The September 2006 issue of Pharmacotherapy [26(9): 1269-1272] published a paper by William R. Wolowich, Pharm.D., Alisha M. Perkins, M.D., and John J. Cienki, M.D., titled "Analysis of the Psychoactive Terpenoid Salvinorin A Content in Five Salvia divinorum Herbal Products." The analytical results reported were intriguing. However, an obvious error in the construction of their study caused the researchers to fabricate product claims, and then complain that the products did not live up to these claims! The researchers conclude by warning that the dangers of mislabeled products may "pose a potential risk of both misuse and overdose." Frankly, I am shocked that a paper that so clearly manufactured the figures it presents, ever made it into publication.

The paper begins with a brief, well-informed history of the ethnographic and contemporary use of Salvia divinorum, followed by a description of the plant's pharmacology. A couple of minor errors are presented: when comparing the potency of salvinorin A to other entheogens, they give the dose of mescaline as "100 mg." Mescaline doses generally start at two to four times that much. As the source for their dosage information, they strangely cite a paper (Kalant 2001) that primarily deals with the topic of MDMA, and which never mentions any dosage amount for mescaline. Wolowich et al. similarly low-ball the "5 mg" dose they give for psilocybin. Ott describes psilocybin as "psychoptic above 10 mg" (Ott 1996), while the Shulgins list a dose range of "10-20 mg" (Shulgin & Shulgin 1997). Strangely, again, the dose-response study that they cite (Hasler et al. 2004) doesn't ever describe research done with a "5 mg" dose of psilocybin; rather, it presents a range moving from placebo (0 µg/kg body weight) to "very low dose" (45 µg/kg), "low dose" (115 µg/kg), "medium dose" (215 µg/kg), and "high dose" (315 µg/kg). For a person weighing 70 kilograms (154 pounds), the doses of psilocybin used would have ranged from about 3 mg to about 22 mg. It is impossible to determine how the authors obtained the "100 mg" and "5 mg" doses that they present. And while these dose errors are minor and not particularly relevant to the research being reported on, they are nevertheless the sort of thing that should have been caught in a peer-review process.

Wolowich et al. mention how another research paper described a rat model, used to screen for efficacy of antidepressant drugs, that suggested salvinorin A produced depressant effects. Wolowich et al. then segued this into a mention of "the case of a salvia user who committed suicide," noting that the "potential for depression caused by salvia consumption must...be considered and needs further investigation." While it is possible that Salvia divinorum use might contribute to or cause depression (and rats aren't really talking about it), it seems worth mentioning that humans have anecdotally reported that S. divinorum can have antidepressant effects (Hanes 2001; Hanes 2003; Siebert 2002; Siebert 2007).

The stated objective of their study was to "determine the content of the hallucinogen salvinorin A in a variety of Salvia divinorum herbal products and to compare the content with the label claims of potency and purity." The researchers ran five commercially available products through high performance liquid chromatography and thin-layer chromatography-gas chromatography-mass spectroscopy. Two of the samples were found to be adulterated with vitamin E, and one sample was adulterated with caffeine. And while each of the samples contained salvinorin A, the levels they contained were surprisingly low.

The paper claims that: "The purveyors of salvia use nomenclature intended to imply standardization; 1x is the potency of the natural product." The paper then further cites an Internet users forum claim that 1x is "equivalent to a salvinorin A content of 2.5 mg/g." And it also cites a salvia users guide explaining that "10x means 10 times the potency of 1x." Using these figures, the paper's authors then manufacture a specific amount of salvinorin A that they misleadingly propose should be present in assorted commercial products, either plain dried leaf, or fortified leaf marked 5x, 10x, or 20x. (Fortified leaf products are created by using a solvent to create a liquid extract of the leaf material, then depositing that liquid onto some smaller amount of non-extracted leaf material and allowing it to dry.)

The trouble is that the "X" nomenclature used by many purveyors of Salvia divinorum products is not intended to imply standardization. Rather, it only claims to present information about how potent the product is relative to the potency of the product's starting material. The potency of natural products can vary dramatically; in some cases a correctly identified botanical may contain virtually none of the chemical targeted for medicinal or recreational use. This is a well known issue in the herbal medicine industry.

The authors of this paper were unquestionably aware of the fact that the salvinorin A content of Salvia divinorum, specifically, can be highly variable. Why? Because the very analytical method that they cited as having used to extract and quantify the salvinorin A from their samples (Gruber et al. 1999) reported finding a salvinorin A content in botanical samples ranging from 0.63 mg/g to 3.70 mg/g--they mention these figures in their own paper. (The low-end figure, which was actually "<0.63 mg/g" in the original paper, can probably be chucked, since it was taken from a stem analysis; the low-end figure for a leaf sample was 0.89 mg/g.) Of the 20 unfortified leaf samples tested, ten of them had more than 2.5 mg/g and ten of them had less than this; and yes, the average salvinorin A content was 2.45 mg/g. How, then, do these figures mean that all Salvia divinorum leaves can be considered to uniformly contain 2.5 mg/g? Taking a median or a mean and applying it to non-standardized commercial products is, quite simply, a flawed approach. The problem is magnified as the "X" factor increases. Based on the figures from Gruber et al., a 5x product might be estimated at containing anywhere from 4.45 mg/g to 18.5 mg/g, and a 20x might contain from 17.8 mg/g to 74 mg/g. These figures clearly indicate that within the known range of salvinorin A content, one could quite possibly obtain a 5x product that was stronger than a 20x product. How could the paper's authors, and any peer reviewers, have possibly missed this fact?

The paper's authors repeatedly describe the products that they analyzed as having a "label claim (mg/g)" related to how much salvinorin A each product contained. But, unless any of the products actually did make claims about the amount of salvinorin A they contained, then the approach that these authors have taken is a fiction. (I twice e-mailed the paper's primary author, William R. Wolowich, asking whether or not any of the products that they tested specifically listed a claim on its label regarding how much salvinorin A it contained. Unfortunately, he never responded to my inquiries. Considering the mathematical gymnastics which the paper's authors present in order to create their "standardized" label claims, it seems highly unlikely that any of the products they tested actually made any specific salvinorin A content claims, since--if they had--then these claims could have much more easily been taken directly from the labels themselves.)

Inspired from a description of such preparations in Dale Pendell's Pharmako/Poeia, the first of the "X-strength" Salvia divinorum extracts appeared on the commercial market in November of 1997. By the summer of 1999, products were being marketed that actually did contain a standardized salvinorin A content. Vendors in this new market realized (in relatively short order) that natural products can be highly variable in their chemistry, and they responded by extracting and purifying the salvinorin A, then redepositing it onto leaves in measured amounts. Although the "X-strength" method of naming Salvia divinorum products largely continued, some of these products now specifically state how much salvinorin A they contain. But not all of them. There are still a number of crude, non-standardized extracts available, and strangely Wolowich et al. appear to have focused their attention entirely on these, while misrepresenting them as if they made claims to be standardized.

Nevertheless, the salvinorin A content of the products they analyzed does seem oddly low. The five products they tested were comprised of one sample of unfortified leaves (0.408 mg/g), two "5x" products (0.126 & 1.137 mg/g), one "10x" (0.951 mg/g), and one "20x" (0.461 mg/g). The plain leaves had less than half the salvinorin A content of the lowest figure previously reported (Gruber et al. 1999), and the "20x" only had slightly more than the plain leaves. Although a pool of five products is not a huge sampling, it is somewhat surprising that all of these products had such a low salvinorin A content. It is also odd that two of the samples contained vitamin E and one of them contained caffeine. Unfortunately, the paper did not mention any brand names for the products that they tested; hence, the reported results lack any teeth, so far as consumer advocacy goes. If anything, the results presented make a good argument for consumers to only purchase standardized extracts of Salvia divinorum from reputable vendors. Perhaps Wolowich et al. might consider a future study wherein they subject a number of those products that do make claims of standardization to analysis, in order to determine how accurately these products are labeled.

So what the heck is with the low-potency Salvia divinorum products that Wolowich et al. analyzed?

It would be interesting to know if the "1x" sample that they tested was whole leaf, or whether it was sifted/crushed leaf. (This was another question that I asked Wolowich, which went unanswered.) Despite the fact that salvinorin A content can be variable, the results reported just seem too damn low. One can only speculate:

1) The researchers' testing procedure was flawed. (This strikes me as the most likely scenario.) Having an independent group retest the same material could determine if this was the case. Hopefully reference samples of the materials tested were retained and are available for independent testing; if not, the product brand names should at least be provided to anyone who wishes to perform verification tests.

2) The salvinorin A content of Salvia divinorum leaves turns out to frequently be much lower than has previously been thought based on analysis. Lower levels might be related to overharvesting, or harvesting at the wrong time of year.

3) Companies creating assorted X-strength fortified leaf products are not using the appropriate solvents or extraction techniques to increase the potency of their products.

4) Some manufacturers have purposefully been producing weaker products, perhaps due to concerns about being sued if someone has a bad experience from consuming a full strength product.

5) One or more major Salvia divinorum wholesalers in Mexico is diluting their leaves with some other inert botanical before shipping them out. At least with any "1x" material, it should be possible to determine if this is the case by utilizing microscopic analysis, as previously described in The Entheogen Review (see "Establishing Identification Methods for Ethnobotanical Medicine" by Sidney Sudberg and Elan M. Sudberg, TER 12(2): 45-48).

If a whole lot of subpotent Salvia divinorum extracts were floating around, one might think that there would be some degree of noise about this in the psychonautical community. So far, I have not heard anything. Am I simply too enmeshed in the "grow yer own" crowd? I would be interested to hear any thoughts on this matter from ER readers.