable peak serum urea level and prevent uremic symptoms. In this hypothesis, a weekly Kt/V of 1.67 in CAPD is equivalent to a thrice-weekly HD Kt/V of 0.86. By extrapolating data from the National Cooperative Dialysis Study, Gotch proposed that a weekly Kt/V of 2.0 for CAPD was equivalent to Kt/V 1.0 per session in HD by single-pool urea kinetics and would result in acceptable clinical outcome 4, which is consistent with the hypothetical model proposed by Popovich et al. 1. It is interesting to note that none of the above models considered weekly creatinine clearance, and all recommendations on weekly creatinine clearance for PD adequacy are based on clinical experience rather than theory.
Is the above theoretical prediction supported by clinical studies? A number of early studies used the prospective cohort design to evaluate the association between peritoneal adequacy of dialysis and clinical outcomes 5–9. In general, these reports suggest a better clinical outcome, in terms of mortality and hospitalization, with a higher weekly Kt/V. The acceptable Kt/V in most of these studies conformed with the theoretical construct described above, and ranged from 1.9 to 2.4 6–9. However, the sample size was small in many of the studies, and the duration of follow-up was no more than 12 months in the others. In addition, these studies used univariate statistical analysis and did not control for confounding factors, such as diabetic status, cardiovascular disease, or residual renal function, when the relationship between dialysis adequacy and clinical outcome was analyzed.
In the mid-1990s, the results of three large-scale prospective cohort studies on PD adequacy were published 10–12. In the first one, 68 prevalent CAPD patients were followed for 3 years 10. Patient survival was associated with age, peripheral vascular disease, total Kt/V, and residual renal function. In this study, residual renal function accounted for approximately 20% of the total Kt/V and was a major confounding factor for the statistical analysis. In the second, 201 new CAPD patients were followed for 24 months 11. In this study, statistics were based on retrospective analysis of prospectively collected data, and patient survival was associated with age, diabetes, pre-existing cardiovascular disease, co-morbidity score, serum triglyceride level, and Kt/V. The relative risk of death associated with inadequate dialysis, defined as weekly Kt/V below 1.7, was 1.69. It is important to note that in this study, Kt/V at the beginning of the study was used for survival analysis. Since only new patients were enrolled, most of the patients had substantial residual renal function, and the potential confounding effect was not corrected for in the multivariate analysis. In fact, since the PD prescription was relatively fixed for the enrolled patients, the total Kt/V measured at the beginning of this study reflected the magnitude of residual renal function or the time (early vs. late) of dialysis initiation rather than the dosage of PD throughout the study period.
The most influential prospective study of PD adequacy during that period was the CANUSA study 12, which was a prospective cohort of 680 consecutive new CAPD patients from 14 dialysis centers in North America followed for 24 months. Unlike other cohort studies described above, total Kt/V and nutritional status were treated as time-dependent covariates for multivariate analysis in the CANUSA study. In essence, this approach considered the fluctuation of total Kt/V and nutritional status with time, and attributed any event (e.g., death or technique failure) to the level of Kt/V or nutritional status recorded at the 6-month evaluation preceding the event. The result showed that patient survival was associated with age, diabetes, cardiovascular disease, serum albumin, nutritional status (lean body mass or subjective global assessment score), and dialysis adequacy indices (total Kt/V or creatinine clearance). Technique survival was associated with serum albumin and creatinine clearance, but not Kt/V. Hospitalization was associated with serum albumin, subjective global assessment score, and creatinine clearance. The CANUSA study estimated that a decrease of 0.1 unit Kt/V was associated with a 5% increase in the risk of death, while a decrease of 5 L/week of creatinine clearance was associated with a 7% increase in the risk of death. More importantly, the modeled mortality rate as determined by the Cox proportional hazards model decreased linearly over the range of Kt/V from 1.5 to 2.3, further supporting the hypothesis that a weekly Kt/V of 2.0 and creatinine clearance of 60 L per week were reasonable adequacy targets. Based on these result, the National Kidney Foundation-Dialysis Outcomes Quality Initiative (NKF-DOQI) workgroup recommended a weekly total Kt/V greater than 2.0 for CAPD. The target weekly Kt/V for machine-assisted automated PD (APD) was set to be 2.1 in the NKF-DOQI recommendation, while the total weekly creatinine clearances was 60 L/week/1.73 m2 for both CAPD and machine-assisted PD.
In Hong Kong, CAPD is the first line RRT for all ESRD patients. Patients are only switched to long-term HD when they have ultrafiltration failure or peritoneal sclerosis. In addition, the usage of 2 L exchanges thrice daily, which was considered suboptimal by Western standards, was used as the routine CAPD regime in Hong Kong due to budgetary constraints. Although the practice of routine small volume CAPD may not be ideal, it provides an excellent opportunity to examine the effect of dialysis adequacy on the clinical outcome of a large unselected group of ESRD patients. Around the time of the CANUSA study, our group performed a cross-sectional study of 117 CAPD patients 13. In this study, patients with Kt/V < 1.7 were similar to those with Kt/V ≥ 1.7 in age, duration of CAPD, plasma creatinine, albumin, peritonitis rate, and incidence of hypertension. Patients with Kt/V ≥ 1.7 had higher hemoglobin, normalized protein nitrogen appearance, and more residual renal function. More importantly, the 5-year survival was 79% despite a lower Kt/V, and the employment and rehabilitation status was also satisfactory. This is the first study to suggest that small-volume dialysis may be acceptable in the Asian population with smaller body size and may be the preferable practice in view of the financial constraints.
Two subsequent prospective cohort studies from our group further support the importance of dialysis adequacy in Chinese PD patients. In the first study, we recruited 270 prevalent CAPD patients 14. After following them for 3 years, multivariate analysis showed that the duration of dialysis, diabetes, fat-free edema-free body mass, indices of dialysis adequacy (Kt/V or creatinine clearance), residual renal function, and requirement of a helper for dialysis exchanges were independent factors of patient survival. In this study, every 0.1 increase in weekly Kt/V was associated with a 4% decrease in the relative risk of death. In order to remove the confounding effect of residual renal function, our subsequent study followed 140 anuric CAPD patients for 2 years 15. Again, we found that dialysis adequacy had a significant impact on the clinical outcome of anuric CAPD patients. For two patients who differed only in weekly Kt/V in this study, a 0.1 higher value was associated with a 6% decrease in the relative risk of death. Although the findings of these studies seem to be in line with the previous ones, there is a critical difference: The range of weekly Kt/V tested was approximately 1.5–1.9, and excellent outcome could be achieved with a weekly Kt/V of 1.7.
This seemingly conservative approach of dialysis adequacy was further supported by subsequent randomized control studies. In the ADEMEX study, patients were randomly assigned to receive either a usual volume of PD fluids (2 L exchanges 4 times daily) or an increased dose to maintain total creatinine clearance >60 L/week/1.73 m2 16. The total Kt/V was 2.27 in the intervention group compared to 1.80 in the control group. The 2-year overall patient survival was found to be similar between the two groups (68% and 69%, respectively). In other words, increases in peritoneal small-solute clearances from 1.8 to 2.3 does not improve patient survival, even when the groups are stratified according to potential confounding factors (e.g., age, diabetic status, serum albumin, normalized protein nitrogen appearance, and anuria).
In another study, Lo et al. 17 randomized 320 new CAPD patients with a renal Kt/V less than 1.0 into three Kt/V targets: group A, 1.5–1.7; group B, 1.7–2.0; and group C, above 2.0. After follow-up of 2 years, the overall patient survival was 84.9%. Contrary to the expectation, there was no significant difference in patient survival, serum albumin level, composite nutritional index scores, or hospitalization rate between the three groups. However, patients had more clinical problems and severe anemia when weekly Kt/V was below 1.7. In another retrospective study performed by the same group, 150 anuric PD patients were enrolled 18. The 2- and 5-year survival rates were 88.7% and 66.7%, respectively. In this study, patients with baseline Kt/V below 1.67 had poorer survival than the others, while there seems to be a limit of beneficial effect for Kt/V above 1.80. Based on these data, the authors suggested a minimal Kt/V target of 1.7 and an optimal target of 1.8.
As mentioned above, the target weekly Kt/V for APD was set to be 2.1 in the NKF-DOQI recommendation, probably because of the intermittent nature of treatment. In this regard, the European APD Outcome Study (EAPOS) enrolled 177 patients with residual glomerular filtration rate <1 mL/min/1.73 m2 and were treated with APD 19. In this study, the APD prescription was adjusted at the physician’s discretion to aim for a creatinine clearance of 60 L/week/1.73 m2 and ultrafiltration of more than 750 mL/day during the first 6 months. After 2 years, patient survival was 78% and technique survival was 62%. Notably, the result showed that baseline ultrafiltration, but not creatinine clearance, was associated with patient survival. Similarly, the Netherlands Cooperative Study on the Adequacy of Dialysis (NECOSAD) study recruited 542 new PD patients 20; 166 developed anuria during the first 6 months. Among the anuric patients, small solute clearance was not associated with survival when analyzed as continuous variables. Conversely, when the results were analyzed dichotomously using predefined cut off points, weekly Kt/V <1.5 and creatinine clearance <40 L/week/1.73 m2 were associated with an increase in the relative risk of death. The authors concluded that the lower limits of adequate dialysis for PD is weekly Kt/V 1.5 and creatinine clearance 40 L/week/1.73 m2.
Based on the above studies, the International Society for Peritoneal Dialysis (ISPD) recommends that for small solute removal, the total weekly Kt/V should be at least 1.7 21. According to this guideline, the contributions of renal and peritoneal clearances may be added in the presence of residual renal function, although renal and peritoneal clearances may not be truly additive. However, it is important to appreciate that small solute removal above the recommended level should not be equated with “adequate dialysis.” The latest ISPD guideline does not specify a separate target for creatinine clearance in CAPD. In contrast, due to a more variable relationship between urea and creatinine clearance in APD, an additional target of 45 L/week/1.73 m2 for creatinine clearance is recommended. This guideline has been in place for almost 10 years and there were few new data in this aspect. Instead, recent focus on PD adequacy has moved from small solute clearance to fluid removal and preservation of residual function.
CC Szeto is currently a Professor of Department of Medicine & Therapeutics, The Chinese University of Hong Kong, and an Honorary Consultant Physician of Prince of Wales Hospital, Hong Kong Hospital Authority. His research interests include peritoneal dialysis related infections and vascular disease of renal failure.