Figure

Figure. Study design to assess effects of shock and vibration on last-mile transportation of Ad26.ZEBOV/MVA-BN-Filo regimen Ebola vaccine regimen under refrigerated conditions. Ad26.ZEBOV/MVA-BN-Filo were produced and stored at –85°C to –55°C....

Packaging configurations (material used and shipping and storing conditions) were representative of supplies available in Africa during large vaccination campaigns (Figure; Appendix Table 1). We conducted simulated distribution testing in 2 sequential steps at 2°–8°C to approximate last-mile transport in rural areas that included a shock test followed by a vibration test. The test simulated unpaved roads typical of rural areas, representing long-distance (100-km) travel that could occur between the distribution center and vaccination site, by using a rough-road transport simulation profile.

We dropped vials 9 times in various orientations from heights of 30.5–91.4 cm, depending on the weight of the shipping container (Appendix Table 2, Figure). We determined the vibration sequence by using the International Safe Transit Association’s web-based software application 4AB (https://ista.org/test_procedures.php#enhanced-simulation-section), selecting the truck profile to approximate truck transport in rural areas. During the 36-minute vibration test, vibration intensity increased and the vibration profile shape remained constant. Vials were oriented upside down to simulate a worst-case scenario.

After simulated distribution testing, we froze all 80 vials at –20°C and shipped them back to analytical laboratories, where they were stored at –85°C to –55°C before analysis. We applied assays for indicating critical quality attributes, including appearance, potency, quantity, aggregation, protein impurities, and presence of subvisible particles (Appendix Table 3). We conducted potency testing as described previously (10). We selected test methods and release specifications on the basis of regulatory guidance provided in the International Conference on Harmonisation Q6B and European Pharmacopoeia (12,13).

We compared assay results from the drug product vials with the control (unstressed) vials from the same batch of drug product and the release specifications, noting any deviations that occurred during the study (Appendix). The attributes evaluated did not show substantial differences between stressed and control samples of Ad26.ZEBOV (Appendix Table 4). After the simulation, the appearance of the Ad26.ZEBOV drug product was clear, without coloration or visible particulate matter. Potency, quantity of Ad26.ZEBOV viral particles, and polydispersity were within release specifications. We observed no difference in the average hydrodynamic radius of stressed (54.2 nm) and control (55.2 nm) samples. Those values exceeded the current release specification (<53 nm), which is attributable to the method used for this study having an ≈5 nm bias versus the release method. The percentage of main hexon identified on reverse phase ultra-high performance liquid chromatograms was numerically lower in both stressed (65.06%) and control (64.42%) samples versus the reference batch (75.1%). The percentage of free hexon was similar in stressed (4.3%) samples and the reference batch (4.0%).

We observed no substantial effect on attributes of the MVA-BN-Filo drug product between stressed and control samples after the simulation (Appendix Table 4). The appearance of material in both groups was light yellow and milky, with no visible extraneous particles; product-related particles were present in the stressed vials. Potencies of stressed and control samples were within commercial release specifications. We confirmed the transgene expression of the Zaire Ebola virus (Mayinga), Sudan virus (Gulu), and Marburg virus (Musoke) glycoproteins and the Taï Forest virus nucleoprotein encoded by MVA-BN-Filo in stressed and control samples. Genomic vaccinia DNA contents were identical between the 2 groups. Average particle sizes measured by nanoparticle tracking analysis were numerically higher in stressed (195 nm) and control (191 nm) samples versus the reference batch (157 nm). Average virus particle sizes measured by fluorescence nanoparticle tracking analysis were numerically higher in the stressed samples (453 nm) versus both the control samples (399 nm) and the reference batch (394 nm). Total subvisible particle concentrations measured by microflow imaging were numerically higher in stressed (7.82 × 10⁶ particles/mL) versus control (4.73 × 10⁶ particles/mL) samples, whereas mean subvisible particle sizes were numerically lower in stressed (2.49 μm) versus control (3.26 μm) samples. We identified no substantial differences between stressed and control samples from the particle size distribution graphs of the 3 methods.

The stability of Ebola virus vaccine drug products is critical to ensure that quality remains unaffected in conditions likely encountered during distribution. We have shown that a liquid formulation for Ad26.ZEBOV/MVA-BN-Filo is suitable for mass vaccination in resource-limited regions at risk for outbreaks (10). We assessed the impact of shock and vibration, simulating rough-road transport conditions at 2°‒8°C, on the quality of Ad26.ZEBOV/MVA-BN-Filo.

Product stability can be demonstrated through a quantitative measure of potency, which is linked to vaccine safety and efficacy (14). Potency can correlate with infectious titer and transgene expression and may be negatively affected by aggregate formation (14). Aggregates can affect immunogenicity and subsequently compromise vaccine efficacy (14). We have demonstrated that Ad26.ZEBOV and MVA-BN-Filo undergoing simulated distribution contained similarly sized aggregates and viral particles similar in number and potency to the control samples, suggesting that the immunogenicity of samples exposed to simulated rough-road transport would not differ from the control samples.

One limitation of this study is that the stress that may be encountered during in-country transport was simulated. However, the simulation used in this study provides more realistic exposure to shock and vibration than can be generated by laboratory orbital shakers. Data on stability after agitation of alternative candidate Ebola virus vaccine regimens are limited in this regard (15).

In summary, this study shows that simulated, refrigerated (2°–8°C), rough-road truck transport, representative of the last mile of the supply chain, had no substantial effect on the quality of Ad26.ZEBOV/MVA-BN-Filo compared with the control condition. These findings confirm that refrigerated transport over rough terrain is possible and meets requirements for the challenging rural areas in the cold chain.