Chromatography Forum

As you all know, the cannabis testing industry does not have official methods with well defined QA/QC protocols. Although there are guidelines RE recovery and RPD limits, there are no guidelines as far as passing QC with flags. I am currently involved with a cannabis certified lab in CA and it kills me that in several instances where we have had to re-run batches because our LCS, in just one analyte or two, the recovery is out of the 70-130% recovery limits. And these have been analytes that are ND in the actual samples. In the env field typically you would pass the QC based on the fact that the analyte is not present in the samples and the recoveries were high on the high end and not on the low end. I was not involved in the development of the method and I do not have feedback as to what other analysts/labs are doing in these specific circumstances. I have asked this question to the CA regulatory agency in charge of the labs and I have not been able to get a response other than "read the regulations". Having said this, how have you dealt with situations like this? Have any of you been able to justify out of control recoveries and flag it? Are you all just re-running batches when you could potentially justify an out of control analyte? Have any of you, from OR, WA or CA asked your regulatory agency if you are allowed to pass the batch in instances similar to what I am describing? Please share your experiences or procedures RE QC protocols. Re-running batches is time consuming, costly and the customers ends up waiting longer for their results. Thx in advance for your comments on this topic.

Lack of official methods and protocol was one of the reasons we got out of the Hemp program in our state.

What does the regulation say that you were told to read? Does the regulation list the 70-130% or is that generated from some other means?

What does the regulation say that you were told to read? Does the regulation list the 70-130% or is that generated from some other means?

In CA CCV's, ICV's and LCS's allow have a +/- 30% deviation, 30% RSD for duplicates- this applies to pesticides, potency, metals and residual solvents. I know its pretty wide compared to the environmental field. These are guidelines established by the regulatory agency, CA Bureau of Cannabis Control. IDK what limits Oregon or Colorado have but like I stated in the beginning of the post, due to the nature of matrices many times we encounter ion suppression and we have to scratch the entire batch just because 1 analyte or 2 may be at 68% or 67% recovery and this invalidates the entire batch. I am trying to find out if other analysts in the cannabis field apply flags when QC is minimally/slightly out of control.

Although the Bureau has established the QC limits mentioned above, method development should follow guidelines from any of the following agencies:


(b) To the extent practicable, the laboratory test methods shall comport with the following guidelines:

(1) US Food and Drug Administration’s Bacterial Analytical Manual, 2016;

(2) AOAC International’s Official Methods of Analysis for Contaminant Testing of AOAC International, 20th Edition, 2016; and

(3) United States Pharmacopeia and the National Formulary’s Methods of Analysis for Contaminant Testing, 2016.

+/- 30% is used a lot in the Drinking Water manual and methods from EPA, which is strange since those are normally clean while waste waters are at +/- 20%.

The drinking water limits are wider but normally they don't allow reporting of anything outside those limits, so maybe the CA group is looking at drinking water rules on this one.

Do you run your standards in a blank matrix taken from a known clean sample to calibrate out the matrix enhancement/suppression?

+/- 30% is used a lot in the Drinking Water manual and methods from EPA, which is strange since those are normally clean while waste waters are at +/- 20%.

The drinking water limits are wider but normally they don't allow reporting of anything outside those limits, so maybe the CA group is looking at drinking water rules on this one.

Do you run your standards in a blank matrix taken from a known clean sample to calibrate out the matrix enhancement/suppression?

James_Ball,

Currently we do not do matrix matched calibrations. I have thought about doing MM calibrations but where do I get a clean source? I have thought about selecting the most clean extracts that we process here at the lab and do it that way. I don't think I will find a perfectly clean sample but at least select the best I can find. The alternative would be buying CRM hemp flower and perform an extraction on it. But if I am not mistaken the cost is like $450/gram. But definitely will have to do the matrix matched calibration, need to find out for myself how well it would perform and if it would make a difference.

Possibly have a client grow a few plants where no pesticides are used and let you use those as a source for clean matrix?

The CRMs usually have a known amount of targets in them so that probably wouldn't work.

If you can find a sample that has the lowest amount, maybe if you dilute that extract by 2x or 5x, you would have something close to a matrix. It would have some of the things that cause the enhancement or suppression of the signal and help compensate for that.