Global Vaccine Safety

Global Advisory Committee on Vaccine Safety, 12-13 December 2007

Published in the WHO Weekly Epidemiological Record on 25 January 2008

The Global Advisory Committee on Vaccine Safety (GACVS), an expert clinical and scientific advisory body, was established by WHO to deal with vaccine safety issues of potential global importance independently from WHO and with scientific rigour.1 GACVS held its 17th meeting in Geneva, Switzerland, on 12–13 December 2007.2 The committee discussed a number of general issues relevant to all vaccines as well as a number of vaccine-specific issues. Discussions of vaccine-specific issues pertained both to long-standing vaccines as well as to new vaccines or vaccines still under development. The following issues, among others, were considered.

General issues

Guillain-Barré syndrome and vaccination

Guillain-Barré syndrome (GBS) is a relatively rare (1–2 cases per 100 000 people annually) acute peripheral neuropathy that most frequently occurs without an identified causal event. It can, however, follow some infectious illnesses, of which the most common is caused by Campylobacter jejuni (1 case of GBS per 3000 infectious episodes), probably as a result of an autoimmune response. GBS has also been occasionally observed in a temporal association with vaccination; this association has been considered as causal in cases following Swine influenza vaccine (attributable risk: 9.5 per million doses administered) as well as rabbit brain and other nervous-tissue derived rabies vaccines. Antibodies against autologous gangliosides have been demonstrated in one form of GBS (acute motor axonal neuropathy (AMAN)) that is prevalent in some Asian and Latin American countries. Although such antibodies can be induced by C. jejuni, the immunological pathogenesis of the most frequent form of GBS encountered in USA and Europe (acute inflammatory demyelinating polyradiculoneuropathy (AIDP)) is still unclear and may involve T-cells or antibodies against antigens as yet not identified.

Cases of GBS have been reported in temporal association with other vaccines, including seasonal influenza, tetanus, meningococcal conjugates and diphtheria–tetanus–pertussis (DTP). There is conflicting evidence concerning whether influenza vaccines other than Swine influenza are associated with GBS. To date, a causal relationship between GBS and immunization has not been established, other than for Swine influenza vaccine and nervous-tissue derived rabies vaccines.

GACVS considers that investigation of a possible causal relationship could best be achieved by large-scale studies of the incidence of GBS before and after an immunization programme. All incident cases would need to be carefully ascertained and documented to ensure as accurate a diagnosis as possible and to identify the form of GBS (principally AIDP, or AMAN). Improved understanding of the pathogenesis of all forms of GBS will assist the investigation of possible associations between GBS and immunization. In this context, the collection of serum samples from incident cases of GBS would contribute to the identification of the different forms of the disease and of understanding their possible relationship with vaccines.

Such studies would be particularly helpful in investigating neurological adverse events following immunization (AEFI) that occur in association with pandemic or pre-pandemic influenza vaccines.

Safety of immunization in immunocompromised individuals

The risks of vaccination in immunocompromised individuals encompass a broad spectrum of situations related to the types of immunological impairment and the vaccines being considered. Vaccine efficacy may be reduced among immunocompromised individuals whose response to immunization is weak and who may also face increased risks of AEFI. A review of selected studies on the use of vaccines among immunocompromised individuals revealed that the level of evidence available to estimate efficacy and safety varies substantially depending on the situations being considered.

Recognizing the importance of providing state-of-the-art advice for each situation, GACVS undertook an in-depth review of existing evidence and tasked a working group with oversight of this topic. It is proposed that the safety of vaccines in people living with HIV/AIDS be considered as a first priority. Priority will also be given to live attenuated vaccines, in particular those against measles, yellow fever and rotavirus. Bacille Calmette-Guérin (BCG) vaccine has been considered in a previous GACVS session.3 Among those living with HIV/AIDS, the potential effect of each vaccine on the progression of HIV infection itself will also be considered. The safety of vaccines will be considered separately for those on antiretroviral therapy and for those who have untreated HIV infections.

For each vaccine considered, the level of evidence available, the importance of the public health risk identified and the need for additional studies will be assessed. For vaccines for which sufficient studies are available, systematic reviews will be commissioned. For other situations, guidance will be provided on areas that would benefit from additional studies in order to assess the safety of specific vaccines in particular circumstances that might not have been addressed during the clinical development phases. The subgroup will report to the full committee with a framework for developing advice and will propose a detailed workplan for addressing the 3 priority vaccines identified.

Issues associated with specific vaccines

Safety of yellow fever vaccines

A yellow fever vaccination campaign was conducted in the Ica Region of Peru following the earthquake in September 2007. Four cases of vaccine-associated viscerotropic disease, all fatal, were reported among approximately 40 000 individuals who received one particular lot of yellow fever vaccine. As a result, national authorities suspended the campaign. The rate of vaccine-associated viscerotropic disease was approximately 10 per 100 000 doses administered for this vaccine lot, compared with an expected rate of approximately 0.3 per 100 000 based on previous experience with yellow fever vaccines. A current investigation includes clinical and laboratory evaluation of cases, epidemiological evaluation of AEFI and a review of vaccine production at the manufacturing facility. This investigation should include testing of samples of vaccine remaining from this lot in Peru and ascertainment of AEFI associated with any use of this lot of vaccine elsewhere. Following completion of the investigation, an expert review is recommended in the broader context of all serious AEFI that may follow yellow fever vaccination.

Hepatitis B vaccination and rheumatoid arthritis

GACVS considered the potential association of hepatitis B vaccination (HBV) and rheumatoid arthritis (RA). Prior to previous discussions on this topic held in June 2006, the committee had commissioned a comprehensive literature review. At this meeting, it had reviewed more recent information, particularly on genetic issues.

The literature relating HBV to RA comprises mainly single case reports, case series and a few case–control studies. The published studies are limited and difficult to interpret owing to problems in the methodology and control of confounding. The one high-quality case–control study did not find a statistically significant association, but it had limited power and wide confidence intervals. However, as only a small proportion of cases had received HBV, there is at most a very small contribution of HBV to the incidence of RA. GACVS had also seen the preliminary results of a large study based on the United States Vaccine Safety DataLink (VSD) project in June 2006, which was analysed in several ways, but none showed a significant association between HBV and RA.

It had been suggested that the failure to find any increased risk overall following HBV may be because the adverse effect of vaccination is present only in a small subgroup that might be at increased risk of RA because of their genetic make-up. GACVS considered the preliminary results of genetic analyses conducted using the VSD case–control study, which was able to classify cases and controls according to their HLA DRB1*04 status. The choice of this marker was made largely because it has been described as a biomarker of the genetic susceptibility to develop RA.

GACVS was presented with a preliminary analysis of the interaction between HBV and HLA status in respect of the occurrence of RA. This is the relevant question if the adverse effect is limited to or occurs largely among a particular genetically-determined subgroup. It has the advantage that analysis can be done using cases of RA alone. These are compared between the genetically determined groups, which will not differ in regard to likelihood of HBV exposure. Hence, confounding is not a major problem.

There are various subtypes of HLA DRB1*04, of which 9 occurred in this study; analyses were limited to 2 of the subtypes.

HBV exposure in the 90, 180 and 365 days before onset of RA symptoms was examined. There was no statistically significant evidence of an increased risk in the genetic subgroups examined, and point estimates were less than unity. However, an increased risk in a subgroup could not be excluded because of the low power of the study, given the small numbers of vaccinated cases. This is an inevitable limitation but this also makes it clear that HBV, at most, can make very little impact on the incidence of RA. In addition, whether HLA DRB1*04 is the best genetic marker for increased risk of development of rheumatoid arthritis is still unknown.

GACVS concluded, based on a review of the limited data available, that there was no convincing evidence to support an association between HBV and RA. It will consider the topic further if new findings become available.

Safety of live 14-14-2 Japanese encephalitis vaccine

GACVS reviewed an updated report on the frequency and severity of AEFI reported with the use of live Japanese encephalitis vaccine (JEV) in the Republic of Korea, where it was introduced in 2002 but is used only in the private sector. The recommended schedule is 2 doses administered 12 months apart in children aged 1–2 years and a booster at 6 years of age. Inactivated mouse-brain JEV is used in the national immunization programme. Since the private sector is not linked to AEFI surveillance, no safety data are available on the >1.7 million doses of live JEV distributed since 2002. The national requirement for the introduction of new drugs is to investigate adverse events in detail in 600 subjects. Among 673 children followed up after a dose of live JEV, 318 events were reported in 163 subjects; most AEFI reported were relatively minor, non-neurological and indistinguishable from common childhood illnesses. The review of the safety of live JEV will be finalized in 2008.

GACVS also reviewed a study on co-administration of measles vaccine (MV) with live JEV conducted in the Philippines. Three groups of 223–228 infants (aged 9–11 months) were given either JEV followed by MV a month later, MV followed by JEV a month later or both given simultaneously. The design was to examine non-inferiority of response to MV when co-administered. One month after vaccination, although overall sero-protection was very high in all 3 groups, in the group in which the vaccines were co-administered, the proportion of infants who achieved sero-protection following MV (96%) was slightly lower than in the MV-only group (100%); [difference = –4%, 95% confidence interval, –1, –6]). In addition, the antibody titres were significantly lower in the co-administration group than in those who had received only MV.

These results indicate some interference of live JEV on MV response. Among those who had sero-converted by 1 month, at 1-year follow-up the geometric mean antibody titres were similar in both groups, suggesting that the interference is only temporary and that co-administration of live JEV and MV is acceptable.

Based on earlier reviews and the information from the above studies, GACVS concluded that the short-term safety profile of live JEV appears satisfactory and the vaccine could safely be administered with measles vaccine as of 9 months of age. More investigations are needed to confirm that low-frequency adverse events (especially neurological) do not occur. Since live JEV is currently used in “catch-up” campaigns on many millions of children in Asian countries, the opportunity should be taken to examine if the vaccine-safety profile remains valid in large study populations. Further studies to check if MV effectiveness remains undiminished should be encouraged.

Safety of meningococcal B vaccines

GACVS was presented with data relating to the safety of outer membrane-vesicle based meningococcal B vaccines based on their usage in Cuba, France, New Zealand and Norway. It noted in particular the very carefully designed safety monitoring programme that had been set up in New Zealand to ascertain possible serious AEFI following the vaccination of around 1 million people aged <20 years from July 2004 onwards, including 200 000 vaccinees monitored through linkages to hospital admission records. The various surveillance methods used to detect potentially serious AEFI consistently found no evidence of such effects attributable to vaccination.

The committee also appraised the evidence of vaccine safety derived from studies in Norway, where the vaccine formulation originated. Although the vaccine had never been introduced into routine use in Norway, because of waning of the meningococcal disease epidemic, there had been extensive use of the vaccine in trials among teenagers and young adults. In some age cohorts, up to 40% of the population had been vaccinated. Despite media reports on possible increased risk of myalgic encephalomyelitis (ME), also called chronic fatigue syndrome, the results from these trials provided no specific causes for concern with respect to serious AEFI. The committee noted that a case–control study had been conducted of ME in Norway, involving all 273 cases in whom ME had been diagnosed at the 2 major referral hospitals in Norway, in the 1972–1977 birth cohorts, of whom 201 participated in the case-control study. This study had been prompted by media reports of a possible increased risk of ME associated with the use of meningococcal B vaccine. A random sample of 889 controls was drawn from the general population, of whom 389 participated in the study. About 45% of both case and control groups had received meningococcal vaccine, and the study thus provided no evidence of an increased risk of ME associated with vaccination (relative risk = 1.06, 95% confidence interval, 0.67–1.66).

The Norwegian-type vaccine had also been used in France in a 3-dose schedule to vaccinate around 2700 children aged 12 months to 5 years in 1 administrative region (Department of Seine-Maritime), following an increased incidence of meningococcal disease in that department. It was noted that there had been little prior experience of use of the vaccine in young children in the Norwegian trials. Parents of vaccinees were sent a questionnaire to ascertain possible adverse events. A high proportion of these were returned and 9 serious adverse events were identified, 8 of which were purpura (1 idiopathic thrombocytopenic purpura (ITP), 3 Henoch-Schönlein purpura, 4 febrile purpura) and 7 of which occurred after the second vaccine dose. All of these 7 had received a third dose of vaccine with no reported ill effect. The case of ITP, considered unlikely to be causally related to vaccination, was the only one not to have recovered. The significance of the cases of purpura following vaccination was difficult to evaluate because of the absence of data on background rates of purpura in the general population.

The most widely used meningococcal B vaccine has been that produced in Cuba. Over 55 million doses have been used in Cuba and other Latin American countries over the past 20 years. The committee was presented with data from the original Phase 3 study involving over 100 000 people, in which no evidence of an excess of serious AEFI had been identified in the vaccinated group. Also, no evidence of an excess of serious AEFI had been reported from the extensive use of the vaccine in vaccination programmes in Cuba and other Latin American countries, but the sensitivity of monitoring of serious AEFI with these uses of the vaccine was unclear.

GACVS noted that several new meningococcal vaccines were being developed, at least one of which was closely related to the vaccine used in New Zealand. While it would be important to set up careful safety studies of these new vaccines, the committee was reassured by the absence of evidence of serious AEFI of existing meningococcal B vaccination, which had been assessed particularly carefully in studies in New Zealand and Norway.

Rotavirus vaccines and Kawasaki disease

GACVS was presented with the latest information from the US and the European Union on the potential association between Kawasaki disease (KD) and the administration of rotavirus vaccines. The United States Centers for Disease Control and Prevention defines a case of KD as illness in a patient with fever lasting 5 or more days and the presence of at least 4 of the following 5 clinical signs: rash, cervical lymphadenopathy (at least 1.5 cm in diameter), bilateral conjunctival injection, oral mucosal changes and peripheral extremity changes. Patients whose illness does not meet the above KD case definition but who have fever and coronary artery abnormalities are classified as having atypical or incomplete KD.

The possibility of an association had previously been raised following the observation of a non-statistically significant excess of KD in pre-licensure trials of Rotateq™ vaccine (5 cases in vaccinees versus 1 case in controls) and GACVS had already discussed the possibility of an association.3 In June 2007, the FDA amended the product information for the US to note the occurrence of such cases, but stated that the causality had not been established.

Since June 2007 and as of 14 October 2007, a total of 16 cases of KD (12 confirmed as cases after review) had been reported to the US Vaccine Adverse Events Reporting System (VAERS) in the context of stimulated reporting. The pattern of reporting had not shown a consistent effect in terms of time of occurrence following vaccination, with the onset varying between 0 and 54 days after vaccination. Although this volume of reporting met the criteria of a signal because of the overall rarity of reports of KD, it was not considered that the overall pattern showed evidence of causality and, while underreporting or under-ascertainment is likely, the rate per distributed dose was much lower than the best estimate of the background rate. The reported rate was 1.6 per 100 000 person-years compared with an estimated background rate of 17 per 100 000 person-years among children aged <5 years.

Active surveillance during the first year of life following vaccination with Rotateq™ was implemented within the Vaccine Safety Datalink. Preliminary results indicate that there has been no confirmed case of KD following administration of >125 000 doses. Further work is planned to extend the monitoring.

In the European Union, where Rotarix™ vaccine is licensed and in use, there have been no spontaneous reports of KD, with about 12 million doses administered in 6 million individuals. In ongoing clinical trials, there is a slight excess of cases in the Rotarix™ groups, with a non-statistically significant difference between the Rotarix™ and the placebo arms. Delay between vaccination and onset of KD varied from 2 weeks to 19 months.

The overall conclusion of GACVS was that the evidence for a causal association was not strong and that there was no reason for concern. Consideration was given as to whether there was an urgent need for KD to be included in Brighton Collaboration case definitions. GACVS noted that there is a clear definition using American Heart Association criteria and, while a Brighton definition could be helpful, its development was not a matter of urgency. A major problem was that the necessary investigations were often not carried out, especially in resource-poor settings, and that, even in a clinical trial setting, the descriptions of cases are often too sparse to decide whether the case is definite.

Modus operandi of the committee and additional information

In addition to publications in the Weekly Epidemiological Record, the scope of the committee’s work and past decisions, recommendations and actions as well as its modus operandi, have been published in the American Journal of Public Health.4 More information about the topics discussed in this article as well as the committee’s terms of reference and the work of its subgroups can be found on the GACVS web site at

  • See No. 41, 1999, pp. 337–338.
  • GACVS invited additional experts to present evidence on the safety of yellow fever vaccination and cluster of reported cases of viscerotropic disease in Peru; hepatitis B vaccination and arthritis; review of safety of live Japanese encephalitis vaccine; safety of immunization in the immunocompromised; review of patho-physiology of GBS following immunization; safety of meningococcal B vaccines; and Kawasaki disease in relation to the use of rotavirus vaccines. Depending on the session, these experts were affiliated with the Norwegian Institute of Public Health; the University of Texas Medical Branch; the Agence française de sécurité sanitaire des produits de santé ; the Finlay Institute in Cuba; the Department of Clinical Neuroscience, King’s College London; the Center for Health Research, Atlanta, GA, USA; the University of KwalaZulu-Natal; the Korean Food and Drug Administration; the New Zealand Ministry of Health; the New Zealand Pharmacovigilance Center; the Program for Appropriate Technology in Health; the Centre Medical Universitaire de Genève; the United States Centers for Disease Control and Prevention; the US Food and Drug Administration; and the European Medicines Evaluation Agency.
  • See No. 28/29, 2007, pp. 252–259.
  • A global perspective on vaccine safety and public health: the Global Advisory Committee on Vaccine Safety. American Journal of Public Health, 2004, 94:1926–1931.