Preparing for the Scale Up of Rotavirus Vaccine Introduction in Africa – Establishing Surveillance Platforms to Monitor Disease Burden and Vaccine Impact

Abstract

Countries in Africa have begun introducing rotavirus vaccines in their national immunization programs and wide-scale rollout across the continent is expected over the next few years. In preparation for vaccine introduction, many countries initiated surveillance for rotavirus and other studies to document disease burden, describe the epidemiology, and to monitor circulating rotavirus strains in Africa. Additionally, two countries sought to systematically investigate cases of intussusception, a rare adverse event that has been associated with rotavirus vaccines in some settings. The on-going surveillance is providing data that will serve as a baseline against which the impact of rotavirus vaccines in Africa can be assessed.

Introduction

Rotavirus, the most common cause of severe diarrhea among children <5 years of age worldwide, is responsible for significant mortality and morbidity among children in Africa. In 2008, of the estimated 453,000 global deaths from rotavirus diarrhea in children <5 years of age, more than half (N=230,000) occurred in African children.¹ Five of the 10 countries worldwide with the greatest absolute number of rotavirus deaths and 9 of the 10 countries with the highest rate of rotavirus mortality are in Africa¹. The African Rotavirus Surveillance Network coordinated by the World Health Organization (WHO) Africa Regional Office has played an important role in documenting the rotavirus disease burden in Africa. Established in 2006 with 4 countries conducting rotavirus surveillance, the network has expanded over subsequent years. By 2013, 22 countries in sub-Saharan Africa were conducting rotavirus surveillance using a standardized protocol developed by WHO.² Similarly, in North Africa, four countries are participating in the rotavirus surveillance network coordinated by the Eastern Mediterranean Regional Office and follow the same WHO protocol.

Currently, two live, attenuated, oral vaccines, a pentavalent bovine-human reassortant vaccine (RV5; RotaTeq^® (Merck and Co, Inc, Pennsylvania)) and a monovalent vaccine (RV1; Rotarix^™ (GSK Biologicals, Rixensart, Belgium)) based on a human rotavirus strain, are licensed, pre-qualified by WHO and are available for introduction into national immunization programs of countries.^{3, 4} Clinical trials conducted in 5 countries in Africa found that these vaccines were 49–72% efficacious in preventing severe gastroenteritis in African children during the first year of life.^{5, 6} While the efficacy of these oral vaccines in the African clinical trials was lower than the 85–98% efficacy observed in clinical trials carried out in Europe and the Americas,^{3, 4} the vaccine will prevent a greater number of deaths due to rotavirus diarrhea in Africa and other high disease burden settings than in low-burden, developed countries because of the substantially greater rotavirus mortality rates in Africa. Recognizing this, in 2009, WHO recommended the use of rotavirus vaccines in all countries globally and particularly those countries with high mortality due to diarrhea.⁷

In 2009, South Africa became the first African country to introduce rotavirus vaccine into its national immunization program followed by Morocco which introduced in 2010. In 2011, the Republic of Sudan became the first GAVI-eligible country in Africa to introduce rotavirus vaccine with the support of WHO and partners. By the end of July 2013, 6 additional countries (Ghana, Rwanda, Botswana, Malawi, Tanzania and The Gambia) had introduced rotavirus vaccine into their national immunization programs and one country, Zambia, introduced rotavirus vaccine in one province as part of a demonstration project. In addition these ten countries that introduced rotavirus vaccine, thirteen more African countries have been approved to introduce rotavirus vaccines with the support of GAVI by end of 2014 but this will depend on availability of rotavirus vaccines and actions to overcome other logistical issues such as cold chain expansion. The rapid introduction of rotavirus vaccines in low-income African countries has been facilitated by financial support for vaccine purchase from the GAVI Alliance for eligible countries with a GNI of US$1,550 per capita or below. These countries are required to co-finance for a full dose of rotavirus vaccine for each child, and the remainder of the price is subsidized by the GAVI Alliance.

The GAVI Alliance has also provided financial support to WHO for rotavirus surveillance networks in Africa and in other regions of the world, both to establish baseline data on rotavirus disease burden and viral strain diversity, to coordinate capacity building activities and to help monitor vaccine impact after introduction. This supplement highlights data on rotavirus disease burden in Africa prior to rotavirus vaccine introduction. The studies included in this supplement describe the baseline rotavirus disease burden and strain diversity in many African countries as well as characterize the epidemiology of intestinal intussusception, a rare adverse event associated with rotavirus vaccine.⁸ As rotavirus vaccine is rolled out across Africa, these studies coupled with ongoing sentinel surveillance will provide important data against which the impact of rotavirus vaccine can be measured. Notably, all these early countries (except Botswana) that have introduced rotavirus vaccine in their expanded program on immunization have established robust hospital based sentinel site surveillance infrastructure that can be used to evaluate impact and safety of rotavirus vaccines.

Rotavirus Disease Burden

Sentinel hospital-based surveillance in many countries showed that rotavirus accounted for 21%-56% of diarrheal hospitalizations among children <5 years of age in Africa during the pre-vaccine era.^9–19 A population-based surveillance study at one site in Kenya estimated rotavirus diarrhea resulted in 565 hospitalizations per 100,000 children <5 years of age.¹² In many countries, rotavirus hospitalizations occurred most frequently in young children, with children <1 year of age accounting for a majority of all rotavirus hospitalizations in Ethiopia (56%), Sudan (61%), Ghana (64%), Uganda (65%), Zimbabwe (66%), and Nigeria (77%).^{9, 11, 14–16, 18} (Figure 2). Conversely, in Mauritius, only 21% of rotavirus hospitalizations occurred among children <1 year of age but this may be due to differential treatment seeking behaviors by age with children <1 year of age potentially more likely to seek care at private, non-surveillance hospitals than older children.¹⁷ A birth cohort study in Egypt found that 40% of children had at least one episode of rotavirus by their second birthday.²⁰ Given this high disease burden, particularly among young children, rotavirus vaccines will likely have a rapid and substantial impact on diarrheal disease when introduced into the national immunization programs of these countries.

Figure 2.

Cumulative age distribution of rotavirus hospitalizations by 5 years of age for select countries.

In-hospital deaths due to rotavirus and diarrhea were captured by most of the surveillance sites, although the numbers of deaths were small. These numbers likely underestimate the true rotavirus mortality as children who are brought for care in a timely manner and are appropriately treated usually have good outcomes. However, many deaths due to diarrhea occur in the community and are not captured by hospital-based surveillance. Furthermore, hospitalized children recruited into surveillance are not followed after discharge but have been shown to have higher rates of mortality than healthy children.²¹ A study in Kenya that used verbal autopsy data for all deaths estimated that 140 rotavirus-associated deaths per 100,000 children <5 years of age occurred annually.¹² Given that rotavirus vaccines have proven effective in preventing mortality from diarrhea in Latin American countries that were early vaccine adopters,^22–24 they will help prevent deaths among African children who suffer from severe diarrhea and are unable to access appropriate and/or timely care.

Rotavirus was also detected among children seeking care in outpatient settings. In Kenya, 12%-20% of outpatient visits for diarrhea among children <5 years of age were due to rotavirus.^{12, 25} Rotavirus was also detected in older children and adults who sought outpatient care for diarrhea in Kenya. While these populations had lower rates of disease than among young children, introduction of rotavirus vaccine could have a broad impact if these groups indirectly benefit from reduced circulation of rotavirus in the community. Indirect benefits have been observed in several high and middle income countries using rotavirus vaccine in their national immunization programs, and as vaccines are rolled out in Africa,^26–28 evaluations should specifically assess whether such effects occur in the high burden African setting.

Rotavirus was detected year-round in Africa but disease peaked during the cool, dry months in most countries. However, some variability in the seasonal pattern occurred. For example, Ghana and Mauritius each reported that rotavirus peaked outside the typical season during one year of their multi-year surveillance periods.^{11, 17} Uganda reported two peaks per year both during the rainy season and Sudan also reported two peaks in rotavirus disease.^{15, 16}

In addition to active surveillance, routinely collected health data can also provide insight into the burden and seasonality of diarrhea and rotavirus disease. In Rwanda, data routinely collected through the Health Management Information System (HMIS) showed that diarrheal disease was highly seasonal with hospitalizations and outpatient visits peaking during the cool, dry months, the same months when rotavirus disease peaks in Rwanda.²⁹ Thus, consistent, high quality, routinely collected data can provide another platform against which to gauge the broader impact of rotavirus vaccine on the national diarrhea disease burden, and one would expect this impact to be greater during the months of the year with peak rotavirus activity.

Rotavirus Strain Diversity

Tremendous diversity of circulating rotavirus strains was observed in sub-Saharan Africa. While G1P[8] was the most frequently reported strain detected during 2007 to 2011, it was found in only 18% of rotavirus positive samples from 16 countries followed by G9P[8] (12%) and G2P[4] (9%).³⁰ There was some variability between regions within Africa but each region had a wide range of circulating strains included mixed infections and untypeable strains. Some countries also reported substantial year to year variability in the predominant strain. For example, in Mauritius, G3P[8] accounted for 89% of circulating strains in 2008, G4P[8] for 76% of strains in 2009, and G1P[8] for 90% of strains in 2010.¹⁷ Other countries, such as Ethiopia, Kenya, Uganda, and The Gambia, reported a board range of circulating strains each year with no one strain accounting for greater than 50% of strains detected.^{9, 13, 16, 31}

While these findings support earlier studies showing that diverse rotavirus strains circulate in African populations, it is reassuring to note that clinical trials and post-licensure evaluations of both rotavirus vaccines demonstrate good protection against rotavirus disease caused by various strains, including those that do not match either the G or P type or both of strains included in the vaccines.^{3–6, 32–35} Nevertheless, as vaccines are implemented in national programs, monitoring their impact on circulating rotavirus strains will be important.

Intussusception

While pre-licensure trials did not show a link between current rotavirus vaccines and intussusception,^{3, 4} these vaccines have been associated with a low level risk of intussusception during post-licensure surveillance in some high and middle income countries including the United States, Australia, Mexico, and Brazil.^{8, 36–38} Whether the vaccines will be associated with intussusception in low income countries and, if so, whether the level of risk will differ in these settings from high and middle income countries, is unclear.

While data on intussusception in Africa are sparse, in preparation for vaccine introduction, researchers in Zambia and Rwanda reviewed medical records for children hospitalized with intussusception.^{39, 40} The majority of cases occurred among infants <1 year of age and almost all required surgical intervention. Delays in presentation were common with 56% of children in Zambia and 70% of children in Rwanda presenting over 3 days after symptom onset. Delays in access to care may have contributed to poor outcomes. The case fatality rate was 34% in Zambia and 28% in Rwanda. Similarly, a case report from South Africa highlights the importance of sensitizing African doctors to consider intussusception in their differential diagnosis of cases presenting with the classic triad of symptoms for intussusception (vomiting, abdominal pain, and currant jelly stools) to ensure that intussusception cases are identified and treated promptly.⁴¹

Summary

Active rotavirus surveillance networks, such as those highlighted in the supplement, will play an important role in monitoring the impact of rotavirus vaccine following its introduction. First, the data collected through these platforms can serve as a baseline against which changes in the proportion of diarrhea hospitalizations due to rotavirus can be measured and changes in the age distribution of rotavirus cases, seasonality of rotavirus disease, and circulating strains can be monitored. Similarly, high quality, routinely collected health information system data on all-cause diarrhea hospitalizations can also be used to monitor the impact of rotavirus vaccine at the broader country level to estimate the number of hospitalizations and outpatient visits prevented by vaccine introduction. Second, rotavirus causes disease in older children and adults, albeit at lower levels than in young children. Studies in the Unites States and Australia have shown a decrease of diarrheal disease in these populations following rotavirus vaccine introduction into the infant immunization schedule suggesting that unvaccinated children and adults may be indirectly protected by reduced circulation of disease among young children.^26–28 Monitoring rates of disease in these populations are important to understand the full impact of the vaccine. Finally, as rotavirus vaccines have been associated with a low-level risk of intussusception in some settings, monitoring the risk of intussusception following rotavirus vaccine introduction will be an important priority for countries introducing rotavirus vaccine. Africa has one of the greatest burdens of rotavirus disease globally and rotavirus vaccines are poised to have a substantial impact across the continent. The data reported in this supplement highlight the critical need to establish and maintain robust surveillance that can be used to document the impact of newly introduced vaccines in national immunization programs.

Figure 1:

Rotavirus vaccine introductions in Africa as of July 2013