GB2033081A

GB2033081A - Optical method for determining endotoxin

Info

Publication number: GB2033081A
Application number: GB7932452A
Authority: GB
Original assignee: Baxter Travenol Laboratories Inc
Current assignee: Baxter International Inc
Priority date: 1978-10-06
Filing date: 1979-09-19
Publication date: 1980-05-14
Also published as: FR2438269B1; DE2927103A1; FR2438269A1; IT7925293A0; IT1162772B; GB2033081B

Abstract

In an optical method for determining endotoxin, a sample is contacted with an endotoxin-coagulable protein from Limulus amoebocytes to form a reaction product containing coagulated protein, and the reaction product is contacted with an ionic detergent and a suspending agent, or with an ionic suspending agent which has detergent properties. The amount of suspended matter is determined by a nephelometric, turbidimetric or absorbance procedure.

Description

SPECIFICATION Optical method for determining endotoxin This invention relates to methods for the determination of endotoxin which use as their basis the endotoxin-activated enzymatic coagulation or block polymerization of certain proteins native to Limulua polyphemus blood cells. In particular, this invention relates to those methods in which the Limulus protein coagulation is followed by nephelometry.

The detection of endotoxin in solutions and in biological fluids has long been important in the economics of commerical production of therapeutic intravenous solutions, and is diagnosis and prognosis of human disease. This test has become an important factor in the cost of manufacturing parenteral solutions because the heretofore employed US Pharmacopeia method has required live rabbits and a complex and lengthy assay procedure.

Pyrogens of greatest commercial and disease importance are derived from gram negative bacteria. The terms "endotoxin", and "pyrogen" refer to substances present in the cell walls of bacterial families such as Brucellaceae, Enterobacteriaceae, Pseudomonadeceae, and Spirillaceae. Striking examples of the role of these substances in disease are evident when it is realized that endotoxins are the primary cause of fever, diarrhea, and abortion in diseases such as typhoid, cholera, food poisoning, and Brucellosis. Development of endotoxemia from bacteria residing in the gastrointestinal tract is a prime cause of death in appendicitis or other conditions involving rupture of the colon.

Endotoxin assays using Limulus blood cells extracts or lysates have gained considerable acceptance.

These assays makes use of the phenomenon that an enzyme present in the Limulus lysate is activated by endotoxin, and that once activated the enzyme catalyzes a reaction with other lysate proteins to form a crosslinked or coagulated protein matrix or gel. Since the amount of gel is proportional to the amount of endotoxin originally present, endotoxin in unknown samples may be readily determined.

Originally, formation of a Limulus lysate clot after incubation for a fixed time in the presence of endotoxin was the signal used to measure endotoxin concentration. While this method is generally more sensitive than the US Pharmacopeia rabbit test and is simple to conduct, it is insensitive and subject to considerable variation and error. The test is insensitive because coagulation occurring at low endotoxin levels may be unapparentto visual examination. It is subject to error because of the difficulty in subjectively determining the point at which a clot occurs.

The sensitivity and reproducibility of the endotoxin assay have been improved by the use of various optical methods for detecting lysate precipitation. Some methods measure the reduction in the amount of light which is passed by a suspension of coagulated Limulus protein, i.e., they measure the absorbence or turbidity of the suspension. See for example U.S. Patents 4,038,029 and 3,915,805, and Hollander et al., "Biochem. Med." 15: 28-33 (1976). Other workers have separated the coagulated protein from the reaction mixture, then chromogenically determined the total amount of coagulated protein. See the example in U.K.

Patent Specification 1,499,846.

Another optical method for determining the extend of coagulation is nephelometry. This method is frequently described or suggested in other terms, e.g., determination of light scattering. For example, U.K.

Patent Specification 1,499,846, U.S. Patent 3,915,805 and Levin et al. ("Thromb. Diath. Haemorrh." 19: 186-197 [1968]) all speak in terms of Limulus protein coagulation detection by "light scattering". For the purpose of this application nephelometric methods for the detection of endotoxin are those in which a determination is made of the amount of light reflected by a solution or suspension of coagulated Limulus amebocyte protein at greater than 0 and less than 18Q" from the incident beam of light as a measure of the extent of coagulation of the protein by an endotoxin-activated Limulus amoebocyte enzyme.

Nephelometric methods have not found widespread acceptance in endotoxin detection. In part, this may be attributed to the poor results that have been reported to date when using this technique. The principal problems which are evident are those of sensitivity and reproductibility. Levin metal. (Figure 8) reported no increase in light scattering of Limulus amoebocyte lysate when contacted with 40 pg endotoxin/ml., and not until the endotoxin level was increased to 4000 pg/ml was a lethargic response noted. Similarly, U.S. Patent 3,915,805 reported an endotoxin detection threshold of 500 pg/ml. Additionally, inspection of the specific results reported by this patent discloses considerably irregularity in the plotted points, even at high endotoxin concentrations. While U.K.Patent Specification 1,499,846 and Hollander et al. disclose low endotoxin thresholds the results were obtained using optical density or turbidimetric procedures.

It is an object of this invention to improve the sensitivity and reproducibility of nephelometric endotoxin assays.

The present invention provides an optical method for determining endotoxin wherein a sample thought to contain endotoxin is contacted with an endotoxin coagulable protein from Limulus amoebocytes to form a reaction product containing coagulated protein, wherein the reaction product is contacted with an ionic surface-active suspending agent.

The invention also resides in an optical method for determining endotoxin wherein a sample thought to contain endotoxin is contacted with an endotoxin-coagulable protein from Limulus amoebocytes to form a reaction product containing coagulated protein, wherein the reaction product is contacted with an ionic detergent and a suspending agent.

Reference is now made to the accompanying drawings wherein the sole Figure is a graph which demonstrates the improved sensitivity and reproducibility obtainable using the process of this invention.

Curves A, B, C and D, respectively, delineate a nephelometric determination without the agent of this invention, with suspending agent only, with ionic detergent, and with both ionic detergent and suspending agent.

The ionic detergents for use in this invention will be cationic, anionic or amphoteric, and they all should exhibit certain general characteristics. First, their molecular weight will generally range from 150 to 450.

Second, it is convenient to select detergents which are capable of stopping the coagulation of Limulus protein; the use of such a detergent eliminates any need for a separate step in the method for stopping the reaction. Third, suitable detergents will preferably exhibit a high hydrophylic/lipophylic balance (HLB), generally greater than about 20 and most usually about from 30 to 40. The HLB for a selected detergent or mixture of detergents is generally available from the manufacturer of the detergent or may be readily calculated in accordance with known processes, e.g., as disclosed in Belgian Patent 837,675.

Representative cationic detergents are the alkyl or aryl quaternary ammonium or quarternary pyridine salts such as cetyl trimethylammonium chloride, lauryl trimethylammonium bromide or benzyl triethylammonium chloride which may be used as mixtures.

Exemplary amphoteric detergents are the alkyl sulfonate or sulfate substituted quaternary ammonium salts having the general formulae:

wherein R is alkyl or aryl of from 1 to 9 carbon atoms, and A is a barboxyl, sulfate or sulfonate-substituted alkyl radical of from three to fourteen carbon atoms in which at least one A is sulfate or sulfonatesubstituted. Generally, R is a short chain, normal alkyl such as methyl or ethyl. On the other hand, the A group alkyl radical is ordinarily a long chain, normal alkyl, for example nonyl or dodecyl. These detergents are well known and a number are commerically available. Again, mixtures may be used satisfactorily.

The preferred detergents of this invention are anionic. They are generally the sulfated or sulfonated derivatives of aryl alkanes, alkanes, fatty alcohols, olefins, monoglycerides, polyoxyethylenes, or succinates.

Representative examples of each group are well known and commerically available.

The alkyl, aryl and alkyl sulfonates are especiaily preferred. These compounds are saturated hydrocarbons ordinarily having a total of from 8 to 20 carbon atoms. The alkyl aryl sulfonates contain a saturated hydrocarbon ring of from 5 to 8 carbon atoms which may be optionally hydroxyl or lower alkyl substituted, e.g., benzyl, toluyl, xylyi or hydroxyphenyl. In addition, the aryl group of the alyl aryl sulfonates will be substituted with at least one sulfonated, short chain, normal alkyl having about from three to six carbon atoms. The sulfonyl group is generally located at the opposite end of this alkyl group from the aryl radical. A suitable alkyl aryl sulfonate is a salt of dodecyl benzyl sulfonate.

The alkyl sulfonates are the most preferred anionic detergents. Again, the alkyl group is ordinarily normal, terminally sulfonated and contains from four to fifteen carbon atoms. The preferred detergent of this invention is sodium dodecyl sulfonate. The sodium dodecyl sulfonate is substantially free of sulfhydryl reducing agents such as mercaptoethanol.

The sulfonates or sulfate substituted detergents are generally employed as the alkali metal salts, preferably the sodium salts, while the cationic detergents are used as the halogen salts, e.g., chloride. The selection of ion is not critical so long as the resuiting salt remains a water soluble detergent.

Various species of anionic detergents may be combined for use in this invention, as may different cationic and different amphoteric detergents. In fact, commercially available detergents such as sodium dodecyl sulfonate are often found contaminated with longer or shorter chain sulfonates, but such mixtures are entirely satisfactory. However, detergents of the three classes of anionic, cationic and amphoteric detergents should not be mixed as their charged radicals will tend to neutralize one another with a commensurate loss in detergent activity.

The amount of detergent to be used to achieve the unexpectedly improved results of this information will vary depending upon the detergent selected and the sample to be determined. However, the amount should be sufficient to dissolve whatever coagulated Limulus protein one would expect to encounter with the test samples to be assayed under given reaction conditions. It is not necessary that so much detergent be added that the coagulated protein be converted into a true or molecular solution. Rather, a colloidal solution may be satisfactory so long as the protein aggregates are rendered substantially uniform by the detergent and there are no remaining macroscopic particles or clot fragments in the sample. The coagulated protein is considered to be dissolved when the coefficient of variation among the replicates is below about 15%.The preferred average diameter of the dissolved protein is from 0.1 to 0.4 microns. The quantitative amount of detergent to achieve the above goals is ordinarily at least 0.05% by weight in the reaction product of the Limulus lysate and endotoxin-containing sample. The preferred amount when using sodium dodecyl sulfonate is about 0.16%. the detergent concentration has no real upper limit in achieving the objects of this invention, for after a threshold is reached further additions of detergent yield no significant change in assay performance. For example, a concentration of 1.28% sodium dodecyl sulfonate in the reaction product can be used with an effect substantially equivalent to that achieved with 0.16%. Obviously, upper limits may be established by such considerations as detergent solubility, excessive foaming and nonspecific protein precipitation.However, these factors can be readily determined by the skilled artisan.

The suspending agent may be any emulsifying agent which has heretofore been employed to stabilize pharmaceutical or food emulsions and colloidal solutions. The suspending agent is generally a hydrophylic or hydrocolloid-forming organic macromolecular polymer having an average molecular weight from 1,000 to 20,000. Alternatively, and somewhat less desirably, the suspending agent may be hydrocolloid-forming mineral such as one of the many commercially available clays, e.g., bentonite.

Native or synthetic organic polymers may be used as the suspending agent. However, it is preferred to use synthetic polymers as their supply is reliable and they may be secured in uniform lots. Suitable polymeric suspending agents are carboxymethyl cellulose, polythylene glycol, methylcellulose, polyvinyl-pyrrolidone, polyvinyl alcohol, hydroxyethyl cellulose and carboxypolymethylene, with their alkali metal salts being most desirable. Suspending agent mixtures may be used. The preferred suspending agent is sodium carboxymethyl cellulose.

The concentration of suspending agent needed in the reaction product will be a function of the amount of coagulated Limulus protein expected, the effectiveness of the detergent in producing a uniform molecular or colloidal solution of coagulated protein, the molecular weight of the detergent, the presence of native suspending agents in the test sample and the nature of the suspending agent. Hence, limited routine experimentation may be necessary to determine the optimum concentration for the desired blank and standard curve slope. Generally, a satisfactory amount of suspending agent will yield a blank of less tham about 15% relative light scatter. Suitable concentrations in the reaction product ordinarily should exceed about 0.01% by weight, but additional advantage is rarely obtained with a concentration greater than about 0.3% by weight.The preferred concentration is about 0.06% by weight in the reaction product.

It should be understood that the properties of the detergent and the suspending agent will often overlap to a degree. For example, as the molecular weight of the detergent increases, emulsifying or colloidal suspending properties will appear. Similarly, either the introduction of charged moieties into the macromolecular suspending agent or increases in charge density will affect the degree of detergent exhibited by the agent. It is preferred to employ detergents and surface active agents that carry as little of the properties of the other as possible. However, ionic, surface active, suspending agents may be used, which substantially merge the characteristics of the detergent and surface active agent into a single polymer class.

Generally, these bifunctional polymers are sulfonated or sulfonated hydrocolloid-forming polymers such as cellulose sulfate, sulfonated polyoxyethylene or polystyrolsulfonate having a molecular weight of from 500 to about 5000. A sufficient amount of these agents must be used to dissolve coagulated Limulus protein, including clotted protein, as described above in connection with monofunctional detergent. Again, optimal amounts can be determined by routine experimentation directed at achieving the performance desired for the detergent and suspending agent when supplied as discrete compounds. Usually the amount employed should be greater than 0.05% by weight in the final reaction mixture, with the upper limit being largely a matter of discretion.

The ionic, surface active suspending agent may be supplied as a single aqueous solution containing bifunctional polymer or a mixture of detergent and suspending agent. However, the assay is more versatile if separate aqueous solutions of detergent and suspending agent are contacted with the reaction product separately. Since the detergent is preferably selected for its ability to stop the Limulus protein coagulation reaction it should be contacted with the reaction product only after the reaction has proceeded to the desired stage of completion. On the other hand the suspending agent may be added at any point in the coagulation reaction, either before or after addition of detergent. If the suspending agent is added at the commencement of the reaction it is convenient to supply it as the magnesium or manganese salts. These ions are Limulus coagulation enzyme cofactors and their addition in this form can save a pipetting step. Nonetheless, it is preferred to add the suspending agent after the detergent has reached equilibrium with the coagulated Limulus protein. Addition of detergent may be optionally accompanied by the mechanical disruption of any clotted protein which has formed.

The solutions of detergent and suspending agent are ordinarily water solutions at a pH near neutrality which were passed through 0.45Fm and 1.2cm filters, respectively, before use to remove particulate contaminants that could interfere in the nephelometric determination. It is preferred that the solutions contain alkali metal ions, either by dissociation of detergent and suspending agent salts or by addition of extraneous ions in the form of inorganic salts.

The concentrations of detergent and suspending agent in the reagent may vary considerably as a matter of choice. Since it is disclosed above that comparatively low concentrations of detergent and suspending agent are effective, their concentrations in the added reagent may also be low. Solutions containing detergent from 0.5% to 20% by weight and suspending agent from 0.01%to3% by weight are satisfactory. For example, stock solutions of 1% sodium dodecyl sulfonate and 0.08% sodium carboxymethyl cellulose are sufficiently versatile to be used under most conditions.When using the small sample aliquots, e.g., 0.05 mi, made possible by the practice of this invention it is often necessary to increase the volume of the reaction product before it can be analyzed in commerical nephelometers, which generally require at least one mi. of sample. Thus dilute solutions of detergent and suspending agent perform the combined functions of increasing reaction product volume, dissolving coagulated Limulus protein, stabilizing the colloidal solution of the coagulated Limulus protein and quenching the endotoxinactivated Limulus enzyme responsible for the coagulation.

The foregoing improvements in the nephelometric determination of endotoxin are facilitated by the use of selected nephelometer features. For example, it is preferred that the incident or sampie-illuminating light source of the nephelometer be substantially monochromatic. Ideally, a laser light source can be used for this purpose. The light source should be of a wavelength greater than about 500 mu, in the green to red portion of the visible spectrum, rather than in the blue range used heretofore (Levin et al., supra). A preferred wavelength is 632.8nm. Finally, the photodetector should be mounted to detect forward light scatter, in the order of 100 to 800 and preferably about 30", rather than the 90" scatter conventionally measured.Forward light scatter is defined in terms of the angle distended between the scattered light and the incident beam measured on the side of the sample opposite the light source. A nephelometer having these features is commercially available from the Hyland Diagnostics Division of Travenol Laboratories, Inc., Costa Mesa, California.

While this invention has been described in terms of improvements to nephelometric procedures for the assay of endotoxin it is apparent that the improved replicate reproducibility achieved by the use of an ionic, surface active agent is of advantage in all Limulus endotoxin optical methods wherein light which is scattered or absorbed by the protein particles is detected. This includes not only nepheiometric but turbidimetric or absorbance assays as well, e.g. determination of absorbance at 360nm as measured opposite the Incident light source illuminating the sample. For the purposes of this invention optical methods are defined as nephelometric, turbidimetric or absorbance procedures.

The following specific examples are intended as illustrations but not limitations of the scope of the present invention.

Example 1 This example demonstrates the preparation of a standard curve in a representative nephelometric assay for endotoxin, but without the use of detergent or suspending agent.

Lyophilized Limuluspolyphemus blood cell lysate in pyrogenfree test tubes was prepared following the method of British Patent 1,499,846. The lyophilized lysate was reconstituted by dissolution in 3.0 ml of 0.5 M MgC12. 0.1 ml aliquots of the lysate solution were then withdrawn and added to five sets of pyrogen-free test tubes, each set containing five tubes. 0.1 ml aliquots of 0, 12, 25, 50 and 100 pg/ml water solutions of E. coli 055.B5 endotoxin (Difco Laboratories) were added to the tubes to make up a set of 0, 12,25,50 and 100 pg/ml standards in five replicates each. The mixtures were incubated in a 37"C water bath for 60 minutes. Then 1 ml of 0.9% NaCI was added and any clotted protein mechanically disrupted by rapid vortexing of each tube.The total light scattering of the suspension in each tube was measured by a Hyland PDQ nephelometer having a laser light source of 632.8 mu. The total elapsed time between adding endotoxin and conducting the nephelometric determination was approximately equal for each tube. The results are shown in curve A of Figure 1 wherein the percent relative light scatter (%RLS) is plotted against concentration. The highest and lowest value secured with the five replicates at each endotoxin concentration were plotted and then joined with a bar. As can be seen from this curve, while a simple, unaided nephelometric endotoxin determination using the limulus system can be useful it nonetheless suffers from a number of defects. These include a very high blank or control, considerable replicate sample variation and overall curve nonlinearity.The elevated blank and curve nonlinearity adversely affect the assay sensitivity while the replicate sample variation and or nonrepreducibility introduces error into the system.

Example 2 The method of Example 1 was repeated except that the volumes of sample and lysate solution were reduced to 0.05 ml each and 1 ml of 0.08% sodium carboxymethylcellulose (CMC) suspending agent was added in place of the 0.9% NaCI solution. The sodium carboxymethylcellulose suspending agent was made by taking up commerically available CMC in saline to a concentration of 0.08% by weight and the solution passed through a 1.2 um Millipore filter disc. The results are shown in curve B of Figure 1. As can be seen, the curve nonlinearity and replicate sample nonreproducibilitystill remain serious problems.

Example 3 The method of Example 2 was repeated except that 0.2 ml of a 1.0% water solution of sodium dodecyl sulfate (SDS) was added in place of the 0.08% CMC solution. As can be seen from the results in this Example, graphed in curve C of Figure 1 the detergent solution considerably reduced the replicate sample variation compared to assays either with or without CMC. However, nonlinearity of the standard curve persists when SDS is used alone.

Example 4 This example demonstrates the suprising results obtained by use of both a detergent and a suspending agent in the nephelometric endotoxin assay. SDS was added as in Example 3, but within about 3 minutes thereafter 1 ml of the 0.08% CMC reagent was thoroughly mixed with the reaction mixture and the product assayed nephelometrically as above. The results are shown in curve D of Figure 1. Surprisingly, it was found that the curve exhibits excellent linearity when compared to the results secured with SDS or CMC alone. This phenomenon demonstrates ciassic synergism because the linearity is clearly not an additive function of the effects of SDS and CMC. Further, while SDS used alone does reduce replicate variation, at elevated endotoxin concentrations CMC further reduces this variation.This is particularly surprising because the use of CMC alone demonstrates either no effect or an adverse effect on replicate variation at these endotoxin levels.

Example 5 0.1 ml of test sample was added to 0.05 ml of Limulus lysate that has been reconstituted with 5 ml of 0.05M Tris buffered 0.23% MgC12 to give a pH of 7.4 and an ionic strength of 0.07 moles/liter. This reaction mixture was incubated for 30 minutes at 370C, after which 0.2 ml of 1.0% SDS and 1 ml of 0.8% CMC were sequentially added. The mixture was then vortexed and read in the nephelometer described in Example 1 which was standardized with a known ednotoxin. Table 1 represents typical data obtained when known endotoxin concentrations were assayed using this optimum nephelometric system. As can be observed from the data obtained in three consecutive days, the assays are highly reproducible with an average slope of 1.55 and a coefficient of variation of 8.6% at 50 pg endotoxin/ml. Consistent data of this nature indicate that the nephelometric LAL assay system is a substantial improvement over existing test methods for the quantitation of endotoxin.

TABLE 1 Endotoxin Conc (pg/ml) Sample 0 25 50 100 Slope R2* Day 1 1 1.1 17.9 68.3 153.2 1.583 0.9799 2 4.7 25.6 72.7 150.5 1.505 0.9882 3 0.4 18.2 69.5 157.3 1.631 0.9805 Day 2 1 0.7 25.7 75.0 149.4 1.528 0.9921 2 0.0 26.1 79.4 154.5 1.590 0.9912 3 0.1 26.8 75.2 149.3 1.529 0.9938 Day3 1 1.7 31.1 80.2 151.8 1.533 0.9948 2 1.3 33.2 86.4 150.2 1.519 0.9898 3 0.0 29.7 87.3 151.14 1.555 0.9860 Mean 26.0 77.1 151.9 1.55 0.9879 Standard Deviation 5.2 6.7 2.6 0.04 0.0055 Coeff. of Variation 20% 8.6% 1.7% 2.5% 0.55% *R2 = correlation coefficient indicating the linearity of the assay (1.0 suggests an ideal line).

Claims

1. An optical method for determining endotoxin wherein a sample thought to contain endotoxin is contacted with an endotoxin-coagulable protein from Limulus amoebocytes to form a reaction product containing coagulated protein, wherein the reaction product is contacted with an ionic surface-active suspending agent.

2. A method according to Claim 1 wherein the surface-active suspending agent is macromolecular detergent employed in an amount sufficient to dissolve said coagulated protein.

3. A method according to Claim 2 wherein the macromolecular detergent is an alkali metal salt of polystyrol-sulfonate, cellulose sulfate or sulfonated polyoxyethylene.

4. A method according to Claim 2 wherein the macromolecular detergent has a molecular weight of from 500 to 5,000.

5. An optical method for determining endotoxin wherein a sample thought to contain endotoxin is contacted with an endotoxin-coagulable protein from Limulus amoebocytes to form a reaction product containing coagulates protein, wherein the reaction product is contacted with an ionic detergent and a suspending agent.

6. A method according to Claim 5 wherein the detergent and suspending agent are mixed together and then contacted with the product.

7. A method according to Claim 6 wherein the detergent has an HLB of greater than 30.

8. A method according to Claim 5 wherein the detergent is an alkyl or aryl sulfonate.

9. A method according to Claim 8 wherein the detergent is a sulfated or sulfonated fatty alcohol, olefin, monoglyceride or succinate.

10. A method according to Claim 8 wherein the detergent is sodium dodecyl sulfonate.

11. A method according to Claim 5 wherein the detergent is cationic.

12. A method according to Claim 11 wherein the detergent is a quaternary ammonium salt.

13. A method according to Claim 5 wherein the detergent is amphoteric.

14. A method according to Claim 5 wherein the detergent has a molecular weight of from 150 to 450.

15. A method according to any one of Claims 5 to 14 wherein a sufficient amount of detergent is contacted with said product to yield at least 0.05% of detergent by weight in said product.

16. The method of Claim 10 wherein a sufficient amount of sodium dodecyl sulfonate is contacted with said product to yield about 0.16% by weight of detergent by weight in said product.

17. A method according to any one of Claims 5 to 15 wherein a sufficient amount of detergent is employed to dissolve the coagulated protein.

18. A method according to any one of Claims 5 to 17 wherein the suspending agent is a hydrocolloid-forming mineral.

19. A method according to Claim 18 wherein the suspending agent is bentonite.

20. A method according to any one of Claims 5 to 17 wherein the suspending agent is a hydrocolloid-forming organic polymer.

21. A method according to Claim 20 wherein the suspending agent is alkali metal carboxymethyl cellulose, polyethylene glycol, methylcellulose, polyvinylpyrrolidine, polyvinyl alcohol, hydroxyethyl cellulose or carboxypolymethylene.

22. A method according to Claim 21 wherein the suspending agent is sodium carboxymethyl cellulose.

23. A method according to Claim 20 wherein the polymer has an average molecular weight of between 1,000 and 20,000.

24. A method according to any one of Claims 5 to 3 wherein an aqueous solution of the detergent is contacted with the reaction product.

25. A method according to Claim 24 wherein an aqueous solution of the suspending agent is contacted with the reaction product.

26. A method according to Claim 25 wherein the detergent solution is contacted with the reaction product prior to contact of the reaction product with suspending agent.

27. A method according to Claim 24 wherein the detergent is sodium dodecyl sulfonate and the aqueous solution is substantially free of sulfhydryl reducing agent.

28. A method according to any one of Claims 5 to 27 wherein a sufficient amount of suspending agent is contacted with said product to yield at least 0.01% of suspending agent by weight in said product.

29. A method according to Claim 22 wherein a sufficient amount of sodium carboxymethyl cellulose is contacted with said product to yield about 0.06% by weight of sodium carboxymethyl cellulose in said product.

30. An optical method for determining endotoxin wherein a sample thought to contain endotoxin is contacted with an endotoxin-coagulable protein from Limulus amoebocytes to form a reaction product containing coagulated protein, including the steps of mechanically disrupting said coagulated protein in said product and contacting said disrupted protein with a suspending agent.

31. An optical method for determining endotoxin wherein a sample thought to contain endotoxin is contacted with an endotoxin-coagulable protein from Limulus amoebocytes to form a reaction product containing coagulated protein, including the steps of determining the forward light scattering by said product of a substantially monochromatic incident light wavelength of greater than 500 mu.

32. An optical method for determining endotoxin wherein the reaction product is contacted with a sufficient amount of ionic detergent to dissolve said coagulated protein.