Improving Manufacturing in the

Pharmaceutical Industry through the

Implementation of Continuous Processing

Sarah Hunter, Process Engineer, R&D

Batch Processes have and will continue to fulfill an important role in the manufacture of Active Drug Substance; they allow multiple recipes to be executed in the same core equipment. However, the compromise is, batch fails to deliver the precision, speed of manufacture and intensification of a flow process. Here I present how, at GSK, we are developing flow processes for organometallic chemistry. Using process understanding generated using batch methods, and a combination of insilico simulation and physical demonstration we are able to redefine conventional operating parameters for this type of chemistry and industrialise processes which are not feasible in batch due to risk of scale. We believe this will allow us to deliver more medicines of value to patients, both in affordability and function

1

Flow Chemistry at GSK

Where We Are

2

We want a Manufacturing solution with these attributes...

2003 Started Programme

2016 Gain First Commercial Credibility for 3

Stage Process.

Clinical Campaign of Organometallic Flow

Process Case Study.

2017-2021 Design/Build the Space to house the

next Phase of Projects

Route

selection

Training

New technology

development and

implementation

Process

understanding

Kinetic

experiments

Calorimetry

design

Reactor

library

industrialisation

Analytical

techniques

New

simulation

platform

New

guideline

Strategic

partners

Modular

technology

Process

Simulation

Flow Chemistry at GSK

Pharma has been working on a knowledge and trial and error principal. Development has

been governed by statistical models and small scale increment changes until manufacturing

scale.

GSK is partially converting its API batch portfolio to continuous with a great ambition to reach

up to 50% by 2020.

To support this strategy, simulation has to be integrated at the centre of process development

to;

Accelerate process development

Provide more robustness in the development phase

Ensure and understanding from scale to scale

Batch is following the transition and integrate simulation as part of the workflow.

Where We Aim to Be

3

Organometallic Chemistry

Organometallic chemistry is common place in pharmaceutical synthesis and often presents a challenging

process to execute in batch.

Batch Process

4

E-1

Organometallic Reagent Substrate

Batch is….

Versatile.

Available across

scales.

Familiar to

chemists

Batch is also….

Imprecise.

Hard to scale up.

Labour intensive and

Slow.

Organometallic Chemistry

Flow Process

5

Low Inventory System

Limited

Intermediate

Hold Times

A team at GSK have been developing process workflow for effective design and scale up of

continuous organometallic processes which is currently being applied to a number of

processes in active development.

Minimised

Reagent

Handling

Effective Mass Transfer

How do We Allow the Transition to Happen From Laboratories to Manufacture

6 6

Process Understanding

Industrialisation

Demonstration

Case Study P38 Kinase Inhibition of Inflammatory Pathways

– Rheumatoid Arthritis – Candidate selected 2003

– Atherosclerosis

– Depression

– Neuropathic pain

– Acute Coronary Syndrome - PhIII

– COPD

– Focal Segmental Glomerular Sclerosis

– Stage 1 – Flow Synthesis

Case Study Flow Stage Development

– Evaluation of solubilities, rates and

stabilities required

– Kinetic model to define reactor

characteristics

Flow Batch

0.00

10.00

20.00

30.00

40.00

50.00

60.00

70.00

80.00

90.00

100.00

0 500 1000 1500

%

Time min

Aryl Grignard Stability with Temperature

30°C

40°C

50°C

0.9eq ofiPrMgCl

Process Understanding

Kinetics & Calorimetry

Calorimetry

Deprotonation

Exchange

Kinetics

Main Reaction

Degradation – Batch Time Course

Simulation PFR Model

10

Matching ;

– Temperature profile

– Residence times

– Reagent concentrations & flow rates

Deprotonated

Aryl Grignard

Methyl Grignard

Starting Material

Define process parameter ranges and

the equipment performance;

– Heat Transfer Coefficient

– Surface Area to Volume Ratio

– Micro mixing

– Bodenstein Number

The simulation demonstrates process robustness low sensitivity to parameter inputs allowing

the simplest design concept.

Plug flow reactor with static mixer;

User Requirement Specification

Continuous Process

11

L1 L2

Mixing Zone

τmix<0.5s

Residence Time Zone

τ<0.5s

U ≈ 0 W/m2.K (adiabatic)

Demonstration PFR Model

12

1

2

HXN

3

Demonstration

MeMgCl

feed

Temp

probe

To iPrMgCl

“reactor” – Heat Exchanger

solution feed Temp

probe

Continuous Process

Demonstration

Process Control

– Initial intent: to flow directly into a quench pot containing the trimethyl borate

– Extended addition time gave higher levels of the borinic acid

Work Up

18

Addition time (min) Borinic acid level

HPLC (%a/a)

0.5 1.3

3.5 4.2

8 7.9

Demonstration

Work Up

19

Demonstration

• If k1 and k3 fast but k2 slow?…..

• Rate of addition relative to progression of the “ate” complex controls borinic

levels

• Rapid trimethyl borate addition to the Grignard solution preferable

Simulation Design Space and QbD

1.051.15

1.251.35

1.45

90

92

94

96

98

100

0.00.1

0.2

0.3

0.4

Methyl Grignard charge [eq]

yield [%]

residence time [s]

98-100

96-98

94-96

92-94

90-92

1.05

1.2

1.35

1.5

0

0.01

0.02

0.03

0.04

0.00 0.10 0.20 0.31 0.41Methyl Grignard

charge [eq]Residence time [s]

0.03-0.04

0.02-0.03

0.01-0.02

0-0.01

Impurities [%]

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

1 1.1 1.2 1.3 1.4 1.5

yield

des Bromo

retro Michael

Hexyl Imp

iPr Grignard concentration [mol/L]

Gellingobserved

0.0

0.1

0.2

0.3

0.4

0.5

20 25 30 35 40

yield

des Bromo

retro Michael

Hexyl Imp

CSTR temperature [ C]

Gellingobserved

The simulation is a critical part of

development facilitating;

– Process design

– Risk evaluation

– Specification of deign space

Industrialisation

Scale Independent Design

21

Mixing

Zone

Residence

Time Zone

Scale Up

ν1

ν1

Re

Re

COV

COV

Mixing

Zone

Residence

Time Zone

U

U

Bo

Bo

Industrialisation In Practice

Processes have been demonstrated at production rates of;

• Lab scale ~100g/hr

Isolated Yield 70% th

Isolated Quality 99.7% a/a

• Pilot plant scale ~500g/hr.

Operating as a flow process

• Enabling operating conditions which are not feasible as a

batch operation.

• Eliminating a potential critical quality attribute.

A recent pilot plant campaign

produced 88.2 kg of clinical material.

Yields averaged 84.95% th.

Acknowledgements

Continuous Primary Team

Flavien Susanne, Andrew Rutter, Charles Wade

Project Teams

Robert Whitten, Thoralf Hartwig, Mark Schilling, Lindon Francis, Kathryn Southgate

David Pascoe, Mike Webb, Jason Cooke, Sophie Duffield, Peter Skett, Laura Palmer

23