Abstract
Objective
Peritoneal dialysis (PD) with a 7.5% icodextrin-containing dialysis solution provides prolonged ultra-filtration compared with glucose-based dialysis solutions. Colloid osmosis is the most likely mechanism, but studies in rats suggest it is caused by an increase in osmolality due to degradation of icodextrin. Therefore, human spent dialysate was analyzed with high-performance liquid chromatography (HPLC) using gel permeation size-exclusion chromatography. An increasing peak (with a low molecular weight, < 1000 Da) was observed during the dwell. The aim of this study was to quantitate breakdown products of icodextrin (which could explain this peak) and investigate whether there was a relationship with dialysate amylase concentration and dialysate osmolality.
Design
Long-dwell effluents (dwell time 9.15 – 14.30 hours) obtained from 12 PD patients using a 7.5% ico-dextrin solution during the night were analyzed. The following icodextrin breakdown products were measured: maltotetraose (G4), maltotriose (G3), maltose (G2), and glucose (G1). In 6 of these patients, the sugars maltoheptaose (G7), maltohexaose (G6), and maltopentaose (G5) were also determined in both effluent and plasma. In addition, G4, G3, G2, and G1 were measured in four Wistar rats during a 6-hour dwell study.
Results
In the human studies, the median distribution of the sugars in the effluent was G4, 6.7%; G3, 16.5%; G2, 23.1%; and G1, 53.5%. The osmolality in spent dialysate ranged between 288 and 326 mOsm/kg H2O. The median contribution of the sugars G2 – G4 was 5.4 mOsm/kg H2O. No correlation was present between dialysate osmolality and duration of the dwell (r = –0.04, p = 0.91); nor was there a relation between the concentration of G2 and duration of the dwell (r = 0.50, p = 0.10). No relationship was found between the amount of amylase and the concentration of G2 in the effluent (r = 0.49, p = 0.10), nor between the total concentration of the sugars G2 – G4 in the spent dialysate and dialysate osmolality (r = –0.31, p = 0.33). However, a strong correlation was seen between urea concentration and osmolality (r = 0.85, p < 0.001), and also between sodium concentration and dialysate osmolality in the spent dialysate (r = 0.92, p < 0.0001). The levels of the sugars G2, G3, and G4 in effluent were higher than in unused dialysate, but lower than or similar to plasma levels. Concentrations of the sugars G5, G6, and G7 were lower in spent dialysate than in unused dialysate, and higher than in plasma. In the rat study, dialysate osmolality increased with the duration of the dwell. A clear relationship was present between osmolality and concentration of the sugars G2 – G4 in the effluent. The median amount of amylase in the effluent was 1252 U/L.
Conclusion
A 7.5% icodextrin-based dialysis solution used during the long exchange caused only a slight increase in dialysate osmolality in humans. The osmolality at the end of the dwell in the human situation was dependent mainly on concentrations of the small solutes urea and sodium in the effluent. The contribution of icodextrin degradation products was marginal. In the rat, however, a clear relationship was present between osmolality and icodextrin degradation products in spent dialysate, explaining the increased dialysate osmolality at the end of the dwell. The difference between the two species can be explained by the very high amylase concentrations in the rat, leading to a rapid degradation of icodextrin. The rat is therefore not suitable to study peritoneal fluid kinetics using icodextrin as an osmotic agent.
The aim of the present study was to investigate whether intraperitoneal enzymatic degradation of icodextrin occurs in PD patients, and whether species differences are present between humans and rats.
Methods
Patients
Long-dwell effluents of 12 PD patients were analyzed. All patients used 7.5% icodextrin-containing dialysate for their long dwell on a chronic basis. In 6 patients, a blood sample was drawn (heparinized plasma) for analysis of icodextrin breakdown products. The blood was drawn after drainage of the long-dwell bag.
Rats
Four male Wistar rats (Harlan CBP, Zeist, The Netherlands), median weight 350 g, were investigated during an acute 6-hour standard peritoneal permeability analysis (SPARa). The 6-hour SPARa is an adaptation of the 4-hour SPARa as described by Zweers et al. (6). Dialysate was infused and sampled via a venflon (intravenous needle with a plastic sheet). Fluid was supplemented via an intravenous infusion of 3 mL/hour 0.9% NaCl in a tail blood vessel.
Assays
The icodextrin-containing dialysis solution was Extraneal (Baxter, Utrecht, The Netherlands), which contains 7.5% icodextrin as osmotic agent. The glucose standard was purchased from Merck (Merck, Darmstadt, Germany) and the standards for the 7.5% icodextrin breakdown products from ICN (Costa Mesa, CA, U.S.A.). Osmolality was assessed by a freeze point reduction method using an automated analyzer (EBMC, Kerkdriel, The Netherlands). Sodium was measured using an ion-selective electrode (Hitachi H747, Boehringer Mannheim, Mannheim, Germany). Urea was measured with an enzymatic method (Hitachi H747). Amylase was also measured with an enzymatic method using the substrate p-nitrophenylmaltoheptaoside (Hitachi H747). Analysis of spent dialysate with gel permeation size-exclusion chromatography was done with a polyhydroxyethyl aspartamide column, 200 x 9.4 mm, 30-nm pores (PolyLC Inc., Columbia, MD, U.S.A.). The mobile phase used was 50 mmol/L formic acid. Detection was done with a refractive index (RI) detector (ERC-7510, Erma Inc.; Tokyo, Japan). Quantification of the icodextrin breakdown products [maltoheptaose (G7), maltohexaose (G6), maltopentaose (G5), maltotetraose (G4), maltotriose (G3), maltose (G2), and glucose (G1)] was done according to the high-performance liquid chromatography (HPLC) method described by Honda et al., with some modifications (7). The apparatus used for HPLC was composed of one 300 and one 480 Gynkotek liquid chromatograph pump (Gynkotek, Germering, Germany), a Basic Marathon autosampler (Spark Holland, Emmen, The Netherlands), and an ABI 759A absorbance detector (ABI, Foster City, CA, U.S.A.). For separation of the 1-phenyl-3-methyl-5-pyrazolone (PMP) derivatives, a 10-cm Hypersil BDS 3 C18 column (Chrompack, Bergen op Zoom, The Netherlands) was used. The wavelength for detection was 245 nm. Before derivatization with PMP, one volume of perchloric acid (0.66 mol/L) was added to deproteinize the samples. After centrifugation (10 minutes at 10 000 rpm), 100 μL supernatant was mixed with 50 μL NaOH (4.26 mol/L), 50 μL arabinose (1 g/L), 400 μL H2O, and 600 μL PMP (8.71 g in 100 mL MeOH). Incubation followed for 30 minutes at 70°C, then 50 μL HCl (3.6 mol/L) was added. Thereafter, the incubation mix was extracted three times with CHCl3. A 200-μL aliquot of the reaction mix was put in a clear crimp vial and, after 500 μL phosphate buffer (0.1 mol/L, pH 3.0) was added, the sample was analyzed by HPLC. The mobile phase consisted of 14% acetonitrile in phosphate buffer (0.1 mol/L, pH 3.0) (A), and 22.5% acetonitrile in phosphate buffer (0.1 mol/L, pH 3.0) (B). A linear gradient was used for separation of the sugars, from 30% B to 55% B in 20 minutes. Thereafter the gradient was changed to 30% B to allow equilibration prior to the next injection.
Statistics
Pearson's linear correlation analysis was used for the calculation of correlation coefficients.
Results
Analysis of Human Effluent
When effluent was analyzed with HPLC using gel permeation size-exclusion chromatography, an increasing peak (with a low molecular weight, < 1000 Da) was observed during the dwell (Figure 1). When spent dialysate was spiked with G4, G3, G2, and G1, a rising peak with the same retention time was seen. To quantitate the contribution of icodextrin breakdown products to this peak, the sugars G4 to G1 were measured. Their median distribution was G4, 6.7%; G3, 16.5%; G2, 23.1%; and G1, 53.5%. Dialysate osmolality ranged from 288 to 326 mOsm/kg H2O (median 307 mOsm/kg H2O). The median contribution of the sugars G2 – G4 was 5.4 mOsm/kg H2O (Table 1). No correlation was found between dialysate osmolality and dwell duration (r = –0.04, p = 0.91) (Figure 2), or between the amount of G2 in spent dialysate and dwell duration (r = 0.50, p = 0.10). The dialysate amylase concentration ranged from 3 to 13 U/L. No relationship was found between amounts of amylase and G2 in the effluent (r = 0.49, p = 0.10). No correlation was present between total concentration of the sugars G2 – G4 and osmolality in the spent dialysate (r = –0.31, p = 0.33) (Figure 3). A strong correlation was found between urea concentration and osmolality (r = 0.85, p < 0.001) (Figure 4), and also between sodium concentration and osmolality (r = 0.92, p < 0.0001) (Figure 4).
Analysis of 7.5% icodextrin-containing human spent dialysate (results of one representative patient) and two standards with gel permeation size-exclusion chromatography (RI signal expressed as percentage of maximum): standards, dextran 70, and glucose (open circles); unused dialysate (crosses); dialysate sampled after 1 hour (open triangles); dialysate sampled after 4 hours (open squares); dialysate sampled after 8 hours (asterisks). The arrows show the retention time of maltotetraose (G4), maltotriose (G3), maltose (G2), glucose (G1), and urea).
Osmolality versus dwell time in human spent dialysate. No correlation was present (r = –0.04, p = 0.91).
Total concentration of the sugars maltose (G2), maltotriose (G3), and maltotetraose (G4) versus osmolality in human peritoneal effluent. No correlation was present (r = –0.31, p = 0.33).
Concentrations of sodium (closed triangles) and urea (closed squares) versus osmolality in human spent dialysate. A strong correlation was seen between urea concentration and osmolality (r = 0.851, p < 0.001), and between sodium concentration and osmolality (r = 0.923, p < 0.0001) in peritoneal effluent.
Data of Spent Dialysate of 12 CAPD Patients: Osmolality, Maltotetraose (G4), Maltotriose (G3), Maltose (G2), and Amylase
Comparison of Human Spent Dialysate with Plasma
The mean concentrations of G2 – G7 in plasma and spent dialysate were calculated for 6 patients and compared with concentrations in unused dialysis solution (Figure 5). Higher G2, G3, and G4 levels in effluent and plasma were seen in comparison with unused dialysate. The mean levels of G2, G3, and G4 were lower or similar in effluent than in plasma. Furthermore, the concentrations of G5, G6, and G7 in both plasma and spent dialysate were lower than in unused solution. For G6 and G7, no measurable concentrations could be found in plasma.
Mean (±SEM) concentrations of the sugars maltose (G2), maltotriose (G3), maltotetraose (G4), maltopentaose (G5), maltohexaose (G6), and maltoheptaose (G7) in unused dialysate (open bars), human spent dialysate (n = 6, lined bars), and human plasma (n = 6, dotted bars).
Analysis of Rat Spent Dialysate
Dialysate osmolality increased with dwell duration (Figure 6). Furthermore, a clear relationship was present between osmolality and total concentration of G2 – G4 in the effluent (Figure 7). The dialysate amylase concentration increased to a median level of 1252 U/L after 6 hours.
Mean (±SEM) dialysate osmolality versus dwell time in the rat.
Concentrations of the sugars maltose (G2), maltotriose (G3), and maltotetraose (G4) versus osmolality in peritoneal effluent obtained from individual rats.
Discussion
In the present study, no clinically significant degradation of icodextrin was found during a dialysis dwell in chronic PD patients. We performed this study because data obtained in a rat model suggested that intraperitoneal degradation occurs, explaining the increase in dialysate osmolality during a dwell (5). Indeed, gel permeation chromatography of drained dialysate showed the development of a peak in the MW region below 1000 Da during long dwells, which suggests an elevation in the amount of low MW sugars. This was confirmed when spent dialysate was spiked with G4, G3, G2, and G1. Determination of oligosaccharides in effluent showed higher concentrations of G4, G3, and G2 than in unused dialysate, but in the same ranges as or lower than plasma levels. It appears from these results that peritoneal transport of oligosaccharides from the circulation to the dialysate explains most of the increase in effluent G4 – G2 during the dwell, whereas icodextrin breakdown can explain only a small amount of the increase.
Patients use the icodextrin solution on a chronic basis, that is, once daily for the long dwell exchange. This explains the high plasma concentrations of oligosaccharides. The loss of icodextrin during an overnight dwell averages 20% of the instilled amount (3), probably by uptake into the lymphatic system. From there it will eventually reach the plasma, where local degradation to maltose occurs by circulating amylase (8). Plasma concentrations of oligosaccharides exceeding those in unused dialysate will lead to their diffusion from the circulation to the dialysate.
The contribution of G2 – G4 to the increased effluent osmolality was approximately 20%, while that of urea accounted for approximately 80%. This makes the contribution of oligosaccharides to crystalloid osmosis less important.
The dialysate amylase concentrations were generally below 10 U/L, which are in agreement with previously published values in continuous ambulatory PD (CAPD) patients (9). However, oligosaccharides may interfere with the enzymatic determination of amylase activity. (Icodextrin and p-nitrophenylmaltoheptaoside are both substrates for amylase, and therefore the activity of amylase will probably be underestimated.) The difference between our findings in humans and those reported in rats urged us to repeat the study by Wang et al. (5) in four rats to investigate whether we could reproduce their results. This was indeed the case: a marked rise in dialysate osmolality occurred that could be attributed to a marked increase in G2 – G4 effluent concentrations. This strongly suggests local degradation of icodextrin in rats. The most likely explanation is the 125-fold higher dialysate amylase concentration in these animals. The extremely high serum amylase concentrations in rats, as known from the literature (10), can fully explain the high effluent amylase concentrations in the rats, as amylase diffuses from plasma to peritoneal dialysate. The mean peritoneal amylase clearance in CAPD patients is 0.2 mL/minute (11). High serum amylase levels are not unique to the rat, but are also present in other rodents, such as the mouse and the guinea pig (10). The necessity for high amylase levels might be explained by the diet of these rodents.
It can be concluded that the use of icodextrin as osmotic agent in human PD is not accompanied by an important local degradation of this glucose polymer. In contrast, in the rat, local degradation of icodextrin occurs in the peritoneal cavity, which causes the rise in dialysate osmolality. The difference between the two species can be explained by the very high dialysate amylase concentrations in the rat. Therefore, rodents are not suitable to study peritoneal fluid kinetics using icodextrin as osmotic agent.