Tools and Techniques for Assessment of Containment
And Exposure in The Pharmaceutical Industry
Matthew J. Meiners, CIH
Division Manager
Bureau Veritas North America
2
Containment – Functional Definition
There are three basic elements to pharmaceutical
manufacturing…
► Product
► People
► Facilities and surrounding
environment
For our purposes, containment
is the isolation* of the first of these
(product) from the other two.
3
Why Contain Pharmaceutical Processes?
► Protect workers, facilities and environment from product. Though considered a chemical hazard, pharmaceuticals are unique - they are intended to have biological effects at even very low levels.
► Protect product from the environment
► Prevent production losses of product
4
Containment… To What Level?
Containment (isolation*) is not an absolute, and must be further defined.
► “Zero” can never be assumed, and is impossible to measure.
► Thresholds limiting product migration to the other two elements (people and environment) must be defined.
► Limits, sometimes called “Containment Performance Targets” (CPT), are rationalized and set.
5
Containment Strategies
Containment strategies (equipment and procedures) are selected based upon the “CPT”, with consideration to the underlying process.
Strategy 1: Controlled general ventilation
Strategy 2: Localized capture of airborne particles
Strategy 3: Barrier isolation
Strategy 4: Barrier isolation with closed transfer
Strategy 5: Complete isolation with remote process control.
6
The Containment Challenge….
A paraphrased quote – source unknown…
“A required containment level of 5 nanograms/m3 is the equivalent of one grain of pollen in an average sized living room.”
Containment to and testing against such criteria present significant challenge.
Techniques for Assessing Containment Performance
8
Techniques for Assessing Containment Performance
► Parametric testing
► Visualization
► Tracer gases
► Functional testing with surrogates
► Live process testing (direct analysis for fugitive APIs)
9
Techniques for Assessing Containment Performance
Parametric Testing
► Examples – face velocity (open system) and pressure drop.
► Useful for indicating air handling equipment is operating within specification
► Weakness - Does not necessarily relate to effectiveness of the system
10
Techniques for Assessing Containment Performance
Visualization
► Examples – smoke tubes and candles
► Useful for seeing flow patterns, air movement/direction and can be used to visualize barrier isolation
► Weakness - qualitative
11
Techniques for Assessing Containment Performance
Tracer Gases
► Detectable vapor metered into the system and collected at critical points where containment may be lost
► Examples – ASHREA 110 (SF6), Freon, Cinnamon.
► Can be performed quantitatively, with high sensitivity
► Weakness – Vapors behave differently than aerosols: they disperse evenly and are drawn away from surfaces. Particles settle and create a source of recontamination.
12
Techniques for Assessing Containment Performance
Direct Testing During Manufacturing
► The most accurate measure of containment performance.
► Requires that a sampling and analytical method of adequate sensitivity for the API has been developed, validated and is available.
► Weakness – If relied upon as the initial measure of containment, and performance does not meet specification (CPT), it can result in unacceptable and preventable employee API exposures and facility contamination.
13
Techniques for Assessing Containment Performance
Functional Testing with Surrogates
Substitution of process API with another solid material of similar characteristics, low toxicity, with a sensitive assay.
► Useful in that it allows for equipment/procedural testing of
aerosol containment performance prior to introduction of
API material.
► Can be performed quantitatively, with high sensitivity, and
short of actual sampling of API during manufacturing, is the
best indicator of containment performance
► Sensitive methods are defined and available.
► Can provide “challenge” testing for multi-use equipment
► Provides “standardized testing” for comparison of relative
performance
Surrogate Testing: When and How?
15
Factory Acceptance Testing (FAT)
Standardized performance testing of equipment prior to installation at the site.
16
Site Acceptance Testing (SAT)
The evaluation of newly installed containment equipment and procedures at the site, prior to
introduction of actual process API.
17
Operational Qualification (OQ)
Performance testing of in-use equipment and procedures
18
Elements of a Successful Surrogate Evaluation
#1 - Establish Containment Target (CPT)
► Driven by IH, toxicology, QA or engineering specification
► Often based upon OEL or other health limits
► Consideration of PPE is incorporated if specified
► Incorporation of safety factors is common
19
Elements of a Successful Surrogate Evaluation
#2 - Determine the Means to Simulate the Operation
► Testing facility configuration – goal is to simulate “in-use process train”, incorporating all critical steps in the process
► Multiple operators, assuming sub-optimal conditions
► Don’t rely on a single test – several runs or operations should be tested.
20
Elements of a Successful Surrogate Evaluation
#3 - Selection of Surrogate Material
► No perfect surrogate material, but there are standardized materials – the best choice will depend upon the particular manufacturing process and the situational priority of a variety of variables.
21
Elements of a Successful Surrogate Evaluation
#4 - Develop a Sampling Plan
► Area sample at critical containment points
► Operator exposure sampling
► Task sampling – make sure task duration provides adequate sensitivity for meaningful results!
More sensitive methods (example: LC/MS/MS application vs.. HPLC)
Higher flow rates
Longer sample times (multiple iterations for tasks)
► Consider surface sampling if it adds value
22
Elements of a Successful Surrogate Evaluation
#5 - Perform Simulation and Sampling
23
Elements of a Successful Surrogate Evaluation
#6 - Interpretation of Results
► Report should describe the testing method in detail, with objective interpretation of result. Pictures are helpful.
► Report should demonstrate if the CPT specification is being met
Factors in Which Drive Surrogate Selection
25
Surrogate Selection
Meets Requirements for the Manufacturing Process
► Flowability (bulk powder movement)
► Compactability and adhesion (tableting)
► Solubility (emulsion vs. solution)
26
Surrogate Selection
Relative Hazard or Toxicity of Test (surrogate) Material
► OEL or other health limit
► Process safety concerns
27
Surrogate Selection
Sensitivity of the Sampling and Analytical Method
► Internal or Commercial Lab
► Analytical sensitivity
► Air sampling flow rate
► Duration of task or test
28
Surrogate Selection
Selectivity of the Analytical Method
► Possible interferences in the test environment
► Possible environmental background of surrogate material
29
Surrogate Selection
Quality Concerns
► API vs. non-API
► Cleanability
► Established cleaning procedures
30
Surrogate Selection
Test Material Availability and Cost
► How much test material is needed?
► Cost of bulk material
► “Off the shelf” vs. Custom manufacture (micronization)
31
Surrogate Selection
Appropriate “Containment Challenge”
► “Dustiness” - defined as the propensity of a material to emit dust during manufacturing: may be considered analogous to vapor pressure on a molecular scale.
► Flour vs. Sugar
► Though studied in a variety of ways (gravity dispersion, mechanical dispersion and gas dispersion), a standardized measure of “dustiness” has not been established.
► A variety of individual particle characteristics effect relative dustiness.
32
Surrogate Selection
Factors Effecting the Dustiness of Bulk Powders
► Powder Mass/Bulk Density
► Moisture Content
► Particle size and size distribution
► Particle density
► Flowability and Cohesion
► Particle Shape
33
Material Characteristics Relative to Dustiness
Powder Mass and Bulk Density
► The effect of sample mass on dust generation demonstrates an initial increase in dustiness with increase in mass, followed by a decrease after a critical point is reached.
► The observed decrease in dustiness index with sample weight is a function of both the dust generator itself and a decrease in the volume of entrained air.
► Conclusion: the measuring instrument, process being evaluated and containment system itself can affect “dustiness” index of a bulk powder.
34
Material Characteristics Relative to Dustiness
Moisture Content and Humidity
► Generally, an increase in moisture content for bulk powders results in decreased dustiness.
► Mechanism – Moisture content increased the cohesion of powders as well as particle weight.
► Generally, dustiness indices are highest for dried powders, tested at low humidity.
► Spray dried lactose is an important exception – it absorbs less than 2% moisture up to 75% relative humidity with little change in measured dustiness compared to its respective dried powder.
35
Material Characteristics Relative to Dustiness
Particle Size and Size Distribution
► Characteristic with the greatest
impact on powder dustiness
► Generally, decrease in particle size
results in increase in dustiness
► Powders blended to contain only a
portion of small particles were found to
be nearly as dusty as powders
comprised entirely of small particles
► For most materials, as the proportion of
small particles in a blend diminishes, the
faction of small particles in the generated dust increases.
36
Material Characteristics Relative to Dustiness
True Density
► Particle size, not density, governs the motion of aerosol powders
► Studies have found no clear relationship between density and dustiness
37
Material Characteristics Relative to Dustiness
Flowability and Cohesion
► Measured by Angle-of-Repose
► Subjective
► Found to correlate poorly with dustiness measurements
► No satisfactory relationship has been found between powder cohesion and powder aerosolization.
38
Material Characteristics Relative to Dustiness
Particle Shape
► Important in it’s relation to particle size measurements
► Shape factors (aspect ratio, circularitry, elongation ratio, …) do not directly correlate with dustiness indices of powders
► Materials with significant elongation ratios such as acetaminophen cause correlation problems.
► Pujara (Abbott) has found a good relationship between “empirical shape coefficient” and dustiness by incorporating an elongation ratio into the equation.
39
Material Characteristics – Conclusion
Dustiness of Surrogate Material to API
► A variety of factors affect the “dustiness” of bulk powders
► The best way to determine the relative “dustiness” of surrogate to API would be by direct measurement of the bulk powders by the same technique.
► In the absence of dustiness measurements, the best means of selecting a representative surrogate is by mean particle size, with similar distribution.
Common Process Surrogates
41
Common Process Surrogates
The most commonly used process surrogates in the Pharmaceutical and Containment Industry
► Lactose
► Naproxen Sodium
► Mannitol
► Acetaminophen
► Riboflavin
► Sucrose
► Arizona Road Dust
► Insulin
42
Surrogate Selection Criteria – Relative toxicity
Surrogate
► Lactose
► Naproxen Sodium
► Mannitol
► Acetaminophen
► Riboflavin
► Sucrose
► Arizona Road Dust
► Insulin
Typical OEL
► 10 mg/m3
► 2.5 mg/m3
► 10 mg/m3
► 3 mg/m3
► 5 mg/m3
► 10 mg/m3
► 3-10 mg/m3
► .003-0.3 mg/m3
43
Surrogate Selection Criteria – Analytical Method
Surrogate
► Lactose
► Naproxen Sodium
► Mannitol
► Acetaminophen
► Riboflavin
► Sucrose
► Arizona Road Dust
► Insulin
BVNA LOQ Mass/Sample
► 2.5 ng*
► 0.2 ng
► 1.0 ng
► 0.5 ng*
► 5 ng
► 5 ng
► 10,000 ng
► NA
Method Selectivity
► Good
► Very Good
► Good
► Exceptional
► Good
► Good
► Poor
► Exceptional
44
Surrogate Selection Criteria – Quality Concerns
Surrogate
► Lactose
► Naproxen Sodium
► Mannitol
► Acetaminophen
► Riboflavin
► Sucrose
► Arizona Road Dust
► Insulin
Type
► Excipient
► API
► Excipient
► API
► API
► Excipient
► Dirt
► API
Cleanability
► Highly Soluble
► Highly Soluble
► Highly Soluble
► Modestly Soluble
► Sparingly Soluble
► Highly Soluble
► Insoluble
► Poorly Soluble
45
Surrogate Selection Criteria – Availability and Cost
Surrogate
► Lactose
► Naproxen Sodium
► Mannitol
► Acetaminophen
► Riboflavin
► Sucrose
► Arizona Road Dust
► Insulin
Relative Cost
► 1
► 100-1000
► 5
► 5-10
► 15
► 5
► 10-20
► Very high
Multiple Specification
► Yes – wide variety
► Custom
► Yes – variety
► Custom
► Custom
► Yes – variety
► Yes – variety
► NA
46
Surrogate Selection Criteria – Summing it up…
Selection of the best surrogate material for your application depends upon…
► which criteria are your drivers
► particular surrogate characteristics that may also drive the selection process…
47
Lactose
Selection Drivers ► Least expensive in bulk, wide
variety of specifications
(particle size distribution)
available “off the shelf”
► Low hazard
► Minimal quality concerns,
easily cleanable
► Sensitive and specific
method available (2.5
ng/sample)
► Specified surrogate in ISPE
guidance document
► Dustiness independent of
humidity and moisture
content
Potential Concerns ► Common excipient – may be
used in the test formulation
or facility may have
background levels.
► International shipments can
be complicated by bovine
product import concerns
► May lose specification upon
storage of micronized
material
► Complicated analytical
method limits options for
analysis
48
Naproxen Sodium
Selection Drivers
► Highest sensitivity (0.2
ng/sample)
► Easily cleanable
► Uncomplicated analysis
increases availability of
method
► Widely accepted
surrogate, first alternative
in ISPE Guidance
Document
► Good containment
challenge (high dustiness
index)
Potential Concerns
► High relative material cost
if handling in bulk
► API introduction may
cause quality concerns
► Custom sieving or
micronization needed for
specific size distribution
49
Mannitol and Sucrose
Selection Drivers
► Low cost in bulk, variety of
specifications (particle size
distribution) available
► Low hazard
► Minimal quality concerns,
easily cleanable
► Sensitive and specific
method available (Mannitol
is 1.0 ng/sample)
► Good alternative if Lactose
is present in formulation or
facility
Potential Concerns
► Common excipients – may
be used in the test
formulation or facility may
have background levels.
► Complicated analytical
method limits options for
analysis
50
Acetaminophen (Paracetamol)
Selection Drivers
► Relatively low cost API in
bulk
► Highly sensitive and
specific method available
(0.5 ng/sample)
► Good surrogate for
tableting or tablet handling
activities.
► Was one of the first
commonly used surrogate
materials (EU)
Potential Concerns
► API introduction may
cause quality concerns
► More difficult to clean
► Custom sieving or
micronization needed for
specific size distribution
► Unusual elongation ratio
for crystals
► High cost and limited
options for analysis
(LC/MS/MS)
51
Riboflavin (Sodium Acetate)
Selection Drivers
► Common use for CIP
coverage testing may
eliminate quality concerns
► Surface contamination can
be visualized by black light
Potential Concerns
► API introduction may
cause quality concerns
► More difficult to clean
► Custom sieving or
micronization needed for
specific size distribution
► Low level environmental
background common
52
Standardizing Surrogate Testing
ISPE Guidance Document
► “SMEPAC” product
► Adopted and published by ISPE in 2005
► “Good Practice Guide”
► FAT and SAT – intended to provide standard testing and comparison of relative performance and against common criteria
► Updated (Second Edition) in 2012
53
The Next Step – Ongoing Containment Verification
► FAT and SAT testing is performed in order to predict with some degree of certainty, that CPTs and worker exposures are not exceeding limits.
► FAT and SAT testing is performed on newly installed equipment, in optimal working conditions.
► Performance of containment equipment can degrade or fail over time.
► Ongoing evaluations of containment and employee exposures during routine operations is the only means to be confident that containment is maintained, and potent materials are handled safely!
Process Sampling and Analytical Methods
For Ongoing Containment Verification
55
Introduction
A control specification has been established (CPT).
Are process controls meeting the specification?
→Over full process duration
→Tasks and short duration activities
→Consideration of PPE
The above question should be answered definitively, and control within the specification should be demonstrated on an
ongoing basis.
This requires a sampling and analytical method with adequate sensitivity, accuracy, precision and specificity.
56
Sampling & Analytical Methods - Validation
Methods need to be verified prior to use -“Validation”
Validation Elements:
► Analytical method (ICH Guidelines)
► Collection media selection
► Collection efficiency
► Recovery from collection media
► Stability of collected samples
► Robustness and Ruggedness
57
Validation Spike Prep - Standard Model
► Results of experiments are variable based upon mass level
► Validation must therefore be performed at critical mass levels
► Highly Potent API methods must be validated at very low mass loadings
► Only practical means of low mass spike generation is by direct application of dilute API in solvent
58
Solution Deposition – the standard model
Airborne Particles are partially protected
Surface area increased, material exposed.
59
N
O
F
CH3
NCH3
Case Study - Escitalopram
OEL = 30 mg/m3
0.1x OEL 15 minutes = 90 ng
0.1x OEL 30 minutes = 180 ng
Unacceptable Collection Efficiency using standard Model
60
Initial Escitalopram Experiments with Liquid Spiking
Experiment:
Capped in dark 1 day: 96.3%
Capped in light 1 day: 99.6%
Uncapped 1 day: 58.7%
4 hour air draw: 47.6%
4 hour air draw uncapped 1 day: 23.3%
4 hour air draw uncapped 2 days: 10.8%
4 hour air draw uncapped 7 days: 1.2%
4 hour air draw capped 7 days: 48.4%
Conclusions:
Not light sensitive
Air related
Low concentration contaminant
Contaminant depleted quickly
61
Test Atmosphere
0
100,000
200,000
300,000
400,000
500,000
600,000
700,000
800,000
N2 O2 Ar CO2 Ne He CH4 Kr H2 N2O CO Xe O3
pp
m
62
Possible Culprits
- Nitrogen
- Oxygen
- Argon
- Carbon Dioxide
- Neon
- Helium
- Methane
- Krypton
- Hydrogen
- Nitrous Oxide
- Carbon Monoxide
- Xenon
- Ozone
Concentration
Inert
Most Suspicious
63
The prime suspect: Ozone
- Extremely powerful oxidizing agent
- Half-life of about half an hour at normal atmospheric conditions
- Branches of exploration – solid deposition, susceptible classes,
and prevention.
64
Some Methods Fail to Validate…
“Essentially, all models are wrong… but some are useful”
George E.P Box, Statistician
65
Aerosol Deposition – Actual Sample Collection
Deposition of particles - ng/sample - practicality
66
A New Model - Low Mass Aerosol Spikes
67
Collection onto the Filter
68
Total Dust Sampling vs Cyclones for Deposition?
► 37mm total dust cassette
- Rapid deposition
- Minimal size discrimination
► Cyclone
- Improved precision of mass deposition
- Smaller particles allows for lower mass
- Longer entrainment
- Worst case for degradation and a variable tool for degradation studies
69
Extraction of Small Particles using Cylone
Cyclone
70
Degradation Inversely Related to Particle Size
Grit Pot Filter
71
Comparison of Liquid vs Aerosol Models:
Escitalopram Spikes After Air Challenge
ug recovered ug recovered
R1 0.128 L1 0.126
R2 0.126 L2 0.103
R3 0.108 L3 0.129
R4 0.087 L4 0.114
DE avg. 0.120 Air avg. 0.107
0.0110 0.0192
%RSD 9.21 %RSD 17.9
89.3% Recovery for Aerosol Deposition
15% Recovery for Liquid Deposition
72
Escitalopram Results: Nominal 120 ng/sample
120 ng STD
Immediate Extraction.
Extract after 8 hours
Aerosol spike after 8 hours of air sampling
Liquid spike after 8 hours air sampling
73
A second example…
Antifungal Echinocandin
mg recovered mg recovered
R1 0.1068 L1 0.0826
R2 0.0846 L2 0.0892
R3 0.0874 L3 0.0806
R4 0.0804 L4 0.0934
DE avg. 0.09052 Air avg. 0.0841
0.0103 0.0045
%RSD 11.33 %RSD 5.35
Aerosol Deposition: 92.9% retained following 8 hours air sampling
Liquid Deposition: 20.9% remains following 8 hours air sampling
74
Degradation –Cannon Ball vs Grape Shot Analogy
75
SEM Images – Aerosol vs Liquid Deposition
76
Surface Area and Degradation
Penetration to a depth of 10 nm (0.01 mm)
A 1 mm particle would have 5.9% degradation A 100 mm particle would have 0.06% degradation
0
100,000
200,000
300,000
400,000
500,000
600,000
0 20 40 60 80 100
Mass
Particle Diameter (micrometers)
Inner Volume
Surface Volume
77
Functionality Attributed to Observed Degradation
CH3
CH3
NH
N
F
CH3
CH3
N
ON
78
If Validation by Aerosol Deposition Fails…
Ionic Gel Coated Sampling Filters
► Designed to both chemically and physically protect collected aerosols from oxidation during air sampling
► Aerosol particulates impact on gel during sampling, absorbed into the gel.
► Employed for a growing number of BVNA API applications
► Limitations:
Can’t be used for hydrolysis susceptible APIs
New, proprietary technology – limited availability
79
Questions??
Matthew J. Meiners, CIH
(847) 726.3720
If you would like more information, or
have any questions, contact me.
THANK YOU!