Introduction: This case study describes a validation assignment carried out by a contractor for a client. In addition to discussing the actual assignment and how it was handled, the article includes sections that generally talk about the legislation issues involved in the assignment.
A pharmaceutical company (the client) has purchased a blister-packaging machine, which has not yet been delivered. The intention is to use the machine for packaging of the client’s own non-drug products, but preferably also for packaging of tablets under contract for another pharmaceutical company. The specific tablets fall under the category of drugs, and the other company has stipulated that the machine, in addition to qualification and validation, under- goes cleaning validation. After completion of the cleaning validation, the client would then be audited with a view to entering into a con-tract for packaging of the specific tablets.
The client has no experience with cleaning validation and cannot spare the resources for the task due to the tight time schedule. A contractor is hired to do the job. The analysis part of the cleaning validation was carried out by a laboratory under contract, and is not included in the assignment. Therefore, the analysis aspects are discussed to the extent only that they relate to the assignment. The packaging machine was to be delivered from Italy and was still being constructed when the cleaning validation assignment was started.
21 CFR part 211.34 1 includes a description of contractor work detailing that the client must ensure that contractors are qualified for carrying out the specific task and that they keep records. The FDA will hold the client responsible for all work performed by contractors and will inspect these operations through the client. Furthermore, the FDA will inspect and approve contractor sites for manufacturing, packaging, testing, and holding. EU’s GMP 2 does not describe anything about contractors. The section dealing with QA responsibilities, states that QA has the overall responsibility, which means that the client’s QA function must approve validation performed by the contractor.
The Danish Medicines Agency will, like the FDA, hold the client responsible for all work performed by contractors and will inspect these operations through the client. Also, the Danish Medicines Agency has the option of inspecting the contractor sites for manufacturing, packaging, testing, and holding. Consequently, the contractor cannot be directly responsible for a validation assignment performed. However, the contractor is under obligation as stated in the agreement entered into with the client. And furthermore, contractors can only be interested in performing the job to the clients’ satisfaction. If they do not perform, they will soon lose their clients.
The client has chosen to purchase two identical sets of machine parts (those in direct contact with the product) to be used exclusively for packaging of the specific tablets. The machine parts are thus classified as dedicated equipment. Use of dedicated equipment is normally recommended for substances that are difficult to remove, equipment that is difficult to clean, or products with a high safety risk, e.g., products of a high potency, which may be difficult to detect below an acceptable limit. 3 In this case, dedicated equipment is used to avoid a cleaning validation involving several non-drug products, and thus, a validation matrix that will be difficult to handle. The client chose a simpler and probably also less expensive solution by purchasing dedicated equipment for the drug products. The two sets are labeled so that it is easy to distinguish between the two and to tell them apart from other machine parts in connection with future repairs and adjustments. The two sets are used randomly during cleaning validation, but care must be taken that the validation does indeed include both sets.
Preconditions of a Satisfactory Cooperation
The criteria for a well-run project and for good teamwork are a solid project agreement. Both the overall framework and the details must be in place. Who must do what? Who is responsible for which step of the process? The clients often underestimate the time they must spend on preparing comments, approval activities, and making decisions in cases when the preconditions or other aspects are changed. The overall and basic issues are often covered by the contract or the task specification, which is a brief description of the assignment, who delivers what, specification of price and deadline. Establishing all the details during a meeting with several participants is highly recommended, particularly with participants from the client, as the client is often represented by a proportionally higher number of functions although on a smaller scale.
Comprehensive and approved minutes of this meeting constitute the detailed project agreement. The following was agreed and stipulated in the contract: the contractor was to provide and deliver a prepared protocol including rationales behind the validation concept/worst case. The contractor was furthermore to deliver a test plan, execute the tests, prepare a validation report, and prepare final SOP for cleaning of the machine. The client was to provide the following: a validated method of analysis including sampling method, method of analysis for the detergent, execution of analyses, operator help for running the machine, product to make it possible to run campaigns that reflect the daily production, and resources for commenting on and approving protocol and report.
During the first meeting after the contract was in place, the following distribution of responsibilities was arranged: The client was to supply a copy of a validated method of analysis, secure materials for sampling, submit information about order sizes to use for calculation of acceptance criteria, obtain and submit drawings of the machine to use for calculation of surface area, prepare a draft for a SOP for cleaning of the machine, secure an adequate number of tablets for the validation, distributed on three different batches, and update the time schedule for the project. The contractor was to prepare a protocol to be commented on within a fixed deadline and provide rationales behind selection and acceptance criteria. The time from the first meeting after the contract was in place and until the project was to be finished and the machine ready for production was set to two and a half months. Within this timeframe, the machine was to be installed and IQ, OQ, PQ, and other activities were to take place before the cleaning validation could start.
If for some reason the project exceeds timeframes, budget, or in other ways does not live up to the agreed terms, it is important to be in a position to point out changes and consequences as early in the process as possible. In this particular case, delivery of the machine was delayed. This required revision of the time schedule, and it is especially important to note that all parties involved in commenting and approving activities, the analysis work, the actual test work, and the report preparation, etc. must be notified of the changed time schedule. Beyond the descriptions of responsibilities included in both the agreement and in the meeting minutes, the protocol specifies the distribution of responsibilities applicable in connection with the actual validation.
Calculation of Area in Contact with the Tablets and Definition of Acceptance Limits
The first task was to establish the acceptance limits for the active substance. The analysis laboratory had to start validation of the method of analysis as quickly as possible and therefore needed to know at what level the acceptance limit was going to be specified. To establish the acceptance limit, the total machine surface area in direct contact with the product must be estimated. All parts of the machine to be in contact with the tablets, i.e., the entire tablet infeed system, must be included in the calculation. The tablet infeed system consists of a funnel. The tablets are poured into the funnel, and the funnel directs the tablets to a vibrating plate made of steel provided with small holes.
The vibrating plate leads to a sloping steel funnel, and this funnel again leads to a rotating device made of plastic and plexiglass that distributes the tablets into tracks on an aluminum plate which again leads down to the blister foil. A plate made of plexiglass covers the aluminum plate with tracks - Figure 1. As already mentioned, the packaging machine is still in Italy at the start of the project. For confidentiality reasons, it was not possible to get access to detailed drawings and the exact measurements were therefore not available. The only materials available are a brochure showing parts of the design, and two drawings one of which is showing a few measurements. Together with the brochure and the other drawing, the drawing with the measurements is used for proportional calculations of measurements, which can then be used for estimating the required area.
This is a very uncertain method, but the only method available at this point in time. As the total area in contact with the product is included in the calculation of the acceptance limit in the denominator, it is important to arrive at a worst case calculation to round off to the nearest high measurements and calculations so that the area gets as large as possible and the acceptance limits are tightened. The surface area was calculated to be A cm?. When the machine was installed at the client’s site, the client found that the estimated worst case was indeed adequately covered and that the calculated area was larger than the actual area. The actual area is estimated to be two thirds of the calculated area.
The specific tablets are manufactured in three strengths, the lowest strength being without coloring agents whereas yellow and red coloring agents are added to the medium and highest strength, respectively. The tablets containing the three strengths have the same weight and size. The tablets with the highest strength thus contain more active substance in terms of percentages, and they therefore present the worst case for the validation?. Selection of acceptance criteria included considerations to decomposition products. The client was contacted and expressed that the active substance is very stable and that decomposition of the product during the period from production start to cleaning is deemed unlikely. This also was the reason why the method of analysis was not validated in relation to decomposition products.
Table A. Lowest determined recovery percentages.
When approaching the client, the other pharmaceutical company had stated the acceptance criteria to be fulfilled. The requirement for the cleanliness of the equipment is as follows: No more than 0.1 % of the normal therapeutic dose of any product will appear in the maximum daily dose of the following product³.
The most stringent of the following criteria must be met:
a. No more than 0.1 % of the normal therapeutic dose of any product will appear in the maximum daily dose of the following product.
b. No more than 10 ppm of any product will appear in another product.
c. No quantity of residue should be visible on the equipment after cleaning procedures are performed. Spiking studies should determine the concentration at which most active ingredients are visible.
It must be assumed that the other pharmaceutical company has calculated that (a) is the most stringent acceptance limit for this active substance. Normal therapeutic dose and maximum daily dose are thus included in the calculation of the acceptance limit for the active substance. The normal therapeutic dose varies from patient to patient, and the smallest possible dose is therefore used as worst case. The minimum daily dose (Lowest Daily Dose, LDD) is three tablets of the lowest strength.
LDD = 3 * B µg
= C µg and 0.1 % of C µg = D µg
The maximum daily dose (Highest Daily Dose, HDD) is eight tablets of the highest strength?. The behaviour of any leftover residue after cleaning is in no way predictable. The only predictable thing is that the leftover residue is not distributed evenly onto the tablets that will subsequently be in contact with the equipment. To finish the calculations, it is therefore necessary to assume that a certain, even distribution will take place. In this case, it is assumed that any leftover residue is released in an even distribution on the subsequent tablets. In worst case, the maximum eight tablets that a patient is to take will be contaminated with tablet dust from the last packaged tablets. The smallest order packaged by the client is for 60,000 tablets. If these 60,000 tablets have the highest strength and are taken in a maximum dose, the scenario would be as follows:
The maximum allowed addition to each of the 7,500 daily doses is D µg. This means an allowed maximum of D µg * 7,500 = E µg distributed on the entire machine. This in turn assumes that the residue that might be left over on the machine is transferred and evenly distributed to the next packaging order.
As one of the first tasks, the analysis laboratory performed a recovery test of the active substance on the four materials that are covered by the tab-let feeding system. 100 cm 2 /15.5 inch 2 are applied with a known quantity of active substance in a given concentration. The lowest determined recovery percentages for the four materials are shown in Table A.
The test area is 100 cm?, standard size for sampling areas. A cotton roller moistened with methyl alcohol is used for swabbing the test area which is the same method used for validation of methods of analysis for active substances. As a result, the final calculation of the acceptance limit is:
= the maximum substance quantity allowed in a single sample. Consequently, the acceptance criteria for the test are that all results must be dF.
The results of the chemical tests all fell under the detection limit.
The fact is that the machine is still in Italy at the start of the project also makes determination of hot spots difficult. Hot spots are defined as places where residue tends to collect or as places that are difficult to reach during cleaning. The tablet residue is expected to settle on and in the edges and corners of the vibrating plate. The vibrating plate is provided with holes and the dust from the tablets falls through these holes and is collected in a dust tray. The dust tray is not covered by the validation, since it is not in contact with the product. The device used for distributing tablets in the tracks is a rotating cylinder on which four flat rectangular plastic discs have been firmly affixed. This device is definitely expected to collect dust and to be difficult to clean. It was not the immediate intention to disassemble the device for distribution of the tablets prior to cleaning if at all avoidable, but it was definitely something that might become necessary.
The tracks in the aluminum plate are not expected to be especially dust collecting as the plate is sitting at sloping angle, the tracks may on the other hand be difficult to clean and inspect for visual cleanliness. Ten hot spots are selected, covering all four material types. When the machine was installed at the client’s site and the machine was put in operation, a pre-test was per-formed and the client ascertained that the anticipated dust-collecting spots do indeed match reality. The client furthermore ascertained that it is necessary to completely disassemble the device for distribution of the tablets to obtain an acceptable cleaning result. It was clear that dust collected between the rotating cylinder and the four plastic plates attached to the cylinder. The dust could not be removed without disassembly of the four plastic plates.
As the active substance is not soluble in water or ethyl alcohol, a detergent must be used for cleaning. To this end, a detergent is used containing a quaternary ammonium compound that is very surface-active. Using a detergent adds a new aspect to the process, i.e., the risk of transferring detergent to the tablets and resulting ingestion of detergent in connection with intake of the tablets. When a detergent is used, the validation must thus include inspection of the removal of detergent residue. The client was in charge of a test for removal of the detergent. A “Test kit for determination of cationic detergents (surfactants)” was used. The acceptance criteria and the rationale behind the criteria were included in the overall protocol.
To be in a position to clean the machine and to validate the cleaning, the equipment must be used and contaminated with the active substance. It was discussed how many tablets were needed for the validation. The tablets are relatively expensive and the client therefore wished to minimize the quantity and reuse the tablets instead. The contractor preferred to have a sufficient quantity available, and did not want to reuse the tablets. The maximum speed of the machine is 270 blister cards per minute. Each card contains 15 tablets, which corresponds to 4,050 tablets per minute or 243,000 tablets per hour. The quantity of the tablets used for the validation is equal to running the tablets through the machine three times. A test was not performed to show that the tablets gave off as much dust during the third run as they did during the first run. To avoid puncturing the blister packages to get the tablets out, the tablets are transported without the top foil and in such a way that after they have been placed in the bottom foil they are poured out of the foil and collected. The tablets are then poured into the funnel again and reused for contaminating the machine.
The tablets are run through the machine for one hour. Prior to cleaning with water and detergent, all components are vacuumed for tablet dust as thoroughly as possible. A vacuum cleaner has been purchased for this purpose, dedicated to use for runs with the specific tablets. As vacuum cleaning removed practically all the dust anyway, the amount of dust given off by the tablets was of no importance. And it would therefore have made no difference if a test had been performed to show how much dust the tablets gave off during a third run. In addition, all equipment components are vacuum cleaned at the end of the work day if the machine is to continue packaging tablets from the same batch on the following day. The client and the contractor had agreed that the cleaning method for the new equipment must resemble the method already being used by the client.
The client uses manual washing processes, soap water, and rinses the equipment with running water and a hand shower. It can be difficult to adapt a non-validated washing process like that to a controlled situation. First of all, the solution of the detergent to be used on the new equipment must have a predefined concentration. This was solved with the use of a measuring cup for measuring out the detergent. A large sink was marked to indicate the filling level of the water. As both the processes of measuring out the detergent and the water involve a certain degree of uncertainty, the validation demonstrates that the machine can be cleaned adequately even with a short measure of detergent and can be rinsed adequately after an excess of detergent. Secondly, the process of rinsing with running water is difficult to control.
There may be differences in how hard the water is turned on, the duration of the rinse, and how the equipment is moved around under the water. The temperature of the water may also have an influence on how easy or difficult it is to rinse the equipment to get it clean. It is furthermore necessary to collect the rinse water for purposes of measuring potential leftover residue. Filling the sink with a defined measure of water, rinsing, and repeating the process solved this problem. The last batch of rinse water was used for measuring detergent residue. Apart from the last batch of rinse water, the water used for cleaning and rinsing is city water from a tap producing water with a permanent temperature of approximately 40°C. The last rinse uses RO water that also has a permanent temperature of approximately 40°C.
The client’s internal monitoring system generates a warning if the temperature goes below or above the specified limits. Thirdly, the human factor must always be allowed for when manual procedures are included in the process. The human factor is difficult to handle. It cannot be measured and it may not be the same every day. There are both disadvantages and advantages to manual wash processes. When people performing a task know that they are being watched or supervised, one of two things usually happens. They either get nervous and start fumbling and do things in the wrong sequence, and the result is not necessarily the same as if they had not been watched and could perform the task at their own pace. Or they become very thorough and do things more meticulously than they would under normal circumstances. Experience shows that skilled operators tend to fall under the latter category, whereas a relatively new and inexperienced operator is likely to fall under the first category. Watching or supervising people during such processes also may reveal missing SOP compliance. An operator is well into a wash process, stops, and hesitates a little: “I know that the SOP specifies that I must ..., but that is not possible because ... So I will ...” A variation of this could be: “I know that the SOP specifies that we must ... but we will ...” The first scenario is relatively easy to handle.
The SOP must be revised to match the real world if the real world complies with cGMP. This involves simply asking the operators to draw attention to such instances. The second scenario is a bit more difficult to handle. Why does the SOP and reality not match, when, how, and why did reality change into something other than what is described in the SOP. And last but not least, where else are discrepancies between the SOP and reality likely to be found? The time of day when the cleaning is performed also influences the cleaning process. If the cleaning takes place close to the end of the work day, the operator may tend to be sloppy with the cleaning. Under normal circumstances, two to three operators can easily perform a demanding task and carry on a conversation at the same time. They do not look at each other while they perform the task, but at the task at hand. When an outsider talks to them while they perform the task, they often have eye contact with the person they are talking to out of politeness. These disturbances and stress factors are used to influence the manual cleaning procedure to arrive at worst case. The client had selected a number of operators who were to be permanently assigned to this blister machine.
The operators were chosen based on their knowledge and experience, and they were all good at their work. Subsequently, no new or inexperienced personnel could be engaged in the validation process, but handling of these specific machine parts was relatively new for all involved. During all three cleaning processes, the operators were distracted a stressed caused by questions asked by the contractor and by other interruptions. One of the cleaning processes with subsequent tests took place on a late Friday afternoon just before the operators were to go home for the weekend. (As a minimum, three different operators must clean and rinse the equipment with the tasks preferably distributed on experienced and less experienced/relatively new operators in connection with the three subsequent runs/cleaning processes).
Neither the FDA nor the EU specifies room classification for packaging of tablets. The closest we can get is 211.46 and 3.12. Under normal circumstances, packaging takes place in controlled areas corresponding to class D/100,000. In class D/100,000 rooms, 50-cfu/ plate is the allowed maximum on 55 mm/2.2 inches. contact plates. When working in class D/100,000 rooms, the personnel must cover their hair and beards. Suitable clothing and shoes must be worn. These unwritten rules are followed at the client’s site.
The staff changes to work clothes, including caps, before entering class D production areas. When their work involves physical contact with the equipment, they also put on gloves. The room is monitored at suitable intervals for compliance with class D/ 100,000. The equipment also must live up to class D/100,000 after completed cleaning. Samples are taken on the equipment equal to one sample on each of the four materials. The results for microbial cleanliness must be =50 cfu/plate when the plate has a diameter of 55 mm/2.2 inches.
The other pharmaceutical company mentioned at the beginning of the article expressed their concern in connection with vacuum cleaning being carried out at the end of the work day if the machine was to continue packaging tablets from the same batch on the following day. The concern was directed at microbiological contamination of the equipment from the vacuum cleaner. Microbiological tests were performed on the equipment before and after the vacuum cleaning and the concern was documented as being un-founded. The results of all microbiological tests were found to be below the acceptance limit.
The cleaning validation is valid as long as the same cleaning procedure is used for cleaning as was used for the validation. The smallest packaging order of 60,000 tablets is included in the calculation of the acceptance limit and can therefore not be changed to a smaller quantity without revalidation. In principle, there is no upper limit for the size of the packaging order as it is not included in the calculation of the acceptance limit. But that does not mean that this aspect should not be considered if very large quantities are scheduled for packaging.
The condition of the equipment and the manual cleaning process should be evaluated regularly. New equipment is smooth and polished. During use, the surfaces get worn and may be scratched in connection with handling the packaging processes. Changes may creep in during manual processes. For a start, these changes may be barely discernible, but in the long run, they can change a process completely. Regular tests for verification are recommended.
To perform a validation for a client is not always without problems. For one thing, it is not possible to obtain all relevant information, and information and statements must therefore be used in good faith. The reason for this might be that the client is not interested in revealing sensitive information, that the client does not set sufficient time aside for passing on information to the contractor, or that the client takes things for granted. The situation does not get easier in this case by a third party being involved, a third party who owns the main part of the information that might be of interest to the project.
The client wants to meet the demands made by the other pharmaceutical company, but on the other hand does not want this to cost more than absolutely necessary, as a satisfactory cleaning validation does not automatically lead to a contract with the other pharmaceutical company. The client does not have the experience or the resources to solve the task within the strict timeframe. By hiring a contractor, the client buys the experience that the client does not have available, at the same time binding the contractor to deliver a validation that complies with current requirements, meets a deadline based on certain assumptions, and keeps within a fixed price, also based on certain assumptions. The contractor possesses the knowledge and experience needed to solve such tasks within the given framework.
- FDA: Code of Federal Regulations 21 Part 211 “Current Good Manufacturing Practices for Finished Pharmaceuticals,” US Food and Drug Administration.
- EU: The Rules Governing Medicinal Products in the European Community, Volume IV, “Good Manufacturing Practice for Medicinal Products.”
- PIC/S, Recommendations on Cleaning Validation, August 2001.
- LeBlanc, D.A., “Establishing Scientifically Justified Acceptance Criteria for the Cleaning Validation of APIs,” Pharmaceutical Technology, October 2000, pp 160-68.
- Danish Catalogue listing pharmaceutical products, 2000.
Reprinted from Pharmaceutical Engineering, The Official Journal of ISPE March/April, Vol. 24 No. 2,www.ispe.org
This case study describes a validation assignment carried out by a contractor for a client.