Post-licensure deployment of oral cholera vaccines: a systematic review

Abstract Objective To describe and analyse the characteristics of oral cholera vaccination campaigns; including location, target population, logistics, vaccine coverage and delivery costs. Methods We searched PubMed, the World Health Organization (WHO) website and the Cochrane database with no date or language restrictions. We contacted public health personnel, experts in the field and in ministries of health and did targeted web searches. Findings A total of 33 documents were included in the analysis. One country, Viet Nam, incorporates oral cholera vaccination into its public health programme and has administered approximately 10.9 million vaccine doses between 1997 and 2012. In addition, over 3 million doses of the two WHO pre-qualified oral cholera vaccines have been administered in more than 16 campaigns around the world between 1997 and 2014. These campaigns have either been pre-emptive or reactive and have taken place under diverse conditions, such as in refugee camps or natural disasters. Estimated two-dose coverage ranged from 46 to 88% of the target population. Approximate delivery cost per fully immunized person ranged from 0.11–3.99 United States dollars. Conclusion Experience with oral cholera vaccination campaigns continues to increase. Public health officials may draw on this experience and conduct oral cholera vaccination campaigns more frequently.


Introduction
Vibrio cholerae O1 and O139 causes severe diarrhoea and the main strategies to prevent the disease are to promote hygiene and to ensure safe water and sanitation. These basic needs are often not met in endemic areas with seasonal cholera outbreaks or during man-made or natural disasters in impoverished areas. An additional tool for cholera prevention and control is the oral cholera vaccine. In October 2009, the World Health Organization (WHO) Strategic Advisory Group of Experts on immunization recommended that oral cholera vaccination should be considered as a reactive strategy during outbreaks, in addition to the already recommended preventive use of oral cholera vaccine in endemic areas. 1 A vaccine stockpile was created in 2012, with an initial two million doses to be available mainly for epidemic response in low-income countries. 2 In November 2013, the global alliance for vaccines and immunizations (Gavi Alliance) approved a financial contribution towards the stockpile to expand its use. With the availability of the oral cholera vaccine stockpile, more governments might consider cholera vaccination where needed.
A monovalent inactivated vaccine containing killed whole-cells of V. cholerae serogroup O1 and the B-subunit of cholera toxin was the first oral cholera vaccine to obtain international licensure in 1991 and WHO prequalification in 2001. The vaccine is marketed as Dukoral® (Crucell, Netherlands). Randomized, placebo-controlled trials of earlier versions of Dukoral® in Bangladesh and the current recombinant B-subunit whole cell vaccine in Peru showed that the vaccine is safe and confers an initial protection of approximately 85% in the first months. 3,4 Follow-up studies in Bangladesh estimated a 62% protection during the first year, 57% during the second year and negligible thereafter. 3 During the mid-1980s, the National Institute of Hygiene and Epidemiology in Viet Nam developed an oral cholera vaccine for the country's public health programme. A twodose regimen of a first-generation of monovalent (anti-O1) cholera vaccine had an estimated efficacy of 66% against the El Tor strain of V. cholerae. 5 In 1997, the vaccine was augmented with killed V. cholerae serogroup O139 whole cells to create a bivalent vaccine, 6 which was locally licensed as ORC-Vax™ (Vabiotech, Viet Nam). After changing production procedures in 2009, the vaccine was reformulated and licensed as mORC-Vax™ (Vabiotech, Viet Nam) and is currently used in Viet Nam's public health programme. 7 However, the vaccine is not pre-qualified by WHO.
To make the mORC-Vax™ internationally available, manufacture of the reformulated vaccine was transferred to Shantha Biotechnics Ltd in India, where the national regulatory authority is approved by WHO. 8 This led to the development of Shanchol™, which is the third currently-available oral cholera vaccine. A randomized, placebo-controlled trial in India showed that Shanchol™ is safe and confers 67% protective efficacy against cholera within two years of vaccination, 8 66% at three years 9 and 65% at five years 10 of follow-up. Shanchol™ was licensed in India in 2009 and received WHO pre-qualification in 2011.
Objective To describe and analyse the characteristics of oral cholera vaccination campaigns; including location, target population, logistics, vaccine coverage and delivery costs. Methods We searched PubMed, the World Health Organization (WHO) website and the Cochrane database with no date or language restrictions. We contacted public health personnel, experts in the field and in ministries of health and did targeted web searches. Findings A total of 33 documents were included in the analysis. One country, Viet Nam, incorporates oral cholera vaccination into its public health programme and has administered approximately 10.9 million vaccine doses between 1997 and 2012. In addition, over 3 million doses of the two WHO pre-qualified oral cholera vaccines have been administered in more than 16 campaigns around the world between 1997 and 2014. These campaigns have either been pre-emptive or reactive and have taken place under diverse conditions, such as in refugee camps or natural disasters. Estimated two-dose coverage ranged from 46 to 88% of the target population. Approximate delivery cost per fully immunized person ranged from 0.11-3.99 United States dollars. Conclusion Experience with oral cholera vaccination campaigns continues to increase. Public health officials may draw on this experience and conduct oral cholera vaccination campaigns more frequently.

Oral cholera vaccination campaigns
Stephen Martin et al.
A comparison of the three oral cholera vaccines is shown in Table 1. 11,12 The safety, relative effectiveness and duration of protection of the different types of oral cholera vaccine has previously been reviewed. 13 Here we conduct a systematic review of post-licensure oral cholera vaccines. The objective of the review is to generate information -by describing and analysing the campaigns -that can be used to inform planning for the future use of these vaccines.

Search
We searched the Cochrane database of systematic reviews and its database of abstracts and reviews of effects from 1990 to the present and found no reviews of oral cholera vaccination campaigns.
We conducted a systematic review of published documents on post-licensure vaccination campaigns using one of three oral cholera vaccines following the search and analysis process recommended in the Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines. We searched PubMed and the WHO website using "cholera vaccination", "cholera outbreak response" and "cholera vaccination campaign" as search terms with no date or language restrictions. The bibliographies of the retrieved articles were also screened for relevant papers. Reports, presentations and international organization or company documents were obtained through targeted web searches. We also contacted public health personnel, experts in the field and in ministries of health for further information.
All identified documents in English that described campaigns using oral cholera vaccine were assessed for appropriateness using the following selection criteria. We included all documents describing campaigns using Du-koral® after 1991, ORC-Vax™ after 1997, mORC-Vax™ after 2009 and Shanchol™ after 2009. Campaigns organized either as part of a public health response to endemic or epidemic cholera, pilot campaigns, demonstration projects, assessments of feasibility and acceptability, as well as studies of vaccine effectiveness were included. Each campaign may have more than one reference, describing different aspects of the vaccination (e.g. feasibility, coverage, cost, etc.). We excluded documents describing prelicensure trials, reports on knowledge and perception of cholera and oral cholera vaccines, as well as planning or policy briefs that did not describe actual oral cholera vaccine deployment.
By adhering to the pre-defined inclusion and exclusion criteria, we could make a valid comparison across articles. To assess the broad picture of the vaccine campaigns, we did not exclude any document based on quality or deficiency of reporting. Information from the published and unpublished documents was extracted and entered into a spreadsheet independently by two of the authors and then corroborated and summarized by a third author.

Definitions
Oral cholera vaccine campaigns can either be pre-emptive or reactive. Preemptive or preventive vaccination refers to campaign implementation before a cholera outbreak begins, ideally in conjunction with improved water, hygiene and sanitation. Pre-emptive vaccination may be conducted before the next seasonal outbreak in sites where cholera regularly occurs, in communities adjacent to an area with cholera or during humanitarian emergencies to prevent cholera. Reactive campaigns are those implemented after a cholera outbreak has started and while cholera cases are still being detected in the target population. 14 In areas where cholera tends to occur all year-round, the distinction between pre-emptive and reactive vaccination may be difficult.
The target population was defined as the number of individuals living in a circumscribed area to whom oral cholera vaccine is offered. The target population may be an estimate based on administrative population figures or a more precise figure based on a study census. Coverage was defined as the percentage of the target population who received one dose and two doses (fully immunized) of the vaccine, except when otherwise indicated (i.e. community surveys were used to calculate vaccine coverage in some campaigns particularly when a precise target population number was not known). The approximate total number of oral cholera vaccine doses deployed was defined as the sum of the first and second dose recipients; when data on the first dose recipients were not available, we multiplied the number of fully vaccinated individuals by two. We plotted the number of approximate doses deployed in oral cholera vaccine campaigns by country. Countries were colour-coded by the number of cholera cases reported in 2005, 15   Redlands, USA). Adverse events following immunization were defined as medical incidents that take place after an immunization and cause concern. Adverse events following immunization may be coincidental or causally associated. A serious adverse event following immunization is one that requires hospitalization and/or causes birth defects, permanent damage, or death.
To allow comparison of the expenses for vaccination across various campaigns, the expenses were grouped into the following categories: vaccine and/or international shipment costs, computers and other capital expenses, international consultants, local storage and transport, meetings, social mobilization, training, local salaries, supplies and waste management and the detection and management of adverse events following immunization. The delivery cost per fully immunized person was calculated using the total local expenses (excluding vaccine, international shipment and consultant costs) as the numerator and the number of fully immunized persons as the denominator.

Results
We identified 173 unique documents of potential relevance and 33 of these met the inclusion criteria ( Fig. 1).  In addition, we obtained information about recent campaigns through personal communications with two co-authors (DL and KA). We mapped the approximate number of doses administered in post-licensure oral cholera vaccination campaigns from 1997 to 2014 (Fig. 2) and plotted them by year (Fig. 3). As of August 2014, 280 000 oral cholera vaccine doses from the stockpile were shipped to Ethiopia, 280 000 to Guinea, 400 000 to Haiti and 300 000 to South Sudan. For campaigns with detailed data available, the characteristics and main findings are shown in Table 2 and the vaccination logistics by target population size is shown in Table 3.
The duration of the vaccination campaigns ranged from one to five months and consisted of two rounds at a 10-to 14-day interval (Table 3). Each round took 4 to 15 days. 16,20,23,24,26 A cold chain for vaccine delivery was reportedly maintained at 2-8 °C from storage to administration in Aceh, Indonesia, 24 Beira, Mozambique 20 and Zanzibar, United Republic of Tanzania. 26 In Uganda, the vaccine was maintained at room temperature. 16 Vaccination teams were able to vaccinate 100 to 1735 persons per day. 16,20,23,24,26 Reported adverse events following immunization in Mozambique 20 and Uganda 16 were minor and non-specific. Delivery cost per fully immunized person ranged from 0.53 United States dollars (US$) to US$ 3.66 (Table 4).

ORC-Vax™ and mORC-Vax™
In Viet Nam, an estimated 10.9 million doses of ORC-Vax™ and mORC-VAX™ have been deployed from 1997 to 2013 through targeted mass vaccination or -to children -through the Expanded Programme of Immunization in cholera-endemic regions. [30][31][32][33] Documented coverage during the vaccination of half of the communes in Hue was 79% (118 703/149 557) in 1998 and 75% (103 226/137 082) in the other half in 2000; long term vaccine effectiveness (three to five years after the campaign) was 50%. 30,31 (Table 2).Vaccine coverage was not precisely quantified in the 2008 Hanoi campaign; vaccine effectiveness was 76%. 32 The duration of the vaccination campaigns ranged from two to four weeks with each round taking 3 to 9 days (Table 3). [30][31][32] Mass campaigns are held yearly in Hue and are part of the routine public health provision, requiring minimal additional costs. The delivery cost in Hue during a 2013 campaign was US$ 0.11 per fully immunized person. 33

Shanchol™
Since WHO pre-qualification, Shan-chol™ has been increasingly used in campaigns. [34][35][36][37][38][39][40][41][42][43][44][45][46][47][48] About 2 649 189 doses have been administered in more than 10 campaigns (Table 2; data from the most recent campaigns in Ethiopia, Guinea and Haiti are not yet available), three of which were described as reactive. The percentage of fully immunized persons ranged from approximately 46-85% (Table 2). A study in Odisha, India 2011, found that oral cholera vaccination through the Indian public health system is feasible. 34 The campaign in Dhaka, Bangladesh 2011, includes an assessment of vaccine effectiveness with and without other interventions. 35 The two vaccination campaigns in Haiti in 2012 were pilot projects that paved the way for the launching of a national  cholera vaccination programme integrated in a long-term plan to address water safety and sanitation. [36][37][38][39][40] There was a third campaign in Haiti in 2013 that was part of this plan. Shanchol™ was deployed for pre-emptive vaccination in the Solomon Islands in 2012, following reports of cholera in a nearby area. 41 The vaccination campaign in Thailand, 2012, was conducted to prevent seasonal outbreaks in a stable camp setting. 42 The vaccination campaign in Guinea, 2012, was the first reactive oral cholera vaccine campaign in sub-Saharan Africa and the first time that Shanchol™ was used in an African setting. [43][44][45] The campaigns in Guinea and in Maban county, South Sudan 2013 confirmed that large-scale vaccinations under logistically difficult conditions are feasible. 46,47 The campaign in internally displaced persons camps in South Sudan in 2014, was the first to use the oral cholera vaccine stockpile. 48 The Shanchol™ campaigns were conducted in 1-3 months. [34][35][36][37][38][39][40][41][42][43][44][45][46][47][48] The 2012 Haiti campaign was carried out in two phases due to an overlapping national oral polio vaccination campaign. [36][37][38][39][40] The number of persons vaccinated per day ranged from 774-1150. 35,[43][44][45][46][47][48] No serious adverse events following immunization were reported. In campaigns in Odisha, Dhaka and in Haiti in 2012, acold chain for vaccine was maintained at 2-8 °C from storage to delivery on site. [34][35][36][37][38][39][40] In the campaigns in Guinea and in 2013 in South Sudan cold chain was maintained until the day of vaccination, during which vaccines were transported to vaccination sites and used at ambient temperature [43][44][45][46][47] (Table 3).
The delivery costs of Shanchol™ through the existing government health system in Bangladesh 35 and India 34 were US$ 1.63 and US$ 1.13, respectively, per fully immunized person. The local expenses of reactive deployment in Guinea were US$ 1.97, 45 while costs in Maban, South Sudan were US$ 3.99 per fully immunized person (Table 4). 47

Discussion
We estimate that about 3 175 206 doses of Dukoral® and Shanchol™ have been deployed in vaccination campaigns in areas affected by cholera around the world from 1997 to 2014. Only one country, Viet Nam, incorporates oral cholera vaccination into its public health programme and has used more than 10 million doses since 1997. Recently larger numbers of doses have been deployed in different areas globally but the vaccine is still under-used compared to the 1.4 billion people at risk of cholera in endemic areas. 15 There is a shortage of licensed, WHO-prequalified cholera vaccines to meet global endemic and epidemic needs and insufficient supply is often cited as an obstacle to wider vaccine use. 49 Availability of an oral cholera vaccine stockpile may lead to a larger vaccine supply through more consistent and predictable demands and may help increase vaccine use. Insufficient vaccine supply can be addressed by encouraging manufacturers to increase production capacity.
The deployments of oral cholera vaccine have previously been pre-emptive but recent experiences in Guinea [43][44][45] and Haiti [36][37][38][39][40] have shown that reactive mass vaccinations are feasible. , The number of cases and deaths that can be prevented by reactive vaccination depends on the characteristics of the outbreak, with greatest impact during large and long-lasting outbreaks usually seen in populations with no recent exposure to the disease. 14 With the development of an oral cholera vaccine stockpile and possibility of rapid deployment, increased reactive use of oral cholera vaccine is anticipated.
To be able to compare the campaigns, we calculated the total delivery cost per fully immunized person by excluding the expenditures for vaccine, shipment and technical experts, but the estimates still varied considerably. Deployment costs were lowest in Hue, Viet Nam, where the vaccine is administered routinely through the public health system 30,33 but a similar delivery strategy may not be possible in other choleraendemic areas or during the acute phase of emergencies. The requirement for co-administration of a buffer with the Dukoral® vaccine complicates the delivery of such vaccine and likely increases its delivery costs. Both mORC-Vax™ and Shanchol™ do not require a buffer, which should streamline the delivery and reduce logistical requirements. This analysis has several limitations. First, there was a wide variation in the methods used to calculate coverage and costs in the vaccination campaigns. Some coverage estimations were precise, while others were approximations. Although we attempted to make the costing comparable, the calculated figures should be interpreted with caution. There are large variations in the costing of some items that cannot merely be explained by differences in site conditions and access. There are also local variables such as distance from central storage to the vaccine administration sites, campaign duration and vaccine storage conditions that affect the costs. Variations in campaign logistics also influence the estimates. Differences may also arise from the methods used to calculate expenses. For future campaigns, estimating cost   Oral cholera vaccination campaigns Stephen Martin et al.
using a standardized method would be very useful. Second, reporting was not consistent, as some information about the campaign, such as coverage, delivery, adverse events following immunization monitoring and other details, were not always measured or reported. We obtained the least information on the oral cholera vaccine campaigns in the Comoros and the Solomon Islands. Third, information from the more recent post-licensure vaccination campaigns is not yet available. Updated reporting will be required. Fourth, 24% (8/33) of documents included in the analysis were not published in peer-reviewed journals but were the only available sources of data for some of the vaccination campaigns. Fifth, many of the campaigns were done in collaboration between ministries of health and external health agencies (e.g. Médecins Sans Frontières, WHO, Partners for Health, United States' Centers for Disease Control and Prevention). It will be important to continue to monitor and evaluate future campaigns using vaccine from the stockpile and implemented mainly by ministries of health. Despite these limitations, our findings provide important lessons. The number of oral cholera vaccination campaigns is increasing and experience has been documented in a variety of settings. The increasing use of oral cholera vaccine is reassuring but more needs to be done to encourage its use where needed. Since the creation of the stockpile, a higher number of doses have been used and this increase will likely continue with the availability of an oral cholera vaccine stockpile and as more experience is gained with campaigns. Data from the deployments confirm the effectiveness, safety and feasibility of mass oral cholera vaccination. While the two-dose vaccination schedule may be perceived as an impediment to delivery and coverage, the experience with both Dukoral® and Shanchol™ disproves this perception. In addition, community education on cholera control and distribution of other preventive measures such as soap and chlorine solution were feasibly integrated into recent vaccination campaigns. 35,[37][38][39][43][44][45] We also found that there were substantial differences in how the campaigns were reported making comparisons difficult. A more systematic approach to decision-making -such as a rapid assessment tool -and a standardized method for data collection, monitoring and evaluation should be