Frank
Roschangar
*a,
Juan
Colberg
b,
Peter J.
Dunn‡
c,
Fabrice
Gallou
d,
John D.
Hayler
e,
Stefan G.
Koenig
f,
Michael E.
Kopach
g,
David K.
Leahy
h,
Ingrid
Mergelsberg
i,
John L.
Tucker§
j,
Roger A.
Sheldon
kl and
Chris H.
Senanayake
a
aChemical Development, Boehringer Ingelheim Pharmaceuticals, Ridgefield, Connecticut 06877, USA. E-mail: frank.roschangar@boehringer-ingelheim.com
bPharmaceutical Sciences - Worldwide Research & Development, Pfizer, Groton, Connecticut 06340, USA
cPfizer, Sandwich, UK
dChemical & Analytical Development, Novartis Pharma, 4002 Basel, Switzerland
eAPI Chemistry, GlaxoSmithKline Medicines Research Centre, Stevenage, SG1 2NY, UK
fSmall Molecule Process Chemistry, Genentech, a Member of the Roche Group, South San Francisco, California 94080, USA
gSmall Molecule Design and Development, Eli Lilly and Company, Indianapolis, Indiana 46285, USA
hChemical and Synthetic Development, Bristol-Myers Squibb, New Brunswick, New Jersey 08903, USA
iProcess Chemistry, Merck, Rahway, New Jersey 07065, USA
jProcess Development, Amgen, Thousand Oaks, California 91320, USA
kMolecular Sciences Institute, School of Chemistry, University of the Witwatersrand, Johannesburg, South Africa
lDelft University of Technology, 2628 BL Delft, Netherlands
First published on 22nd November 2016
Green and sustainable drug manufacturing go hand in hand with forward-looking visions seeking to balance the long-term sustainability of business, society, and the environment. However, a lack of harmonization among available metrics has inhibited opportunities for green chemistry in industry. Moreover, inconsistent starting points for analysis and neglected complexities for diverse manufacturing processes have made developing objective goals a challenge. Herein we put forward a practical strategy to overcome these barriers using data from in-depth analysis of 46 drug manufacturing processes from nine large pharmaceutical firms, and propose the Green Aspiration Level as metric of choice to enable the critically needed consistency in smart green manufacturing goals. In addition, we quantify the importance of green chemistry in the often overlooked, yet enormously impactful, outsourced portion of the supply chain, and introduce the Green Scorecard as a value added sustainability communication tool.
First, to tackle metrics unification, we take advantage of the recently introduced Green Aspiration Level (GAL).15 The GAL is not simply another process metric, but rather represents the first substantial green efficiency goal for any manufacturing process to a given drug. It has the distinct advantage over other traditional waste metrics of defining the expected environmental impact for producing any pharmaceutical agent by considering the complexity of its ideality-adjusted synthesis route.16 With this quantifiable descriptor, process greenness becomes measurable, and as such can deliver manageable and credible green chemistry process goals. This is in line with the proven management adage “you can only manage what you can measure”. We note that the GAL assesses process waste and as such intentionally discounts process energy, safety, and environmental impact, which are integral parts of the complementary Life Cycle Analysis (LCA).17–22 In fact, GAL consensus analysis starting points will streamline any LCA approach, and we therefore envision integration of the two concepts in the future.
Second, to realize the complex goal of aligning industry and broadly implementing the GAL methodology, we explored the concept within two leading green chemistry industry consortia. The International Consortium for Innovation & Quality in Pharmaceutical Development (IQ, https://iqconsortium.org/initiatives/working-groups/green-chemistry) and the ACS Green Chemistry Institute Pharmaceutical Roundtable (ACS GCI PR, https://www.acs.org/content/acs/en/greenchemistry/industry-business/pharmaceutical.html) strongly support the Green Aspiration Level as a valuable tool to evaluate best green chemistry choices. Consequently, we decided to showcase the GAL to assess green drug manufacture. For this purpose, we streamlined the methodology for consistency (ESI Discussion 1†) and defined a single smart process complexity-based waste goal (ESI Discussion 2†).
We carried out an in-depth analysis of 46 manufacturing processes from nine large pharmaceutical firms with good distribution across the drug lifecycle in early development, late development, and commercial manufacturing. The results of our analysis are summarized in Table 1 and detailed in ESI Table 1.† Our process waste data correlate well with those from 2008 data by the ACS GCI PR25 (ESI Table 2†) that had been used to establish the original process step waste goal, termed transformation GAL or tGAL. Further coherence with the 2008 data is observed with respect to solvent and water utilization. The ACS GCI PR reported 86% average water and solvent utilization, while our dataset reveals 87–91% (ESI Fig. 2†). Comparison of simple E factor (sEF, ESI eqn (1)†) and cEF values are a helpful indicator for relative solvent use and thus waste reduction potential.
Phase of drug manufacturing | N | Complexity | sEF [kg kg−1] | cEF [kg kg−1] | GAL [kg kg−1] | RPG | % cEF litd | % (organic solvents + water) |
---|---|---|---|---|---|---|---|---|
a RPG calculated based on updated tGAL = 26 kg kg−1. b All figures represent the means. c N = number of manufacturing process data sets. d % waste derived from literature procedures to fulfill $100 per mol process analysis starting point requirement. e Campaigns making drug supplies for up to Phase IIa/Proof of Concept clinical trials. f Campaigns supporting Phase IIb clinical trials up to registration. g Campaigns providing market supplies. | ||||||||
Early developmente | 16 | 9.4 | 90 | 793 | 245 | 49% | 21% | 87% |
Late developmentf | 15 | 8.0 | 34 | 308 | 208 | 96% | 17% | 91% |
Commercialg | 12 | 5.9 | 15 | 156 | 154 | 132% | 20% | 89% |
Diving deeper into our analysis, a dissection of the cEF waste data shows both the expected reduction in process waste and narrowing distribution of cEF with advancing project phase (Fig. 1). The cEF values of three early development projects included in the analysis are outside the display region with values of 1361, 1744, and 2746.
With implementation of the US $100 mol−1 process analysis starting points (ESI Discussion 1†), and the resulting need to consult literature procedures for raw materials that do not meet this requirement, the 2016 data are expected to show more waste per kg of produced drug, and this is principally observed. In fact, without the standardized starting point rule, an average of 20% process waste would have been excluded from our results, and overall process waste understated accordingly. Thus, we deemed it imperative to apply our new commercial process data to update the tGAL from 19 to 26 kg kg−1 and utilize the revised value for Relative Process Greenness (RPG) evaluation (ESI Discussion 2†).
As one would expect, the average RPG (ESI eqn (2)†) significantly improves and increases over the course of development into commercialization from 49 to 96 and then to 132%. The RPG for the average commercial process is 100%. If the GAL goal had been in place prior to our analysis, it would have been realized (RPG ≥ 100%) with 6 of the 12 analyzed commercial processes (ESI Table 1,† entries 32–37). Our data therefore establish the GAL as the first meaningful and achievable SMART (Specific, Measurable, Achievable, Results-based, Time-bound) green chemistry process goal.
Our study also helps refine the scope of the tGAL goal. For example, three late phase and commercial drugs with highly solvent volume-intensive manufacturing processes typically associated with New Biological Entities (NBEs), polypeptides, carbohydrates, and drug conjugates, would have abysmal RPG ranging from 4 to 20% (ESI Table 1,† entries 44–46) when using the new tGAL value. This emphasizes the need for distinct tGAL for these classes of compounds. Of course, the GAL methodology still applies per se, but a new tGAL value must be established in order to enable smart green chemistry process goals of these compound classes. Thus, the tGAL discussed herein is pertinent only to small molecule New Chemical Entity (NCE) drugs.
The fact that we have simplified GAL methodology with a single commercial tGAL goal brings about relatively low RPG values for early and late development drug processes. Since it is motivationally important to maintain achievable goals across all phases and enable managers and teams to set and strive for smart green chemistry goals, communicate green chemistry process achievements, and encourage green chemistry throughout the entire drug development lifecycle, we create a fair phase-dependent ranking system that is described with our communication strategy vide infra.
But how important is it exactly to manage green chemistry in the supply chain? What is the impact of the pharmaceutical manufacturing supply chain on overall drug process waste generation? How green is the supply chain? To our knowledge we are the first to quantitatively address these important questions.
We include literature step waste in external process waste analysis (ESI Fig. 3†). Our results show that 41–61% or about half of the drug manufacturing process waste is generated externally (Table 2). Thus, green chemistry in the supply chain is vitally important. From the data set we also derive that Relative Process Greenness (RPG) of the overall process (internal and external manufacture) is similar to the external, indicating comparable green process performance, and opportunities to improve, within the pharmaceutical firms and their supply chains.
Phase of drug manufacturing | % External waste | External RPG | Overall process RPG |
---|---|---|---|
a All figures represent the means. | |||
Early development | 41% | 58% | 49% |
Late development | 61% | 109% | 96% |
Commercial | 44% | 145% | 132% |
However, until now managing external green chemistry has proven difficult due to the lack of simple and fair benchmarking concepts.29 Undoubtedly our data show that the new GAL methodology is ideally suited to also establish smart external green process manufacturing goals. For example, the GAL and RPG can readily be integrated into procurement systems and category scorecards, possibly via the green scorecard introduced vide infra. This will allow pharmaceutical buyers to monitor and reward supplier performance and improvements against smart green chemistry goals, and catalyze external green chemistry efforts with constructive feedback based on quantitative and standardized RPG data. In addition, suppliers are empowered to quantify their contributions and evaluate their own supply chain. Overall, the GAL methodology has the potential to inspire environmental impact reductions within the supply chain, stimulate external green innovation, and foster best green chemistry practices from “cradle to grave”.
The calculator template has been made available via free download from the IQ GC website (https://iqconsortium.org/initiatives/projects/green-aspiration-level). To use the Green Scorecard calculator, one simply inputs project name and phase, complexity, sEF, and cEF for one or more completed campaign in order to obtain the calculated RPG and its rating. The impact of innovation and waste reductions via RPI, RCI, and PI are auto-calculated if two or more campaigns are entered (ESI Discussion 3†).
The phase-dependent Green Scorecard ratings are derived from our dataset via probability plots of RPG at the 90, 70, and 40 percentile levels (ESI Fig. 5†). We assign four process greenness ratings of excellent = 90 percentile, good = 70 percentile, average = 40 percentile, and below average for the bottom 40% RPG values. This results in the phase-dependent RPG ratings matrix of Table 3.
Thus, the Pradaxa process with a RPG of 222% falls into the commercial 70 percentile, i.e. is estimated greener than 70% of commercial industrial process in terms of waste quantity and rated “good”.
The Green Scorecard will be the integral part of our green chemistry implementation and communication strategy. It holds the potential to cascade broad adoption of the GAL throughout the pharmaceutical industry and its supply chain into process chemists and engineers daily routine, since it forcefully yet fairly communicates the objective value added of green process chemistry via quantitative and goal-driven data, unlike any other green chemistry metric to date.
It is conceivable that with increasing acceptance the GAL will become a mainstream tool beyond pharmaceutical manufacturing and find application as key criterion in green chemistry award selection processes. We consider the GAL methodology complementary to the LCA tool from the ACS GCI PR mentioned vide supra, and can envision integration of both concepts. In fact, incorporation of GAL's consensus analysis starting points will enhance consistency of the streamlined LCA approach. “History teaches us that most breakthrough innovations are … [integrated] … through networks”,31 so we will continue to work through the IQ and ACS GCI PR consortia to foster our vision.
Footnotes |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6gc02901a |
‡ Retired. |
§ Current address: Neurocrine Biosciences, San Diego, California 92130, USA. |
This journal is © The Royal Society of Chemistry 2017 |