 |
||||
| 2007 JASN IMPACT FACTOR 7.111 | HOME AUTHOR INFO EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP | |||
| CURRENT ISSUE | ARCHIVES | JASN Express | ONLINE SUBMISSION | |
|
||||
| 2007 JASN IMPACT FACTOR 7.111 | HOME AUTHOR INFO EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP | |||
| CURRENT ISSUE | ARCHIVES | JASN Express | ONLINE SUBMISSION | |
Peritoneal Dialysis |
Research Service, VA Salt Lake City Health Care System, and Departments of Internal Medicine and Bioengineering, University of Utah, Salt Lake City, Utah.
Correspondence to: Dr. John K. Leypoldt, Dialysis Program, University of Utah, 50 N. Medical Drive East, Salt Lake City, UT 84112-5350. Phone: 801-585-3812; Fax: 801-581-4750; E-mail: Ken.Leypoldt{at}hsc.utah.edu
Abstract
ABSTRACT. The current understanding of the transport pathways that govern solute removal during peritoneal dialysis is reviewed. Diffusive transport rates across the peritoneal membrane for small solutes are slow. Even though the rate of diffusive solute transport decreases with increasing molecular size, large molecules (e.g., albumin) are nevertheless removed from the patient during routine peritoneal dialysis. Recent work has confirmed a previous suggestion that diffusive solute transport is limited by the small area of the peritoneal membrane that participates in the transport process. This small functional area is due to either poor contact of the peritoneal membrane with dialysis solution bathing the peritoneal cavity or to the limited surface area of capillaries that perfuse peritoneal tissues. Convective solute transport during peritoneal dialysis is proportional to the transperitoneal ultrafiltration rate but is less than that expected, because of low solute sieving by the peritoneal membrane and fluid absorption from the peritoneal cavity. Low solute sieving across the peritoneal membrane was first identified in 1966, a phenomenon that is now attributed to the presence of water-only transport pathways mediated by aquaporin-1. Fluid absorption from the peritoneal cavity occurs at the same time as transperitoneal ultrafiltration, but the pathways by which these two processes occur simultaneously remain speculative. This review proposes a novel hypothesis, whereby fluid absorption occurs in areas of the peritoneal membrane that are governed by different physical forces than those governing transperitoneal ultrafiltration. Further understanding of the pathways for fluid and solute transport during peritoneal dialysis will permit improvements in the adequacy of the dialysis dose and the more efficacious use of peritoneal dialysis to treat patients with end-stage renal disease.
The kinetics of solute and fluid removal from the peritoneal cavity during peritoneal dialysis have been studied extensively, yet the precise transport pathways remain incompletely understood. A comprehensive understanding of these processes is of practical significance to the nephrologist in guiding the peritoneal dialysis prescription to achieve an appropriate dose of therapy. Kinetic studies are also of theoretical significance to the understanding of physiology, because they provide insights into the structure of peritoneal tissues. The purpose of this review is to examine the current understanding of peritoneal solute transport, to provide directions for future research.
Figure 1 summarizes the current understanding of the routes for peritoneal solute and fluid transport to and from the peritoneal cavity. Removal of uremic toxins from the patient occurs across the peritoneum or peritoneal membrane by two major mechanisms, diffusion and convection. The rate of solute transport by diffusion is proportional to the solute concentration difference between blood (or, more precisely, plasma water [1]) and the dialysis fluid within the peritoneal cavity. The proportionality factor, which accounts for the extent of contact area between the peritoneum and the dialysis fluid, defines the permeability of the peritoneal membrane. The overall rate of peritoneal solute transport by convection is the difference between two separate pathways. First, convective solute transport can occur from the patient into the dialysis solution; the rate of this transport pathway is proportional to the transperitoneal ultrafiltration rate. The convective proportionality factor is defined as the sieving coefficient of the peritoneal membrane; a sieving coefficient of one indicates no rejection of solute by the peritoneal membrane, whereas a lower value indicates hindrance to solute relative to fluid movement. (In this article, the term "sieving" is used to indicate the ease with which solutes are convectively transported across the peritoneum.) Fluid and solute removal from the patient during peritoneal dialysis are reduced by the absorption of fluid and the accompanying convective solute transport out of the peritoneal cavity that occurs by an incompletely understood mechanism. Fluid transport by this pathway is directed out of the peritoneal cavity and includes both direct absorption into peritoneal tissues and transport into lymphatic vessels. This review will focus on the current understanding of these pathways for peritoneal solute transport; moreover, a novel explanation for fluid and convective solute transport out of the peritoneal cavity will also be proposed.
|
Although several investigators had already examined the kinetics of fluid absorption from the peritoneal cavity during the late 19th century (see below), solute transport from blood to fluids added to the peritoneal cavity was not investigated extensively until 1923. In that year, Putnam (2) first characterized the peritoneum as a simple passive membrane on the basis of observations in nonuremic animals that the rate of solute diffusion between blood and the peritoneal cavity varied inversely with molecular size. Interestingly, this study was reported in the same year that Ganter (3) reported the first attempt at peritoneal dialysis in a human patient.
Another seminal contribution to the understanding of diffusive solute transport across the peritoneal membrane was reported by Boen (4) on the basis of his extensive clinical experience using intermittent peritoneal dialysis. In that report, Boen emphasized the need to calculate clearance to quantify peritoneal transport and reported diffusion and absorption curves for various uremic toxins and electrolytes. That work was significant, because it set the stage for the development of the peritoneal equilibration test by Twardowski et al. (5), which established the first approach for quantification of individual differences in peritoneal transport characteristics. Henderson and Nolph (6) subsequently defined the term "dialysance" for the peritoneal membrane and demonstrated the relationship between clearance and dialysance in describing peritoneal solute transport. Dialysance is equivalent to the mass transfer–area coefficient (MTAC) or permeability-area product (PA) and has been defined by various other symbols in the literature; it is relatively independent of dwell time and is a fundamental measure of the rate of diffusive solute transport across the peritoneal membrane. (Recent work has shown that PA is actually time-dependent [(7]; however, this dependence of PA on time is poorly understood and will be ignored in this discussion.) Babb et al. (8) subsequently compared this diffusive transport parameter for the peritoneal membrane with that for a cellulosic hemodialysis membrane using a number of solutes with different molecular weights. These investigators noted that the decrease in PA with increasing molecular size was more substantial for the hemodialysis membrane, thereby indicating a higher intrinsic permeability or porosity of the peritoneal membrane. The high intrinsic diffusive permeability of the peritoneal membrane has been reported variously by other investigators; indeed, this observation contributed to the generation of the middle molecule hypothesis first proposed by Scribner in 1965 (Meeting of the American Society for Artificial Internal Organs, Atlantic City, NJ).
The identification of PA as the product of two separate terms, peritoneal membrane permeability and peritoneal membrane surface area, is helpful because it emphasizes that solute transport rates across the peritoneal membrane can be influenced separately by either of these parameters. It should be emphasized, however, that determinations of peritoneal solute kinetics by themselves cannot be used to evaluate peritoneal membrane permeability or surface area individually; only the product of these two together can be determined in such experiments. Further understanding of the determinants of peritoneal transport required novel analyses to evaluate the permeability and surface area separately.
The anatomical surface area of the peritoneal membrane is large and was first reported by Wegner in 1877 (9) to approximate skin surface area, 
1.73 m2 for a person who weighs 70 kg. Investigators subsequently measured the surface area of both infants (10) and adults (11) and reported that peritoneal membrane surface area is similar to but somewhat less than that reported originally by Wegner. Those studies were based on direct autopsy measurements of the surface area; however, such measurements do not identify the surface area of the peritoneal membrane actually in contact with the dialysis fluid within the peritoneal cavity—that is, the functional surface area of the peritoneal membrane available for transport.
Estimates of the functional surface area in contact with dialysis fluid have been made by several workers. On the basis of equal values of the PA of the peritoneal membrane and the MTAC of a 1-m2 cellulosic hemodialysis membrane for inulin and knowledge of the higher intrinsic diffusive permeability of the peritoneal membrane than such hemodialysis membranes (see above), Henderson (12) reasoned that the functional surface area of the peritoneal membrane must be substantially <1 m2. Recent work by others suggests that this contention is indeed correct. First, Flessner (13) estimated the functional surface area of the peritoneum indirectly. He first measured the mass transfer coefficient (MTC) of various tissues that make up the peritoneal membrane by the use of a novel diffusion chamber. In the same investigation, Flessner next measured the anatomic surface area for each tissue making up the peritoneal membrane and calculated the MTAC expected if all of the anatomical surface area participated in solute transport under the assumption that MTAC = MTC x A. The calculated MTAC values were greater, at least by a factor of two, than those measured in separate kinetic experiments, which suggests that the functional surface area of the peritoneal membrane that participates in solute transport was less than the anatomical area. Second, investigators from Israel (14) have estimated the surface area of the peritoneal membrane in contact with dialysis solution by applying stereologic techniques to computerized tomography scan images of patients undergoing peritoneal dialysis, during treatment. In the six patients studied, the surface area of the peritoneal membrane in contact with dialysis solution was directly measured as 0.55 ± 0.04 m2. Third, Flessner et al. (15) have recently measured both the surface area of the peritoneal membrane in contact with fluids placed within the peritoneal cavity and the anatomical surface area of the peritoneal membrane in the mouse. One hour after solution was placed within the peritoneal cavity, the contact (or functional) area was, on average, 41% of the anatomical area of the peritoneal membrane.
The above definition of the contact or functional surface area of the peritoneal membrane was used because this parameter can be directly measured. If this definition is used, the permeability of the peritoneal membrane depends on several additional parameters (Table 1). For example, peritoneal membrane permeability depends on the permeability of capillaries that perfuse the various types of tissue surrounding the peritoneal cavity, the diffusivity of solutes within the peritoneal interstitium, and the rate of blood perfusion through peritoneal tissues (16). When the surface area is defined as the contact area between the peritoneal membrane and fluid within the peritoneal cavity, it is interesting to note that peritoneal membrane permeability is also a function of the density of peritoneal capillaries (or capillary surface area per unit volume of peritoneal tissue) that perfuse these tissues. Because changes in capillary surface area may induce alterations in diffusive solute transport rates similar to those induced by changes in peritoneal membrane area in contact with dialysis solution, other investigators have defined an alternative parameter called the "effective surface area" (17). This term assigns any alteration in either peritoneal membrane area in contact with dialysis solution or peritoneal capillary surface area that leads to changes in peritoneal solute transport rates to a change in effective surface area. On the basis of differences in creatinine MTAC values (which primarily reflect differences in effective surface area) among a cross section of patients undergoing peritoneal dialysis, Zweers et al. (18) have recently suggested that such differences are due to local production of vascular endothelial growth factor within the peritoneum. Furthermore, these (19) and other investigators (20) have suggested that increased local vascular endothelial growth factor production within the peritoneum may stimulate neoangiogenesis and lead to enhanced capillary surface area in patients undergoing long-term peritoneal dialysis.
|
Two factors act to limit the potential of convection in increasing solute transport across the peritoneal membrane. First, the rate of convective solute transport does not equal the transperitoneal ultrafiltration rate, because solutes are rejected at the peritoneal membrane such that solute sieving coefficients are less then unity, even for small solutes. Second, transperitoneal ultrafiltration is opposed by fluid absorption from the peritoneal cavity into the adjacent tissues and lymphatic vessels. The mechanisms that govern these limitations are only beginning to be understood; the current understanding of these transport pathways is described separately below.
Determinants of Solute Sieving
The importance of convection in enhancing solute transport during peritoneal dialysis was first demonstrated in 1966 (23). In those experiments, Henderson added urea to hypertonic dialysis solution at the same concentration as that in the plasma water of patients undergoing intermittent peritoneal dialysis. Because diffusive solute transport across the peritoneal membrane depends on the solute concentration difference between that in blood and dialysis solution, this intervention essentially blocked diffusive solute transport; therefore, any solute movement across the peritoneal membrane could be attributed to solute transport accompanying transperitoneal ultrafiltration by convection or solvent drag. The sieving coefficient for urea determined in these experiments was greater than zero, which demonstrated significant convective solute transport during peritoneal dialysis when hypertonic dialysis solutions were used. Surprisingly, however, the urea sieving coefficient was less than one, the value expected for an open or porous membrane, such as that demonstrated when peritoneal membrane permeability from diffusive solute transport measurements was assessed (see above). These observations were subsequently reproduced by Henderson and Nolph (6), who demonstrated approximately equivalent convective solute transport rates for both urea and inulin across the peritoneal membrane. These initial, original studies used dialysis solutions that contained 7% glucose, a concentration that is relatively high and may increase peritoneal membrane permeability (6). Subsequent work by Rubin et al. (24), who performed essentially similar, but more extensive, experiments using dialysis solutions that contained 4.25% glucose, confirmed the conclusions from these early studies. Thus, sieving coefficients for small solutes are less than one, which indicates that small solutes do not freely accompany water across the peritoneal membrane. Furthermore, these observations suggested that convective solute transport across the peritoneal membrane appears to occur across a tight or less porous membrane—a concept distinctly different from that when diffusive solute transport across the peritoneal membrane is analyzed.
A model of overall peritoneal solute transport that could explain both open diffusive and tight convective solute transport characteristics remained a mystery; some attempted to resolve this dilemma by insisting that measured sieving coefficients for small solutes less than one were an artifact of the experiment (25). One the basis of diffusive and convective solute transport measurements for small solutes, such as creatinine, and polydisperse dextrans over the molecular radius range from 13 to 50 Å in a rabbit model of peritoneal dialysis, Leypoldt et al. (26) suggested that transport across the peritoneum was akin to transport across two separate barriers in series. The barrier proximal to blood was postulated to be thin and tight (or restrictive), with a small mean pore size; this barrier was proposed as the capillary wall or endothelium. The distal barrier to blood was postulated to be thick yet open (or unrestrictive), with a large mean pore size; this barrier was postulated as the interstitium of peritoneal tissues. The overall transport properties of such a barrier are complex but have been studied elsewhere (27). Because overall diffusive solute transport across such a barrier is dominated by the thickest barrier (the longest diffusional pathlength), it was proposed that diffusion was limited by the relatively thick but open interstitial tissue layer. Thus, diffusive solute transport would appear to occur across a relatively open or unrestrictive membrane. In contrast, convective solute transport across such a barrier would be dominated by the proximal barrier (27)—that is, the capillary wall, which represents a tight barrier to convective solute transport. Thus, convective solute transport would appear to occur across a relatively tight or restrictive membrane. The transport characteristics of such a dual-barrier peritoneal membrane are similar to those described by a distributive model of peritoneal transport (28).
Subsequent work by Rippe and Stelin (29) provided a rigorous theory based on capillary physiology for understanding solute sieving by the peritoneum. Those workers first described peritoneal solute transport using a two-pore model of capillary transport, similar to that proposed conceptually by others (30). This two-pore model was, however, unable to explain the low sieving coefficients for the peritoneum. On the basis of this discrepancy, then, those investigators proposed the existence of water-only pathways within the peritoneal membrane that allowed transport of only water, not solutes. These water-only channels have been subsequently identified as aquaporin-1 (31) by several investigators (17). Because it provides specific physiologic details regarding peritoneal membrane pore structure, this three-pore model of peritoneal transport has become the paradigm for describing the kinetics of fluid and solute transport during peritoneal dialysis.
The identification and characterization of peritoneal membrane sieving characteristics are currently of high interest, especially those of sodium in patients undergoing peritoneal dialysis who have ultrafiltration failure (32). It has long been known that low sieving coefficients for sodium occur in patients undergoing peritoneal dialysis (33), and recent work has suggested that sodium sieving coefficients vary with the individual transport characteristics of the patient—that is, high-transport patients have high sodium sieving coefficients, and low-transport patients have low sodium sieving coefficients (34). In addition, sodium sieving coefficients decrease during peritonitis (35). It has further been conjectured that an increase in peritoneal membrane sieving coefficients indicates a loss of aquaporin-1–mediated water transport, which can lead to ultrafiltration failure in certain patients (36,37). For example, Ho-dac-Pannekeet et al. (37) reported that low sieving coefficients for sodium were not present in three of eight patients undergoing peritoneal dialysis who presented with overt ultrafiltration failure. This hypothesis is currently under investigation in an ongoing multicenter clinical trial (17).
Recent work in animal models, however, does not support an important role for aquaporin-1 in governing changes in peritoneal sieving coefficients during peritonitis or ultrafiltration failure: rather, the up-regulation of nitric oxide synthase appears to be more important (20,35,38). For example, Combet et al. (35) showed that the expression of nitric oxide synthase, but not aquaporin-1, was up-regulated in a rat model of peritonitis. In that model, sodium sieving coefficients were increased during peritonitis without up-regulation of aquaporin-1 expression, dissociating the expression of aquaporin-1 from acute changes in sodium sieving characteristics of the peritoneum. Additional recent studies by the same group that used biopsy tissue samples from patients undergoing peritoneal dialysis also did not support a major role for changes in aquaporin-1 expression in describing changes in peritoneal sieving or solute transport characteristics (20).
It should be noted that sieving coefficients for the peritoneal membrane depend on several other parameters in addition to the presence of water-only channels, specifically the transperitoneal ultrafiltration rate across the peritoneum. The importance of the transperitoneal ultrafiltration rate in determining solute sieving coefficients was first described by Henderson in 1981 (39) on the basis of the theoretical work by Spiegler and Kedem (40), which can be summarized as follows. When the transperitoneal ultrafiltration rate across a membrane is high, low sieving coefficients are readily observable because of the dominance of convective solute transport across the membrane. When, however, the transperitoneal ultrafiltration rate across a membrane is low, sieving coefficients are higher than expected because rapid transperitoneal diffusion decreases the solute concentration difference across the membrane. In many cases where low sodium sieving coefficients are not observable across the peritoneal membrane, such as in patients undergoing peritoneal dialysis who have high transport characteristics or during episodes of peritonitis, there are enhanced transperitoneal diffusion of sodium and low transperitoneal ultrafiltration rates. Thus, the absence of low sodium sieving coefficients may be the result, not the cause, of low rates of transperitoneal ultrafiltration. Recent work by Zweers et al. (41) has suggested how sodium sieving coefficients can be corrected for diffusion of sodium across the peritoneum, but the accuracy of the proposed correction formulae remains to be validated. Further work regarding the role of aquaporin-1 and nitric oxide synthase, together with accurate separation of sodium diffusive and convective transport properties in governing changes in sodium sieving coefficients, are necessary before the importance of aquaporin-1–mediated water transport can be directly evaluated in patients undergoing peritoneal dialysis who have ultrafiltration failure.
Pathways for Fluid Absorption from the Peritoneal Cavity
During peritoneal dialysis, fluid is lost continuously from the peritoneal cavity both directly into the tissues surrounding the peritoneal cavity and via lymphatic vessels (42). This pathway was originally evaluated by Starling and Tubby (43) over a century ago and has important consequences for patients undergoing peritoneal dialysis because it diminishes overall fluid removal. Although some have emphasized the importance of flow through lymphatic vessels in governing fluid lost from the peritoneal cavity (44), the majority of investigators agree that fluid absorption directly into the tissues surrounding the peritoneal cavity is the predominant pathway (42). There remains difficulty in understanding how significant fluid absorption into peritoneal tissues and transperitoneal ultrafiltration into the peritoneal cavity occur at the same time (Figure 2).
|
The recent study by Flessner et al. (15) suggested a possible explanation for the simultaneous occurrence of transperitoneal ultrafiltration and fluid absorption during peritoneal dialysis. As described briefly above, those investigators observed that the ratio of the area of the peritoneal membrane in contact with dialysis solution to the anatomic area was <50% when measured after 1 h after infusion of a solution (Krebs-Ringers buffer) into the peritoneal cavity. When, however, 24 h were allowed between fluid infusion into the peritoneal cavity and measurement of the contact area of the peritoneal membrane and the solution, essentially 100% of the peritoneal membrane came in contact with the solution within the peritoneal cavity.
A simplistic interpretation of these experiments is shown in Figure 3. After 1 h of dwell within the peritoneal cavity, only a fraction of the anatomical surface area of the peritoneal membrane has come in contact with the dialysis solution (Figure 3A). After 24 h of dwell, the dialysis solution has been able to penetrate all of the various interstices of the peritoneal cavity and thus come in contact with approximately the entire anatomical surface area of the peritoneal membrane. It is important to note that the solution that resides within the interstices of the peritoneal cavity is not easily drained, because the rate of fluid movement from these regions is slow. In essence, this hypothesis proposes that there exist two separate compartments of fluid within the peritoneal cavity: one with a large volume that is easily accessible to be filled and drained and a second one of smaller volume that equilibrates only slowly with the other compartment.
|
|
Specific Contributions of Lee W. Henderson to Peritoneal Dialysis
The contributions of Lee W. Henderson to the development of peritoneal dialysis as a maintenance therapy for patients with end-stage renal disease were several. He contributed, in the early days of peritoneal dialysis at the Peter Bent Brigham Hospital, to several practical aspects of this therapy (49,50). This had a major impact by setting up protocols that provided training to a number of nephrologists who trained at Harvard Medical School during the subsequent years.
As described in this review, the scientific contributions of Henderson in the development of peritoneal dialysis were in providing understanding of peritoneal transport, with respect to both diffusive and convective solute transport pathways. In particular, Henderson was instrumental in describing the peritoneum as a membrane that was amenable to mathematical and physical analysis. This focus on describing the peritoneum as a membrane was a major impetus in understanding peritoneal transport in several ways. First, describing the peritoneum as a membrane provided a simple basis for a number of kinetic models that have been used to describe peritoneal transport physiology and to evaluate a dose of therapy. Second, the characterization of the peritoneum as a membrane permitted the application of irreversible thermodynamics to the peritoneum. Although most are now familiar with such terms (sieving coefficient and reflection coefficient) when characterizing peritoneal transport, it should be noted that these concepts only became cemented in the biologic literature during the 1960s, after the publication of a number of papers and the influential book by Katchalsky and Curran (51). It should be noted that the first edition of this book was published in 1965, only 1 yr before the first description of solute sieving coefficients across the peritoneum by Henderson. Although there are certain experiments that require analysis of the peritoneum in terms more complicated than that of a simple membrane, the characterization of the peritoneum as a dialyzing membrane played a fundamental role in the development of more complex mathematical models of peritoneal fluid and solute transport and in formulating problems for future research.
References
This article has been cited by other articles:
 |
|
 |
|
  S.-K. Shin, C. D. Kamerath, J. F. Gilson, and J. K. Leypoldt Effects of anaesthesia on fluid and solute transport in a C57BL6 mouse model of peritoneal dialysis Nephrol. Dial. Transplant., October 1, 2006; 21(10): 2874 - 2880. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
HOME
CURRENT ISSUE
ARCHIVES
JASN Express
ONLINE SUBMISSION
AUTHOR INFO
EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP |
