|
|
| |||||||||||||||||||||||||||||||||||||||
Background Bismuth subsalicylate is a common constituent of over-the-counter medications for diarrhea. However, it is uncertain whether bismuth offers any more benefit than standard oral rehydration therapy with early feeding.
Methods We conducted a placebo-controlled, randomized trial to evaluate the effect of bismuth subsalicylate (100 or 150 mg per kilogram of body weight per day for up to 5 days) on the duration and volume of acute watery diarrhea in 275 male infants and young boys (mean age, 13.5 months). Serum salicylate and bismuth levels were monitored throughout the study and were also measured two weeks after discharge. All the patients received fluid replacement by the oral route and early feeding of easily digestible foods with high caloric density.
Results Diarrhea stopped within 120 hours of admission in 74 percent of the patients given placebo, 89 percent of those given 100 mg of bismuth per kilogram (P = 0.009 vs. the placebo group), and 88 percent of those given 150 mg of bismuth per kilogram (P = 0.019 vs. the placebo group). As compared with the patients given placebo, those given bismuth had significant reductions in their total stool output (P = 0.015), total intake of oral rehydration solution (P = 0.013), and duration of hospitalization (P = 0.005); there was no significant difference between the two groups given bismuth in these clinical outcomes. All measurements of bismuth and salicylate concentrations in blood were well below concentrations considered toxic. No adverse reactions were seen.
Conclusions Treatment with bismuth subsalicylate decreases the duration of diarrhea and is a safe and effective adjunct to oral rehydration therapy for infants and young children with acute watery diarrhea.
More recently, Soriano-Brucher et al.6 showed that bismuth subsalicylate in a dose of 100 mg per kilogram of body weight per day decreased the duration and frequency of unformed stools in children with acute diarrhea, as compared with placebo. In that study, however, treatment usually included intravenous fluids, even for moderately dehydrated patients. Furthermore, food (mostly as diluted milk formula) was first offered to the patients only late in the course of the treatment. The question remains whether bismuth subsalicylate is effective if incorporated into more standard treatment with fluid replacement by the oral route and early introduction of easily digestible foods with high caloric density. We conducted a placebo-controlled, double-blind, randomized study to evaluate bismuth subsalicylate (100 or 150 mg per kilogram per day) as an adjunct to oral rehydration therapy for infants with acute watery diarrhea.
Methods
Study Population
Boys brought to an oral-rehydration unit (Instituto Nacional de Salud del Nino, Lima, Peru) for the treatment of acute diarrhea between January 1990 and March 1991 were eligible for this study if they were 3 to 59 months old and had passed three or more watery stools in the preceding 24 hours. Only male patients were studied to ensure accurate collection of stools. Patients were excluded from the study if they had blood in their stools (at admission or within the next 24 hours), had had diarrhea for more than five days, had received antibiotics or antidiarrheal medication or any treatment with acetylsalicylic acid in the 72 hours before admission, had clinical evidence of another illness requiring antibiotic therapy (at admission or within the next 24 hours), had severe malnutrition (<60 percent of the value for the 50th percentile for weight for age according to the tables of the National Center for Health Statistics7), had a history of allergy to salicylate or bismuth, or had been exclusively breast-fed. Informed consent was obtained from the parents of each child. The study protocol was approved by the research and ethics committees of the Universidad Peruana Cayetano Heredia, the Instituto Nacional de Salud del Nino, and the Johns Hopkins Committee on Human Volunteers.
Sample Size
The size of the sample was calculated with data from a study of Peruvian children with acute diarrhea8. The mean (±SD) stool volume in that study was 151 ±95 ml per kilogram. The assumptions of a 33 percent reduction in stool volume, a type I error of 0.05, and a power of 90 percent were used in estimating the number of patients required in each study group.
Evaluation of Patients
The patients were admitted to the rehydration ward and placed in metabolic beds to collect stools separately from urine. A complete medical history was obtained, and a standard physical examination was performed that included an assessment of the degree of dehydration9. Fresh stool samples were collected at the time of admission and transported in Cary-Blair medium to the microbiology laboratory (Universidad Cayetano Heredia). Specimens were cultured and examined for the presence of salmonella, shigella, and vibrio species and Escherichia coli on standard direct and enriched enteric mediums10. Campylobacter organisms were selectively cultured on sheep's blood agar containing Butzler supplement, and aeromonas organisms on blood agar with ampicillin after overnight enrichment in alkaline peptone water10. Five colonies of E. coli isolates were serotyped according to standard methods with polyvalent typing serums for the recognition of enteropathogenic E. coli10. Strains of E. coli were not tested for enterotoxigenicity. Rotavirus antigen was identified by enzyme-linked immunosorbent assay11.
Blood samples were obtained from each patient for the measurement of serum salicylate and bismuth at admission; 24, 72, and 120 hours after admission; and two weeks after discharge. Serum salicylate levels were measured with a spectrophotometric method12. Bismuth levels were determined by hydride-generation atomic absorption spectroscopy13. Patients were withdrawn from the study if their serum salicylate levels were 40 mg per deciliter or higher ( 
2.9 mmol per liter). Blood samples were also used to determine the hematocrit, the plasma specific gravity by refractometry, the serum sodium and potassium levels by flame photometry, the serum chloride level by titration, and the total carbon dioxide value by microgasometry. Fluid intake and output were measured and recorded every four hours until the diarrhea stopped.
Stools were collected and weighed in preweighed containers. Urine was collected separately in urine bags. Vomitus was absorbed and weighed in diapers weighed beforehand. Patients were weighed unclothed every 4 hours during the first 24 hours and every day thereafter. Dehydration was expressed as a percentage, with the difference between the weight at admission and the peak weight within 24 hours of admission serving as the numerator and the peak body weight serving as the denominator. The fluid deficit was corrected within four to six hours of admission. Patients with mild or moderate dehydration received ad libitum the oral rehydration solution recommended by the World Health Organization9. Patients with severe dehydration (indicated by two or more of the following signs: inability to drink, no passage of urine for six hours, unconsciousness, rapid and deep breathing, and a rapid and weak pulse) received the oral solution after they received intravenous rehydration. Feeding was recommenced four to six hours after admission in infants who had undergone clinical rehydration. Breast milk or lactose-free formula was used, to provide a daily caloric intake of 100 kcal per kilogram. Infants more than six months old were also given soft foods from the hospital; 10 percent of calories of these foods were proteins (chicken or beef), 45 percent were carbohydrates (starch from noodles, rice, or potatoes), and 45 percent were fats (vegetable oil). The amount of oral rehydration solution required was calculated every four hours on the basis of the volume of fluid lost in the preceding four hours, until the diarrhea stopped. Diarrhea was considered to have stopped when an unformed stool was followed by a formed stool or when no stool was passed for 12 hours.
Drug-Therapy Regimen
Each patient admitted to the study was randomly assigned to one of three study groups: one group received placebo, the second group received 100 mg of bismuth subsalicylate per kilogram per day (the 100-mg group), and the third group received 150 mg of bismuth subsalicylate per kilogram per day (the 150-mg group). Randomization was stratified according to age (3 to 23 months vs. 24 to 59 months). Random permuted blocks of variable size were used. The first dose of each study agent was administered at the time of enrollment, and subsequent doses were given every four hours for five days or until the diarrhea stopped, whichever occurred first. The placebo was identical in appearance and taste to the bismuth. The 100-mg and 150-mg preparations of bismuth contained 17.5 and 26.25 mg of active drug per milliliter, respectively, and all three preparations were administered by weight -- i.e., 0.952 ml per kilogram of the body weight at admission. For example, patients weighing 10 and 20 kg received 9.5 and 19.0 ml of the study preparation, respectively, every four hours. The preparations were readministered if the patients vomited them. Both preparations of bismuth also contained 0.6 mg of sodium salicylate per milliliter and 0.7 mg of salicylic acid per milliliter. Treatment failure was indicated by a recurrence or the continued presence of dehydration of more than 5 percent, a lack of weight gain from the time of admission onward, the development of ileus or severe diarrhea (stool output 10 ml per kilogram per hour) at any time during the treatment period, or the development of bloody stools after the first 24 hours.
Statistical Analysis
The data were first collected in precoded forms and then were entered in a data base organized with a relational model (FoxPro version 1.02) with a data-entry program with on-line checking. Two-way comparisons of base-line and outcome data were made between the three study groups, and three-way comparisons (with time since the start of treatment as the third variable) were made of variables measured several times during follow-up. Chi-square tests were used to compare discrete variables, and one-way analysis of variance with two degrees of freedom was used to compare continuous variables. The variables of stool output and intake of oral rehydration solution did not have a normal distribution; therefore, their values were transformed logarithmically before analysis. The duration of diarrhea was analyzed by estimating the percentage of patients in each group in whom diarrhea stopped by 120 hours after admission or sooner. Curves were constructed with the method of Cutler and Ederer14,15 to compare the duration of diarrhea.
The statistical analyses included all patients who completed the study. If patients were withdrawn before the completion of the study because of treatment failure or at their parents' request, the data on these patients up to the time of withdrawal were included in the analysis. A two-sided P value of less than 0.05 was considered to indicate statistical significance.
Results
Of the 275 patients initially included in the study (91 in the placebo group and 92 in each of the two bismuth groups), 23 were excluded within the first 24 hours (after randomization). These patients were excluded because of blood in the stool (four patients in the placebo group, four in the group that received 100 mg of bismuth and five in the group that received 150 mg of bismuth), the absence of unformed stools (two in the 100-mg group and two in the 150-mg group), excessive vomiting (one in the placebo group), and diagnoses of pneumonia (one in the 150-mg group) and acute otitis media (two in the placebo group, one in the 100-mg group, and one in the 150-mg group). Therefore, 252 patients (84 in the placebo group, 85 in the 100-mg group, and 83 in the 150-mg group) were considered in the final analysis. Only two patients in the 100-mg group and one patient in the placebo group needed initial intravenous therapy because of severe dehydration. Otherwise, only oral rehydration solution was used. No patient refused the study agents, and none died during the study.
Base-Line Comparison
The three study groups were comparable in their demographic and clinical characteristics at admission (Table 1). There were no significant differences among the groups in the distribution of the enteropathogens identified at admission (Table 2). An etiologic agent was identified in 166 of the 249 patients (67 percent) evaluated.
|
|
Diarrhea stopped within 120 hours of admission in 74 percent of the placebo group, 89 percent of the 100-mg group, and 88 percent of the 150-mg group. The difference between each bismuth group and the placebo group in the proportion of patients in whom diarrhea stopped was significant (100-mg group, P = 0.009; 150-mg group, P = 0.019), but the difference between the two bismuth groups was not significant (P = 0.763). The Cutler-Ederer curves for the study population (Figure 1) show that the reduction in the duration of diarrhea became evident the third day after treatment in the bismuth groups and was sustained throughout the period of evaluation. The duration of hospitalization was significantly reduced in both bismuth groups (Table 3). Total stool output was significantly reduced (by about 30 percent) in both bismuth groups (Table 3).
|
|
The total intake of oral rehydration solution was significantly less (about 25 percent lower) in both bismuth groups (Table 3). No significant differences were found among the three groups in the total caloric intake, the mean percentage of weight change (maximal weight 24 hours after admission vs. weight at discharge), or the volume of vomitus (Table 3).
Multivariate Analysis
Because there was a slight although nonsignificant imbalance in the distribution of rotavirus and enteropathogenic E. coli microorganisms among the study groups at base line, a multivariate analysis was performed to control the study outcomes for these two additional factors and their respective interactions. No evidence of confounding or interaction was found.
Salicylate Levels
Differences among the study groups in serum salicylate levels were significant at all times except at admission and two weeks after discharge (Figure 2). The levels in the bismuth groups reached a plateau at 24 hours and decreased toward base line two weeks later. The highest mean level of salicylate was 11.50 mg per deciliter (0.8 mmol per liter) in the 100-mg group and 18.49 mg per deciliter (1.3 mmol per liter) in the 150-mg group at 24 and 72 hours, respectively. None of the patients included in the study were withdrawn because of high salicylate levels ( 40 mg per deciliter).
|
There were no serum samples with bismuth levels of 100 ng per milliliter or more, levels considered potentially toxic16. There were 23 samples in which the bismuth levels were more than 50 ng per milliliter; 10 of these samples were from patients in the 150-mg group: 5 samples were obtained at 24 hours, 1 at 48 hours, 2 at 120 hours, 1 at discharge, and 1 at admission. Five other samples were from patients in the 100-mg group: three samples were obtained at 24 hours, one at 72 hours, and one at follow-up, two weeks later. Bismuth levels were high in eight samples from patients in the placebo group: four samples were obtained at admission, and one sample each at 24, 48, and 72 hours and two weeks after discharge.
Adverse Reactions
Two patients (one in the placebo group and the other in the 150-mg bismuth group) had a rash during the study. They were given an antihistamine in syrup, and the rash cleared rapidly. Neither patient had to be withdrawn from the study.
Discussion
The results of this study show the effectiveness of bismuth subsalicylate as adjunctive therapy to oral rehydration and early continued feeding of children with acute diarrhea. Both the patients given low doses of bismuth and those given high doses had clinically important and statistically significant reductions in the duration of diarrhea and hospitalization, stool output and the volume of diarrheal stool, and the intake of oral rehydration solution. In an earlier clinical trial, Soriano-Brucher et al.6 administered bismuth subsalicylate in a similar dose (100 mg per kilogram per day) to children hospitalized for diarrhea, and found that bismuth treatment also decreased stool frequency and shortened the duration of diarrhea; however, in that study rehydration depended largely on intravenous solutions, with food withheld during the initial days of treatment. Thus, the effectiveness of bismuth subsalicylate as an adjunct to an optimal regimen for diarrhea in children (incorporating oral rehydration therapy and continued feeding with an appropriate diet) was not addressed. Our study found that treatment with bismuth significantly improved the clinical course of disease even when the currently recommended oral treatment regimen was followed.
In our study, comparable reductions in the total stool volume, total stool output, and duration of diarrhea were observed in the groups given two different doses of bismuth subsalicylate. In addition, stool output decreased more rapidly in these groups. Furthermore, the difference in the average total stool output between the placebo group and the group given 100 mg of bismuth represented a reduction of 30 percent, and the difference between the placebo group and the group given 150 mg of bismuth represented one of 33 percent. The reduction in the intake of oral rehydration solution and the duration of hospitalization in the patients given bismuth was related to the reduction in stool output and the duration of diarrhea in these groups. We chose the dose of 150 mg of bismuth per kilogram per day to determine whether a larger dose would achieve a greater reduction in stool output or the duration of diarrhea. We found no significant differences between the effects of the two doses used.
Although bismuth salts are highly insoluble and little bismuth is absorbed across the small intestine, there has been some concern about the potential neurotoxic effect of this drug. Between 1973 and 1980, approximately 1000 cases of bismuth-related encephalopathy were reported in France and Australia,17,18 most of which occurred during long-term treatment with high doses of bismuth subnitrate or bismuth subgallate. There is no evidence that short-term therapy with bismuth subsalicylate presents a similar hazard if the recommended dose is not exceeded. Thus, bismuth subsalicylate seems to be a safe agent. Hillemand et al.16 reviewed the relation between blood bismuth levels and encephalopathy. They concluded that plasma levels persistently above 100 ng per milliliter were potentially toxic and recommended that therapy with a bismuth preparation should be discontinued if serum levels exceeded 100 ng per milliliter. In our study, all the patients who received bismuth subsalicylate had bismuth blood levels well below those considered toxic, and none had adverse reactions. Patients who had detectable serum bismuth levels at admission, as well as patients in the placebo group who had high levels later in the study, may have received preparations containing bismuth before they were studied. The patients' mothers were asked to report the use of such compounds, and administration of these drugs precluded admission to the trial.
Despite the widespread use of bismuth subsalicylate in the treatment of diarrhea, its mechanism of action is not completely understood. Some evidence suggests that it has both antisecretory and antimicrobial properties19,20,21. The literature indicates that possible mechanisms of action are the prevention of attachment of microorganisms to the intestinal mucosa, a direct antimicrobial effect on the gut, inactivation of enterotoxins, and an antisecretory effect of the salicylate present in this drug. Manhart21 demonstrated that bismuth subsalicylate and other bismuth salts can inhibit the growth of a wide variety of enteropathogens in vitro. Graham et al.22 showed that pathogens were isolated less frequently in patients with diarrhea who received bismuth than in those who received placebo. These investigators rarely recovered enterotoxigenic E. coli from the stool of volunteers who ingested these organisms and received bismuth subsalicylate prophylactically, but they did isolate enterotoxigenic E. coli in almost 87 percent of the patients who were treated with placebo. These findings suggest a specific interaction of bismuth and bacteria. In addition to its antibacterial effects, bismuth may influence the course of viral enteric infections. It has been shown to inhibit in vitro replication of four strains of rotavirus in tissue-culture cells and to cause a dose-dependent reduction in the growth of a number of enteric viruses23. Several studies have demonstrated the efficacy of bismuth subsalicylate in the treatment of traveler's diarrhea. DuPont et al.3,24 demonstrated that this drug was significantly superior to placebo in decreasing stool frequency in both the entire study population with diarrhea and subgroups with diarrhea with specific causes. Also, a decrease in symptoms such as diarrhea, nausea, and abdominal pain within 24 hours of therapy was reported significantly more often by the patients given bismuth. These findings were confirmed by Steffen et al.,25 who also showed that bismuth subsalicylate relieved severe illness as effectively as it relieved mild illness.
Although the specific mechanism of action of bismuth is not yet known, it is likely that its effect is due to both antimicrobial and antiinflammatory capabilities4,19,21. Other than some antimicrobial agents used to treat specific conditions such as cholera and shigellosis, antidiarrheal drugs have been considered useless and potentially harmful medicines26. The positive effect of bismuth subsalicylate on the clinical course of diarrhea in our study was seen in patients receiving properly managed fluid replacement and feeding. The cost of adding bismuth treatment to the therapy of infants with acute watery diarrhea could be balanced by savings due to shorter hospitalization. Bismuth subsalicylate, used properly, could be a useful, safe, and cost-effective adjunct to oral rehydration therapy and nutritional therapy in children with acute watery diarrhea. Although the serum bismuth levels in this study remained well below levels associated with toxicity, there is a need to determine the effectiveness of lower doses of bismuth. In addition, the use of a regimen requiring fewer than the six doses per day given the inpatients in this study could improve compliance among outpatients.
Supported by a grant from the International Child Health Foundation (a 501(C) nonprofit organization supported by gifts and grants from individuals, foundations, and corporations) and by funds from the Procter & Gamble Company.
We are indebted to the nursing staff of the Rehydration Unit of the Instituto Nacional de Salud del Nino; to Dr. Khin-Maung-U of the International Child Health Foundation, who assisted in the investigation; and to Dr. William B. Greenough III and Dr. Charles B. Stephensen, who provided excellent reviews of the manuscript and assistance in its preparation.
Source Information
From the Instituto Nacional de Salud del Nino (D.F.-Q.) and Universidad Peruana Cayetano Heredia (E.S.-L., R.L.-B., S.S.-A., M.C.-S., E.E.-M.), Lima, Peru; and the Johns Hopkins University School of Hygiene and Public Health, Baltimore (R.B.S.). Presented in part at the 31st Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, October 1, 1991.
Address reprint requests to Dr. Salazar-Lindo at Universidad Peruana Cayetano Heredia, A.P. 4314, Lima 100, Peru.
References
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
Bismuth in Infants with Watery Diarrhea
Abramson J. S., Givner L. B., Woods C. R., Guiraldes E., Salazar-Lindo E., Sack R. B., Figueroa-Quintanilla D.
Extract |
Full Text
N Engl J Med 1993;
329:1742-1743, Dec 2, 1993.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
