Pierre: We are going to start this new episode in which we are going to talk about rapid microbiology method with a specific topic on how to choose the best rapid microbiology method. And I have a special guest for that, who is Dimitra. So hi Dimitra.
Dimitra: Hi Pierre. Thank you for having me.
Pierre: I'm very pleased to have you here. We met last year in the Congress called Pharma Lab. And the first thing you told me, I have tried almost all the rapid microbiology method in my lab. So I could not believe it at the beginning. say, okay, you look young and it's possible that you've tried maybe two or three of them, but not all of them. then we discussed. And I find out that it's true. You try almost all of them. So I was amazed by that and say, okay, I need to have her in my podcast. But maybe before, can you introduce yourself, Dimitra?
Dimitra: Yeah, so I'm as you said, I'm Dimitra and I'm an ex-academic microbiologist. I'm specialized I was specialized in environmental microbiology for a very long time. and after my academic career I've turned into an industry enthusiast for the last four to five years. well as in terms of education, I from very early on I was interested in microbiology, so I started already from my bachelor studied studying microbes and their applications, then I finalized my PhD in the group of MyGetten here in the Netherlands studying microbial physiology, but for the great chunk of it I was developing methods to detect microorganisms that harbor specific enzymes that are related to the carbon and nitrogen cycles. Then I was very privileged and I joined the lab of Michi Wagner and Holger Dimes in the University of Vienna where I got to play around with very cool spectroscopic instruments, advanced microscopy, genomics, you name it. They have all the cool toys. And there my focus was again on method development actually, more than physiology. So ultimately towards the end of my academic career I acquired quite big funding. for a project that was aimed to develop this new spectroscopy-based methods to detect previously very, very hard to find and culture microorganisms. So my goal was actually a rapid method, the development of a rapid method for a process that normally would take decades, and my goal was to reduce it into weeks. So that is how I ended up I ended my academic career. And after that I moved back in the Netherlands and I joined Cell Point that later became the Galapagos, currently Lakefront Biotherapeutics. there I worked as a principal scientist within both the analytical development and the QC departments, so I had a double role. And there I was responsible for sometimes even co-developing, but evaluating and validating alternative and rapid micromethods. And transferring them and implementing them to the QC setting. And finally, after Galapagos winded down their cell therapy arm, I recently joined Johnson Johnson Innovative Medicine, where I'm very proud to be part of the QC Microbiology department, where I get to experience a very big organization. global one with commercial products but I also get to still do what I like which is implementing rapid micromethods and evaluating them as part of the global work stream. So
Pierre: Okay, great. That's a long experience. And especially in the last few years, you dedicated it to rapid microbiology method applied to pharmaceutical products. And this is what I would like us to talk about today. Maybe before we talk about how to choose the technology, Can we have a quick overview on the technologies that you use?
Dimitra: Sure. I don't know how short we can make it based on your starting thing, but of course we have different assays within a QC microbiology setting, right? let's take it assay by assay. a part that a lot of people forget that we can implement very, very easily, I would say also an alternative or rapid microbiological method is environmental monitoring. There we have a lot of different vendors providing mostly image-based detection. of microcolonies so we can reduce the time to detection from yeah five days or so depending on the process that we have in place to one or two days depending on the species. There as I said we have a lot of alternatives we have pretty easy to implement systems with great primary validation packages with a caveat that some of these instruments do struggle a bit in the case of molds. So there we need to be a bit more careful. but most of them are able to automate incubation and readout and reduce this insane workload that in cases of sterile manufacturing and annex one compliant plans we have thousands of plates per day, right? So this is a very big burden in a QC lab that can very easily be automated. And yeah, the regulatory acceptance is just there, I think for most of the instruments there's no issue to implement from a health authority perspective. And it gives a quite big return of investment, I think, in terms of FT workload reduction.
Pierre: So it's not only quick, it as well reduces the manipulation and the labor.
Dimitra: Of course, because there is no ideally there should not be a manual readoff or some cases for example when a plate is flagged by the instrument as yeah out of specification then there is a manual readout of the operator but for example in a sterile manufacturing setting we expect one percent of the plates to be out of specification and then we can introduce an operator. important there that prior to implementation, QA needs to play a big role to ensure that when we discuss a 4I principle for GMP generated data, that these are mutually understood what a 4I principle means. So yeah. And we have technologies nowadays that also incorporate maybe a bit more sophisticated technologies that also incorporate AI components in the readoff of the plates.
Pierre: OK, so that was for the environmental testing and the use of technology that can detect micro colonies instead of larger colonies.
Dimitra: Yes. Okay.
Pierre: What about bio-burden or finished product testing?
Dimitra: Yes, so bio burden and sterility testing are sterility testing I think is the holy grail, right? So a lot of people are interested in a very f rapid sterility method, especially in the context of more advanced therapies as we have now CGTs and ATMPs. I would first, before we dive into sterility, maybe touch upon bio burden because There we have a bit less to talk about, maybe. bioburden, of course, can be either growth-based. We have some vendors offering growth-based detection and assays, and we have some vendors offering growth-independent detection. So for the growth-dependent, we have Again, incorporation of optical systems, right? Sophisticated microscopic imaging and sometimes a software component which can be either AI or machine learning, etc. And there we have detection, the same principle as what we have with the compendial method for a plate by a burden assay, but Now we reduce the time to result because we do not wait for the colonies to be as big as to be microscopically detected, but the specific analysis can detect the microcolonies in a sub size, which is normally I would say around 300 microns. most of the applications give such a sensitive detection. We have better distinction s sometimes compared to operators of one versus two colonies because some of the vendors offer time-lapse imaging as well, which helps us identify if that's one big thing that we see on a plate came from one or two or multiple colonies. so yeah, we we have again, of course, automation and digitalization and Yeah, that offers a time to result normally within a day or two compared to a five day normally incubation that we have for the compendial methods.
Pierre: So this one is for the growth base and same as what you discussed earlier, it's micro colony based technologies.
Dimitra: Yes. And analysis, image analysis, right? To when we offer a time lapse to see how the colonies develop. and if we have for example spreading microbes, sometimes we we can better decipher
Pierre: It's one of its...
Dimitra: Exactly. Then we go into the discussion of one or more than two hundred C of Us and we can easily make decisions based on the time lapse then on checking the growth of the colony at the end. And of course that is very important, right? Because it might seem for a rapid that two days is not as rapid to some people. but of course in a continuous production setting that's offers real power in terms of decision making because sometimes we move forward with our production whereas we don't have all the data to actually make informed decisions and that can save for a company a lot of money time etc so this has normally a very good return of investment it is since it's the same media the same thing that we detect, we detect microcolonies, so C CFUs. We have the same value that we detect, so it's a CFU unit. we basically measure growth, so all of these are equal to what the compendial method does. That makes it very easy to validate and implement for the health authorities. Okay. then What else do we have? We have growth-independent methods, and there we have a couple of more established players and some new kids in the block. right. so for bioburden we of course have solid phase cytometry, it's been around for decades. it's a single-cell detection, it detects It's normally combined with the fluorescent labeling and it can normally detect active metabolically active microorganisms. If you first filter your product on a membrane and then you have a laser scanning. and there we as I said we have single cell detection, which means that these technologies might even be suitable for products that are inherently less bioburden heavy so we do not fall necessarily in the sterility realm but we do have low bioburden products. there we have the health authorities a bit more sensitized right towards solid phase cytometry. It's a slightly it's relatively mature technology It's been around for a while. and we have a lot of vendors offering such solutions. smaller, bigger, it depends on the on the platform. there the time to result is actually quite impressive because we can have results up to a few minutes, so an hour. normally for bio burden testing to my experience with this newer technologies you do not necessarily need any incubation so the sensitivity is there to allow detection even without a pre-incubation step
Pierre: So we've seen environmental monitoring, bio-burden, for which both of them we need quantitative method. We need to count the actual number of microorganisms. Correct. And then we have the sterility testing, for which the type of test is slightly different because here we don't need to count, we just need to know if it's positive or if it's
Dimitra: Negative. Correct. It's a yes or no test. So we do not
Pierre: Not that simple, but yes.
Dimitra: I I know, unfortunately. So we move from the quantitative test to the qualitative tests. And indeed a sterility method should be absence presence of any potential microorganism. And this is something that I've seen quite few people struggle with, to be honest, because they struggle with a part of any. microorganism. a lot of vendors to my experience of course they provide a lot of primary validation data and of course they have to limit the data production to a point but they check the compendio panels and they say yeah I hands down I can detect everything. Well that's not true. And there we can discuss maybe a bit further down As to what we as customers should do. Then for sterility, we again have growth-dependent and growth-independent methods, right? And some of the growth-dependent methods are very, very mature technologies, they are widely implemented, especially in the context of USP 1071. and EP2627 which are for short they are they give guidelines for short self-life products that is seeing a very big rise especially because of the CGTs and ATMPs currently being produced. so first of all I think the most mature technology out there the one that is very easily accepted Is respiration based assays. Normally, with these assays, we detect active microorganisms that grow, and these require an incubation. Normally, this is seven days. Most of the vendors have validated those incubations for seven days. Some customers, to my knowledge, have revalidated the assays for Shorter incubation time, so five days, for example. Most of these assays are performed with compendial media, so proprietary media, but they mimic the compendial formulations. and they are based on the CO2 production by microorganisms mostly, which subsequently causes a drop of the pH in the medium. That of course causes that causes a change either of the color or emission of fluorescence from a sensor that is normally located at the bottom of these proprietary media bottles. these are highly automated systems, they continuously measure, so they measure normally every ten minutes, and they provide real time data, so bottles that become positive they are flagged normally immediately, so you do not have to wait until the end of incubation, which means that for heavily contaminated samples or for samples that might contain fast growing microorganisms, the re time to result might be less than the five to seven days.
Pierre: Seven days is the time to negative. Time
Dimitra: Yeah, correct.
Pierre: to positive can be everywhere in between.
Dimitra: Yes. and as I said, these are very mature technologies. they are very widely accepted by the pharmacopeyas. some of them have made them compendial for short self-life products and yeah they are clearly mentioned now in USB 72. other growth dependent methods is A to P biluminescence based methods. Again we have multiple vendors. Some of them are smaller, some of them are bigger methods, and there are a lot of customers including big pharma customers that have validated such methods. They provide the option to digitalize, of course, everything. They require less handling from the operators, and there, what we measure is the production of ATP from microbial cells So we have an incubation period And then lies of the cells and detection of the A to B that was produced.
Pierre: So here it's the end point.
Dimitra: Correct. It's always it's an endpoint assay. And although it's a growth based assay, the detection at the end is not C of U, so it's fluorescence units. so that is something to keep in mind for validation purposes and I think the new PDA Tier thirty three provides really great Guidelines on how to do that. Here, caution is that sometimes we have background, especially for matrices that have been either coming from cell cultures or contain cells, we might have quite a lot of background ATP, which we need to make sure that our assay is able to deplete. And also for cell-based. Matrices, it's normally a a part where we either chemically or mechanically remove the mammalian cells from the matrix. And yeah, every vendor has its own solution. but yeah, the options are there and the customers need to decide whether this is acceptable or not. I will say here that Any method that is chemical, of course, we need to check the vendor data and potentially also generate data ourselves to show that this chemical lysis does not influence the microbial burden because we know that yeah it it can influence the outer membrane of microbial cells as well.
Pierre: So you mean you have chemicals that are supposed to kill the mammalian cell and keep alive the bacteria and we need to make sure that it does not kill the bacteria at the same time as the mammalian cell.
Dimitra: Correct. And so far we've seen that in some cases that is the the case.
Pierre: In your personal experience, you mean?
Dimitra: Yes. Okay. we've seen that it's maybe even genus specific. So if we need to have this data, I would advise that this is checked in a wider panel of microorganisms. If the vendor does not provide this data, because a lot of vendors do provide data like that. Okay. We
Pierre: We're going to talk about primary validation later.
Dimitra: Okay, great. So growth-based method also is microcalorimetry, and there we have the detection of heat that is produced during the metabolism of microorganisms when they grow. And this of course means that all microorganisms produce heat when they grow. So these assays can detect a wide array of microorganisms because their detection is universal, so it's not specific. There we have quite a few pros, and which are of course, as I said, it's a growth-based assay, it requires a certain incubation period. And because it's a growth-based essay, the limiting factor is always our most slow growing microorganism. I think most primary validation data that I've seen, C acnes is the one that always causes a problem. That is true for all grace growth-based methods. I have a point which I forgot actually for this respiration-based assays. because there of course the acnes is our growth limiting factor, but caution needs to be also for fungal species. So not yeast but molds because they inherently produce less CO2. Which although they might grow relatively fast compared to C acnes, they do not accumulate so much CO two, so their detection might also be slow. Okay. Just a side note. So you might see that Exactly but you
Pierre: you might see them but they might not...
Dimitra: still get a negative result from the instrument. And this is very crucial because and that all vendors recommend is that at the end of the respiration based assays you perform also a manual check of the bottles for growth because of the molds. Now back to micro calorimetry the time to result is normally anywhere from 48 to 72 hours and that is because of See Acnes. it there some the vendors normally use compendial media and incubation temperatures, which makes it quite easy, I would say, from a regulatory perspective, to present that to a health authority because it's a growth-based assay, it uses the same media, it is yeah, it's it has an incubation period and the only thing that changes compared to the compendial sterility method is how you detect the the growth not what you detect. So I think this is although the technology itself is not as mature, I think that from a regulatory perspective it's should be easier than some other alternative and rapid microbiological methods to implement. And of course there the automation and digitalization and low workload in the lab is also always part, right?
Pierre: For all those technologies, is it direct inoculation or can it be filtered and then the filtered place into the vials or the bottles?
Dimitra: So for this the respiration based is direct inoculation. For microcalorimetry is also direct inoculation. And what else did we talk about?
Pierre: ATP
Dimitra: ATP yeah, most of the cases it's also direct to my experience.
Pierre: So then the limitation is the volume you can...
Dimitra: Correct. And I know for all of the methods, all of the principles that we discussed, that was a a small issue. I think most of the vendors now switch towards using like higher volumes, but of course from an operational perspective if you are Obliged to test a certain amount of volume. Let's say for ATMPs is 1% of a batch, which normally is few mLs, right? that's yeah, that that might be tricky if you can only inoculate, I don't know, a few micro hundreds of microliters per vial. You don't want to have an operator injecting tiny bottles.
Pierre: 10 or 20 bottles, yes.
Dimitra: Correct. that's increases of course mistakes, operational complexity, etc etcetera. So but I'm happy because most of the vendors now provide also alternatives for higher inocular volumes. So I don't see that as a very big issue. Okay.
Pierre: so we talk about growth-based method. Is there any other growth-based method available?
Dimitra: Not that I've worked with, I'm not aware that we are. Okay. Then we can switch to growth independent methods, right? There as before for bioburden, we have solid phase itometry. there we have filtering in the membrane, etc. etc. as we discussed before. fluorescent staining. And normally with the addition of a small pre-incubation. step then the sensitivity of the method increases to the point that it can be considered it can give an LOD of one C of U so it can be considered a true sterility test.
Pierre: Okay.
Dimitra: Of course again it's there is a pre-incubation part normally because we don't necessarily need so much growth but we do need metabolic activity to incorporate the the the dyes they are activity based dyes that to my experience has a time to result of 48 hours like two days basically so much faster than the compendial 14 days again automated digitalized workflows Very minimal operator handlings. part there, of course, whenever there is a fluorescent labeling, again, background is an issue. and for matrices that contain a high concentration of mammalian cells or other cell debris or any other high concentration of organic components. maybe cell lysis step to remove those parts from the matrix will be needed, but that depends of course on the matrix that we're talking about. Yeah.
Pierre: I've seen some vendors mixing solid-phase cytometry with cross-phase methods. So they start with a cross-phase in standard. So they filter, they add the TSB and the FTM, and then they incubate it for a short period of time, like 48, 72, three, four days. And then they use the solid-phase cytometry at the end for
Dimitra: Detection part.
Pierre: the detection, yes.
Dimitra: Yes. So the thing is there that we do not necessarily need growth though. We need incorporation of the dye into the metabolically active cells, and that can be independent of the doubling of the microorganism where we define as growth. It needs to build biomass because normally These dice target universal microbial enzymes. but so it needs to be live and active, but it doesn't necessarily need to grow in that sense. So yeah, that's that's a benefit, and that's why I would put it more towards the growth independent part.
Pierre: Okay.
Dimitra: yeah. there, of course, we have Spectroscopic methods, one that I have close to my heart because I worked a lot during my academic years, Raman spectroscopy and combined with D2O labeling in some cases, so heavy water or deteriorated water to be more scientifically correct, where we have Microorganisms incubated for a very short period of time of like one or two hours in the presence of heavy water. Then this heavy hydrogen gets incorporated into the membranes, into the lipids of the membranes of the cells. And then normally we have a filtration step where all the microorganisms are captured on a filter. which is scanned with a laser, and from this scanning we get Raman spectra from each cell. So we have single cell analysis, the LOD is very low. The time to result in without the incubation with D2O might be an hour, two hours as long as it takes to scan the surface of the membrane. and if we add the pre incubation step, we're talking about excuse me, couple of hours of time to result. There, the positive part is that Raman spectra can provide a sense of ID of the microorganism. so based on the spectra and how they look on the phenotyping region, we can and while we have let's say a database of reference spectra, we can identify the microorganism for root cause analysis purposes on at least the genus level. on the more negative side, although the technology has been quite widespread in food industry, It's not so widely employed, it's not widely employed in pharma. So yeah, the health authorities are yeah, they know it, but they're not as sensitized towards it. as I said, the positive part is that once we incorporate a labeling incubation. Then we also get information as to live debt, which can influence decision making, of course.
Pierre: Actually, I've never heard about this technology and I've never seen it. No, no, It looks quite impressive on paper or at least discussing about it. But I've never seen it live into a lab, even food labs. But I don't know all of the labs in the world. So this is something I will keep in mind.
Dimitra: I've personally worked a lot of it in development during my academic career, but I also I w I also had projects feasibility and co development projects with such companies and vendors. And yeah it's pretty cool.
Pierre: Yeah, it looks cool. I know that Raman, it's not new technology, but it has been quite difficult to implement it into, I would say, routine labs because it generated a lot of data and then you need to analyze them. And now with the AI, it seems to be much easier to deal with all those data. And I can see that many more spectroscopy Raman technologies are coming up because we know how to deal with the data.
Dimitra: Correct and this is another part which yeah, many people underestimate, but whatever has images and spectra that generates a lot, a lot of data. So you need to have the correct stakeholders when you identify a technology that you want to implement. also from other departments such as IT and infrastructure to actually tell you whether you can do it or not. Because in some cases with smaller companies, yeah, having ten terabytes of data being created over one assay that's not feasible. So yeah, I
Pierre: And now in the routine, what you want is something that's plug and play. So that means you put your sample into the technology and the technology tells you it's positive, negative, or if it's 10, 20, 0, CFU.
Dimitra: Yeah, but this this capabilities these technologies have, right? The automation. automation and digitalization, I think most of the vendors that develop ARMMs they strive for these two things. So it's very rare that you will find something that it's insanely manual or requires a lot of intricate sample prep, to be honest.
Pierre: Is there any other technologies you wanted to talk about or you experimented?
Dimitra: I think NAT based for sterility might be a good one. What do you think? We have, as we said before, QPCRs, digital PCRs, normal RT PCRs, but also very highly automated ultra fast PCR platforms. There of course we detect like specific targets and Most of the assays detect RNA gene, so 16S, 23S, 18S, 28S, and ITS regions. depending on the vendor, we have different levels of automation from fully manual PCRs to like fully completely automated, no precision handlings to just plug and play, as you said before. sterility tests, and there we have high specificity, right? We detect very specific genes. Sensitivity can be influenced by the matrix composition. Complex matrices sometimes give quite a bit of interference with a sample prep mostly, that can also influence the LOD at the end. most of the vendors provide really extensive data sets with inclusivity, exclusivity, amongst all the other critical assay attributes. And yeah, the challenge there is the fact that it can detect it detects DNA. And unfortunately for us, microbial DNA is everywhere. And clean rooms and isolators are not DNA-free environments. So ideally we need to be looking to close systems, systems that cannot be influenced by the surroundings, because yeah, good luck with all the false positives that you can get from a dirty DNA dirty environment. we also need to be cautious about the fact that it can detect live but also dead DNA coming from live and dead cells or damaged cells or extracellular DNA present in the matrices. And these are components that with a more sophisticated workflow can be taken into account and in some cases they are taken into account, so specific treatments. are happening during a sample prep to allow distinction from extracellular or DNA coming from dead cells, pre-existing DNA. Another part is reagents in my experience, because a lot of the enzymes even used in PCRs, because they are recombinant enzymes, they are produced by microbes, they are DNA contaminated. So we need to look into if we want to select something like that, we need to look very close into the workflow, into the consumables, into how it will look in our setting and what the risks are for false positives, and how the vendors deal with their consumables and manufacturing. There, however, I would say that These are ultra rapid methods, so time to detection of one hour or two hours or max three hours. And in the context of specific products such as ATMPs again, yeah, this might be a very big positive. And since we can implement assays on a risk-based approach for these products. Maybe the positives on time to result outweigh the the negative.
Pierre: Okay, so it's just a way to find the balance.
Dimitra: Yeah. And at the end maybe we have some more exotic technologies like dielectrophoresis combined with fluorescence labeling for live and dead staining. Yeah this is I I call it exotic, it's a very cool concept of capturing microbes in a The electrophoretic field. It's a specific capture, which is great especially for matrices that contain cells or high concentration of other debris, for example, because it's a universal microbial frequency capturing. So you can only see your tiny little cells being captured in these electrodes. There we have short pre-incubation for an activity for a live dead labeling. so we have simultaneously capturing and detection via fluorescence of live microbial cells. Yeah, it has a very high sensitivity, and it's it's can also be a very automated workflow. And yeah, the time to result normally will be in terms of hours, so two, three hours, and not more. And
Pierre: It's amazing, I've never heard about this technology as well. I'm learning so much.
Dimitra: Yes, and then there are a lot of smaller companies also, more startup companies, that can provide very nice combination of these technologies. So I would say that the sky's the limit and science is doing a lot and now slowly these These technologies are coming to the market and it's a matter of marrying the correct intended use with the correct technology.
Pierre: Okay, that's a perfect transition of what we are going to talk about now we have all those nice technologies and the challenge to pick up the good one, the one that fits the lab needs. And when we prepared that podcast, you told me that one of the most important things to do is before selecting the technology or just to help you choose the right technology, you need to build a user requirement specification. And this is something that most of the labs are not doing or doing very quickly. And when we discussed, you told me that it's very important to spend time on it. So we are going to spend a little bit of time on this one. And I would like to know for you what is a good user requirement specification.
Dimitra: Yes, a URS is a requirement to have, and that makes it sometimes turns it into just another paperwork we need to fill in many cases, right? It's another tick box to have, but in my opinion, this is not how it should be viewed, and I'm glad with the peer PDA tier thirty-three it also emphasizes that because A URS gives you the fundamental requirements from a scientific, an operational perspective, from a quality perspective, regulatory perspective, even and quite important from a business perspective as to what you exactly need and what is needed from your assay. So, in my opinion, this is a very very good tool to define the question you're asking the problem you're trying to solve am I trying to automate my or am I trying to release my badge faster right so there is a a a good place to define the the question and then try to in as much detail specify what success would look like for you so what exactly would you need it to do? And that is a great then document to start looking into technologies and vendors that can provide the answers to these requirements and not the other way around. Because I see in a lot of cases we first find a vendor and then we build our URS based on the technology, but I think it should be the other way around. so a good URS will have to have of course the Question defined. I need a sterility test for the release of my batch. And then the scientific requirements, the LOD, the LOQ, if we're talking about bioburden, for example, or specificity, sensitivity, inclusivity, exclusivity, all the critical attributes of the assay that we care about, how much volume. It can handle, for example, and then we go to the operational part. And then the operational part is very, very important. We want to make sure that we specify: hey, I need a closed system. In terms, as we discussed before, if we have an NIT-based sterility test, that's a given that one might ask for a closed consumable or a closed workflow or if you want to not necessarily perform your sterility in a grade B in a BC in a grade B. You can specify what you would require for the assay to be able to do. How many handlings? How easy will it be? How how would the sample prep look like not in terms of specifying I want cell lices and then DNA extraction, etc. But I want my operators to do maximum three handlings and only one of them should be metered. Or I don't want any metered. I don't want them to use any pipettes. I want them to just squeeze something in and then the consumable does its job. You know, all these things the ease of use, the throughput The amount of samples per day, all of these things are crucial from an operational perspective. Then we have compliance, data integrity, CFR part 21 CFR Part 11 compliance, computerized system validation, which a lot of people forget as well. then we move to the vendor, And of course, in terms of compliance, compliance to the regulatory like requirements, right? moving to the vendor, we have vendor capabilities. Do is it important for us to for the vendors to have a quality system? Do we need them to be ISO certified? Do they have the manufacturing capabilities that we require? If I want to implement 40 instruments, can they provide me with 40 instruments? what about the consumables? Can they produce consumables? Do they outsource production? can they service? Can they provide technical support? And in case of big companies that have worldwide presence, do we want them to be able to service the whole world and provide support in the whole world? because then a lot of the smaller vendors unfortunately lose traction, right? And from a business perspective, yes, how much is it gonna cost? Do I care how much my consumables will cost? do I care about the overall supply chain? Can it sustain If I implement an alternative or rapid method, and three years later the business continuity of the vendor is gone, what happens, right? So all these things need to be checked and of course return of investment, right? Because overall the company will make an investment, which is normally quite big in terms of FTEs, money to validate and implement to acquire validate and implement so people want to see making their money back. So does it make sense for us to implement? I think these questions are quite crucial to have as specific as possible defined in a URS.
Pierre: So it's good to lose some time at the beginning of the project in order to make that document, in order to save some time later on in the project itself.
Dimitra: I don't think it's lost time because in my experience and this comes from academic incident, but we acquire a very, very big instrument, right? It does what it says, we we know that this is what we want, it comes in and it cannot fit through the door. Because we never put dimensions as a critical part in our URS. So what happens? We need to demolish a wall to make it come in the lab. So ultimately the investment becomes so much bigger that one might have questioned, do we actually need this one? Or can we get an alternative one? So even the most even the things that seem less critical, like how big the instrument, and in some cases, okay, space is a constriction. So some customers Want to have specific dimensions or stackable equipment, but also things like if I have a worldwide company and I have I don't know a location which is up high in the Alps and a location which is here in the Netherlands below sea level, does my ass is my assay robust enough? Can it handle, for example, altitude differences or temperature differences and most of the vendors play around with these attributes but in some cases when we have extreme locations that even might be a decisive factor so i think all these need to be super good defined before we actually try to look around and shop.
Pierre: Okay, that makes sense. I like those small examples that you gave that really make sense because you can visually see them. The system does not fit the door and I've seen that in the past. So it's a story, not only in your lab, but in many other labs. So what you mentioned is that the rapid microbiology method journey never starts with a technology itself. It always starts with a user requirement specification.
Dimitra: Correct.
Pierre: Okay. I've got one more question about URS. Who should be part of building it? Is it the QC, the QA, the IT? Who should be part of it?
Dimitra: Great question. The answer is as many as possible. because in my experience implementing an ARMM is not necessarily a micro department decision. It requires a lot of different stakeholders, like you mentioned, some operations, micro, QA, IT, purchasing. procurement, so all of these regulatory product quality all these different departments need to be part of a URS and I mean the definition the initial definition can be done by one team but then the final URS needs to be approved by as many stakeholders as possible because one of the great things with having such a strong URS is that approved by multiple stakeholders is that they are all aware. Right? So this is a great communication tool. And ideally you want in your team also someone from management that has decision making capabilities.
Pierre: OK, so you talk about primary validation. This is something that is very important as well. What is a primary validation?
Dimitra: Well, the primary validation is a data package, and this data package is supposed to show us that the method that we are looking at scientifically actually does what it's supposed to do. So it's a compilation of all the scientific data that one would require to make sure that the method performs. that is normally provided by most of the vendors nowadays. In some cases, primary validation data can be acquired by a combined effort of the vendor and the customer, but most of the vendors come with at least some data packages. these data packages are created in the absence of a specific matrix, so in the absence of product. although more and more of the vendors do create data with representative matrices or matrix proxies And this data can really really accelerate the validation efforts of the customers because to an extent they can be leveraged to show that the method is really working. And then we as customers in order to validate on our side and implement, we need to show that not only it generically works, but it also works in the presence of our product. And we need to show equivalency, of course.
Pierre: Should we look at it before we buy the instrument or after?
Dimitra: Well that's a good question. I would if given the opportunity, I will always say you need to look at it beforehand. Most of the vendors can supply at least some data beforehand. Maybe not the complete data package, but at least some data in one way or another are shared before purchasing.
Pierre: Okay, because that could be part of the decision. Okay.
Dimitra: Correct. Confidence, right? If I don't believe in the technology that I'm gonna invest few thousands to millions of Euros or dollars, then yeah, it's a bit hard to sell it also. So having these packages really greatly facilitates our efforts to implement these methods.
Pierre: Okay. I was wondering just to, are almost at the end of this podcast. And if you had one or two advice that you could give to people that would like to invest into rapid micro-budget meter, you gave many advice into that podcast. But if there's something like a key takeaway or new type of advice that you would give, it's now.
Dimitra: Yes. I think we kind of touched upon it, but first thing maybe is to indeed build the most concrete and detailed URS you can build prior to start shopping around for a technology. Not even a vendor, a technology. So be as specific to the even tiniest detail that you can add in there. as you said, not A vendor will never fully 100% match your URS. That's the dream scenario. But at least this way you can identify the risks that you see, right? Like what is it's easier to to have to see something if you have an a better overview. then the second one would be to indeed have as many stakeholders as possible. when defining the URS because that makes it makes the communication and alignment much easier. It's also easier for the downstream to actually sell the project to higher management if I'm being honest. Because of course everyone is in favor of innovation and speed and improvements but it needs to make sense also from a business perspective and it's good that we look at it from different points of view. As you said, huge data sets might be great for me because I do not care about storage, but IT will say something different. So all of these things, all of these stakeholders needs need to come together. And the third one is to actually not only look at speed. I think yes a very reduced time to result is absolutely great who doesn't want results faster but the most crucial part is to produce data that are of higher quality and that makes make us make better decisions because ultimately in a pharma setting we influence the safety and the health of a lot of people. So speed is not all the only critical attribute that we should look into these assays. We need to look at compliance, we need to look at reducing errors, reducing false positives and negatives. So all of these play an equally important role in decision making when we choose the correct ARM.
Pierre: That's maybe why we talk more and more about alternative microbiological methods rather than rapid microbiological methods.
Dimitra: Correct, yes.
Pierre: of data integrity and automation,
Dimitra: Correct. And I think from a health and safety perspective we need something that enables us to make better decision, not something that it's just faster. So
Pierre: Right, I like it. So I've got one last question for you. It's an easy one. if you had to go to a congress or conference, which one would you advise to go for somebody that wants to implement in the coming weeks, months, year, a rapid microbiology method? Where can you get the information from?
Dimitra: Yes, that's a good one actually. I really like conferences because it gives you food for thought. so if you want to have great discussions on the scientific principles and see all the latest scientific parts, but also have quite heavy regulatory background, I would say the PDA micro conference. Normally it's in October. in the US. It's a great one. And then we go the next month of November in Pharma Lab where we also met. There, I really like this conference because you have really a lot of vendors being present, exhibiting, you can actually see the equipment, you can have great discussions, and you again have a lot of regulatory presence from USP delegates, EDQM delegates XFDA delegate. So provides a lot of good opportunity for great discussions, seeing actually the technology itself, some diamonds also. also the ECA microbiology conference, which is way more dedicated. normally it's in March. think
Pierre: think everybody has the chance to go out of the lab, I think we need some of them and it's just a matter of picking up the good one.
Dimitra: These three, maybe also the EDQM workshops, that is a great opportunity to see what is coming up with the certification system of the EDQM. I think that is in mid October. so yeah, if you are planning to implement an ARMM, maybe it's worth to be there.
Pierre: Yeah, that's true. And something that I like about the conferences is the chit chat in between the conferences where you can discuss with other microbiologists to better understand what they went through. And when you discuss with other microbiologists, you find out that we all have the same type of problem. And connecting, it's a good way as well to share practice. So I like that very much.
Dimitra: Like it too. I mean it brought us to this point. So
Pierre: Exactly.
Dimitra: yeah, I think it's a very good opportunity to also indeed have these talks and see what other people did in some cases. If you are a bit on the younger side, I mean I gain a lot of insights every time I visit these conferences, so and In some cases, you know there is a social setting, it's easier to ask USP and EDQM people some tougher questions and you might get an answer.
Pierre: No, that's true. And you might end up in a weird podcast as well, going there. I know it's a perfect opportunity. I know it's not easy to get out of the lab because of the workload, because of the costs, and everything else. But that's a really good opportunity as well to save some time later.
Dimitra: Correct.
Pierre: OK, Dimitra, thank you very much for your time and sharing all your experience with Rapid Microbology Method.
Dimitra: Thank you very much Pierre as well. I hope it's useful.
Pierre: At least it was useful for me. I'm sure it will be useful for many other microbiologists. So thank you very much. And maybe in the next one, we can talk about endotoxins because this is something else we can talk about later on. Okay. So thank you very much and see you soon. Bye bye.
Dimitra: Bye.
Dimitra: Hi Pierre. Thank you for having me.
Pierre: I'm very pleased to have you here. We met last year in the Congress called Pharma Lab. And the first thing you told me, I have tried almost all the rapid microbiology method in my lab. So I could not believe it at the beginning. say, okay, you look young and it's possible that you've tried maybe two or three of them, but not all of them. then we discussed. And I find out that it's true. You try almost all of them. So I was amazed by that and say, okay, I need to have her in my podcast. But maybe before, can you introduce yourself, Dimitra?
Dimitra: Yeah, so I'm as you said, I'm Dimitra and I'm an ex-academic microbiologist. I'm specialized I was specialized in environmental microbiology for a very long time. and after my academic career I've turned into an industry enthusiast for the last four to five years. well as in terms of education, I from very early on I was interested in microbiology, so I started already from my bachelor studied studying microbes and their applications, then I finalized my PhD in the group of MyGetten here in the Netherlands studying microbial physiology, but for the great chunk of it I was developing methods to detect microorganisms that harbor specific enzymes that are related to the carbon and nitrogen cycles. Then I was very privileged and I joined the lab of Michi Wagner and Holger Dimes in the University of Vienna where I got to play around with very cool spectroscopic instruments, advanced microscopy, genomics, you name it. They have all the cool toys. And there my focus was again on method development actually, more than physiology. So ultimately towards the end of my academic career I acquired quite big funding. for a project that was aimed to develop this new spectroscopy-based methods to detect previously very, very hard to find and culture microorganisms. So my goal was actually a rapid method, the development of a rapid method for a process that normally would take decades, and my goal was to reduce it into weeks. So that is how I ended up I ended my academic career. And after that I moved back in the Netherlands and I joined Cell Point that later became the Galapagos, currently Lakefront Biotherapeutics. there I worked as a principal scientist within both the analytical development and the QC departments, so I had a double role. And there I was responsible for sometimes even co-developing, but evaluating and validating alternative and rapid micromethods. And transferring them and implementing them to the QC setting. And finally, after Galapagos winded down their cell therapy arm, I recently joined Johnson Johnson Innovative Medicine, where I'm very proud to be part of the QC Microbiology department, where I get to experience a very big organization. global one with commercial products but I also get to still do what I like which is implementing rapid micromethods and evaluating them as part of the global work stream. So
Pierre: Okay, great. That's a long experience. And especially in the last few years, you dedicated it to rapid microbiology method applied to pharmaceutical products. And this is what I would like us to talk about today. Maybe before we talk about how to choose the technology, Can we have a quick overview on the technologies that you use?
Dimitra: Sure. I don't know how short we can make it based on your starting thing, but of course we have different assays within a QC microbiology setting, right? let's take it assay by assay. a part that a lot of people forget that we can implement very, very easily, I would say also an alternative or rapid microbiological method is environmental monitoring. There we have a lot of different vendors providing mostly image-based detection. of microcolonies so we can reduce the time to detection from yeah five days or so depending on the process that we have in place to one or two days depending on the species. There as I said we have a lot of alternatives we have pretty easy to implement systems with great primary validation packages with a caveat that some of these instruments do struggle a bit in the case of molds. So there we need to be a bit more careful. but most of them are able to automate incubation and readout and reduce this insane workload that in cases of sterile manufacturing and annex one compliant plans we have thousands of plates per day, right? So this is a very big burden in a QC lab that can very easily be automated. And yeah, the regulatory acceptance is just there, I think for most of the instruments there's no issue to implement from a health authority perspective. And it gives a quite big return of investment, I think, in terms of FT workload reduction.
Pierre: So it's not only quick, it as well reduces the manipulation and the labor.
Dimitra: Of course, because there is no ideally there should not be a manual readoff or some cases for example when a plate is flagged by the instrument as yeah out of specification then there is a manual readout of the operator but for example in a sterile manufacturing setting we expect one percent of the plates to be out of specification and then we can introduce an operator. important there that prior to implementation, QA needs to play a big role to ensure that when we discuss a 4I principle for GMP generated data, that these are mutually understood what a 4I principle means. So yeah. And we have technologies nowadays that also incorporate maybe a bit more sophisticated technologies that also incorporate AI components in the readoff of the plates.
Pierre: OK, so that was for the environmental testing and the use of technology that can detect micro colonies instead of larger colonies.
Dimitra: Yes. Okay.
Pierre: What about bio-burden or finished product testing?
Dimitra: Yes, so bio burden and sterility testing are sterility testing I think is the holy grail, right? So a lot of people are interested in a very f rapid sterility method, especially in the context of more advanced therapies as we have now CGTs and ATMPs. I would first, before we dive into sterility, maybe touch upon bio burden because There we have a bit less to talk about, maybe. bioburden, of course, can be either growth-based. We have some vendors offering growth-based detection and assays, and we have some vendors offering growth-independent detection. So for the growth-dependent, we have Again, incorporation of optical systems, right? Sophisticated microscopic imaging and sometimes a software component which can be either AI or machine learning, etc. And there we have detection, the same principle as what we have with the compendial method for a plate by a burden assay, but Now we reduce the time to result because we do not wait for the colonies to be as big as to be microscopically detected, but the specific analysis can detect the microcolonies in a sub size, which is normally I would say around 300 microns. most of the applications give such a sensitive detection. We have better distinction s sometimes compared to operators of one versus two colonies because some of the vendors offer time-lapse imaging as well, which helps us identify if that's one big thing that we see on a plate came from one or two or multiple colonies. so yeah, we we have again, of course, automation and digitalization and Yeah, that offers a time to result normally within a day or two compared to a five day normally incubation that we have for the compendial methods.
Pierre: So this one is for the growth base and same as what you discussed earlier, it's micro colony based technologies.
Dimitra: Yes. And analysis, image analysis, right? To when we offer a time lapse to see how the colonies develop. and if we have for example spreading microbes, sometimes we we can better decipher
Pierre: It's one of its...
Dimitra: Exactly. Then we go into the discussion of one or more than two hundred C of Us and we can easily make decisions based on the time lapse then on checking the growth of the colony at the end. And of course that is very important, right? Because it might seem for a rapid that two days is not as rapid to some people. but of course in a continuous production setting that's offers real power in terms of decision making because sometimes we move forward with our production whereas we don't have all the data to actually make informed decisions and that can save for a company a lot of money time etc so this has normally a very good return of investment it is since it's the same media the same thing that we detect, we detect microcolonies, so C CFUs. We have the same value that we detect, so it's a CFU unit. we basically measure growth, so all of these are equal to what the compendial method does. That makes it very easy to validate and implement for the health authorities. Okay. then What else do we have? We have growth-independent methods, and there we have a couple of more established players and some new kids in the block. right. so for bioburden we of course have solid phase cytometry, it's been around for decades. it's a single-cell detection, it detects It's normally combined with the fluorescent labeling and it can normally detect active metabolically active microorganisms. If you first filter your product on a membrane and then you have a laser scanning. and there we as I said we have single cell detection, which means that these technologies might even be suitable for products that are inherently less bioburden heavy so we do not fall necessarily in the sterility realm but we do have low bioburden products. there we have the health authorities a bit more sensitized right towards solid phase cytometry. It's a slightly it's relatively mature technology It's been around for a while. and we have a lot of vendors offering such solutions. smaller, bigger, it depends on the on the platform. there the time to result is actually quite impressive because we can have results up to a few minutes, so an hour. normally for bio burden testing to my experience with this newer technologies you do not necessarily need any incubation so the sensitivity is there to allow detection even without a pre-incubation step
Pierre: So we've seen environmental monitoring, bio-burden, for which both of them we need quantitative method. We need to count the actual number of microorganisms. Correct. And then we have the sterility testing, for which the type of test is slightly different because here we don't need to count, we just need to know if it's positive or if it's
Dimitra: Negative. Correct. It's a yes or no test. So we do not
Pierre: Not that simple, but yes.
Dimitra: I I know, unfortunately. So we move from the quantitative test to the qualitative tests. And indeed a sterility method should be absence presence of any potential microorganism. And this is something that I've seen quite few people struggle with, to be honest, because they struggle with a part of any. microorganism. a lot of vendors to my experience of course they provide a lot of primary validation data and of course they have to limit the data production to a point but they check the compendio panels and they say yeah I hands down I can detect everything. Well that's not true. And there we can discuss maybe a bit further down As to what we as customers should do. Then for sterility, we again have growth-dependent and growth-independent methods, right? And some of the growth-dependent methods are very, very mature technologies, they are widely implemented, especially in the context of USP 1071. and EP2627 which are for short they are they give guidelines for short self-life products that is seeing a very big rise especially because of the CGTs and ATMPs currently being produced. so first of all I think the most mature technology out there the one that is very easily accepted Is respiration based assays. Normally, with these assays, we detect active microorganisms that grow, and these require an incubation. Normally, this is seven days. Most of the vendors have validated those incubations for seven days. Some customers, to my knowledge, have revalidated the assays for Shorter incubation time, so five days, for example. Most of these assays are performed with compendial media, so proprietary media, but they mimic the compendial formulations. and they are based on the CO2 production by microorganisms mostly, which subsequently causes a drop of the pH in the medium. That of course causes that causes a change either of the color or emission of fluorescence from a sensor that is normally located at the bottom of these proprietary media bottles. these are highly automated systems, they continuously measure, so they measure normally every ten minutes, and they provide real time data, so bottles that become positive they are flagged normally immediately, so you do not have to wait until the end of incubation, which means that for heavily contaminated samples or for samples that might contain fast growing microorganisms, the re time to result might be less than the five to seven days.
Pierre: Seven days is the time to negative. Time
Dimitra: Yeah, correct.
Pierre: to positive can be everywhere in between.
Dimitra: Yes. and as I said, these are very mature technologies. they are very widely accepted by the pharmacopeyas. some of them have made them compendial for short self-life products and yeah they are clearly mentioned now in USB 72. other growth dependent methods is A to P biluminescence based methods. Again we have multiple vendors. Some of them are smaller, some of them are bigger methods, and there are a lot of customers including big pharma customers that have validated such methods. They provide the option to digitalize, of course, everything. They require less handling from the operators, and there, what we measure is the production of ATP from microbial cells So we have an incubation period And then lies of the cells and detection of the A to B that was produced.
Pierre: So here it's the end point.
Dimitra: Correct. It's always it's an endpoint assay. And although it's a growth based assay, the detection at the end is not C of U, so it's fluorescence units. so that is something to keep in mind for validation purposes and I think the new PDA Tier thirty three provides really great Guidelines on how to do that. Here, caution is that sometimes we have background, especially for matrices that have been either coming from cell cultures or contain cells, we might have quite a lot of background ATP, which we need to make sure that our assay is able to deplete. And also for cell-based. Matrices, it's normally a a part where we either chemically or mechanically remove the mammalian cells from the matrix. And yeah, every vendor has its own solution. but yeah, the options are there and the customers need to decide whether this is acceptable or not. I will say here that Any method that is chemical, of course, we need to check the vendor data and potentially also generate data ourselves to show that this chemical lysis does not influence the microbial burden because we know that yeah it it can influence the outer membrane of microbial cells as well.
Pierre: So you mean you have chemicals that are supposed to kill the mammalian cell and keep alive the bacteria and we need to make sure that it does not kill the bacteria at the same time as the mammalian cell.
Dimitra: Correct. And so far we've seen that in some cases that is the the case.
Pierre: In your personal experience, you mean?
Dimitra: Yes. Okay. we've seen that it's maybe even genus specific. So if we need to have this data, I would advise that this is checked in a wider panel of microorganisms. If the vendor does not provide this data, because a lot of vendors do provide data like that. Okay. We
Pierre: We're going to talk about primary validation later.
Dimitra: Okay, great. So growth-based method also is microcalorimetry, and there we have the detection of heat that is produced during the metabolism of microorganisms when they grow. And this of course means that all microorganisms produce heat when they grow. So these assays can detect a wide array of microorganisms because their detection is universal, so it's not specific. There we have quite a few pros, and which are of course, as I said, it's a growth-based assay, it requires a certain incubation period. And because it's a growth-based essay, the limiting factor is always our most slow growing microorganism. I think most primary validation data that I've seen, C acnes is the one that always causes a problem. That is true for all grace growth-based methods. I have a point which I forgot actually for this respiration-based assays. because there of course the acnes is our growth limiting factor, but caution needs to be also for fungal species. So not yeast but molds because they inherently produce less CO2. Which although they might grow relatively fast compared to C acnes, they do not accumulate so much CO two, so their detection might also be slow. Okay. Just a side note. So you might see that Exactly but you
Pierre: you might see them but they might not...
Dimitra: still get a negative result from the instrument. And this is very crucial because and that all vendors recommend is that at the end of the respiration based assays you perform also a manual check of the bottles for growth because of the molds. Now back to micro calorimetry the time to result is normally anywhere from 48 to 72 hours and that is because of See Acnes. it there some the vendors normally use compendial media and incubation temperatures, which makes it quite easy, I would say, from a regulatory perspective, to present that to a health authority because it's a growth-based assay, it uses the same media, it is yeah, it's it has an incubation period and the only thing that changes compared to the compendial sterility method is how you detect the the growth not what you detect. So I think this is although the technology itself is not as mature, I think that from a regulatory perspective it's should be easier than some other alternative and rapid microbiological methods to implement. And of course there the automation and digitalization and low workload in the lab is also always part, right?
Pierre: For all those technologies, is it direct inoculation or can it be filtered and then the filtered place into the vials or the bottles?
Dimitra: So for this the respiration based is direct inoculation. For microcalorimetry is also direct inoculation. And what else did we talk about?
Pierre: ATP
Dimitra: ATP yeah, most of the cases it's also direct to my experience.
Pierre: So then the limitation is the volume you can...
Dimitra: Correct. And I know for all of the methods, all of the principles that we discussed, that was a a small issue. I think most of the vendors now switch towards using like higher volumes, but of course from an operational perspective if you are Obliged to test a certain amount of volume. Let's say for ATMPs is 1% of a batch, which normally is few mLs, right? that's yeah, that that might be tricky if you can only inoculate, I don't know, a few micro hundreds of microliters per vial. You don't want to have an operator injecting tiny bottles.
Pierre: 10 or 20 bottles, yes.
Dimitra: Correct. that's increases of course mistakes, operational complexity, etc etcetera. So but I'm happy because most of the vendors now provide also alternatives for higher inocular volumes. So I don't see that as a very big issue. Okay.
Pierre: so we talk about growth-based method. Is there any other growth-based method available?
Dimitra: Not that I've worked with, I'm not aware that we are. Okay. Then we can switch to growth independent methods, right? There as before for bioburden, we have solid phase itometry. there we have filtering in the membrane, etc. etc. as we discussed before. fluorescent staining. And normally with the addition of a small pre-incubation. step then the sensitivity of the method increases to the point that it can be considered it can give an LOD of one C of U so it can be considered a true sterility test.
Pierre: Okay.
Dimitra: Of course again it's there is a pre-incubation part normally because we don't necessarily need so much growth but we do need metabolic activity to incorporate the the the dyes they are activity based dyes that to my experience has a time to result of 48 hours like two days basically so much faster than the compendial 14 days again automated digitalized workflows Very minimal operator handlings. part there, of course, whenever there is a fluorescent labeling, again, background is an issue. and for matrices that contain a high concentration of mammalian cells or other cell debris or any other high concentration of organic components. maybe cell lysis step to remove those parts from the matrix will be needed, but that depends of course on the matrix that we're talking about. Yeah.
Pierre: I've seen some vendors mixing solid-phase cytometry with cross-phase methods. So they start with a cross-phase in standard. So they filter, they add the TSB and the FTM, and then they incubate it for a short period of time, like 48, 72, three, four days. And then they use the solid-phase cytometry at the end for
Dimitra: Detection part.
Pierre: the detection, yes.
Dimitra: Yes. So the thing is there that we do not necessarily need growth though. We need incorporation of the dye into the metabolically active cells, and that can be independent of the doubling of the microorganism where we define as growth. It needs to build biomass because normally These dice target universal microbial enzymes. but so it needs to be live and active, but it doesn't necessarily need to grow in that sense. So yeah, that's that's a benefit, and that's why I would put it more towards the growth independent part.
Pierre: Okay.
Dimitra: yeah. there, of course, we have Spectroscopic methods, one that I have close to my heart because I worked a lot during my academic years, Raman spectroscopy and combined with D2O labeling in some cases, so heavy water or deteriorated water to be more scientifically correct, where we have Microorganisms incubated for a very short period of time of like one or two hours in the presence of heavy water. Then this heavy hydrogen gets incorporated into the membranes, into the lipids of the membranes of the cells. And then normally we have a filtration step where all the microorganisms are captured on a filter. which is scanned with a laser, and from this scanning we get Raman spectra from each cell. So we have single cell analysis, the LOD is very low. The time to result in without the incubation with D2O might be an hour, two hours as long as it takes to scan the surface of the membrane. and if we add the pre incubation step, we're talking about excuse me, couple of hours of time to result. There, the positive part is that Raman spectra can provide a sense of ID of the microorganism. so based on the spectra and how they look on the phenotyping region, we can and while we have let's say a database of reference spectra, we can identify the microorganism for root cause analysis purposes on at least the genus level. on the more negative side, although the technology has been quite widespread in food industry, It's not so widely employed, it's not widely employed in pharma. So yeah, the health authorities are yeah, they know it, but they're not as sensitized towards it. as I said, the positive part is that once we incorporate a labeling incubation. Then we also get information as to live debt, which can influence decision making, of course.
Pierre: Actually, I've never heard about this technology and I've never seen it. No, no, It looks quite impressive on paper or at least discussing about it. But I've never seen it live into a lab, even food labs. But I don't know all of the labs in the world. So this is something I will keep in mind.
Dimitra: I've personally worked a lot of it in development during my academic career, but I also I w I also had projects feasibility and co development projects with such companies and vendors. And yeah it's pretty cool.
Pierre: Yeah, it looks cool. I know that Raman, it's not new technology, but it has been quite difficult to implement it into, I would say, routine labs because it generated a lot of data and then you need to analyze them. And now with the AI, it seems to be much easier to deal with all those data. And I can see that many more spectroscopy Raman technologies are coming up because we know how to deal with the data.
Dimitra: Correct and this is another part which yeah, many people underestimate, but whatever has images and spectra that generates a lot, a lot of data. So you need to have the correct stakeholders when you identify a technology that you want to implement. also from other departments such as IT and infrastructure to actually tell you whether you can do it or not. Because in some cases with smaller companies, yeah, having ten terabytes of data being created over one assay that's not feasible. So yeah, I
Pierre: And now in the routine, what you want is something that's plug and play. So that means you put your sample into the technology and the technology tells you it's positive, negative, or if it's 10, 20, 0, CFU.
Dimitra: Yeah, but this this capabilities these technologies have, right? The automation. automation and digitalization, I think most of the vendors that develop ARMMs they strive for these two things. So it's very rare that you will find something that it's insanely manual or requires a lot of intricate sample prep, to be honest.
Pierre: Is there any other technologies you wanted to talk about or you experimented?
Dimitra: I think NAT based for sterility might be a good one. What do you think? We have, as we said before, QPCRs, digital PCRs, normal RT PCRs, but also very highly automated ultra fast PCR platforms. There of course we detect like specific targets and Most of the assays detect RNA gene, so 16S, 23S, 18S, 28S, and ITS regions. depending on the vendor, we have different levels of automation from fully manual PCRs to like fully completely automated, no precision handlings to just plug and play, as you said before. sterility tests, and there we have high specificity, right? We detect very specific genes. Sensitivity can be influenced by the matrix composition. Complex matrices sometimes give quite a bit of interference with a sample prep mostly, that can also influence the LOD at the end. most of the vendors provide really extensive data sets with inclusivity, exclusivity, amongst all the other critical assay attributes. And yeah, the challenge there is the fact that it can detect it detects DNA. And unfortunately for us, microbial DNA is everywhere. And clean rooms and isolators are not DNA-free environments. So ideally we need to be looking to close systems, systems that cannot be influenced by the surroundings, because yeah, good luck with all the false positives that you can get from a dirty DNA dirty environment. we also need to be cautious about the fact that it can detect live but also dead DNA coming from live and dead cells or damaged cells or extracellular DNA present in the matrices. And these are components that with a more sophisticated workflow can be taken into account and in some cases they are taken into account, so specific treatments. are happening during a sample prep to allow distinction from extracellular or DNA coming from dead cells, pre-existing DNA. Another part is reagents in my experience, because a lot of the enzymes even used in PCRs, because they are recombinant enzymes, they are produced by microbes, they are DNA contaminated. So we need to look into if we want to select something like that, we need to look very close into the workflow, into the consumables, into how it will look in our setting and what the risks are for false positives, and how the vendors deal with their consumables and manufacturing. There, however, I would say that These are ultra rapid methods, so time to detection of one hour or two hours or max three hours. And in the context of specific products such as ATMPs again, yeah, this might be a very big positive. And since we can implement assays on a risk-based approach for these products. Maybe the positives on time to result outweigh the the negative.
Pierre: Okay, so it's just a way to find the balance.
Dimitra: Yeah. And at the end maybe we have some more exotic technologies like dielectrophoresis combined with fluorescence labeling for live and dead staining. Yeah this is I I call it exotic, it's a very cool concept of capturing microbes in a The electrophoretic field. It's a specific capture, which is great especially for matrices that contain cells or high concentration of other debris, for example, because it's a universal microbial frequency capturing. So you can only see your tiny little cells being captured in these electrodes. There we have short pre-incubation for an activity for a live dead labeling. so we have simultaneously capturing and detection via fluorescence of live microbial cells. Yeah, it has a very high sensitivity, and it's it's can also be a very automated workflow. And yeah, the time to result normally will be in terms of hours, so two, three hours, and not more. And
Pierre: It's amazing, I've never heard about this technology as well. I'm learning so much.
Dimitra: Yes, and then there are a lot of smaller companies also, more startup companies, that can provide very nice combination of these technologies. So I would say that the sky's the limit and science is doing a lot and now slowly these These technologies are coming to the market and it's a matter of marrying the correct intended use with the correct technology.
Pierre: Okay, that's a perfect transition of what we are going to talk about now we have all those nice technologies and the challenge to pick up the good one, the one that fits the lab needs. And when we prepared that podcast, you told me that one of the most important things to do is before selecting the technology or just to help you choose the right technology, you need to build a user requirement specification. And this is something that most of the labs are not doing or doing very quickly. And when we discussed, you told me that it's very important to spend time on it. So we are going to spend a little bit of time on this one. And I would like to know for you what is a good user requirement specification.
Dimitra: Yes, a URS is a requirement to have, and that makes it sometimes turns it into just another paperwork we need to fill in many cases, right? It's another tick box to have, but in my opinion, this is not how it should be viewed, and I'm glad with the peer PDA tier thirty-three it also emphasizes that because A URS gives you the fundamental requirements from a scientific, an operational perspective, from a quality perspective, regulatory perspective, even and quite important from a business perspective as to what you exactly need and what is needed from your assay. So, in my opinion, this is a very very good tool to define the question you're asking the problem you're trying to solve am I trying to automate my or am I trying to release my badge faster right so there is a a a good place to define the the question and then try to in as much detail specify what success would look like for you so what exactly would you need it to do? And that is a great then document to start looking into technologies and vendors that can provide the answers to these requirements and not the other way around. Because I see in a lot of cases we first find a vendor and then we build our URS based on the technology, but I think it should be the other way around. so a good URS will have to have of course the Question defined. I need a sterility test for the release of my batch. And then the scientific requirements, the LOD, the LOQ, if we're talking about bioburden, for example, or specificity, sensitivity, inclusivity, exclusivity, all the critical attributes of the assay that we care about, how much volume. It can handle, for example, and then we go to the operational part. And then the operational part is very, very important. We want to make sure that we specify: hey, I need a closed system. In terms, as we discussed before, if we have an NIT-based sterility test, that's a given that one might ask for a closed consumable or a closed workflow or if you want to not necessarily perform your sterility in a grade B in a BC in a grade B. You can specify what you would require for the assay to be able to do. How many handlings? How easy will it be? How how would the sample prep look like not in terms of specifying I want cell lices and then DNA extraction, etc. But I want my operators to do maximum three handlings and only one of them should be metered. Or I don't want any metered. I don't want them to use any pipettes. I want them to just squeeze something in and then the consumable does its job. You know, all these things the ease of use, the throughput The amount of samples per day, all of these things are crucial from an operational perspective. Then we have compliance, data integrity, CFR part 21 CFR Part 11 compliance, computerized system validation, which a lot of people forget as well. then we move to the vendor, And of course, in terms of compliance, compliance to the regulatory like requirements, right? moving to the vendor, we have vendor capabilities. Do is it important for us to for the vendors to have a quality system? Do we need them to be ISO certified? Do they have the manufacturing capabilities that we require? If I want to implement 40 instruments, can they provide me with 40 instruments? what about the consumables? Can they produce consumables? Do they outsource production? can they service? Can they provide technical support? And in case of big companies that have worldwide presence, do we want them to be able to service the whole world and provide support in the whole world? because then a lot of the smaller vendors unfortunately lose traction, right? And from a business perspective, yes, how much is it gonna cost? Do I care how much my consumables will cost? do I care about the overall supply chain? Can it sustain If I implement an alternative or rapid method, and three years later the business continuity of the vendor is gone, what happens, right? So all these things need to be checked and of course return of investment, right? Because overall the company will make an investment, which is normally quite big in terms of FTEs, money to validate and implement to acquire validate and implement so people want to see making their money back. So does it make sense for us to implement? I think these questions are quite crucial to have as specific as possible defined in a URS.
Pierre: So it's good to lose some time at the beginning of the project in order to make that document, in order to save some time later on in the project itself.
Dimitra: I don't think it's lost time because in my experience and this comes from academic incident, but we acquire a very, very big instrument, right? It does what it says, we we know that this is what we want, it comes in and it cannot fit through the door. Because we never put dimensions as a critical part in our URS. So what happens? We need to demolish a wall to make it come in the lab. So ultimately the investment becomes so much bigger that one might have questioned, do we actually need this one? Or can we get an alternative one? So even the most even the things that seem less critical, like how big the instrument, and in some cases, okay, space is a constriction. So some customers Want to have specific dimensions or stackable equipment, but also things like if I have a worldwide company and I have I don't know a location which is up high in the Alps and a location which is here in the Netherlands below sea level, does my ass is my assay robust enough? Can it handle, for example, altitude differences or temperature differences and most of the vendors play around with these attributes but in some cases when we have extreme locations that even might be a decisive factor so i think all these need to be super good defined before we actually try to look around and shop.
Pierre: Okay, that makes sense. I like those small examples that you gave that really make sense because you can visually see them. The system does not fit the door and I've seen that in the past. So it's a story, not only in your lab, but in many other labs. So what you mentioned is that the rapid microbiology method journey never starts with a technology itself. It always starts with a user requirement specification.
Dimitra: Correct.
Pierre: Okay. I've got one more question about URS. Who should be part of building it? Is it the QC, the QA, the IT? Who should be part of it?
Dimitra: Great question. The answer is as many as possible. because in my experience implementing an ARMM is not necessarily a micro department decision. It requires a lot of different stakeholders, like you mentioned, some operations, micro, QA, IT, purchasing. procurement, so all of these regulatory product quality all these different departments need to be part of a URS and I mean the definition the initial definition can be done by one team but then the final URS needs to be approved by as many stakeholders as possible because one of the great things with having such a strong URS is that approved by multiple stakeholders is that they are all aware. Right? So this is a great communication tool. And ideally you want in your team also someone from management that has decision making capabilities.
Pierre: OK, so you talk about primary validation. This is something that is very important as well. What is a primary validation?
Dimitra: Well, the primary validation is a data package, and this data package is supposed to show us that the method that we are looking at scientifically actually does what it's supposed to do. So it's a compilation of all the scientific data that one would require to make sure that the method performs. that is normally provided by most of the vendors nowadays. In some cases, primary validation data can be acquired by a combined effort of the vendor and the customer, but most of the vendors come with at least some data packages. these data packages are created in the absence of a specific matrix, so in the absence of product. although more and more of the vendors do create data with representative matrices or matrix proxies And this data can really really accelerate the validation efforts of the customers because to an extent they can be leveraged to show that the method is really working. And then we as customers in order to validate on our side and implement, we need to show that not only it generically works, but it also works in the presence of our product. And we need to show equivalency, of course.
Pierre: Should we look at it before we buy the instrument or after?
Dimitra: Well that's a good question. I would if given the opportunity, I will always say you need to look at it beforehand. Most of the vendors can supply at least some data beforehand. Maybe not the complete data package, but at least some data in one way or another are shared before purchasing.
Pierre: Okay, because that could be part of the decision. Okay.
Dimitra: Correct. Confidence, right? If I don't believe in the technology that I'm gonna invest few thousands to millions of Euros or dollars, then yeah, it's a bit hard to sell it also. So having these packages really greatly facilitates our efforts to implement these methods.
Pierre: Okay. I was wondering just to, are almost at the end of this podcast. And if you had one or two advice that you could give to people that would like to invest into rapid micro-budget meter, you gave many advice into that podcast. But if there's something like a key takeaway or new type of advice that you would give, it's now.
Dimitra: Yes. I think we kind of touched upon it, but first thing maybe is to indeed build the most concrete and detailed URS you can build prior to start shopping around for a technology. Not even a vendor, a technology. So be as specific to the even tiniest detail that you can add in there. as you said, not A vendor will never fully 100% match your URS. That's the dream scenario. But at least this way you can identify the risks that you see, right? Like what is it's easier to to have to see something if you have an a better overview. then the second one would be to indeed have as many stakeholders as possible. when defining the URS because that makes it makes the communication and alignment much easier. It's also easier for the downstream to actually sell the project to higher management if I'm being honest. Because of course everyone is in favor of innovation and speed and improvements but it needs to make sense also from a business perspective and it's good that we look at it from different points of view. As you said, huge data sets might be great for me because I do not care about storage, but IT will say something different. So all of these things, all of these stakeholders needs need to come together. And the third one is to actually not only look at speed. I think yes a very reduced time to result is absolutely great who doesn't want results faster but the most crucial part is to produce data that are of higher quality and that makes make us make better decisions because ultimately in a pharma setting we influence the safety and the health of a lot of people. So speed is not all the only critical attribute that we should look into these assays. We need to look at compliance, we need to look at reducing errors, reducing false positives and negatives. So all of these play an equally important role in decision making when we choose the correct ARM.
Pierre: That's maybe why we talk more and more about alternative microbiological methods rather than rapid microbiological methods.
Dimitra: Correct, yes.
Pierre: of data integrity and automation,
Dimitra: Correct. And I think from a health and safety perspective we need something that enables us to make better decision, not something that it's just faster. So
Pierre: Right, I like it. So I've got one last question for you. It's an easy one. if you had to go to a congress or conference, which one would you advise to go for somebody that wants to implement in the coming weeks, months, year, a rapid microbiology method? Where can you get the information from?
Dimitra: Yes, that's a good one actually. I really like conferences because it gives you food for thought. so if you want to have great discussions on the scientific principles and see all the latest scientific parts, but also have quite heavy regulatory background, I would say the PDA micro conference. Normally it's in October. in the US. It's a great one. And then we go the next month of November in Pharma Lab where we also met. There, I really like this conference because you have really a lot of vendors being present, exhibiting, you can actually see the equipment, you can have great discussions, and you again have a lot of regulatory presence from USP delegates, EDQM delegates XFDA delegate. So provides a lot of good opportunity for great discussions, seeing actually the technology itself, some diamonds also. also the ECA microbiology conference, which is way more dedicated. normally it's in March. think
Pierre: think everybody has the chance to go out of the lab, I think we need some of them and it's just a matter of picking up the good one.
Dimitra: These three, maybe also the EDQM workshops, that is a great opportunity to see what is coming up with the certification system of the EDQM. I think that is in mid October. so yeah, if you are planning to implement an ARMM, maybe it's worth to be there.
Pierre: Yeah, that's true. And something that I like about the conferences is the chit chat in between the conferences where you can discuss with other microbiologists to better understand what they went through. And when you discuss with other microbiologists, you find out that we all have the same type of problem. And connecting, it's a good way as well to share practice. So I like that very much.
Dimitra: Like it too. I mean it brought us to this point. So
Pierre: Exactly.
Dimitra: yeah, I think it's a very good opportunity to also indeed have these talks and see what other people did in some cases. If you are a bit on the younger side, I mean I gain a lot of insights every time I visit these conferences, so and In some cases, you know there is a social setting, it's easier to ask USP and EDQM people some tougher questions and you might get an answer.
Pierre: No, that's true. And you might end up in a weird podcast as well, going there. I know it's a perfect opportunity. I know it's not easy to get out of the lab because of the workload, because of the costs, and everything else. But that's a really good opportunity as well to save some time later.
Dimitra: Correct.
Pierre: OK, Dimitra, thank you very much for your time and sharing all your experience with Rapid Microbology Method.
Dimitra: Thank you very much Pierre as well. I hope it's useful.
Pierre: At least it was useful for me. I'm sure it will be useful for many other microbiologists. So thank you very much. And maybe in the next one, we can talk about endotoxins because this is something else we can talk about later on. Okay. So thank you very much and see you soon. Bye bye.
Dimitra: Bye.
