19

Lyophilizer Qualification:
Some Practical Advice

Thorsten Fischer
Aventis Behring GmbH,
Marburg, Germany

I. INTRODUCTION

The project described here was initiated by a change control request one
year before starting commissioning. The change request identified that an
existing lyophilizer required equipping with a new control system, an
automatic filter integrity test system, and associated new pipework.
Before starting the project the contributing departments of engineering,
production, qualification, and quality assurance, analyzed the qualifica-
tion status of the system. It was found that the implementation of the
defined changes would have continued into a major change of the whole
system. In this start-up meeting it was determined that a new
qualification with the steps IQ (Installation Qualification), OQ (Opera-
tional Qualification), and PQ (Performance Qualification) was required.
Each protocol was to be pre- and post-approved by all the departments
involved.
The qualification exercise included parts of computer and software
validation as well as equipment qualification. To ensure a coherent
qualification it was determined that portions of the computer and software
validation would be subject to an interim approval. This would
allow specified computerized systems validation tasks to be completed
prior to making equipment operational, and subsequently performance
qualification (6,7).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 II. PROCEDURE

The documents described, namely, System Qualification Procedures (SQPs)

and Standard Operating Procedure (SOP) in Figure 1 contain a generic risk
assessment for determination of the validation scope and effort on the
equipment lyophilizer. They were also used as a generic validation master
plan for the project. Management of the project resources, costs, and
deadlines was performed with a model created in a standard project planner.
In advance of the qualification a User Requirement Specification
(URS) was written by the production and engineering departments.
The vendor started detail engineering after approval of the Functional
Specification (FS).
The lyophilizer was installed and thoroughly tested during the
commissioning and start-up phase.
The qualification activities, as set out in Figure 1, were executed with
the support of the vendor and one personnel resource according to Aventis
Behring standard qualification procedures for equipment.
Qualification document development is based upon the use of
templates that are customized for particular systems and equipment. The
SQPs and SOP govern the use of these templates.

A. Installation Qualification
The process for equipment qualification is documented in the Aventis
Behring (AB) ‘‘System Qualification Procedure’’ for lyophilizers.

Figure 1 Creation of qualification protocol.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Upon completion of the installation and the cycle development phases
the IQ phase was initiated. Normally the approach to the IQ phase is similar
for each piece of equipment and would differ only in system-specific details.
The IQ verified the following information (testing items 1 to 5 are
defined in an SQP for the IQ template):
1. Mechanical equipment installation: verification of the authorized
drawings (P&IDs, system schematics, as-built drawings) and
component lists including the verification that utilities are
installed as required.
2. Mechanical component installation: verification of secondary
system components on compliance with specifications.
3. Instruments specifications and calibration: verification of critical
and non-critical instruments.
4. Equipment documentation: verification of the technical docu-
mentation used during qualification on compliance with basic
Good Manufacturing Practice (GMP) rules (6,7).
5. Control system: verification of automation and software parts
by performance of additional testing as described below (testing
5.1 to 5.8 is defined in an SOP for qualification of automated
systems):
5.1. Electrical installation: verification of authorized circuit
diagrams including the component list.
5.2. System specifications and drawings: verification of engi-
neering documents as per user requirements, functional
specification, software design specification, software inte-
gration plan, etc., on compliance with the built system.
5.3. Hardware installation: verification of secondary automa-
tion system components (main components as installed
input/output cards, temperature controller, etc.).
5.4. Wiring termination and pneumatic circuit check: verifica-
tion that the wiring and pneumatic circuit are installed
properly.
5.5. Input/output (I/O) devices: verification that the I/O devices
comply with manufacturers’ specification.
5.6. Configurable switches: verification of the positions of
configurable switches (e.g., Dual In-line Package switches).
5.7. Software installation: verification of each installed software
item.
5.8. Automation system documentation: verification of the
technical documentation, used during qualification, on
compliance with basic GMP rules (6,7).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 While executing the electrical installation testing at 5.1 the testing described
in 5.2 to 5.6 was also performed with the support of the vendor.
6. System specific testing: verification of aspects defined within the
scope of the system qualification procedure for lyophilizers:
6.1. Determination of system equivalence: the equivalence of
the present hard- and software with other comparable
qualified systems was evaluated within the scope of the IQ.
Based on that assessment further additional testing
activities were identified with the support of the automa-
tion group. The evaluation was documented within the
scope of the IQ protocol.
6.2. Chamber door: visual verification of the correct installa-
tion of the chamber door and the door sealing.
6.3. Room separation: verification that the lyophilizer was
correctly installed into the cleanroom wall and no visual
defects were identified in the seals between the lyophilizer
front plate and the cleanroom wall.
6.4. Shelves: verification that all shelves were correctly installed
and showing no visual damage such as a leakage of heat
transfer fluid.

1. Results
In total 16 groups of tests were performed and documented over a 6 week
period. The IQ established that the new and existing equipment had been
adequately installed and met the relevant acceptance criteria.
Several deviations were identified and promptly resolved. Current
documentation requirements meant that the documents for existing parts
of the lyophilizer needed substantial supplements and the creation and
provision of these were the main issue. Other deviations were of such a
minor type that OQ was started without delay and no amendments or
addenda to the IQ were required.
The equipment vendor provided execution support. Test execution
management and elements of technical support were provided by one AB
staff resource. This included solution of deviations, management of
document revisions/updates, and data review.

B. Operational Qualification
As a prerequisite for operational qualification of the equipment it was
required, as part of the Validation Plan, that major parts of the computer

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 and software validation (CSV) be completed. The CSV testing approach
is defined by a separate SOP for Qualification of Automated Systems.
To achieve CSV a plan for testing the function of the control system
(operator panel, switches, power failure test, etc.) and the visualization
program (input, functions, alarm messages, etc.) was prepared and
executed with the support of the vendor. This control and visualization
functionality was provided through an optional personal computer (PC)
connected directly to the control system using a programmable logic
controller (PLC).
The following testing was performed for the computer and software
validation.
1. Radio frequency interference: verification that the control system
has a Certificate of Exportability (CE) that is adequate to show
that the system is shielded for interferences.
2. Operator devices (switches, indicator lights installed at the
control cabinets): verification that each operator interactive
device and indicator light operate as specified.
3. Operator panel: verification of the panel control functions.
4. Security and access: verification that the access is password
protected and is active for each specified level.
5. Visualization system: verification of the accuracy and complete-
ness of each screen (equipment schematics, tables, curves, trending,
etc.) and the control functions (alarm messages etc.).
6. Alarms: verification of all process-relevant alarms and messages.
Testing is performed just below, above, and at the limit.
7. Critical parameters: verification of the programmed critical
parameters against specification (e.g., program cycle parameters
for time, temperature, pressure, alarm).
8. Power failure: verification of the system condition after a power
failure and recovery of the system.
9. Backup: verification that for each software item active on the
system there is a backup copy to restore the system.
Further automation testing was defined within the scope of the system-
specific qualification procedure and included the following:
10. Network connection: the functioning of the network card was
verified. The system boundary ended at the bus connection.
Network qualification was performed in a separate protocol and
was executed by an automation specialist.
11. Chart recorder: verification of the correct color, description, and
range for each channel.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 12. Recipes: verification that each automated cycle was programmed
within its limits in the form of a recipe (lyophilization,
sterilization, etc.) and complies with the specification. This test
is an additional system-specific test for critical parameters (see 7
above).

1. Interim Result
Tests 1 to 12 were executed in a 2 week period. As noted above, interim
approval was required and obtained before the further equipment quali-
fication could continue.

2. Equipment Operational Qualification

Operational equipment testing was performed for the vacuum, temperature,
and ventilation system, the stoppering function, and the sterilization
process. The qualification of the lyophilizer’s Sterilization-in-Place (SIP)
system is performed similar to an autoclave and uses biological indicators
and accumulated lethality (F0) determination (2,5).
Additional testing as part of the OQ included:
13. Program cycles: verification of each programmed cycle to ensure
that no further cycle development was required; testing was
performed by running each programmed cycle (one recipe for
lyophilization, SIP cycle, filter integrity test, etc.). The system
qualification procedure describes the identification of the most
appropriate lyophilization cycle for this test. Testing of the
lyophilization cycle for this step is combined with test 18.
14. Vacuum pumps: verification that each pump system is able to
reach the lowest pressure, defined in any lyophilization recipe.
15. Vacuum leakage test for chamber and condenser: verification of
the system integrity, through a leakage rate test. A leakage rate of
less than 102 mbar L/s was chosen.
The leakage rate is determined in an empty lyophilizer with
running of the condenser cooling. Before starting the test phase
it is recommended to evacuate the lyophilizer (chamber and
condenser) several hours (e.g., 4 h) to remove the gas (e.g., below
102 mbar). After stopping the vacuum pumps the test phase
begins when pressure is stabilized and has reached 102 mbar.
Sometimes virtual leaks might cause a pressure increase
directly after stopping the pumps; therefore a stabilization
phase is recommended, before starting the test. We have
observed that in the range of 102 to 101 mbar a representative

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 proportional pressure increase can be monitored with a
conductivity detector (vacuum meter based on the Pirani
principle (1)). For the test procedure see also Figure 2.
16. Silicon oil pumps: verification of the heat transfer media recycle
pump redundancy to ensure that the pumps can replace each
other if one of them fails. The test was performed by monitoring
the pressure in the pipe before and after failure of a pump.
17. Condenser chilling test: verification of the system cooling rate,
normally 2
C/min, by monitoring the temperature at the
condensing coils during cooling from 0
C to 60
C (1).
18. Ice capacity: verification that the condenser was able to hold an
ice capacity representing a worst-case condenser load. The test is
performed with the shortest lyophilization recipe that can be used
to load the freeze-dryer with the vial format giving the worst-case
load (maximum water to sublime).
19. Automatic stoppering: verification that the lyophilizer can close
the vials under vacuum conditions with stoppers. Closure of the
vials has been verified with a vacuum tester.
20. Shelf temperature distribution: verification of the temperature
homogeneity across all shelves at one time. The shelf temperature
mapping is by use of a large number of thermocouples (in our
case, five per shelf each, e.g., 70 thermocouples per measurement
event), one of the most complex, time-consuming, and error-
prone measurements in equipment qualification. It requires an
experienced protocol executer for calibration, proper placement

Figure 2 Scheme of vacuum leak rate test procedure.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 and examination of the thermocouples and their monitored data.
Monitoring of the temperatures was recorded with a calibrated
data acquisition system, temperature reference bath, Resistance
Temperature Detector (RTD), and type K thermocouple. The
thermocouples were calibrated before measurement and rever-
ified after the measurement. Use of an automated system is
recommended. As the flange is key to a successful qualification it
should be made part of the User Requirement that the vendor
provide, and show, correct operation of a special vacuum and
pressure (for sterilization) resistant flange–thermocouple system.
To verify the temperature profile the thermocouple should have
good contact with the shelf surface. By the use of lab lifts and a
compressible block of isolation material a close surface contact
can be provided. See Figure 3.
The temperatures were monitored at the worst-case tempera-
tures of all recipes (e.g., at –45
C, þ10
C, þ60
C) according a
qualification recipe. See Figure 4.
Documentation of the heat transfer from the silicon oil via the
shelf to the surface the temperature probes at the silicon oil inlet
and outlet pipe were considered an important operational
characteristic. This was recorded with the installed calibrated
temperature probes.
Results of the measurements are shown in Figure 5: Graph
A shows the total run and graphs B, C, and D give a focus at the
temperature hold phases.
Figure 6 shows in graph A the results of the measurements
across the top shelf and graph B across the bottom shelf at
minimum temperature. In graph A one thermocouple shows
the effect of an insufficient contact to the shelf. This effect is
not visible under vacuum conditions because the compressible
isolation material is enlarging its volume under the vacuum
and pushing the thermocouple closer to the surface of the shelf.
21. Air ventilation system: verification that the air filter can be
automatically tested for integrity. Modern lyophilizers are
equipped with an automatic filter–testing unit that verifies the
filter integrity by either the water intrusion or the forward flow
test method.
22. Sterilization of chamber and condenser: verification of the eight
coldest spots within the chamber. These points were then to be
used for execution of the performance qualification. Testing
was performed using methods similar to those used for an
autoclave (2,4,5).

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Figure 3 Thermocouple placement for the shelf temperature distribution studies.

Figure 4 Scheme of the qualification recipe.

The thermocouples:
were positioned throughout the chamber and within the anticipated
hot and cold regions, at the steam inlet (air filter downstream side),
and at the condensate drain;

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 were held in positions with heat- and water-resistant tape;
were positioned at the extremities of the chamber and the condenser,
as close as possible to the equipment surface;
had two-point calibration with a midpoint verification prior to the
measurement;
were subjected to midpoint verification after the measurement.

Figure 5 Results of the shelf temperature distribution study.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Figure 5 Continued

Due to the age of the equipment and the fact that the shelves are pushed
from the bottom to the top, the condenser could not be entered for
measurements in the condenser. An acceptable compromise in this case may
be the evaluation of the installed condenser drain temperature and
condenser pressure probes to ensure that they accurately measure saturated
steam conditions.
For new lyophilizers it is recommended that the condenser is
positioned either below the chamber or can be entered at minimum by a

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Figure 6 Temperature distribution across the top and the bottom shelves at
minimum temperature.

man-hole. For this case, as part of the health and safety assessment for
equipment entry, the following points should be considered:
ventilation of the condenser during installation of the thermocouples;
isolation or disconnection of any heating and cooling media;
at minimum, one extra person for safety;
extra lighting.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Because of the short time scheduled for the qualification the empty chamber
temperature distribution study was combined with the empty chamber
temperature penetration study. These studies differ only in the usage of
biological indicator (BI), an F0 (accumulated lethality during the SIP hold
period) calculation, and the number of thermocouples. By using the same
amount of BIs as thermocouples, the studies could be performed together.
A separate performance qualification was established as unnecessary.
This approach was documented in the qualification plan.
The testing verified the following items:
Temperatures above or equal to 121
C are recorded during the SIP
hold period at all times (see Fig. 7).
Measured temperatures and pressures complied with saturated steam
conditions. These were confirmed by an external data acquisition
system (external pressure and temperature sensor) (see Fig. 8).
Homogeneous temperature profile: all temperatures measured inside
the chamber were within a range of 2
C (see Fig. 7).
Overkill approach provided by the SIP cycle. For the verification the
BIs had to fulfill the criteria described below.
Selection of an Adequate BI of Bacillus stearothermophilus for Determination
of an Effective Overkill (Biological Evidence). Targets:
1. (a) Verification of a 12 log reduction.
(b) Use of an adequate BI equivalent to a specific theoretical
biological lethality (Fbio theo), i.e., Fbio theo 12 min.
Calculation is performed with the certified population
(N0), the population at the end of the SIP cycle (N), and the
certified decimal reduction time (D) according to the
formula.

F ¼ ðlog N0  log N ÞDZ Z

T ¼ nDT

To calculate an ideal BI, based on an overkill requiring a 12 log

reduction,

F ¼ ðlog N0  log NÞDzT

¼ ðlog 1012  log 108Þ1:0 min ¼ 12  1:0 min ¼ 12 min

Coherence is valid for the lethality F. F ¼ 12 describes the

criteria to be passed before using the BIs. An example for a
purchasable BI is as follows:
 
Fbio, theo ¼ 2:2 min log 106  log 100 ¼ 13:2 min 12 min

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 2. A vendor certificate with certified population N and D values is
available.
3. Verification of the biological lethality (Fbio) by use of the BI with
a specific N and D. Fbio for determination of the minimum F0
value is Fbio 15 min. The basis for Fbio is (3)

ATUN BI
 BI   
Fbio ¼ F121:1
C ¼ D121:1
C log N0  log NFATUN

where:
NATUN
F ¼ realistic end population after sterilization, if all used
BIs show no growth (all test units are negative
(ATUN)) (3)
NBI
0 ¼ certified starting population of the BI
   
Model :Fbio ¼ 1:0 min log 1  1012  log 1  103 ¼ 15 min
   
Example :Fbio ¼ 2:2 min log 1  106  log 1  103 ¼ 19:8 min

To ensure that conducting the BI and thermographic tests in

parallel was acceptable it was essential that each of the three
criteria be passed, before using the BI. Thus Fbio has to be
reached by or exceed the thermographic calculated F0 value. The
calculation follows following formula (5):
X
F0 ¼ t ½10 expðT  Tb Þ=Z 

where:
T ¼ instantaneous temperature
Tb ¼ base temperature (121.1
C)
Z ¼ value or temperature coefficient (10
C)
t ¼ time interval
Testing typically is performed multiple times. In the case of the initial
qualification of the sterilization the testing was conducted in triplicate.

3. Results
In total 22 tests were performed for execution of the operational qualifi-
cation in a total time of 8 weeks. The execution required one full-time and
one part-time person. Much of the work was conducted by supervised
contractors who had received specific training by AB staff.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 Figure 7 Theoretical sterilization phase.

Figure 8 Graphs showing the results of the measurements.

Only minor deviations, relating to control system hardware installa-

tion and operational configuration backup, were identified and these were
promptly resolved. The support equipment for temperature mapping and BI
usage operated well and valuable lessons on probe placement, equipment
sealing, and data investigation were learned.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 C. Final Report
It was recognized at the beginning of the work that the testing would be
complicated and to make the results readily understandable to a reviewer a
suitable Final Report (FR) structure would be required. To ease the
workflow the summaries were generated for each set of test results
documented and used as attachments to the FR.
A proper FR also enables a quick review of the system history, and
what eventually might be necessary for periodic requalification of the
system.

Figure 9 Project flowchart.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.

 D. Overall Conclusion
The engineering phase of the project for decommissioning and removal of
the old equipment and installation of the new equipment including
commissioning and start-up testing on site took 8 weeks. This was followed
by the qualification activities in a time of 14 weeks.
After an additional 4 weeks for process validation the upgraded
system was completely validated and could be used for routine production.
A project flowchart is shown in Figure 9.
The main difficulties encountered in the execution of the project
concerned the old documentation of the existing system. One principal
recommendation of the project team members, including the vendors, was
that review of more focused documentation should be completed during the
commissioning and start-up phase to ensure that what is installed is properly
documented.
For future work it was established that a clear understanding of roles
and responsibilities would smooth progress and ensure that vendor and
owner team members better understood other participants’ problems and
needs.

ACKNOWLEDGMENTS

The author would like to acknowledge the cooperation and support of

colleagues at Aventis Behring GmbH, Marburg. In particular Mr. Mathias
Klein, Director of Facility Qualification, for his council and direction and
Mr. Maurice Shakeshaft of CB Automation Ltd, who assisted in the
preparation of the English version of this chapter.

BIBLIOGRAPHY

1. Oetjen, Georg-Wilhelm. Gefriertrocknen. Weinheim: VCH Verlag, 1997.

2. FDA, Guide to Inspections of Lyophilization of Parenterals, 15.02.01.
3. IJ Pflug, KD Evans. PDA Journal of Pharmaceutical Technology 54(2) (March–
April), 2000.
4. DIN, Sterilisation Desinfektion Sterilgutversorgung, 1988, Beuth, 2nd ed.
5. PDA, Validation of Steam Sterilization, PDA Technical Report No. 1-Revision,
Draft No. 5, January 1999.
6. FDA, 21 CFR Parts 210 and 211, Current Good Manufacturing Practices for
Finished Pharmaceuticals.
7. Auterhoff, Gert. EG-Leitfaden einer Guten Herstellungspraxis für Arzneimittel.
5th ed. Aulendorf: ECV Verlag, 1998.

Copyright 2004 by Marcel Dekker, Inc. All Rights Reserved.