Secondary Logo

Journal Logo


Development of a Contamination Free 6 Valve Injector Inline Monitoring System for Endotoxin Measurement in Dialysate

Miyasaka, Takehiro*; Matsuda, Yasuko*; Sakai, Kiyotaka*; Mochizuki, Seiichi; Tanaka, Shigenori

Author Information
  • Free


Recently, dialysate has been widely used as a replacement solution in hemodiafiltration (HDF). Consequently, inflow of endotoxin into patients’ blood may occur due to use of an endotoxin contaminated replacement solution 1. Endotoxin stimulates monocytes and macrophages, resulting in release of cytokines such as tumor necrosis factor-α (TNF-α). 2,3 Excessive cytokine influx into blood thus causes dialysate resistant complications, including dialysis amyloidosis, 4,5 carpal tunnel syndrome, 6–8 and hypotension. 9,10 Use of ultrapure dialysate (endotoxin free) reduced stimulation of cytokine release, resulting in reduced β2- microglobulin synthesis and improved survival rate. 6-8 Accordingly, endotoxin levels in dialysate have been extensively studied. 11–15 The limulus amebocyte lysate (LAL) assay, a reaction of endotoxin with the LAL reagent, 16,17 is commonly used for endotoxin measurement; however, in this method, time consuming batch steps are necessary. Therefore, it would be of clinical value to develop an easy endotoxin monitoring system.

There have been previous studies on development of inline endotoxin monitoring systems, mostly by our group. Himi et al.18, Yotsumoto et al.19, and Yoshimi et al.20 developed a highly sensitive continuous endotoxin monitoring system in a single tube in which the LAL reagent and sample solution were mixed and reacted without atmospheric contamination. However, baseline absorbance was unstable due to the influence of the LAL reagent and endotoxin remaining in the tube. Aoyagi et al.21 could inhibit formation of endotoxin residue by using hydrophobic Teflon tubes in place of hydrophilic silicone tubes and by modifying the tube cleaning method. Moreover, reagent consumption was remarkably reduced by pulse injection of the reagent. Mizumoto et al.22 applied the same system and reached the detection limit of 50 EU/L at a reaction time of 45 minutes. However, dispersion of the reagent in the tube caused incomplete mixing with an endotoxin sample solution. Iijima et al.23 and Miyasaka et al.24 thus shortened the reaction circuit and proposed a stopped flow measurement system. In this system, the detection limit was 80 EU/L at a reaction time of 10 minutes, and 40 EU/L at 20 and 30 minutes. Thus, it is important to suppress intratube dispersion to improve the detection limit with shorter measuring time.

Endotoxin exists in the air. Monitoring devices are thus easily contaminated by endotoxin adsorption, causing errors in measurements. Accordingly, devices must be kept endotoxin free. Use of a new endotoxin free device for each measurement is not practical from an economic viewpoint. Instead, complete replacement of the solution in the circuit with another endotoxin sample solution would be more practical and efficient.

On the basis of this background, in the present study, we aimed at developing an inline endotoxin monitoring system with high sensitivity, short measuring time, and low cost. First, to simplify the system configuration, we used a high performance liquid chromatogram (HPLC) system, which can control excessive dispersion of the reagent; second, to eliminate possible endotoxin contamination, we proposed a 6 valve injector inline monitoring system (Figure 1). The merits of this new 6 valve system are as follows: (1) by introducing the LAL reagent into the circuit by means of a specific tube for the reagent, the system is completely sealed and endotoxin free, and (2) the stopped flow system allows all solutions to be well mixed.

Figure 1
Figure 1:
Six valve injector inline endotoxin monitoring system. LAL, limulus amebocyte lysate; UV/VIS, ultraviolet/visible;p-NA, p-nitroaniline.

Experimental Design

Monitoring System Configuration.

Figure 1 shows a schematic of our newly developed 6 valve injector inline monitoring system, consisting of a 6 valve injector (Teflon Rotary Valve 5011; Rheodyne, Cotati, CA), a reaction chamber (Urtrograd 11300 mixer; LKB, Tokyo, Japan), a LC-10Ai pump (Shimadzu, Kyoto, Japan), and an ultraviolet/visible SPD-10Avi spectrophotometer (cell volume,16 μl; Shimadzu). A poly-ether-ether-ketone (PEEK) chemical resistant tube (inner diameter, 0.5 mm) was used. The cylindrical reaction chamber (inner diameter, 6 mm; depth, 21 mm; inner volume, 250 μl) with a stirring bar was housed in a CO-8010 column oven (Toso, Tokyo, Japan) and kept at 40°C. Both the upper and lower sides of the chamber were connected with tubes (inner diameter, 0.5 mm). Endotoxin concentration was measured as absorbance (at 405 nm) of p-nitroaniline (p-NA), a stable end-product of the reaction between the LAL reagent and endotoxin.

Optimization of Operating Conditions Using a Stable End-Product Solution

Injection volume of a p-NA standard solution.

After filling the circuit with distilled water, the valve was switched to introduce 50 to 250 μl of 0.324 mM p-NA into the reaction chamber at a flow rate of 0.5 ml/min. Here, p-NA solution was used as a standard solution to exclude possible reaction effects. Distilled water and the p-NA solution were then mixed for 5 minutes. By introducing distilled water, the mixed solution was pushed into the detector and absorbance was measured. This process was repeated three times, and the results were presented as mean ± standard deviation. Dispersion coefficient was calculated from the absorbance ratio of the original p-NA solution and the mixture based on the previous approach. 25,26

Flow rate and mixing.

After filling the whole circuit with distilled water, the valve was switched to introduce 125 μl of 0.756 mM p-NA solution into the reaction chamber at a flow rate of 0.5 ml/min. Distilled water and the p-NA solution were then mixed for 5 minutes. By introducing distilled water, the mixed solution was pushed into the detector at a flow rate of 0.5 to 1.0 ml/min, and absorbance was measured. Absorbances with and without mixing after restarting flow were also compared to evaluate the mixing effect. This process was repeated three times, and the results were presented as means ± standard deviations.

Evaluation of System Performance Using an Endotoxin Standard Solution

Preparation of an endotoxin standard solution.

USP reference standard endotoxin (USPRSE; Seikagaku, Tokyo, Japan) produced from Escherichia coli O113:H10 was dissolved in either distilled water or dialysate (AP-3P Kindaly solution; Fuso Pharmaceuticals, Osaka, Japan). Endotoxin concentration of the USPRSE solution measured with a microplate reader (Wellreader; Seikagaku) was taken to be the “endotoxin concentration” in Figures 2 and 3.

Figure 2
Figure 2:
Time course of microplate reader monitored concentration of endotoxin leaving the 6 valve injector inline monitoring system. Endotoxin concentration in endotoxin units (EU)/L: open squares, 0→160; filled squares, 160→0; open triangles, 60→160; filled triangles, 160→60; open circles, 0→60; and filled circles, 60→0.
Figure 3
Figure 3:
Relationship between peak absorbance and standard endotoxin (USPRSE) concentration in distilled water (open circles, r2 = 0.99) and dialysate (filled circles, r2 = 0.99) (n = 3). Endotoxin concentration (endotoxin units, EU) was measured by the microplate reader.

Preparation of the LAL reagent.

To improve the stability of the endotoxin detection reagent (Endospecy ES-50 M set; Seikagaku), we modified the preparation method of the LAL reagent described in the manufacturer’s instructions. Normally, the LAL reagent is dissolved in a 2.8 ml of Tris-HCl buffer solution (pH 8.0), with autocoloring of the reagent beginning immediately after addition of the Tris-HCl buffer solution. We dissolved the LAL reagent in 2.8 ml of distilled water in which autocoloring of limulus reagent does not occur. Finally, the LAL reagent in distilled water and Tris-HCl buffer solution were mixed in the reaction chamber. We called this the “improved” LAL reagent. Improved LAL reagents were stable for more than 8 hours.

Confirmation of endotoxin free operation.

Distilled water was introduced at a flow rate of 2 ml/min. The USPRSE solutions (60 and 160 EU/L) and distilled water were then in turn introduced into the circuit at a flow rate of 2 ml/min. Each solution was then replaced with the next sample solution, and samples were obtained upon exiting from the circuit (waste in Figure 1). Endotoxin concentrations of these samples were measured using the microplate reader array.

Measurement of endotoxin concentration in the standard solution.

Table 1 shows the entire process. The sample solutions (USPRSE solution and USPRSE dialysate [AP-3P Kindaly]) were introduced at a flow rate of 2 ml/min for more than 10 minutes, and the flow was stopped after complete replacement of the solution in the circuit. Next, 50 μl of the LAL reagent solution and 0.4 M Tris-HCl buffer solution were introduced, and the flow was stopped when the LAL reagent solution and buffer solution entered the reaction chamber. The improved LAL reagent reacted with a sample solution for 20 minutes, and the reactant was introduced into the detector at a flow rate of 0.5 ml/min. This process was repeated three times, and the results presented as means ± standard deviation.

Table 1
Table 1:
Operation Sequence and Operating Conditions for Endotoxin Samples

Results and Discussion

Because an HPLC system can control dispersion of the reagent with a narrow tube, we developed an endotoxin monitoring system with an autoinjector for first trial. However, in the autoinjector inline monitoring system, the injection needle and autoinjector were found to be contaminated by endotoxin in the air. Thus to eliminate endotoxin contamination, a 6 valve injector inline monitoring system was developed (Figure 1). This new system design makes the circuit more airtight and prevents endotoxin contamination (Figure 2), while allowing easy operation, because each solution is introduced simply by switching the valve.

Optimized Operating Conditions

First, to investigate dispersion effects of the p-NA injection volume, peak absorbance of p-NA was compared between the stock p-NA solution and p-NA solution diluted with distilled water. Figure 4 shows the relationship between the dispersion (dilution) coefficient and p-NA injection volume; the dispersion coefficient decreased with increasing p-NA injection volume. When a mixed solution flows during reaction, medium dispersion is suitable. 26 In this system, an injection volume of 60–180 μl provided dispersion coefficient of 3 to 10, which is seen in medium dispersion. 26

Figure 4
Figure 4:
Relationship between dispersion coefficient and injection volume (n = 3).

Second, effects of flow rate and mixing condition after restarting flow were studied. Figure 5 shows the relationship between flow rate and peak absorbance of p-NA with and without mixing. The peak absorbance of p-NA was 1.80 × 106 (± 0.06 × 106) with mixing and 1.64 × 106 (± 0.06 × 106) without mixing. The absorbance was nearly independent of flow rate in the range of 0.5 to 1.0 ml/min. Peak absorbance with mixing after restarting the flow diameter was slightly larger than that without mixing. In this system, the flow was suddenly enlarged from a tube (0.5 mm in diameter) to the reaction chamber (6 mm in diameter), and the LAL reagent is of relatively high viscosity. Accordingly, flow channeling may occur and in fact was observed (data not shown). Therefore, it may have required a long time to completely wash out the LAL reagent from the chamber (longer retention time), and thus the LAL reagent may have been diluted further, resulting in the decrease in absorbance.

Figure 5
Figure 5:
Effects of flow rate and mixing on peak absorbance (n = 3). Continuous mixing (filled circles) and mixing only when the flow was stopped (open circles).

Because the absorbance was not influenced by the flow rate in the flow range of 0.5 to 1.0 ml/min (Figure 5), the flow rate can be set freely in this range. However, the LAL reagent is foamy; thus, flow rate was set as slow as possible, but pressure must be kept higher than 10 kgf/cm2 for stable pump operation. In contrast to a conventional HPLC system in which a separation column is connected between a pump and a spectrophotometer, and inline pressure is reduced by the column before the spectrophotometer, in this system the spectrophotometer is connected directly in series with the pump without a separation column so that operating pressure must be kept at < 10 kgf/cm2, the maximum resistant pressure of the sample cell in the spectrophotometer. Actual pump pressure was 8 kgf/cm2 (flow rate, 0.5 ml/min), 11 kgf/cm2 (0.7 ml/min), and 17 kgf/cm2 (1.0 ml/min). Accordingly, an optimal flow rate was determined to be 0.5 ml/min.

Solution Replacement Time and Confirmation of Contamination Free Operation

Replacement time of endotoxin standard solution in the monitoring system was investigated. Figure 2 shows the time course of endotoxin concentration at the exit of the circuit measured by the microplate reader using the LAL reagent and USPRSE solution. After 10 minutes, endotoxin concentration reached a steady level, indicating complete replacement of the endotoxin sample solution. Flowing distilled water (0 EU/L endotoxin) for more than 20 minutes made the circuit endotoxin free. This result also showed that the system was contamination free. Our new system did not require changing a tube from a sample solution bottle to a LAL reagent bottle; we simply changed the valve from the sample solution tube to the LAL reagent tube, allowing us to eliminate endotoxin contamination completely.

Measurement with the Endotoxin Standard Solution

Endotoxin concentrations in the USPRSE water solution and USPRSE dialysate were measured using the improved LAL reagents with a reaction time of 20 minutes. Figure 3 shows calibration curves for the USPRSE water and USPRSE dialysate. A good linear correlation between peak absorbance and endotoxin concentration was obtained in the range of 0 to 70 EU/L (r2 = 0.99; experimental error, 10%) for the USPRSE water solution and in the range of 0 to 30 EU/L (r2 = 0.99; experimental error, 10%) for the USPRSE dialysate. Sensitivity was estimated to be approximately 1 EU/L, equivalent to the conventional batch method using a microplate reader and considered to be practical for clinical use. The optimized operating conditions are summarized in Table 1.


A new inline endotoxin monitoring system was developed, and endotoxin contamination was completely eliminated using a 6 valve injector. Endotoxin concentration can be measured using the improved LAL reagent with a sensitivity of 1 EU/L, equivalent to the conventional microplate reader. This system allows easy operation and is suitable for monitoring endotoxin concentration in a dialysis line.


This study was supported in part by Grant-in-Aid 12780654 for Encouragement of Young Scientists and by Grant-in-Aid 13680961 for Scientific Research both from the Ministry of Education, Culture, Sports, Science, and Technology, Japan.


1. Stiller S, Mann H, Brunner H: Backfiltration in hemodialysis with highly permeable membranes. Contr Nephrol 46: 23–32, 1985.
2. Rietschel ET, Brade H: Bacterial endotoxins. Sci Am August: 267: 54–61, 1992.
3. Mege JL, Sanguedolce MV, Purgus R, et al: Chronic and intradialytic effects of high-flux hemodialysis on tumor necrosis factor-α production: Relationship to endotoxins. Am J Kidney Dis 20: 482–488, 1992.
4. Drueke TB: β2-Microglobulin and amyloidosis. Nephrol Dial Transplant 15: (Suppl.) 17–24, 2000.
5. Gejyo F, Suzuki S, Arakawa M: Dialysis-related amyloidosis and clinical significance of extracorporeal removal of β2-microglobulin. Ther Apher 1: 126–128, 1997.
6. Baz M, Durand C, Ragon A, et al: Using ultrapure water in hemodialysis delays carpal tunnel syndrome. Int J Artif Organs 14: 681–685, 1991.
7. Kleophas W, Haastert B, Backus G, Hilgers P, Westhoff A, van Endert G: Long-term experience with an ultrapure individual dialysis fluid with a batch type machine. Nephrol Dial Transplant 13: 3118–3125, 1998.
8. Schiffl H, Lang SM, Stratakis D, Fischer R: Effects of ultrapure dialysis fluid on nutritional status and inflammatory parameters. Nephrol Dial Transplant 16: 1863–1869, 2001.
9. Henderson LW, Koch KM, Dinarello CA, Shaldon S: Hemodialysis hypotension: The interleukin hypothesis. Blood Purif 1: 3–8, 1983.
10. Shaldon S: The interleukin-1 hypothesis: An update. Blood Purif 6: 162–163, 1988.
11. Baurmeister U, Travers M, Vienken J, et al: Dialysate contamination and back filtration may limit the use of high-flux dialysis membranes. ASAIO Trans 35: 519–522, 1989.
12. Harding GB, Klein E, Pass T, Wraight R, Million C: Endotoxin and bacterial contamination of dialysis center water and dialysate: A cross sectional survey. Int J Artif Organs 13: 39–43, 1990.
13. Gault MH, Duffett AL, Murphy JF: Purchase LH: In search of sterile, endotoxin-free dialysate. ASAIO J 38: M431–M435, 1992.
14. Bambauer R, Schauer M, Junk WK, Daum V, Vienken J: Contamination of dialysis water and dialysate. A survey of 30 centers. ASAIO J 40: 1012–1016, 1994.
15. Ledebo I, Nystrand R: Defining the microbiological quality of dialysis fluid. Artif Organs 23: 37–43, 1999.
16. Obayashi T, Tamura H, Tanaka S, et al: A new chromogenic endotoxin-specific assay using recombined limulus coagulation enzymes and its clinical applications. Clin Chim Acta 149: 55–65, 1985.
17. Tanaka S, Iwanaga S: Limulus test for detecting bacterial endotoxins. Methods Enzymol 223: 358–364, 1993.
18. Himi N, Yoshimi Y, Kanamori T, Sakai K: An endotoxin (ET) sensor for monitoring of dialysate concentration (in Japanese). Jpn J Artif Organs 21: 902–904, 1992.
19. Yotsumoto M, Himi N, Yoshimi Y, Kanamori T, Sakai K: Development of a medical sensor for monitoring of endotoxin concentration in dialysate (in Japanese). Jpn J Artif Organs 22: 1036–1039, 1993.
20. Yoshimi Y, Yotsumoto M, Sakai K: A novel method for the continuous measurement of endotoxin concentration. ASAIO J 41: 68–71, 1995.
21. Aoyagi S, Yoshimi Y, Sakai K, Aketagawa J: Development of a new endotoxin sensor with intermittent injection of limulus reagent for continuous monitoring of dialysate fluid. ASAIO J 41: M601–M604, 1995.
22. Mizumoto D, Yoshimi Y, Sakai K, Aketagawa J, Tanaka S: Optimal design of a continuous measurement system for endotoxin concentration in dialysate fluid (in Japanese). Jpn J Artif Organs 25: 98–101, 1996.
23. Iijima K, Miyasaka T, Yoshimi Y, et al: Development of a continuous system for monitoring endotoxin concentration in dialysate fluid (in Japanese). Jpn J Artif Organs 27: 524–527, 1998.
24. Miyasaka T, Iijima K, Sakai K, et al: Development of a stopped flow reactor for the rapid high sensitivity measurement of endotoxin concentration in dialysate. ASAIO J 45: 104–108, 1999.
25. Riley C, Rocks BF, Sherwood RA: Controlled-dispersion flow analysis, flow-injection analysis applied to clinical chemistry. Anal Chim Acta 179: 69–80, 1986.
26. Ruzicka J, Hansen EH: Flow Injection Analysis. New York, Wiley-Interscience, 1981.
Copyright © 2002 by the American Society for Artificial Internal Organs