Application of Biofluorescent Particle Counters for Real-Time Bioburden Control in Aseptic Cleanroom Manufacturing
Round 1
Reviewer 1 Report
The publication is interesting. The basic problem of the reader here is the lack of a diagram of the measuring system and an up-to-date diagram of the conducted research because the mentioned issue refers the reader to "paid publications". In the description of Figure 2 there is a list of measuring instruments, using abbreviations -- this is very well readable for the authors of the publication, for readers probably not.
I agree with authors that: “A direct comparison of BFPC-measured values with CFUs counted on agar plates or stripes is hardly possible, as described by several scientists (19, 24–26). The reason for this is the dependency on possible growth on agar plates with CFU methods.” -- without specifying the exact parameters of the BFPC system, in the area of particle detection methods (florescence excitation: wavelength and power of the source, detector sensitivity and a set of test molecules) such a comparison does not make any sense. Therefore, please provide the indicated parameters, because according to me, the basic molecules well detected by fluorescent systems are polycyclic (3-5 rings) aromatic hydrocarbons present in the air mainly as a result of using wheeled transport (tires, exhaust gases, etc.) Of course, some "fungi" can be fluorescent, but bio-fluorescence is not excited "universally". Thus, having received the indicated data, I will be able to correctly assess the quality of the results presented in the work. Besides, I don't fully understand the contents of Table 3. "Particles >1.0 μm/m³" Particles according to me have characteristic dimensions and concentrations, free path length, chemical activity -- so please explain the idea clearly. I know you can correlate everything with everything, but is that what the reader means?
Author Response
Author's Reply to Reviewer 1
We highly appreciate the reviewer’s feedback and comments. We revised our manuscript according to your recommendations.
With respect to your comment that “The basic problem of the reader here is the lack of a diagram of the measuring system”, we added a more detailed description of the system in section 2 and referred to an illustration given in a brochure of the BioTrak® manufacturer.
For the conducted research we added a table with the test sequence and setup to better illustrate the complete experimental study. We assume that the “paid publictions” you refer to, are the ISPE guidelines in the 2nd and 3rd edition, which are not available in an open access version. Though these guidelines provide valuable recommendations for pharmaceutical companies and are frequently bought and used by manufacturers, they are still only non-regulatory guidelines. The valid authority requirements are given in Annex 1 (EU) and the aseptic guideline (FDA). Hence, we just highlighted the fact that even this widely used guideline withdraw the recommendation of ACR 20 in ISO 8 requirements in their latest edition, leaving the FDA aseptic guideline as the only guiding document still including a recommended ACR value for ISO 8. We would not be surprised if the FDA dropped this guiding value in a future revision of the aseptic guideline.
Re-reading our description of figure 2, we agree with the reviewers comment that readers might have troubles with following the abbreviations. Consequently, we revised that description.
We furthermore appreciate the reviewers comment to “provide the indicated parameters, because according to me, the basic molecules well detected by fluorescent systems are polycyclic (3-5 rings) aromatic hydrocarbons present in the air mainly as a result of using wheeled transport (tires, exhaust gases, etc .) Of course, some "fungi" can be fluorescent, but bio-fluorescence is not excited "universally".”
On this, we added a more descriptive paragraph in section 2 with specific data on the measuring system and its capability to distinguish between viable and non-viable particles. As in cleanroom operations the incoming air is HEPA-filtered and MCPs are only measured larger than 1.0 µm, it can be assumed that all detected viables in a cleanroom have biological sources.
Our correlations list for particles >1.0 µm is meant to show the validity of our measurements. As microbes typically connect to particles, a correlation can be expected. All three possible comparisons show a good correlation. If they did not, the experimental design and results would have been questionable.
Thank you again for your valuable input which surely improves the readability and preciseness of our work.
Best regards,
Detlef Behrens
Reviewer 2 Report
The study demonstrates in an effective and transparent way under what conditions a cleanroom can be operated and how the air quality can be monitored real time in a reliable way by a biofluorescent particle counter.
Author Response
Authors Reply to Reviewer 2
We are very grateful for the reviewers comments and are very happy that all rating questions have been answered with “yes”. In a revised version we have included improvement suggestions by other reviewers. This should help readers, who are not familiar with pharmaceutical cleanroom operations, to better understand some terms and statements.
Best regards,
Detlef Behrens
Reviewer 3 Report
Dear authors,
Congratulations for your work on this important topic. I find your work on the critical issue of cleanroom HVAC for pharmaceutical and other uses to be reasonably well-presented, and this research most certainly should be published.
However, I strongly feel that drawing a conclusion that cleanroom air change rates could be very greatly lowered industry-wide, solely on the basis of experiments performed in a single training chamber, is definitely an overreach. Usage of air change rate as the sole cleanroom HVAC parameter under consideration makes the assumption -- which is unfortunately entirely unfounded in many real-world industry examples -- that there is perfect, or nearly-perfect, air mixing within a cleanroom itself. In many real-world cleanroom examples, cabinets, desks, benches, equipment, etc, can (and are, in many real-word cases) unfortunately be placed so that air pockets can (and do) get partially trapped behind them -- so that the airflow rate of the cleanroom at large is vastly greater than the airflow rate within such air regions. Specifying ACRs that are 20 h-1 (or even above), while being perhaps approximately a factor of 2 excessive for a completely perfectly-air-mixed cleanroom, is ultimately, to some extent, partially to accommodate the fact that in many real-world cases, with real equipment placed where it sometimes needs to be in some real-world cleanrooms, the assumption of completely perfect air mixing throughout the cleanroom would be, and is, a very incorrect one.
Thus, the logical next step would be to do a study of usage of BFPCs, when in parallel with presently-installed systems, in many different real-world industry cleanrooms. Then, after such a future much larger study, conclusions can (and should) be drawn about suggested industry cleanroom ACRs, assuming of course that such a larger study is done well.
This present work provides very good, strong motivation for such a larger study. However, what the present work most definitely does not do -- contrary to a number of statements within the paper -- is to provide sufficient motivation for changing ACRs industry-wide without the need for such a follow-on larger, real-world, study.
I posit that the above important caveat should be carefully stated within a revised draft (and the statements within the paper that indicate that one could just change ACR recommendations industry-wide, without having a larger follow-on study similar to that described above, should be carefully modified).
Then, following such small but important modifications, the paper most certainly should be published.
Author Response
Author's Reply to Reviewer 3
We first thank the reviewer for the congratulations. Furthermore, we express our thanks for the comments and fully agree with the suggestions for improvement. We revised our manuscript accordingly and removed or changed statements which were obviously ‘over-enthusiastic’. The changes have been made in the abstract as well as in the discussion and conclusion section.
With respect to the mentioned air pockets due to non-perfect air mixing, we share this observation. It must be ensured that all areas in a cleanroom will not exceed regulatory limits, regardless of the actual air change rate at this point. Still, we are confident that even these areas would meet regulatory requirements with lower ACRs, as particles and MCPs would simply sediment and become uncritical for ISO 8 manufacturing if not kept airborne by turbulence.
As our research continues, we recently conducted a further study in a real-world production room which will be published in September. In this study, similar results have been found.
Moreover, we agree with the reviewer that more data from a long-term study in real-world production environment is required before a strong, scientifically based recommendation to reduce ACRs can be made. On this comment, we are happy to announce that we are currently planning such a study in co-operation with a big bio-pharmaceutical company. This study shall commence in October and last for at least six months with parallel measurement of MCP, particles, and CFU.
Thank you again for your valueable feedback and input.
Best regards,
Detlef Behrens
Reviewer 4 Report
General comment:
The paper “Application of biofluorescent particle counters for real-time bioburden control in aseptic cleanroom manufacturing” by Behrens et al. presents the experimental study of existing conditions in a pharmaceutical cleanroom under different operators’ garments and different air change rates, comparing the measurements of biofluorescent particle counter, microbe carrying particles (MCPs) and Colony Forming Units. The paper is well structured, the methodology is clearly explained and the results are interesting and well commended. A minor revision is recommended. Below specific comments useful for improving the manuscript are listed.
Specific comments:
Lines 37-38: specify the limit values ​​indicated by the regulations for CFU. Since they are mentioned here for the first time, it would be appropriate to move the Line 315 quote here.
Line 221: the authors should specify the data aggregation period. From the graphs, I deduce that the average values ​​and statistics are calculated over time intervals of 1 minute, but how often are the raw data acquired?
Line 224: why does the number of particles decrease in the start phase? Authors should add a comment about it.
Line 222: it would be interesting to see in Figure 4 also the results of the 19G case to understand how the different garments can affect the particles in the various phases of the experiment
Figure 4b: why in the NCG case during the start phase the number of particles does not decrease but remains constant?
Line 288: the text states that spray disinfectant is used up to test #7 but from table 2 it appears that only tests #4-#7 are affected by this problem. What about tests #1-#3?
Line 392: the authors should add a comment about the influence of the initial conditions occurring in a real cleaning room: what is the expected initial particle concentration? And how can it affect the levels reached in the operation phases?
Author Response
Author's Reply to Reviewer 4
We thank the reviewer for the feedback and recommendations for improving our manuscript. We revised the paper accordingly.
With respect to the specific comment: “Lines 37-38: specify the limit values indicated by the regulations for CFU. Since they are mentioned here for the first time, it would be appropriate to move the Line 315 quote here.”, we added the CFU limit value at this point.
Thank you for the recommendation to add the aggregation period of the system. To describe the BioTrak® measurement system in more detail we added a paragraph below figure 2 (line 180 in the revised version). Here we also included the sampling volume and the aggregation period.
To your question “Line 224: why does the number of particles decrease in the start phase? Authors should add a comment about it.”: We added a comment in the respective paragraph that this is caused by sedimentation as no HVAC turbulence would keep particles airbourne.
Regarding your comment “Line 222: it would be interesting to see in Figure 4 also the results of the 19G case to understand how the different garments can affect the particles in the various phases of the experiment”: We appreciate this comment and agree that the comparison of I8G and I9G is interesting. However, to not overload Figure 4 with more information we only added a reference to the supplementary material and added an additional figure there which compares I8G/I9G particles.
To your question: “Figure 4b: why in the NCG case during the start phase the number of particles does not decrease but remains constant?” we revised the paragraph to better describe the activities during the starting phase. The operators had to switch on the measuring systems first before leaving the room. This activity, in some experiments, was tracked already by the BioTrak as an increase in particles/viables. In the next minutes the values decreased due to sedimentation. At the end of the start phase, with opening of the airlock door, particles and viables increased again as expected.
To answer your question “Line 288: the text states that spray disinfectant is used up to test #7 but from table 2 it appears that only tests #4-#7 are affected by this problem. What about tests #1-#3?”: You are correct: Only tests 4 – 7 are actually affected. As two different disinfectant bottles were available (liquid and spray), the operators used any of these up to test 7. For the first 3 tests, this was only the liquid bottle. Only after we observed the effect of the spray bottle use on particle measurement, we removed this bottle to eliminate this unwanted effect.
Thank you for your recommendation for Line 392: “the authors should add a comment about the influence of the initial conditions occurring in a real cleaning room: what is the expected initial particle concentration? And how can it affect the levels reached in the operation phases?” We added a paragraph to explain the study limitations at this point. As our research continues, we recently conducted a further study in a real-world production room which will be published in September. In this study, similar results have been found.
We are very grateful for your comments and recommendations. We believe that these helped considerably to improve the quality of our manuscript.
Round 2
Reviewer 1 Report
I like corrected version.
