Tag Archives: analyses

By Dr. Edward F. Askew

Editor’s Note: The views expressed in this article are the author’s opinions based on his experience working in the laboratory industry. This is an opinion piece in a series of articles designed to highlight the potential problems that clients may run into with labs.

In the last two articles, I discussed the laboratory’s first line of defense (e.g. certification or accreditation) paperwork wall used if a grower, processor or dispensary (user/client) questioned a laboratory result and the conflicts of interest that exist in laboratory culture. Now I will discuss the second line of defense that a laboratory will present to the user in the paperwork wall: Quality Control (QC) results.

Do not be discouraged by the analytical jargon of the next few articles. I suggest that you go immediately to the conclusions to get the meat of this article and then read the rest of it to set you on the path to see the forest for the trees.

QC in a laboratory consists of a series of samples run by the laboratory to determine the accuracy and precision of a specific batch of samples. So, to start off, let’s look at the definitions of accuracy and precision.QC Charts can provide a detailed overview of laboratory performance in a well-run laboratory.

According to the Standard Methods for the Examination of Water and Wastewater:

Accuracy: estimate of how close a measured value is to the true value; includes expressions for bias and precision.

Precision: a measure of the degree of agreement among replicate analyses of a sample.

A reputable laboratory will measure the Accuracy and Precision of QC samples in a batch of user samples and record these values in both the analytical test report issued to the user and in control charts kept by the laboratory. These control charts can be reviewed by the user if they are requested by the user. These control charts record:

Accuracy (means) chart: The accuracy chart for QC samples (e.g., LRB, CCV, LFBs, LFMs, and surrogates) is constructed from the average and standard deviation of a specified number of measurements of the analyte of interest.

Precision (range) chart: The precision chart also is constructed from the average and standard deviation of a specified number of measurements (e.g., %RSD or RPD) for replicate of duplicate analyses of the analyte of interest.

Now, let’s look at what should be run in a sample batch for cannabis analyses. The typical cannabis sample would have analyses for cannabinoids, terpenes, microbiological, organic compounds, pesticides and heavy metals.

Each compound listed above would require a specific validated analytical method for the type of matrix being analyzed. Examples of specific matrixes are:

Cannabis buds, leaves, oil
Edibles, such as Chocolates, Baked Goods, Gummies, Candies and Lozenges, etc.
Vaping liquids
Tinctures
Topicals, such as lotions, creams, etc.

Running QC analyses does not guarantee that the user’s specific sample in the batch was analyzed correctly.

Also, both ISO 17025-2005 and ISO 17025-2017 require the use of a validated method.

ISO 17025-2005: When it is necessary to use methods not covered by standard methods, these shall be subject to agreement with the customer and shall include a clear specification of the customer’s requirements and the purpose of the test and/or calibration. The method developed shall have been validated appropriately before use.

ISO 17025-2017: The laboratory shall validate non-standard methods, laboratory-developed methods and standard methods used outside their intended scope or otherwise modified. The validation shall be as extensive as is necessary to meet the needs of the given application or field of application.

Validation procedures can be found in a diverse number of analytical chemistry associations (such as AOACand ASTM) but the State of California has directed users and laboratories to the FDA manual “Guidelines for the Validation of Chemical Methods for the FDA FVM Program, 2nd Edition, 2015”

The laboratory must have on file for user review the following minimum results in an analytical statistical report validating their method:

accuracy,
limit of quantitation,
ruggedness,
precision,The user must look beyond the QC data provided in their analytical report or laboratory control charts.
linearity (or other calibration model),
confirmation of identity
selectivity,
range,
spike recovery.
limit of detection,
measurement uncertainty,

The interpretation of an analytical statistical report will be discussed in detail in the next article. Once the validated method has been selected for the specific matrix, then a sample batch is prepared for analysis.

Sample Batch: A sample batch is defined as a minimum of one (1) to a maximum of twenty (20) analytical samples run during a normal analyst’s daily shift. A LRB, LFB, LFM, LFMD, and CCV will be run with each sample batch. Failure of any QC sample in sample batch will require a corrective action and may require the sample batch to be reanalyzed. The definitions of the specific QC samples are described later.

The typical sample batch would be set as:

Instrument Start Up
Calibration zero
Calibration Standards, Quadratic
LRB
LFB
Sample used for LFM/LFMD
LFM
LFMD
Samples (First half of batch)
CCV
Samples (Second half of batch)
CCV

The QC samples are defined as:

Calibration Blank: A volume of reagent water acidified with the same acid matrix as in the calibration standards. The calibration blank is a zero standard and is used to calibrate the ammonia analyzer

Continuing Calibration Verification (CCV): A calibration standard, which is analyzed periodically to verify the accuracy of the existing calibration for those analytes.

Calibration Standard: A solution prepared from the dilution of stock standard solutions. These solutions are used to calibrate the instrument response with respect to analyte concentration

Laboratory Fortified Blank (LFB): An aliquot of reagent water or other blank matrix to which known quantities of the method analytes and all the preservation compounds are added. The LFB is processed and analyzed exactly like a sample, and its purpose is to determine whether the methodology is in control, and whether the laboratory is capable of making accurate and precise measurements.

Laboratory Fortified Sample Matrix/Duplicate (LFM/LFMD) also called Matrix Spike/Matrix Spike Duplicate (MS/MSD): An aliquot of an environmental sample to which known quantities of ammonia is added in the laboratory. The LFM is analyzed exactly like a sample, and its purpose is to determine whether the sample matrix contributes bias to the analytical results. The background concentrations of the analytes in the sample matrix must be determined in a separate aliquot and the measured values in the LFM corrected for background concentrations (Section 9.1.3).Laboratories must validate their methods.

Laboratory Reagent Blank (LRB): A volume of reagent water or other blank matrix that is processed exactly as a sample including exposure to all glassware, equipment, solvents and reagents, sample preservatives, surrogates and internal standards that are used in the extraction and analysis batches. The LRB is used to determine if the method analytes or other interferences are present in the laboratory environment, the reagents, or the apparatus.

Once a sample batch is completed, then some of the QC results are provided in the user’s analytical report and all of the QC results should be recorded in the control charts identified in the accuracy and precision section above.

But having created a batch and performing QC sample analyses, the validity of the user’s analytical results is still not guaranteed. Key conclusion points to consider are:

Laboratories must validate their methods.
Running QC analyses does not guarantee that the user’s specific sample in the batch was analyzed correctly.
QC Charts can provide a detailed overview of laboratory performance in a well-run laboratory.

The user must look beyond the QC data provided in their analytical report or laboratory control charts. Areas to look at will be covered in the next few articles in this series.

June 19, 2018

From The Lab

Quality Plans for Lab Services: Managing Risks as a Grower, Processor or Dispensary, Part 2

By Dr. Edward F. Askew

No Comments

In the previous article, I discussed the laboratory’s first line of defense (e.g. certification or accreditation) when a grower, processor or dispensary (user) questions a laboratory result. Now let us look behind this paperwork wall to the laboratory culture the user will encounter once their complaint is filtered past the first line of defense.

It is up to the client (processor, grower or dispensary) to determine the quality of the lab they use.In an ISO 17025 (2005 or 2017) and TNI accreditation, the laboratory must be organized into management, quality and technical areas. Each area can overlap as in the ISO 17025-2017 standard or be required to remain as separate sections in the laboratory as in the ISO 17025-2005 or TNI 2009 standards. ISO 17025 standards (e.g. 2005 and 2017) specifically require a separation of monetary benefits for laboratory results as it applies to the technical staff. This “conflict of interest” (CoI) is not always clearly defined in the laboratory’s day-to-day practices.

One example that I have experienced with this CoI separation violation goes back to my days as a laboratory troubleshooter in the 1990s. I was called into a laboratory that was failing to meet their Department of Defense (DoD) contract for volatile organic hydrocarbon analyses (VOAs) of soil samples by purge trap-gas chromatography-mass spectroscopy. I was required to “fix” the problem. What I determined was:

The analytical chemists performing the VOAs analyses were high school graduates with no coursework in chemistry or biology.
There was no training program in place for these analysts in instrument use, instrument troubleshooting and interpretation of the analytical results.
The only training the analysts received was for simple instrument set-up and basic instrument computer software use. (e.g. Push this button and send results to clerks)
Clerks with a high school degree and no analytical chemistry training in the business office generated the final reports and certified them as accurate and complete.

None of the staff was technically competent to perform any in-depth VOAs analytical work nor was the clerical staff competent to certify the results reported.

When I pointed out these discrepancies to the laboratory management, they declined to make any changes. The laboratory management had a direct monetary interest in completing all analyses at the lowest costs within the time limit set by DoD. If the laboratory did not complete the analyses as per the DoD contract, DoD would cancel the contract and not pay the laboratory.

The DoD, in a “Double Blind” test sample, later caught this laboratory.. A Double Blind test sample is used to check to see if the laboratory is performing the tests correctly. The laboratory does not know it is a test sample. So if the laboratory is cheating, they will be caught.This does not mean that all laboratories have staff or management issues

Once the laboratory was caught by DoD with the Double Blind, laboratory management claimed they were unaware of this behavior and management fired all analytical staff performing VOAs and clerical staff reporting the VOAs results to show DoD that it was a rogue group of individuals and not the laboratory management. The fired staff members were denied unemployment benefits as they were fired with cause. So, the moral to this story is if the analytical staff and specifically the clerical staff had wanted to hold the laboratory management accountable for this conflict of interest, they may have been fired, but without cause. The staff would have kept their reputation for honesty and collected unemployment benefits.

I have witnessed the “CoI above repeatedly over the last 30+ years both in laboratories where I have been employed and as a consultant. The key laboratory culture problems that lead to these CoI issues can be distilled into the following categories:

Financial CoI: In the financial CoI, the laboratory management must turn out so many analytical test results per day to remain financially solvent. The philosophical change that comes over management is that the laboratory is not producing scientific results, but is instead just churning out tests. Therefore, the more tests the laboratory produces, the more money it makes. Any improvement in test output is to be looked upon favorably and anything that diminishes test output is bad. So, to put this in simple terms: “The laboratory will perform the analyses quickly and get the report sent to the user so the laboratory can be paid. Anything that slows this production down will not be tolerated!” To maximize the Return on Investment (RoI) for the laboratory, management will employ staff that outwardly mirrors this philosophy.
I Need This Job CoI: This is the CoI area that poor quality lab technical staff and clerical staff most readily falls into. As outlined in the example above, both the analytical staff and clerical staff lacked the educational credentials, the technical training to be proficient in the use of the analytical instruments, ability to identify problems performing the analytical methods or complications in reporting analytical results. That means they were locked into the positions they held in this specific laboratory. This lack of marketable skills placed pressure on these staff members to comply with all directives from management. What happened to them in the end was regrettable, but predictable. Management can prey on this type of staff limitation.
Lack of Interest or Care CoI: This form of CoI is the malaise that infects poor quality laboratories, but can reach a level in management, quality and technical areas as to produce a culture where everyone goes through the moves, but does not care about anything but receiving their paycheck. In my many years of laboratory troubleshooting this type of CoI is the most difficult to correct. Laboratories where I had to correct this problem required that I had to impress on the staff that their work mattered and that they were valued employees. I had to institute a rigorous training program, require staff quality milestones and enforce the quality of work results. During my years of laboratory troubleshooting, I only had to terminate three laboratory staff for poor work performance. Unfortunately after I left many of these laboratories, management drifted back to the problems listed above and the laboratory malaise returned. This proves that even though a laboratory staff can achieve quality performance, it can quickly dissolve with lax management.

So, what are the conclusions of this article?

Laboratory culture can place profit over scientific correctness, accuracy and precision.
Laboratory management sets the quality of staff that determines the analytical results and report quality the user receives.
Laboratory quality can vary from acceptable performance to unacceptable performance over the lifetime of the laboratory depending on management.
This does not mean that all laboratories have staff or management issues. It is up to the client (processor, grower or dispensary) to determine the quality of the lab they use.

The next article in this series will introduce the user to the specific Quality Control (QC) analyses that an acceptable laboratory should perform for the user’s sample. These QC analyses are not always performed by accredited laboratories as the specific state that regulates their cannabis program does not require them. The use of these QC samples is another example of how laboratory’s with poor quality systems construct another paper work wall.

November 30, 2016

Steep Hill, ACCL Find Pesticides in Over 50% of Cannabis Samples

By Aaron G. Biros

4 Comments

On Election Day, voters in California passed Proposition 64, establishing a recreational cannabis market and regulatory environment. While the state won’t issue the first licenses under the new regulatory scheme until 2018, the medical cannabis industry is already well established.

Steep Hill Labs, Inc., based in Berkeley, California, found in October that 84.3% of samples submitted tested positive for pesticide residue, according to a press release. The announcement came before Election Day, but is particularly eye opening given the massive new market created overnight by Prop 64.

Particularly concerning is their detection of Myclobutanil, which was found in more than 65% of samples submitted to the lab. According to the press release, when Myclobutanil is heated (i.e. smoked or vaporized), it is converted to Hydrogen Cyanide, which is extraordinarily toxic to humans and can be fatal in higher doses.

Reggie Gaudino, Ph.D., vice president of scientific operations and director of genetics at Steep Hill Laboratories. (photo credit: Preston Gannaway) — Reggie Gaudino, Ph.D. (photo credit: Preston Gannaway)

According to Reggie Gaudino, Ph.D., vice president of science, genetics and intellectual property at Steep Hill, their more recent study shows they detected pesticides in roughly 70% of the samples they received and 50% of those contained Myclobutanil. Gaudino says that up to a third of those samples would have failed under Oregon’s regulatory standards.

If a lab test were failed, it would contain pesticides at or higher than the required action level. Oregon’s action level, or the measured amount of pesticides in a product that the OHA deems potentially dangerous, for Myclobutanil is 0.2 parts-per-million (PPM). Steep Hill’s instrumentation has a method detection limit down to the parts-per-trillion (PPT) level, which is a more precise and smaller amount than Oregon’s action level.

“Those in the cannabis community who feel that all cannabis is safe are not correct given this data – smoking a joint of pesticide-contaminated cannabis could potentially expose the body to lethal chemicals,” says Jmichaele Keller, president and chief executive officer of Steep Hill. “As a community, we need to address this issue immediately and not wait until 2018.”

Potentially harmful pesticides, and specifically Myclobutanil, have been detected in Colorado and Washington’s recreational markets on a number of occasions, proving this is a widespread issue. Steep Hill’s release suggests that California regulators take a look at Oregon’s pesticide regulations for guidance when developing the regulatory framework.

What’s even more troubling is that not all laboratories have or had the capability of detecting pesticides at sufficiently low levels and because of this, other labs had significantly lower rates of pesticide detection, suggesting possible inconsistencies in testing methods, instrumentation, sample preparation or other variations. During a 30-day period in late September and early October, Steep Hill found, using publicly available data, or data from contracted testing, that other labs were only reporting between 3% and 21% pesticide detection.

Examination of cannabis prior to testing- credit Steep Hill Labs, Inc.

It is important to note that those samples were not identical and there could be a great degree in variation on the quality of samples sent to different laboratories, so it is not an entirely accurate comparison. Steep Hill does however detect pesticides down to the parts-per-trillion level, whereas many common methods for detecting pesticides look at the parts-per-billion level.

Reggie Gaudino says the Association of Commercial Cannabis Laboratories (ACCL) is using this data to work with Steep Hill and a number of other labs to address these issues. “As a member of the ACCL, and after discussion with ACCL, we have agreed that all future discussion of this issue should not include laboratory names, as this is about educating the industry in general, and making sure all members of the ACCL are developing the best possible methods for detecting pesticides,” says Gaudino. “The ACCL has responded to this data, by inquiring on a larger, industry-wide basis, which represents a better picture of the issue, rather than only in California’s still-technically unregulated market.” The important message is this is a major issue that needs addressing urgently. “As such, the troubling issue remains, across the larger ACCL membership, there is still detection of pesticides in at least 50% of the cannabis being tested.”

According to Jeffrey Raber, Ph.D., president of the ACCL, the industry is experiencing a pesticide problem, but it is very difficult to quantify. “It is fair to say that around 50% of the cannabis being tested contains pesticides, but we really don’t know that exact number until a much more comprehensive statistical analysis is performed,” says Raber. “We agree this is a big problem and that it needs to be addressed, but we are not sure just how big of a problem it really is.” With so much variation in labs in a state where not everyone is required to test products, it is very difficult to pin down how consistent lab results are and how contaminated the cannabis really is. “If all of the labs had the same methodology, samples and shared statistical analyses for a real study then we can look at it closely but it seems we are a ways off from that. I can say confidently however that this is a pretty significant problem that needs addressing.”

Still, Steep Hill detecting pesticides in a majority of their samples and some labs finding as little as 3% should raise some eyebrows. “Unfortunately, our recent study discovered that 84.3% of the samples assessed by our triple quadrupole mass spectrometer contained pesticides,” says Keller. “As of today, this tainted product could be sold in most dispensaries throughout the state of California without any way of informing the patients about the risks of pesticide exposure.”

These findings could mean potentially enormous health risks for medical and recreational cannabis consumers alike, unless regulators, labs and growers take quick action to address the problem.