Landscape and opportunities for active pharmaceutical ingredient manufacturing in developing African economies

Darren L. Rileya, Ian Strydoma, Rachel Chikwambab and Jenny-Lee Panayides *b
aDepartment of Chemistry, Faculty of Natural and Agricultural Sciences, University of Pretoria, Pretoria, 0028, South Africa
bCouncil for Scientific and Industrial Research, Brummeria, 0184, South Africa. E-mail: JPanayides@csir.co.za

Received 1st October 2018 , Accepted 16th January 2019

First published on 16th January 2019


Africa is one of the world's fastest growing economies, with South Africa having the fifth highest worldwide pharmaceutical expenditure per capita. In recent years, several companies have considered regional pharmaceutical production but have failed to make the investment, in stark contrast to the massive growth in pharmaceutical production in other BRICS countries. Major constraints identified have been the small local market, lack of skills, and an export-averse culture, which have prevented regional manufacturers from achieving the economies of scale that are essential to survive in a global market. In contrast, the pharmaceutical industry is undergoing a revolutionary change in manufacturing, with the potential to switch from batch manufacturing to continuous flow processing. The possibility of applying this new pharmaceutical business model in emerging markets will open the door for dramatic changes in regional commercial manufacturing. Advances in cloud computing, automation and system unification are paving the way for continuous active pharmaceutical ingredient production with integrated digital connectivity. This review will highlight the opportunities that exist in the localization of cutting-edge manufacturing technologies; in order to show the potential application of fundamental process research key production examples relevant to the region will be provided.


image file: c8re00236c-p1.tif

Jenny-Lee Panayides

Dr Jenny-Lee Panayides is a research group leader in Pharmaceutical Technologies at the Council for Scientific and Industrial Research (SA), where she is spearheading the organisation's pharmaceutical development program. Her research focuses on the application of high-throughput screening for hit identification, investigation of novel solid-supported catalysts to process development, and on the translation of batch and flow syntheses to stimulate the local production of active pharmaceutical ingredients. Jenny-Lee holds a PhD in the fields of organic chemistry and microbiology from the University of the Witwatersrand (SA) and a certification in project management from the University of Pretoria (SA).


1. Introduction

1.1 Pharmaceutical need on the African continent

Africa is one of the world's fastest growing economies, with gross domestic product (GDP) growth predicted to reach USD 3.3 trillion by 2020. However, Africa has a disproportionate burden on disease, being home to 75% of the world's HIV/AIDS infections, 90% of malaria deaths, high rates of tuberculosis resistance and many other infectious diseases that cause substantial morbidity and mortality. The impact of infectious diseases is mostly prevalent in sub-Saharan Africa, with the North African regions having disease profiles similar to those of industrialised countries, where cardiovascular disease, diabetes and cancer dominate. Non-communicable diseases (NCDs) are becoming increasingly prevalent on the continent given the demographic changes that are taking place, with predictions that the NCDs will overtake infectious diseases as the leading cause of death in Africa by 2030.1

The pharmaceutical industry market value rose from USD 4.7 billion in 2003 to USD 20.8 billion in 2013, with growth predicted to reach USD 40–60 billion by 2020 and 160.7 billion by 2024 at a compounded annual growth rate (CAGR) rate of 20.4% over the forecast years.1,2 More than 300 drug companies are active in drug production across Africa, including multinationals such as Johnson & Johnson (South Africa) and Sanofi (Algeria), and incentives have been provided to build manufacturing plants, including tax exemptions, reduced land prices, and others (Algeria, Nigeria, South Africa). Factors currently driving growth in this sector include the rise of major cities, expansion in healthcare capacity across the continent and the maturing of the business environment.

Over the last few decades, several private sector companies have actively considered regional active pharmaceutical ingredient (API) production but have almost without exception failed to make an investment and to build local capabilities. This experience is in stark contrast to the massive growth in API production in other countries, including India, China and Brazil. The major constraints to growth in Africa have been a small local market,3a a lack of skilled entrepreneurs,3b,c the scarcity of skilled human capital required by the API industry,3b,c poor logistics infrastructure and long lead times,3d,e an immature regulatory system,3f,g and an export-averse culture;3c,h which together have prevented regional API manufacturers from achieving the necessary economies of scale and competitive positions that are essential to survive in a global market (Table 1).3i Nearly all of the APIs (with a few exceptions in South Africa, Egypt, and Ghana) used in Africa are imported, and the quality standards to which the manufacturers adhere varies significantly between and within countries.1a

Table 1 SWOT analysis of the African pharmaceutical manufacturing landscape1–3
STRENGTHS WEAKNESSES
• Heightened political interest to support local API production • Regulatory oversight varies, it is mostly insufficient to protect citizens and allow fair competition
• Demonstrated that high quality production of essential medicines is feasible in Africa • Currently heavily dependent on imports, which usually benefit from export incentives
• Strengthening of legal and regulatory frameworks • Non-conductive environment, policy incoherence and absent sector development strategies
• Existence of scientific and technical institutions with the relevant skills in a number of countries • High utility costs and unreliable supply
• Increasing political stability and rapid economic development provide a strong basis for development of the industry • Limited availability of affordable finance of sufficient magnitude and duration constrains the sector's ability to invest in necessary upgrades
• Developments in health insurance schemes and health systems with increasing investment in public health • Small domestic market and fragmented regional markets hamper the ability to manufacture efficiently
• High disease burden linked to a high demand for drugs, stimulating low-cost generic medicines • For many companies the systems, processes, premises and equipment at plant level are insufficient to ensure quality of production
• Rapid urbanisation, sedentary lifestyles, and dietary trends will lead to long-term demand for chronic and lifestyle-related disease pharmaceuticals • Limited supporting industries for local production of APIs, packaging and excipients; availability of services to maintain and repair equipment; and to conduct trials to attest to the suitability of products
• Some countries have well established pharmaceutical sectors and regulatory institutions • WHO pre-qualification scheme covers only a limited range of products, the quality assurance for most medicines is under the authority of the national medicines regulatory authorities
• Comparatively low labour costs, especially in the face of rising Indian labour costs
• Some countries are already exporting to industrialized markets

OPPORTUNITIES THREATS
• Awareness of importance of local production amongst international partners and donors, and growing commitment to support this agenda • Without holistic and robust approach, efforts from committed manufacturers will be undermined by ongoing penetration of counterfeit medicines
• Evidence suggests local manufacturers can compete with imports and produce high-quality products given a conductive environment • An ongoing lack of evidence as to the prevalence and impact of sub-standard and counterfeit medicines could undermine political will to develop the industry
• Creating more favourable market conditions will encourage investment in the sector • Development of the sector requires coordination across public health and industrial development components
• Development of the sector and requisite legislation will enable TRIPS flexibilities to be utilized • Without concerted action, reliance on Asian imports may increase and quality of products may be increasingly suspect
• Increased prevalence of lifestyle diseases can be more comprehensively addressed through local production (growing market opportunity) • If donor assistance plateaus a possible increase in sub-standard products for pandemic diseases could jeopardize the health of patients and lead to accelerated increase in drug resistance
• Training programmes and institutions exist but need to be expanded and scaled up • Donor initiatives that can inadvertently have detrimental consequences for the African pharmaceutical industry
• Potential for countries to export to other markets
• Ongoing work under the African medicines regulatory harmonization initiative could lead to defragmented markets on the sub-regional level
• Strong interest from international generic industry


In response to these challenges and the forecasted reductions in the level of healthcare donations support, the 54-member states of the African Union (AU) ratified the Pharmaceutical Manufacturing Plan for Africa (PMPA) in 2007.1 The PMPA aims to benefit all the member states through improved access to affordable, safe and efficacious essential medicines, independent of donations. The PMPA is premised on the belief that industrial development of the pharmaceutical sector will contribute to improved public health outcomes, provided that the development of the industry is based on the principle that all manufacturers supplying pharmaceuticals on the continent should ultimately meet international standards of production. The context in which pharmaceutical manufacturing is envisioned to take place will include a system of manufacturers, national medicine regulatory authorities, various government ministries, trade associations, and an array of distribution channels.1

Over the last decade, however, limited progress has been made towards the implementation of pharmaceutical manufacturing on the continent. Through the World Health Organization (WHO) department of Public Health Innovation and Intellectual Property, a study was conducted to analyse how industry, trade and health policies can work together towards shared goals to both further economic development and improve public health.1b The study highlights the challenge governments face in balancing the availability of quality-assured medical products that meet priority public health needs with ensuring that products are acceptable and affordable. Imports of affordable medicines are one solution, but the report highlights a growing trend towards domestic production – with case studies from countries that are building a viable local manufacturing industry (including Ethiopia and Uganda) providing evidence to support this trend.1c Through additional funding from the European Union, additional support was provided by the WHO to launch the National Strategy and Plan of Action for Pharmaceutical Manufacturing Development in Ethiopia.1d

While progress has been made in implementing aspects of the PMPA, critical components, such as attainment of international standards of Good Manufacturing Practices (GMP) and building of the requisite capacity in the industry, are lagging behind. This is because these mandates require, among other things, the physical modification of most existing manufacturing facilities, construction of new structures, procurement, installation and qualification of GMP-compliant equipment and building of human capital in technology and innovation.1e These activities are capital intensive and require long-term low-cost financing, which currently falls outside the scope of most facilities of the African banking and financial systems. The full and timely implementation of PMPA is therefore at risk of failure. As such, there have been calls for the establishment of the Fund for African Pharmaceutical Development (FAP-D) to address the critical issue of access to capital by the industry and enable full and successful implementation of the PMPA, which were endorsed by the Executive Council of the African Union (decision EX.CL/Dec.970 (XXXI), July 2017).1e The African Union, in partnership with the African Development Bank and the African Import and Export Bank, held a three-day consultation event in August 2018 to discuss practical steps towards the creation of a fund to support pharmaceutical manufacturing in Africa.1f It is envisioned that the FAP-D will provide affordable financing for activities geared toward attainment of GMP, capacity building and enabling growth of the African pharmaceutical sector, as well as providing technical advisory services, overseeing efficient use of financing, supporting partnerships and collaborations to enable private sector engagement and seeking solutions through African innovation.

In addition, the Department of Trade and Industry of the African Union Commission organized the 7th Strategic Stakeholders Retreat on Industry under the theme “Implementation Review of the Plan of Action for the Accelerated Industrial Development of Africa”. The implementation of the PMPA business plan requires the coordination of many actors across the pharmaceutical manufacturing system; therefore the need for the establishment of a dedicated task force to focus on implementation of the PMPA was highlighted.1g

1.2 Current trends, drivers and challenges for regional API production

Recently, the Southern African Development Community (SADC) industrialization strategy was developed as an inclusive long-term modernization and economic transformation scheme that enables substantive and sustained rising of living standards, intensifying structural change and engendering a rapid catch up of the SADC countries with industrializing and developed countries. It is anchored on three interdependent and mutually supportive strategic pillars: (i) industrialization as a champion of economic transformation, (ii) enhancing competitiveness, and (iii) deeper regional integration.4,5

The SADC region has a well-established manufacturing industry and is relatively well constructed in most areas including the procurement of APIs, packaging, labelling, distribution and the sale of pharmaceuticals. However, this has not translated to pharmaceutical manufacturing strength. Notably, South Africa is the only country in the SADC region that currently meets the GMP standards of the WHO. Furthermore, the SADC free trade area ensures tariff-free access to neighboring markets which continue to rely on South Africa as their main source of pharmaceutical products.6

1.3 Detailed analysis of the South African pharmaceutical sector

In terms of human health, the South African pharmaceutical market is the largest in sub-Saharan Africa, with one of the highest worldwide expenditures on pharmaceuticals per capita.1 In the 2016 financial year, pharmaceutical sales in South Africa were estimated at USD 3.25 billion (including all taxes and mark-ups), contributing 1.08% to the South African GDP.7 The sector has a two-tier market structure, private and public, and due to the limited investment in public sector healthcare, a large disparity exists between the two markets. According to recent statistics, in 2015 the private medical sector accounted for 84.5% of the total pharmaceutical market in South Africa, while serving only 16% of the population; with the public sector contributing 15.5% while serving 84% of the total population.7

South Africa has a well-established capacity in the procurement and distribution of pharmaceuticals. The local manufacture of APIs and biopharmaceuticals is almost non-existent, and current data indicate that 37% of the total pharmaceutical market value is derived from the total value of the APIs; however only about 2.3% of these are locally produced.8 The situation for anti-retrovirals (ARVs) is much worse, as there is currently no local ARV API production and the country has suffered from critical drug shortages.9 In contrast, the local formulation sector is comparatively strong; 59% by value of the total pharmaceutical market is locally formulated.

There is a large number of registered manufacturing entities in South Africa, including a mixture of large sites producing several thousand doses per year and smaller, relatively unsophisticated facilities. Most major pharmaceutical multinational corporations have a presence in South Africa and have been focused predominantly on marketing branded, innovative and generic drugs to the private sector. Pfizer, GlaxoSmithKline and Sanofi have local manufacturing facilities, whilst the other international players are present through representative offices and exports.7 The local pharmaceutical industry is relatively well developed and mostly focused on the production of generic products. Local production is driven by two key players, Aspen Pharmacare and Adcock Ingram, both of which are GMP certified and have WHO pre-qualification status.7

South Africa's pharmaceutical market is dominated by prescription drug spending, including patented and generic drug expenditure, which accounts for approximately 88% of the total market. Over-the-counter (OTC) spending represents the remainder of 12%. Of the three subdivisions, patented medicines represent the largest single segment of the total market by value, at approximately 53.7% (2016), followed by generic drugs with 34.4% of total medicine spend.7 Prospects for spending on pharmaceuticals are largely limited to the private sector, which is coming under increasing cost containment measures. Soaring unemployment in SA has drastically lowered the purchasing power, but more significantly has also removed a sizable portion of patients that were previously covered by health insurance. The public sector, funded by the state, purchases the majority of all medicines in volume terms, but this represents a much less significant market share in value terms. Most of the publicly purchased medicines are also generics.

Pharmaceutical exports, whether at the level of API or formulated medicines, account for only 4% of the total local formulated product output. As a result, there is a huge trade imbalance, with the pharmaceutical industry being the fifth largest contributor to South Africa's trade balance deficit, and providing a large foreign exchange burden for the sector (Fig. 1).6 South Africa's pharmaceutical exports into the region are however expected to become increasingly competitive, driven by increased productivity and a favourable exchange rate for trade, making South Africa well positioned to exploit the surrounding highly import-dependent African markets.7


image file: c8re00236c-f1.tif
Fig. 1 Pharmaceutical trade forecast.6

While South Africa's largest killer remains the human immunodeficiency virus infection and acquired immune deficiency syndrome (HIV/AIDS) epidemic, in terms of pharmaceutical sales, the biggest sellers are treatments for diseases such as heart disease and hypertension, which predominantly affect a richer population profile. Nearly 100 people die of heart attacks or strokes each day in South Africa, although this figure pales in comparison to the estimated 1000 people that die of AIDS. More than 7 million adults smoke, around 6.3 million suffer from hypertension and 5 million have high cholesterol. According to recent statistics, 9.3 million disability-adjusted life years (DALYs) were lost to non-communicable diseases in 2016, compared to 9.5 million DALYs lost to communicable diseases.7 This demonstrates a consistently strong demand for medications and significant commercial opportunities for drugmakers over the long term. South Africa's evolving demographic and epidemiological profile will provide increased revenue earning opportunities for pharmaceutical companies, particularly those producing non-communicable disease treatments. Despite this, the existing healthcare inequalities in the country will hinder much of the population's access to these essential medicines, and downward pressure on the rand will weigh on the interest of multinational drugmakers. As such, the ability to innovate the manufacturing and distribution channel is critical for development.

1.4 Policy and regulatory opportunities in the sector

The main piece of patent legislation in South Africa is the Patent Act of 1978.10 Under this act patents have a lifespan of 20 years, but after three years an annual renewal fee becomes payable. In addition, the Medicines and Related Substances Act 101 of 1965 (amended in 2014) is in place which focuses on governance pertaining to the supply, registration, licensing and management of medicines within South Africa.11 Despite these provisions, patent protection has been the cause of friction between the South African government and the international drug industry.7 Under the existing patent law, exclusive rights are granted to the inventor/assignee for a limited period of time, in exchange for the public disclosure of the invention. In recent times however, it has become the practice of a number of innovator companies to extend the patent term of their innovative molecules to maintain market dominance. This practice unfairly blocks generic competition and keeps drug prices high, often limiting the patient's ability to access life-saving medications. South Africa currently has plans to amend the intellectual property (IP) law to bring into force public health safeguards that are absent in the current patent system. These changes would, among other initiatives, ensure that companies cannot unfairly extend monopolies by simply changing the formulation or combining two medicines into a single tablet, and registering a new patent for this obvious change. It is expected that South Africa's new IP policy would modernize the country's laws to align with those of other middle-income countries and international norms and would set a positive example for other African countries who are also involved in patent law reform processes.

South Africa is a signatory to the World Trade Organisation's (WTO) Trade-Related Aspects of Intellectual Property Rights (TRIPS) agreement, which offers underutilized opportunities to assist and promote the development of pharmaceutical manufacturing and to contribute to improved public health outcomes. TRIPS attempts to strike a balance between the long-term social objective of providing incentives for future inventions and creation and the short-term objective of allowing people to use existing interventions and creations. The TRIPs agreement obliged all WTO members to provide patent protection for pharmaceuticals, defined the exclusive rights conferred to patent owners, limited the possible exceptions to such rights, and determined the conditions for granting of compulsory licenses. It also introduced, for the first time in an international agreement, the obligation to protect data against unfair competition.12 The TRIPS agreement essentially does this by allowing for compulsory licensing, whereby a government can allow someone else to produce the patented product/process without the consent of the patent owner.13 Unfortunately, the TRIPS flexibilities have generally not been utilized; there are limited links between industry and academia, and collaboration between companies in Africa, as well as with international manufacturers, is quite rare.

In addition, in 2005 the WTO approved legal status when they adopted the protocol amending the TRIPS agreement to provide an additional legal pathway for WTO members with insufficient or no manufacturing capacity in the pharmaceutical sector to access medicines. Unlike other flexibilities in the TRIPS agreement, the Paragraph 6 System was devised as a new mechanism for countries without previous domestic experience to draw upon. This mechanism entitles WTO members to grant a special type of “trade-related” compulsory license permitting the production of medicines exclusively for export to meet the needs of other WTO members.14 These underutilized opportunities offer substantial potential benefits, including expansion of the current range of locally manufactured products.1

2. Emergence of continuous manufacturing

Over the course of the last 150 years, synthetic and process chemistry has become largely reliant on batch processing technologies, which although well refined by today's standards, still suffer from certain inherent inefficiencies which are intrinsically linked to having to perform batch-by-batch production. In contrast, over the last 15–20 years the use of flow technology in which reactions are performed continuously in flowing streams has been shown to overcome many of these inefficiencies, resulting in a rapidly growing interest particularly with respect to its use as a tool for the manufacture of pharmaceuticals. Recently, the global flow chemistry market size was valued at USD 1023.7 million in 2016 and was projected to witness considerable growth as a result of rising penetration of the product in pharmaceutical applications.15

The technology at this stage is disruptive in nature and as a result, buy-in from countries with established pharmaceutical manufacturing sectors has been slow, despite advantages afforded by the technology. The technology provides a high level of safety as reactions are occurring on a micro scale and the system operates under steady-state conditions,16–18 which is likely to have a positive impact on the market growth. In addition, properties such as less space occupancy, controlled environment, improved mixing efficiency, intensified heat transfer, low maintenance, and reduced reaction steps are further anticipated to propel the demand.16 The technology, in addition, is often regarded as green technology on account of its superior and environmentally friendly properties which typically include a reduced reactor footprint and improved space-time.

Looking forward, the increasing focus towards cost-effectiveness, safety, and environmental regulations coupled with the CAGR of over 10% observed in the pharmaceutical sector (2016–2025)15 is expected to have a positive impact on the demand for the continuous production of fine and specialty chemicals. In addition, the recent approval and promotion of the use of the technology by the United States Food and Drug Administration (FDA) for the manufacture of APIs under cGMP conditions will further act to push the adoption of the technology. Furthermore, the adoption of the technology is also currently being advocated by such bodies as the Royal Society of Chemistry and the Green Chemistry Society.19 That being said, shortcomings of flow technology, particularly those linked to technical issues such as the handling of solid materials, costly capital, research and development costs, regulatory issues and existing investment in batch-based plants are anticipated to restrain the flow chemistry market.

Regionally, Africa is in a unique position in that there is in general limited existing pharmaceutical manufacturing infrastructure to hinder the adoption of a new disruptive technology. In very much the same way as Africa was able to leapfrog traditional fixed-line telecommunications infrastructure and move straight into cellular networks, existing pharmaceutical manufacturing constraints are viewed as an opportunity to leapfrog existing batch-only manufacturing. As a result, the adoption and use of flow technologies has received widespread interest and has been identified as a critical development area at both industrial and government levels within the SADC region and in particular within South Africa where there is currently a strong push for the localization of pharmaceutical manufacturing.

3. Review of existing process routes of high relevance to the African continent

The United Nations (UN) estimated the Africa region's 2015 population at close to 1 billion people, with the WHO statistics indicating 9.2 million deaths from all causes in 2015 in the region (note that WHO's Africa region consists of 47 of 54 countries on the continent). The WHO classifies the causes of deaths into three broad groups: (i) communicable, maternal, newborn and nutritional conditions; (ii) non-communicable or chronic diseases, and (iii) those caused by injuries.20 In 2015, communicable conditions accounted for 5.2 million of deaths (56.4%), including lower respiratory tract infections (1 million), HIV/AIDS (760[thin space (1/6-em)]000), diarrhoeal diseases (643[thin space (1/6-em)]000), tuberculosis (434[thin space (1/6-em)]000) and malaria (403[thin space (1/6-em)]000). Non-communicable diseases accounted for 3.1 million deaths (33.5% of all deaths), including stroke (451[thin space (1/6-em)]000), ischaemic heart disease (441[thin space (1/6-em)]000), and cirrhosis of the liver (174[thin space (1/6-em)]095). Deaths caused by injury accounted for 930[thin space (1/6-em)]000 of the total deaths in the region (10.1%), including deaths from road injuries (269[thin space (1/6-em)]000), interpersonal violence (102[thin space (1/6-em)]000), and self-harm (87[thin space (1/6-em)]000) (Fig. 2).20
image file: c8re00236c-f2.tif
Fig. 2 Diseases causing the most deaths in Africa (2015).

Over the last two decades, the African continent has undergone major demographic, social, and economic changes, which in turn have led to significant lifestyle shifts.21 All of these trends are the perfect backdrop for a rapid transition, resulting in a triple burden of malnutrition, infectious diseases, and non-communicable diseases exploding on the continent.22 This has been observed by a shift in the leading causes of death, with stroke and ischaemic heart disease having replaced tuberculosis and malaria in the top five primary causes of death on the continent since 2010.

In this section, an overview highlighting selected process routes to diseases with a high African relevance is given. The section will focus largely on routes to drugs making use of continuous and flow-based technologies and, where appropriate, contrasting these approaches against existing batch routes. The reader is further referred to a number of recent reviews describing in more detail flow routes for the preparation of APIs.23–25

3.1 HIV/AIDS

HIV targets the immune system and acts by impairing and ultimately destroying the function of immune cells.26 As a result, afflicted individuals become immunodeficient over time, resulting in increased susceptibility to secondary infections. Left untreated, the disease progresses to AIDS over the course of 2 to 15 years, earmarked by the development of secondary infections, diseases and cancers.26 Africa is the most heavily affected region in the world, and in 2015 approximately 26 million people were living with HIV in Africa, with sub-Saharan Africa being most heavily affected, accounting for 65% of all new infections and 70% of all AIDS-related deaths.26 The structures of a selection of current frontline HIV drugs including lamivudine 1, zidovudine 2, abacavir 3, nevirapine 4, efavirenz 5, atazanavir 6 and tenofovir disoproxil fumarate 7 are shown in Fig. 3. The reader is referred to the review by Watts and co-workers for further insight into batch-based process routes to HIV drugs.27
image file: c8re00236c-f3.tif
Fig. 3 Structures of a selection of HIV drugs.
Lamivudine. Lamivudine 1 is an HIV-1 and -2 reverse transcriptase inhibitor marketed by GlaxoSmithKline that is commonly used in combination with other antiretrovirals such as zidovudine 2 and abacavir 3.27 Several process routes for lamivudine 1 have been developed, almost all involving the initial synthesis of a 1,3-oxathiolane core commonly followed by coupling to silylated cytosine. The requisite stereochemistry is typically established using either a chiral auxiliary in the form of L-menthol, enzyme-catalyzed transformations or enzymatic resolution.27 An efficient synthetic approach to lamivudine 1 was reported by Whitehead and co-workers in 2005 wherein a Vorbrüggen reaction with L-menthyl glyoxylate 8 and 1,4-dithiane-2,5-diol 9 followed by selective crystallization from n-hexane afforded 1,3-oxathiolane 10 as the desired pure isomer.28 Subsequent chlorination afforded 11 which underwent N-glycosidation with silylated cytosine giving 12. Finally, reductive cleavage of the auxiliary with sodium borohydride afforded 1 in an overall yield of 43% (Scheme 1).
image file: c8re00236c-s1.tif
Scheme 1 Whitehead approach: (i) toluene, Δ, (ii) 9, (iii) n-hexane, NEt3 crystallisation (80%, 3 steps), (iv) SOCl2, DMF, DCM, (v) NEt3, toluene, (vi) NEt3, H2O, n-hexane (66%, 3 steps), (vii) NaBH4, EtOH, 83%. Caso approach: (vi) (a) AcOH, toluene, Δ, (b) 9, Δ, (c) 0 °C, NEt3 n-hexane, 60%, (vii) Ac2O, Py, rt, 91%, (viii) −20 °C, NEt3 n-hexane (recrystallisation), 42%, (ix) (a) polymethylhydrosiloxane/I2, CH2Cl2, (b) silylated cytosine, DCM, rt, 90%, >99% ee, (v) NaBH4, EtOH. Watts approach: (vi) a) n-butanol, cat. AcOH, 9, 110 °C, 82% or solvent free, 9, 110 °C, 93%, (b) 0 °C, 1% NEt3 in heptane, (vii) Ac2O, NaHCO3, CH3CN, rt, 93%, (viii) −20 °C, NEt3 n-hexane (recrystallisation), 46%, (ix) pyH·OTf, silylated cytosine, CH3CN, 80 °C, 82%, (v) K2HPO4/H2O, NaBH4/MeOH, rt, 88%.

More recently, Caso and co-workers modified the approach by removing the undesirable chlorination with thionyl chloride, instead choosing to acetylate the racemic version alcohol 10 prior to selective recrystallization with n-hexane.29 N-Glycosidation with silylated cytosine was then achieved in the presence of silane and iodine. Finally, reductive cleavage of the chiral auxiliary with sodium borohydride afforded 1. In this case the authors achieved an overall yield of 18.7% (Scheme 1). Watts and co-workers re-looked at the Whitehead process and redesigned the Vorbrüggen reaction, affording 82% conversion when using n-butanol as a solvent and catalytic acetic acid, or 93% under solvent- and catalyst-free conditions.30 Thereafter, they chose to follow the acetylation approach of Caso, affording 13 in similar yield and achieved N-glycosidation in the presence of pyridinium triflate, affording 1 in an overall yield of 28.7% (Scheme 1).

The Watts group subsequently demonstrated a modified three-stage semi-continuous flow-synthesis utilizing a similar approach (Scheme 2).31 L-Menthyl glyoxylate 8 and 1,4-dithiane-2,5-diol 9 were reacted by passage through a 2 mL glass reactor heated to 90 °C. The effluent stream was fed through a 10 psi back pressure regulator (BPR) and mixed with an acetic anhydride/pyridine solution at ambient temperature in two 0.2 mL glass reactors connected in series. The approach afforded the desired acetylated 1,3-oxathiolane 13 in 95% yield with a residence time of 9.7 min.


image file: c8re00236c-s2.tif
Scheme 2 Watts three-stage flow synthesis of lamivudine 1.

In the second stage the N-glycosidation of 13 is achieved by mixing 13 with pyridinium triflate both in acetonitrile in a 0.2 mL glass static mixer. The reaction stream was then mixed with silylated cytosine in a T-piece mixer prior to heating at 80 °C in a polytetrafluoroethylene (PTFE) tubular reactor (3.8 mL), affording racemic 12 in 95% yield with a residence time of 8.4 min. An off-line recrystallization (1% triethylamine/hexane) afforded the desired isomer of 12 in 46% yield.

The chiral auxiliary was removed by mixing 12 with K2HPO4/H2O in a 0.2 mL glass static chip which was then combined with a stream of sodium borohydride in basic water in a T-piece mixer prior to maturation in two 0.2 mL glass reactors connected in series, affording lamivudine 1 in 94% yield with a residence time of 3.3 min. The process represents a rapid 21.4 min (excluding off-line processing in stage 2) synthesis of lamivudine 1 in an overall yield of 41%. Notably, the process afforded comparable yields to that of Whitehead and co-workers without having to resort to the use of undesirable thionyl chloride.

Nevirapine. Nevirapine 4 is a non-nucleoside reverse transcriptase inhibitor (NNRTI) originally developed by Boehringer Ingelheim Pharmaceuticals.32 The drug's mode of action involves the inhibition of the essential viral reverse transcriptase enzyme. Several process routes have been developed, most of which are variants of each other. The patented Boehringer Ingelheim process details the typical approach used (Scheme 3) wherein an amide coupling between 2-chloro-4-methylpyridin-3-amine 14 and 2-chloronicotinoyl chloride 15 under basic conditions affords 16; thereafter an ipso-substitution with cyclopropyl amine 17 gives 18, and base mediated cyclisation affords nevirapine 4 in 83% yield (287.2 kg scale, >9 h reaction time).32 Subsequent routes largely utilize analogous approaches with subtle changes in the choice of reagents. In a few instances the ordering of reactions has been modified with the ipso-substitution taking place before the amide coupling and cyclisation.
image file: c8re00236c-s3.tif
Scheme 3 Boehringer Ingelheim process: (i) base, (ii) CaO, (iii) NaOH.

In 2013 McQuade and co-workers demonstrated a flow synthesis of 2-bromo-4-methylnicotinonitrile 19 which can be used as an analogous starting material in the synthesis of nevirapine 4 (Scheme 4).33 The synthesis involves an initial Knoevenagel condensation between acetone 20 and malononitrile 21 which was achieved by passing a mixture of the two materials through a packed bed reactor (PBR) containing aluminium oxide at 95 °C followed by a PBR housing 3 Å molecular sieves at 20 °C (∼2 min residence time) affording 22. Thereafter, the reaction stream was reacted with dimethyl formamide–dimethyl amide and acetic anhydride in dichloromethane in a T-piece mixer, affording 23, prior to maturation in a coil reactor heated to 95 °C (4 min residence time) which afforded the cyclisation. Finally, the reaction mixture was collected in HBr/acetic acid, facilitating the bromination to give 19 in 69% yield.


image file: c8re00236c-s4.tif
Scheme 4 Flow synthesis of important intermediates in the synthesis of nevirapine 4.

More recently in 2017 Gupton and co-workers developed a complete flow synthesis of nevirapine starting from 2-chloro-4-methylnicotinonitrile 24 (Scheme 5).34 Initially a 3 M solution of 24 and a 3 M solution of sodium hydride both in diglyme were reacted in a spinning disk reactor at 95 °C. The reaction mixture was then condensed with methyl 2-(cyclopropylamino)nicotinate 25 in a continuously stirred tank reactor (CSTR) prior to passage through a PBR housing sodium hydride at 165 °C affording nevirapine 4 in 92% overall yield. The process showed significant improvements linked to yield, process intensification and waste generation when benchmarked against the Boehringer Ingelheim approach.


image file: c8re00236c-s5.tif
Scheme 5 Gupton flow-based synthesis of nevirapine 4.
Efavirenz. Efavirenz 5 is a NNRTI developed in the Merck Laboratories for the treatment of HIV-1.35 The drug has a stereogenic quaternary carbon group with an (S) configuration. Traditionally, most synthetic approaches start from 1-(2-amino-5-chlorophenyl)-2,2,2-trifluoroethanone 26 or a structurally analogous species and are then subjected to a stereoselective condensation either mediated using chiral reagents or chiral protecting groups. Thereafter, a phosgene-mediated cyclisation affords efavirenz 5. The Merck process route (Scheme 6) describes the process starting from p-chloroaniline 27 which in four steps is converted to 26 (84.4% yield across four steps).35 Thereafter, the aryl amine is protected, followed by a stereoselective condensation with lithiated cyclopropylacetylene 28 to afford 29 (91–93% yield, >99.5% ee). Cyclisation with phosgene under basic conditions facilitates the cyclisation to 30 (95%) and finally deprotection with ceric ammonium nitrate affords efavirenz 1 in an overall yield of 62% across the seven steps.
image file: c8re00236c-s6.tif
Scheme 6 (i) NaOH (aq), MTBE, 97%, (ii) (a) n-BuLi, TMEDA-MTBE, (b) CF3CO2Et, (iii) HCL-HOAc, (87%, 2 steps), (iv) NaOAc (aq), MTBE, (v) cat. TsOH, p-methoxybenzyl, CH3CN (90%, 2 steps), (vi) THF, (vii) COCl2, NEt3, toluene, 95%, (viii) CAN, EtOAc/H2O, 76%.

In recent times, both Seeberger and Watts have turned their attention to the development of a flow-based synthesis of efavirenz. In the case of the Seeberger team a concise three-stage semi-flow synthesis of rac-efavirenz 5 was realized (Scheme 7).36 In the first stage, 1,4-dichlorobenzene 31 was ortho-lithiated by reaction with n-BuLi in a coil reactor cooled to −45 °C (4 min residence time) and subsequently quenched with trifluoroacetylmorpholine 32 in a second coil reactor again at −45 °C (13.3 min residence time), connected in series to a third coil reactor at −45 °C (2 min residence time). Thereafter, the reaction was quenched by passage through a PBR housing anhydrous silica at −10 °C affording 33 in 87% yield. The second stage utilized an analogous approach to that in stage 1. In this instance, cyclopropylacetylene 34 was lithiated in a coil reactor at −20 °C (1 min residence time) prior to reaction with 33 in a second coil reactor again at −20 °C (2 min residence time); the reaction mixture was then collected and processed off-line, affording 35 in 92% yield.


image file: c8re00236c-s7.tif
Scheme 7 Seeberger semi-continuous synthesis of efavirenz 5.

Finally, the third stage, cyclisation, was achieved by passing a mixture of 35, Cu(OTf)2 and trans-N,N′-dimethyl-1,2-cyclohexanediamine (CyDMEDA) across a PBR housing celite, Cu(0) and NaOCN (60 min residence time) affording rac-efavirenz 5 in 62% yield after off-line processing. Although the process is not stereoselective it should be noted that it was conducted as a proof-of-concept study and in three steps represents the shortest reported route to efavirenz 5. The copper-catalyzed formation of an aryl isocyanate and subsequent intramolecular cyclisation is further desirable as it negates the need for using toxic phosgene.

In the case of the route designed by Watts and co-workers, N-Boc-4-chloroaniline was lithiated in a microreactor utilizing n-BuLi to afford an analogous trifluorinated precursor.37 The route notably demonstrated an improvement in yield from 28% to 70% when utilizing flow technologies.

Atazanavir. Atazanavir 6 is a protease inhibitor which binds to the active protease HIV site, preventing it from cleaving the pro-form of viral proteins into the working machinery of the virus.38 Originally approved for medical use in 2003, atazanavir 6 is now listed on the WHO list of essential medicines. The original Bristol-Myers Squibb Company patent described the synthesis of atazanavir from hydrazinecarboxylate 36 through a condensation with epoxide 37 to afford the protected triamine 38 (Scheme 8).39 Subsequent deprotection under acidic conditions afforded triamine 39, which was then subjected to an N-(3-dimethylaminopropyl)-N1-ethylcarbodiimide hydrochloride/1-hydroxybenzotriazole hydride mediated amide coupling with 40 to afford atazanavir 6.
image file: c8re00236c-s8.tif
Scheme 8 Bristol-Myers Squibb synthesis of atazanavir 6.

The Kappe group has successfully developed flow-based syntheses for both the hydrazinecarboxylate 36 precursor and a synthetic equivalent of epoxide 37, compound 41 (Scheme 9).40,41 In this first instance 36 was prepared in 3 steps in an overall yield of 74% by employing Suzuki–Miyaura cross-coupling chemistry.40 A solution of 2-bromo-pyridine 42, 4-formyl-phenylboronic acid 43 and tetrakis was mixed with potassium phosphate tribasic using a T-piece mixer; thereafter maturation in a coil reactor at 150 °C (20 min residence time) afforded the crude coupled biaryl system 44. The crude material underwent a manual partitioning and phase separation. The Suzuki–Miyaura product 44 was then mixed with trimethylsilyl triflate and tert-butylcarbazole successively using sequential T-piece mixers prior to heating in a coil reactor at 50 °C (residence time 8 min) affording hydrozone 45. Thereafter a liquid–liquid extraction with aqueous potassium carbonate using a FLLEX module removed any remaining acid. Finally, the mixture was subjected to a flow hydrogenation using an H-Cube reactor fitted with a 10% Pd/C cartridge. The desired hydrazinecarboxylate 36 precursor was obtained in 100% yield.


image file: c8re00236c-s9.tif
Scheme 9 Kappe group flow-based routes to advanced intermediates for the synthesis of atazanavir 6.

In the second instance, 41 was prepared by reacting N-Cbz-phenylalanine 46 and tributylamine with ethyl chloroformate 47 in a T-piece mixer followed by maturation in a coil reactor (7 min residence time) preparing anhydride 48.41 The anhydride 48 was then reacted with anhydrous diazomethane (generated by reaction of diazald and potassium hydroxide) in a tube-in-tube reactor (5 min residence time) followed by a tube reactor (27 min residence time), both at ambient temperature. The resulting α-diazo ketone 49 was finally reacted with ethereal hydrochloric acid in a tube reactor (13 min residence time) at 0 °C affording 41 in 87% yield with an overall residence time of only 52 min and complete retention of stereochemistry.

Abacavir. Abacavir 3 is a nucleoside analogue reverse transcriptase inhibitor (NRTI) originally approved by the FDA in 1998. Two predominant synthetic strategies have been developed; in the first case, investigators have focused on the construction of the cyclopentene ring stereoselectively on highly substituted pyrimidines before cyclisation to afford the desired nucleobase.27 In the second case, the cyclopentene ring is introduced stereoselectively directly onto the nucleobase. In almost all instances the cyclopropyl ring is introduced in the final step through an ipso-substitution of chlorine on the nucleobase.27

To the best of our knowledge, a flow-based synthesis has not been reported to date in the literature; however, regionally the iThemba Pharmaceuticals group from South Africa developed an enantioselective synthesis of abacavir utilizing an enzymatic resolution approach (Scheme 10).42 Cyclopentadiene 50 and chlorosulfonyl isocyanate were reacted and subsequently tosylated to afford 51. Thereafter a palladium mediated coupling with 2,6-dichloropurine 52 afforded 53, which was subsequently methylated by treatment with methyl iodide to afford 54. Interestingly, the use of 2,6-dichloropurine 52 as a source of the nucleobase was atypical with all other routes reported using similar approaches using 2-amino-6-chloropurine with the required amine functionality at the 2-position already in place. As a result, the final stages of the synthesis required a regioselective addition of cyclopropyl amine 17 at the 6-position followed by amination at the 2-position. In the latter case this was conveniently realized using 4-methoxybenzylamine as an alternative to ammonia. The process notably avoided several toxic and hazardous reagents like sodium azide and tin(II) chloride commonly seen in other approaches.


image file: c8re00236c-s10.tif
Scheme 10 (i) Chlorosulfonyl isocyanate, lipase B, (ii) tosyl chloride, n-BuLi, THF, 81%, (iii) Pd(dba)3, THF, (iv) MeI, K2CO3, acetone, 74%, (v) (a) NaBH4, IPA/THF, 86%, (b) 17 ethanol, 50 °C, 73%, (vi) 4-methoxybenzylamine, DMSO, 190 °C, 90%, (vii) TFA, 73% Δ.
Tenofovir disoproxil fumarate. Tenofovir disoproxil fumarate (7, TDF) and structural analogue tenofovir alafenamide fumarate (TAF) are pro-drugs of the NRTI tenofovir 55 developed for the treatment of HIV/AIDS and hepatitis B.43 Despite its status as a front-line HIV treatment, an efficient high-yielding process route to TDF 7 has to date not been realized. Gilead originally patented a process in 1998 which afforded TDF in 13% overall yield;44 in 2010 this was improved to 24% by the Clinton Health Access Initiative (CHAI) (Scheme 11).45 In the CHAI approach adenine 56 is condensed with R-propylene carbonate 57 affording (R)-9-[2-(hydroxyl)propyl]adenine (HPA) 58 in 66% yield. Thereafter, alkylation mediated by magnesium tert-butoxide (MTB) with tosylated hydroxymethylphosphonate diester 59 affords diethylester (PPA) 60 in 90% conversion (stage 2a). Crude PPA 60 is then hydrolysed by treatment with TMSCl/NaBr affording the active drug tenofovir 55 in an isolated yield of only 59% (stage 2b). The pro-drug form is then obtained by alkylative esterification using chloromethyl isopropyl carbonate 61 in the presence of triethylamine and tetrabutylammonium bromide followed by treatment with fumaric acid, affording TDF 7 in 62% yield (stage 3). The last stage represents a significant improvement over the analogous stage in the Gilead route (35%).
image file: c8re00236c-s11.tif
Scheme 11 CHAI approach: (i) 53 (1.32 eq.), NaOH (cat), DMF, 120 °C, 22 h, 66%, (ii) 55 (1.5 eq.) Mg t-OBu (3 eq.), NMP, 70–74 °C, 7 h, (iii) TMSCl/NaBr, NMP, 75 °C, 16 h (59% 2 steps), (iv) 57 (5 eq.), NEt3 (4 eq.), TBABr (1 eq.), NMP, cyclohexane, 50 °C, 5.5 h, 69%. iThemba approach: (i) 53 (1.32 eq.), NaOH (cat), DMF, 120 °C, 2.5 h, 65%, (ii) 55 (2.5 eq.), MeMgCl (1 eq.), t-BuOH (1 eq.), cyclohexane (min), 75 °C, 4.5 h, 85%, (iii) 33% HBr/AcOH, 75 °C, 3 h, 67%, (iv) 57 (5 eq.), NEt3 (4 eq.), TBABr (1 eq.), NMP, cyclohexane, 50 °C, 5.5 h, 69%.

To date process development has been largely hampered by the choice of base and solvent in stage 2a, and although the use of MTB allows high conversions, it is expensive, its performance is dependent on the quality of the MTB, and the magnesium salt by-products form a sticky hydroscopic salt cake which is difficult to process, necessitating the need to telescope the material directly into stage 2b.43 The hydroscopic nature of the salts further complicates the use of moisture-sensitive hydrolysis reagents like TMSBr or TMSCl/NaBr commonly employed in the process.43 Despite extensive investigation the identification of a more suitable base has proven challenging with almost all alternatives affording moderate yields at best.43

In 2016, Riley and co-workers reported an efficient process route to TDF which largely overcomes the difficulties associated with the use of MTB in the second stage (Scheme 11).43 Starting from adenine 56, a simple exchange of the catalyst from sodium hydroxide to potassium hydroxide greatly improved the rate of reaction, affording pure HPA 58 in 65% yield (2.5 h vs. 22 h). In attempts to identify a replacement for MTB it was noted that Grignard reagents afforded conversion albeit poorly (<28% PhMgCl); this conversion however increased dramatically to 76% on the addition of tert-butanol. The stage was ultimately optimized to afford pure PPA 56 in 85% yield (94–98% purity); notably the approach was also conducted under close to solvent-free conditions and an extractive work-up/purification was developed which allowed the isolation of pure PPA 60. In doing so, the challenges associated with the cost and use of MTB were negated. PPA 60 was then hydrolysed in a separate step utilizing HBr/acetic acid to afford tenofovir 55 in 73% yield. When coupled with the improvements developed for stage 3 as reported in the CHAI route the process afforded TDF 7 in a comparably overall yield of 23% vs. 24%. Furthermore, the use of cheap reagents made the iThemba Pharmaceuticals route more economical than both the Gilead and the CHAI routes.

3.2 Tuberculosis

Tuberculosis (TB) is the ninth leading cause of death worldwide, with 25% of all TB deaths (2.5 million in 2016) occurring within the African region.46 TB additionally adds to the HIV/AIDS burden in Africa in that it is the leading killer of HIV-positive people, accounting for 40% of all HIV-related deaths.46 Furthermore, the advent of multidrug-resistant TB (MDR-TB) continues to grow as a public health crisis and health security threat with ∼18% of all African TB cases displaying resistance to rifampicin, the most effective first-line drug.46,47

Currently there are five first-line TB drugs, isoniazid 62, pyrazinamide 63, ethambutol 64, streptomycin 65 and rifampicin 66 (Fig. 4), and a further 21 drugs used in combination treatments.47 To date, the development of process routes to TB drugs utilizing modern manufacturing approaches is limited and in the context of this review we have chosen to highlight process routes pyrazinamide 63, ciprofloxacin 67 and linezolid 68.


image file: c8re00236c-f4.tif
Fig. 4 Structures of a selection of anti-TB drugs.
Ciprofloxacin. Ciprofloxacin 67 is on the WHO list of essential medicines and is an anti-bacterial that is used as a second-line TB treatment for drug-resistant variants of the disease.47 First patented in 1983 by Bayer (Scheme 12), the process involves the reaction of 2,4-dichloro-5-fluorobenzoyl chloride 69 with a malonic ester followed by β-decarboxylation and Dieckmann-type condensation with ethyl orthoformate to afford 70.48 Thereafter, treatment with cyclopropylamine 17 and base-mediated cyclisation afforded 71. Finally, hydrolysis and SNAr displacement with piperazine afforded ciprofloxacin 67 in 49% yield with an overall process time of >24 h.
image file: c8re00236c-s12.tif
Scheme 12 (i) CH2(CO2Et)2, MgOEt, 50–60 °C, 45 min, (ii) cat. p-TolSO3H, H2O, 82% (2 steps), (iii) HC(OEt)3, Ac2O, Δ, 2 h, (iv) EtOH, 1 h, 92% (2 steps), (v) K2CO3, DMF, 140–145 °C, 2 h, 90%, (vi) cat. conc. H2SO4, HOAc/H2O, 1.5 h, Δ, 93%, (vii) DMSO, 140 °C, 2 h, 78%.

This was further improved by the MacDonald group, who in 1996 followed an analogous approach (Scheme 13) utilizing solid-phase organic synthesis by employing a tetrabenzo[a,c,g,i]fluorene Wang resin.49–51 Interestingly, the chosen resin has a high affinity for charcoal as a support, allowing the use of solution phase chemistry with purification facilitated through absorption/desorption by simply switching between polar and non-polar solvents. To its advantage, the approach afforded overall improvement in the yield to 57% and greatly simplified heterogeneous purification; however, this required a process time of ∼100 h.


image file: c8re00236c-s13.tif
Scheme 13 (i) (a) K+CH2(CO2Et)(CO2), MgCl2, NEt3, CH3CN, rt, 2.5 h, (b) rt, 16 h, 94%, (ii) DMAP, toluene, Δ, 40 h 70%, (iii) Me2NCH(OMe)2, THF, rt, 24 h, (iv) THF, rt, 20 h, (v) TMG, THF, Δ, 20 h, 67% (3 steps), (vi) pyridine, Δ, 6 h, 76%, (vii) 90% TFA/H2O, CH2Cl2, 57%.

In 2017 Jensen, Jamison and co-workers tackled the process and reported on an elegant approach wherein ciprofloxacin 67 was accessed under continuous flow conditions (Scheme 14).52 The process involved the telescoping of six reactions in five reactor modules without the need for the separation or isolation of intermediates. Furthermore, the process represented the longest linear telescoped flow process wherein the reaction stream is not held at any point.52


image file: c8re00236c-s14.tif
Scheme 14 MIT flow synthesis of ciprofloxacin hydrochloride 80.

Following a similar approach to that of the MacDonald group, the Jensen group elected to acylate vinylogous carbamate 72 with acyl chloride 73, affording 74 (reactor 1); in doing so two steps that are present in the batch routes were removed. Subsequent treatment with cyclopropyl amine 17 afforded 75 (reactor 2); however, competitive fluorine displacement with the associated dimethylamine by-product necessitated the addition of a quenching stream of acetyl chloride 76 (reactor 3), converting the dimethylamine to N,N-dimethylacetamide (DMA) which proved innocuous in downstream steps. An SNAr displacement of fluorine with piperazine 77 in the presence of 1,8-diazabicyclo(5.4.0)-undec-7-ene followed (reactor 4), affording 78. The solubility of 78 was found to be low, resulting in precipitation; fortunately, nucleation of crystals took between 1 and 2 min on cooling and as such simple insulation of the reagent lines and connectors between reactors 4 and 5 was sufficient to prevent reactor fouling. The treatment with sodium hydroxide in reactor 5 afforded the sodium salt of ciprofloxacin 79 which when acidified post-reactor afforded ciprofloxacin hydrochloride 80 in an overall yield of 60%. Notably, although the overall yield of the process was comparable to that of the batch processes, the flow alternative had a residence time of only 9 min with no manual work-up or purification steps prior to the final acidification, filtration and washing.

Linezolid. Linezolid 68 is used for the treatment of infections caused by Gram-positive bacteria that exhibit resistance to other antibiotics, and although indicated for the treatment of pneumonia and skin/soft tissue infections, linezolid is also used as part of second-line combination therapies for the treatment of TB.53 Several syntheses are reported, most of which are simple variations of each other. The drug was originally developed by Brickner and co-workers at Pharmacia & Upjohn Laboratories (now part of Pfizer) who were able to demonstrate an eight-step synthesis on a 100 kg scale (Scheme 15).53 3,4-Difluoronitrobenzene 81 was treated with morpholine 82 in the presence of diisopropylamine (DIPA), affording a regioselective SNAr displacement of fluorine; thereafter, reduction of the nitro group and subsequent CBz addition afforded 83. A Manninen cyclisation with (S)-3-chloro-1,2-propanediol 84 gave intermediate 85; thereafter, the alcohol group was converted to an amine by nosylation, followed by treatment with excess ammonium hydroxide. Finally, the free amine was acetylated affording linezolid 68 in an impressive 65% overall yield.
image file: c8re00236c-s15.tif
Scheme 15 (i) DIPA, EtOAc, Δ, 97%, (ii) HCO2NH4, Pd/C, THF/MeOH, 96%, (iii) CBzCl, NaHCO3, aq. acetone, 99%, (iv) Li t-BuO, 84%, (v) NsCl, NEt3, DCM, (vi) NH4OH excess, (vii) Ac2O, Py.

A complete flow-based synthesis of linezolid 68 has, to the best of our knowledge, not yet been reported; however, the Kappe team in their development of a hydrazine-mediated reduction of nitro functionalities using colloidal iron nanoparticles demonstrated an alternative approach for the nitro reduction step (Scheme 16).54 In the process the aryl nitro 86 in methanol is combined with hydrazine hydrate (1.7 eq.) in the presence of catalytic colloidal Fe(acac)3 nanocrystals (1.5 mol%) prior to passage through a stainless-steel coil reactor (1.5 min residence time, 10 mL min−1) heated to 170 °C. The reaction stream was then passed through a water-cooled heat exchanger prior to exiting the reactor through a 34-bar back pressure regulator, affording the pure reduced linezolid precursor 87 in 99% conversion (95% isolated yield) corresponding to a production rate of 30 g h−1. The iron nanocrystals were collected magnetically from the output stream and resuspended by sonication to regenerate the highly active colloidal catalyst.


image file: c8re00236c-s16.tif
Scheme 16 Kappe flow synthesis of the intermediate used in the synthesis of linezolid 68.
Pyrazinamide. Pyrazinamide 63 is an antitubercular agent that is used in combination with other TB drugs including isoniazid 62 and rifampicin 66; the drug is typically used during the initial two months of treatment to reduce the duration of treatment required.46 The Ley group has on several occasions demonstrated the synthesis of pyrazinamide 63 through a simple single-step hydration of pyrazinecarbonitrile 88.55–57 The process is notable in that although nitrile hydrations are well known, atom-economical and industrially relevant, the transformation is today still in many cases challenging to achieve.

In 2014 the Ley group demonstrated the synthesis of pyrazinamide 63 by the hydration of pyrazinecarbonitrile 88 in a PBR housing manganese dioxide (Scheme 17).55 A solution of 0.5 M pyrazinecarbonitrile 88 in a 10[thin space (1/6-em)]:[thin space (1/6-em)]1 ratio of H2O[thin space (1/6-em)]:[thin space (1/6-em)]isopropyl alcohol was passed through a PBR housing 2.5 g manganese dioxide and heated to 98 °C (flow rate of 1.5 mL min−1), affording pyrazinamide 63 in a 99% isolated yield. The reaction was scaled to 400 mmol utilizing a single 2.5 g MnO2 PBR affording 396 mmol pyrazinamide, equating to a production rate of 2.21 g h−1 g−1 of catalyst. The process further showed only negligible leaching of Mn over several hours with a residual concentration of only 0.06 ppm detected by ICP-MS. In the same year the group also showed a machine-assisted multistep flow synthesis of pyrazinamide 63 using a PBR housing anhydrous zirconia (discussed in detail in section 4.2).56


image file: c8re00236c-s17.tif
Scheme 17 Ley flow-synthesis of pyrazinamide 63.

Finally, in 2018 the group revisited the manganese dioxide approach, this time demonstrating the synthesis with a telescoped plug flow crystallization (Scheme 18).57 The approach utilized a novel tubular flow crystallisation system employing a tri-segmented flow in which the reaction stream is segmented with air and an immiscible carrier fluid. In practice, a 0.28 M solution of pyrazinecarbonitrile 88 in water was reacted by passage at 1 ml min−1 across a PBR housing MnO2 heated to 80 °C. Thereafter, the reagent stream was passed through a back-pressure regulator prior to segmentation in a heated (70 °C) cross-piece mixer. The segmented flow was then passed through a tri-segmented tubular crystallizer (KRAIC) in the form of three coil reactors (1/8′′ ID FEP tubing, 15 m) prior to the separation of the carrier fluid and online filtration. It was found that the process as described suffered from reactor blockages; however, this was remedied through the insertion of a section of jacketed tubing between the segmentation module and the KRAIC module cooled to 10 °C. The more rapid cooling resulted in the formation of a greater number of nuclei thereby reducing the crystallite size and preventing blocking. The process afforded the isolation of pure pyrazinamide 63 in 53% yield with a residence time of 32 min 17 s. Of additional interest, was that pyrazinamide 63 was obtained exclusively as the γ-polymorph which previously was only observed with non-aqueous solvents/solvent mixtures.


image file: c8re00236c-s18.tif
Scheme 18 Ley flow-synthesis of pyrazinamide 63 incorporating tubular flow crystallisation.

3.3 Malaria

The number of global reported malaria cases exceeded 214 million in 2015 with 90% of all deaths occurring in Africa, placing a great burden on the continent.58 Chloroquine 89, discovered in 1934, was until recently still the most widely used antimalarial; however, the emergence of drug resistance has rapidly decreased its effectiveness. That being said, chloroquine 89 is still the first-line antimalarial in sub-Saharan Africa. The rapidly observed drug resistance has led to the continued development of several different antimalarials over the course of the last century including amodiaquine 90, pyrimethamine 91, proguanil 92, mefloquine 93, atovaquone 94, primaquine 95 and artemisinin 96 (Fig. 5). The batch and flow process routes to several antimalarials will be discussed in this review; however, the reader is further referred to the review by Browne and co-workers which showcases the latest advances in the flow production of antimalarials and illustrates how machine-assisted distributed chemical manufacturing can be used for on-demand production of antimalarials.59
image file: c8re00236c-f5.tif
Fig. 5 Structures of a selection of antimalarial drugs.
Artemisinin. As concerns regarding antimalarial drug resistance increase, artemisinin-based combination therapies (ACTs) have gained momentum as a first-line defense for uncomplicated malaria.60 Artemisinin 96 is extracted from the sweet wormwood plant (Artemisia annua) which is cultivated in many countries for this purpose.61 The complexity of the molecule excludes an economically viable total synthesis of the API at this stage, despite several reports in the literature.62–66 The cost drivers of artemisinin 96 therapy are fueled by unstable supply and harvest yield and shortages have resulted in a surge of substandard, spurious, falsified, falsely labeled, and counterfeit (SSFFC) medicines. Alternatives to isolating the API include the semi-synthesis of artemisinin from artemisinic acid 97 (AA) or dihydroartemisinic acid 98 (DHAA) which can be obtained by extraction from the plant, or in the case of 97 by an engineered strain of yeast, Saccharomyces cerevisiae (discussed in more detail in section 4.1). Several practical semi-synthetic routes have been developed;67,68 however, these are still challenging to scale-up, most notably as the large-scale generation of singlet oxygen in traditional batch reactors is problematic.

A typical semi-synthesis (Scheme 19) can be realized wherein DHAA 98 is derived by reduction of AA 97, which reacts with singlet oxygen to yield the tertiary allylic hydroperoxide 99. A Hock cleavage follows with subsequent allyl migration to yield an enol 100. Reaction with triplet oxygen yields a hydroperoxide 101 that leads to the target artemisinin 96 after a series of condensations.


image file: c8re00236c-s19.tif
Scheme 19 Semi-synthesis of artemisinin from plant- or yeast-derived artemisinic acid.

The development and implementation of a photochemical batch reactor to produce singlet oxygen is severely limited by the fall-off in light penetration as the distance from the light source increases. As a result, Seeberger and co-workers turned to flow chemistry (Scheme 20).69 The group was able to overcome the challenges associated with the production of singlet oxygen by using transparent tubing wrapped around a glass plate, mounted in close proximity to a light-emitting diode (LED). The use of a narrow transparent tubing meant that the drop-off in light intensity with distance was negligible, and the use of a “cool” LED light source prevented overheating issues typically associated with batch photochemical reactors. The process at hand performs better at lower temperatures and to further reduce the temperature the tubing was submerged in a cooling bath, providing a process solution that is not easily achieved under traditional batch conditions.


image file: c8re00236c-s20.tif
Scheme 20 Seeberger flow synthesis of artemisinin.

The optimised continuous process (Scheme 20) involved a feed of DHAA 98, a photosensitizer (dicyanoanthracene, DCA) and trifluoroacetic acid (TFA) as acid catalyst in toluene, mixing with 2 equivalents of oxygen (adjusted by a flow controller) in a T-mixer. The mixture is cooled to −20 °C to allow for optimal photooxidation. The temperature of the exiting mixture is adjusted to room temperature and complete conversion was achieved by simply adjusting the length of tubing used. A BPR was employed at the end of the flow line to keep the system pressurized and improve oxygen solubility. The reaction output underwent a simple aqueous extraction and the resultant crude product was purified by recrystallisation. A yield of 65% was obtained and translated to an impressive 165 g artemisinin 96 per day, considering the total volume of the reactor (47.5 mL) a space–time yield of 3500 kg m−3 per day was realised.69 This process has been evaluated techno-economically, and clear advantages to continuous pharmaceutical manufacturing (CPM) have been presented by Gerogiorgis.70 The comparison between the batch production and continuous production showed a capital expenditure (CapEx) and operating expenditure (OpEx) saving of up to 19.6% and 29.3%, respectively, with a 20 year plant lifetime cost saving of up to 20.1%. In addition, the environmental factors computed illustrate significant advantages in terms of waste reductions and material efficiencies.70

The process developed further allows the facile production flexibility to continuously synthesize four artemisinin derivatives, including artesunate 102, artemether 103, artemotil 104 and dihydroartemisinin 105 (Scheme 21). To achieve this, the system is expanded to include three additional modules after the photooxidation; a reduction, derivatization and purification module.


image file: c8re00236c-s21.tif
Scheme 21 Seeberger four-module system to produce artemisinin derivatives.
OZ439. Increasing pressure due to growing drug resistance has created a lot of interest in drug discovery programs both within Africa and internationally.71–73 A series of recently developed novel antimalarials includes the hit compound OZ277 and lead compound OZ439 106, currently in phase IIa clinical trials.74–76 The preparation of these compounds has several drawbacks, including the use of expensive precursors or genotoxic materials.

The Ley group has developed a flow-batch hybrid process yielding a robust and efficient route to OZ439 106 from [1,1′-biphenyl]-4,4′-diol 107 (Scheme 22).77 The approach involves three flow-based stages (a selective partial hydrogenation, acetylation of a phenol group and a Griesbaum co-ozonolysis). The first stage involved the selective reduction of one of the phenol groups in 107 through the use of a HEL FlowCAT reactor module housing a trickle bed containing Pd/C (20 mol%) heated to 100 °C with a H2 (g) flow rate of 100 mL min−1. A 3-minute residence time followed by off-line purification by recrystallisation afforded 4-(4-hydroxyphenyl) cyclohexanone 108 in 58% yield with a capacity to produce 0.4 g h−1.


image file: c8re00236c-s22.tif
Scheme 22 Ley batch-flow hybrid synthesis of OZ439. Batch stages: (i) NaOH, HSO4NBu4, CH3CN, 30 min, rt, (ii) 4-(chloroacetyl)morpholine, 12 h, 60 °C, (iii) Zn(OAc)2 (10 mol%), (EtO)3SiH, THF/CH2Cl2 (1[thin space (1/6-em)]:[thin space (1/6-em)]1), 22 h, 60 °C.

In the second stage, a straightforward acetylation of the phenolic proton was achieved in near-quantitative yield by exposing a stream of 108, triethylamine and DMAP to acetic anhydride in a coil reactor. Thereafter in the third stage, the acetylated 109 was subjected to Griesbaum co-ozonolysis reaction by treatment with oxime and a stream of generated ozone in a 2 mL coil reactor. The reaction mixture was then passed into a vessel and degassed with argon affording the desired 1,2,4-trioxolane moiety 110 in a yield of 70%, which translates to a production of 1.9 g h−1 (45.6 g d−1). Traditionally, the removal of the acetyl group and generation of a phenoxide anion that is reacted with 4-(2-chloroethyl)morpholine (a known genotoxic agent) afforded OZ439 106 directly. Ley and co-workers instead opted to react the phenoxide anion with 2-chloro-1-morpholinoethanone (a safer amide intermediate) to afford 111 in a yield of >99%. Finally, reduction of amide 111 afforded OZ439 106 in 86% yield.77

Hydroxychloroquine. Despite drug administration policy changes between the 1940s and 2000s due to the emergence of chloroquine-resistant strains of the malaria parasite, the use of hydroxychloroquine (HCQ) 112 and chloroquine 89 remains important on the African continent. Recently, a process for the synthesis of HCQ utilizing flow technology has been reported that affords a significant improvement in yield over the current commercial batch process.78

The commercial synthesis of HCQ's 112 largest cost contributor is the synthesis of the key intermediate 5-(ethyl(2-hydroxyethyl)amino)-pentan-2-one 113 (Scheme 23). Starting from 4-chloro-2-pentanone 114, the ketone is protected by treatment with 1,2-ethanediol 115 under acidic conditions; thereafter an SN2 displacement of the chlorine with 2-(ethylamino)ethanol 116 followed by deprotection of the ketone under acidic conditions afforded 113. The protection/deprotection step has been a source of inefficiency and although alternative routes utilising multi-transition-metal-catalyst systems to achieve a direct substitution have been reported, these are still not efficient enough.79


image file: c8re00236c-s23.tif
Scheme 23 (i) HO(CH2)2OH, cyclohexane, 80 °C, 90%, (ii) HO(CH2)2NHEt, toluene, 110 °C, 83%, (iii) HCl/H2O, 40 °C.

In the continuous process reported by Gupton and co-workers (Scheme 24), α-acetyl butyrolactone 117 is reacted neat with 55% hydroiodic acid in a 10 mL tube reactor heated to 80 °C. A 5-minute residence time afforded 4-iodo-2-pentanone 118 in an isolated yield of 89%. Excess hydroiodic acid was removed in-line by introducing a stream of methyl tert-butyl ether in hexane and saturated sodium bicarbonate followed by a membrane-based extraction using a Zaiput separator. The resultant 4-iodo-2-pentanone 118 is reacted with 2-(ethylamino)ethanol 116 in a 10 mL tube reactor heated to 80 °C followed by a PBR housing potassium carbonate heated to 100 °C, affording 119. The reaction stream is then treated with ammonium hydroxide in THF and passed through a 10 mL tube reactor followed by a PBR housing potassium carbonate, both heated to 100 °C, affording oxime 120. Thereafter, a RANEY®-Ni hydrogenation in a CSTR afforded 121 in an overall telescoped yield of 68%. The authors note that 121 could have been produced directly from 119 by reductive amination but opted to avoid an in-line solvent swap from THF to the required protic solvent. The output of the telescoped process can then be fed into a second CSTR reactor affording the coupling of 121 with dichloropurine 122 to produce HCQ 112 in 78% yield.


image file: c8re00236c-s24.tif
Scheme 24 Gupton flow synthesis of hydroxychloroquine 112.

3.4 Antibiotics and antifungals

Infectious diseases including the neglected tropical diseases remain the greatest disease burden in the low-income African economies.80 Antibiotic resistance has been steadily increasing over the last few decades and there is a continual need for the development of new antimicrobial agents. In particular, the threat of Gram-negative bacteria that are resistant to multiple antibiotics has been highlighted by the WHO. There are four major classes of antifungal agents available to clinicians today, and with careful use and management of underlying diseases, success in treatment of invasive fungal diseases is possible, but concerns are mounting for the need for new antifungal agents. Amphotericin B is still the broadest anti-fungal available; the azoles have matured from imidazole to the safer and broader spectrum triazoles. However, the azoles generally have complicated drug–drug interactions and possess many side effects. The third class of antifungal agents is the newest and includes the echinocandins (caspofungin, micafungin, and anidulafungin), with minimal drug–drug interactions. Finally, flucytosine is a “niche” drug with limited uses and distribution worldwide.81
Antifungal – flucytosine. Flucytosine 123 is an antifungal used broadly for the treatment of severe Candida infections and cryptococcosis. In terms of the latter. cryptococcal meningitis (CM) is responsible for 20% of all HIV/AIDS-related mortalities worldwide82,83 and is the leading cause of meningitis in sub-Saharan Africa, accounting for 625[thin space (1/6-em)]000 mortalities.84,85 CM is a particularly heavy burden for the African continent, where 70% of patients diagnosed with CM in sub-Saharan Africa die, compared to 9% in the developed nations.82 A concern is that although the WHO recommends the use of flucytosine 123 in combination with amphotericin B as first-line defense against CM,86 flucytosine 123 is not registered for use in many African countries.82 Flucytosine 123 notably has a high manufacturing cost which reduces the availability of the API on the continent.

The current manufacturing route requires a 4-step transformation from 5-fluorouracil that involves a chlorination, amination and hydrolysis (Scheme 25).87–89 The first stage of the batch route is challenging and low yielding. The process involves dissolving cytosine 124 in formic acid and passing fluorine gas (10% in nitrogen) through the solution at room temperature. Excess fluorine gas was required to afford full conversion due to purification challenges. The batch process is further plagued by poor selectivity with around 56% of the material produced consisting of undesired difluorinated and related side products. Subsequent stages are moderate to high yielding, affording flucytosine 123 in 38% overall yield in 160 min.


image file: c8re00236c-s25.tif
Scheme 25 (i) F2(g), (ii) POCl3, 85%, (iii) NH3, 79%, (iv) H+, H2O, 96%.

Researchers from Sanofi Aventis and Durham University proposed a one-step continuous flow synthesis of flucytosine 123 (Scheme 26) and have illustrated the advantages of flow processes over traditional batch production with higher yields, reduced reaction time and easy translation from batch to flow to pilot scale with a microreactor setup. In the process, fluorine and nitrogen gas are mixed with a stream of cytosine dissolved in formic acid (introduced by a syringe pump) in a T-mixer prior to passage through a 1 m stainless-steel coil reactor at ambient temperature followed by a gas scrubber. The process afforded flucytosine 123 in 66% yield in 90 min. An analogous scaled pilot process was developed; the process involved swapping the tubular reactor for a series of 6 process (mixing) plates providing a 16 m channel length with a 61 mL reactor volume. The process afforded flucytosine 123 in 83% yield (99.9% purity by HPLC) and was able to produce 58 g of material in only 60 minutes.


image file: c8re00236c-s26.tif
Scheme 26 Lab and pilot plant scale flow synthesis of flucytosine 123.
Antibiotic – cefotaxime. The antibiotic class with the highest production volume is that of the β-lactam antibiotics, which include the penicillin and the cephalosporin groups. In most cases, the active antibacterial substances are produced by semi-synthetic methods from a feedstock of 6-aminopenicillanic acid or 7-aminocephalosporanic acid 126 (7-ACA) for penicillin and cephalosporin C derivatives, respectively.90 A green process utilizing biocatalysts has been developed for production of 7-ACA 126, with cephalosporin C as starting material (derived from a fermentation process).91 Derivatisation at position 3′- and 7′ of cephalosporin C and 7-ACA 126 increases the antibacterial activity and yields the target APIs that include cefotaxime 127, a cephalosporin-based antibiotic that has recently been translated to a flow synthetic process.92

Current methods to prepare cefotaxime 127 suffer from low atom economy and excessive waste generation.93–95 In an attempt to reduce the waste generated, a new batch route was developed and optimized (Scheme 27). The approach involved the activation of the carboxylic acid group of (Z)-2-(2-amino-1H-imidazol-4-yl)-2-(methoxyimino)-acetic acid 128 by treatment with 4-toluenesulfonyl chloride 129 (chosen as an economical reagent that provides non-toxic toluenesulfonic acid as a by-product) in dimethylacetamide (DMAA), affording 130 in quantitative yield. Subsequent amide coupling of 130 and 7-ACA 126 in methanol (chosen as a green alternative to dichloromethane) affords cefotaxime 127 in 95% yield.


image file: c8re00236c-s27.tif
Scheme 27 (i) TsCl, NEt3, DMMA, −10 °C, 1.5 h, 100%, (ii) 126, NEt3, MeOH, −30 °C, 0.5 h, 95%.

A potential flow translation (Scheme 27) of the amide coupling was envisaged during the batch developmental stages and as such efforts were made to try to ensure homogeneity. Unfortunately, 130 is not completely soluble in DMAA and as such a solution was required which could handle particulate matter. Thus a meso-scale tubular reactor with channel diameters in the millimeter range was utilized to prevent blockages. In addition, peristaltic pumps were chosen based upon their superior ability to pump suspensions. As the starting reagents 126 and 130 were heat sensitive, the stocks were stored at 0 °C; in addition the stocks were continuously stirred to prevent sedimentation. The feedstocks were mixed in a polypropylene mixer (Y-shaped) and allowed to react in a 10 mL coil reactor submerged in a cooling bath (−10 °C). A short residence time of only 1 minute afforded cefotaxime 127 in a yield of 81%. Despite lower yields, the flow route afforded a lower reaction time (1 min vs. 30 min) and a higher reaction temperature/lower energy demand (−10 °C vs. −30 °C). In terms of space–time yield the flow approach showed a clear advantage with a yield of ∼3000 kg m3 h−1 (400 times greater than the batch process).

3.5 Diabetes

Africa has historically suffered under a burden of infectious diseases; however, with an increasing population age and a steadily rising middle class, NCDs are starting to contribute more to morbidity and mortality statistics. A key NCD that has seen rapid growth across Africa is type 2 diabetes mellitus.96–98
Vildagliptin. Vildagliptin 131 (Novartis Pharma) is an oral antihyperglycemic drug of the dipeptidyl peptidase-4 (DPP-4) inhibitor class (Scheme 28).99 The current batch synthesis of vildagliptin 131 involves three steps, L-proline amide 132 is condensed with chloroacetyl chloride 133 affording the N-acylated adduct 134. The Vilsmeier reagent 135 is then used to induce dehydration, yielding cyanopyrrolidine 136. Subsequent SN2 displacement of the chloride with hydroxy amino adamantane 137 in the presence of a base affords vildagliptin 131. The use of the Vilsmeier reagent 135 as the dehydration reagent comes with considerable health and safety considerations with respect to transportation, storage and handling. The Vilsmeier reagent 135 is highly reactive and moisture-sensitive and large-scale use is accompanied by rapid increases in temperature and pressure requiring care to prevent exposure and runaway reactions.100–105
image file: c8re00236c-s28.tif
Scheme 28 Batch synthesis of vildagliptin 131.

Continuous flow processes offer attractive means to safely use the Vilsmeier reagent 135 as the reagent can be generated and consumed in-line. Novartis have developed a continuous flow process that demonstrates the advantages of flow manufacturing when utilizing the Vilsmeier reagent 135.99 A batch-flow hybrid process was found to be the easiest and most viable route to access vildagliptin 131 from an operational point of view (Scheme 29). The reaction of L-proline amide 132 and chloroacetyl chloride 133 was performed under batch conditions. Thereafter the crude reaction mixture was pumped through a coil reactor heated to 22 °C and mixed with freshly generated Vilsmeier reagent 135. The generation of the Vilsmeier reagent 135 was achieved by reaction of neat POCl3 (oxalyl chloride was avoided due to violent reaction with gassing out of CO2 and CO observed initially) and DMF pumped through a perfluoroalkoxy (PFA) reactor coil (1.6 mm, 4.5 mL) after combining in a PFA T-piece. The reaction temperature was estimated at 40 °C by controlling the external temperature of the coils at 22 °C taking into consideration the calculated adiabatic temperature rise (ΔTad) of 16 °C [at 4.8 M], below the onset temperature of 67 °C. The authors noted that these conditions are acceptable due to the minimal reactor volume (4.5 mL coil), minimal residence time and the built-in software safety features of the instrument (such as automatic shutdown and automatic instrument cooling on temperature and pressure fluctuations beyond defined parameters) preventing the risk of runaway reactions and exposure.


image file: c8re00236c-s29.tif
Scheme 29 Novartis batch-flow hybrid synthesis of vildagliptin 131.

The streams of crude 134 and the Vilsmeier reagent 135 were mixed in a tubular PFA T-piece mixer (PFA was chosen for its resistance to corrosion) and passed through a PFA coil (50 mm, 50 mL) containing glass beads (2 mm) for efficient mixing, resulting in a residence time of 90 s at a flow rate of 35 mL min−1 (at 22 °C). The resulting stream flowed into a second batch reactor, where solvent reduction, crystallisation and filtration were used to purify 136 (with a space–time yield of 5.8 kg h−1 L−1) for the next batch reaction, where the solid was re-dissolved in DMF, combined with 137 and K2CO3 to provide vildagliptin 131 in a yield of 79%.

4. Emerging and disruptive technologies to enable the next production revolution

The next production revolution entails a confluence of technologies ranging from a variety of digital technologies (e.g. 3D printing, Internet of Things, advanced robotics) to new materials (e.g. nano-based) to new processes (e.g. data-driven production, artificial intelligence, synthetic biology).106 As these technologies have an impact on the production and the distribution of goods and services, they will have far-reaching consequences for productivity, skills, income distribution, well-being and the environment.

4.1 Bio-based conversions

Biocatalysis is rapidly evolving into a key technology for the discovery and production of chemicals, especially in the pharmaceutical industry, where high-yielding chemo-, regio-, and enantioselective reactions are critical. Biocatalysis has moved through three distinct phases over the past century. The first, in the 1920s, revolved around using naturally occurring biocatalysts to bring about transformations. In the 1980s, the second phase brought with it early protein engineering techniques that could expand the substrates of biocatalysis to non-natural compounds. However, their applications for the drug industry were largely underexplored until the recent technological breakthroughs in large-scale DNA sequencing, robust protein expression systems, metabolic engineering and directed evolution.107 These advances ushered in the third phase during this century, which optimised biocatalysis further, fully introducing the field into industry. We now stand on the edge of the fourth and most exciting phase, artificial intelligence (AI) assisted biocatalysis, where with the help of advanced robotics, the computational power of high-throughput screening can be increased to previously unthought-of levels.

There are two different types of biocatalysts currently in use, whole cell and purified protein biocatalysts. Whole cell catalysis uses the whole organism to conduct the transformation, while purified protein biocatalysis uses an extracted protein without the cell present.108 In whole cell biocatalysis, unless the protein is excreted from the organism, or displayed on the surface of the organism, the substrate must enter into the cell for it to be transformed, and the product must also exit the organism after the reaction. The advantages of whole cell biocatalysis include: (i) the cost is lower than using purified proteins, (ii) the reaction does not need to be supplemented with expensive co-factors, as they are already contained within the cell, and (iii) the reaction is performed in its native environment. The disadvantages of whole cell biocatalysis include: (i) the cell membrane limits penetration of the substrate and product making the reaction slower compared to purified protein, (ii) the cell can have undesirable metabolic pathways creating toxic by-products or unwanted chemicals, (iii) the process is more complicated, so research and development can take longer, and (iv) from a process standpoint, the reaction chamber is fouled from cellular debris and residual product, which can hinder downstream processing.108

Purified protein is typically obtained via bacterial expression but can also be sourced from yeast or mammalian cell lines. In this process, a host cell is transformed with a plasmid encoding the gene for the required protein. When the cells are grown, the protein of interest is expressed using the native cellular machinery. After protein expression, cells are broken down to release the protein into solution, and the protein is isolated by centrifugation and purified by chromatography or precipitation. Working with purified enzymes have a number of advantages, including (i) they are incredibly specific, therefore reducing by-product formation, (ii) the substrate is required to diffuse only into the active site of the protein and not through the cell membrane, and (iii) the concentration of the desired enzyme is higher compared to the same mass of whole cells. Purified proteins also have several disadvantages, including (i) the purification process can be expensive, (ii) sometimes over-expressed proteins are not correctly folded, (iii) proteins often require cellular environments or additional co-factors for high activity, and (iv) proteins can be unstable outside of the cell.108

Historically, the most popular enzymes used for chemical synthesis are lipases, esterases, proteases, acylases and amidases, among others. Recently, several recombinant biocatalysts have been discovered and isolated, significantly expanding the toolbox for biotransformations. Based on the reactions they catalyze, enzymes can be broadly classified into six major categories (Table 2).109 It has been estimated that about 60% of biotransformations currently rely on the use of hydrolases, followed by oxidoreductases at 20%.110 Recently, C–C bond forming and oxygenation enzymes have been employed to catalyze reactions with very high reaction efficiency and very low waste generation, underlining the potential of emerging enzymes. Enzymes can replace conventional catalysts in drug manufacturing and give significant gains in terms of reduced raw materials and other inputs by improving efficiency and allowing reactions to take place under milder conditions. In addition, biocatalytic steps often offer a powerful solution to the problem of controlling stereochemistry.

Table 2 Broad enzyme classifications109
Enzyme class Examples Reactions catalysed
Hydrolases Lipase, protease, esterase, nitrile, nitrile hydratase, glycosidase, phosphatase Hydrolysis reactions in water
Oxidoreductases Dehydrogenase, oxidase, oxygenase, peroxidase Oxidation or reduction
Transferases Transaminase, glycosyltransferase, transaldolase Transfer of a group between molecules
Lyases Decarboxylase, dehydratase, deoxyribose-phosphate aldolase Nonhydrolytic bond cleavage
Isomerases Racemase, mutase Intramolecular rearrangement
Ligases DNA ligase Bond formation requiring triphosphate


There are many potential sources of innovative enzymes, including organisms from extreme environments and genomic databases. In addition, enzyme developer companies are already investing in technology that can predict the ordering of amino acid protein sequencing. Additionally, with new predictive models to understand how enzymes are organised, a lab worker could consider a new enzyme's uses before it has even been created, which could vastly shorten the timescale of drug development, getting drugs to market far sooner. With the added benefit of enabling companies to create new proteins that hit higher performance targets than ever thought possible, this advancing technology is certain to drive biocatalysis in drug manufacturing for the foreseeable future. A selection of four examples linked to process routes covered in section 3.1 is given below to highlight the opportunities for using bio-based transformations in the preparation of APIs of relevance to the African continent.

Example 1: lamivudine (HIV/AIDS). In addition to the batch and flow-based syntheses discussed in section 3.1.1, lamivudine 1 has been the subject of several syntheses involving bio-based conversions. In most instances the chiral 1,3-oxathiolane is introduced by selective crystallization from a racemic mixture,111,112 or using enzymatic hydrolysis/acetylation of other stereoisomers.113–116 The enantioselective synthesis of lamivudine 1 was initially described by Liotta and co-workers in 1992 when a separation of the enantiomers was required for biological evaluation (Scheme 30).113 The process involved the coupling of acetylated 1,3-oxathiolane 138 with silylated cytosine affording racemic O-acetylated lamivudine 139. An enantioselective hydrolysis of 139 was then performed using pig liver esterase in 60% conversion and >99% ee.
image file: c8re00236c-s30.tif
Scheme 30 Synthesis of lamivudine 1 using biotransformations.

Several biosynthetic approaches have since been developed; notably in 1995 Cousins and co-workers described a selective reduction of racemic oxathiolane propionate 140 with a lipase from Mucor mienhei, affording the (−)-enantiomer 141 in 48% conversion but with only a moderate 70% ee (Scheme 30).115 In the same year Milton and co-workers, starting from racemic α-acetoxysulfide 142, used a lipase from Pseudomonas fluorescens, affording 143 as a single enantiomer in 49% yield but in an improved 95% ee (Scheme 30).114 In both instances the reliance on kinetic resolution necessitated the loss of >50% of the material.

Finally, in 2013 Ramström and co-workers reported an enzyme-catalyzed dynamic kinetic resolution (DKR) with reversible formation of the intermediate stereoisomers (Scheme 30).116 Surfactant-treated subtilisin (STS) Carlsberg was used to effect an enzymatic DKR between 1,4-dithiane-2,5-diol 9 and benzoyl protected aldehyde 144, affording oxathiolane 145 racemically (89% yield, 82% ee); thereafter two chemical transformations afforded lamivudine 1 in an overall yield of 40%.

Example 2: atazanavir (HIV/AIDS). A biocatalytic process was developed for the preparation of a key chiral intermediate required for the synthesis of atazanavir 6 (Scheme 31).117,118 The diastereoselective microbial reduction of key intermediate 146 to homochiral chlorohydrin (1S,2R)-[3-chloro-2-oxo-1-(phenylmethyl)propyl]carbamic acid, 1,1-dimethyl-ethyl ester (1S,2R)-147, has been achieved. Several microbial cultures were evaluated to afford the reduction with three strains of Rhodococcus giving >90% yield with a diastereomeric purity of >98% (99.4% ee). Rhodococcus erythropolis SC 13845 and Rhodococcus sp. 16002 were grown in a 25 L fermentor for 48 h after which time the cells were suspended in a 70 mM potassium phosphate buffer (pH 7.0). The resulting cell suspensions were used to carry out the diastereoselective reduction of 146 affording 147. An efficient single-stage fermentation–biotransformation process was developed for the reduction of ketone (S)-146 with cells of Rhodococcus erythropolis SC 13845. A reaction yield of 95%, diastereomeric purity of 98.2% and ee of 99.4% for (1S,2R)-147 were obtained.
image file: c8re00236c-s31.tif
Scheme 31 Synthesis of chiral intermediates used in the preparation of atazanavir 6.

The synthesis of atazanavir 6 also required the conversion of (S)-tert-leucine 149 (Scheme 31), via an enzymatic reductive amination using leucine dehydrogenase from Thermoactinomyces intermedius affording keto acid 149. The leucine dehydrogenase from T. intermedius was cloned and over-expressed in E. coli. The transformation required the addition of ammonia and NADH as a cofactor; the NAD produced during the reaction was then converted back to NADH with formate dehydrogenase from Pichia pastoris (also cloned and over-expressed in E. coli). The process afforded 149 in a yield of >95% (>99.5% ee).118

Example 3: abacavir (HIV/AIDS). The key (1R,4S)-azabicyclo[2.2.1]hept-5-en-3-one intermediate 150 used in the manufacture of abacavir can be accessed utilizing enantiocomplimentary microorganisms Rhodococcus equi NCIB 40213 and Pseudomonas solanacearum NCIB 40249.119 Mutant strains of P. solanacearum NCIB 40249 hyper-expressing gamma-lactamase afford a highly enantioselective kinetic resolution process with substrate concentrations of >100 g L−1 (Scheme 32, where R = H).119 The process was developed to run using whole cells as the enzyme was too unstable to isolate; however, difficulties with downstream processing prompted investigations that identified a new lactase that was stable enough to isolate and clone. The recombinant lactamase then afforded a highly efficient process using 500 g L−1 substrate concentrations with a simplified work-up and purification.
image file: c8re00236c-s32.tif
Scheme 32 Bio-based synthesis of abacavir 3.

Mahmoudian and co-workers,120,121 hypothesized that Boc-protection of the racemic 150 would activate the amide bond towards nucleophilic attack. The group screened numerous commercial hydrolases and identified savinase, which proved to be highly enantioselective towards hydrolysis of the undesired enantiomer, leaving the desired enantiomer in 50% conversion (>99% ee) (Scheme 32, where R = t-BuOC(O)O). Savinase is a serine-type protease and is produced by submerged fermentation of recombinant alkalophilic Bacillus lentus. It is cheaply available in bulk as it is an active ingredient used in the detergent industry.122

Readers are further referred to the review by Song, which comprehensively provides an overview of the use of bio-based transformations in the synthesis of anti-HIV drugs.123

Example 4: artemisinin (malaria). A semi-synthetic synthesis of artemisinin 96 required engineering to afford the production of the intermediate artemisinic acid 97 in yeast (Saccharomyces cerevisiae). The process relied on the discovery of Artemisia annua genes encoding for artemisinic acid 97 biosynthetic enzymes and the use of synthetic biology to engineer yeast metabolism.124

The first synthesis amenable to scale-up was developed by Amyris chemists (Scheme 33).125 As the photochemical production of singlet oxygen was challenging and as its implementation would add significant capital costs (section 3.1.3), the photogeneration of singlet oxygen was replaced by a chemical disproportionation of concentrated hydrogen peroxide using group VI metal salts.126,127 To improve the process safety profile, air was used instead of oxygen and a benzenesulfonic acid/Cu(II) Dowex resin was used as the catalyst, replacing the expensive copper triflate traditionally utilised. The 4-step approach gave the desired target in 40% overall yield.


image file: c8re00236c-s33.tif
Scheme 33 Amyris and Sanofi semi-syntheses of artemisinin 96. Amyris approach: (i) RhCl(PPh3)3 cat, toluene, H2(g), 25 bar, 99%, (ii) SOCl2, toluene, MeOH, NEt3, 85%, (iii) H2O2, Na2MoO4, 1,2-butanediol, (iv) DOWEX resin, Cu(OSO2Ph)2, air, CH2Cl2, 23% (2 steps) (19% over all steps). Sanofi approach: (i) RuCl2[(R)-dtbm-Segphos](DMF)2, MeOH, NEt3, 22 bar, 99%, (ii) EtOC(O)Cl, K2CO3, CH2Cl2, 20 °C, 100%, (iii) Hg vapour lamp, TPP/air-CH2Cl2, TFA, −10 °C, (iv) Hock cleavage cyclisation, NaHCO3, H2O (55% across all steps).

The technology was then transferred to Sanofi and extensive work was undertaken to develop an industrial process. A series of improvements were made; however, it soon became apparent that the approach had reached its performance limits and was not economical enough to warrant commercialisation. The Sanofi chemists ultimately returned to the original photochemical approach,128,129 and a resulting semi-batch tetraphenylporphyrin dye sensitized photochemical process in a custom-designed photochemical reactor was developed. The process is currently the only industrial-scale semi-synthesis of artemisinin producing up to 60 tons per year of the target molecule in ∼55% overall yield from artemisinic acid 97.128

As these selected examples show, bio-based conversions are emerging as one of the greenest technologies for chemical synthesis.130,131 Specifically, biocatalysis can prevent waste generation by using catalytic processes with high stereo- and regioselectivity, prevent or limit the use of hazardous organic reagents by using water as the green solvent, design processes with higher energy efficiency and safer chemistry by conducting reactions at room temperature under ambient atmosphere, and increase atom economy by avoiding extensive protection and deprotection steps. To truly realize the promise of this emerging technology for the pharmaceutical industry, it is essential to integrate enzymatic transformations with modern chemical research and development at the retrosynthetic level to deliver efficient and practical synthetic sequences with fewer synthetic steps and significantly reduced waste streams.132,133

Although several large-scale industrial biocatalytic processes are operated in continuous mode (in a series of stirred tanks or in packed beds) for good economic reasons, the vast majority are still operated in fed-batch mode, mainly because of the necessity of using multipurpose plants. Nonetheless, recent developments in continuous-flow chemistry are stimulating applications in biocatalytic processes.134 A particular challenge is how to run multiphasic reactions in flow mode (especially with gases). This is an active area of research in all areas of synthesis. Biocatalytic examples have focused mainly on oxidations, which can be carried out in tube-in-tube reactors, where the outer tube can supply oxygen (under pressure) to the biocatalyst in the inner tube.134 Further applications of continuous-flow biocatalysis are expected to be forthcoming, as this mode of operation results in smaller production units, significant productivity increases (eliminating downtime), and reduced inventories, affording more sustainable processes.134

For more in-depth information, an excellent review exploring continuous flow biocatalysts with emphasis on new technology, enzymes, whole cells, co-factor recycling, and immobilization methods for the synthesis of pharmaceuticals, value-added chemicals, and materials has recently been published by Britton and co-workers.108 In addition, detailed reviews have been prepared on the application of bioconversion approaches in continuous flow135 and the application of reaction engineering to biocatalysis.136

4.2 Industrial digital transformation

The utility of flow chemistry for the synthesis of small molecules and single-stage transformations has grown rapidly since the early 2000s; however, the development of integrated multi-stage syntheses of complex molecules has proved to be significantly more challenging. Several factors have contributed to these challenges, including the need to develop new approaches to downstream processing and purifications, the need to effect efficient in-line solvent swaps, the complex timing of sequential reagent additions into a dispersing product stream and the concurrent automation of multiple different instruments. As such, the development of a fully integrated and automated flow chemistry platform capable of overcoming these difficulties has started receiving significant interest over the course of the last ten years.

The integration of standard in-line analytical technologies such as NMR,137,138 MS,139 HPLC,137 Raman,140 UV/vis and IR141,142 has proven to be an extremely valuable tool to understand and monitor flow systems in real time and has played a critical role in the development of integrated multi-stage syntheses. Concurrently, there has been a growing development of flexible, low-cost, often bespoke in-line processing tools aimed at providing solutions to the complexities associated with in-line downstream processing and purifications.143 The integration of these and other technologies with traditional flow approaches and in-line automation is currently a driving force in the “synthesis 4.0” revolution.

The Ley group has been instrumental in developing solutions to many of these difficulties. The group notably published a report in 2011 which detailed a robust approach for the accurate sequential addition of reagents, wherein inline IR and UV monitoring was used to develop intelligent pumping methods.141 The approach involves the use of an inline monitor (IR or UV) to determine the concentration of a reagent stream in real time, the output of which is used to control the pumping units. The approach thereby allows accurate addition of downstream reagents based upon the inline concentration of the reagent stream regardless of dispersion (Scheme 34).


image file: c8re00236c-s34.tif
Scheme 34 Ley approach to intelligent pumping using in-line IR or UV detectors.

The group then applied a similar principle using a “computer vision” in the development of a continuous-flow liquid–liquid extractor.144 The approach involves the use of a Python™ computer control script and a webcam which monitors the interface between the organic and the aqueous phases in an extractor by tracking the relative up or down movement of a coloured float that sits at the interface (Scheme 35). The pump rate is then dynamically controlled to keep the float within a predefined range. The approach was later adopted to allow multiple liquid–liquid separations in flow.145 The reader is further referred to the review by the group highlighting the use of cameras as an enabling technology in organic synthesis.146


image file: c8re00236c-s35.tif
Scheme 35 Ley approach to automated continuous-flow liquid–liquid extractor using “computer vision”.

The realization of a fully integrated and automated flow synthesis requires not only advanced in-line monitoring, but also a means of interfacing between different instruments. The Ley group proposed the use of scripted control algorithms, wherein a specification for each class of device (e.g. pumps, in-line sensors, heaters) would afford a flexible means of integrating devices and building control system interfaces.56 The approach is demonstrated in the synthesis of pyrazinamide 63 and piperazine-2-carboxamide 151 which incorporates the use of cheap low power Raspberry Pi® computers and Python™ scripting.

The automated synthesis of pyrazinamide 63 (Scheme 36) involved an initial hydration of pyrazinecarbonitrile 88 by passage over a PBR housing hydrous zirconia followed by a BPR, affording quantitative conversions in aqueous ethanol when the PBR was heated to 100 °C (20 min residence time). To facilitate the processing of large amounts of material an in-line IR was connected after the BPR and a Python™ script was written which facilitated the switching between collection and waste if the IR absorbance fell above or below a predefined absorbance value. In addition, the script also facilitated switching from collection to waste in the instance of severe pressure and/or temperature fluctuations. In doing so, a manual intervention to correct the fluctuation could be undertaken and once the system was back within the defined range the scripted program would continue. Furthermore, integration through the Internet of Things (IoT) allowed remote monitoring of the process and included the generation of SMS notifications when the reactor switched from collection to waste, allowing the operator to then effect a manual intervention. In the second stage a reduction of the pyrazine ring system to 151 was afforded using an H-Cube®; unfortunately, the flow rate required for the second stage was significantly slower than that required for the first stage. A semi-continuous process was however developed by collecting the runoff from the first stage in a reservoir and then using a webcam and a “computer vision” approach to monitor the solvent level in the reservoir. Again, a Python™ script was used to automate the movement of material from the reservoir, thereby ensuring that the reservoir did not run dry.


image file: c8re00236c-s36.tif
Scheme 36 Ley fully automated synthesis of pyrazinamide 88 and pyrazine 151.

The realization of an intelligent, self-controlling platform for integrated continuous reactions requires an advanced integration of the chemistry, engineering and information systems.147 The three systems can be viewed as a pyramid wherein the apex of the pyramid is represented by chemical transformations, the successful implementation of which are built on engineering processes such as quenches, extractions solvent switches, filtrations etc. In a fully integrated system the engineering processes themselves are built on top of information processes which represent the control system incorporating aspects such as reservoir control, product detection, phase-separation controllers, valve controllers etc.

Ley and co-workers demonstrated this logic in an elegant five stage synthesis of 2-aminoadamantane-2-carboxylic acid 152.147 The process incorporates five reactions (Grignard, Ritter, cyclisation, hydrolysis and ozonolysis), the five reactions required, in addition to the chemical transformation itself, three quench stages, two extractions, two solvent swaps, a solvent removal and a filtration. The control system involved the acquisition of data from in-line monitors (IR and webcams), which was then utilized to dynamically control the required eleven pumping units (Scheme 37).

image file: c8re00236c-s37.tif
Scheme 37 Ley five stage synthesis of 2-aminoadamantane-2-carboxylic acid 152. Engineering processes: step 1 (quenching, extraction, solvent swap), step 2 (quenching and filtration), step 3 (extraction and solvent swap), step 4 (quenching), step 5 (solvent removal).

The control of chemical reactions in flow systems has been further extended to the cloud with the development of modular software that allows reaction monitoring via the internet on almost any web-enabled device. The Ley group highlighted this in the development of their cloud-based LeyLab software which allows communication between users and equipment through an internet browser.148 The software, notably through design of experiment, not only allows online automation but also autonomous self-optimization of reaction conditions. The principle was exemplified through a three-dimensional and five-dimensional optimization of a heterogeneous catalytic reaction and an Appel reaction.

Looking forward, the evidence suggests that the use of flow technologies will be a critical tool in the fully integrated and automated synthetic platform of the future. That being said, one cannot be solely dependent on a single piece or type of technology, and the development of automated platforms consisting of batch and flow-based modules is even more attractive, allowing one access to the best of both worlds.149,150 In recent times the development of such integrated batch–flow hybrid approaches has started drawing attention. Ley and co-workers notably described a fully automated batch–flow hybrid synthesis of 5-methyl-4-propylthiophene-2-carboxylic acid, a precursor to a novel breast cancer drug candidate.149 The process required the development of bespoke batch-based glassware which could be incorporated onto a commercial flow reactor platform.

Recently the use of in-line analysis in flow systems has been extended to the acquisition of biological data,151 pointing towards a future where integrated drug discovery and biological screening will become commonplace. Several approaches involving flow-based array chemistry platforms being linked to off-line purification and screening have been shown.152–154 In addition, approaches allowing in-line biological assessment have been demonstrated, where the Ley group used frontal affinity chromatography while developing a series of GABAA agonists,155,156 and by researchers from Roche who designed a completely integrated synthesis and bioassay platform.157 In the latter case a series of β-secretase inhibitors were prepared, purified in-line by HPLC, quantified by an evaporative light scattering detector and subjected to a Taylor dispersion prior to a fluorescence dequenching assay performed on a micro-chip. The process notably reduced the synthesis/biological screening time from several days to 1 h.

In 2017, scientists at Eli Lilly & Co. achieved a flow chemistry manufacturing milestone when they implemented a multi-step kilogram-scale synthesis of prexasertib monolactate monohydrate under continuous-flow cGMP conditions.158 The process incorporated eight continuous unit operations affording 3 kg of the target per day (a total of 24 kg were produced, which were suitable for human clinical trials). The fully automated process integrated the use of small continuous reactors, extractors, evaporators, crystallizers and filters, and its successful implementation was linked to advances in chemistry, engineering, analytics, process modelling and equipment design.

The cGMP production was dependent on integrated process analytical technology (PAT) which was used extensively during both the development and the implementation of the cGMP process. The PAT tools included online HPLC and refractive index measurements. In addition, temperature, pressure and mass flow rate monitoring were conducted by a distributed control system (DCS). The online HPLC data were not used for cGMP decision-making and forward-processing; decisions were instead reached using manual, offline testing. The online PAT instead provided feedback on minor adjustments to the process parameters, thereby allowing the process performance to be kept on target. Furthermore, the PAT was not integrated with the DCS to allow feedback control based on the analytical data.

5. Conclusions

Over almost two decades of development, flow chemistry has evolved from a novel synthesis concept into a powerful and versatile platform for continuous manufacturing of APIs with high productivity, a small manufacturing footprint, and reduced cost and waste. A new ambitious goal is now centered on integrating the entire pharmaceutical manufacturing process, from raw materials to final dosage forms, into a continuous flow process. Scientists from Massachusetts Institute of Technology (MIT) are leading this effort. In 2013, a research team at MIT (sponsored by Novartis) showcased the proof of concept by synthesizing aliskiren from advanced intermediates to final tablets in a continuous flow process. The manufacturing process was performed in a compact plant module (2.4 m × 7.3 m × 7.3 m) where flow synthesis, purification, formulation and tableting were fully integrated.159 In 2016, the same team developed a refrigerator-sized (1.0 m × 0.7 m × 1.8 m), reconfigurable manufacturing platform that can perform multistep synthesis, work-up, purification (e.g. crystallization), formulation and real-time monitoring. Four drugs were tested on the platform from commercially available starting materials to finished forms at a rate of hundreds to thousands of oral or topical liquid doses per day. Moreover, the final products met US Pharmacopeia standards.160

More recently, two remarkable FDA approvals have heralded a manufacturing paradigm shift towards continuous manufacturing. The first was for Vertex's Orkambi (lumacaftor/ivacaftor for cystic fibrosis) in 2015 as the first New Drug Application (NDA) approval using a continuous manufacturing technology for production. The second FDA approval was for Johnson & Johnson's Prezista (darunavir for HIV) in 2016 as the first NDA supplement approval for switching from batch manufacturing to continuous manufacturing.161 The FDA has been a strong advocate of continuous manufacturing since the launch of the Pharmaceutical cGMP initiative in 2002. According to the agency, there are no regulatory hurdles for implementing continuous manufacturing; however, there is a lack of experience.

Africa has a unique position in that the region has limited existing pharmaceutical manufacturing infrastructure that would hinder the adoption of a new technology. It is the opinion of the authors that the region has the opportunity to leapfrog the existing batch manufacturing technology and move directly into the modern continuous flow manufacturing approaches. There is a clear business case for regional API manufacturing companies to adopt new technologies and embrace digital disruption; these technologies can effectively limit the impact of critical drug shortages, protect against foreign exchange fluctuations and help to generate competitive differentiation. It is envisioned that the region can initially focus on the implementation of multi-step continuous flow synthesis, inclusion of green chemistry approaches, the use of robotics, process automation and the application of AI for process optimization, minimization of waste and energy consumption, and continuous monitoring systems in newly established API manufacturing facilities. However regional pharmaceutical companies will need to expand their skill sets and become deeply familiar with the world of software, data and service to cross the digital divide.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors would like to gratefully acknowledge Dr Boitumelo Semete-Makokotlela and Dr Dharmarai Naicker from the Council for Scientific and Industrial Research (CSIR, South Africa) for many helpful discussions and insightful inputs. This work was supported through the CSIR Parliamentary Grant and Researcher Accelerator Awards and the National Research Foundation (NRF, South Africa) Thuthuka grant (number 106959). Competitive Support for Unrated Researchers grant (number 116303). Opinions expressed in this publication and the conclusions arrived at are those of the authors and are not necessarily attributed to the CSIR or NRF.

References

  1. (a) African Union, Pharmaceutical manufacturing plan for Africa: Business Plan, 2012 Search PubMed; (b) WHO Policy Brief, Department of Public Health Innovation and Intellectual Property, Local Production for Access to Medical Products: Developing a Framework to Improve Public Health, 2011 Search PubMed; (c) WHO Report, Promoting and Supporting Local Manufacturing of Quality Medical Products in Developing Countries: The Business Case for Improving Access, 2011 Search PubMed; (d) Government of Ethiopia, National strategy and plan of action for pharmaceutical manufacturing development in Ethiopia (2015–2025), 2015 Search PubMed; (e) African Union, Position Paper on the Establishment of a Fund for the Development of the African Pharmaceutical Manufacturing Sector, Second Session of the Specialised Technical Committee on Health, Population and Drug Control (STC-HPDC-2), Ethiopia, 20–24 March 2017 Search PubMed; (f) African Union, Concept Note: Fund for African Pharmaceutical Development (FAP-D), August 2018, Cairo, Egypt, (https://au.int/en/newsevents/20180807/fund-african-pharmaceutical-development-fap-d-meeting-7-9-august-2018-cairo, accessed 21 December 2018) Search PubMed; (g) African Union, The 7th Strategic Stakeholders Retreat on Industry, February 2018, Nairobi, Kenya, (https://au.int/en/pressreleases/20180219/7th-strategic-stakeholders-retreat-industry, accessed 21 December 2018) Search PubMed.
  2. (a) T. Holt, M. Lahrichi, J. Mina and J. S. da Silva, Insights into pharmaceuticals and medical products - Africa: A continent of opportunity for pharma and patients, McKinsey & Company, 2015 Search PubMed; (b) African Pharmaceutical Market - Industry Analysis and Forecast; Goldstein Research, 2018 Search PubMed.
  3. (a) www.pharmexec.com/country-report-south-africa-0?pageID=4 (accessed 19 December 2018); (b) www.pharmexec.com/country-report-south-africa-0?pageID=2 (accessed 19 December 2018); (c) https://www.fin24.com/Companies/Health/more-pharmaceutical-skills-in-sa-will-increase-competition-counter-collusion-20170702 (accessed 19 December 2018); (d) J. H. Havenga, Z. P. Simpson, A. De Bod and N. M. Viljoen, South Africa's rising logistics costs: An uncertain future, Journal of Transport and Supply Chain Management, 2014, 8(1), 155 Search PubMed; (e) http://belfreight.com/news/addressing-the-challenges-of-transport-logistics-in-south-africa-10052016/ (accessed 19 December 2018); (f) https://www.bizcommunity.com/Article/196/733/171992.html (accessed 19 December 2018); (g) www.engineeringnews.co.za/article/the-south-african-pharmaceutical-regulatory-environment-2018-06-20/rep_id:4136) (accessed 19 December 2018); (h) https://tradingeconomics/south-africa/balance-of-trade (accessed 19 December 2018); (i) D. R. Walwyn, Briefing Note for the Pharmaceutical Industry: Proposed Support for the Local Manufacture of Active Pharmaceutical Ingredients, 2008 Search PubMed.
  4. SADC, SADC industrialization strategy and roadmap 2015–2020, 2015 Search PubMed.
  5. J. Olympio, G. Tomasin and J. Daniels, Formulation of industrialization program for the SADC secretariat, 2017 Search PubMed.
  6. http://www.itac.org.za/upload/Industrial%20Policy%20Action%20Plan%20Part%202%2017-18%20-%2019-20.pdf (accessed on 25 January 2019).
  7. BMI Research, S. A. Pharmaceuticals and healthcare report, Q3, 2015 Search PubMed.
  8. Figures sourced from Quantec.
  9. (a) T. Molelekwa, Essential medicines shortage hit Gauteng (Health-e News), 6 July 2017, https://www.health-e.org.za/2017/07/06/essential-medication-shortage-gauteng-hospitals-clinics/ (accessed 22 September 2018) Search PubMed; (b) G. Makam, Medicines stockout crisis in North West (Health-e News), 19 April 2018, https://www.health-e.org.za/2018/04/19/medicines-stockout-crisis-in-north-west/ (accessed 22 September 2018). Vol. 2018 Search PubMed.
  10. Republic of South Africa Government Gazette, Patents Act No 57 of 1978, 17 May 1978, https://www.gov.za/sites/default/files/Act%2057%<?pdb_no 20of?>20of<?pdb END?>%201978.pdf (accessed 22 September 2018) Search PubMed.
  11. (a) Republic of South Africa Government Gazette, Medicines and Related Substances Act 101 of 1965, 19 June 1965, http://www.hpcsa.co.za/Uploads/editor/UserFiles/downloads/legislations/acts/medicines_and_related_sub_act_101_of_1965.pdf (accessed 22 September 2018) Search PubMed; (b) Republic of South Africa Government Gazette, Medicines and Related Substances Amendment Bill, 20 February 2014, http://pmg-assets.s3-website-eu-west-1.amazonaws.com/bill6-2014_1.pdf (accessed 22 September 2018) Search PubMed.
  12. C. M. Correa, Public Health Reviews: Implications of bilateral free trade agreements on access to medicines, WHO Bulletin, 2006, vol. 84, 5, pp. 399–404 Search PubMed.
  13. WTO OMC, Fact Sheet: TRIPS and pharmaceutical patents, 2006 Search PubMed.
  14. R. Kampf, WTO Economic Research and Statistics Division, Special Compulsory Licenses for Export of Medicines: Key Features of WTO Member's Implementing Legislation, 2015 Search PubMed.
  15. Grand View Research, Flow Chemistry Market Analysis by Application, by Reactors, by Region and Segment Forecasts 2014–2025, March 2017 Search PubMed.
  16. B. Gutmann, D. Cantillo and C. O. Kappe, Continuous-Flow Technology—A Tool for the Safe Manufacturing of Active Pharmaceutical Ingredients, Angew. Chem., Int. Ed., 2015, 54(23), 6688–6728 CrossRef CAS PubMed.
  17. D. J. Ager, Managing Hazardous Reactions and Compounds in Process Chemistry, in Managing Hazardous Reactions and Compounds in Process Chemistry, American Chemical Society, 2014, vol. 1181, pp. 285–351 Search PubMed.
  18. P. J. Harrington and E. Lodewijk, Twenty Years of Naproxen Technology, Org. Process Res. Dev., 1997, 1(1), 72–76 CrossRef CAS.
  19. C. Jiménez-González, P. Poechlauer, Q. B. Broxterman, B.-S. Yang, D. a. Ende, J. Baird, C. Bertsch, R. E. Hannah, P. Dell'Orco, H. Noorman, S. Yee, R. Reintjens, A. Wells, V. Massonneau and J. Manley, Key Green Engineering Research Areas for Sustainable Manufacturing: A Perspective from Pharmaceutical and Fine Chemicals Manufacturers, Org. Process Res. Dev., 2011, 15(4), 900–911 CrossRef.
  20. P. Jansen van Vuuren, FACTSHEET: Africa's leading causes of death, Africa Check, 14th August 2017, https://africacheck.org/factsheets/factsheet-africas-leading-causes-death/ (accessed 22 September 2018) Search PubMed.
  21. The World Bank, The Global Burden of Disease: Main Findings for Sub-Saharan Africa, 4 September 2013 Search PubMed.
  22. J. I. Davies, A. J. Macnab, P. Byass, S. A. Norris, M. Nyirenda, A. Singhal, E. Sobngwi and A. S. Daar, Developmental origins of health and disease in Africa-influencing early life, Lancet Glob. Health, 2018, 6(3), e244–e245 CrossRef PubMed.
  23. (a) R. Porta, M. Benaglia and A. Puglisi, Flow Chemistry: Recent Developments in the Synthesis of Pharmaceutical Products, Org. Process Res. Dev., 2016, 20(1), 2–25 CrossRef CAS; (b) M. Baumann and I. R. Baxendale, The synthesis of active pharmaceutical ingredients (APIs) using continuous flow chemistry, Beilstein J. Org. Chem., 2015, 11, 1194–1219 CrossRef CAS PubMed; (c) R. O. M. A. de Souza and P. Watts, Flow processing as a tool for API production in developing economies, J. Flow Chem., 2017, 7(3–4), 146–150 CrossRef; (d) P. Bana, R. Örkényia, K. Lövei, A. Lakó, G. I. Túrós, J. Éles, F. Faigl and I. Greiner, The route from problem to solution in multistep continuous flow synthesis of pharmaceutical compounds, Bioorg. Med. Chem., 2017, 25(23), 6180–6189 CrossRef CAS PubMed.
  24. (a) J. M. D. Souza, R. Galaverna, R. A. A. N. D. Souza, T. J. Brocksom, J. C. Pastre, R. O. M. A. D. Souza and K. T. D. Oliveira, Impact of continuous flow chemistry in the synthesis of natural products and active pharmaceutical ingredients, An. Acad. Bras. Cienc., 2018, 90, 1131–1174 CrossRef PubMed; (b) J. Britton and C. L. Raston, Multi-step continuous flow synthesis, Chem. Soc. Rev., 2017, 46, 1250–1271 RSC.
  25. (a) R. Gérardy, N. Emmanuel, T. Toupy, V.-E. Kassin, N. N. Tshibalonza, M. Schmitz and J.-C. M. Monbaliu, Continuous Flow Organic Chemistry: Successes and Pitfalls at the Interface with Current Societal Challenges, Eur. J. Org. Chem., 2018, 2018(20–21), 2301–2351 CrossRef; (b) M. B. Plutschack, B. Pieber, K. Gilmore and P. H. Seeberger, The Hitchhiker's Guide to Flow Chemistry, Chem. Rev., 2017, 117(18), 11796–11893 CrossRef CAS PubMed.
  26. WHO Africa HIV/AIDS https://afro.who.int/health-topics/hivaids (accessed 29 September, 2018) Search PubMed.
  27. D. Mandala, W. A. Thompson and P. Watts, Synthesis routes to anti-HIV drugs, Tetrahedron, 2016, 72(24), 3389–3420 CrossRef CAS.
  28. M. D. Goodyear, M. L. Hill, J. P. West and A. J. Whitehead, Practical enantioselective synthesis of lamivudine (3TC™) via a dynamic kinetic resolution, Tetrahedron Lett., 2005, 46(49), 8535–8538 CrossRef CAS.
  29. M. F. Caso, D. D'Alonzo, S. D'Errico, G. Palumbo and A. Guaragna, Highly Stereoselective Synthesis of Lamivudine (3TC) and Emtricitabine (FTC) by a Novel N-Glycosidation Procedure, Org. Lett., 2015, 17(11), 2626–2629 CrossRef CAS PubMed.
  30. D. Mandala and P. Watts, An Improved Synthesis of Lamivudine and Emtricitabine, ChemistrySelect, 2017, 2(3), 1102–1105 CrossRef CAS.
  31. D. Mandala, S. Chada and P. Watts, Semi-continuous multi-step synthesis of lamivudine, Org. Biomol. Chem., 2017, 15(16), 3444–3454 RSC.
  32. R. F. Boswell, B. F. Gupton and Y. S. Lo Method for making nevirapine, US6680383B1, 2002.
  33. A. R. Longstreet, S. M. Opalka, B. S. Campbell, B. F. Gupton and D. T. McQuade, Investigating the continuous synthesis of a nicotinonitrile precursor to nevirapine, Beilstein J. Org. Chem., 2013, 9, 2570–2578 CrossRef PubMed.
  34. J. Verghese, C. J. Kong, D. Rivalti, E. C. Yu, R. Krack, J. Alcázar, J. B. Manley, D. T. McQuade, S. Ahmad, K. Belecki and B. F. Gupton, Increasing global access to the high-volume HIV drug nevirapine through process intensification, Green Chem., 2017, 19(13), 2986–2991 RSC.
  35. M. E. Pierce, R. L. Parsons, L. A. Radesca, Y. S. Lo, S. Silverman, J. R. Moore, Q. Islam, A. Choudhury, J. M. D. Fortunak, D. Nguyen, C. Luo, S. J. Morgan, W. P. Davis, P. N. Confalone, C.-Y. Chen, R. D. Tillyer, L. Frey, L. Tan, F. Xu, D. Zhao, A. S. Thompson, E. G. Corley, E. J. J. Grabowski, R. Reamer and P. J. Reider, Practical Asymmetric Synthesis of Efavirenz (DMP 266), an HIV-1 Reverse Transcriptase Inhibitor, J. Org. Chem., 1998, 63(23), 8536–8543 CrossRef CAS.
  36. C. A. Correia, K. Gilmore, D. T. McQuade and P. H. Seeberger, A Concise Flow Synthesis of Efavirenz, Angew. Chem., Int. Ed., 2015, 54(16), 4945–4948 CrossRef CAS PubMed.
  37. S. Chada, D. Mandala and P. Watts, Synthesis of a key intermediate towards the preparation of efavirenz using n-butyllithium, J. Flow Chem., 2017, 7(2), 37–40 CrossRef.
  38. P. Rivas, J. Morello, C. Garrido, S. Rodríguez-Nóvoa and V. Soriano, Role of atazanavir in the treatment of HIV infection, Ther. Clin. Risk Manage., 2009, 5, 99–116 CAS.
  39. S. Kim, B. T. Lotz, M. F. Malley, J. Z. Gougoutas, M. Davidovich and S. K. Srivastava, Process for preparing atazanavir bisulfate and novel forms, US8513428B2, 2004.
  40. L. Dalla-Vechia, B. Reichart, T. Glasnov, L. S. M. Miranda, C. O. Kappe and R. O. M. A. de Souza, A three step continuous flow synthesis of the biaryl unit of the HIV protease inhibitor Atazanavir, Org. Biomol. Chem., 2013, 11(39), 6806–6813 RSC.
  41. V. D. Pinho, B. Gutmann, L. S. M. Miranda, R. O. M. A. de Souza and C. O. Kappe, Continuous Flow Synthesis of α-Halo Ketones: Essential Building Blocks of Antiretroviral Agents, J. Org. Chem., 2014, 79(4), 1555–1562 CrossRef CAS PubMed.
  42. G. A. Boyle, C. D. Edlin, Y. Li, D. C. Liotta, G. L. Morgans and C. C. Musonda, Enantioselective synthesis of the carbocyclic nucleoside (−)-abacavir, Org. Biomol. Chem., 2012, 10(9), 1870–1876 RSC.
  43. D. L. Riley, D. R. Walwyn and C. D. Edlin, An Improved Process for the Preparation of Tenofovir Disoproxil Fumarate, Org. Process Res. Dev., 2016, 20(4), 742–750 CrossRef CAS.
  44. M. N. Arimilli, K. C. Cundy, J. P. Dougherty, C. U. Kim, R. Oliyai and V. J. Stella, Antiviral phosphonomethyoxy nucleotide analogs having increased oral bioavarilability, US5922695A, 1998.
  45. D. H. B. Ripin, D. S. Teager, J. Fortunak, S. M. Basha, N. Bivins, C. N. Boddy, S. Byrn, K. K. Catlin, S. R. Houghton, S. T. Jagadeesh, K. A. Kumar, J. Melton, S. Muneer, L. N. Rao, R. V. Rao, P. C. Ray, N. G. Reddy, R. M. Reddy, K. C. Shekar, T. Silverton, D. T. Smith, R. W. Stringham, G. V. Subbaraju, F. Talley and A. Williams, Process Improvements for the Manufacture of Tenofovir Disoproxil Fumarate at Commercial Scale, Org. Process Res. Dev., 2010, 14(5), 1194–1201 CrossRef CAS.
  46. WHO Africa Tubercolosis (TB) https://afro.who.int/health-topics/tuberculosis-tb (accessed 29 September 2018) Search PubMed.
  47. https://www.tbfacts.org/tb-drugs/ (accessed 29 September 2018).
  48. J. J. Li, D. S. Johnson, D. R. Sliskovic and B. D. Roth, Antibacterials, in Contemporary Drug Synthesis, Wiley-Interscience, 2004 Search PubMed.
  49. A. A. MacDonald, S. H. DeWitt, E. M. Hogan and R. Ramage, A solid phase approach to quinolones using the DIVERSOMER® technology, Tetrahedron Lett., 1996, 37(27), 4815–4818 CrossRef CAS.
  50. A. M. Hay, S. Hobbs-Dewitt, A. A. MacDonald and R. Ramage, Use of tetrabenzo[a,c,g,i]fluorene as an anchor group for the solid/solution phase synthesis of Ciprofloxacin ®, Tetrahedron Lett., 1998, 39(47), 8721–8724 CrossRef CAS.
  51. A. M. Hay, S. Hobbs-Dewitt, A. A. MacDonald and R. Ramage, Use of Tetrabenzo[a,c,g,i]fluorene as an Anchor Group for the Solid/Solution Phase Synthesis of the Quinolone Antibacterial, Ciprofloxacin, Synthesis, 1999, 11, 1979–1985 CrossRef.
  52. H. Lin, C. Dai, T. F. Jamison and K. F. Jensen, A Rapid Total Synthesis of Ciprofloxacin Hydrochloride in Continuous Flow, Angew. Chem., Int. Ed., 2017, 56(30), 8870–8873 CrossRef CAS PubMed.
  53. S. J. Brickner, M. R. Barbachyn, D. K. Hutchinson and P. R. Manninen, Linezolid (ZYVOX), the First Member of a Completely New Class of Antibacterial Agents for Treatment of Serious Gram-Positive Infections, J. Med. Chem., 2008, 51(7), 1981–1990 CrossRef CAS PubMed.
  54. D. Cantillo, M. M. Moghaddam and C. O. Kappe, Hydrazine-mediated Reduction of Nitro and Azide Functionalities Catalyzed by Highly Active and Reusable Magnetic Iron Oxide Nanocrystals, J. Org. Chem., 2013, 78(9), 4530–4542 CrossRef CAS PubMed.
  55. C. Battilocchio, J. M. Hawkins and S. V. Ley, Mild and Selective Heterogeneous Catalytic Hydration of Nitriles to Amides by Flowing through Manganese Dioxide, Org. Lett., 2014, 16(4), 1060–1063 CrossRef CAS PubMed.
  56. R. J. Ingham, C. Battilocchio, J. M. Hawkins and S. V. Ley, Integration of enabling methods for the automated flow preparation of piperazine-2-carboxamide, Beilstein J. Org. Chem., 2014, 10, 641–652 CrossRef PubMed.
  57. C. D. Scott, R. Labes, M. Depardieu, C. Battilocchio, M. G. Davidson, S. V. Ley, C. C. Wilson and K. Robertson, Integrated plug flow synthesis and crystallisation of pyrazinamide, React. Chem. Eng., 2018, 3, 631–634 RSC.
  58. World Health Organization, World malaria report 2015, World Health Organization, 2016 Search PubMed.
  59. J. L. Howard, C. Schotten and D. L. Browne, Continuous flow synthesis of antimalarials: opportunities for distributed autonomous chemical manufacturing, React. Chem. Eng., 2017, 2(3), 281–287 RSC.
  60. J. A. Flegg, C. J. E. Metcalf, M. Gharbi, M. Venkatesan, T. Shewchuk, C. Hopkins Sibley and P. J. Guerin, Trends in Antimalarial Drug Use in Africa, Am. J. Trop. Med. Hyg., 2013, 89(5), 857–865 CrossRef PubMed.
  61. N. J. White, Qinghaosu (artemisinin): the price of success, Science, 2008, 320(5874), 330–334 CrossRef CAS PubMed.
  62. G. Schmid and W. Hofheinz, Total synthesis of qinghaosu, J. Am. Chem. Soc., 1983, 105(3), 624–625 CrossRef CAS.
  63. J. Yadav, R. S. Babu and G. Sabitha, Stereoselective total synthesis of (+)-artemisinin, Tetrahedron Lett., 2003, 44(2), 387–389 CrossRef CAS.
  64. X. X. Xu, J. Zhu, D. Z. Huang and W. S. Zhou, Total Synthesis of Arteannuin and Deoxyarteannuin, Chem. Inform., 1986, 17(24) CAS.
  65. J. Yadav, B. Thirupathaiah and P. Srihari, A concise stereoselective total synthesis of (+)-artemisinin, Tetrahedron, 2010, 66(11), 2005–2009 CrossRef CAS.
  66. L. Hsing-Jang, Y. Wen-Lung and Y. C. Sew, A total synthesis of the antimarial natural product (+)-qinghaosu, Tetrahedron Lett., 1993, 34(28), 4435–4438 CrossRef.
  67. R. J. Roth and N. Acton, A simple conversion of artemisinic acid into artemisinin, J. Nat. Prod., 1989, 52(5), 1183–1185 CrossRef CAS.
  68. J. I. Turconi, F. d. r. Griolet, R. Guevel, G. Oddon, R. Villa, A. Geatti, M. Hvala, K. Rossen, R. Göller and A. Burgard, Semisynthetic artemisinin, the chemical path to industrial production, Org. Process Res. Dev., 2014, 18(3), 417–422 CrossRef CAS.
  69. D. Kopetzki, F. Lévesque and P. H. Seeberger, A continuous-flow process for the synthesis of artemisinin, Chem. – Eur. J., 2013, 19(17), 5450–5456 CrossRef CAS PubMed.
  70. H. G. Jolliffe and D. I. Gerogiorgis, Plantwide design and economic evaluation of two Continuous Pharmaceutical Manufacturing (CPM) cases: Ibuprofen and artemisinin, Comput. Chem. Eng., 2016, 91, 269–288 CrossRef CAS.
  71. K. Chibale, How Africa is helping expand the global antimalarial drug pipeline: health research, Quest, 2016, 12(3), 22–23 Search PubMed.
  72. J. Okombo and K. Chibale, Recent updates in the discovery and development of novel antimalarial drug candidates, Med. Chem. Commun., 2018, 9(3), 437–453 RSC.
  73. J. N. Burrows, K. Chibale and T. N. C. Wells, The state of the art in anti-malarial drug discovery and development, Curr. Top. Med. Chem., 2011, 11(10), 1226–1254 CrossRef CAS PubMed.
  74. E. L. Flannery, A. K. Chatterjee and E. A. Winzeler, Antimalarial drug discovery—approaches and progress towards new medicines, Nat. Rev. Microbiol., 2013, 11(12), 849 CrossRef CAS PubMed.
  75. J. L. Vennerstrom, S. Arbe-Barnes, R. Brun, S. A. Charman, F. C. Chiu, J. Chollet, Y. Dong, A. Dorn, D. Hunziker and H. Matile, Identification of an antimalarial synthetic trioxolane drug development candidate, Nature, 2004, 430(7002), 900 CrossRef CAS PubMed.
  76. S. A. Charman, S. Arbe-Barnes, I. C. Bathurst, R. Brun, M. Campbell, W. N. Charman, F. C. Chiu, J. Chollet, J. C. Craft and D. J. Creek, Synthetic ozonide drug candidate OZ439 offers new hope for a single-dose cure of uncomplicated malaria, Proc. Natl. Acad. Sci. U. S. A., 2011, 108(11), 4400–4405 CrossRef CAS PubMed.
  77. S.-H. Lau, A. Galván, R. R. Merchant, C. Battilocchio, J. A. Souto, M. B. Berry and S. V. Ley, Machines vs malaria: a flow-based preparation of the drug candidate OZ439, Org. Lett., 2015, 17(13), 3218–3221 CrossRef CAS PubMed.
  78. E. Yu, H. P. R. Mangunuru, N. S. Telang, C. J. Kong, J. Verghese, S. E. Gilliland III, S. Ahmad, R. N. Dominey and B. F. Gupton, High-yielding continuous-flow synthesis of antimalarial drug hydroxychloroquine, Beilstein J. Org. Chem., 2018, 14, 583–592 CrossRef CAS PubMed.
  79. H. L. Yu, F. Ning, Z. Zheng, Q. Yu, X. Niu and C. Li, Simple and green method for synthesizing 5-(N-ethyl-N-2-hydroxyethylamino)-2-amylamine, 2015 Search PubMed.
  80. (a) P. J. Hotez, M. Alvarado, M.-G. Basáñez, I. Bolliger, R. Bourne, M. Boussinesq, S. J. Brooker, A. S. Brown, G. Buckle and C. M. Budke, The global burden of disease study 2010: interpretation and implications for the neglected tropical diseases, PLoS Neglected Trop. Dis., 2014, 8(7), e2865 CrossRef PubMed; (b) C. J. Murray, T. Vos, R. Lozano, M. Naghavi, A. D. Flaxman, C. Michaud, M. Ezzati, K. Shibuya, J. A. Salomon and S. Abdalla, Disability-adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010, Lancet, 2012, 380(9859), 2197–2223 CrossRef PubMed.
  81. J. R. Perfect, Is there an emerging need for new antifungals?, Expert Opin. Emerging Drugs, 2016, 21(2), 129–131 CrossRef PubMed.
  82. A. Loyse, H. Thangaraj, P. Easterbrook, N. Ford, M. Roy, T. Chiller, N. Govender, T. S. Harrison and T. Bicanic, Cryptococcal meningitis: improving access to essential antifungal medicines in resource-poor countries, Lancet Infect. Dis., 2013, 13(7), 629–637 CrossRef PubMed.
  83. R. Rajasingham, R. M. Smith, B. J. Park, J. N. Jarvis, N. P. Govender, T. M. Chiller, D. W. Denning, A. Loyse and D. R. Boulware, Global burden of disease of HIV-associated cryptococcal meningitis: an updated analysis, Lancet Infect. Dis., 2017, 17(8), 873–881 CrossRef PubMed.
  84. B. J. Park, K. A. Wannemuehler, B. J. Marston, N. Govender, P. G. Pappas and T. M. Chiller, Estimation of the current global burden of cryptococcal meningitis among persons living with HIV/AIDS, AIDS, 2009, 23(4), 525–530 CrossRef PubMed.
  85. D. Armstrong-James, G. Meintjes and G. D. Brown, A neglected epidemic: fungal infections in HIV/AIDS, Trends Microbiol., 2014, 22(3), 120–127 CrossRef CAS PubMed.
  86. World Health Organization, WHO model list of essential medicines, 20th list (March 2017, amended August 2017), 2017 Search PubMed.
  87. A. Kleemann, J. Engel, B. Kutscher and D. Reichert, Pharmaceutical Substances, 2009: Syntheses, Patents and Applications of the most relevant APIs, Georg Thieme Verlag, 2014 Search PubMed.
  88. D. Robert, 2-Chloro-4-amino-5-fluoropyrimidine, US3185690A, 1965.
  89. D. Robert, Processes and intermediates for deoxyfluorocytidine, US3040026A, 1962.
  90. F. Von Nussbaum, M. Brands, B. Hinzen, S. Weigand and D. Häbich, Antibacterial natural products in medicinal chemistry—exodus or revival?, Angew. Chem., Int. Ed., 2006, 45(31), 5072–5129 CrossRef CAS PubMed.
  91. H. Groeger, M. Pieper, B. Koenig, T. Bayer and H. Schleich, Industrial landmarks in the development of sustainable production processes for the β-lactam antibiotic key intermediate 7-aminocephalosporanic acid (7-ACA), Sustainable Chem. Pharm., 2017, 5, 72–79 CrossRef CAS.
  92. M. Pieper, M. Kumpert, B. König, H. Schleich, T. Bayer and H. Gröger, Process Development for Synthesizing the Cephalosporin Antibiotic Cefotaxime in Batch and Flow Mode, Org. Process Res. Dev., 2018, 22(8), 947–954 CrossRef CAS.
  93. J. L. Zhang, W. Qian and Y. Yuhong, Baolin and J. Gao, Method for preparation of Cefotaxime acid, CN102702230, 2012.
  94. D. W. Miao, J. Longke, J. Xu and J. X. Zeng, Process for preparation of Cefotaxime, CN101560217, 2009.
  95. H. L. Fu and X. Tian, Environment- friendly method for preparation of third- generation cephalosporins and removing byproduct 2-mercaptobenzothiazole, CN102453041, 2012.
  96. Y. Yako, J. Echouffo-Tcheugui, E. Balti, T. Matsha, E. Sobngwi, R. Erasmus and A. Kengne, Genetic association studies of obesity in A frica: a systematic review, Obes. Rev., 2015, 16(3), 259–272 CrossRef CAS PubMed.
  97. Y. Y. Yako, M. Guewo-Fokeng, E. V. Balti, N. Bouatia-Naji, T. E. Matsha, E. Sobngwi, R. T. Erasmus, J. B. Echouffo-Tcheugui and A. P. Kengne, Genetic risk of type 2 diabetes in populations of the African continent: A systematic review and meta-analyses, Diabetes Res. Clin. Pract., 2016, 114, 136–150 CrossRef CAS PubMed.
  98. NRFCA Group, Trends in obesity and diabetes across Africa from 1980 to 2014: an analysis of pooled population-based studies, Int. J. Epidemiol., 2017, 46(5), 1421–1432 CrossRef PubMed.
  99. L. Pellegatti and J. r. Sedelmeier, Synthesis of vildagliptin utilizing continuous flow and batch technologies, Org. Process Res. Dev., 2015, 19(4), 551–554 CrossRef CAS.
  100. D. Levin, Potential toxicological concerns associated with carboxylic acid chlorination and other reactions, Org. Process Res. Dev., 1997, 1(2), 182–182 CrossRef CAS.
  101. J. Grimm, C. Maryanoff, M. Patel, D. Palmer, K. Sorgi, S. Stefanick, R. Webster and X. Zhang, Reaction Safety: a critical parameter in the development of a scaleable synthesis of 2, 3-bis-chloromethylpyridine hydrochloride, ACS Publications, 2002 Search PubMed.
  102. A. Miyake, M. Suzuki, M. Sumino, Y. Iizuka and T. Ogawa, Thermal hazard evaluation of Vilsmeier reaction with DMF and MFA, Org. Process Res. Dev., 2002, 6(6), 922–925 CrossRef CAS.
  103. M. Bollyn, Thermal Hazards of the Vilsmeier− Haack Reaction on N,N-Dimethylaniline, Org. Process Res. Dev., 2005, 9(6), 982–996 CrossRef CAS.
  104. M. Stare, K. Laniewski, A. Westermark, M. Sjögren and W. Tian, Investigation on the Formation and Hydrolysis of N, N-Dimethylcarbamoyl Chloride (DMCC) in Vilsmeier Reactions Using GC/MS as the Analytical Detection Method, Org. Process Res. Dev., 2009, 13(5), 857–862 CrossRef CAS.
  105. U. C. Dyer, D. A. Henderson, M. B. Mitchell and P. D. Tiffin, Scale-up of a vilsmeier formylation reaction: use of hel auto-mate and simulation techniques for rapid and safe transfer to pilot plant from laboratory, Org. Process Res. Dev., 2002, 6(3), 311–316 CrossRef CAS.
  106. OECD, Enabling the Next Production Revolution: The Future of Manufacturing and Services - Interim Report, Meeting of the OECD Council at Ministerial Level, Paris, 1–2 June 2016 Search PubMed.
  107. J. Tao and J.-H. Xu, Biocatalysis for the Pharmaceutical Industry, John Wiley & Sons, Asia, 2009 Search PubMed.
  108. J. Britton, S. Majumdar and G. A. Weiss, Continuous flow biocatalysis, Chem. Soc. Rev., 2018, 47(15), 5891–5918 RSC.
  109. K. Tipton and S. Boyce, Bioinformatics, 2006, 16, 34–40 CrossRef.
  110. A. J. J. Straathof, S. Panke and A. Schmid, The production of fine chemicals by biotransformations, Curr. Opin. Biotechnol., 2002, 13(6), 548–556 CrossRef CAS PubMed.
  111. J.-Z. Li, L.-X. Gao and M.-X. Ding, The chemical resolution of racemic cis-2-hydroxymethyl-5-(cytosine-1′-yl)-1,3-oxathiolane (BCH-189)—One direct method to obtain lamivudine as anti-HIV and and anti-HBV agent, Synth. Commun., 2002, 32(15), 2355–2359 CrossRef CAS.
  112. B. N. Roy, G. P. Singh, D. Srivastava, H. S. Jadhav, M. B. Saini and U. P. Aher, A Novel Method for Large-Scale Synthesis of Lamivudine through Cocrystal Formation of Racemic Lamivudine with (S)-(−)-1,1′-Bi(2-naphthol) [(S)-(BINOL)], Org. Process Res. Dev., 2009, 13(3), 450–455 CrossRef CAS.
  113. L. K. Hoong, L. E. Strange, D. C. Liotta, G. W. Koszalka, C. L. Burns and R. F. Schinazi, Enzyme-mediated enantioselective preparation of pure enantiomers of the antiviral agent 2′,3′-dideoxy-5-fluoro-3′-thiacytidine (FTC) and related compounds, J. Org. Chem., 1992, 57(21), 5563–5565 CrossRef CAS.
  114. J. Milton, S. Brand, M. F. Jones and C. M. Rayner, Enantioselective enzymatic synthesis of the anti-viral agent lamivudine (3TC™), Tetrahedron Lett., 1995, 36(38), 6961–6964 CAS.
  115. R. P. C. Cousins, M. Mahmoudian and P. M. Youds, Enzymic resolution of oxathiolane intermediates — an alternative approach to the anti-viral agent lamivudine (3TC™), Tetrahedron: Asymmetry, 1995, 6(2), 393–396 CrossRef CAS.
  116. L. Hu, F. Schaufelberger, Y. Zhang and O. Ramström, Efficient asymmetric synthesis of lamivudine via enzymatic dynamic kinetic resolution, Chem. Commun., 2013, 49(88), 10376–10378 RSC.
  117. R. N. Patel, Biocatalytic Synthesis of Chiral Intermediates, Food Technol. Biotechnol., 2004, 42(2), 305–325 CAS.
  118. R. N. Patel, L. Chu and R. Mueller, Diastereoselective microbial reduction of (S)-[3-chloro-2-oxo-1-(phenylmethyl)propyl]carbamic acid, 1,1-dimethylethyl ester, Tetrahedron: Asymmetry, 2003, 14(20), 3105–3109 CrossRef CAS.
  119. J. P. Adams, A. J. Collins, R. K. Henderson and P. W. Sutton, Biotransformations in small-molecule pharmaceutical development, in Practical methods for biocatalysis and biotransformations, ed. J. Whittall and P. W. Sutton, Wiley, UK, 2010 Search PubMed.
  120. M. Mahmoudian, Biocatal. Biotransform., 2000, 18(2), 105–118 CrossRef CAS.
  121. M. Mahmoudian, A decade of biocatalysis at Glaxo Wellcome, in Biocatalysis in the pharmaceutical and biotechnology industries, ed. R. N. Patel, CRC Press, USA, 2007 Search PubMed.
  122. Novozyme: Savinase 6T, Lipolase 100T, Novo Nordisk enzyme data sheet, 1996 Search PubMed.
  123. Y. Wang, Y. Cao, Y. Li, J. Jin, J. Li and H. Song, Asymmetric biosynthesis of intermediates of anti-HIV drugs, Tetrahedron: Asymmetry, 2017, 28(6), 745–757 CrossRef CAS.
  124. S. H. Kung, S. Lund, A. Murarka, D. McPhee and C. J. Paddon, Approaches and Recent Developments for the Commercial Production of Semi-Synthetic Artemisinin, Front. Plant Sci., 2018, 9(87) Search PubMed.
  125. C. J. Paddon, P. J. Westfall, D. J. Pitera, K. Benjamin, K. Fisher, D. McPhee, M. D. Leavell, A. Tai, A. Main, D. Eng, D. R. Polichuk, K. H. Teoh, D. W. Reed, T. Treynor, J. Lenihan, H. Jiang, M. Fleck, S. Bajad, G. Dang, D. Dengrove, D. Diola, G. Dorin, K. W. Ellens, S. Fickes, J. Galazzo, S. P. Gaucher, T. Geistlinger, R. Henry, M. Hepp, T. Horning, T. Iqbal, L. Kizer, B. Lieu, D. Melis, N. Moss, R. Regentin, S. Secrest, H. Tsuruta, R. Vazquez, L. F. Westblade, L. Xu, M. Yu, Y. Zhang, L. Zhao, J. Lievense, P. S. Covello, J. D. Keasling, K. K. Reiling, N. S. Renninger and J. D. Newman, High-level semi-synthetic production of the potent antimalarial artemisinin, Nature, 2013, 496, 528 CrossRef CAS PubMed.
  126. K. Boehme and H. D. Brauer, Generation of singlet oxygen from hydrogen peroxide disproportionation catalyzed by molybdate ions, Inorg. Chem., 1992, 31(16), 3468–3471 CrossRef CAS.
  127. V. Nardello, S. Bogaert, P. L. Alsters and J.-M. Aubry, Singlet oxygen generation from H2O2/MoO42−: peroxidation of hydrophobic substrates in pure organic solvents, Tetrahedron Lett., 2002, 43(48), 8731–8734 CrossRef CAS.
  128. J. Turconi, F. Griolet, R. Guevel, G. Oddon, R. Villa, A. Geatti, M. Hvala, K. Rossen, R. Göller and A. Burgard, Semisynthetic Artemisinin, the Chemical Path to Industrial Production, Org. Process Res. Dev., 2014, 18(3), 417–422 CrossRef CAS.
  129. A. Burgard, T. Gieshoff, A. Peschl, D. Horstermann, C. Keleschovsky, R. Villa, S. Michelis and M. P. Feth, Optimisation of the photochemical oxidation step in the industrial synthesis of artemisinin, Chem. Eng. J., 2016, 294, 83–96 CrossRef CAS.
  130. N. Ran, L. Zhao, Z. Chen and J. Tao, Recent applications of biocatalysis in developing green chemistry for chemical synthesis at the industrial scale, Green Chem., 2008, 10(4), 361–372 RSC.
  131. P. A. Anastas and J. C. Warner, Green Chemistry: Theory and Practice, Oxford University Press, USA, 1998 Search PubMed.
  132. J. L. Tucker, Green Chemistry, a Pharmaceutical Perspective, Org. Process Res. Dev., 2006, 10(2), 315–319 CrossRef CAS.
  133. L. Zhao and N. Ran, Recent Advances in Developing Chemoenzymatic Processes for Active Pharmaceutical Ingredients, Org. Process Res. Dev., 2007, 11(2), 259–267 CrossRef.
  134. R. A. Sheldon and J. M. Woodley, Role of Biocatalysis in Sustainable Chemistry, Chem. Rev., 2018, 118(2), 801–838 CrossRef CAS PubMed.
  135. L. Hajba and A. Guttman, Continuous-Flow Biochemical Reactors: Biocatalysis, Bioconversion, and Bioanalytical Applications Utilizing Immobilized Microfluidic Enzyme Reactors, J. Flow Chem., 2016, 6(1), 8–12 CrossRef CAS.
  136. R. H. Ringborg and J. M. Woodley, The application of reaction engineering to biocatalysis, React. Chem. Eng., 2016, 1(1), 10–22 RSC.
  137. M. V. Gomez and A. de la Hoz, NMR reaction monitoring in flow synthesis, Beilstein J. Org. Chem., 2017, 13, 285–300 CrossRef CAS PubMed.
  138. P. Giraudeau and F.-X. Felpin, Flow reactors integrated with in-line monitoring using benchtop NMR spectroscopy, React. Chem. Eng., 2018, 3(4), 399–413 RSC.
  139. D. L. Browne, S. Wright, B. J. Deadman, S. Dunnage, I. R. Baxendale, R. M. Turner and S. V. Ley, Continuous flow reaction monitoring using an on-line miniature mass spectrometer, Rapid Commun. Mass Spectrom., 2012, 26(17), 1999–2010 CrossRef CAS PubMed.
  140. T. A. Hamlin and N. E. Leadbeater, Real-time Monitoring of Reactions Performed Using Continuous-flow Processing: The Preparation of 3-Acetylcoumarin as an Example, J. Visualized Exp., 2015, 105, e52393 Search PubMed.
  141. H. Lange, C. F. Carter, M. D. Hopkin, A. Burke, J. G. Goode, I. R. Baxendale and S. V. Ley, A breakthrough method for the accurate addition of reagents in multi-step segmented flow processing, Chem. Sci., 2011, 2(4), 765–769 RSC.
  142. C. F. Carter, H. Lange, S. V. Ley, I. R. Baxendale, B. Wittkamp, J. G. Goode and N. L. Gaunt, ReactIR Flow Cell: A New Analytical Tool for Continuous Flow Chemical Processing, Org. Process Res. Dev., 2010, 14(2), 393–404 CrossRef CAS.
  143. S. V. Ley, D. E. Fitzpatrick, R. J. Ingham and R. M. Myers, Organic Synthesis: March of the Machines, Angew. Chem., Int. Ed., 2015, 54(11), 3449–3464 CrossRef CAS PubMed.
  144. M. O'Brien, P. Koos, D. L. Browne and S. V. Ley, A prototype continuous-flow liquid–liquid extraction system using open-source technology, Org. Biomol. Chem., 2012, 10(35), 7031–7036 RSC.
  145. D. X. Hu, M. O'Brien and S. V. Ley, Continuous Multiple Liquid–Liquid Separation: Diazotization of Amino Acids in Flow, Org. Lett., 2012, 14(16), 4246–4249 CrossRef CAS PubMed.
  146. S. V. Ley, R. J. Ingham, M. O'Brien and D. L. Browne, Camera-enabled techniques for organic synthesis, Beilstein J. Org. Chem., 2013, 9, 1051–1072 CrossRef CAS PubMed.
  147. R. J. Ingham, C. Battilocchio, D. E. Fitzpatrick, E. Sliwinski, J. M. Hawkins and S. V. Ley, A Systems Approach towards an Intelligent and Self-Controlling Platform for Integrated Continuous Reaction Sequences, Angew. Chem., Int. Ed., 2015, 54(1), 144–148 CrossRef CAS PubMed.
  148. D. E. Fitzpatrick, C. Battilocchio and S. V. Ley, A Novel Internet-Based Reaction Monitoring, Control and Autonomous Self-Optimization Platform for Chemical Synthesis, Org. Process Res. Dev., 2016, 20(2), 386–394 CrossRef CAS.
  149. D. E. Fitzpatrick and S. V. Ley, Engineering chemistry: integrating batch and flow reactions on a single, automated reactor platform, React. Chem. Eng., 2016, 1(6), 629–635 RSC.
  150. N. C. Neyt and D. L. Riley, Batch–flow hybrid synthesis of the antipsychotic clozapine, React. Chem. Eng., 2018, 3(1), 17–24 RSC.
  151. M. Baumann, Integrating continuous flow synthesis with in-line analysis and data generation, Org. Biomol. Chem., 2018, 16(33), 5946–5954 RSC.
  152. B. Desai, K. Dixon, E. Farrant, Q. Feng, K. R. Gibson, W. P. van Hoorn, J. Mills, T. Morgan, D. M. Parry, M. K. Ramjee, C. N. Selway, G. J. Tarver, G. Whitlock and A. G. Wright, Rapid Discovery of a Novel Series of Abl Kinase Inhibitors by Application of an Integrated Microfluidic Synthesis and Screening Platform, J. Med. Chem., 2013, 56(7), 3033–3047 CrossRef CAS PubMed.
  153. W. Czechtizky, J. Dedio, B. Desai, K. Dixon, E. Farrant, Q. Feng, T. Morgan, D. M. Parry, M. K. Ramjee, C. N. Selway, T. Schmidt, G. J. Tarver and A. G. Wright, Integrated Synthesis and Testing of Substituted Xanthine Based DPP4 Inhibitors: Application to Drug Discovery, ACS Med. Chem. Lett., 2013, 4(8), 768–772 CrossRef CAS PubMed.
  154. A. Baranczak, N. P. Tu, J. Marjanovic, P. A. Searle, A. Vasudevan and S. W. Djuric, Integrated Platform for Expedited Synthesis–Purification–Testing of Small Molecule Libraries, ACS Med. Chem. Lett., 2017, 8(4), 461–465 CrossRef CAS PubMed.
  155. L. Guetzoyan, N. Nikbin, I. R. Baxendale and S. V. Ley, Flow chemistry synthesis of zolpidem, alpidem and other GABAA agonists and their biological evaluation through the use of in-line frontal affinity chromatography, Chem. Sci., 2013, 4(2), 764–769 RSC.
  156. L. Guetzoyan, R. J. Ingham, N. Nikbin, J. Rossignol, M. Wolling, M. Baumert, N. Burgess-Brown, C. M. Strain-Damerell, L. Shrestha, P. E. Brennan, O. Fedorov, S. Knapp and S. V. Ley, Machine-assisted synthesis of modulators of the histone reader BRD9 using flow methods of chemistry and frontal affinity chromatography, Med. Chem. Commun., 2014, 5(4), 540–546 RSC.
  157. M. Werner, C. Kuratli, R. E. Martin, R. Hochstrasser, D. Wechsler, T. Enderle, A. I. Alanine and H. Vogel, Seamless Integration of Dose-Response Screening and Flow Chemistry: Efficient Generation of Structure–Activity Relationship Data of β-Secretase (BACE1) Inhibitors, Angew. Chem., Int. Ed., 2014, 53(6), 1704–1708 CrossRef CAS PubMed.
  158. K. P. Cole, J. M. Groh, M. D. Johnson, C. L. Burcham, B. M. Campbell, W. D. Diseroad, M. R. Heller, J. R. Howell, N. J. Kallman, T. M. Koenig, S. A. May, R. D. Miller, D. Mitchell, D. P. Myers, S. S. Myers, J. L. Phillips, C. S. Polster, T. D. White, J. Cashman, D. Hurley, R. Moylan, P. Sheehan, R. D. Spencer, K. Desmond, P. Desmond and O. Gowran, Kilogram-scale prexasertib monolactate monohydrate synthesis under continuous-flow CGMP conditions, Science, 2017, 356(6343), 1144–1150 CrossRef CAS PubMed.
  159. S. Mascia, P. L. Heider, H. Zhang, R. Lakerveld, B. Benyahia, P. I. Barton, R. D. Braatz, C. L. Cooney, J. M. B. Evans, T. F. Jamison, K. F. Jensen, A. S. Myerson and B. L. Trout, End-to-End Continuous Manufacturing of Pharmaceuticals: Integrated Synthesis, Purification, and Final Dosage Formation, Angew. Chem., Int. Ed., 2013, 52(47), 12359–12363 CrossRef CAS PubMed.
  160. A. Adamo, R. L. Beingessner, M. Behnam, J. Chen, T. F. Jamison, K. F. Jensen, J.-C. M. Monbaliu, A. S. Myerson, E. M. Revalor, D. R. Snead, T. Stelzer, N. Weeranoppanant, S. Y. Wong and P. Zhang, On-demand continuous-flow production of pharmaceuticals in a compact, reconfigurable system, Science, 2016, 352(6281), 61–67 CrossRef CAS PubMed.
  161. S. Lee, MIT-CMAC 2nd International Symposium on Continuous Manufacturing of Pharmaceuticals, 26–7 Sept 2016 Search PubMed.

This journal is © The Royal Society of Chemistry 2019