Drug interactions with Natural Health Products: A Discussion Paper (November 2001)

3.0 Biomedical Issues

The potential for NHPs to alter the effect of conventional (or pharmaceutical) medications is beginning to be recognized by the public as well as the scientific community. Health care practitioners, from both conventional and complementary health care teams, are often required to assess the clinical impact of combining an NHP with a prescription or nonprescription medicine. A number of resources for information regarding the use of herbal medicines are available. However, depending upon the resources consulted, the quality of information regarding drug interactions associated with a particular NHP can vary significantly. For example, the potential of St. John's wort to interact with a number of drugs has been documented in both the scientific literature and the public press (Appendix 1). Yet some texts do not list any interactions between St. John's wort and conventional drugs or report that there are "no known interactions."(4, 5) Other references suggest that there is a "potential" for interactions if St. John's wort is taken with MAOIs (6) or state that there are "many potential interactions" and may list pertinent drugs.(7, 8) Sometimes the differences in the information provided occur because references were published before the interaction information was available; however, this is not always the case. This puts health care professionals in the difficult position of having to decide whether the reference they consult for information regarding NHP/drug interactions is as accurate and complete as possible.

Review of current state of knowledge

A survey of primary literature identified a number of articles, written for the health-care professionals that describe NHP/drug interactions.(9-21) The most useful of these references include tables that list NHP/drug interactions. While many of these NHP/drug interaction tables are extensive, what may not be immediately apparent is that very little of the information is based on rigorous scientific evaluation of actual interactions. Many entries are based on case reports of "possible," often unconfirmed, interactions and others are noted as "theoretical."

Our current understanding of NHP/drug interactions is largely a result of theoretical predictions based on the pharmacological actions of known chemical constituents of NHPs. NHP/drug interactions, like drug/drug interactions, can occur via a number of mechanisms.(22) In the classical sense, one drug may interact chemically with another, changing the structure and/or activity of one or both of the drugs. Very few, if any, NHP/drug interactions that involve a true chemical interaction have been identified. Drug interactions can also occur through pharmacodynamic and pharmacokinetic processes. Pharmacodynamic interactions affect drug action in a qualitative manner, while pharmacokinetic interactions are those in which one drug affects the absorption, distribution, metabolism or excretion of another. The significance of each of these types of interactions is illustrated below using examples of reported herb/drug interactions.

a) Pharmacodynamic interactions - when an NHP with a particular physiological effect is combined with a conventional medication possessing a similar action, the resulting physiological effect may be greater than expected, or desired. Conversely, an NHP may have activity that counteracts that of the conventional medication, reducing the expected response. The majority of pharmacodynamic interactions listed for NHPs are based on theoretical predictions of pharmacological activities of known constituents.

An example of a pharmacodynamic interaction between an NHP and a conventional medication was documented in a recent case report.(23) A 46-year-old woman was prescribed warfarin, digoxin and furosemide to treat heart problems. Her blood clotting, assessed by prothrombin time (PT) and international normalized ratio (INR) values, was stabilized in the appropriate range using warfarin. Without warning, both PT and INR values increased substantially, indicating a significant decrease in the ability of her blood to clot. Upon questioning, the patient revealed that she had recently begun taking dong quai, an herb used to treat menopausal symptoms. When she discontinued the dong quai, her clotting values returned to the desired level. Dong quai has been documented to contain a number of coumarins as well as ferulic acid, all of which can affect the clotting process. Coumarins and other constituents known to affect clotting are widely distributed throughout the plant kingdom. Therefore, most medicinal herbs known to contain these compounds have been identified as having the "potential" to interact with anticoagulant therapy, regardless of whether there have been any documented interactions. Since the chemical content of herbal products may vary both qualitatively and quantitatively, the magnitude of these interactions is difficult to predict. However, since warfarin and other anticoagulant therapies are considered to have low therapeutic indexes, it is wise to avoid combining anticoagulant therapies with any NHP that could theoretically affect clotting. Several good references that identify NHPs with a theoretical and/or documented risk of interacting in a pharmacodynamic manner with anticoagulant therapy are available.(19, 24-26)

Pharmacodynamic interactions between NHPs and conventional medications may be of particular concern for patients using conventional drug therapy for the management of chronic conditions such as hypertension, diabetes or depression. Several case reports of pharmacodynamic interactions resulting from the combination of NHPs and anti-hypertensive therapy have been documented.(12) As well, a number of theoretical interactions involving NHPs and blood pressure medication have been proposed based on the understanding that many herbs, such as garlic and ginseng, contain chemicals that are able to affect blood pressure, either directly through effects on the heart or blood vessels, or indirectly as a result of diuretic effects.(12, 27) The clinical significance of these interactions is difficult to predict. However, it is possible that a patient who maintains blood pressure control through the use of a prescription medication may experience undesired consequences if s/he takes an herb that also affects blood pressure.

Similar results are predicted for diabetic patients who combine hypoglycemic therapy with one of a number of herbs known to affect blood sugar levels.(27) American ginseng (Panax quinquefolius, also known as Canadian ginseng) recently received much public attention when its ability to reduce blood sugar levels was reported. In a small clinical trial, P. quinquefolius reduced blood sugar in both diabetics and nondiabetics.(28) However, the therapeutic response associated with adding a ginseng product to an antidiabetic treatment protocol is unpredictable because of inconsistencies in the content of commercially available ginseng products. Variation in product content can result from different processing, extraction and formulating parameters. One of the principle reasons for such variation between commercial products is that the common name "ginseng" refers to at least three different plants that all have different chemical or pharmacological profiles. Each species has been implicated in a number of common, as well as unique, herb/drug interactions (Appendix 2). It is interesting to note that some of the interactions reported for ginseng appear to be contradictory. For example, ginseng can increase or decrease blood clotting and can have similar effects on blood sugar.(25, 29) Definitive scientific evaluation is required to validate these findings and to determine if they are dose-related responses or if different components of ginseng are responsible for the opposing actions.

A final example of a pharmacodynamic interaction between an NHP and a conventional drug is provided by a case report of a patient who combined paroxetine (a selective serotonin reuptake inhibitor or SSRI) with St. John's wort in the management of depression. The clinical syndrome experienced by the patient resembled sedative/hypnotic intoxication and was believed to be a result of the additive effects of the two preparations.(30)

b) Pharmacokinetic interactions are those that affect how the body is able to handle a drug: its absorption, distribution, metabolism and elimination.

i) absorption - NHPs high in fibre (e.g., psyllium) and mucilages (e.g., flaxseed) may alter the absorption of conventional drugs.(8, 31) NHPs containing tannins may form precipitates with certain drugs, which can also reduce absorption. In addition, calcium, iron and iodine can chelate with some drugs to prevent their absorption. Some herbs may stimulate blood circulation and digestive secretions, resulting in enhanced absorption.

ii) metabolism - The discovery that grapefruit juice can alter the metabolism of some drugs increased our awareness and understanding of how substances other than conventional drugs can affect the fate of a drug in the body. Grapefruit juice has been shown to inhibit cytochrome P-450, thereby preventing the metabolism and subsequent elimination of a number of drugs.(32, 33) This means that some drugs accumulate in the body and cause adverse effects. Grapefruit juice also appears to affect the P-glycoprotein pump that is responsible for transporting many drugs that are metabolized by cytochrome P-450. The potential exists for other NHPs to have similar effects on the cytochrome P-450 group of isozymes.

For example, St. John's wort was found to reduce levels of indinavir when given to healthy volunteers.(34) It was proposed that St. John's wort was able to enhance the metabolism of this protease inhibitor by inducing CYP3A4, a cytochrome P450 enzyme. The ability of St. John's wort to induce CYP3A4 was validated in a human trial where volunteers took St. John's wort for 14 days and showed increased levels of CYP3A4 levels.(35) Since it is estimated that more than 73 drugs are metabolized by CYP3A4, many theoretical interactions have since been predicted to occur with St. John's wort, based on its ability to induce this enzyme.(35) For example, a letter was issued from Health Canada warning physicians of the potential impact of using St. John's wort in combination with oral contraceptives, anticonvulsants, immunosuppressants, antidepressants or anticoagulants.(36) The letter was careful to note that these were "potential" interactions, based on those seen with St. John's wort and indinavir, and that no studies on the specified classes of drugs had been conducted. Similar warnings were sent from the FDA in the US, the UK Committee on Safety of Medicines and the European Medicines Evaluation Agency.

It should be noted that the clinical significance of these potential interactions with St. John's wort remains unclear. For example, should patients taking oral contraceptives be advised to avoid St. John's wort? Three cases of adverse effects resulting from concurrent use of oral contraceptives and St. John's wort have been reported.(37) All cases involved breakthrough bleeding resulting from what appeared to be lowered ethinyl estradiol levels. It is hypothesized that estrogen metabolism was increased as a consequence of enzyme induction caused by St. John's wort. Given the large number of women who use oral contraceptives, it is clearly important to determine the risk of contraceptive failure if oral contraceptives are taken with St. John's wort.

Additional metabolic studies with St. John's wort suggest that it is also able to induce CYP3A2 (in rats) and intestinal P-glycoprotein/MDR1 (in rats and humans).(38) Hyperforin, a component of St. John's wort, has also been found to bind to the pregnane X receptor (PXR) in a cell based assay.(39) PXR is a hormone receptor that binds to the CYP3A4 promoter and thereby induces CYP3A4. A recent paper reported the results of a trial (in humans) that assessed the short term versus long-term effects of St. John's wort on CYP450.(40) The study found no effect on metabolism when St. John's wort was given in a single dose with a single dose of the CYP3A4 substrate. A significant pharmacokinetic interaction did occur when St. John's wort was administered for 14 days prior to giving a CYP3A4 substrate.

It has been suggested that other NHPs may also affect metabolic enzymes. Budzinski et al. evaluated 21 commercial herbal products and 13 pure plant products for their ability to inhibit CYP3A4 in vitro.(41) Significant inhibition of CYP3A4 was demonstrated by at least 75% of the commercial products and 50% of the pure compounds. These results show the potential for many herbal products to have a significant impact on the metabolism of CYP3A4 as well as other isozymes. Additional research is confirming this. For example, the Chinese herb Angelica dahurica was recently found to affect the metabolism of two substrates for CYP450, tolbutamide and diazepam.(42) In one final example, Foster et al. evaluated the effects of various formulations, brands and production lots of garlic on human cytochrome P450 2C9*1, 2C9*2, 2C19, 2D6, 3A4, 3A5 and 3A7. (43) When tested in vitro, garlic affected cytochrome P450 2C, 2D and 3A mediated metabolism in an inhibitory fashion, with one exception. The paper notes the variabilities inherent in natural products and the difficulties this and other factors present for their pharmacological evaluation. The authors also point out that when garlic is taken with other drugs or herbal products, its effects may be inhibited or enhanced.

Gaps in Knowledge

Health care practitioners, from both conventional and complementary health care teams, are becoming increasingly vigilant with respect to identifying potential interactions between NHPs and conventional drugs. Resources that help in the identification of potential interactions are now available. What remains unclear is how well theoretical predictions of interactions translate into detectable interactions in the clinical setting.

The following issues are key 'gaps' in our current understanding of NHP/drug interactions:

a) Which NHPs are likely to be involved in clinically important interactions with conventional drugs?

Many theoretical NHP/drug interactions have been reported, based on known activity of constituents contained in NHPs. Does this mean that the NHP will produce an interaction in vivo? Is the constituent that is implicated in the theoretical interaction present in a sufficient quantity to actually produce an interaction? Compiling validated case reports of NHP/drug interactions in a systematic manner would help to identify important interactions that occur frequently. A proactive approach to identifying potential interactions is needed. In some instances - in examining pharmacokinetic interactions with St. John's wort, for example - human trials have been conducted. While valuable, this is an expensive and labour intensive method. Other in vitro methods could be used to screen products for pharmacodynamic and pharmacokinetic interactions. Significant interactions could then be investigated more thoroughly.

b) How can consistency in pharmacological activity be maintained from product to product?

If a clinically significant NHP/drug interaction is identified, it is important that the magnitude of the interaction is predictable, no matter which preparation of a particular NHP is used. Due to the complexity of many NHPs, standardization to a particular 'active' or 'marker' compound(s) does not ensure that the total pharmacological activity of an NHP will be consistent between products. This issue was best illustrated with the discovery that hyperforin and not hypericin, the compound to which many St. John's wort products are standardized, likely accounts for the effects produced by the herb on CYP450.(8) Determining the complex chemical profile of each NHP is a daunting task. An option might be to standardize a product to a pharmacological activity using an in vitro assay. This too has limitations, as not all of the pharmacological activities associated with NHPs are easily assessed in vitro. Other cost-effective methods of ensuring quality, consistent products must be developed.

Key Challenges

Challenges in identifying and understanding NHP/drug interactions from a biomedical perspective include:

• translating 'theoretical' NHP/drug interaction information into clinically relevant data
• standardizing the quality of NHP products. It is essential to ensure that marketed products contain the plant (or other source material) identified on the label in a form that has been standardized to consistent activity. Without an assurance that what is on the label is what is in the bottle, understanding NHP/drug interactions is meaningless.

On a final note, NHP/drug interactions should not always be viewed as harmful. In some cases, NHPs have been shown to help to mitigate or prevent adverse effects associated with drugs. For example, aromatic herbs such as ginger may help reduce or prevent drug-induced nausea;(44) co-enzyme Q10 has been used to correct a deficiency caused by conventional HMG-CoA reductase inhibitors;(45, 46) milk thistle may be able to prevent the liver toxicity associated with some drugs;(47) and the diarrhea associated with antibiotics may be decreased by the co-administration of Lactobacillus acidophilus.(48) Clearly, synergy between NHPs and drugs could be therapeutically advantageous if the magnitude was predictable and reproducible. There is much work to be done in this area.