Fentanyl or Not? Re-analysis of Samples #5776 and #5779

As covered last week in Fentanyl Test Strips, Hot Spots, and Unhomogenized Samples, two samples reported in the November 6, 2017 batch of EcstasyData results were submitted to the lab with notes by the senders saying they had used fentanyl field tests on their sample before sending it.

The original GC/MS results for these samples (#5776 and #5779) did not reveal fentanyl. After writing about the problems associated with non-homogeneous samples and “hot spots”, we decided to use this as an opportunity to re-test both of these samples. This time, we used up 100% of what was sent by dissolving all of the sample (and not just a small portion), to make sure not to miss any potential “hot spots” in the original. That is a different method than our normal sampling protocol, meant to verify the no-fentanyl results.

On November 12, the lab reported back that the re-analysis of one of the samples returned a different result than the original analysis. The other sample’s result did not change.

Re-tested Using Modified Sample Prep Method

With most EcstasyData samples, there is material left over after an analysis. Whatever is not destroyed in the GC/MS process is stored for secure disposal after one year. This permits the lab to re-test a sample, potentially several times, if circumstances call for a re-test.

A modified method was used to re-test samples #5776 and #5779: For each of the re-tests, all the material, including the capsule, that was left over after the original analysis was placed in solvent to produce a consistent liquid sample, which was then run through the normal GC/MS process. This eliminated the “sub-sample of a sub-sample” condition that we described in the previous article.

Heroin Sample 5779: Positive fentanyl test-strip confirmed by GC/MS

The sample whose result did change after re-analysis was a powder represented as heroin.

The person who submitted the sample had noted the ‘Sample tested positive on both ‘DanceSafe’ and another brand of Fentanyl test strips (or cassette / dip card). Would like to know if these immunoassay fentanyl tests actually work.’ While the first GC/MS analysis did not detect fentanyl (even though a lot of the material was prepped for testing and the sample appeared homogenous), the second analysis that followed the method above did detect fentanyl.

This discrepancy in results can most likely be attributed to a hot spot or spots in the powder.

The lab also identified several other substances that they did not report in the first analysis: trace amounts of codeine, 4-ANPP, papaverine, and very small amounts of actylecodeine and 6-monoacetylmorphine. The additional very small findings are normal minor components of poorly-cleaned heroin produced from natural poppy resin. When we re-test a sample looking for very potent substances like fentanyl, we take a closer look at the trace and near-trace tiny “noise” bumps in the GC readouts.

Although we do report trace substances in most cases, street heroin is a good example of the type of material that often contains a lot of “noise” because it’s not pure, nor is it a combination of drugs; it’s a partially-synthesized natural product with lots of leftovers.

Because of the issue of fentanyl and fentanyl analogs showing up in heroin, EcstasyData has obtained a number of NPS-fentanyl analog standards and will be taking extra care to look for them in future heroin and opioid samples. Please include on your submission forms if you have concerns about fentanyl in your sample so the lab tech knows to look at what we normally consider “noise” in messy samples, and to set up the Gas Chromatograph run so that we make sure he can differentiate noise from signal at the time points where fentanyl and the known fentanyl analogs come out of the column.

MDMA Sample 5776: False-positive fentanyl test-strip

The sample whose result did not change with re-analysis was for a powder represented as MDMA. Only MDMA, with no traces of fentanyl or other compounds, was detected using the method described above. The person who submitted the sample had noted a ‘strip test tested positive for Fentanyl’ prior to sending their sample in. This strip-test result was not confirmed by GC/MS.

Evolving and Improving Procedures

Although we run one of the best analytical projects of its kind in the world, this is a great example of the many types of errors and misses that can occur, and we take steps like those described here, to verify our results and change our procedures to improve the accuracy over time.

— Sylvia and Earth

Fentanyl Test Strips, Hot Spots, and Unhomogenized Samples

With fentanyl and fentanyl analogs haunting the opioid crisis in North America, some harm reduction field workers and users have been experimenting with what cost-effective reagent-based field tests might have to offer. One method that has been explored is the repurposed fentanyl urinalysis test-strip, where rather than dipping the test-strip in urine, it is dipped in a solution of the drug itself. A panel presentation covering the topic of drug checking and the opioid crisis was held at Drug Policy Alliance’s 2017 Reform Conference.

Since two samples included in the November 6, 2017 batch of EcstasyData results came with notes saying the sender had used a fentanyl test-strip on their sample before sending it, we’re starting to look at what that means for how we report EcstasyData results in such cases.

The question was posed by one sender,  “Would like to know if these immunoassay fentanyl tests actually work.” Like many Yes/No questions that people have about drug analysis, the answer is a combination of “it depends” and “it’s complicated”.

Hot Spots
In answering the question, it helps to remember that not all powders will be completely evenly homogenized and may contain “hot spots” with uneven concentrations of a given chemical in sub-parts of the larger amount of powder or crushed crystalline material.

Powders and tablets that have more than one component to them aren’t always evenly mixed. Sometimes there’s a higher concentration of a drug in one or several areas. Think of a burrito that has hot salsa in one end of the burrito but not the other. If you bite into the burrito on one end, it’s spicy. On the other end, it’s not. Or a chocolate chip cookie — the chocolate chips might not be evenly distributed in the cookie. We’ve discussed homogenization in EcstasyData samples before when talking about traces of drugs in samples sent to the lab.

Hot spots are a bigger deal when it comes to fentanyl drugs or other similarly potent drugs, that are active at doses below a milligram but are sold in powders weighing hundreds of milligrams or grams.

Sub-samples of Sub-samples
When a powder sample is received by the EcstasyData lab, the lab tech preparing it for analysis first does a very simple partial homogenization by shaking the sample container or stirring it a little before extracting a sub-sample with a clean metal or plastic spoon. The sub-sample is then dissolved in a solvent and the technician confirms that the material dissolves completely or may take additional steps to get a fully dissolved, consistent liquid sample.

The dissolved sub-sample is then inserted into the testing equipment (the GC/MS). If fentanyl is present in that solution, then fentanyl will show up on the test, and we report it. If fentanyl is only present in a part of the sample that was not dissolved into solution, the fentanyl can not be detected.

So there are at least two steps of sub-sampling that occur before a tiny amount of material reaches the GC/MS:
1) The sender takes a sub-sample out of their stash/bag/jar at home and sends it to the lab.
2) The lab tech takes a sub-sample of that sub-sample to dissolve and inject into the GC/MS.

Someone using test-strips on powders or tablets at home is facing a similar issue of potential hot spots.

Can’t Be Sure
If a fentanyl test-strip is used to check a drug sample and the result is consistent with the presence of fentanyl (a so-called “positive”), there are several possible explanations, some including fentanyl being present, and some where there is not fentanyl present (“false positive”). A test-strip can also be difficult to read.

Some Positives are False Positives
A positive result with a test-strip can mean fentanyl is present in the sample, either deliberately, or by contamination. Tablets might have come into contact with the substance and have a tiny residue left that could trigger a positive field test.

An unknown and unknowable number of other conditions can cause false positives on urine drug screen strip-tests. Non-fentanyl drugs might trigger the field test to show positive. The problem of false positives is the reason urine strip-tests alone are insufficient to prove someone has taken a given drug in legal or employment contexts. Positives from urine screens are always double checked using an advanced technique such as GC/MS to confirm or exclude the simpler, cheaper strip test.

Test-Strip Outcome Can Be Hard to Interpret
Test-strips and other field tests often have a wide range of possible strengths of color changes. It is common to get results that are not 100% clear in what they mean.

EcstasyData can’t comment on whether fentanyl test-strips in general are useful in detecting fentanyl in drug samples. Each situation is unique.

We are keeping an eye on this topic.

— Sylvia and Earth

PS: In August 2017, Erowid confirmed that the DanceSafe fentanyl test strips give false positives for buprenorphine. When mixed at a concentration of .01mg per ml water. We have another draft post that hasn’t seen proper review yet that goes over this in detail.

Marquis and LSD: Is color change visible?

Short Summary: Marquis, Mecke, Mandelin, and Simons are not reliable tools for confirming or disproving the presence of LSD in blotter, liquid, gel, or tablet forms.

 Figure 1 Confirmed LSD blotter produces little or no reaction to Marquis, Mecke, and Mendelin when tested on white porcelain.
Figure 1 Confirmed LSD blotter produces little or no reaction to Marquis, Mecke, and Mendelin when tested on white porcelain.

Long Version: We get asked a lot about how specific drugs react to the various drug-detection field tests like Marquis, Mecke, Mandelin, Simons, Robadope, Ehrlich’s, etc. In most cases, it’s a matter of getting some of the pure target compound, a fresh set of reagents, and doing a little testing, photographing, and documentation..

However, with drugs active below 1mg, such as LSD, this may not be so simple. Because the amount of target compound is often very small, the reactions can be altered, slowed, or blocked by tiny amounts of other substances present. In the case of LSD blotter paper, where the amount of LSD on a 1/4″ square (6mmx6mm) is usually at or below 100 micrograms, the paper and the ink on the paper are far more likely to be the cause of a color change than the LSD itself. With liquid LSD, the alcohol carrier can dilute the response enough that no color change is visible. Thus, a color change or the lack of color change can be due to the form in which the substance is being tested.

LSD is said to create an olive green or black reaction with a Marquis reagent test. Organizations that sell reagent tests such as Dancesafe and Bunk Police report that LSD has an olive-black reaction with a Marquis test. This may be based upon sources such as this Department of Justice article stating that LSD causes an “Olive black” result. [ Fatah A. “Color Test Reagents/Kits for Preliminary Identification of Drugs of Abuse” (2000) ].

Figure 2 LSD sample confirmed by GC/MS produces no response to Marquis, Mecke, or Mandelin reagents.  Black squares shown to illustrate blotter location do not indicate a “black” color reaction to these tests.
Figure 2 LSD sample confirmed by GC/MS produces no response to Marquis, Mecke, or Mandelin reagents. Black squares shown to illustrate blotter location do not indicate a “black” color reaction to these tests.

However, the results of our own lab’s field tests on samples confirmed to contain LSD using GC/MS show varying responses to Marquis and other field tests. When tested with a Marquis reagent, most of these samples showed no reaction or only a very slight reaction and none produced an olive green or black reaction. For examples of these test results see Figure 1 and Figure 2. Furthermore, the blotter paper itself—not the LSD—may be causing these slight reactions.

You can see a list of numerous LSD samples on EcstasyData along with their field test color changes here:


Despite claims that LSD produces an olive green or black reaction to Marquis, evidence collected and published through EcstasyData does not support this. Furthermore, a 1979 paper by by Johns et al is consistent with these EcstasyData results. The researchers tested a wide variety of drugs from that era with nine different reagent tests, including Marquis and Ehrlich (Table 1). This paper reports that LSD only reacts with Ehrlich and not with any of the other reagents. Interestingly, the researchers used dry LSD from Sandoz Laboratories. Other publications claiming that LSD produces olive green or black reactions to Marquis do not report the source or form (i.e. liquid, crystal, or blotter) of the LSD tested. The results of the Johns et al., study suggests that confirmed pure crystal LSD has no reaction with a Marquis test. Below is a summary of the results:

Compound Reagent test Color reaction Testing surface
LSD Marquis no reaction Porcelain
LSD Mecke no reaction Porcelain
LSD Madelin no reaction Porcelain
LSD Cobalt thiocyanate no reaction Porcelain
LSD Dille-Koppanyi no reaction Porcelain
LSD Ehrlich purple Porcelain
LSD Froehde no reaction Porcelain
LSD Sulfuric acid no reaction Porcelain
LSD Nitric acid no reaction Porcelain

Although other sources say that LSD causes an olive-black reaction with a Marquis reagent test, this study, along with the EcstasyData results, suggests that LSD may not cause an olive-black reaction with a Marquis test—at least not on blotter paper in the amounts it is commonly sold. Based on this evidence, the Marquis is not a reliable for test for LSD. On the other hand, the Ehrlich may be an alternative test for this purpose based on the findings published by Johns et al. The Ehrlich reagent produces a color change when LSD or another indole is present. However, like all reagent tests, an Ehrlich is not a confirming test, it is just another in a wide array of rule in / rule out tests that can help confirm or deny the presence of LSD in a given sample.

Thanks to W and Shaolin at the DMT Nexus for their work on this issue.