Journal of Global Infectious Diseases

: 2018  |  Volume : 10  |  Issue : 4  |  Page : 206--211

Action of colloidal bismuth hydroxide gel and its commercial cream on enteropathogens and colonizers of the gastrointestinal tract

Tomás Subils, Cecilia Casabonne, Agustina González, Virginia Aquili, Claudia E Balagué 
 Department of Microbiology, Faculty of Biochemical and Pharmaceutical Sciences, National University of Rosario, Suipacha, Rosario, Santa Fe, Argentina

Correspondence Address:
Dr. Tomás Subils
Suipacha 531 S2002lRK, Rosario, Santa Fe


Background: Acute diarrheal diseases constitute a world public health problem because they are the second cause of death in children under 5 years of age. Colloidal bismuth hydroxide gel (CBHG) is an active ingredient in low-cost, antidiarrhetic drugs for oral use; it does not inhibit intestinal motility, and it features very low intestinal absorption of <1%. Materials and Methods: We analyzed the sensitivity by determining the minimum inhibitory concentration (MIC) and the minimum bactericidal concentration (MBC); the effect on bacterial growth by studying the specific growth velocity and the generation time in growth curves; and bacterial attachment by counting viable plaques, of enteropathogenic Escherichia coli, shigatoxigenic E. coli O157:H7, Klebsiella pneumoniae, Salmonella spp., and Shigella flexneri in the commercial cream (Chobet® bismuth cream with pectin [CBCHP]), its active ingredient (CBHG), and its excipients (E) separately. Results: CBCHP: MIC 6–10 mg/ml and MBC 7.5–15 mg/ml of bismuth; CBHG: MIC 6–10 mg/ml of bismuth. E: No inhibition was observed at the concentration studied in this study. At very low subinhibitory concentrations of CBCHP and CBHG, there was already evidence of a significant decrease in growth, which could not be recorded for E. CBCHP and CBHG presented an elevated capacity for bacterial displacement, significantly greater than E. Conclusions: We believed that the results obtained in this study are very promising from the treatment standpoint, as a possible treatment for cases of diagnosis or suspicion of bacterial gastroenteritis. The antimicrobial and attachment effects of CBCHP are exclusively due to its active ingredient CBHG; these effects are promoted in the presence of E.

How to cite this article:
Subils T, Casabonne C, González A, Aquili V, Balagué CE. Action of colloidal bismuth hydroxide gel and its commercial cream on enteropathogens and colonizers of the gastrointestinal tract.J Global Infect Dis 2018;10:206-211

How to cite this URL:
Subils T, Casabonne C, González A, Aquili V, Balagué CE. Action of colloidal bismuth hydroxide gel and its commercial cream on enteropathogens and colonizers of the gastrointestinal tract. J Global Infect Dis [serial online] 2018 [cited 2020 Apr 9 ];10:206-211
Available from:

Full Text


Acute diarrheal diseases constitute a world public health problem, especially in developing countries, and they are the second cause of death in children under 5 years of age. It is calculated that 1.5 billion episodes of acute diarrhea occur each year, causing the death of between 1.5 and 2 million children under 5 years of age.[1]

In Latin America, epidemiological data showed 13.5 million episodes of persistent diarrhea each year with 15% mortality. Enteric infection can be attributed to different infectious agents, and the epidemiology depends on the host, season of the year, and sanitation conditions.[2]

The reported rates of acute diarrhea in Argentina during the year 2010 were 3,034/100,000 inhabitants.[3] The bacterial etiological agents most often involved in infectious diarrheas include the bacteria Escherichia coli (E. coli), Salmonella, Shigella, Campylobacter, Yersinia, and Aeromonas.[4] In Argentina, according to data published in 2010, in 18.13% of 23,747 patients studied by stool culture, it was possible to identify more often: Shigella spp. (n = 2,728) and E. coli (n = 671) with predominance of Shigella flexneri and enteropathogenic E. coli (EPEC).[5],[6] Shigella flexneri is the predominant species in the world, mainly in developing countries.[7] The high incidence in these countries is due to the lack of drinking water, poor sanitation, malnutrition, and the high cost of antibiotic treatment.[8]

On the other hand, the colonization of the gastrointestinal tract after a primary infection by microorganisms such as β-lactamases and carbapenemase-producing Klebsiella pneumoniae (K. pneumoniae) is a risk factor for presenting a second infection, and furthermore, the carriers may serve as an important reservoir for dissemination of these microorganisms in health-care facilities.[9],[10]

The indiscriminate use of antibiotics in cases of diarrhea can increase the risk of adverse events, as in the case of Hemolytic Uremic Syndrome due to the shigatoxigenic E. coli (STEC),[11] or it can produce bacterial resistance.[12] The development and propagation of resistance to antibiotics are considered as a threat to world public health.[1]

Medications that contain bismuth have been used extensively in gastroenterology.[13],[14] Colloidal bismuth hydroxide is not salt, so it contains no associated radicals such as salicylate, which is responsible for the adverse effects and contraindications of bismuth subsalicylate.[15] Recently, studies have demonstrated that the use of colloidal bismuth hydroxide exhibits antibacterial activity, and also inhibits the activity of the pathogenic factor, like Shiga toxin in STEC.[14],[15],[16],[17],[18]

Many epidemiological studies have confirmed the efficacy of treatments with bismuth compounds for the prevention of traveler's diarrhea, particularly for the various virotypes of E. coli, Salmonella spp., and Shigella spp;[19],[20] for the treatment of acute diarrhea from rotavirus and enterotoxigenic E. coli in children;[21],[22],[23] and as a supplement in the treatment of gastric and duodenal ulcers caused by Helicobacter pylori,[24],[25] among other uses. Despite this vast and extensive background in the use of bismuth compounds for the prevention and treatment of diarrhea, there is still controversy regarding their complete mechanism of action. Some studies have indicated that bismuth compounds inhibit the intestinal secretion caused by toxins of Vibrio cholerae and enterotoxigenic E. coli[26] and decrease the cell invasion of enteroinvasive E. coli.[27] Furthermore, bismuth enhances the opsonophagocytosis of K. pneumoniae, thereby reducing the expression of the capsule,[28] and reversibly represses the expression of fimbriae in enteropathogenic and uropathogenic E. coli.[29] A study by Brogan et al. showed dithiol bismuth to be an inhibitor of the Rho protein of E. coli, an essential protein that controls the expression of several genes in Gram-negative bacteria. Thus, the inhibitory and antibacterial activity of bismuth compounds is believed to be the result of multiple mechanisms.[30]

The aim of this study was to seek potential therapeutic strategies for the treatment of bacterial enterophatogens and colonizers of the gastrointestinal tract by studying the effect of colloidal bismuth hydroxide gel (CBHG) and its commercial creman CBCHP in clinical and subclinical concentrations on the viability and on the removal of the most common enteropathogenic strains in our population.

The aim of this study was to seek potential therapeutic strategies for the treatment of bacterial enterophatogens and colonizers of the gastrointestinal tract by studying the effect of clinical and subclinical concentrations of CBHG and its commercial creman CBCHP on the viability and the removal of the most common enteropathogenic strains in our population.

 Materials and Methods

Bacterial strain and bismuth compounds

The strains of EPEC, STEC O157:H7, K. pneumoniae, Salmonella spp, and S. flexneri that are used in the present study were isolated at Provincial Hospital of Rosario and characterized phenotypically and genetically at our laboratory. In addition, we used the reference strain E. coli ATCC 25922 as a control.

We studied Chobet® bismuth cream with pectin (CBCHP) (SOUBEIRAN CHOBET, S. R. L., Buenos Aires, Argentina) because this is the only medicinal specialty with bismuth hydroxide gel (CBHG) as active ingredient with a concentration equivalent to 30 mg/ml of metallic bismuth. In addition, we recorded their excipients (E) whose composition consists of pectin, glycerin, methylparaben, propylparaben, essence of raspberries, vanillin, sodium saccharin, red amaranth, and water (SOUBEIRAN CHOBET, S. R. L., Buenos Aires, Argentina).

Determination of sensitivity to bismuth compounds

We determined the minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) of CBHG, CBCHP, and E of the strains mentioned above.[31],[32] We started from 0.5 McFarland bacterial suspensions. For the MIC test, MH-agar plates with different concentrations of bismuth were seeded by drop. We worked with dilutions of 1/50, 1/25, 1/15, 1/10, 1/5, and 1/3, which correspond to bismuth concentrations of 0.6, 1.2, 2, 3, 6, and 10 mg/ml, respectively. Meanwhile, in the case of MBC, MH-broth tubes with serial dilutions of bismuth were inoculated, incubated overnight at 37°C and grown in MH-agar plates in the absence of bismuth. The dilutions used were 1/32, 1/16, 1/8, 1/4, and 1/2, which are equivalent to bismuth concentrations of 0.94, 1.88, 3.75, 7.5, and 15 mg/ml, respectively. The E was evaluated at the same dilutions as for bismuth compounds.

Effect on bacterial growth

The bacterial culture was incubated with and without bismuth compounds at different subinhibitory concentrations. The dilutions used were 1/500, 1/250, 1/100, and 1/50, which are equivalent to bismuth concentrations of 0.06, 0.12, 0.3, and 0.6 mg/ml, respectively. We measured the optical density at 630 nm, at different periods of growth (0, 2, 3, 4, 6, and 24 h), and in comparison to control samples without viable bacteria.

In all cases, to evaluate the growth, the following parameters were analyzed: specific growth velocity (μ) = ln(At2/ At1)/T and time of generation (G) = ln2*T/ln(At2/At1), where At1 is the absorbance in time 1, At2 is the absorbance in the subsequent time during the logarithmic phase and T is the time in hours.[33]

Determination of binding

The bacterial culture (1010 bacteria/ml) was incubated with and without bismuth compounds at a subinhibitory concentration (2 mg/ml) in LB medium at 37°C with shaking. Samples were then collected at specific time intervals (6, 24, and 168 h) to determine the total number of viable organisms present in the suspension. Additional samples were removed and centrifuged at 190 × g for 5 min, and they were counted on agar plates.[32] Preliminary experiments indicated that this centrifugation did not pellet bacteria free in suspension, but these conditions were sufficient to pellet CBHG. Bacteria remaining in the supernatant were considered unbound, and those present in the pellet were considered bound to CBHG.

Statistical analysis

All tests were performed independently three times, and the data were statistically evaluated by variance analysis, followed by the Tukey-Kramer multiple comparison test.


Sensitivity to bismuth compounds

When we performed the inhibition assays with CBCHP, we observed a MIC corresponding to 6 mg/ml of bismuth for the Salmonella spp., S. flexneri, EPEC and STEC and 10 mg/ml for the K. pneumoniae. We observed an MBC corresponding to 15 mg/ml of bismuth for the Salmonella spp., K. pneumoniae, EPEC and STEC and 7.5 mg/ml for the S. flexneri. In the case of CBHG, we observed a MIC corresponding to 10 mg/ml of bismuth for the Salmonella spp., K. pneumoniae, EPEC and STEC and 6 mg/ml for the S. flexneri. However, MBC was higher than 15 mg/ml. We did not observe a MIC or an MBC in any strain when we used the E alone at the studied concentrations. For the control strain E. coli ATCC 25922 we only observed a MIC of 6 mg/ml of bismuth with CBCHP [Table 1].{Table 1}

Effect on bacterial growth

Then, we evaluated the effect of the drugs on bacterial growth. For this purpose, we worked at low, subinhibitory concentrations (0.06, 0.12, 0.3, and 0.6 mg/ml) of CBCHP, CBHG, and E to determine the scope of the inhibition observed in the previous section [Figure 1] and [Figure 2].{Figure 1}{Figure 2}

For all strains, a significant decrease in bacterial growth was observed in the presence of bismuth compounds (CBCHP y CHBG) in comparison to the untreated, which could be appreciated due to a significant decrease in the specific growth velocity (P < 0.05), and an increase in generation time (P < 0.05). No significant difference was observed in the presence of E in comparison with the untreated.

The inhibitory effect was increased with the bismuth concentration tested. Almost all strains demonstrated the same behavior toward bismuth, and significant inhibition of growth was observed starting from the lowest concentration (0.06 mg/ml), with the exception of S. flexneri, which showed a significant inhibition starting from 0.3 mg/ml [Figure 1] and [Figure 2]c.

The inhibitory effect was significantly superior (P < 0.05) with CBCHP in comparison with CHBG in almost all strains, with the exceptions of S. flexneri and EPEC, which did not show significant differences between compounds.

Bacterial binding to colloidal bismuth

Finally, we tested the in vitro binding capacity that compounds showed on enterobacteria. CBHG and CBCHP exhibited a high bacterial attachment, from 94% to 97% and from 86% to 90%, respectively [Table 2]. These values were not modified significantly over time (P > 0.05) and they were significantly greater (P < 0.05) than that shown with excipients (36%–59%) in all studied strains.{Table 2}

No significant differences were observed in binding between CBHG and CBCHP. E. coli ATCC 25922 exhibited less attachment for all compounds tested (from 45% to 55%), and it was the only strain that did not show significant difference between bismuth compounds and the excipients. No significant difference was observed in the number of viable microorganisms attached without bismuth compounds under applied conditions.


Bismuth salts have been used for more than two centuries to treat different gastrointestinal pathologies. In spite of this, their mechanisms of therapeutic action continue to be a controversial subject. CBCHP is a low-cost, antidiarrhetic drug for an oral use that demonstrates multiple benefits in treating gastrointestinal infections. The most important benefit is that bismuth is available in colloidal form (CBHG) and pectin as active ingredients and not in the form of salts such as salicylate, subsalicylate, or subcitrate.[21],[22],[23],[25],[28],[29] CBHG does not inhibit intestinal motility and demonstrates very low intestinal absorption at <1%.

In previous studies of CBGH, we demonstrated an antimicrobial activity toward STEC and its main factors of virulences, very promising results in the prognosis of hemolytic uremic syndrome.[18] As a result of this work, we base our study on the analysis of the effect of CBHG, within the usual dosage, on Gram-negative enterobacteria strains frequently involved in gastrointestinal infections in our population.

In the first place, we evaluated the sensitivity of the bacteria to CBCHP, to its active principle CBHG and their excipients. The MIC is the most commonly used indicator in selecting an antimicrobial treatment. In all cases tested, the values of MIC were from 6 to 10 mg/ml of bismuth concentration. It was observed that K. pneumoniae was the strain that showed a slightly greater resistance, and S. flexneri slightly less than the rest of the studied strains, a similar result to that presented in other studies.[34] The standard concentration of bismuth in CBHG and CBCHP is 30 mg/ml, 3–5 times higher than the MICs detected. Hence, the strains were sensitive to CBHG treatment in concentrations lower than those normally used in treatment and in concentrations that it reaches in vivo in the intestine (10.8 mg/ml of bismuth).

The inhibitory scope of CBCHP and CHBG observed was high enough since it significantly inhibited bacterial growth in all strains at very low, subinhibitory concentrations of bismuth. At a concentration of 0.3 mg/ml of bismuth, all strains showed a reduction in growth velocity, while bismuth subsalicylate alone does not have a significant inhibitory effect on growth.[34]

In addition, other authors found comparable results regarding the viability of cells of Helicobacter pylori[35] and E. coli.[36] This bactericide action of the compound could be due to the loss of function of the cell membrane followed by the inhibition of ATP.[32]

The values MIC, MBC, and bacterial growth obtained showed a greater sensitivity to CBCHP, and no inhibitory effect was observed in the presence of only the excipients (E), which demonstrated that this effect was due exclusively to the active principle CBHG and that there was a cooperative effect between CBHG and the excipients that promote this effect.

Moreover, CBHG and CBCHP have demonstrated a great capacity for binding and displacement of the studied strains in solution. Both compounds demonstrated similar capacity of binding bacterial cells and was significantly greater than the capacity of the excipients. This demonstrated that this capacity for displacement was due mainly to the colloidal form of bismuth in the active ingredient CBHG.

The reference strain E. coli ATCC 25922, which lacks pathogenic factors, showed a significantly decreased capacity for displacement. Hence, we could infer that the presence of superficial structures, as pathogenic surface factors present in pathogenic cells, promote the attachment and displacement by colloidal bismuth.

The virulence of the bacterial strains was directly related with their capacity to be mobilized and attach to the epithelial gastrointestinal walls.[37],[38],[39] The mechanical attachment to the compounds used probably obstruct the bacterial attachment to the gastrointestinal epithelium, thus preventing colonization and favoring the rapid expulsion of the pathogens from the area.


The aim of this study was to find potential therapeutic strategies for the treatment of bacterial gastroenteritis by studying the effect of CBHG in clinical and subclinical concentrations on the viability and on the removal of the most frequent enteropathogenic strains in our population.

We believe that the results obtained with CBHG are very promising from the treatment standpoint. In this respect, its wide antimicrobial spectrum on the most frequent gastrointestinal pathogens and its capacity for bacterial displacement establish it as a treatment option in cases of diagnosis or suspicion of bacterial gastroenteritis.


We gratefully acknowledge Provincial Hospital of Rosario for providing isolates.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.


1World Health Organization. World Health Statistics 2011. World Health Organization; 2011. Available from:
2Giugno S, Oderiz S. Bacterial etiology of acute diarrhea in pediatric patients. Acta Bioquím Clín Latinoam 2010;44:63-9.
3Pan American Health Organization. Health in the Americas. Argentina: Pan American Health Organization; 2012. Available from: &view=article&id=18&Itemid=19&lang=en.
4González Corona E, Cunil Romero S. Diagnosis and treatment of persistent diarrhea in the service of acute diarrheic diseases. Medisan 2002;6:42-9.
5Casabonne C, González A, Aquili V, Balagué C. Prevalence and virulence genes of Shigella spp. Isolated from patients with diarrhea in Rosario, Argentina. Jpn J Infect Dis 2016;69:477-81.
6MSAL. Department of Epidemiology. Ministry of Health of Argentina. Diarrhea Surveillance Report 2010/2011; 2011.
7Vinh H, Nhu NT, Nga TV, Duy PT, Campbell JI, Hoang NV, et al. A changing picture of shigellosis in Southern Vietnam: Shifting species dominance, antimicrobial susceptibility and clinical presentation. BMC Infect Dis 2009;9:204.
8Jennison AV, Verma NK. Shigella flexneri infection: Pathogenesis and vaccine development. FEMS Microbiol Rev 2004;28:43-58.
9Jacoby GA, Munoz-Price LS. The new beta-lactamases. N Engl J Med 2005;352:380-91.
10Bilavsky E, Schwaber MJ, Carmeli Y. How to stem the tide of carbapenemase-producing enterobacteriaceae? Proactive versus reactive strategies. Curr Opin Infect Dis 2010;23:327-31.
11Wong CS, Jelacic S, Habeeb RL, Watkins SL, Tarr PI. The risk of the hemolytic-uremic syndrome after antibiotic treatment of Escherichia coli O157:H7 infections. N Engl J Med 2000;342:1930-6.
12Fernández-Riverón F, López Hernández J, Ponce-Martínez LM, Machado Betarte C. Resistencia bacteriana. Rev Cubana Med Mil 2003;32:44-8.
13Lambert JR, Midolo P. The actions of bismuth in the treatment of Helicobacter pylori infection. Aliment Pharmacol Ther 1997;11 Suppl 1:27-33.
14Kliegman RM. Nelson Textbook of Pediatrics. 18th ed. Elsevier: Barcelona; 2008.
15Lesa AM, Montrull HL, Nates SV, Bornoroni E, Gómez Molina A, Simes JC, et al. Use of a compound based on bismuth hydroxide gel with pectin in acute diarrhea. Med Infant 2004;11:180-4.
16Mahony DE, Woods A, Eelman MD, Burford N, Veldhuyzen van Zanten SJ. Interaction of bismuth subsalicylate with fruit juices, ascorbic acid, and thiol-containing substrates to produce soluble bismuth products active against Clostridium difficile. Antimicrob Agents Chemother 2005;49:431-3.
17Wu CL, Domenico P, Hassett DJ, Beveridge TJ, Hauser AR, Kazzaz JA, et al. Subinhibitory bismuth-thiols reduce virulence of Pseudomonas aeruginosa. Am J Respir Cell Mol Biol 2002;26:731-8.
18Subils T, Casabonne C, Balagué C. The inhibitory effect of colloidal bismuth hydroxide gel on Escherichia coli O157:H7 and on the activity of Shiga toxins. BMC Res Notes 2014;7:875.
19Steffen R, DuPont HL, Heusser R, Helminger A, Witassek F, Manhart MD, et al. Prevention of traveler's diarrhea by the tablet form of bismuth subsalicylate. Antimicrob Agents Chemother 1986;29:625-7.
20DuPont HL, Ericsson CD, Johnson PC, Bitsura JA, DuPont MW, de la Cabada FJ, et al. Prevention of travelers' diarrhea by the tablet formulation of bismuth subsalicylate. JAMA 1987;257:1347-50.
21Soriano-Brücher H, Avendaño P, O'Ryan M, Braun SD, Manhart MD, Balm TK, et al. Bismuth subsalicylate in the treatment of acute diarrhea in children: A clinical study. Pediatrics 1991;87:18-27.
22Figueroa-Quintanilla D, Salazar-Lindo E, Sack RB, León-Barúa R, Sarabia-Arce S, Campos-Sánchez M, et al. A controlled trial of bismuth subsalicylate in infants with acute watery diarrheal disease. N Engl J Med 1993;328:1653-8.
23Chowdhury HR, Yunus M, Zaman K, Rahman A, Faruque SM, Lescano AG, et al. The efficacy of bismuth subsalicylate in the treatment of acute diarrhoea and the prevention of persistent diarrhoea. Acta Paediatr 2001;90:605-10.
24Coudron PE, Stratton CW. Factors affecting growth and susceptibility testing of Helicobacter pylori in liquid media. J Clin Microbiol 1995;33:1028-30.
25Marcus EA, Sachs G, Scott DR. Colloidal bismuth subcitrate impedes proton entry into Helicobacter pylori and increases the efficacy of growth-dependent antibiotics. Aliment Pharmacol Ther 2015;42:922-33.
26Ericsson CD, Evans DG, DuPont HL, Evans DJ Jr., Pickering LK. Bismuth subsalicylate inhibits activity of crude toxins of Escherichia coli and Vibrio cholerae. J Infect Dis 1977;136:693-6.
27Gump DW, Nadeau OW, Hendricks GM, Meyer DH. Evidence that bismuth salts reduce invasion of epithelial cells by enteroinvasive bacteria. Med Microbiol Immunol 1992;181:131-43.
28Domenico P, Salo RJ, Straus DC, Hutson JC, Cunha BA. Salicylate or bismuth salts enhance opsonophagocytosis of Klebsiella pneumoniae. Infection 1992;20:66-72.
29Kunin CM, Hua TH, Guerrant RL, Bakaletz LO. Effect of salicylate, bismuth, osmolytes, and tetracycline resistance on expression of fimbriae by Escherichia coli. Infect Immun 1994;62:2178-86.
30Brogan AP, Verghese J, Widger WR, Kohn H. Bismuth-dithiol inhibition of the Escherichia coli rho transcription termination factor. J Inorg Biochem 2005;99:841-51.
31Clinical and Laboratory Standards Institute. Performance Standards for Antimicrobial Susceptibility Testing. Informational Supplement. Document M100S. 27th ed. Wayne PA: Clinical and Laboratory Standards Institute; 2017.
32Sox TE, Olson CA. Binding and killing of bacteria by bismuth subsalicylate. Antimicrob Agents Chemother 1989;33:2075-82.
33Bren A, Hart Y, Dekel E, Koster D, Alon U. The last generation of bacterial growth in limiting nutrient. BMC Syst Biol 2013;7:27.
34Domenico P. Method and Composition for Inhibiting Bacteria. United States Patent RE37, 793. Washington DC: U.S. Patent and Trademark Office; 2002.
35Worku ML, Sidebotham RL, Walker MM, Keshavarz T, Karim QN. The relationship between Helicobacter pylori motility, morphology and phase of growth: Implications for gastric colonization and pathology. Microbiology 1999;145(Pt 10):2803-11.
36Alharbi SA, Mashat BH, Al-Harbi NA, Wainwright M, Aloufi AS, Alnaimat S, et al. Bismuth-inhibitory effects on bacteria and stimulation of fungal growth in vitro. Saudi J Biol Sci 2012;19:147-50.
37Law D. Virulence factors of Escherichia coli O157 and other Shiga toxin-producing E. coli. J Appl Microbiol 2000;88:729-45.
38Lawlor MS, O'connor C, Miller VL. Yersiniabactin is a virulence factor for Klebsiella pneumoniae during pulmonary infection. Infect Immun 2007;75:1463-72.
39Josenhans C, Suerbaum S. The role of motility as a virulence factor in bacteria. Int J Med Microbiol 2002;291:605-14.