Treatment of multiple hereditary osteochondromas of the forearm in children: a study of surgical procedures.

We have evaluated the clinical outcomes of simple excision, ulnar lengthening and the Sauvé-Kapandji procedure in the treatment of deformities of the forearm in patients with multiple hereditary osteochondromas. The medical records of 29 patients (33 forearms) were reviewed; 22 patients (22 forearms) underwent simple excision (four with ulnar lengthening) and seven the Sauvé-Kapandji procedure. Simple excision increased the mean supination of the forearm from 63.2 degrees to 75.0 degrees (p = 0.049). Ulnar lengthening did not significantly affect the clinical outcome. The Sauvé-Kapandji technique improved the mean pronation from 33.6 degrees to 55.0 degrees (p = 0.047) and supination from 70.0 degrees to 81.4 degrees (p = 0.045). Simple excision may improve the range of movement of the forearm but will not halt the progression of disease, particularly in younger patients. No discernable clinical or radiological improvement was noted with ulnar lengthening. The Sauvé-Kapandji procedure combined with simple excision of osteochondromas can improve stability of the wrist, movement of the forearm and the radiological appearance.

HPLC and HPCE analysis of microcystins RR, LR and YR present in cyanobacteria and water by using immunoaffinity extraction.

Microcystins, which have their origin in species of cyanobacteria present in freshwaters, have recently been found to be important contaminants of the aquatic environment at trace levels. HPLC and HPCE with UV detection have been applied in the determination of such toxic compounds. Immunoaffinity chromatography for the selective extraction and clean-up of microcystins has been successfully applied to different matrices. Simple protocols for unambiguous determination of these toxins are presented and the immunoaffinity clean-up is compared with conventionally used solid phase extraction procedures. The development and optimisation of an on-line preconcentration procedure based on field amplified sample stacking for the analysis of microcystins by HPCE in the micellar electrokinetic chromatography mode is also described, using borate buffer with the anionic surfactant SDS, as separation electrolyte. Results indicate that sub-nanogram/gram content of microcystins can be detected in water samples, while sub-microgram/gram concentrations can be determined in algae samples.

Survey of bottled drinking waters sold in Canada for chlorate, bromide, bromate, lead, cadmium and other trace elements.

Mineral, spring and other bottled drinking waters sold in Canada in the winter of 1995-96 were surveyed for chlorate, bromide, bromate, Cr(VI), Li, B, Al, Mn, Cu, Zn, Sr, Ba, Be, V, Cr, Co, Ni, As, Se, Mo, Ag, Cd, Sb, Tl, Pb, Na, K, Ca and Mg. Chlorate and bromide were determined by ion chromatography (IC) with conductivity detection, Cr(VI) by IC with colorimetric detection, bromate by solvent extraction and gas chromatography (GC), trace elements by inductively coupled plasma mass spectrometry (ICPMS), and Na, K, Ca and Mg by flame atomic absorption spectrometry (FAA). Most chemicals in the 199 samples analysed were well within national and international drinking water guidelines. World Health Organization and/or Canadian drinking water guidelines were exceeded for B (22 samples), Al (9), Cr (1), Mn (5), Ni (1), As (10), Se (24) and Pb (1). Bromate levels are reported for information purposes and are considered as the maximum concentrations in the samples. In three distilled water products, unexpectedly high concentrations of Cu (88-147 micro g l(-1)) and Ni (16-35 micro g l(-1)) were found, and a comparison of distilled and non-distilled waters from two of the brands suggested the likely cause to be contamination during the distillation process. Li concentration in one sample was at a therapeutic dose and could pose an overdose risk to individuals on Li medication.

Determination of microcystins in blue-green algae, fish and water using liquid chromatography with ultraviolet detection after sample clean-up employing immunoaffinity chromatography.

Anti-microcystin LR immunnoaffinity cartridges were evaluated for their ability to selectively remove microcystins from extracts of blue-green algae, fish and water samples for subsequent analysis by liquid chromatography with UV absorbance detection at 238 nm. Blue-green algae and fish samples were extracted with 75% methanol in water. A portion of the extract was diluted and passed through an immunoaffinity cartridge. Water samples were applied directly to the cartridge. The cartridge was rinsed with water and 25% methanol in water. The microcystins were eluted with 80% methanol in water containing 4% acetic acid. It was found that the cartridges were effective in isolating the microcystins from blue-green algae, fish and water samples, resulting in extracts that were clean enough to enable direct LC-UV detection down to approximately 0.03 microg/g in the blue-green algae and fish samples, and as low as 0.02 ng/ml for water samples. The cartridges were found to have a capacity of approximately 200 ng each for a mixture of microcystins RR, YR, LR and LA, or as much as 525-800 ng for individual compounds. Recoveries trough the complete analytical procedure ranged from 64 to 115% (all values) with an overall average of approximately 80% at spiking levels of 0.5-4.0 microg/g for the microcystins in blue-green algae. The average recoveries (n=8) from spiked (0.1-0.5 microg/g) fish samples were 73% for RR, 79% for YR, 81% for LR and 77% for LA, while from the spiked (2.0-0.04 ng/g) tap and river water samples (n=6), recoveries were 78% for RR, 86% for YR, 94% for LR and 89% for LA.

Comparison of liquid chromatography/mass spectrometry, ELISA, and phosphatase assay for the determination of microcystins in blue-green algae products.

More than 100 samples of blue-green algae products (consisting of Aphanizomenon, Spirulina, and unidentified blue-green algae) in the form of pills, capsules, and powders were collected from retail outlets from across Canada. The samples were extracted with 75% methanol in water and centrifuged to remove solids. Aliquots of the extracts along with spiked blank sample extracts were sent to each participating laboratory and independently analyzed for microcystins by enzyme-linked immunosorbent assay (ELISA), protein phosphatase inhibition assay, and by liquid chromatography-tandem mass spectrometry (LC-MS/MS) after sample cleanup using C18 solid-phase extraction. The results obtained by ELISA and LC-MS/MS agreed very well over a concentration range of about 0.5-35 microg/g. The colorimetric phosphatase results generally agreed with the other 2 methods. While the 2 biochemical assays measured total microcystin content compared with a standard of microcystin LR, the LC-MS/MS method measured specific microcystins (LA, LR, RR, YR) using external standards of these for identification and quantitation. Microcystin LR was found in all positive samples by LC-MS/MS. Microcystin LA was the only other microcystin found in the samples analyzed. These 2 microcystins represent essentially all the microcystins that were present in the extracts. Otherwise, the LC-MS/MS results would have been significantly lower than the results of the biochemical assays had other unknown microcystins been present.

Quantitative determination of paralytic shellfish poisoning toxins in shellfish by using prechromatographic oxidation and liquid chromatography with fluorescence detection.

The prechromatographic oxidation LC method developed by Lawrence [J. Assoc. Off. Anal. Chem. 74, 404-409(1991)] for the determination of paralytic shellfish poisoning (PSP) toxins has been tested for the quantitative determination of PSP toxins in shellfish. All aspects of the method were studied and modified as necessary to improve its performance for routine regulatory purposes. The chromatographic conditions were changed to shorten analysis time. The oxidation reaction was tested for repeatability and the influence of the sample matrix on quantitation. An important part of the study was to quantitatively evaluate an ion exchange (-COOH) cleanup step using disposable solid-phase extraction cartridges that separated the PSP toxins into 3 distinct groups for quantitation, namely the C toxins, the GTX toxins, and the saxitoxin group. The cleanup step was very simple and used increasing concentrations of aqueous NaCl for elution of the toxins. The C toxins were not retained by the cartridges and thus were eluted unretained with water. The GTX toxins (GTX1 to GTX6 as well as dcGTX2 and dcGTX3) eluted from the cartridges with 0.05M NaCl while the saxitoxin group (saxitoxin, neosaxitoxin, and dcsaxitoxin) required 0.3M NaCl for elution. Each fraction was analyzed by LC after oxidation with periodate or peroxide. All of the compounds could be separated and quantitatively determined in spiked samples of mussels, clams, and oysters. The nonhydroxylated toxins could be quantitated at concentrations as low as about 0.02 microg/g (2 micro/100 g) of tissue while the hydroxylated toxins could be quantitated at concentrations as low as about 0.1 microg/g (10 microg/100 g). Average recoveries of the toxins through the complete cleanup procedure were 85% or greater for spiked extracts of oysters and clams and greater than 73% for mussels.

Effect of pH on the oxidation of paralytic shellfish poisoning toxins for analysis by liquid chromatography.

The effect of pH on the oxidation of individual PSP toxins using both periodate and peroxide oxidations was studied. It was found that the optimum pH for individual toxins varied considerably. For periodate oxidations, pH 8.2 produced the maximum yield of fluorescent products for neosaxitoxin and GTX1/GTX4 while the non-hydroxylated toxins (saxitoxin, GTX2/GTX3, decarbamoyl saxitoxin, GTX5) showed optimum pHs from about pH 10-11.5. Neosaxitoxin and GTX1/GTX4 did not produce significant fluorescent oxidation products with peroxide oxidation at any of the pHs studied (pH 8.2-12.8). The non-hydroxylated toxins all showed optimum pHs above pH 12 with peroxide oxidation. Yields of fluorescent products of these toxins decreased substantially at pHs below pH 12. Neosaxitoxin and GTX1/GTX4 each produced three product peaks at pH 8.2 with periodate oxidation. There was no pH where these toxins produced predominantly a single oxidation product. Decarbamoyl saxitoxin always produced two oxidation products with both oxidation reactions at the pHs studied. However, the relative yields of the products changed with pH. At low pH the second eluting product predominated, while at higher pH values the first eluting product predominated. This pattern was observed for both oxidation reactions. The other non-hydroxylated toxins produced mainly single unique products with both oxidation reactions over the pH range studied. No single pH was found optimum for the oxidation of both hydroxylated and non-hydroxylated toxins without a significant compromise in yield of oxidation products. This has implications for the post column oxidation liquid chromatographic methods, since small changes in pH of the post column oxidant can both positively and negatively affect the yields of oxidation products of toxin mixtures leading to increased error in the subsequent quantitation of these compounds.

Evaluation of silica- and sepharose-based immunoaffinity sorbents for sample cleanup in determination of fumonisins B1 and B2 in corn products.

Anti-fumonisin B1 polyclonal antibodies were isolated from the serum of rabbits, immobilized onto the surface of glutaraldehyde-activated silica or Sepharose CL-4B particles, and placed into empty small plastic solid-phase extraction cartridges. The immobilized antibodies were evaluated for their ability to retain fumonisin B1 and fumonisin B2. Cartridge capacity and elution conditions were determined, and the results were compared to those obtained with a commercially available cartridge. The cartridges, which were tested for their effectiveness to isolate the fumonisins from extracts of corn flour and nacho chips, detected fumonisins down to levels of about 20 ng/g. However, additional cleanup was required for detection at lower concentrations. With the use of a strong anion-exchange cartridge as a preliminary cleanup before immunoaffinity chromatography, the detection limit reached 2-5 ng/g in the products tested. The silica sorbent material exhibited strong interactions with the fumonisins, requiring acidified ethanol-water mixtures for elution and resulting in an additional degree of selectivity in isolating fumonisins from sample extracts. The silica-based immunoaffinity cartridges were successfully reused more than 10 times; the Sepharose-based cartridges were less robust. Liquid chromatography with fluorescence detection was used after prechromatographic derivatization with o-phthaldialdehyde-mercaptoethanol.

Effect of temperature and solvent composition on extraction of fumonisins B1 and B2 from corn products.

Fumonisins B1 and B2 were extracted from naturally contaminated corn products by using different extraction solvent compositions (methanol-water, acetonitrile-methanol-water, ethanol-water, and 100% water) and a range of temperatures from ambient to 150 degrees C. Ground samples of several corn products and 1 rice sample were mixed with an adsorbent material (Hydromatrix), and the fumonisins were extracted in 2 sequential 5 min static extractions at various temperatures. The combined extracts were cleaned up and analyzed by reversed-phase liquid chromatography with fluorescence detection after o-phthaldialdehyde-mercaptoethanol derivatization. The results showed a clear influence of temperature and solvent composition on recovery of fumonisins from some matrixes. With acetonitrile-methanol-water (1 + 1 + 2) the quantity of fumonisins extracted from naturally contaminated taco shells almost tripled in going from 23 degrees to 80 degrees C, and increased by another 30% when ethanol-water (3 + 7) was used as extraction solvent at 80 degrees C. Similar results were obtained with nacho chips. These effects were less pronounced with cornmeal, and small differences due to temperature and solvent composition were observed for corn flakes and rice. The ethanol-water extraction solvent combinations were specifically evaluated in an effort to use the cheapest, least toxic, and most environmentally friendly solvents for organic residue analysis. At 80 degrees C, ethanol-water combinations performed equally or better than methanol-water (8 + 2) or acetonitrile-methanol-water (1 + 1 + 2), combinations which are commonly used for fumonisin extractions. Even 100% water was successful for extracting fumonisins from the products, except for rice. However, increased amounts of water created technical problems and required an increased amount of Hydromatrix in the samples prior to extraction.

Congenital radius deficiency: radiographic outcome and survivorship analysis.

The experience with congenital radius deficiency, or radial hemimelia, at the Shriners' Hospital for Children, Los Angeles Unit, was reviewed. A cohort of 29 limbs in 23 patients was identified with an average follow-up period of 50 months. Radiographic parameters were assessed using the hand-forearm angle, hand-forearm position, and ulnar bow. We compared radialization to modified centralization, assessed the efficacy of ulnar osteotomy, and assessed the effect of age, preoperative deformity, ulnar osteotomy, and Bayne's type on the final result. Revisions were noted and a survivorship analysis performed. The cohort had statistically significant correction of hand-forearm angle and hand-forearm position. Radialization was similar to modified centralization in the final outcome. Ulnar osteotomy was an efficacious way to correct ulnar deformity. Age, preoperative deformity, performance of an ulnar osteotomy, and Bayne's type did not affect the final wrist position. Survivorship analysis was performed using revision as the end point, with a survivorship rate at 5 years of 67%. Significant risk factors for revision included radial or positive hand-forearm angle and young age at the time of the index procedure. There was a suggestion that small postoperative hand-forearm position, or radial translation, increased the risk of revision. Preoperative deformity, performance of an ulnar osteotomy, and Bayne's type did not affect the risk of revision. These data offer support for the hypothesis that a more ulnar translation and an ulnar angulation of the wrist is a means of reducing the radial lever arm and thus the incidence of deformity recurrence and need for revision.

Determination of benzidine in the food colours tartrazine and sunset yellow FCF, by reduction and derivatization followed by high-performance liquid chromatography.

Free and bound benzidine, a non-sulphonated aromatic amine (NSAA), were determined in the food colours tartrazine and sunset yellow FCF. Bound amines were released by reducing with sodium dithionite, then total NSAAs were extracted into chloroform, transferred to aqueous acid solution and diazotized with sodium nitrite before coupling with 2-naphthol-3,6-disulphonic acid, disodium salt (R-salt). Coloured benzidine and aniline derivatives (BZDRS and ANRS) were analysed using reversed-phase ion pair high-performance liquid chromatography (HPLC) and an absorbance detector set at 548 nm. Levels of total benzidine were similar to those reported in studies conducted in the USA, and ranged from < 5 to 270 ng/g. Total aniline was also determined (0.2-188 micrograms/g). Recoveries of benzidine with this method were found to be lower than expected (average ca 46%), but were reproducible. Detection limits were 15-20 ng BZDRS/g (3-4 ng benzidine/g). Mass spectrometry (LC-MS) was evaluated for identifying and determining purity of the standards, but difficulties were encountered when the methodology was applied to commercial samples.

Determination of clenbuterol in beef liver and muscle tissue using immunoaffinity chromatographic cleanup and liquid chromatography with ultraviolet absorbance detection.

Clenbuterol, a beta-agonist, was determined in samples of beef liver and muscle. The method employed an acidic aqueous extraction followed by protein precipitation. The supernatant liquid was passed through a weak cation-exchange cartridge and then through a commercially available immunoaffinity cartridge. Clenbuterol was eluted from the immunoaffinity cartridge with 80% ethanol in water. The eluate was concentrated and analysed directly by reversed-phase liquid chromatography using gradient elution and UV detection at 245 nm. Detection limits were estimated to be 0.3 ng g-1 clenbuterol. A single immunoaffinity cartridge was used for ten sample extracts with no significant loss in capacity. No organic solvents other than ethanol and methanol were employed in the procedure. Recoveries of clenbuterol from samples of beef liver and muscle spiked at 2 and 5 ng g-1 carried through the entire procedure were 63 +/- 11% (range, 53-74%) compared to pure standards. Absolute recoveries of pure standards (30 ng clenbuterol) carried through the same analytical steps were 70 +/- 5% (n = 6), the losses being primarily due to the ion-exchange step.

Development of a manganese dioxide solid-phase reactor for oxidation of toxins associated with paralytic shellfish poisoning.

Using manganese dioxide as a solid-phase oxidant, a post-column HPLC reactor for paralytic shellfish poison toxins was constructed and evaluated. Operating parameters such as reaction temperature and pH, flow-rate, reactor column size, and MnO2 particle size were studied. Based on a 3:1 signal-to-noise ratio the limits of detection for the non-hydroxylated toxins were in the sub-nanogram range, while the hydroxylated toxins were ten times less detectable. The limit of detection for saxitoxin was about 0.1 ng per injection. The repeatability of replicate injections was < or = 10% (relative standard deviation). The system was used to analyze extracts of shellfish and plankton yielding results that were in agreement with those obtained by established analytical methods. Extracts of shellfish contaminated at the regulatory limit of 0.8 microgram/g were successfully analyzed.

Determination of decarbamoyl saxitoxin and its analogues in shellfish by prechromatographic oxidation and liquid chromatography with fluorescence detection.

Oxidation and chromatographic conditions for detecting the decarbamoyl analogues of several paralytic shellfish poison (PSP) toxins were studied. Prechromatographic oxidation with periodate or hydrogen peroxide under slightly alkaline conditions was used as previously reported for the parent PSP toxins. Both periodate and hydrogen peroxide oxidations produced 2 fluorescent products separable by liquid chromatography for each decarbamoyl (dc) toxin (dc-saxitoxin, dc-neosaxatoxin and dc-gonyautoxins 2 and 3). Decarbamoyl saxitoxin produced the same 2 products as did dc-neosaxitoxin but in different ratios. One of these products was the same as the one obtained with neosaxitoxin after periodate oxidation. Decarbamoyl gonyautoxins 2 and 3 (together) produced 2 products, one of which was the same as the major product obtained with gonyautoxins 1 and 4 (together) after periodate oxidation. Decarbamoyl gonyautoxins 1 and 4 were not available for study. The method was used to detect dc-saxitoxin and dc-gonyautoxins 2 and 3 in shellfish extracts.

High-performance liquid chromatographic separation of carminic acid, alpha- and beta-bixin, and alpha- and beta-norbixin, and the determination of carminic acid in foods.

During a study of natural food colours, a simple and reliable high-performance liquid chromatography system was developed for use with cochineal and annato. An isocratic mobile phase, consisting of methanol and 6% aqueous acetic acid, resolved bixin and norbixin, while a gradient system was used to separate carminic acid and the annato compounds. The carminic acid contents of cochineal extract, carmine and carmine hydrosoluble were determined using an isocratic mobile phase (40:60, v/v). The detection limit for carminic acid in the various products was approximately 100 ng/g. Carminic acid was determined quantitatively in fruit beverages, yogurt and candies. It was demonstrated that, because of decomposition, carminic acid was not suitable for use in candies when manufacturing temperatures above 100 degrees C were required. Most membrane filters are not suitable for use with cochineal solutions, but a cellulose membrane filter did not adsorb carminic acid and was used successfully to remove impurities from water-based cochineal products and food extracts containing carminic acid.

Liquid chromatographic determination of okadaic acid and dinophysistoxin-1 in shellfish after derivatization with 9-chloromethylanthracene.

The reagent 9-chloromethylanthracene was evaluated for derivatization of the diarrhetic shellfish poisons, okadaic acid and dinophysistoxin-1 (DTX-1), to form fluorescent products separable by liquid chromatography. The toxins were reacted with the reagent in acetonitrile in the presence of tetramethylammonium hydroxide for 1 h at 90 degrees C. The products were purified by using two silica solid-phase extraction cartridges before being determined by reversed-phase liquid chromatography with fluorescence detection. The results are comparable to those obtained using 9-anthryldiazomethane (ADAM) for okadaic acid and DTX-1 in mussel tissue. Detection limits were estimated to be about 70-100 ng/g hepatopancreas (equivalent to 12-20 ng/g whole tissue) for each toxin.

Ion chromatographic determination of cyanide released from flaxseed under autohydrolysis conditions.

Flaxseed is increasingly being used in some food products because of its high content of alpha-linolenic acid and dietary fibre. However, flaxseed contains cyanogenic glycosides which release toxic hydrogen cyanide in the presence of water (autohydrolysis). A method for estimation of cyanide in flaxseed under these conditions is described. The determination is carried out by homogenizing the sample with water, letting it stand, filtering it through a membrane and then injecting the filtrate into an HPLC system consisting of an anion exchange column and an electrochemical (amperometric, oxidation) detector. The homogenate is analysed at various intervals until a maximum value of cyanide is observed. The cyanide content of ten cultivars of flaxseed, when analysed by this method, was found to range from 124 to 196 micrograms/g. The release of cyanide showed a maximum at about 3 h of hydrolysis. Virtually no cyanide was detected on boiling the homogenate or the flaxseed before determination.

Evaluation of a postcolumn electrochemical reactor for oxidation of paralytic shellfish poison toxins.

A liquid chromatographic method using a postcolumn electrochemical reactor that oxidizes paralytic shellfish poison toxins to fluorescent derivatives has been developed. Several experimental parameters, including pH and oxidation potential, were investigated. For nonhydroxylated toxins, the sensitivity improved with increasing pH and voltage. At optimum operating conditions, the sensitivity for saxitoxin and gonyautoxins 2 and 3 was an order of magnitude greater than that for neosaxitoxin and B1 and 2 orders of magnitude greater than that for B2. The limit of detection for saxitoxin was 0.10 ng (signal-to-noise ratio, 3:1). Electrochemical oxidation products were similar to those formed in the prechromatographic periodate oxidation method. Shellfish and plankton extracts were analyzed with the electrochemical system, and results agreed well with those obtained with established methods. Shellfish samples contaminated at the regulatory limit of 0.8 microgram/g were readily analyzed by the method.

Evaluation of prechromatographic oxidation for liquid chromatographic determination of paralytic shellfish poisons in shellfish.

A liquid chromatographic (LC) method employing prechromatographic oxidation for the determination of paralytic shellfish poison (PSP) toxins was evaluated. A number of changes to an earlier method resulted in improved separation and quantitation of most PSP analogues. Modification of the periodate oxidation reaction for the N-hydroxy-containing toxins led to improved sensitivity and stability of the products, enabling automated overnight analyses. These changes also enabled quantitation of gonyautoxins 1 and 4 (together) in the presence of gonyautoxins 2 and 3. Decarbamoylsaxitoxin can be identified and quantitated after peroxide oxidation. A cleanup step using a strong-anion-exchange column removed the C toxins and B-2 from the extracts and enabled a more accurate quantitation of gonyautoxins 1 and 4 and neosaxitoxin. Chiral chromatography, employing a reversed-phase column and chiral mobile-phase additives (copper-proline complex), was briefly evaluated for the separation of the oxidation products of the isomer pairs, gonyautoxins 1 and 4 and gonyautoxins 2 and 3. A comparison of the method with the mouse bioassay for the determination of PSP in lobster hepatopancreas (58 samples) showed a reasonable correlation (0.90) over a concentration range of 40-500 micrograms/100 g (saxitoxin equivalents), although the LC results were consistently higher than the mouse bioassay values by about 40%.

Determination of annatto in high-fat dairy products, margarine and hard candy by solvent extraction followed by high-performance liquid chromatography.

Utilizing solvents such as ethanolic aqueous ammonia, petroleum ether, hexane and chloroform, annatto components alpha- and beta-norbixin and alpha- and beta-bixin were extracted from cheese, butter, margarine and hard candy. After transferring the extract into a solution of aqueous acetic acid in methanol, bixin and norbixin were determined quantitatively using high-performance liquid chromatography (HPLC) and an absorbance detector set at 500 nm. Recovery of norbixin from spiked cheese samples averaged 92.6% over a range of 1 to 110 micrograms/g. Commercial cheese samples were found to contain 1.1-68.8 microgram/g total norbixin, and two samples also contained 5.1-5.6 micrograms/g total bixin. Samples of uncoloured butter were spiked with bixin and recovery averaged 93.2% over a range of 0.1 to 445 micrograms/g. Levels of 0.2 microgram/g total bixin and 0.91 microgram/g total norbixin were found in one commercial butter sample; the others contained trace levels of both compounds. Hard candies were prepared in the laboratory and recovery studies conducted. Recovery of norbixin averaged 88%.