## Abstract

ABSTRACT. Large dialysate volumes are often required to increase solute clearance for peritoneal dialysis patients. The resulting increase in solute clearance might be attributable to an increased plasma-to-dialysate concentration gradient and/or to an increased effective peritoneal surface area. One of the factors affecting the latter is the peritoneal surface area in contact with dialysate (PSA-CD). The aim of this study was to estimate the change in PSA-CD after a 50% increase in the instilled dialysate volume for patients undergoing peritoneal dialysis. PSA-CD was estimated by using a method applying stereologic techniques to computed tomographic (CT) scans of the peritoneal space. The peritoneal cavity of 10 peritoneal dialysis patients was filled with a solution containing dialysate, half-isotonic saline solution, and contrast medium. Peritoneal function tests and CT scanning of the abdomen were performed twice for each patient (with an interval of 1 wk), after instillation of a 2- or 3-L solution. Scanning of thin helical CT sections was performed, and 36 random sections of the abdomen were obtained after reconstruction. A grid was superimposed on the sections. The surface area was estimated by using stereologic methods. After instillation of the 2-L solution, the volume of the peritoneal solution at the time of CT scanning was 2.32 ± 0.05 L. The PSA-CD was 0.57 ± 0.03 m^{2}, ranging from 0.41 to 0.76 m^{2}. The use of the 3-L solution increased the peritoneal volume by 46 ± 2%. PSA-CD increased by 18 ± 2.3% to 0.67 ± 0.04 m^{2} (range, 0.49 to 0.84 m^{2}; *P* < 0.01). Creatinine mass transfer increased from 112 ± 10 mg to 142 ± 11 mg (*P* < 0.0001). The slope of the change of the plasma-to-dialysate creatinine concentration gradient with time decreased from −2.26 ± 0.23 × 10^{−2} to −1.97 ± 0.16 × 10^{−2} (*P* = 0.01). *K*_{BD-0} (permeability-surface area product or mass area transfer coefficient at time 0 of the dwell) increased from 10.6 ± 0.7 to 13.6 ± 1.2 ml/min (*P* < 0.02). These data demonstrate that increasing the instilled dialysate volume by 50% for peritoneal dialysis patients results in significant increases in the PSA-CD and *K*_{BD}. E-mail: uzig@clalit.org.il

Large dialysate volumes are often required to increase solute clearance for patients undergoing peritoneal dialysis (1,2). Theoretically, the increased solute clearance obtained with the use of large volumes (2–4) could be attributable either to an increased blood/dialysate solute concentration gradient or to an increase in the effective peritoneal surface area. Experimental studies have demonstrated that increasing the instilled dialysate volume from 2 to 3 L results in inconsistent changes in small-solute mass transfer area coefficients among peritoneal dialysis patients (3,4). A 250% increase in dialysate fill volume did not affect the diffusive mass transfer coefficients for urea and glucose in rats (5). The mass transfer area coefficient, or permeability-surface area product, is the solute clearance via diffusion per unit of concentration gradient (*K*_{BD}). *K*_{BD} provides information on the effective peritoneal surface area, *i.e*., the vascular peritoneal surface area (6,7) times the effective peritoneal solute permeability. The vascular peritoneal surface area is itself dependent on the peritoneal surface area in contact with dialysate (PSA-CD) and on the density of the pores of the perfused capillaries.

Studies in rodents by Flessner and co-workers (8,9) demonstrated that the PSA-CD during dialysis exchange was significantly less than that of the anatomic peritoneum and that this contact area could be increased with animal agitation or administration of surfactant (9). In a subsequent study, those authors demonstrated that increasing the peritoneal surface contact area with a surfactant resulted in increased peritoneal transport (10). A study performed with human subjects in our laboratory (11) confirmed the findings of Flessner’s group in rodents, demonstrating that the PSA-CD during a 2-L dialysis exchange was approximately 0.55 m^{2}. This value is lower than the estimates of the anatomic peritoneal surface area obtained from *post mortem* studies in human subjects (12–14). To estimate the PSA-CD, we have developed a method applying stereologic techniques to computed tomographic (CT) imaging of the peritoneal membrane (11). CT peritoneography allows visualization of the peritoneal space (15,16). The boundaries of the peritoneal space represent the mesothelial aspect of the peritoneal membrane in contact with dialysate. The surface area of these boundaries is the PSA-CD. Stereologic techniques enable quantitative estimation of three-dimensional structures from two-dimensional sections, irrespective of their shape (17). Surfaces in three dimensions are observed as lines in two dimensions. Therefore, a method of quantifying the lines in two-dimensional sections would permit estimation of the corresponding surface area. We have validated the application of these methods to CT imaging (11).

Because the peritoneum is not used at its maximal capacity during 2-L dialysis exchanges (8–11), we hypothesized that increasing the dialysate fill volume from 2 to 3 L would result in an increase in the peritoneal contact area. In addition, we performed solute transport studies, in an attempt to clarify the relationships between peritoneal volume, surface area, and diffusion capacity.

## Materials and Methods

### Patients and Dialysis

Ten peritoneal dialysis patients participated in the study. They were 55 ± 6 yr of age (range, 38 to 78 yr). Three patients were female and seven were male. All except two had been treated with peritoneal dialysis for >1 yr (median, 3.2 yr; range, 0.8 to 4.7 yr). Five patients were undergoing chronic ambulatory peritoneal dialysis, and five were undergoing automated peritoneal dialysis. All patients were treated with a large dialysate volume (≥2.5 L). The height and weight of the patients were 1.67 ± 0.04 m and 74 ± 3 kg, respectively. Their body surface area was 1.82 ± 0.04 m^{2}. Studies were performed in the morning, after drainage of the fluid from the peritoneal cavity. Patients were studied twice (with an interval of 1 wk), after instillation of a 2-L solution (first study) or a 3-L solution (second study). Each 1 L of solution contained 790 ml of 1.36% Dianeal glucose (Teva Medical, Ashdod, Israel), 160 ml of 0.45% NaCl, and 50 ml of contrast material (Ultravist 300, containing 623 mg/ml iopromide; Schering, Berlin, Germany). 2.5 μCi of radiolabeled serum ^{125}I-albumin (RISA) (Seralb-125; CIS Biointernational, France) was added for estimation of the peritoneal dialysate volume (18). This addition was performed after priming of the bag with 1 ml of a salt-free 20% albumin solution, to minimize adhesion of the radiolabeled albumin to the surfaces of the plastic material. Peritoneal function studies were followed by CT imaging of the peritoneal space.

### Study 1

The 2-L solution was infused through an infusion set. The patients rolled from side to side, and 200 ml were drained out and mixed. Ten milliliters were sampled for radioactivity counting and measurement of creatinine levels (time 0), and the remainder of the solution was instilled back into the peritoneal cavity. Blood was sampled at 0, 90, and 180 min, and dialysate was sampled eight times, at 4, 30, 60, 90, 120, 150, 180, and 205 min. The last sample was drawn just before CT scanning. Series of helical scans of the patient were then obtained, from the level of the diaphragm to the level of the symphysis pubis (Helicat II CT scanner; Marconi, Haifa, Israel). Scan parameters were as follows: for thin patients, slice width, 2.7 mm; slice interval, 1.4 mm (50% overlap); pitch, 1.5; for other patients, slice width, 3.2 mm; slice interval, 1.6 mm (50% overlap); pitch, 1.0. Image parameters were as follows: window center, 0; window width, 350. Dialysate and blood samples were withdrawn before and after each series. After scanning, the images were reconstructed and sets of 36 systematic, isotropic, uniform, random sections of the whole peritoneum were obtained.

### Study 2

The same steps were followed as in study 1, using 3 L of solution.

### Data Analysis

#### Estimation of the Intraperitoneal Dialysate Volume.

The intraperitoneal dialysate volume was calculated by using RISA as a volume marker, with a correction applied for its elimination from the peritoneal cavity (18), based on the timed dialysate samples obtained during the peritoneal function tests and CT scanning. The repeated sampling was used to estimate the changes in intraperitoneal fluid volume during the peritoneal function tests and CT scanning (which lasted approximately 40 min). The dialysate volume during CT scanning was calculated as the mean of the four measurements. The peritoneal space volume at time *t* of each sampling [*V*_{D}(*t*)] was calculated as follows:

where *V*_{A}(*t*) is the apparent volume of dialysate calculated from the dilution of RISA at time *t* and CF(*t*) is a correction factor at time *t*.

where *V*_{A}(*T*) is the apparent volume of dialysate estimated from the dilution of RISA at the end of the dwell time (*i.e*., at time *T*), *V*_{out} is the volume of dialysate drained after termination of the dwell, and *V*_{res} is the residual dialysate volume after drainage, calculated from the equation

where RISA_{DIAL-T} is the concentration of RISA at the end of the dwell time, *V*_{rinse in} is 1 L of fresh dialysate (1.36%, without RISA), and RISA_{DIAL-rinse out} is the concentration of RISA in the drained rinsing fluid.

#### Stereologic Measurements.

A transparent counting grid formed of test lines (multipurpose test system) (17) was placed over the CT images. The test line length, after correction for the scale, was 13.9 mm for patients 1 to 3, 13.5 mm for patients 4 and 5, and 13.0 mm for patients 6 to 10. The PSA-CD was estimated by counting the number of intersections of the test lines with the boundaries of the peritoneal fluid and the number of lines passing over the reference volume. The intraperitoneal dialysate volume was used as the reference volume. The ratio estimate of the PSA-CD (*S*_{V-PSA-CD}) was generated by using the relationship *S*_{V-PSA-CD} (m^{2}/m^{3}) = 2*I*/*L*, where *I* is the number of intersections with the surface of interest and *L* is the total length of test lines within the reference volume, after correction for the scale of the print. The ratio estimate of the PSA-CD was converted to an absolute value (PSA-CD) by multiplication by the peritoneal volume (*V*_{perit}), as follows: PSA-CD (m^{2}) = *S*_{V-PSA-CD} (m^{2}/m^{3}) × *V*_{perit} (m^{3}).

#### Peritoneal Function Tests.

Dialysate creatinine concentrations were corrected for the effects of the glucose concentration by using the formula Cr_{DIAL} = (1.021 × Cr_{DIAL-meas}) − (0.00008 × Gluc_{DIAL}) − 0.065, where Cr_{DIAL} is the corrected dialysate creatinine concentration (mg/dl), Cr_{DIAL-meas} is the measured dialysate creatinine concentration (mg/dl), and Gluc_{DIAL} is the dialysate glucose concentration (mg/dl). The plasma creatinine concentration was calculated as the mean of the plasma creatinine concentrations (mg/dl) measured at 0, 90, and 180 min.

*K*_{BD}, *i.e*., the rate of creatinine transfer from plasma to dialysate via diffusion per unit of plasma-to-dialysate concentration gradient, is not constant during the dialysis dwell. It varies with time, being higher during the initial phase of the dwell (19–21) and decreasing exponentially with time (20). In this study, *K*_{BD} was calculated for time 0 of the dwell (*K*_{BD-0}) from the regression equation of log *K*_{BD} with time

Where a is the regression coefficient. Equation 4 indicates that, at *t* = 0, log *K*_{BD} equals the constant of the regression equation (log *K*_{BD-0} = constant). Therefore, *K*_{BD-0} = *e*^{constant}.

*K*_{BD(t)} was determined during seven time periods, the first starting at 4 min and the last ending at 205 min; the sampling times are listed above. Each period was defined by the times of two consecutive samplings of dialysate. The time *t* for each *K*_{BD} determination was defined as the average time of each time period. *K*_{BD(t)} was calculated as the rate of creatinine mass transfer from plasma to the peritoneal solution via diffusion (Cr_{diff}) divided by the mean concentration gradient between plasma and dialysate during period *n* (Cr_{P-D gradient}).

Cr_{P-D gradient} was calculated as Cr_{P-D gradient} = Cr_{PL} (mg/ml) − Cr_{DIAL} (mg/ml), where Cr_{PL} is the plasma creatinine concentration and Cr_{DIAL} is the mean of the dialysate creatinine concentrations at the beginning and the end of period *n*.

The creatinine mass cleared via diffusion (Cr_{diff}) was calculated for each period as Cr_{diff} = Cr_{total} − Cr_{net uf} + Cr_{abs}, where Cr_{total} is the creatinine mass transferred from plasma to dialysate during period *n*, Cr_{net uf} is the creatinine mass transferred from blood to the peritoneal cavity by convection via net ultrafiltration during period *n*, and Cr_{abs} is the creatinine mass transferred from plasma to the peritoneal cavity via diffusion and eliminated from the peritoneal cavity via tissue absorption and lymphatic vessels during period *n*.

Cr_{total} was calculated as Cr_{total} = [Cr_{DIAL-final} (mg/ml) × *V*_{DIAL-final} (ml)] − [Cr_{DIAL-initial} (mg/ml) × *V*_{DIAL-initial} (ml)], where Cr_{DIAL} is the creatinine concentration in the dialysate and *V*_{DIAL} is the dialysate volume, obtained from Equation 1. Initial and final refer to the values at the beginning and the end of period *n*, respectively.

Cr_{net uf} was calculated as Cr_{net uf} = [*V*_{DIAL-final} (ml) − *V*_{DIAL-initial} (ml)] × Cr_{PL} (mg/ml) × Siev_{CR}, where Cr_{PL} is the plasma creatinine concentration and Siev_{CR} is the sieving coefficient for creatinine. An assumed value of 0.55 was used for Siev_{CR}; this value is the mean of the minimal and maximal values noted in animal studies for sieving coefficients for small solutes (range, 0.3 to 0.8) (22).

Cr_{abs} was calculated as Cr_{abs} = Cr_{DIAL} (mg/ml) × *V*_{abs-n} (ml), where Cr_{DIAL} is the mean creatinine concentration in the dialysate during period *n*. *V*_{abs-n} is the volume of dialysate absorbed into peritoneal tissues and lymphatic vessels during period *n*. *V*_{abs-n} was calculated as

RISA_{DIAL-initial} and RISA_{DIAL-final} refer to the dialysate concentrations of RISA at the beginning and the end of period *n*, respectively. RISA_{DIAL-mean} refers to the mean concentration of RISA during period *n*.

The creatinine mass transfer was calculated as the sum of Cr_{total} for all *n* periods. The change in Cr_{P-D gradient} with time was estimated by linear regression analysis of Cr_{P-D gradient} *versus* time during the *n* periods and was expressed as the slope of the regression line.

### Statistical Analyses

Data are expressed as mean ± SEM. The significance of differences between the 2-L and 3-L groups was evaluated by using a two-tailed paired *t* test.

Informed consent was obtained from all participants. The study was approved by the Rabin Medical Center Ethics Committee.

## Results

### Peritoneal Membrane Surface Area

After instillation of the 2-L solution, the volume of peritoneal solution at the time of CT scanning was 2.32 ± 0.05 L. *S*_{V-PSA-CD} was 2.43 ± 0.12 m^{2}/m^{3}. PSA-CD was 0.57 ± 0.03 m^{2}, ranging from 0.41 to 0.76 m^{2}.

Use of the 3-L solution increased the peritoneal volume by 46 ± 2%, to 3.38 ± 0.06 L (*P* < 0.0001). *S*_{V-PSA-CD} decreased by 19 ± 3%, to 1.97 ± 0.10 m^{2}/m^{3} (*P* < 0.001). PSA-CD increased by 18 ± 2%, to 0.67 ± 0.04 m^{2} (range, 0.49 to 0.84 m^{2}; *P* < 0.001) (Figure 1).

### Peritoneal Function Tests

Creatinine mass transfer was 112 ± 10 mg with the 2-L solution and increased to 142 ± 11 mg with the 3-L solution (*P* < 0.0001). This represents the solute mass transferred from blood to the peritoneal space via diffusion and convection, not including the absorbed solute mass. The slope of the change in the plasma-to-dialysate creatinine concentration gradient with time was −2.26 ± 0.23 × 10^{−2} with the 2-L solution and decreased to −1.97 ± 0.16 × 10^{−2} with the 3-L solution (*P* = 0.01). This decrease is attributable to an enlarged distribution volume for creatinine with the 3-L solution, resulting in a wider plasma-to-dialysate creatinine concentration gradient during the dwell time. *K*_{BD-0} was 10.6 ± 0.7 ml/min with the 2-L solution and increased to 13.6 ± 1.2 ml/min with the 3-L solution (*P* < 0.02), reflecting the increased diffusing capacity of the peritoneal membrane.

## Discussion

This study demonstrated that increasing the peritoneal dialysate volume by 46% resulted in an 18% increase in PSA-CD. PSA-CD was estimated by using a method previously validated by our group (11). The present estimate of PSA-CD with a 2-L solution (0.57 m^{2}) confirmed the findings of our initial study (0.55 m^{2}) (11). The larger number of patients studied revealed large interindividual variations (0.41 to 0.76 m^{2}).

PSA-CD was determined as *S*_{V} times the actual dialysate volume. The latter variable was also used for the estimation of *K*_{BD-0}. The actual dialysate volume was assessed with the tracer dilution technique, using RISA as a macromolecular marker. This method assumes the disappearance of RISA to be linear during the dwell time. In fact, Zakaria and Rippe (23) demonstrated that the disappearance of RISA was faster during the first 30 min of the dwell period and then followed a nearly linear course. Those authors demonstrated that the assumption of linear disappearance led to a slight overestimation of the dialysate volume at the beginning of the dwell time (approximately 4% in rats) and this overestimation decreased with time. Therefore, the use of this method for the estimation of K_{BD-0} and PMSA-CD leads to minimal overestimation of these parameters.

The results of this study confirm that the PSA-CD after instillation of a 2-L peritoneal solution is smaller than the surface area of the anatomic peritoneum. The latter value was estimated in early studies to range between 1.8 and 2.1 m^{2} (12,13) and was determined in more recent studies to approximate 0.8 to 1 m^{2} (13,14). The results presented here are in accordance with rodent studies reported by Flessner *et al*. (8–10), which demonstrated that only approximately 25 to 40% of the peritoneal surface area was in contact with dialysate after instillation of a dialysis solution scaled to 2 to 3 L for a 70-kg human subject. Interestingly, those experimental data and the findings of our study confirm the theoretical estimations proposed in 1973 by Henderson (24), who reasoned that the functional surface area of the peritoneum must be substantially less than 1 m^{2}. In light of these experimental studies and the difference observed between the PSA-CD with a 2-L dialysis solution and the surface area of the anatomic peritoneum (11), we hypothesized that increasing the peritoneal volume would result in a larger PSA-CD.

This study demonstrates that, after a 50% increase in the instilled dialysate volume, the peritoneal fluid volume increased by 46%. This difference is accounted for by the presence of a residual dialysate volume. PSA-CD increased by 18%, reaching 0.67 m^{2}, which is still less than the estimated surface area of the anatomic peritoneum (12–14). This increase was associated with an increase in the small-solute *K*_{BD-0}. *K*_{BD}, the permeability-surface area product or mass area transfer coefficient, is the solute clearance via diffusion per unit of concentration gradient. It is not influenced by the decrease in the solute concentration gradient that occurs during dialysis (7). However, it varies with time, being higher during the initial dwell phase (19–21). Carlsson and Rippe (21) demonstrated, in a rat model, that peritoneal arteriolar vasodilation is a major factor accounting for this early increase, although the authors could not exclude a role played by the macroscopic stirring that occurs when fluid is instilled into the peritoneal cavity. In this study, the *K*_{BD} values reported are those calculated at time 0 of the dwell (*K*_{BD-0}).

The changes in *K*_{BD} reflect changes in membrane and non-membrane factors affecting the effective surface area and the effective permeability of the peritoneal membrane. Therefore, the increase in PSA-CD associated with the increase in the instilled volume may not be the only factor accounting for the increase in *K*_{BD-0}. Other factors, affecting the effective permeability, might also account for the augmented *K*_{BD-0} associated with the increase in volume. One of these factors might be a differential increase in the contact area of the visceral and parietal peritoneum. It was demonstrated that the effective permeability of the visceral peritoneum is poor, compared with that of the parietal peritoneum (25–29). Zakaria *et al*. (29) suggested that the relatively low efficiency of the visceral peritoneum with respect to transport might be attributable to the presence of unmixed or poorly mixed fluid trapped in the many pouches formed by the complex shape of the peritoneal space (30). This assumption is supported by the findings of improved contact area (9) and solute clearance (29,31) resulting from agitation or vibration applied to the abdomen of experimental animals, suggesting that lack of mixing is a main factor limiting solute transport by the visceral peritoneum. Despite these limitations, the data collected in this study may provide a unique method to estimate *in vivo* the effective peritoneal surface area and effective permeability determining *K*_{BD} in human subjects.

This study demonstrates that the 3-L solution increases the peritoneal surface area exposed to dialysate. However, this large volume does not make contact with the entire anatomic peritoneal surface. This suggests that an additional increase in dialysate volume might recruit a larger peritoneal surface area, resulting in a higher *K*_{BD}. However, the associated increase in hydrostatic pressure in the peritoneal space would be expected to increase fluid absorption by peritoneal tissues and lymphatic vessels, thus decreasing net ultrafiltration (3,5,32–36) and consequently solute removal. Nevertheless, the fact that the PSA-CD with the 3-L solution was still less than the estimated anatomic peritoneal surface area suggests that methods should be sought to provide an additional increase in PSA-CD without increasing the hydrostatic pressure inside the peritoneal cavity. Flessner *et al*. (9,10) demonstrated that surfactant enhanced the PSA-CD in mice and rats, probably by decreasing the surface tension between peritoneal surfaces. This effect was associated with an increase in solute mass transfer (10). The use of surfactants for human subjects (if such treatment is proved to be nontoxic) might provide a method for enhancing PSA-CD without increasing hydrostatic pressure.

In addition to peritoneal dialysis, the intraperitoneal administration of solutions has been used for regional chemotherapy for the treatment of metastatic cancer (37,38). Such concentrations are not achievable via intravenous administration without severe systemic side effects (39). The rationale behind this treatment is that it provides high tissue concentrations of the therapeutic agent. Two physiologic determinants of the efficacy of this treatment are surface exposure (*i.e*., PSA-CD) and tissue penetration (40). The 18% augmentation in PSA-CD after an increase in the volume of the peritoneal solution from 2 to 3 L suggests that the use of large volumes may improve the efficiency of intraperitoneal chemotherapy. The increase in peritoneal hydrostatic pressure associated with the use of large-volume solutions is a limiting factor in peritoneal dialysis, resulting in decreases in net ultrafiltration and solute removal (3,5,32–36). This effect would not be a drawback for peritoneal chemotherapy, because such treatment is not aimed at removing fluid and solutes. In fact, as demonstrated by Flessner (41), the increased pressure associated with large volumes might be an advantage in intraperitoneal regional therapy, because increased hydrostatic pressure results in improved tissue penetration of macromolecules. Therefore, the administration of intraperitoneal regional therapy with large-volume solutions might improve efficiency by increasing both surface exposure and tissue penetration.

In conclusion, this study demonstrates that increasing the instilled dialysate volume from 2 to 3 L among patients undergoing peritoneal dialysis increases the PSA-CD. This increase results in an augmented *K*_{BD}. These findings might have important applications in the fields of peritoneal dialysis and intraperitoneal chemotherapy.

## Acknowledgments

We thank Miki Grazutis and Miriam Yuran, radiographers of the Department of Radiology, Rabin Medical Center-Golda Campus, for invaluable help in performing this study.

- © 2002 American Society of Nephrology