Skip to main content
NCBI home page
Scientific Reports logoLink to Scientific Reports
. 2024 Apr 18;14:8931. doi: 10.1038/s41598-024-59755-4

A meta-analysis on the potency of foot-and-mouth disease vaccines in different animal models

Jiao Jiao 1,2,3, Peng Wu 1,2,3,✉
PMCID: PMC11026367  PMID: 38637656

Abstract

Whether mice can be used as a foot-and-mouth disease (FMD) model has been debated for a long time. However, the major histocompatibility complex between pigs and mice is very different. In this study, the protective effects of FMD vaccines in different animal models were analyzed by a meta-analysis. The databases PubMed, China Knowledge Infrastructure, EMBASE, and Baidu Academic were searched. For this purpose, we evaluated evidence from 14 studies that included 869 animals with FMD vaccines. A random effects model was used to combine effects using Review Manager 5.4 software. A forest plot showed that the protective effects in pigs were statistically non-significant from those in mice [MH = 0.56, 90% CI (0.20, 1.53), P = 0.26]. The protective effects in pigs were also statistically non-significant from those in guinea pigs [MH = 0.67, 95% CI (0.37, 1.21), P = 0.18] and suckling mice [MH = 1.70, 95% CI (0.10, 28.08), P = 0.71]. Non-inferiority test could provide a hypothesis that the models (mice, suckling mice and guinea pigs) could replace pigs as FMDV vaccine models to test the protective effect of the vaccine. Strict standard procedures should be established to promote the assumption that mice and guinea pigs should replace pigs in vaccine evaluation.

Keywords: Vaccines, Pigs, Foot and mouth disease, Meta-analysis

Subject terms: Immunology, Infectious diseases, Viral infection

Introduction

Foot and mouth disease virus (FMDV) belongs to Picornaviridae, which is a single-stranded positive-sense RNA virus of the genus Aftab1. FMD is listed among the highly contagious diseases in animals and is endemic in Africa, most of Asia, the Middle East, and parts of South America2. FMD endemic regions contain three-quarters of the world’s FMD-susceptible livestock and most of the world’s poorer livestock keepers3.

Vaccines play an important role in controlling FMD4. There are serological tests, virus neutralization tests, and enzyme-linked immune sorbent assay (ELISA) methods to evaluate the immune efficacy of FMD vaccines, but the most reliable method is the in vivo protection test to determine the 50% protective dose or the protective rate of systemic hoof infection 5. Efficacy tests of other target animals (such as sheep, goats, or buffaloes) and the use of different methods have not been standardized (OIE Manual Terrestrial)6. It would be very valuable to verify the expected protection rate of a vaccine with cattle and to estimate the possibility that cattle can resist 10,000 infective doses after one vaccination7,8. However, it is difficult to use cattle when evaluating the efficacy of a vaccine. Cattle need many people for their care, they are dangerous, and they are expensive. Particularly in the exploratory stage of vaccine research, the laboratory stage, a new evaluation model would be beneficial for the development of new vaccines6. Different animal models are usually used in the research of FMDV vaccines9. The models used to evaluate laboratory vaccine effects include guinea pigs, mice, and suckling mice10. When studying the protective efficacy of vaccines, mice and guinea pigs are often used as substitutes for pigs11. The use of mice and guinea pigs simplifies the experimental process12. As a model animal, mice have incomparable advantages13, such as simple operations, and a large number of reports with considerable data regarding mice as FMD vaccine models14,15. However, the major histocompatibility complex (MHC) of mice and guinea pigs is very different from that of pigs16,17, and some animal models may not be appropriate for the vaccine evaluation of pigs18,19.

At present, there are no related literature reports on the correlation between the results of mice and pigs for FMDV vaccines. The ultimate goal of this meta-analysis study was to explore the rationality of replacing large animals with small animal models for vaccine testing. A meta-analysis can increase the credibility of the conclusion and support the analysis of controversial arguments20. A meta-analysis increases the statistical efficiency that a single experiment does not have, and has guiding significance for follow-up clinical experiments21. We summarized previous experimental data by employing statistical methods to avoid using and injuring a large number of animals. To clarify the possibility of using different animal models instead of pigs for FMD potency studies, a meta-analysis was performed in the present study.

Materials and methods

Literature search strategy

For this systematic review with meta-analysis (JJ and PW) searched literature published from January 1995 to August 2023. The databases PubMed, China Knowledge Infrastructure (CNKI), EMBASE, and Baidu Academic were used to search for FMDV models. The keywords were as follows: “FMDV, “mice,” “guinea pigs,” “pig or swine,” and “vaccine.” Efforts were made to include relevant gray literature, but none was found.

Inclusion and exclusion criteria

The inclusion standards were as follows: ① published Chinese and English literature on FMDV immune animal models; ② studies that included more than two animal models; ③ documents that included challenge potency (direct potency, not only serology) studies with FMDV; ④ the number of animals in the experiment was reported accurately in the literature; and ⑤ published studies and gray literature dated from January 1995 to August 2023.

The exclusion standards were as follows: ① systematic reviews without animal experiments; ② FMD models were not included; ③ when other reports provide the same data, the latest published data will be taken into account; and ④ the literature did not include a clear number of experimental animals.

Data extraction

Two researchers performed preliminary screening by reading titles and abstracts. Then, we read the full text and selected documents for further analysis according to the inclusion and exclusion criteria. Any differences of opinion were settled through discussion. Data were extracted independently and entered into a specially designed data extraction table. The extracted data included the first author, publication time, number of animals, number of protected animals, and other similar information. "Event" referred to the number of protected animals. The database was built using Microsoft Office Home and Student 2021 software.

Statistical analysis

Meta-analyses were performed using Review Manager 5.4 software (RevMan 5.4) provided by the Cochrane Collaboration. Statistical heterogeneity was quantified using the tau parameter that estimates the dispersion of the true treatment effects across the studies. Combined effect sizes and 95% confidence intervals (CI) were calculated using a random-effects model. The random-effects model used built-in modules in RevMan 5.4 software. The Mantel–Haenszel method was used to analyze the combination of effects. A funnel plot was used for the visual (and fully subjective) investigation of possible small-study effects. For data analysis, the groups were divided by different animal models. To study the protective effects of the different models, we conducted an analysis comparing the swine group with the control group. We conducted a non-inferiority analysis of the data. Non-inferiority was investigated by JMP software. The non-inferiority boundary value was set to 0.5. We used X to fit Y for non-inferiority tests. Through the relationship between the upper and lower limits of 90% difference and the boundary value, the result could be directly judged.

Results

Identified study reports

The literature was searched and screened (Fig. 1). A total of 2861 literature reports were retrieved from PubMed, CNKI, EMBASE, and Baidu Academic. After removing 23 duplicate articles and reading titles and abstracts, 189 articles met the inclusion criteria. A total of 14 articles were included in the meta-analysis.

Figure 1.

Figure 1

Open in a new tab

Flowchart of included and excluded trials.

Characteristics of the reports

Table 1 shows the features of the selected studies. A total of 869 animals were included in the meta-analysis. The animals in this research included mice, guinea pigs, and pigs. The research period was from 1997 to 2023, and it included 14 studies (Table 1).

Table 1.

Characteristics and summary findings of the selected studies.

Author Year Treatment1 Event1 n1 Treatment2 Event2 n2 Treatment3 Event3 n3
1 Gisselle N. Medina 2023 Swine 19 28 Mice 12 12 22
2 Ji-Hyeon Hwang 2021 Swine 2 4 Mice 23 30 23
3 Hyundong Jo 2021 Swine 12 16 Mice 39 50 24
4 Yanmei Dong 2015 Guinea pigs 34 60 Swine 8 15 15
5 Teresa Rodrı guez-Calvo 2010 Mice 37 37 Swine 12 14 14
6 Carolina Cubillos 2008 Guinea pigs 10 10 Swine 10 10 11
7 Houhui Song 2005 Swine 8 10 Mice 59 90 25
8 Jun 2005 Mice 116 134 Swine 2 2
9 Guangjin Li 2004 Swine 15 20 Guinea pigs 12 12 26
10 Ligang Wu 2003 Swine 3 3 Suckling mice 15 20 Guinea pigs 29 38 27
11 EWC chan 2001 Suckling mice 12 12 Swine 12 15 28
12 MA Kuprianova 2000 Swine 3 6 Guinea pigs 24 47 29
13 Quanxing Xu 1998 guinea pigs 69 77 Swine 52 65
14 Yongjin You 1997 swine 7 12 Guinea pigs 12 20
Open in a new tab

Meta-analysis

The results of the forest plot showed statistically non-significant differences between different animal models (mice, suckling mice, and guinea pigs) and swine with FMDV [MH = 0.69, 95% CI (0.43, 1.10), P = 0.12] (Fig. 2). The forest plot showed that the protective effects in pigs were statistically non-significant from those in mice [MH = 0.56, 95% CI (0.20, 1.53), P = 0.26] (Fig. 2A).

Figure 2.

Figure 2

Open in a new tab

Forest plot.

The results showed that the protective effects of guinea pigs were statistically non-significant from those of pigs [MH = 0.67, 95% CI (0.37, 1.21), P = 0.18] (Fig. 2B). There were statistically non-significant differences between swine and suckling mice [MH = 1.70, 95% CI (0.10, 28.08), P = 0.71] (Fig. 2C). At present, there were only two articles on the relationship between swine and suckling mice.

The forest plot clearly showed serious statistical heterogeneity with study results pointing to different directions. The result of I2 was not consistent with the forest map. Although the value of I2 was small, it also had serious statistical heterogeneity. There were few relevant literature reports because the extraction standard of the meta-analysis required that two controlled experiments must appear in the same article.

A funnel plot was used for the visual (and fully subjective) investigation of possible small-study effects (Fig. 3). Overall, the plot resembled a funnel chart. However, the funnel charts of the three subgroups were not ideal by themselves. The reason may be that there were too few studies included in the subgroups, and the subgroups were not suitable for use in funnel charts.

Figure 3.

Figure 3

Open in a new tab

Funnel plot.

Non-inferiority test could provide a hypothesis that the models (mice, suckling mice and guinea pigs) could replace pigs as FMDV vaccine models to test the protective effect of the vaccine (Fig.4). Through meta-analysis, we found that there was some heterogeneity in this study (Fig. 2). Even though the null hypothesis was rejected in all tests, the results should be interpreted with caution due to the substantial statistical heterogeneity observed in the forest plot (Fig. 4).

Figure 4.

Figure 4

Open in a new tab

Non-inferiority plot. (A) Non-inferiority was tested with mice and pigs. (B) Non-inferiority was tested with guinea pigs and pigs. (C) Non-inferiority was tested with suckling mice and pigs. (D) Mice, guinea pigs, and suckling mice was made non-inferiority test to pigs. When the blue line (90% confidence interval) is included in the blue interval (upper and lower bounds), a non-inferiority conclusion could be drawn. When the red line (90% confidence interval) is not included in the blue interval (upper and lower bounds), a non-inferiority conclusion cannot be drawn.

Discussion

FMD is a highly contagious and destructive virus30. There are very strict restrictions on FMD experiments, and the requirements for the laboratory are also very high31. These existing conditions restrict the development of experiments and the acquisition of data on FMD. A meta-analysis assumes that the processed data are normally distributed32. In principle, the data should conform to a normal distribution32. The occurrence of zero events has a great impact on META-analysis33. We have tried our best to collect appropriate data.

As model animals, mice have the advantages of clean genetic backgrounds, easy breeding, and simple acquisition14,15. Compared with pigs, mice are more accessible12. It is easy to administer vaccines and drugs to mice by injection13. The injection dose for mice is less than that for pigs, which is more suitable for preliminary research. However, the MHC of mice and pigs is different16,17. Antibodies against the same antigen are also different18,19. The forest plot showed that the protective effects on pigs were statistically non-significant from those of mice [MH = 0.56, 95% CI (0.20, 1.53), P = 0.26] (Fig. 2A).

We innovatively compared different models, which also involved heterogeneity of methods34. Although clinical and methodological heterogeneity was always present, in many studies, mice and guinea pigs were used instead of pigs to evaluate vaccine effects. Although different methods increase heterogeneity, a scientific selection of indicators can reduce heterogeneity as much as possible, so that the results of the two models tend to be similar. We made a direct comparison between mice and pigs, guinea pigs and pigs, and suckling mice and pigs. There was no comparison between mice, suckling mice, and guinea pigs directly. Network meta-analysis (NMA) may help to directly compare different models35. To visually investigate small-study effects in NMA, Chaimani and colleagues developed a tool36,37. Mavridis et al. extended the Copa selection model for publication bias to NMA38. A transitivity assumption is the cornerstone of NMA; it posits that the comparisons do not differ beyond the interventions compared39. However, the different models we studied were not applicable to NMA. We chose RevMan to perform a traditional meta-analysis.

There are some limitations in this study. There are many guidelines for performing a meta-analysis40. A meta-analysis has comprehensive and objective advantages, including data integration41. There may be some heterogeneity and deviations in any research42. First, the inconsistent dosages administered to animals may affect the experimental results, leading to heterogeneity. Second, a funnel plot was used for the visual (and fully subjective) investigation of possible small-study effects. In this study, reducing the occurrence of deviations was of prime importance. Some of the retrieved data may be ignored, such as data in different languages, from different databases, and using different keywords. Inclusion and exclusion criteria may also lead to bias, and deviations may also appear at different steps in the process. However, according to the funnel chart, the bias was within the acceptable range.

In this study, to the best of our knowledge, a systematic review and meta-analysis were used for the first time to analyze the immune effects of different FMD animal models. Non-inferiority test can provide a hypothesis that the models (mice, suckling mice and guinea pigs) can replace pigs as FMDV vaccine models to test the vaccine protection effect. Reasonable selection of animal models can not only reduce the use of experimental animals but also promote the evaluation of vaccine effects, thus improving the protective effects of the vaccine. It is very valuable to compare the effects on a small animal model with the effects on pigs. Our experiment results will improve the rationality of the model. Furthermore, the cost of vaccine research and development is reduced. Animal models have accelerated the speed of vaccine development. Whether the results of the model can be used as an OIE standard still needs further research and efforts.

Conclusion

In conclusion, non-inferiority test could provide a hypothesis that the models (mice, suckling mice and guinea pigs) could replace pigs as FMDV vaccine models to test the protective effect of the vaccine. Strict standard procedures should be established to promote the assumption that mice and guinea pigs should replace pigs in vaccine evaluation.

Acknowledgements

We thank LetPub (www.letpub.com) for its linguistic assistance during the preparation of this manuscript. This research was funded by Shihezi university (Grant no. RCZK202048) and (Grant no. CXBJ202105).

Author contributions

P.W. designed the manuscript. P.W. and J.J. searched documents and extracted data. J.J. operated the software. P.W. and J.J. wrote the main manuscript. All authors examined the manuscript.

Data availability

All data generated or analyzed during this study were included in this published article.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher's note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  • 1.Zhao FR, Xie YL, Liu ZZ, et al. Transcriptomic analysis of porcine PBMCs in response to FMDV infection. Acta Trop. 2017;173:69–75. doi: 10.1016/j.actatropica.2017.05.009. [DOI] [PubMed] [Google Scholar]
  • 2.Hammond JM, Maulidi B, Henning N. Targeted FMD vaccines for Eastern Africa: The AgResults foot and mouth disease vaccine challenge project. Viruses. 2021;13(9):1830. doi: 10.3390/v13091830. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Knight-Jones T, McLaws M, Rushton J. Foot-and-mouth disease impact on smallholders—What do we know, what don't we know and how can we find out more. Transbound. Emerg. Dis. 2017;64(4):1079–1094. doi: 10.1111/tbed.12507. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Di Giacomo S, et al. Assessment on different vaccine formulation parameters in the protection against heterologous challenge with FMDV in cattle. Viruses. 2022;14:1781. doi: 10.3390/v14081781. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Edwards S. OIE standards for vaccines and future trends. Rev. Sci. Tech. 2007;26(2):373–378. doi: 10.20506/rst.26.2.1749. [DOI] [PubMed] [Google Scholar]
  • 6.Barnett PV, Geale DW, Clarke G, Davis J, Kasari TR. A review of OIE Country status recovery using vaccinate-to-live versus vaccinate-to-die foot-and-mouth disease response policies I: Benefits of higher potency vaccines and associated NSP DIVA test systems in post-outbreak surveillance. Transbound. Emerg. Dis. 2015;62(4):367–387. doi: 10.1111/tbed.12166. [DOI] [PubMed] [Google Scholar]
  • 7.Maradei E, La Torre J, Robiolo B, et al. Updating of the correlation between lpELISA titers and protection from virus challenge for the assessment of the potency of polyvalent aphtovirus vaccines in Argentina. Vaccine. 2008;26(51):6577–6586. doi: 10.1016/j.vaccine.2008.09.033. [DOI] [PubMed] [Google Scholar]
  • 8.Periolo OH, Seki C, Grigera PR, et al. Large-scale use of liquid-phase blocking sandwich ELISA for the evaluation of protective immunity against aphthovirus in cattle vaccinated with oil-adjuvanted vaccines in Argentina. Vaccine. 1993;11(7):754–760. doi: 10.1016/0264-410X(93)90261-U. [DOI] [PubMed] [Google Scholar]
  • 9.Li P, Huang S, Zha J, et al. Evaluation of immunogenicity and cross-reactive responses of vaccines prepared from two chimeric serotype O foot-and-mouth disease viruses in pigs and cattle. Vet. Res. 2022;53(1):56. doi: 10.1186/s13567-022-01072-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Wu P, et al. Layered double hydroxide nanoparticles as an adjuvant for inactivated foot-and-mouth disease vaccine in pigs. BMC Vet. Res. 2020;16:474. doi: 10.1186/s12917-020-02689-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Cubillos C, de la Torre BG, Jakab A, et al. Enhanced mucosal immunoglobulin A response and solid protection against foot-and-mouth disease virus challenge induced by a novel dendrimeric peptide. J. Virol. 2008;82(14):7223–7230. doi: 10.1128/JVI.00401-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Ren ZJ, Tian CJ, Zhu QS, et al. Orally delivered foot-and-mouth disease virus capsid protomer vaccine displayed on T4 bacteriophage surface: 100% protection from potency challenge in mice. Vaccine. 2008;26(11):1471–1481. doi: 10.1016/j.vaccine.2007.12.053. [DOI] [PubMed] [Google Scholar]
  • 13.Balani S, Nguyen LV, Eaves CJ. Modeling the process of human tumorigenesis. Nat. Commun. 2017;8:15422. doi: 10.1038/ncomms15422. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Rodríguez-Calvo T, Ojosnegros S, Sanz-Ramos M, et al. New vaccine design based on defective genomes that combines features of attenuated and inactivated vaccines. PLoS One. 2010;5(4):e10414. doi: 10.1371/journal.pone.0010414. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Dong YM, Zhang GG, Huang XJ, Chen L, Chen HT. Promising MS2 mediated virus-like particle vaccine against foot-and-mouth disease. Antiviral Res. 2015;117:39–43. doi: 10.1016/j.antiviral.2015.01.005. [DOI] [PubMed] [Google Scholar]
  • 16.Grafen A. Of mice and the MHC. Nature. 1992;360(6404):530. doi: 10.1038/360530a0. [DOI] [PubMed] [Google Scholar]
  • 17.Manning CJ, Wakeland EK, Potts WK. Communal nesting patterns in mice implicate MHC genes in kin recognition. Nature. 1992;360(6404):581–583. doi: 10.1038/360581a0. [DOI] [PubMed] [Google Scholar]
  • 18.Xiong X, Tian S, Yang P, et al. Emerging enterococcus pore-forming toxins with MHC/HLA-I as receptors. Cell. 2022;185(7):1157–1171.e22. doi: 10.1016/j.cell.2022.02.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Chen FX, Tang J, Li NL, et al. Novel SLA class I alleles of Chinese pig strains and their significance in xenotransplantation. Cell Res. 2003;13(4):285–294. doi: 10.1038/sj.cr.7290173. [DOI] [PubMed] [Google Scholar]
  • 20.Jiao J, Wu P. A meta-analysis: The efficacy and effectiveness of polypeptide vaccines protect pigs from foot and mouth disease. Sci. Rep. 2022;12(1):21868. doi: 10.1038/s41598-022-26462-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Stroup DF, et al. Meta-analysis of observational studies in epidemiology: A proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000;283:2008–2012. doi: 10.1001/jama.283.15.2008. [DOI] [PubMed] [Google Scholar]
  • 22.Medina GN, Spinard E, Azzinaro PA, et al. Deoptimization of FMDV P1 region results in robust serotype-independent viral attenuation. Viruses. 2023;15(6):1332. doi: 10.3390/v15061332. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Hwang JH, Lee G, Kim A, et al. A vaccine strain of the A/ASIA/Sea-97 lineage of foot-and-mouth disease virus with a single amino acid substitution in the P1 Region That Is adapted to suspension culture provides high immunogenicity. Vaccines (Basel). 2021;9(4):308. doi: 10.3390/vaccines9040308. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Jo H, Kim BY, Park SH, et al. The HSP70-fused foot-and-mouth disease epitope elicits cellular and humoral immunity and drives broad-spectrum protective efficacy. NPJ Vaccines. 2021;6(1):42. doi: 10.1038/s41541-021-00304-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Song H, Wang Z, Zheng D, et al. A novel mucosal vaccine against foot-and-mouth disease virus induces protection in mice and swine. Biotechnol. Lett. 2005;27(21):1669–1674. doi: 10.1007/s10529-005-2727-4. [DOI] [PubMed] [Google Scholar]
  • 26.Li G, Chen W, Yan W, et al. Comparison of immune responses against foot-and-mouth disease virus induced by fusion proteins using the swine IgG heavy chain constant region or beta-galactosidase as a carrier of immunogenic epitopes. Virology. 2004;328(2):274–281. doi: 10.1016/j.virol.2004.07.025. [DOI] [PubMed] [Google Scholar]
  • 27.Wu L, Jiang L, Zhou Z, et al. Expression of foot-and-mouth disease virus epitopes in tobacco by a tobacco mosaic virus-based vector. Vaccine. 2003;21(27–30):4390–4398. doi: 10.1016/S0264-410X(03)00428-6. [DOI] [PubMed] [Google Scholar]
  • 28.Chan EW, Wong HT, Cheng SC, et al. An immunoglobulin G based chimeric protein induced foot-and-mouth disease specific immune response in swine. Vaccine. 2000;19(4–5):538–546. doi: 10.1016/S0264-410X(00)00186-9. [DOI] [PubMed] [Google Scholar]
  • 29.Kuprianova MA, Zhmak MN, Koroev DO, Chepurkin AV, Volpina OM, Ivanov VT. Synthetic peptide constructs on the basis of immunoactive fragments of the A22 strain VP1 of the foot-and-mouth disease virus. Russian J. Bioorgan. Chem. 2000;26(12):832–837. doi: 10.1023/A:1026614622241. [DOI] [Google Scholar]
  • 30.Li K, et al. Virus-host interactions in foot-and-mouth disease virus infection. Front. Immunol. 2021;12:571509. doi: 10.3389/fimmu.2021.571509. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Theerawatanasirikul S, et al. Small molecules targeting 3C protease inhibit FMDV replication and exhibit virucidal effect in cell-based assays. Viruses. 2023;15:1887. doi: 10.3390/v15091887. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Jackson D, White IR. When should meta-analysis avoid making hidden normality assumptions. Biom. J. 2018;60(6):1040–1058. doi: 10.1002/bimj.201800071. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Efthimiou O. Practical guide to the meta-analysis of rare events. Evid. Based Ment. Health. 2018;21(2):72–76. doi: 10.1136/eb-2018-102911. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Ades, A.E., Welton, N.J., Dias, S., Phillippo, D.M., Caldwell, D.M. Twenty years of network meta-analysis: Continuing controversies and recent developments. Res. Synth. Methods. (2024). [DOI] [PubMed]
  • 35.Nikolakopoulou A, et al. Living network meta-analysis compared with pairwise meta-analysis in comparative effectiveness research: empirical study. BMJ. 2018;360:k585. doi: 10.1136/bmj.k585. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Beguelin, A., Dongarra, J.J. Graphical development tools for network-based concurrent supercomputing. 435–444 (1991).
  • 37.Chaimani A, Salanti G. Using network meta-analysis to evaluate the existence of small-study effects in a network of interventions. Res. Synth. Methods. 2012;3(2):161–176. doi: 10.1002/jrsm.57. [DOI] [PubMed] [Google Scholar]
  • 38.Mavridis D, Welton NJ, Sutton A, Salanti G. A selection model for accounting for publication bias in a full network meta-analysis. Stat. Med. 2014;33(30):5399–5412. doi: 10.1002/sim.6321. [DOI] [PubMed] [Google Scholar]
  • 39.Salanti G. Indirect and mixed-treatment comparison, network, or multiple-treatments meta-analysis: Many names, many benefits, many concerns for the next generation evidence synthesis tool. Res. Synth. Methods. 2012;3(2):80–97. doi: 10.1002/jrsm.1037. [DOI] [PubMed] [Google Scholar]
  • 40.Higgins J, Green S. GSe, Cochrane Handbook for Systematic Reviews of Interventions. Naunyn-Schmiedebergs Archiv für experimentelle Pathologie und Pharmakologie. 2011;5(2):S38. [Google Scholar]
  • 41.Arya S, Schwartz TA, Ghaferi AA. Practical guide to meta-analysis. JAMA Surg. 2020;155:430–431. doi: 10.1001/jamasurg.2019.4523. [DOI] [PubMed] [Google Scholar]
  • 42.Li H, Shih MC, Song CJ, Tu YK. Bias propagation in network meta-analysis models. Res. Synth. Methods. 2023;14:247–265. doi: 10.1002/jrsm.1614. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

All data generated or analyzed during this study were included in this published article.

Articles from Scientific Reports are provided here courtesy of Nature Publishing Group

RESOURCES