Pandemic-response adenoviral vector and RNA vaccine manufacturing

Abstract

Rapid global COVID-19 pandemic response by mass vaccination is currently limited by the rate of vaccine manufacturing. This study presents a techno-economic feasibility assessment and comparison of three vaccine production platform technologies deployed during the COVID-19 pandemic: (1) adenovirus-vectored (AVV) vaccines, (2) messenger RNA (mRNA) vaccines, and (3) the newer self-amplifying RNA (saRNA) vaccines. Besides assessing the baseline performance of the production process, impact of key design and operational uncertainties on the productivity and cost performance of these vaccine platforms is quantified using variance-based global sensitivity analysis. Cost and resource requirement projections are computed for manufacturing multi-billion vaccine doses for covering the current global demand shortage and for providing annual booster immunisations. The model-based assessment provides key insights to policymakers and vaccine manufacturers for risk analysis, asset utilisation, directions for future technology improvements and future epidemic/pandemic preparedness, given the disease-agnostic nature of these vaccine production platforms.

Fig. 1. Computational framework for uncertainty quantification in vaccine production platforms.

The aim of this approach is to evaluate process performance under uncertainty and variability resulting from both the design and operation of the new vaccine production platform technologies. The uncertainty is propagated from input factors to key performance indicators (KPIs) via the mathematical model. Then, the KPI variation ranges are apportioned back to each input factor as sensitivity indices. Input factors include scale of production process, batch failure rate, titre/yield in the production bioreactor, cost of labour, drug substance amount per dose, and cost of quality control. KPIs include capital investment cost requirements, operating costs, number of batches produced per year, amount of drug substance produced per batch, amount of drug substance produced per year, number of doses produced per batch, number of doses produced per year, and production cost per dose.

Fig. 2. Global sensitivity analysis of multiple input factors for AVV, mRNA and saRNA platforms on key performance indicators (KPIs)^–.

The input factors are scale, titre, failure rate, CleanCap purchase price (for mRNA and saRNA only), labour cost, drug substance (AVV or RNA) amount per dose, and quality control (QC) cost. The KPIs are capital costs (CapEx), operating costs (OpEx), number of batches produced per year, amount per batch, amount per year, number of doses per batch, number of doses per year (annual production amounts), and cost per dose. a–h, i–p, q–x: vaccine drug substance production performance for AVV, mRNA and saRNA platforms. a, b, i, j, q, r: first-order effect (S_i) and total effect (S_t) sensitivity indices for the KPIs versus each input factor for the AVV, mRNA and saRNA platforms. Large S_i and S_t values indicate a strong impact of given input factor (X-axis) on KPI (Y-axis), while low S_i and S_t values indicate a weaker dependence of the KPI on the input factor. c–h, k–p, s–x: magnitude of the random co-variation of AVV, mRNA and saRNA drug substance annual production amounts, and cost per dose against production scale, titre, and drug substance amount per dose. Dots clustered around a narrower band indicate that the given input factor explains most of the KPI variance, while dots spread out over a wider band suggest that the input parameter explains little or none of the KPI variance.

Fig. 3. Comparison of the AVV, mRNA and saRNA vaccine production platforms under the uncertainty scenarios in Table 1.

a Violin plots of the required times for producing 1 billion doses of AVV, mRNA and saRNA vaccine drug substance. b Violin plots of the number of vaccine doses produced per year and unit bioreactor working volume. The box and whiskers inscribed within each violin plot depict the interquartile range and full percentile range (excluding outliers), respectively, and the median value is indicated by the white dot inside each box. The bottom 5% and top 5% of all values were excluded from all violin plots to enable a better visualisation of the region of interest around the box plot—cf. Supplementary Fig. 4 for violin plot comparisons on the same x-axis and including the full data range.

Fig. 4. Projection of the resources and capacity needed to produce 1 billion COVID-19 vaccine doses per year using the AVV, mRNA and saRNA platform technologies.

The analysis considers the same uncertainty scenarios as in Table 1, and assumes AVV vaccines filled into 10-dose vials and mRNA/saRNA vaccines filled into 5-dose vials. a Violin plots of the capital costs (CapEx), both without and with fill-to-finish. b Violin plots of the operating costs (OpEx), both without and with fill-to-finish. c Violin plots of the required production process scales, expressed per unit bioreactor working volume. d Violin plots of the required numbers of batches. The box and whiskers inscribed within each violin plot depict the interquartile range and full percentile range (excluding outliers), respectively, and the median value is indicated by the white dot inside each box. The bottom 5% and top 5% of all values were excluded from all violin plots to enable a better visualisation of the region of interest around the box plot—cf. Supplementary Fig. 5 for violin plot comparisons on the same y-axis and including the full data range.