Introduction
Recently, there has been an increased focus on contamination control activities for pharmaceuticals due to updated requirements in the European Union’s Annex 1 Update. (EU, 2022). The recent update of Annex 1 (ISO 14644-1, 2022) has specified the need for a contamination control strategy for both sterile and non-sterile products. Recently, a contamination investigation was conducted by a group of investigators from the Centers for Disease Control and Prevention, the Food and Drug Administration, the University of California San Francisco School of Medicine, and the Utah Department of Health (collectively designated as the “regulators”). The investigation was instigated to investigate post-transfusion sepsis that was attributable to bacterial contamination (i.e., polymicrobial contaminated apheresis platelet components) in a platelet collection set manufactured in the United States. The investigation was conducted during the time from May 2018 until December 2022 and was subsequently published. (Kracalik, et al., 2023). In the referenced article, the appendices are useful in understanding the sampling procedures and the organism identification procedures utilized by the regulators.
The contaminants found in these products included Acinetobacter calcoaceticus-baumannii complex (designated ACBC) and staphylococcus saprophyticus, which were isolated from both patients and components. The ACBC was found in six patients and six transfused platelets samples. Staphylococcus saprophyticus was found in four patients and four transfused platelets samples. Seven sepsis cases were identified from six platelet donations from six different donors in patients from six different states. Unfortunately, three patients died. (Kracalik, et al., 2023)
During the investigation (more than a year after the sepsis events), the regulators conducted sampling at a variety of donation centers, hospitals, and manufacturing locations (affected by the incidences of contaminated products). Whole-genome sequencing was performed on the majority of the contaminants, and it indicated that environmental isolates from one manufacturer’s location were closely related genetically to patient and platelet isolates, suggesting to the authors of the report that the manufacturer was the most probable source of recurrent polymicrobial contamination. (Kracalik, et al., 2023)
This seems like a nice and tidy investigation that was completed which found a definitive root cause for the contamination event. Individuals in pharmaceutical facilities recognize that it is often difficult to find a definitive root cause, especially based on environmental monitoring data. As is often the case, one might say, not so fast! As we perform the contamination investigation, do we have responsibilities to do more than just find the genetically similar organism in the facility?
This paper identifies some additional considerations that may need to be addressed before assuming that the probable root cause has been found. It also provides food for thought on when it is appropriate to close out an investigation.
Increased scrutiny of this investigation led to many additional questions and considerations that were not addressed in the published report. Additionally, the authors of this paper found some practices used by the regulators to be quite different from what we experienced when working in pharmaceutical facilities (this term is used generically to be both pharmaceuticals and biological products).
Consideration of Other Factors as a Source of Contamination
As part of the regulatory investigation, environmental sampling was conducted at blood establishments and healthcare facilities that reported septic transfusion reactions. At the device manufacturing facilities, environmental sampling was collected according to the facility’s established environmental monitoring program. All samples were sent for whole genome sequencing.
Whole-genome sequencing indicated that environmental isolates from a manufacturer’s location were closely related genetically to patient and platelet isolates, suggesting to the authors of the report that the manufacturer was the most probable source of recurrent polymicrobial contamination. (Kracalik, et al., 2023). However, the most closely related organism at the manufacturing site was identified by a less stringent identification method. (Note: The report states “a subset of the manufacturer’s environmental isolates were sent to CDC for Species verification and WGS.” It also states that “ one ACBCisolatese cultured from...manufacturing facility underwent multilocus sequence typing at an outside lab. Results were sent back to CDC for additional analysis, but the isolate was not available for WGS.”
Interestingly, the same organisms linked to post-transfusion sepsis were found in environmental samples gathered from all the healthcare facilities. It is not clear to the reader of the regulatory paper for the discrediting of these samples. Hospitals were not considered as the potential contamination source as evidenced by the contamination not being the source and was not pursued further even though these events occurred during the pandemic CDC reported a higher incidence of hospital-acquired infections during this timeframe. (CDC, 2020, 2021). Should this have been investigated further? From a pharmaceutical company point of view when conducting the investigation we would have also considered the hospitals as a potential source and at minimum would likely have concluded at this point that either the manufacturer or the hospital could have been considered the source of the contamination.
What other factors should be considered? Sterilization Malfunction? Handling and Storage? Environmental Factors? Human Error? Contaminated Blood?
The contaminated products considered in this investigation were terminally sterilized in a steam sterilization process (as indicated by other details found on product information websites).
The chance of a terminally sterilized product from a manufacturer being non-sterile is remarkably low. Manufacturers of medical products must demonstrate the microbiological safety of their products. Validation of the sterilization cycle is performed using biological indicators and may also include chemical indicators to demonstrate that there would be a very low probability of a non-sterile unit. This is expressed as a Sterility Assurance Level (SAL) of 1 x 10-6 and sterilizers are regularly requalified to confirm they are operating within the approved parameters. (FDA, 2024). A SAL of 10-6 indicates that post sterilization, the probability of a non-sterile unit is one in a million. To ensure the sterility of the finished product, additional steps are routinely taken such as conducting sterility tests before each batch is released or confirming the sterility results via the parametric release sterility indicating test, if parametric release has been FDA approved. Parametric release is a method, that must be FDA-approved if used, that replaces the sterility test method with a tightly controlled manufacturing process that assures the sterility assurance level will be acceptable if these validated parameters are met. The packaging of the product is also verified to maintain the sterility of the product throughout its shelf life (i.e., container closure integrity or package integrity testing).
Manufacturers are further required to comply with 21 CFR 211, the FDA’s Current Good Manufacturing Practices, which outline the requirements for controlling microbiological contamination. Given these stringent manufacturing requirements for sterile products, it would take a catastrophic failure of the sterilization cycle and/or extreme contamination of the production environment to yield a nonsterile product. One of the perplexing observations made regarding the regulatory report is that there is no explanation of the sterilization cycle used for the product. No comments indicated any issues with the sterilization cycles. RA review of the efficacy of the sterilization cycles to prevent contamination of products would have been a key component of this investigation at a pharmaceutical facility.
For the specific products in the regulatory paper, there is no consideration of the additional procedures conducted to reduce the likelihood of contamination, e.g., the use of an e-beam sterilization cycle after an additional manufacturing step conducted at the manufacturer, and pathogen inactivation steps used with the product (in the machine used for preparing the product for administration). The use of at least three different steps to ensure the “sterility” of the product further reduces the likelihood of contamination for the finished product. If contamination occurred, the contaminating organism (which was not a spore-former) would have to be resistant to at least three different types of sterilization methods. Furthermore, if the contamination happened before sterilization or in the first sterilization cycle there is no rationale provided to support that the organism could still grow after the other two cycles. From a pharmaceutical perspective, we likely would have looked at/for studies to show that the contaminant could survive all the various inactivation procedures. The organisms found as contaminants are not spore-forming, so all of the resistance required for inactivation would seem to be diminished.
In the unlikely event, that the sterile processed product becomes contaminated, one must consider how long can bacteria survive in the product. Bacterial contamination of sterile products is dependent on the type of contamination, the nutrient source available, and the environmental conditions to which the product is subjected. The growth of bacteria follows an exponential pattern with a lag time before contamination can be detected. In addition, growth conditions (e.g., temperature, oxygen, nutrients, moisture) must be available. There is no description of these types of evaluations being conducted or available in the literature for these contaminants in the regulatory report. In a pharmaceutical setting, we would be unlikely to declare a documented root cause without being able to show how the organism got into the product and how long it could potentially survive in the product under all the harsh conditions to which it was subjected.
Once a sterile medical product is received at a hospital, there may be increased handling of the product. According to the CDC in their Guideline for Disinfection and Sterilization in Healthcare Facilities, it states that “the contamination of sterile items is event-related, and the probability of contamination increases with increased handling”. (CDC, 2008) There can be opportunities for contamination each time the product is handled from human contact or handling, packaging damage because of handling, and even improper storage. The regulatory report provided does not include details of how these types of concerns were reviewed and/or addressed relative to the contamination event. In a pharmaceutical facility, an extensive review of the packaging would have been conducted to look for small nonintegral packaging leaks, which are not addressed in the report.
It was not clear in the report why was the hospital environment not considered as a source of contamination. The hospital environment appears to be a more likely source for polymicrobial contamination given its role in hospital-acquired infections (HAI). The timeline of these events (post-transfusion sepsis), particularly during the pandemic, is a significant factor. The CDC COVID-19 Impact Reports for the years 2020 and 2021 as well as publications in journals like Infection Control & Hospital Epidemiology, and the American Journal of Infection Control reported a continued rise in healthcare-associated infections (HAI) during 2020 and 2021. In fact, according to a report by the CDC in 2021, “the increase in HAI was particularly noticeable in central line-associated bloodstream infections, catheter-associated urinary infections, ventilator-associated events and methicillin-resistant Staphylococcus aureus bacterium” (CDC, 2021). The pandemic placed extraordinary pressures on hospitals with supply shortages, increased patient load, overworked staff, and perhaps a lapse in personal hygiene and disinfection procedures of devices and equipment between uses.
The CDC has acknowledged that although the blood supply is safe in the United States, bacterial contamination can be difficult to detect. Blood donors are screened through a health assessment and testing the blood sample for infectious diseases. However, in very rare instances a donor’s blood can be contaminated but because the level of infection is low, it would not be detected through routine screening. Bacterial contamination of platelets is especially susceptible because platelets are stored at 22°C which benefits bacterial growth. (CDC, Blood Safety).
When an Organism is Implicated in a Contamination Event, Is Location Important?
In the referenced investigation, samples were collected from many different locations. Extensive detail is provided about the various samples collected and how similar they may or may not be to the sepsis reaction organisms. However, little to no explanation is provided as to whether it is logical or likely that a contamination found in this location could be the source of the contamination event.
If one has a large production facility, say with multiple production lines, does a consideration in the investigation need to include where the organism is found? Is it enough to say that the organism is genetically similar (or the same) as the product contaminant, without understanding how the organism got from where it was located to the manufacturing area or process where the product was supposedly contaminated? As an example, if there are seven production lines and the genetically similar organism is found on line 7, but the believed contaminated product was produced on line 1, is it appropriate to end the investigation at this step saying that this organism is the product contaminant in question? Does one need to consider whether the production steps/processes would provide a method of transport from one production line to another, e.g., use of carts, transfer of materials, or the like? What if you can’t figure out a pathway for the organism to get to the production environment of the contaminated product?
Do we have a responsibility in an investigation to identify or evaluate how the genetically similar organism could have gotten from line 7 to line 1 before claiming it to be the contaminant? Alternatively, is having the contaminant somewhere in the facility sufficient to claim it is the most probable source of the contamination, even if you cannot conceive of a method of transfer of the organism to the affected production line or area?
When an Organism is Implicated in a Contamination Event, Is the Concentration (Amount) of the Organism Detected Important?
Assuming an organism is isolated which is identified as the same or like a contaminant, does the amount of the organism found matter? Is finding a single CFU in a year sufficient to say that this organism is responsible for the contamination event? Does it make sense that the level of contamination found can support a claim of gross or medium or any contamination of a product? Maybe just a claim of some contamination?
The Parenteral Drug Association owns and operates a Training and Research Institute (TRI) in Bethesda, MD. John Lindsay and David Matsuhiro coordinated the Aseptic Processing Training Program for several years. A key component of this course included executing a media fill-in for each one-week-long session of the course. The instructors for the course thought it would be a good idea to have a media fill that fails in a way where it would allow the students to find a clear, assignable cause for the failure (Lindsay, 2005).
As such, the instructors inoculated different sites along the manufacturing process with contaminants at low levels, defined as 10 – 100 organisms. Instead of surprising students with a failed media fill, they had instructors with an acceptable media fill trying to figure out what went wrong (Lindsay, 2005).
For the next media fill the instructors tried inoculating surfaces with higher levels of concentration of contaminants, that is 100 – 1000 organisms. Once again, the instructors were surprised, not the students (Lindsay, 2005).
Inoculated sites in the early studies included locations like the isopropyl alcohol spray bottle solutions used to disinfect surfaces and operator’s gloves during processing, stopper hoppers, and other product contact surfaces, immediately close to the filling needles, and the like. Many of the sites selected were considered “high-risk” sites by regulatory investigators in the presentations given. One of the more interesting studies was conducted with the operators not wearing any type of cleanroom garb while performing the media fill. Even though this media fill passed, it is not a recommendation to remove cleanroom garb (Lindsay, 2005).
Finally, a study was conducted contaminating sites with up to 10,000 cfu. This study finally resulted in some detectable contamination in the media fill. (The contamination level was only a small number of positives, e.g., 1-2 contaminated units.) These studies were media fills, for which the media used instead of the product would be growth-promoting and conducive to growth (Lindsay, 2005). Depending upon the formulation of the specific product, it may be even more difficult to contaminate the product formulation. There are product formulations that may be bacteriostatic or bactericidal to potential contamination. The product formulation needs to be considered in assessing the likelihood of contamination.
In addition to the concentration of the contaminant being important, one should also recognize that the design of modern air handling systems includes pushing air away from things blowing directly into the pharmaceutical containers making it even more difficult for environmental isolates to get into product containers.
When considering the PDA studies, are presented only to show that it is not as easy for an environmental contaminant to physically get inside the container, grow, and contaminate the product. It is not advocating the performance of poor cGMP procedures.
Even if the organisms could successfully navigate their way into the container, the organisms would require sufficient nutrients to grow and thrive until the time the product is used for its intended purpose.
If one is to consider that a contaminant got inside the product, would it be necessary to show the ability of the contaminant to survive in the product or would it be sufficient to show it stays in bacteriostasis? Do you need studies to show how long it can survive or would have to survive to be a true patient contaminant? The same types of concerns about growth may need to be evaluated for the container, e.g., can the organism survive in the container? If so, how long can it survive, and under what specific conditions? In claiming an organism is the probable cause of a contamination event, how much information is needed to support this evaluation?
How Closely Does an Organism Need to Match to Be the Probable Cause of an Event?
In considering genetic matches and whether you have the same organism or not, you need to explain your rationale for “equivalence”. Does the testing for the test organism and the “probable source” be performed on the same machine or only the same type of machine? Should the identifications use the same technology for the identification method? For example, are whole genome sequencing and multi-locus sequence typing testing sufficiently close enough to make an identification that the probable source was found when the reference organism and the probable source are tested on the two different technologies? What basis for this determination can be utilized? It has been well established that different technologies used for identification, e.g., Vitek, Gas Chromatography, and API strips can yield different identifications.
Whole-genome sequencing (WGS) or Next-Generation Sequencing (NGS) is represented as key for identifying the source of the contamination event. When using these technologies, does it have to match 100%? If it does match 100%, is that suspicious? How close is close enough? Does the time between testing these organisms (e.g., time of contamination and finding the isolate two or three years later make a difference?
It is common in most pharmaceutical companies to follow up on the investigation of organisms to perform an analysis of the specific strain involved. Even if one looked at every base in the genome and had a 100% match, one appropriately should say that the two organisms are indistinguishable based on the genes analyzed. It would not be appropriate to say they are identical since there can be differences at the epigenetic level, e.g., methylation.
MMulti-locusSequence Typing (MLST) was used for the sequencing of 400–500 DNA base pair fragments of seven housekeeping genes to allow small variations within a species to be detected.
When a Potential Contaminating Organism Match is Found – How Important is Timing?
In the investigation conducted by the CDC, referenced in this article, the identification of the organism that was claimed to be the most probable cause of the contamination was sampled and identified months after the sepsis events occurred (and was not found within the same production area as the affected “contaminated” product.)
Consider the case where sampling is performed many months after the initial contamination event occurred. In this consideration, say that an organism found somewhere in the facility is a strong genetic match to the contaminating organism. Does that automatically mean that it is the most probable cause of the contamination? However, in this evaluation, environmental monitoring performed at the time of batch production did not even have the same genus and species or the supposed contaminant present during batch production. What is the scientific rationale for claiming that the organism is a contaminant? How do you even know that this specific organism was present at the time the potentially contaminated batches were produced? Was the organism found in an area where it is believed that it could have gotten into the product?
Being a genetic match to an organism does not guarantee there aren’t other organisms out there that could be a “match”. For example, could you have closely related organisms present at the hospital/transfusion sites as well (which in this case, there were)?
There can be many different sources of organisms that could have a similar or closely related identification (like from a variety of sites). In the CDC investigation, is it appropriate to disregard closed related organisms at other sites?
If you are making a claim an isolate is the cause of the contamination, what is the rationale for accepting the timing of the isolation and identification.?For example, if it is isolated many months after the “contamination event”, how do you correlate finding the organism at another time long after the contamination event to when the contamination supposedly happened?
Adaptability of the Microorganism
Another consideration is that the Acinetobacter baumanni complex is that it is described as an organism with an “adaptive nature of this organism is unrivaled…” There are a variety of different ways for this type of organism to acquire or adapt, e.g., horizontal gene transfer, or natural transformation. (Castanheira, et al., 2023) This adaptive nature of the Acinetobacter baumannii complex makes it difficult to correlate organisms isolated and identified months apart, e.g., how do we know whether the organism adapted to have the so-called “novel” clusters months later?
If Multiple Organisms are Cited as Contaminants, Do You Need to Recover All of Them to Claim the Probable Cause?
In the cited article by the CDC, the sepsis was claimed to be “polymicrobial.” When you have a polymicrobial contaminant, and you find one “potential source,” is that sufficient to claim you found the source? Does it have an effect if you only find one organism, when multiple organisms were found in the contaminated patients? What should be considered by the investigator if one of the organisms is a commonly found organism like Staphylococcus? Does that change your determination? The goal of an investigation of a polymicrobial contamination event is to identify the root cause of the common source of the contamination event. Were all microbial isolates in this contamination event assessed by use of Whole Genome Sequencing? This includes environmental samples from the manufacturing facility and other potential areas where organisms found by the investigator could be isolated that contributed to this contamination event. If only one of the organism types is found, a risk assessment should be conducted to decide if the bacteria isolated represents the source of the polymicrobial contamination. In the cited report, the Staphylococcus contamination is not addressed as being found in the manufacturing facility cited. Should this still be a definitive root cause if only one organism was found?
What Parts of the Manufacturing Process Need to be Considered in Determining the Probable Root Cause?
The product in question in the cited report was terminally sterilized at the manufacturing site before commercial distribution (apheresis bags). In the United States, large-volume parenterals are expected to be terminally sterilized except in some cases where there is a risk associated with using terminal sterilization, e.g., the manufacture of antineoplastic drugs. It is customary practice and expectation to validate steam sterilization cycles using biological indicators, which are a more stringent challenge than the use of routine plant bioburden. There are a variety of biological indicators commonly used in the United States, e.g., Geobacillus stearothermophilus, Clostridium sporogenes, Bacillus smithii, and Bacillus subtilis 5230. While the thermal resistance value for these biological indicators varies from approximately 0.5 minutes at 121°C to approximately 2.5 minutes at 121°C, all the thermal resistance values are much higher than that of routine bioburden. (Agalloco, 2017)
Within the cited investigation report, there is no indication of any testing or evaluation to support a claim that the contaminating organisms if present in the facility were able to survive the terminal sterilization process. DData could have been used by the investigator in this evaluation to include information like a determination of the thermal resistance of the probable contaminating organisms to show that it was more resistant than the biological indicators used to qualify the terminal sterilization cycle and could potentially survive the sterilization cycle.
Even if one were to assume that both contaminating organisms were present in the facility, in a location where they could get into the potentially contaminated product, and could survive in the product, it could not be the source of the contaminated product unless the organisms were able to withstand a properly validated steam sterilization cycle. Failure to show that it could survive the sterilization cycle and/or that the sterilization cycle was not properly validated and operating in a state of control, significantly reduces the likelihood of these organisms being the probable root cause of the contamination event. In the specific products of the investigation, after steam sterilization, the product was also subjected to an E-beam sterilization cycle, and a pathogen inactivation process at the hospital where the product was administered. It is highly unlikely that a non-spore-forming organism could survive all of these harsh processes.
Conclusion
While finding closely related organisms in a facility can significantly aid in the determination of the most probable root cause, it alone is not sufficient to identify the root cause of the contamination event. It also needs to be reasonable to believe that the organism can survive the manufacturing process and has a route to get into the product. Some of the examples provided in this document provide additional information in assessing when you have appropriately completed your contamination investigation. In this particular example, the “source” of the contamination was identified long after the contaminated product was manufactured. In most pharmaceutical facilities, it would be unlikely to look at subsequent monitoring results to “backtrack” the contamination to a previous contamination event. It may be useful to incorporate considerations within your contamination control strategy or other documents on the conduct of investigations. It could be extremely useful to guide when you have sufficiently investigated, e.g., the types of considerations provided in this article to close out the investigation in your site’s applicable procedures. One might also want to consider what type of post-production actions are taken in investigating product complaints, e.g., how long should data be looked at after the product is made.
Literature Cited
Agalloco, J.P. (2017) Kill the Bioburden, Not the Biological Indicator. BioPharm International 30 (4): 50-52. Downloaded from: Kill the Bioburden, Not the Biological Indicator (biopharminternational.com) on January 9, 2024.
EU (2022) European Commission. “EU Guidelines for Good Manufacturing Practice for Medicinal Products for Human and Veterinary Use, Annex 1: Manufacture of Sterile Medicinal Products.” Brussels: European Commission, August 22, 2022
Kracalik, I.; Kent, A.G.; Villa, C.H.; Gable, P.; Annambhotla, P.; McAllister, G.; et al. (2023) Posttransfusion Sepsis Attributable to Bacterial Contamination in Platelet Collection Set Manufacturing Facility, United States. Emerg Infect Dis. 2023;29(10):1979-1989. https://doi. org/10.3201/eid2910.230869. Posttransfusion Sepsis Attributable to Bacterial Contamination in Platelet Collection Set Manufacturing Facility, United States - Volume 29, Number 10—October 2023 - Emerging Infectious Diseases journal - CDC
Lindsay, J. (2005) Chapter 14 Environmental Impact on Media Fills. In Moldenhauer, J., Ed. (2005) Environmental Monitoring – A Comprehensive Handbook Volume 2. PDA/DHI. Bethesda. 187 – 216.
CDC COVID-19 Impact on HAIs in 2020 Downloaded from http://www. cdc.org CDC Healthcare-associated Infections (HAI)>All Data>Data Portal on January 13, 2024
CDC COVID-19 Impact on HAIs in 2021Downloaded from http://www.cdc.org CDC Healthcare-associated Infections (HAI)>All Data>Data Portal on January 13, 2024
CDC (2020) 2020 National and State Healthcare-Associated Infections Progress Report Downloaded from http://www.cdc.org January 13, 2024.
CDC Bacterial Contamination of Platelets. Downloaded from http://www.cdc.org CDC>Blood Safety>Diseases and Organisms on January 13, 2024
FDA, January 2024. Submission and Review of Sterility Information in Premarket Notification (510(k)) Submissions for Devices Labeled as Sterile Guidance for Industry and Food and Drug Administration Staff
21 CFR, FDA Current Good Manufacturing Practices
CDC (2008), Sterilizing Practices: Guideline for Disinfection and Sterilization in Healthcare Facilities. Downloaded from http://www. cdc.org January 13, 2024
CDC Bacterial Contamination of Platelets, Downloaded from http://www.cdc.orgCDC>Blood Safety>Diseases and Organisms January 13, 2024
ISO 13485 Contamination Control Procedure & Product Cleanliness
EU CGMP Annex 1 2022
Author Details
Victoria Galliani, Mary Griffin, and Jeanne Moldenhauer (jeanne@exellpharma.net), Excellent Pharma Consulting, Inc., Lake Villa, IL
Publication Details
This article appeared in American Pharmaceutical Review:Vol. 28, No. 2March 2025Pages: 30-35
Subscribe to our e-newsletters.
Stay up to date with the latest news, articles, and events. Plus, get special
offers from American Pharmaceutical Review delivered to your inbox!
Sign up now!