Here’s how we tested some of the leading brands of St. John’s wort, the popular herbal antidepressant.
We went to a CVS store in Cambridge, Mass., and bought the following products: CVS’ house brand; Natrol; NatureMade; Nature’s Resource; Quanterra; and YourLife. In addition, we obtained a bottle of Herbalife, which is sold privately through distributors.
We placed several pills of each brand in two sets of unmarked, identical-looking bottles and coded each brand. The CVS brand, for instance, became BG-001 (for Boston Globe), Herbalife became BG-002, etc.
We then sent one set of coded samples to an herbal testing company in Ithaca, N.Y. called Paracelsian, Inc. and hired it to do biological assays. The goal was to see whether each product worked – as judged by its ability to block the “reuptake” (or absorption) of both serotonin and dopamine, two neurotransmitters involved in depression, in rat brain cells in a test tube.
We sent the other samples to a testing company called PhytoChem Technologies, Inc. in Chelmsford, Mass., and hired it to do a chemical analysis of each of the products. The goal was to see if the products had as much hypericin, an ingredient in St. John’s wort, as the labels indicated.
In addition, unbeknownst to either company, we included an eighth set of pills in each set, placebos, or “dummy” pills, supplied to us by the Massachusetts College of Pharmacy. This was a way of testing the tests – a placebo should not pass the biological assay (and it didn’t) nor should it have a chemical profile like that of real St. John’s wort products (and again, it didn’t.)
After we got the results, we analyzed them with the help of independent scientists, including Paul S. Blum, a neuroscientist in Cambridge, Mass., who helps drug companies develop new drugs, and Timothy Maher, director of pharmaceutical sciences at Massachusetts College of Pharmacy.
We also discussed our results with the manufacturers of all the products we studied and herbal industry spokespeople.
Details of the testing methods: In the biological analysis, Paracelsian compared each product to Prozac and Zoloft (prescription antidepressants known as selective serotonin reuptake inhibitors) and to a “reference” herbal product, the Perika brand of St. John’s wort, which has been shown to inhibit the reuptake of the neurotransmitters serotonin and dopamine in rat studies, and to reduce mild to moderate depression in some people.
Prozac and Zoloft work by blocking the reuptake into brain cells of serotonin, and to a lesser extent, by blocking dopamine as well. Paracelsian’s assumption was that if St. John’s wort works, it may act by this same mechanism, that is, blocking neurotransmitter reuptake.
The Paracelsian scientists took cells from rats’ brains and put them in a test tube with a sample of St. John’s wort. They then added serotonin and calculated how much was left in the test tube after several minutes. The more that is left in the test tube, the better the antidepressant effect.
The Paracelsian team then repeated this for dopamine.
In the chemical analysis, PhytoChem did two tests on each sample – a spectrophotometric assay (called DAC 1991) and HPLC, or high performance liquid chromatography.
In the first, PhytoChem passed light through each sample.
Because scientists know what wavelengths of light hypericin absorbs, they can calculate how much hypericin is in a product – the more light absorbed, the more hypericin the product contains. For each sample, PhytoChem did this test twice and then averaged the results. PhytoChem also compared each result to the amount of light absorbed by a known control substance.
In the second test, the HPLC, PhytoChem injected each sample into a stainless steel column packed with silica (like powdered glass or sand). The silica was then treated with a special surface coating. The individual substances in St. John’s wort were then separated based on how well they stuck to this coating.
As individual compounds are separated out and leave the column, they can be detected by the absorption of light. The resulting pattern of peaks exiting the column at particular times and different intensities of light absorption produce a “chromatogram” – essentially a “fingerprint” of the constituents of a particular plant. The “fingerprint” of St. John’s wort is clearly different from that of, say, ginseng.
(We used only the first type of chemical testing, the spectrophotometric method, to gauge the amount of hypericin in each product. We did not use the HPLC test results to calculate hypericin; we used it only in a qualitative sense. It showed that all of our sample products except the placebo had the typical St. John’s wort fingerprint.)
Strengths of our study: We used two different companies for biological and chemical testing and independent consultants to help analyze the results. We blinded the testing companies to the identity of the products. We also did not tell them we had submitted placebos. And we bought the products as any consumer might – in a popular drug store.
Drawbacks of our study: Our study was not a clinical trial in humans, we did not re-confirm our results with other testing companies, and we did not test all available brands, including two major products, Kira and Movana.
The Paracelsian assay may not test the correct mechanism for St. John’s wort. While some scientists think St. John’s wort works by blocking serotonin reuptake, others are unconvinced. That means some samples that flunked the Paracelsian test might pass a test based on another mode of action.
The PhytoChem analysis isn’t perfect, either. It tested our samples for hypericin, which scientists used to think was the active ingredient in St. John’s wort. Now, some scientists think the active ingredient may be a chemical called hyperforin.
Still, hypericin is the ingredient manufacturers use to standardize products and it’s the one that’s touted on the labels.