Global Vaccine Safety

Yellow fever vaccine safety during mass immunization campaign in sub-Saharan Africa

Extract from report of GACVS meeting of 12-13 June 2013, published in the WHO Weekly Epidemiological Record on 19 July 2013

The introduction of the yellow fever 17D vaccine in the 1930s provided an effective preventive measure resulting in a significant decline of the disease. However, there has been a resurgence of yellow fever resulting from changes in population dynamics, urbanization, deforestation coupled with other agricultural and developmental activities, climate changes and a decline in population immunity. In 2006, the Yellow Fever Initiative led by the WHO in partnership with UNICEF and GAVI was launched to control this resurgence in order to reduce the risk of epidemics in sub-Saharan Africa. This entailed a mass preventive vaccination campaign against the backdrop of other WHO-UNICEF yellow fever control strategies.

The recent (2007–2010) preventive yellow fever (YF) vaccination campaigns in West and Central African countries provided an opportunity for surveillance of AEFIs, thus further characterizing the safety of the vaccine. Nine countries were involved – Benin, Burkina Faso, Cameroun, Guinea, Liberia, Mali, Senegal, Sierra Leone and Togo. GACVS reviewed the recently published findings of surveillance of AEFIs conducted during those campaigns. In all, 38 million doses of the vaccine were administered, and 3116 AEFIs were observed (2952 non-serious and 164 serious). Of the serious AEFIs, 22 were classified as related to the YF vaccine and 142 were not related; of the 22, 6 clinical cases resembled acute neurotropic disease (YEL-AND), 5 clinical cases resembled acute viscerotropic disease (YEL AVD) and a further 11 involved hypersensitivity reactions. The attack rates per 100 000 vaccinated people obtained from the study therefore were 0.016, 0.013 and 0.029 for YEL-AND, YEL-AVD and hypersensitivity reactions respectively. These rates were lower than those seen with recipients of a first dose of YF vaccine in more developed settings. The median time to onset (days) was observed as 8 (YEL-AND), 4 (YEL-AVD) and 1.8 (hypersensitivity reactions). Vaccine virus identification, however, was not successful for acute cases.

The authors noted the challenges and limitations of the study to account for a relatively low specificity and sensitivity of active case findings due to many coincidental cases, operational issues with inadequate collection, storage and transportation of specimens, poor laboratory and investigation facilities, cultural practices hindering post-mortem examination, misclassification of cases, and inadequate prioritization of pharmacovigilance, amongst others. Despite the limitations, the authors concluded that the study has had impact on the countries in ensuring a proactive surveillance system as well as reinforcing the safety profile of YF vaccine in this particular setting.

GACVS noted the enormous challenges of conducting a pharmacovigilance study in a resource-limited setting. The committee suggested that enhanced vaccine safety monitoring, including additional resources to provide adequate capacity and expertise, should be included in planning vaccination campaigns. It observed that the criteria for case definition were very strict and difficult to apply appropriately in such settings. It therefore suggested that more operational criteria could be proposed that would be adapted to local clinical practice or that additional dedicated efforts be conducted to meet existing criteria. There is also a need to put standard operating procedures or tiered instructions in place to strengthen pharmacovigilance and address technical and logistic issues. GACVS also recommended that clinical and laboratory findings even if limited be more systematically correlated with post-mortem examinations.