Kochman

It is still prudent to run some kind of control when a donor/patient types as AB pos. While it it is true that the new monoclonals are less likely to react non-specifically due to things like polyagglutination, there are occassional reports of false positives due to a component of the reagent diluent. The ideal control would be the reagent diluent itself but I'm not certain it is always available. As far as following the manufacturers' package insert, don't assume that if the manufcturer does not describe such a control that it is not necessary. Remember, it might be required by some local/state/federal agency even if it's not in the PI.

The old blood bank inspection check list used to allow something similar to CAP's allowance to use patient results to show appropriate reactivity, i.e., it allowed comparing the forward and reverse groups and showing they gave the same interpretation as evidence that the Anti-A, Anti-B, and reverse grouping cells were working appropriately. That wasn't allowed for Anti-D and the others However, CLIA places more empahsis on ensuring that QC testing ensures that reagents are working at the limits of detection and thus capturing those occassional weaker types. I understand your confusion related to apparent differences in criteria and hope we (FDA) can address this in the near future.

Actually, you should use diluted monoclonal antisera only on those QC cells that are positive for the desired antigen. Negative QC cells should be tested with undiluted reagents. Patient samples should not be diluted. Sorry I wasn't clearer. Sheryl Kochman

The "obscure FDA document" referred to above is The Guide to Inspection of Blood Banks and is one that all blood collection facilities should have available to them. It is available at: http://www.fda.gov/ora/inspect_ref/igs/blood.html Another useful document that applies to Licensed and Unlicensed Blood Banks, Brokers, Reference Laboratories, and Contractors is the Compliance Program for Blood and Blood Products. It, and other documents of potential interest, are available at: http://www.fda.gov/cber/cpg/cpg.htm Sheryl Kochman

Because monoclonal antisera are usually very strong, it is NOT appropriate to use undiluted monoclonal antisera for QCing cells of any type. Using undiluted monoclonal antisera can give you a false sense of security that the cell possesses the antigen desired at levels that will react with patient serum/plasma. If you don't want to use stored plasma containing antibodies, one suggestion is to dilute monoclonal antisera to achieve a 2+ or 3+ reaction with in-date panel cells. If the expired panel cells have deteriorated, you should be able to tell that based on a reduction in the reaction grade. Yes this method has to be validated but the time spent up front could be well worth it if you prevent missing a clinically significant antibody because the cells were not strong enough to react with the patient’s sample. It is also important to run a NEGATIVE control with expired panel cells to ensure that they have not become contaminated, particularly with microbes. Sheryl Kochman Chief, Devices Review Branch FDA/CBER/OBRR/DBA

I think you'll find that reagent manufacturers recommend the use of a positive and a negative control each day of use. Note also that FDA reviews these recommendations.

Those devices that have already been cleared are posted at the CBER website: http://www.fda.gov/cber/products/510ksoft.htm FDA cannot disclose the existence of or discuss pending submissions as this information is considered confidential.

Manufacturers and users need to be aware that systems like those described by Richard Kriozere are medical devices because they are intended to prevent disease (transfusion reaction) by preventing incompatible transfusions. They also meet the definiton of Blood Establishment Computer Software. Therefore, any of these systems in interstate commerce or where there is interstate transfer of or access to the data requires a 510(k).

I still maintain (very strongly) that some sort of comparison study, whether you call it validation or not, is neccessary in this situation, particularly if you are a transfusion service. Yes, the reagents are approved by FDA (in my Branch) but FDA requirements are minimum requirements. Some manufacturers simply meet those minimums, others exceed them. If you are used to a certain performance level, you need to know if you can expect that same performance level with the new reagent. For example, if treatment decisions are based on the strength of an antibody titrated and tested by the AHG method, you need to know if the new reagent will give the same endpoints as the one you have been using. If what you were using exceeds FDA's minimum standards, you might see a higher titration endpoint with your old reagent than you would with a new reagent that simply meets minimum requirements. So, if you intervene at a particular titration endpoint, this endpoint might have to be revised to be appropriate for the new reagent. I'm happy to talk or correspond with anyone who wants additional rationale on this topic in the interest of improving their operations. Sheryl Kochman Chief, Devices Review Branch OBRR/CBER/FDA 301-827-6123 kochman@cber.fda.gov

Terry You state "We do not track donors, perform electronic crossmatches or anything that needs to be validated according to NY-DOH." This causes me concern because there are FDA requirements for use of software in the blood bank, whether you are only registered or are licensed. Please see 21 CFR 211.68 and, more importantly, Compliance Program, Chapter 42 - Blood and Blood Products, Inspection of Licensed and Unlicensed Blood Banks, Brokers, Reference Laboratories, and Contractors- 7342.001 (Issued 7/18/2003), Attachment H at http://www.fda.gov/cber/cpg/7342001bld.htm#attH. This is the guidance that invesitagtors use when conducting inspections of blood establishments. They will be looking to "Determine which operations are computerized and how the user validated the computer system on-site to demonstrate that it performs the intended functions accurately and reliably." Others who have used Off-the-Shelf Software (OTSS) (as software such as Microsoft Access is called) have found that it does not always perform as they expected it would. Elizabeth, The following quotes, also from the Compliance Program referenced above might help you convince your pathologist that his idea is a bad one. "All software, including software developed in-house, used to manufacture blood and blood components, or to maintain data for making decisions about donor (suitability) eligibility, release products for transfusion or further manufacture are devices under Section 201(h) of the FD&C Act. The device provisions such as: registration as a device manufacturer, product listing, medical device reporting and compliance with the quality system regulation and pre-market notification 510(k) or application apply to the device software manufacturer." "FDA has previously advised blood banks to transition to either a cleared software product or to one for which a manufacturer was actively pursuing clearance or to submit a plan to convert its in-use computer software." "Blood establishments that developed software for their own use and that did not ship it interstate are subject to the Quality System Regulation requirements (21 CFR 820), the CGMPs for Blood and Blood Components (21 CFR 606), and the CGMP’s for Finished Pharmaceuticals (21 CFR 211)." "FDA deems a blood establishment that develops software for its own use as a medical device manufacturer and; therefore, subject to the Quality System Regulation." I hope this helps! Sheryl Kochman Chief, Devices Review Branch DBA/OBRR/CBER/FDA

Ortho AutoVue is not available in the US nor have the claims made for it been reviewed by FDA. Furthermore, specific performance would depend on the source of the Anti-D in the gel card.

Some food for thought: The manufacturers' package insert should tell you what kind of control to run. It is important to note that Anti-D reagents have additives that help the cells resuspend and give cleaner reactions. If you make your own control, it won't have these additives whereas the manufacturers' reagent-matched control should. Also, using the manufacturers' reagent-matched control can save you some headaches, like how to QC it, how long you can use it, etc. This will all be in the labeling.

Back in February, I posted this: This response was based on information provided to FDA by the manufacturer (Gamma) at the time of licensure (1992). Specifically, the manufacturer included the following limitation in it's package insert: "Examples of pure IgG4 subclass antibodies may not be detected by this reagent. Note, however, that pure IgG4 antibodies are very uncommon." For the most up-to-date information on the specific performance characteristics of any reagent, consult the current package insert or contact the manufacturers' technical support group.

There is an on-going discussion on the CBBS web-site too. See: http://www.cbbsweb.org/enf/rh_testing_dvi.html

I'm surprised to hear people indicating that their reagent manufacturers' package inserts do not include directions to use both a positive and a negative control, unless these are for reagents that are not U.S. licensed. During the review and aproval of U.S. licensed reagents, the QC directions are reviewed to ensure that they require use of appropriate positive and negative controls. If anyone knows of a an example where this is not the case, we sure would like to hear about it. kochman@cber.fda.gov

First, keep in mind that the concept of validation is based on the fact that no two testing facilities are exactly alike. Validation is supposed to show that you can make the reagent work properly (meet performance specifications/expectations) in your own facility, with your own staff, your own instruments, your own adjunct reagents (for example - saline), following your SOPs, etc. It is difficult to give detailed guidance on validation when someone asks since, as stated above, every situation is different. In general, you should think about your extremes and test at each extreme. For example, if you know your "room temperature" varies between 18 and 25 C, you should test at 18 and at 25. If testing works at both extremes, it is presumed it will work in between. Likewise you should consider things like run/batch sizes (smallest and largest), incubator temperature variability/stability, size/volume and number of sample drops, sample age (freshest and oldest), etc. The list is too long to cover here and not all variables are applicable to each validation situation. That's why you should sit down and develop a validation plan first and then go about your testing. In this specific case (validation of change to monoclonal anti-human globulin), consider doing parallel antibody screening and identification using the range of antibodies you normally see or "worst case" antibodies like anti-Fyb, anti-Jka and anti-Jkb. Also consider parallel DATs and weak D tests. If you do both hand washing and machine washing, test using both methods. And use samples that are the oldest your SOP allows.

Sorry, but checking with a researcher at a reagent manufacturing facility will not always give you the best regulatory answer. The concept of validation is based on the fact that no two testing facilities are exactly alike. Validation is supposed to show that you can make the reagent work properly (meet performance specifications/expectations) in your own facility, with your own staff, your own instruments, your own adjunct reagents (for example - saline), following your SOPs, etc. I know that CAP expects you to validate and I assume JAHCO expects it, too. It is difficult to give detailed guidance on validation when someone asks since, as stated above, every situation is different. In general, you should think about your extremes and test at each extreme. For example, if you know your "room temperature" varies between 18 and 25 C, you should test at 18 and at 25. If testing works at both extremes, it is presumed it will work in between. Likewise you should consider things like run/batch sizes (smallest and largest), incubator temperature variability/stability, size/volume and number of sample drops, etc. The list is too long to cover here and not all variables are applicable to each validation situation. That's why you should sit down and develop a validation plan first and then go about your testing. In this specific case (validation of change to monoclonal anti-human globulin), consider doing parallel antibody screening and identification using the range of antibodies you normally see or "worst case" antibodies like anti-Fyb, anti-Jka and anti-Jkb. Also consider parallel DATs and weak D tests. If you do both hand washing and machine washing, test using both methods. Hope this helps instead of scaring you.

Sorry to respond to this so late but I'm new to the site and think this is a very important topic. First, why should QC give 1-3+ and not 4+? Because when you get a 4+ reaction, all you know is that you will detect strong reactions. You don't know if your system (reagents + method + equipment + staff) will detect weak reactions as are often seen in patient testing. Second, regarding dilution of reagents for QC - be careful! Reagent manufacturers add things to their reagents to make them avid (give strong results) and stable. If you dilute them too much, you can make them very unstable. I always recommend that if you are going to dilute reagents you should use saline with some amount of bovine albumin (3-6%) in it. Lastly, if you can't make QC work, it might not be a problem with the reagents - it might be a problem with your system and the QC is telling you EXACTLY what you need to know, i.e., maybe you're doing something wrong. This might make me wildly unpopular, but please at least think about it.

It is critically important to read the ENTIRE package insert for reagents and not just the "directions for use part" when making decisions about switching reagents and validation. The brand of anti-human globulin has not been mentioned but, did you know that Gamma's monoclonal Anti-IgG will not detect antibodies that are subclass 4 whereas rabbit Anti-IgG does? Probably not a big deal and maybe even desirable but you should know these kinds of things BEFORE you put a reagent into use so you don't have to figure them out during testing. A thought - "doing nothing" (for any issue) is usually not acceptable.

Depending on the monoclonal reagent you are using, it could be different than the rabbit material you are used to. I would recommend doing some level of validation.

If you haven't done so already, you should first check the instructions provided by the manufacturer with the reagents and with the instruments. If these are not clear, you should consult the manufacturer.