Environmental Monitoring of cGMP Cleanroom Facilities for Manufacturing of Cellular Therapy Products in an Academic Hospital Setting

Abstract

As the field of cell and gene therapy continues to grow, so too must the infrastructure and regulatory guidance supporting the manufacture of these potentially life-saving products – especially early-phase products manufactured at an increasing number of academic or hospital-based facilities providing decentralized (or point of care) manufacturing. An important component of current Good Manufacturing Practices, including those regulating cell and gene therapies, is the establishment of an effective environmental monitoring (EM) program. While several guidelines for establishing an EM program are available, these guidelines do not specifically address the unique aspects of manufacturing cell and gene therapy products and they do not provide real-world evidence demonstrating the effectiveness of the program. Here, we describe the establishment and evolution of an EM program in a cell therapy manufacturing facility at an academic hospital. With 10 years of EM data, we analyze the effectiveness for identifying trends in environmental conditions and highlight important findings, with the aim of providing practical evidence and guidance for the development of future early-phase EM programs.

INTRODUCTION

Background

Over the past decade, the field of cell and gene therapy (CGT) has seen remarkable progress in translating benchtop research into bedside clinical application. To date, the US Food and Drug Administration (FDA) has approved 36 cell and gene therapies, however, the large number of clinical trials in the pipeline and high rates of clinical success observed thus far, suggests these numbers will continue to increase at a rapid pace¹. By 2025, the FDA estimates they will be approving 10–20 CGT products per year². As the field continues to grow and expand, so too must the infrastructure and regulatory guidance supporting the manufacture of these CGT products, especially early-phase products that are not manufactured by large pharmaceutical companies for licensed products.

Cellular therapy manufacturing for clinical trials according to current good manufacturing practices (cGMP), regulated by section 351 of the Public Health Services Act, is mostly limited to a handful of large industry and academic institutions capable of meeting cGMP regulations. An important, yet resource intensive, component of cGMP as outlined in the Code of Federal Regulation Title 21 Section 211, is the establishment of a “system for monitoring environmental conditions”³. While several guidelines for establishing an environmental monitoring (EM) program are available from the FDA⁴, United States Pharmacopoeia (USP)^5,6, and the European Commission^7,8, these guidelines are largely written for pharmaceutical processes and do not specifically address many of the unique aspects present in manufacturing cellular therapy and cell-based gene therapy products. Such unique aspects include the often small, single batches of patient-specific products, the fact that products are living drugs – thus precluding sterilization – and inherent patient-to-patient variability necessitating flexible manufacturing processes. Importantly, given the special considerations of CGT, there is a strong desire for decentralized and even point-of-care manufacturing^9,10. As a result of the increase in available CGT products and the desire to develop in-house manufacturing capabilities, there has been an increase in the number of facilities performing these functions; and as these facilities continue to proliferate, the type, frequency, and evaluation of EM methods continues to become more diverse. However, to our knowledge, there are limited available reports on the set up, maintenance, and effectiveness of EM programs in CGT manufacturing facilities.

As early adopters of the promise of cell-based therapies, the Cellular Therapy Laboratory (CTL) at Children’s National Hospital was established in 2013, and since its inception has manufactured over 600 products under investigational new drug (IND) applications in support of over 25 early-phase clinical trials. Here, for the first time, we describe the establishment and evolution of an EM program in a CGT manufacturing facility and assess its effectiveness by evaluating nearly 10 years of real-world data, ultimately with the purpose of providing practical evidence and guidance for the development of future early-phase EM programs.

Design of environmental monitoring program.

Environmental monitoring is a program designed to detect changing trends and patterns of air quality, microbial counts, and microflora growth within controlled environments. Key components of an effective environmental monitoring program include viable and non-viable particle monitoring, surface microbiological monitoring, temperature and humidity controls, personnel monitoring, and facility maintenance (Figure 1). The results obtained provide essential information about overall cleanroom operations, aide in contamination control, and identify any threats to the purity of the manufactured clinical drug products.

Figure 1. Overview of operational controls and monitoring of cleanrooms.

Schematic representation of the key components to consider when designing an effective environmental monitoring program for a cell and gene therapy manufacturing facility.

The environmental monitoring program in the CTL at Children’s National Hospital (CNH) was designed in accordance with the International Standards Organization (ISO) classification of clean rooms^11,12, FDA⁴ and USP guidelines, as well as academic guidance, all with the intention of supporting phase 1 and phase 2 studies¹³. The sampling points and frequency in the clean rooms were derived from the requirements stated under USP⁶, initial clean room qualification results, and our own internal risk assessments. These assessments considered factors such as high-traffic areas, personnel flow, equipment location, and facility layout to guide the designation of sampling locations within the facility. Alarm limits for air quality and surface contamination, above which facility operations must halt until rectified, were established based on the ISO standards^11,12 and USP guidelines. Alert limits, above which Quality Assurance and Facility Director are to be notified to determine appropriate remedial action plan, were established based on the initial qualification of the clean rooms as well as the available historical EM data to alert when environmental monitoring samples are approaching alarm limits.

The types of EM testing performed in the EM program include non-viable particle counts, viable particle counts, and surface monitoring (i.e., touch plates). Non-viable testing was performed to verify adequate control over particles that are not living, which may include dead microorganisms, in accordance with the area classification and can also be associated with the number of air changes per hour. Viable monitoring was performed to verify adequate control over living microorganisms (e.g. bacteria, molds, and fungi) in accordance with the area classification. Surface monitoring is performed to verify that the microbial levels on tabletops, counters, floors, and equipment adhere to the area classification requirements. These tests were carried out on a weekly basis during both static and dynamic manufacturing conditions.

Overview of cGMP Cleanroom facilities.

The cGMP facilities at CNH consist of two separate processing spaces located on floors 3.5 and 5 of the CNH Main Building (Figure 2). The facility on floor 3.5 (3.5GMP) consists of one ISO 8 classified suite (424 ft²) and one ISO 7 suite (90 ft²). The ISO 8 suite contains one 4-foot and one 6-foot A2 ISO 5 biosafety cabinet, while the ISO 7 room has one 4-foot A2 ISO 5 biosafety cabinet (Figure 2A). The ISO 7 lab is positively pressured in relation to the ISO 8 lab and the ISO 8 airlock is positively pressured in relation to the ISO 8 lab and the corridor.

Figure 2. Schematic layout of GMP cleanroom facilities.

Scale representation of the floor 3.5 GMP (A) and 5 GMP (B) cleanroom facilities located at Children’s National Hospital. Environmental monitoring sampling locations for non-viable (grey triangle), viable (yellow circle), and surface (orange square) samples are displayed and individually numbered. Arrows indicate direction of air flow due to pressure differential between rooms.

The facility on floor 5 (5GMP) is a newer facility opened in January 2021 and consists of one ISO 8 anteroom (134 ft²) that serves as a bi-directional gowning room and two ISO 7 suites (318 ft²) (Figure 2B). Both ISO 7 cleanrooms are positively pressured in relation to the ISO 8 anteroom, which is positively pressured in relation to the unclassified space. Each ISO 7 lab is equipped with one 6-foot A2 ISO 5 biosafety cabinet for open manipulation of cells.

Clean, oil-free dry compressed air is supplied to the facility from the hospital building’s system. Each of the biosafety cabinets is supplied with a compressed air drop. An air handling unit supplies the 3.5 facility with 100% outside air that is HEPA filtered prior to supplying to the classified rooms and two exhaust fans provide 100% exhaust. Similarly, two separate air handling units supply the ISO 7 cleanrooms and ISO 8 anteroom on 5^th floor with 100% outside air that is HEPA filtered prior to supplying the classified rooms. The ISO 8 and ISO 7 rooms in the 3.5 facility operate at least 15 and 39 air changes per hour, respectively while the ISO 8 and ISO 7 rooms in the 5th floor facility operate at least 35 and 72 air changes per hour, respectively. The CTL facilities are registered with the FDA as a tissue establishment to support the processing of Human Cells, Tissues, or Cellular and Tissue-Based Products (HCT/P or Type 361 products). They are also accredited by the Foundation for the Accreditation of Cellular Therapy (FACT) for processing both minimal and more than minimally manipulated products. The clean rooms are certified annually by a third-party vendor.

In total the cGMP cleanroom facilities at CNH have supported the manufacturing of 526 products (as of December 2022) for various phase 1 clinical trials including the product types listed in Table 1. The majority of the manufacturing was performed in open culture systems within biosafety cabinets. In the case of mesenchymal stromal cell manufacturing, a semi-automated bioreactor, the Quantum Cell Expansion System, was utilized¹⁴. Time requirements for manufacturing these products varied from 1–4 weeks depending on the product type.

Table 1.

Type of Products Manufactured in cGMP Cleanroom Facilities at Children’s National Hospital

Product Type	Manufacturing Time	Automated vs Open System
Tumor-Associated Antigen-Specific T cells (TAA-Ts)15–18	3–4 weeks	Open system
Virus Specific T Cells (VST)19–21	2–3 weeks	Open system
Mesenchymal Stromal Cells (MSCs)14	1–2 weeks	Semi-automated system

METHODS

Surface Microbiological Monitoring:

The total surface bacterial and fungal counts were assessed using Replicate Organism Detection and Counting (RODAC) plates (diameter: 55 mm, surface area: 24 cm²) containing irradiated tryptic soy agar with lecithin and polysorbate 80 (BioMerieux, Maryland, USA). Each plate was briefly touched to representative surfaces – i.e. biosafety cabinets, worktables, upper horizontal surface of refrigerators, centrifuges, computers, floors, and inside incubators (Figure 2). For bacterial and fungal detection, plates were incubated at 30–35°C for 7 days, and plates were assessed for growth on day 2–4 and day 7.

Viable Particle Monitoring:

The total viable bacterial and fungal counts in circulating air were analyzed using the SAS Super 100 particle counter (Bioscience International, Maryland, USA). A total of 1000 liters of air was aspirated at designated sampling points throughout the cleanroom facilities (Figure 2). The resulting laminar air flow was directed onto RODAC plates for bacterial/fungal capture and culture. Plates were then incubated at 30–35°C for 7 days and assessed for growth on day 2–4 and day 7.

Non-Viable Particle Monitoring:

The total non-viable particle counts were carried out using the MET one HHPC 3+ non-viable particle counter (Beckman Coulter, California, USA). Air samples were collected for 10 minutes at 100 liters/minute at each of the established sampling location throughout the cleanroom facilities (Figure 2) and the number of non-viable particles ≥0.5μm per cubic meter of air was analyzed.

Identification of positive environmental monitoring samples.

All touch plates and viable particle count plates identified as positive for microbial growth were shipped overnight at ambient temperature to Charles River Laboratory (Newark, Delaware, USA) for further identification. The total time from sample inoculation to arrival was ≤ 24 hours. Matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) mass spectrometry was utilized to generate unique protein spectral fingerprints assessed by AccuPro-ID MALDI-TOF microbial identification service for bacteria and yeast and AccuFUN-ID for microbial identification for filamentous fungi. Turnaround time for each service was 5 days and the final report included a list of closest matched species according to protein fingerprint similarity.

Sterility testing of final products.

All cell therapy products manufactured in our cleanroom facilities underwent microbiologic testing (aerobic, anaerobic, and fungal) prior to release for clinical administration in accordance with the Chemistry, Manufacturing, and Controls section of the IND application with the FDA. Approximately 0.5 mL of final drug product was inoculated into each of a BacT/ALERT^® FA Plus, BacT/ALERT^® FN Plus (bioMérieux), and fungal isolator tube (Wampole Laboratories, Cranbury, NJ). BacT/ALERT^® bottles were cultured for 14 days at 37°C and monitored for bacterial growth using the BacT/ALERT^® VIRTUO^® (bioMérieux) automated blood culture system. Samples from fungal isolator tubes were plated onto Saboraud Dextrose agar plates and incubated at 30°C for 21 days. Positive cultures were further identified by standard biochemical assays in a CLIA-certified hospital microbiology laboratory. These sterility testing methods were previously qualified to demonstrate consistent detection of microorganisms at low bioburden.

RESULTS

Environmental monitoring sample positivity rates.

Between October 2013-December 2022 a total of 9210 environmental monitoring samples were collected in accordance with our established program. These included 3780 (41.0%) surface touch plate samples, 2550 (27.7%) viable particle air samples, and 2880 (31.3%) non-viable particle air samples (Tables 2–3 and Figure 3). Of the 3780 surface touch plates, 0.59% (5/846) of ISO 5, 0.96% (14/1463) of ISO 7, and 2.72% (40/1471) of ISO 8 samples exceeded the established alert/alarm limits. Touch plate sampling revealed that the floor was the most highly positive area of the cleanroom facilities accounting for 66% (39/59) of the total samples exceeding alert/alarm limits. Specifically, the high-traffic areas at the entrance of the doorways to the ISO 8 and ISO 7 facilities – sample IDs 22 and 18 in Figure 2A, respectively – accounted for approximately half of the floor samples exceeding alarm/alert limits.

Table 2.

Summary of touch plate samples and positivity rates within cleanroom facilities

Year	ISO 5					ISO 7					ISO 8
	# of Samples	Alert (≥1 CFU/plate)		Alarm (≥3 CFU/plate)		# of Samples	Alert (≥6 CFU/plate)		Alarm (≥10 CFU/plate)		# of Samples	Alert (≥12 CFU/plate)		Alarm (≥20 CFU/plate)
2013	12	1	8.33%	0	0.00%	23	0	0.00%	1	4.35%	19	0	0.00%	1	5.26%
2014	30	0	0.00%	0	0.00%	106	2	1.89%	1	0.94%	162	4	2.47%	14	8.64%
2015	48	0	0.00%	0	0.00%	94	0	0.00%	0	0.00%	181	1	0.55%	1	0.55%
2016	52	1	1.92%	0	0.00%	111	0	0.00%	1	0.90%	151	0	0.00%	1	0.66%
2017	44	0	0.00%	0	0.00%	106	0	0.00%	5	4.72%	131	0	0.00%	17	12.98%
2018	49	2	4.08%	0	0.00%	100	0	0.00%	3	3.00%	134	0	0.00%	0	0.00%
2019	115	1	0.87%	0	0.00%	114	0	0.00%	0	0.00%	143	0	0.00%	0	0.00%
2020	127	0	0.00%	0	0.00%	88	0	0.00%	1	1.14%	115	0	0.00%	0	0.00%
2021	169	0	0.00%	0	0.00%	337	0	0.00%	0	0.00%	173	0	0.00%	1	0.58%
2022	200	0	0.00%	0	0.00%	384	0	0.00%	0	0.00%	262	0	0.00%	0	0.00%
TOTAL	846	5	0.59%	0	0.00%	1463	2	0.14%	12	0.82%	1471	5	0.34%	35	2.38%

Table 3.

Summary of viable particle samples and positivity rates within cleanroom facilities

Year	ISO 5			ISO 7					ISO 8
	# of Samples	Alarm (≥1 CFU/m3)		# of Samples	Alert (≥6 CFU/m3)		Alarm (≥10 CFU/m3)		# of Samples	Alert (≥60 CFU/m3)		Alarm (≥100 CFU/m3)
2013	8	0	0.00%	22	1	4.55%	0	0.00%	24	0	0.00%	0	0.00%
2014	36	0	0.00%	55	0	0.00%	0	0.00%	84	0	0.00%	0	0.00%
2015	41	0	0.00%	41	0	0.00%	0	0.00%	78	0	0.00%	0	0.00%
2016	40	0	0.00%	40	0	0.00%	0	0.00%	71	0	0.00%	0	0.00%
2017	44	0	0.00%	50	0	0.00%	0	0.00%	81	0	0.00%	1	1.23%
2018	41	0	0.00%	82	0	0.00%	6	7.32%	90	0	0.00%	2	2.22%
2019	67	0	0.00%	67	0	0.00%	1	1.49%	95	0	0.00%	0	0.00%
2020	69	0	0.00%	68	0	0.00%	1	1.47%	69	0	0.00%	0	0.00%
2021	124	0	0.00%	261	0	0.00%	2	0.77%	116	0	0.00%	0	0.00%
2022	199	2	1.01%	278	0	0.00%	0	0.00%	209	0	0.00%	0	0.00%
TOTAL	669	2	0.30%	964	1	0.10%	10	1.04%	917	0	0.00%	3	0.33%

Figure 3. Longitudinal analysis of non-viable particle counts in cleanroom facilities.

Data are presented as median particle counts aggregated quarterly with error bars displaying the interquartile range for (A) ISO 7 and (B) ISO 8 classified cleanroom facilities. Red lines represent the ISO standard alarm limits for the respective classified space, and the yellow lines represent our alert limits established to determine when corrective action is required. Arrows with horizontal dashed lines represent timing of HEPA filter replacement at the terminal air vents supplying cleanroom facilities.

Of the 2550 viable particle air samples taken, only 16 (0.63%) exceeded the established alert/alarm limits (Table 2). Notably, 6 (37.5%) of the 16 out of specification results were the initial detection and follow-up analysis from an isolated event within the incubators in the ISO 7 facility (sample ID 13, Figure 2A) that occurred over a 2-week period in 2018. Of the 2880 non-viable particle air samples taken, 0 (0.0%) exceeded the alert/alarm limits. While non-viable particles remained considerably below ISO standards for ISO 7 and ISO 8 classification¹¹, there was a notable decrease in non-viable particle counts in both ISO 7 and ISO 8 facilities observed around Q2 2016 (Figure 3). This decrease coincided with the replacement of the HEPA filters at terminal vents supplying air to the facilities and this pattern of decreased non-viable particle counts following HEPA filter replacement continued for subsequent HEPA filter replacements.

Speciation of environmental monitoring samples.

To establish a compendium of organisms within our cleanroom facilities, beginning December 2020, all EM samples positive for bacterial/fungal growth were analyzed by MALDI-TOF mass spectrometry for genus/species-level identification. Of the 164 touch plates and 107 viable air particle samples analyzed, a total of 398 isolates were identified (Figure 4), belonging to 35 bacteria and 4 fungi genus. Of the 382 bacterial isolates identified, 96.6% (369/382) were identified as Gram-positive bacteria with the most abundant being Staphylococcus species with 126 isolates, Bacillus species with 76 isolates, Micrococcus species with 64 isolates, and Corynebacterium with 39 isolates. The remaining 3.3% (13/382) of isolates were identified as Gram-negative bacteria consisting mostly of Aureimonas and Pseudomonas species. A small number of fungal isolates were identified, consisting of Aspergillus fumigatus with 13 isolates, Penicllium, Chaetomium, and Epicocum with 1 isolate each. Interestingly, 63% (10/16) of filamentous fungi were found specifically in 5GMP ISO 7 and ISO 8 cleanrooms between March-April 2021, and 1 occurrence in December 2021. These fungal isolates were present in both airborne and surface samples from six sampling locations within 5GMP: sample ID 7 (floor next to door), ID 8 (top of centrifuge), ID 9 (floor under passthrough), ID 20 (top of bench), top incubator and bottom incubator (Figure 2B). The most common identified microorganisms stratified by cleanroom classification (ISO 5, ISO 7 and ISO 8) are displayed in Figure 4B. Overall, the frequency of identified bacterial/fungal isolates remained relatively consistent over the two-year time frame for which speciation data was available (December 2020-December 2022).

Figure 4.

Catalogue of all organisms identified from GMP cleanroom surface and airborne samples (2020–2022). (A) A total of 398 isolates were identified belonging to 35 bacteria genus and 4 fungi kingdoms. The data presents the frequency of each isolate by individual species between 2020 and 2022. (B) The data presents the frequency of each isolate by individual species (more than 1%) in the various cleanroom between 2020 and 2022.

The greatest number of samples with microbial growth were identified in ISO 8 cleanrooms, with a total of 234 isolates (Supplemental Figure 1). 52.6% (123/234) were identified from viable air particle samples with the highest rates of positivity occurring at sampling sites near the microscope, sample ID 3 (3.5GMP-ISO 8); the computer, sample ID 6 (3.5GMP-ISO 8), and the area between ISO 7 doors, sample ID 19 (5GMP-ISO 8). Of the 123 isolates, 35.7% (44/123) Staphylococcus species identified from viable particle samples were present inside ISO 8 cleanrooms. The most common strains found in airborne sampling were Coagulase-negative staphylococcus (CoNS) including S. capitis, S. epidermidis and S. hominis which are known commensals of the skin, anterior nares, and ear canals of humans²². 47.4% (111/234) were identified from touch plate surface sampling with the highest rates of positivity occurring at sampling sites on the floor next to the gowning bench, sample ID 22 (5GMP-ISO 8); the sink, sample ID 9 (3.5GMP-ISO8) and the area on top of the table, sample ID 21 (5GMP-ISO8).

A total of 162 isolates were identified in ISO 7 cleanrooms (Supplemental Figure 2). 59.3% (96/162) were identified from touch plate samples. Of the 96 isolates, 38.5% (37/96) were identified as Bacillus species, predominantly B. aryabhattai, B. megateium, B. beringensis, B. cereus, B. licheniformis. These species were frequently found at sampling sites near the floor of BSC, sample ID 33 (5GMP-ISO 7); centrifuge, sample ID 34 (5GMP-ISO 7); top of fridge, sample ID 35 (5GMP-ISO7); and the door handle, sample ID 18 (3.5GMP-ISO 7).

There were 2 occurrences of microbial isolates detected in separate ISO 5 BSCs via viable particle air sampling. The identified bacteria were 1 Staphylococcus hominis organism and 1 Terribacillus goriensis/saccharophilus organism. Viable particle sampling took place during dynamic testing while active manufacturing was ongoing in the biosafety cabinet.

Positive environmental monitoring samples did not correlate with positive product sterility results.

During the time frame included in this analysis, 526 cellular therapy products were manufactured in the cleanroom facilities. Of those, 7 (1.3%) products were identified as positive for microbial growth during manufacturing and 3 (0.6%) during post-infusion analysis of the final product cryovial after thawing and transferring to a syringe for infusion. Identified organisms included 3/12 (25%) Acinetobacter baumannii, 3/12 (25%) unidentified Gram-positive bacillus, 1/12 (8.3%) Staphylococcus warneri, 1/12 (8.3%) Cutibacterium acnes, 1/12 (8.3%) Staphylococcus lugdunensis, 1/12 (8.3%) Comamonas acidovorans, 1/12 (8.3%) Gram-variable bacillus, and 1/12 (8.3%) not identified organism. Interestingly, the 3 positive post-infusion samples were all found to contain A. baumannii, which was not detected in any of the cleanroom facility environmental monitoring samples. Overall, positive product sterility testing results did not associate with out of specification EM samples.

DISCUSSION

The goal of an effective EM program is to track and trend key environmental attributes to identify any potential issues in the manufacturing facility prior to any impact on clinical products. Here we provide an overview of the establishment and evolution of our EM program and present nearly 10 years of real-world data with the intent of guiding the development of EM programs in early-phase CGT programs. These data are based on the needs of our facility and program – including the type and complexity of the cells manufactured, the design of the facility, the number of personnel in the clean room, the types of microbes present in the environment, and the degree of open manipulations being performed. Nevertheless, the information provided here can be used as a framework from which to build a program specific to each facility. Each program would then naturally evolve and adapt to address deviations, excursions, and areas of concern to ensure compliance with cGMP and ultimately ensure the safety of the final drug products delivered to patients.

The demonstrated utility of our EM program for effectively detecting environmental trends is shown in Table 1. In 2014, touch plate floor samples identified a sudden increase in viable organisms. This prompted an investigation which led to the identification of cleanroom-specific shoes as the source of contamination. Based upon these findings, the policy was revised to ensure routine disinfection of cleanroom-specific shoes, and ultimately shoe covers were required in all areas of the facility regardless of classification. In another example, monitoring of air quality identified an upward trend in non-viable air particles in both ISO 7 and ISO 8 classified spaces in Q1 2016 (though counts were still below the alert limit; Figure 3). Continued monitoring identified marked decreases in non-viable particle counts correlating with HEPA filter changes. This data helped support a new policy outlining the assessment of HEPA filters at the time of cleanroom recertification and their replacement annually if necessary. Furthermore, the existing institutional policy on HEPA filters was impacting operations within our facility, therefore, we decided to create an independent policy for managing HEPA filters within our facility.

A final example of EM trends used to identify areas of improvement occurred in 2018 when in-process samples from two consecutive products were identified as positive for bacterial growth, sparking a root cause investigation. EM data were evaluated and used to identify or eliminate potential root causes. Interestingly, touch plate samples from inside the incubator were within expected limits. However, additional viable air particle samples were taken from within the incubator as part of the investigation and counts were found to be outside the acceptable range (Table 2). Ultimately, the humidity chamber of the incubator was identified as the source of contamination. As a corrective action, multiple components of the incubator including CO₂ filter, humidity chamber filter, water bottle, and HEPA air filters were replaced annually. Because touch plates alone were unable to detect the contamination, we amended our EM sampling plan to routinely include viable air particle testing inside the incubators.

Characterization of all identified organisms present within cleanroom facilities provides vital information for any effective EM program. This is useful in identifying any trends in the type or frequency of microorganisms identified in the facility, whether there are any adverse trends, as well as the effectiveness of cleaning and disinfection in terms of microbiological testing⁴. An EM qualification study provides an ideal opportunity to initiate a compendium of organisms identified in the facility that is routinely updated thereafter²³.

EM programs at academic institutes may vary depending on several factors including the type of product manufactured, type of facility (classified vs non-classified), and phase of manufacturing. In our experience, most academic cGMP facilities perform some kind of environmental monitoring based on their facility and operations and should have performed a risk analysis to determine the extent of EM needed in addition to other operational controls. A facility that processes type 361 products (for instance, routine products for hematopoietic stem cell transplantation) in a non-classified space may not perform as extensive of monitoring compared to a facility processing type 351 (GMP) products. Moreover, EM frequency may vary between programs – daily, weekly, monthly, semi-annually, or even annually – but in our experience, our weekly EM schedule tends to be more frequent than many of our peers which tend to perform EM on a monthly or semi-annual basis¹⁰. Nevertheless, it is important to qualify the EM program at each facility and regularly assess its effectiveness.

While the EM program described herein has been able to detect excursions and mitigate risks, as with all cGMP, it is an evolving program that is guided by data, trends, and deviations. As such, there are limitations to our program and the data as they are outlined here. For one, counter to some peer academic institutions, we do not routinely perform real-time monitoring during manufacturing. Because dynamic monitoring requires active processing, it is not always possible to align the Quality Control/EM schedule with the manufacturing schedule, and thus we designed our program to monitor on a weekly basis. Additionally, although we were able to detect many mold and fungal species by incubating the media plates at 30–35°C, some environmental contaminants grow better at lower temperatures, thus these microorganisms may be missed by our current methods²⁴. The use of dual temperature incubations may expand the breadth of environmental organisms detected and is currently under evaluation at our facility. Lastly, we currently use the BacT/ALERT^® system for sterility testing of the final drug products, relying upon the core services available through our hospital’s microbiology lab. Although this method has been qualified for the types of products manufactured in our facility, it is not the compendial method listed in USP <71>²⁵.

Consistent with the current literature, we have not seen a direct correlation between positive sterility results of the product and out of specification environmental monitoring events²³. One exception was the incident in which the humidity chamber of an incubator was found to be the source of product contamination and expanded EM was utilized for root cause investigation purposes. While EM alone is not sufficient to protect products from contamination – rather having all operational controls (Figure 1) in place helps minimize any product sterility failure events – EM provides data on the functional state of operational controls within the facility and is a valuable tool when investigating, and ideally mitigating, any sterility failures²³.

In conclusion, we report here the establishment, ongoing development, and evaluation of effectiveness of an EM program in an academic CGT manufacturing facility. The program has successfully identified excursions in the EM program and in one case led to the root cause identification of successive positive sterility events. While effective, the program itself has undergone several modifications and continues to evolve based on trends, changes to procedures, facility parameters, and regulatory guidance.