Search

In a field trial carried out in 1985 in Matlab, Bangladesh, the authors evaluated whether subjects who developed Vibrio cholerae 01 infections during the first year after earlier receipt of B subunit-killed whole cell (BS-WC) or killed whole cell-only (WC) oral cholera vaccines exhibited deficient serum vibriocidal immune responses to these infections. After severe V. cholerae 01 infections (n = 70) in subjects > 5 years of age, the age group in which both vaccines were efficacious, a 6.5 geometric mean-fold rise of serum vibriocidal antibodies was observed among vaccinees, compared with an 18.6 geometric mean-fold rise in placebo-recipients (p < 0.01). Depressions of serum vibriocidal responses among vaccinees were even more marked after asymptomatic infections (n = 30): a 1.1 geometric mean-fold rise in vaccinees versus a 5.9 geometric mean-fold rise in placebo-recipients (p < 0.01). The authors conclude that subjects who failed to be protected by BS-WC and WC, despite being in the age group for which these vaccines were protective, exhibited poor immune responses even to the vigorous stimulus of natural infection. These findings raise the possibility that immune hyporesponsiveness may limit the potential efficacy attainable by cholera vaccines in populations with endemic cholera.

The recent spread of cholera to Latin America, together with the persistent burden of this disease in Asia and Africa, have stimulated efforts to evaluate new cholera vaccines in field settings. Although the standard experimental paradigm for vaccine field trials is well established, the success of these trials will also depend on suitable consideration of the epidemiology of cholera and of cholera vaccination in the setting under study. Epidemiological studies done in Bangladesh emphasize the importance of appreciating the poorly predictable, multifocal occurrence of cholera in estimating a probable incidence of cholera for a field trial. They also underscore how the filtering effect of enrolling subjects into a prospective trial can dramatically reduce the available population for study, and can yield a study sample whose expected risk of cholera differs markedly from that for the source population. Finally, the data highlight the subtle effects that the mode of surveillance and the choice of an outcome definition can have upon protective efficacy, and emphasize the need for subgroup analyses that address the distinctive variations in vaccine protection that may occur in subjects differing in age and in ABO blood groups, and in subjects exposed to classical versus El Tor cholera.

In a randomized, placebo-controlled field trial of B subunit-killed whole cell (BS-WC) and killed whole cell only (WC) inactivated oral cholera vaccines in rural Bangladesh, active surveillance of selected neighborhoods during the first year after vaccination identified 127 Vibrio cholerae O1 infections among 3285 three-dose recipients. For each vaccine, protective efficacy was greater against symptomatic (57%, P < .05 for BS-WC; 58%, P < .05 for WC) than against asymptomatic infections (46%, P < .05 for BS-WC; 32%, P = .09 for WC), and protection against each grade of infection was demonstrable for both the classical and El Tor biotypes. Although vaccine protection against symptomatic infections was evident in both young children and older persons, only persons vaccinated at age > 5 years were protected against asymptomatic infections. These results suggest that the inactivated oral vaccines acted both to protect against intestinal colonization by V. cholerae O1 and to interrupt the pathogenic sequence of established infections.

The authors estimated the incidence rates of cholera and death between 1985 and 1988 for 32,642 age- and sex-eligible persons who did not participate in a randomized, placebo-controlled field trial of killed oral cholera vaccines in rural Bangladesh. As compared with 20,744 placebo recipients, the relative risk of cholera for all nonparticipants, adjusted for potentially confounding demographic variables, was 1.20 (95% confidence interval (CI) 1.03-1.41); this adjusted relative risk reflected elevated adjusted relative risks in nonparticipants who were medically ineligible (RR = 1.65; 95% CI 1.22-2.22) or refused to participate (RR = 1.19; 95% CI 1.01-1.41), but not in persons absent at the time of vaccination (RR = 1.00; 95% CI 0.78-1.28). The adjusted relative risk of death was also elevated in nonparticipants as compared with placebo recipients (RR = 1.28; 95% CI 1.10-1.48), with the same pattern of adjusted relative risks for different categories of nonparticipants: for ineligible subjects, 2.64 (95% CI 2.12-3.29); for refusers, 1.20 (95% CI 1.02-1.41); and for absentees, 0.95 (95% CI 0.75-1.22). The authors concluded that nonparticipation was associated with clinically cogent adverse health outcomes, but that the magnitude of these associations varied according to the reason for nonparticipation. These findings underscore the caution required in assessing vaccine efficacy with controls who are not vaccinated because of choices made by patients or vaccinators.

The relationship of serum vibriocidal (VC) and IgG anti-cholera toxin (CT) antibodies to the risk of cholera was evaluated during the first year of follow-up of recipients of three oral doses of B subunit (BS)-whole-cell vaccine, whole-cell vaccine, or Escherichia coli K12 strain placebo in Bangladesh. Acute sera from 121 cholera patients were compared with sera from 2592 contemporaneous community controls. Each doubling of VC titer was associated, on average, with a 22%-47% reduction of cholera risk in the three groups. In contrast, in the two groups that did not receive BS, anti-CT titers were directly associated with cholera and thus served as markers of higher cholera risk. Each vaccine conferred approximately 65% protective efficacy against cholera, but antibody titers did not correlate with vaccine efficacy, indicating that serum VC and anti-CT antibodies are poor markers of the longitudinal pattern of vaccine efficacy.

To test the hypothesis that clinical Vibrio cholerae O1 infections protect against recurrent cholera, treated cholera episodes in a rural Bangladesh population of 188,153 people who were followed between 1985 and 1988 were analysed. Of the 2214 people with initial episodes of cholera, 7 had a second episode. The incidence of cholera was 61% lower in subjects who had had an earlier episode than in those without such an episode. Whereas initial episodes of classical cholera were associated with complete protection against subsequent cholera, initial episodes of El Tor cholera were associated with negligible protection.

In a field trial conducted in Bangladesh, ingestion of either B subunit-killed whole cell (BS-WC) or killed whole cell (WC) oral cholera vaccines by mothers was associated with a 47% reduction of the risk of cholera in their non-vaccinated children aged under 36 months. Because vaccine-induced breast-milk immunity seemed a possible explanation for these findings, we evaluated anti-lipopolysaccharide (LPS) and anti-cholera toxin (CT) IgA antibody responses in breast milk collected during the trial from 53 lactating women who ingested three doses of BS-WC, WC, or an Escherichia coli K12 strain (K12). Despite induction of moderate vibriocidal (1.4 to 2.0-fold) and anti-CT (4.5-fold) serum antibody responses, the vaccines did not elicit significant rises of anti-LPS or anti-CT IgA breast-milk antibodies. The failure of the vaccines to elicit significant levels of breast-milk anti-cholera antibodies suggests an alternative explanation for protection of young children by maternal vaccination, such as interruption of maternal-child transmission of Vibrio cholerae 01.

The association between breast feeding and the risk of severe cholera was examined in a case-control study of rural Bangladeshi children under 36 months of age who were studied in 1985-1986 during a field trial of killed oral cholera vaccines. A total of 116 cases who were treated for severe cholera were compared with 464 age-matched community controls without severe cholera. Overall, the odds ratio relating breast feeding to severe cholera (0.30, p less than 0.0001) reflected a 70% reduction in the risk of severe cholera among breast-fed children. The estimated reduction of risk declined with age, but was clearly evident in children up to 30 months of age. Although the association between breast feeding and a reduced risk of severe cholera was not significantly greater in children of mothers who had received cholera vaccine than in children whose mothers had received placebo during the trial, maternal vaccination per se was suggestively associated with a reduced risk of severe cholera in their nonvaccinated children (odds ratio = 0.53, p = 0.05). These results indicate that breast feeding was associated with a substantial reduction of the risk of severe cholera and raise the possibility that vaccination of mothers may provide protection to their young children in endemic settings.

The protective efficacy (PE) of B subunit killed whole-cell (BS-WC) and killed whole-cell-only (WC) oral cholera vaccines was assessed in a randomised double-blind field trial among children aged 2-15 years and women over 15 years in rural Bangladesh. Among the 62 285 subjects who received three doses of BS-WC, WC, or Escherichia coli K12 strain placebo, cumulative PE at 3 years of follow-up was 50% for BS-WC and 52% for WC. PE was similar against severe and non-severe cholera, but was significantly lower in children who were vaccinated at 2-5 years (26% for BS-WC; 23% for WC) than in older persons (63% for BS-WC; 68% for WC). Among persons vaccinated at 2-5 years, protection at 4-6 months of follow-up was similar to that for older persons, but rapidly waned thereafter and was not evident during the third year of follow-up. In contrast, persons vaccinated at older ages were protected even in the third year of follow-up (PE 40% for BS-WC; 62% for WC). PE was substantially higher against classical cholera (58% for BS-WC; 60% for WC) than against El Tor cholera (39% and 40%).

Although gastric acid is thought to be an important host defense against certain enteric infections, field studies of the role of gastric acid in preventing enteric infections have been hampered by the lack of a suitable non-invasive test. Because low gastric acid output (GAO) is an established risk factor for cholera, we assessed after validation, whether a new non-invasive test which estimates GAO by measuring breath hydrogen excess after ingestion of magnesium and a stimulant of gastric acid secretion, could discriminate between persons at high and at low risk of developing cholera. Fifteen age-matched pairs, participants in the field trial of two oral cholera vaccines in rural Bangladesh, were tested. In each pair the "case" was a person who had recovered from severe cholera at least 6 months before testing and the "control" was a person who resided in the home of a cholera patient but remained uninfected. The stimulated breath hydrogen was higher in controls (median hydrogen excess = 369 mumol/80 min) than in cases (median hydrogen excess = 150 mumol/80 min) (p less than 0.05) and was higher in controls in 12 out of 15 pairs. The results, which are consistent with past invasive assessments of the association between hypochlorhydria and cholera, suggest that this non-invasive test may be useful in evaluating GAO in epidemiological field studies.

The association between undernutrition and the risk of colonization and disease with Vibrio cholerae O1, concentrations of salivary IgA and the serologic response to infection and to orally administered cholera B subunit were examined prospectively in a family study in Bangladesh. Children ages 1 to 8 years who were family contacts of patients hospitalized with culture-confirmed cholera were visited within 24 hours of the hospitalization and daily for 10 days, queried for the presence of diarrhea and cultured for V. cholerae O1. On Day 1 each child was weighed and saliva was collected to measure total IgA. On Days 1 and 21 blood was taken to assess vibriocidal and antitoxin titers, and on Days 1 and 2 B subunit or placebo was given orally as part of a trial to look for a toxin-blocking effect. Of 412 children enrolled in the study 35% (143) became infected with V. cholerae O1 and 49% (70) of these developed diarrhea. Undernutrition, defined in a child as weight less than 70% of the Harvard reference weight-for-age, was not associated with colonization, disease or the duration or severity of cholera. Moreover well-nourished children did not differ from undernourished children in their concentrations of salivary total IgA, initial serum antitoxin or vibriocidal antibodies or in their serologic response to colonization, disease or B subunit. The immune system in its response to cholera appears to be quite resistant to nutritional insults. The good antitoxin response to B subunit among undernourished children is of particular importance in considering the use of future oral cholera vaccines in areas where such undernutrition is common.

The cholera experience of a sample of families in a rural area of Bangladesh is reported in relation to water supply and use. Tanks were the primary source for 65% of families, canals for 20% and the river for 14%. The highest attack rate was associated with access to canal water (13%). Attack rates did not vary markedly according to the purpose for which a source was used. The importance of cultural patterns in water use is identified.

A review of the incidence of cholera from 1964 through 1974 in Matlab, Bangladesh, revealed that among the villages several had very high incidence-rates. Investigation indicated that high cholera-rates in two of these villages were probably related to water contamination from a nearby cholera hospital established in 1963. The data imply that heavy contamination can overcome any immunity resulting from repeated exposure.

In rural Bangladesh, family contacts of patients with cholera were studied prospectively to examine whether protection against colonization and disease due to Vibrio cholerae O1 was associated with circulating antibodies to V. cholerae. Family contacts (1,071) of 370 patients with cholera were visited daily for 10 days, cultured for V. cholerae, and queried about diarrhea. Sera collected on days 1 and 21 were assayed for vibriocidal antibodies, IgG and IgA antibodies to cholera toxin, and IgG antibodies to lipopolysaccharide (LPS). Vibriocidal titers of greater than or equal to 20 present in 50% of contacts by 20 years of age were associated with protection against both colonization and disease. An elevated level of IgG antitoxin was not associated with protection against colonization or disease but was the most sensitive indicator of recent symptomatic cholera and of immune response to the oral immunogen B subunit. IgG antibody to LPS and IgA antitoxin were of little value in predicting colonization or disease.

We assessed the protective efficacy (PE) of three doses of B subunit-killed whole cell (BS-WC) and killed whole cell-only (WC) oral cholera vaccines in a randomized, double-blind trial among 62,285 children and women residing in rural Bangladesh. After one complete year of surveillance, 110 cases of cholera were detected in the placebo group, 52 in the WC group (PE, 53%; P less than .0001), and 41 in the BS-WC group (PE, 62%; P less than .0001). Protection was greater for BS-WC recipients than for WC recipients only during the initial eight months of observation. Both vaccines conferred equivalent protection against cholera associated with life-threatening dehydration and against less severe cholera. High-grade, sustained protection was observed in persons vaccinated when older than five years; in younger persons protection was transient. We conclude that BS-WC and WC vaccines confer significant protection against cholera, particularly in persons vaccinated when older than five years.

The impact of B subunit killed whole-cell (BS-WC) and killed whole-cell-only (WC) oral cholera vaccines was assessed in a randomised double-blind trial in rural Bangladesh. 62,285 children aged 2-15 years and women aged over 15 ingested three doses of one of the vaccines or placebo. During the first year of follow-up there was a 26% reduction of all visits for treatment of diarrhoea in the BS-WC group and a 22% reduction in the WC group. The reduction of all admissions for fatal or severely dehydrating diarrhoea was 48% in the BS-WC group and 33% in the WC group. Overall mortality rates were 26% lower in the BS-WC group and 23% lower in the WC group during the first year, and reductions of mortality were observed only in women vaccinated at ages over 15 years. However, no differences in cumulative mortality were evident at the end of the second year of surveillance.

At the Matlab Hospital of the International Centre for Diarrhoeal Disease Research, Bangladesh, the authors examined the blood groups of patients hospitalized between January and September 1979 for diarrheal disease due to a variety of bacterial and viral agents. A significant association was identified only for cholera, in which cholera patients were twice as likely to have blood group O and one-ninth as likely to have blood group AB as community controls. A follow-up study of family contacts of cholera patients, carried out between September 1980 and July 1982, indicated that blood group did not affect an individual's risk of having a culture-proven infection with V. cholerae 01 but was directly related to the severity of disease. Individuals with the most severe diarrhea compared with those with asymptomatic infection were more often of blood group O (68% versus 36%, p less than 0.01) and less often of AB (0% versus 7%, p less than 0.01). It was not possible to identify the molecular basis for this genetically related protection using biologic models of cholera that are currently available. The constant selective pressure of cholera against people of O blood group may account in part for the extremely low prevalence of O group genes and the high prevalence of B group genes found among the people living in the Gangetic Delta.

In 1979, an outbreak of plasmid-borne, multiply drug-resistant Vibrio cholerae serogroup O1 biotype El Tor (V. cholerae O1) occurred in the Matlab area of Bangladesh. The outbreak could have resulted from the introduction into the area of a single resistant strain or from multiple conjugations of drug-sensitive V. cholerae O1 with C plasmids in other environmental flora. Resistant strains were phage typed to determine their relatedness, and plasmid studies were conducted to determine the frequency of C plasmids in nonvibrio flora of family contacts of cholera patients. Forty-one (85%) of 48 resistant strains of V. cholerae O1 examined belonged to two closely related phage types new to the area, whereas 59 drug-sensitive strains from the same period were primarily of two different phage types. Group C plasmids were in nonvibrio strains from five of 36 family contacts of patients with drug-resistant cholera but none of 82 family contacts of patients with sensitive cholera. This outbreak most likely began from the introduction into the area of a single, multiply drug-resistant strain of V. cholerae O1. C plasmids detected in the nonvibrio flora of family contacts probably came from the resistant strain of V. cholerae O1.

We performed a prospective study to examine whether the IgA antibodies against cholera that are present in breast milk protect breast-fed infants and children against colonization with Vibrio cholerae 01 and disease. Among families of patients with cholera, we collected breast milk from mothers who had not had diarrhea in the previous week and monitored them and their breast-fed children for cholera colonization and diarrhea for 10 days. Breast milk was assayed for IgA antibodies to cholera toxin and lipopolysaccharide. Ninety-three mother--child pairs were studied; 30 infants became colonized with V. cholerae 01 and disease developed in 19. There were no differences between the antibody levels in milk fed to children who became colonized and in milk fed to children who did not. However, among the children who became colonized, those who had diarrhea drank breast milk containing significantly lower levels of both kinds of cholera antibodies than were present in the milk consumed by children who had no symptoms. We conclude that breast-milk antibodies against cholera do not appear to protect children from colonization with V. cholerae 01 but do protect against disease in those who are colonized.

The epidemiology of Vibrio cholerae O1 El Tor in rural Bangladesh was examined with a new phage-typing system for characterization of individual strains. During a two-year period, 537 strains were typed with a set of standard and experimental phages. Four major and many minor patterns were identified, some associated with discrete outbreaks of disease and others persisting for the entire period. One outbreak-related phage type was associated with a C plasmid bearing multiple drug resistance. The yearly cholera seasons began with the emergence of strains of many different phage types in many different locations; this pattern is consistent with an environmental reservoir for V cholerae. Each patient with cholera excreted only one phage type over time, whereas some family members were infected with strains of different phage types. Phage typing can be a useful marker with which the spread of cholera strains can be traced in endemic settings.

The protective efficacy of oral B subunit killed whole-cell (BS-WC) and killed whole-cell (WC) cholera vaccines was assessed in 63 498 Bangladeshi children aged 2-15 years and women aged over 15 years. Each received three doses of BS-WC, WC, or placebo in a randomised, double-blinded fashion. Surveillance for cases seeking medical care up to six months after the third dose revealed 26 cases of confirmed cholera in the placebo group, 4 cases in the BS-WC group (protective efficacy 85%; p less than 0.0001), and 11 cases in the WC group (protective efficacy 58%; p less than 0.01). For each vaccine protective efficacy was consistent in different age-groups (2-10 years versus greater than 10 years) and for different severities of cholera.

A randomized, controlled field trial was performed to test the ability of B subunit, the nontoxic, binding portion of cholera toxin, to block the toxin receptors (GM1 ganglioside) in the small intestine and thereby prevent diarrhea in individuals infected with Vibrio cholerae O1. Of 1,922 family contacts of 370 index patients selected randomly to receive orally on two successive days either B subunit (low dose, 1.0 mg; high dose, 5.0 mg) or placebo, 190 were asymptomatically infected on day 1 or day 2 of the study and within 24 hr of receiving B subunit. During the first 24-hr period of follow-up, the relative risk of disease among contacts receiving B subunit versus placebo was 0.18 for the low dose (P = .08) and 0.50 for the high dose (P = .22). Subsequently the relative risk increased toward 1.0 and was at no single point significantly reduced, although in five of the six follow-up periods the risk of disease was less in the B subunit group.

Within an area of Bangladesh in which the incidence of cholera was high, use of sanitary pipe wells did not protect against cholera or related non-cholera diarrhoeas because well users also used contaminated water sources regularly enough to maintain high infection-rates. Protection was found to correlate with education and wealth.