Introduction
How can the time between a cleanroom power outage and loss of environmental control in the critical area be determined? Once the power is restored, how can the time it takes to recover the desired environmental conditions be determined?
An interruption of power supply to the HVAC systems may produce a “loss of control” which can be defined as a breech in the integrity of the controlled areas in sterile manufacturing. Appropriate steps to be taken during and after an interruption of air supply to the aseptic processing area (APA). (These steps should be placed into site procedures before the studies recommended in this document are executed.)
Qualification studies should be carried out to define a time limit after which the controlled environment reaches or exceeds action levels for particulates, temperature, humidity, and pressure differential.
To assist in meeting these requirements, this guidance will provide recommendations:
* To assist in the qualification of an “in control” time period after power loss to a controlled environment.
* To assist in the qualification of a time period for full recovery of the controlled
environment following power restoration.
It should be noted that this guidance is intended for Aseptic Processing Areas which are defined as those controlled environments consisting of Grade A or Grade B classifications. The recommendations of this document can be applied to other classifications (e.g., Grade C) depending upon a risk analysis by the site.
In addition, this guidance is primarily concerned with pressure drops in a controlled environment due to a power outage to the central HVAC system. A Power failure occurrence in an individual air handling unit is beyond the scope of this document.
Recommendations & Rationale for Recommendations
The most effective way to deal with interruptions in power supply to air handling systems is to prevent them from happening in the first place. Production loses from power outages can be minimized or eliminated by supplying critical equipment with power from generators, automatic transfer switches, and uninterruptible power supplies (UPS).
However, in situations where such equipment is not installed, a contingency plan for handling power outages to air handling systems is needed. It is important to consider three questions when developing a procedure to deal with power interruptions to air handling equipment.
What should be done during the interruption?
Most importantly APA personnel should act to minimize movement, especially into or around Grade A areas. It is also important to avoid opening of doors to areas of lower classification. Further, it is critical that APA personnel be trained in these procedures so there is no need to leave the APA.
What length interruption will require re-sterilization of product contact parts, unused closures, and unused containers?
This question requires generation of data to answer. Setting a time limit without supporting data will risk product contamination or the unnecessary destruction of sterile product.
Determination of a time limit consists of extensive environmental monitoring after the APA power has been interrupted and the critical air handling systems ceases to function. This may also includes interruptions to the air handling systems of areas adjacent to the APA that would be affected by the interruption.
The essential monitors to include in the determination are total airborne non-viable particulates, viable quantitative (active) air and/or viable passive air (settling plates), and pressure differentials. Additional parameters such as temperature and humidity should be tested if they are critical to the process.
The Non-viable monitors and pressure differential monitors are the most sensitive indicators for detecting any changes in the quality of the APA environment. The state of microbiological control in an APA can be directly correlated from these monitors.
Although the viable monitors (active and passive) are not as precise a tool for this study, these measurements are also essential to provide a complete understanding of the total APA environment After testing, the monitor(s) that reached the predetermined acceptance criteria in the shortest time will define your power interruption time interval. After the power interruption time limit has been established, it should be integrated into process simulations (i.e. media fills) for an added level of assurance.
A power interruption time interval study should also reflect “in use” APA conditions according to procedures that define the actions of personnel present during a power interruption. For example, the power interruption study should take into account the potential maximum number of personnel that may be present in the APA and what actions the personnel would perform during the interruption, such as exiting the APA or reducing all movements to nominal tasks.
Air Quality
Non-viable particulate and viable microbiological air tests used to determine a power interruption time limit are quite similar to the qualification studies used to classify the area. The monitors are placed in the controlled area while the air handling system is operational. At time zero, power to the air handling system is shut down and the air is continuously monitored for a defined length of time.
For non-viable particulate monitoring, the number of locations can be determined by taking the square root of the total controlled surface area.
It is also prudent to place additional particulate monitors at critical locations based on a risk analysis.
The number and placement of passive and/or active viable microbiological monitors can be determined by risk analysis. The risk analysis placement should include locations that have a high potential for microbial presence based upon areas that are difficult to clean, historical data, production activities, previous qualification data, traffic patterns, and line configurations.
Uniform coverage of the controlled area may also be utilized by the inclusion of additional monitoring points to ensure that the viable particulate profile is accurately represented throughout the area.
Acceptance criteria for air quality tests should be no higher than the action levels employed during routine production.
Pressure differentials and smoke studies
Another key parameter in defining a power interruption time limit is the flow of air through the classified area. Pressure differentials are ideal for this purpose since data can be monitored continuously and can give an indication of the potential change in the flow of particulates from adjacent areas. As in air quality tests discussed above, pressure differentials are monitored after the air handling system is shut down and the time measured until the differential reaches a predetermined acceptance criterion.
However, unlike the air quality test, the routine production action levels may not be appropriate, especially in controlled environments with openings (e.g. mouseholes). The minimum acceptance criterion should be pressure equilibrium between the test area and an adjacent controlled area of lower classification (i.e. no reversal of air
flow).
As with air quality testing, a safety margin should be built into the acceptance criteria (i.e. requiring slight positive pressure in room being tested). For example, if non-viable particulates levels reach the action limits for the room at a pressure differential of 2 Pascals, then the acceptance criteria should have a minimum pressure differential greater than 2 Pascals.
Further assurance can be achieved in controlled environments with numerous openings via the use of smoke studies. Careful analysis of smoke patterns around doors and openings can ensure that air does not flow from an adjacent area of lower classification into the area being tested.
Temperature and humidity
If temperature and humidity in the controlled area are critical process parameters (e.g. lyophilization), and then these should also be included in determining the time limit during a power interruption. Testing can be performed concurrently with air quality tests or continuously throughout the outage period and the acceptance criteria should be no higher than the action levels employed during routine production. A conservative approach would be to assign lower acceptance criteria in order to provide a margin of safety.
Process simulations (Media Fills)
After the power interruption time limit has been established, it can be qualified by incorporation into process simulations. Like the other tests mentioned above, it would be appropriate to incorporate power outages into process simulations only during qualification or requalification of the cleanroom (i.e. after major changes). As is the case with routine production, processing should be completed for units exposed to the environment during the power interruption with subsequent segregation of these units from the rest of the batch. Positive units from the segregated portion of the batch do not indicate failure of the entire media fill provided that written procedures and batch documentation are adequate to describe the associated clearance during routine production. It does however, indicate a need to revisit the data generated during the power interruption time limit determination and potentially perform this work again with more stringent acceptance criteria. It may also indicate a need for retraining of personnel. It should also be noted that releasing the portion of a product batch segregated during a routine production power outage would be extremely difficult to justify until the time limit has been qualified successfully with process simulations.
Revalidation
A minimum of one validation study should be performed for the initial determination of a power interruption time limit. A revalidation should be conducted if significant changes are made to the HVAC system that may effect the initial validation.
How long will it take to return the APA to a controlled state?
This is another question that requires data to answer. The Recovery Test is the standard method for determining the time interval for a controlled environment to return to its specified cleanliness class after being exposed to a source of airborne particulates6. It is essential that a recovery test be performed after the power outage study unless recovery time data was generated during the initial qualification of the APA. In the recovery test, a particulate source (smoke or aerosol) is generated from the center of a predefined grid area in the room until the particle count is above the controlled environment’s at-rest level.
After the particle generator is shut off, the particle concentration should be allowed to decrease to a point (e.g., 1 minute) where the counter will not be saturated with smoke or aerosol. Particle measurements are subsequently taken until the original at-rest air particulate level is reestablished. The recovery time is defined as the time interval between the particulate source shut off and when the particle levels return to the original state. In routine production, product processing should not restart until after the qualified recovery time has been exceeded and the requisite sanitization, sterilization, and HVAC/environmental monitoring steps (e.g., total particulates, temperature, humidity, & pressure differentials) have been completed.
Case Study Example
The following is an example for conducting a power outage study for a class 10,000 (100 at rest) APA. This is only one example of how to perform this type of study as various approaches can be utilized according to the needs of the facility. In order to define the duration of the study, and thus the maximum power outage time limit, continuous real time monitoring was needed. This included real time pressure differential monitoring and non-viable particulate monitoring.
The length of the study was determined when the pressure equilibrium reaches a _Pof 2 Pascals or when the concentration of non-viable particulates reaches the at rest level of 100. (Note: Instead of ending the study and restoring power to the HVAC at pressure equilibrium (_P=0 Pascals), an added safety margin of 2 Pascals was arbitrarily established.) The number of non-viable particulate monitor locations was based upon a risk assessment by the site and by previous APA qualification data. For the viable active and passive monitors, the acceptance criteria were based upon the action levels established for the “at rest” room. These monitors were placed in locations previously established through routine and initial qualification studies. The measurements from the viable active and passive monitors are not as critical for this type of study because pressure equilibrium or non-viable particulate levels will most likely surpass the acceptance criteria long before the viable count acceptance criteria is exceeded.
However, this data is still needed in order to provide a complete understanding of the APA environment.
One hour prior to power shut down, non-viable particulate monitoring, pressure differential monitoring, active viable air monitoring, and passive monitoring with Tryptic Soy Agar settling plates were conducted in order to establish a normal operations baseline. The pressure differential monitors and the non-viable particulate monitors were to remain in continuous operation before, during, and after the power shut down. All of the active monitor strips and passive settling plates were collected just prior to initiating power shut down (time zero) and new active and passive monitors were added so they could be utilized throughout the remainder of the study.
At time zero, power to the HVAC systems was discontinued and all production personnel had exited the APA according to procedure. Because the exact length of the shut down time was not apparent, it was decided to remove and activate new viable active and passive monitors in 5 minute increments from time zero of the study. The study was ended and the power supply to the HVAC systems was restarted after it was found that the non-viable particulates had reached 100 within 15 minutes after power shut down.
However, the pressure differential had reached only 5 Pascals within this 15 minute time frame. This was well above the established pressure differential safety margin of 2 Pascals. The viable active and passive monitors were subsequently incubated and were all found to be within the pre-established action level acceptance criteria. From this study, it can be concluded that the area meets the Class 10,000 (100 at rest) acceptance criteria within 15 minutes after shut down of power to the HVAC systems.
Because the point of failure was determined at 15 minutes, a safety margin was incorporated by deducting an additional 5 minutes from the total maximum power outage time limit of 10 minutes.
After power was restored to the HVAC systems, the pressure differentials reached the normal operation parameters of 12.5 Pascals and the non-viable particulates returned to acceptable at rest levels within 10 minutes.
An SOP can be established from this data describing how long the APA meets acceptance criteria during a power interruption and what actions need to be taken.