Efficacy of a single-dose regimen of inactivated whole-cell oral cholera vaccine: results from 2 years of follow-up of a randomised trial

Transcription

1 Efficacy of a single-dose regimen of inactivated whole-cell oral cholera vaccine: results from 2 years of follow-up of a randomised trial Firdausi Qadri, Mohammad Ali, Julia Lynch, Fahima Chowdhury, Ashraful Islam Khan, Thomas F Wierzba, Jean-Louis Excler, Amit Saha, Md Taufiqul Islam, Yasmin A Begum, Taufiqur R Bhuiyan, Farhana Khanam, Mohiul I Chowdhury, Iqbal Ansary Khan, Alamgir Kabir, Baizid Khoorshid Riaz, Afroza Akter, Arifuzzaman Khan, Muhammad Asaduzzaman, Deok Ryun Kim, Ashraf U Siddik, Nirod C Saha, Alejandro Cravioto, Ajit P Singh, John D Clemens Lancet Infect Dis 2018; 18: Published Online March 14, S (18) See Comment page 591 International Centre for Diarrhoeal Disease Research Bangladesh, Dhaka, Bangladesh (F Qadri PhD, F Chowdhury MPH, A I Khan PhD, A Saha MMed, M T Islam MPH, Y A Begum PhD, TR Bhuiyan PhD, F Khanam MSc, M I Chowdhury MPH, A Kabir MSc, A Akter MPH, A Khan MBBS, M Asaduzzaman MPH, A U Siddik MSS, N C Saha MSc, Prof J D Clemens MD); Department of International Health, Johns Hopkins School of Public Health, Baltimore, MD, USA (M Ali PhD); International Vaccine Institute, Seoul, South Korea (J Lynch MD, J-L Excler MD, D R Kim MSc); Vaccine Development Global Program PATH, Washington, DC, USA (T F Wierzba PhD); The Institute of Epidemiology, Disease Control and Research, Dhaka, Bangladesh (I A Khan MPH); Department of Public Health and Hospital Administration, National Institute of Preventive and Social Medicine, Dhaka, Bangladesh (Prof B K Riaz MPH); Department of Public Health, Universidad Nacional Autonoma de Mexico, Mexico City, Mexico (Prof A Cravioto PhD); MSD Wellcome Trust Hilleman Laboratories, New Delhi, India (A P Singh MD); Department of Epidemiology of the Center for Global Infectious Diseases, UCLA Fielding School of Public Health, Los Angeles, CA, USA (Prof J D Clemens); and Department of Medicine, Korea University School of Medicine, Seoul, South Korea (Prof J D Clemens) Summary Background A single-dose regimen of inactivated whole-cell oral cholera vaccine (OCV) is attractive because it reduces logistical challenges for vaccination and could enable more people to be vaccinated. Previously, we reported the efficacy of a single dose of an OCV vaccine during the 6 months following dosing. Herein, we report the results of 2 years of follow-up. Methods In this placebo-controlled, double-blind trial done in Dhaka, Bangladesh, individuals aged 1 year or older with no history of receipt of OCV were randomly assigned to receive a single dose of inactivated OCV or oral placebo. The primary endpoint was a confirmed episode of non-bloody diarrhoea for which the onset was at least 7 days after dosing and a faecal culture was positive for Vibrio cholerae O1 or O139. Passive surveillance for diarrhoea was done in 13 hospitals or major clinics located in or near the study area for 2 years after the last administered dose. We assessed the protective efficacy of the OCV against culture-confirmed cholera occurring days after dosing with both crude and multivariable per-protocol analyses. This trial is registered at ClinicalTrials.gov, number NCT Findings Between Jan 10, 2014, and Feb 4, 2014, people were randomly assigned to receive either vaccine or placebo, of whom ( vaccine recipients and placebo recipients) were included in the per-protocol analysis. 287 first episodes of cholera (109 among vaccine recipients and 178 among placebo recipients) were detected during the 2-year follow-up; 138 of these episodes (46 in vaccine recipients and 92 in placebo recipients) were associated with severe dehydration. The overall incidence rates of initial cholera episodes were 0 22 (95% CI 0 18 to 0 27) per person-days in vaccine recipients versus 0 36 (0 31 to 0 42) per person-days in placebo recipients (adjusted protective efficacy 39%, 95% CI 23 to 52). The overall incidence of severe cholera was 0 09 (0 07 to 0 12) per person-days versus 0 19 (0 15 to 0 23; adjusted protective efficacy 50%, 29 to 65). Vaccine protective efficacy was 52% (8 to 75) against all cholera episodes and 71% (27 to 88) against severe cholera episodes in aged 5 years to younger than 15 years. For aged 15 years or older, vaccine protective efficacy was 59% (42 to 71) against all cholera episodes and 59% (35 to 74) against severe cholera. The protection in the older age groups was sustained throughout the 2-year follow-up. In younger than 5 years, the vaccine did not show protection against either all cholera episodes (protective efficacy 13%, 68 to 25) or severe cholera episodes ( 44%, 220 to 35). Interpretation A single dose of the inactivated whole-cell OCV offered protection to older children and adults that was sustained for at least 2 years. The absence of protection of young children might reflect a lesser degree of pre-existing natural immunity in this age group. Funding Bill & Melinda Gates Foundation to the International Vaccine Institute. Copyright 2018 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. Introduction Despite progress in understanding of the epidemiology, pathogenesis, and treatment of cholera, and in providing access to clean water and adequate sanitation, Vibrio cholerae O1 continues to cause large outbreaks and remains endemic in many parts of the world, especially in the least privileged populations. 1 3 A major advance to address this persisting problem was the development and creation of a global stockpile of inexpensive, safe, and effective inactivated whole-cell oral cholera vaccines (OCVs). These OCVs are given as a two-dose regimen and confer protection against cholera for at least 5 years after dosing. 4 To date, two such vaccines, identical in composition but produced under different trade names (Shanchol by Shantha Biotechnics, Hyderabad, India, and Euvichol by Eubiologics, Seoul, South Korea) have been used in the stockpile. They have been deployed to control cholera in humanitarian crises, outbreaks, and Vol 18 June 2018

2 Research in context Evidence before this study To assess previous evidence on the clinical protection by a single dose of inactivated whole-cell oral cholera vaccine (OCV), we searched for publications in English on PubMed published between Jan 1, 1986, and April 1, 2017, using the search terms [ oral cholera vaccine OR cholera vaccine ] AND [ efficacy OR effectiveness ]. We identified six studies that assessed protection by a single dose. All but one of these studies tested a multidose regimen, with secondary analyses of individuals who received only a single dose. An early randomised trial of a three-dose regimen of two different inactivated whole-cell-based OCVs (whole cell only and whole cell with B subunit of cholera toxin) found that a single dose conferred 12 24% protection against cholera over 3 years of follow-up, depending on the vaccine. Other studies, conducted during mass immunisation campaigns, reported that the protection by a single dose was 33 67% at follow-up intervals ranging from 3 months to 3 years. None of these secondary analyses yielded estimates of protection whose 95% CIs excluded 0%. The only study designed with a primary objective of evaluating the protection by a single dose of OCV assessed a mass immunisation campaign with a single dose of the OCV Shanchol in South Sudan in a population that had experienced repeated cholera outbreaks in the years before the campaign. Although the study reported vaccine protective effectiveness as 87% (95% CI ), it was limited by its observational design, by a high proportion of in whom vaccine histories were based on recall alone, by a small number of cholera outcome events, and by only 2 months of post-vaccination follow-up. Added value of this study To our knowledge, our study is the only evaluation of an inactivated whole-cell-based OCV using a randomised, double-blind design powered to assess protection by a single dose. Moreover, our study is the only evaluation adequately designed to evaluate short-term (6 months) and long-term (2 years) protection, and provided estimates of protective efficacy at 2 years with sufficiently narrow confidence limits to exclude low protective efficacy in older children and adults, as well as clinically meaningful protection in children vaccinated at ages younger than 5 years. The post-hoc results of longer-term follow-up in our study show that vaccine protection is sustained for 2 years in people vaccinated at ages of 5 years and older, and confirm the findings of our earlier analysis of 6-month follow-up in which no protection was observed in children vaccinated when younger than 5 years. Implication of all the available evidence The use of a single-dose regimen for older children and adults or of a two-dose regimen despite a likelihood that circumstances could prevent delivery of a complete two-dose regimen to many targeted individuals (such as in humanitarian crises), might be rational. Our findings provide indirect support to the notion that boosting after the standard two-dose primary series with inactivated whole-cell-based OCV might require only a single dose. This possibility should be evaluated in future studies. Evaluations of either a single-dose or two-dose regimen of the inactivated whole-cell OCV in settings without endemic cholera should be investigated. Correspondence to: Dr Firdausi Qadri, International Centre for Diarrhoeal Disease Research (icddr,b), Mohakhali, Dhaka 1213, Bangladesh fqadri@icddrb.org cholera-endemic settings in several African countries, one country in the Middle East (Iraq), several countries in Asia, and in Haiti. 5 Evaluations of OCV deployments from the stockpile have found vaccination to be feasible and efficacious when given in a two-dose regimen. However, in humanitarian crises, natural calamities, and in resourcepoor settings with epidemic and endemic disease, giving the second dose can be challenging and immediate vaccine protection after the initial dose is highly desirable. 6,7 Because of these considerations and of data suggesting that the first dose of these OCVs elicits a robust serum vibriocidal antibody response, we did a large-scale, placebo-controlled randomised trial of a single dose of Shanchol in urban Bangladesh. 8 A single dose conferred 40% protection against all episodes of cholera and 63% protection against severe episodes during 6 months of follow-up. 9 Here, we report protective efficacy of OCV 2 years after vaccination. Methods Study design and This placebo-controlled, double-blind, individually randomised clinical trial of the inactivated whole-cell OCV Shanchol was done in children and adults in the urban slums of Dhaka, Bangladesh. 9 A census was done and then updated between Sept 10 and Dec 9, 2013, to enumerate and households and to ascertain geographical coordinates of each residence 10 before the start of the vaccination. Each individual in the study was identified by a unique ID that remained unchanged throughout the study period. The population was updated every 6 months, including at the end of the 2-year followup period, through demographic surveillance. Participants were eligible for inclusion if they gave written informed consent, were at least 1 year of age, were not pregnant by history (obtaining verbal history of last menstrual period), were apparently healthy, and had no history of receipt of OCV (self-reported when inquired by the study volunteer using the census questionnaire). The study was approved by the institutional review boards of the International Centre for Diarrhoeal Disease Research (icddr,b) and the International Vaccine Institute (IVI) and was monitored by a data and safety monitoring board. Written informed consent was obtained from all or their parents or guardians (for younger than 18 years) and assent was also obtained from all aged years. Vol 18 June

3 Randomisation and masking Randomisation of to vaccine or placebo was done on the basis of the census. Vaccine and placebo single-dose vials were randomly ordered within blocks of six, given consecutive unique numbers, and placed in boxes. The oral placebo was an identically appearing liquid with inert constituents. Teams were instructed to deliver a dose of vaccine or placebo to each successive eligible participant according to the numerical order of the vial in the box. The identities of the numbered vials were kept by designees at the manufacturer and IVI who were not otherwise involved in the trial. The on-site principal investigator also had the identity of each number in a sealed envelope, with the provision that opening of the envelope for unblinding of individual cases should be done only if needed for clinical management. No envelope was opened during the 2-year follow-up period. Procedures Each dose of OCV contained 1500 lipopolysaccharide (LPS) ELISA units of V cholerae O1 and 600 of O139 LPS (ratio of 5:2). A mass vaccination campaign was done from Jan 10 to Feb 4, 2014, to administer a single dose of OCV or oral placebo. At the time of vaccination, the ID given to each participant was checked and completeness of ingestion of the administered dose was recorded in a vaccine record book. Participants were monitored for 30 min after vaccination and then via passive surveillance for adverse events for 28 days. Active surveillance for adverse events was done for a subset of 6021 for 28 days after vaccination. Adverse events, both overall and severe, occurred at similar frequencies in the two groups. Detailed analyses of adverse events, which revealed no safety concerns, have been reported previously. 9 Passive surveillance for diarrhoea was initiated at two icddr,b hospitals and 11 major clinics located in or near the study area from Jan 10, 2014, for 2 years after the last administered dose. A diarrhoeal visit was defined as a hospital or clinic visit in which the patient reported having at least three loose or liquid stools in the 24 h before presentation; or at least one bloody loose or liquid stool; or one, two, or an indeterminate number of loose or liquid stools with some sign of dehydration according to the WHO criteria. 11 Severe diarrhoea, defined as a cholera episode presenting with severe dehydration in at least one constituent visit of the episode, was assessed on the basis of clinical features at the time of presentation for care. 11 Patients from the study area reporting with diarrhoea were identified as study at the patient registration desk at the surveillance site by their household ID cards (if presented at the time of registration) or by a computerised search of the database in the absence of the ID card. Patients were examined and treated by trained physicians and either stool samples or rectal swabs were collected after obtaining consent. The specimens were transported in Cary-Blair media to the icddr,b laboratory, where the specimens were tested for V cholerae by serogroup, biotype, and serotype, using conventional methods The ID of each patient whose faecal specimen yielded V cholerae O1 or V cholerae O139 was confirmed by a home visit within 14 days of pathogen isolation. Diarrhoeal visits within 7 days from the discharge date of the previous visit were considered a single diarrhoea episode, with the onset for the first visit being recorded as the onset of the episode. Outcomes The primary endpoint was an episode of non-bloody diarrhoea for which the onset was at least 7 days after vaccination, a faecal culture was positive for V cholerae O1 or O139, and the home visit confirmed that the person had sought care for diarrhoea on the date of presentation. Severe cholera and adverse events following dosing were secondary endpoints. Statistical analysis For the primary analysis, we estimated that at least individuals would be needed in each group, assuming a cholera risk of 0 56 episodes per 1000 people, a true vaccine protective efficacy of at least 50% during the first 6 months of follow-up, a loss to follow-up of 25% annually, and exclusion of a lower limit of a one-sided 95% CI of 10% with 80% power. The primary analysis as previously reported 9 assessed the vaccine protective efficacy during the first 6 months of follow-up. This secondary analysis of protective efficacy was done per protocol, and included only first episodes of cholera that began days after vaccination among who had ingested a full dose. We also did a post-hoc, intention-to-treat analysis that also included all individuals who received at least one dose of study treatment, and included the first episodes of cholera that occurred between 1 day and 730 days after the first dose. Prespecified individual-level and community-level baseline variables assumed to be potentially related to the risk of cholera were compared between vaccine recipients and placebo recipients using the χ² test (or Fisher s exact test) for categorical variables and Student s t test (or the Mann-Whitney U test for variables not following a normal distribution) for continuous variables. In crude analyses, we estimated rate ratios of first cholera episodes, divided by person-time of follow-up, in vaccinees versus placebo recipients and calculated p values and 95% CIs with test-based methods. 15 We also compared the risk of first episodes of cholera between vaccinees and placebo recipients with Kaplan-Meier survival analysis, censoring who migrated out or died, whichever came first. We used Cox proportional hazard models to estimate multivariable hazard ratios of first cholera episodes in vaccine recipients versus placebo recipients, adjusting for Vol 18 June 2018

4 covariates in the models, by exponentiating the coefficient for vaccination status, an independent variable, and estimated the p value and 95% CIs for the hazard ratio estimates using the SE of the fitted value of the coefficient for the vaccination variable. In these models, we considered as candidate covariates only those baseline variables known or assumed to affect the risk of cholera on clinical, biological, or epidemiological grounds. To avoid overfitting these models, we used a backward elimination algorithm to select covariates associated with time to event at p<0 10 and followed the rule of requiring at least ten events per independent variable in the model. 16 Vaccine protective efficacy was calculated as (1 hazard ratio) 100. We evaluated multivariable hazard ratios of initial episodes of all types of cholera days after vaccination. We also analysed the data by age group at vaccination (<5 years, 5 years to <15 years, and 15 years) as prespecified in the study protocol, as well as a post-hoc analysis of younger than 5 years compared with those at least 5 years of age. Heterogeneity of vaccine protective efficacy between age groups was assessed by analysing two-way interaction terms between the vaccination and age group variables in the models. Heterogeneity of vaccine protective efficacy over time was assessed by evaluating the fulfilment of the proportional hazards assumption for the exposure. All analyses were evaluated at significance level p=0 05 (two-tailed) and were done with SAS version 9.4. This study was registered with ClinicalTrials.gov, number NCT received vaccine aged <5 years old aged 5 years to <15 years aged 15 years or older assessed for eligibility were age eligible 488 excluded from per-protocol analysis 33 received two doses 53 had irregular randomisation number 402 received incomplete dose 6051 were not age eligible randomly assigned to vaccine or placebo were not included declined to participate 12 were pregnant* 1 was severely ill 464 had previously received oral cholera vaccine assigned to vaccine assigned to placebo 325 excluded from per-protocol analysis 36 received two doses 80 had irregular randomisation number 209 received incomplete dose received placebo 9765 aged <5 years old aged 5 years to <15 years aged 15 years or older Role of the funding source The funder of the study had no role in the study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication migrated out during follow-up 5706 aged <5 years old aged 5 years to <15 years aged 15 years or older 286 died during follow-up 8 aged <5 years old 12 aged 5 years to <15 years 266 aged 15 years or older migrated out during follow-up 5441 aged <5 years old aged 5 years to <15 years aged 15 years or older 285 died during follow-up 11 aged <5 years old 14 aged 5 years to <15 years 260 aged 15 years or older Results Between Jan 10 and Feb 4, 2014, people were randomly assigned to receive either vaccine or placebo, of whom were included in the per-protocol analysis. Of these, migrated out and 571 died during the 2-year follow-up (figure). A total of 287 first episodes of cholera were detected during the 2-year follow-up period, of which 138 episodes were severe. 96 episodes of cholera and 25 episodes of cholera with severe dehydration occurred among younger than 5 years and 191 episodes of cholera and 113 episodes of cholera with severe dehydration occurred among at least 5 years of age. 286 (>99%) with cholera had at least three loose or liquid stools, and only one had an indeterminate number of loose or liquid stools with at least one sign of severe dehydration according to the WHO criteria. No episodes began during the 6 days following dosing. All cases were V cholerae O1 El Tor biotype: Figure: Trial profile *The low number of pregnant women represents those presenting for vaccination at the vaccination centre; most pregnant women did not present for vaccination because they were told not to do so at the time of the census. Refers to duplicated randomisation number on different vaccine vials, absence of a sticker number for a vaccinated individual in the vaccination register, a sticker number that was out of range ( ), or two different sticker numbers on the same vial. 273 (95%) isolates were Ogawa serotype and the remainder were Inaba serotype. 255 (89%) of the cholera episodes were treated at the icddr,b s two hospitals and 32 (11%) were treated at 11 major clinics inside or nearby the study area. 136 (99%) patients with episodes of severe cholera were admitted to the icddr,b s hospitals. The monthly distribution of all analysed cholera episodes is shown in the appendix (pp 9, 12). Each year of the study exhibited a bimodal seasonal pattern of cholera, with peaks in spring (April May) and autumn (August October). Baseline characteristics were similar in each group (table 1). The baseline features of who See Online for appendix Vol 18 June

5 Vaccine recipients Placebo recipients Individual-level variables Number of Age on the day on which 23 9 (16 1) 24 1 (16 1) vaccine or placebo was received, years Female (53 9%) (53 8%) Male (46 1%) (46 2%) Diarrhoea in the previous (10 4%) (10 3%) 6 months Number of residents in the 4 75 (1 8) 4 74 (1 8) household Living in own house (17 8%) (17 8%) Duration living in the study 44 3 (77 7) 44 4 (77 9) area, months Living in household with 8492 (8 3%) 8503 (8 3%) improved water source* Living in households using (59 4%) (59 4%) treated water for drinking Living in households with (72 7%) (72 4%) sanitary toilets Living in household in which (94 1%) (94 0%) residents washed hands with soap and water Monthly per-capita (1651 8) (1549 3) expenditure of household, Bangladeshi taka Distance from household to (620 8) (623 1) the nearest health facility, m Distance from household to (766 0) (763 5) the previous intervention area, m Community-level variables People living in households 73 6% (22 0) 73 5% (22 0) using sanitary toilets People living in households 94 0% (7 7) 94 0% (7 8) in which residents washed hands with soap and water People living in households 8 3% (11 3) 8 3% (11 3) with an improved water source Population density per 100 m² 5 6 (3 2) 5 6 (3 2) Data are n, n (%), or mean (SD). *An improved water source was defined as a tap, well, or hand pump. Water that was boiled, filtered, or chlorinated was considered to have been treated. US$ Bangladeshi taka. The previous intervention area is the area in which an oral cholera vaccine campaign was conducted in Community-level variables considered people living within a prespecified 100 m radius around the household, as measured by a geographical information system. Table 1: Baseline individual-level and community-level characteristics of vaccine and placebo recipients in the per-protocol analysis migrated out or died during the 2-year follow-up period also seemed well balanced between the two groups (appendix). The overall incidence rates of initial cholera episodes were 0 22 (95% CI 0 18 to 0 27) per persondays in vaccine recipients versus 0 36 (0 31 to 0 42) per person-days in placebo recipients (adjusted protective efficacy 39%, 95% CI 23 to 52; table 2). The overall incidence of severe cholera was 0 09 (0 07 to 0 12) per person-days versus 0 19 (0 15 to 0 23; adjusted protective efficacy 50%, 29 to 65). The incidence rates of cholera episodes were similar in vaccinees and placebo recipients younger than 5 years (table 2; appendix). In this age group, the adjusted protective efficacy was 13% ( 68 to 25) against all cholera episodes and 44% ( 220 to 35) against severe cholera episodes. However, in aged 5 years to younger than 15 years at vaccination, adjusted protective efficacy was 52% (8 to 75) against all cholera and 71% (27 to 88) against severe cholera. Likewise, for aged 15 years or older, protective efficacy was 59% (42 to 71) against all cholera and 59% (35 to 74) against severe cholera (table 2). In a post-hoc analysis of all older children and adults (age 5 years), the incidence of all cholera episodes was significantly lower for vaccine recipients than for placebo recipients (table 2). In a post-hoc analysis of aged 5 years or older, the event-free survival curves show sustained vaccine protection against cholera and severe cholera over the 2-year follow-up (appendix pp 10 11). The survival curve among younger than 5 years showed no significant vaccine protection (appendix pp 13 14). In multivariable models that adjusted for important covariates in the 5-years-or-older age group, protective efficacy was 57% (42 69; p<0 0001) against all cholera episodes and 62% (43 75; p<0 0001) against severe cholera episodes during 2 years of follow-up (table 2). Estimates of vaccine protective efficacy over 2 years of follow-up were significantly higher in people aged 5 years or older than in younger children both for all episodes of cholera and for severe episodes of cholera (table 2). In a post-hoc analysis of older children and adults by year of follow-up, vaccine protective efficacy was 50% (95% CI 23 67) against all cholera episodes and 54% (16 75) against severe cholera episodes in the first year of follow-up, and 64% (43 77) against all cholera episodes and 67% (43 81) against severe cholera episodes during the second year of follow-up. No heterogeneity of vaccine protection with time following vaccination was noted either for all cholera episodes or for severe cholera episodes (table 2). In the post-hoc intention-to-treat analysis, vaccine protective efficacy was 39% (95% CI 23 52) against cholera and 50% (29 65) against severe dehydrating cholera (appendix p 3). Discussion Our study found that a single dose of the inactivated whole-cell OCV provides overall protection in all age groups of 39% at 2 years of follow-up. Our analyses by age group showed no significant protective effect in children younger than 5 years; however, a single dose of the vaccine administered to older children and adults Vol 18 June 2018

6 conferred 57% protection against all cholera episodes and 63% protection against cholera with severe dehydration over 2 years of follow-up. The lower protective efficacy in young children might be explained by their lower prevaccination exposure to and immunological priming by natwural cholera infections, as well as the nature of the protective antigen in the OCV, which might require immunological priming. Levels of protective efficacy were lower both in young children and in older people in this trial than in a placebo-controlled, randomised clinical trial 4 of a two-dose regimen of the same OCV after 2 years (49% for children younger than 5 years and more than 63% in the older age group), 17 but were still substantial in size for older children and adults. Of studies that have assessed the clinical protection by inactivated whole-cell-based OCVs against naturally occurring cholera, using either experimental or nonexperimental study designs, we identified six that assessed protection by a single dose. All but one of these studies tested a multidose regimen, with secondary analyses of individuals who received only a single dose owing to failure to complete the two-dose regimen. An early randomised trial 18 of a three-dose regimen of two different inactivated whole-cell-based OCVs (whole cell only and whole cell with B subunit of cholera toxin), both of which had a lower O1 serogroup LPS content than Shanchol and did not contain whole-cell O139, found that a single dose conferred 12 24% protection against cholera over 3 years of follow-up, depending on the vaccine. Other studies, which were observational analyses of mass immunisation programmes with Shanchol or of the whole-cell-with-b-subunit vaccine Dukoral, reported that protection by a single dose was 33 67% at follow-up intervals ranging from 3 months to 3 years None of these analyses yielded estimates of protection whose 95% CIs excluded 0% to assess protection by a single dose. The only study designed with a primary objective of evaluating the protection by a Vaccine recipients Placebo recipients Protective efficacy* Number of Cholera episodes per person-days of follow-up Incidence per person-days Number of Cholera episodes per person-days of follow-up Incidence per person-days Crude estimate p value Adjusted estimate All cholera episodes All / / % < % < (0 18 to 0 27) (0 31 to 0 42) (23 to 52) (23 to 52) Age group <5 years / (0 85 to 1 46) / (0 73 to 1 32) 13% ( 69 to 24) % ( 68 to 25) years / (0 10 to 0 17) 5 years to <15 years / (0 07 to 0 19) 15 years / (0 10 to 0 18) Duration of follow-up among aged 5 years 180 days / (0 07 to 0 19) days / (0 08 to 0 21) days / (0 11 to 0 28) days / (0 05 to 0 19) Duration of follow-up among 5 years or older 365 days / (0 08 to 0 17) days / (0 10 to 0 20) / (0 25 to 0 36) / (0 16 to 0 33) / (0 27 to 0 40) / (0 20 to 0 38) / (0 13 to 0 30) / (0 35 to 0 62) / (0 21 to 0 44) / (0 19 to 0 31) / (0 31 to 0 49) 57% (42 to 69) 51% (7 to 74) 59% (42 to 71) 58% (24 to 76) 37% ( 19 to 67) 62% (34 to 78) 67% (29 to 84) 50% (23 to 67) 64% (43 to 77) < % (42 to 69) % (8 to 75) < % (42 to 71) % (24 to 76) % ( 20 to 67) % (34 to 78) % (30 to 84) p value <0 0001; p for subgroups < % (23 to 67) < % < (43 to 77) (Table 2 continues on next page) Vol 18 June

7 Vaccine recipients Placebo recipients Protective efficacy* Number of Cholera episodes per person-days of follow-up Incidence per person-days Number of Cholera episodes per person-days of follow-up Incidence per person-days Crude estimate p value Adjusted estimate (Continued from previous page) Severe cholera All individuals / / % % (0 07 to 0 12) (0 15 to 0 23) (29 to 65) (29 to 65) Age group <5 years / (0 19 to 0 53) / (0 12 to 0 42) 44% ( 220 to 36) % ( 220 to 35) years / (0 05 to 0 10) 5 years to <15 years / (0 02 to 0 11) 15 years / (0 05 to 0 12) Duration of follow-up among aged 5 years 180 days / (0 02 to 0 11) days / (0 03 to 0 13) days / (0 06 to 0 21) days / (0 02 to 0 13) Duration of follow-up among aged 5 years 365 days / (0 03 to 0 10) days / (0 05 to 0 14) / (0 15 to 0 23) / (0 10 to 0 25) / (0 15 to 0 25) / (0 11 to 0 25) / (0 04 to 0 15) / (0 22 to 0 44) / (0 14 to 0 33) / (0 09 to 0 18) / (0 20 to 0 35) 62% (43 to 75) 71% (27 to 88) 59% (35 to 74) 70% (29 to 87) 20% ( 103 to 68) 63% (27 to 82) 74% (29 to 90) 54% (16 to 75) 67% (43 to 81) < % (43 to 75) % (27 to 88) % (35 to 74) % (29 to 87) % ( 103 to 68) % (27 to 82) % (30 to 90) % (16 to 75) % (43 to 81) p value <0 0001; p for subgroups *Two-tailed 95% CIs and p values are given for the analyses. Adjustments are described in the appendix. Calculated for the overall interaction between vaccine or placebo recipients and the subgroup variable in the model. p value is for 5 years vs 5 years. Calculated for the overall interaction between vaccine or placebo recipients and the subgroup variable in the model. p for subgroups is for 5 years to <15 years vs 15 years. Constancy of vaccine protection over time. The number of outcomes was insufficient for the adjusted model. Table 2: Incidence of cholera and severe cholera, and the protective efficacy of the vaccine in per-protocol analyses single dose of OCV assessed a mass immunisation campaign with a single dose of Shanchol in South Sudan in a population that had experienced repeated cholera outbreaks in the years before the campaign. 23 Although the study reported vaccine protective effectiveness as 87% (95% CI ), it was limited by its observational design, by a high fraction of in whom vaccine histories were based on memory recall alone, by a small number of cholera outcome events, by only 2 months of post-vaccination follow-up, and by the fact that a case-control analysis of the same population detected substantially lower protective efficacy. To our knowledge, the study reported herein is the only doubleblind, randomised controlled trial to date that has evaluated the protective efficacy of a single dose of inactivated whole-cell OCV as its primary objective. Our study had several potential limitations. First, the trial was done in a cholera-endemic setting where the population is continually exposed to cholera and thereby acquires natural immunity. Thus, our findings might not be generalisable to populations without natural immunity to cholera. More research is needed on the performance of both single-dose and two-dose regimens of inactivated whole-cell OCVs in such populations. Second, we used a passive surveillance system for detecting cholera cases, which might have missed some cases, particularly those that were less severe and thereby provided estimates of overall vaccine protection that were weighted towards Vol 18 June 2018

8 protection against severe cholera. Third, our estimates of protective efficacy were for the compliant study population in per-protocol analysis; however, estimates in the intention-to-treat analysis, which considered all who ingested a dose regardless of the amount ingested and the first episode of cholera with onset from the day 1 after the first dose, were similar to those for the per-protocol analysis. Our findings have several implications. The failure of a single dose of OCV to protect young children against cholera contrasts with the robust serum vibriocidal antibody response observed 14 days after the receipt of a first dose by young children. 8 However, titres of serum vibriocidal antibodies, the mainstay immunological test used in clinical evaluations of OCVs, are not immunological correlates of protection in cholera endemic settings: robust serum vibriocidal antibody responses were seen in a large-scale field trial 24 of the live OCV CVD103 Hg-R, but the vaccine was not found to protect against cholera in Jakarta, Indonesia, and postvaccination titres of serum vibriocidal antibodies have been shown to fall rapidly despite continuing protection against cholera by inactivated whole-cell OCVs in Bangladesh. 25 Although our results cannot be readily extrapolated to populations without endemic cholera, and do not support routine dosing with a single dose of this OCV in young children in cholera-endemic settings, the findings do have practical relevance. The global stockpile of inactivated whole-cell OCV has often been deployed in settings with natural or man-made humanitarian disasters, in which completion of a two-dose regimen could be challenging. Our results indicate that when such emergencies occur in populations with endemic cholera, the challenges of successfully com pleting a two-dose regimen should not deter deployment of OCV, because older children and adults will be protected for at least 2 years by a single dose. Also, because the global production of inactivated whole-cell OCV is limited, a scarcity of doses might make it desirable in choleraendemic settings to deploy a single-dose regimen to older children and adults, reserving a two-dose regimen for young children. Finally, because the age-related differences in vaccine protection in our study might have been due to a greater pre-existing anticholera natural immunity in the older, our findings provide indirect support to the notion that boosting after the standard two-dose primary series with inactivated wholecell OCVs might require only a single dose. This hypothesis should be evaluated in future studies. Contributors FQ, MAl, AC, APS, TFW, and JDC designed the study. FQ, JDC, AIK, FC, AS, IAK, and MIC were involved with implementation, delivery, and supervision of the study. NCS, AUS, and AKa collected and managed the data system and FQ, JDC, JL, DRK, and MAl analysed the data. YAB, FK, and TRB participated in the microbiological part of the study. All authors participated in the writing of the manuscript and had full access to the data in the study. All authors revised drafts and approved the final version of the manuscript. Declaration of interests We declare no competing interests. Acknowledgments This study was supported by a grant from the Bill & Melinda Gates Foundation to the International Vaccine Institute (IVI) and by core grants from the Governments of Bangladesh, Canada, Sweden, and the UK to the International Centre for Diarrhoeal Disease Research (icddr,b) and from the governments of South Korea and Sweden to IVI. We are grateful to the members of the data and safety monitoring board, the staff members of icddr,b and IVI for their work in the field and laboratory and for data management and the residence of Mirpur for their participation in the trial. References 1 Nair GB, Faruque SM, Bhuiyan N, Kamruzzaman M, Siddique A, Sack DA. New variants of Vibrio cholerae O1 biotype El Tor with attributes of the classical biotype from hospitalized patients with acute diarrhea in Bangladesh. J Clin Microbiol 2002; 40: Ali M, Nelson AR, Lopez AL, Sack DA. Updated global burden of cholera in endemic countries. PLoS Negl Trop Dis 2015; 9: e WHO, UNICEF. Progress on drinking water and sanitation: joint monitoring programme update. Geneva: World Health Organization and the United Nations Children s Fund, Bhattacharya SK, Sur D, Ali M, et al. 5 year efficacy of a bivalent killed whole-cell oral cholera vaccine in Kolkata, India: a cluster-randomised, double-blind, placebo-controlled trial. Lancet Infect Dis 2013; 13: Desai SN, Pezzoli L, Martin S, et al. A second affordable oral cholera vaccine: implications for the global vaccine stockpile. Lancet Glob Health 2016; 4: e Parker LA, Rumunu J, Jamet C, et al. Adapting to the global shortage of cholera vaccines: targeted single dose cholera vaccine in response to an outbreak in South Sudan. Lancet Infect Dis 2017; 17: e Hsiao A, Desai SN, Mogasale V, Excler J-L, Digilio L. Lessons learnt from 12 oral cholera vaccine campaigns in resource-poor settings. Bull World Health Organ 2017; 95: Kanungo S, Paisley A, Lopez AL, et al. Immune responses following one and two doses of the reformulated, bivalent, killed, whole-cell, oral cholera vaccine among adults and children in Kolkata, India: a randomized, placebo-controlled trial. Vaccine 2009; 27: Qadri F, Wierzba TF, Ali M, et al. Efficacy of a single-dose, inactivated oral cholera vaccine in Bangladesh. N Engl J Med 2016; 374: Khan IA, Saha A, Chowdhury F, et al. Coverage and cost of a large oral cholera vaccination program in a high-risk cholera endemic urban population in Dhaka, Bangladesh. Vaccine 2013; 31: WHO. The treatment of diarrhoea: a manual for physicians and other senior health workers. Geneva: World Health Organization, Qadri F, Azim T, Chowdhury A, Hossain J, Sack RB, Albert MJ. Production, characterization, and application of monoclonal antibodies to Vibrio cholerae O139 synonym Bengal. Clin Diagn Lab Immunol 1994; 1: Qadri F, Jonson G, Begum YA, et al. Immune response to the mannose-sensitive hemagglutinin in patients with cholera due to Vibrio cholerae O1 and O0139. Clin Diagn Lab Immunol 1997; 4: Monsur KA. A highly selective gelatin-taurocholate-tellurite medium for the isolation of Vibrio cholerae. Trans Royal Soc Trop Med Hyg 1961; 55: Rothman KJ. Epidemiology: an introduction. Oxford: Oxford University Press, 2012: Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol 1996; 49: Sur D, Lopez AL, Kanungo S, et al. Efficacy and safety of a modified killed-whole-cell oral cholera vaccine in India: an interim analysis of a cluster-randomised, double-blind, placebo-controlled trial. Lancet 2009; 374: Clemens JD, Sack DA, Harris JR, et al. Field trial of oral cholera vaccines in Bangladesh: results from three-year follow-up. Lancet 1990; 335: Vol 18 June

9 19 Khatib AM, Ali M, von Seidlein L, et al. Effectiveness of an oral cholera vaccine in Zanzibar: findings from a mass vaccination campaign and observational cohort study. Lancet Infect Dis 2012; 12: Wierzba TF, Kar SK, Mogasale VV, et al. Effectiveness of an oral cholera vaccine campaign to prevent clinically-significant cholera in Odisha State, India. Vaccine 2015; 33: Luquero FJ, Grout L, Ciglenecki I, et al. Use of Vibrio cholerae vaccine in an outbreak in Guinea. N Engl J Med 2014; 370: Ivers LC, Hilaire IJ, Teng JE, et al. Effectiveness of reactive oral cholera vaccination in rural Haiti: a case-control study and bias-indicator analysis. Lancet Glob Health 2015; 3: e Azman AS, Parker LA, Rumunu J, et al. Effectiveness of one dose of oral cholera vaccine in response to an outbreak: a case-cohort study. Lancet Glob Health 2016; 4: e Richie E, Punjabi NH, Sidharta Y, et al. Efficacy trial of single-dose live oral cholera vaccine CVD 103-HgR in North Jakarta, Indonesia, a cholera-endemic area. Vaccine 2000; 18: Sack D, Clemens J, Huda S, et al. Antibody responses after immunization with killed oral cholera vaccines during the 1985 vaccine field trial in Bangladesh. J Infect Dis 1991; 164: Vol 18 June 2018