By Judy Foreman, Globe Staff, 01/10/00

We thought it would be easy.

After all, we had just two seemingly simple questions: Does St. John's wort, the popular herbal adtidepressant on which Americans spend $250 million a year, work - at least on rat brain cells in a test tube? And do the product labels accurately reflect what's inside the tablets?

The path toward answers proved tortuous indeed.

We hired two companies, Paracelsian, Inc. of Ithaca, N.Y., and PhyttoChem Technologies Inc., of Chelmsford, Mass., to independently test seven St. John's wort products we purchased and repackaged into bottles coded by number. The companies didn't know it at the time, but we also sent each one an eight product, a placebo or inert drug, supplied to us by the Massachusetts College of Pharmacy.

We found that there was considerable chemical and biological variation among the products tested.

We'll give you tyhe bottome line next - but read on, because the caveats are important.

On the basis of the PhytoChem analysis, only one product, Nature's Resource, lived up to the standard claim on product lables that the products contain 0.3 percent of hypericin, a substance once thought to be the active ingredient in St. John's wort. (Now, to be the active ingredient, but the industry continues to standardize products to 0.3 percent of hypercin.)

Four other products, Natrol, NatureMade, Herbalife, and YourLife, were lower in hypericin, containing 0.28 percent, 0.27 percent, 0.25 percent, and 0.25 percent, respectively - less than their labels claimed.

With prescription drugs, the US Food and Drug Administration allows products to contain slightly less (10 percent below) to slightly more 10 percent above) than the contents stated on the label.

With diatary supplements, the FDA insists that they contain at least 100 percent of what's declared on the label, but because the agency does not require supplements to meet the same strigent requirements for safety and efficacy that it does for drugs, it does not specify what tests should be used to measure herbal ingredients or which labs should do them.

Although many of the products we tested fell short of their labeling numbers using the PhytoChem test, they might have passed with other testing methods in other labs.

One product, Quanterra, contained almost no hypericin, but its label makes no claim that it does.

According to the Paracelsian biological assays, only two products, Quanterra and NatureMade, passed the company's "BioFIT" test for their ability to block the reuptake of both serotonin and dopamine, two neurotransmitters involved in depression.

(Abnormalities in the reuptake, or absorption, of serotonin and other neurotransmitters are believed to be a major cause of depression; many prescription antidepressants work by blocking the reuptake of serotonin into brain cells, thus leaving more in the synapse, or gap, between cells.)

John Cardellina, vice president for botanical sciences at the Council for Responsible Nutrition, which represents 100 manufacturers of dietary supplements, said the Globe's chemical analysis actually "doesn't look bad for the industry." While only one product had as much hypericin as it listed on the label, "everybody was close to the mark. There's not much to complain about."

The biological assay, he felt, was more controversial. Paracelsian's BioFIT test is "not yet an accepted practice or marker," said Cardellina. One limitation is that, by measuring serotonin and dopamine reuptake, the test focuses on only "one of multiple mechanisms of action by which St. John's wort works." Another, of course, is that it is conducted in a test tube and does not involve human subjects.

Still, the fact that only two products - NatureMade and Quanterra - passed the BioFIT test is noteworthy. "Frankly, I'd expect to see more activity in that assay," Cardellina said.

When we shared our findings with manufacturers of the products we tested, some reacted with vigorous criticism.

Although its products fared better than most, Pharmavite, which makes NatureMade and Nature's Resource, noted there is always variability in chemical tests of herbal products. The company added it does "not believe that the data provided offers a reliable indicator of the quality of any of the brands tested."

It also said it had "significant concerns about the manner in which the tests were conducted and the apparent reliance on limited and possibly unsubstantiated test methodology in forming any conclusions about product quality."

On the other hand, Michael Cleary, director of product development for Natrol, Inc., welcomed our findings. "I think it's pretty healthy" to do such a study, he said. "There was not any hint of any unfairness in any of this."

Furthermore, he said, "To tell the truth, we expect people to pull our product off the shelf, and if they have it analyzed, to find what the label claims."

And to many of those familiar with the largely unregulated herbal industry, including the heads of the two testing companies we hired, our findings appeared to come as no shock at all.

"It's not surprising that few of these compounds passed," said Timothy Maher, director of pharmaceutical services at the Massachusetts College of Pharmacy, one of our independent data reviewers.

"People are buying these products and not always getting what they pay for," added Bernie Landes, the CEO of Paracelsian, one of the testing companies.

Even Robert Barry, the president of PhytoChem but a strong critic of our study, said that in the industry as a whole, more "rigorous study needs be done - 80 to 90 percent of what's out there has not been subjected to tests to see whether they provide a physiological effect."

Even if, as some in the industry point out, the bioassay used by Paracelsian measures only one of several possible mechanisms of action of St. John's wort, and even if the chemical analysis by PhytoChem had built-in limitations, our study nevertheless shows that quality control in herbal products is a big problem.

"You have demonstrated quite nicely with this study that there is lot of variation in St. John's wort products. I am concerned about patients self-treating a serious illness like depression with products whose contents they cannot count on," said Dr. Serena Koenig, a physician at Brigham and Women's Hospital in Boston who studies drug and herbal remedy interactions.

"Quality control is the number one issue in the herbal industry, no question about it," acknowledged Mark Blumenthal, executive director of the American Botanical Council in Austin, Texas, a nonprofit research and education organization funded in part by the herbal industry.

A big part of the problem, he added, "is the lack of federal or industry policies, regulations, or requirements that stipulate a particular method of analysis for SJW [St. John's wort] and other herbs."

The FDA does not require that herbal remedies, which are sold as dietary supplements, be approved for safety or efficacy before they are marketed. And it allows them to stay on the market unless there is evidence of harm.

Using PhytoChem's data, we did our own calculation and found that one product, Nature's Resource, actually had slightly more hypericin, 0.31 percent, than its label claimed. We calculated the percent hypericin in product extracts so consumers could compare this to the 0.3 percent hypericin stated on most labels.

But PhytoChem became uneasy. In a Dec. 17 letter, PhytoChem president Barry expressed "very grave concerns regarding any conclusions that may be drawn from the very limited testing that our company was asked to peform on the Saint [sic] John's wort products supplied by the Boston Globe."

In interviews and in writing, Barry also expressed his strong belief that the proper methodology would have been to have both the chemical and the biological evaluations "conducted by three independent laboratories," with at least one being a lab with no commercial interest in product testing.

Both PhytoChem and Paracelsian have a commercial interest in testing herbal products. Paracelsian plans to use its bioassay called BioFIT to provide a seal of approval for manufacturers to put on product labels. PhytoChem tests products under development for herbal and pharmaceutical companies.

Still, others in the herbal industry saw things more positively.

"I think of all the analyses done by the media, this is one of the more intriguing approaches because you have asked for chemical analysis and have looked at the emerging technology for a biological assay of the physiological activity" as well, said Cardellina of the Council for Responsible Nutrition.

"By pairing the content [chemical testing] and the activity question [biological testing], you did a good thing. . .I don't think you can go beyond that without running a clinical trial."

Dr. Scott Ewing, director of the depression and anxiety disorders clinic at McLean Hospital in Belmont, Mass., and a St. John's wort researcher, wondered why the BioFIT assay looked at dopamine, a neurotransmitter that may not play as big a role in depression as serotonin, and not at epinephrine, which does play a role. (Curiously, most products we tested fared better on the dopamine test than on the serotonin assay, though it's not clear why.)

In any case, says Ewing, the BioFIT data on dopamine "may not have any real, practical significance in terms of antidepressant effects."

Perhaps also puzzling was the fact that there was little overlap between the biological and chemical results. Paul Blum, another of our independent data reviewers and a neuroscientist and pharmaceutical consultant in Cambridge, Mass., did a statistical analysis of the correlation of the chemical and biological findings and found "the two tests don't correlate well with each other."

This could be explained, however, if the chemical for which PhytoChem tested, hypericin, is not the active ingredient in St. John's wort and some other ingredient, such as hyperforin, is.

Some of the variance in hypericin content that we found may be due to differences in raw materials and testing methods, noted Blumenthal of the American Botanical Council.

In its Dec. 9 letter to us, PhytoChem explained that it had used a spectrophotometric method called DAC 1991 to test for hypericin. Some companies test their products using an earlier test, DAC 1986. And it makes a difference.

"While not all the products passed the PhytoChem test based on the DAC 91 method, all but one would have passed by the DAC 86 standard, except for the CVS product, which comes very close," said Blumenthal.

"Today, you can call up supply houses that sell St. John's wort extract and you can buy two grades, one at 0.3 percent hypericin as determined by the DAC 1986 method, and the other, by the DAC 1991 method, which costs more. Therein lies part of the problem."

Because the price difference can be as much as 30 percent, some manufacturers may buy material that meets only the easier, 1986 standard, "especially with no industry or government-mandated standards for one or the other," said Blumenthal.

In part because of such problems, the industry is currently moving away from DAC testing toward a more precise, quantitative version of high-performance liquid chromatography testing, said Cardellina of the nutrition council.

Some of the product variance we found may also be due to other factors. St. John's wort "is sensitive to extremes of temperature," said Ewing, the McLean psychiatrist. It's also sensitive to light and to humidity. That means samples "left for many months on pharmacy shelves tend to degrade and therefore lose their potency."

Ultimately, the solution to quality control problems in the herbal industry is probably tighter regulation - of both manufacturing and labelling. If the government won't do it, consumer pressure may force the industry to do so.

In fact, a Denver firm, Industrial Laboratories, has formed the Institute for Nutraceutical Advancement, which is funded by the herbal industry to validate methods for testing products.

Blumenthal, of the American Botanical Council, welcomes such efforts. Hopefully, he said, "one of the results of this Globe study will be to hasten the adoption by industry organizations of uniform methods of labelling standardized products - an issue that the industry has been working on for some time."

All content herein is © Globe Newspaper Company and may not be republished without permission. If you have questions or comments about the archives, please contact us at any time.