Introduction

Foot-and-mouth disease (FMD), one of the most contagious animal diseases, occurs in most parts of Africa, Asia, the Middle East, and South America1. More than 100 species of wild, laboratory, or domesticated animals have been infected with the FMD virus (FMDV) naturally or experimentally2. In endemic areas alone, the production losses and vaccination costs caused by FMD amount to US$6.5 to US$21 billion annually3.

Meta-analysis can be used to compare differences in the effects of various FMD vaccines to identify the most effective vaccines4. Network meta-analysis (NMA) is superior to the traditional paired meta-analysis method, as it evaluates the efficacy of interventions within a single framework, improves accuracy, compares intervention pairs that have never been directly compared in experiments, and provides the level of interventions according to their effectiveness5.

FMD is highly contagious, which hinders the performance of virus challenge studies, leading to a relative lack of data on virus challenges. Therefore, the data of virus challenge reported in the literature is very valuable6. As a step toward filling this research gap, this meta-analysis compared the effects of different vaccines. This study used Bayesian network meta-analysis to compare the protective effects of different FMD vaccines and identify the most effective vaccine.

Materials and methods

Search strategy

This study was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement guidelines. The literature retrieved in this meta-analysis was collected by searching databases from their inception until August 2024 and evaluated by three researchers (PW, ZQL, and JJ). “King of Medical Literature” software was used to search for and delete duplicate files. The National Library of Medicine (Medline via PubMed) and Embase databases were searched using the keywords (FMDV AND vaccine) OR (FMDV AND protect). The China National Knowledge Infrastructure and Wanfang Data databases were searched for studies on FMDV vaccines using the keywords “FMDV,” “protect,” and “vaccine” using King of Medical Literature software.

Inclusion and exclusion criteria

The following inclusion criteria were used: (1) the virus serotype must be type O. There are several reasons why FMDV serotype O is included. There is no cross protection between different serotypes of FMDV7. First, the FMDV serotype O serotype has the widest global distribution, covering multiple regions, including Asia, Africa, Europe, and the Americas, and as such, its harm to livestock far exceeds that of other serotypes8,9. Second, the structure of FMDV serotype O is the most unstable and easily degraded, and its vaccine immune effect is the worst10,11. Third, among FMD vaccines, O serum vaccine has enough data to meet the requirements of high-quality meta-analysis12,13,14. (2) the vaccine must be an FMD vaccine; and (3) the vaccine protective effects against FMDV must be evaluated and include challenge potency studies (direct potency studies, not only serology studies) with FMDV. The following exclusion criteria were used: (1) the carrier was not related to FMD; (2) the results did not provide the necessary basic data; and (3) the number of challenged animals was not recorded.

Data extraction and summation

Two researchers conducted a preliminary screening by reading the titles and abstracts of the retrieved studies before reading the full text. They then made their selection according to the inclusion and exclusion criteria. JJ and ZQL extracted the data, and if they disagreed about any aspect of the extraction, PW made the final decision. The data extracted included the first author’s name, publication date, and total number of animals in the trials.

Statistical analysis

R 4.3.0 language can be used to conduct Bayesian meta-analysis using JAGS 4.3.0 software. In R4.3.0, the “meta,” “grid,” and “net meta” libraries were used. The occurrence of zero events leading to bias was caused by the small number of animals involved. When the number of events was zero, the total number of events increased by 0.01. When all cases had events, 0.01 was subtracted from the total number of events. The random-effects model was used for meta-analysis to calculate the risk ratio along with a 95% confidence interval for dichotomous results15. The calculation of the random-effects model was actively performed using R software16. Where applicable, the results from the individual trials estimated were presented in a forest plot. The different FMD vaccine groups were compared in R4.3.0. Tau2 was used to quantify heterogeneity.

Results

Study identification

Figure 1 presents the flowchart of the article selection process. Of the 944 articles screened, 76 were relevant for full-text review. These articles contained a total of 29 studies that underwent meta-analysis.

Fig. 1
figure 1

Flowchart of the article selection process.

Full size image

Study characteristics

The selected articles included 29 studies focusing on FMD using the FMDV strain17. All studies evaluated the protective effects of FMD vaccines and performed virus challenge studies (Table 1).

Table 1 Protective effects of FMD vaccines18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43.
Full size table

Meta-analysis

The network of FMD vaccines is presented in Fig. 2. The results indicated that the number of comparisons made with blank groups, inactivated vaccines, recombinant virus vaccines, and Escherichia coli vaccines was the largest. All groups were directly compared with the blank control group.

Fig. 2
figure 2

Network of FMD vaccines.

Full size image

Furthermore, the quantifying heterogeneity between the groups was low (tau2 = 0.000, tau = 0.000, and I2 = 0.0% [0.0%; 24.6%]) (Fig. 3).

Fig. 3
figure 3

Forest plot.

Full size image

A funnel plot was developed for visual investigation of possible small-study effects. Overall, the plot resembled a funnel chart. Funnel plot analysis revealed that the bias between the groups was within an acceptable range (Fig. 4). The result of heterogeneity testing was small, with tau2 = 0.000.

Fig. 4
figure 4

Funnel plot.

Full size image

The results of the meta-analysis indicated that FMD vaccines produced through different methods provide different degrees of protection, with the following ranking from highest to lowest efficacy: inactivated vaccine > mRNA vaccine > E. coli vaccine > plant vaccine > recombinant virus vaccine > phage vaccine > synthesize vaccine > DNA vaccine > negative control (Table 2).

Table 2 Ranking of FMD vaccines based on their efficacy.
Full size table

Figure 5 shows that research has primarily focused on inactivated, phage, and peptide vaccines and that relatively little research has investigated plant-derived vaccines. Among all vaccines, inactivated FMD vaccines are the most extensively used.

Fig. 5
figure 5

Heat map.

Full size image

Discussion

Currently, the primary method of preventing animals from being infected with FMD virus is through vaccination. With the continuous progress of biotechnology, new vaccines are constantly being developed and undergoing continuous research and further development. The performance of FMD vaccines is inconsistent, and the identification of the most effective vaccine remains controversial. This study compared the data on all vaccines with known efficacy, particularly vaccines that have not undergone direct comparison, to determine their value and guide their follow-up in vaccine development. The inactivated vaccine (0.728) has the highest score, indicating its effectiveness and reliability. It is a traditional vaccine type, in which the virus is chemically inactivated but retains its ability to induce an immune response. Its high score suggests widespread use, proven efficacy, and safety in controlling FMD outbreaks44. However, it may require cold chain storage and multiple doses for sustained immunity45. mRNA vaccines (0.6774) represent a modern approach, leveraging genetic technology to stimulate immune responses. Although mRNA vaccines scored slightly less than inactivated vaccines, they offer advantages such as rapid development and scalability. Their lower score might reflect challenges in stability, delivery, or limited long-term data in FMD applications46. E. coli vaccines (0.6559) use E. coli as a vector to express FMD antigens. Their scores indicate moderate effectiveness, likely due to their cost-effectiveness and ease of production47,48. However, they may face limitations in antigen presentation or immune response robustness compared with other methods. DNA vaccines (0.2142) have the lowest score, indicating limited success in FMD applications. Although they offer simplicity in design, they often struggle with low immunogenicity, delivery efficiency, and regulatory hurdles. Inactivated vaccines remain the most effective option for FMD, while mRNA and E. coli vaccines show promise as modern alternatives49,50. Plant and recombinant virus vaccines offer innovative approaches but face practical challenges51,52. Phage, synthesized, and DNA vaccines are currently less effective, highlighting the need for further research and development53,54.

FMD is highly contagious, which hinders the performance of virus challenge studies, leading to a relative lack of data on virus challenges. As a step toward filling this research gap, this meta-analysis compared the effects of different vaccines. Under Bayesian mediation analysis, the inference is straightforward and exact, which makes it appealing for studies with small samples55, and this is one of the most widely used methods56,57. The advantages of meta-analysis include the provision of a comprehensive retrieval strategy and qualification criteria for retrieval research. Nevertheless, the use of meta-analysis in this study posed several limitations. The possibility of publishing studies with undesired results is low, which could lead to higher vaccine protection data58. First, FMDV is highly infectious, which limits the number of animals that can be used for experiments. Accordingly, having insufficient data, such as in studies related to new vaccines, increases the deviation of the results. Second, there may be bias in the funnel plot used for the visual (and fully subjective) investigation of possible small-study effects59. Third, a source of heterogeneity may be related to the dose of the vaccine; some vaccines are dose-dependent, only exerting fully protective effects at high doses60.

Only FMDV serotype O was included for several reasons. Although there are some data on FMDV serotype A, more research is needed to meet the requirements of meta-analysis61,62. For these reasons, it has the most important value for the study of the FMD vaccine. NMA can contribute to resolving controversies, reducing the reliance on laboratory animals, avoiding duplication of work, and guiding future research directions. NMA helps researchers study important and previously unanswerable questions, which has contributed to the rapid increase in the number of studies using NMA in the biomedical literature4. To our knowledge, our study is the first to use NMA to investigate FMD vaccines, making our study design innovative and our findings significant.

Conclusions

In this meta-analysis, 29 studies were evaluated to analyze the efficacy of FMD vaccines. The findings revealed that the inactivated vaccine provides the best protection among the different types of vaccines. Based on these findings, we recommend using inactivated vaccines as controls in the development of novel vaccines, as they achieved the highest efficacy among all evaluated vaccine types.