Comparison of Peritoneal Dialysis With Other Treatments for Acute Kidney Injury

Solute and Volume Control and Overall Outcomes

Many have favored CRRT/IHD/SLED over PD because of presumed superiority in solute, volume, and metabolic control. However, a number of studies have documented excellent solute and volume control with PD. Chitalia et al. compared continuous equilibrated peritoneal dialysis (CEPD) with tidal peritoneal dialysis (TPD) in patients with hypercatabolic AKI. In this study, patients were classified as having mild, moderate, or severe hypercatabolic AKI according to severity of catabolism, which was based on estimation of excess urea nitrogen appearance (UNA). Patients with mildly to moderately hypercatabolic AKI were subsequently randomly assigned to undergo TPD or CEPD with a total volume of approximately 26 L per session in both arms. TPD was associated with lower postdialysis BUN and creatinine concentrations and higher creatinine and urea clearances than CEPD. The per-session and weekly Kt/V urea achieved with TPD were 0.34 ± 0.14 and 2.43 ± 0.87, respectively—significantly higher than those with CEPD. Ultrafiltration was also superior with TPD as compared with CEPD (3 L vs. 2 L per day). Others have targeted even higher clearances using high-volume peritoneal dialysis. Investigators from the same institution in Brazil targeted a per session Kt/V and 0.60 and 0.65 in separate studies, and achieved a Kt/V of 0.51 and 0.55, respectively. In both studies, rapid correction in metabolic derangement while maintaining an ultrafiltration rate of roughly 2L/day occurred.

Clearances with continuous-flow peritoneal dialysis (CFPD) may be superior to those in TPD and CFPD and may be particularly useful in AKI. In this modality, large volumes of intraperitoneal fluid are replenished continuously with dialysis solution in constant contact with the membrane. There are two modes of CFPD, continuous infusion and removal of sterile dialysate solution (called single-pass CFPD) and recirculation of the dialysate (in which the peritoneal effluent, and not blood, is “dialyzed” through a hemodialysis or CRRT circuit against an external dialysate and then returned to the patient). Obviously, this technique requires either two catheters or a catheter with dual lumens. Physical separation of catheter lumens is important to minimize recirculation. Urea clearances of 30 to 50 mL/min usually are achieved with this technique, with the potential for even greater clearances. Such clearances are equal or superior to those achieved with CRRT. Unfortunately, no studies have directly compared CFPD with other modalities in patients with AKI. The overall cost of CFPD also may be limiting and may preclude its routine use in AKI, particularly at centers with limited resources. Further studies of the use of CFPD in AKI are required for this promising PD modality.

Volume overload is encountered commonly in critically ill patients with AKI. Hypotension often limits the rate at which fluid can be removed, making IHD less desirable for many nephrologists. Although most clinicians would prefer CRRT in these cases, PD offers an alternative. Exchanges using 4.25% dextrose can yield approximately 1 L per 4-hour exchange and even more with hourly exchanges, so rates of sodium and water removal can be high with PD. Hypernatremia is a risk with rapid PD ultrafiltration, because of sodium sieving, but not with IHD or CRRT. However, low-sodium PD solutions can be prepared to yield very high rates of ultrafiltration with balanced Na and water removal.

Several studies have compared outcomes between patients with AKI managed with PD and IHD or CRRT. One of the first randomized trials comparing PD to continuous venovenous hemofiltration (CVVH) in the treatment of AKI came from Vietnam, and it demonstrated an increased risk of mortality in the group receiving PD (17/36 [47%] vs 5/34 [15%] in the CVVH group). However, the applicability of this study has been questioned for several reasons. The PD techniques used in this study were suboptimal, specifically rigid PD catheters, and open drainage was employed. In addition, the population studied is unique and may not be generalizable because it consisted mainly of patients with AKI in the setting of Plasmodium falciparum malaria infection (68%), and the CVVH treated group had a much lower mortality rate than would be expected in this type of population with AKI (only 15%). In contrast to these early findings, subsequent randomized trials did not demonstrate differences in mortality between PD and IHD or CRRT. In 2008 a randomized controlled trial from Brazil comparing high-volume PD (36–44 L/day in 18–22 exchanges/day) to daily HD in 120 patients with AKI from ATN demonstrated similar mortality and renal recovery rates (high volume PD [HV-PD] 58% mortality vs. daily HD [D-HD] 53% mortality p = .48, HV-PD 83% renal recovery vs. D-HD 77% renal recovery p = .84). Furthermore, in a small study from India randomizing 50 patients with AKI to PD or CVVHDF, mortality was high in both groups, but not significantly different statistically (PD 18/25 vs. CVVHDF 21/25, respectively [ p = .49]). Most recently, Chionh et al. analyzed these randomized trials along with several other observational cohort studies in a systemic review of studies focusing on PD in AKI. This review found no significant differences in mortality between AKI patients treated with PD and extracorporeal blood purification therapies (i.e., IHD or CRRT). In the seven cohort studies included in this systemic review (270 patients treated with PD and 444 patients treated with HD/CRRT) the odds ratio for mortality was 0.96 (95% CI 0.53–1.71). Similarly, in the combined analysis of the four randomized trials (three of which were discussed above) comparing PD (125 patients combined) to either HD or CRRT (123 patients combined), no significant difference in mortality was identified (OR was 1.50, 9% CI 0.46–4.86). There was significant heterogeneity between the results of the randomized trials (I ² = 73%, p = .03). Although these data suggest clinically significant outcomes in patients with AKI treated with PD are likely similar to those outcomes in patients with AKI treated with HD or CRRT, many of these studies have limitations, including small sample size and methodologic concerns. Consequently, additional larger high-quality randomized controlled trials on PD in AKI would be very useful and informative.

The optimal dose of dialysis for PD in AKI has not been studied extensively. Ponce et al. compared high-intensity PD (target Kt/V of 0.8 per session) versus low-intensity (target Kt/V of 0.5 per session). The investigators found similar metabolic control and mortality at 30 days in both arms. However, the achieved per session Kt/V in the high intensity arm was 0.59, whereas in the low-intensity arm the achieved Kt/V was much closer to prescribed (0.49). Whether the outcomes would have improved if clearances were closer to those prescribed, perhaps using CFPD, requires further study. The achieved weekly Kt/V was similar to that achieved by Gabriel et al. comparing acute PD and intermittent HD. This has prompted some to recommend a minimum weekly Kt/V of 3.5, whereas others feel a weekly Kt/V of 2.1 will be adequate for most patients. Clearly, the metabolic demand of the individual pain will guide therapy until more studies are available to answer these questions.

Cost

Chitalia et al. found TPD and CEPD to be relatively inexpensive. They reported that (in 2002) a 12-hour treatment with TPD cost approximately $160 and 48 hours of CEPD cost $175; these figures would yield weekly rates of approximately $1100 for TPD and $600 for CEPD. In comparison, the weekly cost of SLED is about $1200 to $1400, that of CRRT with heparin anticoagulation about $2000 to $2500, and that of CRRT with citrate anticoagulation about $2500 to $3500. The significantly higher costs of CRRT are secondary to greater nursing time, higher equipment costs, and larger need for anticoagulation.